Correlates of variability in killer whale stereotyped call repertoires.

Abstract

Killer whales vocalisations include repertoires of stereotyped call types (Ford 1984). There is strong evidence that these vocalisations are learnt (Hoelzel and Osborne 1986; Bain 1989; Deecke et al. 2000; Yurk et al. 2002). Call types can be group specific or shared amongst a number of groups, depending upon the social structure of the population or the call type (Ford 1991; Deecke 2003). It is thought that these call types function in the group cohesion and coordination (Hoelzel and Osborne 1986; Ford 1989 1991; Miller 2000, 2002). Some call types contain two overlapping, independently modulated, components each having different transmission properties (Miller 2002), these call types have a higher estimated active space than single- component call types (Miller 2006).This thesis investigates the evolution of these call type repertoires, focusing on call type usage and structure of the Southern Resident population over a period of 27 years, but including comparisons with other populations. I present evidence of hetero-specific mimicry and further evidence for vocal production and usage learning in killer whales. I compared the relative frequency of use of call types between two time periods (1977-81 & 2001-2003) and between contexts, such as direction changes with directional travel and multi-pod aggregations with single pods. I found a strong correlation of relative call type usage for each pod between the two time periods and each pod was easily acoustically distinguishable from the other two pods in both periods. The implications of these results for a role of call type repertoires in kin recognition are discussed. The least cohesive pod produced a significantly higher proportion of two-component call types than the other two more cohesive pods. Lone whales separated from their pod also used a rare subset of two-component call types rather than their pod's main call types. In recordings of multi-pod aggregations I recorded a high proportion of the same subset of two-component call types not commonly produced by any of the three pods individually, these call types were used in significantly higher proportions when all three pods were converging or socialising rather than travelling. These contextual correlates suggest that call types are selectively used and shared between groups based on their transmission properties. Each of the Southern Resident pods, J к and L, were found to increase the duration of their primary call type 10-15% in the presence of vessel noise in recordings made between 2001 & 2003. This response was not detected in recording from two earlier time periods, (1977 or 1989-1992). This change in behaviour correlated with an increase over the past decade in vessel presence around this population and may be an anti-masking strategy. I also compared the range and mean minimum and maximum fundamental frequency of the call types within the repertoires of six North Pacific killer whale populations. There was a degree of homogeneity in the range of call type fundamental frequencies within the repertoires of populations of the same ecotype, but differences between ecotypes. Offshore call types generally had a higher pitch fundamental frequency than transient or resident call types. All three resident populations had call types in their repertoires that had a maximum fundamental frequency 3 kHz higher than found in any transient call type.