Peak spawning for cod normally occurs between February and March, but the peak can vary from year to year (Smith, 1983). In warm to moderate winters, peak spawning can occur as early as December but, during unusually harsh winters, it can be delayed until the end of March (Smith et al., 1981). Optimum spawning and hatching temperatures are between 5-7oC (Heyerdahl and Livingstone, 1982). Haddock spawning occurs during late March or early April (Overholtz, 1987) although the onset and duration of spawning appears to be associated with increasing water temperature (Marak and Livingstone, 1970). On average, peak spawning probably occurs around the first week in April. Although spawning is widespread in shoaler waters, the northeastern part of Georges Bank is an important spawning area for both species (Figure 8). Eggs generally drift south and west in the clockwise gyre (Lough, 1984) and the larvae hatch in 2-3 weeks at typical early spring temperatures (Laurence and Rogers, 1976). Developmental rate is similar for both species as they grow through larval and juvenile stages (Fahay, 1983). The transition from pelagic to predominantly demersal life normally occurs by mid-summer, 3-4 months from hatching, when they attain a size of 4-6 cm (Lough et al., 1989). Juveniles typically are associated with pebble-gravel substrate which probably reduces predation and provides an abundance of epibenthic prey (Lough et al., 1989). Growth is rapid for the 0-group fish and sexual maturity commences at age 2 at approximately 40 cm (Overholtz, 1987; O'Brien, 1990). Year-class strength appears to be set by the end of their first year of life; however, the timing and causes of first-year mortality for a given year class are difficult to determine and consequently our understanding is limited (Fogarty et al., 1987).

The early life history of cod and haddock has been divided into six stages (Figures 7b, 8, 9) based on our knowledge of their timing and location on Georges Bank, vertical distribution, growth, principal prey, and selected environmental factors. The six stages are: (1) Egg, 1.5 mm diameter, (2) Early Larva, 2-8 mm, (3) Late Larva, 9-13 mm, (4) Early Pelagic Juvenile, 14-29 mm, (5) Late Pelagic Juvenile, 30-49 mm, and (6) Recently Settled Juvenile, 50-69 mm. Peak spawning for cod occurs about a month earlier than haddock, but overlaps with that of haddock. Haddock spawning is more discrete in time and space and their larvae develop when stratification begins and the bank gyre intensifies. The mortality rate of cod eggs has been reported to average 22%/day (Daan, 1981). Upon reaching Stage 6, cohort abundance has decreased about 4-5 orders of magnitude at an average mortality rate of 6-8%/day (Lough, 1984). Growth for both species has been described by Gompertz-type curves based on daily growth increments of otoliths (Bolz and Lough, 1988). Growth rate increases from 0.13 mm/day at hatching to about 1.0 mm/day at age 100 days. In well-mixed waters during early spring, eggs and larvae (stages 1-3) are broadly distributed throughout the water column with their vertical distribution centered at mid-depth (20-30 m ; Lough and Potter, in press). Late pelagic juveniles (stage 5) are located progressively deeper in the water column; near 40 mm, most juveniles are associated with the bottom. Larvae tend to be found deeper by day and shallower by night; the larger fish have a greater vertical range. Recently-settled juveniles (stage 6) remain on the bottom by day and migrate 3-5 m into the water column at night (Lough et al., 1989). While the diel vertical migration of cod appears to be strongly related to the light-dark cycle, haddock behavior can be more complex.

Egg, larval, and juvenile patches have been identified on Georges Bank. They have been tracked and studied as they were advected along the southern flank in a sheared flow field (Lough and Bolz, 1989). Some variable part of the population is recirculated around the western end of the Bank through the Great South Channel, but in unusual years, a large portion of the larvae may be transported westward into the Mid-Atlantic Bight (Polacheck et al., 1992). Nevertheless, Lough and Bolz (1989) found evidence for continuous recruitment of cod and haddock to the shoal central part of the Bank. Retention appears to be enhanced by residing near bottom in areas less than 70 m deep and interacting with cross-isobath currents. The shelf/Slope Water front intersects the bottom at about 80 m and Gulf Stream warm-core eddies moving near the southern edge can play an important role in the movement of water on and off the shelf, and the entrainment and transport of organisms (Lough, 1982; Joyce and Wiebe, 1983; Wroblewski and Cheney, 1984; Joyce et al., 1992). The southeast flank of Georges Bank is particularly vulnerable to advective exchange/loss during periods of strong southward wind stress (Walford, 1938; Cohen, et al., 1986). The generalized gyral pattern for the six life stages best fits haddock. The distribution pattern of cod is more widespread since their spawning is more protracted and earlier in the winter, when there is more wind mixing and advection, and the mean circulation gyre is weaker. Some surveys appear to show a resident larval cod population on eastern Georges Bank.

Water column stratification over the southern flank of Georges Bank has been shown by Buckley and Lough (1987) to have a significant influence on the feeding, growth, and survival of cod and haddock larvae. In May 1983, larvae and their prey were found concentrated in the thermocline region. A comparison of well-mixed versus stratified sites showed that recent growth based on RNA/DNA ratio analysis was higher at the stratified site. At the well-mixed site where prey biomass was lower, 50% of the haddock were in a starved condition. Cod collected at the same site were in better condition and growing faster than haddock. Cod larvae appear to be better adapted as a winter species when prey densities generally are lower; haddock larvae, better adapted to spring conditions, require higher prey densities which are concentrated by spring stratification (Lough, 1984). However, on eastern Georges Bank, Perry and Neilson (1988) found that in late June 1985, prey zooplankton biomass was lower at a stratified site compared to a well-mixed site. This may have been due to differential rates of local growth caused by different food environments in well-mixed vs. stratifed conditions.

The prey of larval cod and haddock consist of the developing zooplankton on Georges Bank. Yolk sac and first-feeding larvae (Stage 2) prey primarily on small plankton such as copepod nauplii, phytoplankton, and lamellibranch larvae (Kane, 1984; Buckley and Lough, 1987; Auditore et al., 1992). Both species of fish eat diatoms and Peridinium although these items represent a larger percentage of the diet of haddock than cod. With the exception of the yolk-sac stage, both cod and haddock feed on the same species of prey throughout their early life and select prey that are numerically dominant. Prey size plays an important role as larger size prey generally are selected as the fish grow larger. During the spring on Georges Bank, larvae (stages 2 and 3) prey on the increasing populations of copepod nauplii, copepodites, and adults of Calanus finmarchicus, Pseudocalanus spp., Oithona similis, and Centropages typicus. As pelagic and recently-settled juveniles (stages 4-6), they shift prey selection to epibenthic, swarming populations that also undergo diel vertical migrations such as mysids (Neomysis americana), amphipods (Gammarus annulatus, Themisto gaudichaudii, T. compressa), chaetognaths (Sagitta elegans), and euphausiids (Meganyctiphanes norvegica) (Auditore et al., 1988; Perry and Neilson, 1988; Lough et al., 1989).

Predation on pre-recruit stages can substantially reduce both absolute recruitment levels and the resilience of the population to exploitation (Fogarty et al., 1987). Its importance as a dominant source of pre-recruit mortality in many fish populations is now widely appreciated (Hunter, 1981; Sissenwine, 1984; Houde, 1987, 1989; Bailey and Houde, 1989). Further, predation can play a critical role in compensatory processes through several possible pathways including: (1) density-dependent growth coupled with size dependent predation (Beverton and Holt, 1957; Shepherd and Cushing, 1980; Werner and Gilliam, 1984; Miller et al., 1988; Anderson, 1988; Beyer, 1990); (2) density-dependent predation (i.e., predator switching and selective predation on abundant prey); and (3) directly through cannibalism. The Georges Bank fish community has undergone profound changes during the last three decades (Fogarty et al., in press). As the traditionally harvested species of principal groundfish and flounders declined sharply under intense exploitation, the abundance of other demersal species (small sharks and skates) have increased dramatically in the last decade. Spiney dogfish now may have replaced silver hake as the primary fish predator (Sissenwine and Cohen, 1991). Dogfish, silver hake, and squid also are known to prey on juvenile fish (Edwards and Bowman, 1979). Pelagic fish species, notably mackerel and herring, have increased in abundance so that Georges Bank is currently considered a predator-controlled system (Michaels, 1991; Fogarty et al., in press).

Recent studies have shown that a significant biomass of mackerel (and possibly herring) migrate along the southern flank of Georges Bank during late-April through early-May (the same time and place as identified in this document for intense study of advection and stratification). Significant mortality of larval cod and haddock may result from predation by mackerel at this time (Michaels, 1991). It is hypothesized that in years with warmer temperature anomalies on southern Georges Bank during spring, the mackerel migration could proceed across the southern portion of the Bank (rather than around the Bank as in cool years), resulting in a greater probability of encounter with the drift of cod and haddock larvae from the northeastern region of the Bank. Therefore, climatic warming could possibly result in the poor recruitment of Georges Bank cod and haddock due to shifts in the distribution and migrations of planktivorous fish.

There are not sufficient data to adequately identify and quantify egg and larval mortality due to predation, let alone partition mortality rates among the various predator groups on Georges Bank. Numerous studies document the consumption of fish eggs and larvae by a wide variety of invertebrates and fish. Hunter (1984) summarized mortality rates of pelagic eggs and concluded that 98% of cod eggs in the North Sea probably are consumed by predators before hatching. Möller (1984) provides evidence that the jellyfish Aurelia aurita caused a significant decline in the larval herring population of the Kiel Fjord. Fish eggs and larvae can comprise a large part of the natural diet of other passive predators such as medusae and ctenophores (especially Pleurobrachia) (Fraser, 1962; Purcell, 1985). More active invertebrate predators such as chaetognaths, large copepods, amphipods, isopods, and euphausiids may contribute to egg and larval losses. Older cod larvae and pelagic juveniles are likely predators of larval cod (Laurence et al., 1981). Hunter (1984) concluded that herring and other planktivorous fishes were major consumers of fish eggs and larvae.