L964. Fishes Inhabiting the , :ies on Texas and , Between logy EI Paso and the Pecos Confluencel 3. ~aniza- by Clark Hubbs 2, Robert Rush Miller3, > • 2 2 Robert J. Edwards, Kenneth W. Thompson, 224 70. Edie Marsh, Gary P. Garrett, Gary L. Powell, ack D. J. MorrisS, and Robert W. Zerr5 :l New 54. ation parian t Abstract --The fishes of the middle part of the Rio Grande can be divided into three faunal assemblages: The saline Rio Grande fauna (made up of widely distributed and salt tolerant species) upstream from the Conchas confluence; the Rio Conchos-Rio Grande fauna (mostly 974. south Texas and Mexican species) in the Rio Grande between the Conchos dan and Pecos; the tributary creek fauna (Chihuahuan species plus some values. derivatives) that depend on tributary creeks for all or part of their RM-1l7, life history stages. Endangered species are found in the last assem­ blage but two presumed endangered species (Notropis ~imus and Scaphi­ rhynchus platyrynchus) seem to have been eliminated already. ntory Mexico. uq. ,

INTRODUCTION A .sco The fishes of the Rio Grande (Belcher, a large number of collections made from the Rio Catron fig. 3) have been intermittently studied Grande in Big Bend National Park between 1954 on for the past 130 years. Reasonably extensive and 1976. .hed ms. reports exist for Colorado (Beckman, 1952), New Mexico (Koster, 1957), and the Rio Grande down­ The Rio Grande "enters" Texas as a small stream from its confluence with the Pecos River stream most or all of which is diverted to irri­ . Vege­ (Trevino-Robinson, 1959). No comparable summar­ gate fields south and east of El Paso. Commonly, :tates. ization exists for the intervening segment, al­ the stream is dry over much of the distance be­ II Soc., Miller (1977) treated the Mexican part tween El Paso and Ft. Hancock. Southeast of middle Rio Grande basin. Proposals to this town the valley narrows and the ground water about 300 kilometers of the river surfaces to form a salty stream. The river re­ .n the to designate another 200 kilometers as a mains small for the next 300 km until" it "receives" wild river underscored the absence of a summar­ the Rio Conchos. Small volumes of water are ization of the fish fauna. The bulk of this added by small salt laden springs (such as Indian paper is a report on fishes collected on two Hot Springs) and fresh tributary creeks (such as recent visits to the Rio Grande in the two pro­ Capote Creek). These increases are commonly ex­ ject areas. We also include a summarization of ceeded by losses from evaporation or irrigation diversions. Drastic increases in flow periodi­ cally follow intense desert rains. These torrents 1 :he Contributed paper, Symposium on the soon subside and the Rio Grande again becomes 1m. Importance, Preservation and Management of the a small, sometimes intermittent stream. This Riparian Habitat, July 9, 1977, Tucson, Arizona. pattern is of long duration as Emory (1859) re­ 2Department of Zoology, The University of ported periodic dry stream beds and occasional Texas, Austin, Texas 78712 severe flooding and Thomas (1963) reported high salinities in the Rio Grande in 1936. This 3Museum of.Zoology, The University of Mich­ reach of the river has been extensively impacted igan, Ann Arbor, Michigan 48109. by human activities. Much of the flow (and most 4Texas Water Development Board, Austin, of the low saline water) is diverted at or , 78701 north of El Paso. The northwestern 150 kilometers 5 Texas Parks and Wildlife Department, have be.en leveed and channelized. A 16 kilometer in, Texas 78701 segment around the Conchos confluence has also

91 been leveed and channelized. The intervening 300 kilometers has been proposed for channel "rectification" and is extensively leveed already.

Much of the flow of the Rio Grande east of the Conchos is dependent upon that "tributary". Historically, the contribution of the Conchos has been considerably greater than that of the Rio Grande above the confluence and that dif­ ference has been magnified by the Rio Grande diversions upstream. Present flow rates depend Hot Spri ngs chiefly on releases from Luis L. Leon Reservoir; at the time of our visit on 18 March, 1977, the Conchos flow was nearly 2 orders of magnitude greater than that of the Rio Grande. Strangely, Form the man-made conjunction has the Conchos enter­ ing at a right angle; in effect a forced right angle turn of the huge stream where it enters the small stream. Between the end of its recti­ fied channel below Presidio and the upper part of (just upstream from the Pecos confluence), the Rio Grande has not been Marov substantially impacted by human activities. Can The major items are the stream measurement weirs just below Alamito Creek and just above Amistad Reservoir, river fords at Stillwater Crossing and Boquillas. and a bridge near Stillwater

Crossing. Other impacts are indirect such as Slillw, minor irrigation diversions, overgrazing, exotic Crossir plants and fishes, pesticides washed from nearby fields, leaching from mine tailings, etc. Huch of this distance is little_disturbed and one can see the diverse geology and magnificent land :'Capote Cr. formations. 13 Fj COLLECTION SITES CANOE LAR I A Most of the newly reported locality records Their widE are based on two collecting trips, 14-18 March would be E and 3-7 April, 1977. Collections were concen­ ment. The trated in the channelization and wild river is an exot segments, respectively. Previously, only one tributed I sample had been obtained from each of those species <1 reaches. The 1977 and previous (1954) locations were coIl, are plotted on figures 1 and 2. The bulk of the former W.lE 1954 (and subsequent) Rio Grande collections likely the were from the Big Bend National Park and have ever those been reported in Hubbs (1958}, Hubbs and Wauer doubtedly, (1973) and Hubbs and Williams (in press}. Rio) is ar would be e because of RESULTS and consec The 15 collections from the Rio Grande fish. In west of the influence of the irrigation water Cr. Ictalurus from the Rio Conchos that enters the Rio Grande in a small reach. Tb between Stations 15 and 16 contain 11 fish 10 k .. species (Table 1). The redundant nature of not previc these samples is reflected by the presence As this fi of 7 fishes (Dorosoma cepedianum, Cyprinus it is like descendant carpio, Notropis lutrensis, Carpiodes carpio, Figure l.--Location of collection stations in Ictalurus punctatus, Gambusia affinis, and the Rio Grande. from and adjacent to the pro­ The Lepomis cyanellus) in 9 or more collections. posed channelization. Presidio

92 \ Terrell Co. \ Vol Verde Co.

Sanderson LANGTRY Canyon #

Son

RIO GRANDE Panther r Canyon" Brewster Co.

24

Canyon" de Tule \ Canyon de Son Rocendo

IOkm

r. Figure 2.--Location of collection stations in the Lower Canyons of the Rio Grande.

Their widespread abundance suggests that they 11 were not taken upstream but 2 from there were would be expected anywhere in this stream seg­ absent. We expect that increased effort would The second listed species (Cyprinus carpio) have produced an Ictalurus punctatus, but that is an exotic but the others are all widely dis­ Pimephales vigilax is not present. Nine of the tributed native fishes. Two of the other four additional 11 species, Astyanax mexicanus species (Lepomis megalotis and Morone chrysops} (as A. fasciatus), Hybopsis aestivalis, Notropis were collected at widely separated sites. The , Notropis braytoni, Notropis jemezanus, former w~s found at very brushy sites. It is Pimephales promelas, Campostoma ornatum, Pylo­ likely that this species can be obtained wher­ dictis olivaris, and Lepomis macrochirus, were ever those conditions prevail. The latter (un­ reported from the Big Bend region by Hubbs doubtedly, derived from fishes stocked near Del (1958). 'One exception, Cyprinodon eximius Rio) is an open-water top carnivore. This fish has subsequently been reported from Terlingua would be expected to be sparsely distributed creek by Miller (1977). The other, Menidia because of dependence upon a complex food chain beryllina, is undoubtedly derived from descendants and consequently high primary productivity per of bait-released stocks now abundant in Amistad fish. In a similar way, a large fish like Reservo~r. We expect that Menidia (a euryhaline Gr. Ictalurus furcatus would be expected to be rare species) will soon spread and become abundant in a small stream like the Rio Grande in this in the saline Rio Grande waters upstream from reach. The last species Pimephales vigilax, bas the Conchos confluence. not previously been taken east of Val Verde Co. As this fish is commonly used as a bait minnow, The distinct difference between the Rio it is likely that the samples obtained are Grande fishes on either side of the Conchos ns in descendants from escaped bait. confluence is reflected by similar differences e pro- between the fishes inhabiting the tributary The six collections from the vicinity of creeks, Capote and Alamito (stations 13 and Presidio (16B on Table 1) contain 20 species; 19, respectively).

93 C::hemical a Table- l.--Numhers of fishes collected from the Rio Crande from and adjacent to the proposed dmnnf!liz.ation between creeks are Presidio and El Pa;c;o is slightl near the m Stations Spec tes A 3 ~ 7 8 12 10 13 11 14 15 16 18 17 19 20 B difference tions into Doroso~~rY11in!. 2 4 4 10 1 1 1 The f Horon'!. .'=.!'.!.l.!~ 1 1 38 105 168 11 ~cm~~ ~I!llus. 4 8 11 23 29 fishes of 2 1 .,bepomis !,egalo~ The 1954 s ~mis--_. macrocnirus__. tive of th H' 1.7 1.3 1.5 0.4 1.5 0.9 1.1 1.8 1.4 1..4 1.7 0.3 2.1 l.8 2.0 2.2 1.6 0.5 1.6 2.3 % Introduced 65 18 12 1 9 6 74 19 1 18 26 5 44 53 8 12 8 1 2 The 1 79 316 248 151 161 445 500 78 381 527 133 Fishl s~tne hour 46 101 119 92 193 156 136 166 351 Grande Can Thirteen ( Astyanax ~ pha1es E.E!? jemezanus, toni, lcta Gaiiibus~ were found and report National P Cyc1eptus TabJe 2--List of fishes collected from the Low~r Canyons of the Rio Grande>. tae from t 'b"e""a seaso Stations reported f ~!~---- 21 22 ...... ::2-"-J_.:.2C!4_.:.2~5 26 27 28 29 3",0,---",3.=.1_~32 33 are also a 1.~sosteus ~.!:!!!. 1 (C). All D0ro50m~ cepedlantut! 15 3 21 11 5 1 stream wer Cycleptus elongatus 108 15 6 7 I, 19 1 Carpiodes carpio 5 1 2 8 31 the bulk a lctiobus bubalus 1 1 is unlike1 Astyanax mexicanus 13 1 collected .cyprin!!!', .care..!.!!. 3 1 1 Rhynichthys cataractae 188 182 209 92 87 77 47 184 141 191 168 850 167 water (and Hybopsis aes~~ 4 13 36 8 6 7 4 6 3 3 11 9 9 mouth of 1 Pimephillel!. promelas 1 1 many young Notropis ..£hihuahu!!. 1 Notropis j emezanus 3 574 20 19 3 3 48 5 16 6 12 11 lections f Notropis .!!!..trenlEtis 6 1 19 1 5 112 4 6 2 12 29 37 1 are likely Notropls ~tonl 2 11 3 15 9 4 19 6 6 10 25 breeding s Ictalurus ~cta!.!:!!. 1 7 2 2 3 2 1 2 Ictalurus furc3tus 13 10 6 4 3 8 4 69 just east Pylodlctu&i olivaris 4 3 6 c1uded one Fundulus kansae 2 Gambuaia Si'"ITilIs 3 4 Rhinichthy Henidl! bcryTfina 2 served. B Micropterus ~ide8_ 1 1 supported ~Dl1s cyanellus Lp.pollis mac.roc.hirus the Rio GI in Big Ben been taken The absenc bubalus an lections n ling artif very diffj

94 ~hemical and physical conditions in the two commonly occupied by these (especially Ictiobus creeks are reasonably similar - Capote and }licropterus) an~ our downstream samples were is slightly smaller and has been reported dry sufficiently infrequent that chance occurrence in -----. near the mouth. We attribute the fish faunal the Presidio samples is likely. Two species 20 B differences to the impact of seasonal migra­ (Menidia bery11ina and Lepomis cyane11us) were tions into the Rio Grande as reported for Ter­ collected near Presidio but not reported from Big lingua Creek by Hubbs and Wauer (1973). Those Bend National Park by Hubbs (1958). It is un­ 1 salty Rio Grande waters at the mouth of Capote likely that the former existed in the region 13 Creek may exclude the typical Rio Grande tribu­ before 1960 as Tilton and White (1964) showed tary creek fauna from any upstream tributary. 1 44 that this fish was then just being distributed 63 36 Regardless of the cause, these fishes were not across Texas. Hubbs and Eche11e (1972), docu­ 22 found in Capote Creek. We looked carefully for mented a similar and recent spread of this fishes in the waters of Indian Hot Springs to fish in the Pecos Basin. Lepomis cyane1lus is 16 determine if an endemic fauna were there. These now known from Big Bend National Park (Hubbs warm, salt-laden springs were fish1ess. We did and Williams, in press), supporting Hubbs' note Gambusia affinis was concentrated in the 2 5 (1958) prediction that it existed within the 1 1 warm outflow waters emptying into the colder park. Similar to Menidia audens, Fundulus 1 Rio Grande waters during our March visit. kansae has recently been introduced into the 28 14 region. Its introduction and subsequent spread The faunistic difference between the was reported by Hubbs and Wauer (1973). The fishes of the two segments is of long duration. fish from the Rio Grande at the mouth of Mara­ The 1954 samples (A and B) are well representa­ villas Cr. surely reflects an additional tive of the faunal units found in 1977 samples. spread. We herein also report the presence of 6 2.3 Fundulus kansae in McKinney Spring in Big Bend 2 2 The 13 collections from the Lower Rio ~7 133 National Park. It is likely that the speci­ Grande Canyons contained 23 species (Table 2). mens of Micropterus sa1moides reflect a modest Thirteen (Dorosoma cepedianum, Carpiodes carpio, population of indigenous fishes that can serve Astyanax mexicanus, Hybopsis aestivalis, Pime­ as a recreational resource. We expect that pha1es promelas, Notropis chihuahua, Notropis largemouth bass occur throughout the Rio Grande jemezanus, Notropis 1utrensis, Notropis bray­ east of the Conchos confluence (and also in toni, Icta1urus furcatus, Py10dicitis olivaris, much of the Conchos system). Gambusia affinis, and Lepomis macrochirus) were found in the collections near Presidio The 23 species extensively overlap those and reported from the Rio Grande in Big Bend reported from the Big Bend by Hubbs (1958) who National Park (Hubbs, 1958). The absence of recorded 7 additional fishes. Four of them, Cyc1eptus e10ngatus and Rhinichthys catarac­ [Dionda episcopa, Gambusia gaigei, Lepomis tae from the upstream stations is likely to (~obryttus) gulosus, and Lepomis micro­ ~a seasonal artifact because both have been 10phusl were recorded as inhabiting small clear reported from the Rio Conchos. Both species tributaries and would .be rare or absent in the are also absent in the August 1954 collection river proper. Moxostoma congestum was subse­ (C). All of the Cyc1eptus collected down­ quently reported from Tornillo Cr. by Hubbs and stream were young of the year. Similarly, Wauer (1973). Three fishes, Anguilla rostrata, the bulk of the Rhinichthys were young. It Hybognathus p1acitus, and Ap10dinotus grunniens) is unlikely that adult Cyc1eptus would be were reported from Big Bend by Hubbs (1958), collected with the seines used in such high but not obtained in the 1977 samples. The first water (and none were). Samples taken near the is catadromous and upstream migrants would be mouth of Tornillo Creek in April commonly have unlikely to pass Falcon Dam (much less Amistad); many young Cyc1eptus but no adults are in col­ samples have not been obtained since Falcon was lections from that spot. Similarly, Rhinichthys filled. The other two would be expected to occur are likely to be most abundant just after the in the area. Aplodinotus could easily have been breeding season. Our station 20 (Rio Grande overlooked but the absence of Hybognathus is just east of the mouth of A1amito Creek) in­ inexplicable. cluded one fish tentatively identified as a Rhinichthys that escaped prior to being pre­ served. Rhinichthys abundance in the area is DISCUSSION supported by its presence in a collection from the Rio Grande just upstream from Mariscal Canyon The fishes inhabiting the Rio Grande in in Big Bend National Park. Specimens have also west Texas can be placed in three faunal assem­ been taken from the Conchos system in Chihuahua. blages: Saline Rio Grande fauna, Rio Conchos­ The absence of Lepisosteus ~, Ictiobus Rio Grande fauna, Tributary Creek fauna. lUba1us and Micropterus sa1moides in the col­ lections near Presidio is likely to be a samp­ The Saline Rio Grande Faunal assemblage ling artifact. The high-water flows made it is dominated by four wide spread species, very difficult to sample deep-water environments Dorosoma cepedianum, Cyprinus carpio, Notropis

95 duction: lutrensis, and Lepomis cyanellus. The limited tributaries in Coahuila) that support the former diversity (Shannon H' values are generally occurrence of shovelnose sturgeon in the river. well below 2) seems to be due to harsh condi­ Notropis simus has been recorded from the Rio tions - salinity and periodic interrupted stream Grande in New Mexico and downstream from Del flows. The latter may be most critical as the Rio but the collections preceded or were at a We 1 fish present are ones expected in pools in similar time interval as the first collections ment by west Texas streams. Our repeated efforts in from our study area. We doubt that Notropis in plam riffles were generally unproductive. This assem­ simus now lives in the Lower Canyons of the Rio Garrett, blage has been impacted by human activities. Grande and suggest that work to ascertain if it Dwane K: Certainly the three exotics (Cyprinus, Morone, still exists concentrate on the lower Rio fishes. Pimephales) must have some impact. It is We have no suggestions as to the conditions Pollard likely that Cyprinus has depressed Carpiodes that may have led to the extinction or subs vided ht abundance but the impact of Morone and Pime­ decline of these two fishes that once were part' pass act phales is difficult to assess, and the absence of this faunal assemblage. Both species are William of prior studies makes any conclusions con­ monly found on listings of endangered species Swanson jectural. N. simus may be extinct in U.S. waters. Its ning ph, sence in Trevino-Robinson's collections is par­ Wildlife The Rio Conchos - Rio Grande faunal assem­ ticularly alarming as m6st Texas records are gave pel plage is made up of those species living in the from that stream segment. The New Mexico re­ Rio Grar Rio Grande and not dependent upon tributary creeks cords are from the Rio Grande in areas that now for a part of their life history. The abundance have reduced flow or are dry. of these fishes is not correlated with the pre­ sence of tributary flows. Typical fishes of this The tributary creek faunal assemblage i~ assemblage are Notropis jemezanus, !. lutrensis, made up of a group of fishes that spends all or Beckman, !. braytoni, Rhinichthys cataractae, Hybopsis a substantial fraction of their time in the Cole aestivalis, Ictalurus punctatus, Ictalurus tributaries. Three species (Notropis lutrensis, Belcher, furcatus, Pylodictis olivaris, Dorosoma cepe­ Pimephales promelas, Notropis braytoni) may of t dianum, Cycleptus elongatus, and Carpiodes occur in the creeks or Rio Grande. Except for 29. carpio. Seven other fishes (Lepisosteus ~, the first, they are seldom found far from the Cope, E. Ictiobus bubalus, Pimephales promelas, Men­ creek mouth. Three (Moxostoma congestum upon idia beryllina, Micropterus salmoides, Aplodin­ Carpiodes carpio, Cycleptus elongatus) are tion otus grunniens, and Hybognathus nuchalis) are creek inhabitants only as young and the adults New reasonably abundant in the Rio Conchos-Rio Grande may be found with equal abundance elsewhere 1872 faunal assemblage. in the Rio Grande. Eleven species (Cyprinodon Expl eximius, Campostoma ornatum, Notropis chihuahua, Surv Only one (Menidia beryllina) of those 18 Fundulus kansae, Astyanax mexicanus, Gambusia Emory, W species is introduced. Its impact is not yet affinis and the sunfishes, Lepomis gulosus, Stat fully assessed as its entry into the region is cyanellus, microlophus, macrochirus, and mega­ so recent that its abundance may be in a growth lotis) are most commonly collected in creeks phase. It is not likely that this quiet water but have been found in the Rio Grande. The euryhaline form will become excessively abundant first six are listed by relative frequency of in the fresh-flowing waters of the Rio Grande. creek vs. river abundance. Hubbs and Wauer Rhinichthys cataractae is not only a prominent (1973) had reported that this assemblage moved member of this faunal assemblage, it also seems out of the creeks seasonally but our 1977 sam­ to be absent or very scarce in adjacent areas. ples of the first two are the first demonstrat This population is isolated from other stocks by of fish that must have moved into the river. the saline and frequently dry Rio Grande up­ Samples of the five sunfishes are sufficiently stream from Presidio. It is likely that it re­ infrequent that definite patterns are difficult presents a race adapted to deep canyons with re­ to ascertain. Two species (Gambusia gaigei, latively warm water. Essentially, a collection Dionda episcopa) are limited to the tributary from Texas with numerous Rhinichthys and/or waters. The former is on all lists of endan­ Cycleptus is likely to be from the Rio Grande gered fishes; its status has been discussed between Presidio and Amistad Reservoir. The recently by Hubbs and Williams (in press). Rio Conchos - Rio Grande faunal assemblage will The fishes in the tributary creek assemblage often be supplemented by representatives from often present special problems. Three of them the tributary creek faunal assemblage. (Cyprinodon eximius, Campostoma ornatum, Notropis chihuahua) are commonly found on Two fishes (Scaphirhynchus platyrynchus endangered species listings as their U.S. dis­ and Notropis simus) may once have inhabited tribution is restricted to the creek mouths. the Rio Conchos - Rio Grande faunal assemblage. These areas should be watched with care to Scaphirhynchus was reported from the Rio Grande reduce the possibility of extermination of near Albuquerque by Cope and Yarrow (1875). this fragile assemblage. The spread of the We have obtained hearsay reports of a sturgeon introduced Fundulus kansae is of primary con­ from near Dryden Crossing (and also from Mexican cern (Hubbs and Wauer, 1973). Future intro­

96 ductions of bait minnows should be avoided. Hubbs, Clark. 1958. List of fishes known or he former expected to belong to the fauna of Big Bend e river. National Park. Mimeograph report to Big he Rio ACKNOWLEDGMENTS Bend National History Association. m Del ______, and Antony A. Echelle. 1973. En­ e at a We have benefitted from the encourage­ dangered non-game fishes in the Upper Rio ections ment by and discussion with James E. Johnson Grande Basin. In: Endangered Vertebrates ropis in planning our program. David S. Marsh, Linda in the Southwes~ William C. Huey (ed.) . the Rio Garrett. Danny Swepston, Floyd Potter, and New Mexico Game and Fish:147-l67 • ,in if it DWane Kippes participated in collecting the ______, and Roland Wauer. 1973. Seasonal ~io Conchos •.... fishes. Ross Foster, John Vandertulip, and changes in the fish fauna of Tornillo Creek, :ions pollard Rogers were hospitable hosts, pro­ Brewster County, Texas. The Southwestern vided helpful information, and permitted tres­ Nat. 4:375-370. ,ere party': pass access to or from collection sites. ______, and John G. Williams. in press. A William Provine, Floyd Potter, and W. S. review of circumstances affecting the abun­ Swanson helpfully participated in the plan­ dance of Gambusia gaigei, an endangered ning phases of our studies. Texas Parks and fish endemic to Big Bend National Park. Wildlife and Departamento de Pesca de Mexico In: Proceedings of the First Conference gave permission to collect fishes from the on Scientific Research in the National Rio Grande. Parks. Robert Linn and George Sprugle, (eds.) U.S. Government Printing Office. Koster, W.J. 1957. Guide to the Fishes of LITERATURE CITED New Mexico. Universi~y of New Mexico Press. Miller, R.R. 1977. Composition and deriva­ Beckman, W. C. 1952. Guide to the fishes of tion of the native fish fauna of the Chi­ Colorado. Univ. Colo. Mus. Leaflet 11. huahuan desert region. In: Transactions, Belcher, R. C. 1975. The geomorphic evolution Symposium on the Biological Resources of of the Rio Grande. Baylor Geol. Stud., Bull. the Chihuahuan Desert Region, U.S. and 29. Mexico. R.H. Wauer and D.H. Riskind (eds.). Cope, E.D., and R.C. Yarrow. 1875. Report National Park Service. 11m upon the collections of fishes made in por­ Thomas, H.E. and others. 1963. Effects of are tions of Nevada, Utah, California, Colorado, drought in the Rio Grande Basin. U.S. Geol. e adults New Mexico and Arizona, during the years 1871, Surv. Prof. Pap. 372-D-Dl-59. "here 1872, 1873, and 1874. Rept. Geog. and Geol. Tilton, J.E., and R.L. White. 1964. Meni­ prinodon Expl. and Surv. W 100th Merid. (Wheeler dia from several central Texas impoundments. chihuahua, Survey) 5:637-703. Texas J. Sci. 16:120. ambusia Emory, William C. 1857. Report of the United Trevino-Robinson, D. 1959. The ichthyofauna osus, States and Mexican Boundary Survey. Vol. I. of the lower Rio Grande, Texas and Mexico. nd mega­ Copeia 1959:253-256. creeks The ency of Wauer ge moved 977 sam­

aigei, ibutary endan­ ussed ss). mblage of them

~, on .S. ouths. e to n of f the ry con­ intro­

97