Mollet et al : Reproductive biology of Isurus oxyrinchus 



303 



cally significant differences between hemispheres 

 (Neter and Wassermann, 1974). The condition factor 

 (CF=total masslTL'^, in units of kg/ni'*) was used as 

 a quaHtative estimate of the size of the yolk-filled 

 stomach in lamnid and alopiid embryos. 



The calculation of gonadosomatic index (GSI) 

 (ovary mass as percentage of total mass) and hema- 

 tosomatic index (HSI) (liver mass as percentage of 

 total mass) required a mass estimate if the shark 

 was not weighed. For a few sharks only the mass 

 was available and a TL estimate was required. We 

 estimated mass or TL (from FL) of western North 

 Atlantic and Southern Hemisphere sharks from the 

 equations of Casey and Kohler (1992) and Stevens 

 (1983), respectively. Only two dimensions (length 

 and width) were available for the ovaries of western 

 North Atlantic sharks and the mass had to be esti- 

 mated. The estimates were calibrated with an ovary of 

 dimensions 19.6x8.7cmandmassof0.630kg(Fig. 2C 

 in Stevens, 1983). The resulting GSI values proved 

 to be comparable with those obtained from weighed 

 ovaries of mature Southern Hemisphere sharks. 



Size at maturity 



In the absence of embryos or egg cases, the matu- 

 rity of nonpregnant females was determined from 

 GSI, maximum ova diameter (MOD), oviducal gland 

 diameter (measured in western North Atlantic spec- 

 imens only), uterus width, and absence of a hymen. 

 We fitted a logistical model y=[l+e-'°**'^'']-i to our 

 binomial maturity data (immature=0, mature=l) of 

 female specimens with TL > 2.0 m (Wilkinson, 1986; 

 Welch and Foucher, 1988). Two meaningful parame- 

 ters to characterize maturation, namely median TL- 

 at-maturity {MTL=-<i/b} and slope at MTL (S=b/4), 

 can be expressed in terms of a and b as given. An 

 i^-test was used to check for statistically significant 

 difTerences between hemispheres (Neter and Was- 

 sermann, 1974). 



Litters, gestation, and parturition 



Data collected on litters included number of embryos 

 , length and mass of embryos, maternal length and 

 mass, capture date and location, and sea surface 

 temperature (Table 1). Data for 21 litters, either not 

 previously published or published with errors, are 

 presented. We fitted a power regression to the litter 

 size and maternal length data. For two recently fer- 

 tilized South African fish, the number of uterine 

 egg cases containing a single blastodisc ovum was 

 assumed to be the litter size. The gestation period 

 was estimated from seven litters from the western 

 North Atlantic, one from the Gulf of Guinea, one 



from the Mediterranean Sea, two from Japan, seven 

 from Australia, four from South Africa, and four 

 from Brazil (n=26). The litters from Hawaii (no. 10 

 in Table 1) and the Gulf of Mexico (no. 35) were 

 not included because of uncertain length data. We 

 defined early-, mid-, and near-term litters as having 

 embryo size between to 20, 20 to 45, and 45 to 70 

 cm TL, respectively. Embryos of a term litter had 

 TLs between 65-75 cm. Capture dates and TL esti- 

 mates of 188 age-O-i- fish from the western North 

 Atlantic (extrapolations based on Fig. 3 in Pratt and 

 Casey [1983] [n=45J), from California (Cailliet et al., 

 1983; Cailliet^) (n=16), and from New South Wales, 

 Australia (Pepperell, 1992; Pepperell*') (/! = 119), and 

 South Africa (n=8), were also considered in the esti- 

 mation of size at birth and time of parturition. 



Capture dates from Northern and Southern Hemi- 

 sphere fish were combined on a single time scale 

 based on seasons. For example, spring was defined 

 as 21 March to 21 June in the north and from 23 Sep- 

 tember to 23 December in the south. Seasons were 

 further divided into three periods of equal length 

 to define the terms early, mid, and late (e.g. early 

 spring in the north lasts from 21 March to 20 April ). 



To support the combination of Northern and South- 

 ern Hemisphere data for the shortfin mako, we 

 examined sandtiger shark (Carcharias taurus) data 

 from South Africa (Bass et al., 1975) and the east- 

 ern United States (Gilmore et al., 1983). These data 

 were also used to support our use of a linear regres- 

 sion to estimate gestation and parturition from the 

 slope of the embryo TL versus seasonal time regres- 

 sion, as well as size at birth. Bass et al. (1975) 

 reported that 100-cm neonate sandtiger sharks were 

 born from June through August after a gestation 

 of 8-9 months. Gilmore et al. (1983) reported that 

 100-cm neonates were born from December through 

 March after 9-12 months of gestation. We analyzed 

 reported length-at-capture data by using a linear 

 regression which yielded similar estimates for time 

 of parturition and gestation period from intercept 

 and slope: 



South Africa: Jul. -Aug. parturition, 10 months ges- 

 tation, n=27,r2=0.93; 



Eastern United States: Jan.-Feb. parturition, 8 

 months gestation, n = 14, /•'-=0.83; 



Combined data: midwinter parturition, 9 months 

 parturition, n=Al, r'-=0.93. 



^ Cailliet, G. M. 1997. Moss Landing Marine Laboratory, PO 

 Box 530. Moss Landing, CA 95039. Personal commun. 



'' Pepperell, J. G. 1997. Pepperell Research and Consulting 

 Pty, PO Box 818. Caringbah, NSW 2229, Australia. Personal 

 commun. 



