854 



Fishery Bulletin 92(4), 1994 



dorsal spine which was considered most suitable for 

 ageing (also see Litvinov, 1990). All ridges on the 

 surface of the mantle covering the spine were counted 

 by a single reader with a low-power dissecting scope 

 and methods described by Ketchen (1975) and 

 Beamish and McFarlane (1985). No ridges were 

 grouped and counted as a single "annulus" as was 

 done in several earlier studies (e.g. Holden and Mead- 

 ows, 1962). Loss of increments due to abrasion of the 

 tip of the spine has been reported for other species of 

 Squalus (e.g. Ketchen, 1975), but this did not occur 

 in the present study. For example, when a worn spine 

 was observed, the worn area was confined to the re- 

 gion of the spine tip having a diameter of <3 mm. 

 However, unworn spines having spine base diameters 

 of <3 mm had not yet formed any increments (the 

 spine base diameter of a late-term uterine embryo 

 was 2.0 mm). 



Validation of the annual nature of spine increment 

 formation (following the methods described by 

 McFarlane and Beamish [1986, 1987]) was not pos- 

 sible. Therefore increment counts from spines re- 

 ported here must be considered as tentative esti- 

 mates of age. Nonetheless, validation studies con- 

 ducted on S. acanthias have verified that spine in- 

 crements do represent annual marks (Beamish and 

 McFarlane, 1985; Tucker, 1985). 



Individual length-at-age data for each sex were fit- 

 ted with a nonlinear estimation procedure (Wilkin- 

 son, 1988) to the von Bertalanffy growth model 

 (Ricker, 1975): 



L, =L„(l-e<-*"-'" ,) ), 



where L, is length at age, L^ is asymptotic length, k 

 is the growth coefficient, and r is the theoretical age 

 when L ( = 0. 



Reproductive data were first collected from shark 

 specimens during the summer of 1986. For females, 

 counts were made of mature ovarian eggs (greater 

 than about 2.5 cm diameter), candled embryos (i.e. 

 gelatinous uterine capsules containing embryos in 

 early stages of development), and embryos free in 

 the uteri. Females possessing any or all of these re- 

 productive products were considered sexually ma- 

 ture. Sex and TL of uterine embryos were also re- 

 corded. Ketchen ( 1972) determined that a period of 

 rapid increase in clasper length indicates the onset 

 of sexual maturity in male S. acanthias. A similar 

 allometric growth phase between clasper length and 

 TL was observed for S. mitsukurii. Thus, the right 

 clasper length was measured to the nearest 1 mm 

 from the body juncture to the clasper tip. 



To determine the size at 50% sexual maturity for 

 both sexes, data were fitted to the logistic function 



by using an iteratively weighted (i.e. inverse of vari- 

 ance) nonlinear estimation procedure (Wilkinson, 

 1988) and evaluated at 50% (Somerton, 1980). The 

 logistic equation is defined as 



Y=l/(l+Ae BX ), 



where Y is the proportion of animals sexually ma- 

 ture, X is the midpoint of a length class, and A and B 

 are parameters defining the curve. Before the logis- 

 tic equation was fitted to length data for males, the 

 data were transformed to approximate the logistic 

 pattern of growth. For example, clasper length (CD 

 plotted against fish TL (not shown) produced a 

 roughly sigmoid curve, although at large and small 

 fish sizes, clasper length continues to increase with 

 size of the fish. To flatten the ends of the curve as 

 required by the logistic equation, clasper length was 

 expressed as a proportion offish TL and normalized 

 to values between 1 and (i.e. with a=CL/TL, the 

 quantity Y'=(a-a min )/(a max -a min ) was plotted against 

 fish TL). 



To obtain information on feeding habits, the stom- 

 ach contents from 251 S. mitsukurii caught on 

 longlines and 42 fish caught in gill nets during the 

 summer of 1986 were examined. Longline-caught fish 

 ranged from 20.6 to 79.5 cm TL ( mean length L =50.4 

 cm) and gillnet-caught fish ranged from 49.5 to 78.9 

 cm TL ( L =65.5 cm). Samples from the bottom 

 longline and gillnet gear were compared to examine 

 the potential feeding bias that might occur if attrac- 

 tion to baited longlines varied as a function of stom- 

 ach fullness for S. mitsukurii. 



For most of the fish, stomachs were extracted upon 

 capture and preserved in 10% formalin until exami- 

 nation in the laboratory. However, if time permitted, 

 stomachs were examined at sea for the presence of 

 food items; empty stomachs were noted and discarded 

 (those with food items were saved). In the labora- 

 tory, samples were sorted, counted, and identified to 

 the lowest possible taxon. Food items were sorted to 

 taxa, blotted dry, and weighed to the nearest 0.1 g. 



Stomach contents data were analyzed for frequency 

 of occurrence, numerical abundance, and gravimet- 

 ric proportions of prey items to quantitatively de- 

 scribe the diet and feeding habits of S. mitsukurii 

 (Hyslop, 1980). Prey items that were attributed to 

 the presence of the research vessel (i.e. bait, galley 

 refuse, or processed fish offal) were not included in 

 the analyses; stomachs containing only those items 

 were considered empty. To examine diel feeding be- 

 havior, samples were grouped by time of capture: 

 0600-1200 (n=39), 1200-1800 (ra=86), 1800-2400 

 (n=64 ), and 0000-0600 ( n =62 ). Chi-square ( y? ) analy- 

 ses of 2 x 2 contingency tables were used to test for 



