FISHERY BULLETIN: VOL. 83, NO. 3 



Figure 4. -Variation in the len^h of vectors derived 

 from a circular distribution test for different 

 periodicities in the settlement and fertilization data 

 presented in Figures 2 and 3 (upper graph). Note that a 

 period of about 15 d provides the largest vector value 

 beyond a 1-d period (Rayleigh test statistic z = 28.2 for 

 settlement and 10.8 for fertilization, where 2(p^Q qi) = 

 4.6). Lower graph is a plot of vector values from ar- 

 tificial data generated by repeating the analysis for 2, 4, 

 and 8 cycles (artificial data were 0, 0, 0. 0, 25, 50, 75, 

 100, 75, 50, 25, 0, 0, 0, 0). Note how an increased 

 number of cycles or periods narrows the peak for the 

 dominant 15-d period and introduces more harmonics. 

 The settlement and fertilization data (upper graph) 

 represent 16 cycles for a period of 15 d. The height of 

 the harmonics after 15 d probably reflects basic sam- 

 pling noise and the presence of smaller weekly peaks 

 (see Figs. 2, 3). 



•— FERTILIZATION 

 O— SETTLEMENT 



10 15 20 

 PERIOD IN DAYS 



25 30 



Table 1.— Dominant periods derived from a circular distribution test of the settlement 

 and fertilization data presented in Figures 2 and 3. Significance symbols are ** = P< 

 0.01; z values are Rayleigh test statistic (Batschelet 1965). 



for settlement and at least 9 short-period pulses for 

 fertilization (Table 2). 



Timing of Settlement and Fertilization 



To estimate the phase of settlement and of fertili- 

 zation to the lunar cycle, the delay in days from the 



nearest full or new moon was determined for each 

 event (Table 2). The mean phase delay for fertiliza- 

 tion was 5.4 d, and for settlement 7.3 d. Clearly both 

 fertilization and settlement are more often associ- 

 ated with the quarter moons than they are with the 

 full or new moons. The overall relation to the lunar 

 cycle is revealed more clearly by matching each in- 



418 



