910 



Fishery Bulletin 97(4), 1999 



ID 



5 



20 

 30 

 40 

 50 

 60 

 70 

 80 

 90 

 100 



D16 



361 



□20 



22 

 1% 



D18 



9 



013 



42 

 -50 

 ■M 

 361 



AD16Q 



017 



24 



-11 

 37 

 « 



AD14 



014 



45 

 50 



•37 

 39 

 12 



377 



015 



-17 



■20 

 -6 

 -23 

 -14 

 102 



016 

 □20 

 018 

 013 

 017 

 014 

 015 



AD18(? 



AD20C88) AD 13d 



AD15 



AD17 



Figure 6 



Top: dendrogram illustrating intrapod groups 

 and relationships among individuals with CAI val- 

 ues calculated at the population level. (Relation- 

 ships among intrapod groups shown in Fig. 2.1 

 Middle: matrix showing CAI values between 

 pairs of individuals calculated at the pod level 

 for AD16 pod. Numbers at the base of each col- 

 umn in the matrix are the number of valid photo 

 sequences in which that individual appears. Bot- 

 tom: inferred genealogical trees. 



Of the 45 proposed maternal lineages (Figs. 3-12), 

 25 included two generations, 16 included three gen- 

 erations, and four had four generations. In all but one 

 of the four-generational trees, the oldest matriarch died 

 during the study. We identified four matrilineal 

 intrapod groups that appeared destined to die out. 

 These were the single adult male, AB3 ( Fig. 4); the AB 10 

 subpod which consisted of the apparently postrepro- 

 ductive female AB 10 and 3 adult male sons, AB4,AB5, 

 and ABll (Fig. 6); the apparently postreproductive 

 AJ12 and her adult male son, AJ16 (Fig. 9); and the 

 apparently postreproductive AN 34 and her adult male 

 son, AN21 (Fig. 10). Three of the remaining matrilin- 

 eal groups had produced two reproductive females in 

 one generation and were growing, and eight had pro- 

 duced one reproductive female and were relatively 

 stable. The fate of the other matrilineal groups will be 

 determined later as offspring bom during the study 

 mature and their reproductive potential is realized. 



Discussion 



The strength of bonds between male offspring and 

 their mothers was highly variable (Table 4). Simple 

 regression analysis indicated that the strength of 

 bonds tended to diminish with age, but not signifi- 

 cantly (Fj ^.= 1.62; P=0.215). 



Direct observations indicated that resident killer 

 whales in Prince William Sound do not associate ran- 

 domly with one another, but rather tend to associate 

 with specific individuals. Statistical analysis of as- 

 sociations in photographic sequences supports this 



