58 



Fishery Bulletin 98(1) 



21-28 individual ova, in the uteri of a female that con- 

 tained four embryos about 4.3^.7 cm long. Duhamel 

 and Ozouf-Costaz (1982) found 102 nonfertile eggs in 

 the uteri of their Kerguelen Island female, which car- 

 ried embryos of 9.6-10.4 mm. No photographs were 

 taken of the eggs, and it is unclear whether they were 

 egg capsules or individual ova (DuhameP). 



The emerging picture for oophagous sharks is that 

 large numbers of nutritive egg capsules accumulate 

 in the uteri during the early stages of gestation, and 

 they are rapidly consumed as the embryos grow large 

 enough to puncture and eventually swallow them ( Gr- 

 uber and Compagno, 1981; Otake and Mizue, 1981; 

 Gilmore, 1983; Gilmore et al., 1983; Mollet et al."). 

 Ovulation peaks during midgestation, but full egg cap- 

 sules are rarely found in the uteri, probably because 

 they are eaten soon after entering the uteri. During 

 later gestation, ovulation ceases and embryos metabo- 

 lise their accumulated stomach contents for energy, 

 growth, and storage in the liver. 



Litter size, embryonic growth, and gestation 



In 36 out of 40 of our pregnant females, litter size 

 was four, with a mean of 3.85 embryos. Litters of 

 two or three were occasionally recorded. In the North 

 Atlantic, Shann (1911, 1923) stated that litters com- 

 monly consisted of two embryos (range 1-4), but his 

 data almost certainly included several partial litters 

 (see Table I in Shann, 1923). Templeman ( 1963) found 

 three, four and four embryos in his three litters, and 

 Gauld ( 1989) found four to be the most common num- 

 ber of embryos in a litter, with a mean of 3.7 (n = 12).^- 

 One litter of five has also been reported ( Bigelow and 

 Schroeder, 1948). Our sample of Southern Hemisphere 

 litters is the largest yet assembled, and Gauld's ( 1 989 ) 

 is the largest fi-om the North Atlantic. Both samples 

 had very similar mean numbers of embryos. We con- 

 clude that litter size is usually four, but smaller lit- 

 ters are occasionally found; litters larger than four are 

 extremely rare. Some litters with fewer than four em- 

 bryos were probably incomplete. Abortion of embryos 

 during capture is common among nonlamnid sharks, 

 but it is difficult to imagine midterm embryos with 

 grossly distended abdomens being aborted, even when 

 the mother is compressed in a trawl net. One of our 

 New Zealand longline litters containing two midterm 

 embryos, 37.5 and 38.3 cm long, was presumably com- 

 plete. Abortion of near-term embryos is quite possible. 



'^ Gilmore's Table 1 contains a number of errors. The reference to 

 Aasen (19661 in the Lamna nasus section should presumably be 

 Aasen (19631; the two Swenandcr il907) litters contained four 

 embryos each rather than 2; and the '219-cm TL shark with four 

 embryos attributed to Nakaya ( 19711 was actually a male, and 

 no embryos were mentioned by Nakaya. , 



Linear regressions fitted to the length-month data 

 for both hemispheres suggest that embryos have a 

 rapid growth rate of about 7 cm per month, but there 

 was much unexplained variability (Fig. 7). The esti- 

 mated growth rate is almost twice that of shortfin 

 mako embryos (3.7 cm per month) (Mollet et al.^'). 

 The length of the gestation period appears to be about 

 8-9 months in both hemispheres. During July-Sep- 

 tember, Aasen ( 1963) found no embryos in the north- 

 west Atlantic despite examining hundreds of mature 

 females. He argued that the gestation period was 8 

 months, and that the females he examined were un- 

 dergoing a short rest period between pregnancies. Our 

 interpretation agrees with that of Aasen ( 1963). 



A contrary hypothesis involving a gestation period 

 of 1-2 years has been advanced by Shann ( 1923 ) and 

 Gauld (1989). They argued that the high variability 

 in embryo length and the apparent presence of two 

 cohorts of embryos were inconsistent with a gestation 

 period of less than one year. We cannot rule out their 

 hypothesis, and we are conscious that our Southern 

 Hemisphere data are limited in seasonal scope and 

 that pooling data across locations and years is not de- 

 sirable. However, we believe the data from both hemi- 

 spheres are most consistent with a gestation period 

 of less than one year. The implied rapid embryonic 

 growth rate is not unreasonable given the abundant 

 embryonic food supply and the relatively high growth 

 rates of postnatal juveniles discussed above. 



The high variability in the embryo length data 

 might be explained by an extended mating period. 

 In the northwest Atlantic, Aasen (1963) found males 

 with seminal vesicles that were filling at the end of 

 August, indicating that mating would begin in Sep- 

 tember. Pratt ( 1993) reported a mature female caught 

 in October with moderate amounts of spermatozoa in 

 the oviducal gland and with ft-esh vaginal abrasions. A 

 mature male with haematose claspers was caught on 

 the same longline, providing strong evidence of mat- 

 ing in October. Gauld ( 1989) found females with fresh 

 bite marks, thought to be inflicted during mating, on 

 females near the Shetland Islands in December-Jan- 

 uary. These observations suggest that mating lasts 

 from September to January in the North Atlantic. 



If our hjTDothesis of rapid embryonic growth and 

 high intracohort variability is correct, parturition 

 probably peaks in June— July (winter) in the Southern 

 Hemisphere and possibly extends ft-om April to Sep- 

 tember. Parturition in the Northern Hemisphere may 

 peak around two months earlier (spring-summer). The 

 presence of distinct length modes in juvenile length-fre- 

 quency distributions from New Zealand, Australia, and 

 the northwest Atlantic (Aasen, 1963 ) confirms that par- 

 turition is restricted to part of the year, rather than oc- 

 curring year-round. Svetlov ( 1978) reported the capture 



