(Perrin et al. 1976). If it be assumed that some 

 factor in delphinid growth is unaccounted for in 

 the Sacher-Staffeldt model and that gestation 

 time for S. longirostris is underestimated to a 

 similar degree (11.5 mo minus 10.0 mo/ 11. 5 mo, or 

 13%), an adjusted Sacher-Staffeldt estimate of 

 10.54 mo is obtained. 



The estimate based on length at birth ( 10.74 mo) 

 and the adjusted Sacher-Staffeldt estimate (10.54 

 mo) are close to each other, and a rounded off aver- 

 age of the two estimates, 10.6 mo, is used below in 

 the analyses of reproduction. Making the assump- 

 tion that fetal growth follows a pattern similar to 

 that in S. attenuata, i.e., that t in Laws' fetal 

 growth equation L = a(t - t ) is approximately the 

 same proportion of gestation time as in S. at- 

 tenuata, or 0.135 t g (Perrin et al. 1976), a fetal 

 growth curve can be estimated (Figure 15). The 

 slope of the linear portion of the curve is 8.367 

 cm/mo, as compared with 8.283 cm/mo estimated 

 for S. attenuata (Perrin et al. 1976). 



Postnatal Growth 



We found it impossible to estimate postnatal 

 growth rates by the usual method of following the 

 seasonal progression of length modes, for the 

 reasons discussed above. We deduced an estimate 

 of growth rate during the first 10 to 11 mo by 

 application of the equation, log Y = 0.99 \ogX - 

 1.33, expressing an inferred relationship in 

 toothed cetaceans between length at birth (X 

 above) and the difference (Y above) between the 

 growth rates during the linear phases of fetal and 

 early postnatal growth (Perrin et al. 1976). The 

 predicted difference based on length at birth of 77 

 cm is 3.60 cm/mo. Subtraction of this from the 



80 -_ _ Average Jength_at birUi 



FISHERY BULLETIN: VOL. 75. NO. 4 



fetal linear growth (estimated above) of 8.37 

 cm/mo yields an estimate of average growth rate 

 during the first 10 to 11 mo after birth of 4.77 

 cm/mo. If this is taken as an estimate of average 

 growth rate during the first year, predicted length 

 at 1 yr is 134 cm. This method overestimates 

 length at 1 yr to some unknown, but slight extent, 

 as growth is only approximately linear in the first 

 year. 



We examined the relationship between length 

 and number of postnatal dentinal growth layers in 

 the teeth for 183 males and 250 females (Figure 

 16). The occurrence in the samples of length-layer 

 data for relatively more females than males with 

 more than about 12 layers is accounted for by the 

 fact that the sample of males selected for tooth- 

 sectioning was stratified entirely by length, 

 whereas the sample of females was stratified by 

 length in juveniles and by number of ovarian cor- 

 pora in adults. All females with more than 10 

 ovarian corpora were included in the sample, in 

 addition to randomly selected, corpora-stratified 

 subsamples of females with <10 corpora. 



We fit growth curves to the data (to single-layer 

 incremental means), using a two-cycle version of 

 the Laird growth model (see Perrin et al. 1976 for 

 discussion of the model). Juvenile males and 

 females were considered jointly. The fit was forced 

 through the origin (zero growth layers, and esti- 

 mated length at birth of 77 cm), and asymptotic 

 length (La,) was estimated as the average length of 

 animals with 13 or more layers (Loo for 12 males = 

 179.46 cm and for 60 females = 170.91 cm), fixing 

 the upper ends of the two curves of the second 

 cycle. The simultaneous iterative fitting proce- 

 dure arrived at 4.111 growth layers (rounded off to 

 4 below) as the age at which convergence of the 

 three curves (estimated onset of a secondary 

 growth spurt) yields the best fit (Figure 16). Esti- 

 mated length at this age is 156.85 cm. The Laird/ 

 Gompertz model (Laird 1969) is 



L(t) = L exp \±\ l-exp( 



""'HI 



FIGURE 15. — Estimated fetal growth curve for the eastern spin- 

 ner dolphin. 



736 



where L = length in centimeters 

 t = age 



L = length at age zero 



a = specific rate of exponential growth 

 a = rate of decay of exponential growth. 



A form of the model generalized to the present 

 case of more than one cycle is 



