Meckel et al : Evasive behavior of Stenella attenuata and 5. longirostris 



699 



index contours were approximately parallel to the 60% 

 contour and in addition formed a triangle-shaped area off 

 the coast of Guatemala. Evasion indices 40'7( and lower 

 extended south and offshore west of Mexico (Fig. 1). There- 

 fore, the 607c , 509< , and 40% contours were considered the 

 limits between three areas with different evasive behavior 

 of the northeastern offshore spotted dolphin for our study. 

 Three areas were identified (shaded areas in Fig. 1); 



1) Low evasion: south of Mexico (30% to 40% estimated 

 evasion index by set) 



2) Medium evasion: coastal area south of Guatemala (50% 

 to 55% estimated evasion index by set) 



3) High evasion: mouth of the Gulf of California (60% and 

 higher estimated evasion index by set) 



Not all data between these contours were used because 

 of difficulty in extracting data throughout the geographic 

 range and because sample sizes for each area (n>100) 

 seemed to be adequate for the analysis. Average mean eva- 

 sion indices differed significantly between the three areas 



(Ki-uskal-WaUis, ;),=206, ;i.,=lll. 



=491; total 7)=808. 



//=93.13, 2 df, /'<0.0001). Hence, nonparametric multiple 

 comparisons between all pairs of areas were executed, and 

 all pairs were found to be significantly different (Table 2). 

 According to these results, there seemed to be a spatial 

 pattern in evasive behavior (measured in our study as the 

 estimated median evasion index by set in each evasion 

 area) of northeastern offshore spotted dolphins during 

 fishing operations of the Mexican fleet from 1992 to 1995. 



Differences between stocks 



The eastern spinner dolphin seemed to evade capture more 

 effectively than the northeastern offshore spotted dolphin 

 in evasion area 3 when estimated evasion indices by set for 

 both stocks were compared (Fig. 2). The differences were 

 significant according to the Kolmogorov-Smirnov one-sided 

 test for two independent samples (spinner: ;!=275, spot- 

 ted: «=489; D=0.2031, P<0.001). 



With respect to the evasive strategies of these two 

 stocks, evasion index when the dolphins escaped under 

 the net was compared by set. Eastern spinner dolphins 

 apparently escaped more effectively under the net than 



northeastern offshore spotted dolphins (spinner «=177, 

 spotted ?! = 125, D=0.4097, P<0.001, Kolmogorov-Smirnov 

 two-sample test; Fig. 3). 



Dispersion behavior of northeastern offshore spotted 

 and eastern spinner dolphins is presented for evasion 

 area 3 (mouth of the Gulf of California, Fig. 4). Herds of 

 both stocks tended to disperse from one set stage to the 

 next; during encirclement, grouping code 3 (herd divided 

 into more than 3 subgroups) had increased and was the 

 most frequently recorded grouping code for northeastern 

 offshore spotted dolphins in area 3 during our study. 

 Northeastern offshore spotted herds tended to be more 

 fragmented than eastern spinner herds before chase and 

 to disperse in greater numbers during subsequent set 

 stages (Fig. 4). Logistic regression analysis revealed a 

 reliable full model (x2=60.209. P<0.001, df 3, n=544), i.e. 

 the outcome of stock was predicted by the three indepen- 

 dent variables (set stages). The prediction of stock out- 

 come might further be interpreted as a difference between 

 northeastern offshore spotted and eastern spinner dol- 

 phins with respect to dispersion behavior in our study. 



In contrast, both stocks presumably escaped only on 

 rare occasions by leaping out of the net because eastern 

 spinners did so in only 5 of 275 sets (1.82%) and north- 

 eastern offshore spotted dolphins in 3 of 489 sets (0.61%) 

 during the study period. Estimated evasion indices by set 

 could not be compared because the samples were too small 

 for any statistical test. 



Discussion 



Spatial patterns in evasive behavior 



An apparent significant geographic difference in evasive 

 behavior of northeastern offshore spotted dolphins was 

 found between three areas in the EPO (Mexican fleet data, 

 1992-95, Fig. 1, Table 2). The lowest evasion area was 

 located south of Mexico (area 1), the medium evasion area 

 was in a relatively small coastal area south of Guatemala 

 (area 2), and the highest evasion area was in the Gulf of 

 California mouth (area 3, Fig. 1). About twenty years ago, 

 fishermen noticed that dolphins were more difficult to cap- 

 ture in some areas than in others. Certain coastal dolphin 



