censused weekly during most of the year and daily 

 during the breeding season. Regular weekly cen- 

 suses of other pinnipeds have been conducted 

 since 1972, and as far back as 1970 irregular 

 counts were made. Since we found shark activity 

 to be seasonal (see below), we included comments 

 on the seasonal changes in sea-surface tempera- 

 ture and salinity. Such information was derived 

 from daily readings made at noon (P.s.t.). 



Results 



We recorded events involving sharks 58 times 

 between 9 September 1970 and 9 February 1979. 

 Of these, 37 were definite observations of sharks 

 eating pinnipeds (i.e., "shark attacks") and the 

 remainder (21) were mostly of sharks seen within 

 1 km of the island. In definite shark attacks, we 

 were first (36 of 37 times) alerted to the incident 

 by a flock of gulls (mostly the western gull, Larus 

 occidentalis), hovering above a large bloody area 

 (5 m^) in the water. Usually we saw the shark's 

 head and dorsal and caudal fins which offered 

 clues to species identification and estimation of 

 size and number. We often saw the pinniped prey 

 as well. On five occasions only an area of bloody 

 water and the hovering gulls were observed. 

 These, too, were likely shark attacks on pinnipeds, 

 but we did not include them in the 37 known 

 attacks. Observations, from discovery of the gull 

 fiock to dissipation of blood and gulls, lasted from 

 3 to 15 min. If the carcass was not consumed in 

 that time, then its disappearance was probably 

 due to sinking. In the areas of most shark/ 

 pinniped interactions the water was 4-12 m deep. 



In 20 interactions we saw the pinniped prey 

 sufficiently well for a positive identification. Two 

 involved sea lions, one or perhaps two involved 

 harbor seals, and the remainder involved elephant 

 seals. Based on size, we could tell that seven 

 elephant seals were young individuals, about 3 yr 

 old or less, and an eighth was an adult female. On 

 rare occasions, we have observed sea lions with 

 obvious, fresh shark-bite wounds. However, the 

 location of most shark attacks in the vicinity of the 

 elephant seal hauling out areas (Figure 1) further 

 supports what other data indicate, that at South- 

 east Farallon, sharks more frequently ate ele- 

 phant seals than other pinnipeds. 



In 30 instances, the white shark, Carcharodon 

 carcharias, was identified as the species seen 

 preying on seals. All were at least 3 m long and 

 most about 3.5-5 m long. In 16 of 21 nonattack 



observations within 1 km of the island, the shark 

 was also identified: 14 involved white sharks and 

 two involved the blue shark, Prionace glauca. 

 How many sharks were present at any one time is 

 not known. On at least three occasions two sharks, 

 and once three sharks, simultaneously fed on one 

 pinniped. On 8 January 1976, a 3 m long white 

 shark was caught in Fisherman's Bay and 7 d later 

 two larger white sharks were seen on the opposite 

 side of the islands. 



Sharks were more abundant or more active 

 during the late fall and winter over the 9 yr 

 (Figure 2). The number of attacks in December 

 and January was perhaps artifically low (see 

 below) because on many days during those months 

 few if any gulls were present to alert observers. 

 The possibility that attacks were missed was 

 particularly likely in December 1976 and January 

 1977. Sharks were known to be present then 

 because several seals hauled out with fresh shark- 

 bite wounds and part of another was seen floating 

 in the water. Yet no attacks were seen. The timing 

 of greatest shark activity corresponded closely 

 with the Davidson Current period (October- 

 February) which, as described by Bolin and Abbott 

 (1963), is characterized by slowly declining sea- 

 surface temperatures and salinities and the ap- 

 pearance of a northward flowing countercurrent 

 (shoreward of the south-flowing California Cur- 

 rent). White sharks were thus present (or at least 

 active, see below) when waters were warm but not 

 necessarily the warmest (Figure 2). Blue sharks, 

 on the other hand, were definitely most abundant 

 during the warm oceanic period (July -September), 

 when the California Current slackens, allowing 

 warm saline oceanic waters to flow shoreward. 

 They were observed commonly but only at 3 km or 

 more away from the island. Few were involved in 

 the observations reported here. 



The timing of most shark attacks also cor- 

 responded closely with the late autumn peak in 

 elephant seal numbers (Figure 2). Each year the 

 elephant seals reached maximum numbers twice, 

 during midspring and again during late fall (Le 

 Boeuf et al. 1974). A third, smaller peak occurred 

 during the winter breeding season. Only two 

 shark attacks, both involving elephant seals, were 

 observed during the upwelling period (March- 

 July), when the California Current flows most 

 strongly and temperatures reach their lowest. 

 One of these attacks occurred during the spring 

 peak in elephant seal numbers. During the 

 fall, when most shark attacks were observed. 



942 



