From boat based observations, surface durations of about 3 sec were 

 strongly correlated with a single breath. Longer surface durations sometimes 

 correlated with two breaths, one at the beginning and one at the end of the 

 surfacing. Occasionally, and especially at night, we recorded surface durations 

 Icisting many seconds, to > 1 min. Sometimes a dolphin had its transmitter 

 antenna continuously above water for as long as 3 min. From the tracking 

 vessel, such protracted surface durations were observed in dolphins 

 motionless at the surface and in dolphins foraging in very shallow water 

 (> 0.5 m). 



ASSOCIATIONS AMONG INDIVIDUALS 



Frequent changes in group sizes and affiliations occurred among these 

 radio-tagged, freeze-branded, and other recognizable dolphins. Almost all 

 affiliations for freeze-branded dolphins seen more than once were below 

 0.200 on the half-weight index of association, indicating that none of the adult 

 freeze-branded dolphins were dose associates. No difference was found in the 

 number of affiliates across sex, pregnant vs. non-pregnant, 

 with-calf vs. without-calf, or age for non-calves with ^ five sightings. Both 

 male and female dolphins tended to have more male affiliates than female, 

 and males tended to have more affiliates over-all (for non-calf freeze-branded 

 aiumals seen ^ five times and with ^ four freeze-branded affiliates). Sample 

 sizes were insufficient to show potential affiliation differences by age or 

 reproductive classes. 



Affiliations between freeze-branded dolphins only, as judged by 

 occurrence within the same group in a sighting, were weak except for mother- 

 calf pairs (Fig. 10). The mean level of association was 0.12 ± 0.027 (SD) (for 

 non-calves seen ^ five rimes, n = eight males, six females). Table 4 shows the 

 14 dolphins and their mean index values by sex of affiliate and overall. 



Thirty-six of a possible 91 (39%) pairwise combinatioris were sighted. Forty 

 percent (6/15), 43% (12/28), and 75% (18/24) of pairwrise combinations were 

 sighted for female-female, male-male, and opposite-sex affiliations, 

 resp>ecrively. Resighrings of pairs were low: two female-female pairs were 

 seen twice, one male-male pair was seen twice, cmd six opposite-sex pairs were 

 seen two to three times. 



The mean index value {i.e., level of association) did not differ among 

 male-male associations, female-female associations, opposite-sex associations, 

 or associations overall (Kruskal-Wallis, df = 3, P = .45), though male-female 

 pairs showed greater variability in level of association (Table 5). 



16 



