Table 1. — Replicate tow data for analysis of variance of 

 catches of Penaeus duorarum protozoeae, Tortugas Shelf, 

 Florida, August 1962 to October 1964 



Table 2. — Summary of analysis of variance of estimates of 

 abundance of protozoeal Penaeus duorarum under 10 m.- 

 surface area on the Tortugas Shelf, Florida, August 1962 

 to October 1964 



in addition to the normal sources of variation, 

 such as patchiness of plankton, random varia- 

 tions in catch, and errors in the actual counting 

 and subsampling procedures, the estimates in- 

 corporate variation in flowmeter performance and 

 average depth over which the oblique tows were 

 made. More important than these factors, how- 

 ever, is the fact that the analysis is based not 

 upon paired hauls, but upon replicates. RepUca- 

 tion was necessary because of the hmitations of 

 the winch used during the work. One sample was 

 taken while the vessel traveled in a given direc- 

 tion and the other while the vessel moved over 

 the station in the reverse direction. Such a sampling 

 method must result in greater variabiUty than 

 would be obtained from simultaneous hauls. 



Errors involved in integrating estimates over 

 time and space have been discussed by a number 

 of authors; the reader is referred to the works of 

 Taft (1960), Saville (1964), and EngUsh (1964). 



I The large value of F indicates that significantly more of the total varia- 

 tion IS due to diflerences between stations than between replicates (F <0.01) . 



On four occasions during this investigation, 

 several sampling stations were occupied over 24- 

 hour periods. No consistent differences were ob- 

 served between night and daytime catches, and 

 variability appeared to be within the margins 

 expected according to the foregoing analysis. 

 The oblique-tow method apparently sampled all 

 layers of the water column and counteracted 

 samphng problems caused by diurnal vertical 

 migrations. 



SEASONAL ABUNDANCE OF LARVAE 



The equation developed for Pacific sardine by 

 Sette and Ahlstrom (1948) and Ahlstrom (1954) 

 has been appUed to the counts of first protozoeae 

 of P. duorarum to obtain estimates of larval 

 production in the sampling area during the period 

 covered by this survey. 



The equation is C = S(CiW,ti), in which 



C=the total number of first protozoeae pro- 

 duced in the survey area during the 

 period ti 



C,=L,/D 



Li = estimated number of first protozoeae per 

 10 m.^ of surface area at the ith station 



D = the duration of the first protozoeal stage 

 (3 days) 



Wi = the area represented by station i 

 t, = the time weighting given to station i 

 (equal to one-half the time elapsing since 

 the preceding occupancy plus one-half the 

 time elapsing before the succeeding oc- 

 cupancy) 



The basic data for the analysis are given in 

 tables 3 and 4. The area represented by each 

 sampling station is 320 km.^ (see Sette and Ahl- 

 strom, 1948, for rationale regarding the area 

 represented by a station) and the time weighting, 

 ti, is usually about 14 days. 



168 



U.S. FISH AND WILDLIFE SERVICE 



