HOUDE: VITAL RATES. AND ENERGETICS OF MARINE FISH LARVAE 



vation that mortality rates potentially can be 

 derived from relatively easy to obtain informa- 

 tion on larval growth rate and its variability 

 (Houde 1987). Before such estimates are possi- 

 ble, it will be necessary to explicitly determine 

 the gi'owth rates and mortality rates of many 

 species to estabhsh how reliable this approach 

 might be. 



Differences in spawning strategies of marine 

 fishes may have evolved as a consequence of the 

 different constraints on growth and survival of 

 larvae from high and low latitudes. The con- 

 straints for low-latitude larvae, i.e., high mortal- 

 ity rates, high gi'owth rates with attendant short 

 stage duration, and required high ingestion, sug- 

 gest that few larval cohorts will find the neces- 

 sary local conditions conducive for grow1;h and 

 survival. In most cases cohorts will starve or be 

 eaten. It is hypothesized that protracted spawn- 

 ing, serial spawning and frequent batch produc- 

 tion of eggs, a common strategy in the tropics, 

 will insure that some larval cohorts hatch during 

 those brief periods when conditions favor the 

 high gi'owth rates that promote survival. In high 

 latitudes, where spawning often is temporally 

 and spatially confined, larvae have different con- 

 straints. There, both mortality and gi'owth rates 

 tend to be low, ingestion is relatively low, but 

 stage duration is long and potentially very vari- 

 able. Under such circumstances small changes in 

 either mortality rates or growth rates can have 

 major impacts on recruitment potential 

 (Shepherd and Gushing 1980; Houde 1987). And, 

 long larval stage durations also provide ample 

 time for density-dependent mechanisms to de- 

 velop which may regulate abundance and 

 dampen fluctuations in stocks that originated 

 from one or a few batch spawnings that occurred 

 during a brief time. Under these conditions the 

 timing of spawning (Gushing 1975) and the selec- 

 tion of favorable spawning sites (lies and Sinclair 

 1982) by adults are critical to the recruitment 

 success of a cohort. 



Although density-dependent regulation in 

 early life often is assumed, there is relatively 

 httle evidence that it does in fact play a major 

 role in the egg and larval stage. Jones (1973). 

 Gushing and Harris (1973), Ware (1975), and 

 Shepherd and Gushing (1980) have modeled the 

 recruitment process, demonstrating how den- 

 sity-dependent mortality and/or gi"owth can reg- 

 ulate abundance. They have argued that regula- 

 tion may be most effective in the larval stage. In 

 support of those arguments, Savoy and Grecco 

 (1988) have demonstrated that density-depen- 



dent mortality during the egg and larval stage 

 may play a significant role in the regulation of 

 anadromous American shad, Alosa sapidissi)7ia, 

 populations. Based on analyses of larval life his- 

 tory characteristics reported here, if density- 

 dependent regulation is significant, it seems 

 more hkely to be effective in high latitudes than 

 in tropical seas. The long stage duration and its 

 potential variability, caused by varying tem- 

 perature or food availability, may promote com- 

 petition or allow predators to aggi'egate, favor- 

 ing density-dependent control. In contrast, 

 because larval stage durations in tropical seas 

 are short and less variable, the probabilities of 

 competitive or predator-mediated, density- 

 dependent effects seem less likely. 



The production of multiple cohorts during pro- 

 tracted spawning by tropical fishes and by many 

 summer spawners in higher latitudes is a bet- 

 hedging strategy that will allow some daily- 

 produced cohorts to experience conditions favor- 

 able for survival. Lambert (1984) and Lambert 

 and Ware (1984) have proposed that single 

 batch, demersal spawners in high latitudes (e.g., 

 herring and capelin, Mallotus villosus) are more 

 likely to produce easily discernible cohorts of 

 larvae than are summer-spawning pelagic 

 species (e.g., Atlantic mackerel, Scomber scom- 

 brus, and white hake, Urophycis tenuis) in the 

 same region because the demersal spawners are 

 characterized by waves of females that deposit 

 eggs at discrete time intervals. They argued 

 that, in the cases of hening and capeHn, widely 

 spaced cohort production represented bet-hedg- 

 ing by reducing potential intraspecific competi- 

 tion among larvae and by assuring that cohorts 

 of prey would develop with cohorts of fish larvae 

 (Jones 1973; Jones and Hall 1974). Lambert and 

 Ware (1984) believed that Atlantic mackerel and 

 white hake females spawned every 2-3 days and 

 that spawning in such species would appear to be 

 continuous during their summer spawning 

 season. In agi'eement with that argument, it is 

 hypothesized here that teleost stocks spawning 

 at high temperatures are more likely to produce 

 daily cohorts of eggs than are those stocks 

 spawning at low temperatures because the 

 larvae of warm-water stocks have short stage 

 durations and are constrained by the necessity 

 for high growth and by their high mortality 

 rates. This point is supported by Lambert and 

 Ware's (1984) figure 4, in which they show that 

 when larval gi'owth rates are high, as they are in 

 the tropics and in many high-latitude summer 

 spawners, the predicted time period between 



489 



