r le X 75^5 Special Publications Museum of Texas Tech University Number 46 12 May 2003 Aquatic Fauna of the Northern Chihuahuan Desert BRAR . *(**%«'*. #1 /& ' V -t AUG 1 3 201 HARVAR JNIVER. - SSft- "• . -- <^,T Contributed Papers from a Special Session within the Thirty-Third Annual Symposium of the Desert Fishes Council HELD 17 November 2001 AT Sul Ross State University, Alpine, Texas Edited By GaryP. Garrett and Nathan L. Allan Cover illustration of Notropis braytoni Jordan & Evermann, commonly known as sardinita Tamaulipeca or Tamaulipas shiner, was prepared by Patty R. Manning of Alpine, Texas. The reference specimen was 40 mm standard length, captured from Terlingua Creek, near Terlingua, Brewster County, Texas on 10 February 1993. Special Publications Museum of Texas Tech University Number 46 Aquatic Fauna of the Northern Chihuahuan Desert Contributed Papers from a Special Session within the Thirty-Third Annual Symposium of the Desert Fishes Council Gar y P. Garrett and Na than L. Allan Heart of the Hills Fisheries Science Center, Texas Parks and Wildlife Department and Austin Ecological Services Field Office, U. S. Fish and Wildlife Service Layout and Design: Gary P. Garrett, Nathan L. Allan, and Jacqueline B. Chavez Cover Design: Christopher W. Hoagstrom Copyright [2003], Museum of Texas Tech University All rights reserved. No portion of this book may be reproduced in any form or by any means, including electronic storage and retrieval systems, except by explicit, prior written permission of the publisher. This book was set in Times New Roman and printed on acid-free paper that meets the guidelines for permanence and durability of the Committee on Production Guidelines for Book Longevity of the Council on Library Resources. Printed: 12 May 2003 Library of Congress Cataloging-in-Publication Data Special Publications, Number 46 Series Editor: Robert J. Baker Aquatic Fauna of the Northern Chihuahuan Desert, Contributed Papers from a Special Session within the Thirty-Third Annual Symposium of the Desert Fishes Council Gary P. Garrett and Nathan L. Allan ISSN 0149-1768 ISBN 1-929330-06-5 Museum of Texas Tech University Lubbock, TX 79409-3191 USA (806)742-2442 Contributors Chihuahuan Desert Research Institute Desert Fishes Council Sul Ross State University Texas Parks and Wildlife Department Texas Tech University The Nature Conservancy U.S. Fish and Wildlife Service World Wildlife Fund TEXAS PARKS & WILDLIFE ML ejSF Nitre} Conservancy WWF Aquatic Fauna of the Northern Chihuahuan Desert Table of Contents Forward Groundwater systems feeding the springs of West Texas 1 J. M. Sharp, Jr., R. Boghici, and M.M. Uliana Contemporary water supply in West Texas as an example for the northern Chihuahuan Desert 13 K. Urbanczyk Hydrology and geomorphology of the Rio Grande and implications for river rehabilitation 25 J.C. Schmidt, B.L. Everitt, and G.A. Richard Gammarid amphipods of northern Chihuahuan Desert spring systems: an imperiled fauna 47 B. K. Lang, V. Gervaiso, D.J. Berg, S.I. Guttman, N.L. Allan, M.E. Gordon, and G. Warrick Declining status of freshwater mussels in the Rio Grande, with comments on other bivalves 59 R.G. Howells Fish assemblages of the Rio Conchos Basin, Mexico, with emphasis on their conservation and status 75 R.J. Edwards, G.P. Garrett, and E. Marsh-Matthews Historical and recent fish fauna of the lower Pecos River 91 C. W. Hoagstrom Pupfishes of the northern Chihuahuan Desert: status and conservation 1 1 1 A. A. Echelle, A.F. Echelle, S. Contreras Balderas, and Ma. de L. Lozano Vilano Spring-endemic Gambusia of the Chihuahuan Desert 127 C. Hubbs My favorite old fishing holes in West Texas: where did they go? 135 J.F. Scudday Aquatic conservation and The Nature Conservancy in West Texas 141 J. Karges Innovative approaches to recover endangered species 1 5 1 G.P. Garrett Forward The twelve papers presented in this volume are a collection of proceedings from a special session within the 33rd Annual Symposium of the Desert Fishes Council (DFC), held on November 17-20, 2001, at Sul Ross State University, Alpine, Texas. It was the desire of the local DFC planning committee (co-chaired by Nathan Allan and Chris Hoagstrom) that the meeting should have • a session dedicated to the issues of special conservation concern for the region. We invited speakers from a variety of disciplines with expertise on aquatic issues in the northern Chihuahuan Desert. Our intent was to expand the scope of interest from fishes to include other aquatic biota and to provide context for conservation considerations. We felt strongly that the information presented at the special session must be captured in a proceedings document to provide a single source of written information on aquatic biota of this region. No volume could be considered a comprehensive assessment of such a vast topic, but we intended to include as much information as possible to provide a realistic overview of the biota of the region and the conservation issues we face. The papers are clearly biased towards fishes, because of our bias as fishery biologists. We are keenly aware that plants, amphibians, reptiles, snails and other groups are equally dependent on the waters of the desert and their conservation is equally significant. All of the invited speakers were offered the opportunity to contribute their manuscripts to this publication and 12 of the 14 chose to do so. The geographic extent of the “northern Chihuahuan Desert” was not intended to be a well- defined boundary, but rather a general guide for the papers presented. Many of the papers are focused on the Trans Pecos area of West Texas. This is because of the experience of the authors solicited for inclusion, not necessarily any statement of priority of this area within the entire Chihuahuan Desert. We are especially honored by the contributions of two of the eminent biologists of the northern Chihuahuan Desert, Dr. Clark Hubbs and Dr. James F. Scudday. Their early work and insight helped build the groundwork for most of us today. Dr. Hubbs began his work at the University of Texas at Austin in 1948 and is now Professor Emeritus. He remains an active researcher and continues to contribute to the conservation of fishes. In recent years his emphasis has been primarily on spring fishes and correlates with chemical parameters. As with so much of his previous endeavors, this work will very likely provide a baseline for many researchers for years to come. Dr. Scudday retired as a biology professor from Sul Ross State University in 1996, after 33 years on the faculty. A native of Fort Stockton, Texas, Dr. Scudday is a notable authority on vertebrates of Trans Pecos, Texas, largely because of his lifetime of experience there. Although much of his paper is anecdotal, the picture he paints of the Trans Pecos area more than 50 years ago is a stark contrast to the area today and provides an important perspective that is relevant to the remainder of the volume. This volume also includes a number of esteemed scientists who have spent decades researching zoology, ecology, and geology in the Trans Pecos and adjacent regions. A primary goal was to assemble a large amount of the institutional knowledge currently available for the region. If nothing else, this volume should serve as a valuable source for information on a wide array of topics related to aquatic resource conservation and should stand as a useful literary reference for future investigators. We wish to offer our sincere appreciation to everyone who volunteered his or her time and expertise to help complete this document. A special thanks to Dr. Gary Garrett for serving as the managing editor of the proceedings- we doubt he had any idea what he was getting into when he took the job. All the papers have been peer reviewed by at least three other persons. All of the authors also served as reviewers for other papers. We also thank Tim Bonner (Southwest Texas State University), Ray Mathews (Texas Water Development Board), Gordon Linam (Texas Parks and Wildlife Department), and Kirk Winemiller (Texas A&M University) for serving as reviewers of various papers. Thanks also to the DFC Executive Committee (especially Dean Hendrickson and Phil Pister) for giving us latitude to work “outside the box” of the normal DFC meetings. We appreciate the Museum of Texas Tech University (especially Robert Baker and Jackie Chavez) for being willing to include this document within the Special Publication series. Finally, we recognize the organizations (and the people who made it happen) which provided financial support for both the meeting in Alpine and this publication: West Texas Program Office of The Nature Conservancy (John Karges); World Wildlife Fund, Chihuahuan Desert Program (Jennifer Atchley); Desert Fishes Council (Phil Pister); Chihuahuan Desert Research Institute (Cathryn Hoyt); and U.S. Fish and Wildlife Service, New Mexico Fishery Resources Office (Jim Brooks) and Austin Ecological Services Office (Bill Seawell). Nathan Allan Christopher Hoagstrom Co-Chairs, Local Committee, DFC 2001 Groundwater Systems Feeding the Springs of West Texas John M. Sharp, Jr., RaduBoghici, and Matthew M. Uliana Abstract Major existing and former springs of the north¬ ern Trans-Pecos, Texas, include the Balmorhea Springs (San Solomon, Phantom Lake, Giffin, and East and West Sandia) in Reeves and Jeff Davis Counties and Comanche Springs, Leon Spring, and Diamond-Y Springs in Pecos County. Understanding the regional groundwater flow systems that feed or fed these springs is needed to manage regional water resources, including the springs that provide islands of aquatic habitat. Some springs have ceased to flow or now flow at greatly diminished rates. Data indicate that spring discharges have been gradually declining for at least the last 100 years. In addition, groundwater ex¬ traction for municipal, domestic, and irrigation uses threatens continued spring flows. The individual groundwater basins are connected through regional flow systems in fractured, karstic carbonate rocks. Regional fracture trends connect the major recharge and discharge areas and localize discharge from car¬ bonate aquifers. Analysis of fracture systems allows interpretation of regional flow systems and regional- scale permeability. Recharge is from fractures in the highlands, losing streams on proximal portions of allu¬ vial fans, irrigation return flow, and interbasin flow. Discharge is to the springs, by wells, and in the past to the Pecos River. 87Sr/86Sr ratios and other chemical and isotopic data confirm the inferred regional flow systems and suggest that some of the springflow re¬ charged during the Pleistocene. The groundwater sys¬ tem is evolving because of both climatic trends and anthropogenic effects. Introduction Trans-Pecos Texas encompasses the general area of Texas west of the Pecos River (Figure 1) and is the most southeastern portion of the Basin and Range physiographic province in the United States. It has a subtropical semiarid climate. Average annual precipi¬ tation is less than 300 mm, and precipitation increases with increasing elevation (e.g., Schuster, 1996). With the exception of a few significant springs and the brack¬ ish Pecos River and Rio Grande, surface water re¬ sources are minimal. A number of individual ground- water basins form parts of regional groundwater flow systems. Regional-scale structural features create a template for fractures and karst features that control the flow systems. The regional flow systems, in turn, discharge at springs that provide unique wetland habi¬ tats for endangered aquatic species. Through wells and spring flow, these systems have been developed to meet much of the area’s municipal, domestic, and agricultural needs. Brune’s (1981) compendium listed the impor¬ tant springs of the region. This study concentrates on those in the northern Trans-Pecos associated with endangered species and the groundwater systems as¬ sociated with these springs. Figures 1 and 3 show the locations of the springs. The unique biotas that inhabit the spring systems suggest persistent, long-term springflow discharge. However, data indicate that dis¬ charge from these springs may have been in decline for the past 100 years and, possibly, for the past sev¬ eral thousand years (Hall, 1990; Sharp et al., 1999; Musgrove, 2000) as is this case for most of the south¬ western U. S. A. Increased groundwater pumpage for municipal and agricultural use has hastened or caused springflow declines and changes in the regional flow systems. Proper management of the groundwater in this region is important to maintain spring flows and the habitats 2 Special Publications, Museum of Texas Tech University Figure 1 . Study area is located in northern Trans-Pecos Texas and southestem New Mexico. The Salt Basin, Wildhorse Flat, Lobo Flat, and the Toyah Basin are clastic basin fills; the Davis Mountains are volcanic rocks. The major springs are delineated. Giffin Spring is adjacent to San Solomon Springs in Toyahvale and East and West Sandia Springs are in the town of Balmorhea. Phantom Lake Springs is closely linked to the springs in Toyahvale. Leon Springs formerly discharged between Comanche and Diamond Y Springs. for the endangered species and to meet the region’s municipal and agricultural needs. Herein we review the regional hydrostratigraphy and structural geology and discuss the hydrostratigraphy, hydrogeology, structural geology, regional hydrogeology, fracture controls on permeability, some springwater chemistry, and implica¬ tions for spring discharge and water resource manage¬ ment. Hydrostratigraphy Geologic units range in age from Precambrian to Holocene, but the most significant units hydrogeologically are: 1) the Permian shelf, reef, and basinal sediments of the Delaware Basin; 2) Cretaceous carbonates; 3) Ter¬ tiary igneous rocks in the Davis and Barilla Mountains; and 4) Cenozoic alluvium. Permian rocks are subdi¬ vided into three hydrostratigraphic facies (Figure 2) with highly different hydraulic properties (Hiss, 1980; Nielson and Sharp, 1985; Boghici, 1997; Mayer and Sharp, 1998; Boghici and van Broekhoven, 2001). The Guadalupian shelf margin (reefal facies) provides excellent aquifers as exemplified by the Capitan aquifer, which has high porosity and permeabilities that are the result of exten¬ sive karstification (e.g., Carlsbad Caverns and associ¬ ated caves in southeast New Mexico, the Apache Moun¬ tains, and the Glass Mountains). Aquifers in the Per- Sharp et al. — Groundwater Systems Feeding the Springs of West Texas 3 WEST EAST Permeability dependent on tracture porosity Permian Shelf High Permeability Guadalupian Shell Margin 11 - A - 1 r _BON£ SPRINGS HUECO (Wolfcamp Series) Low Permeability Basin Facies /s_ CARLSBAD U*tIsL ‘ SEVEN RIVERS GOAT SEEP VICTORIO PEAK IN GROUP Figure 2. Permian hydrostratigraphic facies (after Nielson and Sharp, 1985). The stippled pattern denotes sandstones and carbonates overlying basinal evaporite-rich rocks; the crosshatched pattern denotes highly permeable reefal facies rocks of the shelf margin. mian shelf facies have highly variable fracture -depen¬ dent permeability. Karstification is controlled by the frac¬ ture set characteristics (e.g., the Bone Spring aquifer that supplies Dell City). The basin fill facies rocks, in¬ cluding the Rustler Formation generally possess low per¬ meability and extensive evaporite deposits. They form aquifers with low permeabilities, poor quality water, and low well yields. The Rustler provides brackish water to wells and along a deep fault system to Diamond Y Springs. Lower Cretaceous rocks crop out in the eastern, southeastern and southwestern parts of the Toyah Basin and on the Diablo Plateau (Figure 1). These rocks rep¬ resent mostly marginal, near-shore, and marine facies. They supply irrigation and livestock wells with fresh to slightly brackish water (Ogilbee et al., 1962; Boghici, 1997). The Cretaceous Edwards-Trinity is the most important aquifer in the Diamond Y area. It underlies most of Pecos County, as well as parts of Reeves, Culberson, and Jeff Davis Counties (Anaya, 2001). The more permeable units in the Edwards-Trinity are the lower Cretaceous sands and limestones, which are hydrauli¬ cally connected with the overlying Pecos Cenozoic allu¬ vial aquifers of the Coyanosa and Toyah Basins. In some locales, Cretaceous carbonate units are juxtaposed with Permian reefal rocks and form parts of the same flow system, such as the one that flows to Balmorhea and the Toyah Basin (LaFave, 1987; Uliana and Sharp, 2001). Phantom Lake Springs issue from a cave opening in Lower Cretaceous limestone. The Cretaceous carbon¬ ates of the Diablo Plateau support a regional aquifer and a less extensive perched aquifer. These rocks can be extremely transmissive because of fracture and solution porosity (Scalapino, 1950; Kreitler and Sharp, 1990). Groundwater flows to the northeast and discharges into the Salt Basin, where it evaporates in gypsum flats. Tertiary igneous rocks are mostly ash-flow tuffs and lava flows that overlie the Cretaceous rocks. The volcanic rocks of the Davis Mountains do not contain significant regional aquifer systems (Hart, 1992; Chastam- Howley, 2001), but runoff from them contributes to re¬ charge of the Toyah Basin alluvial aquifer (LaFave and Sharp, 1987; Uliana, 2000) and, presumably, other sur¬ rounding basins. Thick Cenozoic fluvial siliciclastic deposits occur in the Toyah and Salt Basins. Other recent deposits in¬ clude fluvial terraces, playa muds and evaporites, aeolian deposits, and colluvium. Aquifers in these units provide significant amounts of water to wells and municipalities in the Toyah Basin and in Wild Horse Flat (Gates et al., 1980; Sharp, 1989; Ashworth, 1990). The Toyah Basin was formed by dissolution of underlying Permian evapor¬ ites and is filled with up to 470 m of Cenozoic alluvium (Maley and Huffmgton, 1953). The alluvial sediments filling in the Salt Basin reach a thickness of over 750 m (Gates et al., 1980). Groundwater divides separate the basin into three different flow systems. The northern and the middle areas are closed basins with recharge occurring through the bounding faults on the east and west of the basin, and discharge through the gypsum flats or vadose playas in the center of the basin (Boyd, 1982; Nielson and Sharp, 1985). The Wild Horse Flat part of the basin, however, lacks playas. In addition, the oldest mapped potentiometric surface (Nielson and Sharp, 1985; Sharp, 1989) indicates interbasin flow through the Apache Mountains towards the east. 4 Special Publications, Museum of Texas Tech University ) j Figure 3. Regional flow systems of Trans-Pecos Texas (Sharp, 2001). WH and LF denote Wild Horse Flat and Lobo Flat of the Salt Basin. Springs are denoted by letters - A, Phantom Lake Spring; B, San Solomon and Giffin Springs; C, East and West Sandia Springs; D, Leon Springs; E, Comanche Springs; F, Diamon-Y Springs; and G, Indian Hot Springs. A, D, and E no longer flow. The regional flow systems are: 1 and 2, the discharge at the Fabens artesian zone and Indian Hot Springs (G), respectively; 3, Eagle Flat - Red Light Draw flow system; 4, Sacramento Mountains - Dell City flow system; 5, flow systems in the Capitan Reef; 6, eastward flow in the Delaware Basin, perhaps discharging at Diamond-Y Springs (F); 7, the Salt Basin - Toyah Basin - Pecos River system that also feeds Balmorhea Springs (A, B, and C); and 8, speculative eastwards extensions of this last flow system. Structural Geology Trans-Pecos Texas has been subjected to at least five major tectonic episodes that formed fault zones and structural trends that have been repeatedly reactivated throughout the history of the area. Precambrian com- pressional events generated the basic northwesterly trends that are still predominant and that influenced sub¬ sequent structures. The early Pennsylvanian Ouachita collision was responsible for the formation of the Dela¬ ware Basin and the thick sequences of Permian sedi¬ mentary rocks that form a significant part of the re- Sharp et al. — Groundwater Systems Feeding the Springs of West Texas 5 gional groundwater systems. The morphology of the Delaware Basin has influenced later fault patterns. For example, major faults in the Apache Mountains and the Guadalupe Mountains run parallel to the Permian paleoreef front. Later events include the Mesozoic rifting of the ' Gulf of Mexico, the early Cenozoic Cordilleran/Laramide Orogeny, and Eocene Basin and Range extension. These events caused the repeated reactivation of the earlier structural features. Faulting is active today (Goetz, 1977). Details of the region’s tectonic/structural setting are given in Dickerson and Muehlberger (1985) and Muehlberger and Dickerson (1989). Regional Hydrogeology Groundwater production in Trans-Pecos Texas is concentrated in the Permian and Cretaceous rocks and in the Cenozoic alluvial fill in the Salt and Toyah basins (Davis and Leggat, 1965; Couch, 1978; Hiss, 1980; Ashworth, 1990; Sharp, 2001). Evidence exists for sig¬ nificant groundwater flow in fractures and in karst con¬ duits (LaFave and Sharp, 1987; Mayer and Sharp, 1998; Uliana, 2000; Sharp, 2001). Both local and regional flow systems exist in the area. Figure 3 shows the major re¬ gional flow systems inferred for northern Trans-Pecos Texas. The major springs discharge both from regional and local flow systems associated with Permian/Creta¬ ceous carbonate rocks (White et al., 1941; LaFave and Sharp, 1987; Uliana and Sharp, 2001). The flow sys¬ tems that feed the currently active springs harboring endangered aquatic species (San Solomon, Giffin, East and West Sandia, and Diamond Y Springs) are only two on the documented, probable, or inferred regional flow systems in northern Trans-Pecos Texas and southeast¬ ern New Mexico. The other flow systems may have once discharged to springs (i.e., Crow Springs near Dell City) that have also have once harbored unique biota. All these regional flow systems are in fractured carbonate rocks. Fracture Controls on Groundwater Flow The regional flow systems are controlled by frac¬ ture systems. The density and orientation of fractures in the Trans-Pecos region reveal a definite relationship between the regional structural trends and the fracture orientations. LaFave and Sharp (1987) and Uliana (2000) document fracture orientations in the Apache and Dela¬ ware Mountains that follow the prevalent N10W orien¬ tation of regional structural trends. Mayer (1995) and Uliana (2000) used aerial photos and field studies to map lineaments and fractures. They document the correla¬ tion between fault orientations and the regional struc¬ tural grain. Fault patterns in the Salt Basin (Goetz, 1977) show a similar correlation between fracture patterns and the regional structural grain. Fracturing of the carbonate rocks influenced their subsequent karstification. Nielson and Sharp (1985), LaFave and Sharp (1987), and Uliana (2000) document regional connections between Wildhorse Flat and the Toyah Basin, including the springs at Balmorhea. Frac¬ turing of Permian reefal rocks and karstification create a high permeability zone along the Stocks Fault-Rounsaville Syncline trend (Figures 1 and 3). Hydraulic head data show that water discharged from Wildhorse Flat at one end of the structural trend and flowed into the Toyah Basin along its southwestern edge, near the springs. Additional evidence supporting the regional flow hypoth¬ esis includes the lack of discharging playas in the Wildhorse Flat section (Figure 1), fracture trends in the Apache Mountains, and isotopic data (Uliana, 2000; Uliana and Sharp, 2001). The northern parts of the Salt Basin contain extensive vadose playas that are primary natural discharge features. The lack of playas in Wildhorse Flat is consistent with groundwater discharge by interbasin flow through the Apache Mountains. Uliana and Sharp (2001) examined the^Sr/^Sr distributions in Trans-Pecos groundwater. High 87Sr/86Sr values occur at the up-gra¬ dient end of the flow system; these are caused by ground- water interaction with Precambrian rocks (including clasts in alluvial fans) along the west edge of Wildhorse Flat. 87Sr/86Sr values decrease along the hypothesized regional flow paths. This suggests that the high 87Sr/86Sr values in the Wildhorse Flat groundwater equili¬ brate with the Permian and Cretaceous carbonates and fluid mixing with other waters recharged or displaced along the flow path. Phantom Lake Springs formerly issued from an opening in Cretaceous limestone. This opening leads into a network of caverns. Cave divers 6 Special Publications, Museum of Texas Tech University have mapped and surveyed over 2300m and measured groundwater discharge in the caverns. Their map (Tucker, 2000, reproduced in Uliana, 2000, p. 21) indi¬ cates that the cave network follows a linear trend that parallels the Stocks Fault - Rounsaville Syncline trend. In January 1999, Phantom Lake Springs ceased to flow for the first time in probably at least 200 years (Figure 4). A similar high-permeability structural trend (the Otero Break) connects the Sacramento River recharge area in southern New Mexico and the Dell City, Texas, irrigation district (Mayer and Sharp, 1998). This inferred conduit flow (Flow system 4 on Figure 3) is confirmed by both geochemical and hydraulic head data. This flow system formerly discharged at Crow Springs in the Salt Basin (Ashworth, 2001). Flow system 3 from Eagle Flat is presumed to discharge to the Rio Grande (Darling and Hibbs, 2001). Indian Hot Springs (G on Figure 3) discharge from carbonate rocks and the flow is through fractured carbonate rocks in the U. S. A., and, perhaps, Mexico. Although undocumented, similar regional flow systems may exist in southern Trans-Pecos Texas and Northern Mexico. Comanche Springs (Figures 1 and 3) formerly is¬ sued from a 673-m long, fracture-controlled cave formed along a N60-65W trending joints (Boghici, 1997; Veni, 1991). Dye tracer tests conducted in the 1950’s dem¬ onstrated that groundwater was flowing to Comanche Springs at rates of up to 3.2 km/day (Sparks, cited in Veni, 1991). A structural low in the permeable Creta¬ ceous limestones (the Belding-Coyanosa trough) extends some 64 km to the southwest of Comanche Springs connecting them with the recharge areas in the vicinity of the Glass Mountains (Boghici, 1997; Boghici and Van Broekhoven, 2001). Darling (1997), Darling et al. (1995), and Darling and Hibbs (2001) examined oxygen, deuterium, and car¬ bon isotopes in the groundwater in the western portions of the study area (in particular, in flow systems 2 and 3 on Figure 3). These data suggest recharge during a cooler and wetter Pleistocene climate. Groundwater ages of 8,000 to 50,000 years are indicated. Balmorhea springs show little seasonal variation. Their waters are brackish and slightly higher than mean annual surface tempera¬ ture. Isotopic data, although sparse, support Darling’s results, indicating long residence times (Uliana, 2000; Figure 4. Spring flow hydrographs of Comanche, Phantom Lake, San Solomon, and Giffin Springs. Data are from U. S. Geological Survey, U. S. Bureau of Reclamation, and Groundwater Field Methods Class (1990, 1992, 1995, and 1996). Sharp et al. — Groundwater Systems Feeding the Springs of West Texas 7 Uliana and Sharp, 2001). In addition, the unique spring ecosystems suggest that spring flow and spring chemis¬ try have remained relatively stable for extended periods of time. Such stability in a semi-arid zone implies a re¬ gional flow system because local flow systems show greater seasonal variations. Consequently, a few years of drought would not be suspected to impact spring flows significantly. Spring Discharge and Water Resource Management Figure 4 shows the discharge of the four of the largest springs in the area. Early (before 1920) data are less reliable, but several trends are evident. First, springflows appear to have been declining prior to the development of extensive irrigation in the 1940’s. This is consistent with longer-term climatic studies that sug¬ gest a drying of the region (cited in Kreitler and Sharp, 1990). Second, there is variability in discharge, particu¬ larly, when average annual discharges are low. This is caused by normal climatic variability and may not be indicative of long-term sustained trends. The variations probably reflect short-term variability in groundwater re¬ charge. Some later data on San Solomon, Giffin, and Phantom Lake Springs are point (in time) measurements taken by the Groundwater Field Methods Classes (1990, 1992, 1995, and 1996). These are not annual averages, but reflect the short term variability. Third, San Solomon and Giffin Springs show a remarkably steady annual flow regardless of climatic variability. Flow has been rela¬ tively constant since the 1930’s even through the major drought of 1947-1955. Four, Comanche and Phantom Lake spring discharges began to decline at greater rates when groundwater pumpage for irrigation commenced. Comanche Springs ceased to flow about 1961 , and Phan¬ tom Lake Springs largely ceased to flow in January 1999. Balmorhea Springs. — San Solomon, Giffin, and Phantom Lake springs are in close proximity and are probably fed from the same fractured, karstic conduit (Uliana and Sharp, 2001). The conduit system, which is accessed at the Phantom Lake Spring orifice, was gauged in 1997 by Tucker (2000). His preliminary data showed that, at that time, only a small percentage of the flow in the conduit discharged at Phantom Lake Springs. The bulk presumably discharges to San Solomon and Giffin springs, but this is not yet documented. Phantom Lake Spring fed a cienega and canal system that harbored en¬ dangered fish species. Inconsistencies in spring dis¬ charges fed from the fractured, karstic conduit trend are typical of karstic systems. The decline in Phantom Lake springflow may be caused by several factors. First, a regional lowering of the water table is documented by comparing recent data from The University of Texas Groundwater Field Meth¬ ods classes (1990, 1992, 1995, and 1996) and the U. S. Bureau of Reclamation with the earliest known data (White et al., 1941). These show that the water table near Balmorhea has been lowered by some combination of long-term climatic change and, probably more impor¬ tantly, regional pumping for irrigation and municipal uses. This pumping could include both local well effects and the regional lowering of heads in the Toyah Basin. A sec¬ ond possibility is that the “plumbing” itself has changed because of either opening of fractures/conduits because of dissolution, tectonic activity, or changing sediment storage in the conduits that could increase or decrease the permeability of various branches of the conduit. Major pulses in turbidity of spring discharge have been ob¬ served (White et al., 1941; Kreitler and Sharp, 1990). Loss of recharge because of domestic wells in the Davis Mountains is another potential, but probably secondary, effect. Finally, the beheading of the regional flow sys¬ tem because of pumping in Wildhorse Flat may have an effect. Water recharging Wildhorse Flat now supplies the communities of Van Horn and Sierra Blanca, as well as local irrigators of cotton, pecans, hay, and vegetables (Nielson and Sharp, 1985; Kreitler and Sharp, 1990) so that discharge from Wildhorse Flat via interbasin flow has been diminished, if not stopped. At present, there are insufficient data to differentiate between these hy¬ potheses or eliminate any of them. Pecos County Springs. — The cessation of flow at Comanche and Leon Springs is closely correlated in time with the onset of irrigation pumpage, although, quality long-term discharge data are not available for Leon Springs. Diamond Y Springs, however, continue to flow. The discharge data for the Diamond Y Springs are sparse: only four measurements between 1943 and 1987, and only the main spring was gauged. Based on their re- 8 Special Publications, Museum of Texas Tech University sponse to rainfall, Veni (1992) indicated the existence of two distinct groups of springs and seeps. The springs’ response suggests slight to moderately extensive flow conduits feeding the Diamond Y Springs system. Dia¬ mond Y Springs discharge a low to moderately saline Na-Ca-Cl-S04 type of water that is similar in composi¬ tion to waters from the Rustler aquifer near Fort Stock- ton. The main processes affecting the water chemistry are: calcite, dolomite, halite, and gypsum dissolution and/ or precipitation, and ion exchange between calcium, mag¬ nesium, and sodium (Boghici, 1997, 1999). Stable and radiogenic isotope analyses in the Dia¬ mond Y Springs suggest that evaporation and water mix¬ ing processes are important controls on the spring water chemistry. Oxygen- 1 8 and deuterium data indicate that Diamond Y waters are meteoric in origin; the data are distributed along an evaporation line according to spring discharge and pool size - the larger the discharge and the pool, the closer they resemble the main spring composi¬ tion. The two main springs, Diamond Y main spring and Euphrasia spring, show tritium and 14C data indica¬ tive of a mixing of older and recent waters (Boghici, 1997, 1999). This occurs elsewhere in the study region (e.g., LaFave and Sharp, 1987; Uliana, 2000; Darling, 1 997; Darling et al., 1 995) and is supported by geochemi¬ cal mass-balance modeling (Boghici, 1997). These are all consistent with the model of Diamond Y spring water being the product of mixing between Rustler Formation waters and recent local rain falling directly on the springs’ pools. There is apparently discharge of older waters from the Rustler Formation, perhaps up along a fault trend, into Diamond Y Draw (Figures 1 and 3). Regional discharge, coupled with limited pumping from nearby wells to lower water tables, can explain the steadiness of these springs discharge. The spring flows then sink in the streambed and may reappear at downstream as seeps after undergoing various levels of evaporative concen¬ tration. Discussion and Conclusions All data and existing models support the hypoth¬ esis of extensive regional flow systems in Trans-Pecos Texas. Analyses of the fracture zones and the regional structural trends indicate that those trends influence, and probably control, the flow paths in this regional system. We also observe extensive structural features that con¬ nect recharge and discharge areas over great distances. Analyses of the fracture zones indicate that their orienta¬ tions are controlled by pre-existing structural trends that have been reactivated over the history of this area. These demonstrate a pattern where ancient structural trends create the templates for fractures and karst patterns that control the development of a regional groundwater sys¬ tem. We hypothesize that this pattern may be repeated in carbonate systems in other semi-arid parts of the world. This has implications for our understanding of ground- water flow systems in regions that are usually depen¬ dent on groundwater for irrigation and municipal needs. Regional flow systems connect individual ground- water basins. Fractures and subsequent karstification follow the structural trends and control the location of the major natural recharge and discharge areas (springs). At least 5 tectonic events and physical stratigraphic vari¬ ability have led to a complex set of fracture domains. Mapping of these domains demonstrates how they have controlled the development of the regional hydrogeologic system, including the karstification. Furthermore, the fractured (and karstic) systems inherent variability makes it difficult to predict the response of the groundwater systems to anthropogenic stresses and climatic variabil¬ ity. Detailed hydrogeologic assessments are required to utilize the region’s groundwater resources and yet main¬ tain critical environmental habitat. Natural tracer tests using strontium isotopes (^Sr/^Sr ratios) and fracture trace/intensity mapping are demonstrated to be promis¬ ing assessment techniques of regional flow systems in this area or similar hydrogeological settings. Diminishing spring flows or their cessation dem¬ onstrates the potential threats to remaining spring sys¬ tems and their unique biota. Cessation is caused by a combination of climatic and regional factors. Increased groundwater extraction for irrigation and municipal use is the obvious cause, but other factors may also be im¬ portant. These include the long-term pattern of increas¬ ing aridity in the region, possible alterations in the car¬ bonate system permeability by either tectonic or sedi¬ mentation effects, localized pumping effects, reduction of recharge to the regional flow systems, and changes in the regional flow system boundaries, such as the be¬ heading of the Wildhorse Flat - Toyah Basin flow sys- Sharp et al. — Groundwater Systems Feeding the Springs of West Texas 9 tern. However, it seems clear in general that rates of springflow decline have increased with groundwater extractions for agricultural, municipal, and domestic use during the past century. Municipalities (e.g., El Paso, Pecos, Fort Stock- ton, and Midland-Odessa) may need increased water re¬ sources. Irrigated agriculture is an economic mainstay for the area, and although the trends for irrigated agri¬ culture are highly dependent upon economic conditions, the long-term needs for agricultural products are inferred to remain steady or increase. Increased utilization of groundwater will draw upon groundwater in storage both on a cyclical basis depending upon normal climatic vari¬ ability and on a long term trend that could lead to overexploitation. Increased understanding of the regional flow systems, including fracture controls, the nature of the recharge, and the flow paths, is needed to manage these resources. It may be possible to design pumping strategies that minimize the effects on natural spnngflows and yet meet projected demands. In addition, identifica¬ tion of key recharge areas and a priori analysis of frac¬ ture systems to identify fracture hydraulic domains may make it possible to maintain the spnngflows in the face of present or increased levels pumping (White et al., 1 94 1 ; Mayer and Sharp, 1996; Uliana, 2000). We suggest that similar analyses and methodologies may prove of value in studies of spring systems in other areas of the south¬ western United States and northern Mexico or other ar¬ eas where regional flow systems exist in carbonate aqui¬ fers in semi-arid and arid zones. Acknowledgements This research was partially supported by the Chev- of the Geology Foundation of The University of Texas, ron Centennial Professorship and the Owen-Coates Fund Literature Cited Anaya, R. 2001. An overview of the Edwards-Trinity aquifer system, central-west Texas. Pp. 100-1 19, in Aquifers of West Texas (R. E. Mace, W. F. Mullican, III, and E. S. Angle, eds.), Texas Water Development Board, Austin, Texas, Re¬ port 356, 272 pp. Ashworth, J. B. 1990. Evaluation of ground-water resources in parts of Loving, Pecos, Reeves, Ward, and Winkler Coun¬ ties, Texas. Texas Water Development Board Report 317, 51 pp. Ashworth, J. B. 2001. Bone Spring B Victorio Peak aquifer of the Dell Valley region of Texas. Pp. 135- 152, in Aquifers of West Texas (R. E. Mace, W. F. Mullican, III, and E. S. Angle, eds.), Texas Water Development Board, Austin, Texas, Re¬ port 356, 272 pp. Boghici, R. 1997. Hydrogeological investigations at Diamond Y Springs and surrounding area, Pecos County, Texas. Un¬ published Master’s Thesis, University of Texas at Austin, 120 pp. Boghici, R. 1 999. Changes in groundwater conditions in parts of Trans-Pecos Texas, 1988-1998. Texas Water Development Board, Austin, Texas, Report 348, 29 pp. Boghici, R. and Van Broekhoven, N. G 2001. Hydrogeology of the Rustler Aquifer, Trans-Pecos Texas. Pp. 207-225, in Aquifers of West Texas (R. E. Mace, W. F. Mullican, III, and E. S. Angle, eds.), Texas Water Development Board, Austin, Texas, Report 356, 272 pp. Boyd, F. M., 1982, Hydrogeology of the Northern Salt Basin of West Texas and New Mexico. Unpublished Master’s The¬ sis, University of Texas at Austin, 135 pp. Brune, G 1981. Springs of Texas (vol. 1). Branch-Smith, Inc., Fort Worth, Texas, 565 pp. Chastain-Howley, A.. 2001. Igneous aquifers of far West Texas. Pp. 175-189, in Aquifers of West Texas (R. E. Mace, W. F. Mullican, III, and E. S. Angle, eds.), Texas Water Develop¬ ment Board, Austin, Texas, Report 356, 272 pp. Couch, H. E. 1978. Study of the lower Cretaceous and associated aquifers in the Balmorhea district of Trans-Pecos, Texas. Texas Department of Water Resources, Unpublished Re¬ port, 97 pp. Darling, B. K. 1997. Delineation of the Ground-Water Flow Systems of the Eagle Flat and Red Light Basins of Trans- Pecos, Texas. Unpublished Ph.D. Dissertation, University of Texas at Austin, 179 pp. Darling, B. K. and Hibbs, B. J. 2001. The aquifers of Red Light Draw, Green River Valley and Eagle Flat. Pp. 226-240, in Aquifers of West Texas (R. E. Mace, W. F. Mullican, III, and E. S. Angle, eds.), Texas Water Development Board, Austin, Texas, Report 356, 272 pp. 10 Special Publications, Museum of Texas Tech University Darling, B. K., Hibbs, B. J., Dutton, A. R., and Sharp, J. M., Jr. 1995. Isotope hydrology of the Eagle Mountains area, Hudspeth County, Texas: Implications for development of ground-water resources. Pp. SL12-SL24 in Water Resources at Risk (W. R. Hotchkiss, J. S. Downey, E. D. Gutentag, and J. E. Moore, eds.). American Institute of Hydrology, Minneapolis, Minnesota. Davis, M. E. and Leggat, E. R. 1965. Reconnaissance investiga¬ tion of the ground-water resources of the Rio Grande Basin, Texas. Texas Water Commission Bulletin 6502, 199 pp. Dickerson, P. W. and Muehlberger, W. R. (eds.). 1985. Structure and Tectonics of Trans-Pecos Texas. West Texas Geological Society Field Conference, Publication 85-81, 278 pp. Gates, J. S., White, D. E. Stanley, W. D., and Ackermann, H. D. 1 980. Availability of fresh and slightly saline ground water in the basins of westernmost Texas. Texas Department of Water Resources, Report 256, 108 pp. Goetz, L. K. 1977. Quaternary Faulting in Salt Basin Graben. Unpublished Master’s Thesis. University of Texas at Aus¬ tin, 136 pp. Groundwater Field Methods Class. 1990. Hydrology of springs and shallow ground water near Balmorhea, Texas. Unpub¬ lished Report, University of Texas at Austin, 11 pp+. Groundwater Field Methods Class. 1 992. Water characteristics of springs and groundwater near Balmorhea, Texas. Unpub¬ lished Report, University of Texas at Austin. Groundwater Field Methods Class. 1995. Characterization of the ground-water resources in the Toyah Basin near Balmorhea, Texas. Unpublished Report, University of Texas at Austin, 17+ pp. Groundwater Field Methods Class. 1996. Hydrologic investiga¬ tion in the Toyah Basin, Balmorhea, Texas. Unpublished Report, University of Texas at Austin, 31+ pp. Hall, S. A. 1990. Channel trenching and climatic change in the southern U. S. Great Plains. Geology, 18:342-345. Hart, M. A. 1992. The hydrogeology of the Davis Mountains, Texas. Unpublished Master’s Thesis, University of Texas at Austin, 158 pp. Hiss, W. L. 1980. Movement of ground waters in Permian Guadalupian aquifer systems, southeastern New Mexico and western Texas. Pp. 289-294, in Trans-Pecos Region: New Mexico Geological Society 31st Field Trip Guide¬ book, (P. W. Dickerson and J. M. Hotter, eds.). Kreitler, C. W., and Sharp, J. M., Jr. 1990. Hydrogeology of Trans-Pecos Texas. University of Texas at Austin, Bureau Economic Geology, Guidebook No. 25, 120 pp. LaFave, J. I., 1987. Groundwater flow delineation in the Toyah Basin of Trans-Pecos Texas. Unpublished Master’s The¬ sis, University of Texas at Austin, 159 pp. LaFave, J. 1., and Sharp, J. M. Jr. 1987. Origins of ground water discharging at the springs of Balmorhea. West Texas Geo¬ logical Society Bulletin, 26:5-14. Maley, V. C. and Huffington, R. M. 1953. Cenozoic fill and evaporite solution in the Delaware Basin, Texas and New Mexico. Geological Society of America Bulletin, 64:539- 546. Mayer, J. R. 1995. The role of fractures in regional groundwater flow: field evidence and model results from the basin and range of Texas and New Mexico. Unpublished Ph. D. dis¬ sertation, University of Texas at Austin, 221 pp. Mayer, J. R. and Sharp, J. M., Jr. 1998. Fracture control of regional ground-water flow in a carbonate aquifer in a semi- arid region. Geological Society America Bulletin, 110:269- 283. Muehlberger, W. R. and Dickerson, P. W. 1989. Structure and Stratigraphy of Trans-Pecos Texas. American Geophysical Union Field Trip Guidebook, T317, 199 pp. Musgrove, M. 2000. Temporal links between climate and hydrol¬ ogy: insights from central Texas cave deposits. Unpub¬ lished Ph. D. dissertation, University of Texas at Austin, 432 pp. Nielson, P. D. and Sharp, J. M., Jr. 1985. Tectonic controls on the hydrogeology of the Salt Basin, Trans-Pecos Texas. Pp. 231-235, in Structure and Tectonics of Trans-Pecos Texas (P. W. Dickerson and W. R. Muehlberger, eds.), West Texas Geological SocietyField Conference, Publication 85-81 . Ogilbee, W., Wesselman, J. B., and Ireland, B. 1962. Geology and ground-water resources of Reeves County, Texas. Texas Water Commission Bulletin 6214, 1 : 1 93 and 2:245. Scalapino, R. A. 1950. Development of ground water for irrigation in the Dell City area, Hudspeth County, Texas. Texas Board Water Engineers Bulletin, No. 5004, 41 pp. Schuster, S. K. 1996. Water resource and planning assessment in San Solomon Springs and associated spring systems sur¬ rounding Balmorhea, Texas. Unpublished Master’s Thesis, University of Texas at Austin, 102 pp. Sharp, J. M., Jr. 1989. Regional ground-water systems in northern Trans-Pecos Texas. Pp. 123-130 in Structure and Stratigra¬ phy of Trans-Pecos Texas (W. R. Muehlberger and P. W. Dickerson, eds.), 28th International Geological Congress Field Trip Guidebook T317. _ . 200 1 . Regional groundwater flow systems in Trans-Pecos Texas. Pp. 41-55, in Aquifers of West Texas (R. E. Mace, W. F. Mullican, III, and E. S. Angle, eds.), Texas Water Development Board, Austin, Texas, Report 356, 272 pp. Sharp, J. M., Jr., Uliana, M. M., and Boghici, R.. 1999. Fracture controls on regional groundwater flow in a semiarid environ¬ ment and implications for long-term maintenance of spring flows. Water 99 Joint Congress, Institute of Engineers, Brisbane, Australia, 2:1212-1217. Tucker, A. W. 2000. Phantom Lake Cave. The Texas Caver, Texas Speleological Association, p. 4-7. Sharp et al. — Groundwater Systems Feeding the Springs of West Texas 11 Uliana, M. M.. 2000. Delineation of regional groundwater flow paths and their relation to regional structural features in the Salt and Toyah Basins, Trans-Pecos Texas. Unpublished Ph. D. dissertation, University of Texas at Austin, 215 pp. Uliana, M. M., and Sharp, J. M., Jr. 2001 . Tracing regional flow paths to major springs in Trans-Pecos Texas using historical geochemical data. Chemical Geology, 179: 53-72. Addresses of Authors: John M. Sharp, Jr. Department of Geological Sciences [Cl 100] Jackson School of Geosciences The University of Texas Austin, TX 787 12-1 101 Radu Boghici Texas Water Development Board P.O.Box 13231 Austin, TX 787 11-3231 Veni. G 1991. Delineation and preliminary hydrogeologic inves¬ tigation of the Diamond Y Spring, Pecos County, Texas. Unpublished Report, The Nature Conservancy of Texas, San Antonio, Texas, 1 10 pp. White, W. N., Gale, H. S., and Nye, S. S. 1941. Geology and ground-water resources of the Balmorhea area, western Texas. U. S. Geological Survey Water-Supply Paper, 849- C:83-146. Matthew M. Uliana Department of Biology Southwest Texas State University 601 University Drive San Marcos, TX 78666 Contemporary Water Supply in West Texas as an Example for the Northern Chihuahuan Desert Kevin Urbanczyk Abstract Water supply in the northern Chihuahuan Desert region occurs as both surface water and groundwater sources. The Rio Grande basin is the most significant of the surface water sources. This basin includes both the Rio Conchos and the Pecos River systems. The Rio Grande Compact, the 1944 Treaty with Mexico regarding deliveries from the Rio Conchos to the Rio Grande, and the Pecos River Program, governs water use in the basin. Groundwater resources in west Texas include several distinct aquifers. The Texas Water Development Board has categorized these aquifers as Major or Minor, based upon their size, geographic lo¬ cation, and geologic structure. Population growth pre¬ dictions suggest large increases in the west Texas re¬ gion, mostly in the El Paso area. These increases will place a significant strain on the available water sup¬ plies, and further use of both surface and groundwa¬ ter sources will place a strain on the water available for wildlife. Within Texas, the Far West Texas Water Planning Group suggests several strategies for dealing with both the predicted population increase, and the decrease in available water due to drought and over¬ use of certain aquifers. These strategies include con¬ version of surface water appropriations from agricul¬ tural to municipal use, desalination, and interbasin trans¬ fer. Of these, the desalination and interbasin transfer of groundwater could significantly impact the ability of the groundwater to sustain wildlife populations. Introduction The northern Chihuahuan Desert encompasses parts of northern Mexico and the southwestern United States. As in all populated arid regions, water supply in this area is an essential issue for human habitation of the region. Complex legislation exists to control water use and, particularly in regions with expanding populations, overconsumption of water resources not only degrades the quality of life for humans, but sig¬ nificantly alters the natural ecosystems of this sensi¬ tive desert region. This paper summarizes water supply in this arid region. Water “supply” is intimately associated with water “demand”. This supply and demand concept is complex, and involves not only municipal water use, but also agricultural and industrial use. State, inter¬ state and international government legislation is required to attempt to manage the use of this precious resource. Therefore, a summary of water supply necessarily requires the inclusion of both a discussion of demand and of legislation governing water use. Water supplies in the northern Chihuahuan Desert occur as both sur¬ face and as groundwater. These two distinct sources of water occur in different areas, and are legislated in manners that also are quite different. Figure 1 includes the Rio Grande drainage basin and major political boundaries. The Rio Grande drain¬ age basin begins in southern Colorado, flows through New Mexico, enters Texas near El Paso, and contin¬ ues to the Gulf of Mexico where it forms the bound¬ ary between the United States and Mexico. Signifi¬ cant tributaries to the Rio Grande in the northern Chihuahuan Desert region include the Rio Conchos (confluence located at Presidio, Texas and Ojinaga, Mexico), and the Pecos River (confluence near Langtry, Texas). 14 Special Publications, Museum of Texas Tech University Albuquerque ■cLBluff Presidio La^Boquilla NUEVO LEO! v /-'••y--'/ 1 05 "W lOO'W *5‘l 5 COLORADO X I"- . ■■ ■ "w.i -I- ■■■— ■■■— , *"fyp NEW MEXIQO A K "U I Rio Grande Drainage Basin soo 100 2C0 400 Kj Kloroetars ■f* ~y 7 ) r r J / J /CU Elephant Butte yJ Ha hallo ,/ n V TEXAS Cities o Dams jo- SI- CHIHUAHUA \ 2S‘N— ft- -36-N -JO-N T - 1 - 105 *W IQCrW Figure 1 . Rio Grande drainage basin with selected cities and dams. Basin boundaries and drainage patterns are from Texas Natural Resources Conservation Commission, and are available at http://www.riogrande.org/programs/gis/gisdata.htm (10/22/01). Much of the content of this paper utilizes data avail¬ able for the state of Texas. These data were produced in association with the development of the 1997 State Wa¬ ter Plan, and the development of the draft version of the 2002 State Water Plan. The data represent the most complete compilation of water resources information for any section of the Chihuahuan Desert. Discussions of areas outside of Texas are included where appropriate and available. Urbanczyk — Contemporary Water Supply in West Texas 15 Water Law Water laws in Texas are similar to those in most other western states, but may differ from laws in Mexico. In Texas, water is classified as to where it physically occurs: percolating groundwater, underground streams, diffuse surface water, and streamflow (Wurbs et al., 1994). Of these, percolating groundwater and stream flow are the two significant water sources present in the northern Chihuahuan Desert. Groundwater has been historically governed by the “Right (or Rule) of Cap¬ ture” doctrine. According to this doctrine, a landowner has the right to use or sell all of the water that can be captured from beneath a property (Wurbs et al., 1994). Stream flow is governed by the “Prior Appropriation” doctrine. Surface water (in Texas) is publicly owned, and permits must be obtained to use surface water. The permitting, which is controlled by agencies such as the Texas Commission on Environmental Quality (TCEQ, formerly Texas Natural Resource Conservation Com¬ mission, TNRCC), is generally based upon who received the rights first (“first come, first serve”). Diffuse sur¬ face water is water such as return flow from an irri¬ gated property. It is the property of the landowner until it reaches a watercourse. Underground streams are present in the northern Chihuahuan Desert. These feed water sources such as San Salomon springs in the Balmorhea area (Sharp et al., this volume). Surface Water Resources The Rio Grande basin encompasses approximately 180,000 square miles (466,000 square kilometers, Fig¬ ure 1). It is best considered as two separate basins - an upper basin that ends at Presidio, Texas, and a lower basin that extends from there to the Gulf of Mexico. Additionally, the Pecos River and the Rio Conchos can be considered part of the lower basin. Most of the flow in the upper basin is due to precipitation in southern Colo¬ rado and northern New Mexico (Wilson, 2000). This flow is impounded in a series of reservoirs. These in¬ clude Elephant Butte (constructed in 1916; 2.11 million acre-feet (MAF) capacity) and Caballo (1938; 0.331 MAF) in southern New Mexico (Figure 1). As an ex¬ ample of evaporation rates. Elephant Butte and Caballo account for 85% of the 0.34 MAF/year that evaporate off of New Mexican Rio Grande reservoirs (Wilson, 2000). Dams in the Chihuahuan Desert region of the lower basin include La Boquilla (1916; 2.34 MAF) and Luis L. Leon (1968; 0.29 MAF) on the Rio Conchos (Mexico), and Red Bluff (1921 ; 0.3 1 MAF) on the Pecos River (Texas). The International Boundary and Water Commis¬ sion (IBWC) has operated a series of gages in the Rio Grande basin since 1880 for some locations. Figure 2 shows the annual runoff past a series of these gages (IBWC; data from http://www.ibwc.state.gov/wad/ rio_grande.htm). Several features are apparent upon observation of these data. First, there exists a general decline in discharge from Elephant Butte down river to Caballo, then to El Paso, and finally to Ft. Quitman. There have been periods of no flow in the Rio Grande at and below Ft. Quitman, particularly during and after the drought of the 1950s. Second, the buffering effect of the installation of Elephant Butte dam on the discharge at El Paso is evident. Note the oscillation (higher high flows and lower low flows) evident in the early 1900s, which disappears after the construction of Elephant Butte Dam in 1916. Third, significant, yet declining, flow is pro¬ vided by the Rio Conchos near Presidio. Below the confluence of the two rivers, the Rio Conchos provided 83% of the total downstream flow during the period 1961 to 1999, but only 55% during the period 1992 to 1999 (Brock et al., 2001). This demonstrates a considerable decline in the amount of water added to the main chan¬ nel via the Rio Conchos, a symptom of both Mexican water use and long-term drought. Fourth, from the mid- 1990s on, the flow in both the Rfo Conchos and the Rio Grande has diminished to protracted low levels not seen since the drought of the 1950s. For a complete over¬ view of the hydrology of the Rio Grande, see Schmidt, et al. (this volume). Legislative action that governs appropriation of Rio Grande waters in the Chihuahuan Desert region begins with the 1902 (extended in 1905 to include western Texas) Reclamation Act (Littlefield, 2000). This act authorized the constmction of Elephant Butte Dam and reservoir. Waters stored in the reservoir would be diverted to users through a system referred to as the Rio Grande Project. This 1905 law was the first interstate allocation of any river mandated by Congress. In 1906, an agreement was made between the U.S. and Mexico, which required that 60,000 acre-feet (AF) of water be allocated to Mexico Annual Runoff in Million Acre-Feet 16 Special Publications, Museum of Texas Tech University 1900 1920 1940 1960 1980 2000 Figure 2. Annual runoff for selected sites along the Rio Grande in the northern Chihuahuan desert region (IBWC, 2002). The y-axis is fixed with a range of 0 to 2.5 MAF to aid in visual comparison between graphs. Note that for certain years the El Paso and down¬ stream from Presidio graphs exceeded this value. Urbanczyk — Contemporary Water Supply in West Texas 17 annually at a location upstream from Ciudad Juarez (Eq¬ uitable distribution of waters of the Rio Grande Conven¬ tion between the United States and Mexico, available at: http://www.ibwc. state. gov/FORAFFAI/ 1906_convention.HTM). Elephant Butte Dam was con¬ structed in 1916, primarily for irrigation and flood con¬ trol. Part of the appropriations from this project included enough water to irrigate 88,000 acres (35,600 ha) in New Mexico, and 67,000 acres (27,000 ha) in western Texas. Appropriation of water between Colorado, New Mexico, and Texas is governed by the Rio Grande Com¬ pact of 1939 (Rio Grande Compact, reprinted in NMWRRI, 2000). This compact defines the obligations of Colorado and New Mexico to deliver water ultimately to the Elephant Butte reservoir (from there to be distrib¬ uted via the Rio Grande Project). The Rio Grande Com¬ pact requires minimum discharges be maintained at a series of gauging stations along the course of the upper Rio Grande. A complex set of equations was established to determine the amount of flow to be delivered from Colorado to New Mexico, and on to Elephant Butte. Total available water was therefore variable as discharge var¬ ied, but guaranteed flow to both New Mexico and Texas. Approximately 2.5 MAF of water are available for use within the Rio Grande Compact states annually. Most of this water is used to irrigate nearly 1,000,000 acres (405,000 ha) in the upper Rio Grande basin. Approxi¬ mately 600,000 acres (243,000 ha) are in the San Luis valley of Colorado; additional irrigated acreage is located in the “middle” Rio Grande valley in New Mexico. Ap¬ proximately 300,000 AF/year are used in the “middle” Rio Grande valley (Wilson, 2000), while 60 to 80% of flow at the Otowi gage (near Santa Fe) must be by¬ passed to Elephant Butte Reservoir due to Compact re¬ strictions. Downstream from Elephant Butte, the Rio Grande Project (agricultural irrigation) and the delivery to Mexico (1906 Treaty) consume most (or all) of the remaining flow. Only in wet years is there any expected (or actual) flow downstream from Ft. Quitman. In fact, the average flow at Ft. Quitman is 140,000 AF/year, only 5% of the total water supply in the upper Rio Grande basin. The river is completely appropriated. New Mexico is currently in compliance with the Rio Grande Compact with regards to the delivery of water to Texas. It had accrued a total of 529,000 AF deficit to Texas after an all-time low of 19,000 AF stored in Elephant Butte in 1951 (Mutz, 2000). This deficit was erased by 1972, and New Mexico has been in compliance with the Rio Grande compact since then. There continues to be diminished flow in the Rio Grande channel downstream from Ft. Quitman until the confluence with the Rio Conchos (Figure 2). At this point, the Rio Grande once again becomes a perennial stream, with most of the water provided by the Rio Conchos. Figure 2 demonstrates the impact the Rio Conchos has on the main Rio Grande channel. Note the significantly different character of the annual runoff curve seen in the Rio Grande downstream from the Rio Conchos confluence when compared to the Rio Grande upstream from Presidio. Also apparent on this Figure is the diminished flow provided by the Rio Conchos since the mid 1990s. Appropriation of the Rio Conchos, and delivery of water to Texas, is governed by the 1944 Treaty between Mexico and the United States (Treaty Between the United States of America and Mexico, avail¬ able at http://www.ibwc.state.gov/FORAFFAI/ treaties.HTM). In summary, Mexico is entitled to two thirds of the flow reaching the main channel of the Rio Grande through a series of rivers and streams, the larg¬ est of which is the Rio Conchos. This is subject to the U.S. right to an average of 350,000 AF/year in cycles of five consecutive years. According to this agreement, Mexico is currently in a deficit situation (Kelly, 2001). As of October 1997, Mexico owed the U.S. 1.024 MAF, a figure that is twice the deficit incurred by Mexico dur¬ ing the drought of the 1950s. An additional 0.48 MAF deficit has been added as of early 2000 (Brock et al., 2001). According to the 1944 Treaty, Mexico is obli¬ gated to repay the water debt by October 2002. There does exist some question about this, though, because a provision exists in the treaty to alter the deliveries during “extraordinary drought”. The exact definition of “ex¬ traordinary drought” is not made, and the drought of the mid-1990s might qualify. Another water source in the western Texas region is the Pecos River. In its western Texas reach, flow in the Pecos River is controlled by releases from the Red Bluff Reservoir. The delivery of water from New Mexico is the primary control on storage in the Red Bluff Reser¬ voir. The Pecos River Program (NMSEO, 1998) allo¬ cates Pecos River water between New Mexico and Texas. In 1988, the U.S. Supreme Court determined that New Mexico had underdelivered an average of 1 0,000 AF/year during the period 1950 to 1983. New Mexico agreed to pay $14 million to Texas to eliminate this defi¬ cit. Today, average daily discharges along the Pecos River downstream from Red Bluff Reservoir vary from 4 to 15 cubic-feet per second (0.1 to 0.4 cubic-meters per second) (FWTRWPG, 2001). 18 Special Publications, Museum of Texas Tech University Groundwater Resources Groundwater is another major source of water for the northern Chihuahuan Desert region. Groundwater is inherently more difficult to study than surface water, particularly for estimating total resources. Aquifers are recharged by precipitation that infiltrates into the ground, by losing streams, by inflow from adjacent aquifers, and by irrigation return flow. Aquifer types in the general northern Chihuahuan Desert region include bolson type aquifers, alluvial aquifers, limestone aquifers, and igne¬ ous aquifers (Mace et al., 2001). The following discus¬ sion focuses primarily on the western Texas aquifers with information pertaining to the development of the 2002 State Water Plan for Texas. This plan will super¬ sede the existing 1997 plan. The 2002 plan is the first to be adopted since the passage of Senate Bill 1, which has allowed for more public participation in the production of regional water plans. The portion of Texas that in¬ cludes the northern Chihuahuan Desert is located in Re¬ gion E of the 2002 State Water Plan (Figure 3). This region includes the counties of El Paso, Hudspeth, Culberson, Presidio, Jeff Davis, Brewster, and Terrell. Associated with the water planning and the preparation of the 2002 State Water Plan has been research and data gathering by the Texas Water Development Board (TWDB). These data are used in the following sum¬ mary of the most detailed assessment of water resources in the northern Chihuahuan Desert. Aquifer Details The TWDB has formally designated several aqui¬ fers in western Texas (TWDB, 1997; Figure 3). Note that in Figure 3, several of these aquifers overlap be¬ cause the Figure includes both surface and subsurface spatial locations for the aquifers. These include the Hueco-Mesilla Bolson, Cenozoic Pecos Alluvium, and Edwards-Trinity (major aquifers), and the Bone-Spring Victorio Peak, Capitan Reef Complex, West Texas Bolsons, Igneous, Rustler, and Marathon (minor aqui¬ fers) (Mace et al., 2001). The Hueco-Mesilla Bolson and West Texas Bolson aquifers are located in sedimen¬ tary deposits associated with Basin and Range type ex- tensional tectonics typical of the southwestern United States. This type of geologic activity produces linear mountain ranges separated by linear basins, which fill with sedimentary deposits as the mountain ranges rise over time. These linear basins are typically fertile sources of groundwater with recharge of water occurring pri¬ marily at the margins of the basins. Like other ground- water sources in this arid area, these aquifers are not recharged at rapid rates. The TWDB data compilation (TWDB, 1997) includes an estimate of “sustainable” water supply based upon precipitation and recharge rates for selected aquifers, and also includes an estimate of total useable (non-saline) water in storage for each of the aquifers. These figures for the Hueco-Mesilla Bolson aquifer are 0.024 MAF/year recharge, and 9 MAF stor¬ age, and 0.024 MAF/year recharge and 7 MAF storage in the West Texas Bolson aquifers. Recharge rates of 1% are estimated for the west Texas Bolson aquifers. The Edwards-Trinity aquifer extends only partly into the Chihuahuan Desert region. This large aquifer system is located in Cretaceous limestone. It extends eastward into central Texas where it is connected to the Ogallala and the Edwards aquifers. It has approximately 145 MAF in storage, with an effective recharge rate of 0.776 MAF/yr. The Cenozoic Pecos Alluvium aquifer is located in alluvial deposits of the Pecos River. This aquifer has an estimated recharge rate of 0.071 MAF/year, and 9.5 MAF in storage (available non-saline water). More than 200 feet (61 m) of water level declines have occurred in this aquifer in Reeves and Pecos counties. Groundwater that once contributed base flow to the Pecos River now flows in the subsurface to areas with heavy withdrawals. The Bone Spring-Victorio Peak aquifer is located in joints, fractures, and cavities in Permian limestone beds. Recharge estimates are 0.09 MAF/year (annual recharge and irrigation return flow), and no total storage estimates are available. The Capitan Reef Complex aqui¬ fer is also located in Permian limestone, including the Capitan reef and reef talus. These limestone beds are commonly very porous (vuggy) and in extreme condi¬ tions may be cavernous (Carlsbad Caverns are located in this type of rock). Recharge estimates are 0.012 MAF/year, and total storage is estimated to be 0.385 MAF. Urbanczyk — Contemporary Water Supply in West Texas 19 Figure 3. Major and minor aquifers of western Texas (TWDB, 2001b). The Region E counties are labeled in the top map. 20 Special Publications, Museum of Texas Tech University The Igneous aquifers are located in Brewster, Presidio, and Jeff Davis counties. These aquifers are located in Tertiary (approximately 30 to 40 million years old) volcanic and volcaniclastic deposits, and associated recent alluvial sediments. They have estimated recharge rates of 0.014 MAF/year. The TWDB estimates total recharge rates in the aquifer system to be 2.5% of pre¬ cipitation. The Marathon aquifer is located in Paleozoic ma¬ rine sediments. This aquifer has an estimated recharge of 0.018 MAF/year, and a recharge rate of 2.5% of total precipitation. The Rustler aquifer exists in up to 500 feet (152 m) of basinal limestone, dolomite, and evaporites repre¬ senting the demise (drying up) of the Permian reef ba¬ sin. The evaporite beds formed as the inland bay was cut-off from seawater circulation with the open ocean. Estimated recharge rates are 0.004 MAF/year. Water in this aquifer is not suitable for human consumption due to the high total dissolved solids (up to 6,000 mg/L). Social Impacts and Supply vs. Demand Predictions The population of Texas is expected to double in the next 50 years, from 21 million in 2000, to 40 million in 2050 (TWDB, 2001a). The Region E population is expected to increase from 800,000 to 1,587,097. The project population for El Paso county alone is 1 ,536,423, a 99% increase over the year 2000 census. By the year 2050, 38 percent of Texas’ population will need to re¬ duce demand or develop addition resources to meet pro¬ jected demands during drought conditions (TWDB, 2001a). Agriculture, which is currently the largest wa¬ ter user in the state, will be surpassed by the combina¬ tion of municipal and manufacturing demand. Impacts of this on western Texas include decreased irrigation due to depletion of groundwater resources and increase of groundwater use by urban centers such as El Paso. The TWDB recommends that the state Legislature es¬ tablish protection of rural-community access to local water resources. They also recommend the use of groundwater models to evaluate the long-term sustain¬ able levels of groundwater aquifers. Water demand in Texas is projected to increase from the current 17 MAF per year to 20 MAF in the year 2050. Region E demand is expected to increase from 0.509 to 0.586 MAF/year. Municipal demand is expected to increase by 67%, while irrigation demand is expected to decrease 12%. This decline will be due to more efficient irrigation systems and canal delivery systems, declining groundwater sup¬ plies, and the transfer of groundwater rights to munici¬ pal use as population increases. Current per capita wa¬ ter use throughout the state is 181 gallons/person/day (gpd), with a range of 275 gpd in Richardson and a low of 120 in Killeen (El Paso ranks near the bottom, using 144 gpd). Conservation efforts are expected to reduce the average per capita use to 159 gpd by 2050. Total water supply projections (surface and ground, from existing sources) indicate an expected decrease of 18%, from 17.8 MAF/year in 2000 to 14.5 MAF/year in 2050 (TWDB, 2001b). Total groundwater availability is estimated by the Regional Planning groups to be 14.9 MAF/year. Total groundwater supplies (water acces¬ sible with existing infrastructure) are estimated to be 8.8 MAF/year in 2000, and are projected to decline 18% by 2050. Note that these “supply” estimates include the water available with existing infrastructure, and differ from the recharge and total storage estimates listed in the aquifer descriptions above. Statewide, groundwater constituted 50% of the total water supply in 2000 and is projected to provide the same in 2050. In Region E, groundwater constituted 79% of the total water supply in 2000, and is projected to supply 88% by the year 2050. These estimates again pertain to existing sources, and are skewed by the anticipated depletion of the usable portion of the Hueco-Mesilla bolson aquifer by the year 2030. The depletion of this major aquifer will create a critical need to find other sources to meet the projected growing demands for water. El Paso will clearly be un¬ able to meet water demands by 2030 considering the existing supply. The Region E planning group recom¬ mends the following strategies to increase supply: 1) obtain additional surface water from conservation sav¬ ings in irrigation; 2) purchase irrigation rights; 3) reuse; 4) desalinate; 5) purchase and use groundwater from outside of El Paso County. The impacts of groundwater transfers from rural counties will become a critical is¬ sue. Urbanczyk — Contemporary Water Supply in West Texas 21 Related Information from Northern Mexico Only about one quarter of 60 aquifers in Chihua¬ hua have been studied in any detail (Kelly, 2001), and most water level measurements were suspended in 1990. The Mexican National Water Commission (Comision Nacional de Aguas (CNA)) has identified several over- exploited aquifers which are listed in Table 1. Currently, only 1% of the wells have any type of metering (CNA, 1997). Table 1. Major over-exploited aquifers in the Rio Conchos basin ( data from CNA, 1997). Aquifer Total Annual Pumping (MAF) Total Annual Recharge (MAF) % Over- Exploitation Chihuahua-Sacremento 0.102 0.045 127% Jimenez-Camargo 0.475 0.361 88% Parral-Valle de Verano 0.026 0.021 21% Tabaloapa-Aldama 0.054 0.045 19% Long-term Prognosis The long-term prognosis for the west Texas and northern Chihuahuan Desert region is difficult to deter¬ mine. The projected population growth and anticipated depletion of groundwater reservoirs are critical water resource issues that need careful attention. Anthropo¬ genic manipulation of surface and groundwater sources has and will further impact the ability of the water sources to sustain wildlife populations. This issue is compounded by the fact that we are in a significant drought. It is clear that we are using several of the groundwater sources at levels that are not sustainable. This has led policy makers to develop strategies to try to increase future municipal water supply. Some strategies proposed in west Texas, such as converting irrigation appropriations to municipal supply or increasing the amount of “reused” water, will not likely diminish the amount of available surface water for support of wild¬ life. However, desalination and interbasin transfer of groundwater could impact local surface water resources. Sharp et al. (this volume) suggest that there are pro¬ longed groundwater flow paths from locations such as the Wild Horse basin aquifer near Van Horn that likely feed spring systems such as San Salomon at Balmorhea. The export of water from the Wild Horse basin aquifer could, therefore, impact flow rates from these springs. Detailed studies such as these should be done to evaluate the impact of interbasin transfers of groundwater. Careful attention must be given to compliance with respect to legislation and treaties governing deliveries of surface water between states and countries. It is impor¬ tant that we recognize how the effects of the current drought impacts Mexico’s ability to deliver water via the Rio Conchos. That deficit is definitely not the only sur¬ face water supply problem in west Texas. The over appropriation of the Rio Grande is at fault, also. There are no easy solutions to any of our water resource is¬ sues. 22 Special Publications, Museum of Texas Tech University Conclusion The information presented in this paper is repre¬ sentative of the water supply issues in the arid northern Chihuahuan Desert. Surface water resources are very limited, and are appropriated to the point of severe dam¬ age to natural ecosystems such as the Rio Grande. Many argue that we can adhere to legislation such as the Rio Grande Compact without the ecological destruction that currently occurs, and will likely occur in the future (Har¬ ris, 2000). As population increases in this region, there will be a continuous shift to more dependence on ground- water rather than surface water to meet the growing demand. A necessary consideration will be the sustain¬ able limits to groundwater withdrawals that will not com¬ pletely mine these limited resources and will allow for the presence of both humans and wildlife. Literature Cited Brock, L., Kelly, M.E., and Chapman, K. 2001 . Legal and institu¬ tional framework for restoring instream flows in the Rio Grande: Fort Quitman to Amistad. Texas Center for Policy Studies, document presented to the World Wildlife Fund, http://www.texascenter.org/publications/instreamflow.pdf (10/22/01). CNA (Comision Nacional del Agua). 1 997. Programa hydraulico de gran vision, Estado de Chihuahua ( 1 996-2020). Comision Nacional del Agua, http://www.sequia.edu.mx/plan-hidra/ (10/22/01). FWTRWPG (Far West Texas Regional Water Planning Group). 2001. Executive Summary, Region E Far West Texas Re¬ gional Water Plan, http://www.twdb.state.tx.us/assistance/ rwpg/main-docs/regional-plans-index.htm( 10/22/01). Harris, S. 2000. Opportunities and constraints for environmental enhancement and recreation along the Rio Grande. The Rio Grande Compact: It’s the Law. Proceedings of the 44th Annual New Mexico Water Conference, New Mexico Water Resources Research Institute, WRRI Report #3 1 2, Decem¬ ber 2-3, 1999, Santa Fe, New Mexico, 5 pp. IBWC (International Boundary and Water Commission). 2002. Historical Rio Grande flow conditions, http:// www.ibwc.state.gov/wad /histflol.htm (05/10/02). Kelly, M. E. 2001. The Rio Conchos: A preliminary overview. Texas Center for Policy Studies, Austin, Texas, 27 pp. Littlefield, D.R. 2000. The history of the Rio Grande Compact of 1938. The Rio Grande Compact: It’s the Law. Proceedings of the 44th Annual New Mexico Water Conference, New Mexico Water Resources Research Institute, WRRI Report #3 1 2, December 2-3, 1 999, Santa Fe, New Mexico, 8 pp. Mace, R. E., W. F. Mullican III, and E. S. Angle (editors). 2001. Aquifers of West Texas. Report 356. Texas Water Devel¬ opment Board, Austin, Texas, 272 pp. Mutz, P. 2000. Post Compact delivery of water by New Mexico. The Rio Grande Compact: It’s the Law. Proceedings of the 44?1 Annual New Mexico Water Conference, New Mexico Water Resources Research Institute, WRRI Report #3 1 2, December 2-3, 1999, Santa Fe, New Mexico, 4 pp. http:// wrri.nmsu.edu/publish/watcon/proc/proc44/contents.html (10/22/01). NMSEO (New Mexico State Engineers Office). 1998. Office of the State Engineer, Interstate Stream Commission, 1997- 1998 Annual Report, http://www.seo.state.nm.us/publica- tions/97-98-annual-report/index.htm ( 1 0/22/0 1 ). NMWRRI (New Mexico Water Resources Research Institute). 2000. Reprint of the Rio Grande Compact The Rio Grande Compact: It’s the Law. Proceedings of the 44lh Annual New Mexico Water Conference, New Mexico Water Resources Research Institute, WRRI Report #3 1 2, December 2-3 , 1 999, Santa Fe, New Mexico, 12 pp. TWDB (Texas Water Development Board). 1997. Water for Texas: A consensus based update to the State Water Plan; Volume II Technical Planning Appendix, Texas Water Development Board, Document # GP-6-2 . TWDB (Texas Water Development Board). 2001a. Draft, State Water Plan, Water for Texas 2002, draft version available for public comment at http://www.twdb.state.tx.us (10/22/01). TWDB (Texas Water Development Board). 2001b. Major and minor aquifers of Texas, http://www.twdb.state.tx.us/ gis_data /aquifers/ (05/10/02). Wilson, L. 2000. Surface water hydrology of the Rio Grande Basin. The Rio Grande Compact: It’s the Law. Proceedings of the 44th Annual New Mexico Water Conference, New Mexico Water Resources Research Institute, WRRI Report #312, December 2-3, 1999, Santa Fe, New Mexico, 12 pp. Wurbs, R.A., Sanchez-Torres, G, and Dunn, D.D. 1994. Reser¬ voir/river system reliability considering water rights and water quality. Technical Report No. 165, Texas Water Re¬ sources Institute, Austin, Texas. Urbanczyk — Contemporary Water Supply in West Texas 23 Address of author: Kevin Urbanczyk Department of Earth and Physical Sciences Sul Ross State University Alpine, TX 79832 e-mail address: kevinu@sulross.edu Hydrology and Geomorphology of the Rio Grande and Implications for River Rehabilitation John C. Schmidt, Benjamin L. Everitt, and GigiA. Richard Abstract The Rio Grande watershed includes a northern and southern branch that have very different hydro- logic regimes. The natural flood regime of the north¬ ern branch is snowmelt driven, and that of the south¬ ern branch, the Rio Conchos, is driven by summer rainfall. Downstream from the confluence of the two branches, near Presidio, Texas, the natural pattern of high and low flow was dominated by runoff from the Conchos basin between July and the following March prior to the construction of large dams. Dams and diversions greatly altered the natural hydrologic regime of both branches. The magnitude of the 2-year recur¬ rence flood of the Rio Grande at El Paso, on the north¬ ern branch, declined by 76% after 1915. The magni¬ tude of the 2 -year recurrence flood downstream from Presidio was reduced by 49% after 1915. Dams and diversions have also significantly al¬ tered the natural sediment flux, and significant geo- morphic adjustments of the channel have resulted. The northern branch includes reaches where degradation or aggradation has occurred during the past century. Reaches immediately downstream from dams have degraded beds and narrowed widths. Further down¬ stream, the channel bed has aggraded, and the channel width has narrowed. Channelization and levee con¬ struction have occurred in some of these same river segments. Restoration, defined as returning an ecosystem to a close approximation of its condition prior to dis¬ turbance, is impossible on the main stem of the Rio Grande because of current institutional demands on stream flow and the extent of alteration of the flood- plain. Rehabilitation, defined as returning essential physical and ecological functions to a degraded eco¬ system, is a more appropriate goal for the Rio Grande. In light of the diverse styles of twentieth-century chan¬ nel adjustments that have occurred throughout the basin, different river segments must be assigned dif¬ ferent rehabilitation goals. Introduction The Rio Grande has the second longest river course and had the sixth largest mean sediment dis¬ charge in North America before the continent was settled extensively by Europeans (Meade et al., 1990). Human activity has disrupted the natural flux of water and sediment. Large dams store floods for subse¬ quent diversion, and these dams also trap sediment. The total volume of stream flow has been reduced, and the magnitude of floods in some parts of the Rio Grande have been reduced by more than 50%. Meade et al. (1990) estimated that annual sediment delivery to the Gulf of Mexico decreased from about 30 x 106 tonnes in 1700 to about 0.8 x 106 tonnes in 1980. These changes have caused significant adjustments of the channel of the Rio Grande. Historically, the Rio Grande had a mobile bed and erodible banks, and the channel changed from year to year. Today’s channel is smaller, more stable, changes less from year to year, and infrequently inundates its former floodplain. The riverine ecosystem has adjusted to these changes in ways that do not benefit some native spe¬ cies. Inundation of the floodplain, which now occurs rarely in some segments, is necessary for recruitment in the riparian forest that lines the Rio Grande (Moles et al., 1998). Non-native salt cedar ( Tamarix sp.) has widely colonized abandoned alluvial surfaces of the once-wider channel. The endangered Rio Grande sil- 26 Special Publications, Museum of Texas Tech University very minnow ( Hybognathus amarus ) is adapted to the fonner wide shallow braided channel and associated habi¬ tats, and its population has declined greatly in response to channelization and diminished flows. The purpose of this review paper is to describe hydrologic and geomorphic conditions of the river dur¬ ing the past century and to summarize changes in the water and sediment flux. We describe some of the geo¬ morphic adjustments of the channel and its floodplain that have occurred during the past century, emphasizing channel changes downstream from the large dams on the northern branch: Cochiti Dam and Elephant Butte and Caballo dams. These changes in hydrology, sedi¬ ment transport, and physical characteristics of the chan¬ nel and floodplain affect the aquatic and riparian ecosys¬ tem of the river. We conclude by commenting on the implications of these physical changes to development of a basin wide strategy for rehabilitating physical at¬ tributes and processes of the riverine ecosystem. Description of the Rio Grande Drainage Basin The hydrologic regime of the Rio Grande down¬ stream from Presidio, Texas, results from the combined flow of northern and southern branches of the river (Fig¬ ure 1). The drainage basin of the northern branch, called the Rio del Norte by Spanish explorers, comprises about two-thirds of the total watershed area upstream from Presidio. The flow of this branch, called the Rio Grande in the United States and the Rio Bravo in Mexico, is pri¬ marily contributed by snowmelt in the southern Rocky Mountains, and this branch had its annual peak flow in late spring, prior to the construction of dams. The Rio Conchos, whose headwaters are in the Sierra Madre Occidental, is the southern branch. Although the Rio Conchos basin is smaller than that of the northern branch, its mean annual runoff is much larger, and this branch has its maximum flows in late summer. Several names are used to describe the different parts of the northern branch. The basin upstream from Elephant Butte Reservoir was referred to as the Upper Basm by Dortignac (1956) and the Northern Rio Grande by Graf (1994). Scurlock (1998) defined the segment between the Rio Chama and Elephant Butte Reservoir as the Middle Basin and many studies of this segment refer to it as the Middle Rio Grande, distinguishing it from the Upper Rio Grande that occurs upstream from the Rio Chama. For our purposes, we use the term northern branch when referring to the entire basin upstream from the Rio Conchos, and we refer to shorter river reaches by specific geographical names. The river flows through a series of structural ba¬ sins, where the alluvial valley is very wide, separated by intervening canyons where the valley is narrow. The occurrence of wide alluvial valleys and intervening nar¬ row canyons is important in analyzing channel adjust¬ ment to the regulation of stream flow and sediment flux. Rivers typically have lower gradients in wide alluvial val¬ leys where they have large floodplains and meandering channels. Typically, channels narrow to greater extents in alluvial segments with flat gradients, and channel ad¬ justments are less in narrow canyons (Grams and Schmidt, 2002). The northern branch’s headwaters are in the San Juan, Sangre de Cristo, and Jemez mountains of Colo¬ rado and New Mexico (Figure 1). The most upstream beginnings of stream flow occur near Stoney Pass in the San Juan Mountains. The Rio Grande leaves the San Juan Mountains near Del Norte, Colorado, and enters the San Luis Valley of south-central Colorado. This val¬ ley is a deep structural basin at the northern end of the Rio Grande Rift that is filled with more than 9,000 m of alluvium. The Rio Grande has a low gradient and has not significantly incised its channel through these sedi¬ ments. Thus, the Rio Grande is easily diverted onto adjoining valley lands here, and irrigation is extensive. South from the San Luis Valley, the Rio Grande enters a narrow canyon through the Taos Plateau - the Canon del Rio Grande. Further downstream are Espanola Basin, White Rock Canyon, and the Santo Domingo- Albuquerque-Belen basin. The large basins of central New Mexico have been aggrading for as much as 1 1 ,000 years (Sanchez and Baird, 1997), and the Rio Grande channel is not significantly incised into the sediments of the alluvial valley. Aggradation in these basins continues to the present. Two basins in southern New Mexico - Engle and Las Palomas Valleys - are partially inundated by Elephant Butte and Caballo reservoirs, respectively. The releases from these reservoirs are diverted for agri¬ culture in the Mesilla and El Paso/Juarez valleys further downstream. Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 27 Rio Grande-Rio Bravo Drainage Basin/ Cuenca del Rio Grande-Rio Bravo Figure 1 . Map showing entire drainage basin of the Rio Grande/Rio Bravo. The El Paso/Juarez valley is about 136 km long, 16 km wide in places, and extends downstream to approxi¬ mately Fort Quitman, Texas (Stotz, 2000). Downstream from Fort Quitman, the Upper Canyon segment includes 200 km where there are 11 different canyons and as many intervening alluvial valleys. The longest individual canyon is 14.7 km long, and canyon reaches comprise about 24% of the Upper Canyon segment upstream from Candelaria. The Upper Canyon segment also includes the Presidio Valley, which is about 120 km long between Candelaria and Presidio. The Presidio Valley is less than 5 km wide. The river is mostly channelized and leveed here. The Rio Grande is joined by the Rio Conchos near Presidio. Downstream from the Rio Conchos, the Rio Grande flows through alternating alluvial and confined reaches in the Big Bend section, including four narrow canyons that are popular for recreational boating - Colo¬ rado, Santa Elena, Mariscal, and Boquillas canyons (Aulbach and Gorski, 2000). The Lower Canyons ex¬ tend to the headwaters of Amistad Reservoir (Aulbach and Butler, 1998). 28 Special Publications, Museum of Texas Tech University Downstream from Amistad Reservoir, the Rio Grande exits its canyons and flows across the Gulf Coast piedmont. With the added contributions of the Pecos and Devils Rivers, it still occasionally lives up to its names “Grande” (Big) and “Bravo” (Wild). Peak flows from occasional autumn hurricanes exceed 25,000 mVs. Downstream from Laredo, Texas, the Rio Grande wan¬ ders across its delta plain of fine-grained alluvial depos¬ its. The History of Water Development Agricultural use of the Rio Grande in New Mexico began in pre-history (Table 1). Pueblo peoples were utilizing ditch irrigation on a limited scale at the time of Spanish exploration in 1591 (Scurlock, 1998). Graf (1994) speculated that, “Diversion works on the main stream probably consisted of brush and boulder struc¬ tures . . . [that] probably washed away with each spring flood.” Spanish and Mexican settlers in New Mexico expanded irrigation on floodplains and terraces of the Rio Grande, and the area of irrigated farming steadily increased in New Mexico until it reached a peak of 50,500 ha in 1880 (Sorenson and Linford, 1967, cited by Scurlock, 1998). Ditch irrigation began in the mid- 1600s in the El Paso/ Juarez Valley and direct diversions of the main channel in this valley were underway by at least the late 1700s (Stotz, 2000). Water was being diverted from the Rio Conchos for use at the presidio in the Presidio Valley by 1750. Of the 63 dams built in the northern branch water¬ shed prior to 1916, 48 were in Colorado, and their pur¬ pose was to facilitate irrigation in the San Luis Valley. Between 1855 and 1893, 8 dams were built there whose cumulative reservoir storage was 4.08 x 106 m3 (data from U. S. Army Corps of Engineers, 1996). Between 1894 and 1915,55 more dams were built in the northern branch watershed, and the cumulative reservoir storage increased more than 100 times to about 486 x 106 m3. Depletions of stream flow caused by irrigation with¬ drawals have been substantial for more than a century. Kelley (1986) estimated that more than half the summer stream flow from central and northern New Mexico be¬ tween 1890 and 1893 was consumed by irrigation. Kelley (1986) also estimated that 74% of the Rio Grande’s stream flow was lost to seepage, evapotranspiration, and irrigation between the Mesilla Valley in southern New Mexico and Presidio during the same period. Without irrigation, Kelley (1986) estimated that losses would only have been about 35%. Between 1936 and 1953, the average annual depletion in the San Luis Valley was 9.9 x 108 m3, and annual depletions ranged from about 6.2 x 108 m3 in dry years to more than 12.3 x 108 m3 in wet years. Depletions in central New Mexico were of a similar magnitude during this period (Thomas et al., 1963). Elephant Butte Dam was completed in 1916, and had an initial capacity of about 2.93 x 109 m3. The dam was built to control floods and ensure the delivery of irrigation water to southern New Mexico and to Mexico. At the time of completion, Elephant Butte Reservoir had a capacity of 2.5 times the mean annual discharge and was the largest reservoir in the world. Its construction increased the total reservoir storage in the basin by more than 6 times to 3,390 x 106 m3 (Figure 2). Small reservoirs, low head main stem diversion structures, levees, and channelization works were built throughout central New Mexico in the 1920s (Scurlock, 1998). These construction activities were directed by the Middle Rio Grande Conservancy District, organized in 1925. Diversion dams directed stream flow into ex¬ tensive irrigation canals at Cochiti, Angostura, Isleta, and San Acacia. The construction of levees to prevent avul¬ sions into surrounding agricultural lands along the river exacerbated the aggradation by confining sediment depo¬ sition to a smaller area (Scurlock, 1998; Sanchez and Baird, 1997). The construction of levees, begun in the 1920s, became a comprehensive channelization scheme that was completed in central New Mexico by the early 1960s (Graf, 1994). El Vado Dam on the Rio Chama was completed in 1935 for flood control and irrigation supply. Caballo Dam, immediately downstream from Elephant Butte, was completed in 1938, and total basin wide reservoir stor¬ age increased to 4.37 x 109 m3. Together, Elephant Butte and Caballo completely stored the annual snowmelt flood in every year between 1915 and 1941, and there were no flood releases downstream. The years 1941 and 1942 had unusually large runoff, however, and the dams and levees of that time were not able to control those floods. Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 29 Table 1. Dams and other structural modifications in the Rio Grande basin upstream from Amistad Reservoir. Date Event 1200- 1850s 1659 .1899 1916 1925 1925-1935 1926 1933 1935 1938 1938 1941 and 1942 1940s 1950s and later 1950s? 1963 1967 1969 1971 1971 1973 Pueblo, Spanish, and Mexican temporary diversion structures in the Rio Grande channel in New Mexico with gradual expansion of irrigated area in central New Mexico Founding of mission at Paso del Norte, temporary diversion and headgate constructed Cordoba Island cut-off, El Paso-Juarez Elephant Butte Dam completed Middle Rio Grande Conservancy District organized Diversion dams at Cochiti, Angostura, Isleta, and San Acacia completed, 290 km of riverside drains and 260 km of interior drains constructed in irrigated fields of central New Mexico Salt cedar planted for erosion control in Rio Puerco basin Channelization through Mesilla Valley to El Paso El Vado Dam on the Rio Chama completed Caballo Dam completed Rectification and channelization, El Paso to Ft Quitman Large floods cause 27 levee breaks near Albuquerque Rio Puerco sediment control structures and revegetation Channelization of the Middle Rio Grande Sediment control dams on tributary arroyos between Elephant Butte and Fort Quitman Abiquiu Dam completed Settlement of the “Chamizal” boundary dispute and construction of concrete-lined channel separating El Paso and Juarez Amistad Dam completed Heron Dam completed Transbasin diversion from the San Juan River Cochiti Dam completed YEAR Figure 2. Graph showing time series of cumulative reservoir storage in the northern branch. 30 Special Publications, Museum of Texas Tech University Abiquiu Dam was built on the Rio Chama in 1963 as part of the Colorado River Storage Project, and di¬ versions from the San Juan River into the Chama ba¬ sin began in 1971. Today, Abiquiu is the second larg¬ est dam in the northern branch watershed. Cochiti Dam on the Rio Grande, located 65 km upstream from Albuquerque, was completed in 1973. It provides the largest flood control storage volume on the northern part of the main stem (Bullard and Lane, 1993). The dam was completed in November 1973 for flood control and sediment detention (U.S. Army Corps of Engineers, 1978) and traps virtually the en¬ tire sediment load from upstream (Dewey et al., 1979). Large dams have also been constructed on the Rio Conchos, creating La Boquilla Reservoir in 1913, Francesco I. Madera Reservoir in 1947, and Luis L. Leon Reservoir in 1967. The largest reservoirs in the Rio Grande basin are located downstream from Presidio: Falcon (completed in 1954; 3.18 x 109 m3) and Amistad (completed in 1969; 5.13 x 109 m3). The cumulative size of Amistad and Falcon reservoirs is greater than the total storage of all the reservoirs of the northern branch, illustrating the substantially greater stream flow that is regulated in the downstream parts of the Rio Grande/ Rio Bravo. Treaties The international position of the Rio Grande has played a significant role in its physical and hydrologic history, as well as its cultural history (Mueller, 1975). Under the Treaty of 1848, the segment between El Paso and the Gulf of Mexico was made the boundary between the two countries. Mapping of the river boundary was completed in 1852 (Emory, 1857). The active nature of the river was not anticipated, and within 30 years parts of the river had wandered kilometers from its 1852 course and dozens of oxbows were abandoned, making redefinition of the boundary necessary. The Treaty of 1884 included specific language providing for a move- able boundary, following the natural migration of the channel by erosion and accretion, but remaining fixed in the abandoned channel in the event of avulsion. The treaty also provided for additional mapping of channel changes, and restricted artificial modification of the chan¬ nel. The Treaty of 1970 provided for the first complete mapping of the 2000-km river boundary since 1 852. The treaty strengthened restrictions against artificial modifi¬ cation to include levees on the flood plain that might raise flood heights on the opposite bank. The Water Treaty of 1906 apportioned the flow of the northern branch, and provided for storage and deliv¬ ery of Mexico’s allotment via the Rio Grande Project. The 1944 Water Treaty allocated the water of the Rio Grande downstream from Presidio and gave the Inter¬ national Boundary and Water Commission authority to oversee measurement and distribution of stream flow. The treaty provided for the construction of international storage reservoirs. Reflecting the wartime emphasis on agriculture and industry, the treaty established the fol¬ lowing priority for use of stream flow: domestic and municipal uses, agricultural and stock-raising, hydroelec¬ tric power generation, other industrial uses, navigation, fishing and hunting, and other beneficial uses. Hydrology of the Basin Prior to 1915 The records of floods and droughts on the north¬ ern branch are preserved in the journals and notes of explorers and residents of the basin. Scurlock (1998) determined that there were at least 50 major floods ex¬ ceeding 280 m3/s in New Mexico between 1849 and 1942 and 51 floods in the El Paso/Juarez Valley since 1846. Twice as many known floods occurred in the 1800s than in the 1600s or 1700s. Scurlock (1998) and Stotz (2000) suggested that environmental degradation may have contributed to the increase in flood frequency in the 1800s, but Graf (1994) suggested that regional cli¬ mate change was a more likely cause. The largest flood occurred in 1 828 and had an estimated discharge of about 2,830 m3/s. During this flood, the entire Rio Grande valley was inundated from Albuquerque to at least El Paso. Other very large floods occurred in 1872 and 1884. Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 31 The gauged flow of the Rio Grande prior to 1915 reflected the impacts of irrigation withdrawal in the San Luis Valley and central New Mexico. The northern branch flooded in late spring, with a secondary peak in summer (Scurlock, 1998). The magnitude and average duration of the spring snowmelt flood increased in the down¬ stream direction between the San Juan Mountains and central New Mexico, as reflected in the difference be¬ tween measurements near Del Norte, at Embudo, and at Otowi Bridge (Figure 3). Between central New Mexico and El Paso, the magnitude of the snow melt flood did not increase, however, because there are no other large tributaries that drain high mountain ranges with signifi¬ cant annual snow fall. Thus, the magnitude of the spring snowmelt flood at Otowi Bridge was nearly the same magnitude as at El Paso (Table 2). Prior to 1915, the reach between El Paso and Presidio was a losing stream due to seepage losses, evapo- transpiration, and irrigation diversions (Kelley, 1986). The entire flow was sometimes diverted at El Paso, resulting in occasional dewatering of the river downstream (Everitt, 1993). The magnitude of the 2-year recurrence flood, prior to 1915, decreased from 209 to 122 m3/s between El Paso and the Rio Conchos (Table 3). In those years when the annual peak flow at El Paso was less than 1 00 mVs, no snowmelt flood peak reached the Rio Conchos. In years of greater snowmelt runoff, the magnitude of the peak flow at the Rio Conchos was never more than 90% of that measured at El Paso, and typically occurred 7 to 10 days after the peak had passed El Paso. The only times when stream flows at Presidio were significantly larger than at El Paso were in the late summer and early fall when flood flows were triggered by rainfall in the downstream parts of the basin. Table 2. Summary of hydraulic characteristics of the Rio Grande at selected gauging stations in late 1800s and early 1900s, before completion of Elephant Butte Dam. Gauging station location and period of record Median annual maximum mean daily discharge, in cubic meters per second Median date of the annual maximum mean daily discharge Mean annual discharge, in cubic meters per second1 Number of days whose median discharge exceeded twice the mean annual discharge near Del Norte (1/1/1890-5/31/1890; 7/1/1890-9/30-1896; 1/1/1904-12/31/1906; 1/1/1908-9/30/1915) 107.3 June 13 27.2 60 at Embudo (1/1/1889-3/31/1904; 9/1/1912-9/30-1915) 121.7 June 5 29.5 62 at Otowi Bridge (2/1/1895-12/31/1906) 146.9 June 4 37.3 61 at El Paso (5/10/1889-6/30/1893; 1/1/1897-12/31/1897; 2/1/1898-9/30/1915) 147.0 June 18 35.1 66 above Rio Conchos, near Presidio (1/23/1900-1/31/1900; 2/23/1900-2/28/1900; 3/23/1900-3/31/1914) 77.0 May 26 22.4 63 below Rio Conchos, near Presidio (5/1/1900-5/31/1914) 219.0 September 6 72.1 50 at Langtry (5/1/1900 - 9/30/1913) 694 August 14 70 — 1 computed as the mean of all days when measurements were made 32 Special Publications, Museum of Texas Tech University Figure 3. Graph showing median hydrographs of mean daily discharge of six gauging stations of the Rio Grande for varying periods in the late 1 800s and early 1 900s. See Table 2 for periods of record for each station. Table 3. Magnitude of floods of different recurrences, upstream and downstream from the Rio Conchos. Discharge, in cubic meters per second, of the annual maximum mean daily dis¬ charge, for the indicated period at the indicated location 1.25 yr 2 yr 5 yr 10 yr 1898-1916 at El Paso 98 209 378 484 above Rio Conchos, near Presidio 52 122 244 330 below Rio Conchos, near Presidio 217 567 1160 1545 1916-1996 at El Paso 32 51 100 124 above Rio Conchos, near Presidio 15 34 70 99 below Rio Conchos, near Presidio 126 288 661 1023 The natural hydrology of the Rio Grande changed dramatically downstream from the Rio Conchos (Figure 3). The Rio Conchos’ hydrology is entirely determined by rainfall, which is greatest in late summer and early fall in the Sierra Madre Occidental. This watershed yields the bulk of its natural stream flow between July and the following March (Table 4). Prior to 1915, the magni¬ tude of peak flows downstream from the Rio Conchos was approximately four times what they were upstream (Table 3). During September, when the Rio Conchos reached its annual maximum discharge, approximately 93% of the lower Rio Grande’s total monthly flow came from the Rio Conchos. Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 33 Table 4. Mean monthly discharge of the Rio Grande/Rio Bravo upstream and downstream from the Rio Conchos, near Presidio, 1901-1913. Month Mean monthly discharge above Rio Conchos, in cubic meters per second Mean monthly discharge below Rio Conchos, in cubic meters per second Percentage of mean monthly discharge of the lower Rio Grande/Rio Bravo that originated in the northern branch October 218 1002 22 November 125 592 21 December 117 537 22 January 101 298 34 February 95 362 26 March 139 297 47 April 233 291 80 May 723 795 91 June 1013 1273 80 July 543 1478 37 August 205 1828 11 September 207 2797 7 Channel Characteristics in the 1800s and Early 1900s The northern branch was an aggrading stream whose braided channel was constantly shifting (Graf, 1994). Large loads of sandy sediment and widely fluc¬ tuating flows caused the channel to be very wide and relatively shallow. As described in the El Paso/Juarez Valley by Major O.H. Ernst of the Army Engineer Corps in 1896 (cited by U.S. Department of State, 1903), “The size and character of the [Rio Grande] are ever varying, and its requirements as to form and dimension of bed vary equally. The river’s work of altering its bed to suit the necessities of the moment is never ending.” Channel change data for the part of the Rio Grande that is the international boundary demonstrate that the channel was very active and migrated rapidly across its sandy flood plain by both lateral erosion and avulsion (Mueller, 1975). Channel avulsions that were typically meander cutoffs during floods were most common in the wide alluvial valleys. In central New Mexico, the channel was generally straight with numerous braided channels. In the El Paso area, the channel had a meandering course at flood stage, had a braided channel at low flows, and changed course frequently. There is some evidence that the Rio Grande near El Paso had a narrow sinuous channel in early his¬ toric times, suggesting that the wide shallow channel of the late 1 800s was perhaps the result of a “metamorpho¬ sis” (Schumm, 1969) resulting from the flood of 1828. In central New Mexico in 1944, the Rio Grande at base flow was described by Rittenhouse (1944, p.150) as a ...“winding, elongated sand flat, averaging about 200-300 yards in width. One or more small low-water channels meander over the sand flat, re-working the de¬ posits in it. At high stages the entire sand flat, as well as the adjacent floodway area beyond the low banks, is under water. Between large floods the width of the sand flat is decreased by growth of cottonwoods and salt ce¬ dars. These may be removed or the entire channel shifted during high flows.” Rittenhouse (1944) also noted that the floodway was nearly 1 km wide. Lateral movements of the Rio Grande downstream from Cochiti Dam between 1918 and 1935 averaged 20 to 35 m/year (Richard, 2001). 34 Special Publications, Museum of Texas Tech University Hydrologic Changes in the Basin Changes to the hydrology of the Rio Grande since 1915 have been profound. Peak discharges declined upstream and downstream from Elephant Butte Dam. Changes upstream from the dam are probably due to regional climate change as well as changing patterns of irrigation diversion. These changes greatly diminished the magnitude and duration of the annual peak flood and changed the season in which these floods occur. The net effect of all changes has been to make the magnitude of the annual floods more similar throughout the north¬ ern branch (Figure 4). In fact, the average flood at Del Norte is now larger, on average, than the magnitude of floods at El Paso. Ainsworth and Brown (1933) summarized the ef¬ fect of the recently-constructed Elephant Butte Dam on the downstream water and sediment flux: “Elephant Butte Dam and Reservoir have retained the entire flow of the Rio Grande entering the reservoir during the period of operation, 1916 to date [1932]. Release of water has been entirely under control and predicated on irrigation demand [and] exceeds 2,000 second-feet for only short intervals. Practically all the silt (20,000 acre-feet annu¬ ally) entering the reservoir from upper river sources has been retained above the dam.” These changes are illustrated by the median hydrograph for the period 1924 to 1940 for the reach between El Paso and Presidio (Figure 5). The well-de¬ lineated spring snowmelt peak was eliminated, and mod¬ erate flows at El Paso extended between April and Sep¬ tember. These stable flows facilitated efficient agricul¬ tural water withdrawal in the El Paso/Juarez valley, as is evident in the difference between stream flow measured at El Paso and at Fort Quitman. These changes caused the magnitude of annual floods to be reduced by about 65 to 75% for the flows in the El Paso/Juarez valley (Table 3). In contrast, the magnitude of flood peaks downstream from the Rio Conchos only decreased by between 33 and 49%, because the magnitude of flood control provided by reservoirs in the Conchos basin is not nearly as great as in the northern branch. D 03 O X > 33 D m 2 o c ro o m —I m 33 03 33 m 33 03 m o o Figure 4. Graph showing the time series of annual maximum mean daily discharge at four gauging stations along the northern branch. Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 35 o o LU CO DC LLI CL c n DC LU LJJ O CD Z> o LU (3 DC < X o CO Q Figure 5. Graph showing median hydrographs of mean daily discharge for the Rio Grande for the period 1924-1940 at El Paso, at Fort Quitman, and above Rio Conchos, near Presidio. Cochiti Dam affects the annual floods and the sedi¬ ment input to the reach directly downstream. Cochiti Dam operations reduce those few floods exceeding 142 m3/s, resulting in a 38% decrease in annual floods from the pre-dam (1895 to 1973) to post-dam (1974 to 1995) period at the Cochiti gage, just downstream from the dam. Further downstream at the Albuquerque gage, the impact is diminished and the annual flood was only re¬ duced by 4% following completion of the dam. The duration of peak flows increased 60 to 130% between the same time periods (Richard, 2001). Completion of Cochiti Dam resulted in a 99% reduction in sediment concentration flowing into the channel downstream. Upstream from the dam, at the Otowi Bridge, the sus¬ pended-sediment concentration also declined around this time, and thus some aspect of reduced sediment con¬ centrations may be due to regional climate and land use change. Suspended-sediment transport increases down¬ stream from the dam due to re-supply of fine sediment from tributaries and/or erosion of the bed and banks. As a result, the post-dam reduction in annual mean sus¬ pended sediment concentration at the Albuquerque gage is 78% (Richard, 2001). Resulting Channel Changes Today, the northern branch between Cochiti Dam and Presidio can be divided into two long segments - one segment affected by the existence and operations of Cochiti Dam and the other affected by the existence and operations of Elephant Butte and Caballo dams. In the two segments, the reach nearest the dam has experi¬ enced bed degradation and coarsening of bed material. Further downstream in each segment, the channel has aggraded in reaches where the combined influx of sedi¬ ment from tributaries exceeds the diminished transport capacity of the river. The degrading reach downstream from Cochiti Dam probably extends to San Acacia, al¬ though smaller diversion dams at Angostura, Isleta, and San Acacia complicate this longitudinal pattern. The ag¬ grading reach extends from there to the head of Elephant Butte Reservoir. The degrading reach downstream from Caballo Dam once extended to the Mesilla Valley, but channelization has obliterated this evidence. The chan¬ nel has significantly aggraded downstream from El Paso, but a natural channel only exists downstream from Fort 36 Special Publications, Museum of Texas Tech University Quitman. Within each segment, reaches have received different cultural treatments, in terms of direct manipu¬ lation to the channel and floodplain (Table 5). The description of channel change between Cochiti Dam and Amistad Reservoir is based on an unusually comprehensive set of geomorphic data. The combina¬ tion of severe flooding and sedimentation between Cochiti and Elephante Butte, along with irrigation needs in the middle Rio Grande valley in the early 1900s prompted state and federal agencies to begin intensive surveys of the river. These surveys include cross-section surveys beginning in 1918, bed material sampling beginning in the 1930s, suspended sediment sampling beginning in the 1940s, and aerial photography (Leon et al., 1999). Changes along the international boundary are monitored by the International Boundary and Water Commission. Table 5. Summary of cultural impacts to the Rio Grande between Cochiti Dam and Amistad Reservoir. Reach Flood Regime Cultural Treatment From To Pre-1915 Post- 191 5 Channelization Regulation Depletion Cochiti Elephant Butte spring spring moderate?? moderate minor Caballo El Paso spring summer moderate extreme moderate El Paso Fort Quitman spring summer extreme extreme extreme Fort Quitman Candelaria spring summer minor extreme extreme Candelaria Presidio spring summer moderate? extreme extreme Presidio Amistad Reservoir summer summer none moderate moderate Channel Changes Downstream from Elephant Butte and Caballo Dams Completion of Elephant Butte Dam caused the chan¬ nel to degrade immediately downstream from the dam. Further downstream, the channel began to shrink in size in the El Paso/Juarez valley, because the low-gradient channel could not transport this delivered load, nor the load sluiced to the channel from irrigation channels or delivered naturally from ephemeral tributaries. Ainsworth and Brown (1933) reported that: “Silt carried in suspen¬ sion past El Paso now varies from 0.03 percent to 1.5 percent by volume, depending upon the ratio of arroyo runoff to reservoir releases. But by far the greater part of the material transported is sand traveling probably as bottom load. This is either scoured from the riverbed or from the arroyo fans which are annually replenished by run-off from the summer rains. The controlled flow in the river is successively depleted for irrigation use at the various diversion points along its course, at each of which, through operation of skimming weirs and sand sluiceways, a great part of the sand is returned to the riverbed. The ordinary flow of the river by El Paso is not capable of transporting the load of sand and silt an¬ nually brought down the river from above.” “Peak flows at El Paso since Elephant Butte Dam, as a result of the above factors, are of annual occur¬ rence, and while usually under 4,000 second-feet, have amounted to 13,500 second-feet. The peak of these floods is sharp, lasting but a few hours so that the total acre-footage passed is low. The short duration of these summer floods precludes, as to the valley below El Paso, any lasting scouring action or long distance transporta¬ tion of the accumulated deposits. Their action is more to carry the sand scoured from the bed over banks onto the flood plain, which is thus being constantly elevated. A general lowering of the river bed above El Paso has taken place [while] a general filling of the riverbed below El Paso has taken place. Narrowing of the normal chan¬ nel has progressively occurred both above and below El Paso. This effect is most marked below El Paso where the normal channel has only about one third its fomier width.” “River gradients have been but little disturbed ex¬ cept where cut-offs have been made and in the immedi¬ ate reaches above diversion dams or above plugs (of sediment) deposited by side flow and except for the reach of river immediately below the International Dam where filling has resulted in an increase in gradient from 2.45 feet per mile in 1917 to 3.00 feet per mile in 1932. How¬ ever, decreasing gradients due to increasing river lengths are apparent below El Paso where the natural length has been undisturbed by cut-offs.” Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 37 “River length above El Paso has apparently been slightly shortened by natural processes and reduced about five miles by artificial cut-offs. River length below El Paso has been lengthened by nearly 20 percent (com¬ pared to 1907) and by about 4 percent (compared to 1917) in those reaches where neither cut-offs have been made or avulsions occurred. River length above and below El Paso, when compared to valley axial length, has a ratio of 1.21:1 and 1.91:1, respectively.” “The processes of the adjustment of the bed of the river to the new conditions of flow are not complete ....” Channel shrinking on the Rio Grande provided the first clear evidence that large main stem dams in the western United States do not necessarily provide down¬ stream flood control, and under certain conditions may actually increase flood risk due to diminished channel capacity. It spawned a short-lived discussion in the en¬ gineering literature of the 1920s and 1930s regarding the long-term effect of structural methods of flood control (Lawson, 1925; Stevens, 1938). The problem, from the point of view of water supply and flood control, was summarized in the Joint Report of the Consulting Engi¬ neers, International Boundary Commission, on rectifica¬ tion of the Rio Grande (IBWC, 1933): “Notwithstand¬ ing the fact that the present total amount of sediment annually carried through this valley by the Rio Grande is only a very small percentage of that carried previous to the construction of the Elephant Butte Dam, the absence of the former large scouring floods has resulted in the silting up of the river channel to a point where rainfall discharges from arroyos entering the river between El¬ ephant Butte and El Paso-Juarez menace the improved and developed properties of both cities and valley lands. Only large floods of destructive proportions are capable of eroding accumulations of sediment as they now oc¬ cur in the meandering channel.” This report and the subsequent analysis of Ainsworth and Brown (1933) provided justification for straightening and channelizing the river from Elephant Butte to Fort Quitman. The channelization was begun about 1933 and essentially completed in 1938. This reach of river is now artificially maintained as a water delivery and drainage canal. Studying the remaining unchannelized reach be¬ tween Fort Quitman and Presidio, Everitt (1993) con¬ cluded that the physical changes in the channel repre¬ sented a complex chain of responses driven by deposi¬ tion of excess sediment which the depleted river was no longer able to transport. He proposed a three-stage model for channel evolution. This model provides a concep¬ tual basis for evaluating the relationship among the inter¬ dependent variables of declining stream flow, decreased flood magnitude and duration, floodplain aggradation, and channel capacity. Everitt (1993) termed changes in channel width, channel depth, and channel cross-section area as “first- order responses” which began immediately after the flow regime changed, as excess sediment was deposited within the abandoned, oversized channel. Once channel ca¬ pacity was reduced, over-bank flow resumed. These “second-order responses” included meander cutting, changes in the relationship between the main channel and its tributaries, and readjustment of channel gradient. Deposition of sediment within the pre-dam chan¬ nel of the Rio Grande occurred between 1915 and 1925 in the upstream end of the El Paso/Juarez valley (Ainsworth and Brown, 1933). The channel shrank in cross-sectional area, and overbank flooding did not oc¬ cur during this time. Downstream from Fort Quitman, a similar pattern of infilling without overbank flooding occurred between 1915 and 1932. Photographs of the river taken during the U.S. Geological Survey hydro- graphic survey of 1901 depict a broad, shallow, sand- bedded channel downstream from Fort Quitman. Maps of the pre-dam river show a channel about 100 m wide. Aerial photographs taken in 1928 show that the channel had narrowed to about 30 m. Today, some of this old channel survives as oxbows that are lined with very old cottonwoods. Beginning in 1925 near El Paso/Juarez and begin¬ ning in 1932 downstream from Fort Quitman, the channel’s flood capacity had sufficiently decreased such that the lower magnitude floods of this period again be¬ gan to overtop its banks, depositing fine sediment across the valley floor (Figure 7). Floodplain inundation is a necessary process to cause meander cutoffs, and cut¬ offs became a renewed geomorphic process that had not occurred since 1915. Cutoffs occurred in each year in which flood discharges were large in relation to the shrunken channel. The process of channel shrinkage was reversed in 1941 and 1942 when there were unusually large releases 38 Special Publications, Museum of Texas Tech University from Elephant Butte reservoir. Losses were not great during these floods and the peak discharge at Presidio, upstream from the Rio Conchos, was 145 m3/s, which would have been about a 3 -year recurrence flood prior to 1915. The high flows reestablished a larger channel cross-section that was narrower and deeper than the pre-dam channel (Figure 7). These changes occurred by erosion of the channel bed and by deposition of new floodplain sediments. Channel infilling resumed after 1943 and contin¬ ued downstream from Fort Quitman until 1963 (Figure 7). After 1963, sufficient aggradation had occurred that floodplain inundation and meander cutoffs again began to occur. In 1970, the Rio Grande channel was between 10 and 15 m wide (Everitt, 1993). The channel of the Rio Grande between Fort Quitman and Presidio is now about 90% smaller than the channel that existed in 1900. Thus, the Rio Grande channel decreased in size by aggradation of the stream bed and deposition of bars inset within the former active channel and by floodplain deposition. Everitt (1993) concluded that the post- 1970 channel of the Rio Grande was approaching a balance between discharge and channel capacity. Thus, with the resumption of over bank flow the valley floor had resumed its function of storage of floodwater and sedi¬ ment. Thus, deposition of sediment in the Rio Grande valley since about 1970 has occurred by concurrent depo¬ sition in the channel and on the floodplain such that the relationship between the two geomorphic features re¬ mains the same, and the geomorphic functionality of the river is relatively unchanged. The contrast between the channelized reach be¬ tween El Paso and Fort Quitman, and the natural reach from Fort Quitman and Candelaria, illustrates the conse¬ quences of different cultural treatments. Both reaches were initially similar in physical geometry and hydrol¬ ogy, and both experienced similar changes in flow re¬ gime following construction of Elephant Butte Dam. The channelized reach resembles a drainage ditch, dewatered much of the year, separated from its flood plain by steep banks, and with floodplain vegetation artificially main- United States Estados Unidos Mexico Distance in Meters Distancia en Metros Figure 6. Graph showing composite cross sections of the Rio Grande channel at the gauging station above Rio Conchos, near Presidio, 1933-1974 (from Everitt, 1993, fig. 3). Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 39 tained by mowing. Beyond the levees, what remains of the cottonwood gallery forest is cut off from the river, its seeds falling on barren ground. Former tributaries have been cut off by erosion-control dams, isolating the river from its watershed. The riverine landscape has lost the physical continuity that once provided migration routes for riparian plants and animals, and the dynamic nature that once provided cycling and storage of water and nutrients. Downstream from Fort Quitman, although severely depleted in stream flow, the river continues to be a func¬ tioning part of the landscape. Channel dimensions in some reaches have adjusted to the altered discharge so that a smaller river flows in a smaller channel in a rela¬ tively broader flood plain. As in pre-dam times, the river continues to meander in some places forming oxbows and in others braiding and forming islands, maintaining the topographic irregularities that provide habitat diver¬ sity. The mosaic of landscape elements necessary for the foundation of a healthy riparian ecosystem is still present. Brushy banks with fallen trunks provide shaded scour holes for fish. Natural levees pond flood water beyond the channel, allowing it to percolate slowly back to the river, maintaining the shallow alluvial ground-wa¬ ter system and prolonging base flow. Broad overflow lands spread and filter water during high stages and flush accumulated salt from the soil. Here the river landscape retains both its longitudinal and lateral continuity, although there have been profound changes in the vegetation. Cochiti to Elephant Butte Changes in the Rio Grande between Cochiti and Bernalillo are similar to the pattern of bed degradation and narrowing that occurred soon after completion of Elephant Butte Dam. Richard’s (2001) study of adjust¬ ments of the Rio Grande between Cochiti and Bernalillo demonstrated that continued lateral adjustments, includ- RIO GRANDE/RIO BRAVO ABOVE RIO CONCHOS NEAR PRESIDIO, TEXAS RIO GRANDE/RIO BRAVO ARRIBA DEL RIO CONCHOS, CERCA DE PRESIDIO, TEXAS 100 50 0 (t> £0 Q Q. 3" CD £D — 3 P 3 tt> 2. CD (— > O v 2. o O' CD 3 CD H CO a ® o r--4' 5' 3 03 3 C /> < CD CD o. > Q_ CD CD^ 0) O 5' £ 3 o CD CD 3 ro 3 ro Figure 7. Graph showing time series of change in annual discharge and cross section at the gauging station above Rio Conchos, near Presidio, 1933-1974 (from Everitt, 1993, fig. 2). 40 Special Publications, Museum of Texas Tech University ing narrowing and decreased lateral migration rates, ments occurred following construction of the dam in 1973. primarily sand. The sandy bed of the channel responded to the temporal variability in sediment inputs by alternat¬ ing aggradation and degradation; the net sum was gradual aggradation of the bed. Bed degradation began after Cochiti Dam was completed and up to 1 .9 m of bed erosion occurred between 1972 and 1998. Following dam construction, the bed material between Cochiti and Bernalillo coarsened to gravel and cobbles (Richard, 2001). Measurements from historic maps and aerial pho¬ tographs indicate that as the peak discharges began to decrease (ca. 1930s) due to natural and anthropogenic factors the channel responded by narrowing, simplify¬ ing and reducing its rate of lateral migration (Figure 8). The width of the Rio Grande between Cochiti and Bernalillo decreased 60% prior to dam construction and by 1992 the channel was 70% narrower than in 1918. Also between 1918 and 1 992 the channel planform shifted from a multi-thread braided configuration toward a single- occurred between 1918 and 1992. More rapid vertical adjust- Prior to dam construction, the bed of the channel was thread pattern as the number and size of mid-channel bars and islands decreased (Figure 9). Following dam construction, a meandering pattern became more pro¬ nounced as the sinuosity of the channel downstream from Cochiti increased slightly (Richard, 2001). Richard (2001) concluded that attempts to “stabi¬ lize” the Cochiti reach of the Rio Grande through flood control, sediment detention, channelization, and bank stabilization succeeded in reducing the dynamism in both the inputs to the reach and in the responding form of the channel. Average lateral movements of the channel de¬ creased from 27 m/year in 1918 to 5 m/year in 1992, and the active channel width has remained less than 100 m since 1985. Incision of the channel bed following construction of the dam disconnected the channel from the floodplain. The resulting narrow and deep configu¬ ration of the channel and reduced peak flows creates a situation in which even the highest flows no longer achieve bankfull conditions (Richard, 2001). Discussion Implications of Hydrologic and Geomorphic History for River Restoration or Rehabilitation We have shown that the physical attributes of the Rio Grande — its hydrology, sediment load, channel di¬ mensions, and temporal variability of channel location - are much different than they were a century ago. Some segments are still evolving in response to past alteration in flood regime, depletion of flow, and deposition of sedi¬ ment in reservoirs. The aquatic and riparian ecosystems are also much different than they were a century ago. Ecosystem change is driven by the following variables: 1) Change in climate that affects the runoff and sediment flux. 2) Change in hydrologic and sediment regime caused by human activities. 3) Changes in the physical structure of the channel and floodplain. These changes are caused by changes in the flux of water and sediment and by direct manipula¬ tion of the channel or floodplain. The fundamental at¬ tributes of physical structure are the cross-sectional form of the channel, the characteristics of the bed material and how it is organized, channel planform, channel gra¬ dient, and the relationship between the channel and its alluvial valley. These changes not only affect the distri¬ bution of aquatic habitats and the exchange rate of sedi¬ ment between channel and alluvial valley but also the characteristics of nutrient spiraling. 4) Introduction of exotic species. In addition to many naturalizing herbs and grasses, there are 3 woody exotics that are expanding their range at the expense of native vegetation along the Rio Grande: Russian olive ( Eleagnus angustifolia L.) in the upstream part of the northern branch, salt cedar in the southern part of the northern branch and the upstream part of the lower Rio Grande, and giant reed ( Arundo donax) downstream from Presidio. 5) The internally-driven dynamics of ecosystems that cause some species to replace others over time. Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 41 Cochiti Gage - Peak Flows ■ 1 Oto\vi/ Cochiti Susp Sed Cone » - “Albuquerque Susp Sed Cone Figure 8. Summary of changes in the Rio Grande between Cochiti Dam and Bernalillo from 1918 to 1992. Of these input variables, climate change is beyond the control of Rio Grande managers. Eradication of non¬ native species invasions and vegetation manipulation that alters the trajectory of ecosystem change are extremely difficult, although large-scale eradication of salt cedar has been conducted in parts of the Pecos River alluvial valley. It is only by manipulating the runoff regime, sedi¬ ment regime, or physically altering the channel or flood¬ plain that we can alter the balance among competing species in a functioning ecosystem. We know we can do this, because we have already performed experiments on the Rio Grande. A hundred years of data on the 6 reaches of Table 5 provide case studies of how local riparian communities respond to different kinds of treat¬ ment under different local circumstances. Elsewhere, stream flow and sediment fluxes are being altered by dam reoperations in order to alter down stream ecosys¬ tems. 42 Special Publications, Museum of Texas Tech University Pre-dam Post-dam Figure 9. Planform maps of the active channel of the Cochiti reach for 191 8 through 1992. Other Goals for Environmental Management Environmental management of the Rio Grande must be grounded in establishment of a set of well-defined goals for the future trajectory of ecosystem change on each segment. Ecosystem restoration is defined as “the return of an ecosystem to a close approximation of its condition prior to disturbance” (National Research Coun¬ cil, 1992). There also are other possible goals for aquatic ecosystems, including reclamation, rehabilitation, miti¬ gation, and creation (National Research Council, 1992). Reclamation is the process of adapting a wild or natural resource to serve a utilitarian human purpose. Thus, this term is reserved for activities such as converting native floodplain ecosystems to agricultural uses. Reha¬ bilitation is a term used primarily to indicate putting a natural resource back into good condition or working order. Mitigation is typically defined as alleviating any or all detrimental effects arising from a specific human activity. Creation is the bringing into being of a new ecosystem that previously did not exist at the site. Envi¬ ronmental management goals might include any of those listed above. Choice of goals, on a segment by segment basis, is an effort in policy development, will inevitably be based on dialogue among river stakeholders, and is necessarily political. Disturbances to the hydrologic regime of the Rio Grande began hundreds of years ago and are significant. Restoration of the Rio Grande’s northern branch to its condition in 1900 would require dam decommissioning and the abandonment of most irrigated agriculture in southern Colorado, New Mexico, the El Paso/Juarez valley, and the Presidio valley. Restoration would also require removal of levees, rehabilitation of channelized sections, and relocation of large numbers of people from Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 43 the historic floodplain of the rivers, especially in El Paso and Cuidad Juarez. Political consensus to undertake such a comprehensive program of river restoration probably does not exist in either the United States or Mexico. Thus, it is essentially impossible to restore most of the river. Goals for the Rio Grande might include (1) reha¬ bilitation to some post-1915 condition, although the chan¬ nel was not in equilibrium with its floodplain for most of this time, (2) rehabilitation so that the channel and allu¬ vial valley have a broader suite of ecological processes and attributes similar to the pre-disturbance river, (3) mitigation by maintenance of a new ecosystem, with or without salt cedar, that is adjusted to a specified range of flood flows and annual flows, (4) mitigation to the level of ecosystem function necessary to recover endangered species, or (5) acceptance of the riverine ecosystem as it is today. The identification of the appropriate goal depends on a precise identification of the natural and human values that would be improved and degraded if the present ecosystem were changed. There is probably no single environmental man¬ agement goal that is appropriate for all of the Rio Grande. Each goal described above is associated with its own economic cost, and achievement of political consensus on any environmental management goal is difficult to achieve. Knowledge of the magnitude of twentieth cen¬ tury environmental change does not necessarily mean the trajectory towards restoration will follow the same path and the trajectory of historical change. Where chan¬ nels have significantly narrowed, become disconnected from their floodplains, and overgrown with salt cedar, the question remains whether reintroducing more natu¬ ral water and sediment fluxes will immediately reverse undesired historical changes. Restoration science is not yet able to predict these trajectories of system recovery. In the face of such uncertainty, pursuing uniform basin-wide rehabilitation goals is essentially impossible. Is it better to ask where in the basin can undesired his¬ torical changes be efficiently reversed? Where in the basin will the native riverine ecosystem respond most favorably to reversal of historical changes in the physi¬ cal environment? Where can the historical changes of water and sediment flux be feasibly reversed and at what political cost? Where is the greatest need for rehabilita¬ tion? Answers to each of these questions can only be provided by considering what is feasible and possible in each segment of the river. Only then can one examine how much water is available for redistribution to envi¬ ronmental objectives and develop an allocation system that recognizes the needs of the natural riverine ecosys¬ tem and the physical template within which it has devel¬ oped. Conclusion The extent of changes to the water and sediment flux are so great, and the extent of changes to the physi¬ cal system of the channel and floodplain are so exten¬ sive, that comprehensive restoration of the Rio Grande is impossible. Priorities must be established wherein dif¬ ferent environmental management goals are established for different segments of the river system. Establish¬ ment of these priorities is inevitably a political process, wherein the role of the scientific community is to present a clear picture of the magnitude of transformation of the present riverine ecosystem from its pre-disturbance con¬ dition and the activities necessary to rehabilitate the eco¬ system to varying degrees or to reverse undesired changes in the physical or ecological system. Acknowledgements Critical review of a draft manuscript was provided Data from the National Inventory of Dams was graciously by Kevin M. Urbanczyk and an anonymous reviewer. provided by W. L. Graf. 44 Special Publications, Museum of Texas Tech University Literature Cited Ainsworth, C. M. and F. P. Brown. 1933. Report on the changes in regimen of the Rio Grande in the valleys below since the construction of Elephant Butte Dam, 1917-1932. Unpub¬ lished report of the United States Section, International Boundary and Water Commission, El Paso, Texas. Aulbach, L. F. and J. Butler. 1998. The lower canyons of the Rio Grande. Wilderness Area Map Service, Houston, Texas, 94 pp. Aulbach, L. F. and L. C. Gorski. 2000. The upper canyons of the Rio Grande. Wilderness Area Map Service, Houston, Texas, 93 pp. Bullard, K. L. and W. L. Lane. 1993. Middle Rio Grande peak flow frequency study. U.S. Bureau of Reclamation, Albu¬ querque, New Mexico. Dewey, J. D., F. E. Roybal, and D. E. Funderburg. 1979. Hydro- logic data on channel adjustments 1970 to 1975, on the Rio Grande downstream from Cochiti Dam, New Mexico be¬ fore and after closure. U. S. Geological Survey Water Re¬ sources Investigations 79-70, Albuquerque, New Mexico. Dortignac, E. J. 1956. Watershed resources and problems of the upper Rio Grande basin. USDA Rocky Mountain Forest and Range Experiment Station report, Fort Collins, Colo¬ rado, 107 pp. Emory, W. H. 1857. Report on the United States and Mexican boundary survey, made under the direction of the Secretary of the Interior. 34th Congress, 1st Session, Senate Exec. Doc. No 106. Everitt, B. 1993. Channel response to declining flow on the Rio Grande between Ft. Quitman and Presidio, Texas. Geomor¬ phology, 6:225-242. Graf, W. L. 1 994. Plutonium and the Rio Grande - environmental change and contamination in the nuclear age. Oxford Uni¬ versity Press, New York, New York. Grams, P.E. and Schmidt, J.C. 2002. Streamflow regulation and multi-level floodplain formation: channel narrowing on the aggrading Green River in the eastern Uinta Mountains, Colo¬ rado and Utah. Geomorphology, 44:337-360. IBWC. 1933. Joint report of consulting engineers, in Rectification of the Rio Grande: Convention between the United States of America and the United States of Mexico, U.S. Depart¬ ment of State, Mexico City. Kelley, P. 1986. River of lost dreams: navigation on the Rio Grande. University of Nebraska Press, Lincoln. Lawson, L. M. 1925. Effects of Rio Grande storage on river erosion and deposition. Engineering News-Record, 95 (10):372-374. Leon, C., G Richard, T. Bauer, and P. Julien. 1999. Middle Rio Grande, Cochiti to Bernalillo bridge, hydraulic geometry, discharge and sediment data base and report. Colorado State University, Fort Collins. Meade, R. H., T. R. Yuzyk, and T. J. Day. 1990. Movement and storage of sediment in rivers of the United States and Canada. Pp. 255-280, in Surface Water Hydrology (M. G Wolman and H. C. Riggs, eds.), The Geology of North America, vol. O-l, The Geological Society of America, Boulder, Colo¬ rado. Moles, M. C., Jr., C. S. Crawford, L. M. Ellis, H. M. Valett, and C. N. Dahm. 1998. Managed flooding for riparian ecosystem restoration. Bioscience 48(9):. Mueller, J. E. 1975. Restless River. Texas Western Press, El Paso, 155 pp. National Research Council. 1992. Restoration of aquatic ecosys¬ tems. National Academy Press, Washington, D.C., 552 pp. Richard, G A. 2001. Quantification and prediction of lateral channel adjustment downstream from Cochiti Dam, Rio Grande, NM. PhD dissertation, Colorado State Univer¬ sity, Fort Collins. Rittenhouse, G 1944. Sources of modem sands in the middle Rio Grande valley, New Mexico. Journal of Geology, 52:1 45- 183. Sanchez, V. and D. Baird. 1997. River channel changes down¬ stream from Cochiti Dam, middle Rio Grande, New Mexico. In Proceedings of the Conference on Management of Land¬ scapes Disturbed by Channel Incision. University of Mis¬ sissippi, Oxford. Schumm, S. A. 1969. River Metamorphosis. Journal of the Hydraulics Division, 95:255-273. Scurlock, D. 1998. From the Rio to the Sierra: an environmental history of the middle Rio Grande basin. USDA Forest Service General Technical Report RMRS-GTR-5, 440 pp. Sorenson, E. F. and D. Linford. 1967. Rio Grande basin - settle¬ ment, development, and water use. Pp. 143-168, in Water Resources of New Mexico: Occurrence, Development, and Use. New Mexico State Engineer, State Planning Office, Santa Fe. Stevens, J.C. 1938. The effect of silt removal and flow regulation on the regimen of the Rio Grande and Colorado Rivers. American Geophysical Union Transactions, 653-656. Stotz, N. G 2000. Historic reconstruction of the ecology of the Rio Grande/Rio Bravo channel and floodplain in the Chihuahuan Desert: Desert Scribes, Las Cruces, report pre¬ pared for Chihuahuan Desert Program, World Wildlife Fund, 158 pp. Thomas, H. E. and others. 1963. Effects of drought in the Rio Grande basin. U.S. Geological Survey Professional Paper 372-D, 59 pp. U. S. Army Corps of Engineers. 1978. Cochiti Lake, Rio Grande basin, New Mexico, water control manual, appendix C to Rio Grande basin master water control manual. Albuquer¬ que, New Mexico. Schmidt et al. — Hydrology and Geomorphology of the Rio Grande 45 U. S. Army Corps of Engineers. 1996. Water control infrastruc¬ ture: national inventory of dams [CD-ROM]. Federal Emer¬ gency Management Agency, Washington, D. C. Addresses of Authors: John C. Schmidt Department of Aquatic, Watershed, and Earth Resources Utah State University Logan, Utah 84322-5210 e-mail: jschmidt@cc.usu.edu Benjamin L. Everitt U.S. Department of State. 1903. Proceedings of the International Boundary Commission, United States and Mexico, equable distribution of the waters of the Rio Grande. Gigi A. Richard Department of Physical and Environmental Sciences Mesa State College Grand Junction, CO 81501 e-mail: grichard@mesastate. edu Utah Division of Water Resources Salt Lake City, Utah e-mail: beneveritt@utah.gov Gammarid Amphipods of Northern Chihuahuan Desert Spring Systems: an Imperiled Fauna Brian K. Lang, Vivianaluxa Gervasio, David J. Berg, Sheldon I. Guttman, Nathan L. Allan, Mark E. Gordon and Gordon Warrick Abstract Gammarid amphipods of the Gammarus pecos complex Cole, 1985 are restricted to euryhaline desert spring systems in the Pecos River Valley of New Mexico and Texas. Within the past 35 years, com¬ plete loss and diminution of spring flows, exacerbated by regional drought conditions and local groundwater withdrawals, are implicated in the extirpation of 2 iso¬ lated populations of G. desperatus in New Mexico, and the dramatic decline of the Gammarus sp. form “C” from the Phantom Lake Spring system in Texas. The distribution and abundance of gammarid amphipods was determined from benthic samples collected in May 2001 from 7 sites in the study area. Gammarid densi¬ ties were highest in shallow, low velocity habitats with aqueous silts. Three new gammarid populations were documented from 2 springs in Reeves County, Texas, and 1 site in Eddy County, New Mexico. This study illustrates the need for comparative morphological and biochemical genetic studies of the G pecos complex to clarify outstanding taxonomic relationships within this group. Introduction Members of the Gammarus pecos complex Cole, 1985 represent endemic species geographically iso¬ lated in euryhaline desert spring systems of the Pecos River Valley of New Mexico and Texas. These fresh¬ water amphipods are likely derived from a widespread progenitor marine amphipod that was isolated inland during recession of the Late Cretaceous epieric sea (ca. 66 mya) (Bousfield, 1958; Holsinger, 1976). Spe- ciation within this complex likely occurred as a result of local adaptive variation in response to ecological constraints imposed by diverse aquatic environments on amphipod populations further isolated during pro¬ gressive climatological changes that ensued in the late Pleistocene to early Holocene. Such models of island and vicariant biogeography are proposed for a diver¬ sity of invertebrate taxa within arid ecosystems of southwestern North America (e.g., Peracarida crusta¬ ceans [freshwater isopods, see Bowman, 1981; hyalellid amphipods, see Thomas et al., 1994]; proso- branch snails, see Hershler et ah, 1999; pulmonate land snails, see Bequaert and Miller, 1973, Metcalf and Smartt, 1997). Based on percent similarities in Mann- Whitney U tests for 20 morphological traits from 7 populations of gammarid amphipods, Cole (1985) identified the G. pecos complex as a group of morphologically similar species that are endemic to Chihuahuan Desert spring systems of the Pecos River Valley in southeastern New Mexico and western Texas. Morphological character combinations unique to this group include: non- calceolate antennae with spine(s) on the first pedun¬ cular article of the antennule; mandibular palps bear¬ ing C-setae; setiferous coxal plates I-IV; and narrow oostegites (brood plates). Currently this complex is understood to consist of 3 described species (Cole and Bousfield, 1970; Cole, 1976, 1981), 2 populations representing undescribed species, and 2 morphotypes of undetermined taxonomic affinity (see Cole, 1985). Cole (1985) emphasized numerous unresolved taxonomic relationships within this complex based on marked intra- and inter-population variation of mor¬ phological characters and body sizes among West Texas gammarids. 48 Special Publications, Museum of Texas Tech University Formulation of effective conservation measures or intensive ecological assessments would be prema¬ ture until outstanding taxonomic affinities of the G. pecos complex have been clarified. Morphological assessment of the G. pecos complex, considered in combination with ongoing genetic studies, can fill a critical gap in our current knowledge of these endemic taxa: “Which species or significant population segments merit conservation management?” Resource agencies charged with protection and stewardship of these amphipods have no baseline data from which to as¬ sess conservation status and threats, or to prescribe management options for this group. Amphipods are vital links in aquatic food webs and play critical roles in nutrient processing of aquatic ecosystems (Gee, 1988; Pennak, 1989). Due to their acute sensitivity to aquatic conditions (Covich and Thorpe, 1991), gammarid amphipods can be consid¬ ered ecological indicators of ecosystem health (Lackey, 1995) and integrity (Callicott, 1994). The current rate of impediment of amphipods of the G. pecos complex is alarming and provides testimony to the deterioration of aquatic conditions of the Pecos River Basin of New Mexico and Texas. No gammarids were observed during preliminary status surveys of Phantom Lake Spring and its canal system in March and June 2000, which suggested that Gammarus hyalelloides Cole, 1976 and the undescribed Gammarus sp. form “C” were possibly extirpated from this aquatic system. Although the location of Cole’s (1985) Gammarus sp. form “M” population could not be reconciled from locality descriptors while in the field, the ample evidence of spring head drying and defunct irrigation canal systems observed in the gen¬ eral vicinity of Phantom Lake Spring suggested that this gammarid morphotype may be extirpated as well. These preliminary findings prompted this study. Specific causes for the suspected extirpation of Gammarus sp. form “C” from the aquatic environs of Phantom Lake Spring appeared directly related to re¬ duced spring discharge. Diminution of Phantom Lake Spring discharge has been attributed to depletion of the Toyah Basin Aquifer from groundwater withdraw¬ als under ongoing regional drought conditions (Ashworth et ah, 1997; Schuster, 1997; Sharp et ah, 1999; Allan, 2000). Similar factors were largely re¬ sponsible for the extirpation of 2 isolated populations of Gammarus desperatus Cole, 1981 in New Mexico (Cole, 1981, 1985). We present preliminary findings from a macroinvertebrate survey to document the current sta¬ tus of gammarid amphipods from desert spring sys¬ tems in the Pecos River Valley of New Mexico and Texas. This study illustrates the need for compara¬ tive morphological and biochemical genetic studies of the G. pecos complex. Study Area and Methods The distribution and abundance of gammarid amphipods was determined from benthic samples col¬ lected in May 2001 from 7 sites in the study area of southeastern New Mexico and West Texas (Figure 1). Amphipod densities were determined from 3 random benthic collections per site using a stainless steel mesh sampler (sample area = 80 cm2). Benthos were pre¬ served in 95% ethanol for enumeration by taxa in the lab. Habitat parameters measured at each sample site included water depth (± 1 cm; metric topset rod), ve¬ locity (cm/sec; Marsh-McBimey Flowmate®), and substrata following the general scheme of Cummins’ (1962) index of dominant grain size as clay, silt, sand, gravel, and mixed. Vegetation, including submerged macrophyte beds and detrital plant debris, was included as a “substrate” category since amphipods of this com¬ plex typically frequent this cover type. Physicochemical data (i.e., water temperature [°C], pH, dissolved oxy¬ gen [DO; mg/1, % saturation], salinity [ppt], specific conductance [pS/cm] and total dissolved solids [TDS; ppt]) presented herein were compiled from published accounts and agency reports due to equipment failure in the field. Live Gammarus for ongoing morphological and biochemical genetic studies were collected with a dip net by sweeping submerged macrophyte beds, by siev¬ ing aqueous silt, or from the underside of coarse sub¬ strata by washing into a tray. To quantify genetic and morphological variation within a gammarid population, Lang et al. — Gammarid Amphipods of Northern Chihuahuan Desert Springs 49 Figure 1. Historic and current distribution of gammarid amphipods of the Gammarus pecos complex Cole, 1985 in the Pecos River Valley of New Mexico and Texas. Population designations from modified map per Cole (1985:): “D ”, G. desperatus, Roswell, Chaves Co., NM (extant at Bitter Lake Na¬ tional Wildlife Refuge, extirpated from Lander Springbrook and North Spring); “E”, Gammarus sp., “in Carlsbad” or “in Carlsbad Caverns National Park”, Eddy Co., NM; “P”, Gammarus pecos, Diamond Y Spring, Pecos Co., TX; “S”, Gammarus sp., San Solomon Spring, Toyahvale, Reeves Co., TX; “H”, Gammarus hyalelloides, Phantom Lake Spring, Jeff Davis Co., TX; “C”, Gammarus sp., Phantom Lake Spring canal system, Jeff Davis Co., TX; “M”, Gammarus sp., “ 3.5 miles west of Toyahvale” or “350 m north of Phantom Lake Spring", Jeff Davis Co., TX. Numeric population designa¬ tions: 1, Gammarus sp., Giffin Spring, Toyahvale, Reeves Co., TX; 2, Gammarus sp., East Sandia Spring, Reeves Co., TX(0.5 mi. eastofBrogado, TX, and 0.3 mi. south of state route 17); 3, Gammarus sp., Sitting Bull Spring, Eddy Co., NM. 50 Special Publications, Museum of Texas Tech University representative voucher material was collected from diverse habitat types employing subsampling in large spring systems (e.g., Diamond Y Preserve, San Solomon Spring) or from populations with geographi¬ cally isolated populations (i.e., Bitter Lake National Wildlife Refuge). This sampling protocol yielded ad¬ equate voucher material for intra- and interspecific genetic and morphological studies from multiple subsamples for 7 gammarid populations. At least 50 amphipods from each collecting locality were frozen in liquid nitrogen immediately upon collection for ge¬ netic study. Specimens retained for morphological analysis (N = 50 amphipods/locality) were preserved in 95% ethanol. Results Our survey documented 3 previously undetec¬ ted gammarid populations from Giffin and East Sandia springs, Texas, and Sitting Bull Spring, New Mexico, that morphologically are referable to the G. pecos com¬ plex (Cole, 1981, 1985). It remains unclear to us whether the Sitting Bull Spring gammarid population is referable to Cole’s (1985) Gammarus sp. form “E”, or may actually represent a previously undocumented population, as Cole listed two possible localities for the latter taxon: “A third unnamed population occurs [or once occurred] in Carlsbad Cavern National Park, or the town of Carlsbad...” Notwithstanding, prelimi¬ nary analysis indicates that the Sitting Bull Spring popu¬ lation appears morphologically distinct from the near¬ est gammarid populations in New Mexico (G. desperatus) and Texas (G pecos). Morphological stud¬ ies are ongoing to describe the Sitting Bull Spring gammarid, and to compare within- and among-popu- lation variation of morphological characters in this spe¬ cies complex. No gammarids were observed in Phantom Lake Spring or in the downstream canal system during the May 200 1 inventory. All lateral canals in the immedi¬ ate area were either dry or dysfunctional; this further confirmed our March and June 2000 observations that Gammarus hyalelloides and the morphotype Gammarus “C” were in all likelihood extirpated from the Phantom Lake Spring system. However, sampling during a visit in November 2001 by Lang and Berg revealed G. hyalelloides from the cave mouth pool, where inten¬ sive surveys in March and June 2000, and May 2001, failed to yield even the slightest evidence of an extant gammarid population. We also found live G. hyalelloides in hypogean habitats at 3-4 meters inside the cave’s entrance during the November 2001 sur¬ vey. Table 1 presents amphipod densities (± SE) and flow conditions (depth and velocity; ± SE) measured at sample sites. Physicochemical parameters com¬ piled from published records and agency reports (Table 2) show highly variable chemical environments with ionic concentrations of these fresh to moderately sa¬ line desert spring systems determined largely by the underlying karst stratigraphy. Amphipod densities were highest in the Diamond Y Draw spring system (Euphrasia Spring, Gammarus pecos , = 8,042 amphipods/m2) where mean water depths and velocities measured at 3 sites ranged from 0.06-0.16 m and 0.03-0.06 m/sec., respectively. The gammarid population in the upper rheocrene of Sitting Bull Spring was the least dense with 125 amphipods/ m2. Low flow conditions in the main canal of San Solomon Spring ( velocity = 0; depth = 0.09 m ± 0.02 SE) likely account for the high mean density of Gammarus sp. form “S” (6,833 amphipods/m2) dur¬ ing seasonal draw down for swimming pool mainte¬ nance at Balmorhea State Park. Mean density of G desperatus (575 amphipods/m2) from Sago Spring is considered lower than expected since this estimate was derived from artificial tile samples (Lang, unpublished data). Hyalellid amphipods are referred to herein as Hyalella sp. since ongoing taxonomic research con¬ tinues to identify genetically and morphologically dis¬ tinct populations once considered as a single ubiqui¬ tous species, Hyalella (Hyalella) azteca, in the south¬ west United States (see references in Baldinger et ah, 2000; Duan et ah, 2000). Hyalella occurred syntopically with Gammarus sp. at East Sandia (1,083 Hyalella sp./m2) and Euphrasia (125 Hyalella sp./m2) springs (Table 1). Densities of Hyalella sp. were high- Lang et al. — Gammarid Amphipods of Northern Chihuahuan Desert Springs 51 Table 1. Mean desnity of hyalellid and gammarid amphipods (no./m2) and mean water depth (m) and mean velocity (m/sec.) measured at benthic sample sites (N = 36 grabs) in desert spring systems of the Pecos River valley of New Mexico and Texas, May 2001. Amphipod Amphipod Water Water Taxon Site Subsample Density (X±SE) Depth (X±SE) Velocity (X±SE) Texas Gammarus hyalelloides Phantom Lake Spring spring head pool 0 0.40 ± 0.04 no measurable flow Gammarus sp. form “C” Phantom Lake Spring lateral canal 0 dry - Gammarus sp. form “S” San Solomon Spring swimming pool 0 0.18 ± 0.04 no measurable flow Gammarus sp.form “S” San Solomon Spring main canal 6833 ± 5416 0.09 ± 0.02 no measurable flow Gammarus sp.form “S” San Solomon Spring cienega outflow - - - Gammarus sp. Giffin Spring spring head pool 0 0.26 ± 0.06 no measurable flow Gammarus sp. Giffm Spring rheocrene 1167 ± 730 0.19 ± 0.09 0.12 ± 0.04 Gammarus sp. East Sandia Spring rheocrene 4625 ± 804 0.10 ± 0.03 0.14 ± 0.08 Hyalella sp. »» 99 99 »> 1083 ± 1083 » 99 Gammarus pecos Diamond Y Draw rheocrene near source 2208 ± 1585 0.16 ± 0.04 0.03 ± 0.02 Gammarus pecos Diamond Y Draw Euphrasia Spring 8042 ± 7229 0.06 ± 0.00 0.06± 0.03 Hyalella sp. 99 99 99 99 99 125 ± 72 »» Gammarus pecos Diamond Y Draw John’s Pool 2708 ± 1381 0.09 ± 0.02 0.04± 0.01 New Mexico Gammarus sp. Sitting Bull Spring lower rheocrene 4542 ± 2361 0.09 ± 0.01 0.09 ± 0.02 Gammarus sp. Sitting Bull Spring upper rheocrene 125 ± 72 0.12 ± 0.03 0.02 ± 0.01 Hyalella sp. 99 99 99 99 99 3167 ± 1601 99 »» Gammarus desperatus* Bitter Creek Lost River pool 4250 ± 683 0.16 ± 0.04 0.12 ± 0.03 Gammarus desperatus * Bitter Creek Sago Spring head 575 0.18 ± 0.04 0.08 ± 0.06 Gammarus desperatus* Bitter Creek Sago Spring run 3750 ± 735 0.07 ± 0.06 0.18 ± 0.05 Gammarus desperatus Bitter Creek Unit 6 167 ± 110 0.11 ± 0.02 0.16 ± 0.04 *Density estimates derived from previous study (Lang, unpubl. data). Table 2. Physicochemical data for desert spring systems in the Permian Basin where gammarid amphipods of the Gammarus pecos complex Cole, 1985 occur in New Mexico and Texas. Site Temperature °C pH units Specific Conductance pS/cm Salinity ppt TDS mg/1 DO % saturation DO mg/1 Lander Springbrook, Chaves Co., NM1,2 18-22 7. 1-7.2 — 4. 4-5. 7 — — — North Spring, Chaves Co., NM2 19.0-20.5 7.2 17,600 5.7 — — — BLNWR, Bitter Creek, Chaves Co., NM 7.6-25.2 6. 9-8. 7 8044-10,538 4. 9-5. 3 5. 5-6. 7 8.0-191.1 0.8-15.9 BLNWR, Sago Spring, Chaves Co., NM 16.0-21.5 7. 0-7. 3 6500-8100 4.3-5. 1 5. 1-5.8 10.0-185.0 2.5-14.8 BLNWR, Unit 6, Chaves Co., NM 22.8 7.3 5367 2.9 3.5 70.2 7.4 Sitting Bull Spring, Eddy Co., NM 14.0-18.5 — — — — — — Diamond Y Draw (Sta. 2), Pecos Co., TX3 20.6 7.1 — 3.6 — — 6.7 East Sandia Spring, Reeves Co., TX3 19.8 7.2 — 2.2 — — 8.2 Giffin Spring, Reeves Co., TX 15.5-17.5 — — — — — — San Solomon Spring, Reeves Co., TX 25.0 7.2 — 1.8 — — 1.9 PLS, Jeff Davis Co.,TX3 24.7 6.9 — 1.9 — — 1.9 PLS, below weir (Sta. 1), Jeff Davis Co., TX3 19.1 6.5 — 4.2 — — 8.8 1 Noel (1954); - Cole (1981); 3 Hubbs (2001) BLNWR = Bitter Lake National Wildlife Refuge; PLS = Phantom Lake Spring 52 Special Publications, Museum of Texas Tech University est (3,167 amphipods/m2) at the upstream sample site in Sitting Bull Spring. There was no evidence of geo¬ graphic overlap between these amphipod genera at the downstream site in this system. We note similar dis¬ tributional patterns of amphipods at Bitter Lake Na¬ tional Wildlife Refuge (BLN WR) where G. desperatus shows affinities for shallow lotic habitats near ther¬ mally stable spring sources, while Hyalella occurs in more lentic eurythermal waters and deep aqueous silts. The abundance of amphipod taxa was vari¬ able across the spectrum of substrate types sampled (N = 36 benthic grabs) throughout the study area (Fig¬ ure 2). Sampling inefficiency likely accounts for the high outlying densities of Gammarus spp. in clay (22,500 amphipods/m2) and gravel (18,280 amphipods/m2) substrata since sample effort in these indurate sub¬ strate types was limited to a single benthic grab. Small sample size (N = 3) may likewise explain the high den¬ sity of Gammarus (3,667 amphipods/m2) found in sand. Non-indurate substrata afforded by loosely compacted, aqueous silts (N = 16 grabs) and vegetation (i.e., mac¬ rophyte beds, detritus; N = 7 grabs) were inhabited by both amphipod genera, but to a lesser degree than mixed substrates (N = 8 grabs) where Hyalella (1,078 am¬ phipods/m2) and Gammarus (2,406 amphipods/m2) were abundant. Gammarid amphipods were particu¬ larly abundant beneath large stones in lotic habitats that were difficult to sample by our quantitative meth¬ ods. Discussion All desert spring systems surveyed in this study occur within an area once overlain by Permian seas (Hills, 1942) where the underlying geology of these karst lands consists of dissolute evaporite rock (White et al., 1995; Martinez et al., 1998). Several basins in the lower Pecos River Valley are described in the study area, namely the Roswell Artesian Basin of New Mexico, the Toyah Basin of West Texas (Hill, 1900), and the Delaware Basin in both states (Lang, 1937). Gammarid populations of this complex occur in Gammmarus spp. (2406) Hy della sp. (1078) x mixed 22% vegetation Gammarus spp. (22,500 *) Gammarus spp. (563) Hya! el la sp . (258) 19% gravd sand s' Gammarus spp. (1696) 3% 8% Hyalella sp. (18) Gammarus spp i.(18,280*) Gammarus spp . (3667) Figure 2. Mean density of Amphipoda (#/m2) by percentage substrate types sampled (n=36 benthic grabs) in desert spring systems surveyed in the Pecos River Valley of New Mexico and Texas, May 2001. (*Denotes sample bias from a single benthic grab.) Lang et al. — Gammarid Amphipods of Northern Chihuahuan Desert Springs 53 sulfatochloride, carbonate-rich spring systems of vari¬ able salinities and alkalinities over small spatial scales that Cole (1985) considered “ffesh-to-miohaline” (Table 2). There are numerous reports that such spatial di¬ versity of hydrochemical habitats and lithological fea¬ tures in karst basins of the eastern United States and Central Europe leads to habitat segregation, migration barriers, and geographical differences in genetic and morphological variation among gammarid amphipods in discrete, highly-structured habitats (Gooch and Hetrick, 1979; Foeckler and Schrimpff, 1985; Gooch, 1989, 1990; Sarbu et al., 1993). Similar physical and ecological constraints may be responsible for specia- tion within the G. pecos complex. Within the past 35 years, diminution or complete loss of spring flow, exacerbated by regional drought conditions and local groundwater withdrawals, is im¬ plicated in the extirpation of 2 isolated populations of gammarid amphipods in New Mexico (G. desperatus from Lander Springbrook and North Spring [Cole, 1981, 1985]), and Gammarus sp. form “C” from the seasonally dewatered and impoverished aquatic con¬ ditions of the Phantom Lake Spring canal system in Texas (Allan, 2000). Gammarid amphipods in the Phantom Lake Spring canal were previously docu¬ mented in 1995 (Winemiller and Anderson, 1997). The enigmatic recurrence of G. hyalelloides at Phantom Lake Spring pool in November 2001, where repeated survey effort from March 2000 to May 2001 implied that the species was possibly extirpated, may be attributed to a number of factors. Preliminary sur¬ vey effort focused on epigean habitats at the spring head pool since access to hypogean waters was lim¬ ited. However, recent scuba surveys did not docu¬ ment gammarid amphipods from hypogean habitats deep within the Phantom Lake Spring cave system where several aquatic troglobites (i.e., the isopods, Lirceolus cocytus [Lewis, 2001] and Cirolanides texasensis, and an undescribed cave-adapted amphi- pod) are reported (J. Krejca, University of Texas at Austin, pers. comm.). Aquatic surveys may have sim¬ ply over-looked shallow lotic habitats near the cave’s entrance that interface between epigean waters and deeper hypogean habitats. Similar cave streams are colonized by Gammarus minimus in highly structured karst areas of the eastern United States (Holsinger and Culver, 1970; Gooch and Hetrick, 1979). Secondly, it is possible that under diminished spring flows and im¬ poverish aquatic conditions of Phantom Lake Spring (Allan, 2000), that G. hyalelloides was either at very low densities or actually extirpated from epigean habi¬ tats during the period March 2000 to May 2001. Ef¬ forts to sustain the endangered fish community in Phan¬ tom Lake Spring through installation of a pumping system commenced in June 2000 with construction of a sand-bag dam in the cave mouth to maintain ad¬ equate surface water elevation in the spring pool. The pump was activated in May 2001 and began cycling water from the cave to the spring pool. Such conser¬ vation practices may have serendipitously prevented the extirpation of G. hyalelloides at Phantom Lake Spring by replenishing the epigean pool population with individuals of a hypogean source population of G. hyalelloides from inside the cave. The March 2000 Sandhill Fire at BLNWR has resulted in demonstrable effects on abiotic conditions (e.g., depressed dissolved oxygen levels [0.7 mg/1], increased temperature, spikes in salinity and pH) of Bitter Creek, which harbors 1 of 3 known G desperatus populations on BLNWR (Lang 200 1 ). Preliminary find¬ ings implicate increased levels of post-fire polycyclic aromatic hydrocarbons as a possible adverse short¬ term effect on localized populations of G. desperatus in Bitter Creek. All described amphipods of this com¬ plex are considered Federal species of concern. Gammarus desperatus is the only species that has for¬ mal protection as a state Endangered species (NMGF Regulation 657); and has been recently proposed for Federal listing as endangered under the Endangered Species Act (67 Federal Register 6459-6479). Although direct causes responsible for the extir¬ pation of gammarid amphipods from North Spring and the seasonal disappearance from the Phantom Lake Spring system remain undetermined, altered hydro- logic conditions of regional groundwater aquifers af¬ fecting these desert springs are implicated. Reduced flow regimes can significantly alter the physicochemi¬ cal balance of a lotic ecosystem (Hynes, 1970), re¬ sulting in an altered hydrologic regime, disruption of complex ecological relationships, and adverse conse¬ quences for aquatic fauna of desert spring ecosys¬ tems (Bowles and Arsuffi, 1993; Mehlhop and Vaughn, 1994; Hubbs, 1995, 2000, 2001). 54 Special Publications, Museum of Texas Tech University Local adaptive variation (genetic and morpho¬ logical) of gammarid amphipods is reported from di¬ verse aquatic ecosystems in response to local envi¬ ronmental conditions and regional geohydrology (Gooch and Hetrick, 1979; Gooch, 1989, 1990). In some instances morphological differentiation in hyalellid and gammarid amphipods has been attributed to ge¬ netic differentiation (Kane et al., 1992; Sarbu et ah, 1993; Muller et ah, 1999; Duan et ah, 2000). Physi¬ ologically stressful conditions, frequently caused by anthropogenic activities, can select for certain geno¬ types in hyalellid and gammarid amphipods (Guttman, 1994; Duan et ah, 1997; Hogg et ah, 1998, 1999). Such selection can result in highly localized popula¬ tions consisting of stenotropic demes acutely adapted to a narrow range of abiotic conditions (e.g., physico¬ chemical gradients) and hydrologic regimes. Under extreme environmental stress isolated amphipod popu¬ lations may lack sufficient genetic diversity to cope with stochastic fluctuations in the aquatic environment and are susceptible to extirpation events. From a conservation perspective, genotypic vari¬ ability provides populations with genetic plasticity to cope with environmental perturbations. Amphipod species with low genotypic variation and marked mor¬ phological differentiation are usually highly adapted to specific natural habitat conditions (e.g., G. pecos com¬ plex), but may be less capable of surviving long-term environmental change (Foeckler and Schrimpff, 1985; Duan et ah, 1997; Hogg et ah, 1998, 1999). Evolu¬ tionary consequences must be considered in the evalu¬ ation of natural or anthropogenic environmental stres¬ sors in order to direct conservation efforts that focus on not only a typological approach for species protec¬ tion, but also consider important population segments as significant evolutionary units that likewise merit preservation (Rojas, 1992; Nee and May, 1997). Cole (1985) documented marked overlap in body size and morphological characters between geographi¬ cally proximate gammarid populations “M”, “S” and “C” of West Texas. The numerous unresolved taxo¬ nomic affinities within the G. pecos complex highlight the need for phylogenetic assessment of this group focusing on comparative morphological and biochemi¬ cal genetic studies. Recent genetic studies of hyalellid amphipods (Guttman, 1994; Duan et ah, 1997; Tho¬ mas et ah, 1994; Hogg et ah, 1998; McPeek and Well¬ born, 1998) indicated that lack of phylogenetic research has impeded studies of Hyalella comparative biology, biogeography, and evolution (Duan et ah, 2000). Morphological discrimination of amphipod spe¬ cies requires analysis of whole animal morphology, proportional metrics of diagnostic character combina¬ tions and structural meristics of sexually mature males and females to distinguish cryptic species (Holsinger, 1967, 1976; Muller et ah, 1999). We are currently en¬ gaged in such studies, and limit our comparison of characters examined from new gammarid populations discovered during this survey (i.e., Giffin, East Sandia, and Sitting Bull springs) with traits diagnostic for the complex (see Cole, 1985). Biochemical genetic analy¬ ses using allozyme electrophoresis to compare within- and among-population variation are ongoing. While preliminary results of these genetic studies indicates significant within-population variation with 75% of the populations showing heterozygote deficiencies (Gervasio, unpubl. data), ongoing analyses will char¬ acterize the partitioning of genetic variation across spatial scales. Once taxonomic boundaries of this complex are identified, resource agencies and private land stew¬ ards will have a solid baseline for evaluating broad- scale environmental trends and assessment of threats due to environmental degradation in the Pecos River Valley of New Mexico and Texas. These data are es¬ sential for stewardship of aquatic resources in the Chihuahuan Desert ecoregion. Acknowledgments This research was funded by the: United States Fish and Wildlife Service (Office of Federal Aid, Re¬ gion 2 [FWS Agreement No. 1448-20 1 8-00-J6 1 8]), World Wildlife Fund, Committee on Faculty Research at Miami University, and New Mexico Department of Game and Fish. The Nature Conservancy (John Karges), United States Forest Service, Bureau of Rec¬ lamation, and private landowners facilitated access for field surveys. The Texas Parks and Wildlife Depart¬ ment (David Riskind, Tom Johnson) authorized a sci¬ entific study permit. Richard Worthington and Shan¬ non Knapp provided voucher material for this study. Lang et al. — Gammarid Amphipods of Northern Chihuahuan Desert Springs 55 Literature Cited Allan, N. L. 2000. Deterioration of Phantom Lake Spring, Jeff Davis County, Texas. Proceedings of the Desert Fishes Council, 32:50-51. Ashworth, J. B., D. B. Coker, and W. Tschirhart. 1997. Evalua¬ tion of diminished spring flows in the Toyah Creek Val¬ ley, Texas. Open File Report 97-03. Texas Water Devel¬ opment Board, Austin, 12 pp. Baldinger, A. J., W. D. Shepard, and D. L. Threloff. 2000. Two new species of Hyalella (Crustacea: Amphipoda: Hyalellidae) from Death Valley National Park, California, U.S. A. Proceedings of the Biological Society of Washing¬ ton, 1 13:443-457. Bequaert, J. C. and W. B. Miller. 1 973. The mollusks of the arid southwest with an Arizona checklist. University of Ari¬ zona Press, Tucson, 271 pp. Bousfield, E. L. 1958. Fresh-water amphipod crustaceans of glaciated North America. The Canadian Field-Naturalist, 72:55-113. Bowles, D. E. and T. L. Arsuffi. 1993. Karst aquatic ecosys¬ tems of the Edwards Plateau region of central Texas, USA: a consideration of their importance, threats to their exist¬ ence, and effort for their conservation. Aquatic Conser¬ vation of Marine and Freshwater Ecosystems, 3:317- 329. Bowman, T. E. 1981 . Thermo sphaeroma milleri and T. smithi, new sphaeromatid isopod crustaceans from hot springs in Chihuahua, Mexico, with a review of the genus. Jour¬ nal of Crustacean Biology, 1 : 1 05- 1 22. Callicott, J. B. 1994. A brief history of American conservation philosophy. Pp. 10-14, in Biodiversity and management of the Madrean Archipelago: the sky islands of south¬ western United States and northwestern Mexico (Wallace W. Covington and Leonard F. DeBano, eds.). U. S. Forest Service, Rocky Mountain Forest and Range Experiment Station. General Technical Report RM-GTR-264, 669 pp. Cole, G A. 1976. A new amphipod crustacean, Gammarus hyalelloides n. sp., from Texas. Transactions of the Ameri¬ can Microscopical Society, 89:80-85 Cole, G A. 1981. Gammarus desperatus, a new species from New Mexico (Crustacea: Amphipoda). Hydrobiologia, 76:27-32. Cole, G A. 1985. Analysis of the Gammarus-pecos complex (Crustacea: Amphipoda) in Texas and New Mexico, USA. Journal of the Arizona-Nevada Academy of Science, 20:93- 103. Cole, G. A. and E. L. Bousfield. 1970. A new freshwater Gammarus (Crustacea: Amphipoda) from western Texas. American Midland Naturalist, 83:89-95. Covich, A. P. and J. H. Thorp. 1991. Crustacea: introduction and Peracarida. Pp. 665-689, in Ecology and Classifica¬ tion of North American Freshwater Invertebrates (J. H. Thorp and A. P. Covich, eds.). Academic Press, Inc., San Diego, California, 911pp. Cummins, K. W. 1962. An evaluation of some techniques for the collection and analysis of benthic samples with spe¬ cial emphasis on lotic water. American Midland Natural¬ ist, 67:477-504. Duan, Y., S. I. Guttman, and J. T. Oris. 1997. Genetic differen¬ tiation among laboratory populations of Hyalella azteca'. implications for toxicology. Environmental Toxicology and Chemistry, 16:691-695. Duan, Y., S. I. Guttman, J. T. Oris, and A. J. Bailer. 2000. Genetic structure and relationships among populations of Hyalella azteca and H. montezuma (Crustacea: Amphipoda). Journal of the North American Benthological Society, 19:308-320. Foeckler, F. and E. Schrimpff. 1985. Gammarids in streams of Northeastern Bavaria, F. R. G. II. The different hydrochemical habitats of Gammarus fossarum Kock, 1835 and Gammarus roeseli Gervais, 1835. Archiv fur Hydrobiologia, 104:269-286. Gee, J. H. 1988. Population dynamics and morphometries of Gammarus pulex L.: evidence of seasonal food limita¬ tions in a freshwater detritivore. Freshwater Biology, 19:333-343. Gooch, J. L. and S. W. Hetrick. 1979. The relation of genetic structure to environmental structure: Gammarus minus in a karst area. Evolution, 33:192-206. Gooch, J. L. 1 989. Genetic differentiation in relation to stream distance in Gammarus minus (Crustacea: Amphipoda) in Appalachian watersheds. Archiv fur Hydrobiologia, 114:505-519. Gooch, J. L. 1990. Spatial genetic patterns in relation to re¬ gional history and structure: Gammarus minus (Amphipoda) in Appalachian watersheds. American Midland Naturalist, 124:93-104. Guttman, S. I. 1994. Population genetic structure and ecotoxicology. Environmental Health Perspectives, 102 (Supplement 1 2):97-l 00. Hershler, R., H. P. Liu, and M. Mulvey. 1999. Phylogenetic relationships within the aquatic snail genus Tryonia : im¬ plications for biogeography of the North American South¬ west. Molecular Phylogenetics and Evolution, 13:377- 391. Hill, R. T. 1900. The physical geography of the Texas region. United States Geological Survey Topographic Folio, 3:1- 2. United States Government Printing Office, Washing¬ ton, D.C. 56 Special Publications, Museum of Texas Tech University Hills, J. M. 1942. Rhythm of Permian seas-a paleographic study. Bulletin of the American Association of Petro¬ leum Geologist, 26:217-255. Hogg, I. D., C. Larose, Y. de Lafontaine, and K. G Doe. 1998. Genetic evidence for a Hyallela species complex within the Great Lakes-St. Lawrence River drainage basin: Im¬ plications for ecotoxicology and conservation biology. Canadian Journal of Zoology, 76:11 34- 1 1 40. Hogg, I. D., J. M. Eadie, and Y. de LaFontaine. 1999. Passive dispersal among fragmented habitats: the population ge¬ netic consequences for freshwater and estuarine amphi- pods. Pp. 307-326, in Crustaceans and the Biodiversity Crisis (Frederick R. Schram and J. C. von Vaupel Klein, eds.). Proceedings of the Fourth International Crusta¬ cean Congress, Volume 1 . Amsterdam, Netherlands. Holsinger, J. R. 1967. Systematics, speciation, and distribution of the subterranean amphipod Genus Stygonectes (Gammaridae). United States National Museum Bulle¬ tin, 259, 172 pp. Holsinger, J. R. 1976. The freshwater amphipod crustaceans (Gammaridae) of North America. U. S. Environmental Protection Agency, Water Pollution Control Research Series, 1 8050 ELD04/72, Cincinnati, Ohio, 89 pp. Holsinger, J. R. and D. Culver. 1970. Morphological variation in Gammarus minimus (Amphipoda, Gammaridae) with emphasis on subterranean forms. Postilla, No. 146. Hubbs. C. 1995. Springs and spring runs as unique aquatic systems. Copeia, 1995:989-991. Hubbs, C. 2000. Gambusia affinis survival in a hostile environ¬ ment. The Southwestern Naturalist, 45:521-522. Hubbs. C. 2001 . Environmental correlates to the abundance of spring-adapted versus stream-adapted fishes. Texas Jour¬ nal of Science, 53:299-326. Hynes, H. B. N. 1970. The ecology of running waters. The University of Toronto Press, 555 pp. Kane, T. C., D. C. Culver, and R. T. Jones. 1992. Genetic structure of morphologically differentiated populations of the amphipod Gammarus minus. Evolution, 46:272- 278. Lackey, R. T. 1995. Ecosystem health, biological diversity, and sustainable development: research that makes a differ¬ ence. Renewable Resources Journal, Summer 1995:8-13. Lang, W. B. 1937. The Permian formations of the Pecos Valley of New Mexico and Texas. Bulletin of the American Association of Petroleum Geologist, 216:833-898. Lang, B. K. 2001 . Status of aquatic mollusks of New Mexico. Annual Performance Report, E-20-8, submitted to the Office of Federal Aid, U. S. Fish and Wildlife Service, Albuquerque, New Mexico. Lewis, J. J. 2001. Three new species of subterranean asellids from western North America, with a synopsis of the species of the region (Crustacea: Isopods: Asellidae). Texas Memorial Museum, Speleological Monographs, 5: 1-15. Martinez, J. D., K. S. Johnson, and J. T. Neal. 1998. Sinkholes in evaporite rocks. American Scientist, 86:38-5 1 . McPeek, M. A. and G A. Wellborn. 1998. Genetic variation and reproductive isolation among phenotypically diver¬ gent amphipods populations. Limnology and Oceanog¬ raphy, 43:1162-1169. Mehlhop, P. and C. C. Vaughn. 1994. Threats to and sustainability of ecosystems for freshwater mollusks. Pp. 68-77, in Sustainable Ecological Systems: Implementing an ecological approach to land management. U. S. Forest Service, General Technical Report RM-247. Metcalf, A. L. and R. A. Smartt. 1997. Land snails of New Mexico: a systematic review. Pp. 169, In Land Snails of New Mexico (A. L. Metcalf and R. A. Smartt, eds.). New Mexico Museum of Natural History and Science Bulletin 10, Albuquerque, New Mexico, 145 pp. Muller, J., E. Partsch, A. Link, and A. Seitz. 1999. Differentia¬ tion of two cryptic Gammarus fossarum types in a con¬ tact area: morphology, habitat preference, and genetics. Pp. 279-286, in Crustaceans and the Biodiversity Crisis (Frederick R. Schram and J. C. von Vaupel Klein, eds.). Proceedings of the Fourth International Crustacean Con¬ gress, Volume 1. Amsterdam, Netherlands. Nee, S. and R. M. May. 1997. Extinction and the loss of evolutionary history. Science, 274:692-694. Noel, M. S. 1954. Animal ecology of a New Mexican springbrook. Hydrobiologia, 6 : 1 20- 135. Pennak, R. W. 1989. Fresh-water invertebrates of the United States: Protozoa to Mollusca. 3rd edition. John Wiley & Sons, Inc., New York, New York, 628 pp. Rojas, M. 1992. The species problem and conservation: what are we protecting? Conservation Biology, 6:170-178. Schuster, S. K. 1997. Hydrogeology and local recharge analysis in the Toyah Basin Aquifer. Unpublished Master’s The¬ sis, Geological Sciences, University of Texas at Austin, 130 pp. Sarbu, S., Thomas C. Kane, and D. C. Culver. 1993. Genetic structure and morphological differentiation: Gammarus minus (Amphipoda: Gammaridae) in Virginia. American Midland Naturalist, 129:145-152. Sharp, J. M., Uliana, M.M., and R. Boghici. 1999. Fracture controls on regional groundwater flow in a semiarid envi¬ ronment and implications for long-term maintenance of spring flows. Water 99 Joint Congress, Institute of Engi¬ neers, Brisbane, Australia, 2:1212-1217. Lang et al. — Gammarid Amphipods of Northern Chihuahuan Desert Spring 57 Thomas, E. P., D. W. Blinn, and P. Keim. 1994. Genetic and behavioral divergence among desert spring amphipod populations. Freshwater Biology, 38:138-143. Winemiller, K. O. and A. A. Anderson. 1997. Response of endangered desert fish populations to a constructed ref¬ uge. Restoration Ecology, 5:204-213. Addresses of Authors: Brian K. Lang Conservation Services Division New Mexico Department of Game and Fish P. O. Box 25112 Santa Fe, NM 87504 e-mail: blang@state.nm.us VlVIANALUXA GERVASIO Department of Zoology Miami University Oxford, OH 45056 e-mail: vgervasio@libero.it David J. Berg Department of Zoology Miami University Hamilton, OH 45011 e-mail: bergdj@muohio.edu Sheldon I. Guttman Department of Zoology Miami University Oxford, OH 45056 e-mail: guttmasi@muohio.edu White, B. W., D. C. Culver, J. S. Herman, T. C. Kane, and J. E. Mylroie. 1995. Karst lands: the dissolution of carbonate rock produces unique landscapes and poses significant hydrological and environmental concerns. American Sci¬ entist, 83:450-459. Nathan L. Allan U. S. Fish and Wildlife Service Ecological Services Office 10711 Burnet Road # 200 Austin, TX 78758 e-mail: nathan_allan@fws.gov Mark E. Gordon New Mexico Museum of Natural History and Science 1801 Mountain Road Albuquerque, NM 87101-1375 e-mail: gordonme@juno.com Gordon Warrick Bitter Lake National Wildlife Refuge 4065 Bitter Lakes Road Roswell, NM 88201-007 e-mail: gordon_warrick@fws.gov Declining Status of Freshwater Mussels in the Rio Grande, with Comments on Other Bivalves Robert G. Howells Abstract Freshwater mussels (Family Unionidae) are one of the fastest declining faunal groups in North America due, in part, to their sensitivity to environmental deg¬ radation and modification. There has been a notewor¬ thy paucity of data on unionid assemblages through¬ out the Rio Grande drainage despite its being a unique region of zoogeographic overlap between southern and northern faunas. Therefore, selected historic collec¬ tions, state surveys by Texas Parks and Wildlife (TPW) 1992-1997, and federally-funded work by TPW and New Mexico Department of Game and Fish 1998-2001 were reviewed to provide a better understanding of mussel status in the system. At least 16 species of unionids occurred in the Rio Grande drainage of Texas, New Mexico, and Mexico. All have been dramatically reduced in both abundance and distribution in recent decades. Only six native species have been found alive within the past ten years along with two others that are apparent introductions. Among taxa endemic to the Rio Grande, Salina mucket ( Potamilus metnecktayi ) and Mexican fawnsfoot ( Truncilla cognato ) have not been found alive since 1972 (though recently dead valves of the former collected in the late 1990s sug¬ gest it may still survive) and living or recently dead Rio Grande monkeyface ( Quadrula couchiana) were last documented in 1898. The remaining taxa, includ¬ ing some that may be abundant elsewhere, also appear to have disappeared from the system over the past 10 to 100 years. Factors contributing to this decline in¬ clude natural and anthropogenic desertification, water and land management practices, habitat modification, pollution, siltation, and increased salinity (in some ar¬ eas). Unfortunately, there is little indication status of unionids in the Rio Grande will improve in the future. Introduction The Rio Grande of the United States and Mexico is one of the longest rivers in North America and is an area of overlap for many northern and southern faunal groups (Neck, 1982; Neck and Metcalf, 1988). De¬ spite the political, ecological, and economical impor¬ tance of the area, relatively little effort had been di¬ rected at understanding freshwater mussels (Family Unionidae) of the Rio Grande (Neck and Metcalf, 1988). Freshwater mussels are one of the most rapidly de¬ clining faunal groups in North America (Neves, 1993; Williams et al., 1993) including within the Rio Grande (Howells and Garrett, 1995; Howells andAnsley, 1999). Their sensitivity to environmental disturbance and deg¬ radation in conjunction with development of the re¬ gion now places them at the pinnacle of ecological concern. Historical information comes from Dali (1896) and others who reported on early boundary surveys. Strecker (1931), Neck and Metcalf (1988), Taylor (1967, 1997), and Johnson (1998) discussed these and other historic museum collections from the drainage. Cockerell (1902) reported finding unionids in the Pecos River drainage of New Mexico; Metcalf (1982) dis¬ cussed recent, subfossil, and fossil records in the drain¬ age basin; Metcalf (1974) commented on the Pecos River fauna; Metcalf and Stem (1976) briefly men¬ tioned Rio Grande fauna; Neck (1984) commented on declining mollusks in Texas; and Neck and Metcalf (1988) reviewed mussels of the lower Rio Grande. Metcalf and Smartt (1972) discussed introduced mol¬ lusks. Several studies examined benthic invertebrates in general (Metcalf, undated, 1966; Bane and Lind, 60 Special Publications, Museum of Texas Tech University 1978; Davis, 1980a, 1980b, 1980c). Nonetheless, data from all these sources on the unionids of the Rio Grande was limited. Recently, Texas Parks and Wildlife’s (TPW) Heart of the Hills Fisheries Science Center (HOH) initiated statewide surveys of this group, including sites within the Rio Grande drainage (Howells, 1994a, 1994b, 1995, 1996a, 1996b, 1996c, 1997a, 1997b, 1997c, 1998a 1998b; Howells and Garrett, 1995; Howells et al., 1996, 1997). In 1998, federally-funded work was jointly initiated by TPW and New Mexico Department of Game and Fish to further examine the mussel fauna of the system (Howells, 1999, 2000, 2001a; Howells and Ansley, 1999; Lang, 2000). Johnson (1999) provided an analysis of the unionid fauna of the system, based on his interpretation of museum specimens and the published literature. Howells (2001b) summarized pre¬ vious records and recent TPW surveys. Collectively, from these sources, a picture of the unionid faunal composition, abundance, and distribution within the Rio Grande Basin has started to emerge. This paper presents a condensed summary of previously published reports and surveys. Materials and Methods Standard mussel sampling protocols employed in HOH work were presented in Howells (1995, 1996a). Specimen counting methods follow Howells (1995, 1996a) and include single valves and matched pairs of valves as one individual, two unmatched valves (one valve from each of two different animals) as two individuals, etc. (note that a valve or half of a shell originated from a single animal). Shell condition ter¬ minology follows (Howells, 1995, 1996a) and is pre¬ sented in Table 1 . Taxonomy follows Howells et al. (1996) and Turgeon et al. (1998), except for the com¬ mon name for Disconaias conchos presented here. Details relating to specific sampling and collection sites were presented in Howells (2001b) as were maps to all locations. Physicochemical aspects of the Rio Grande within Texas were examined by reviewing selected water quality records of the U.S. Geological Survey for water years: 1968 (USGS, 1968), 1975 (USGS, 1975), 1986 (Buckner et al., 1986), 1996 (Gandara et al., 1996), 1999 (Gandara et al., 1999), and 2000 (Gandara et al., 2001). Data reported from these sources were averaged by site for all years combined, years for all sites combined, or both in an effort to identify temporal and upstream-to-downstream pat¬ terns. Precipitation data were obtained from the Texas Office of State Climatologist, College Station, Texas, and monthly and yearly totals were summed and de¬ cade (or partial decade) averages obtained. Unless otherwise stated, recent survey work was largely confined to U.S. and boundary waters. Cur¬ rent status of bivalves in tributaries of the Rio Grande or other drainage basins in Mexico therefore remains unknown. Results and Discussion Species Accounts At least 1 6 taxa of unionids occurred in the Rio Grande drainage of Texas, New Mexico, and Mexico (Howells et al., 1996). Three species are endemic to this drainage basin. Among the freshwater mussels, two are apparently recent introductions, as are Asian clams ( Corbicula spp.; Family Corbiculidae), and an¬ other may have been lost, then reintroduced. Addi¬ tionally, at least four fingernail clams (Family Sphaeriidae) have also been reported in the system. Tampico pearlymussel ( Cyrtonaias tampicoensis ) The native range of Tampico pearlymussel ex¬ tends from the Brazos River, Texas, (Howells et al., 1996) south to at least the Rio Papalopan system, Vera Cruz, Mexico (Johnson, 1999). Upstream records extend to the eastern boundary of Big Bend National Park (Howells, 1994b), with subfossil and fossil re¬ mains in the Pecos River upstream to Chaves and Eddy Howells — Declining Status of Mussels in the Rio Grande 61 Table 1. Shell condition terminology as presented by Howells (1995, 1996a). Note that it is not usually possible to determine exactly how long a freshwater mussel shell has been dead and many variables can impact specimen condition and rate of disintegration. Nonetheless, it is often useful to apply qualitative terms to help estimate time-since-death. Term Definition Very recently dead Soft tissue remains attached to the shell or valve; shell is in good conditiion, essentially as it would bein a living specimen; internal and external colors are not faded. Recently dead No soft tissue remains, but shell otherwise in good condition (looking like a living speciment that had been killed and cleaned); internally nacre is glossy and without evidence of algal or other staining, calcium deposition, or external erosive effects; internal and external colors are not significantly faded. Relatively-recently dead Shell is in good condition, but internally nacre is losing its glossy nature; algal or other staining, calcium deposition, or external erosive effects (or some combination of these) is evident on the nacre; internal and external colors are often somewhat faded. Relatively-long dead Similar to above, but more pronounced; shell epidermis often has sections beginning to flake. Long dead Shell shows signs of internal and external erosion, staining, calcium deposition, or some combination of these; most or all of the internal coloration and glossy nature of the nacre has faded (especially in species with colored nacre); shell epidermis often has major sections absent, or if present, clearly aged and flaking. Very-long dead Shell shows significant signs of internal and external erosion, staining, and calcium deposition more widely pronounced than above; coloratioin is often faded white or nearly so; often with relatively little epidermis remaining; for specimens in particularly erosive environments (scoured substrates, low pH waters, etc.), internal (e.g., shell teeth) and external features (e.g., pustules, ridges) are often weathered and smoothed, or otherwise exfoliated; shells are often chalky, brittle, and crumbling. Subfossil Shells have little or no epidermis remaining; nacre is faded white and entire shell is often bleached white; shell sometimes shows signs of erosion, staining, or calcium deposition of recent origin; shells are typically chalky and powdery to the touch, and often brittle and crumbling. counties, New Mexico (Metcalf, 1982). It appears to be the most abundant unionid remaining in the Rio Grande. In the 1990s, populations were found in Amistad, Falcon, and Casa Blanca reservoirs, and in resacas, canals, oxbows, and reservoirs in Hidalgo and Cameron counties, Texas. However, the species is now apparently absent from areas upstream of Amistad Reservoir and Texas tributaries downstream to Starr County. Drought conditions, starting in mid- 1995 and continuing into 1996 and beyond, also eliminated large numbers from Falcon and Amistad reservoirs as wa¬ ter levels fell. Conchos disk {Disconaias conchos) This species was first described by Taylor (1997) from 1969 collections near Saucillo, Rosetilla, and Julimes in the Rio Conchos, Chihuahua, and a 1937 collection at Villa Juarez in the Rio Sabinas, Coahuila, Mexico. It is apparently endemic to those systems and not known from the main stem of the Rio Grande or other tributaries. During desert fishes sur¬ veys by the HOH staff in 1994, 10 unmatched valves were found in the Rio Conchos near Julimes, but were listed as unidentified (Howells and Garrett, 1995; Howells, 1996a). Among these, all were long dead except a single juvenile that was relatively-recently dead. Status of this species is uncertain, but the 1994 juve¬ nile is the only suggestion the species may still sur¬ vive. Yellow sandshell ( Lampsilis teres) Yellow sandshell ranges throughout much of the central U.S. and is often common elsewhere in Texas (Howells et al., 1996). In the Rio Grande drainage, it 62 Special Publications, Museum of Texas Tech University has been documented in the Pecos River, Val Verde County, Texas (Johnson, 1999); the Rio Grande near Del Rio, Val Verde County, downstream to Cameron County, Texas (Strecker, 1931); and the Rio Sabinas, Coahuila (Johnson, 1999); and Rio Salado near Anahuac, Nuevo Leon (Metcalf, 1982), Mexico. Neck and Metcalf (1988) described it as common in resacas in the Lower Rio Grande Valley and second in abun¬ dance only to Tampico pearlymussel. However, HOH surveys found only: 1) one subfossil valve in the Rio Grande downstream of Falcon Dam, Starr County, 1994 (Howells, 1996a); 2) a very-long dead valve in an oxbow pond, La Coma Tract - Lower Rio Grande National Wildlife Refuge, Hidalgo County, 1999 (Howells, 2000); and 3) a recently dead specimen in Elm Creek, Eagle Pass, Maverick County, 1992 (Howells, 1994a). A Late Archaic archeological site at Eagle Pass examined in 1995 (dated to 2,200-1,200 years before present) revealed yellow sandshell to have been the most abundant unionid material recovered (Howells, 1998c). Unfortunately, when the Elm Creek site was reexamined in 1996, the stream bottom was covered over 1 m deep in soft silt that had eliminated all unionids (Howells, 1997a). Despite a history of wide distribution in the system, the 1992 Elm Creek specimen was the only suggestion living animals might remain, but none have been located since. Washboard (Megalonaias nervosa) The range of this large waterbody species in¬ cludes the central United States and lower river reaches in Texas from the Red River to the Rio Grande (Howells et al., 1996). In the Rio Grande, fossil shells have been found as far upstream as the Pecos River, Eddy County, New Mexico (Metcalf, 1982). Records in Texas range from a fragment at Elm Creek, Maverick County (Metcalf, 1982) and Las Moras Creek, Kinney County (Strecker, 1931) to fresh shells in the Rio Grande at Chapeno downstream of Falcon Dam in the mid-1970s (Neck and Metcalf, 1988). Records from Mexico include the Rio Sabinas (Johnson, 1999) and fossil material from the Rio Salado near Anahuac, Nuevo Leon, and sub fossil specimens near Villa Juarez, Coahuila (Metcalf, 1992). Neck and Metcalf (1988) indicated washboard was rare in the Rio Grande and possibly extinct. Collections by HOH produced only a single very-long dead shell near Chapeno. However, D. Kumpe (South Padre Inland, Texas; pers. comm.) reported being given a fresh specimen taken from the lower Rio Grande and seeing washboard shell frag¬ ments in dredge spoils at the mouth of the Edinburg Canal, Hidalgo County, Texas, in the mid-1990s. Ad¬ ditionally, P.D. Hartfield (U.S. Fish and Wildlife Ser¬ vice, Jackson, Mississippi; pers. comm.) interviewed a commercial musseler in the 1990s who was in pos¬ session of washboards he claimed had been recently collected in the lower Rio Grande Valley of Texas. The collection also contained Tampico pearlymussels (that could only have come from Texas or Mexico). HOH collections in Starr, Hidalgo, and Cameron counties failed to find fresh or living washboards, but some deepwater areas of the main channel remain to be ex¬ amined, as does the main stem run between Amistad Dam and the Falcon Reservoir. Texas hornshell ( Popenaias popeii ) Though sometimes reported as endemic to the Rio Grande (Howells et al., 1997), Texas hornshell historically ranged south along the Mexican coastal systems at least to the Rio Cazones, Vera Cruz (Johnson, 1999). It has been found upstream in the Rio Grande to sites just downstream of Big Bend, Brewster County, Texas (Howells, 1 999); upstream in the Pecos River to Chaves and Eddy counties, New Mexico (Cockerell, 1902; Metcalf, 1982); in the Dev¬ ils River upstream of Dolan Springs, Val Verde County, Texas (Howells, 2000); as well as several Mexican tributaries of the lower Rio Grande (Johnson, 1999). A shell found in the Llano River in 1971 at Castell, Llano County (Ohio State University Museum collec¬ tion), and another recently dead shell taken in the South Concho River in 1992 at Cristoval, Tom Green County (N. Strenth, Angelo State University, San Angelo, Texas; pers. comm.), both in the Colorado River drain¬ age of Texas, are enigmatic. However, numerous col¬ lections throughout this system have failed to find evi¬ dence of other specimens. The only records in the past decade of living or recently dead specimens in¬ clude one recently dead shell in 1 992 from the mouth of San Francisco Creek, Brewster and Terrell coun¬ ties, Texas (Howells, 1994a); three recently dead shells in 1998 found in the Rio Grande between Big Bend and San Francisco Creek, Brewster County, Texas (Howells, 1999); and an extant population in the Black River (Pecos River tributary), Eddy County, New Mexico (Lang, 2000). The Black River population in New Mexico and a possible relict population in the Rio Grande downstream of Big Bend may be the only sur¬ viving Texas hornshell in the United States. Howells — Declining Status of Mussels in the Rio Grande 63 Salina Mucket {Potamilus metnecktayi) Salina mucket has been included under both Disconaias and Potamilus (Howells et ah, 1996). Johnson (1998) redescribed endemic animals from the Rio Grande as Potamilus metnecktayi and seems to have combined a number of other related species to the south of the Rio Grande under Disconaias disca. In the Rio Grande, the species has been documented from just upstream of Boquillas Canyon, Brewster County, Texas (Howells, 1999), and from fossil mate¬ rial from the Pecos River, Eddy County, New Mexico (Metcalf, 1982), downriver to below Falcon Dam, Starr County, Texas (Neck and Metcalf, 1988), and the Rio Salado of Tamaulipas, Coahuila, and Nuevo Leon, Mexico (Johnson, 1999). Living specimens were found in 1968 in the lower Pecos River, Val Verde County, Texas (Johnson, 1999) and fresh shells were found in the Rio Grande west of Del Rio, Val Verde County, in 1972 (Metcalf, 1982). No living speci¬ mens have been documented since. Two recently dead specimens were found at Dryden Crossing, Terrell County, Texas, in 1992 (Howells, 1994a); 11 recently dead and two relatively recently dead shells or valves were collected in 1998 between Dean Canyon, Brewster County, and 4.4 km downstream of San Fran¬ cisco Creek, Terrell County (Howells, 1999); and a single relatively recently dead valve was found just upstream of Boquillas Canyon in 1999 (Howells, 2000), all in the Rio Grande. In nearly 30 years, these are the only records that suggest the Salina mucket is still ex¬ tant. Bleufer {Potamilus purpuratus ) The native range of bleufer extends from the Guadalupe-San Antonio drainage of Texas north and east (Howells et al., 1996). Previous reports of this species in the Rio Grande represent misidentified Tampico pearlymussels (Howells, 1997c). However, shells and living specimens were found in Amistad Reservoir, Val Verde County, Texas, in 1994 and 1995 where it appears to have been introduced (Howells, 1997c). It was documented there again in 1998 (Howells, 1999). Giant floater {Pyganodon grandis) Though widely distributed and often common in Texas and elsewhere in the United States (Howells et al., 1996), there have been few reports of giant floater from the Rio Grande. Records include the El Toro Cement Agency Lake, El Paso County, Texas, 1969 (specimens at the U.S. National Museum and Philadel¬ phia Academy of Science); Granjeno Lake (Strecker, 1931) and canals at Mercedes (Ellis et al., 1930), both Hidalgo County, Texas; and fossil material from Billingslea Draw near Toyah in the Pecos River drain¬ age, Reeves County, Texas (Metcalf, 1982). None were taken in any HOH surveys conducted since 1992; however, in 2000, Lang (2000) found living specimens in the Pecos River upstream of Avalon Reservoir and in the main stem at Flume Park Carlsbad, Eddy County, New Mexico. Metcalf and Smartt (1972) considered giant floater to have been introduced at the El Paso site, as did Lang of the New Mexico specimens. The El Paso site has not been reexamined in many years and the status of that population is undetermined. Other than the population in New Mexico (possibly non-na¬ tive stock), there is no other recent confirmation of the species in the Rio Grande drainage. Southern mapleleaf {Quadrula apiculata) Although southern mapleleaf is common throughout most of Texas and other Gulf states (Howells et al., 1996), it is absent from the fossil record (Metcalf, 1982), historical, and Mexican collections from the Rio Grande. Neck and Metcalf (1988) sug¬ gested it may be an introduction. It is known to occur in Lake Casa Blanca, Maverick County; Falcon Reser¬ voir, Zapata County; and in the Rio Grande, canals, resacas, oxbows, and impoundments in Starr, Hidalgo, and Cameron counties, Texas (Neck and Metcalf, 1988; Howells, 1996a, 1997a, 1999, 2000). Rio Grande monkeyface {Quadrula couchiana ) This endemic species has been reported as fossil remains as far upstream in the Pecos River as Eddy County, New Mexico, and the Pecos River near the mouth of Hackberry Draw, Ward County, Texas (Metcalf, 1982), with other material from the Devils River (Johnson, 1999) and downstream in the Rio Grande tributary of Las Moras Creek, Fort Clark (Brackettville), Kinney County, Texas (Strecker, 1931). Other subfossil and fossil specimens have been found at sites in the Rio Salado, Coahuila and Nuevo Leon, 64 Special Publications, Museum of Texas Tech University Mexico (Metcalf, 1982). A record from the Nueces River drainage, Zavalla County, Texas, is thought to be spurious (Johnson, 1999). This species may have been last seen alive or recently dead in the 1898 col¬ lections in Las Moras Creek (see Taylor, 1967). Neck (1984) reported D.W. Taylor had told A.L. Metcalf that he had found living specimens in the Rio Conchos, Chihuahua, Mexico; however, no subsequent published record has been presented and none were found there during HOH surveys of this system in the 1990s. Oth¬ erwise, only fossil material has been found in the past century and the species may well be extinct. False spike ( Quincuncina mitchelli) This species is known only from two disjunct populations, one in Central Texas and the second in the Rio Grande drainage (Howells et al., 1996). Living specimens have not been found in Central Texas since the late 1970s (Howells etal., 1997). Nearly all records from the Rio Grande are those of Metcalf (1982) who documented subfossil and fossil specimens from the Pecos River, Eddy County, New Mexico; the Pecos River drainage of Reeves and Ward counties, Texas; and in Mexico in the Rio San Juan of Nuevo Leon, and Rio Salado, Nuevo Leon and Coahuila, Mexico. Johnson (1999) noted an additional collection from the Rio Salado, Tamaulipas, in the University of Michi¬ gan Museum of Zoology, but did not comment on the condition of the specimen. Surveys by HOH have failed to find even fossil fragments of false spike in the Rio Grande and the species is likely extinct in the ba¬ sin. Two recently dead valves found in the lower Guadalupe River drainage in 2000 (Howells, 2001) are the only suggestion the Central Texas population has not been lost as well. Lilliput ( Toxolasma parvus) Taxonomic confusion among species in this ge¬ nus has been problematic, particularly with historic collections. Lilliput occurs throughout most of Texas and much of the United States (Howells et al., 1996), but in the Rio Grande, it is restricted to the lower reaches where it has been rather rare. Reports include Delta Reservoir, Hidalgo County (H.D. Murray; col¬ lection now at the Philadelphia Academy of Science) and resacas of the lower Rio Grande (Neck and Metcalf, 1988). However, during HOH surveys collections in¬ cluded only a single relatively recently dead valve in 1994 from Lake Casa Blanca, Maverick County (Howells, 1996a), and four living specimens in 1995 from Falcon Reservoir, Zapata County (Howells, 1996b). No other occurrences have been documented since. Texas lilliput ( Toxolasma texasiensis) Although Turgeon et al. (1998) and Williams et al. (1993) recognized lilliput as well as Texas lilliput and western lilliput (T. mearnsi), genetic analysis (Howells, 1997b) failed to distinguish the latter taxon as a distinct species (Howells, 1997b). Howells et al. (1996) included western lilliput under Texas lilliput. Often locally abundant in much of Texas (Howells et al., 1996), Texas lilliput has only rarely been reported in the Rio Grande. Historic collections, all in Texas, include the mouth of the Devils River, Val Verde County, and Las Moras Creek, Fort Clark (Brackettville), Kinney County (Strecker, 1931; Taylor, 1967). Texas lilliput was third in abundance in archeological excavations along Elm Creek, Maverick County (Howells, 1998c), where its sexually dimorphic shells were readily ap¬ parent. However, it was not found in HOH survey sites throughout the basin examined during the past decade. Mexican fawnsfoot ( Truncilla cognata) This Rio Grande endemic is known only from a small number of specimens. Originally described from the Rio Salado, Nuevo Leon, Mexico in 1860, it has also been taken from the same river in the state of Tamaulipas (U.S. National Museum; Johnson, 1999). Metcalf (1982) reported a specimen of probably fossil origin from the Rio Salado, Nuevo Leon. Metcalf (1982) also noted finding fresh shells in 1968 in the lower Pecos River and in 1972 in the Rio Grande west of Del Rio, both in Val Verde County, Texas. C.M. Mather (University of Science and Arts of Oklahoma, Chickasha; pers. comm.) collected a weathered valve in the Rio Grande 72 km west of Laredo in 1975. A relatively-long dead and deformed valve found in Fal¬ con Reservoir in 1996 and initially attributed to this species (Howells, 1997b) may be a misidentification. None have been documented otherwise since the 1972 Metcalf collection. Howells — Declining Status of Mussels in the Rio Grande 65 Pondhorns (sp.?) ( Uniomerus tetralasmus - U. declivis) A pondhom, given as Unio manibius, was taken in an early boundary survey at Rio Agualeguas, P'untiagudo, near General Trevino, Nuevo Leon, Mexico, in the 1850s (see photograph in Johnson, 1999). Morrison (1976) synonymized Unio minibus and Uniomerus declivis, and he and Frierson (1903) both considered U. tetralasmus and U. declivis to be valid species. However, Johnson (1999) combined them. Neck (1987) considered both the Rio Agualeguas specimen and others from Baffin Bay drainage, the next system to the north, to be U. tetralasmus. Many Texas Uniomerus populations are intermediate between the two species (HOH unpublished data), as are both Rio Agualeguas and Baffin Bay examples, and their true taxonomic affinities remain unclear. Neck and Metcalf (1988) indicated specimens collected in the 1 920s in Cottingham Resaca, Brownsville, Texas, were present in the Corpus Christi Museum collection; how¬ ever, this collection could not be located for examina¬ tion. Regardless, Neck (1987) concluded that no mem¬ bers of this genus were currently known from the lower Rio Grande Valley. Recent work supports the conclusion that all pondhorns of any species appear to have been eliminated from the system. Paper pondshell ( Utterbackia imbecillis ) Paper pondshell is often common in most Texas drainages and elsewhere in the United States. In the Rio Grande, it has been documented from Matamoros in the mid- 1800s, Tamaulipas, and Rio Salado near Anahuac, Nuevo Leon, Mexico (Johnson, 1999). His¬ toric sites in Texas include: Brownsville, Cameron County (Johnson, 1999); San Lorenzo Creek, Webb County (Johnson, 1999); Las Moras Creek, Kinney County (Strecker 1931); Rio Grande upstream of the Rio Conchos in 1979, Presidio County (Metcalf, un¬ dated); and Beaver Lake on the upper Devils River, Val Verde County (Strecker, 1931). More recent Texas records include: an oxbow pond and Sapo Lake, Hidalgo County (Howells, 2000); Falcon Reservoir, Zapata County (Howells, 1998a, 1999); Lake Casa Blanca, Maverick County (Howells, 1994a, 1996a, 1997a); the lower Devils River, Val Verde County (Howells, 1996a); and Lake Balmorhea, Reeves County (Howells, 1998a, 1999). Two lots of paper pondshell taken by R.D. Camp early in the past century upstream of El Paso in the San Jose River near San Rafael, Valencia County, New Mexico are present in the Cor¬ pus Christi Museum collection (J. Deisler-Seno, Cor¬ pus Christi Museum; pers. comm.). Small popula¬ tions probably persist in ponds, impoundments, and backwaters from the lower Rio Grande Valley upstream to the lower Devils River and in spring runs at Balmorhea of the Pecos River drainage. Status of the species in Mexico and New Mexico is uncertain. Species Reports of Doubtful Validity Threeridge (Amblema plicata ) was reported by Dali (1896), as Unio undulatus, from Kinney County based on a single “badly broken, and much worn right valve.” In as much as even fossil specimens are oth¬ erwise lacking from the system, it seems likely the valve was either a misidentified washboard or the col¬ lection locality was incorrect. Round pearlshell ( Glebula rotundata) was listed by Simpson (1914) and Strecker (1931), based on an earlier report by Conrad (1855), from the Rio Grande. The species is other¬ wise not known from areas south of Green Lake at the mouth of the Guadalupe River and reports of round pearlshell from the Rio Grande are likely misidentified Tampico pearlymussel (Howells et al., 1996). Although it is probably not part of the Rio Grande assemblage, P.D. Hartfield (U.S. Fish and Wildlife Service, Jack- son, Mississippi; pers. comm.) reported seeing this species among others a commercial musseler claimed to have taken from the Rio Grande (as discussed above). Nonetheless, recent survey efforts failed to find it. 66 Special Publications, Museum of Texas Tech University Freshwater Bivalves In Other Families Asian clam ( Corbicula fluminea), first found near El Paso in the mid-1960s (Metcalf, 1966), occurs throughout the drainage basin (Howells et al., 1996). A second possible species in the genus has also been reported from the Rio Grande (Hillis and Patton, 1981 ; Hillis and Mayden, 1985). Among the fingernail clams, several have been documented including: long fmgemailclam Musculium transversum, ubiquitous peaclam Pisidium casertanum, striated fmgemailclam Sphaerium striatinum, mottled fmgemailclam Eupera cubensis (Metcalf, undated; Davis, 1980a, b, c). Asian clam is an undesirable exotic, but its status in the Rio Grande has rarely been quantified. Fingernail clams are even less well studied than unionids. Possible Reasons for Decline Climate and Weather The Rio Grande drainage basin has been natu¬ rally changing from a cooler, more -moist climate to a warmer and dryer situation. Wilkins (1992) described a general climatic trend toward warming and drying since the end of the Wisconsin glaciation about 1 8,000 years ago. Smith and Miller (1986) discussed a re¬ gional shift from woodlands to grasslands and deserts from about 11,500 years ago. Loss of species like washboard from upriver areas in New Mexico and its range reduction to the lower-most Rio Grande reflects this ongoing climatic change. Similarly, Rio Grande monkeyface and false spike may have been in a mode of natural decline prior to European impact. Recent precipitation pattern changes may also play a role in mussel decline. Long-term precipitation records for selected sites in Texas yield 10-year aver¬ ages demonstrating a pattern of increasing rainfall at Brownsville (61.7 cm in the 1950s to 83.0 cm in the 1990s) and Laredo (112.0 cm in the 1940s to 133.6 cm in the 1990s). However, this pattern was less evi¬ dent at Del Rio and was not found at El Paso. In addition to increasing net rainfall in some areas, weather records also indicate a shift to fewer light and moder¬ ate showers and more severe storm events (MICRA, 1995). Thus, recent decades have experienced more major storms producing scouring floods that directly destroy unionids and alter habitat, with long periods of insufficient precipitation in between resulting in dewa¬ tering losses. Freshwater mussels need stable condi¬ tions to survive and prosper. Current precipitation trends are clearly unfavorable. Water Flow Patterns Human impact appears to be the major reason for the massive reduction in mussel faunal diversity and abundance in the Rio Grande. Overgrazing, land clearing, construction of impervious surfaces, and other anthropogenic modifications have also contrib¬ uted to increased runoff during storms and additional scouring and riverbed modifications. Increased groundwater pumping, in turn, reduces spring flows and subsequently river water levels during dry peri¬ ods. All of these factors can contribute to negative impacts on unionids. Flow rates reflect an interrelationship of both precipitation and water retention or releases from im¬ poundments, as well as other factors. The U.S. Na¬ tional Park Service (NPS, 2001) reported 69-86% of the water in the Rio Grande downstream from Presidio originated from the Rio Conchos, Mexico, and that although a treaty between the U.S. and Mexico de¬ fined allotments related to annual flows, the treaty did not specify release schedules for Mexican rivers. Mean annual flows by site are least at El Paso and Brownsville, but greatest between Langtry and Rio Grande City (Table 3). Mean annual flows during the years exam¬ ined display a pattern of decrease over time from the Rio Grande downstream of Amistad dam to Brownsville (Table 3). Flows at Laredo in 1975 averaged 149 m3/ s, but dropped in 2000 to 43.2 mVs. At Brownsville, flows were 100 m3/s in 1975, but fell to 1 m3/s in 1999 and 5 m3/s in 2000. In 2001, flow downstream of Howells — Declining Status of Mussels in the Rio Grande 67 Brownsville stopped and freshwater failed to reach the Gulf of Mexico at times. Historic droughts have also reduced or eliminated flows in the Rio Grande as in 1952, 1955, 1957, and 1958 when the riverbed was dry at Johnson Ranch, Big Bend National Park (NPS, 2001). Past dewatering likely reduced or eliminated some unionid populations and the current pattern of flow rate decline poses an increasing threat again. However, large-volume releases below mainstem dams, which may actually occur during drought conditions, can cause scouring-related damage to mussels and aquatic habitat at downstream sites as well. Nutrient Loads Mean annual levels of total phosphorus between El Paso and Brownsville and in the Pecos River at Langtry were below 0.5 mg/L (Table 2). Similarly, sulfate ranged from 161 -290 mg/L in the main channel of the Rio Grande, was low in the Devils River (9 mg/ L), but averaged over 2,000 mg/L in the upper and central Pecos River of Texas (Table 2). Among all years combined, nitrate levels averaged below 1 .4 mg/ L in the main channel and the Pecos River at Langtry, except at Brownsville where it slightly exceeded 1.6 mg/L (Table 2). Interestingly, nitrate concentrations show a significant pattern of decline over time. In 1968, levels were very high in the Rio Grande at Langtry (3.2 mg/L), Laredo (3.7 mg/L), and Brownsville (6.0 mg/L), but decreased to < 1 .0 mg/L at these same sites in 2000, with a decline to 0.2 mg/L at Brownsville (Table 3). Although there is little pub¬ lished information associating nutrient levels and fresh¬ water mussels, it seems unlikely that concentrations currently found in the Rio Grande would be problem¬ atic (but with some concern about sulfate in the upper and central Pecos River). Table 2. Mean values for selected physicochemical parameters at locations on the Rio Grande, Pecos River, and Devils River, Texas, obtained from measurements presented in U.S. Geological Survey reports for water years 1968 (USGS, 1968), 1975 (USGS, 1975), 1986 (Buchner et al., 1986), 1996 (Gandaraetal., 1996), 1999 (Gandara et al., 1999), and 2000 (Gandara et al., 2001). Values for Amistad, Falcon, and Anzalduas dams were actually taken in the Rio Grande downstream of those structures. Parameter Rio Grande - Main channel El Paso Langtry Amistad Dam Laredo Falcon Dam Rio Grande City Anzalduas Dam Brownsville Flow rate (m3/s) 22.2 52.1 45.6 83.1 63.5 68.2 34.6 33.0 Total phosphorus (mg/L) 0.41 0.9 <0.01 0.13 0.06 - - 0.15 Sulfate (mg/L) 289 290 214 161 200 209 233 233 Nitrate (mg/L) 0.89 1.35 0.22 1.14 0.40 - - 1.62 Chloride (mg/L) 148 117 141 116 116 136 167 181 Conductivity (pS/cm) 1368 1214 1103 959 971 1065 1020 1322 Suspended sediments (mg/L) 535 2685 5 137 20 - - 64 Dissolved solids (mg/L) 863 772 663 598 589 645 730 761 Turbidity (NTU) 140 775 1 53 8 - - 27 Orla Pecos River Girvin Langtry Devils River Comstock Flow rate (m3/s) 2.1 1.0 9.5 6.4 Total phosphorus (mg/L) - - 0.04 - Sulfate (mg/L) 2012 3477 467 9 Nitrate (mg/L) - - 0.92 0.94 Chloride (mg/L) 3477 5818 777 15 Conductivity (pS/cm) 12651 21585 34551 381 Suspended sediments (mg/L) - - 12 22 Dissolved solids (mg/L) 9742 12953 2038 214 Turbidity (NTU) - - 10 5 68 Special Publications, Museum of Texas Tech University Table 3. Annual mean values for flow rate, nitrate concentration, and suspended sediments at locations on the Rio Grande, Pecos River, and Devils River, Texas, obtained from measurements presented in U.S. Geological Survey reports for water years 1968 ( USGS , 1968), 1975 (USGS, 1975), 1986 (Buchner et al, 1986), 1996 (Gandara et al., 1996), 1999 (Gandara et al., 1999), and 2000 (Gandara et al., 2001). Values for Amistad, Falcon, and Anzalduas dams were actually taken in the Rio Grande downstream of those structures. Year 1968 1975 1986 1996 1999 2000 Flow Rate (m3/s) Rio Grande - El Paso - - 27.2 16.9 19.9 24.8 Rio Grande - Langtry - 54.8 52.0 35.2 65.3 53.4 Rio Grande - Amistad Dam - - 80.4 46.2 30.4 25.5 Rio Grande - Laredo - 148.9 110.9 46.8 65.8 43.2 Rio Grande - Falcon Dam 98.4 61.4 32.7 61.5 Rio Grande - Rio Grande City - - 97.4 64.0 56.6 54.9 Rio Grande - Los Ebanos - - 98.3 30.4 41.7 38.1 Rio Grande -Anzalduas Dam - - 39.0 33.4 34.8 31.1 Rio Grande - Brownsville 81.1 99.5 8.2 3.1 1.0 5.1 Pecos River - Orla 2.4 2.6 0.8 2.5 _ _ Pecos River - Langtry 0.5 2.0 7.2 16.0 0.6 0.7 Devils River - Comstock - - 6.0 6.9 3.6 4.0 Nitrate (mg/L) Rio Grande - Langtry 3.2 1.5 1.1 0.7 0.8 0.8 Rio Grande - Laredo 3.7 0.6 0.4 - 0.7 0.5 Rio Grande - Brownsville 6.0 0.3 0.9 0.1 0.4 0.2 Pecos River - Langtry - 1.8 0,6 0.3 0.4 0.9 Devils River - Comstock - - 1.2 0.7 - - Suspended sediments (mg/L) Rio Grande - Langtry _ 654 2364 4113 3609 Rio Grande - Falcon Dam - - 122 29 19 20 Rio Grande - Brownsville - - 120 63 47 64 Chloride and Conductivity Chloride concentrations and associated conduc¬ tivity values may have a direct impact on mussel pres¬ ence or absence at some locations in the basin. Aver¬ ages among years produce chloride values from El Paso to Brownsville of 116-181 mg/L and conductiv¬ ity values of 959-1,368 pS/cm that are typically great¬ est at up- and downstream sites (Table 2). In the Devils River, chloride averaged only 1 5 mg/L and con¬ ductivity 381 pS/cm (Table 2). However, chloride concentrations in the upper (Orla), central (Girvin), and lower (Langtry) Pecos River were dramatically elevated to 3,477, 5,818, and 777 mg/L, respectively (Table 2). Similarly conductivity at these same sites was 12,651, 21,585, and 34,551 pS/cm, respectively (Table 2). Thus the Pecos River is the major source of el¬ evated chloride and conductivity values in the system. Natural salt seeps and deposits are present in the area, but groundwater pumping that has reduced freshwa- Howells — Declining Status of Mussels in the Rio Grande 69 ter input, long periods of reduced precipitation, and brines from oil and gas drilling operations likely con¬ tribute to current saline conditions. Chloride and con¬ ductivity levels in the mainstem Rio Grande are prob¬ ably not limiting (though upper lethal limits for most unionids remain undefined). However, levels in the upper and central reaches of the Pecos River in Texas are probably sufficiently high as to preclude long-term mussel survival. At most sites in the Pecos River, Texas, even disturbance-tolerant Asian clam is not present. In addition to the dam at Red Bluff Reservoir, salt waters probably genetically isolate the Texas homshell population in the Black River, New Mexico, from any survivors in the Rio Grande downstream of Big Bend. Total Dissolved Solids and Sediment Load Total dissolved solids average from El Paso to Brownsville range from 589-863 mg/L, are lower in the Devils River (214 mg/L), but dramatically elevated in the Pecos River (2,038-12,953 mg/L)(Table 2). Mean suspended sediment levels (mg/L) by location from upstream to downstream in the Rio Grande were 535 (El Paso), 2,685 (Langtry), 5 (Rio Grande down¬ stream of Amistad Dam), 137 (Laredo), 20 (down¬ stream of Falcon Dam), and 64 (Brownsville), and were 12 (Pecos River, Langtry) and 22 (Devils River, Comstock)(Table 2). The increase in suspended sedi¬ ment agrees with field observations of heavy recent, silt deposition at and upstream (in the Rio Grande) of the Pecos River mouth. Indeed, comparing annual means indicates suspended sediment loads typical of the system in general in the Rio Grande at Langtry in 1986 (654 mg/L), but with levels elevating in the late 1990s to 2,364 (1996), 4,113 (1999), and 3,609 mg/L (2000)(Table 3). U.S. Geological Survey (Gandara et ah, 2001) data from a sampling station in the Rio Grande just downstream of the Rio Conchos indicated an average suspended sediment load in 2000 of 169 mg/L, lower than location averages elsewhere in the mainstem of the Rio Grande. The Rio Conchos, then, appears not to be the main source of the suspended sediments that have increased between Presidio and Langtry. Turbidity measurements follow the same pattern found with suspended sediments (Table 2). Although the primary source of increased sediment is unclear from recent freshwater mussel survey work, certainly overgrazing, development, and other typical anthropogenic factors contribute. It is noteworthy the only areas in the Rio Grande of Texas where Texas homshell and Salina mucket may still be surviving are in or just upstream of the major site of silt deposition. Deep soft silts are unacceptable mussel habitat (Howells et al., 1996). Finally, despite heavy silt loads and deposition at Langtry, Amistad Reservoir appears to block much of this material from progressing down¬ stream, as does Falcon Reservoir to a lesser extent downriver. Although this may benefit mussels still surviving in the river below these dams, it suggests a questionable future for reservoir populations. Reservoir Construction and Pollution Reservoir construction and resultant impounded waters may eliminate some mussels and their natural habitat while creating additional habitat for other spe¬ cies (Howells et al., 1996). Indeed, neither Texas homshell nor any of the endemic unionids in the Rio Grande are known from reservoirs and may well re¬ quire flowing water situations. Conversely, Tampico pearlymussel adapts well to reservoirs, despite having evolved in riverine environments (Howells, 1996c; Howells et al., 1996). Even when unionids are able to survive in impounded waters, water management prac¬ tices can still be destructive. Rapidly fluctuating wa¬ ter levels and long periods of dewatering are common sources of mortalities. These problems were particu¬ larly evident in Amistad and Falcon reservoirs when water levels began to decline substantially in mid- 1995. Reservoirs may also block movement of host fishes required for parasitic unionid larvae. Host fishes uti¬ lized by Tampico pearlymussel, yellow sandshell, wash¬ board, bleufer, giant floater, lilliput, Texas lilliput, pondhom (sp.?), and paper pondshell (Howells, 1997d; Howells et al., 1996), as well as Texas homshell (Gor- 70 Special Publications, Museum of Texas Tech University don, 2000) are probably not seriously limited by dams in the basin. However, hosts required by Conchos disk, Salina mucket, southern mapleleaf, Rio Grande monkeyface, and Mexican fawnsfoot are unknown; thus host availability issues cannot be addressed. Pollution and habitat modification associated with human development have probably impacted unionid fauna of the Rio Grande as they have throughout the country, but documentation locally is largely lacking. However, some areas reportedly have a high potential for chemical impacts (Kelly and Reed, 1998). Tre¬ mendous human population growth and human devel¬ opment has occurred in the drainage basin over the past 30 years (Kelly and Reed, 1998). The North American Free Trade Agreement (NAFTA) of recent years has no doubt enhanced the speed and extent of this development. It is not apparent, however, that any NAFTA-related development has ever specifically considered impacts on unionids. Commercial Harvest Freshwater mussels have supported important commercial shell fisheries elsewhere in Texas and the U.S. (Howells, 1993). Historically, the U.S. military harvested local shells for buttons early in the past cen¬ tury in Cameron County, Texas (Neck, 1990), and a button factory operated briefly at Mercedes, Hidalgo County, Texas (Garrett, 1929). Pearl harvesters that seek Tampico pearlymussel in the Colorado and Brazos drainages, Texas (Howells, 1993, 1996c) appear not to focus similar harvest efforts on the Rio Grande. A survey of both resident and non-resident mussel li¬ cense holders in Texas (Howells, 1993) found no re¬ spondents indicating they took mussels from the Rio Grande. There is no indication that commercial har¬ vest for shells or pearls is, or has ever been, economi¬ cally important or contributed to the decline of the fauna locally. Summary All unionid species in the Rio Grande drainage basin have been dramatically reduced in abundance and distribution both historically and in recent years. Among the 1 6 freshwater mussel taxa reported in the system, two are likely extinct (Rio Grande monkeyface and false spike). Seven other taxa have not been docu¬ mented in recent years and are either extinct or re¬ duced to very low numbers (Conchos disk, Mexican fawnsfoot, yellow sandshell, washboard, lilliput, Texas lilliput, and pondhom (sp.?). Texas homshell is still extant in a short stretch of the Black River in southern New Mexico and a small relict population may also be present in the Rio Grande downstream of Big Bend. Similarly, Salina mucket may still survive in low num¬ bers just downriver of Big Bend. Tampico pearlymussel, paper pondshell, and introduced south¬ ern mapleleaf are currently maintaining populations in Texas waters, and a bleufer population introduced in Amistad Reservoir is presumed to be present as well. Among other bivalves, Asian clams are widely distrib¬ uted and even abundant in some areas. Fingernail clams are still present, but current species, abundance, and distribution are poorly defined. Unfortunately, eco¬ logical sensitivity of unionids, projected future devel¬ opment within the drainage, and general disinterest among regulatory authorities suggest an extremely dim future for this unique faunal group. Addendum Since this manuscript was drafted, unionid col¬ lections in 2002 in the Rio Grande in Webb County have included one living and several recently dead shells or valves each of washboard and Texas homshell, as well as several recently dead shells and valves of Mexi¬ can fawnsfoot. Additional survey efforts are planned to better define the status of these species in the Webb County area. Howells — Declining Status of Mussels in the Rio Grande 71 Literature Cited Bane, C. A. and O. T. Lind. 1978. The benthic invertebrate standing crop and diversity of a small desert stream in the Big Bend National Park, Texas. The Southwestern Natu¬ ralist, 23:215-226. Buckner, H. D., E. R. Carrillo, and H. J. Davidson. 1986. Water resources data. Texas. Water Year 1986. U.S. Geological Survey Water-Data Report TX-86-3, Austin, Texas. Cockerell, T. D. A. 1902. Unio popeii, Lea, in New Mexico. The Nautilus, 16:69-70. Conrad, T. A. 1855. Descriptions of three new species of Unio. Proceedings of the Academy of Natural Sciences of Phila¬ delphia, 7:256. Dali, W. H. 1896. Report on the mollusks collected by the International Boundary Commission of the United States and Mexico, 1 892- 1 894. Proceedings United States Na¬ tional Museum, 19:333-379, plates 31-33. Davis, J. R. 1980a. Species composition and diversity of benthic macroinvertebrates in the upper Rio Grande, Texas. The Southwestern Naturalist, 25:137-150. _ . 1980b. Species composition and diversity of benthic macroinvertebrate populations of the Pecos River, Texas. The Southwestern Naturalist, 25:241-256. _ . 1980c. Species composition and diversity of benthic macroinvertebrates of Lower Devil’s River, Texas. The Southwestern Naturalist, 25:379-384. Ellis, M. M., A. D. Merrick, and M. D. Ellis. 1930. The blood of North American fresh-water mussels under normal and adverse conditions. Bulletin of the U.S. Bureau of Fish¬ eries, 46:509-542. Frierson, L. S. 1 903. The specific value of Unio declivis. Say. The Nautilus, 17:49-51. Gandara, S. C., W. J. Gibbons, F. L. Andrews, R. E. Jones, and D. L. Barbie. 1996. Water resources data. Texas. Water Year 1996. U.S. Geological Survey Water-Data Report TX-96-3, Austin, TX. Gandara, S. C., W. J. Gibbons, D. L. Barbie, and R. E. Jones. 1999. Water resources data. Texas. Water Year 1999. U.S. Geological Survey Water-Data Report TX-99-5, Austin, TX. Gandara, S. C., W. J. Gibbons, and D. L. Barbie. 2001. Water resources data. Texas. Water Year 2000. U.S. Geological Survey Water-Data Report TX-00-05, Austin, TX. Garrett, B. 1929. Pearl buttons from Valley clams. Monty’s Monthly, 11(6):46-18. Gordon, M. E. 2000. Determination of glochidial hosts for the Texas homshell, Popenaias popeii (Lea, 1 857) (Mollusca: Bivalvia: Unionoidea). Final Report for New Mexico Department of Game and Fish, Contract 99-515-59, Al¬ buquerque. 6 pp. Hillis, D. M. and R. L. Mayden. 1985. Spread of the Asiatic clam Corbicula (Bivalvia: Corbiculidae) into the New World tropics. The Southwestern Naturalist, 30:454- 456. Hillis, D. M. and J. C. Patton. 1981. Morphological and elec¬ trophoretic evidence for two species of Corbicula (Bivalvia: Corbiculidae) in North America. American Midland Naturalist, 108:74-80. Howells, R. G 1993. Preliminary survey of freshwater mussel harvest in Texas. Texas Parks and Wildlife Department, Management Data Series 100, Austin. 30 pp. _ . 1 994a. Preliminary distributional surveys of freshwater bivalves in Texas: progress report for 1992. Texas Parks and Wildlife Department, Management Data Series 105, Austin. 16 pp. _ . 1994b. Mussel survey sites: July and August 1994. Info-Mussel Newsletter, 2(8): 1-5. _ . 1995. Distributional surveys of freshwater bivalves in Texas: progress report for 1993. Texas Parks and Wild¬ life Department, Management Data Series 119, Austin. 46 pp. _ . 1996a. Distributional surveys of freshwater bivalves in Texas: progress report for 1994 . Texas Parks and Wild¬ life Department, Management Data Series 120, Austin. 153 pp. _ . 1996b. Distributional surveys of freshwater bivalves in Texas: progress report for 1995. Texas Parks and Wild¬ life Department, Management Data Series 125, Austin. 41 pp. 1996c. Tampico pearlymussel ( Cyrtonaias tampicoensis ): shades of the Old West. American Con- chologist, 24(2):24-26. _ . 1997a. Distributional surveys of freshwater bivalves in Texas: progress report for 1996. Texas Parks and Wild¬ life Department, Management Data Series 144, Austin. 52 pp. _ . 1 997b. Taxonomic status of western lilliput ( Toxolasma mearnsi). Triannual Unionid Report, (11 ):32. _ . 1997c. Range extension for b\eufer Potamilus purpuratus (Bivalvia: Unionidae) in Texas. The Texas Journal of Science, 49:79-82. _ . 1997d. New fish hosts for nine freshwater mussels (Bivalvia: Unionidae) in Texas. The Texas Journal of Science, 49:225-258. _ . 1 998a. Distributional surveys of freshwater bivalves in Texas: progress report for 1997. Texas Parks and Wild¬ life Department, Management Data Series 147, Austin. 24 pp. 72 Special Publications, Museum of Texas Tech University _ . 1 998b. Freshwater mussels of the Rio Grande with special emphasis on the Big Bend Region. Prepared for a National Park Service training class, Texas Parks and Wild¬ life Department, Ingram. 19 pp. _ . 1998c. Freshwater mussel (Unionidae) shell remains. Pp. 164-175 and 324-338, in 41 MV1 20: A stratified Late Archaic site in Maverick County, Texas (B. J. Vierra). The University of Texas at San Antonio, Center for Ar¬ cheological Research, Report 25 1 . 347 pp. _ . 1 999. Distributional surveys of freshwater bivalves in Texas: progress report for 1998. Texas Parks and Wild¬ life Department, Management Data Series 161, Austin. 28 pp. _ . 2000. Distributional surveys of freshwater bivalves in Texas: progress report for 1999. Texas Parks and Wild¬ life Department, Management Data Series 170, Austin. 49 pp. _ . 200 1 a. Distributional surveys of freshwater bivalves in Texas: progress report for 2000. Texas Parks and Wild¬ life Department, Management Data Series 187, Austin. 43 pp. _ . 2001b. Status of freshwater mussels of the Rio Grande, with comments on other bivalves. Texas Parks and Wild¬ life Department, Austin. 81 pp. Howells, R. G and S. P. Ansley. 1999. Recent freshwater- mussel surveys of the Rio Grande in the Big Bend Re¬ gion. Triannual Unionid Report, (17):7. Howells, R. G and G P. Garrett. 1995. Freshwater mussel surveys of the Rio Grande tributaries in Chihuahua, Mexico. Triannual Unionid Report, (8): 10. Howells, R. G, C. M. Mather, and J. A. M. Bergmann. 1997. Conservation status of selected freshwater mussels in Texas. Pp. 1 17-129 in Conservation and management of freshwater mussels II: initiatives for the future (K. S. Cummings, A. C. Buchanan, C. A. Mayer, and T. J. Naimo, eds). Proceedings of a UMRCC symposium, St. Louis, Missouri, October 1995. Upper Mississippi River Con¬ servation Committee, Rock Island, Illinois. 293 pp. Howells, R. G, R. W. Neck, and H. D. Murray. 1996. Freshwa¬ ter mussels of Texas. Texas Parks and Wildlife Press, Austin. 218 pp. Johnson, R. I. 1998. A new mussel, Potamilus metnecktayi (Bivalvia: Unionidae) from the Rio Grande System, Mexico and Texas with notes on Mexican Disconaias. Occa¬ sional Papers on Mollusks, Museum of Comparative Zoology, 5(76):427-455. _ . 1999. Unionidae of the Rio Grande (Rio Bravo del Norte) System of Texas and Mexico. Occasional Papers on Mollusks, Museum of Comparative Zoology, 6(77): 1 - 65. Kelly, M. and C. Reed. 1 998. The Texas/Mexico border region. Borderlines, 6(3): 1-2. Lang, B. K. 2000. Status of the Texas homshell and other native freshwater mussels (Unionoidea) in the Rio Grande and Pecos River of New Mexico and Texas. New Mexico Department of Game and Fish, Performance Report E- 38-3, Santa Fe. Metcalf, A. L. Undated. Survey of the mollusks and larger crustaceans of the Rio Grande Valley, Alamito Creek, Presidio County, Texas and Esperanza, Hudspeth County, Texas. University of Texas at El Paso, Department of Biological Sciences. 42 pp. _ . 1966. Corbicula manilensis in the Mesilla Valley of Texas and Mexico. Nautilus, 80( 1 ): 1 6-20. _ . 1 974. Fossil and living freshwater mussels (Unionacea) from the Pecos River, New Mexico and Texas. Bulletin of the American Malacological Union, 1973:47-48. _ . 1982. Fossil Unionacean bivalves from three tributaries of the Rio Grande. Pp. 43-48 in Proceedings of the Sym¬ posium on recent Benthological investigations in Texas and adjacent states (J. R. Davis, ed.). Texas Academy of Science, Austin. 278 pp. Metcalf, A. L. and R. A. Smartt. 1972. Records of introduced mollusks: New Mexico and Texas. The Nautilus, 85:144- 145. Metcalf, A. L. and E. M. Stem. 1976. Notes on Unioniacean mollusks of the Rio Grande System, United States and Mexico. Bulletin of the American Malacological Union, Inc., 1976:42-42. MICRA (Mississippi Interstate Cooperative Resource Asso¬ ciation). 1995. Climate change. River Crossings, 4(5): 15. Morrison, J. P. E. 1976. Species in the genus Uniomerus. Bulletin of the American Malacological Union, 1 976: 1 0- 11. NPS (U.S. National Park Service). 2001. The Rio Grande: the desert’s lifeblood. NPS, Big Bend National Park, Big Bend National Park, Texas. 4 pp. Neck, R. W. 1 982. Preliminary analysis of the ecological zooge¬ ography of the freshwater mussels of Texas. Pp. 33-42 in Proceedings of the Symposium on recent Benthological investigations in Texas and adjacent states (J. R. Davis, ed.). Texas Academy of Science, Austin. 278 pp. _ . 1984. Restricted and declining nonmarine mollusks of Texas. Texas Parks and Wildlife Department, Technical Series 43, Austin. 17 pp. _ . 1987. Freshwater bivalves of the Baffin Bay drainage basin, southern Texas. The Texas Journal of Science, 39:177-182. _ . 1990. Button shells from Fort Brown, Texas. Appendix IV. Pp. 147 in Archeological investigations at Fort Brown (41CF96), Cameron County, Texas (S.B. Carlson, J. Saunders, F. Winchell, and B. Aiken). Archeological Re¬ search Lab, Texas A&M University, Report of Investiga¬ tions 1 1 , Brownsville, Texas. Howells — Declining Status of Mussels in the Rio Grande 73 Neck, R. W. and A. L. Metcalf. 1988. Freshwater bivalves of the Lower Rio Grande, Texas. The Texas Journal of Sci¬ ence, 40:259-268. Neves, R. L. 1993. A state of the unionids address. Pp. 1-lOz/j Conservation and management of freshwater mussels (K.S. Cummings, A.C. Buchanan, and L.M. Koch, eds.). Proceedings of a UMRCC Symposium, October 1992, St. Louis, Missouri. Upper Mississippi River Conser¬ vation Committee, Rock Island, Illinois. 189 pp. Simpson, C. T. 1914. A descriptive catalogue of the naiades, or pearly fresh-water mussels. B. Walker, Detroit, Michi¬ gan. 1540 pp. Smith, M. L. and R. R. Miller. 1986. The evolution of the Rio Grande basin as inferred from its fish fauna. Pp 457-485 in The zoogeography of North American fishes (C. Nocutt and E. Wiley, eds.). John Wiley and Sons, New York, New York. 866 pp. Strecker, J. K. 1931. The distribution of the naiades or pearly fresh-water mussels of Texas. Baylor University Mu¬ seum, Special Bulletin 2. 69 pp. Taylor, D. W. 1967. Freshwater mollusks collected by the United States and Mexican Boundary Surveys. The Ve- liger, 10(2): 152-158. _ . 1997. A new mussel, Disconaias conchos (Bivalvia: Unionidae) from Rio Conchos of the Rio Grande System, Mexico. Occasional Papers on Mollusks, Museum of Comparative Zoology, 5(75):4 19-425. Turgeon, D .D., J. F. Quinn, Jr., A. E. Bogan, E. V. Coan, F. G Hoehberg, W. G Lyons, P. M. Mikkelsen, R. J. Neves, C. F. E. Roper, G Rosenberg, B. Roth, A. Scheltema, F. G Thompson, and J. D. Williams. 1998. Common and scientific names of aquatic invertebrates from the United States and Canada: mollusks. American Fisheries Soci¬ ety, Special Publication 26, Bethesda, Maryland. 526 pp. USGS (U. S. Geological Survey). 1968. Water resources data for Texas. Part 2. Water quality records. U. S. Geological Survey, Austin, Texas. _ . 1975. Water resources data for Texas. Volume 3. U. S. Geological Survey Water-Data Report TX-75-1, Austin, Texas. Wilkins, K. T. 1 992. Mammalian paleofaunas of Central Texas from the late Wisconsinan glacial period to the latest Ho¬ locene. The Texas Journal of Science, 44:263-28 1 . Williams, J. W., M. L. Warren, Jr., K. S. Cummings, J. L. Harris, and R. J. Neves. 1993. Conservation status of freshwater mussels of the United States and Canada. Fisheries, 18(9):6-22. Address of author: Robert GL Howells Texas Parks and Wildlife Department Heart of the Hills Fisheries Science Center HC 7, Box 62 Ingram, Texas 78025 e-mail: robert.howells@tpwd. state. tx. us Fish Assemblages of the Rio Conchos Basin, Mexico, with Emphasis on Their Conservation and Status Robert J. Edwards, GaryP. Garrett, and Edie Marsh-Matthews Abstract The Chihuahuan Desert region contains a num¬ ber of unique aquatic environments, but with few ex¬ ceptions, these have been little studied. We sampled the Rio Conchos Basin in 1994 and 1995 to assess the status of the fishes of this region. Most sites showed some degree of human-induced impacts. A number of potentially threatened fishes were either abundant at only a few sites or rare or absent throughout the lo¬ calities sampled. Comparisons with collections taken during the 1950s indicate that the basic fish fauna is largely intact. However, there appears to be dimin¬ ished relative abundances of “large river” forms in fa¬ vor of non-natives (primarily an African cichlid, Oreochrornis aureus ) and “quiet-water” native fishes. This change seems related to decreased flows and regu¬ lated flow regimes from dams in the basin. Resumen La region del desierto de Chihuahuan contiene un numero de ambientes acuaticos unicos, pero con pocas anomalias, estos se han estudiado poco. Porque muchos de los pescados en la region se piensan para ser amenazados con la extincion o han ido extintos, muestreamos el lavabo de Rio Conchos en 1994 y 1995 para evaluar el estatus de los pescados de esta region. La mayoria de los sitios mostraron un cierto grado de impactos humano-inducidos. Un numero de pescados potencialmente amenazados eran o abundantes en solamente algunos sitios o raro o ausente a traves de los lugares muestreo. Las comparaciones de nuestros datos a las colecciones tomadas durante los anos 50 indican que mientras que la fauna basica de los pescados en la region es en gran parte intacta, aparece ser reducciones en formas del “no grande” y aumenta las formas del ambiente introducido de los pescados y de la “reservado-agua”. Estos cambios aparecen relacionados a disminuido y los regimenes regulados del flujo. Introduction The limited aquatic habitats of the Chihuahuan Desert have undergone substantial anthropogenic modi¬ fication in the last hundred years, including reduced water quality, diversion of surface water, overdrafting of groundwater, channelization, impoundment, and introduction of non-native species (Miller and Chemoff, 1979; Propst and Stefferud, 1994; TNRCC, 1994; IBWC, 1994; Lee and Wilson, 1997; Edwards et al., 2001). Impacts from these modifications are only now being documented and few baseline data exist concerning the ecological requirements for most of the aquatic species. Approximately half of the native fishes of the Chihuahuan Desert are threatened with extinction or are extinct (Hubbs, 1990). Documented extinctions from the northern Chihuahuan Desert include Maravillas red shiner ( Cyprinella lutrensis blairi), phantom shiner ( Notropis orca), Rio Grande bluntnose shiner ( Notropis simus simus), and Amistad gambusia ( Gambusia amistadensis ) (Miller et al., 1989; Hubbs, this volume). Some noteworthy extirpations include Rio Grande shiner ( Notropis jemezanus ) from the New Mexico portion of the Rio Grande (Propst et al., 1987) and Rio 76 Special Publications, Museum of Texas Tech University Grande silvery minnow ( Hybognathus amarus), Rio Grande cutthroat trout ( Oncorhynchus clarki virginalis) and blotched gambusia ( Gambusia senilis) in Texas (Bestgen and Platania, 1990, 1991; Hubbs et al., 1991) as well as others. The status of a number of fishes in the northern Chihuahuan Desert is poorly understood, particularly for the Mexican portion of their ranges. Endemic species other than fishes also are being lost from this area (Howells and Garrett, 1995; Howells, this volume; Lang et al., this volume). In this paper, we present data on fish collections from 14 localities in the Rio Conchos Basin in 1994 and 1995. We compare our results with a series of collections taken from the basin nearly 40 years ear¬ lier, and we comment on the status of several imper¬ iled fishes for which the Rio Conchos Basin repre¬ sents a significant portion of their geographic distribu¬ tion. Study Area The Rio Conchos receives its water from a se¬ ries of tributaries originating in the Sierra Madre Occi¬ dental in Chihuahua, Mexico, along with a number of springs, seasonal rains and periodic tropical storms (Tamayo and West, 1964). The climate ranges from subhumid and temperate in the Sierra Madre Occiden¬ tal to semiarid and warm in the central Plateau and warm and arid in its northern-most reaches; tempera¬ tures often exceed 40°C and precipitation averages about 315 mm/year, with greater amounts in the moun¬ tain areas and lesser amounts in the central and north¬ ern portions of the state (Kelly, 2001). The Mexican state of Chihuahua has over two million people, with 80% living in small to medium sized towns and the remainder centered in the cities of Juarez and Ciudad Chihuahua. Anthropogenic impacts range from de¬ forestation and mining in the Sierra Madre, impound¬ ments on both large and small streams throughout the region, manufacturing in the cities, and agriculture throughout the central Plateau, especially surrounding Ciudad Chihuahua (Comision National del Agua, 1997). Only the largest cities have wastewater treatment plants and most rural areas lack even basic sewage collec¬ tion and disinfection facilities (Kelly, 2001 ). Some major streams (for example, the Rio Florido) are severely impacted with high levels of oil, fecal coliform bacte¬ ria, discharges from chemical plants, and pollution from agricultural return flows (Comision Nacional del Agua, 1997). The Rio Conchos is the primary source of water for the Rio Grande downstream of Presidio, Texas. The average annual flow of the Rio Grande immediately upstream of its confluence with the Rio Conchos is approximately 72 thousand acre-feet (8.9 x 107 m3), whereas the Rio Conchos contributes an annual average of 779 thousand acre-feet (9.6 x 108 m3), far exceeding the input from any other tributary of the Rio Grande System (IBWC, 1990; Eaton and Hurlburt, 1992; TNRCC, 1994). Materials and Methods We collected fishes from 14 locations (combined into 1 1 stations) throughout the Rio Conchos System of Chihuahua (Figure 1). The Rio Parral was not sampled because locals advised us that the waters were too polluted to even safely wade in. Our station desig¬ nations, specific sampling localities and sampling dates were as follows: Station 1, springs at Ojo Talamantes, 16 km NE of Allende, 6 August 1994; Station 2, the Rio Conchos at Valle de Zaragosa, 6 August 1994; Sta¬ tion 3, Rio San Pedro S of Satevo at Highway 24 cross¬ ing, 6 August 1994; Station 4, Rio Santa Isabela, 20 km downstream from Riva Palacio, 7 August 1994; Station 5, Rio Santa Isabela immediately upstream from Riva Palacio, 7 August 1994; Station 6, Rio Conchos, immediately downstream from Julimes, 5 August 1994; Station 7, a series of three collections at the springs and outflows at San Diego de Alcala, 24 October 1995; Stations 8a and 8b, Rio Chuviscar near San Diego de Alcala, 4 August 1994 (Station 8a) and 24 October 1995 (Station 8b), respectively, during and after con¬ struction of a water pipe to draw water from the stream; Station 9, headwaters of Rio Chuviscar, near High¬ way 160, approximately 15 km S of Namiquipa, 23 October 1995; Station 10, Rio Conchos at Cuchillo Edwards et al. — Fish Assemblages of the Rio Conchos Basin, Mexico 77 Figure 1 . Map of Rio Conchos study area. Numerals indicate most current collection sites and letters indicate comparative historical collection sites. See text for collections localities. Parado, about 30 km from the confluence with the Rio Grande, 3 August 1994; and, Station 11, springs at Ojo del Arrey, approximately 4 km NE of Angostura, 10 August 1994. The Rio Florido was not sampled because it was dry or too seriously polluted through¬ out much of its course. We intensively sampled contiguous segments of habitats with seemingly pristine conditions so as to represent, to the greatest degree possible, the natural biota. We selected multiple sampling sites at each lo¬ cation and collected at each site until we detected no additional changes in species occurrence and relative abundances in the sample. In general, all available microhabitats were sampled roughly in proportion to their occurrence. Sites were sampled with seines 3 m (3 mm mesh) to 10 m (6 mm mesh) long or electrofishing in all available habitats. At most sites waters were too shallow for effective electrofishing, and seining was the major method. At each location, all specimens collected were identified and enumer¬ ated. A representative subsample of each species (ex¬ cept those not allowed by permit) was retained. Our data are presented as relative abundances to facilitate comparisons. We compared our data to museum col¬ lection records (Texas Natural History Collections, The University of Texas at Austin) of a series of collec¬ tions taken by Clark Hubbs and Victor Springer in June and July 1954 and by Hubbs and Oscar Wiegand in December 1954 from similar areas (and using seines) throughout the Rio Conchos basin. We collected in all available habitats within a fewer number of larger sites with generally three to six teams of collectors at each station, whereas the 1954 collectors sampled at a greater number of smaller habitats, in part, because only two people were sampling (C. Hubbs, University of Texas at Austin, pers. comm.). We combined a number of the 1954 sampling stations into a smaller series of stations in order to be more comparable to our collection data. The stations created for the 1954 comparisons (Figure 1) were as follows. Station a, Rio Florido, at Highway 45 crossing, 17.6 km ESE of 78 Special Publications, Museum of Texas Tech University Villa Ocampo, 27 June 1954; Rio Florido at Espirito Santo, 8 km ESE of Villa Ocampo, 27 June 1954, a tributary of the Rio Florido, 24.6 km ESE of Parral, 27 June 1954, and the Rio Florido 0.8 km W of Villa Ocampo, 27 June 1954; Station b, Rio Florido at Guadalupe, 22.4 km E of Parral, 26 June 1954; Sta¬ tion c, a series of collections at the springs and associ¬ ated outflows near the Ojo de la Hacienda Delores in¬ cluding the Rio Valle de Allende, at Valle de Allende, 27 June 1954, El Ojo de la Hacienda Dolores, 8 km S of Jimenez and 3.2 km SE of Dolores, 30 June 1954, an irrigation ditch draining El Ojo de la Hacienda Dolores, 30 June 1954, El Ojo Almoloya, 3.2 km WofEstacion Troya, 30 June 1954, an irrigation ditch at Highway 45, 1.6 km SE of Bachimba, 1 July 1954, Rio Parral, 3.2 km W of Parral and approximately 0.5 km W of the railroad bridge, 26 June 1954, Rio Valle de Allende, 1.6 km W of Valle de Allende at the small dam, 30 December 1954, a ditch near El Ojo Almoloya, 2.4 km W of Troya, 31 December 1954 in an irrigation ditch near Ojo Hacienda Dolores at points 3.2 and 6.4 km S of Jimenez, 31 December 1954, El Ojo de La Haci¬ enda Dolores, 8 km S of Jimenez and 3.2 km SE of Dolores, 3 1 December 1954, and Ojo Hacienda Dolores, 8 km S of Jimenez, 31 December 1954; Station d, Ojo de San Gregorio, 0.5 km W of San Gregorio and 19.2 km ENE of Parral, 31 December 1954; Station e, Rio Conchos at Camargo, 25 June 1954, Rio Conchos at La Cruz (several different sites), 25 June 1954, Rio Conchos, 1.6 km N of Saucillo, 28 June 1954, Rio Conchos at highway crossing, 19.2 km SW of Camargo, a tributary of the Rio Conchos, 1.6 km E of San Francisco de Condios and 24 km SW of Camargo, both on 28 June 1954; Station f, Rio Conchos at Saucillo on the E channel at ford, 25 June 1954; Sta¬ tion g, Rio San Pedro, 1.6 km SW of Meoqui, 24 June 1954, Rio San Pedro at Meoqui, 30 December 1954; Station h, Rio Conchos at Julimes, 24 June 1954; Sta¬ tion j, was the Rio San Pedro, 0.4 km SW of confluence with Rio Conchos, 24 June 1954, Rio Sacramento 1.6 km N of Ciudad Chihuahua, 1 July 1954; Station k, Rio Conchos, 32 km W of Rio Grande confluence, 14 June 1954; Station 1, Rio Conchos, 1 km. upstream of confluence with Rio Grande, 13 June 1954, Rio Conchos, 6.4 km W of Rio Grande confluence at Si¬ erras Navas, 14 June 1954; Station m, Rio del Carmen at El Carmen, 1 July 1954; Station n, Rio Santa Maria at Buenaventura, 2 July 1954. Results 1994-1995 Collections Our collections yielded 18,371 specimens repre¬ senting 37 species (Table 1). Streams within the Rio Conchos Basin are characterized by a relatively large minnow (Cyprinidae) component that includes Cyprinella lutrensis, Notropis chihuahua, Notropis braytoni, Notropis jemezanus, Campostoma ornatum, Macrhybopsis aestivalis, and Rhinichthys cataractae. Other typical species were Scartomyzon austrinus, Ictalurus lupus , Astyanax mexicanus, Cyprinodon eximius, and Gambusia senilis. Several wide-ranging species were found, including Dorosoma cepedianum, Pimephales promelas, Ictalurus furcatus, Pylodictis olivaris, Gambusia speciosa, Lepomis cyanellus, L. megalotis, L. macrochirus, and Microptems salmoides. As expected, mainstem localities contained more spe¬ cies and tributary streams contained a subset of the total species complement. The springs at Ojo de Talamantes (Station 1) and those at Ojo del Arrey (Station 1 1) were quite different from the other localities sampled. At the former site, an undescribed species of Gambusia accounted for over 70% of the fishes captured, while at the latter site, G. speciosa and an undescribed species of Cyprinodon accounted for more than 95% of the fishes captured. Each of these springs has been substan¬ tially modified. The springs at Talamantes have been transformed into an aquatic tourist park and impounded with a low concrete dam and the springs at Ojo del Arrey have been developed into a swimming pool, with a small outflow spring run. The headwaters of the Rio Chuviscar (Station 9), the Rio San Pedro (Station 3) and the Rio Chuviscar at San Diego de Alcala (Station 8b) all contained large Edwards et al. — Fish Assemblages of the Rio Conchos Basin, Mexico 79 S -3 5 C5 3 a o • **»* a 50 Os Os i >»• On Os »'«( a <3 Q £ a O O <2 so M- ‘n Os •"-I t>0 a • *»»* a a, a. •v^ 8 .a a 3 ^ ^ .SP O k, £- o ^ 3 $ o • «« • c< ~a "3 £ a o &a ■ a 3^ <3 s«a 3 "3 st- •o Os so a o • o "o o . 3 o <4) -a “a <-> <4) fts e5 S' as x> oo . O' . °° . ™ C3 (N - 00 1 1 — 1 ^11111111111 1 1 1 1 m 1 1 \ 1 - d d loll 1 o 1 1 o d O ! a) On ON cn ■*t 1 1 1 ~“ V ■ 1 <1 1 ^ 1 1 - I | 1 I [ | I so I I 1 d O s 6 1 1 1 a- 1 o 1 o o 1 1 d >n 1 o 1 1 ^ 1 *n 1 m in o I — a) oo . 3) . O' 00 so os cn ai oo OO a) . ^ m Tt , 'O o ra - - sd co rs o o — * => r' <=> 1 1 d 1 1 o 1 1 o co d d sd o 1 1 o d 1 o d 1 1 o ci d o 3 .so 3 5 a a. ■a o' -k 43 It CD ^ 3 a a -« 3 a § 3 3^ a -a -3 3 so to S. 5. c> c> d 4s o o So so -5 <4> a a 3 1 1 <4) 5 a 14) 3 •»** 3 2 3 So 3 .9 ■§ So 3 It a §J o ■§ ^ 0, 3. a vj C" O 0 3^ S, O So 2, so 3 ■5 *§ 5 -ago •9 -2 tj ■S S- 2 •a d 3 k U s. 3 i a k S sT 3, 3 > ■V* *t a So -C> a a a a Sr 3 -a ■§ -§ ■§ N co 3 *»«» 3 *«* •2 •! 3 a a a a 3, 3. 3, 3 s •S ^ a ~3 3. a CO Gq Co 3 3 3 ~C) *>C) -O S 5 5 3 3 3 0^0 ^ 3. 3. so So S) 3 .3 .3 s? sj 5 3 3 3 -a -a -a 5 S S 3 3 3 O O CD 80 Special Publications, Museum of Texas Tech University vs so o — — o © —I c- 00 o SO 00 -ct vs — Os — o — — Tf CN (N O O — — — m © © © o oo a a 3 r x 3 cj 3 C) -2- © ■3 3 © X 1 2 g3 3 5 £ r~ o - — — SO N0q o p « s Q. 3 00 — U Z Oh U (N O c _o — . — J~. O C/1 ‘5 o a. oo oo r-' . ^ m (N vs rn uS 1 (N 1 | [ (N 3 s. -3 X .“0 ,co O O ■5 Co 3 - 3 CD 3 <3 N CD s CD 3 CD 3 3 a 5 o 8, 8 ?> s? ' Q -S X X to © O ■§ 2- CS o d fs\ I*- rv ^ .§ ^ ^ a, ft; 2- ® r13 r3 U s 5 © ■v* 8 Sv 3 5 «o a §■ 3 Q Cj co a, 03 ^ a 8 «5 ^ c, ? 3 3 3 3 3 • 3 3 3 CD "3 ~3 3 a a ^ ■**•» X ^ ^ Edwards et al. — Fish Assemblages of the Rio Conchos Basin, Mexico 81 in o Tf t-~ t'- 00 SO — O (N SO o © OO © c in m as °® — o d 00 (N . o oc r- ni 1 so £ T3 U « T! I- u s 3 O O. 3 CO 0.-0 O Vi O Vh iO o c C/5 c 1) o JO E *— ___ Vl^ 00 2 Oh C/5 .22 © Oh >> <>5 o 05 3 •*«* co O .Co 6- 5 a a a Hs S' 5 S3 a c o a 5 o ■»-* n in Tt co — < ni n O O Tf <0n co , r- . ^ CN . °o . 00 -pi P r' — 1 m | | | 1 1 1 1 1 5 O C Co O Q 5o to O -Si & £ S 3 "Q §• §- 5§ -2 N N J £ C o a -a a ■»— o o -5 -a: a a ^ ^ a. ^ U O to 3 -C> 3 <1 to 3 -O O a a 3 ^ a a -2 bo ^ -C> C ^ s -2^ 5) ti (o to 3 3 3 to Co JiT O O 5j e c a a a -O 53 52 5 I t ~3 ^ 5J a ^ a o $> 5 ^3 53 o' &0 S .Co .Co .Co 5 5 5 000 O, sa. Ci, 5T -J O O co 52 .to O ■3 3 53 -5 1 - g>3 c\ Co a a §• N- 3 3 bO c, to 5 o 3 cx. -Si 3 ^ 3 ■*-« o - a -5 o -3 | 51 O N cx ^ Lq x O (N m o o S 0 m r- o -00^ >n — ns = u « 3 5J 5J D. D 3 eo 0.-0 cj co O id tc; w C O C U l- C u 3! Edwards et al. — Fish Assemblages of the Rio Conchos Basin, Mexico 83 Gambusia senilis populations but with relatively few additional species, a situation that is typical of head¬ water creeks or degraded and polluted conditions. The species diversity was higher at the Rio San Pedro site than the other two localities; however, all were small streams dominated by G. senilis. The lower reaches of Rio Conchos tributaries (Stations 4, 5, 7, and 8a) had relatively large numbers of N.jemezanus, G. senilis, and usually Campostoma ornatum. These sites differed from each other in that the Rio Santa Isabella localities also contained large numbers of Codoma ornata and our only captures of Gila pulchra , whereas the two sites near San Diego de Alcala had relatively large numbers of Dionda episcopa and either C. eximius (Station 8a) or an undescribed species of Cyprinodon (Station 7). The hot spring outflows at the San Diego de Alcala loca¬ tion (Station 7) contained our largest collection of Lepomis macrochirus , an introduced species account¬ ing for 18% of the fish collected at that site. The Rio Conchos immediately above Presa de Boquilla (Station 2) had relatively large numbers of C. lutrensis, N. chihuahua, C. ornatum and G. senilis and this area also contained P. promelas, A. mexicanus, L. cyanellus, and the introduced Oreochromis aureus. Although this area has been impacted greatly by ur¬ banization and influences from the reservoir, introduced species were less abundant than at some of the other localities. The downstream and middle Rio Conchos sta¬ tions (6 and 10) were heavily impacted by channelization and degraded water quality from agri¬ cultural inputs. The collections were dominated by C. lutrensis, which accounted for a third to about 60% of the total fish captures at these sites. Also present were the introduced Menidia beryllina and some of the more typical Chihuahuan Desert fishes, such as N. chihuahua, N. braytoni, Scartomyzon austrinus and several catfish, including Ictalurus furcatus, I. lupus, I. punctatus, and Pylodictis olivaris. These sites also produced large numbers of the introduced Oreochromis aureus, especially the site near Julimes (Station 6), which was dominated by this species. Comparison with 1954 Data To contrast changes in fish community compo¬ sition that have occurred in the half-century since the 1954 collections were taken, we summarized collec¬ tions from Rio Conchos stations, Rio San Pedro sta¬ tions and Rio Florido stations taken during the 1954 series of collections and our present samples. The high degree of overall similarity in species occurrences indicates that the fish fauna in the Rio Conchos basin is still intact (Table 2). Some notable changes have occurred. Introduced fishes have long been known from the Rio Conchos. For example, the common carp ( Cyprinus carpio) was present in the early col¬ lections. However, blue tilapia ( Oreochromis aureus ) were not in the system in the 1950s, but are now widely found throughout and dominate the fish assemblages at some localities. There appears to be a change in other elements of the fish assemblages, as well, possi¬ bly in response to lessened water flows in the basin. There appears to be a loss of minnow diversity (10 species commonly found in 1954 versus 5 in the present study) and a diminution of species commonly found in large river systems ( Lepisosteus osseus, Notropis braytoni, N. jemezanus, Cycleptus elongatus, Aplodinotus grunniens, Micropterus salmoides, Lepomis macrochirus, and L. megalotis). In contrast, there appear to have been increases in smaller stream forms such as Cyprinodon eximius and Gambusia senilis both of which are indicative of diminishing and regu¬ lated flows. As these latter two species are of conser¬ vation concern, their increased populations in our present collections could be tenuous. Further declines in streamflow could negatively impact these species. The change in the fish communities of the Rio Florido is dramatic. The fish assemblage in this stream was quite similar to the mainstem Rio Conchos in 1954, but in our survey the river was dry. Our two samples from the Rio Chuviscar (Sta¬ tions 8a and 8b) in 1994 and 1995, indicate how swiftly noticeable changes can occur in fish assemblages. During our initial sampling, workers were laying a water pipe as part of a system to pump water out of the 84 Special Publications, Museum of Texas Tech University Table 2. Comparison of changes in fish communities between 1954 and 1994-1995 for selected streams in the Rio Conchos basin. An asterisk (*) indicates that the stream was dry. Rio Conchos Rio San Pedro Rio Florido Species 1954 1994-95 1954 1994-95 1954 1994-95 Lepisosteus osseus 0.60 - - - - - - * Dorosoma cepedianum 1.02 — — - — — * Campostoma ornatum — 1.04 0.39 14.42 8.97 * Codoma ornata - - 0.18 1.46 20.29 8.64 ♦ Cyprinella lutrensis 9.47 3.31 0.39 — 22.69 ♦ Cyprinus carpio 0.81 — > — — — * Dionda episcopa 6.21 7.23 3.51 — 13.58 ♦ Gila pulchra — — — - 6.01 3.30 * Macrhybopsis aestivalis 1.78 — - — - — — * Notropis braytoni 9.10 — — - — — * Notropis chihuahua 8.79 14.24 4.07 40.99 5.99 * Notropis jemezanus 15.21 — • — - 0.40 0.35 ★ Pimephales promelas 9.77 0.26 3.02 2.40 8.38 * Rhinichthys cataractae 0.12 — — - — ■ 2.21 • * Carpiodes carpio 1.12 — — — 2.40 * Catostomus conchos — 0.04 — . — 2.84 * Cycleptus elongatus 0.24 — — ■ — - — * Ictiobus bubalus — 0.02 — - - — ★ Scartomyzon austrinus 0.43 0.04 — — — * Astyanax mexicanus 7.74 3.90 1.46 — 4.97 * Ameiurus melas — — — 0.13 — * Ictalurus furcatus 0.28 — • — — — * Ictalurus lupus — - 0.02 — — • — - * Ictalurus punctatus — — 0.39 — — - * Pylodictis olivaris 0.40 — — — — * Cyprinodon eximius 1.67 15.46 17.25 — 0.18 ♦ Gambusia speciosa 1.42 — — — — ♦ Gambusia senilis 4.19 26.97 57.06 10.95 13.62 ♦ Gambusia sp. 2 — 23.96 — - - — * Lepomis cyanellus 0.32 0.02 — — - — * Lepomis macrochirus 9.89 0.18 1.16 — - - * Lepomis megalotis 2.58 0.02 1.16 2.00 1.88 * Micropterus salmoides 6.57 2.22 8.67 — — ★ Etheostoma australe/pottsi 0.23 — — 2.40 — ★ Aplodinotus grunniens 0.04 — — — — * Oreochromis aureus — 0.89 — — - — * Number of collections 13 3 3 1 5 * Number of species 26 19 13 10 15 0 stream. A year later, the pump had been installed and the stream was visibly altered, showing numerous ef¬ fects of the construction activity. Relative abundances of C. ornatum, D. episcopa, N. chihuahua, and C. eximius were lower and those for C. lutrensis and G. senilis were notably higher during the second visit. Pimephales promelas, A. mexicanus, L. cyanellus and O. aureus also had somewhat increased relative abun¬ dances following the perturbation. The species show¬ ing increased relative abundances are characteristic of either highly degraded areas or small headwater streams and generally are considered to be more tolerant of extreme environmental conditions. Edwards et al. — Fish Assemblages of the Rio Conchos Basin, Mexico 85 Status of Species A number of fishes inhabiting the Chihuahuan Desert region have been proposed for listing as endan¬ gered or threatened species of the U.S. or Mexico. Based on our results we provide additional observa¬ tions on the status of some of these species. The Mexican stoneroller ( Campostoma ornatum ) occurs in numerous localities in the Big Bend region of Texas and northwestern Mexico, including the rios Conchos, del Fuerte, Casas Grandes, del Carmen, Yaqui, Papigochic, Sonora, Nazas, Piaxtla and Trujillo (Burr, 1976). Although it occurs throughout the Rio Conchos basin, we only found it abundant in the Rio Santa Isabella. Some populations are seemingly ephem¬ eral, particularly in highly impacted habitats such as Rio Chuviscar. In 1994, this species had a relative abundance of 2.4%, but in 1995 no specimens were obtained. Contreras-B. (1977) reported it extirpated from the Rio Chihuahua (= Chuviscar) and the Rio Conchos at Camargo, citing the loss of well-oxygen¬ ated, clear, moving water flowing over sand and gravel bottoms due to lowered water tables, siltation and sew¬ age effluent. Our Rio Conchos sample at Julimes was downstream of Camargo and we did not obtain C. ornatum. However, our Rio Conchos sample at Valle de Zaragosa is upstream of Camargo and there we obtained 31 specimens. Our collections support a threatened status for this species and agree with many of the designations and reasons for this status given by various governmental agencies and researchers (Miller, 1972; Williams etal., 1989;Hubbs etal., 1991; Texas Organization for Endangered Species, 1995; CONAB IO, 1997). The Chihuahua shiner ( Notropis chihuahua) is listed as Threatened by the Texas Parks and Wildlife Department, Miller (1972), Hubbs et al. (1991) and the Texas Organization for Endangered Species (1995). Our findings agree with Burr’s (1980) assessment that the species occurs sporadically in Texas in the Big Bend region of the Rio Grande, but is currently abun¬ dant in tributaries of the Rio Conchos. Previous find¬ ings from studies in the Big Bend region of the Rio Grande range from absence of the species (Platania, 1990; IBWC, 1994) to a relative abundance of less than approximately 1% (Hubbs and Wauer, 1973; Hubbs et al., 1977; Bestgen and Platania, 1988). The Rio Grande shiner ( Notropis jemezanus) is listed as Rare by Mexico (CONABIO, 1997), Threat¬ ened by the Texas Organization for Endangered Spe¬ cies (1995), Threatened by Hubbs et al. (1991) and Special Concern by Williams et al. (1989). The his¬ toric range included the Rio Grande, Pecos River (New Mexico and Texas), and rios Conchos, San Juan and Salado drainages of Mexico (Gilbert, 1980). Hubbs (1940) noted that the species was “characteristic of the Rio Grande and its tributaries in New Mexico, Texas and northeastern Mexico” and Trevino-Robinson (1959) found the species throughout the middle Rio Grande of the Texas-Mexico borderlands, almost to the mouth of the Rio Grande during her studies in the 1950s. However, the range of N. jemezanus in the Rio Grande and Pecos River has declined dramatically, (Edwards and Contreras-B., 1991; Hubbs et al., 1991; Edwards et al., 2001). This species is part of a main¬ stream Rio Grande-Rio Conchos faunal assemblage that is not dependent on tributaries (Hubbs et al., 1977). It is typically found in large, open rivers over sand and gravel (Gilbert, 1980) where current flows keep the substrate clean from accumulated silt. Our collec¬ tions support the Threatened status designation given this species by Hubbs et al. (1991). The headwater catfish ( Ictalurus lupus) is listed as Rare by Mexico (CONABIO, 1997), Watch List by the Texas Organization for Endangered Species (1995), and Special Concern by Williams et al. (1989) and Hubbs et al. (1991). In our study we found this little known species only in the Rio San Pedro and the Rio Conchos at Cuchillo Parado, Julimes and Zaragosa, where it was always in low abundance. These results support the Rare (and Watch List) status designations previously given to this species. Hubbs et al. (1991) listed the undescribed Chi¬ huahua catfish (Ictalurus sp.) as Special Concern. Very little is known about this very cryptic and rare species and none were obtained in our collections. It occurred historically in the Rio Grande basin of New Mexico and Texas, the Rio Conchos basin in Chihuahua and the Rio San Fernando in Tamaulipas. The absence of this species in our samples may indicate that the spe¬ cies is in greater danger of extinction than generally understood. 86 Special Publications, Museum of Texas Tech University The Conchos pupfish ( Cyprinodon eximius) is listed as Threatened by the Texas Parks and Wildlife Department, the Texas Organization for Endangered Species (1995), Williams et al. (1989), Hubbs et al. ( 1 99 1 ), and Mexico (CON ABIO, 1 997). Historically this species was widely distributed, occurring in the upper Rio Conchos and Rio Sauz in Chihuahua and Alamito, Terlingua and Tomillo creeks and Devils River in Texas (Miller, 1976, 1981; Hubbs and Echelle, 1973; Minckley, 1980; Hubbs et al., 1991). The population in Dolan Creek, a tributary of the Devils River, was extirpated in 1958 and successfully reestablished in 1979 (Garrett, 1980; Hubbs and Garrett, 1990; Garrett et al., 1992). In our surveys, the species was abun¬ dant in the Rio Chuviscar and occurred in low num¬ bers in headwater streams and tributaries of the Rio Conchos. The population in the Rio Sauz Basin may have been extirpated (Echelle et al., this volume). Tire blotched gambusia ( Gambusia senilis) is listed as Threatened by Miller (1972), Mexico (CONABIO, 1997), and Texas Parks and Wildlife Department, Spe¬ cial Concern by Williams et al. (1989) and Extirpated in Texas by Hubbs et al. (1991) and the Texas Organi¬ zation for Endangered Species (1995). The historic range of the blotched gambusia includes the Rio Conchos Basin and Devils River (Hubbs, 1958; Guillory, 1980). Although Hubbs and Springer (1957) reported its range as the Rio Conchos downstream as far as Julimes, our collections at Julimes contained no G. senilis. However, an abundant population was present farther downstream in the Rio Chuviscar and the spe¬ cies almost completely dominates the fish community in the headwaters northwest of Ciudad Chihuahua. In general, we found G. senilis abundant and widely dis¬ tributed in our Mexican samples. The Texas popula¬ tion was isolated by Amistad Reservoir and ultimately eliminated (Hubbs and Echelle, 1973; Hubbs et al., 1991). The Rio Grande Fishes Recovery Team has recommended reestablishment of the Texas popula¬ tion in Devils River State Natural Area from stocks in the Rio Chuviscar. A Threatened status seems appro¬ priate for the existing populations of this species. Discussion Desert ecosystems are easily perturbed and of¬ ten slow to recover. Entrenchment of streams from erosion due to overgrazing and deforestation (Ohmart and Anderson, 1982), introductions of exotic species, and extinction of native species may all cause perma¬ nent damage to these systems. While other perturba¬ tions such as pollution, reduced groundwater, and dam construction are theoretically recoverable, the return to a pristine state is unlikely. Anthropogenic changes in the Rio Conchos ba¬ sin have been going on since the mid- 1800s (Miller, 1961, 1977) but the effects have been compounded over time and are now becoming dramatic. Our sur¬ vey indicates detrimental impacts on the fish assem¬ blages of the Rio Conchos in the past 40 years. In the early part of the 20th century it was apparent that wa¬ ter was becoming a major problem in Chihuahua as extensive irrigation projects were initiated (Tamayo and West, 1964). Brand (1937) noted for northwestern Chihuahua that “the increasing use of spring and river water for irrigation on the haciendas and colonias of the region has contributed markedly to the lessened flow of the rivers in their lower courses.” At least 30 springs have gone dry in Chihuahua and Coahuila and river discharges of the Rio Nazas, Bolson Mayran, Rio Aguanaval, Bolson Viesca, Rio de Nadadores, Rio Saltillo, Rio Salinas, Rio del Carmen and the middle Rio Grande are reduced (Contreras-B. and Lozano-V., 1994). Under these conditions of decreased flow, droughts are even more devastating because of in¬ creased groundwater pumping for agricultural and municipal use. Such extreme conditions favor more tolerant species often at the expense of less widespread species. Tributary creeks tend to be impacted more severely, yet these areas are critical to the breeding and rearing of young of many of the indigenous spe¬ cies including C. ornatum and N. chihuahua in the Rio Grande (Hubbs and Wauer, 1973). One factor noticed during our collections in the Rio Conchos Basin was the great efforts toward mod¬ ernization of the infrastructure in Chihuahua including its highways and municipal facilities. Few areas have escaped this leap toward modernization. However, the magnitude of change for the natural aquatic sys- Edwards et al. — Fish Assemblages of the Rio Conchgs Basin, Mexico 87 terns is sometimes quite dramatic. Many springs, es¬ pecially those located near human habitation, have been modified or are being modified into swimming areas or spas. In the larger municipalities there is a reason¬ ably strong message of water conservation, but the surrounding countryside shows many signs of increas¬ ing use of water consumptive measures such as flood irrigation and spray-water delivery systems. The Rio Conchos Basin has been impacted by centuries of human habitation. Exploitation of limited resources, particularly groundwater pumping, has de¬ graded that environment, caused extirpation and ex¬ tinction of species and, ultimately, loss of habitat and ecosystems (Smith and Miller, 1985). We suspect that the fish assemblages of this region are indicators of the overall integrity of the ecosystem. The few re¬ maining relatively pristine localities need careful man¬ agement if they are to be preserved. Acknowledgments We appreciate the able assistance of Nora Valdes and the Valdes family, T. Castillo, D. Cobb, W. Cobb, L. Garcia, L. Garrett, J. Hernandez, V. Hodges, C. Hubbs, J. Marquart, D. Riskind, A. Sudyka, D. VanMeter, B. Wienecke, and D. Wilson in helping with the logistics for this study and with the field collec¬ tions. We thank the government of Mexico for pro¬ viding us with our collecting permit. Partial funding was provided by the U.S. Fish and Wildlife Service and the Texas Parks and Wildlife Department and we thank them for their contributions. We thank J. Ortega for translating our abstract into Spanish. Finally, this manuscript benefited greatly from suggestions of anonymous reviewers. Literature Cited Bestgen, K. R. and S. P. Platania. 1988. The ichthyofauna and aquatic habitats of the Rio Grande from the New Mexico- Texas border to Big Bend National Park. Final Report to the U.S. Fish and Wildlife Service, Office of Endangered Species, Albuquerque, New Mexico. 55 pp. _ . 1 990. Extirpation of Notropis simus simus (Cope) and Notropis orca Woolman (Pisces: Cyprinidae) from the Rio Grande in New Mexico, with notes on their life his¬ tory. Occasional Papers, Museum of Southwestern Biol¬ ogy, 6: 1-8. _ . 1991. Status and conservation of the Rio Grande silvery minnow, Hybognathus amarus. The Southwestern Natu¬ ralist, 36:225-232. Brand D. D. 1937. The natural landscape of northwestern Chihuahua. University of New Mexico Bulletin Geologi¬ cal Series, 5:1-74. Burr, B. M. 1976. A review of the Mexican stoneroller, Campostoma ornatum Girard (Pisces: Cyprinidae). Transactions of the San Diego Society of Natural His¬ tory, 18:127-144. _ . 1 980. Notropis chihuahua Woolman. Pp. 25 1 , in Atlas of North American Freshwater Fishes (D. S. Lee et al., eds.), North Carolina State Museum of Natural History, Raleigh. Burr, B. M. and R. L. Mayden. 1981. Systematics, distribution and life history notes on Notropis chihuahua (Pisces: Cyprinidae). Copeia, 1981 :255-265. CONABIO (Comision Nacional para el Conocimiento y Uso de la Biodiversidad). 1997. Oficio No. DOO.750,- 1415/ 97), la revision de la NOM-ECOL-059-1994, Norma Oficial Mexicana NOM-ECOL-059- 1 994, que determina las especies y subespecies del flora y fauna silvestres tenestres y acuaticas en peligro de extincion, amnazadas raras y las sujetas a proteccion especial y que establece especificaciones para su proteccion, Publicada en el D.O.F. de fecha 16 de mayo de 1994. Comision Nacional del Agua (CNA). 1997. ProgramaHidraulico de Gran Vision del Estado de Chihuahua 1996 - 2020. Subdireccion General de Programacion Gerencia Regional Norte Gerencia Estatal de Chihuahua, Contrato SGC- GRN-CHIH-96-82A, 923 pp. + 30 appendices. Contreras-B., S. 1977. Speciation aspects and man-made com¬ munity composition changes in Chihuahuan Desert fishes. Pp. 405-432, in Transactions of the Symposium on the Biological Resources of the Chihuahuan Desert Region, United States and Mexico (R. H. Wauer and D. H. Riskind, eds.), National Park Transaction and Proceedings Series, Department of the Interior, Washington, D.C. 88 Special Publications, Museum of Texas Tech University Contreras-B., S. and M. L. Lozano-V. 1994. Water, endangered fishes, and development perspectives in arid lands of Mexico. Conservation Biology, 8:379-387. Eaton, D. J. and D. Hurlburt. 1 992. Challenges in the binational management of water resources in the Rio Grande/Rio Bravo. Austin: U.S. -Mexican Policy Studies Program, Lyndon B. Johnson School of Public Affairs, The Univer¬ sity of Texas at Austin. U.S. -Mexican Policy Report No. 2. Edwards, R. J. and S. Contreras-B. 1991. Historical changes in the ichthyofauna of the lower Rio Grande (Rio Bravo del Norte), Texas and Mexico. The Southwestern Naturalist, 36:201-212. Edwards, R. J., G P. Garrett, and E. Marsh-Matthews. 2002. An ecological analysis of fish communities inhabiting the Rio Conchos basin. (Analisis ecologico de las comunidades de peces que habitan la cuenca del Rio Conchos.) In: Libro Jubilar en Honor al Dr. Salvador Contreras-Balderas (Maria de Lourdes Lozano-Vilano, ed.), pp. 43-61. Garrett, G. P. 1980. Update on some of the protected and endangered fishes of Texas. Proceedings of the Desert Fishes Council, 1 1 :34-36. Garrett, G P., R. J. Edwards, and A. H. Price. 1992. Distribu¬ tion and status of the Devils River minnow, Dionda diaboli. The Southwestern Naturalist, 37:259-267. Gilbert, C. R. 1980. Notropis jemezanus (Cope), Rio Grande shiner. Pp. 279, in Atlas of North American Freshwater Fishes (D. S. Lee et al., eds.), North Carolina State Mu¬ seum of Natural History, Raleigh. Guillory, V. 1 980. Gambusia senilis Girard, Blotched gambu- sia. Pp. 546, in Atlas of North American Freshwater Fishes (D. S. Lee et al., eds.), North Carolina State Mu¬ seum of Natural History, Raleigh. Howells, R. G and G P. Garrett. 1995. Freshwater mussel surveys of Rio Grande tributaries in Chihuahua, Mexico. Triannual Unionid Report, 8:10. Hubbs, C. 1958. Gambusia senilis from the Devil’s River, Texas, an addition to the fish fauna of the United States. Copeia, 1958:239. _ • 1990. Declining fishes of the Chihuahuan Desert. Pp. 89-96, in Third Symposium on Resources of the Chihuahuan Desert Region, United States and Mexico (A. M. Powell, R. R. Hollander, J. C. Barlow, W. B. McGillivray, and D. J. Schmidly (eds.), Chihuahuan Desert Research Institute, Alpine, Texas. Hubbs, C. and A. A. Echelle. 1973. Endangered non-game fishes in the Upper Rio Grande Basin. Pp. 147-167, in Endan¬ gered Vertebrates in the Southwest (W. C. Huey, ed.), New Mexico Department of Game and Fish, Santa Fe. Hubbs, C., R. J. Edwards, and G P. Garrett. 1991. An anno¬ tated checklist of the freshwater fishes of Texas, with keys to identification of species. Texas Journal of Sci¬ ence, Supplement, 43:1-56. Hubbs, C. and G. P. Garrett. 1990. Reestablishment of Cyprinodon eximius (Cyprinodontidae) and status of Dionda diaboli (Cyprinidae) in the vicinity of Dolan Creek, Val Verde Co., Texas. The Southwestern Natural¬ ist, 35:446-478. Hubbs, C., R. R. Miller, R. J. Edwards, K. W. Thompson, E. Marsh, G P. Garrett, G L. Powell, D. J. Morris, and R. W. Zerr. 1 977. Fishes inhabiting the Rio Grande between New Mexico and the Pecos confluence. Pp. 91-97, in Importance, Preservation and Management of Riparian Habitat: A symposium (R. R. Johnson and D. A. Jones, eds.), USDA Forest Service, General Technical Report, RM-43. Hubbs, C. and V. G Springer. 1957. A revision of the Gambusia nobilis species group, with descriptions of three new species, and notes on their evolution. Texas Journal of Science, 9:279-327. Hubbs, C. and R. Wauer. 1973. Seasonal changes in the fish fauna of Tomillo Creek, Brewster County, Texas. The Southwestern Naturalist, 17:375-379. Hubbs, C. L. 1940. Fishes from the Big Bend region of Texas. Transactions of the Texas Academy of Science, 23:3-12. IBWC (International Boundary and Water Commission) . 1 990. Flow of the Rio Grande and related data. Water Bulletin No. 6 1 , El Paso, Texas. _ . 1994. Binational study regarding the presence of toxic substances in the Rio Grande/Rio Bravo and its tributar¬ ies along the boundary portion between the United States and Mexico. Final Report, September 1994, 250 pp. Kelly, M. E. 2001 . The Rio Conchos: A preliminary overview. Texas Center for Policy Studies, Austin, Texas, 27 pp. Lee, R. W. and J. T. Wilson. 1997. Trace elements and organic compounds associated with riverbed sediments in the Rio Grande/Rio Bravo basin, Mexico and Texas. U.S. Geological Survey Fact Sheet, FS-098-97, 6 pp. Miller, R. R. 1961. Man and the changing fish fauna of the American Southwest. Papers of the Michigan Academy of Science, Arts, and Letters, 46:365-404. _ . 1972. Threatened freshwater fishes of the United States. Transactions of the American Fisheries Society, 1 01 :239- 252. _ . 1976. Four new pupfishes of the genus Cyprinodon from Mexico, with a key to the C. eximius complex. Bulletin of the Southern California Academy of Science, 75:68-75. _ . 1977. Composition and derivation of the native fish fauna of the Chihuahuan Desert region. Pp. 365-382, in Transactions of the Symposium on the Biological Re¬ sources of the Chihuahuan Desert Region, United States and Mexico, (R. H. Wauer and D. H. Riskind, eds.), Na¬ tional Park Transaction and Proceedings Series, Depart¬ ment of the Interior, Washington, D.C. Edwards et al. — Fish Assemblages of the Rio Conchos Basin, Mexico 89 _ . 1981. Coevolution of deserts and pupfishes (Genus Cyprinodon ) in the American Southwest. Pp. 39-94, in Fishes in North American Deserts (R. J. Naiman and D. L. Soltz, eds.), John Wiley and Sons, New York. Miller, R. R. and Chemoff, B. 1979. Status of populations of the endangered Chihuahua chub, Gila nigrescens, in New Mexico and Mexico. Proceedings of the Desert Fishes Council, 1 1:74-84. Miller, R. R., J. D. Williams, and J. E. Williams. 1 989. Extinc¬ tions of North American fishes during the past century. Fisheries 14, 22-39. Minckley, W. L. 1980. Cyprinodon eximius Girard, Conchos pupfish. Pp. 496 in Atlas of North American Freshwater Fishes (D. S. Lee et al., eds.), North Carolina State Mu¬ seum of Natural History, Raleigh. Ohmart, R. D. and B. W. Anderson. 1982. North American desert riparian ecosystems. Pp. 433-466, in Reference Handbook on the Deserts of North America (G L. Bender, ed.), Greenwood Press, Westport, Conneticut. Platania, S. P. 1990. The ichthyofauna of the Rio Grande drainage, Texas and Mexico, from Boquillas to San Y gnacio. Final Report to the U.S. Fish and Wildlife Ser¬ vice, Office of Endangered Species, Albuquerque, New Mexico. Propst, D. L., G L. Burton, and B. H. Pridgeon. 1987. Fishes of the Rio Grande between Elephant Butte and Caballo reservoirs, New Mexico. The Southwestern Naturalist, 32:408-411. Propst, D. L. and J. A. Stefferud. 1994. Distribution and status of the Chihuahua chub (Teleostei: Cyprinidae: Gila nigrescens ), with notes on its ecology and associated species. The Southwestern Naturalist, 39:224-234. Addresses of authors: Robert J. Edwards Department of Biology The University of Texas-Pan American Edinburg, Texas 78539 e-mail: redwards@panam.edu Smith, M. L. and R. R. Miller, R.R. 1985. Conservation of desert spring habitats and their endemic fauna in northern Chihuahua, Mexico. Proceedings of the Desert Fishes Council, 13:54-63. Tamayo, J. L. and R. C. West. 1964. The hydrography of Middle America. Pp. 1-570, in Handbook of Middle American Indians, Vol. I. Natural Environment and Early Cultures (R. Wauchope and R. C. West, eds.), University of Texas Press, Austin. TNRCC (Texas Natural Resource Conservation Commission). 1994. Regional Assessment of Water Quality in the Rio Grande Basin including the Pecos River, the Devils River, the Arroyo Colorado and the Lower Laguna Madre. Texas Natural Resource Conservation Commission, AS-34, 377 pp. + appendices. Texas Organization for Endangered Species. 1995. Endangered, threatened, and watch list of vertebrates of Texas. Texas Organization for Endangered Species Publication, 10:1- 22. Trevino-Robinson, D. 1959. The ichthyofauna of the Rio Grande, Texas and Mexico. Copeia, 1959:253-256. Williams, J. E., J. E. Johnson, D. A. Hendrickson, S. Contreras- B., J. D. Williams, M. Navarro-M., D. E. McAllister, and J. E. Deacon. 1989. Fishes of North America endan¬ gered, threatened or of special concern: 1989. Fisheries, 14:2-20. Edie Marsh-Matthews Department of Zoology The University of Oklahoma Norman, Oklahoma 73019 e-mail: emarsh@ou.edu Gary P. Garrett HOH Fisheries Science Center Texas Parks and Wildlife Department Ingram, Texas 78025 e-mail: gary.garrett@tpwd. state, tx. us Historical and Recent Fish Fauna of the Lower Pecos River Christopher W. Hoagstrom Abstract The lower Pecos River extends 770 km, cross¬ ing the Permian Basin and Edwards Plateau from 17 km northwest of Carlsbad, New Mexico to the Rio Grande, near Langtry, Texas. Recent (1991 to 1999) fish collections were depauperate compared with his¬ torical collections from the area. Recent composition of the ichthyofauna was divisible into three assem¬ blages, associated with the following river segments: between Brantley and Red Bluff dams, Red Bluff Dam to Live Oak Creek confluence, and Live Oak Creek to the Rio Grande, but this segregation was not evident within historical collections. Historically, 27 native fish species occurred in all three segments, whereas in re¬ cent collections only nine native fishes occurred in all three. Recent native fish species richness was re¬ duced between 47 and 54% (by segment) from his¬ torical collections. Riverine species were poorly rep¬ resented in recent collections. Seven native and one nonnative fishes historically represented the genus Notropis, but only three species of the genus were found in recent collections. Incidental stockings and bait bucket releases established normative euryhaline fishes (. Fundulus grandis, Menidia beryllina ) that rep¬ resented significant proportions of recent collections. A similar introduction resulted in replacement of Cyprinodon pecosensis by a hybrid swarm (C. pecosensis x C. variegatus). Normative game fishes represented a minor portion of recent collections, de¬ spite concerted efforts to establish several species. Diminishing springflows, normative fish introduction and spread, and toxic algal blooms further threaten native fish populations, while habitat and water quality deterioration favor normative species to the detriment of natives. Appreciation of lower Pecos River histori¬ cal significance to native Rio Grande fishes is impor¬ tant for promoting conservation of remnant native spe¬ cies assemblages both within the lower Pecos River and throughout the Rio Grande basin. Introduction The degraded condition of the lower Pecos River in recent decades was evidenced by elevated salinity (Davis, 1987), toxic algal blooms (James and De la Cruz, 1989; Hubbs, 1990; Rhodes and Hubbs, 1992), and replacement of a native pupfish by a hybrid swarm with a normative congener (Echelle and Connor, 1989). Nineteenth century accounts of the lower Pecos River (Pope, 1854; Dearen, 1996) are very different from more recent accounts (Grozier et al., 1966; Davis, 1980), suggesting that the historical fish fauna may have also been different from the present fauna. This paper summarizes historical fish records, compares them with recent (1991 to 1999) records, and pro¬ vides a historical perspective on habitat and faunal conditions recently observed. Study Area The Pecos River is the largest Rio Grande tribu¬ tary in the United States. The mainstem extends 1,490 km, from more than 3,960 m above sea level in the Sangre de Cristo Mountains, New Mexico, to roughly 305 m above sea level at the river mouth on the Texas- Coahuila border (Figure 1). In the Pecos River drain¬ age, cold- and cool-water streams are present in mon¬ tane headwaters of the mainstem and tributaries. The middle and lower Pecos basins (Figure 1) include warm-water habitats such as the Pecos River 92 Special Publications, Museum of Texas Tech University Lower Pecos River Figure 1. Map of Lower Pecos River, New Mexico and Texas, with the three fish faunal segments delineated. Also included are selected reservoirs, tributaries, and cities. mainstem, spring-fed tributaries, and spring-fed/flood- plain wetlands. This paper is solely concerned with the lower Pecos River mainstem. The lower Pecos River extends roughly 770 km, crossing the Permian Basin for about 575 km (Hill, 1996) and then the Edwards Plateau for 195 km (King, 1935). The Pecos River, Permian Basin section be¬ gins 17 km northwest of Carlsbad, New Mexico (Kelley, 1971; Bachman, 1980), where the river traverses Barrera del Guadalupe, extending down¬ stream to the Edwards Plateau, near Iraan, Texas (Anaya, 2001). Within this stretch, the Pecos River crosses a series of alluvial basins (Maley and Huffington, 1953; Jones, 2001). Downstream, the Pecos River is incised within the Edwards Plateau and confined by limestone cliffs (Thomas, 1972). Hoagstrom — Historical and Recent Fish Fauna of the Lower Pecos River 93 The lower Pecos River receives surface inflow from the middle Pecos River (Figure 1) and from local tributaries that originate in mountains to the west (Guadalupe, Delaware, Apache, Davis, Barrilla, Del Norte, and Glass mountains) (Hill, 1996). However, even the largest western tributaries (e.g., Dark Can¬ yon, Black River, Delaware River, Salt Creek, Toyah Creek, Coyanosa Draw, Tunis Creek) sustain only in¬ termittent surface flow. Headwater flows normally sink into the ground at the base of the mountains, but rise to the surface downstream where water-bearing strata outcrop or encounter impermeable strata (Brune, 1981). A number of significant aquifers, present throughout the lower Pecos River basin, interact hy- drologically (discharge or recharge) with surface wa¬ ters (Richey etal., 1985; Hill, 1996; Mace etal., 2001). An exceptional example is the regional flow system that extends west of the surface drainage boundary and apparently distributes groundwater to multiple Pecos River tributaries by interconnection of the fol¬ lowing aquifer basins: Ryan Flat, Lobo Flat, Salt Ba¬ sin, Apache Mountains, Balmorhea Basin, and Toyah Basin (Sharp, 2001). The east edge of the lower Pecos River drainage is bounded by the Southern High Plains from which no major tributaries enter (Lee, 1925), because of high percolation rates into surficial sands (Jones, 2001). Within the Edwards Plateau, spring- fed tributaries (e.g., Live Oak Creek, Independence Creek, Howard Creek) join the river from both east and west (Brune, 1981). The lower Pecos River ter¬ minates at the Rio Grande, near Langtry, Texas (Fig¬ ure 1). The lower Pecos River forms the boundary be¬ tween the Kansan, Balconian, and Chihuahuan biotic provinces, while the Navahonian province delineates the northwestern boundary and the Tamaulipan prov¬ ince extends up the Rio Grande to near the Pecos River confluence (Blair, 1950). Fishes representative of all five provinces occupy the lower Pecos River, account¬ ing for a relatively diverse native fish fauna (Hubbs, 1957; Smith and Miller, 1986). Most of the lower Pecos River Basin lies within the Chihuahuan province where many native species are threatened or endan¬ gered (Edwards et al., 1989). History According to historical accounts, the lower Pecos River was deep and swift, with steep, unstable banks, a shifting sand substrate, and abundant quicksand (Pope, 1854; U.S. Geological Survey, 1900; Brune, 1981; Leftwich, 1987; Dearen, 1996). In 1854, Pope reported, “The Pecos traverses its valley in a very tor¬ tuous course, and with a current of about two and a half miles to the hour, and from five to twenty feet depth of water.” Rapids or falls were present wher¬ ever the Pecos River encountered bedrock or where tributaries discharged gravel and boulders (Pope, 1854; Hufstetler and Johnson, 1993; Dearen, 1996). Pecos River water was often turbid and had relatively high mineral content, giving the river a reputation for hav¬ ing only bad water. Some springs in the area were very salty or sulphurous (Pope, 1854; Brune, 1981) and saline wetlands were common within the flood¬ plain (Schroeder and Matson, 1965; Leftwich, 1987), but Pope ( 1854) noted “Although the water of the Pecos is somewhat salty. . . the use of it has not been fol¬ lowed by any injurious consequences to the health, of a serious character.” Anglo-American settlement along the lower Pecos River began in the 1 860s and intensified with the com¬ ing of railroad and irrigation companies (Lingle and Linford, 1961). Large-scale water development be¬ gan at the head of the lower Pecos River with Avalon (1891) and McMillan (1893) dams (U.S. Geological Survey, 1900; Freeman and Bolster, 1911; Grover et al., 1922; Meinzer et al., 1926; U.S. National Resources Planning Board, 1942). Immediately thereafter, nu¬ merous river diversions were established between Carlsbad, New Mexico, and Girvin, Texas (Taylor, 1902; Grover et al., 1922; U.S. National Resources Planning Board, 1942). Sediment deprivation result¬ ing from McMillan and Avalon dams changed Pecos River substrate from sand to bedrock (in swift areas) and silt (in slow areas). With capture and diversion of surface flows, groundwater springs became the pri¬ mary source of flow in the mainstem lower Pecos River (Taylor, 1902; Grover et al., 1922; Robinson and Lang, 1938). For example, in May 1918, the Pecos River gained 47.8 cubic feet per second (cfs) (1.35 cms) between the New Mexico/Texas border and 94 Special Publications, Museum of Texas Tech University Girvin, Texas (Grover et al., 1922). In 1925, W.T. Lee observed “[the Pecos River] is a stream of con¬ siderable size at all times. . . Records of two gauging stations near Carlsbad, 2.5 miles apart, show that 80 second-feet of water enters the river in this distance.” Once the lower Pecos River was fragmented, each river segment developed water quality and flow char¬ acteristics specific to local conditions (i.e., the conti¬ nuity of the mainstem river environment was disrupted) (U.S. National Resources Planning Board, 1942). Additional mainstem reservoirs were established during a second development period, beginning in the 1930s (Lingle and Linford, 1961). Red Bluff Dam (1936), near the New Mexico/ Texas border, supplied seven irrigation districts in Texas and Sumner Dam (1937) of the middle Pecos River supplied Carlsbad Irrigation District. Both facilities captured floodwa- ters and sediments. Even with new storage facilities, expanding development and drought caused irrigators to increasingly rely on groundwater, particularly after 1942 (Thomas et al., 1963; West and Broadhurst, 1975). As a result, groundwater flow within adjoining aquifers was altered (Thomas et al., 1963; Mace et al., 2001). Particularly heavy groundwater pumping in the Permian Basin altered groundwater flow-paths, virtually eliminating historical base-flow gains in most areas and causing significant base flow losses in some (Grozier et al., 1966; West and Broadhurst, 1975; Hiss, 1980; Brune, 1981). Davis (1987) summarized a dramatic, human- induced increase in total dissolved solids (i.e., salinity) between 1938 and 1981. Flood control (Howard, 1942; Davis, 1980, 1987), stratification in impound¬ ments and riverine pools (Davis, 1980, 1987), saline aquifer intrusion (Hood, 1963; Havenor, 1968; Jones, 2001), irrigation return flows (Robinson and Lang, 1938; U.S. National Resources Planning Board, 1942; LaFave, 1987; Ashworth, 1990; Mace et al., 2001), oil field pollution (Wiebe et al., 1934; Campbell, 1959; Grozier et al., 1966; Ashworth, 1990), and Tamarix (Davis, 1987), each contributed to salinity increase. Because each of these factors was initiated prior to 1935, when the first water quality investigations were conducted (Robinson and Lang, 1938; U.S. National Resources Planning Board, 1942), pre-development salinity is unknown, but was presumably lower than first recorded in 1935. For purposes of this paper, the lower Pecos River was divided into three segments (Figure 1): 1) Carlsbad segment: McMillan Dam (replaced by Brantley Dam in 1988) to Red Bluff Dam; 2) Toyah segment: Red Bluff Dam to Live Oak Creek confluence; and 3) Edwards segment: Live Oak Creek confluence to the Rio Grande. Segments were distinguishable from each other with respect to flow regime, geomorphology, and water chemistry (Hillis, unpublished; Davis, 1980, 1987; Sublette et al., 1990; Rhodes and Hubbs, 1992). Historical Fish Surveys Ichthyological surveys and summaries seldom have considered the lower Pecos River as a unit. Miller (1977) provided the only species list for fishes spe¬ cific to the area. Important summaries included Evermann and Kendall (1894), who listed fishes taken from the Pecos River basin during railroad and bound¬ ary surveys, and Smith and Miller (1986), who listed fishes native to the entire Pecos River basin and dis¬ cussed their zoogeographic origins. This paper pri¬ marily follows Smith and Miller (1986) in designation of fish species as native (but see Table 1). Historical lower Pecos River fish surveys were not equal among segments. The Carlsbad segment was most heavily surveyed and included the earliest lower Pecos River collections (Pope [1854] at Dela¬ ware River confluence [Evermann and Kendall, 1 894]). Extensive New Mexico Department of Game and Fish (NMDGF) surveys between 1955 and 1970 captured fish using a fish barrier trap and gill-nets below McMillan Dam (Navarre, 1959, 1960; Little, 1961b, 1963a, 1963b). Gill-nets were also used on the mainstem river between Tansill Dam in the city of Carlsbad and Red Bluff Reservoir (Little, 1964c, 1964d, 1965), and within Avalon Reservoir, Carlsbad Munici¬ pal Lake, and Red Bluff Reservoir (Little, 1960a, 1960b, 1961a, 1964b, 1964c). Fishes were also salvaged from irrigation canals (Little, 1964a). General fish commu¬ nity surveys, using seines, were conducted by Koster and associates, University of New Mexico (Koster, Hoagstrom — Historical and Recent Fish Fauna of the Lower Pecos River 95 Table 1. Native fishes known from the mainstem lower Pecos River. Inclusion of species as native follows Smith and Miller (1986) except where noted (numbered footnotes). Recent status of native fishes, based on 1991 to 1999 collections is given: Thriving = frequent and widespread collections in high number (thousands); Stable = reproductive populations in moderate numbers (hundreds); diminished = range reduced, occurrence in moderate numbers; Tenuous - few collections, low numbers ( <25 ); ? = absent from collections but possible via dispersal from Rio Grande; ?? = undetermined due to difficulty in identification and/or lack of documentation; Absent = missing from recent collections, with the year of most recent collection from the lower Pecos River given for each species. The known historical and recent distribution of each species is given by segment. Lettered footnotes provide references of taxonomic interest. Names and taxonomic order follow Mayden et al. (1992). Species Recent Status Historical Distribution Recent Segment Distribution Atractosteus spatula' ABSENT - 1958 EDWARDS? Unknown, possible in Edwards Lepisosteus oculatus 2 ABSENT - 1958 TO YAH - EDWARDS? Unknown, possible in Edwards Lepisosteus osseus stable ALL All, abundant in Carlsbad Anguilla rostrata ABSENT - 1948 ALL - Dorosoma cepedianum THRIVING ALL All, abundant in Carlsbad & Toyah Campostoma anomalum * ABSENT - 1958 EDWARDS - Cyprinella lutrensis STABLE ALL All, abundant in Carlsbad Cyprinella proserpina stable TO YAH - EDWARDS Edwards, common Dionda episcopa diminished ALL Edwards, associated with springs Hybognathus amarus ABSENT - 1963 CARLSBAD - TOYAH - Macrhybopsis a. aestivalis 1 TENUOUS ALL Toyah & Edwards, rare Notropis amabilisw TENUOUS ALL Edwards, uncommon Notropis braytoni TENUOUS ALL Toyah & Edwards, rare Notropis buchanani ABSENT - 1965 EDWARDS - Notropis jemezanus ABSENT - 1987 ALL - Notropis l. ludibundus c TENUOUS CARLSBAD - TOYAH Toyah, rare Notropis orca 4 ABSENT - 1940 EDWARDS - Notropis simus pecosensis d ABSENT - 1987 CARLSBAD - Phenacobius mirabilis 5 ?? ? - Pimephales promelas STABLE ALL Carlsbad, abundant Pimephales vigilax STABLE TOYAH - EDWARDS Edwards, common Rhinichthys cataractae’ ABSENT - 1947 CARLSBAD - Carpiodes carpio elongatusc DIMINISHED ALL Carlsbad, common Cycleptus elongatus{ DIMINISHED CARLSBAD, EDWARDS Carlsbad, mostly between Brantley & Avalon dams, possible in Edwards Ictiobus bubalus TENUOUS ALL Carlsbad, rare, possible in Edwards Ictiobus niger6 ?? ? Unknown, confusing taxonomy Scartomyzon congestus DIMINISHED ALL Carlsbad & Edwards, common in Carlsbad Astyanax mexicanus DIMINISHED ALL Carlsbad & Edwards, uncommon Ictalurus furcatus ^ ABSENT - 1958 ALL Unknown, possible in Edwards Ictalurus lupus ?? ALL Unknown, difficult identification and taxonomy Ictalurus punctatusA DIMINISHED ALL All, most common in Carlsbad Pylodictis olivarisA TENUOUS ALL Carlsbad & Toyah, rare Fundulus zebrinus ABSENT - 1993 ALL - Lucania parvah THRIVING ALL All, most abundant in Toyah Cyprinodon pecosensis ABSENT - 1994 ALL - Gambusia afftnis1 THRIVING ALL All, abundant in Carlsbad & Toyah Gambusia speciosa STABLE EDWARDS Edwards Lepomis cyanellus DIMINISHED ALL All, common in Carlsbad Lepomis gulosus DIMINISHED ALL Carlsbad Lepomis macrochirusA STABLE ALL Carlsbad & Toyah, common in Carlsbad Lepomis megalotis DIMINISHED ALL All, rare in Toyah Micropterus salmoidesA DIMINISHED ALL All, uncommon Etheostoma grahami TENUOUS EDWARDS Edwards, uncommon Etheostoma lepidum ABSENT - 1992 CARLSBAD - 96 Special Publications, Museum of Texas Tech University Table 1. (cont.) Species Recent Status Historical Distribution Recent Segment Distribution Percina macrolepida TENUOUS CARLSBAD - TOYAH Carlsbad, between Brantley & Avalon dams, Black River confluence Aplodinotus grunniens TENUOUS TOYAH - EDWARDS Toyah, rare, possible in Edwards Cichlasoma cyanogidtatum TENUOUS TOYAH - EDWARDS Toyah & Edwards, rare 'Not included by Smith and Miller ( 1 986), but known from at least Amistad Reservoir (J.F. Scudday, Sul Ross State University, pers. comm.), possibly upstream to New Mexico (Hubbs, 1957). 2Not included by Smith and Miller ( 1 986), but well known from the lower Pecos River throughout Texas (e.g., Campbell, 1 959; Hillis, unpublished). incorrectly listed as extinct by Smith and Miller (1986; Hubbs et al., 1991). “Not included by Smith and Miller (1986), but known from a collection by R.M. Bailey at U.S. Highway 90 bridge (Chemoff et ah, 1982). 5Listed as native by Smith and Miller (1986), but nonnative by Sublette et ah (1990). Very few historical records (see Campbell, 1959). 6Listed by Koster (1957) and also by Smith and Miller (1986). The author has found no reference to extant museum specimens or credible accounts. 7Not included by Smith and Miller (1986), but considered native to the Pecos River (Rauchenberger, 1989 and many others). aRio Grande speckled chub, following Eisenhour ( 1 997). bPecos River variant (low mean scale radii counts), following Cobum ( 1 982). cUse of nominal subspecies name follows Tanyolac (1973) as modified by Mayden and Gilbert (1989). dPecos bluntnose shiner, following Chemoff et ah (1982). 'Slender carpsucker, following Hubbs and Black ( 1 940). •Rio Grande basin variant, following Burr and Mayden (1999) and Buth and Mayden (2001). gRio Grande basin variant (unique spotting and head shape), see Garrett and Edwards (2001). hPecos River race, following Hubbs and Miller (1965). ‘Native to Pecos basin, but may be accidental or introduced in lower Pecos River. ANative species that were also introduced as game-fish, likely from sources outside the lower Pecos River basin 1957; Museum of Southwestern Biology records) and, subsequently, by Sublette and associates, Eastern New Mexico University (Sublette, i 975 ; Hatch, 1985; Sublette et ah, 1990). Comprehensive information on rare native fishes of New Mexico was provided by Hubbs and Echelle (1972), while Hatch (1985) and Sublette et al. (1990) summarized historical fish col¬ lections for that state. Historical fish surveys were least extensive in the Toyah segment (Table 2). Collections, primarily by Bailey and others (University of Michigan Museum of Zoology records), Texas Parks and Wildlife De¬ partment (TPW; Campbell, 1959), and Hubbs and Springer, University of Texas at Austin (UT-Austin; Hubbs, 1954, 1957; Texas Natural History Collection records), provided data on fish distribution between 1940 and 1960. Later collections by Davis in 1976, Texas Department of Water Resources (Davis, 1981), Hillis in 1979, and Rhodes et al. in 1987-1988, both of UT-Austin (Hillis, unpublished; Hillis et ah, 1980; Rhodes and Hubbs, 1992; Texas Natural History Col¬ lection records), and Linam and Kleinsasser, TPW in 1987 (Linam and Kleinsasser, 1996) provided more recent information, but only Linam and Kleinsasser surveyed upstream of Girvin, Texas. Historical Toyah segment fish surveys relied entirely on seines. Historical fish surveys from the Edwards seg¬ ment were less extensive than the Carlsbad segment, but greater than from the Toyah segment (Table 2). Except for a TPW survey (Campbell, 1959), surveys were primarily conducted by UT-Austin associates (Hubbs, 1954, 1957; Trevino-Robinson, 1955, 1959; Hillis et ah, 1980; Rhodes and Hubbs, 1992; Texas Natural History Collection records) and Tulane Uni¬ versity associates (Tulane University Museum of Natu¬ ral History records) (Table 3). Historical Edwards segment fish surveys relied on seines except for three gill net collections reported by Campbell (1959). Hoagstrom — Historical and Recent Fish Fauna of the Lower Pecos River 97 Materials and Methods The historical (1939-1990) fish fauna of the lower Pecos River was summarized using published litera¬ ture, agency reports, museum records (Museum of Southwestern Biology, University of Michigan Museum of Zoology, Texas Natural History Collection, Tulane University Museum of Natural History), and personal communications (G.P. Garrett, TPW; J.P. Karges, The Nature Conservancy; S.P. Platania, Museum of South¬ western Biology; J.F. Scudday, Sul Ross State Uni¬ versity). Occurrence of fishes in historical collec¬ tions was tabulated for each of the three lower Pecos River segments. Recent (1991-1999) records reported by Hoagstrom (1994), Larson (1996), and Garrett ( 1 997), along with unpublished data from the NMDGF/ U.S. Fish and Wildlife Service (FWS), and D.M. Hillis of UT- Austin, were compared to pre-1991 records. This paper was primarily concerned with fish community surveys from the mainstem lower Pecos River, but species specific studies (e.g., Echelle and Echelle, 1978; Albeit, 1982; Humphries and Miller, 1982; Hatch et al., 1985; Kelsch and Hendricks, 1990) pro¬ vided supplemental information. Additionally, studies focused on fishes of lower Pecos River tributaries (e.g., Stevenson and Buchanan, 1973; Kennedy, 1977; Cowley and Sublette, 1987; Propst, 1992) added in¬ sight for interpretation of historical and recent fish dis¬ tributions. Results Overall, the effort extended for recent collecting was similar to historical surveys of a given time period (Tables 2 and 3). Recent Carlsbad segment surveys were, in essence, a continuation of traditional NMDGF surveys (Propst, 1992; NMDGF and FWS unpublished data) with increased emphasis on native fishes such as Cyprinodon pecosensis (Echelle et al., 1997; Hoagstrom and Brooks, 1999) and Cycleptus elongatus (Propst, 1999). These efforts were similar to histori¬ cal surveys, with omission of the fish barrier trap and surveys of Avalon and Red Bluff reservoirs, but with addition of boat-mounted electrofishing between Brantley Dam and Avalon Reservoir and within lower Carlsbad Lakes (Propst, 1992). Recent Toyah seg¬ ment surveys were comparable to historical surveys (Table 2), but added gill-net sampling at a few loca¬ tions (Hoagstrom, 1994; Larson, 1996). Recent Edwards segment surveys were least extensive com¬ pared to historical surveys (the main difference being lack of intensive sampling at two sites sensu Rhodes and Hubbs, 1992; Table 3). However, recent and his¬ torical data from the Edwards segment were consid¬ ered at least marginally comparable, particularly be¬ cause 1997 collections conducted by D.M. Hillis, UT- Austin, constituted a partial replication of his 1979 collections. Forty-five native fish species have been reported from historical surveys of the lower Pecos River (not including the unsubstantiated Phenacobius mirabilis and Ictiobus niger, Table 1). Twenty-seven (60%) of these occurred in all three segments. Nine others (20%) were found in two different segments (3 in Carlsbad and Toyah, 5 in Toyah and Edwards, 1 in Carlsbad and Edwards). The remaining nine native fish species (20%) were only present in a single seg¬ ment (3 in Carlsbad, 6 in Edwards). Total number of native species was 34, 35, and 39 in the Carlsbad, Toyah, and Edwards segments, respectively. Recent collections included 30 native fish spe¬ cies (Table 1) with 14 historical inhabitants absent and one (. Ictalurus lupus) uncertain, because of difficulties with identification (Yates et al., 1984; Kelsch and Hendricks, 1986). Only nine (30%) of the 30 remnant native species were found in all three segments, while seven (23%) were taken from two segments (2 in Carlsbad and Toyah, 3 in Toyah and Edwards, 2 in Carlsbad and Edwards). The remaining 14 native spe¬ cies (47%) were only present in collections from a single segment (5 in Carlsbad, 3 in Toyah, 6 in Edwards). The recent total of native species per seg¬ ment was 18, 16, and 18 in Carlsbad, Toyah, and Edwards segments, respectively. Special Publications, Museum of Texas Tech University Hoagstrom — Historical and Recent Fish Fauna of the Lower Pecos River <5 *--c a ‘R Q On ^ O § S* S S ^ a .2 4) a 3 ‘ sa, - O a 'o -r* '--i ^ 4) _ 4^ 4, *-a ^ > 4j 4> _ § ^ £ §• 1 £, .to a cj_ *2 -2 ‘o 4> f | .'o ■». O' *o J. o 5 £ -a > ^ s* * I I ^ 8 | a ^ -2 O jk - <>3 * a O - .2 a £ 2 5 o 5 a si § ^ o' ° 4 S 4. So g a o 2 ^ -3 a no N?* b | ^ ^ ^ 5 I ^ ,a ^ v. so, I r 5 o <0 4> ■2 a I I o £ 2 O S to s 4) <5 5bd^ a *§ ^ s 1 .2 o. | -S S «S S S ! s •! 8 ! s i 'g 1 1 T! *2 2 ^ n2 o 2f ^ a £* 2 > £ ^ 2 c 2 a | ]3 .5 § •S c v» 2 43 -S "2 Si ~a ~q .2 a ^ a •£* i~ -k, ^3 ia ^ ^ § -2 * ^ 2 I, ^ to 5. £ O^ 4) 3. L -->4) *> 2 a £ ^ o 2 "* ■e ^ _ o -J. ~a -r* ^ ^ '■* P*S o o g Pn, o C P .o 2 O ^ CN -g (2^ o ® S § ON o o -~ n — . s; to C 2 o o ^ a S o 2 5 s •>: ^ 5 g 2 ^ » ->2 o to | s to s; o •^««* o 2 o o -« to =o a K tu to •*« O c ^ *° a cu =o ^ , -c 5 ^ o a o o R -« -v 5 5 G o\ fN >* a a o o -a to 5^ o a to sa, 2 -2 ^ ^ o ^ l? a ^ -a ^ ^3 > »* ~o to % (U > u. 3 GO to 'to > CO 4) -*— * C/5 T3 O ‘C 'co > CO o -*-* "c/5 T3 O -C -4—* 4) >* 4) > t- 3 o n co ON CO O •— ■ m oo m c u e D. O 4> > 4> Q 4—4 CO o cu 4-V ^ 4) 4) .3 4) .3 .3 co .3 co co p; co o o _ ; p w w o ire ON ON c o co c S o aL i o _ic > 4) 4> OJO C On On •r io in O.ON ON ^ Z! Z! "S 73 OO X) JO J3 Cu Q. X U CJ C/3 So rn ON NO OO ■a ^ 33 4) C co & •8 | 2 a Garrett 1991 Minnow seine/Electrofisher 1 1 (1.0/site) Larson 1994 Minnow seine/Gill Net 1 1 (1.0/site) Recent Seine 11 11 (1.0/site) Hillis 1997 Seine 9 9 (1.0/site) 100 Special Publications, Museum of Texas Tech University Table 4. Introduced fishes and native/nonnative hybrids known from the mainstem lower Pecos River including Arnistad Reservoir, which inundates the Pecos River-Rio Grande confluence. Historical and recent distribution of each species is given by segment. ? = No locality information. Names and taxonomic order of North American freshwater fishes follow Mayden et al. (1992). Other names follow Robins et al. (1991). Species Historical Distribution Recent Distribution Dorosoma petenense CARLSBAD Carlsbad, uncommon Carassius auratus TOYAH - EDWARDS - Ctenopharyngodon idella CARLSBAD Carlsbad, Carlsbad Lake Cyprinella venusta TOYAH - EDWARDS Edwards, common Cyprinus carpio ALL All, abundant in upper segment Hybognathus placitus CARLSBAD Carlsbad, rare Notemigonus crysoleucas CARLSBAD Carlsbad, rare Notropis girardi CARLSBAD - Catostomus commersoni* CARLSBAD - Ameiurus melas ALL Carlsbad, rare Ameiurus natalis TOYAH - EDWARDS - Esox lucius EDWARDS - Oncorhynchus mykiss CARLSBAD - Menidia beryllina ALL All, abundant in Carlsbad, common in Toyah, uncommon in Edwards Fundulus grandis ALL All, common in Carlsbad & Toyah Cyprinodon pecosensis x C. variegatus ALL All, abundant in Toyah, rare/localized elsewhere Gambusia geiseri EDWARDS Edwards segment, uncommon Morone chrysops ALL Carlsbad, rare Morone saxatilis CARLSBAD - Morone saxatilis x M. chrysops CARLSBAD - Ambloplites rupestris CARLSBAD - Lepomis auritus EDWARDS - Lepomis humilis ? - Lepomis microlophus EDWARDS Edwards Micropterus dolomieu EDWARDS - Micropterus punctulatus CARLSBAD Carlsbad Pomoxis annularis ALL Carlsbad Pomoxis nigromaculatus CARLSBAD - Perea flavescens CARLSBAD - Stitzostedion canadense EDWARDS - Stitzostedion vitreum CARLSBAD Carlsbad, rare Cynoscion nebulosus CARLSBAD - Micropogonias undulatus CARLSBAD - Pogonias cromis CARLSBAD - Sciaenops ocellatus CARLSBAD - TOYAH - Paralichthys lethostigma CARLSBAD - *Native to Pecos River headwaters, presumed introduced to the lower Pecos River. Ten native species appeared to be thriving or have stable populations during recent collections (Table 1). Status of the remaining 20 is either diminished or tenu¬ ous (diminished species were sporadic in occurrence and/or restricted in distribution; tenuous species were very rare). Macrhybopsis a. aestivalis , Notropis l. ludibundus, and Aplodinotus grunniens were each rep¬ resented by a single individual. The number of native species missing from recent collections was 16, 19, and 20 per segment (Carlsbad, Toyah, and Edwards respectively), representing fish species richness re¬ ductions of 47, 54, and 51% respectively. Thirty-six introduced fish species were reported from historical lower Pecos River surveys (Table 4). Seven of these (19%) were known from all segments Hoagstrom — Historical and Recent Fish Fauna of the Lower Pecos River 101 and four (11%) were found in two segments (1 in Carlsbad and Toyah, 2 in Toyah and Edwards). Twenty-four introduced species (69%) were restricted to a single segment (18 to Carlsbad, 6 to Edwards). Total introduced species per segment was 26, 11, and 16 (Carlsbad, Toyah, and Edwards, respectively). Historical distribution of Lepomis humilis was not re¬ ported (Campbell, 1959). Sixteen introduced fish species were present in recent collections (Table 4). Four (25%) were found in all segments, but the remaining 12 (75%) were re¬ stricted to a single segment (9 to Carlsbad, 3 to Edwards). Five of these ( Hybognathus placitus, Notemigonus crysoleucas, Ameiurus melas, Morone chrysops, Stitzostedion vitreum) were rare, likely rep¬ resenting bait bucket releases (first two) or strays from reservoirs or tributaries (last three). The recent total of introduced species per segment was 13, 4, and 7 (Carlsbad, Toyah, and Edwards, respectively). Discussion Comparison of historical and recent native fish species composition indicated significant decline. His¬ torically, the three segments had similar fish species richness and pre-dam (i.e., pre-systematic fish sur¬ vey) similarity between segments was likely even greater. The lower Pecos River did not sustain a com¬ mercial fishery (as did large rivers elsewhere in North America) so the public may have been relatively un¬ aware of what fishes were present, causing large-river fishes (e.g., Atractosteus spatula , Lepisosteus oculatus, Anguilla rostrata, C. elongatus, and A. grunniens ) to be poorly documented in historical accounts. For ex¬ ample, Hubbs (1957) believed Lepisosteus platostomus, collected by Pope in 1 854 (see Evermann and Kendall, 1894) represented A. spatula, whereas Sublette et al. (1990) suggested the species captured was L. oculatus. Following the first development period (after 1930), L. oculatus occurred further upstream than A. spatula (Campbell, 1959; J. F. Scudday, Sul Ross State Uni¬ versity, pers. comm.), suggesting Pope’s specimen was more likely L. oculatus , but human-caused changes in flow regime, channel sediment, and water quality made the middle Twentieth Century lower Pecos River much different from that of 1854 (see above). Thus, it is possible both gar species were present in New Mexico prior to impoundment, neither being documented. Researchers active throughout the second de¬ velopment period noted the decline of lower Pecos River fishes. During a 1947 visit to Malaga Bend, Koster complained, “Collecting to date has been dis¬ appointing. Fish are scarce. Many species which are known from both above and below are seemingly ab¬ sent from this lower stretch of the Pecos” (from field notes 1939-1955). The dramatic decline of Notropis species serves as an example. Notropis jemezanus, once widespread (Campbell, 1959; Trevino-Robinson, 1955; Sublette et al., 1990), was absent from recent collections (possibly persisting in Independence Creek [Garrett, 1997; Karges, The Nature Conservancy, pers. comm.]). Notropis l. ludibundus disappeared from the Carlsbad segment before 1975 (Sublette, 1975) and is known from a single recent Toyah segment specimen (Hoagstrom, 1994). Notropis amabilis and N. braytoni, historically found as far upstream as Roswell, New Mexico (Hatch, 1985; Platania, 1996), are rare in recent Toyah and Edwards segment collec¬ tions. Notropis girardi, introduced to the Carlsbad segment around 1978 (Bestgen et al., 1989), was ap¬ parently never established. Notropis simus pecosensis now is restricted to the middle Pecos River between Sumner and Brantley dams (Sublette et al., 1990; Propst, 1999), but the nominal collection of this subspecies was made in 1 854 by Pope (Chemoff et al., 1982) who surveyed the Pecos River from Black River confluence, downstream to Emigrant Crossing near Barstow, Texas (Pope, 1854). Although the exact location of Pope’s N. s. pecosensis collection was unspecified (Evermann and Kendall, 1894; Platania, 1995), this record indicates that N. s. pecosensis historically inhabited the lower Pecos River when it was deep and swift, with a sand bed. By the time subsequent fish surveys were con¬ ducted, the lower Pecos River had become salty, slug¬ gish, and silty. 102 Special Publications, Museum of Texas Tech University Notropis orca and N. buchanani were rare in his¬ torical lower Pecos River collections, but common in the adjacent Rio Grande (Trevino-Robinson, 1955, 1959; Chemoff et al., 1982). Their abundance in the pre-impoundment lower Pecos River is unknown, but their disappearance from the drainage may have fore¬ shadowed their decline in the Rio Grande. Notropis orca is now extinct (Chemoff et ah, 1982; Bestgen and Platania, 1990; Hubbs et ah, 1991), and N. buchanani is rare within the Rio Grande (Platania, 1990; Edwards and Contreras-Balderas, 1991). Thus, the more recent decline of N. braytoni and N. jemezanus from the lower Pecos River may justify increased con¬ cern for Rio Grande populations of these species. Similar to developments for irrigated agriculture (Taylor, 1902; President’s Water Resources Policy Commission, 1950; Lingle and Linford, 1961), attempts to establish productive sport fisheries in the lower Pecos River did not meet expectations (Campbell, 1 959; Navarre, 1959, 1960; Little, 1961b, 1963a, 1963b, 1964c, 1964d, 1965). Because game-fish manage¬ ment efforts were largely initiated subsequent to ma¬ jor water development, conditions that prevented fish¬ ery success were not solely attributable to geological or human-induced factors, but represented a combi¬ nation of both (Campbell, 1959; Little, 1964c; Davis, 1987). Poor and unstable water quality (Little, 1964d; Larson, 1996; Linam and Kleinsasser, 1996) and toxic algal blooms (Rhodes and Hubbs, 1992) significantly impacted game fish success. As a result of dramatic human-caused changes, the historical status of native game fishes and potential of the pre-impoundment lower Pecos River to support nonnatives will never be known. Some native species present in historical col¬ lections may not have been established within the pre¬ impoundment lower Pecos River. For example, Rhinichthys cataractae, was common in the upper Pecos River, but infrequent in the lower Pecos River (Miller, 1977; Sublette et al., 1990). This species may have colonized tailwaters of McMillan and Avalon res¬ ervoirs during floods or with human aid. Both tailwater reaches were eventually dewatered, so absence of R. cataractae from recent collections is not surprising. Etheostoma lepidum typically inhabits small streams with dense vegetation (Hubbs et al., 1953; Cowley and Sublette, 1987) and it was never abun¬ dant in the mainstem Pecos River (Hubbs and Echelle, 1972). The species may have entered the Pecos River from tributaries, colonizing spring-fed areas that domi¬ nated the mainstem after the first development period. These darters probably declined during the second development period as spring flows were depleted. Similarly, Balconian fishes ( Campostoma anomalum , Cyprinella proserpina, Dionda episcopa, N. amabilis, N. 1. ludibundus, Pimephales vigilax, Scartomyzon conges tus, Etheostoma grahami) and Tamaulipan fishes (N. braytoni, Cichlasoma cyanoguttatum ) were most abundant in spring-fed tributaries (Rhodes and Hubbs, 1992) and their decline from the lower Pecos River coincided with spring flow depletion (Linam and Kleinsasser, 1996). Cyprinodon pecosensis and Fundulus zebrinus were probably uncommon in the pre-impoundment Pecos River mainstem, most likely occupying saline tributaries and floodplain wetlands (Hoagstrom and Brooks, 1999). Both species proliferated in the lower Pecos River as it was dewatered (Campbell, 1959), but loss of floodplain wetlands eventually restricted them to the mainstem and a few persistent tributaries (Hoagstrom and Brooks, 1999). Subsequently, mainstem populations were decimated by introduced congeners. A C. pecosensis x C. variegatus hybrid swarm replaced C. pecosensis (Echelle et al., 1987; Echelle and Connor, 1989; Wilde and Echelle, 1992), and F. zebrinus was replaced by F. grandis (Hoagstrom, 1994). Cyprinodon pecosensis (3 locations) and F. zebrinus (3 locations) persist in off-channel locations, with the largest populations of both species inhabiting upper Salt Creek, Culberson and Reeves counties, Texas (N.L. Allan, FWS; A. A. Echelle, Oklahoma State University; G.P. Garrett, TPW; J.P. Karges, The Na¬ ture Conservancy, pers. comm.). In each lower Pecos River segment, recent native fish species richness was roughly half of his¬ torical richness. Resultant fish communities repre¬ sented a response to the water quality, physical habi¬ tat, and flow regime of each segment. Differences among segments were exacerbated by toxic algal blooms and physical barriers. Native species persis¬ tent in the Carlsbad segment included lentic freshwa¬ ter fauna (e.g., Dorosoma cepedianum, Lepisosteus osseus , centrarchids), generalist freshwater fauna (e.g., Cyprinella lutrensis, Pimephales promelas, Menidia Hoagstrom — Historical and Recent Fish Fauna of the Lower Pecos River 103 beryllina, Gambusia affinis ), and riverine catostomids (e.g., Carpiodes carpio elongatus, C. elongatus, Ictiobus bubalus). The Carlsbad segment was heavily impacted by impoundment, agriculture, and urbaniza¬ tion, with significant impacts from diversion, oil field pollution, groundwater pumping, and upstream devel¬ opment (U.S. National Resources Planning Board, 1942; Thomas et al., 1963; Davis, 1987), readily ac¬ counting for recent absence of 16 native species. Relatively high recent introduced species richness (n=13) is attributable to presence of three large reser¬ voirs, absence of toxic algal blooms (except in Red Bluff Reservoir), and persistence of spring flows in the city of Carlsbad, Black River, and Delaware River. Euryhaline fishes (e.g., D. cepedianum, C. pecosensis x C. variegatus, F. grandis, Lucania parva, G. affinis) dominated the Toyah segment, which was not surprising in light of numerous impacts that con¬ centrated salts therein, including upstream develop¬ ment (e.g., Carlsbad segment), local groundwater with¬ drawal, oil field pollution, mainstem and tributary di¬ version, and agriculture (Taylor, 1902; U.S. National Resources Planning Board, 1942; Grozier et ah, 1966; Mace et ah, 2001). Toxic algal blooms further im¬ pacted Toyah segment fishes (Rhodes and Hubbs, 1992; Hoagstrom, 1994). Failure of introduced game fishes (none taken in recent surveys) is attributable to absence of large reservoirs, absence of significant spring inflows, and toxic algal blooms. Severity of direct and indirect impacts on the Toyah segment and absence of redeeming habitat features (e.g., persistent springs) clearly account for recent absence of 19 na¬ tive species. The Edwards segment supported freshwater generalists (e.g., C. lutrensis, C. venusta, P vigilax, G. affinis) and spring-dwelling specialists (e.g., C. proserpina , D. episcopa, E. grahami). It is possible that the Edwards segment retains species not taken in recent surveys, because recent surveys were not ex¬ tensive (Table 3) and did not include gill-net or boat- mounted electrofishing collections, increasing the like¬ lihood that riverine species (e.g., L. oculatus, C. elongatus, I. bubalus, I. furcatus, A. grunniens) could have escaped detection. Even so, severe upstream impacts (e.g., Carlsbad and Toyah segments) could account for the recent absence of 20 native species, especially in combination with toxic algal blooms that reduce short-term species richness and possibly cause long-term reductions (Rhodes and Hubbs, 1992). Downstream impoundment of the Edwards segment by Amistad Reservoir could also have facilitated spe¬ cies loss (Winston et al., 1991; Wilde and Ostrand, 1999; Lienesch et al., 2000). Persistence of normative game fishes in the Edwards segment may be attrib¬ uted to persistent spring flows and Amistad Reservoir. Minckley (1965) suggested that the lower Pecos River was incidentally stocked withM beryllina along with estuarine sport-fishes (e.g., Sciaenops ocellatus). Subsequently, M. beryllina dispersed be¬ yond the lower Pecos River (Hubbs and Echelle, 1972), currently ranging as far as 1 88 km upstream of Brantley Dam (New Mexico Fishery Resources Office, FWS, unpublished data). Incidental stockings of this sort probably introduced estuarine invertebrates (Davis, 1987) and could have played a role in establishing C. variegatus ( C . variegatus were reported by Campbell in 1959 in the same time period and vicinity of S. ocellatus introductions, but no voucher specimens are available). Successful introduction of Cyprinodon variegatus to the lower Pecos River most likely oc¬ curred after 1980, with establishment and spread fa¬ cilitated by bait-bucket release (Echelle et al., 1987; Echelle and Connor, 1989; Hubbs et al., 1991; Wilde and Echelle, 1992; Echelle et al., 1997; Echelle et al., this volume). Genetic evidence suggests the species first colonized Red Bluff Reservoir via introduction from Lake Balmorhea, Reeves County, Texas (Childs et al., 1996). The Lake Balmorhea C. variegatus popu¬ lation was established from an unknown source prior to 1968 (Stevenson and Buchanan, 1973) and persists today (Echelle et al., this volume). It also served as the source for a recent C. variegatus introduction to Dia¬ mond Y Spring (Echelle and Echelle, 1997). The point of introduction for F. grandis has not been specifically investigated, but establishment in the Edwards seg¬ ment (Hillis et al., 1980; Hubbs, 1982; Rhodes and Hubbs, 1992; Linam and Kleinsasser, 1996) and Carlsbad segment (Propst, 1992), prior to expansion into Toyah segment (Hoagstrom, 1994; Larson, 1996) suggests at least two separate introductions. 104 Special Publications, Museum of Texas Tech University Conclusions The historical fish fauna of the lower Pecos River included many riverine forms and was not fragmented by physical barriers. Early Anglo-American develop¬ ment (1885 to 1929) resulted in capture and diversion of surface waters and alluvial sediments, after which river substrate changed and local groundwater springs became the primary source of river flow. A second phase of Anglo-American development (1930 to 1970) reduced inflow from groundwater, in many cases di¬ recting groundwater flow away from the river, while additional dams furthered river fragmentation. As a result, lower Pecos River hydrology, geomorphology, and water chemistry were dramatically altered from a natural condition that was never quantitatively de¬ scribed. In response to human induced changes, each river segment developed a distinctive fish community composed of tolerant native and introduced fishes. Native riverine fishes declined from all segments be¬ cause mainstream habitats were universally impacted. Spring and wetland fishes retreated to areas sustaining substantial spring inflows (e.g., Black River, Salt Creek, Independence Creek, lower Pecos River Edwards seg¬ ment). The Toyah segment, was most dramatically impacted by development. As a result, this segment suffered the greatest percent native species richness reduction and did not sustain introduced game fishes. Intentional game fish stockings met with only short-term success while incidental stocking and bait- bucket release established normative euryhaline fishes. Success of game species was likely limited by unfa¬ vorable water quality, degraded habitat, and toxic algal blooms. These same factors apparently favored unin¬ tentionally introduced euryhaline fishes. Additional normative fish introductions and spread of locally es¬ tablished nonnatives continue to threaten native lower Pecos River fishes. Because of numerous, large-scale impacts that have altered the mainstem lower Pecos River, chal¬ lenges for native fish conservation are many and great. Proponents of aquatic habitat restoration and native fish protection will benefit from an appreciation of former fish community diversity and severity of im¬ pacts that changed the lower Pecos River. While it is appropriate to consider the entire lower Pecos River as a significant component of the historical range of many native Rio Grande basin fishes, it may be opti¬ mistic to expect restoration of all or even a few miss¬ ing species (in light of prevailing conditions). How¬ ever, a number of persistent native species (including unique Pecos River and Rio Grande forms, Table 1) would benefit from immediate population assessment and conservation activity. Improvement of conditions to preserve these species may also facilitate voluntary re-establishment of fishes occupying the adjacent Rio Grande, Pecos River tributaries, or the middle Pecos River. At least, recognition of rapid and ongoing deci¬ mation of native lower Pecos River fishes should in¬ crease awareness of the general imperilment of Rio Grande basin fishes, focusing attention on and elevat¬ ing prioritization of waters where remnant native fish assemblages persist. Acknowledgments I thank D.M. Hillis, D.A. Hendrickson, and J. Rosales of the Texas Natural History Collection for loan of Dr. Hillis’ 1997 collections. Museum records were provided by S.P. Platania and A.M. Snyder (MSB), A. A. Anderson (TNHC), and H.L. Bart and N. Rios (TUMNH). S.P. Platania and A.M. Snyder also pro¬ vided copies of W.M. Rosters’ field notes. Recent fish collections in New Mexico were conducted by D.L. Propst and J.E. Brooks (whom I thank for free use of their data), assisted by A.L. Hobbes, R.D. Larson, J.R. Smith, J.E. Davis, N.L. Allan, B.G Wiley, T.J. Koschnick, M.R. Brown, C.M. Williams, R.A. Maes, and J.A. Jackson. R. Fletcher, L. Villareal, P. Diaz, N.L. Allan, and C.S. Berkhouse assisted with the author’s collections from Texas. Many generous in¬ dividuals including N.L. Allan, P.J. Connor, D. Dobos- Bubno, G.P. Garrett, Carl W. Hoagstrom, J.P. Karges, R. Kleinsasser, GL. Larson, R.D. Larson, G.W. Linam, Hoagstrom — Historical and Recent Fish Fauna of the Lower Pecos River 105 S.P. Platania, D.L. Propst, J. Rosales, A.M. Snyder, and P.L. Tashjian have given me information. Helpful critique of manuscript drafts was provided by G.L. Larson, G.W. Linam, L.L. Paskus, A. Roberson, K. Winemiller, and two anonymous reviewers. Susan Maestas provided frequent logistical and technical as¬ sistance and K.R. Winter graciously prepared the map (Figure 1). Most of all, I thank J.F. Scudday for bring¬ ing me to the Pecos, J.E. Brooks for providing moral, logistical, and financial support since that time, and N.L. Allan and G.P. Garrett for organizing and facilitat¬ ing this special publication. Literature Cited Albeit, P. S. 1982. The ecology of the Pecos pupfish, Cyprinodon pecosensis Echelle and Echelle, in the lower Pecos River drainage of New Mexico (Cyprinodontiformes; Cyprinodontidae). Unpublished Master’s Thesis, East¬ ern New Mexico University, Portales. 67 pp. Anaya, R. 2001. An overview of the Edwards-Trinity aquifer system, central-west Texas. Pp. 100-119, in Aquifers of West Texas (R.E. Mace, W.F. Mullican III, and E.S. Angle, eds), Texas Water Development Board Report 356, Aus¬ tin. 272 pp. Ashworth, J. B. 1990. Evaluation of ground-water resources in parts of Loving, Pecos, Reeves, Ward, and Winkler coun¬ ties, Texas. Texas Water Development Board Report 317, Austin. 51pp. Bachman, G O. 1980. Regional geology and Cenozoic history of Pecos region, southeastern New Mexico. U.S. Geo¬ logical Survey Open-File Report 80-1099, 1 16 pp. Bestgen, K. R. and S. P. Platania. 1990. Extirpation of Notropis simus simus (Cope) and Notropis orca Woolman (Pisces: Cyprinidae) from the Rio Grande in New Mexico, with notes on their life history. Occasional Papers, Museum of Southwestern Biology, 6:1-8. Bestgen, K. R., S. P. Platania, J. E. Brooks, and D. L. Propst. 1989. Dispersal and life history traits of Notropis girardi (Cypriniformes: Cyprinidae), introduced into the Pecos River, New Mexico. American Midland Naturalist, 122:228-235. Blair, F. W. 1950. The biotic provinces of Texas. Texas Journal ofScience, 2:93-1 17. Brune, G 1981. Springs of Texas, Volume 1. Branch-Smith, Inc., Fort Worth, Texas. 499 pp. Burr, B. M. and R. L. Mayden. 1 999. Anew species of Cycleptus (Cypriniformes: Catostomidae) from Gulf Slope drain¬ ages of Alabama, Mississippi, and Lousiana, with a re¬ view of the distribution, biology, and conservation status of the genus. Alabama Museum of Natural History, Bul¬ letin, 20:19-57. Buth, D. G and R. L. Mayden. 2001. Allozymic and isozymic evidence for polytypy in the North American Catostomid genus Cycleptus. Copeia, 2001 : 899-906. Campbell, L. 1959. Basic survey and inventory of species present in the Pecos River of Texas. Dingell-Johnson Project F-5-R-5, Job B-13, Texas Game and Fish Com¬ mission, Austin. 3 1 pp. Chemoff, B., R. R. Miller, and C. R. Gilbert. 1982. Notropis orca and Notropis simus, cyprinid fishes from the Ameri¬ can Southwest, with description of a new subspecies. Occasional Papers of the Museum of Zoology No. 698, University of Michigan, Ann Arbor. 49 pp. Childs, M. R., A. A. Echelle, and T. E. Dowling. 1996. Devel¬ opment of the hybrid swarm between Pecos pupfish (Cyprinodontidae: Cyprinodon pecosensis ) and sheeps- head minnow ( Cyprinodon variegatus ): a perspective from allozymes and mtDNA. Evolution. 50:2014-2022. Cobum, M. M. 1982. Anatomy and relationships of Notropis atherinoides. Unpublished Ph.D. dissertation, Ohio State University, Columbus. 384 pp. Cowley, D. E. and J. E. Sublette. 1987. Distribution of fishes in the Black River drainage, Eddy County, New Mexico. The Southwestern Naturalist, 32:213-221. Davis, J. R. 1980. Species composition and diversity of benthic macroinvertebrate populations of the Pecos River, Texas. The Southwestern Naturalist, 25:241-256. Davis, J. R. 1981. Diet of the Pecos River Pupfish, Cyprinodon pecosensis (Cyprinodontidae). The Southwestern Natu¬ ralist, 24:535-540. Davis, J. R. 1987. Faunal characteristics of a saline stream in the northern Chihuahuan Desert. Contributed papers of the second symposium on resources of the Chihuahuan Desert region, United States and Mexico. Chihuahuan Desert Research Institute, Alpine, Texas, 1 8 pp. Dearen, P. 1996. Crossing Rio Pecos. Texas Christian Univer¬ sity Press, Fort Worth, 196 pp. Echelle, A. A. and P. J. Connor. 1989. Rapid, geographically extensive genetic introgression after secondary contact between two pupfish species ( Cyprinodon , Cyprinodontidae). Evolution, 43:717-727. Echelle, A. A. andA. F. Echelle. 1978. The Pecos River pupfish, Cyprinodon pecosensis n. sp. (Cyprinodontidae), with comments on its evolutionary origin. Copeia, 1 978:569- 582. 106 Special Publications, Museum of Texas Tech University Echelle, A. A. andA. F. Echelle. 1997. Genetic introgression of endemic taxa by non-natives: a case study with Leon Springs pupfish and sheepshead minnow. Conservation Biology, 1 1:153-161. Echelle, A. A., A. F. Echelle, and D. R. Edds. 1987. Population structure of four pupfish species (Cyprinodontidae: Cyprinodon ) from the Chihuahuan Desert region of New Mexico and Texas: allozymic variation. Copeia, 1 987:668- 681. Echelle, A. A., C. W. Hoagstrom, A. F. Echelle, and J. E. Brooks. 1997. Expanded occurrence of genetically introgressed pupfish (Cyprinodontidae: Cyprinodon pecosensis x variegatus) in New Mexico. The Southwestern Natural¬ ist, 42:336-339. Edwards, R. J. and S. Contreras-B. 1991. Historical changes in the ichthyofauna of the Lower Rio Grande (Rio Bravo del Norte), Texas and Mexico. The Southwestern Naturalist, 36:201-212. Edwards, R. J., G Longley, R. Moss, J. Ward, R. Matthews, and B. Stewart. 1989. A classification of Texas aquatic com¬ munities with special consideration toward the conserva¬ tion of endangered and threatened taxa. Texas Journal of Science, 4 1:23 1-240. Eisenhour, D. J. 1 997. Systematics, variation, and speciation of the Macrhybopsis aestivalis complex (Cypriniformes: Cyprinidae) west of the Mississippi River. Unpublished Ph.D. dissertation, Southern Illinois University, Carbondale, 260 pp. Evermann, B. W. and W. C. Kendall. 1 894. The fishes of Texas and the Rio Grande Basin, considered chiefly with refer¬ ence to their geographic distribution. U.S. Fish Commis¬ sion, Bulletin, 12:57-126 with plates. Freeman, W. B. and R. H. Bolster. 1911. Surface water supply of the United States, 1909, Part VIII. Western Gulf of Mexico. U.S. Geological Survey Water-Supply Paper 268, 107 pp. Garrett, G P. 1997. Chihuahuan Desert fishes status survey. Final Report, Grant Number E-l-8, Endangered and Threatened Species Conservation, Project E410. Texas Parks and Wildlife Department, Ingram, 17 pp. Garrett, G P. and R. J. Edwards. 2001. Regional ecology and environmental issues in West Texas. Pp. 56-65, in Aqui¬ fers of West Texas (R.E. Mace, W.F. Mullican III, and E.S. Angle, eds), Texas Water Development Board Re¬ port 356, Austin, Texas, 272 pp. Grover, N. C., G A. Gray, and C. E. Ellsworth. 1922. Surface water supply of the United States, 1918, Part VIII. West¬ ern Gulf of Mexico basins. U.S. Geological Survey Wa¬ ter-Supply Paper 478, 106 pp. Grozier, R. U., H. W. Albert, J. F. Blakely, and C. H. Hembree. 1 966. Water-delivery and low-flow studies, Pecos River, Texas, quantity and quality, 1964 and 1965. Texas Water Development Board Report 22, Austin, Texas, 21 pp. Hatch, M. D. 1985. The native fish fauna of major drainages east of the Continental Divide in New Mexico. Unpub¬ lished Master’s Thesis, Eastern New Mexico University, Portales, 33 pp, plus appendix. Hatch, M. D., W. H. Baltosser, and C. G Schmitt. 1985. Life history and ecology of the bluntnose shiner ( Notropis simus pecosensis ) in the Pecos River of New Mexico. The Southwestern Naturalist, 30:555-562. Havenor, K. C. 1 968. Structure, stratigraphy, and hydrogeology of the northern Roswell Artesian Basin, Chaves County, New Mexico. State Bureau of Mines and Mineral Re¬ sources, New Mexico Institute of Mining and Technol¬ ogy, Socorro, NM, Circular 93, 30 pp. Hill, C. A. 1 996. Geology of the Delaware Basin, Guadalupe, Apache, and Glass mountains, New Mexico and west Texas. Society for Sedimentary Geology, Permian Basin Section-SEPM, Publication No. 96-39, 480 pp. Hillis, D. M. Unpublished. The heterothermic vertebrates of the lower Pecos River drainage, Pecos, Crockett, Terrell, and Val Verde counties, Texas. Unpublished Manuscript, 50 pp, plus attachments. Hillis, D. M., E. Milstead, and S. L. Campbell. 1980. Inland records of Fundulus grand is (Cyprinodontidae) in Texas. The Southwestern Naturalist, 25:271-272. Hiss, W. L. 1980. Movement of ground water in Permian Guadalupian aquifer systems, southeastern New Mexico and western Texas. New Mexico Geological Society Guidebook, 3 1 “Field Conference, Trans-Pecos Region:85- 90. Hoagstrom, C. W. 1994. Relative abundance and niche parti¬ tioning in the fish community of the Pecos River, Pecos County, TX. Unpublished Master’s Thesis, Sul Ross State University, Alpine, Texas. 136 pp. Hoagstrom, C. W. and J. E. Brooks. 1999. Distribution, status, and conservation of the Pecos pupfish, Cyprinodon pecosensis. Technical Report No. 2. New Mexico De¬ partment of Game and Fish, Santa Fe, 76 pp. Hood, J. W. 1963. Saline ground water in the Roswell Basin, Chaves and Eddy counties New Mexico, 1958-1959. U.S. Geological Survey Water-Supply Paper 1539-M, 46 pp. Howard, C. S. 1 942. Quality of stored water available for use in the lower basin of the Pecos River, Texas. American Geophysical Union, Transactions, 23:593-598. Hubbs, C. L. and J. D. Black. 1940. Status of the Catostomid fish, Carpiodes carpio elongatus (Meek). Copeia, 1940:226-230. Hubbs, C. L. and R. R. Miller. 1965. Studies of Cyprinodont fishes. XXII. Variation in Lucania parva, its establish¬ ment in western United States, and description of a new species from an interior basin in Coahuila, Mexico. Mu¬ seum of Zoology, University of Michigan, Miscellaneous Publications, No. 127, 110 pp. Hoagstrom — Historical and Recent Fish Fauna of the Lower Pecos River 107 Hubbs, C. 1954. Corrected distributional records for Texas fresh-water fishes. Texas Journal of Science, 6:277-291 . Hubbs, C. 1957. Distributional patterns of Texas fresh-water fishes. The Southwestern Naturalist, 2:89-104. Hubbs, C. 1982. Occurrence of exotic fishes in Texas waters. Pearce-Sellards Series, No. 36, 1 8 pp. Hubbs, C. 1990. Declining fishes of the Chihuahuan Desert. Pp. 89-96 in Third Symposium on Resources of the Chihuahuan Desert Region, U.S. and Mexico (A.M. Powell, R.R. Hollander, J.C. Barlow, W.B. McGilivray, and D.J. Schmidly, eds). Chihuahuan Desert Research Institute, Alpine, Texas. Hubbs, C. and A. A. Echelle. 1 972. Endangered non-game fishes of the upper Rio Grande Basin. Pp. 147-167 in Sympo¬ sium on rare and endangered wildlife of the southwestern United States. New Mexico Department of Game and Fish, Santa Fe, New Mexico. Hubbs, C., R. J. Edwards, and G P. Garrett. 1991. An anno¬ tated checklist of the freshwater fishes of Texas, with keys to identification of species. Texas Journal of Sci¬ ence, Supplement, Vol. 43(4), 56 pp. Hubbs, C., R. A. Kuehne, and J. C. Ball. 1953. The fishes of the upper Guadalupe River, Texas. Texas Journal of Science, 5:216-251. Humphries, J. M. and R. R. Miller. 1982. Taxonomic status, ecology, and distribution of the headwater catfish, Ictalurus lupus. Final Report, Contract No. 14-16-0002- 81-210. Museum of Zoology, University of Michigan, Ann Arbor, Michigan, 45 pp. Hufstetler, M. and L. Johnson. 1993. Watering the land, the turbulent history of the Carlsbad Irrigation District. Na¬ tional Park Service, Rocky Mountain Region, Denver, 180 pp. James, T. L. and A. De La Cruz. 1989. Pyrmnesium parvum Carter (Chrysophyseae) as a suspect of mass mortalities of fish and shellfish communities in western Texas. Texas Journal of Science, 4 1 :429-430. Jones, I. C. 2001. Cenozoic Pecos alluvium aquifer. Pp. 120- 1 34, in Aquifers of West Texas (R.E. Mace, W.F. Mullican III, and E.S. Angle, eds), Texas Water Development Board Report 356, Austin, Texas, 272 pp. Kelley, V. C. 1971. Geology of the Pecos country, southeastern New Mexico. New Mexico Bureau of Mines and Min¬ eral Resources, Memoir 24, 78 pp., with plates. Kelsch, S. W. and F. S. Hendricks. 1986. An electrophoretic and multivariate morphometric comparison of the American catfishes Ictalurus lupus and I. punctatus. Copeia, 1986:646-652. Kelsch, S. W. and F. S. Hendricks. 1990. Distribution of the headwater catfish Ictalurus lupus (Osteichthyes: Ictaluridae). The Southwestern Naturalist, 35:292-297. Kennedy, S. E. 1977. Life history of the Leon Springs pupfish, Cyprinodon bovinus. Copeia, 1977:93-103. King, P. B. 1935. Outline of structural development of Trans- Pecos Texas. American Association of Petroleum Geolo¬ gists, Bulletin, 19:221-261. Koster, W. J. 1939-1955. Field Notes. Museum of Southwest¬ ern Biology, University of New Mexico, Albuquerque. Koster, W. J. 1957. Guide to the fishes of New Mexico. University of New Mexico Press, Albuquerque, 1 16 pp. LaFave, J. I. 1987. Groundwater flow delineation in the Toyah Basin of Trans-Pecos Texas. Unpublished Master’s The¬ sis, University of Texas, Austin, 159 pp. Larson, G L. 1996. Fish community structure in relation to water quality and habitat in the upper Pecos River, Texas. Texas Natural Resource Conservation Commission, Aus¬ tin, Texas, 28 pp. Lee, W. T. 1925. Erosion by solution and fill. U.S. Geological Survey Bulletin, 760: 107-121. Leftwich, W. J. 1987. Tracks along the Pecos. Elliot Printing Co., Pecos, Texas, 74 pp. Lienesch, P. W., W. I. Lutterschmidt, and J. F. Schaefer. 2000. Seasonal and long-term changes in the fish assemblage of a small stream isolated by a reservoir. The Southwestern Naturalist, 45:274-288. Linam, G W. and L. J. Kleinsasser. 1996. Relationship between fishes and water quality in the Pecos River, Texas. River Studies Report No. 9, Resource Protection Division, Texas Parks and Wildlife Department, Austin, Texas, 10 pp. Lingle, R. T. and D. Linford. 1961. The Pecos River Commis¬ sion of New Mexico and Texas. A report of a decade of progress: 1950-1960. The Rydal Press, Santa Fe, New Mexico, 284 pp. Little, R. G 1960a. Periodic check of Carlsbad Municipal Lake. Job Completion Report, Federal Aid Project F-22-R-1, Work Plan 4, Job No. I- 1 , 4 pp. Little, R. G 1960b. Periodic check of Red Bluff Reservoir. Job Completion Report, Federal Aid Project F-22-R-1 , Work Plan 4, Job No. I- 1 , 6 pp. Little, R. G 1961a. Periodic check of Avalon Reservoir. Job Completion Report, Federal Aid Project F-22-R-2, Work Plan 4, Job No. I- 1 , 3 pp. Little, R. G 1961b. Periodic check of fish populations between McMillan and Avalon reservoirs. Job Completion Re¬ port, Federal Aid Project F-22-R-2, Work Plan 4, Job No. 1-1, 5 pp. Little, R. G 1 963a. Periodic check of fish populations between McMillan and Avalon reservoirs. Job Completion Re¬ port, Federal Aid Project F-22-R-4, Work Plan 4, Job No. M, 4 pp. 108 Special Publications, Museum of Texas Tech University Little, R. G 1963b. Routine periodic check of Pecos River fish barrier and trap no. 1. Job Completion Report, Federal Aid Project F-22-R-4, Work Plan 4, Job No. 1-1,4 pp. Little, R. G 1964a. Avalon Canal fish salvage. Job Completion Report, Federal Aid Project F-22-R-5, Work Plan 4, Job No. 1-1, 2 pp. Little, R. G 1964b. Periodic check of Avalon Reservoir. Job Completion Report, Federal Aid Project F-22-R-5, Work Plan 4, Job No. 1-1,3 pp. Little, R. G 1 964c. Periodic check of Carlsbad Municipal Lake. Job Completion Report, Federal Aid Project F-22-R-5, Work Plan 4, Job No. 1-1, 3 pp. Little, R. G 1964d. Resurvey of the Pecos River from the Texas-New Mexico line north to Tansill Dam in Carlsbad, New Mexico. Job Completion Report, Federal Aid Project F-22-R-4, Work Plan 4, Job B-l-A(l), 27 pp. Little, R. G 1965. Resurvey of the Pecos River from the Texas- New Mexico line north to Tansill Dam in Carlsbad, New Mexico. Job Completion Report, Federal Aid Project F- 22-R-6, Work Plan 4, Job B-l-A(3), 6 pp. Mace, R. E., W. F. Mullican III, and E. S. Angle (editors). 2001 . Aquifers of West Texas. Texas Water Development Board Report 356. Texas Water Development Board, Austin, Texas, 272 pp. Maley, V. C. and R. M. Huffmgton. 1963. Cenozoic fill and evaporate solution in the Delaware Basin, Texas and New Mexico. Geological Society of America, Bulletin, 64:539- 545. Mayden, R. L., B. M. Burr, L. M. Page, and R. R. Miller. 1992. The native freshwater fishes of North America. Pp. 827- 863 in: R.L. Mayden (editor). Systematics, historical ecology, & North American freshwater fishes. Stanford University Press, Stanford, California. Mayden, R. L. and C. R. Gilbert. 1989. Notropis ludibundus (Girard) and Notropis tristis (Girard), replacement names for N. stramineus (Cope) and N. topeka (Gilbert) (Teleostei:Cypriniformes). Copeia, 1989:1084-1089. Meinzer, O. E., B. C. Renick, and K. Bryan. 1926. Geology of No. 3 reservoir site of the Carlsbad irrigation project, New Mexico, with respect to water-tightness. U.S. Geo¬ logical Survey Water-Supply Paper 580-A, 39 pp. Miller, R. R. 1977. Composition and derivation of the native fish fauna of the Chihuahuan Desert region. Pp. 365-381 in: R.H. Wauer and D.H. Riskind (editors). Symposium on the biological resources of the Chihuahuan Desert Re¬ gion, United State and Mexico. U.S. Department of the Interior, National Park Service Transactions and Proceed¬ ings Series, No. 3. Minckley, W. L. 1965. Records of atherinid fishes at inland localities in Texas and Northern Mexico. Great Basin Naturalist, 25:73-76. Navarre, R. J. 1959. Pecos River fish barrier and trap no. 1 1958-1 959. Job Completion Report, Federal Aid Project F-10-R-4, Work Plan 4, Job No. Cl-2-2, 24 pp. Navarre, R. J. 1960. Pecos River fish barrier and trap no. 1 1958-1959. Job Completion Report, Federal Aid Project F-10-R-4, Work Plan 4, Job No. C-l-2-3, 25 pp. Platania, S. P. 1990. The ichthyofauna of the Rio Grande drainage, Texas and Mexico, from Boquillas to San Ygnacio. Report to U.S. Fish and Wildlife Service, Ar¬ lington, Texas, 83 pp. Platania, S. P. 1995. Distribution, relative abundance, and con¬ servation status of Pecos bluntnose shiner, Notropis simus pecosensis. Report to New Mexico Department of Game and Fish, Santa Fe, 32 pp. Platania, S.P. 1996. Preliminary systematic and taxonomic analysis of Rio Grande shiner, Notropis jemezanus, and the Notropis ( Notropis ) complex in the Rio Grande basin. Division of Fishes, Museum of Southwestern Biology, Department of Biology, University of New Mexico, Al¬ buquerque, New Mexico, 47 pp. Pope, J. 1 854. Report of exploration of a route for the Pacific Railroad, near the thirty-second parallel of north latitude, from the Red River to the Rio Grande. Explorations and Surveys for a Railroad from the Mississippi River to the Pacific Ocean. War Department. Washington, 50 pp. plus appendices. President’s Water Resources Policy Commission. 1950. Ten rivers in America’s future. No. 4. The Rio Grande. Vol¬ ume 2, The Report of the President’s Water Resources Policy Commission, U.S. Government Printing Office, Washington, 67 pp. Propst, D. L. 1992. Inventory and monitoring of warmwater stream fish communities in New Mexico. Project Comple¬ tion Report 1 July 1989 - 30 June 1992, Federal Aid in Fisheries Project FW-17-RD, 143 pp. Propst, D. L. 1999. Threatened and endangered fishes of New Mexico. Technical Report No. 1. New Mexico Depart¬ ment of Game and Fish, Santa Fe, 84 pp. Rauchenberger, M. 1 989. Systematics and biogeography of the genus Gambusia (Cyprinodontiformes: Poecilidae). American Museum Novitates, No. 295 1 , 74 pp. Rhodes, K. and C. Hubbs. 1992. Recovery of Pecos River fishes from a red tide fish kill. The Southwestern Natu¬ ralist, 37:178-187. Richey, S. F., J. G Wells, and K. T. Stephens. 1985. Geohydrol¬ ogy of the Delaware Basin and vicinity, Texas and New Mexico. U.S. Geological Survey Water-Resources Inves¬ tigations Report 84-4077. Denver, Colorado, 99 pp. Hoagstrom — Historical and Recent Fish Fauna of the Lower Pecos River 109 Robins, C. R., R. M. Bailey, C. E. Bond, J. R. Brooker, E. A. Lachner, R. N. Lea, and W. B. Scott. 1991. Common and scientific names of fishes from the United States and Canada. Special Publication 20, American Fisheries Soci¬ ety, Bethesda, Maryland, 183 pp. Robinson, T. W. and W. B. Lang. 1938. Geology and ground- water conditions of the Pecos River Valley in the vicinity of Laguna Grande de la Sal, New Mexico with special reference to the salt content of the river water. 1 2th and 1 3'h Biennial Reports of the State Engineer of New Mexico, Santa Fe: 77-100. Schroeder, A. H. and D. S. Matson. 1965. A colony on the move, Gaspar Castano de Sosa’s Journal, 1590-1591. Alphabet Printing Co., Salt Lake City, 196 pp. Sharp, J. M., Jr. 2001 . Regional groundwater flow systems in Trans-Pecos Texas. Pp. 41-55, m Aquifers of West Texas (R.E. Mace, W.F. Mullican III, and E.S. Angle, eds), Texas Water Development Board Report 356, Austin, Texas, 272 pp. Smith, M. L. and R. R. Miller. 1986. The evolution of the Rio Grande Basin as inferred from its fish fauna. Pp. 457-485 in: C.H. Hocutt and E.O. Wiley (editors). The zoogeograpy of North American freshwater fishes. John Wiley & Sons, New York. Stevenson, M. M. and T. M. Buchanan. 1973. An analysis of hybridization between the cyprinodont fishes Cyprinodon variegatus and C. elegans. Copeia, 1973:682-692. Sublette, J. E. 1975. A survey of the fishes of the Pecos, Canadian and Arkansas drainages in New Mexico. Re¬ port to New Mexico Department of Game and Fish, Santa Fe, 15 pp. plus appendices. Sublette, J. E., M. D. Hatch, and M. Sublette. 1990. The fishes of New Mexico. University of New Mexico Press, Albu¬ querque, 393 pp. Tanyolac, J. 1973. Morphometric variation and life history of the cyprinid fish Notropis stramineus (Cope). Univer¬ sity of Kansas Museum of Natural History, Occasional Papers, No. 12:1-28. Taylor, T. U. 1902. Irrigation systems of Texas. U.S. Geologi¬ cal Survey Water-Supply Paper 71, 129 pp. Thomas, H. E., T. G McLaughlin, I. J. Winograd, E. D. Gordon, C. S. Conover, and L. J. Bjorklund. 1963. Effects of drought in the Rio Grande Basin, drought in the south¬ west 1942-56. U.S. Geological Survey Professional Pa¬ per 372-D, 59 pp. Thomas, R. G 1972. The Geomorphic evolution of the Pecos River system. Baylor Geological Studies, Bulletin No. 22. Baylor University, Waco, Texas. 40 pp. Trevino-Robinson, D. 1955. The ichthyofauna of the lower Rio Grande river, from the mouth of the Pecos to the Gulf of Mexico. Unpublished Master’s Thesis, University of Texas, Austin, 107 pp. Trevino-Robinson, D. 1959. The ichthyofauna of the Lower Rio Grande, Texas and Mexico. Copeia, 1959:253-256. U. S. Geological Survey. 1900. Operations at river stations, 1 899, a report of the division of hydrography, Part III. U.S. Geological Survey Water-Supply Paper 37, 298 pp. U. S. National Resources Planning Board. 1942. The Pecos River joint investigation, reports of participating agen¬ cies. U.S. Government Printing Office, Washington, 407 pp. West, S. W. and W. L. Broadhurst. 1975. Summary appraisals of the nation’s ground- water resources-Rio Grande re¬ gion. U.S. Geological Survey Professional Paper 81 3-D, 39 pp. Wiebe, A. H., J. G Burr, and H. E. Faubion. 1934. The problem of stream pollution in Texas with special reference to salt water from oil fields. Transactions of the American Fish¬ eries Society, 64:81-86. Wilde, G R. and A. A. Echelle. 1992. Genetic status of Pecos pupfish populations after establishment of a hybrid swarm involving and introduced congener. Transactions of the American Fisheries Society, 121:277-286. Wilde, G R. and K. G Ostrand. 1999. Changes in the fish assemblage of an intermittent prairie stream upstream from a Texas impoundment. Texas Journal of Science, 51:203-210. Winston, M. R., C. M. Taylor, and J. Pigg. 1991. Upstream extirpation of four minnow species due to damming of a prairie stream. Transactions of the American Fisheries Society, 120:98-105. Yates, T. L., M. A. Lewis, and M. D. Hatch. 1 984. Biochemical systematics of three species of catfish (genus Ictalurus ) in New Mexico. Copeia, 1984:97-101. 110 Special Publications, Museum of Texas Tech University Address of author: Christopher W. Hoagstrom U.S. Fish & Wildlife Service New Mexico Fishery Resources Office 2105 Osuna Rd. NE Albuquerque, NM 87113 Current address: Department of Wildlife and Fisheries Sciences South Dakota State University Northern Plains Biostress Building, Room 138 Brookings, SD 57007 e-mail: hoagstrom@brookings. net PUPFISHES OF THE NORTHERN ChIHUAHUAN DESERT: STATUS AND CONSERVATION Anthony A. Echelle, Alice F. Echelle, Salvador Contreras Balderas, andMa. de Lourdes Lozano Vilano Abstract Twelve species of pupfish (Genus Cyprinodori) generally are recognized in the northern Chihuahuan Desert. Eight of these are restricted to relatively small spring systems, whereas the remaining four occur in springs and riverine situations. The present abundance and distribution of pupfishes in the region is only a remnant of what must have been present prior to an¬ thropogenic watershed deterioration and depletion of groundwater. Today, most spring-dwelling pupfishes are succumbing to losses of springflows, primarily as a result of pumping of groundwater, and the riverine species are adversely affected by a diversity of an¬ thropogenic factors. The diversity of both groups has declined as a result of introgressive hybridization with a non-native pupfish, the wide ranging coastal species C. variegatus. The rapidity with which native stocks can be lost as a result of such hybridization is dramati¬ cally illustrated by events following the introduction of C. variegatus into the Pecos River Basin in the 1960s. A similar threat is posed by transport of any non-na¬ tive pupfish into waters occupied by an endemic pupfish. Introduction In this paper, we review the history, current sta¬ tus, and conservation efforts for the 12 species of pupfish (Cyprinodontidae: Cyprinodori) generally rec¬ ognized in the northern Chihuahuan Desert. The ge¬ nus ranges west to east from the Death Valley System to the West Indies and north to south in coastal waters from Cape Cod, Massachusetts to Venezuela. In the past century, seven of the 25 species of Cyprinodon in the arid southwest have been driven to extinction in the wild. The first of these events occurred sometime between 1903 and 1953 when C. latifasciatus of the Parras Basin in Coahuila disappeared as a result of a variety of anthropogenic factors, including destruc¬ tion of springs (Miller, 1964). Most recent extinctions occurred in the 1980s and early 1990s with the loss of five Cyprinodon species in the Sandia and Potosi ba¬ sins of southern Nuevo Leon, all as a result of ground- water pumping and resultant loss of springs (Contreras- B and Lozano-V, 1996), including a complex of small to large springs that was unknown to ichthyologists until 1983 (Lozano-V and Contreras-B, 1993). The seventh known species of Cyprinodon to go extinct in historic times is the Monkey Springs pupfish, which was extirpated as a result of groundwater pumping and other human activities (Minckley, 1973; Minckley et ah, 2002). Although extinct in the wild, three of the seven species are being maintained in various aquarium facilities (Lozano-V and Contreras-B, 1993). In the following accounts, we present a basin- by-basin account of the pupfishes in the northern Chihuahuan Desert, eight of which are restricted to single springs or spring systems, whereas the other four occur in riverine situations and associated springfed waters. Regardless of extent of range, how¬ ever, the pupfishes of the northern Chihuahuan Desert are declining. This reflects a variety of anthropogenic factors, but loss of springs, apparently as a result of groundwater pumping, and genetic introgression by non-native pupfish are the most immediate causes of concern. 112 Special Publications, Museum of Texas Tech University Pecos River Basin Cyprinodon bovinus (Leon Springs pupfish). — The waters supporting C. bovinus occur in Diamond Y Draw (= Leon Creek in some publications), a “flood” tributary of the Pecos River that has rarely, if ever, reached the river in historic times. Until 1965, the spe¬ cies was unknown to science except for the original collection, in 1 85 1 , of 1 6 specimens from Leon Springs near the present Fort Stockton, Pecos County, Texas. In 1965, W. L. Minckley found it approximately 15 km downstream of Leon Springs. In the intervening time, an area immediately downstream of, and fed by, Leon Springs had been dammed (about 1910) and the resulting “Lake Leon” was stocked with carp and game fish. In 1938, Carl Hubbs failed to collect the species in the Leon Springs area. By 1958, Leon Springs had gone dry because of over-pumping of the aquifer (Brune, 1975). The species was declared extinct by the late 1950s (Hubbs, 1957; Miller, 1961). Its redis¬ covery was verified in a re-description of the species (Echelle and Miller, 1974), and it was listed as a feder¬ ally endangered species in 1980, with most of the oc¬ cupied areas designated as critical habitat (Federal Register, 45:54678). Since 1990, most of the area oc¬ cupied by C. bovinus has received protection as a pre¬ serve of 607 hectares (Diamond Y Spring Preserve) managed by The Nature Conservancy of Texas. Since its re-discovery, C. bovinus has occupied two separate systems of surface water separated by 1-2 km of dry land: an “upstream watercourse” re¬ ceiving flow from Diamond Y Spring and several smaller springs and seeps, and a “downstream water¬ course” receiving flow from Euphrasia Spring, sev¬ eral small seepage springs, and groundwater seepage. Both of the two watercourses are 2-4 km long, with length of watercourse varying considerably over the years. The primary threats to C. bovinus include pollu¬ tion, loss of habitat, and introgressive hybridization with an introduced congener. Both watercourses sup¬ porting C. bovinus occur in the Fort Stockton Oil and Gas Field, an area of intense petrochemical extraction activity for about 50 years. Thus, there is a continuing threat of pollution (Kennedy, 1977; Gehlbach, 1981). In 1974, the Soil Conservation Service (Natural Re¬ sources Conservation Service) constructed an earthen berm that protects the headpool of Diamond Y Spring from surface spills (Hubbs 1980). In 1992, a ruptured pipeline released crude oil in a nearby area, creating sufficient concern that the oil company dug a trench downslope of the spill to help slow contamination of the watercourse. In recent years, both watercourses supporting C. bovinus have been 1-2 km shorter than they were in the 1970s (Hubbs et al., 1978; Echelle and Echelle, 1980). This may be the combined result of the drought conditions in the region during the past five years and continued over-pumping of groundwater in the basin. In addition to the reduced size of the watercourses, densities of pupfish now seem lower than they were in the past. In the 1970s, the pupfish was abundant in a diversity of open-water situations in Diamond Y Draw (Kennedy, 1977; Hubbs et al., 1978; Echelle and Echelle, 1980). Reduced densities seem to reflect a loss of open-water habitat as a result of encroachment by bulrush because of reduced water flow. The problems posed by introductions of non¬ native pupfish emerged as the most important imme¬ diate threat to C. bovinus shortly after R. D. Suttkus collected C. variegatus from the lower watercourse in 1974. By January 1976, hybrid morphotypes occurred throughout the lower watercourse. This led to an in¬ tensive effort to eliminate hybrids, with some atten¬ tion to protecting the invertebrate community (Hubbs et al., 1978). The effort included treatment of the lower watercourse with rotenone, re-introducing C. bovinus from the upper watercourse, and subsequent seining and selective removal of suspected hybrids (Hubbs, 1980). The absence of morphological traits indicating hybrids (Hubbs, 1980) and the absence of alleles of C. variegatus in a genetic survey of the population (Echelle et al., 1987) indicated that the renovation was suc¬ cessful in restoring C. bovinus to the lower water¬ course (for a more detailed review see Minckley et al., 1991). In retrospect, it was extremely fortunate that, during the 1976 renovation, C. bovinus from the up¬ per watercourse was used to establish a captive stock at Dexter National Fish Hatchery and Technology Cen¬ ter in New Mexico (DNFH). EcHELLE ET AL. - PUPFISHES OF THE NORTHERN CfflHUAHUAN DESERT 113 A second introduction of C. variegatus into Dia¬ mond Y Draw in the late 1980s or early 1990s led to contamination of both the upper and the lower water¬ course of Diamond Y Draw (Echelle and Echelle, 1997). In response, the U.S. Fish and Wildlife Service (FWS) approved and, with help from various agencies and the Rio Grande Fishes Recovery Team, implemented a plan to restore the native pupfish genome. The reno¬ vation occurred in two phases that differed in approach. Antimycin A was used to eliminate all fish (including two non-natives, Gambusia geiseri and C. carpio) from the Diamond Y Spring outflow in August 1998. Prior to the renovation, large samples of the native fishes G. nobilis and L. parva were removed and transported alive to DNFH; after dissipation of the toxin, they were released back into the watercourse. Similar precau¬ tions were taken to protect known invertebrate spe¬ cies of concern. To provide additional protection for the remainder of the fauna, two small areas support¬ ing pupfish were left untreated; one of these was con¬ taminated by C. variegatus, but the population seemed sufficiently small that it was decided to try diluting the introgressed genome by releases of pure C. bovinus from DNFH. Renovation of the downstream water¬ course was initiated in March 2000 and involved re¬ moval of all pupfish captured by intensive seining, dip¬ netting, and trapping. Both watercourses received large numbers of pupfish from the captive DNFH stock of C. bovinus immediately after renovation and in the fol¬ lowing year. Subsequent genetic surveys indicated that the renovation efforts were largely successful in both watercourses (AAE and AFE, unpublished data). There was no evidence of introgression in the upper water¬ course except for the small population not treated with ichthyotoxins. Levels of introgression in the lower watercourse were reduced to possibly acceptable lev¬ els. Further releases of C. bovinus from the DNFH stock are planned for the future, with emphasis on the known areas of persistent contamination. Since establishing the Diamond Y Preserve, The Texas Nature Conservancy (TNC) has attempted to restore the watershed to more natural conditions. Their activity includes renovation of old oil/gas pads, man¬ agement of livestock to reduce their impact on the aquatic system, and using tractor equipment in 2000 to uproot all salt cedars ( Tamarix sp.) in the lower watercourse. TNC is searching for appropriate ways to improve conditions for the pupfish without com¬ promising the habitat for other rare, native species. Cyprinodon pecosensis (Pecos pupfish). — The geographic range of C. pecosensis once included sa¬ line floodplains and springs, lakes, gypsum sinkholes, and other waters associated with the Pecos River from Bitter Lake National Wildlife Refuge and Bottomless Lakes State Park near Roswell, New Mexico, down¬ stream for about 650 river-km to the mouth of Inde¬ pendence Creek in Texas (Echelle and Echelle, 1978). The present distribution represents less than 20% of the historic range. The only known natural population remaining in Texas (but see below) is in a portion of Salt Creek, a saline tributary of the Pecos River just south of the New Mexico border. In New Mexico, the species occurs most abundantly in saline waters of the Pecos River floodplain near Roswell, primarily on the Bitter Lake National Wildlife Refuge and Bottom¬ less Lakes State Park. There is only sketchy knowledge of the distribu¬ tion and abundance of Pecos pupfish prior to the ex¬ tensive habitat alteration that had occurred by 1950, but anthropogenic factors have undoubtedly caused a considerable loss of habitat for the species. The fac¬ tors contributing to such losses were reviewed by Hoagstrom and Brooks (1999). Native Americans used water from the Pecos River for agriculture in the head¬ waters of the river prior to 1 600, but human-induced alterations of aquatic habitats in the basin probably escalated considerably in the late 1800s. Since that time, a variety of anthropogenic factors drastically re¬ duced the habitat available to pupfish. These include construction of dams on the mainstem of the river and tributary streams, introduction of salt cedar, over¬ pumping of aquifers, over-grazing by domestic ani¬ mals, erosion, pollution, and draining of wetlands (Hoagstrom, this volume). An apparently indirect ef¬ fect of human activity are fish kills that have occurred sporadically in the lower Pecos River in Texas since the 1950s, including several kills from 1985 to 1989 that extended over hundreds of river kilometers (Rhodes and Hubbs, 1992; Childs et al., 1996). The latter series of kills apparently resulted from toxins released during blooms of a chrysophyte alga ( Prymnesium parvum) that probably reflect a response to anthropogenic nutrient enrichment (Rhodes and 114 Special Publications, Museum of Texas Tech University Hubbs, 1992). Although not documented in detail, such kills undoubtedly resulted in temporary depletions of the pupfish population. Until recently, and despite losses of habitat and extensive fish kills, C. pecosensis seemed reasonably secure because of its relatively large range and locally high abundances. However, the status of the species changed abruptly with the introduction, sometime be¬ tween 1980 and 1984 (Echelle and Connor, 1989), of C. variegatus into the basin, probably in Red Bluff Reservoir on the New Mexico/Texas boundary (Childs et al., 1996). Collections of pupfish from the Pecos River at four Texas localities in March 1980 showed no morphological evidence of influence by C. variegatus. But less than five years later, in August 1984, collections of six specimens each from two sites separated by about 200 river-km comprised C. pecosensis x variegatus hybrids (Echelle et ah, 1987). The genotypes for allozyme loci indicated that samples consisted of individuals that were minimally second- generation hybrids. A broader geographic survey in 1985 detected locally panmictic hybrid populations throughout ap¬ proximately 430 river-km of the Pecos River in Texas, where, depending on locality, alleles typical of C. variegatus represented 18 to 84 percent of the ge¬ nome (Echelle and Conner, 1989). Subsequent moni¬ toring revealed that hybrids were ubiquitous in the Pecos River and peripheral waters (reservoirs, irriga¬ tion canals, and gravel pits) in an area that extended downstream from near Loving, New Mexico to at least the vicinity of Pandale, Texas, approximately 55 km downstream of the historic range of C. pecosensis (Wilde and Echelle, 1992, 1997; Echelle et ah, 1997; AAE and AFE, unpubl. data). Childs et ah (1996) sug¬ gested that the genetic structure of the hybrid swarm is explained by genetic swamping, possibly mediated by selection for C. variegatus or C. pecosensis x variegatus hybrids during a period of increasing popu¬ lation size, such as those that would have followed the fish kills mentioned earlier in this account. The role of selection is being confirmed experimentally by J. Rosenfeld (pers. comm.) in A. Kodric-Brown’s labo¬ ratory at the University of New Mexico. Until recently, the population in Salt Creek, a tribu¬ tary of the Pecos River near the New Mexico/Texas boundary, was considered effectively free of contami¬ nation except near the mouth of the creek where there were low frequencies of alleles typical of C. variegatus. However, in March 2001, morphological and genetic evidence of hybrids extended approximately 19 river- km upstream in Salt Creek (AAE and AFE, unpubl.). Pure populations of Texas stocks of C. pecosensis ex¬ ist only in two artificial ponds supporting stocks trans¬ planted from Salt Creek (Garrett, this volume) and, as of March 2001, in a headwater reach of Salt Creek (AAE and AFE, unpubl. data). Endangered species status was proposed for C. pecosensis in 1998, primarily because of the threat from hybridization (Federal Register, 63:4608). In 2000, the proposal was withdrawn (Federal Register 65:14513) in response to a Conservation Agreement between the states of New Mexico and Texas, the FWS, and the U.S. Bureau of Land Management (Federal Register, 65:71424). In this agreement, “The signatory agen¬ cies . . . made commitments to protect known extant populations of pure Pecos pupfish, expand the distri¬ bution of the species within its native range by estab¬ lishing new populations, and to prohibit the use of . . .” C. variegatus as baitfish in the Pecos River area. To date, various measures have been taken to protect populations, including, among other proactive measures, constructing fish barriers in two locations, initiating a study of the life history of the pupfish, en¬ acting the necessary baitfish regulations, and estab¬ lishment of two captive populations of the Texas stock in artificial ponds (Garrett, this volume). Cyprinodon elegans (Comanche Springs pupfish). — This pupfish is known only from springfed systems in two separate flood tributaries of the Pecos River in Trans-Pecos Texas, Comanche Draw and the Toyah Creek Basin. Unlike the other two pupfishes endemic to the Pecos River drainage, which occur in saline to moderately saline waters, C. elegans is known only from relatively fresh waters (about 1-3 ppt total dissolved solids). The species was described from 32 specimens taken in 1851 at Comanche Springs, Fort Stockton, Texas. Because of over-pumping of ground- water, the six large springs of the Comanche Springs complex were dry by 1956 (Brune, 1981; Scudday, this volume) and the morphologically divergent (Echelle, 1975) local population of C. elegans was extirpated (Hubbs, 1957; Miller, 1961). ECHELLE ET AL. - PuPFISHES OF THE NORTHERN ChIHUAHUAN DESERT 115 The species still exists about 90 km to the west in the Toyah Creek Basin near Balmorhea and Toyahvale, Reeves and Jeff Davis counties, where there is a system of three large artesian springs (Phan¬ tom Lake, San Solomon, and Giffin springs) and smaller springs that feed the irrigation canals for Reeves County Water Improvement District No. 1 (Garrett and Price, 1993; Garrett, this volume). Genetic and morphological studies indicate that the Phantom Lake Spring population is divergent from the populations in waters fed by Giffin and San Solomon Springs (Echelle, 1975; Echelle et al., 1987). The species occurs in a small segment (< 1 km) of Toyah Creek, a large, swimming pool fed by San Solomon Spring at Balmorhea State Park, and two semi-natural refugia at the park. The remainder of the species is almost entirely confined to a system of earthen and concrete irrigation canals that is fed pri¬ marily by the artesian springs and serves approximately 2428 ha of agricultural land (LaFave and Sharp, 1987). During cooler months of the year, much of the flow is diverted into Lake Balmorhea, an artificial reservoir. The canal system supporting C. elegans extends through an area about 3 to 4 km wide and 1 5 km long, but the species is primarily restricted to areas of more permanent water in the main canals, in sections where the current is slower and the substrate more heterog¬ enous. The springs now supporting C. elegans origi¬ nally would have fed large, marshy habitats (cienegas) that drained into Toyah Creek. However, such cienegas would have been eliminated by the development of the system of irrigation canals for agriculture (Garrett, this volume). Traces of canals built by Native Americans occur in the vicinity of San Solomon Springs (Brune, 1975, 1981), but large-scale diversion of springflows probably started in the 1870s when the area was de¬ veloped for agriculture to supply the military at Fort Davis (Young et al., 1993). Such activity culminated with the amalgamation, in 1914, of local canal compa¬ nies into the Reeves County Water Improvement Dis¬ trict No.l and reconstruction of the canal system by the U.S. Bureau of Reclamation in 1946 (Young et al., 1993). A variety of other fishes are known from the area, including at least eight introduced species, most of which are restricted primarily to Lake Balmorhea where they were released either to support the sport fishery or as accidents associated with release of gamefishes. Since the 1960s, Lake Balmorhea has sup¬ ported a dense, non-native population of Cyprinodon variegatus, probably as a result of accidental trans¬ port (Stevenson and Buchanan, 1973). The presence of an introduced population of C. variegatus in Lake Balmorhea threatens existing popu¬ lations of C. elegans with both hybridization and com¬ petition for resources. A hybrid zone between the two occurs in an earthen canal where Lake Balmorhea re¬ ceives flow from the irrigation system (Stevenson and Buchanan, 1973; A. F. Echelle and Echelle, 1994). Upstream migration of C. variegatus from the reser¬ voir into most of the spring system is precluded by physical barriers. However, in the summer of 1988, C. variegatus had moved by way of a recently dug canal from the vicinity of the lake into East Sandia Spring, a small spring near the periphery of the spring- system in the region. A large sample of the introduced species was taken from the headpool of that spring in July 1988 (A. F. Echelle and Echelle, 1994), but some¬ time after that, and for unknown reasons, C. variegatus disappeared from the spring. To our knowledge there is no historical record of C. elegans from the headpool of East Sandia Spring, but on 20 March 2001 two specimens were found just downstream of a small culvert in the outflow. There is some indication of male sterility among hybrids in Lake Balmorhea, but the pres¬ ence of backcross progeny demonstrates that genetic introgression by C. variegatus is a serious concern should C. variegatus gain access to areas outside of the lake (A.F. Echelle and Echelle, 1994). In 1998, Texas Parks and Wildlife Department expended con¬ siderable effort to eliminate C. variegatus from the lake; although more than 5 million C. variegatus were initially eliminated, the fish quickly reestablished in the lake (Garrett, this volume). The ultimate threat to C. elegans is habitat loss due to declining springflows. Periodic losses of pup fish due to management of the irrigation canals supporting the species (Davis, 1979) is a relatively trivial matter. Some springs in the Balmorhea-Toyahvale area have gone dry and flows from all springs in the area have declined since the early 1900s (Brune, 1981; A. F. Echelle et al., 1989). The two largest of the existing springs, Phantom Lake and San Solomon springs, showed a steady decline in recent decades (A. F. Echelle et al., 1989; Schuster, 1997) and surface flow from the former ceased in 2000 (N. Allan, pers. comm.). In 116 Special Publications, Museum of Texas Tech University 2001, the Bureau of Reclamation and the U.S. Fish and Wildlife Service installed a pump to maintain the small pool at the head of Phantom Lake Spring, but this is a short-term solution that is unlikely to support the fish populations if water levels continue to decline (N. Allan, FWS, pers. comm.). Cyprinodon elegans has been a federally listed endangered species since 11 March 1967 (Federal Register, 32:4001). Three aquatic refugia have been constructed for C. elegans and other endemic forms. Two using flows from San Solomon Spring were con¬ structed at Balmorhea State Park by the Texas Parks and Wildlife Department; one, constructed in 1974, is a meandering, slow-flowing channel about 120 m long (Echelle and Hubbs, 1978) and the other is an artifi¬ cial, 1-ha cienega constructed in 1996 (Garrett, this volume). A refugium at Phantom Lake Spring was built in 1993 through the cooperation of several state and federal agencies. For this refugium, some of the water emerging from Phantom Cave, a hillside cavern, is di¬ verted from an irrigation canal into an artificially con¬ structed channel and side-pool habitat that is about 110 m long (Young et al., 1993). In the Phantom Lake Spring refugium, the abundance of the pupfish peaked rapidly in the first year and then declined somewhat with increasing growth of vegetation, possibly as a result of the elimination of open patches of bottom substrate for spawning (Winemiller and Anderson, 1997). This refugium has been ineffective since 1999 as a result of the loss of springflows. However, the two refugia at Balmorhea State Park support large popu¬ lations of C. elegans in semi-natural settings (Garrett, this volume). Captive stocks of the Phantom Lake Spring population have been maintained at the Uvalde National Fish Hatchery in south-central Texas since 1990. Rio Conchos-Middle Rio Grande Cyprinodon eximius (cachorrito del Conchos, Conchos pupfish). — Cyprinodon eximius comprises at least four forms that may deserve taxonomic recogni¬ tion (Miller, 1976). These include populations in the Rio Conchos and Rio Sauz (Chihuahua), a form in Devils River (Val Verde County, Texas), and a form in Rio Grande tributaries upstream of Devils River (Presidio and Brewster counties, Texas, and Chihua¬ hua). A protein electrophoretic survey provided ge¬ netic support for the distinctiveness of the Devils River population from other populations of the species (the Rio Sauz population was not examined). The Devils River population was fixed for unique alleles at three of 30 loci examined; a population from Alamito Creek, a more upstream tributary of the Rio Grande, was allozymically similar to two samples from the Rio Conchos (Echelle and Echelle, 1998). An undescribed pupfish in Ojo de Villa Lopez, an isolated spring near the Rio Florido (Rio Conchos drainage), may deserve species-level recognition (Contreras-B, 1991), but in overall appearance it resembles C. eximius and is ge¬ netically similar to that species (Echelle and Echelle, 1998). The various forms of C. eximius occupy a vari¬ ety of habitats ranging from constant temperature springs to eurythermal marshes and riverine situations. The riverine forms can occur in relatively diverse as¬ semblages of native fishes. For example, samples taken in 1901 from the Rio Chihuahua at Chihuahua City, and from the Rio Conchos at Camargo, Chihuahua, Mexico produced 12 and 16 species, respectively (Contreras-B, 1977), and, in 1994-1995, the species was taken with 8 to 19 species at four sites in the Rio Conchos Basin (Edwards et al., 2001; Edwards et al., this volume). Although general collections from a site can produce a diversity of fishes, the pupfish tends to occupy shallow, quiet waters with relatively few spe¬ cies (Davis, 1980; Valdes-Cantu and Winemiller, 1997). Possibly because of differences in fish assemblage complexity, the species is rare in the mainstem of the Rio Grande and more common in tributaries. In a se¬ ries of 1 8 collections downstream of the mouth of the Rio Conchos, Hubbs et al. (1977) found only a single specimen in the mainstem and this was at a site just downstream of Alamito Creek, where the species is more abundant. In a 1 994- 1995 survey of 1 1 localities in the Rio Conchos Basin, Edwards et al. (2001) found C. eximius abundant at a site in the Rio Chuviscar and present, but less abundant, at one location each in the Rio Conchos and two tributaries, Rio San Pedro and Rio Santa Isabela. The Rio Florido was dry at sites visited EcHELLE ET AL. — PlJPFISHES OF THE NORTHERN CfflHUAHUAN DESERT 117 during that survey, whereas, in 1989, a collection of C. eximius was taken from that river near Villa Lopez (Oklahoma State University Collection of Vertebrates, catalog number 18240). Populations associated with tributaries of the Rio Grande between the Rio Conchos and Devils River are restricted primarily to the down¬ stream termini of small streams, although specimens are occasionally taken from associated waters in the Rio Grande (Hubbs et al., 1977) . The tributary streams are particularly vulnerable to habitat loss from dewa¬ tering as a result of upstream impoundments or pump¬ ing of groundwater. Construction of Amistad Reser¬ voir in the 1960s inundated most sites of historic oc¬ currence for the population in Devils River, and it was considered extirpated until its rediscovery in an 1 1 -km reach of Devils River at the headwaters of the reser¬ voir (Davis, 1980). In 1979, specimens from this area were transported to Dolan Creek in a successful ef¬ fort to re-establish the population in Dolan Springs, a small (< 1 km) springfed area of historic occurrence in the Devils River drainage approximately 25 km up¬ stream of the reservoir (Hubbs and Garrett, 1990). The population of C. eximius in the Rio Sauz Basin once was considered extinct because of drying of the habitat (Miller, 1961), but it was collected in 1964 from an impounded, possibly springfed, pond 75 km S of El Sueco, Chihuahua (Minckley and Koehn, 1965). In 1968 and 1975 it was collected in Estacion Sauz and springs near Laguna Encinillas, but, in 1 995 the former locality was dry, as was most of the sur¬ rounding area (SCB, unpubl.). The present status of this population is unknown. Cyprinodon macrolepis (cachorrito escamudo, largescale pupfish). — This species is endemic to a rather large springfed pool, Ojo de la Hacienda Dolores, and its outflow, 12.5 km S-SW of Jimenez, Chihua¬ hua, Mexico. Miller (1976) noted that the outflow “no doubt” once connected with the Rio Florido, which supports C. eximius. Now the outflow has been highly modified into a number of irrigation distributaries and is isolated from the river. There is evidence from pro¬ tein electrophoresis for past hybridization and genetic interaction between C. macrolepis and the population of C. eximius in the Rio Florido (Echelle and Echelle, 1998), but the two species have maintained marked differences in color pattern and morphology (Miller, 1976). During our visit in 1989, the species was com¬ mon in the spring and the outflow immediately down¬ stream of the headpool. The spring headpool is a lo¬ cally popular recreation area and has been modified for swimming, with the edges shored up by rock walls. We are unaware of any analysis of the trend of the hydrograph for the springs. Cyprinodon pachycephalus (cachorrito cabezon, bighead pupfish). — Two populations of “big-headed” pupfish occur in separate thermal-spring systems of the Rio Conchos Basin (Minckley and Minckley, 1986): one described as C. pachycephalus in Banos (= Ojo) de San Diego, a small springfed system tributary to the Rio Chuviscar, 57 km E of Ciudad Chihuahua, Chihuahua, Mexico, and an undescribed population in a spring near the Rio Conchos at Julimes, 22 km SSE of Banos de San Diego. The taxonomic status of the latter population has not been determined, but they may represent the same species (Minckley and Minckley, 1986). Banos de San Diego was described by Smith and Chemoff (1981) and Minckley and Minckley (1986). It comprises a small system of thermal springs that emerges on a small hilltop, and, prior to human alter¬ ations, must have emptied into the nearby Rio Chuviscar (Minckley and Minckley, 1986). The sys¬ tem includes several small, commercially operated swimming pools and baths. Outlets from the springheads coalesce to form a small, partially braided, stream channel about 2 m wide and less than 3 cm deep. The spring run has been widened in places to form bathing pools, and, in 1971, it emptied into an artificial pond adjacent to the Rio Chuviscar. How¬ ever, by 1980 the spring outflow was diverted into a canal for irrigation (Smith and Chemoff, 1981). Water temperature is 43.8°C to 44.0°C in the springheads and cools as it moves downstream. The pupfish oc¬ curs throughout the springfed system alongside a pos¬ sibly undescribed species (Contreras-B, 1991) related to the blotched gambusia ( Gambusia senilis). Gonadal condition and other observations led Smith and Chemoff (1981) to conclude that C. pachycephalus and the local form of Gambusia are not stressed by the high temperatures and that the two exhibit “the highest long-term temperature tolerances known for any teleost.” 118 Special Publications, Museum of Texas Tech University The only other fishes reported from the Banos de San Diego system are the more widespread pupfish C. eximius, a few putative C. eximius x C. pachycephalus hybrids, and longear sunfish Lepomis megalotis, all of which were restricted primarily to the tail water pond in 1971 (Minckley and Minckley, 1981); none of these were reported by Smith and Chemoff (1981) from their visit in 1980. However, collections from the tailwater pond by one of us (SCB) in 1982 and by his student, Hector Leal Sotelo, in 1984 in¬ cluded 16 species, including two non-natives, Cyprinus carpio and Ameiurus melas. The status of this species is precarious because of its restricted habitat and the general trend toward loss of springflows as a result of over-pumping of groundwater in arid regions of the southwest (Contreras-B and Lozano-V, 1 994). Human modifica¬ tions of the habitat did not seem to threaten the spe¬ cies in 1989, when AAE and AFE visited Banos de San Diego. However, the present status of the species and the undescribed Julimes population is not well known. Tularosa Basin Cyprinodon tularosa (White Sands pupfish). — Cyprinodon tularosa is restricted to the Tularosa Ba¬ sin in New Mexico where it occupies four isolated bodies of water, three on the White Sands Missle Range (Malpais Spring, Mound Spring, and Salt Creek) and one on Holloman Air Force Base near Alamogordo, Otero County. The Mound Spring and Lost River popu¬ lations both represent translocations by local people in the late 1960s and early 1970s (Pittenger and Springer, 1999). Genetic analyses indicate that Salt Creek was the source for both introductions (Stockwell et al., 1998). Pittenger and Springer (1999) determined from historical records that Malpais Spring and Salt Creek have been modified by human activities, but Springer (FWS, pers. comm.) suggests that this might not have been associated with notable loss of habitat for the pupfish. The present, sharply incised nature of upper Salt Creek apparently occurred as a result of over- grazing and gully erosion sometime after the late 1 800s (Pittenger and Springer, 1999; Springer, FWS, pers. comm.). Habitats of C. tularosa show a wide range in salinity (1.5-60 ppt), temperature (3.0°C to 33.4°C) and environmental stability (Stockwell and Mulvey, 1998). Stockwell and Mulvey (1998) demonstrated a correlation between salinity and genotype at an allozyme locus that appears to reflect adaptation to salinity dif¬ ferences among habitats. In addition, parasitic infec¬ tion in the White Sands pupfish is apparently a func¬ tion of salinity (Stockwell et al., 1998). Physa, the intermediate host of diplostome trematodes, cannot tolerate salinities above 9 ppt. As a result, pupfish in the relatively fresh Malpais Spring were highly infested (up to 100%) with white grub {Posthodiplostomum minimum), whereas the pupfish in Salt Creek were free of this parasite. No other fish species occur in habitats support¬ ing C. tularosa (Miller and Echelle, 1975). However, isolated waters in the vicinity of present populations do support introduced species: western mosquitofish ( Gambusia affinis ) at three sites, largemouth bass ( Micropterus salmoides) at two, and goldfish ( Carrasius auratus) at one (Pittenger and Springer, 1999). Cyprinodon tularosa is considered a federal spe¬ cies of concern (Federal Register, 50:64481). Malpais Spring (266-ha wetland), Salt Creek (33-km stream) and Lost River (5 -km stream) all support substantial pupfish populations, whereas the population in Mound Spring (two small ponds) is relatively small (Pittenger and Springer, 1999). Threats include introduction of exotic species, dewatering, and pollution, as well as disruption of the habitat by feral horses and off-road vehicle use. A conservation team organized in 1994 recommended establishment of additional populations in the Tularosa Basin. Stockwell et al. (1998) recom¬ mended that the Malpais Spring and Salt Creek forms be treated as separate units of conservation because loss of either one would result in a marked decrease in genetic diversity of the species. As previously men¬ tioned, the Salt Creek form is represented in Mound Spring and Lost River. The Malpais Spring population, however, apparently has not been replicated. ECHELLE ET AL. - PUPFISHES OF THE NORTHERN CfflHUAHUAN DESERT 119 Laguna De Guzman Basin Cyprinodon fontinalis (cachorrito de Carbonera). — Cyprinodon fontinalis is known only from a series of five springs and their outflows near Ejido Rancho Nuevo in the Bolson de los Muertos of the Guzman Basin in northwestern Chihuahua. The springs are described in some detail and mapped by Smith and Miller (1980, 1981). They are separated by a maximum distance of only about 5 km. Ojo de Carbonera, the only spring habitat relatively unmodi¬ fied by humans, includes a complex of spring-heads and flows about 100 m (less than 10 cm deep) before entering irrigation canals. In this system, C. fontinalis occurs most abundantly in small solution holes and along undercut banks of the spring outflow. The spe¬ cies also occurs in irrigation ditches and in four im¬ pounded springs. The only native fish sympatric with C. fontinalis is the largemouth shiner ( Cyprinella bocagrande ), which is known only from Ojo Solo (Chemoff and Miller, 1982), one of the five major springs supporting the pupfish. Black bullhead ( Ameiurus melas ) and western mosquitofish ( Gambu - sia affinis ) have been introduced into the area (Smith and Miller, 1980). The species probably declined after human modi¬ fications to the habitat, including introductions of mosquitofish ( Gambusia affinis ), black bullhead (Ameiurus melas), and largemouth bass ( M . salmoides), and construction of irrigation ditches and small impoundments of the springs (Smith and Miller, 1981). In the late 1970s, pupfish were more abundant in the relatively undisturbed Ojo de Carbonera than elsewhere in the area (Smith and Miller, 1980). The species was abundant at Ojo de Carbonera when we visited the area in 1989. An important threat is the pos¬ sibility for increased pumping of groundwater for irri¬ gation, a factor that has contributed to failure of springflows in nearby areas (Smith and Miller, 1981). Cyprinodon albivelis (cachorrito dorsal blanca, whitefm pupfish). — This species occurs in the west¬ ern highlands of Chihuahua, primarily in the headwa¬ ters of the Rio Papigochic, a tributary of the Rio Yaqui on the Pacific versant, and in a small springfed situa¬ tion, Ojo de Arrey, in the Rio Santa Maria sub-basin of the Laguna de Guzman Basin. There is some question regarding whether the latter population is native or a result of a modern introduction from the Rio Papigochic, but Minckley et al. (2002) suggested, on the basis of genetic considerations (Echelle and Dowling, 1992; Echelle and Echelle, 1993a) and gill raker counts, that it probably is a native population. The species was widespread and abundant in the Rio Papigochic system in 1978 (Hendrickson et al., 1981) and in 1982 and 1986 (SCB and MLLV). In 1989, it was locally abundant at a site in the Rio Papigochic and at two different springs in the Ojo de Arrey system (AAE and AFE), and, according to Minckley et al. (2002), P. J. Unmack found the latter population “intact in 1 999.” Minckley et al. (2002) noted that “. . . local stocks . . . [probably] have disappeared due to increased human activities, . . . [but they had] no reason to believe the species is as yet in jeopardy.” Cyprinodon pisteri (cachorrito de Guzman, Guzman pupfish). — This species (= “Palomas pupfish” in some publications; e.g., Echelle and Echelle, 1998) occurs relatively widely in the Guzman Basin, and ap¬ parently introduced populations once occurred just across the U.S. /Mexico boundary in New Mexico (Minckley et al., 2002). The Lago de Guzman com¬ plex comprises the rios Casas Grandes, Santa Maria, del Carmen (= Rio Santa Clara), and the Laguna Bustillos Basin. In March 1990, Propst and Stefferud (1994) surveyed these basins for the occurrence of Chihuahua chub (Gila nigrescens ), and collected the pupfish at one or more sites in each, except that the species was absent from their nine sample sites in the Rio Santa Clara. Minckley et al. (2002) noted that “the species is catholic in habitat, occupying springs, marshes (cienegas), shorelines and cutoff channels of rivers and creeks, even colonizing ephemeral canals and ditches along roadsides.” Anthropogenic factors such as habitat destruc¬ tion, degradation and fragmentation, pollution, and nonnative species have adversely affected native aquatic communities over a large portion of the Guzman Basin (Propst and Stefferud, 1994). Minckley et al. (2002) observed that, although the Guzman pupfish is relatively widespread, losses have occurred with the drying of a diversity of aquatic habitats at lower elevations in the Guzman Basin. They cite Brand (1937) as follows: “The increasing use of spring and river water for irrigation in the haciendas and colonias 120 Special Publications, Museum of Texas Tech University of the region has contributed markedly to the lessened flow of the rivers in their lower courses. Many of the abundant springs that fed this system ~80 years ago have failed.” However, some losses of habitat, includ¬ ing, by about 1975, the type locality, a springfed cienega near Las Palomas, are attributable to over-pumping of groundwater on both sides of the international bound¬ ary (Minckley et al., 2002). CUATRO ClENEGAS Cyprinodon atrorus (cachorrito del Bolson, bandfin pupfish). — Cyprinodon atrorus occurs prima¬ rily in physicochemically variable, often ephemeral habitats in the Bolson de Cuatro Cienegas de Carranza (Cuatro Cienegas Basin), Coahuila, Mexico. However, the species also occurs in stable, springhead environ¬ ments in the southeastern end of the Cuatro Cienegas Basin, where the other pupfish of the basin (C. bifasciatus ) is absent (Arnold, 1972). The two Cuatro Cienegas pupfishes hybridize in areas where they come into contact either naturally or because of human-con¬ structed irrigation canals (Miller, 1968; Minckley, 1977). However, such hybridization is restricted to local zones of contact and, outside these zones, the two species are maintaining their morphological and genetic dis¬ tinctiveness (Echelle and Echelle, 1998; E. Carson, pers. comm.). The species can be locally very abundant, but is considered rare because of its restricted distribution (Williams et al., 1989). The amount of habitat available for pupfish must have been much more extensive at the turn of the century, when Cuatro Cienegas was a closed basin with no outflow (Rodriguez Gonzales, 1926, as cited by Calegari, 1997). Now, however, ca¬ nals transport water outside the valley (Minckley, 1969, 1977). Canalization of springs and their outflows for agriculture and industry, mostly outside the basin (Contreras-B, 1991), undoubtedly has reduced the amount of habitat available to this fish. In November 1994, the federal government designated Cuatro Cienegas as a National Protected Area, affording some protection for the system. The species remains abun¬ dant, despite a long history of human activity in the basin, including agriculture, gypsum mining, tourism, and a recent increase in manufacturing plants (Calegari, 1997). Uncontrolled tourism and continued economic pressures on the human population pose a variety of threats (Contreras-B, 1991; Calegari, 1997), and there is some concern that groundwater pumping is threat¬ ening spring flows by causing the water table to fall (Grail, 1995). The ultimate impact of introduced spe¬ cies, including African cichlids ( Oreochromis sp. and Hemichromis guttatus ), water hyacinth ( Eichornia crassipes), and Asian snail ( Melanoides tuberculata), remains to be seen. Cyprinodon bifasciatus (cachorrito de Cuatro Cienegas, twoline pupfish). — Cyprinodon bifasciatus occupies an arc of constantly warm springs and their outlet pools and streams in the valley floor around the northern tip of Sierra de San Marcos, an area of karst topography where there are cenote-like, springfed sink¬ holes (“pozas”) that range up to 200 m in diameter and more than 10 m deep (Miller, 1968; Minckley, 1969, 1977). Ecologically, the species is largely segregated from C. atrorus, the other pupfish of the Cuatro Cienegas Basin, but they meet and hybridize in periph¬ eral areas (see account for C. atrorus ). The species is considered vulnerable to extinc¬ tion because of its restricted distribution and rather specialized habitat requirements (Williams et al., 1989). Some of the pozas have been developed for picnicking and swimming, but these activities do not seem to pose serious threats to the species. See the account for C. atrorus for additional comments on introductions of non-native species and other human activities in the basin. Discussion None of the 12 pupfishes endemic to the north¬ ern Chihuahuan Desert, as defined for purposes of this symposium, has gone extinct. However, two spe¬ cies, C. bovinus and C. elegans have declined in range by at least 50% and, without intensive management, native stocks of the former would have been lost to ECHELLE ET AL. PUPFISHES OF THE NORTHERN CHIHUAHUAN DESERT 121 introgressive hybridization. The latter factor has elimi¬ nated native stocks of C. pecosensis over about 80% of its historic range. With one major exception, most of the remaining species have undergone various de¬ grees of range contraction as a result of a variety of anthropogenically induced losses and physical alter¬ ations of habitat. The exception is C. tularosa, which is confined within a military reservation with highly restricted access. The range of this species has ex¬ panded within its general area of endemicity as a result of transplantations by local people (Pittenger and Springer, 1999) into previously unoccupied waters in the Tularosa Basin. It is well understood that introduced non-native fishes are a major factor in the decline of native fishes in southwestern North America (Moyle et al., 1986; Allendorf and Leary, 1988; Propst et al., 1992) includ¬ ing northern Mexico (Contreras-B et al., 1976; Contreras-B and Escalante, 1994; Contreras-B, 2000). Correspondingly, depletion of pupfish populations via competition and predation by non-natives has been in¬ ferred for various situations in southwestern North America (Soltz and Naiman, 1978; Schoenherr, 1981). However, the potential for losses as a result of genetic introgression by non-native pupfish may have been underestimated until recently. The rapidity with which this factor can cause losses of native stocks is dra¬ matically demonstrated by events following the intro¬ duction, in the 1960s (Stevenson and Buchanan, 1973), of C. variegatus into Lake Balmorhea. Apparently be¬ cause of physical barriers, the locally endemic pupfish, C. elegans, has been little affected, but the presence of a dense, nearby population of the non-native poses a continual threat for this species (A. F. Echelle and Echelle, 1994) and other endemic pupfishes of the re¬ gion. Genetic markers indicate that the Lake Balmorhea population of C. variegatus has been the source for subsequent introductions into both Diamond Y Draw (Echelle and Echelle, 1997) and the Pecos River (Childs et al., 1996) which support, respectively, the endemic species C. bovinus and C. pecosensis. As described in our accounts for the pupfishes of the Pecos River Basin, the resulting rate and extent of genetic introgression has generated a great deal of concern and costly, some¬ times futile, management activity for the endemic pupfish species. Smith (1981) argued convincingly that pupfishes in desert environments are hardy generalists ultimately derived from estuarine ancestors pre-adapted to sur¬ vive the Post-Pleistocene desiccation of pluvial Pleis¬ tocene waters (Miller, 1981) of southwestern North America. After surviving and thriving during the post- Pleistocene expansion of desert, the pupfishes in the Chihuahuan Desert have been exposed to dramatically rapid anthropogenic reductions in habitable surface waters since the early 1800s. Since those years, arid grasslands were replaced by shrub desert over large portions of the desert southwest, an effect largely at¬ tributable to Anglo-American settlement, agriculture, and domesticated livestock (York and Dick-Peddie, 1969; Gehlbach, 1981; Hendrickson and Minckley, 1984). The rate of desertification, with losses of springs and cienegas and reduced persistence of natural stream habitat, was hastened by the more recent advent of mechanized construction of dams, canals, and water diversions, and groundwater pumping. Groundwater pumping may be the ultimate threat to most of the spring-dwelling pupfishes of the Chihuahuan Desert. In the past few decades, ground- water pumping has delivered the coupe de grace to many springs in the northern Chihuahuan Desert (Scudday 1977, this volume; Brune, 1981; Contreras- B and Lozano-V, 1994), and springs of the region con¬ tinue to fail (Contreras-B and Lozano-V, 1994; Sharp et al., this volume). In southern Nuevo Leon, just south of the region treated in our species accounts, five spe¬ cies of Cyprinodon and the monotypic genus Megupsilon, the closest relative of Cyprinodon (Echelle and Echelle, 1993b; Parker and Komfield, 1995), were driven to extinction in the wild within 1 5 years of the advent of intensive groundwater pumping (Contreras- B and Lozano-V, 1994, pers. observ.). In the northern Chihuahuan Desert, groundwater pumping seems di¬ rectly responsible for losses of two of three morpho¬ logically divergent populations of C. elegans and sig¬ nificant losses of populations of C. bovinus and the Guzman pupfish. In addition, there almost certainly have been losses of populations that may have occu¬ pied now-dry springs whose original faunas were un¬ known to science (W. L. Minckley, pers. comm.; SCB and MLLV, unpubl. data). Persistence of habitats suitable for the great va¬ riety of often locally endemic aquatic and semi-aquatic organisms in desert ecosystems (Williams et al., 1985) is especially threatened when drought conditions are superimposed onto a landscape already depleted of water resources by human activities. This probably explains the recent failure of Phantom Lake Spring 122 Special Publications, Museum of Texas Tech University Figure 1. Basins and springs mentioned in the text. 1 = Diamond Y Draw, 2 = Toyah Creek, 3 = Devils River, 4 = Alamito Creek, 5 = Banos de San Diego, 6 = Ojo de la Hacienda Dolores, 7 = Rio Sauz, 8 = Tularosa Basin, 9 - 13 = Laguna de Guzman Basin, 9 = Bolson de los Muertos, 10 = Rio Santa Maria, 11 = Rio Casas Grandes, 12 = Rio del Carmen, 13 = Laguna Bustillos, 14 = Rio Papigochic, 15 = Cuatro Cienegas. and the decline in surface waters in Diamond Y Draw, both of which are described earlier in this paper. It remains to be seen whether surface waters in these situations will rebound when the present drought ends. Regardless, with time and continued over-pumping of groundwater, these spring systems and others will fail as part of a trend of such failures in the past few de¬ cades (Brune, 1975, 1981; Contreras-B and Lozano- V, 1994; Sharp et al., this volume). This can be avoided only if society decides that these aquatic ecosystems are worth preserving, perhaps for their own sake or for their aesthetic appeal or what they signal about water availability for future human needs. ECHELLE ET AL. - PUPFISHES OF THE NORTHERN CfflHUAHUAN DESERT 123 Acknowledgments We thank N. Allan, J. Brooks, R. Edwards, G. Garrett, C. Hubbs, J. Karges, W. Minckley, S. Norris, D. Propst, and C. Springer for comments on portions of this manuscript and/or lengthy discussions of pup fish and their problems. Literature Cited Allendorf, F. W. and R. F. Leary. 1988. Conservation and distri¬ bution of genetic variation in a polytypic species, the cutthroat trout. Conservation Biology, 2:170-184. Arnold, E. T. 1972. Behavioral ecology of pupfishes (Cyprinodontidae, genus Cyprinodon ) from northern Mexico. Ph.D. Thesis, Arizona State University, Tempe. Brand, D. D. 1937. The natural landscape of northwestern Chihuahua. University of New Mexico Bulletin 316, Geological Series 5:1-74. (Cited from Minckley et al., 2002). Brune, G 1975. Major and historical springs of Texas. Texas Water Development Board, Report 189:1-94. _ . 1981. Springs of Texas. Vol. I. Branch-Smith, Inc., Fort Worth, Texas. Calegari, V. 1997. Environmental perceptions and local conser¬ vation efforts in Cuatro Cienegas, Coahuila, Mexico. Unpublished Master’s Thesis, University of Texas at Austin. Chemoff, B. and R. R. Miller. 1982. Notropis bocagrande, a new cyprinid fish from Chihuahua, Mexico, with com¬ ments on Notropis formosus. Copeia, 1982:514-522. Childs, M. R., A. A. Echelle, and T. E. Dowling. 1 996. Develop¬ ment of the hybrid swarm between Pecos pupfish (Cyprinodontidae: Cyprinodon pecosensis ) and sheeps- head minnow ( Cyprinodon variegatus ): a perspective from allozymes and mtDNA. Evolution, 50:2014-2022. Contreras-B., S. 1977. Speciation aspects and man-made com¬ munity composition changes in Chihuahuan Desert fishes. Pp. 405-431, in Transactions of the Symposium on the Biological Resources of the Chihuahuan Desert Region, United States and Mexico (R. H. Wauer and D. H. Riskind, eds.), National Park Transactions and Proceedings Series, Department of the Interior, Washington, D.C. _ . 1991. Conservation of Mexican freshwater fishes: some protected sites and species, and recent federal legislation. Pp. 191-197, in Battle Against Extinction: Native Fish Management in the American West (W. L. Minckley and J. E. Deacon, eds.), University of Arizona Press, Tucson. Contreras-B., S. 2000. Annotated Checklist of Introduced Inva¬ sive Fishes in Mexico, with Examples of some Recent Introductions. Pp. 33-54 in Nonindigenous Freshwater Organisms, Vectors, Biology, and Impacts (R. Claudi and J. H. Leach, eds.). Lewis Publishers, Boca Raton, Florida. Contreras-B., S. and M.A. Escalante. 1984. Distribution and known impacts of exotic fishes in Mexico. Pp. 102-130, in Distribution, Biology, and Management of Exotic Fishes (W. R. Courtenay, Jr., and J.R. Stauffer, Jr., eds.), Johns Hopkins University Press, Baltimore, Maryland. Contreras-B., S. and M. L. Lozano- V. 1994. Water, endangered fishes, and development perspectives in arid lands of Mexico. Conservation Biology, 8:379-387. _ . 1996. Extinction of most Sandia and Potosi valleys (Nuevo Leon, Mexico) endemic pupfishes, crayfishes and snails. Ichthyological Exploration of Freshwaters, 7:33- 40. Contreras-B., S., V. Landa-S., T.Villegas-G, and G Rodriguez- O., 1976. Peces, Piscicultura, Presas, Polucion, Planificacion Pesquera y Monitoreo en Mexico, o la Danza de las P. Memorias, Simposio sobre Pesquerias en Aguas Continentales, Mexico. Instituto Nacional de Pesca, Subsecretaria de Pesca, Secretaria de Industria y Comercio, Mexico. Davis, J. R. 1979. Die-offs of an endangered pupfish, Cyprinodon elegans (Cyprinodontidae). The Southwestern Natural¬ ist, 24:534-536. _ . 1980. Rediscovery, distribution, and populational sta¬ tus of Cyprinodon eximius (Cyprinodontidae) in Devil’s River, Texas. The Southwestern Naturalist, 25:81-88. Echelle, A. A. 1975. A multivariate analysis of variation in an endangered fish, Cyprinodon elegans, with an assessment of populational status. Texas Journal of Science, 26:529- 538. Echelle, A. A. and P. J. Connor. 1989. Rapid, geographically extensive genetic introgression after secondary contact between two pupfish species ( Cyprinodon , Cyprinodontidae). Evolution, 43:717-727. Echelle, A. A. and T. E. Dowling. 1992. Mitochondrial DNA variation and evolution of the Death Valley pupfishes (Cyprinodon, Cyprinodontidae). Evolution, 46: 1 93-206. Echelle, A. A. and A. F. Echelle. 1 978. The Pecos River pupfish, C. pecosensis n. sp. (Cyprinodontidae), with comments on its evolutionary origin. Copeia, 1978:569-582. _ . 1980. Status of the Pecos gambusia. Endangered Spe¬ cies Report No. 10. U. S. Fish and Wildlife Service, Albu¬ querque, New Mexico, 73 pp. 124 Special Publications, Museum of Texas Tech University _ . 1993a. Allozyme perspective on mitochondrial DNA variation and evolution of the Death Valley pupfishes (Cyprinodontidae: Cyprinodon ). Copeia, 1993:275-287. _ . 1 993b. Allozyme variation and systematics of the New World cyprinodontines (Teleostei: Cyprinodontidae). Biochemical Systematics and Ecology, 21 :583-590. _ . 1997. Genetic introgression of endemic taxa by non¬ natives: a case study with Leon Springs pupfish and sheep- shead minnow. Conservation Biology, 1 1:153-161. _ . 1998. Evolutionary relationships of pupfishes in the Cyprinodon eximius complex (Atherinomorpha: Cyprinodontiformes). Copeia, 1998:852-865. Echelle, A. A. and C. Hubbs. 1 978. Haven for endangered pupfish. Texas Parks and Wildlife Magazine, 36:9-11 . Echelle, A. A. and R. R. Miller. 1974. Rediscovery and rede¬ scription of the Leon Springs pupfish, Cyprinodon bovinus, from Pecos County, Texas. The Southwestern Naturalist, 19:179-190. Echelle, A. A., A. F. Echelle, and D. R. Edds. 1987. Population structure of four pupfish species (Cyprinodontidae: Cyprinodon ) from the Chihuahuan desert region of New Mexico and Texas: allozymic variation. Copeia, 1 987:668- 681. Echelle, A. A., C. W. Hoagstrom, A. F. Echelle, and J. E. Brooks. 1997. Expanded occurrence of genetically introgressed pupfish (Cyprinodontidae: Cyprinodon pecosensis x variegatus ) in New Mexico. The Southwestern Natural¬ ist, 42: 336-339. Echelle, A. F. and A. A. Echelle. 1994. Assessment of genetic introgression between two pupfish species, Cyprinodon elegans and C. variegatus (Cyprinodontidae), after more than 20 years of secondary contact. Copeia, 1994:590- 597. Echelle, A. F., A. A. Echelle, and D. R. Edds. 1 989. Conservation genetics of a spring-dwelling desert fish, the Pecos gam- busia ( Gambusia nobilis, Poeciliidae). Conservation Bi¬ ology, 3:159-169. Edwards, R. J., G P. Garrett, and E. Marsh-Matthews. 2002. An ecological analysis of fish communities inhabiting the Rio Conchos basin. (Analisis ecologico de las comunidades de peces que habitan la cuenca del Rio Conchos.) In: Libro Jubilar en Honor al Dr. Salvador Contreras-Balderas (Maria de Lourdes Lozano-Vilano, ed.), pp. 43-61. Garrett, GP. and A.H. Price. 1993. Comanche Springs pupfish (Cyprinodon elegans) status survey. Texas Parks and Wildlife Department, Section 6 Final Report, Project No. E-l-4, Job No. 2.1, Austin, 20 pp. Gehlbach, F. R. 1981. Mountain Islands and Desert Seas: A natural history of the U.S. -Mexican borderlands. Texas A&M University Press, College Station. Grail, G 1995. Cuatro Cienegas: Mexico’s desert aquarium. National Geographic, 188(October): 85-97. Hendrickson, D. A. and W. L. Minckley. 1984. Cienegas — van¬ ishing climax communities of the American southwest. Desert Plants, 6:131-175. Hendrickson, D. A., W. L. Minckley, R. R. Miller, D. J. Siebert, and P. H. Minckley. 1981. Fishes of the Rio Yaqui basin, Mexico and United States. Journal of the Arizona-Ne- vada Academy of Science, 15:65-106. Hoagstrom, C. W. and J. E. Brooks. 1999. Distribution, status, and conservation of the Pecos pupfish, Cyprinodon pecosensis. Technical Report No. 2, New Mexico De¬ partment of Game and Fish, Santa Fe, New Mexico, 76 pp. Hubbs, C. 1957. Distributional patterns of Texas fresh-water fishes. The Southwestern Naturalist, 2:89-104. _ . 1 980. Solution to the C. bovinus problem: eradication of a pupfish genome. Proceedings of the Desert Fishes Council, 10:9-18. Hubbs, C. and G. P. Garrett. 1990. Reestablishment of Cyprinodon eximius (Cyprinodontidae) and status of Dionda diaboli (Cyprinidae) in the vicinity of Dolan Creek, Val Verde Co., Texas. The Southwestern Natural¬ ist, 35:446-448. Hubbs, C., T. Lucier, E. Marsh, G P. Garrett, R. J. Edwards, and E. Milstead. 1978. Results of an eradication program on the ecological relationships of fishes in Leon Creek, Texas. The Southwestern Naturalist, 23:487-496. Hubbs, C., R. R. Miller, R. J. Edwards, K. W. Thompson, E. Marsh, G P. Garrett, G L. Powell, D. J. Morris, and R. W. Zerr. 1977. Fishes inhabiting the Rio Grande, Texas and Mexico, between El Paso and the Pecos confluence. Pp 91-97, in Symposium Proceedings: Importance, Pres¬ ervation and Management of the Riparian Habitat (R. R. Johnson and D. Jones, eds.), General Technical Report RM-43, U.S.D.A. Forest Service, Washington, D.C. Kennedy, S. E. 1977. Life history of the Leon Springs pupfish, Cyprinodon bovinus. Copeia, 1977:93-103. LaFave, J. I. and J. M. Sharp, Jr. 1987. Origins of ground water discharging at the springs of Balmorhea. West Texas Geo¬ logical Survey Bulletin, 26:5-14. Lozano- V., M. L. and S. Contreras-B. 1993. Four new species of Cyprinodon from southern Nuevo Leon, Mexico, with a key to the C. eximius complex (Teleostei: Cyprinodontidae). Ichthyological Exploration of Fresh- waters, 4:295-308. Miller, R. R. 1961. Man and the changing fish fauna of the American southwest. Papers of the Michigan Academy of Science, Arts, and Letters, 46:365-404. _ . 1964. Redescription and illustration of Cyprinodon latifasciatus, an extinct cyprinodontid fish from Coahuila, Mexico. The Southwestern Naturalist, 9:62-67. EcHELLE ET AL. PUPFISHES OF THE NORTHERN CHIHUAHUAN DESERT 125 _ . 1968. Two new fishes of the genus Cyprinodon from the Cuatro Cienegas Basin, Coahuila, Mexico. Occasional Papers of the Museum of Zoology, University of Michi¬ gan, 659:1-15. _ . 1976. Four new pupfishes of the genus Cyprinodon from Mexico, with a key to the C. eximius complex. Bul¬ letin of the Southern California Academy of Sciences, 75:68-75. _ . 1981. Coevolution of deserts and pupfishes (genus Cyprinodon) in the American Southwest. Pp. 39-94, in Fishes in North American Deserts (R. J. Naiman and D. L. Soltz, eds.), Wiley, New York. Miller, R.R. and A. A. Echelle. 1975. Cyprinodon tularosa, a new cyprinodontid fish from the Tularosa Basin, New Mexico. The Southwestern Naturalist, 19:365-377. Minckley, W. L. 1969. Environments of the Bolson of Cuatro Cienegas, Coahuila, Mexico, with special reference to the aquatic biota. University of Texas at El Paso Science Se¬ ries, 2:1-65. _ . 1973. Fishes of Arizona. Arizona Department of Fish and Game, Phoenix. _ . 1977. Endemic fishes of the Cuatro Cienegas Basin, Northern Coahuila, Mexico. Pp. 383-404, in Transac¬ tions of the Symposium on the Biological Resources of the Chihuahuan Desert Region, United States and Mexico (R. H. Wauer and D. H. Riskind, eds.), National Park Transactions and Proceedings Series, Department of the Interior, Washington, DC. Minckley, W. L. and R. K. Koehn. 1965. Re-discovery of the fish fauna of the Sauz Basin, northern Chihuahua, Mexico. The Southwestern Naturalist, 10:313-315. Minckley, W. L., G K. Meffe, and D. L. Soltz. 1991. Conserva¬ tion and management of short-lived fishes: The cyprin- odonts. Pp. 247-282, in Battle Against Extinction: Native Fish Management in the American West (W. L. Minckley and J. E. Deacon, eds.), University of Arizona Press, Tucson. Minckley, W. L., R. R. Miller, and S. M. Norris. 2002. New pupfishes (Teleostei, Cyprinodontidae, genus Cyprinodon ), two from Chihuahua, Mexico, and another from Arizona, USA. Copeia, 2002:687-705. Minckley, W. L. and C.O. Minckley. 1986. Cyprinodon pachycephalus, a new species of pupfish (Cyprinodontidae) from the Chihuahuan Desert of north¬ ern Mexico. Copeia, 1986:184-1 92. Moyle, P. B., H. W. Li, and B. A. Barton. 1986. The Franken¬ stein effect: impact of introduced fishes on native fishes in North America. Pp. 4 1 5-426, in Fish Culture in Fisher¬ ies Management (R. H. Stroud, ed.), American Fisheries Society, Bethesda, Md. Parker, A. and I. Komfield. 1995. Molecular perspective on evolution and zoogeography of cyprinodontid killifishes (Teleostei: Atherinomorpha). Copeia, 1995:8-21. Pittenger, J. S. and C. L. Springer. 1999. Native range and con¬ servation of the White Sands pupfish ( Cyprinodon tularosa). The Southwestern Naturalist, 44:157-165. Propst, D. L. and J. A. Stefferud. 1994. Distribution and status of the Chihuahua chub (Teleostei: Cyprinidae: Gila nigrescens), with notes on its ecology and associated species. The Southwestern Naturalist. 39:224-234. Propst, D. L., J. A. Stefferud, and P. R. Turner. 1992. Conserva¬ tion and status of Gila trout, Oncorhynchus gilae. The Southwestern Naturalist, 37:1 17-125. Rodriguez-G, J. 1926. Geografia del estado de Coahuila. Soc. Edicion y Libreria Franco-Americana, D.F. (Cited from Calegari, 1997). Rhodes, K. and C. Hubbs. 1992. Recovery of Pecos River fishes from a red tide fish kill. The Southwestern Naturalist, 37:178-187. Schoenherr, A. A. 1981 . The role of competition in the replace¬ ment of native fishes by introduced species. Pp. 173- 203, in Fishes in North American Deserts, (R. J. Naiman and D. L. Soltz, eds.), Wiley, New York. Schuster, S. K. 1997. Hydrogeology and local recharge analysis in the Toyah Basin Aquifer. Unpublished Master’s The¬ sis, University of Texas at Austin, 130 pp. Scudday, J. F. 1 977. Some recent changes in the herpetofauna of the northern Chihuahua Desert. Pp. 5 1 3-522, in Transac¬ tions of the Symposium on the Biological Resources of the Chihuahuan Desert Region, United States and Mexico, (R. H. Wauer and D. H. Riskind, eds.), National Park Transactions and Proceedings Series, Department of the Interior, Washington, DC. Soltz, D. L. and R. J. Naiman. 1978. The Natural History of Native Fishes in the Death Valley System. Natural His¬ tory Museum of Los Angeles County, Science Series 30. Smith, M. L. 1981. Late Cenozoic fishes in the warm deserts of North America: a reinterpretation of desert adaptations. Pp. 11-38, in Fishes in North American Deserts (R. J. Naiman and D L. Soltz, eds.), Wiley, New York. Smith, M. L. and B. Chemoff. 1981. Breeding populations of cyprinodontoid fishes in a thermal stream. Copeia, 1981:701-702. Smith, M. L. and R. R. Miller. 1980. Systematics and variation of a new cyprinodontid fish, Cyprinodon fontinalis, from Chihuahua, Mexico. Proceedings of the Biological Soci¬ ety of Washington, 93:405-416. _ . 1981. Conservation of desert spring habitats and their endemic fauna in northern Chihuahua, Mexico. Proceed¬ ings of the Desert Fishes Council, 13:54-63. Stevenson, M. M. and T. M. Buchanan. 1973. An analysis of hybridization between the cyprinodont fishes Cyprinodon variegatus and C. elegans. Copeia, 1973:682-692. 126 Special Publications, Museum of Texas Tech University Stockwell, C. A. and M. Mulvey. 1998. Phosphogluconate de¬ hydrogenase polymorphism and salinity in the White Sands pupfish. Evolution, 52:1856-1860. Stockwell, C. A., M. Mulvey, and A. G Jones. 1998. Genetic evidence for two evolutionarily significant units of White Sands pupfish. Animal Conservation, 1:213-225. Valdes Cantu, N.E. and K.O. Winemiller. 1997. Structure and habitat associations of Devils River fish assemblages. The Southwestern Naturalist, 42:265-278. Wilde, G R. and A. A. Echelle. 1992. Genetic status of Pecos pupfish populations after establishment of a hybrid swarm involving an introduced congener. Transactions of the American Fisheries Society, 121:277-286. _ . 1997. Morphological variation in intergrade pupfish populations from the Pecos River, Texas, U.S. A. Journal of Fish Biology, 50:523-539. Williams, J. E., D. B. Bowman, J. E. Brooks, A. A. Echelle, R. J. Edwards, D. A. Hendrickson, and J. J. Landye. 1985. Endangered aquatic ecosystems in North American deserts with a list of vanishing fishes. Arizona-Nevada Academy of Science, 20:1-62. Addresses of authors: Anthony A. Echelle Zoology Department Oklahoma State University Stillwater, OK 74078 e-mail: echelle@okstate.edu Alice F. Echelle Zoology Department Oklahoma State University Stillwater, OK 74078 email: aechelle@Juno.com Williams, J. E., J. E. Johnson, D. A. Hendrickson, S. Contreras- B., D. J. Williams, M. Navarro-M., D. E. McAllister, and J. E. Deacon. 1989. Fishes of North America: Endan¬ gered, threatened, or of special concern: 1989. Fisheries, 14:2-20. Winemiller, K. O. and A. A. Anderson. 1997. Response of en¬ dangered desert fish populations to a constructed refuge. Restoration Ecology, 5:204-214. York, J. C. and W. A. Dick-Peddie. 1969. Vegetation changes in southern New Mexico during the past hundred years. Pp. 157-166, in Arid Lands in Perspective. (W. G McGinnies and B. J. Goldman, eds.), University of Arizona Press, Tucson. Young, D. A., K. J. Fritz, G P. Garrett, and C. Hubbs. 1993. Status review of construction, native species introduc¬ tions, and operation of an endangered species refugium channel, Phantom Lake Spring, Texas. Proceedings of the Desert Fishes Council, 25:22-25. Salvador Contreras Balderas Laboratorio de Ictiologia Facultad de Ciencias Biologicas Universidad Autonoma de Nuevo Leon Apartado Postal 425 San Nicolas de Los Garza N.L., Mexico 66450 e-mail: saconbal@axtel.net Ma. de Lourdes Lozano Vilano Laboratorio de Ictiologia Facultad de Ciencias Biologicas Universidad Autonoma de Nuevo Leon Apartado Postal 425 San Nicolas de Los Garza N.L., Mexico 66450 e-mail: marlozan@ccr.dsi.uanl.mx Spring-Endemic Gambusia of the Chihuahuan Desert Clark Hubbs Abstract Spring endemic fishes are restricted to the vicin¬ ity of spring outflows where stenothemial conditions prevail. Stream fishes occupy downstream locations that are substantially more eurythermal. Those cir¬ cumstances prevail throughout the Chihuahuan Desert where Gambusia senilis interacts with G. hurtadoi and G. alvarezi and Gambusia affinis interacts with G. nobilis and G. gaigei. Gambusia amistadensis once occurred in and was restricted to Goodenough Spring where it interacted with G. affinis in the Rio Grande. All evidence indicates that similar interactions occur with other fishes, amphipods, crayfish, and sala¬ manders. Introduction The fishes of the Chihuahuan Desert require ad¬ equate quantity and quality of water. In as much as few lakes occur there, the need is for lentic environ¬ ments. Throughout most of the southwest United States and northwest Mexico, human desires often negatively impact stream flows. Problems associated with all the fishes can be illustrated by problems with members of the genus Gambusia. Many of the fishes are spring endemics that require natural spring flow volumes (Hubbs, 1996, 2001). Often the problem is intrageneric. For example, Gambusia senilis and the affinis species group are stream fishes and G. nobilis, G. gaigei, G. hurtadoi, and G. alvarezi are spring endemics. When spring flow volumes decrease, the stream species gain area at the expense of the spring endemics. Stream species have a competitive advan¬ tage away from springs and spring species have a com¬ petitive advantage in springs. The primary driving factor is stenothermal vs. eurythermal conditions. Similar interactions occur within cyprinids, centrarchids, and percids in the Chihuahuan Desert. Likewise in Ne¬ vada Crenichthys is adapted to stenothermal, warm, low dissolved oxygen springs and many native min¬ nows are better as stream fishes. Crenichthys can survive in below 0.7 ppm O, at 37°C (Sumner and Sargent, 1940; Hubbs and Hettler, 1964). It is unlikely that members of the genus Cyprinodon have similar needs for stenothermal water, as Cyprinodon elegans populations increased at Phantom Cave Spring when spring flow declined (Hubbs, 2001). Nevertheless, Cyprinodon clearly needs adequate volumes of water even if it is not stenothermal. The needs of spring fishes for quantities of spring flow has also been reported from Alabama (Howell and Black, 1976), Arkansas (Robison and Buchanan, 1988), New Mexico (Hubbs and Echelle, 1972), Okla¬ homa (Matthews et al., 1985), China, France, Italy, Croatia (Maurice Kottetot, European Ichthyological Congress, pers. comm.), Australia, Brazil, Congo, and Iran. I consider spring fish to be defined as species found wholly or most often in spring heads and/or in the immediate zone downstream (i.e., spring runs) and seldom found elsewhere in the drainage. Many other aquatic organisms have similar spring vs. stream abun¬ dances including amphipods (Bowles and Arsuffi, 1993), aquatic plants (Emery, 1967) salamanders (Tupa and Davis, 1976; Chippendale et ah, 1993) and cray¬ fish (Hubbs, 2001). I suspect the entire aquatic biota is also involved but commonly not as well known. 128 Special Publications, Museum of Texas Tech University Examples in Gambusia Gambusia nobilis. — The Pecos gambusia is a relatively robust Gambusia , with an arched back and a caudal peduncle depth that is approximately two-thirds of the head length. The margins of the scale pockets are outlined in black and spots are normally absent on the caudal fin, however, sometimes a faint medial row of spots may be present. The dorsal fin has a sub- basal row of spots. Females have a prominent black area on the abdomen that surrounds the anus and anal fin. The male gonopodium has a number of unique features including elongated spines on ray 3, small rounded hooks on the tips of rays 4p and 5a, and an elbow on ray 4a consisting of 3 or 4 fused segments located opposite the serrae of ray 4p. Gambusia nobilis generally have 8 dorsal, 12 pectoral, 9 to 10 anal, and 6 pelvic rays (Hubbs and Springer, 1957; Koster, 1957; Bednarz, 1975; Echelle and Echelle, 1986). Popula¬ tions in Toyah Creek (Texas) and Blue Spring (New Mexico) were found to be the most diverse morpho¬ logically and genetically and the Toyah Creek popula¬ tion had the greatest genetic heterogeneity (Echelle and Echelle, 1986; Echelle etal., 1989). Gambusia nobilis was described by Baird and Girard (1853) based on material from Leon and Comanche springs, Pecos County, Texas. Leon Springs was later designated the type locality (Hubbs and Springer, 1957). The species is endemic to the Pecos River basin in southeastern New Mexico and western Texas and originally ranged from near Fort Sumner, New Mexico to the area around Fort Stockton, Texas. At present, the species is re¬ stricted to four main areas, two in New Mexico and two in Texas. Populations live in various springs and sinkholes in Bitter Lake National Wildlife Refuge, near Roswell, New Mexico; Blue Spring, east of Carlsbad Caverns National Park, New Mexico; the Diamond Y springs and draw (= Leon Creek), near Fort Stockton, Texas; and the Balmorhea springs complex and Toyah Creek near Balmorhea, Texas. Extirpated populations include the Pecos River near Fort Sumner and North Spring River in New Mexico, and Leon and Comanche springs, in Texas. Those populations in Comanche Springs were extirpated when the spring dried (Hubbs and Springer, 1957). The population in Leon Springs was eliminated by impoundment (Miller et al., 1991). Those populations in the Balmorhea springs complex and the Diamond Y region are most abundant in steno¬ thermal waters. Where suitable habitats exist, Pecos gambusia populations can be dense. An estimated 27,000 indi¬ viduals inhabit the Bitter Lake National Wildlife Refuge area, and 900,000 inhabit Blue Spring (Bednarz, 1975, 1979) . Approximately 100,000 Pecos gambusia are estimated to inhabit the Balmorhea springs complex and more than 100,000 in the Diamond Y springs and draw. Pecos gambusia primarily inhabit stenothermal springs, runs, spring-influenced marshes (cienegas), and irrigation canals carrying spring waters. Some populations however, are also known from areas with little spring influence; these habitats generally have abundant overhead cover, and include sedge covered marshes and gypsum sinkholes (Echelle and Echelle, 1 980) . One or two other Gambusia may also be found in association with G. nobilis. Where the western mosquitofish {G. affinis ) is found, G. nobilis inhabits stenothermal waters and G. affinis is most often found in eurythermal habitats. Where together, the Pecos gambusia is much more likely to be found associated with vegetation or in deeper waters, while G. geiseri tends to be at the surface or in open water over non- vegetated substrates (Hubbs et al., 1995). Another spring endemic, G. geiseri was introduced into Pecos gambusia habitat for “mosquito control.” This was a serious error as the native G. nobilis is at least as good at mosquito control. Pecos gambusia feed relatively non-selectively, consuming a diversity of food types, including amphipods, dipterans, cladocerans, filamen¬ tous algae, arachnids and mollusks (Hubbs et al., 1978; Winemiller and Anderson, 1997). Gambusia nobilis produce live young. Bednarz (1979) reported that the number of embryos was related to female size and that the mean number of embryos was 38 in the Blue Spring population. Hubbs (1996) found that the birth weight of Pecos gambusia from Texas populations ranged between 35 and 50 mg and females had an interbrood interval of 52 days. Hybrids between G. nobilis and G. affinis or G. geiseri are occasionally found, especially in habitats where one of the species is rare (Hubbs and Springer, 1957). Pecos gambusia face severe threats from spring flow declines and habitat modification throughout their range. In parts of their range, cienegas, presumed to have supported large numbers of G. nobilis, were drained and spring flows diverted for irrigation. San Hubbs — Spring-Endemic Gambusia of the Chihuahuan Desert 129 Solomon Springs has been modified into a spring-fed swimming pool. During 1998 and 1999, Phantom Lake (part of the Balmorhea springs complex) dried twice and after the second drying, Phantom Lake Spring ceased to flow and ultimately the Phantom Lake Spring refuge canal dried. During May 2000, a pump was installed by the U.S. Bureau of Reclamation to insure continuous flow in the upper spring pool. Additional stresses on the population may occur through compe¬ tition with the introduced G. geiseri. Efforts have been made to improve habitat in the Balmorhea area. A small refugium canal was constructed in 1974 in Balmorhea State Park (Echelle and Hubbs, 1978). In 1993, the Bureau of Reclamation constructed a modified 1 10-m canal at Phantom Lake Spring (Young et al., 1994) with sloped, sinuous sides to resemble a portion of a cienega. That canal is now dry and all fish dead as Phantom Lake Spring no longer flows. After Phantom Lake Spring ceased flow and the only water remaining was in the “spring pool” that is artificially maintained by a pump, G. nobilis declined precipitously, G. affinis increased, and G. geiseri remained rather stable (Hubbs, 2001). In cooperation with local residents and farm¬ ers, in 1996 the construction of the 1-ha San Solomon Cienega was completed (McCorkle et ah, 1998; Garrett, this volume). This wetland is situated within the bound¬ aries of the original, natural cienega on state park land. Designed to resemble and function like the original cienega, the native fish fauna, including G. nobilis , has flourished near the inlet where stenothermal conditions prevail. The park refugium canal, and the San Solomon Cienega have increased numbers and security for the species, but each remains dependent on spring flows. The Diamond Y introductions of G. geiseri were eliminated by efforts to remove exotic Cyprinodon variegatus that formed massive hybrid swarms with the endemic C. bovinus that threatened the survival of the latter (Hubbs et al., 1978). A byproduct of those control efforts was the eradication of exotic G. geiseri from the Diamond Y system. Equivalent collecting effort at Diamond Y Spring before eradication had equal numbers of G. geiseri and G. nobilis. After eradica¬ tion, G. nobilis had abundances similar to the sum of the two species abundances. Where G. geiseri was absent G. nobilis dominated stenothermal water and G. affinis dominated eurythermal waters. In the Balmorhea area three species are involved. In general the 2 spring species dominate the stenothermal water and G. affinis dominates in the eurythermal waters. There is a strong tendency for G. nobilis to be re¬ stricted to the most stenothermal waters, and G. geiseri to be common in waters of intermediate thermal varia¬ tion. Additional protection for G. nobilis stems from its presence in Balmorhea State Park, Bitter Lake Na¬ tional Wildlife Refuge, Diamond Y Draw which is owned by The Nature Conservancy of Texas, and an intro¬ duced stock of Pecos gambusia in artificial pools at the Living Desert State Park near Carlsbad, New Mexico. Gambusia gaigei. — The color pattern of the Big Bend gambusia is weaker than that of the other mem¬ bers of the G. nobilis species group. The ground color is silvery with an iridescent blue overtone. There is considerable yellowish-orange on the clear areas of the unpaired fins. The markings on the margins of the scale pockets are faint. There are none anterior to the anus or below the eye. The middorsal streak and dark coloration of the neurocranium cover and obscure the scale-pocket markings, but the postanal streak does not obscure these markings. The faint, broad lateral band often obscures the scale-pocket markings on one scale row. There are a few dark crescents on the scale rows surrounding the lateral band. The anal spot in females is restricted to the anus. The dorsal has a subbasal row of black spots and a dark margin. The caudal has no dark markings. The anal of females has a dark margin; that of males is grayish. There is only a trace of a dark chin bar. Except for the suborbital bar there is no darkening of the lower parts posterior to the eye. The Big Bend gambusia was described in Hubbs (1929) based on specimens collected by F. M. Gaige from “a marshy cattail slough fed by springs, located close to the Rio Grande at Boquillas, Brewster County, opposite the Mexican village of the same name.” Ap¬ parently this was the largest river-side spring in the Big Bend region (Hubbs, 1940). Subsequently in June 1954 numerous specimens were obtained from Gra¬ ham Ranch Warm Springs (Hubbs and Springer, 1957). Graham Ranch Warm Springs (now known as Spring 4) is the largest spring near Boquillas but is 1 km west of Boquillas. Other springs existed at Boquillas that may have been the source of Fred Gaige ’s captures. Those springs dried in 1954 and no longer contain fish. It is more likely, however, that the original col¬ lections came from Graham Ranch Warm Springs. 130 Special Publications, Museum of Texas Tech University Although Big Bend gambusia were abundant in Graham Ranch Warm Springs and the newly con¬ structed adjacent “kiddie fishing pool”, they were scarce by 1956 and the previously scarce G. affinis very abundant. Consequently, renovation efforts were initiated 9 October 1956. Intensive seining obtained 24 individuals and the area was treated with rotenone and fewer than 12 other Big Bend gambusia (and thou¬ sands of G. qffinis) killed (Hubbs and Broderick, 1963). The 24 remaining individuals were placed in 5 loca¬ tions: 1) Boquillas Spring, Glenn Springs and, stock tank along the Glenn Springs Road (14 individuals); 2) metal tank near the Park Headquarters (6 individuals) and; 3)4 individuals taken to Austin. The fish placed at location 1 were never seen again. The fish at loca¬ tion 2 flourished until a cold day that killed them all. One of the 4 remaining fish died in Austin but the other 3(1 female and 2 males) were returned to the park in a newly constructed pond where they flourished (Hubbs and Broderick, 1963). At the same time the Rio Grande Village Camp Ground was established near the Graham Ranch Warm Spring (and the existing 4 springs renamed 1-4). Trees were planted to shade the camp ground and watered from the Rio Grande. That water drained into the Big Bend gambusia pond. Gambusia affinis got into those irrigation ditches and subsequently into the Big Bend gambusia pond. When the pond was examined 1 6 April 1960 only 27 individuals were obtained. All were taken to Aus¬ tin. Half of them were sent to the University of Michi¬ gan for insurance against extinction. Both cultures flourished, and after a second refuge pond (using Spring 1 water) was constructed, were returned (9 August 1966) to the park (Hubbs and Williams, 1976). With 2 intervals of extreme scarcity, it is not surpris¬ ing that the surviving population is homozygous (Echelle and Echelle, 1991). Minor problems (intro¬ duction of green sunfish, Lepomis cyanellus, and mi¬ nor mortality due to an extremely cold day) did not threaten survival (Hubbs and Williams, 1976). Subse¬ quently all Spring 4 water was diverted to the Camp¬ ground and the kiddie fishing pool dried. Later a flash flood ran through the Spring 1 refuge pool into the Spring 4 pool, reintroducing the Big Bend gambusia into its original habitat. Increased pumpage caused the Spring 1 refuge pool to overflow through its drain pipe. Fish were transferred from the refuge pool into the overflow ditch and both flourish. Eventually, wa¬ ter leaked from the Spring 4 pool and some Big Bend gambusia occurred in a large beaver pond between Spring 4 and the Rio Grande. Big Bend gambusia flour¬ ishes in the 2 spring pools and the Spring 1 drainage ditch. A small population occurs in the Beaver Pond. A small number of G. affinis occurs in the Spring 4 pool. The relative number of G. gaigei and G. affinis is correlated with the thermal stability of the two lo¬ calities. G. gaigei dominates in stenothermal water and G. affinis dominates in the eurythermal water (Hubbs, 2001). In Chihuahua, Mexico a similar pattern emerges but much less data on environmental conditions are available. Gambusia hurtadoi. — The Dolores gambusia lives in El Ojo de la Hacienda Dolores that has a large spring pool ca 20 X 50 m with a dense population of G. hurtadoi as well as an endemic Cyprinodon. Both endemics also flourish in the irrigation ditches leaving the spring pool. During wet weather, spring waters empty into the Rio Florido where G. senilis is the only gambusia present. The color pattern of the Dolores gambusia is darker and has more iridescent blue than that of other members of the G. nobilis species group. The ground color is iridescent bluish-silver. There is less orange on the body than in G. alvarezi. The clear areas on the median fins are yellowish orange. The markings on the margins of the scales are dark but diffuse, and are often obscured by other more prominent markings. There are no marks anterior to the anus and none be¬ low the preopercle. The lateral diffusion of the dark middorsal streak often reaches the lateral band. The thin postanal streak does not obscure the scale-pocket markings. The lateral band is broad and dark. The dark crescents on the side are concentrated along the lateral band and often cover and obscure it. They follow the scale-pocket margins. The anal spot is not restricted to the anus. The dorsal fin has a subbasal row of spots, which are darker than the darkened margin. The caudal has no dark markings. The anal fin of females has a dark margin; that of males is gray¬ ish. The prominent dark chin bar is often interrupted medially. The suborbital bar is dark. Hubbs — Spring-Endemic Gambusia of the Chihuahuan Desert 131 Three color variants have been noted. These may be designated Spotted, Gray, and Golden. All three lack the dorsal streak, the post-anal streak, and the lateral band. Spotted also has no dark markings on the scale-pockets and no suborbital bar; the color pat¬ tern on the body consists chiefly of the crescents which are often grouped and not concentrated along the lo¬ cation of the lateral band. Gray lacks the crescents typical of Spotted. Its coloration somewhat resembles that of G. geiseri except that the scale-pocket mark¬ ings are more diffuse. Golden has no dark markings except the dark margin of the dorsal, three small patches of scale-pocket margins in front of the dorsal and above the pectoral bases, the spotting of the peri¬ toneum, and the dark eye. In life it is an attractive golden yellow. There is no iridescent blue. All three color phases were present in a collection made by Clark Hubbs and Oscar f. Wiegand at El Ojo de la Hacienda Dolores on December 31, 1954. The preserved col¬ lection contained a random sample of 1 ,342 individu¬ als. Later a single Golden individual 16 mm long was collected. Approximately 500 more specimens were collected in the subsequent work. Sixteen Spotted and thirteen Grays have been noted in the preserved sample. Although 502 of the specimens are longer than 20 mm, none of the color variants are. It is pos¬ sible that they have a reduced survival rate. This hy¬ pothesis is supported by our failure to bring the Golden individual back to Austin alive. Gambusia alvarezi. — The color of the San Gregorio gambusia is more yellow and orange than that of any other members of the G. nobilis species group. The ground color is yellow-orange. There is little iridescent blue on the body. All clear areas on the fins are yellow-orange. The diffuse dark markings on the margins of the scales are often obscured by more prominent markings. There are no marks anterior to the anus or below the eye. The middorsal streak is dark, but its lateral diffusion does not reach to the lateral band. The dim postanal streak often obscures the scale-pocket markings. The lateral band is broad and dark. The dark crescentic marks on the side are concentrated along the lateral band. They are not nu¬ merous and more than three are seldom interconnected. The lateral band can easily be traced through the areas between the scattered groups of crescents. The black spot around the anus is large. The dorsal fin has a subbasal row of spots and a darker margin. The cau¬ dal has no dark markings. The anal fin of females has a dark margin; that of males is uniformly grayish. The prominent dark chin bar is often interrupted medially. The suborbital bar is dark. Similarly, El Ojo de San Gregorio has a dense population of G. alvarezi. El Ojo de San Gregorio has much less volume than El Ojo de la Hacienda Dolores and virtually no spring pool. The water eventually flows (or flowed) into the Rio Parral once occupied by G. senilis. Unfortunately the Rio Parral is severely polluted with mining wastes that preclude contact be¬ tween the two species. Color differences were maintained in laboratory- reared stocks of G. alvarezi , G. gaigei, and G. hurtadoi. Gambusia amistadensis . — The Amistad gambu¬ sia was a member of the Gambusia senilis species group and is closely related to G. hurtadoi , G. alvarezi, G. gaigei, and G. senilis (Rauchenberger, 1989). The species is characterized by its relatively slender body, terminal mouth with numerous teeth on each jaw, and males having long serrae on ray 4p of the gonopodium. Preserved specimens have strong crosshatching and numerous darkly pigmented crescent-shaped spots on their scale margins. The mid-dorsal stripe is narrow and the lateral stripe is broad. A short, dusky subocular bar is present. Adult females have a permanent me¬ dian dark anal spot (Peden, 1973). The Amistad gambusia was originally described from Goodenough Springs (29°32’10"N, 10°15’10"W) in Val Verde County, Texas. The original range of the species included the headsprings and the 1 .3-km spring run downstream to its confluence with the Rio Grande (Peden, 1973). The species became extinct in the wild when Goodenough Springs, once the third largest spring system in Texas, was inundated by Amistad Reservoir when the dam gates were closed in 1968 (Peden, 1973; Brune, 1981). Goodenough Springs and its warm spring run rapidly flowed over limestone gravel and sand substrates along its course to the Rio Grande. Waters originated in the relatively large Edwards-Trinity aquifer (Peckham, 1963) and main¬ tained flow rates of approximately 2,000-4,000 cubic liters per second (Brune, 1981). The type locality and habitat for the Amistad gambusia is now under ap¬ proximately 30 m of water and former spring open¬ ings may now be recharge zones (Peden, 1973; Brune, 1981). 132 Special Publications, Museum of Texas Tech University Little is known concerning the food habits of the Amistad gambusia; however, the gut contents of 10 paratypes examined by Peden (1973) contained mostly unidentified items, some insect fragments and traces of filamentous algae. Other fishes co-occurring with the Amistad gambusia prior to the inundation of its habitat included: Astyanax mexicanus (Mexican tetra), Macrhybopsis aestivalis (speckled dace), Cyprinella lutrensis (red shiner), C. venusta (blacktail shiner), C. proserpina (proserpine shiner), N. braytoni (Tamaulipas shiner), N.jemezanus (Rio Grande shiner), Cycleptus elongatus (blue sucker), Ictalurus punctatus (channel catfish), Ameiurus melas (black bullhead), Pylodictis olivaris (flathead catfish), Gambusia affinis (western mosquitofish), Micropterus salmoides (largemouth bass), Lepomis cyanellus (green sunfish) and Cichlasoma cyanoguttatum (Rio Grande cichlid) (Peden, 1973). Observations in aquaria by Peden (1970, 1973) indicated that male courtship appeared similar to that found in other poeciliids and that pregnant female Gam¬ busia amistadensis gave birth to their young in veg¬ etated areas. Of 10 female paratypes examined by Peden (1973), the mean size was 29.8 mm SL (range = 25.9-34.6 mm SL) and 7 contained 5 to 11 (mean 8.9) embryos in each ovary while the other 3 females contained 1 to 7 eggs. Culture populations of G. amistadensis were maintained until the late 1970s at the University of Texas at Austin and at the U.S. Fish and Wildlife Service’s endangered species culture facility in Dexter, New Mexico (Hubbs and Jensen, 1984). These popula¬ tions were contaminated by western mosquitofish {Gambusia affinis), which eliminated the G. amistadensis in these cultures prior to 1983 (Hubbs and Jensen, 1984). I suspect this species would not be extinct if those refugia had been maintained as steno¬ thermal environments. Literature Cited Baird, S. F. and C. Girard. 1 853. Description of new species of fishes collected by Mr. John H. Clark on the U.S. and Mexican Boundary Survey, under Lt. Col. Jas. D. Gra¬ ham. Proceedings of the Academy of Natural Science, Philadelphia, 6:387-390. Bednarz, J. 1975. A study of the Pecos gambusia. Endangered Species Program, New Mexico Department of Game and Fish, Santa Fe, New Mexico, 30 pp. Bednarz, J. 1979. Ecology and status of the Pecos gambusia, Gambusia nobilis (Poeciliidae), in New Mexico. The Southwestern Naturalist, 24:31 1-322. Bowles, D. E. and T. L. Arsuffi. 1993. Karst aquatic ecosys¬ tems of the Edwards Plateau region of central Texas, USA: a consideration of their importance, threats to their exist¬ ence, and efforts for their conservation. Aquatic Conser¬ vation of Marine and Freshwater Ecosystems, 3:317- 329. Brune, G 1981. Springs of Texas. Vol. 1. Branch-Smith, Inc., Fort Worth, Texas. Chippindale, P. T., A. H. Price, and D. M. Hillis. 1993. Anew species of perennibranchiate salamander (Euryces: Plethodontidae) from Austin, Texas. Herpetologica, 49:248-259. Echelle, A. A. and A. F. Echelle. 1980. Status of the Pecos gambusia, Gambusia nobilis. U.S. Fish and Wildlife Ser¬ vice, Albuquerque, New Mexico, Endangered Species Report, 10:1-73. Echelle, A. A. and A. F. Echelle. 1986. Geographic variation in morphology of a spring-dwelling desert fish, Gambusia nobilis (Poeciliidae). The Southwestern Naturalist, 31:459-468. Echelle, A. A. and A. F. Echelle. 1991. Conservation genetics and genetic diversity in freshwater fishes of western North America. Pp. 141-153, in Battle Against Extinction: Na¬ tive Fish Management in the American West (W. L. Minckley and J. E. Deacon, eds.), University of Arizona Press, Tucson. Echelle, A. A., A. F. Echelle, and D. R. Edds. 1 989. Conserva¬ tion genetics of a spring-dwelling desert fish, the Pecos gambusia ( Gambusia nobilis, Poeciliidae). Conservation Biology, 3: 159-169. Echelle, A. A. and C. Hubbs. 1978. Haven for endangered pupfish. Texas Parks and Wildlife Magazine, June:9-1 1. Emery, W. H. P. 1 967. The decline and threatened extinction of Texas wild-rice {Zizania texana Hitchs). The Southwest¬ ern Naturalist, 12:203-204. Hubbs — Spring-Endemic Gambusia of the Chihuahuan Desert 133 Howell, W. M. and A. Black. 1976. Status of the watercress darter, Etheostoma nuchale. Proceedings of the South¬ eastern Fishes Council, 1:1-3. Hubbs, C. 1996. Geographic variation of Gambusia life history traits. Proceedings of the Desert Fishes Council, 27: 1 -21 . Hubbs, C. 2001 . Environmental correlates to the abundance of spring-adapted and stream-adapted fishes. Texas Journal of Science, 53:299-326. Hubbs, C. and H. J. Broderick. 1963. Current abundance of Gambusia gaigei, an endangered fish species. The South¬ western Naturalist, 8:46-48. Hubbs, C. and A. A. Echelle. 1 972. Endangered non-game fishes of the upper Rio Grande Basin. Pp. 147-167, in Endan¬ gered vertebrates in the Southwest (W. C. Huey, ed.), New Mexico Department of Game and Fish, Santa Fe, New Mexico. Hubbs, C., A. F. Echelle, and G Divine. 1995. Habitat parti¬ tioning by two congeners ( Gambusia geiseri and Gam¬ busia nobilis ) at Balmorhea State Park. Texas Journal of Science, 47:325-326. Hubbs, C. and W. F. Hettler. 1964. Observations on the toler¬ ance of high temperatures and low dissolved oxygen in natural waters by Crenichthys baileyi. The Southwestern Naturalist, 9:245-248. Hubbs, C. and B. L. Jensen. 1984. Extinction of Gambusia amistadensis, an endangered fish. Copeia, 1984:529-530. Hubbs, C., T. Lucier, E. Marsh, G P. Garrett, R. J. Edwards, and E. Milstead. 1978. Results of an eradication program on the ecological relationships of fishes in Leon Creek, Texas. The Southwestern Naturalist, 23:487-496. Hubbs, C. and V. G Springer. 1957. A revision of the Gambusia nobilis species group, with descriptions of three new species, and notes on their variation, ecology, and evolu¬ tion. Texas Journal of Science, 9:279-327. Hubbs, C. and J. G Williams. 1976. A review of circumstances affecting the abundance of Gambusia gaigei, an endan¬ gered fish endemic to Big Bend National Park. Pp. 63 1 - 635, in Proceedings of the First Conference on Scientific Research in the National Parks, Vol. I. (Robert M. Linn, ed.). Hubbs, C. L. 1929. Studies of the fishes of the order Cyprinodontes VIII. Gambusia gaigei, a new species from the Rio Grande. Occasional Papers of the Museum of Zoology University of Michigan, 198, 11 pp. Hubbs, C. L. 1940. Fishes from the Big Bend region of Texas. Transactions of the Texas Academy of Science, 1 940: 1 - 12. Koster, W. J. 1957. Guide to the Fishes of New Mexico. Uni¬ versity of New Mexico Press, Albuquerque, New Mexico, 116 pp. Matthews, W. J., J. J. Hoover, and W. B. Milstead. 1985. Fishes of Oklahoma springs. The Southwestern Natural¬ ist, 30:23-32. McCorkle, R., G Garrett, and D. Riskind. 1998. An aquatic phoenix rises. Texas Parks and Wildlife Magazine, Feb¬ ruary: 36-43. Miller, R. R., C. Hubbs, and F. H. Miller. 1991. Ichthyological exploration of the American West: The Hubbs-Miller era, 1915-1950. Pp. 19-40, in Battle Against Extinction: Na¬ tive Fish Management in the American West (W. L. Minckley and J. E. Deacon, eds.), University of Arizona Press, Tucson. Peckham, R. C. 1963. Summary of the ground-water resources in the Rio Grande Basin. Texas Water Commission, 63- 05:1-16. Peden, A. E. 1 970. Courtship behavior of Gambusia (Poeciliidae) with emphasis on isolating mechanisms. Unpublished Ph.D. Dissertation, The University of Texas at Austin, Texas, 274 pp. Peden, A. E. 1 973. Virtual extinction of Gambusia amistadensis n. sp., a poeciliid fish from Texas. Copeia, 1973:210- 221. Rauchenberger, M. 1 989. Systematics and biogeography of the genus Gambusia (Cyprinodontiformes: Poeciliidae). American Museum Novitates, 295 1 : 1-74. Robison, H. W. and T. M. Buchanan. 1988. Fishes of Arkansas. University of Arkansas Press, Fayetteville, xvii +536 pp. Sumner, F. B. and M. C. Sargent. 1940. Some observations of the physiology of warm spring fishes. Ecology, 21:45- 54. Tupa, D. D. and W. K. Davis. 1976. Population dynamics of the San Marcos salamander, Eurycea nana Bishop. Texas Journal of Science, 32:179-195. Winemiller, K. O. and A. A. Anderson. 1997. Response of endangered desert fish populations to a constructed ref¬ uge. Restoration Ecology, 5:204-213. Young, D. A., K. J. Fritz, G P. Garrett, and C. Hubbs. 1994. Status review of construction, native species introduc¬ tions, and operation of an endangered species refugium channel, Phantom Lake Spring, Texas. Proceedings of the Desert Fishes Council, 25:22-25. Address of author: Clark Hubbs Integrative Biology The University of Texas at Austin Austin, TX 78712 My Favorite Old Fishing Holes in West Texas: Where Did They Go? James F. Scudday Abstract I provide a historical perspective of water re¬ sources that were present in Trans-Pecos Texas, par¬ ticularly Pecos County, from 1940-1980, with notes on the fish and other aquatic organisms that existed in these springs, creeks, lakes and rivers. Recreational fishing for game fish was the impetus for visiting these fishing holes. This in turn led to a young man devel¬ oping an interest in the natural history of the region. By the late 1950s, the springs were nearly all gone, wetland marshes disappeared, creeks disappeared, and the rivers looked more like sluggish creeks. The drought of the 1950s, coupled with the boom in pump irrigation projects, spelled the doom for this aquatic waterland. The current drought, in combination with ongoing water development projects, has further de¬ graded the aquatic environment. The future existence of native Chihuahuan Desert fishes under the current trends is unlikely. Comanche Springs I was fortunate as a youth growing up in Ft. Stockton, Texas, to be able to do so in a somewhat Tom Sawyerish way. My father was an avid outdoorsman, and he often took my brother and me with him on his many fishing excursions throughout west Texas. Fort Stockton was the center of an as¬ tonishing number of fishing holes in an otherwise desert environment. Natural springs gushed forth from nu¬ merous locales in all directions from Fort Stockton. The largest series of springs were right in the town of Fort Stockton. The entire Comanche Springs com¬ plex consisted of numerous springs along a nearly half- mile stretch of an arroyo bordering the east side of town. The largest single spring was the Chief Spring, a large spring that flowed an average of 35 million gallons (132 million liters) per day. The combined flow of the Chief with the other lesser springs amounted to an outflow of 66 million gallons (250 million liters) per day. This water was tapped early in the history of the area for irrigation of croplands north and northeast of town. As much as 6,200 acres (2,509 ha) of land was eventually watered from Comanche Springs (Brune, 1975). A municipal swimming pool was constructed around the Chief Spring during the 1930s, and a city park developed around the arroyo bottom to encom¬ pass all the other free-flowing springs. Canals from all these springs converged into one large canal just below the swimming pool. Native species of fish such as Mexican tetra (Astyanax mexicanus ), Comanche Springs pupfish ( Cyprinodon elegans), mosquitofish ( Gambusia affinis and/or perhaps G. nobilis), roundnose minnow ( Dionda episcopa) and channel catfish ( [Ictalurus punctatus), along with the common snapping turtle ( Chelydra serpentina ) and the yellow mud turtle ( Kinosternon flavescens), co-existed in this abundance of water. Largemouth bass ( Micropterus salmoides ) and exotic common carp ( Cyprinus carpio) were introduced into these springs at random by vari¬ ous individuals over time, and provided some fishing for youngsters and a few adults. These fish were very difficult to catch, probably because of all the natu¬ ral food already available. Every October, after the crops had matured, the irrigation canal gates were closed and all the water from the Comanche Springs was diverted down the old Comanche Creek channel. This was done to allow the irrigation district to clean and repair the concrete canals that delivered water to the various farming ar¬ eas. This abundance of water flowing down the old historical Comanche Creek bed re-created the histori¬ cal flow of Comanche Creek and its marsh-like condi- 136 Special Publications, Museum of Texas Tech University tions. An astounding number of fall migrating water- fowl was attracted to this wetland each year for the several months it lasted. The flow of water was usu¬ ally turned back into the irrigation canals by the end of December. All the springs of the Comanche Springs com¬ plex ceased to flow during the great drought of the 1950s. This drought, coupled with the increasing trend of clearing new farmland, and the drilling of irrigation wells that tapped the upstream aquifer of the springs, resulted in the loss of flow of nearly all the springs in Pecos County by 1957. Even now, if rainfall is abun¬ dant over a large area of the recharge region, the springs may flow just enough during the mid-winter months to send a small stream of water down the old creek bed for several months. I believe this last occurred in 1990 - 1991. When the pumps start again for a new growing season, the flow quickly stops. In about five years, the farming situation in central Pecos County shifted from a farming economy based on free flow¬ ing spring water, to one based on expensive pumping of water from the ground. Leon Springs Leon Springs, located about 8 miles (12.8 km) west of Ft. Stockton, was another large spring com¬ plex that was allied with the Comanche Springs com¬ plex. The Leon Springs once flowed down its own northward drainage called Leon Creek. It too dwindled away at the same time as the Comanche Springs com¬ plex. A general trend by the 1920s was for landown¬ ers that had a spring on their property to construct stone and earthen dams across the streambeds below the springs to hold water in a reservoir. Many of these spring-fed reservoirs were stocked with game fish to provide private fishing holes for the landowners and their friends. A large dam across Leon Creek formed a sizable lake, aptly named Leon Lake. The water in Leon Lake was used to irrigate a great many acres of land just north of the lake called the Webb Farms. Today, water pumped from the aquifer below the non¬ flowing springs irrigates about 100 acres (40.5 ha) of pecan trees on the old Webb Farm lands. In the 1940s, fishing rights at Leon Lake were restricted to members who paid dues into a fishing club. The fishing was great. The lake was full of largemouth bass and channel catfish. Native fishes were pretty well confined to the upper springs and the canal below the dam, where fisherman seined for baitfish. Mexican tetras were the preferred bait, but roundnose minnows, pupfish and killifish ( Fundulus ) were almost as good. Now I know the correct names of those baitfish, but back then we knew them as shin¬ ers, stripers, pigfish and zebras respectively. These same genera were also abundant in the cement-lined irrigation canals below Comanche Springs, and they were the source of baitfish for anyone preparing to go fishing within the region. Comanche and Leon springs had a combined flow of 40,000 acre-feet (49,000,000 cubic meters) in 1946 (Texas Water Development Board, 1984). Today it is zero. Other Springs Most of the springs in the area were tied into spring complexes interconnected by underground channels of the same aquifer. It has been historically noted that the headwaters of Comanche Springs flowed about 4 miles (6.4 km) down the Comanche Creek bed, where the water disappeared into the ground. Several kilometers below this point a string of at least eight named springs were located, each separated by 6 to 10 miles (9 to 16 km). These springs flowed from beneath rock ledges or gushed forth from what was described as “deep holes in the ground”. Each of these springs formed extensive marshes, but then the water disappeared into the ground downstream just as it did with Comanche Springs water. Muskrats were apparently plentiful in these marshes and in the Pecos River as well (Bailey, 1905). There are few records of fish at these springs. A local resident who once lived near Casa Blanca Spring is quoted as saying that as a young boy he caught fish Scudday — My Favorite Old Fishing Holes in West Texas: Where Did They Go? 137 on many a morning from the creek fed by Casa Blanca Spring for his family’s breakfast (Adams, 1984). The kinds of fish caught were not mentioned. The only spring of this chain I had any personal experience with was San Pedro Springs. This spring was deep and some of the channels leading from it were deep and wide. The San Pedro Land Company owned the land and irrigated several hundred acres (81-120 ha) of land by irrigation ditches. San Pedro Springs was a great fishing hole in the 1940s for large- mouth bass and several of the smaller centrarchid spe¬ cies. The landscape around the springs greatly re¬ sembled the Diamond Y Spring area today. I cannot recall any small, non-game fish present at San Pedro Springs. All these springs have been dry since the late 1950s. It is interesting to speculate what species of fish might have occurred in this string of springs in the 1800s. Leon Creek had a few small springs along its course, but the Diamond Y Spring system was the largest, and consisted of a large head spring and sev¬ eral smaller springs along its creek bed. The Diamond Y main spring still flows today, although the flow is reduced from what it flowed in the past. We fished at Diamond Y a few times, but it was never as produc¬ tive as other fishing holes. The spring itself was not good for fishing, but deep holes of water along the creek bed sometimes produced a nice catch. In the 1980s, I was surprised to find one of my old fishing holes still held quite a bit of water. Drilling for oil in the marshy area below the springs could have proved disastrous in the late 1980s, but fortunately, the dam¬ age was minimal. East of town, Tunas Springs was a favorite hang¬ out. The Tunas Springs (earlier known as Escondido Springs) consisted of a series of three springs along Tunas Creek. The headspring (or West Spring) is¬ sued forth from below a limestone ledge and ran east¬ ward for about 8 km until it disappeared into the creek bed, but it kept a number of deep holes filled with water. Catching fish in these deep holes was spo¬ radic, but it was a beautiful place to while away a summer day, and if fishing wasn’t good, hunting ar¬ rowheads was. A stagecoach stand was built here long ago close to the spring. During the 1930s, the highway department constructed a roadside park along old U.S. Highway 290 (now Interstate Highway 10) on the hill just above the springs, and built a replica of the stage stand at the park, uphill from its original lo¬ cation. The park is located about 32 km east of Fort Stockton on Interstate Highway 10. Irrigation wells drilled just west of the springs tapped that aquifer too, and the springs have been dry since the late 1950s. The only small fish I remember from this area prob¬ ably were Dionda, stripers in our fish vocabulary of the time. Pecos River and Tributaries Farther east, the Pecos River and two of its tribu¬ taries, Live Oak Creek and Independence Creek, were two of our favorite get-away-from-home-and-campout places to go fishing. The mouths of both these creeks could be real hot spots for catching fish, and great places to romp and play in the water. On my first trip to the mouth of Live Oak Creek, a memorable catch was made. A large freshwater eel was caught on a trotline set in the Pecos River one night. An old river fisherman by the name of Doss seemed to know all about eels. He informed us that he would fry the eel for supper that evening, and if any were left over we would have it for breakfast. However, he said, it would have to be refried for breakfast because fried eel turned raw when it got cold. I believe I tasted it that night, but not for breakfast. A number of large channel cat¬ fish were caught every time we fished that area. Live Oak Creek originates in Sutton County, passes just west of old Fort Lancaster, and empties into the Pecos River just east of Sheffield, Texas. The only fish I have a record of from Live Oak Creek is Fundulus zebrinus. These are in the Sul Ross State University (SRSU) collection and they were seined from beneath the Interstate Highway 10 bridge that spans the creek just east of the Pecos River crossing. Independence Creek lies in Terrell County, and in my memory, it is the most beautiful place within 300 km of the Pecos River. Our whole family made an annual pilgrimage in the summer for a number of 138 Special Publications, Museum of Texas Tech University years to camp for a week or more on the Lindsey Hicks ranch on the upper end of the creek. My brother and I were about 12 and 13 when we first made the trip from Fort Stockton. The purchase of the Hicks’ Ranch by Pinky Roden ended our family camping trips. Later in life I met Joe Chandler who owned the lower end of Independence Creek. My field zoology classes were always welcome at the Chandler Ranch and the Sul Ross Vertebrate Collection contains numerous bio¬ logical materials from that area. Here the Trans-Pecos copperhead (. Agkistrodon contortrix pictigaster) was first discovered by Dr. Frank Blair (Gloyd and Conant, 1943). Later, the first record of barking frogs (Eleutherodactylus augusti ) from Trans-Pecos Texas was reported from the limestone canyons above the confluence of Independence Creek and the Pecos River (Scudday, 1965). Specimens of the river carpsucker (Carpi odes carpio) from the creek and the Pecos River, and a single specimen of the blue sucker ( Cycleptus elongatus ) from the Pecos River just above the mouth of Independence Creek are deposited in the Sul Ross Vertebrate Collection. The creek itself supports a rich fish fauna, including the Rio Grande cichlid 0 Cichlasoma cyanoguttatum). If there was only one place you could visit in west Texas for overall beauty and biological diversity, go to Independence Creek. Below the mouth of Independence Creek, the Pecos River begins to cleanse itself as it flows through high limestone canyons and picks up additional fresh spring water. There was one place downstream from Independence Creek where we had to lower our camp¬ ing gear and ourselves by rope down a bluff to the riverbank. Trotline fishing for catfish was very good, and fishing with rods and reels during the day pro¬ vided a nice catch of a somewhat flat, shiny fish the men called a gaspergou. This was my first encounter with what I later learned is more correctly called the freshwater drum (A plodinotus grunniens ) Another good fishing place on the Pecos River was somewhere between Grandfalls and Imperial. I recall it was a large pool with water moving very slowly through it. My most impressive memory is of numer¬ ous gars swimming lazily on the surface of the water. My brother and I thought it looked like a great swim¬ ming hole, but the gars were too intimidating. Here was my second experience with catching gaspergou and my first introduction to the white bass ( Morone chrysops ). Phantom Lake Perhaps our all time favorite fishing hole was Phantom Lake near Balmorhea. Here the water poured forth from a large cave opening in the hillside to the west, forming a small lake of probably less than 3 ha. There was a low bluff on the north side and the water spread out to the south and east of the bluff. Catfish, largemouth bass and crappie (Pomoxis) were the main fish caught, but a variety of sunfish could also be caught. Dad’s primary interest was the largemouth bass, but my brother and I concentrated on the crap¬ pie. We thought they were the best eating fish of all the fish varieties we had sampled. As youngsters do, we would get tired of fishing and look for something else to do for a while. When exploring above the bluff, we found a large crevice that seemed to drop into a black hole. It was scary looking, and we decided to bring a flashlight with us the next time we came to Phantom Lake, which we did. When we shined the light into the hole, we saw there was a large flat boulder just below the opening. After dropping some rocks into the hole, we saw nu¬ merous bats flying over the big boulder. We were determined that some day when we got older we would come back and explore that hole with some proper equipment. I guess we were about 13 and 14 then. Three years later, about 1946 or 1947, we had our driver’s licenses, the war was over, and young men our age had lots of freedom. My brother and I, with three or four of our high school buddies, returned to Phantom Lake with flashlights, ropes, and leather gloves. We had no problem getting down onto the flat boulder, off the boulder, and onto the floor of what looked like a dry, underground stream. We went what seemed to be downstream a short distance when we came to a narrow crevice that led off to our left. Be- Scudday — My Favorite Old Fishing Holes in West Texas: Where Did They Go? 139 yond the entrance of the crevice, we could hear the sound of rushing water. The slenderest one of us decided he would try to follow the crevice. He quickly returned and said he went only about 1 5 meters when he encountered a running stream. He also stated that the crevice was wide enough for any of us to make it through, and it widened even more when it reached the water so at least two of us could stand there at one time. The most fascinating thing about the excursion was finding that shining the beam of a flashlight into the water attracted great numbers of small catfish right up to you. One could actually reach into the water and pick up a fish. The small fish appeared to be blind, having what appeared to be white membranes over the eyes, yet they could detect the light of the flashlight beam in the water and move toward it. None of the fish appeared to be more than about five to six inches in length, and all had darkly pigmented skin. Our last visit to the site was about 1948. More recently, I have been back to that area twice, once in the mid-1970s, and again in the 1980s. I thought for sure that I could go right to that opening, but I was unable to locate it. The lake itself has been long gone because of the loss of water volume. Man has altered the area around Phantom Lake Spring over time, perhaps in trying to coax more water from the springs and to discourage SCUBA divers from enter¬ ing the large spring opening. The first measure was to place a heavy grate over the large opening from which Phantom Spring emerged to keep divers from entering the spring. Several divers have died while trying to explore the origin of the spring. Clark Hubbs recently (November, 2001) informed me that the Reeves County Water Improvement District also filled in the openings thorough which we once entered the dry portion of the cavern, and that Phantom Lake Spring had ceased flowing completely in 1999. Red Bluff Lake Another of my favorite fishing holes was Red Bluff Lake on the Pecos River along the New Mexico state line. It is difficult to look at Red Bluff Lake today and realize what a great fishing lake this once was during the 1940s. The salt content of the water in the reservoir was much lower than it is today. Some of my earliest memories of Red Bluff are of standing be¬ low the spillway of Red Bluff dam and catching fish as torrents of foaming water poured over the spillway and down the Pecos River. I have no idea how long it’s been since water has gone over the spillway of Red Bluff dam. Fishing for largemouth bass was al¬ ways good in the lake itself using the baitfish we brought with us from Fort Stockton. Some men pre¬ ferred to grapple for big catfish with their hands by diving under water and feeling beneath the large rocks that formed the riprap on the lakeside of the dam. Probably because of the drought and the decrease of flow from the Pecos River, the lake was getting too salty for freshwater fish by the mid 1950s. Sometime after that, the Texas Parks and Wildlife Department began to experiment with stocking saltwater fish into the lake. As I remember, red drum ( Sciaenops ocellatus) and striped bass ( Morone saxatilis) were introduced, and perhaps other species. I never fished Red Bluff after about 1949, but I did hear that the redfish did quite well for a time. Rio Grande The Rio Grande was fished for catfish with trotlines, rod and reel, and jug floats. Channel catfish and flathead catfish ( Pylodictis olivaris ) were about all I remember catching. The Rio Grande flowed an abundance of clear water, unless it rained upriver. The fish we caught in the 1940s and 1950s were very good to eat. Channel catfish between 1 and 2 pounds (0.5 and 0.9 kg) were primarily caught in rapids on rod and reel, while it was not unusual to catch flathead catfish up to 9 kg on the trotlines and jugs. One of the best days catch I made with a rod and reel was about 1943 just upstream from the little village of Ruidosa, Presidio County, Texas. Good rapids were there then from which the village got its name. Today, the Rio Grande no longer flows consistently through that stretch of the river. The river does begin to flow again below Special Publications, Museum of Texas Tech University 140 Presidio where water from the Rio Conchos coming in from Mexico reconstitutes the Rio Grande. From this point downriver, the water is so low and polluted with chemicals and sewer effluent, the Big Bend Na¬ tional Park has posted signs warning visitors not to eat any fish they catch from the Rio Grande. Beavers ( Castor canadensis) still occur along some stretches of the Rio Grande bordering Brewster and Presidio counties. Muskrats once occurred along parts of the Pecos River and in the marshes of the Comanche, Leon, and San Soloman spring complexes (Bailey, 1905). The Pecos River muskrat ( Ondatra zibethicus ripensis) was described by Vernon Bailey in 1902, and may still exist today in irrigation drainage ditches in El Paso County and possibly southeastern New Mexico (K. Holmes, 1970; J. Holmes, 1970). Conclusion Now, in the year 2001, another devastating drought has plagued this region since 1992, a drought I consider worse than the one in the 1950s. Ongoing development of massive pump-irrigation projects and the acquisition of water rights by large cities contin¬ ues at a rapid pace in this desert region. Water levels in underground aquifers are dropping rapidly. The Rio Grande no longer flows into the Gulf of Mexico. The few springs that survived the 1950s drought are now disappearing. The U.S is threatening to sue Mexico for water owed the U.S., and Texas and New Mexico have continued to do court battle over the water of the Pecos River. Will we see more drastic water wars in the future? Given current trends, and based on per¬ sonal observations of declines during the past six de¬ cades, the eventual loss of the remaining Chihuahuan Desert fishes seems inevitable. Literature Cited Adams, M. O. 1984. Pecos County has surprising number of spring systems. Pp. 16-18 in Pecos County History. (Marsha Lea Daggett, ed.). Staked Plains Press, Canyon, Texas. Bailey, V. 1902. Fiber zibethicus ripensis. Proceedings of the Biological Society of Washington, 15:119. Bailey, V. 1905. Biological Survey of Texas. North American Fauna 25, 225 pp. Brune, G 1975. Major and Historical Springs of Texas. Texas Water Development Board Report 189, Pp. 19-22 as excerpted in Pecos County History (Marsha Lea Daggett, ed.). Staked Plains Press, Canyon, Texas. Gloyd, H.K. and R. Conant. 1943. A synopsis of the American forms of Agkistrodon (copperheads and moccasins). Bulletin ofthe Chicago Academy of Science, 7:147-170. Holmes, J. 1970. A 1970 Definition of Ondatra zibethicus ripensis in the vicinity of Clint, Texas. Unpublished Master’s Thesis, Sul Ross State University, Alpine, Texas, 35 pp. Holmes, K. 1970. The ecological status, present and future, of Ondatra zibethicus ripensis in the lower Rio Grande Val¬ ley of El Paso County, Texas. Unpublished Master’s Thesis, Sul Ross State University, Alpine, Texas, 74 pp. Scudday, J.F. 1 965. Eleutherodactylus latrans in Terrell County, Texas. The Southwestern Naturalist, 10:78. Texas Water Development Board. 1984. Water resources of Pecos County. Pp. 13-24. in Pecos County History (Marsha Lea Daggett, ed.). Staked Plains Press, Canyon, Texas. Address of author: James F. Scudday 514 E. Harriet Alpine, TX 79830 email: jimscudday@brooksdata.net Aquatic Conservation and The Nature Conservancy in West Texas JohnKarges Abstract Aquatic biodiversity in the Chihuahuan Desert portion of West Texas is high and of increasing con¬ servation concern because of the high incidence of endemism, limited range distributions for both species and natural communities and the limited areas where surface waters still occur. Surface waters include remaining intact or restorable reaches of principal river systems, perennial or permanent-pool streams, and isolated springs with their attendant outflows and marshes. Some aquatic sites are well known and con¬ served by either government agencies or private con¬ servation organizations, while other areas are either not protected or perhaps not even identified. Over the last decade, The Nature Conservancy has invested considerable money, time and resources in conserving specific, critical areas harboring rich aquatic biodiversity in the northern Chihuahuan Desert by pur¬ chasing preserves and partnering on adjacent lands with conservation easements to provide permanent protection of rare aquatic areas. The Conservancy’s protection and conservation efforts include identify¬ ing and mapping the distribution of rare aquatic spe¬ cies, assemblages and communities within a landscape context (ecoregional planning), long-term landsite pro¬ tection and research, monitoring and stewardship man¬ agement on the site and its biological elements. In addition, research and planning for groundwater is¬ sues of depletion and delineation of watersheds and recharge zones is crucial to sustainable, long-term conservation of these imperiled systems (site conser¬ vation planning). The Nature Conservancy presently is involved at five sites in West Texas with aquatic conservation elements, and through ecoregional plan¬ ning will identify additional “action sites” to protect the aquatic biodiversity of the Chihuahuan Desert. Introduction A recurrent and contemporary conservation is¬ sue throughout deserts involves aquatic biota and the impacts of declining surface water availability on popu¬ lations, species, assemblages of organisms or ecologi¬ cal function and integrity of entire systems (Rinne and Minckley, 1991). Throughout continental aridlands, surface waters frequently harbor endemic, rare or lim- itedly distributed species and in many cases, unique aquatic natural communities. The water may be flow¬ ing or stationary and either ephemeral or perennial, depending upon climatic cycles, seasonal durations, topography and underlying geology. Surface water bodies of springs and seeps, creeks, rivers, marsh¬ lands and playa basins are widely scattered across the northern Chihuahuan Desert landscape in Trans-Pecos Texas. With gradual but significantly increased aridification of the Chihuahuan Desert region since the end of the Pleistocene, surface waters have been greatly reduced. Consequently in many cases of isolated springs and perennial, relictual stream reaches, organ¬ isms isolated from wider, formerly contiguous distri¬ butions have differentiated into distinct genetic enti¬ ties, in some instances at the full species level, but often at least to recognizable populations or subspe¬ cies level. The rapid diminution and sometimes total loss of surface water in the past century due to a combination of climatic and anthropogenic factors is an additional and much more contemporary threat to the desert’s aquatic biodiversity. Aside from Chihuahuan Desert annual or seasonal droughts (typically from mid-win¬ ter to mid-summer), relatively long-term droughts of variable duration (up to nearly a decade in some cases), 142 Special Publications, Museum of Texas Tech University have occurred periodically since the inception of record keeping in the late 1 800s. This coupled with increased human demands and water extraction have contrib¬ uted to the measurable depletion of most surface spring systems in the Chihuahuan Desert (Hubbs, 1995), and has even included total elimination of some historically robust springs and outflows (e.g., Comanche Springs at Fort Stockton, Pecos County, see also Scudday, this volume). The threat to biodiversity by water reduction and over-harvest has already been manifested in the extirpation of the Comanche Springs pupfish ( Cyprinodon elegans ) at the type locality (U.S. Fish and Wildlife Service, 1981). In fact, within the Chihuahuan Desert portion of Texas, 56.4 percent (22) of the 39 total native fish species are of conservation concern as endangered, threatened or declining by the Texas Organization for Endangered Species (Edwards et al., 1989). An additional human-induced threat to regional aquatic biodiversity is the introduction of com¬ petitive exotics, which can seriously impact or degrade ecological integrity and even species’ genetic purity through hybridization (Hubbs, 1990). Among the types of systems and surface fea¬ tures containing identified “hotspots” of aquatic biodiversity within the northern Chihuahuan Desert are relatively intact or potentially restorable reaches of the major rivers, some remaining perennial or permanent pool streams fed by either springflows or runoff drain¬ age from mountainous basins; and desert springs and their associated streams and marshes (cienegas). The major river system of the region is the mainstem of the Rio Grande (Rio Bravo del Norte) and its signifi¬ cant tributaries, the Rio Conchos in Chihuahua, the Pecos River of New Mexico and Texas, and the Dev¬ ils River entirely within Texas. Perennial streams are scarce and widely distributed. Some of the tributaries of the Devils and Pecos rivers in the eastern Trans- Pecos are spring-fed from aquifers in the limestone matrix of the Stockton Plateau. This is an ecotonal region of plants, animals and biotic communities be¬ tween the eastern Chihuahuan Desert and the western Edwards Plateau. Farther west in the Trans-Pecos, streams in Big Bend National Park and Big Bend Ranch State Park contain permanent reaches of surface wa¬ ter, fed by subirrigated flow and augmented by peri¬ odic precipitation related episodes of flash-flooding. These surface pools and runs can be extremely vari¬ able in size, depth, distribution and extent, but as long as some permanent water remains throughout the year, refugia populations of aquatic species persist through seasonal droughts and even flourish during pluvial cycles and events. The last category of surface water types is the remaining desert springs and their atten¬ dant outflow runs and marshland systems (Hubbs, 1 995) that typically harbor endemics or extremely lim¬ ited range taxa. These spring systems often are indi¬ vidually distinctive in water volume and chemical pa¬ rameters. Considered cumulatively across the desert, they include a wide range of dissolved mineral con¬ tents, thermal conditions and regimes and even sub¬ terranean aquifer linkages between other nearby springs. History and Focus of The Nature Conservancy in Aquatic Conservation in West Texas Since 1990, The Nature Conservancy has oper¬ ated a West Texas Program and office within the re¬ gion, primarily to facilitate stewardship, administra¬ tion and other organizational functions. The Nature Conservancy’s conservation actions are predicated on its mission “to preserve the plants, animals and natural communities that represent the diversity of life on Earth by protecting the lands and waters they need to sur¬ vive” and specifically within the region, the biodiversity of the Trans-Pecos and northern Chihuahuan Desert. Because much of the endemism and biological rich¬ ness of the desert occurs at aquatic sites, the Conser¬ vancy focuses its conservation attention and resources on remaining springs and creeks with rare or endemic species and intact ecological function. Typically, sites were identified through the Natural Heritage Program database system of cataloguing and mapping biodiversity, specifically rare species’ occurrences, across each state. Several of the conservation sites for The Nature Conservancy within the Trans Pecos region are established primarily for preservation of aquatic resources and systems, including endemics. These sites may be either the Conservancy’s preserves owned in fee-title or private partnership lands upon which the Conservancy holds conservation easements with protective but consensual legally binding deed Karges — Aquatic Conservation and The Nature Conservancy in West Texas 143 restrictions. At some sites (see specific details be¬ low), notably Independence Creek and the Devils River, the Conservancy uses a combination of fee-owner¬ ship and easements on lands owned by others to af¬ fect conservation across a broader spatial scale than could be done by using only a single conservation tool. At others, The Nature Conservancy owns the preserve entirely, for example, Diamond Y Spring Preserve and Sandia Springs Preserve. Ecoregional Planning for the Chihuahuan Desert In the past few years, The Nature Conservancy has refined its conservation vision to focus on eco¬ logically functional conservation areas within ecoregions (The Nature Conservancy, 2001). Afunc¬ tional conservation area includes either focal species, natural communities or entire ecological systems, and the supporting ecological processes necessary to sus¬ tain them over the long term (Poiani and Richter, 2000). This planning effort is called ecoregional planning and a designed conservation portfolio represents the full distribution and diversity of conservation elements (in¬ cluding native species, natural communities and eco¬ logical systems) within the ecoregion. Ecoregional planning is designed to maximize conservation of biodiversity within and across ecoregions, while opti¬ mizing critical resource allocation toward site-specific conservation actions by the Conservancy and its con¬ servation partners. The basic steps in ecoregional planning are: 1) identifying the species, communities and ecological systems, 2) setting specific goals for the number and distribution of these conservation elements to be cap¬ tured by the portfolio, 3) assembling information and relevant data on the location and quality (contempo¬ rary status) of conservation elements, 4) designing a network of conservation areas that most effectively meets the goals, and 5) selecting the highest priority conservation areas in the portfolio for the Conservancy’s action and involvement. Although The Nature Conservancy has just initiated the Chihuahuan Desert ecoregional planning process, most of the aquatic elements of conservation concern (species, assemblages and communities) have been identified and many occur at what will be un¬ equivocal portfolio sites for conservation action. The sites where the Conservancy is currently involved within the ecoregion reaffirm that early assessments about where biodiversity occurs on the land (and wa¬ ters!) and past decisions about whether the Conser¬ vancy should work there, were in alignment with what is now termed ecoregional planning and “Conserva¬ tion by Design”. Future Conservancy “action sites” will undoubtedly include additional intact perennial or permanent pool stream segments and more mountain ranges with isolated springs with aquatic endemics or high composite conservation values based on species and communities. Diamond Y Spring Preserve This 607-ha (1,502-acre) preserve, located about 20 km NNW of Fort Stockton, Pecos County, con¬ tains the sole naturally occurring population of the en¬ dangered Leon Springs pupfish ( Cyprinodon bovinus). Originally described in the 1850s from the now oblit¬ erated Leon Springs area due west of Fort Stockton, the pupfish was not substantiated for nearly a century and was presumed extinct. Then in 1965, W. L. Minckley and W. E. Barber collected Leon Springs pupfish from Diamond Y Draw (see Echelle and Miller, 1974 for the historical account). With the rediscovery of the pupfish, academic and agency attention focused on the Diamond Y Spring area, and the importance of the site soon was found to include other aquatic taxa as well, including the federally endangered Pecos gam- busia ( Gambusia nobilis) and an obscure species of salt-tolerant sunflower (Pecos or puzzle sunflower, Helianthus paradoxus ) growing in permanently hydric soils along the stream course. In addition, three spe¬ cies of aquatic and littoral zone snails also occur at the preserve. Two are endemic, the Diamond Y spring snail ( Tryonia circumstriata, formerly T. adamantina, 144 Special Publications, Museum of Texas Tech University Hershler et al., 1999) and the Gonzalez spring snail ( Tryonia stocktonensis). The third species, Pecos assiminea {Assiminea pecos) has a wider distribution but is still essentially endemic to the middle Pecos River basin and was recently proposed for listing as feder¬ ally endangered. Other groups that may be represented as endemics or at least genetically distinct forms in the Diamond Y Spring system include the amphipod ge¬ nus Gammarus and an undescribed crayfish, both of which are under investigation currently. Diamond Y Spring is a solution cavity in the middle of an alluvial basin with a stream outflow into an alkaline marshland. The stream flows across the preserve to the confluence with Leon Creek which ends there as a named topographic feature. Diamond Y Draw continues northeastward through the preserve and across Pecos County to the Pecos River. There are two primary stretches of surface water with the rare, aquatic species, one from the spring to and be¬ yond the Leon Creek confluence, and a second reach about 2 km downstream to the north. Both reaches are augmented by small peripheral springs and seeps along their borders, and some of these harbor the en¬ demic invertebrate fauna. The Nature Conservancy acquired Diamond Y Spring Preserve in 1990, and began to initiate on-site stewardship almost immediately. The site is in an ac¬ tively producing oil and gas field. Energy production companies have been partners by providing support¬ ing funds for the purchase of the preserve and helped with certain safeguards for the protection of the sur¬ face waters from any contaminants that had threat¬ ened or plagued the site in the past. The measures included decommissioning buried corrosible metal pipe¬ lines in areas adjacent to vulnerable aquatic resources and their replacement with synthetic surface lines that could be easily monitored and repaired if necessary, as well as emergency shut-off valves installed at both sides of any creek crossings. At production sites, oil well pads were bermed to sufficiently contain any po¬ tential contaminant spill volume prior to detection. A matching grant in the mid-1990s from an energy pro¬ ducer and the National Fish and Wildlife Foundation provided the mechanism to remove some abandoned pad sites and their raised access roads from Diamond Y Draw which actually had impeded surface water flow through the marshland. An urgent conservation issue arose when genetic contamination of the Leon Springs pupfish was de¬ tected with measurable introgression and both eco¬ logical and genetic competition from the introduced congeneric sheepshead minnow (C. variegatus ) (Echelle and Echelle, 1997; Echelle et al., this volume). This genetic contamination had been addressed at Dia¬ mond Y Spring in the mid-1970s (Hubbs et al., 1978), but apparently persisted even after an intensive attempt to eradicate the exotic sheepshead minnows and hy¬ brids. Another effort to eradicate the sheepshead min¬ now genome was initiated in the late 1990s and is pres¬ ently ongoing. Investigators representing the Rio Grande Fishes Recovery Team outlined and imple¬ mented a restoration plan. The plan’s sequential ac¬ tions addressed the hybridization and competition is¬ sues through elimination of introgressed genomes and replacement with genetically pure Leon Springs pupfish stock from the reserve population held at Dexter Na¬ tional Fish Hatchery in Dexter, New Mexico. Prelimi¬ nary results on the effectiveness of the genetic resto¬ ration are promising (A. A. Echelle, Oklahoma State University, pers. comm.). For additional discussion on the conservation and ecology of the Pecos gambu- sia and the elimination of the competitive exotic largespring gambusia ( Gambusia geiseri ), in the Dia¬ mond Y Spring system see Hubbs (this volume and Echelle et al. this volume). Another conservation concern of paramount ur¬ gency is the pervasive threat to groundwater availabil¬ ity and spring discharge that sustains the suite of en¬ demic and rare obligate aquatic species and communi¬ ties at Diamond Y. Although the Conservancy has been effective at securing the immediate land around the spring and the watercourse with surface water, the issue of recharge and discharge volume is much larger than the presently protected land-base. Topographi¬ cally, Leon Creek enters the property with minimal (and sporadic) surface runoff and some subirrigated flow through the valley alluvium. The creek headwa¬ ters and surface drainage basin begin at the now dry Leon Springs west of Fort Stockton and courses north¬ eastward for approximately 1 6 km before entering the preserve. Surface flow of Leon Creek is ephemeral and seasonally episodic in the last km before the confluence with Diamond Y Draw, which does have permanent discharge from Diamond Y Spring and pe¬ ripheral seeps (Veni, 1991). Surface water presently Karges — Aquatic Conservation and The Nature Conservancy in West Texas 145 only extends for perhaps another one km before dis¬ appearing in the alluvium. A. A. Echelle and S. E. Kennedy (pers. comm.) recall much more extensive surface flow during the 1970s, including relatively broad standing water “flats” downstream, depicted in Kennedy’s (1977) map of the upstream segment. Emergent flows again appear on the surface approxi¬ mately 2-3 km downstream with peripheral spring dis¬ charge supplementing the subirrigated alluvial flows in the valley floor at the northeastern end of the preserve. The Conservancy and its conservation partners from agencies and academia face a tremendous chal¬ lenge with the problematic perpetuation of the spring discharge and therefore the sustenance of the aquatic system and species. Groundwater recharge and aqui¬ fer interactions still are imperfectly defined for the entire area (Veni, 1991; Boghici, 1997; Boghici and Van Broekhoven, 2001), and long-term measures to en¬ sure that discharge is sufficient to maintain the system’s functional integrity and conservation values are not in place. There are crucial adjacent tracts the Conser¬ vancy should protect at the immediate preserve bound¬ ary but this does not adequately protect a system fed by an ostensibly large but beleaguered off-site aquifer and potentially huge recharge zone that likely covers hundreds of square kilometers. These broader ground- water issues will be addressed through the Conservancy’s Site Conservation Planning process, a formal assessment of the conservation elements and threats with a strategic plan for permanent and effec¬ tive conservation as well as through species specific Recovery Plans. However, each of these planning ef¬ forts will have to take on the complicated and contro¬ versial issue of groundwater that will have far-reach¬ ing political and economic implications. The Conservancy’s on-site stewardship has in¬ cluded aggressive salt cedar ( Tamarix ramosissima) and mesquite ( Prosopis glandulosa ) control along the watercourse to reduce water consumption by these invasive species, both of which are relatively vora¬ cious water users. With the addition of prescribed fire, mechanical and judicious chemical treatment of these trees, this effort is to reduce as much superflu¬ ous water consumption as possible and retain it in the surface water segments of Diamond Y Draw. Sandia Spring Preserve At the northeastern outwash plain of the Davis Mountains is a complex, interrelated system of springs that include Phantom Lake, San Solomon, Giffin and Sandia springs as well as some relatively smaller seeps. Phantom Lake Spring is of grave conservation con¬ cern currently with virtually no outflow and conse¬ quently critical endangerment to the suite of aquatic rarities found in the immediate cave entrance, includ¬ ing Comanche Springs pupfish, Pecos gambusia and the invertebrate assemblage of snails and amphipods. San Solomon Spring in Balmorhea State Park has a strong discharge still, although heavily manipulated into a recreational impoundment and channelized outflows serving as pupfish refugia including a flow-through demonstration cienega for public interpretation. An¬ other portion of the spring complex just east of Balmorhea is East and West Sandia springs on 97 ha (240 acres) owned by the Conservancy. West Sandia Spring is a tiny spring, often with no apparent flow. It emanates from a cavity in the alluvial valley floor and the outflow channel is only several hundred meters in length with no known connection to either a larger drainage or even the irrigation canal system that sur¬ rounds it. Currently, no species of conservation con¬ cern are known from this system. However the larger East Sandia Springs, located several hundred meters to the east, does harbor species of concern including snails, amphipods, puzzle sunflower and Pecos gam¬ busia with historical occurrences of Comanche Springs pupfish. It also contains aggressive native predatory species like green sunfish ( Lepomis cyanellus), as does the San Solomon system, but they are relatively in¬ consequential to the rare fish species (Hubbs, 1993). At Sandia Springs, the critical issues are the same as those facing all of the other springs in the Balmorhea complex as well as at Diamond Y Spring. Exotic saltcedar invasion draining shallow groundwater re¬ serves and the possibility of sheepshead minnow ge¬ netic contamination of pupfish are two localized threats to the system’s integrity. The mature saltcedars have been removed mechanically but rapid and recurrent 146 Special Publications, Museum of Texas Tech University recruitment will require punctual, intensive and con¬ tinuous control. The overarching conservation di¬ lemma for the entire Balmorhea Spring complex is di¬ minished spring flows, delineation of the recharge zone and replenishment rates of aquifer(s) that contribute to surface discharge. Until these questions are an¬ swered and comprehensive threat abatement strate¬ gies are designed and implemented, most localized and proximate land conservation efforts are myopic stop¬ gap solutions at best. Independence Creek Megasite The Pecos River is divided into two distinct seg¬ ments in Texas from its entry at Red Bluff Reservoir on the New Mexico border to the confluence with the Rio Grande. Across the Pecos Plain (Permian Basin), the Pecos River is a slow, meandering river impacted by dewatering from diversions, groundwater with¬ drawal, saltcedar infestation and erosion (Hoagstrom, 2000). There is virtually no supplementary inflow from tributaries or springs. After entering the limestone canyonlands it has carved downstream of Iraan, it is augmented by very few perennial streams and some relatively small springs. Of these contributory sources, one of the most important in terms of both water qual¬ ity and volume is Independence Creek, which merges with the Pecos River in Terrell County, approximately 45 km south of Sheffield. Independence Creek has long been recognized as a biologically diverse site because of strong peren¬ nial instream flow and its riparian corridor. Ichthy¬ ologists have heralded Independence Creek as a clear, spring-fed refugium for the native Pecos River fish fauna, particularly as a repopulating source of the Pecos River following episodic and lethal mass fish mortalities resulting from red tides related to dinoflagel- late ( Prymnesium parvum ) blooms. These recurrent outbreaks periodically decimate fish populations along stretches of the Pecos River. Independence Creek and only a few other smaller freshwater tributaries between Iraan and the Rio Grande confluence at Lake Amistad are reestablishment sources as native fishes return to the river after the effects of the red tide blooms have abated (Rhodes and Hubbs, 1992). The assem¬ blage of native Pecos River fishes of conservation importance includes the Rio Grande darter ( Etheostoma grahami), proserpine shiner ( Cyprinella proserpina ), headwater catfish ( Ictalurus lupus). The shiner and darter recently have been confirmed in Independence Creek (Valdes, 1994; Karges, field notes, 2000-2001). Independence Creek is only 16 km in length from its outflow source spring to the confluence with the Pecos River, although the catchment basin extends northwestward nearly to Fort Stockton in Pecos County. The site is called a megasite because of the Conservancy’s recognition that it is not sufficient to conserve just a small area immediately around the creek but also an imperative opportunity to work at the land¬ scape scale on a large and significant watershed which affects the creek’s functional integrity. The Nature Conservancy’s conservation actions at Independence Creek began with the 1990 acquisition of a 284-ha (702-acre) conservation easement with private land- owners along the last 3.2 km of the creek including the Pecos River confluence. The easement is struc¬ tured principally to eliminate degradation of the creek and adjacent riparian corridor, and to retain as much instream flow as possible. In 2000, the Conservancy acquired an 3517-ha (8,690-acre) ranch immediately upstream from the easement property, with the pri¬ mary purpose of protecting an additional 4.8 km of surface water. The ranch also contains Caroline Springs, a significant, large volume instream source supplementing the creek. However, the springs have been impounded into two relatively large recreational lakes. These lakes, when combined with an additional source of evaporative loss through a large-scale aerial sprinkler array for lawns and channelized irrigation network for improved pastures, precludes much of the spring waters from rejoining the Independence Creek mainstem. Also, introduced large, predatory fishes including largemouth bass ( Micropterus salmoides ) and channel catfish ( Ictalurus punctatus ) in the reservoirs, could hinder the persistence of the native fishes. The Conservancy is designing a water budget monitoring strategy for assessing water losses between the spring and the creek, with the goal of redirecting Karges — Aquatic Conservation and The Nature Conservancy in West Texas 147 water for conservation of the target species and re¬ storing the strength and integrity of the creek. An additional acquisition of the next upstream ranch adds another 4.4 km of Independence Creek under the Conservancy’s management and stewardship, leaving only the remaining 3.2 km from the headwaters with¬ out conservation protection. Another stewardship ac¬ tion to restore instream water volume along the man¬ aged reach of Independence Creek is the initiation of an aggressive saltcedar eradication campaign and re¬ duction of the native, but pervasive, false willow ( Baccharis neglecta ), both of which rob water from the creek. Dolan Falls/Devils River Megasite The Devils River is the next major Rio Grande tributary east of the Pecos, in the overlap zone be¬ tween the Chihuahuan Desert to the west, the Edwards Plateau to the east and the Tamaulipan Thomscrub ecoregion to the south. This system is defined by a catchment basin that is relatively small on a continen¬ tal scale. It is of considerable significance for both terrestrial and aquatic conservation elements that rep¬ resent the biodiversity where these three ecoregions converge. The surface water of the Devils River ex¬ tends for approximately 90 km from its headwaters to Lake Amistad on the Rio Grande, entirely within Val Verde County. The river is generally characterized by long flat-water pools of varying depths with slow- moving currents punctuated by broad, shallow riffles, stair-step cascades, and channelized flow constrained within the fluted bottom topography of the limestone bedrock. Just above Dolan Falls, Dolan Creek joins the Devils River with a substantial spring-fed contri¬ bution as the largest tributary along the river, although there are numerous spring-fed smaller rivulets and basal springs along much of the river between the headwa¬ ters and Lake Amistad. Neither dams nor pollution have yet impacted the Devils River, and it remains the most intact free-flowing river in West Texas (Harrell, 1978). Texas Parks and Wildlife Department initiated the first conservation land acquisition along the river with a purchase of about 7,689 ha (19,000 acres) in the 1980s creating the Devils River State Natural Area. The purchase included 2 km of riverfront and the head¬ waters of Dolan Creek. Subsequently, The Nature Conservancy purchased the Dolan Falls Ranch, about 7,284 ha (18, 000 acres) in 1 99 1 , protecting Dolan Creek from the headwaters to the confluence. This also en¬ compassed land on either side of Dolan Falls, includ¬ ing nearly 6.4 km of riverfront on the Devils River. The integrity of the hydrology and its diverse biota in the contributing springs, creek and river has been the primary focus of both entities’ land protection efforts. In 2000, the Conservancy also acquired the 8, 903 -ha (22,000-acre) Devils River Ranch downstream from Dolan Falls adjacent to Lake Amistad National Recre¬ ation Area, protecting an additional 19-km segment of the river that harbors the native fish fauna and other aquatic elements. All of the land conservation activity in this area includes terrestrial species and communi¬ ties as well as abating local threats of development and subdivision within the watershed. The Nature Con¬ servancy has developed a site conservation plan for the entire megasite, which includes much of the wa¬ tershed but focuses specifically on the surface waters and immediate drainages of the surface water portion of the Devils River. The formation of Lake Amistad in the late 1960s - early 1970s has impacted the fish fauna of the Trans- Pecos portion of the Rio Grande basin including its Pecos River and Devils River tributaries. Inundation of the confluences as the lake filled fragmented the contiguous distribution of the native guild of river- adapted fishes. Subsequent introduction of exotic predatory species such as striped bass ( Morone saxatilis) and smallmouth bass ( Micropterus dolomieu ) has likely contributed to the decrease of some native species to some extent. Three native fish species are considered extirpated within the system and it is un¬ likely that residual populations or remaining suitable habitats exist in any peripheral areas. Most of the rare species of this riverine fauna require the flowing wa¬ ters found in the Pecos and Devils rivers (Valdes Cantu and Winemiller, 1997). Three other rare fishes are associated with flowing waters of the Devils River, including proserpine shiner, Devils River minnow (. Dionda diaboli ) and Rio Grande darter, and popula- 148 Special Publications, Museum of Texas Tech University tions have been confirmed in the Devils River recently (Garrett et al., 1992). Headwater catfish have also been confirmed in the last two decades from both the lower Pecos River (Kelsch and Hendricks, 1990) and Devils River system (Dolan Springs/Creek in 1980, Garrett et al., 1992). This species is morphologically very similar to the widespread and frequently intro¬ duced channel catfish, but is known principally from clear, spring-fed systems. The Devils River below Dolan Falls is primarily lentic with only short reaches of swift water, riffles and limited suitable habitat for fishes requiring those conditions (Harrell, 1978). The local pupfish has been identified as the Conchos pupfish ( Cyprinodon eximius) but may represent a disjunct, undescribed form that is endemic to the Devils River above the lake (Hubbs et al., 1991). The pupfish has been reestablished in Dolan Creek through introduction from populations in the Devils River (Hubbs and Garrett, 1990) and may be found in lentic pools along the Devils River down¬ stream to at least Pafford’s Crossing (Davis, 1980) and Big Satan Canyon (Karges, unpublished field notes). An additional aquatic species that may prove to be an undescribed endemic is a neotenic salamander in the genus Eurycea. A few specimens are known from springs along Dolan Creek. Most of the on-site stewardship and monitoring actions related to aquatic conservation at Dolan Falls Preserve have been the continuation of periodic sur¬ veys of the fish fauna (Valdes Cantu and Winemiller, 1997; G. P. Garrett, Texas Parks and Wildlife Depart¬ ment, pers. comm.), aquatic invertebrate inventories, and in redirecting visitor use and foot traffic at the fragile micro-habitats at the spring outflows. Madera Canyon Preserve Within the Davis Mountains, few perennial streams remain and of these, Little Aguja Canyon (Jeff Davis County) on the northern slopes is among the most important because of the presence of two aquatic species. The federally endangered Little Aguja pond- weed ( Potamogeton clystocarpus ) is a cryptic species known only from this canyon, in plunge pools and subirrigated permanent pools along the middle and upper reaches of the drainage. Also in these perma¬ nent pools, the Rio Grande chub ( Gila pandora) sur¬ vives as the only relict population remaining in Texas (Miller and Hubbs, 1962). Although the species has not been collected in 1 6 years, 4 individuals were seen in a pool on the Conservancy’s Madera Canyon Pre¬ serve in 2000 (Karges, unpublished field notes) and likely remains in other similar pools throughout the upper and middle sections of the canyon. The Nature Conservancy includes this area in the overall site con¬ servation plan for the Davis Mountains which includes other rare aquatic species including Davis Mountains spring snail (Fontilicella davisi) (Taylor, 1987) and the possibility of reintroducing Rio Grande cutthroat trout ( Oncorhynchus clarki ) to Davis Mountains streams (Garrett and Matlock, 1991) if some highland streamcourses can be restored to perennially flowing montane systems with landscape scale watershed management to restore recharge and instream flows. Dedication This paper and my conference presentation are dedicated to the memory of Dr. W. L. Minckley for his contributions to desert fish research and conser¬ vation. Also, during a Chihuahuan Desert conference in Monterrey, Nuevo Leon (the only time we ever met), he recounted the delightful and intriguing story of his role in the rediscovery of the “extinct” Leon Springs pupfish. Karges — Aquatic Conservation and The Nature Conservancy in West Texas 149 Acknowledgements I thank N. Allan, A. A. Echelle, A. F. Echelle, G. Garrett, C. Hoagstrom, and C. Hubbs for inspiration, challenging “brainstorms,” invitations to participate in fieldwork and sharing their time, knowledge and exu¬ berance for desert fish diversity and its conservation. Literature Cited Boghici,R. 1997. Hydrogeological investigations at Diamond Y Springs and surrounding area, Pecos County, Texas. Un¬ published Master’s Thesis, University of Texas at Aus¬ tin, 120 pp. Boghici, R. and N.G Van Broekhoven. 2001. Hydrogeology of the Rustler Aquifer, Trans-Pecos Texas. Pp. 190-206, in Aquifers of West Texas (R.E. Mace, W.F. Mullican III, and E.S. Angle, eds), Texas Water Development Board Report 356, Austin, Texas, 272 pp. Davis, J. R. 1980. Rediscovery, distribution, and populational status of Cyprinodon eximius (Cyprinodontidae) in Devil’s River, Texas. The Southwestern Naturalist, 25: 81-88. Echelle, A. A. and R. R. Miller. 1974. Rediscovery and rede¬ scription of the Leon Springs pupfish, Cyprinodon bovinus, from Pecos County, Texas. The Southwestern Naturalist, 19: 179-190. Echelle, A. A. and A. F. Echelle. 1 997. Genetic introgression of endemic taxa by non-natives: a case study of Leon Springs pupfish and sheepshead minnow. Conservation Biology, 11:153-161. Edwards, R. J., G Longley, R. Moss, J. Ward, R. Matthews, and B. Stewart. 1989. A classification of Texas aquatic com¬ munities with special consideration toward the conserva¬ tion of endangered and threatened taxa. Texas Journal of Science, 4 1:23 1 -240. Garrett, G P. and G C. Matlock. 1991. Rio Grande cutthroat trout in Texas. Texas Journal of Science, 43: 405-410. Garrett, G P., R. J. Edwards, and A. H. Price. 1992. Distribu¬ tion and status of the Devils River minnow, Dionda diaboli. The Southwestern Naturalist, 37:259-267. Harrell, H. L. 1978. Response of the Devil’s River (Texas) fish community to flooding. Copeia, 1978:60-68. Hershler, R., H.-P. Liu, and M. Mulvey. 1999. Phylogenetic relationships within the aquatic snail genus Tryonia : im¬ plications for biogeography of the North American South¬ west. Molecular Phylogenetics and Evolution, 13:377- 391. Hoagstrom, C. W. 2000. Pecos River fishes. Confluencias, July 2000:3-4. Hubbs, C. 1990. Declining fishes of the Chihuahuan Desert. Pp. 89-96 in Third Symposium on Resources of the Chihuahuan Desert Region (Powell, A. M., R. R. Hol¬ lander, Barlow, J. C., MacGillivray and D. J. Schmidly, eds). Chihuahuan Desert Research Institute, Alpine, Texas. Hubbs, C. 1993. Fishes of the refugium at Balmorhea State Park. Texas Park and Wildlife Department Booklet P44503-002C, Austin, Texas, 8 pp. Hubbs, C. 1995. Springs and spring runs as unique aquatic systems. Copeia, 1995:989-991. Hubbs, C., T. Lucier, E. Marsh, G P. Garrett, R. J. Edwards, and E. Milstead. 1 978. Results of an eradication program on the ecological relationships of fishes in Leon Creek, Texas. The Southwestern Naturalist, 23:487-496. Hubbs, C. and G. P. Garrett. 1990. Re-establishment of Cyprinodon eximius (Cyprinodontidae) and status of Dionda diaboli (Cyprinidae) in the vicinity of Dolan Creek, Val Verde Co., Texas. The Southwestern Natural¬ ist, 35:446-478. Hubbs, C., R. J. Edwards, and G P. Garrett. 1991. An anno¬ tated checklist of the freshwater fishes of Texas, with keys to identification of species. Texas Journal of Sci¬ ence Supplement Vol. 43(4), 56 pp. Kelsch, S. W. and F. S. Hendricks. 1990. Distribution of the headwater catfish Ictalurus lupus (Osteichthyes: Ictaluridae). The Southwestern Naturalist, 35:292-297. Kennedy, S. E. 1 977. Life history of the Leon Springs pupfish, Cyprinodon bovinus. Copeia, 1977:93-103. Miller, R. R. and C. Hubbs. 1962. Gila pandora , a cyprinid new to the Texas fish fauna. Texas Journal of Science, 14:111-113. Poiani, K. and B. Richter. 2000. Functional landscapes and the conservation of biodiversity. Working Papers in Conser¬ vation Science, The Nature Conservancy, Conservation Science Division, Arlington, Virginia, 1:1-11. Rinne, J. N. and W. L. Minckley. 1991. Native fishes of arid lands: a dwindling resource of the Desert Southwest. General Technical Report RM-206. Ft. Collins, CO: United States Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experimental Sta¬ tion, 45 pp. 150 Special Publications, Museum of Texas Tech University Rhodes, K. and C. Hubbs. 1992. Recovery of Pecos River fishes from a red tide fish kill. The Southwestern Natu¬ ralist, 37: 178- 187. Taylor, D. W. 1987. Fresh-water molluscs from New Mexico and vicinity. Bulletin of the New Mexico Bureau of Mines and Mineral Research, 1 16:1-50. The Nature Conservancy. 2001. Conservation by design: a framework for mission success. The Nature Conservancy, Arlington, Virginia, 12 pp. U. S. Fish and Wildlife Service. 1981. Comanche Springs pupfish ( Cyprinodon elegans) recovery plan. United States Fish and Wildlife Service, Albuquerque, New Mexico, 1 7 pp. Valdes, N. E. 1994. Composition and structure of fish assem¬ blages of Chandler Independence Creek Preserve. Un¬ published report to The Nature Conservancy, Fort Davis Texas, 8 pp. Valdes Cantu, N. E. and K. O. Winemiller. 1997. Structure and habitat associations of Devils River fish assemblages. The Southwestern Naturalist, 42:265-278. Veni. G 1991. Delineation and preliminary hydrogeologic in¬ vestigation of the Diamond Y Spring, Pecos County, Texas. Unpublished Report, The Nature Conservancy of Texas, San Antonio, Texas, 110 pp. Address of author: John Karges The Nature Conservancy P.O. Box 2078 Fort Davis, TX 79734 email: jkarges@tnc.org Innovative Approaches to Recover Endangered Species Gary P. Garrett Abstract Texas Parks and Wildlife Department has em¬ barked on a progressive approach to resolving endan¬ gered species issues through involvement with local governments and especially private landowners. Work in the Balmorhea area of West Texas involved local citizens, the City of Balmorhea, three universities, four NGOs, five state and three federal agencies. Together we created a natural cienega, made progress towards eliminating a source of genetic contamination, enhanced nature tourism and began developing an improved sport fishery. We are now working on Conservation Agree¬ ments in West Texas that will enable the resolution of concerns over two candidate fish species and, if suc¬ cessful, will preclude the need to list these species as endangered. With the Devils River minnow, we are working closely with landowners and the City of Del Rio to determine and resolve life history requirements and restore populations to natural levels. In so doing, the cooperating entities will also be protecting the quality of the Devils River and associated streams. A major component of the Pecos Pupfish Conservation Agree¬ ment is to create alternate habitats on private land in Texas. By involving individuals and local governments, we are more likely to achieve long-term benefits for natural resources as well as public health and quality of life. Introduction Approximately 25% of the 169 native freshwa¬ ter Texas fishes are of conservation concern (Hubbs et al., 1991). That is, they are either in danger of extinction or approaching that status. In the Chihuahuan Desert region of Texas, 50% of the native fishes are of conservation concern or already lost to extirpation or extinction (Hubbs et al., 1991). When species are being lost, their predicament is usually in¬ dicative of a larger biological problem. Reversing the trend and avoiding extinctions is usually complex, po¬ litically volatile and expensive. Texas Parks and Wildlife Department (TPW) is working with federal, state and local agencies and es¬ pecially private landowners to conserve natural re¬ sources and resolve endangered species issues. With 97% of the land in Texas privately owned, involve¬ ment with the private sector is often the only way to achieve long-term conservation goals. Herein, I present three case histories of innova¬ tive approaches to resolve specific issues. Although each situation has unique elements, these case histo¬ ries serve to provide examples that have worked and to generate new ideas on how to approach such is¬ sues. Comanche Springs Pupfish and San Solomon Cienega In 1996, a cooperative effort among private, state and federal entities allowed the creation of a desert- wetland (cienega) habitat for two federally endangered fish species, Comanche Springs pupfish ( Cyprinodon elegans ) and Pecos gambusia ( Gambusia nobilis). The primary benefit to the fishes is a “natural” habitat criti¬ cal to their survival, but which had been eliminated through human modifications for recreation and agri¬ culture. Benefits to area residents included: 1) relax¬ ation of some pesticide regulations for farmers, 2) pro¬ tection of the water supply and 3) increased tourism. 152 Special Publications, Museum of Texas Tech University Figure 1 . Project locations in West Texas. Prior to human alterations, Comanche Springs pupfish and Pecos gambusia inhabited two large cienega systems separated by approximately 100 km (Figure 1), one fed by the Balmorhea springs complex (Phantom Lake, San Solomon, Giffin and East Sandia springs), and one by Comanche Springs. San Solomon Springs is the largest spring in the Balmorhea springs complex, producing about 750 liters per second (lps); it is presently the largest spring in the Trans-Pecos and the seventh largest in Texas. Comanche Springs, in Fort Stockton, once flowed at approximately 1,200 lps, but ceased its perennial flows in 1961 due to groundwater pumping (Brune, 1981; Scudday, this volume). More than 100 farmers depended on surface, irrigation waters flowing out of Comanche Springs and the cienega. Groundwater pumping by 1 8 land- owners in an area west of Fort Stockton severely di¬ minished the flows from the spring. The local water district sued these pumpers in 1952 in an attempt to establish their water rights. The pumpers prevailed in the lawsuit by basing their defense on a 1904 case from which had emerged the concept of “right of cap¬ ture.” This concept established that a well owner could pump as much water as desired, regardless of the im¬ pact on the aquifer. This was also the case in which the Texas Supreme Court had determined that the in¬ tricacies of aquifers were so “secret, occult, and con¬ cealed” that it would be impossible to administer a set of rules. Ultimately the flows of Comanche Springs ceased, the cienega dried up, the native flora and fauna disappeared, the surface irrigators lost their farms and their land reverted to desert. Similarly, farmers also diverted water for agri¬ culture from the Balmorhea springs complex and have been doing so since the mid- 1870s (Brune, 1981). In 1915, the Reeves County Water Improvement District No.l (RCWID) was established and, with water from San Solomon and other associated springs, adminis¬ tered irrigation water for 4,900 ha of farmland. Garrett — Innovative Approaches to Recover Endangered Species 153 Cienegas presumed to have supported large numbers of C. elegans and G. nobilis were drained and spring flows were diverted into an irrigation network of con¬ crete-lined canals with swiftly flowing water and dredged, earthen laterals. This habitat is highly un¬ natural, ephemeral and wholly dependent upon local irrigation practices and other water-use patterns. In the 1930s, the Civilian Conservation Corps modified San Solomon Springs into a large swimming pool at Balmorhea State Park. The work of this New Deal program enhanced the park’s visitor services, but fur¬ ther disrupted the natural cienega. Cienegas, and their associated springs, provide habitat for a wide variety of plants and animals, some of which are endemic to these systems (Hendrickson and Minckley, 1984). Not only can cienegas harbor unique species, but also an entire community of inter¬ acting organisms depends on these fragile habitats for survival. This is especially true for the increasingly rare desert fishes. Few cienegas have survived intact to this day. When the original San Solomon cienega was modified, and for the most part destroyed, the only “aquatic habitat” remaining was in the concrete irrigation canals. Although better than no habitat at all, the irrigation canals, at best, provided a tenuous exist¬ ence for some life forms. Some indigenous species, such as the Pecos River muskrat ( Ondatra zibethicus ripensis ), did not adapt and were extirpated. The Comanche Springs pupfish and Pecos gambusia man¬ aged to survive in the irrigation canals, but their num¬ bers were greatly reduced. Because of the loss of most of their natural habitat, both fishes are rare and on the federal and state Endangered Species lists. Previous efforts to improve habitat have occurred in the Balmorhea area. A small refugium canal (120 m) was constructed in 1974 at Balmorhea State Park (Echelle and Hubbs, 1978). During a two-year sam¬ pling study (Garrett and Price, 1993), Comanche Springs pupfish population size in the park refugium canal was estimated to be as low as 968 (May 1990) and as high as 6,480 (September 1990). In 1993, a modified canal was constructed at Phantom Lake Springs by the Bureau of Reclamation (Young et al., 1994). Instead of the original concrete walls, the 110-m canal has sloped, earthen, sinuous sides and was designed to resemble a portion of a cienega. The Phantom Lake Springs Refugium Canal resulted in an increase in local abundance of Comanche Springs pupfish, resulting in an average of 14.7 pupfish/ m2 (Winemiller and Anderson, 1997). Unfortunately, the springs have now failed (Hubbs, 2001) and a pump is needed just to maintain water in a small pool at the spring source (N. Allan, U.S. Fish and Wildlife Ser¬ vice, pers. comm.). People also suffer when their water sources van¬ ish. Farmers who depended on surface irrigation wa¬ ter from Comanche Springs lost their livelihood when the springs went dry. Farmers in the Balmorhea area also rely on surface irrigation from springs, and if the aquifer were diminished, local agriculture would cer¬ tainly suffer. The effects on the rest of the commu¬ nity of Balmorhea would be devastating since they depend on the aquifer and the spring flows for every¬ thing from domestic water to tourism. Although current state law would allow unre¬ stricted pumping from the aquifer that supports the Balmorhea springs complex, one thing that can pre¬ vent overpumping is the Federal Endangered Species Act. The Endangered Species Act protects the fish, the fish need the water, and as long as the water is flowing from the springs it also is available to humans. Through a pragmatic understanding of the basic rela¬ tionship between the natural and human communities, biologists and Balmorhea community leaders chose to work together on a solution that would benefit all con¬ cerned rather than adopt adversarial roles. While the farmers had previously viewed the fish as something that hampered and perhaps threatened their livelihood, the fish may actually be their best insurance for sus¬ tained spring flows. A plan was formulated to create a cienega to look and function like a natural ecosystem. RCWID and the agricultural community it represents agreed to pro¬ vide the essential water needed to create a secure en¬ vironment for the endangered fishes. Water is a rare and precious commodity in West Texas, particularly for farmers, but by each of the users giving up a small amount, they would be providing insurance for future water supplies. An additional benefit for the farmers was that, because of their help in creating preferred habitat for 154 Special Publications, Museum of Texas Tech University the endangered fishes, the Texas Department of Agri¬ culture (TDA), the U.S. Fish and Wildlife Service (USFWS) and the U.S. Environmental Protection Agency (EPA) proposed a plan to allow the benefits of the cienega to offset any potential effects from pesti¬ cide use on farms that could impact the endangered species in the irrigation canals. The fish would have a better place to live and the farmers could continue to raise their crops. Biologists, engineers and resource managers from universities and government agencies joined forces to make the project work. The U.S.D.A. Natural Re¬ source Conservation Service provided soil analysis and, along with staff from the Texas Agricultural Extension Service, TDA, the University of Texas at Austin and the University of Texas-Pan American, gave expert advice on some of the intricacies of the project. The expertise of the Texas Department of Transportation also was crucial. Their surveyors, design engineers and equipment operators transformed biological con¬ cepts into reality. The Texas Department of Criminal Justice provided inmates to build the Observation Deck and retaining walls as well as install the plant materials selected for the initial cienega vegetation restoration. Botanists at Sul Ross State University provided con¬ tainer-grown native plants for the project. The one- of-a-kind window wall was designed, built, transported and installed by a beneficent concrete fabrication com¬ pany located 500 km away. Funding for the San Solomon Cienega was pro¬ vided by grants from the Educational Foundation of America and the National Fish & Wildlife Foundation. Additionally, fabrication costs for the window wall were provided by a TDA grant from the EPA and with con¬ tributions from the Texas Organization for Endangered Species. In 1996, construction of the 1-ha San Solomon Cienega was completed. This wetland is situated on Balmorhea State Park land within the boundaries of the original, natural cienega. As a result, the native fish fauna, including Comanche Springs pupfish and Pecos gambusia, has flourished. This location now provides a natural habitat and contains the largest known concentration of Comanche Springs pupfish. Recent monitoring efforts have resulted in average estimates of the summer population of pupfish in the cienega at 270,000 individuals. Aquatic plants indigenous to cienegas, as well as grasses and shrubs characteristic of the drier aspects of these desert wetland communities, were planted in the cienega and now are well established. Some of the more common species are common cattail ( Typha latifolia ), common reed {Phragmites australis), alkali bulrush ( Scirpus maritimus ), hardstem bulrush ( S . acutus ), Olney bulrush {Scirpus olneyi ), sand spikerush {Eleocharis montevidensis), buttonbush {Cephalanthus occidentalis ), Goodding willow {Salix gooddingii), Rio Grande cottonwood {Populus deltoides var. wislizenii), four-winged saltbush {A triplex canescens ), alkalai sa- caton ( Sporobolos airoides ), big sacaton {Sporobolos wrightii), tobosa {Hilaria mutica ), granjeno ( Celtis pallida) and western soapberry {Sapindus saponaria var. drummondii). Beyond the immediate wetted pe¬ rimeter of the cienega, the habitat grades into a desert plains grassland that was once common to the region. Many species of birds, reptiles and mammals began to use the new wetland almost immediately. These include belted kingfisher (Megaceryle alcyon ), black phoebe {Sayornis nigricans), swallows (Petrochelidon spp.), white-throated swift (Aeronautes saxatalis), green heron {Butorides virescens), swamp sparrow {Melospiza georgiana), yellow-headed black¬ bird {Xanthocephalus xanthocephalus), sora ( Porzana Carolina), yellowthroat ( Geothlypis trichas), blotched watersnake {Nerodia erythrogaster transversa), spiny softshell turtle {Trionyx spiniferus), pond slider {Chrysemys scripta), javelina {Pecari angulatus) and desert cottontail {Sylvilagus audoboni). People of the local and regional community and state park visitors benefit from a living exhibit that shows the importance of the springs and their wet¬ lands for the fishes and other wildlife of West Texas. Because the primary purpose of the cienega is to pro¬ vide desert wetland habitat, visitor access is limited to only a small portion of the total restoration. However, TPW has tried to maximize the aesthetic and educa¬ tional experiences available at locations that are acces¬ sible to the public. The observation deck provides an unobstructed view of most of the above-water por¬ tion of the cienega, and the clear water allows viewing of much of its underwater life. The window wall was custom designed for San Solomon Cienega so that visi¬ tors would have a view that few have seen — life in the cienega as its aquatic residents see it. Garrett — Innovative Approaches to Recover Endangered Species 155 Another attempt to further protect Comanche Springs pupfish was not so successful. Lake Balmorhea is a downstream, 200-ha storage reservoir for irrigation water from the Balmorhea springs com¬ plex and contained an introduced population of sheep- shead minnow (C. variegatus). This species is known to compete and hybridize with Comanche Springs pupfish (Stevenson and Buchanan, 1973; Echelle and Echelle, 1994). In 1998, the RCWID allowed TPW to partially drain the reservoir and attempt to remove all fish by application of rotenone, a fish toxin. The ob¬ jective of the project was to eliminate sheepshead min¬ now from the reservoir in order to remove the threat of hybridization with Comanche Springs pupfish. A large population of sheepshead minnow inhabited the lake; post-rotenone extrapolation of subsamples put the estimate at 5,000,000. Unfortunately and for un¬ determined reasons, some sheepshead minnow sur¬ vived the rotenone treatment. They have since begun repopulating the reservoir. Part of the agreement with RCWID was to restock the reservoir with sport fishes in order to improve tourism in the area. These piscivores should help keep the numbers of sheeps¬ head minnow in check. The creation of the San Solomon Cienega was accomplished through willing participation of diverse entities with a common goal of mutual benefit. During the last decade, USFWS developed a formal method of participation in such projects (for non-listed spe¬ cies) through Conservation Agreements. The success¬ ful cooperation among private and government enti¬ ties in the San Solomon Cienega project served as a precursor to future conservation efforts by TPW. Con¬ servation Agreements are relatively new and some have not worked. Therefore, they are closely scrutinized by not only the USFWS, but also by others from across the political spectrum. In order to be effective, a Con¬ servation Agreement must provide some immediate reduction in threat to the species and provide long¬ term security against extinction. When well designed, they provide a format and incentives for affected enti¬ ties to work together to resolve issues and conserve natural resources. Devils River Minnow Conservation Agreement A Conservation Agreement among the City of Del Rio, TPW and USFWS was implemented in 1998. Due to the cooperative efforts outlined in the Agree¬ ment, Devils River minnow (Dionda diaboli ) was listed as threatened rather than endangered (USFWS, 1999). The Agreement details a 5 -year plan of research and conservation actions that are designed to resolve the threats to the Devils River minnow and lead to its ulti¬ mate de-listing. Benefits include protection of water quality and quantity in the Devils River and adjacent streams for both fish and people and creation of a green belt/stream corridor along San Felipe Creek in the City of Del Rio that not only provides quality habi¬ tat for fishes, but will also provide a nature-friendly, city park and potential for increased tourism. The Devils River (Figure 2) is one of the most pristine rivers in southwestern North America. Due to its geographic location and historic stability, the Devils River sustains many indigenous organisms. It remains relatively unpolluted and undammed and although spring flows have diminished, they are still substantial. Lim¬ ited access has kept the river from being thoroughly studied by the scientific community; however, collec¬ tions in the past decade by Garrett et al. (1992) and others indicate a diminution in abundance of most flow¬ ing-water species, particularly Devils River minnow. A survey in 1953 showed Devils River minnow was the fifth-most abundant fish species at Baker’s Cross¬ ing and the sixth-most abundant fish in the upper Dev¬ ils River (Hubbs and Brown, 1956). In the mid-1970s, Harrell (1978) found it remained the sixth-most abun¬ dant fish (in 72 collections, Harrell averaged 24-25 D. diaboli per collection). In 1988-1989, collections from 25 locations throughout the historic range in the United States yielded a total of only 7 individuals: Devils River = 2; San Felipe Creek = 3; Sycamore Creek = 2 (Garrett et al., 1992). The numbers had declined such that it was rare where it occurred at all and was probably the least abundant of the approximately 30 species that occur in these streams. In 1979, Devils River min¬ now made up 6-18% of the Dionda population at the Head Spring area of San Felipe Creek. In 1989, none were present there (Garrett et al., 1992). 156 Special Publications, Museum of Texas Tech University Figure 2. Range map for Devils River minnow. Garrett — Innovative Approaches to Recover Endangered Species 157 Members of the genus Dionda are specialized for living in spring-fed, flowing waters and are found primarily in Texas and Mexico (Garrett et al., 1992). Devils River minnow is distinguished from other spe¬ cies of the genus by a variety of characters, including distinctive color pattern, narrow head and number of lateral line scales (Hubbs and Brown, 1956). Biochemi¬ cal work in the past decade has further distinguished unique characteristics of this species (Gold et al., 1992; Mayden et al., 1992). The Devils River minnow is known to occur in the Devils River, San Felipe Creek and Sycamore Creek, Val Verde County, Texas. It historically occurred in Las Moras Creek, Kinney County, Texas, but was elimi¬ nated from that locality sometime before 1980 (Smith and Miller, 1986; Garrett et al., 1992). Extirpation was likely due to periodic failure of the springs from drought and groundwater pumping and from modifi¬ cations to the spring for construction and maintenance of a swimming pool. There are also historic records of occurrence in two small streams in Coahuila, Mexico, the Rio San Carlos and Rio Sabinas (Contreras-B. and Lozano-V., 1994). Their current status there is unknown; no collection attempts have been made since the early 1970s. Although very little is known of the ecology of the Devils River minnow, some threats are apparent. Range reduction has occurred by extirpation of the Las Moras Creek population, minimal flows in Sy¬ camore Creek and inundation of the lower Devils River first by Walk and Devils lakes earlier in the 20th cen¬ tury and ultimately Amistad Reservoir in 1968. Many springs in the area have diminished flows and some have totally stopped (e.g., Beaver Springs, Juno Springs and Dead Man’s Hole), thus reducing the overall length of the Devils River as well as the quantity of water flowing in it (Brune, 1981). Many of the area’s peren¬ nial streams, listed by Gray (1919), no longer flow. In the Devils River, U. S. Geological Service data from the Pafford’s Crossing gauging station reveals a gen¬ eral decrease in daily mean discharge for the period between the study by Harrell (1978) and that of Garrett et al. (1992). The Devils River minnow may be suffering from biological threats as well. Numerous exotics have be¬ come established in the area, including common carp (Cyprinus carpio ), Gulf killifish ( Fundulus grandis ), redbreast sunfish ( Lepomis auritus), smallmouth bass ( Micropterus dolomieu ) and blue tilapia ( Oreochromis aureus). Although fishes throughout the Chihuahuan Desert have been negatively impacted by introduced species (Hubbs, 1990) and such factors as predation and competition may be causing negative impacts on the native fish fauna, specific effects on Devils River minnow are not known. Experiments designed to elu¬ cidate these interactions are ongoing. Much of the water for San Felipe Creek comes from two large springs (San Felipe Springs) within the City of Del Rio. The City also gets its municipal water supply from San Felipe Springs. Conserving the quantity and quality of water from the springs is critical for both the Devils River minnow and the citi¬ zens of Del Rio. The USFWS proposed the Devils River minnow for listing as threatened in 1978 with critical habitat proposed for portions of San Felipe Creek and the Devils River. The USFWS withdrew the proposal in 1980 and retained its designation as a candidate spe¬ cies. The USFWS published a new proposal to list the Devils River minnow as endangered in March, 1998. During the period 1997-1998, TPW was work¬ ing with USFWS, Del Rio and private landowners to develop ways to protect the minnow. Landowners and city officials feared repercussions of listing the fish as endangered and came to understand that a co¬ operative approach to restoring and protecting the eco¬ system would be the best for all concerned. The Dev¬ ils River Minnow Conservation Agreement was signed by the USFWS, TPW and the City of Del Rio in Sep¬ tember 1998. The TPW worked closely with city of¬ ficials and local landowners to develop conservation actions that were beneficial to the species. Those ac¬ tions in the Agreement include determining the current status of the species throughout its range, maintaining captive populations for reintroductions in nature, pro¬ tecting the San Felipe Creek watershed, providing tech¬ nical assistance to landowners on riparian protection and management, revising live bait harvest and selling practices in the Devils River area to prevent the fur¬ ther establishment of exotic, aquatic species and addi¬ tional ecological research, including interactions be¬ tween Devils River minnow and smallmouth bass. 158 Special Publications, Museum of Texas Tech University The Conservation Agreement provides a positive incentive for cooperative actions by all parties and yields scientific access to previously unavailable locations. The primary motivation for the Conservation Agree¬ ment was to remove the threats to the Devils River minnow sufficiently so that protection under Federal law was not necessary. The USFWS carefully con¬ sidered the Agreement and to what extent it had been implemented as of the time the listing decision was due. The USFWS concluded that with an accelerated implementation schedule, a listing determination of threatened rather than endangered would be appropri¬ ate. If successful, the Devils River minnow and other aquatic fauna will be protected and the quantity and quality of streams throughout the range will also be insured. Pecos Pupfish Conservation Agreement The Pecos Pupfish Conservation Agreement was initiated in 1999. The parties involved are TPW, New Mexico Department of Game and Fish, New Mexico Department of Agriculture, New Mexico Division of State Parks, Bureau of Land Management (BLM) and USFWS. The Agreement provides conservation mea¬ sures, new bait fish regulations and creation of addi¬ tional habitat. It is designed to preclude the need to list the Pecos pupfish ( Cyprinodon pecosensis) as a feder¬ ally endangered species by reducing threats to the spe¬ cies and establishing populations in newly created habi¬ tats adjacent to the Pecos River in Texas. This effort incorporates the help of private landowners through the federally funded, state administered, Landowner Incentive Program. With financial assistance and bio¬ logical guidance provided by TPW, landowners have modified private ponds to mimic natural desert wet¬ land habitat while still allowing the landowner’s origi¬ nal intentions for the water body. Due to the ability of the Agreement to remove immediate threats to the fish, a proposal to list it as endangered was withdrawn by USFWS in 2000. The Pecos pupfish is endemic to the Pecos River system (Figure 1) from the vicinity of Roswell, New Mexico to the mouth of Independence Creek, Terrell County, Texas (Echelle and Echelle, 1978). It now only occurs in Salt Creek, a small Pecos River tribu¬ tary in Texas, and at Bitter Lake National Wildlife Ref¬ uge and Bottomless Lakes State Park in New Mexico (Hoagstrom and Brooks, 1999; Echelle et al., this vol¬ ume). It also occurs sporadically in the Pecos River upstream of Artesia, New Mexico (Propst, 1999). Pecos pupfish can occur in a variety of habitats and water qualities. It can flourish in locations that fluctuate in water quantity and chemistry, ranging from highly saline sinkholes to typical desert streams. It has been found in locations with dissolved chlorides ranging from 185 mg/1 to 8,940 mg/1 (Davis, 1981). Age at reproductive maturity, ovary size, egg size and egg number vary among populations, and are appar¬ ently associated with population density (Garrett, 1982). The distribution is mostly limited by interspe¬ cific interactions and thus, it is typically found in habi¬ tats that are low in species diversity (Echelle and Echelle, 1978). During 1954, Pecos pupfish was the most abun¬ dant fish in the Pecos River between New Mexico and Sheffield, Texas (Echelle et al., 1997). Abundance has declined dramatically since the early 1980s, when non-native sheepshead minnow was introduced into the Pecos River in Texas. The original introduction appears to have been a baitfish release in Red Bluff Reservoir (Childs et al., 1996). As a result, Pecos pupfish was eliminated from the lower Pecos River upstream to Loving, New Mexico, and replaced with a hybrid swarm (Echelle and Conner, 1989; Wilde and Echelle, 1992; Echelle et al., 1997). Pecos pupfish is listed as “threatened” by Texas and New Mexico and is considered a “species of concern” by the American Fisheries Society (Williams et al., 1989). Their status is due to habitat loss and especially to hybridization with sheepshead minnow. The Texas portion of the Conservation Agree¬ ment consists of establishing off-channel populations near the Pecos River through a cooperative program with private landowners, developing more restrictive bait-fish regulations, attempting to remove sources of sheepshead minnows from West Texas and monitor¬ ing the status of Pecos pupfish in Texas. Progress has been made in all categories. Garrett — Innovative Approaches to Recover Endangered Species 159 New Mexico’s commitment is similar to Texas, with the addition of ensuring the security of popula¬ tions on state park lands. The USFWS and BLM are also protecting populations on federal lands as well as providing some funding. This Conservation Agreement is the most fea¬ sible approach to addressing the biological problems of this fish. Although restrictive rules inherent in en¬ dangered species status are often necessary to pre¬ vent extinction, in the case of the Pecos pupfish, the opposite is true. Creation of new habitat on private lands would likely not be possible if there was a poten¬ tial for negative impacts on private enterprise. With the Conservation Agreement, landowners are willing to have Pecos pupfish stocked in their ponds and fed¬ eral grants can be used to help develop suitable habitat in the ponds. To date, two shrimp farmers, using naturally saline groundwater in West Texas, have joined the program. At each location, a cienega was con¬ structed (1.7 ha and 7 ha) to provide secure habitat and Pecos pupfish from Salt Creek were stocked in 2000 and 2001. The program allows and encourages individuals to participate in the conservation of rare, natural resources without personal risk or liability. Summary In each of the above projects, many of the criti¬ cal conservation actions would not have been possible without the cooperation of public and private entities. These are examples of a positive approach to problem resolution that provides benefits to each of the coop¬ erators. These “win-win” situations are not always available, but are certainly the most desirable. Acknowledgments These projects were the culmination of long hours and hard work by a large group of people, too numer¬ ous to mention individually. However, the contribu¬ tions of David Riskind (San Solomon Cienega), Gary Graham (Devils River minnow), Jennifer Fowler- Propst (Pecos pupfish), Bob Edwards (all projects), Nathan Allan (all projects) and Andy Sansom (vision¬ ary extraordinaire) were essential to the process. Special thanks to Dr. Clark Hubbs, Professor Emeritus of Zoology, University of Texas at Austin, for his inspirational work on desert fishes, and his in¬ sight into the importance of preserving aquatic com¬ munities and their habitats. Literature Cited Brune, G 1981. Springs of Texas. Vol. 1. Branch-Smith, Inc., Fort Worth, Texas. Childs, M. R., A. A. Echelle, and T. E. Dowling. 1996. Devel¬ opment of the hybrid swarm between Pecos pupfish (Cyprinodontidae: Cyprinodon pecosensis) and sheeps- head minnow ( Cyprinodon variegatus ): A perspective from allozymes and mtDNA. Evolution, 50:2014-2022. Contreras-B. S. and M. L. Lozano-V. 1994. Water, endangered fishes, and development perspectives in arid lands of Mexico. Conservation Biology, 8:379-387. Davis, J. R. 1981. Diet of the Pecos River pupfish, Cyprinodon pecosensis (Cyprinodontidae). The Southwestern Natu¬ ralist, 25:535-540. Echelle, A. F. and A. A. Echelle. 1994. Assessment of genetic introgression between two pupfish species, Cyprinodon elegans and C. variegatus (Cyprinodontidae), after more than 20 years of secondary contact. Copeia, 1 994:590- 597. Echelle, A. A. and P. J. Conner. 1989. Rapid, geographically extensive genetic introgression after secondary contact between two pupfish species ( Cyprinodon , Cyprinodontidae). Evolution, 43:717-727. Echelle, A. A. and A. F. Echelle. 1978. The Pecos River pupfish, Cyprinodon pecosensis n. sp. (Cyprinodontidae), with comments on its evolutionary origin. Copeia, 1978:569- 582. 160 Special Publications, Museum of Texas Tech University Echelle, A. A., C. H. Hoagstrom, A. F. Echelle, and J. E. Brooks. 1997. Expanded occurrence of genetically introgressed pupfish (Cyprinodontidae: Cyprinodon pecosensis x variegatus ) in New Mexico. The Southwestern Natural¬ ist, 42:336-339. Echelle, A. A. and C. Hubbs. 1978. Haven for endangered pupfish. Texas Parks and Wildlife Magazine, June:9-1 1 . Garrett, G P. 1982. Variation in the reproductive traits of the Pecos pupfish, Cyprinodon pecosensis. American Mid¬ land Naturalist, 108:355-363. Garrett, G P., R. J. Edwards, and A. H. Price. 1992. Distribu¬ tion and status of the Devils River minnow, Dionda diaboli. The Southwestern Naturalist, 37:259-267. Garrett, G P. and A. H. Price. 1993. Comanche Springs pupfish (' Cyprinodon elegans ) status survey. Final Report, En¬ dangered Species Act, Section 6, Project No. E-l-4. Texas Parks and Wildlife Department, Austin, Texas. Gold, J. R., Y. Li, M. C. Birkner, and J. K. Jenkin. 1992. Chro¬ mosomal NOR karyotypes and genome sizes in Dionda (Osteichthyes: Cyprinidae) from Texas and New Mexico. The Southwestern Naturalist, 37:217-222. Gray, G A. 1919. Gazetteer of streams of Texas. Washington Government Printing Office. Water-Supply Paper 448. Harrell, H. L. 1978. Response ofthe Devil’s River (Texas) fish community to flooding. Copeia, 1978:60-68. Hendrickson, D. A. and W. L. Minckley. 1984. Cienegas - Vanishing climax communities of the American South¬ west. Desert Plants, 6(3): 13 1-1 75. Hoagstrom, C. W. and J. E. Brooks. 1999. Distribution, status, and abundance of the Pecos pupfish, Cyprinodon pecosensis. Technical Report No. 2, New Mexico De¬ partment of Game and Fish, Santa Fe, 76 pp. Hubbs, C. 1990. Declining fishes of the Chihuahuan Desert. Third Symposium on the Resources of the Chihuahuan Desert Region. Chihuahuan Desert Research Institute, Fort Davis, Texas, Pp. 89-96. Hubbs. C. 2001. Environmental correlates to the abundance of spring-adapted versus stream-adapted fishes. Texas Jour¬ nal of Science, 53:299-326. Hubbs, C. and W. H. Brown. 1956. Dionda diaboli (Cyprin¬ idae), anew minnow from Texas. The Southwestern Natu¬ ralist, 1:69-77. Hubbs, C., R. J. Edwards, and G P. Garrett. 1 991 . An annotated checklist of the freshwater fishes of Texas, with keys to identification of species. Texas Journal of Science, Supple¬ mental , Vol. 43(4), 56 pp. Mayden, R. L., R. H. Matson, and D. M. Hillis. 1992. Specia- tion in the North American genus Dionda (Teleostei, Cypriniformes). Pp. 710-746 in Systematics, Historical Ecology and North American Freshwater Fishes (R. L. Mayden, ed.). Stanford University Press. Propst, D. L. 1999. Threatened and endangered fishes of New Mexico. Technical Report No. 1, New Mexico Depart¬ ment of Game and Fish, Santa Fe, New Mexico, 84 pp. Smith, M. L. and R. R. Miller. 1986. The evolution of the Rio Grande basin as inferred from its fish fauna. Pp. 457-485 in The Zoogeography of North American Freshwater Fishes (C. H. Hocutt and E. O. Wiley, eds). John Wiley & Sons, New York. Stevenson, M. M. and T. M. Buchanan. 1973. An analysis of hybridization between the cyprinodont fishes Cyprinodon variegatus and C. elegans. Copeia, 1973: 682-692. USFWS(U.S. Fish and Wildlife Service). 1999. Endangered and threatened wildlife and plants; final rule to list the Devils River minnow as threatened. Federal Register, 64:56596- 56608. Wilde, G R. and A. A. Echelle. 1992. Genetic status of Pecos pupfish populations after establishment of a hybrid swarm involving an introduced congener. Transactions of the American Fisheries Society, 1 2 1 :277-286. Williams, J. E., J. E. Johnson, D. A. Hendrickson, S. Contreras- Balderas, J. D. Williams, M. Navarro-Mendoza, D. E. McAllister, and J. E. Deacon. 1989. Fishes of North America endangered, threatened, or of special concern: 1989. Fisheries (Bethesda), 14(6):2-20. Winemiller, K. O. and A. A. Anderson. 1997. Response of endangered desert fish populations to a constructed ref¬ uge. Restoration Ecology, 5:204-213. Young, D. A., K. J. Fritz, G P. Garrett, and C. Hubbs. 1994. Status review of construction, native species introduc¬ tions, and operation of an endangered species refugium channel, Phantom Lake Spring, Texas. Proceedings ofthe Desert Fishes Council, 25:22-25. Address of author: Gary P. Garrett HOH Fisheries Science Center Texas Parks & Wildlife Department HC 7, Box 62 Ingram, TX 78025 e-mail: gary.garrett@tpwd. state. tx.us