124 



Fisheiy Bulletin 100(1) 



10 



20 



30 



40 



50 



10 



20 30 



Estimated age (yr) 



40 



50 



Figure 7 



Obsei-i'ed gi'owth vectors of marked Molokai green seaturtles between 

 release and recapture (vectors) in relation to the gr-owth predicted by non- 

 parametric models of SCL and age (cui-ves); age was estimated by the 

 two methods: correction factor (A) and spline integration (Bl. The turtle's 

 length at release is assumed to fall on the growth cur\'e. 



effects of fibropapillomatosis and gender on growth rates 

 among the Hawaiian population. 



Age-estimation methods 



A key assumption of both the CF and SI methods is that 

 each estimated humerus growth layer represents 1 year of 

 gi-owth. This assumption has been validated only recently 

 (Hohn and Snover' ). Hawaiian turtles tagged and injected 

 with tetracycline have been recaptured, and these turtles 

 show the appropriate number of LAGs for the years since 

 their receipt of tetracycline. Furthermore, strong support 

 is provided by the consistency of the growth model pre- 

 dictions with obsei-ved gi-owth in tagged Molokai turtles. 

 Additional justifications have been advanced in other stud- 

 ies of seaturtle humerus LAG formation (e.g. Zug and Glor, 

 1998; Coles etal., 2001). 



' Hohn, A., and M. Snover. 2001. Personal commun. Beaufort 

 Laboratory, Southeast Fisheries Science Center. Beaufort, NC. 



Both the CF and SI methods require histological prepa- 

 ration and analysis of humerus sections. But they use the 

 same skeletochronological data in independent and differ- 

 ent ways to estimate the total number of humerus growth 

 layers. The CF method assumes that a constant humerus 

 growth rate (the "correction factor") applies to the resorp- 

 tion core regardless of the diameter of the core and despite 

 the fact that periosteal increment width decreases with 

 length of the turtle (Fig. 2). The correction factor, C. is esti- 

 mated from a subset of the skeletochronological data tak- 

 en from juvenile turtles only, i.e. excluding larger turtles 

 likely to have narrower increments in the outer region of 

 the humerus. Even so, the CF estimates of age for juve- 

 nile turtles appear to be biased upward (Fig. 5), suggest- 

 ing that the constant correction factor also failed to reflect 

 the effect of wider increments deposited in the early years 

 of life. In age estimation, the CF method can be applied 

 only to turtles displaying a complete set of periosteal lay- 

 ers, i.e. distinct LAGs, from the resorption core to the outer 

 margin of the humerus. 



