68 



Fishery Bulletin 100(1) 



August 31 to September 11. Individuals collected by beam 

 trawl on the same day (23 September 1991) had hatched 

 about 2 weeks earlier (from 24 July to 30 August) than the 

 above larvae (Fig. 9). These juveniles appeared to settle 

 as young as 24 days after hatching and at sizes as small 

 as 8.5 mm SL (Table 2). The total hatchmg date distribu- 

 tion for both larvae and newly settled juveniles collected 

 on September 23 reflected a spawning period that ranged 

 from late July to early September and that peaked in late 

 August and early September. 



Settled juveniles with a similar hatching date distribu- 

 tion were identifiable one month later at stations near 

 Beach Haven Ridge, but not at stations farther offshore 

 (Fig. 9, B and C). Fish collected near Beach Haven Ridge 

 on 21 October 1991 had a hatching date distribution with 

 a mode from late August through early September and 

 the overall distribution was skewed to the left. This period 

 was similar to the hatching date distributions for larvae 

 and newly settled fish collected on 23 September 1991. In 

 contrast, fish collected from offshore stations (i.e. stations 

 C and E) on 22 October 1991 were 2-4 weeks older and 

 5-10 mm larger on average (Table 2, Fig. 9). 



Plots of P. carolinus size versus age did not indicate 

 any abrupt change at settlement, specifically for postflex- 

 ion lai^vae and settled juveniles collected on 23 Septem- 

 ber 1991 (Fig. 10). Growth rates for this September collec- 

 tion fitted a hnear model (SL=3. 24-1-0. 229[age|; r2=0.77). 

 Because Prionotus lai-vae hatch at about 3 mm SL (Yus- 

 chak, 1985), this model's y-intercept is biologically realis- 

 tic. Growth rates offish collected in October did not differ 

 significantly between stations (ANCOVA:p7-o6.^, .^=0.13, 

 pro6.,,,,p ,=0.51); therefore the data were pooled. Linear, 

 least squares regression of all data produced an unreal- 

 istic y-intercept (SL=-7.01-H0.382lage]: SE^=2.0; ;--=0.74). 

 This model was rerun after restricting the y-intercept to 

 3 mm and the resulting equation indicated that age-0 P. 

 ca/'olinus continued to grow at about 1 mm every 4 days 

 (SL=3 +0:25l[age\: ;-=0.65) as they had during the lai-val 

 and settlement period. 



Size and age of P. caro/i>H^s juveniles varied significantly 

 along a 12-km transect ( 12-20 m depths; Fig. 1 ). The linear 

 relationship: Hatching age = 17.8 -i- 3.43 x depth; r-=0.35, 

 P<0.01; ri=69) showed that for every two meters change 

 in depth offshore the fish collected were about one week 

 older on average (Fig. 11). Sampling in both 1991 and 1992 

 showed a consistent trend for larger (and presumably old- 

 er) fish to be collected in deeper water in October and No- 

 vember (Fig. 12). After accounting for the effects of depth, 

 or possibly the distance from shore, it appeared that fish 

 reached a larger size in October of 1991 than in 1992 or that 

 larger fish in 1992 were not found in the sampling area. 



Discussion 



Spawning grounds and seasonality of spawning 



Prionotus caro/inus are more abundant than P. evolans 

 in continental shelf habitats whether they are measured 

 as eggs, larvae, juveniles, or adults (Keirans et al., 1986; 



1601 



120- 



8,0 



4.0 



00 



Prionotus carolinus 



Age-0 

 Age-1 + 



1 Jul 



1 Oct 



Figure 6 



Density (geometric mean number of fish per 

 100 m- |±1 standard en-or | ) of different cohorts 

 of postsettlcnient Prionotus carolinus and P. 

 evolans during daylight tows at the landward 

 and seaward stations near Beach Haven Ridge. 

 Note scale differences for each species. 



McBride and Able, 1994; Able and Fahay, 1998; McBride 

 et al., 1998; our present study). The low numbers of P. evo- 

 lans observed in our study may be biased somewhat by our 

 focused effort to sample the continental shelf rather than 

 estuaries. Prionotus evolans reside in shallower, warmer 

 habitats than do P. carolinus during the spawning season 

 (McBride and Able, 1994). If P. evolans spawn to some 

 degi'ee in shallower waters or estuarine habitats, then this 

 would at least partly explain the generally low abundance 

 of P. evolans early life stages in our collections. 



In general, we expect that larval distributions are good 

 predictors of spawning locations for both Prionotus spe- 

 cies because of the short (i.e. about three weeks) lai-val 

 dispersal periods of these species (e.g. Houde and Zastrow 

 11993] reported several shelf species with planktonic du- 

 ration >100 days). In some coastal areas, the distribution 

 of Prionotus spp. eggs and larvae indicates that spawning 

 may be limited to estuaries; however, the abundance of Pri- 

 onotus larvae offshore of New Jersey suggests that spawn- 

 ing by these species occurs outside estuaries as well. For 

 example, Merriman and Sclar ( 1952) did not find Prionotus 



