MORTALITY OF JUVENILE AMERICAN SHAD AND STRIPED BASS PASSED THROUGH OSSBERGER CROSSFLOW TURBINES AT A SMALLSCALE HYDROELJKTRIC SITE Robert B. DuBois Steven P. Gloss

Journal Article 1993 WWRC-93-14

In North American Journal of Fisheries Management

Robert B. DuBois and Steven P. Gloss New York Cooperative Fishery Research Unit Department of Natural Resources Cornell University Ithaca, New York North American Journal of Fisheries Management 13: 178-185, 1993

Mortality of Juvenile American Shad and Striped Bass Passed through Ossberger Crossflow Turbines at a Small-Scale Hydroelectric Site

ROBERTB. DUBOIS’AND STEVENP. GLOSS^ New York Cooperative Fishery Research Unit3 Department of Natural Resources, Cornell University, Ithaca, New York 14853, USA

Abstract. -A full-recoverytechnique was used in mortality experiments conducted with juveniles of American shad Alosa sapidissima and striped bass Morone saxatilis passed through Ossberger crossflow turbines to obtain antecedent information about their fish passage characteristics. Im- mediate turbine-induced mortality was 66% for 85-mm-long (total length) American shad. Turbine- induced mortality of striped bass was significantly related to the total length of the fish and ranged from l6Y0 for 67-83-mm-long fish to 39% for 136-mm-long fish immediately after passage; after 24 h, turbine-induced mortalities of these two size-groups were 61 and 72%, respectively. The mortality of striped bass was not affected by power output (320-600 kW) of the turbine or by turbine size (650 versus 850 kW). Because of high mortality of control fish, the full-recovery technique was not fully adequate for obtaining reliable delayed-mortality estimates for these fragile fish species.

Recent concerted efforts to rehabilitate runs of been fiequently documented (Raney et al. 1954; anadromous fishes in rivers of eastern North Nicholas and Miller 1967; Kynard and Warner America have, by chance, coincided with a resur- 1987). Whereas reasonable mortality estimates are gent interest in small-scale hydropower develop- available for several salmonid species passed ment in this region (Gloss and Wahl 1983). Fish through any of various turbine types (Turbak et passage facilities at many hydroelectric sites are al. 1981; Gloss and Wahl 1983), such data are allowing movement of anadromous fish migrants scarce for nonsalmonids, as is information on the to upper-river spawning areas, but few sites have suitability of associated-mortality testing proce- downstream bypass capabilities to protect mi- dures. Taylor and Kynard (1985) estimated im- grating juveniles, subadults, and postspawn adults mediate turbine-induced mortality of juvenile from entrainment. Total mortality probable for American shad passed through a 17-MW Kaplan fish passed through turbines, particularly juvenile turbine at three levels of power generation. Stokes- migrants, has consequently become an important bury and Dadswell (199 1) estimated the imme- consideration in regional fisheries management diate mortality of several clupeid species passing planning and environmental impact assessment. through a STRAFLO turbine at a low-head tidal- Populations of American shad Alosa sapidissi- power dam. However, in neither case were shad ma are being successfully restored with increasing held for determination of delayed mortality; hence, frequency in U.S. east-coast river systems where it is likely that these results underestimate total hydropower obstacles must be negotiated (Howey turbine-induced mortality. No published docu- 198 1; Moffitt et al. 1982). Migrations of subadults mentation exists for turbine-induced mortality of of striped bass Morone saxatilis into both natal subadult striped bass. and nonnatal rivers (and through fishways) have Estimates of mortality resulting from passage of fish through hydraulic turbines have been based Present address: Department of Natural Resources, on various recovery techniques. These have in- -.. Bureau of Research, Post Office Box 125, Brule, Wis- cluded using returns of adult fish that were marked consin 54820, USA. as juveniles before turbine passage and the partial- Present address: Wyoming Water Research Center, recovery techniques of Schoeneman et al. (196 l), / ,i Post Office Box 3067, University Station, University of as well as full-recovery methods in which all fish Wyoming, Laramie, Wyoming 8207 1, USA. are recovered in nets immediately after they pass The New York Cooperative Fishery Research Unit through turbines (Cramer and Donaldson 1964). is jointly sponsored by the New York State Department of Environmental Conservation, Cornell University, the If returns of adult fish are monitored, several years U.S. Fish and Wildlife Service, and the Wildlife Man- are required to evaluate results, and there are in- agement Institute. herent, potentially severe, problems related to sta- 178 TURBINE-INDUCED MORTALITY 179

tistical reliability. Partial-recovery methods, though potentially valid statistically (Paulik 196 l), require extremely large sample sizes (dependent on anticipated recovery rates) and the testing of numerous assumptions regarding recovery ratios of live and dead fish in experimental and control groups. Full-recovery methods yield near-unity values for recovery rates and ratios, and require substantially smaller sample sizes than do partial- Cylindrical Runner recovery methods to produce a similar degree of ’\ / statistical accuracy. Both recovery methods sub- ject test fish to mortality resulting from stress or FIGURE1. -A horizontal-admission Ossberger turbine mechanical injury, which is a major consideration in cross section. Arrows indicate direction of water flow. in the testing of fragile fish species. Most recovery methodologies were originally developed for use tecedent information about their fish passage char- with emigrating juvenile salmonids. The full-re- acteristics. (No anadromous fish populations were covery net system appears to have the most de- extant in this reach of the Susquehanna drainage.) sirable properties for estimating salmonid mor- Ossberger turbines are of the radial, impulse tality resulting from turbine passage (Turbak et al. type and operate at relatively low speed. The two 1981), but this system has not been adequately guide vanes, which regulate water intake, and the tested with nonsalmonid species. cylindrical runner (Figure 1) are the only moving We used a kll-recovery system originally de- parts. Control of intake flow by partial or complete signed to estimate mortality incurred by juvenile opening of one or both guide vanes allows these salmonids passed through Ossbergercrossflow tur- turbines to develop optimum efficiency over water bines to estimate turbine-induced mortality for flows of 16-100% of each unit’s capacity. The rap- juvenile American shad and striped bass. Mor- idly ascending asymptotic efficiency curve typical tality estimates were derived with a relative re- of Ossberger crossflow turbines is in contrast to covery rate estimator, as described by Ricker (1975) the more parabolic curves of reaction-type tur- and Burnham et al. (1987), to separate mortality bines. The relation between water discharge and caused by turbine passage from that attributable generator output is nearly linear over the plateau to other sources (e.g., damage caused by recovery of operating efficiency, and revolutions of the run- nets). Mortality differences attributable to fish size, ner are constant in this range (Table 1). Spacing turbine size, and power output were examined. between the runner blades is approximately 30 Our objectives addressed two issues of urgent con- mm on the 650-kW unit and 40 mm on the 850- cern to workers assessing the impacts of hydro- kW unit. The clearance between the runner and power development on anadromous fish species: the housing of each is about 3 mm. More complete (1) what mortality can be expected among juvenile descriptions of the study site and turbine were out-migrant American shad and striped bass, and given in Gloss and Wahl (1983). (2) what procedures can be used to estimate mor- tality for fragile fish species. Methods Two general size-groups (mean total lengths, 67- Study Site and Turbine 83 mm and 136 mm) of striped bass were obtained The Colliersville Dam (9.75-m head) on the from the Harrison Lake National Fish Hatchery, north branch of the Susquehanna River in south- Charles City, Virginia. (Hereafter, all fish lengths central New York impounds the 150-hectare are given as total length.) Handling procedures for c Goodyear Lake. The power plant, which was op- erated commercially from 1907 to 1969, was re- furbished during the late 1970s and retrofitted with TABLE1. -Operating characteristics of Ossberger tur- i two Ossberger crossflow turbines of 650 and 850 bines at Goodyear Lake.

~~~ kW. The entire flow available, up to 20 m3/s (the Rated Diameter Maximum maximum discharge capacity of the two turbines), output of runner Revolutions discharge Design is diverted through the powerhouse. These Oss- (kW (m) per minute (m3/s) head (m) berger turbines were the first installed in the USA, 650 1.oo 135 8.5 10 and they provided the opportunity to obtain an- 850 1.25 104 11.5 10 180 DuBOIS AND GLOSS

Introduction Tank

Inside Powerhouse

Headrace

Live Recovery Device

FIGURE2.-Cross section of powerhouse and tailrace at Colliersville Dam, showing apparatus for introduction and recovery of fish. striped bass were identical to those described for All equipment and procedures recommended by Atlantic salmon (Gloss and Wahl1983). American Chittenden (197 1) to minimize excitement-related shad (mean length, 85 mm) used for turbine pas- stress were adhered to, including transport in aer- sage experiments were obtained from the lower ated tanks containing a 0.5% NaCl solution. Delaware River near Byram, New Jersey, during the 198 1 emigration. A single group of 4045-mm- Fish Passage long hatchery-reared American shad (Lamar Fish- The procedures and gear described by Gloss and eries Research and Development Center, U.S.Fish Wahl(1983) for estimating turbine-induced mor- and Wildlife Service, Lamar, Pennsylvania) was tality of juvenile salmonids were adapted for sim- used for control experiments because of difficulty ilar experiments with juvenile American shad and in obtaining adequate numbers of wild American striped bass. Known numbers of test fish (35-1 00, shad when needed. Wild American shad were col- varying with fish size and species) were acclimated lected by seining without beaching the bag of the in an introduction tank on the deck of the pow- seine. Rather, American shad were crowded into erhouse, then released through the bottom of the the bag of the seine with a live-cage designed for tank by way of an introduction hose that emptied holding fish (Buynak and Mohr 1980), which was underwater behind the trash racks and just up- held on its side in shallow water. The seine was stream from the turbine intake (Figure 2). Water worked close to the sides of the cage until most velocity near the end of the delivery tube was 0.3- fish were crowded into the live-cage; the cage was 0.6 m/s (measured with a Marsh-McBirney direct- then turned upright in the water. American shad reading current meter) and theoretically increased were dipped out of the live-cage with plastic buck- toward the turbine. Fish were recovered in a float- ets and dishpans and transferred directly into the ing recovery box similar to that described by Crad- i transport tank; thus, they were never taken from dock (1961), after they passed through a trawl- the water or dip-netted (Chittenden 1971). All sub- shaped recovery net that encompassed the outlet sequent transfers of American shad were made by of the turbine draft tube (mouth of net was entirely slowly dipping of individual fish or small groups above water during discharge). The recovery net with buckets or dishpans; no dip nets were used. was constructed of treated knotless nylon with a TURBINE-INDUCEDMORTALITY 181

mouth opening of 4.6 x 3 m and length of 1 3.5 recovered, we assumed that dead, injured, and un- m. Mesh sizes (stretched measure) were 19 mm in injured fish in this category had equal chances of the mouth section (8 m long), 12 mm in the throat not being recovered. For each trial we calculated section (2.5 m long), and 6.3 mm in the cod end escapement (1 - [number recoverednumber in- (3 m long). The maximum turbine discharge rate troduced]), immediate mortality, and cumulative at which we introduced experimental or control mortality after 24 h. Because control mortality fish was about 8.5 m3/s. Current velocity near the occurred for both species (particularly from the mouth of the recovery net was modest and only recovery process), we used a relative recovery rate reached 0.3 m/s under peak discharge. After pas- estimator as described by Ricker (1 97 5)and Bum- sage, recovered fish were classified as dead, in- ham et al. (1987). The basic structural model for jured, or uninjured (no apparent physical damage calculating turbine mortality (M) was or loss of orientation).Living fish were transferred M = 1 - [(A/R)/(a/r)]; from the recovery box to covered, floating live- cages (Buynak and Mohr 1980)that were anchored A = number of fish alive at recovery after a in quiet water adjacent to the tailrace for obser- turbine passage trial, vations of delayed mortality. Fish were examined R = number of fish recovered after a turbine twice at 24-h intervals. passage trial, a = number of fish alive at recovery after a Control Experiments recovery-net trial, and Each phase of the test procedure for both species r = number of fish recovered after a recovery- was examined for sources of mortality other than net trial. passage through the turbine. In tests simulating Because there was no immediate mortality of introduction, fish were discharged through a de- striped bass during any of the control experiments, livery tube (about the same length and angle as no recovery-net adjustment was necessary for cal- the test introduction hose) into a 1,900-L circular culating their immediate turbine-induced mortal- tank, from which they were recovered and mon- ity. Estimates of delayed (24-h) turbine-induced itored. Mortality attributable to the recovery ap- mortality of striped bass and estimates of imme- paratus was examined by placing fish into the diate turbine-induced mortality of American shad mouth of the recovery net while the turbine was were corrected for recovery-net mortality as in the operating (at the same level of turbine output as above equation. We did not attempt to estimate in passage trials). These fish were subsequently turbine-induced mortality of striped bass beyond recovered from the live-box and treated as during 24 h or delayed turbine-induced mortality of turbine passage trials (they spent the same amount American shad, because when control mortality is of time in the recovery box). Although a seemingly high, as it was for these groups, the relative-re- better experimental design would have been to covery-rate method is not useful. We ignored have simultaneously released treatment and con- holding and handling mortality because there was trol fish, such an approach would have required none initially, nor was there any for striped bass marking fish, which would have contributed an after 24 h. We analyzed percent mortality and re- additional source of stress. Hence, control trials covery rate data with the nonparametric Mann- were run at a separate time. Estimates of mortality Whitney U-test, Kruskal-Wallis test, and Wilcox- attributable to handling and holding procedures on rank-sum test, and we used linear regression were made by acclimating fish in the introduction to examine the relationship between mortality and tank, then recovering and monitoring these fish. power output. A naiysis Results cr Most test and control trials were conducted in triplicate, though several trials were replicated four American Shad or five times. Occasionally, only one or two trials Four batches of wild American shad were sub- 6 were done because of limited numbers of fish or jected to an 8-10-h process involving collection, failure of equipment. We assumed that virtually transportation, and handling. The first batch con- all fish that were released for turbine passage but sisted of 3,000 fish, and although initial mortality were not recovered had escaped upstream before (i.e., mortality at the time of transfer to holding entering the turbine. However, if a few fish did tanks at the site) was only 0.5% in a 700-L tank, pass through a turbine and were not subsequently 99% of the fish died within 24 h. During subse- 182 DuBOIS AND GLOSS

TABLE2.-Mean mortality of American shad and Immediate mortality was also low in recovery- striped bass passed through Ossbergerturbines (data from net experiments (mean, 6%; range, 0-16Yo; seven 650-kW and 850-kW units combined). trials; 30 fiswtrial), but mortality after 24 h was Mean turbine- high (mean, 86%), precluding our ability to esti- induced mate delayed mortality of American shad resulting mortality Size- Mean from turbine passage. The immediate turbine-in- (%) group number duced mortality of American shad through the (total Number offish 24-h length, of per Imme- cumu- 850-kW unit was 66% (Table 2), and power output Species mm) trials ma1 diate lative (range, 320-600 kW) did not have a significant American shad 85 12 42 66 a effect (P> 0.82). Recovery of American shad after Striped bass 136 6 96 39b 72 turbine passage averaged 83%, and recovery-net Striped bass 67-83 27 50 16 61 recovery averaged 75% (see Discussion for the ef- a This mortality value could not be estimated because of high fect of leaf fall on recovery). mortality related to the recovery net. Mean based on only four of the six trials. Striped Bass We had little difficultyin transporting or han- dling the two size-groups of striped bass. Control quent collections we transported smaller batches experiments indicated that only the recovery net (500-1,000 fish) with greater success. Initial post- caused substantial mortality and that this mortality transport mortality was characteristically low, did not occur initially (92% of striped bass used ranging from 1 to 3%. However, losses through 48 in recovery-net trials were recovered). Mean 24-h h consistently exceeded 60%. Mortality after the recovery-net mortality was 2 1% ( 16 trials) and was first 48 h averaged 24Yo daily -an addition of less not influenced by fish size. Although recovery-net than 1940 per day to the cumulative mortality. Be- losses of striped bass were lower than those for cause mortality leveled off after 48 h, American American shad, they still represented a confound- shad were held at least that long before survivors ing factor that complicated our efforts to estimate were used for turbine mortality and recovery-net delayed mortality due to turbine passage. trials. American shad are widely recognized for Striped bass mortality data for various discharge their sensitivity to stresses imposed by virtually rates of the turbines were combined because sig- any type of collection or handling procedures and nificant differences (P < 0.05) were not detected are therefore not readily available for experimen- among them or between the two turbines of dif- tal purposes. Other control mortality experiments ferent capacities. The immediate turbine-induced (handling, holding cages, simulated introduction) mortality for both size-groups combined was 19% were conducted with groups of smaller (4045- (3 1 trials; mean recovery rate, 63%). The delayed mm-long) hatchery-reared American shad. These (24-h) mortality caused by turbine passage of fish were too small to be used for recovery-net striped bass was 63Oh (33 trials). The only turbine- trials (because they would pass through the recov- induced mortality comparisons that were signifi- ery net’s mesh). cantly different (P < 0.05) occurred between the Control experiments with wild American shad 136-mm and 67-83-mm length-groups. Imme- produced substantial mortality only during recov- diate turbine-induced mortality was significantly ery-net trials. In duplicate or triplicate experi- greater (P< 0.01) for the larger size-group (mean, ments (50-100 fish/replicate) where hatchery- 39%; range, 1748%) than for the smaller size- reared American shad were placed in either group (mean, 16%; range, 040%; Table 2). The live-cages, introduction tanks, or the recovery live- delayed (24-h) turbine-induced mortality for the box under discharge conditions for 48 h, mortality larger group (72%) was not significantly different from each source of stress was low (mean, 7%; (P> 0.09) from that for the smaller group (61% 3 range, 0-1 6%; seven trials; 50-100 fish/trial). Sim- Table 2). ulated introduction caused an average loss of 3% through 48 h (range, 04%;four trials; 100 fish/ Discussion li trial). Neither the simulated introduction tech- Our results describing turbine-induced mortal- nique nor the handling and holding procedures ity of juvenile American shad and striped bass alone produced large immediate or delayed mor- paralleled results of a similar study with salmonids tality. by Gloss and Wahl(1983) in showing (1) no change TURBINE-INDUCED MORTALITY 183 in mortality due to power output of Ossberger tur- TABLE3.-Mean recovery of American shad and a bines, and (2) a significant positive relationship size-group of striped bass (total length, 67-83 mm) after between fish size and mortality. The lack of change turbine passage at low, moderate, and high levels of power output. Paired comparisons of 320-kW and 520- in mortality relative to power output is probably 600-kW output tests (P value; Mann-Whitney U-test) due to the flat efficiency curve typical of these showed higher recovery of both species at high output. turbines over a wide range of power outputs. This Numbers of trials (35-50 fishltrial) are in parentheses. efficiency characteristicprovides little opportunity Percent recovery at to mitigate periods of high mortality potential (out- Turbine power output of: migrations) by changing operating conditions. The capac- 320 440 520-600 immediate turbine-induced mortality we estimat- Species ity(kW kW kW kW P ed for American shad (66%) is similar to that re- Americanshad 850 77(2) 77(6) 95(4) 0.07 ported by Taylor and Kynard (1985) for juvenile Striped bass 850 48 (3) 77 (5) 71 (7) 0.03 alosids (62%) passed through Kaplan turbines at Striped bass 650 17 (4) 34 (5) 52 (3) 0.03 their power output of greatest efficiency. The im- mediate and delayed turbine-induced mortality estimates for striped bass we provide are the first for this species passed through any type of turbine. over the first week. If delayed mortality of alosines A substantial body of evidence has now accrued after passage through other types of turbines is (Smith 1960, 1961; Taylor and Kynard 1985; similarly substantial, total turbine-induced losses Stokesbury and Dadswell 1991; this article) doc- could easily exceed 8OYo in some cases. We men- umenting highly variable immediate mortality, tion this point to highlight the need to develop ranging f?om 14 to 82%, for juvenile alosines passed methodologies for estimating thc delayed com- through any of various turbine types. It is not ponent of turbine-induced mortality in addition known why these estimates have been so variable. to the immediate component, not to advocate use Taylor and Kynard (1 985) suggest that such dif- of Smith’s estimate of delayed mortality as a stan- ferences may be species- or age-specific, may be dard. due to experimental design differences, or may be Based on our results, the full-recoverytechnique attributable to turbine design, setting, loading, or has serious limitations for use in estimating total cavitation characteristics. In general, this evidence turbine-induced mortality of fragile fish species indicates that alosines are subject to higher levels because of considerable mortality attributable to of mortality after turbine passage than are salmo- the recovery apparatus. Many hydropower sites nids. However, the lack of reliable estimates of have higher tailrace current velocities than those delayed turbine-induced mortality for alosines experienced in our study. Our failure to obtain raises serious questions about the use of present reliable estimates of delayed turbine-induced mor- estimates for estimation of total turbine-induced tality of American shad under relatively benign mortality. Although it is generally recognized that conditions at a small-scale site suggests that full- available estimates may underestimate turbine-in- recovery techniques are not suitable for this spe- duced mortality because they ignored the factor cies. of eventual death from shock, stress, or injuries Not all fish introduced in turbine passage trials that did not leave visible signs of damage (Stokes- in this study were subsequently recovered, which bury and Dadswell 1991), little information is raises the possibility that uninjured survivors of available about how much delayed mortality may turbine passage may have had a higher probability occur after fish passage through different types of of escaping recovery than injured or dead fish had. turbines. Smith (1960, 1961) is the only worker This possibility is important to consider, because to report estimates of delayed turbine-induced if it had occurred, an overestimation of turbine- mortality for alosines (calculated by subtracting induced mortality would have resulted. Two lines immediate mortality from mean total mortality of evidence support our assumption that we re- after 1 week-an improper calculational ap- covered virtually all fish that entered the turbines. proach), but his results were not subjected to peer First, recovery was higher at high power outputs review and must be cautiously interpreted. His than at low outputs for both striped bass and results do, however, suggest that mortality of alo- American shad (Table 3), suggestingthat fish were sines immediately surviving passage through sometimes able to swim against the current and Smith-Kaplan turbines may have exceeded 30% escape into the headrace, particularly at lower 184 DuBOIS AND GLOSS power outputs. Second, recovery rates of 99-1 00% anonymous reviewers were appreciated. The co- in five trials with striped bass, and recovery rates operation of F. w. E. Stapenhorst, Inc., and Fred exceeding 95% in 25Oh of the trials with American Knott, manager of the Goodyear Lake Station, shad suggest that when fish entered a turbine, re- were instrumental to the project’s success. covery was high. A period of heavy leaf fall from 8 late September to early October represented the References only opportunity for fish entering a turbine to not Blaxter, J. H. S., and D. E. Hoss. 1979. The effect of be recovered (some fish may have been missed rapid changes of hydrostatic pressure on the Atlantic while we sorted through large volumes of leaves hening Clupea harengus L. 11. The response of the in both the recovery box and net). Recovery-net auditory bulla system in larvae and juveniles. Jour- trials, which should have provided a test of our nal of Experimental Marine Biology and Ecology 4 1: 8 7-1 00. recovery assumption, were conducted mostly dur- Burnham, K. P., D. R. Anderson, G. C. White, C. Brown- ing this period of leaf fall (16 of 19 trials), and ie, and K. H. Pollock. 1987. Design and analysis recovery was poorer during these trials than during methods for fish survival experiments based on re- the turbine passage trials. Only 12Oh of the turbine lease-recapture. American Fisheries Society Mono- passage trials were conducted during heavy leaf graph 5. fall. Buynak, G. L., and H. W. Mohr. 1980. Collapsible fish-holding cage. Progressive Fish-Culturist 42:4 1- The relatively close spacing of the runner blades 42. in Ossberger turbines may, in part, explain the Chittenden, M. E. 197 1. Transporting and handling greater mortality to larger fish (increased risk of young American shad. New York Fish and Game mechanical injury); however, our study design did Jo~rnal18: 123-1 28. not specificallyaddress the causes of turbine mor- Craddock, D. R. 196 1. An improved trap for the cap- tality. Cavitation is considered to be the most im- ture and safe retention of salmon smolts. Progres- portant source of mortality to young fish passed sive Fish-Culturist 23: 190-1 92. Cramer, F. K., and I. J. Donaldson. 1964. Evolution through turbines (Muir 1959; Ruggles et al. 1981). of recovery nets used in tests on fish passage through Pressure change appeared to be responsible for hydraulic turbines. Progressive Fish-Culturist 26: 65% ofthe injuries that a STRAFLO turbine caused 36-41. to clupeids (Stokesbwy and Dadswell 1991). In Gloss, S. P., and J. R. Wahl. 1983. Mortality ofjuvenile general, clupeids have shown more evidence of salmonids passing through Ossberger crossflow tur- pressure damage after turbine passage than have bines at small-scale hydroelectric sites. Transactions of the American Fisheries Society 112: 194-200. salmonids(Stokesbury and Dadswell 199l), which Howey, R. G. 1981. A review of American shad res- may result from their having an enclosed swim toration efforts on the Susquehanna River. U.S. Fish bladder that extends into the back of the head and and Wildlife Service, Lamar Information Leaflet 8 1- is in contact with the brain (Blaxter and Hoss 1979). 04, Lamar, Pennsylvania. Kynard, B., and J. P. Warner. 1987. Spring and sum- Acknowledgments mer movements of subadult striped bass, Morone saxatifis, in the Connecticut River. U.S. National Financial support for this study was provided Marine Fisheries Service Fishery Bulletin 85:143- by the U.S. Department of Energy through inter- 147. agency agreement 14-16-0009-80- 1004 with the Moffitt, C. M., B. Kynard, and S. G. Rideout. 1982. U.S.Fish and Wildlife Service (USFWS).We thank Fish passage facilities and anadromous fish resto- William Knapp (USFWS), and Stephen Hilde- ration in the Connecticut River basin. Fisheries (Be- brand and James har (Oak Ridge National Lab- thesda) 7(6):1-10. oratory), for their roles in fostering this research. Muir, J. F. 1959. Passage of young fish through tur- bines. Proceedings of the American Society of Civil Assistance with shad collections given by Arthur Engineers 85:23-46. Lupine and Joseph Miller (USFWS) was greatly Nicholas, P. R., and R. V. Miller. 1967. Seasonal appreciated. John Devore, Thomas French, Tho- movements of striped bass, Roccus saxatilis (Wal- mas Gabuzda, Raymond Mirande, and James baum) tagged and released in the Potomac River, Wahl provided field assistance. Michael Staggs and Maryland, 1959-61. Chesapeake Science 8:102-124. Paul Rasmussen (Wisconsin Department of Nat- Paulik, G. J. 196 1. Detection of incomplete reporting. 4 of tags. Journal of the Fisheries Research Board of ural Resources), and Kenneth Burnham (Colorado Canada 18:817-831. State University), made helpfbl suggestions on data Raney, E. C., W. S. Woolcott, and A. G. Mehring. 1954. analysis. Review comments from James Ander- Migratory pattern and racial structure of Atlantic son, E. Taft, JohnWilliams, Fred Stoll, and several coast striped bass. Transactions of the North Amer- TURBINE-INDUCED MORTALITY 185

ican Wildlife and Natural Resources Conference 19: County, Nova Scotia. Canada Department of Fish- 3 76-396. eries, Halifax. Ricker, W. E. 1975. Computation and interpretation Stokesbury, K. D. E., and M. J. Dadswell. 199 1. Mor- of biological statistics of fish populations. Fisheries tality of juvenile clupeids during passage through a Research Board of Canada Bulletin 1 9 1. tidal, low-head hydroelectric turbine at Annapolis 4 Ruggles, C. P., N. H. Collins, and R. H. Thicke. 198 I. Royal, Nova Scotia. North American Journal of Fish passage through hydraulic turbines. Transac- Fisheries Management 1 1: 149-1 54. tions of the Canadian Electrical Association 20(2)8 1 - Taylor, R. E., and B. Kynard. 1985. Mortality ofju- H- 1 1 1. Montreal. venile American shad and blueback hemng passed < Schoeneman, D. E., R. T. Pressey, and C. 0. Junge. through a low-head Kaplan hydroelectric turbine. 196 1. Mortalities of downstream migrant salmon Transactions of the American Fisheries Society 1 14: at McNav Dam. Transactions of the American 430-435. Fisheries Society 9058-72. Turbak, S. C., D. R. Reichle, and C. R. Shriner. 1981. Smith, K. E. H. 1960. Mortality tests-young salmon Analysis of environmental issues related to small- and gaspereau at Tusket River Power Dam, Yar- scale hydroelectric development. IV fish mortality mouth County, Nova Scotia. Canada Department resulting from turbine passage. Oak Ridge National of Fisheries. Halifax. Laboratory, Environmental Science Division, Pub- Smith, K. E. H. 1961. Mortality tests-yearling gas- lication 1597, Oak Ridge, Tennessee. pereau at Tusket River Power Dam, Yarmouth

i