Zug el a\ Age and growth of Hawaiian Chelonia 



125 



20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 

 Straight carapace length (cm) 



Figure 8 



Variation of the growth rate estimates (Sl-based) by carapace length class. 

 Each box plot shows median (white bar), limits of 2'"' and 3"' quartiles 

 (solid notched box), range (brackets), and outliers (solid lines). 



The SI metho(3 models increment-width variation over 

 the entire distance from humerus center to outer margin 

 by using a nonlinear, nonparametric smoother. The model 

 is estimated from all available sample data with two or 

 more LAGs and associated diameter measurements with- 

 out regard to size of the turtle. The SI model can be ap- 

 plied to age all turtles for which the outside diameter of 

 the humerus section has been measured. 



In the skeletochronological sample we studied, the CF 

 method gave age estimates significantly higher than the 

 SI method for turtles of intermediate length (Table 1), but 

 expected differences for such turtles were no greater than 

 about 2 years. On the other hand, based on current data 

 the CF-based model gives much lower age estimates than 

 the Sl-based model for turtles longer than about 86 cm 

 SCL (Fig. 5). Thus although either method may suffice 

 for aging juvenile Hawaiian green seaturtles, only the SI 

 method appears to provide support for inferences about 

 growth in turtles larger than about 80 cm. The CF method 

 is computationally simpler, because it involves only linear 

 regression rather than fitting and integrating a nonpara- 

 metric smoother, and this consideration may recommend 

 it to some users. 



Based on our experience with Hawaiian green seaturtle 

 data, the main issue in judging the two techniques ap- 

 pears to be the assumption with the CF method that hu- 

 merus growth is linear. In reality, it is curs-ilinear, and the 

 SI method explicitly models this cun'ilinearitv. Moreover, 

 the SI method makes fuller use of available humerus in- 



crement data than the CF method. Further comparisons 

 of the methods with additional skeletochronological data 

 sets are recommended. 



Ecological precis 



1 Chelonia myclas within the Hawaiian Islands is a com- 

 ponent of the larger Indo-Pacific C. inydas gene pool, 

 yet the nesting females of the Hawaiian population 

 comprise a distinct genetic unit and contain a unique 

 mtDNA haplotype (Bowen et al., 1992). 



2 Except for the posthatching pelagic phase, the green 

 seaturtles of the Hawaiian coastal waters are year- 

 round residents, and all known individuals reproduce 

 within the Hawaiian island chain, predominantly on 

 the beaches of the Northwestern Islands at French 

 Frigate Shoals (Balazs, 1998; Wetherall et al, 1999), 



3 Skeletochronological age estimates indicate that Hawai- 

 ian juveniles exit the pelagic phase between the ages of 

 4 to 10 years. 



4 In coastal waters, juveniles 10 years and older possess 

 a relatively constant growth rate until about 28 to 30 

 years (approximately 80 cm SCL), then growth begins 

 to slow as individuals attain sexual maturity. 



5 The mean SCL of nesting females is 92 cm (range 81- 

 106 cm; Balazs, 1980), suggesting ages of 30 or more 

 years at first nesting for some individuals. 



