FISHERY BULLETIN: VOL. 78, NO. 3 



10 15 20 25 10 15 20 25 10 15 20 25 10 15 20 25 

 TEMPERATURE (°C) 



Figure 6. — Average number of ecdyses for all larval stages of 

 Squilla empusa grouped by 16 temperature-salinity combina- 

 tions according to salinity. 



molted rarely and did not survive long at the lower 

 salinities. At higher salinities molting occurred 

 slightly more often and larvae at 10° C-35%o were 

 able to endure the longest of any of the combina- 

 tions. Some of these larvae persisted for as long as 

 47 d, but usually without molting. Since these 

 larvae did not appear to feed and moved only 

 slightly, the low temperature only seemed to delay 

 their deaths. 



DISCUSSION 



The information provided by the RANN survey 

 only loosely delimited the seasonal occurrence of 

 S. empusa larvae in Chesapeake Bay because of 

 infrequent sampling. A weekly sampling program 

 would have been beyond the scope of the investiga- 

 tion considering that the purpose of the RANN 

 study was to survey the entire zooplankton com- 

 munity of the lower Chesapeake Bay over a 2-yr 

 period. Supplemental data taken at Cape Henry 

 combined with the RANN data indicate that the 

 seasonal occurrence of S. empusa larvae extends 

 from late July to early October, a period of about 

 11 wk. Observations made by Brooks (1878) con- 

 cerning the planktonic duration of S. empusa lar- 

 vae are in agreement with the current study, but 

 the time of occurrence was slightly earlier in the 

 previous study. Larvae were present in the 

 plankton from early July through August when 

 Brooks discontinued the study. Eleven weeks is a 



fairly short planktonic duration for stomatopod 

 larvae. The temperate species Oratosquilla 

 oratoria, common in Japan, has a 5-mo duration 

 (Senta 1967), and Pterygosquilla armata 

 schizodontia was discovered to remain in the 

 plankton for up to 9 mo (Pyne 1972). 



Although Brooks (1878) found the .S. empusa 

 larvae in great abundance, sometimes collecting 

 200 or 300 in a single evening from the mouth of 

 the James River, both the RANN data and the 

 Cape Henry data showed that the larvae were 

 never abundant. Because only five of the nine lar- 

 val stages were collected during the RANN survey 

 the abundance values are inordinately low. Ap- 

 parently, the larger larvae are able to avoid the 

 small bongo plankton nets. Great quantities of 

 Stage I larvae were captured throughout the lar- 

 val season but far fewer numbers of Stage II were 

 caught and fewer still Stage III larvae were caught 

 and so on until Stages V-VIII were not collected at 

 all. The large decreases seem to be too great to be 

 accounted for by mortality alone. 



It is possible that the large, quick-moving (pers. 

 obs.) stomatopod larvae could avoid the small 

 mouth of the net which was easily detectable since 

 sampling was conducted during daylight 

 (Fleminger and Clutter 1965; McGowan and 

 Fraundorf 1966; Murphy and Clutter 1972). Olney 

 (1978) also used data collected during the RANN 

 survey and found evidence of avoidance in other 

 large, agile zooplankters, particularly mysids and 

 fish larvae. A Vi m net and night sampling were 

 used during the Cape Henry survey and the elu- 

 sive stages missed by the bongo sampler were cap- 

 tured. Although all nine stages were collected, 

 peak abundance was still slightly lower than the 

 RANN values. The lower value may have resulted 

 from the use of a smaller mesh net, from not 

 employing a flowmeter to obtain better filtration 

 estimates, or from yearly fluctuations in the popu- 

 lation; but, the abundance figures determined by 

 the RANN and Cape Henry surveys are low for 

 an organism that has been considered to be abun- 

 dant in Chesapeake Bay (Brooks 1878; Cowles 

 1930; Wass 1972). 



The RANN data showed the month of maximum 

 abundance to be August, but the Cape Henry data 

 demonstrated a peak abundance in early Sep- 

 tember. Again, this discrepancy may be attributed 

 to normal yearly variation, but it probably re- 

 sulted from the RANN program having sampled 

 only the younger element of the population. 



698 



