Mean index value and number of affiliates were not dependent on age 

 (simple linear regression). However, females with many affiliates had more 

 male affiliates than females with lower numbers of affiliates (P = 0.007, 

 R^ = 0.87, simple linear regression) while maintaining a similar number of 

 female affiliates (P = 0.72, R^ = 0.03). For females, mean index value was not 

 dependent on number of affiliates (split by sex of affiliate: male P = 0.10, 

 R2 = 0.67, female P = 0.72, R^ = 0.04). 



Males with high numbers of affiliates had more male and female 

 affiliates. They had lower index values for female affiliates than males with 

 few affiliates (P = 0.008, R^ = 0.79, simple linear regression), yet showed little 

 difference in association levels with other males (P = 0.72, R^ = 0.02). Males 

 sighted in larger mean group sizes tended to have more female affiliates than 

 those seen in small groups (P = 0.046, R^ = 0.512). 



We examined the first and second highest level affiliates ("1°" and 

 "2°" affiliates, as measured by the tv^o highest association indices) of the eight 

 male and six female non-calves seen ^ five times (Table 6). Sample sizes were 

 too small for statistical tests, so we report here on tendencies. One individual 

 (nnale FB512) had two 1" affiliates (tied index values, a male and a female) and 

 so was counted twice in some of the following analyses. Eleven of 14 1° and 

 2° affiliations were of same-sex pairs (7 male-male, 4 female-female). Seven of 

 eight male-meile 1° and 2* affiliations were of sinular aged dolphins (within 

 1-3 yr). Four of six 1° and 2° female-female associations were of similar aged 

 dolphins (within 1-4 yr). Two males and 1 female had 1° affiliates of opposite 

 sex. Eighteen of 28 affiliations were reciprocated at the 1° or 2° level (i.e., nine 

 pairs of dolphms). Male-female associations tended to be reciprocated as often 

 as same-sex affiliations. 



The 1" reciprocal male-male pair FB502-FB504 had an index value of 0.190 

 (Table 6). For FB504 this value was > 2 SD's above his mean index value for 

 associations with other males, females, or overall (Table 4). For FB502, the 

 0.190 value was < 1 SD above mean, perhaps due to small sample sizes. 

 Similarly, the male-male pair FB518-FB522 shared an index value of 0.114 

 with each other. The index value was > 2 SD above mean for association with 

 other males for FB518 and for associations overall for FB522. FB518 and FB522 

 were each others' second highest affiliates (a reciprocal 2° pair). For both 

 FB518 and FB522, the 1° affiliate was FB521, a female. The FB518-FB521 

 association was reciprocated at an index value of 0.146, and was approximately 

 1 SD above the mean for opposite-sex associations, but > 2 SD above mean for 

 same-sex associations and overall associatiorw, for both individuals. They 

 were seen more with each other than with other freeze-branded dolphins of 

 the same sex, respectively. 



17 



