CHANGES IN THE DEMOGRAPHIC TRENDS OF PINK RIVER DOLPHINS (INIA) AT THE MICROGEOGRAPHICAL LEVEL IN PERUVIAN AND BOLIVIAN RIVERS AND WITHIN THE UPPER AMAZON:MICROSATELLITES AND MTDNA ANALYSES AND INSIGHTS INTO INIA’S ORIGIN - Biology, Evolution and Conservation of River Dolphins Within South America and Asia - page 171

R ESULTS

Microsatellites

All the different procedures applied to these data sets used the 10 microsatellites of this
study. Table 1 shows the imbalance indices (Kimmel et al., 1998). All the values obtained of
 (and ln ) were significantly higher than 1 () or 0 (ln ) thus indicating later population
growth from an initial bottleneck. Morever, all the rivers studied (the total upper Amazon),
showed a similar degree of population expansion after an initial bottleneck. The Marañón
River population had the highest population expansion or it was more recent [ = 3.702, ln
= 1.309; t = 6.179, 9 df, P < 0.01; 99 % confidence interval (2.578 , 4.826); 500 coalescence
based simulations generated an empirical distribution with the ln  values within a 95 %
confidence interval (-0.23, 0.25)].

Table 1. Statistics calculated for the imbalance indices (Kimmel et al., 1998). t =
Student's t test. * Significant tests at P < 0.05. IC = Interval of confidence.

Bolivia
(Mamore River) Statistics Upper Amazon Ucayali River Marañon River Napo-Curaray River

_^
 v 2.03084 5.29446 5.03882 7.28930 4.35980

_^
P o 0.58534 0.40901 0.40051 0.45000 0.46744

^
 2.11694 2.12723 1.92536 3.70188 2.43798

^
ln  0.74997 0.75482 0.65511 1.30884 0.89116

t 5.089* 5.645* 4.603* 6.179* 2.422*

95% IC (1.687/2.547) (1.736/2.519) (1.531/2.319) (2.845/4.559) (1.975/2.900)

99% IC (1.553/2.681) (1.614/2.640) (1.409/2.442) (2.578/4.826) (1.831/3.044)

The test of Zhivotovsky et al. (2000) (Table 2) showed a different picture than the
previous one, indicating, that each one of the tests employed has a different power to detect
diverse demographical changes at different historical moments. This test detected a strong
and significant bottleneck event in the Bolivian population (S k = -2.253; t = - 5.274, 11 df, P
< 0.01; 99 % confidence interval (-3.351 , -1.155)). This significant negative S k was
accompanied by the lowest of the within-population variances found in this study (V = 0.9952
+ 0.2647). The Napo-Curaray rivers, the Marañón River and the overall upper Amazon
showed pink river dolphin populations with a constant demographic size (S k = -0.0001, 0.175
and 0.208, respectively). All these S k values were not significantly different from 0. Their
respective within-population variances (V) were 2.1343 + 0.7786, 3.4587 + 1.7185 and
2.6254 + 0.8257, with the V value for the Marañón River slightly higher than in the other
cases. The Ucayali River yielded a demographic expansion at a significant P < 0.05 level (S k
= 0.350; t = 2.098, 10 df, P < 0.05; 95 % confidence interval (0.023 , 0.677)), but this value
was not significant at a P < 0.01 level (t = 2.098, 10 df, P > 0.01; 99 % confidence interval (-
0.078 , 0.778)) and the within-population variance was similar to that found in the other
Peruvian rivers and in the overall Amazon sample (V = 2.2580 + 0.7985). Therefore, this test
is more powerful than the Kimmel test when detecting stronger bottlenecks in the initial