300 



Fishery Bulletin 98(2) 



litter of 25-30 from the Mediterranean Sea (Sanzo, 

 1912; Mollet et al.i). Pratt and Casey ( 1983 ) reported 

 April parturition at a size of 65-75 cm TL in the 

 western North Atlantic. Stevens (1983) and Cliff et 

 al. (1990) suggested November parturition at about 

 70 cm TL in New South Wales and KwaZulu-Natal, 

 respectively. The data presented by Gilmore (1993) 

 showed no clear time of parturition, but investiga- 

 tion during the course of our study revealed several 

 data errors (see Table 1). Cliff et al. ( 1990) proposed 

 a 6- or 18-month gestation but indicated that 18 

 months appeared more feasible, based on a gesta- 

 tion of a little over 1 year postulated by Pratt and 

 Casey (1983). Pratt and Casey (1983) reported first 

 year growth rates of about 50 cm per year in the 

 western North Atlantic, whereas Cailliet et al. ( 1983 ) 

 suggested growth rates half as fast off California. 

 Recent tagging results from California- and New 

 Zealand (Saul and Holdsworth, 1992; Saul'^) have 

 shown that growth rates in the Pacific are similar to 

 those reported for the Atlantic.^ The size-frequency 

 modes reported by Hanan et al. (1993) and O'Brian 

 and Sunada (1994) confirmed high juvenile gi'owth 

 rates for California sharks.** 



Currently there is growing concern over the extent 

 of pelagic shark catches worldwide, both as targeted 

 fishing and, particularly, as bycatch of high-seas 

 longlining. Further attempts to manage resources 

 of shortfin makos will require, among other demo- 

 graphic data,^ information on reproductive para- 

 meters. Although this species is commonly caught 

 in a variety of fisheries (Bonfil, 1994), few pregnant 

 females have been documented and extensive bio- 

 logical data from a single location do not exist. Given 

 all these factors and assuming reproductive synchro- 

 nism of mating and gestation, spring parturition, 

 and consistency in growth and size at birth of sharks 

 from several widely spaced regions, we decided to 

 combine all available information, adjusting for sea- 

 sonal shifts between hemispheres. Our study docu- 

 ments the reproductive biology of female shortfin 

 mako sharks from around the world. 



' Mollet, H. F., A. D. Tcsti, L. J. V. Compagno, and M. P. P^rancis. 

 1999. Re-identification of a lamnid shark embryo. In review. 



- Laughlin, L. 1997. California Department of Fish and Game, 

 330 Golden Shore, Long Beach, CA 90802. Personal commun. 



■' Saul, P. 1997. Blue Water Marine Research, RD :!, Whanga- 

 rei New Zealand. Personal commun. 



■> Mollet, H. F, and G. M. Cailliet. 1997. Prehminary demo- 

 graphic analysis of the shortfin mako shark, Isuriis oxyrinchu^. 

 Program and Abstract of the American Society of Ichthyologists 

 and Herpelologists ( A.SIH i and American Elasmobranch Society 

 (AE.Si Annual Meetuig, University of Washington, Seattle WA, 

 .June26-July 1, 1997.:i:i6p. 



Materials and methods 



Materials examined 



Reproductive and morphometric data from 95 mature 

 females, including 35 pi'egnant individuals (Table 1), 

 together with data on 450 postnatal fish, including 

 almost 200 age-O-i- fish, were collected from around 

 the world. Much of the data came from the western 

 North Atlantic (largely unpublished, with the excep- 

 tion of data on three pregnant females), from east- 

 ern Australia (all unpublished, except four pregnant 

 females reported by Stevens 1 1983]), and from Kwa- 

 Zulu-Natal, South Africa (either unpublished or pro- 

 viding substantially more detailed information than 

 that reported by Cliff et al. [1990J). For methods of 

 collection, measurement, and assessment of matu- 

 rity see Pratt and Ca.sey ( 1983 ); Stevens ( 1983 ); Cliff 

 et al. ( 1990); and Stevens and McLoughlin ( 1991). 



Morphometries 



We used total length (TL, expressed in meters or 

 cm) for our length measurement following Stevens 

 (1983) and calculated TL of western North Atlantic 

 and some South African sharks from the relation- 

 ship with fork length (FL) (Casey and Kohler, 1992) 

 or precaudal length (PCL) (Cliff et al., 1990), respec- 

 tively. Total length of embryos, if not measured, was 

 estimated from TL-FL ratios (1.156 at TL 60 cm; 

 1.203 at TL 35 cm) and the TL-PCL ratio (1.294 at 

 TL 60 cm) of similar-size specimens. 



We defined power regression as a linear y-on-.v 

 regression of log-transformed data. We used Student 

 residuals and leverage, an index of the leverage of 

 each observation on the size of the mean square 

 error, to carry out linear regression residual analy- 

 ses (Wilkinson, 1986). We fitted power regi'essions 

 to our mass-length data of females weighing more 

 than 55 kg(TL=~2 m) from the western North Atlan- 

 tic and the Southern Hemisphere (no maternal data 

 were available from Brazilian litters) and reported 

 back-transformed equations in the form M (kg)=a 

 TL^ (m). The prepower coefficient "a" predicts the 

 mass of a 1-m-TL shortfin mako because we used 

 meters as the unit of length in all such calculations 

 (Mollet and Cailliet, 1996). We were interested only 

 in possible mass differences of adolescent and mature 

 females from different regions and chose 55 kg as the 

 suitable minimum. A geometrical mean (GM) regres- 

 sion gives the correct functional relationship (Mollet 

 and Cailliet, 1996), but we wanted to compare our 

 results with previously reported back-transformed 

 y-on-.v power regressions. An F-tesi and the linear 

 interaction model were used to check for statisti- 



