Feeding was common in early mornings but tended to decrease 

 throughout the day, being replaced by socializing (Fig. lib). There is a curious 

 increase in "mill" and "unknown" behaviors in mid-day, either of which 

 may be related to forage, rest, or social activities. 



Group sizes were significantly smaller in "channels" (over all habitat 

 types J = 3.5 ± 2.93 (SD) dolphins /group, for "channel" jc = 3.0 ± 2.54 [SD], 

 P < 0.0001, ANOVA, followed by Fisher's LSD). Group sizes were skewed 

 towards smaller groups (median = 2.5, and maximum = 20 dolphins /group 

 over all groups seen). Group sizes tended to be largest in midday (Fig. 12b) and 

 for socializing groups (Fig. 13). No difference in group size was found between 

 age classes, pregnant vs. non-pregnant, vdth-calf vs. without-calf, or sex for 

 freeze-branded dolphins; however, sizes of groups containing males had 

 greater variance than did those containing females (with males 

 J = 7.3 ± 2.25 [SD], n = 8 dolphins; with females x = 6.0 ± 1.40 [SD], n = 8). 

 Biases in behavioral analyses resulting from unequal representation of 

 individual dolphins ii\ the data set can be gauged by sighting frequencies 

 shown in Figure 10. 



Travel direction showed a NE/SW bimodality (Table 8, Figiire 14), 

 probably an artifact of the geography of the study site, which is essentially a 

 corridor with NE/SW orientation. Analysis of behavior by sex, age, pregnancy 

 and with-calf was inconclusive, perhaps due to small sample sizes. 

 Frequencies of behaviors were not significantly different between males and 

 females (P = 0.06, Mann-Whitney U). Excluding May 1993, of all surveys 

 between July 1992 and June 1993 (Table 2), neonates were seen only in the 

 July-August and September 1992; and June 1993 surveys. The proportion of 

 neonates to other dolphins was x = 0.02 ± 0.028 (SD) (n = 8 surveys). 



POPULATION SIZE 



Mark/recapture data indicate that of 409 dolphin sightings during surveys 

 not biased by radio-tracking, there were 31 resightings of freeze-branded 

 animals. Bailey mark /recapture p)opulation size estimates ranged from 101 

 (June 1993) to 434 (October 1992) (Table 9). Over all sbc surveys, representing 

 13 months, the estimate was 218 dolphins. Figure 15 shows the area used for 

 making "recaptures" (resighings) and the area which the 218 dolphins are 

 assumed to occupy. As a comparison, "unrefined" estimates, based on Af = 30 

 with resightings from the entire survey area (all shaded areas in Fig. 4), yield 

 Bailey estimates of 370 (January 1993) to 1,161 (October 1992). 



Our method of approximating geographic population closure gives an 

 indication of the amount of dolphin use of the area, implying that in a year's 



19 



