Feeding and Fate of Wild Larval Razorback Sucker

Environmental Biology of Fishes Vol. 21, No. 1, pp. 59--67, 1988 O Dr W. Junk Publishers, Dordrecht.

Feeding and fate of wild larval razorback sucker

Paul C. Marsh & Daniel R. Langhorst Center for Environmental Studies and Department of Zoology, Arizona State University, Tempe, AZ 85287, U.S.A.

Received 29.10.1986 Accepted 28,4.1987

Key words: Xyrauchen texanus, Imperiled fishes, Colorado River, Zooplankton, Food selection, Predation, Conservation

Synopsis

The razorback sucker (Xyrauchen texanus) is disappearing throughout its native range in the Colorado River basin of western North America. The largest remaining wild population in Lake Mohave, Arizona-Nevada, has shown no recruitment since the 1950s. Although annual spawning is successful and larvae are seasonally abundant, no juveniles have been collected in recent decades. To evaluate the potential role of food availability in determining fate of larvae, fish and zooplankton samples were taken in 1985 from the reservoir and an adjacent, isolated backwater in which larvae were naturally produced. Food availability and primary dietary constituents were similar in both habitats. Reservoir larvae selected Bosmina spp. (Cladocera) and apparently avoided Copepoda, while larvae from the backwater selected Bosmina, but avoided Rotifera. Larvae from both places showed evidence of selection for certain sizes of zooplankters, but preferred sizes differed between habitats. These differences were neither attributable to larval size nor zooplankton community structure. Nutritional factors such as type, number, or size of available foods do not explain disappearance of larval razorback suckers from Lake Mohave, since larvae survive to far greater ages and size in the backwater. Predation by introduced fishes appears a significant cause of larval mortality.

Introduction approximately 12mm total length (TL) by late spring. Explanations for this phenomenon include The razorback sucker, Xyrauchen texanus (Ab- mortality, in-reservoir migration to unknown bott), is a large catostomid fish endemic to the places, and transport out of the system. We exam- Colorado River system of western North America. ined the role of potential nutrient constraints, Once widely distributed and abundant (Holden which could result in high mortality, by comparing 1980, McAda & Wydoski 1980, Minckley 1983), its food availability and diets of razorback sucker lar- numbers and range are now drastically diminished. vae in Lake Mohave, from which they disappear, The largest remaining population in Lake Mohave, and an adjacent, isolated backwater in which they a mainstem Colorado River reservoir in Arizona- survived far longer and to a larger size. Nevada, is comprised of old individuals, with no evidence of successful recruitment since the 1950s (Minckley 1983, McCarthy 1986). Although annual Methods winter-spring spawning produces an abundance of larvae, these disappear for unknown reasons at Lake Mohave (Fig. 1) is immediately downstream 60

HOOVER DAM

~O-Backwater ~Arizona Bay ~._ Hammerhead Cove

5 10 I i km

DAVIS DAM

Fig. 1. Lake Mohave, Arizona-Nevada, showing study areas, and area map (inset). from Hoover (Boulder) Dam and impounded by suckers and their larvae, which are readily ob- Davis Dam. Physical, chemical, and biological fea- served and sampled during calm weather. Larvae tures of the reservoir were described by Allan & are abundant and easily collected from Ham- Roden (1978), Priscu et al. (1980, 1982), and merhead Cove except during turbulence of major Minckley (1983). Our study area was along 5 km of storms. Both areas were regularly visited and shoreline between Arizona Bay and Hammerhead sampled from January to May 1985. Cove (Fig. 1). Arizona Bay supports a large seaso- Larvae are here defined as fish from hatching nal aggregation of spawning adult razorback (6.5-8.6 mm) through development of full compli- 61 ments of fin rays in all fins (35 mm). Specimens the reservoir and backwater was generally at night, examined all had attained active swimming capa- when the small fishes were attracted to a spotlight, bility (middle prolarvae), most were early postlar- captured by dipnet, and preserved in 5% buffered vae, and the largest were middle postlarvae (termi- formalin. Zooplankton in Lake Mohave was col- nology for razorback suckers from Minckley & lected in duplicate Clarke-Bumpus samples (80/z Gustafson 1982). Jaws were formed and mouth and mesh) towed for 5min at 1.3 km hr -~ and 0-3m anus were open, all were thus presumed capable of depth throughout habitats occupied by larval feeding. However, yolk had not been completely suckers. Backwater samples were duplicate, com- assimilated by the smallest individuals, which may posite 38-L grabs concentrated through 80/z mesh. not yet have begun to feed. Zooplankters were preserved in 5% formalin and Adjacent to Arizona Bay is a backwater (Fig. 1) later identified, enumerated (Sedgwick-Rafter separated from the reservoir by a wave-formed spit chamber), and measured (greatest body length and of coarse gravel. This protected habitat was se- width) at 35--450x with a light microscope. Linear lected for rearing razorback sucker in isolation to dimensions were converted to individual volume preclude potential destruction of eggs or larvae by on the basis of similarity to regular geometric sol- introduced fishes, indications of which had been ids. Foods of larval suckers were treated in a simi- observed in Lake Mohave and elsewhere (see be- lar manner after removal from digestive tracts. low). Selection of foods by taxon and size class (volume) The backwater had maximum depths of 1.7 to was evaluated by a linear index: L = r~- Pi, where 3.7 m and surface areas of 0.85 to 2.10 ha at respec- r i and Pi are relative abundances expressed as pro- tive reservoir elevations of 193.5 to 195.5 m, and portion of prey item i in the gut and habitat, respec- was fed by reservoir water that freely exchanged tively (Strauss 1979, 1982). The index ranges from through the porous berm. Water temperature par- -100 to +100. Positive values indicate preference, alleled that of the reservoir, but averaged slightly negative values avoidance or inaccessability, and cooler (-1.1 ° C) in winter and warmer in summer values near zero indicate random feeding. We con- (+1.8°C); thermal differences were most pro- sidered/L/<10 not different from L = 0. nounced on early winter mornings and late summer afternoons. Substrates were coarse gravel around margins and organic silt in deeper areas. Shallow Results regions were occupied by dense stands of water- milfoil (Myriophyllum spicatum) interspersed with Lake Mohave curly leaf pondweed (Potamogeton crispus). Fishes which initially inhabited the backwater were Razorback sucker larvae were first captured from threadfin shad (Dorosoma petenense), common Lake Mohave in Hammerhead Cove on 9 Febru- carp (Cyprinus carpio), channel catfish (Ictalurus ary, approximately 3 weeks after initial observa- punctatus), mosquitofish (Gambusia affinis), tions of spawning at that location. Larvae became largemouth bass (Micropterus salmoides), green abundant within the next few days and remained sunfish (Lepomis cyanellus), and bluegill (L. mac- common into April. Larvae averaged 10.6 + 0.3 rochirus). All are non-native. These were removed (SE) mm TL (n = 410) throughout the period, in- by ichthyocide in autumn 1984, and the backwater dicating continuous spawning and larval produc- stocked in January-February 1985 with 30 female tion. There was no evidence of larval growth since and 150 male adult razorback suckers collected neither mean nor maximum TL increased; the from Arizona Bay. These spawned successfully largest individual captured from the reservoir was and produced larvae that were first captured on 7 12.2 mm. Comparisons with known-age fish reared March 1985 and subsequently monitored through experimentally at similar temperatures (Marsh their disappearance in early April (see below). 1985, Papoulias unpubl, data) indicated maximum Larval sampling in shallow littoral areas of both age of the largest specimens was at most a few weeks. 62

Zooplankton in Lake Mohave was predomi- animals were most abundant in larval diets (Tables nated by Cladocera (Bosmina and Daphnia to- 1, 2), and relative abundances indicated positive gether comprised nearly two-thirds of total num- selection for Bosmina and negative selection (or bers) and Copepoda (about one-third of total, avoidance) of copepods (Table 1). Larval suckers including nauplii, copepodites, and adults). Ro- showed positive selection for zooplankters 400 to tifera were of minor significance (Table 1). Total 999 × 10-4mm 3 and avoidance of those 100 to density averaged 1,554 organisms m -3. These same 399 × 10 -4 mm 3 (Fig. 2A), a pattern nearly identical

Table 1. Relative abundance of predominant zooplankters in water samples (P3 and larval razorback sucker stomachs (r0 and linear index of selectivity (L) in Lake Mohave and the Arizona Bay backwater, Arizona-Nevada, 1985.

Prey items Lake Mohave Arizona Bay Backwater

Pi ri L p~ r i L

Rotifera 5.5 1.2 -4.3 57.9 10.7 -47.2 Cladocera Bosmina 51.0 90.3 39.3 13.4 41.9 28.5 Daphnia 12.8 2.4 - 10.4 12.8 17.7 4.9 (subtotal) (63.8) (92.7) (28.9) (26.2) (59.6) (33.4) Copepoda 30.7 1.2 -29.5 16.0 16.2 0.2 Other <1.0 4.8 4.8 <1.0 13.5 13.5

Total n 345 165 382 1,020 Mean density (no. m 3) 1,554 377

Table 2. Food habits (percent occurrence and mean + SE number of items per stomach that contained food) of larval razorback sucker in Lake Mohave and the Arizona Bay backwater, Arizona-Nevada, 1985.

Prey items Lake Mohave Arizona Bay Backwater

Percent Mean Percent Mean

Rotifera 1.6 1.0 + 0 53 4.2 _+ 7.0 Cladocera Alona - 3 1.0_+ 0 Bosmina 21.8 5.6 _+ 5.0 55 10.4 _+ 14.2 Ceriodaphnia - 35 5.4 _+ 4.5 Chydorus - 5 1.0+ 0 Daphnia 6.4 1.0+0 44 5.5+ 4.9 Macrothr& 1.6 3.0 +_ 1.4 - - Copepoda 2.4 1.0+0 45 4.8_+ 9.5 Ostracoda - 4 4.3 + 4.2 Hydracarina - 3 1.0+ 0 Insecta Chironomidae 5 1.0_+ 0 Trichoptera 2 t.0+ 0 Undetermined 1.6 9 10.7 + 10.8 Organic matter 8.1

Larvae examined 124 75 Number (%) empty 83 (67) 28 (42) 63

~o,c..t.l. A 3,0 2,0 ,,0 B 6- Percentage ,0- 0- ,,o ~o ~o ,,o t 'o ~o a,o 4,0 W.t., Col.~ 3O.- ZC- 2~ 35- wate r Column Lltvll 810mlChl 3~ 40- 3~ 45-- 40-

4S-

x X % 300" 400-

600- 5O0- *0C- e00-

7OC- ?00-

e0C- 800- 000 ~ e0C'- 1000-- 1000-

IOrO00 1100-

IS~000--

20,000--

Fig. 2. Size (volume)-frequency distributions expressed as percentage of zooplankters in the Lake Mohave water column (darkened, left) and in stomachs of larval razorback sucker from the reservoir (open, right); + indicates <1%; index of selection, L, indicated where/L/> about 10 (see text). A. Total zooplankton: water column n = 354 potential food items, larval stomachs n = 165 food items. B. Bosmina: water column n = 176 potential food items, larval stomachs n = 149 food items. to and largely a function of their primary food, Predominant foods of sucker larvae in the back- Bosmina (Fig. 2B). water included rotifers, cladocerans (Bosmina, Ceriodaphnia, and Daphnia) and copepods (Table 2). Dietary diversity was greater than in the reser- Arizona Bay Backwater voir, and included individual animals of larger size (e.g., larval Chironomidae and Trichoptera). Razorback sucker larvae were first captured from However, we found no evidence that mean size of the Arizona Bay backwater on 7 March, 5 to 8 ingested food increased as fish TL increased from weeks after adults were stocked. Numbers in- <12mm to -20mm. Backwater larvae showed creased quickly and remained high into April. positive selection for cladocerans (especially Mean TL of backwater larvae (n = 367) was ini- Bosmina) and strong negative selection for rotifers tially similar to that of Lake Mohave fish (approx- (Table 1). imately 10mm), but increased to >16mm by late Among available foods, backwater larvae ex- April when the largest specimen was 20.3mm, hibited positive selection for zooplankters 50 to similar in size to that achieved by hatchery-reared 399 × 10 -4 mm 3 and negative selection for those 10 razorback in comparable time (Papoulias unpubl. to 14 × 10-4mm 3 (Fig. 3a). Positive selection was data). Larvae in the backwater thus survived to strongest for rotifers 25-29, Daphnia 100-299 and grow nearly twice the size (middle postlarvae) of 1,000~,999, and copepods 100-299 × 10 -4mm 3, the largest (early postlarvae) wild individuals from while strongest negative selection was for rotifers the reservoir. 10-14, Bosmina 100-199, and Daphnia 5,000- Zooplankton species richness was greater in the 14,999 × 10 -4 mm 3 (Figs. 3b-e). backwater than in the reservoir, but the most abundant kinds were the same in both places (Table 2). Rotifers were relatively more numerous (more Discussion than 50%) and copepods less abundant in the backwater (Table 1). Total zooplankton density aver- Zooplankton communities and larval razorback aged 377 animals m -3, only about 25% that of the sucker diets were similar in Lake Mohave and in reservoir. 64

A o.,, • • t =o ,o ,,o 2p B

Percentage

Io- 9- Ii- ii L,,*.l atom,oh, so- 5- al- -i. / 10- ,o- A "o 15-

.;,o~ L~]*I r.4 x 2O" - 200- d I~0,1 ~t 25- £ ~ 400. 30- 60o- *Oo- E= 35- 700- o 800. 49- CO0. I000- II 45- 10.000. 50- 16.00~ ~o,ooo-

C Percentage D Per©enteoe ~o ,? s? ~o ,,o I 1p 2? 3,o ,,o o- 5~ ,,o 3p ~p ,O , tO ~O ~,o ,p °p e,o Water Column Larvsl 8lomeche 5O- 35- Water Column ~_.~srvsl Stomachs

4O- 2O0,- 13.e 45- 3011-

4OO- A ~ lO qr° 500.- -10.2 x eOO- x 200- ,ram ~E ~ 700- ~ 300- E 8011. ~ 400- >o 9011. o 500- > 10O11- ,oo 600"- S000- 10.000- 700- 15,000- 800- 20.00O-

o- e- Io-

~o. 2s- ~c-

x 4o-

20O- ~00,- 400'- SO0- COO- TO0"- SOC-

sooo-

Fig. 3. Size (volume)-frequency distributions expressed as percentage of zooplankters in the Arizona Bay backwater (darkened, left) and in stomachs of larval razorback sucker from the backwater (open, right); + indicates <1% ; index of selection, L, indicated where /L/> about 10 (see text). A. Total zooplankton: water column n = 382 potential food items, larval stomachs n = 1,020 food items. B. Rotifera: water column n = 221 potential food items, larval stomachs n = 109 food items. C. Bosmina: water column n = 51 potential food items, larval stomachs n = 427 food items. D. Daphnia: water column n = 49 potential foods items, larval stomachs n = 181 food items. E. Copepoda: water column n = 61 potential food items, larval stomachs n = 165 food items. 65 the isolated backwater. Taxa and sizes most Down-lake transport may occur, but razorback strongly selected as food by larvae in the backwater populations in the Colorado River below Davis were quantitatively more abundant in the reser- Dam, in downstream Lake Havasu, and below, are voir, yet selection by reservoir larvae for these very small; only sporadic individuals have been same sizes was negative or neutral. This phe- recorded in recent years (Minckley 1983, Ulmer nomenon may have been partly a function of food 1987). If down-lake transport is an important mech- density differences (e.g., Paloheimo 1979, Chesson anism removing young razorback from Lake Mo- 1983). We are aware of no distinctions in behavior, have, it must thus result in high mortality. Validity pigmentation, habitat, etc., between zooplankters of the out-of-system transport hypothesis should be in the two habitats that would help explain ob- evaluated and tested. served differences in selection. Last is mortality. Data continue to accumulate to More importantly, our comparisons indicate strengthen the idea that predation by introduced availability of larval foods of appropriate type, fishes playes an important role in declines of native number, or size does not account for apparent total fishes, and contributes to the continued perilous mortality of larvae in the reservoir. Papoulias (un- status of several species. Unanticipated events publ. data) found that larval razorback survival which terminated the present investigation support (but not growth) was independent of productivity this contention. High reservoir level (196.8m) (level of fertilization) among experimental ponds combined with storm-driven waves on 4 April 1985 at Dexter National Fish Hatchery (DNFH), New to breach the berm that separated the Arizona Bay Mexico. backwater from Lake Mohave. Razorback sucker Declines of razorback sucker and other native larvae were abundant in the backwater and rela- fishes of the Colorado River basin have generally tively large (>20mm TL, middle postlarvae) at been attributed to a suite of physical modifications, that time. Within the next few days, non-native which render habitats inhospitable to life-cycle fishes including threadfin shad, carp, channel cat- completion, or to introduced fishes that have direct fish, largemouth bass, green sunfish, and bluegill or indirect impacts on native kinds (e.g., Miller had invaded, and nearly 40% of green sunfish cap- 1961, 1972, Minckley & Deacon 1968, Johnson & tured over a 24 h period contained an average of Rinne 1982, Minckley 1983). Physical and biolog- four razorback sucker larvae. Larvae could no ical effects often cannot be separated, since major longer be collected by 5 May, when the backwater modifications to the river system and widespread again was isolated, and we attribute total mortality introductions of exotic species were approximately of the stock to predation. Predation on razorback concomitant in time and place. sucker eggs and larvae has further been docu- Explanations for apparent disappearance of lar- mented in Senator Wash Reservoir, California, val razorback sucker from Lake Mohave include and Lake Mohave (Medel-Ulmer 1983, Louder- migration to inaccessible habitats within the reser- milk 1985, Langhorst 1987, Marsh 1987). Marsh & voir, down-lake transport out of the area, and mor- Brooks (unpubl. data) recorded potentially signifi- tality. The first is rejected since extensive collec- cant predation by introduced ictalurid catfishes on tions using gears that have effectively taken juvenile razorbacks stocked into the middle Gila thousands of individuals of other species in the size River, Arizona, and Brooks (unpubl. data) deter- range of juvenile suckers failed to take small razor- mined predation loss of young suckers to be in back. This sampling has been conducted since the direct relation to density of green sunfish in experi- early 1970s and covered all available habitats ex- mental ponds at DNFH. Additional evidence cept the deepest profundal (Minckley 1983, un- comes from study of other fishes in the Region publ. data). These same techniques have suc- (e.g., Meffe et al. 1983, Meffe 1985) and elsewhere cessfully captured juvenile suckers reintroduced (e.g., Selgeby et al. 1978, Lemly 1985, Loftus & into several central Arizona streams (unpubl. Hulsman 1986). data). Most studies of predation on wild razorback 66 sucker larvae have failed to produce quantitative tial, perhaps repeated, removal of non-native data, and thus could only speculate on importance fishes. of predation as a cause of mortality. A major prob- lem has been detection of prey items in stomachs of fishes that masticate or otherwise render foods vi- Acknowledgements sually unrecognizable, or after digestion has pro- ceeded for periods of more than a few hours (un- This work was funded in part by contracts to Ari- publ. data). These difficulties must be resolved if zona State University from U.S. Fish and Wildlife the role of predation is to be determined. The Service Region II, Office of Endangered Species, former may be overcome by use of serological tech- Albuquerque, New Mexico; U.S. Bureau of Recla- niques (Boreham & Ohiagu 1978, Engvall & Perl- mation, Lower Colorado Regional Office, Boulder mann 1972, Heusser et al. 1981, Monroe 1985), but City, Nevada; and Arizona Game and Fish Depart- even these methods may be time-limited by rapid ment (AZGFD), Phoenix. Permits were issued by prey digestion (Theilacker et al. 1986) and thus U.S. National Park Service, Nevada Department require close-order sampling. of Wildlife, and AZGFD. J. Brooks, T. Burke, Investigations to confirm fate of larval razorback and W.L. Minckley participated in various aspects sucker in Lake Mohave have yet to be conducted, of the study; they and W.R. Courtenay, D.L. but may be critical to management and recovery of Galat, and D. Papoulias reviewed the manuscript. the species. If, for example, it were determined that a significant reduction in predation pressure on wild or reintroduced stocks would allow References cited establishment of self-perpetuating populations of the native, then management activities might be Allan, R.C. & D.L. Roden. 1978. Fish of Lake Mead and Lake directed toward reduction of established popula- Mohave. Biol. Bull. No. 7, Nevada Department of Wildlife, tions of known predators. Ichthyocide reclamation Reno. Boreham, P.F.L. & C.E. Ohiagu. 1978. The use of serology in to remove catfishes from selected Arizona streams evaluating invertebrate predator-prey relationships: a re- and subsequent stocking of razorback suckers is an view. Bull. Entomol. Res. 68: 171-194. example of how this management might be imple- Chesson, J. 1983. The estimation and analysis of preference and mented. Such an experiment may be necessary to its relationship to foraging models. Ecology 64: 1297-1304. resolve the elusive question of why stocked juve- Engvall, E. & P. Perlmann. 1978. Enzyme-linked immunosor- bent assay, ELISA. J. Immunol. 109: 129-135. nile razorback suckers have apparently realized Heusser, C.H., J.W. Stocker & R.H. Gisler. 1981. Methods for limited survival (Marsh 1987, unpubl, data) and, at binding cells to plastic: application to solid phase immu- least circumstantially, would contribute toward ex- noassays for cell-surface antigens. Meth. Enzymol. 73: 406- plaining lack of recruitment in Lake Mohave and 417. elsewhere. Holden, P.B. 1980. Xyrauchen texanus (Abbott) humpback sucker, pp. 435. In: D.S. Lee, C.R. Gilbert, C.H. Hocutt, The alternative of planting razorback suckers at R.E. Jenkins, D.E. McAllister & J.R. Stauffer (ed.) Atlas of a size relatively immune to predation may succeed North American Freshwater Fishes, North Carolina State in reestablishing adult stocks, but still fail to insure Museum of Natural History, Raleigh. self-perpetuation if predation on larvae prevents Johnson, J.E. & J.N. Rinne. 1982. The endangered species act recruitment. Unfortunately, reclamation is proba- and southwestern fishes. Fisheries 7: 1-10. Langhorst, D.R. 1987. Larval razorback sucker, Xyrauchen bly infeasible other than in local areas of reservoirs texanus, in Lake Mohave, AZ-NV. Proc. Desert Fishes and remaining mainstream habitat of the Colorado Council 17 (in press). River that comprise much of the original range of Lemly, A.D. 1985. Suppression of native fish populations by the species. Eventual recovery of razorback sucker green sunfish in first-order streams of piedmont North Car- olina. Trans. Amer. Fish. Soc. 114: 705-712. and other imperiled big-river fishes of the South- Loftus, D.H. & P.F. Hulsman. 1986. Predation on larval lake west may best be achieved in backwaters, oxbows, whitefish ( Coregonus clupeaformis) and lake herring ( C. ar- and smaller tributary habitats amenable to substan- tedii) by adult rainbow smelt (Osmerus mordax). Can. J. Fish. Aquat. Sci. 43: 812-818. 67

Loudermilk, W.E. 1985. Aspects of razorback sucker (Xy- Minckley, W.L. & J.E. Deacon. 1968. Southwestern fishes and rauchen texanus, Abbott [sic]) life history which help explain the enigma of 'endangered species'. Science 159: 1424-1432. their decline. Proc. Desert Fishes Council 13: 67-72. Minckley, W.L. & E.S. Gustafson. 1982. Early development of Marsh, P.C. 1985. Effect of incubation temperature on survival the razorback sucker, Xyrauchen texanus (Abbott). Great of embryos of native Colorado River fishes. Southwest. Nat. Bas. Nat. 42: 553-561. 30: 129-140. Monroe, D. 1985. The solid-phase enzyme-linked immunospot Marsh, P.C. 1987. Razorback sucker management in Arizona. assay: current and potential applications. Biochem. Technol. Proc. Desert Fishes Council 17 (in press). 3: 222-229. McAda, C.W. & R.S. Wydoski. 1980. The razorback sucker, Paloheimo, J.E. 1979. Indices of food type preference by a Xyrauchen texanus, in the upper Colorado River basin, 1974- predator. J.Fish. Res. Board Can. 36: 470-473. 1979. U.S. Fish & Wildl. Serv. Tech. Papers 99: 1-15. Priscu, J.C., J. Verduin & J.E. Deacon. 1980. The fate of McCarthy, M.S. 1986. Age of imperiled razorback sucker biogenic suspensoids in a desert reservoir, pp. 1657-1667. In: (Pisces, Catostomidae) from Lake Mohave, Arizona-Ne- H.G. Stephan (ed.) Symp. Surf. Wat. Impoundments ASCE, vada. M.S. Thesis, Arizona State University, Tempe. 38 pp. Minneapolis. Medel-Ulmer, L. 1983. Movement and reproduction of the Priscu, J.C., J. Verduin & J.E. Deacon. 1982. Primary pro- razorback sucker (Xyrauchen texanus) inhabiting Senator ductivity and nutrient balance in a lower Colorado River Wash Reservoir, Imperial County, California. Proc. Desert reservoir. Arch. Hydrobiol. 94: 1-23. Fishes Council 12: 106. Selgeby, J.H., W.R. MacCallum & D.V. Swedberg. 1978. Pre- Merle, G.K. 1985. Predation and species replacement in Amer- dation by rainbow smelt (Osmerus mordax) on lake herring ican southwestern fishes: a case study. Southwest. Nat. 30: (Coregonus artedii) in western Lake Superior. J. Fish. Res. 173-187. Board Can. 35: 1457-1463. Meffe, G.K., D.A. Hendrickson, W.L. Minckley & J.N. Rinne. Strauss, R.E. 1979. Reliability estimates for Ivlev's electivity 1983. Factors resulting in decline of the endangered Sonoran index, the forage ratio, and a proposed linear index of food topminnow Poeciliopsis occidentalis (Atheriniformes; selection. Trans. Amer. Fish. Soc. 108: 344-352. Poeciliidae) in the United States. Biol. Conserv. 25: 135-159. Strauss, R.E. 1982. Influence of replicated subsamples and sub- Miller, R.R. 1961. Man and the changing fish fauna of the sample heterogeneity on the linear index of food selection. American Southwest. Pap. Michigan Acad. Sci., Arts Lett. Trans. Amer. Fish. Soc. 111: 517-522. 46: 365-404. Theilacker, G.H., A.S. Kimball & J.S. Trimmer. 1986. Use of Miller, R.R. 1972. Threatened freshwater fishes of the United an ELISPOT immunoassay to detect euphausid predation on States. Trans. Amer. Fish. Soc. 101: 239-252. larval anchovy. Mar. Ecol. Prog. Series 30: 127-131. Minckley, W.L. 1983. Status of the razorback sucker, Xy- Ulmer, L. 1987. Razorback sucker management in California. rauchen texanus (Abbott), in the lower Colorado River basin. Proc. Desert Fishes Council 17 (in press). Southwest. Nat. 28: 165-187.