644 



Fishery Bulletin 100(3) 



Reproductive seasonality The reproductive period for 

 both species was prolonged and overlapping (Fig. 2). The 

 range of P. carolinus hatching dates was from 19 May 

 to 5 September and P. evolans hatching dates extended 

 from 2 June to 3 September. Median hatching dates for 

 P. carolinus were significantly later than P. evolans i 13 

 August versus 26 July; Wilcoxon 2-sample test: P=0.011). 

 Both species spawned for at least three months but P. 

 carolinus hatching dates were dramatically skewed to 

 the left (i.e. most individuals hatched late in the season). 

 Previous studies independently demonstrated that both 

 species spawn contemporaneously in the New York Bight: 

 Keirans et al. (1986) calculated egg densities, Wilk et al. 

 (1990) examined gonadosomatic indices, McBride and 

 Able (1994) used length-frequency analysis, and McBride 

 et al. (2002) examined larval densities. All demonstrated 

 an extended spawning season for both species from about 

 May to September; furthermore, a bimodal pattern of 

 spawning output was reported for Prionotus spp. by 

 Keirans et al. (1986) and McBride and Able (1994). My 

 sampling procedure (i.e. a stratified, random design with 

 respect to length intervals) would flatten frequency peaks 

 and emphasize the range of hatching dates, a method 

 not well suited for identifying multiple spawning peaks. 

 Moreover, McBride et al. (2002) identified notable geo- 

 graphic variation in spawning seasonality, which could 

 not be separated out in the present study and should be 

 accounted for in future research. Intra-annual reproduc- 

 tive periodicity of Prionotus should also be evaluated by 

 examining gonads for cyclic, group-synchronous oocyte 

 development, which could be an underlying process lead- 

 ing to a protracted spawning period with regularly spaced 

 peaks in production. 



I conclude that interspecific size differences are at least 

 partly the result of interspecific differences in reproduc- 

 tive seasonality (Table 1; i.e. the smaller congener was 

 spawned later in the year). However, Able and Fahay 

 (1998) noted that P. evolans eggs are larger than P. caro- 

 linus eggs; thus size differences exist among embryos. In 

 addition, McBride et al. (2002) showed that interspecific 



size-at-age differences are evident throughout the larval 

 period. 



Age and growth Age-0 P. evolans were larger at a 

 common age than P. carolinus (Fig. 3). On the basis of 

 all individuals examined, P. evolans did not grow at a 

 significantly higher rate (ANCOVA interaction of slopes; 

 P=0.099), but this species was significantly larger than 

 P. carolinus in general (ANCOVA test of intercepts; 

 P=0.0001). The regression slopes, based on all data, were 

 significantly different from zero (P<0.001 ) for both P caro- 

 linus (SL=6.12-K0.323xage; r2=0.62; n=70) and P evolans 

 (SL=3. 52-1-0. 429xage; r2=0.53; n=53). The intercepts for 

 these linear growth models were biologically meaning- 

 ful, particularly that for P. evolans because the intercept 

 was roughly equal to the known size at hatching (3 mm; 

 Able and Fahay, 1998). Prionotus evolans collected at four 

 different stations in Delaware Bay during October 1991 

 deviated strongly from the average growth rates for this 

 species (Fig. 3). Such small, slow-growing fish demon- 

 strated the degree of intraspecific variation possible for P 

 evolans growth rates. 



Age-0 P carolinus were also much smaller on average 

 than P. evolans during winter ( Fig. 4). By February-March, 

 age-0 P. carolinus ranged from 27 to at least 117 mm SL, 

 but this size distribution was strongly bimodal with modes 

 at 42.5 mm and 87.5 mm SL. The age-0 P evolans were all 

 larger than 75 mm SL and appeared to grow even larger 

 than 120 mm SL, which was the cutoff size for collections, 

 so that the interspecific size difference is even greater than 

 that indicated in Figure 4. Apparently all age-1 P. evolans 

 are larger than this cutoff value because none were ob- 

 served in the samples, whereas age-1 P. carolinus ranged 

 from 56 to at least 118 mm. Although age-1 P. carolinus 

 were generally larger than age-0 conspecifics, the sizes of 

 both age classes overlapped in a manner that would con- 

 found length-frequency analyses of this species. 



The intraspecific size variation of age-0 fishes, of both 

 Prionotus species, was spatially correlated. Individually, 



Jun 1 



Aug 1 Sep 1 



Figure 2 



Backcalculated hatching dates (date of capture - 

 [otolith age + 4]) for age-0 Prionotus carolinus 

 (open bars; n=70l and P. evolans (filled bars; n=53) 

 collected in estuarine and continental shelf habi- 

 tats during 1991. 



