Summary: | Dolphins emit whistles for communication among individuals of the same species, transmitting information that enables individual identification and group cohesion. Due to this fact, whistles show characteristics that differentiate them in a species-specific way. An increasing number of studies have used this type of vocalization to identify species, but few of them were carried out in interspecific association contexts. Thus, information about the relationship between these associations and whistles properties represent an important aspect to be understood. In these circumstances, the aim of this work was to classify and characterize the whistles emitted by Globicephala melas (long-finned pilot whale) and Tursiops truncatus (bottlenose dolphin) recorded during an interaction in the Brazilian south continental shelf and slope, South Atlantic Ocean. The acoustic data were opportunistically collected in May 2014 using two hydrophones of an Auset® array (response frequency: 1,592 to 48 kHz) towed through the stern of the R/V Atlântico Sul and coupled to a Fostex FR-2 LE digital recorder (sampling of 48 kHz/24 bits). A total of 418 whistles were analyzed, 117 of which were visually classified as emitted by G. melas and 301 by T. truncatus. The multivariate discriminant analysis showed a correct classification rate of 97.99%. Constant whistles were more commonly produced by G. melas (N = 50, 46.73%), while for T. truncatus the multiple category occurred more frequently (N = 131, 46.13%). The bottlenose dolphin produced whistles with higher mean frequencies than the long-finned pilot whale, in contrast, this species showed a coefficient of variation with higher values for most variables. G. melas whistles varied from 1.64 to 13.93 kHz, while T. truncatus whistles ranged between 5.03 and 23.91 kHz. Both species emitted whistles with near mean duration and number of inflection points above one. Differences were found in whistles acoustic parameters in relation to studies with other populations. We hypothesize that ...
|