AR-318

ICESJ.mar.Sci.,51:91 97. 1994 Differencesbetween marine and freshwaterfish larvae: implications for recruitment

E. D. Houde

Houde, E. D. Differences between marine and freshwater fish larvae: implications for recruitment. ICES J. mar.Sci., 51: 91-97.

Differences in the dynamics (growth and mortality) and energetics properties of marine and freshwater fish larvae have important implications fordetermining the lifestages at which year-class size is established. Aftercorrecting fortemperature effects,marine fish larvae, which typically weigh less at hatch, can be shown to experience higher mortality rates, have higher metabolic requirements, and have longer larval stage durations than do freshwater fish larvae. Growth rates and growth efficiencies are similar for the two categories of larvae. The difference in body size between typically small marine and typically large freshwaterfish larvae is an important factor affecting their dynamics and energetics. Predicted mean survivorship to metamorphosis of a cohort of freshwater fish larvae is 5.30%, while that of marine larvae is only 0.12%. The probability of significant density-dependent regulation during the larval stage could be relatively high for marine species, based upon their life-history properties. The probability that episodic mortalities of larvae will have significantimpacts on recruitment may be higher for freshwater fish than for marine fish.Starvation mortality is more probable formarine larvae, primarily because of their small body size, associated high metabolic demands, and possibly higher ingestion requirements. Because expected mean survivorship of freshwater larvae is 44 times higher than that of marine larvae, juvenile-stage dynamics will be relatively important in controlling regulating recruitment levels of freshwater fishes.In contrast, properties of marine species indicate that larval-stage dynamics will have a greater in0uence on recruitment success.

Key words: larval fish, dynamics, energetics, marine, freshwater, recruitment mechanisms.

Received 7 April 1993; accepted 27July 1993.

E. D. Houde: Chesapeake Biological laboratory, Center for Environmental and Estuarine Studies, The Unil'ersity of Maryland System. PO Box 38, Solomons, Maryland 20688, USA.

Introduction ontogenies vary widely (Balon, 1981). Development during the larval stage may be brief or protracted, result­ Research on fishearly life stages has proliferated in recent ing in either short- or long-stage durations. Because of the years, particularly on marine species. The high fecundities variability in morphologies and body sizes of larvae, as of fishesensure high larval abundances, which arc subject well as stage durations, it is not surprising that many to: (I) density-independent controls that are believed to biological properties differ substantially among species be a primary cause of recruitment variability; and (2) (Miller et al., 1988; Pepin, I 991; Pepin and Myers, 1991). possible density-dependent regulation, which will tend The vital rates (mortality rates, growth rates, stage to stabilize recruitment levels (Cushing, 1974a,b; durations) and energetics of larvae are strongly affected Rothschild, I 986). Growth rates, mortality rates, stage by temperature (Houde, 1989a; Morse, 1989). Tempera­ durations, and energetics properties of teleost larvae, ture and body size probably are the main variables affect­ when compared across taxa and ecosystems (Houde and ing egg and larval stage dynamics (Pepin, 1991) and Zastrow, in press), have many similarities, but they also prey consumption rates (Mackenzie et al., 1990). Their show significantdifferences that have implications for the effectson larval-stage durations may have had important recruitment process, as well as for development of fishery consequences in the evolution of teleost life-history modes research and management plans. (Houde, I989a). Fishes in both marine and freshwater habitats exhibit There are taxa-specific as well as ecosystem-specific diverse morphologies and behaviors. This diversity differences in vital rates and energetics properties of fish extends to the larval stages in which morphologies and larvae (Houde and Zastrow, in press) that could affect

1054 3139I 94 I 010091+07S08.00 0 © 1994 International Council forthe Exploration of theSea 92 £. D. Houde recruitment levels, in addition to influencing dynamics from marine or freshwater systems, acknowledging that and energetics of older life stages. Houde and Zastrow (in such characterization docs not recognize species which press) noted some probable differences in larval-stage arc atypical and whose life histories may deviate signifi­ properties of marine and freshwater fish larvae in a cantly from the norm. Differences between means were broader analysis across aquatic ecosystems and taxa. judged to be significant at a=0.10, rather than at 0.05. In the present paper. I critically compare averaged to increase the power of the tests to detect probable properties of larvae from the combined marine eco­ differences in these quite variable data. which often were systems with properties of larvae from freshwater based upon few observations. ecosystems and discuss the potential for regulation and Regression statistics relating G, z. D. and QO 2 to W0 control of recruitment during the larval and juvenile were calculated to dctermmc if these variables were de­ stages in each system. pendent upon weight at hatching. and if the dependence differedbetween freshwater and marine larvae. Expected numbers of metamorphosed survivors. at mean values of Methods vital rates and ontogcnic attributes. were calculated for fishes from freshwater and marine ecosystems, based Data on larval dynamics (vital rates) and energetics, upon the relationship: categorized by ecosystems and by taxonomic groups. -Z(log. •• -log. W )] _ wm 0 were compiled by Houde and Zastrow (1991, in press). Nm.,-Px N0e [---- ()) Those reports included data on both freshwater and G = marine tclcost larvae. but the analysis and interpretations where N m., number of metamorphosed survivors: did not focus on implications for recruitment in fresh­ N0 = initial number of larvae; and P= probability of water and marine ecosystems. Herc, the analysis is surviving "episodic" mortality events. extended explicitly to examine and compare the relation­ The ratio of N m•• rw=Nm••·MAR for equal N0 was calcu­ ships between size at hatching and vital rate parameters lated to indicate relative stage-specific survivorship. By in the two categories of larvae. Recruitment levels in varying Z. G, and Pin Equation (I). the changes in rates each case arc predicted for mean values of rates estimated or episodic loss probabilities were calculated that are during larval life. and simple simulations arc run to required to cquali,c survivorship at metamorphosis in the demonstrate the extent to which average dynamics in the marine and freshwater ecosystems. Effects of initial egg larval stage must vary to precipitate important changes in or larval numbers (i.e. variability in N 0), caused b) recruitment levels. differencesin mean fecundities between marine and fresh­ Most tcleost taxa for which published data were avail­ water fishes, and the consequences for survivorship at able arc included in the analysis. Exceptions were a few metamorphosis, were considered. Telcost fecundity data taxa of principally freshwater fisheswith unusually large. compiled by Winemiller and Rose (1992) were used for precocial (se11.1u Balon, 1981) hatchlings that have either this purpose. a curtailed or absent larval stage. e.g. Salmonidac, The potential for density-dependent regulation of Acipenseridac, lctaluridae. The variables that were con­ larval survival via growth-rate responses to initial larval = = sidered are: W0 dry weight at hatching (µg): W mei dry abundances was compared for marine and freshwaterfish = weight at metamorphosis (µg): G weight-specific larvae by simulating responses of N m., in Equation (I) to 1 = growth coefficient (µg µg ' day- ): Z instantaneous variability in Ci and N 0• In an example, Ci was reduced by 1 ): = mortality coefficient (day- D larval-stage duration 10, 25, or 50% per five-fold increase in N0• Then, the = (day): 1.c. days to grow from WO to W me,; QO2 weight­ expected survivorships were compared for marine and 1 1 specific oxygen uptake (µI 02 mg- h- ); I= weight­ freshwater larvae to determine their sensitivities to specific food mgcstion (day 1), i.e. fraction of larval body density-dependent growth. = weight consumed daily; and K 1 gross growth efficiency, definedas GI I. Temperature-adjusted mean values of each parameter Results for larvae from five marine ecosystems and freshwater. and from nine broad taxonomic groups. were estimated Marine fish larvae. on average. are smaller at hatch. (Houde and Zastrow. 1991, in press). Adjustments forthe suffer higher mortality rates, have higher weight-specific substantial temperature effects were accomplished via respiration rates, and have longer stage durations than do analysis of covariance. Mean values that arc reported here freshwater larvae (Table I). The I 0-fold difference in for data from pooled marine ecosystems and from frcsh­ mean weight at hatching potentially has important conse­ w�tcr systems are those compiled by Houde and Zastrow quences for larval-stage dynamics and recruitment. (in press). except for minor differences resulting from a Weight-specific growth rates do not differ significantly. few additions or changes to the earlier data set. These nor do gross growth efficiencies, although the data­ values arc referred to as "typical" or "average" for larvae base for the latter is small. The calculated mean. Differenceshet,reen marine and/reslt11·ater/i.1ltlarvae 93

Table I. Temperature-adjusted (by analysis of covariance) means of dynamic and energetics properties of marine and freshwater fish lar,ac. Symbols for variahles arc defined 111 text (sec Methods). Significance 111dicates p :!,'.'Q.10. n number of taxa. s.r. -standard error of the mean.

Manne Fre,h,\atcr

Variahle X S.F n X s.r. n S1gn1ficancc

w. 37.6 6.4 77 359.7 72.8 20 s1g. wmcl 10 845.6 953.0 77 9'176.9 1604.3 16 11.S. G 0.200 0.01 I 80 0.177 0.019 26 n.s. z 0.239 0.021 26 0.160 0.040 7 s1g. D 36.1 I I 79 20 7 I I 15 s1g. 002 5.90 0.38 15 2.79 0.38 15 s1g. K, 0.29 0.03 14 0.32 0.03 8 n.s. I 0.57 O.Q7 14 0.46 0.09 8 11.S.

I 1 Table 2. Expected energy budgets (cal mg day ) of marine and fresh,,atcr fhh larvae. (a) Temperature-adjusted mean energy budgets (from Houde and Zastr

Category Ingestion growth + metabolism + unne + feces

(a) Manne 3.44 LOO + 0.98 + 0.24 + 1.21 Freshwater 2.78 0.89 + 0.47 + 0.19 + 1.23 (b) Manne 3.44 1.00 + 0.98 + 0.24 + 1.21 Freshwater 3.13 1.00 + 0.53 + 0.22 + 1.39 (c) Marine 1.00 0.29 + 0.29 + 0.07 + 0.15 Freshwater 1.00 0.32 + 0.17 + 0.07 + 0.44

weight-specific ingestion rates. which were derived from Applying mean \'alucs (Table I) m Equation (I) indi­ the growth rate and growth efficiency data. are 0.57 for cates that relative survivorship at metamorphosis of a marine larvae and 0.46 for freshwater larvae. These "typical" cohort of freshwater fish is expected to be 44 values were not found to differsignificantly at the a= 0.10 times higher than that of a "typical" cohort of marine level. fish. At metamorphosis. 5.30 °-o of an initial freshwater Weight-specific growth rates (G) for the pooled data cohort is expected to be alive compared to only 0.12% of a of marine and freshwater larvae. and of the freshwater marine cohort. Thus. a much larger fraction of a typical larvae alone. arc significantly correlated with W0 (Fig. freshwater, larval cohort will survive to enter the juvenile Ia). but no relationship between G and WO is apparent for stage. the marine larvae over the relatively small range of W 0 Quite large changes in either Z or G are necessary to that is represented. No significant relationship between equalize expected survivorships at metamorphosis of instantaneous mortality rate (Z) and W0 was demon­ marine and fresh\\aterfish larvae. A 2.3-fold increasein Z strated (Fig. lb). although the highest rates arc reported or decrease in G by a typical cohort of freshwater larvae is for some of the smallest marine larvae. There is also no necessary to reduce its sur\ival at metamorphosis to significant correlation between larval-stage duration (D) 0.12 °10, the estimated survivorship of a typical cohort and W0 (Fig. le), but the longest-stage durations were of marine larvae. Exactly the opposite applies to marine estimated for some of the smallest mannc larvae. The larvae; a typical cohort of marine larvae would need to quite strong relationship between weight-specific oxygen experience 2.3-fold lower Z or higher G to support ° uptake (QO 2) and W0 • which was reported previously for survivorship of 5.30 10 at metamorphosis. the expected combined �annc and freshwater larvae by Houde and level for a typical freshwater cohort. Zastrow (in press) 1s illustrated (Fig. Id). In the present If episodic mortalities of larvae i.e. weather-related analysis. the QO 2 vs. WO relationship was significant for events. arc more prevalent in freshwater than in marine the freshwater larvae. but not for the marine larvae. environments, such losses would tend to equalize 94 E. D. Houde

1 reductions inG are required to generate a similar response - J 0.8 • (a) in freshwater larvae. The result indicates that survival 0.6 • � of marine fish at metamorphosis not only is relatively

ay

d 0.4 more sensitive to external controlling factors that affect

G< .. c,I> mortality rates and stage durations, but that marine

0.2 '4

61

~

.. 0 species also have more potential to regulate their 0.6 abundance in the larval stage. (b) The expected energy budgets formarine and freshwater

)

1

- 0.4 fish larvae, calculated from the temperature-adjusted mean growth rates, ingestion rates, and weight-specific

day ( 0.2 �- oxygen uptakes, arc similar in most respects (Table 2a).

Z �6- . After standardizing budgets to one unit of growth (Table

I~~ 0 �--� LJ 2b), the budgets differ principally in the metabolism 200 I component, which is nearly two times higher for the

E 150 .. smaller marine larvae. Budgets standardized to one unit 100 . of ingestion (Table 2c) show that metabolism is 29% of : ingested energy for marine larvae but is only 17% for

D

50 -,. . . . freshwater larvae. Average assimilation efficiencies

>

~~ 0 " [(Ingestion - Feces)-:- Ingestion] apparently are higher :;- 12 for marine (65%) than for freshwater fish larvae I (d>l ,.c:: (56%), primarily because marine larvae have a higher

1

- 2 8 •• metabolic level, not because they grow faster or more

mg . Q0 "· efficiently.

2 4 • t •

0 • 4 "' • • • •

µI

J.

( 0 200 400 600 800 1000 1200 1400 Discussion Weight at hatch (µ.g dry) It has been presumed since the important contribution by Figure I. The relationships between dry weight at hatching and (a) weight-specific growth rate, (b) instantaneous mortality rate. Hjort (1914) that dynamics in the early larval stage play a (c) larval stage duration, and (d) weight-specific oxygen uptake dominant role in controlling or regulating year-class for marine and freshwater fish larvae. strengths of marine fishes. Numerous studies of concep­ tual, theoretical, and applied nature have supported this assumption (Jones, 1973; Jones and Hall, 1974; Cushing, expected survivorships at metamorphosis in the two sys­ 1974a, 1990; Lasker, I 978; Beyer, 1989; Heath, 1992). tems. But, at the mean Zand G levels of freshwater fish Less emphasis on larval-stage dynamics is apparent in the larvae (Table I), additional episodic losses [P in Equation literature on freshwater fishes. Results of the present (I)] would have to reduce initial cohort abundance by analysis suggest that larval-stage dynamics generally will 97.8% to match the expected survivorship of a typical have less influence on control or regulation of year-class cohort of marine larvae. If episodic mortalities, e.g. strengths in freshwater fishes, and that trophic inter­ caused by aberrant transport or storm events, are an actions, competition for resources, and such factors as important element contributing to mortality of marine overwinter mortality during the juvenile stage, may be the fishlarvae, any such losses would tend to reduce further predominant factors affecting their recruitment levels their expected survivorship at metamorphosis, and thus (Forney, 1976; Mills et al., 1987; Post and Evans, 1989; widen the gap even more between expected survivorships Shuter et al., 1989). of freshwater and marine larval cohorts. The relatively large size al hatching, short-stage Decreases in growth rate, if density-dependent, poten­ duration, and low mortality rate of freshwater fish larvae tially could have relatively strong effects of survivorship favor the probability that recruitment level will be con­ at metamorphosis of marine fish larvae compared to trolled during the juvenile stage. In contrast, marine fish freshwater larvae (Fig. 2). The potential for regulation hatch at relatively small size, have longer-stage durations, of abundance of marine larvae via density-dependent and suffer higher larval mortality rates, indicating a growth is relatively high because their stage durations are higher probability that recruitment levels are controlled

long, their ratio ofG:Zis low, and their ratio ofWme1:W0 primarily in the larval stage. Minor variability in larval is high. Reductions in weight-specific growth rate of growth and mortality rates arc likely to register bigger < 25% per five-fold increase in initial larval abundance effects on recruitment potential of marine fishes because, will induce a strong density-dependent response in rela­ on average, the larval stage duration is 15 days longer tive survival of marine larvae. But, considerably greater in marine species (Table 1). Mean survival through the Differences between marine andfreslnmterfish /arrae 95

5 Marine larvae Freshwater larvae

·;;;00 ..- ··- ··- ··- ··- ··- ··- ··-·· 0 .c 0 ~------~ e- ..... ------0 ' --- - E ' ' -' ...,.,al ' ' ' E ' ' ..., -5 ' ' al ' ' ta ' \ \ > ' ' ·;: ... '\ ::, \ .,"' -10 \ > ·.:; \ _ .,_ .,_ " 0'1 al \ di... \ ---10€7,; ~ \ b.ll - 15 \ ------25q. .3 \ \ --- 50CJ!- \ I -2O'-----L---'---'---'----'-- -'~- - ~-~ - -~- -'--~-~ 1 3 10 30 100 300 1 3 10 30 100 300 Relative No Figure2. The effect of density-dependent growth on relative recruitment potential, indexed as relative survivorship at metamorphosis formarine and freshwater fish larvae. In the sim ulations. weight-specific growth rates were allowed to remain constant (0% decline) or were reduced by IO. 25, and 50°0 per five-fold increase in initial larval numbers.

larval stage is only 0.12% for a typical cohort of ma rine "clutch sizes" and "spawning bouts" in fishes. Their data fish, but is 5.30% for a typical cohort of freshwater fish, indicate that marine fishes. on average, arc approximately suggesting that substantially less scope for control and 11 times more fecund than freshwater fi shes. Data illus­ regulation of abundance may exist in the juvenile stage of trated in Duarte and Alcaraz ( 1989) also suggest at least marine fishes. I 0-fold higher fecundity, on average, in marine fishes. Larvae of teleost fishes arc as diverse as arc adult Even with such a large difference in fecundity, expected telcosts, presenting a spectrum of larva l sizes, morpholo­ survival per adult spawner at the end of the larval stage gies, and ontogenies. But, freshwater species that have [calculated from Equation ( I) with N set at 11 and I been studied have larvae which hatc 0 h at dry weights for marine and freshwater species, respectively] still is exceeding 100 µg (appendix A in Houde and Zastrow, in four times higher for freshwater fishes. The tentative con­ press), while relatively few ma rine larvae hatch at more clusion remains that control of recruitment during the than 100 µg. The difference in initial sizes between juvenile s tage is more probable for freshwater than for freshwater and marine larvae is in part attributable to marine fishes. spawning modes. Virtually all freshwater species are The differences in populatio n dynamics and energetics demersal spawners, producing r elatively few but large properties ofmarine and freshwater fish larvae are partly eggs, while marine species usua lly arc pelagic s pawncrs, attributable to their differences in body size. Peterson a nd producing many but small eggs (Duarte and Alcaraz, Wroblewski (1984) a nd MeGurk ( 1986) demonstrated 1989). If freshwater fi shes that were excluded from this tha t m o rtality rates of marine organisms, including fish analysis, e.g. Salmonidae, Acipenseridac, and lctaluridae, eggs and larvae, decline as size increases. Miller et al. which have large and precocious larvae, had been in­ (1988) and Werner and Gilliam (1984) indicated that cluded, the differences that were demonstrated between many ontogencticand behavioral propertiesoffish larvae freshwater and marine species (Table I) would have been are strongly related to body size. They suggested, and magnified. Beyer ( I 989) argued, that size-dependent mortality and If survival through the larval stage were inversely growth are the decisive factors tha t control the recruit­ proportional to fecundity, then the differences in larval ment process in fishes. Neither Pepin (199 1) nor Houde stage survivorship between marine and freshwater ( 1990) was able to demonstrate that size at hatch of species might not be significant o n a "per adult spawner" marine fish larvae had a significant effect on cumulative or basis. Winemiller and Rose ( 1992) reviewed data on instantaneous mortality during the larval stage, but 96 £. D. Houde neither study included freshwater fish larvae, which Regulation of larval stage abundances, via density­ generally are larger than larvae of marine fishes. Pepin dependent effects on growth and mortality, is proposed to and Myers (199 1) did show that recruitment variability be more likely for marine fish larvae than for freshwater in marine fis h stocks is positively correlated with the dif­ larvae. The longer-stage durations of marine fish larvae ference in size between hatching and metamorphosis imply that even small reductions in growth rates in re­ (1'1L), which they argued was a measure ofstage duration. sponseto increasing larval densities could havea powerful This result, if extended to include freshwater fishes, regulatory effect on recruitment (Ware, 1975; Shepherd suggests that recruitment variability in freshwaterspecies, and Cushing, 1980). The timeframe over which that effect which have relatively small 1'1L and short larval-stage operates will determine the magnitude of its impact. durations, should be lower than in marine species. This Houde (1989b) simulated effects of density-dependent hypothesis remains to be tested. growth on some species of marine and estuarine fish Starvation mortality is predicted to be more likely larvae and found that the biggest effects on recruitment for marine than for freshwater fish larvae when food level were for species with high W me, to WO ratios and low resources are scarce. Ifthe prediction is correct, the prob­ G to Z ratios. In the present analysis, the mean Wme,:W0 ability of death by starvation may be highest for marine for marine fish larvae is 288.4, while that for freshwater larvae that develop at high temperatures (Houde, 1989a). larvae is only 25.8. The mean G:Z for marine fish larvae is Mean weight-specific ingestion, while not demonstrated 0.83, while that for freshwater larvae is 1.12. On average, to be significantly lowerfor freshwater larvae in my analy­ cohorts of marine fish larvae lose biomass during the sis, may in fact be lower (Tables I, 2). Higher required larval stage (G/Z< 1.0), suggesting that they are poten­ ingestion by marine larvae is expected because higher tially more limited by food resources than are freshwater metabolicdemands are imposed by their small body sizes, larvae. In contrast, freshwater larval cohorts accumulate and because neither their weight-specific growth rate nor biomass (G/Z> 1.0) relatively fast during a short-stage gross growth efficiency differed from those characters in duration, which may restrict the potential for density­ freshwater larvae. dependent regulation in the larval stage. Houde and Zastrow (in press) found that mean weight­ Because the relative numbers (per egg or per female) specific oxygen uptake (Q02 ) of marine fish larvae was surviving to enter the juvenile stage arc higher in fresh­ nearly twice that of freshwater larvae. The regression water fishes, their potential for significant adjustment of slopes ofthe relationships between Q02 and temperature year-class si,:es during the juvenile stage is relatively were the same for the marine and freshwater larvae, but great. On average, the relative abundances of freshwater the elevation of the regression was twice as high for the juveniles will be higher and their stage durations longer freshwater species. There also was a strong relationship than those of marine species. Consequently, either coarse betweenQ02 and weight at hatch (Houdeand Zastrow, in or fine controls that arc density-independent. or subtle press) which indicated that freshwater larvae had lower density-dependent regulation acting on freshwater juven­

Q02 because of their relatively large sizes. iles, will have a relatively big impact on recruitment levels. Episodic mortalities of early life stages, often associ­ This prediction docs not exclude the possibility ofsignifi­ ated with weather events, arc more likely to be significant cant control and regulation during the juvenile stage of for freshwater fish. Freshwater nursery habitats generally marine fishes (Sissenwine, 1984), but it suggests that are relatively small, often shallow, and vulnerable to juvenile-stage dynamics will be relatively more important sudden changes in temperature, water levels, pH, oxygen in determining recruitment levels and variability of level, and chemical constituency. Marine nurseries, in a freshwater fishes, while larval-stage dynamics will be relative sense, are less vulnerable to events that may more important in marine species. promote episodic mortalities, except for unfavorable advection which may cause losses ofeggs and la rvae. The Acknowledgments impact oflarge episodic mortalities on recruitment level is often lower than effects due to modest changes in daily Thanks are extended to C. E. Zastrow for assistance in growth or mortality rates (Houde, 1989b). The shorter­ data compilation and graphics. D. H . Secor provided stage durations of freshwater fish larvae, while reducing comments on an early draft of the manuscript. I also the temporal window within which episodes can occur, thank two critical reviewers ofthe submitted manuscript, perhaps more importantly reduce the period of time over L. Crowder and W. Nellen, who were often skeptical of which vital rates may vary to affect survivorship. A the predictions and conclusions in this paper, and who contrasting situation prevails in marine nurseries, i.e. the forced me to be more careful and critical in the revised te_mporal window within which a low-probability episode version. Research was supported in part by the US might occur increases, but the longer-stage durations of National Science Foundation, Biological Oceanography marine fish larvae magnify the probability that control or Program, Grants OCE87-01304 and OCE92-03307. Con­ regulation will act through modest variability in vital tribution no. 25 12 of the Center for Environmental and rates. Estuarine Studies of the University of Maryland System. Differences between marine andfreshwater fish larvae 97

References Mackenzie, B. R., Leggett, W. C., and Peters. R.H. 1990. Esti• mating larval fish ingestion rates: Can laboratory derived Balon. E. K. 1981. Saltatory processes and altricial to precocial values be reliably extrapolated to the wild? Marine Ecology forms in the ontogeny of fishes. American Zoologist, 21: Progress Series, 67: 209 225. 573-596. McGurk. M. D. 1986. Natural mortality of marine pelagic fish Beyer, J.E. 1989. Recruitment stability and survival simple size• eggs and larvae: Role of spatial patchiness. Marine Ecology specific theory with examples from the early life dynamics of Progress Series, 34: 227 242. marine fish. Dana, 7: 45-147. Miller, T. J .. Crowder, L. B., Rice, J. A., and Marschall, E. A. Cushing, D. H. 1974a. The natural regulation offish populations. I 988. Larval size and recruitment mechanisms in fishes: 111 Sea fisheries research, pp. 399-412. Ed. by F. R. Harden• Toward a conceptual framework. Canadian Journal of Jones. J. Wiley & Sons, New York. 510 pp. Fisheries and Aquatic Sciences, 45: 1657 1670. Cushing, D. H. 1974b. The possible density-dependence of larval Mills, E. L.. Forney, J. L., and Wagner, K. J. 1987. Fish mortality and adult mortality in fishes. In The early life history predation and its cascading effect on the Oneida Lake food offish, pp. 103 111. Ed. by J. H. S. Blaxter. Springer-Verlag, chain. In Predation. Direct and indirect impacts on aquatic New York, Heidelberg, Berlin. 765 pp. communities, pp. 118 131. Ed. by W. C. Kerfoot and A. Sih. Cushing, D. H. 1990. Plankton production and year-class University Press of New England, Hanover and London. 386 strength in fish populations: an update of the match/mismatch pp. hypothesis. Advances in Marine Biology, 26: 250 293. Morse, W.W. I989. Catchability, growth and mortality oflarval Duarte, C. M., and Alcaraz, M. 1989. To produce many small or fishes. Fishery Bulletin, US, 87: 417-446. few large eggs: a size-independent reproductive tactic of fish. Pepin, P. I 99 I. The effect of temperature and size on develop• Oecologia, 80: 401-404. ment, mortality and survival rates of the pelagic early life Forney. J. L. 1976. Year-class formation in the walleye history stages of marine fish. Canadian Journal of Fisheries (Sti:ostedion vitreum vitreum) population of Oneida Lake, and Aquatic Sciences, 48: 503 518. New York, 1966--73. Journal of the Fisheries Research Board Pepin, P., and Myers, R. A. 1991. Significance of egg and larval of Canada, 33: 783 792. size to recruitment variability of temperate marine fish. I lea th, M. R. 1992. Field investigations of the early life stages of Canadian Journal of Fisheries and Aquatic Sciences. 48: marine fish. Advances in Marine Biology, 28: 1- 173. 1820 1828. Hjort, J. 1914. Fluctuations in the great fisheries of northern Peterson, I., and Wroblewski. J. S. 1984. Mortality rate of fishes Europe. Rapports et Proces-Verbaux des Reunions du Conseil in the pelagic ecosystem. Canadian Journal of Fisheries and International pour !'Exploration de la Mer, 20: I 228. Aquatic Sciences. 41: 1117 1120. Houde, E. D. 1989a. Comparative growth, mortality and Post, J . R., and Evans, D. 0. 1989. Size-dependent over• energetics of marine fish larvae. Fishery Bulletin, US, 87: winter mortality of young-of-the-year yellow perch (Perea 471-495. jlavescens): laboratory. in situ enclosure, and field exper• Houde, E. D. I989b. Subtleties and episodes in the early life of iments. Canadian Journal of Fisheries and Aquatic Sciences, fishes. Journal of Fish Biology, 35 (Suppl. A): 29- 38. 46: 1958 1968. Houde. E. D. 1990. Temperature-dependent and size-dependent Rothschild, B. J. 1986. Dynamics of marine populations. variability in vital rates of marine fish larvae. ICES CM Harvard University Press, Cambridge, Massachusetts. 277 pp. 1990. L:3, 15 pp. Shepherd, J. G., and Cushing, D. H. 1980. A mechanism fo r Houde, E. D., and Zastrow, C. E. 1991. Ecosystem-specific and density-dependent survival of larval fish as the basis of a stock taxa-specific growth, mortality and energetics relationship of recruitment relationship. Journal du Conseil International teleost larvae. ICES CM 1991 / L: 35, 11 pp. pour !'Exploration de la Mer, 39: 160-167. Houde, E. D., and Zastrow, C. E. Ecosystem• and taxa-specific Shuter, B. J., lhssen. P. E., Wales, D. L., and Snucins, E. J. 1989. dynamic and energetics properties of fish larvae assemblages. The effects of temperature, pH and water hardness on winter Bull. Mar. Sci .• 53. starvation of young-of-the-year smallmouth bass, Micropterus Jones, R. I973. Density dependent regulation of the numbers of dolomieui Lacepede. Journal of Fish Biology, 35: 765 780. cod and haddock. Rapports et Proces-Verbaux des Reunions Sissenwine, M. P. 1984. Why do fish populations vary? In du Conseil International pour !'Exploration de la Mer, 164: Exploitation of marine communities, pp. 59 94. Ed. by R. M. 156 173. May. Springer-Verlag, Berlin. 366 pp. Jones. R .• and Hall. W. B. 1974. Some observations on the Ware, D. M. 1975. Relation between egg size, growth, and population dynamics of the larval stage in the common natural mortality of larval fish. Journal of the Fisheries gadoids. In The early life history of fish, pp. 87 102. Ed. by Research Board of Canada. 32: 2503- 2512. J. H. S. Blaxter. Springer-Verlag, New York Heidelberg Werner, E. E., and Gilliam, J . F. 1984. The ontogenetic niche and Berlin. 765 pp. species interactions in size-structured populations. Annual Lasker, R. 1978. The relation between oceanographic conditions Review of Ecology and Systematics, 15: 393-425. and larval anchovy food in the California Current: identifi­ Winemiller, K. 0., and Rose, K. A. 1992. Patterns of life-history cation of factors contributing to recruitment failure. Rapports diversification in North American fishes: implications for et Proces-Verbaux des Reunions du Conseil International population regulation. Canadian Journal of Fisheries and pour !'Exploration de la Mer. 173: 212 230. Aquatic Sciences, 49: 2196 22 18.

View publication stats