Comparative Population Biology of Four Prions (Genus Pachyptila) from the Indian Ocean and Consequences for Their Taxonomic Status

Abstract

-We studied the morphometrics, breeding biology, genetics, and calls of Pachyptila desolata, P. salvini, P. belcheri, and P. turtur in the southern Indian Ocean. Multivariate analysis of the measurements of live birds revealed some overlap among closely related taxa. Comparisons of biological data provided evidence for ecological segregation between taxa, based mainly on phenology of breeding but also on diet. Calls differed significantly between taxa, and multivariate analysis showed no overlap between sympatric taxa. These data were then compared with data from other taxa (P. crassirostris and macgillivrayi) and from different localities. We support the recognition at the species level of each of the four study taxa. Received 29 June 1989, accepted 6 November 1989. THE SYSTEMATICS of Procellariiformes, especially at the species level, is unclear (see Bourne 1987). This is partly because related species often overlap in their measurements and color patterns which are the most widely used taxonomic characters in the group. The petrels of the genus Pachyptila, the prions, represent a case in point. Prion taxa have usually been divided into two separate groups: the whale-birds, which comprises desolata, salvini, vittata, and belcheri, and the turtur-crassirostris group (Fleming 1941, Harper 1980). Between 1930 and 1950, the six taxa were grouped in one to three genera and in four to six species (Mathews 1912, 1938; Murphy 1936; Falla 1940; Fleming 1941), with many subspecies (19 listed in Murphy 1936). The situation was clarified by Falla (1940). Harper (1980) concluded, after an extensive study of both live (at sea) and museum specimens, that Falla's six species were valid. However, Cox (1980) recognized only three species: vittata (including desolata and salvini), belcheri, and turtur (including crassirostris). His argument was influenced chiefly by the extreme geographic variation within each species, the very large overlap in diagnostic characters (measurements and coloration), and the possible existence of hybrids (between desolata and belcheri on Kerguelen). One reason for the complexity and contradictions in prion taxonomy may be that most authors have worked on beach-washed and museum specimens (Falla 1940, Cox 1980, Harper 1980). Bill measurements, however, vary widely in relation to age of specimens (shrinkage of up to 18% for bill width; Kin-.ky and Harper 1968), age of birds (Richdale 1944, Serventy et al. 1971, Harper 1980), and sample size. As a consequence, some individuals were misclassified or impossible to assign to a particular taxon (Cox 1980). Thus, it appeared that either differences between taxa should be statistically valid (Fullagar 1972, Harper 1980) or the species ranking would have to be abandoned for some taxa (Cox 1980). Taxonomic problems are now addressed by new methods such as electrophoresis (Avise and Zink 1988, Zink 1988), DNA-DNA hybridization (Sibley and Ahlquist 1981, Sibley et al. 1988), behavior (Jouventin 1982, Bretagnolle in press a), and ecology (Mayr 1970). According to the Biological Species Concept, closely related species occupy different ecological niches (Mayr 1982). Investigation of ecological niches therefore provides a powerful tool for detecting sibling species (Mayr 1970). None of these methods has been used on prions. Consequently, we compared the population biology of four taxa in the Indian Ocean. We studied morphometrics of live adult birds, analyzed their breeding biology (nesting habitats, phenology, diet, and distribution at sea), and measured allozyme variation and calls. The calls are particularly relevant for nocturnal burrowing petrels, because they may act as isolating mechanisms (Brooke 1986, Bretagnolle in press b). We extend our results to the broader literature to attempt a clarification of prion taxonomic status. 305 The Auk 107: 305-316. April 1990 This content downloaded from 207.46.13.101 on Sat, 08 Oct 2016 05:38:24 UTC All use subject to http://about.jstor.org/terms 306 BRETAGNOLLE, ZOTIER, AND JOUVENTIN [Auk, Vol. 107