FISHERY BULLETIN: VOL. 83, NO. 4 



160 



E 140 

 u 



X 



I- 



0.040 



Lt = 81.0e 0017 V ^ J 



I I I 

 ND J 



I I I I I I 



FMAMJ 



I I I I i I 

 J ASON D J 



I I I I I 

 FMAMJ 



MONTH 



Figure 6— Mean values of monthly distributions of length from data for the 

 southern offshore spotted dolphin. The curve represents one annual cohort 

 fitted by the Gompertz model of growth. The open circles were not included 

 in the fitting of the curve. NOTE: The equations were fitted using relative 

 time and are not therefore accurate models of growth, lb obtain such growth 

 models, relative time can be converted to absolute time using the estimate of 

 length at birth and the equation refitted to these data. 



ing to follow actual cohorts of animals from in- 

 dividual years as did Perrin et al. (1976) in their 

 analyses. Combining the data from several years in- 

 troduces additional variation in the data if the timing 

 of the calving season varies from year to year, but 

 it increases sample sizes and minimizes bias caused 

 by nonrandom timing of sampling within months. 

 In addition, we are mainly interested in an average 

 growth rate which is best estimated from several 

 years of data. 



Perrin et al. (1976) fitted a linear model to mean 

 lengths estimated in the months of October 1972, 

 January, February, March, and April through June 

 1973 from which they extrapolated to obtain an 

 estimate of length at 1 yr of 147.5 cm for the north- 

 ern offshore spotted dolphin. The authors recognized 

 that this estimate was biased upwards because 

 growth rates of delphinids do decrease in the first 

 year and revised this estimate downwards based on 

 aged specimens. We believe our analyses to be more 



accurate than those of Perrin et al. (1976) and our 

 results to be a substantial improvement. 



Barlow (1984) found an indication of two peak 

 calving seasons for spotted dolphins north of the 

 Equator, in the spring and autumn, but that the 

 animals were born throughout the year. South of the 

 Equator he found a single season peaking around 

 April. These results are similar to ours but the timing 

 of the southern season does not agree The difference 

 can be explained by Barlow's use of Perrin et al.'s 

 (1976) growth curve which predicted a mean length 

 of 138.0 cm for 1-yr-old animals. Our growth equa- 

 tions predict animals of this length to be from 14 

 to 16 mo old. 



The two growth curves fitted to the mean lengths 

 from the two cohorts in the northern region predict 

 different lengths-at-age and are characterized by 

 very different growth rates. It is possible that these 

 two cohorts actually grow at different rates because 

 of environmental factors, but we believe that the dif- 



562 



