FISHERY BULLETIN: VOL. 78, NO. 2 



summer). The secondary importance of distance 

 from shore (negative correlation) was reflected in 

 decreasing abundances with increasing distance 

 from shore. Because all collections were made well 

 within the range of optimum salinities for de- 

 velopment, salinity could have been expected to 

 contribute little to explained variation in abun- 

 dance. 



Results of multivariate analysis of data em- 

 phasize that a bivariate approach to multivariate 

 data can be misleading. Salinity had the highest 

 simple correlation (-0.7086; Table 4) with abun- 

 dance. Consequently, in the usual SPSS stepwise 

 multiple regression procedure (Nie et al. 1975) 

 salinity would be entered as the first variable in 

 the analysis. The proportion of variation (r^, Ir'^) 

 in abundance explained by temperature, distance 

 from shore, salinity, and depth was quite in con- 

 trast to proportions of explained variance when 

 salinity, rather than temperature, was entered 

 first in the multiple regression equation (Table 6). 



The relative importance of temperature, dis- 

 tance from shore, and salinity was best illustrated 

 by partial correlation coefficients (Table 5). This 

 procedure examines correlation between two vari- 

 ables when the effect of other variables is con- 

 trolled. I recommend partial correlation as a pre- 

 liminary step to multiple regression procedures 

 and as more appropriate than bivariate proce- 

 dures. 



This paper reports for the first time the exis- 

 tence of a large population ofCallinectes larvae in 

 offshore shelf waters of the Middle Atlantic Bight. 

 The presence of an offshore population, necessary 

 to the accepted model of larval distribution, has 

 had relatively little documentation. Nichols and 

 Keney (1963 ) found advanced stages of Callinectes 

 as far as 64-97 km offshore, with greatest abun- 

 dance at stations 32 km offshore. Dudley and Judy 



( 197 1) reported zoeae, chiefly stage III and earlier, 

 and a few megalopae at stations 10-13 km 

 offshore. Tagatz (1968) reported a few megalopae 

 as far upstream as 40 km in the St. Johns River, 

 Fla., and Williams (1971) collected megalopae". . . 

 almost the entire length of the [North Carolina) 

 estuary." 



Abundance reported here (Figure 3) is some- 

 what less than that previously reported. Sandifer 



(1972) and Tagatz (1968) reported maximum lar- 

 val abundance of 42,000 and 46,100/100 m^, re- 

 spectively, and Dudley and Judy (1971) reported 

 maximum abundance of 105,000 100 m^. These 

 data, however, included few megalopae. Williams 



(1971) found considerably greater numbers of 

 megalopae than did pi evious workers but reported 

 abundance as numbers per sample (10's-l,000's). 



My results confirm the reported affinity of Cal- 

 linectes larvae, particularly megalopae, for sur- 

 face layers. Previously Sandifer (1972) reported 

 that 89.4% of the Callinectes larvae that he col- 

 lected were from surface samples but reported 

 only three occurrences of megalopae, all in bottom 

 samples. Dudley and Judy (1971) found, overall, 

 more Callinectes larvae in surface (1.0 m) than in 

 subsurface (8.0 m) collections except at offshore 

 (10-13 km) stations. They collected advanced 

 zoeae (their last three stages) only at offshore sta- 

 tions. Tagatz (1968) collected more zoeae at the 

 surface than at the bottom, and Williams (1971) 

 reported Callinectes megalopae to be active at 

 night in surface waters. The results of these 

 studies, however, reflect differences in gear types, 

 mesh sizes, and sampling design; gear specifically 

 designed to sample surface layers was in no case 

 employed. 



Reasons for the affinity of Callinectes and other 

 megalopae for the neuston are not readily appar- 

 ent. Diel increases in abundance in night collec- 

 tions of neuston may indicate a negative photo- 

 tropic response or possibly net avoidance in the 

 daytime. Numerous holoplankters (copepods, etc.) 

 exhibit the same diel pattern, and the upward 

 movement of megalopae may be related to feeding 

 strategies. It is not surprising that of the 

 megalopae collected in this study, the Portunidae 

 (swimming crabs) showed the strongest affinity 

 for the neuston. Megalopae of other crabs, how- 

 ever, such as Cancer, Ocypode, and Dromidia also 

 showed strong surface affinities. 



Spatial distribution of plankton in shelf waters 

 is largely determined by circulation patterns. 

 With cross-shelf flow in the Middle Atlantic Bight 

 offshore at the surface and onshore at depth 

 (Bumpus 1973), coastal organisms in the surface 

 layers would be transported offshore, with the pos- 

 sibility of return at depth. A coastal boundary 

 layer, a band of trapped nearshore flow some 10 

 km wide, has been reported off New Jersey 

 (Csanady 1976). Coastal boundary layers are as- 

 sociated with the upwelling of cold water as a 

 consequence of the offshore movement of surface 

 waters and subsequent thermocline tilt (Csanady 

 1976). Most coastal and estuarine larvae in my 

 collections (species of Uca , Palaemonetes , Libinia, 

 etc.) were infrequent seaward of station CI and 

 are evidently retained within this zone. Late stage 



262 



