Publication date
Introduction
The ability to determine sex in the field for monomorphic bird species using a reliable, inexpensive, and efficient method can help in designing protocols for the study of breeding biology, behaviour, demography, and time budgets (Mallory & Forbes 2005). Although sexual-size dimorphism generally exists in petrels (Procellariiformes) for at least some body measurements (Guicking et al. 2004; Einoder et al. 2008; Bugoni & Furness 2009; Landers et al. 2011), differences often are not overtly obvious (Genevois & Bretagnolle 1995). Consequently, sex frequently needs to be determined using a variety of other techniques, such as cloacal inspection, vocalisations, molecular analyses, or statistical approaches such as discriminant function analysis (Genevois & Bretagnolle 1995; Weidinger & van Franeker 1998; Bertellotti et al. 2002; Mallory & Forbes 2005; Bourgeois et al. 2007). Cloacal inspection can provide a means of identifying sex as the cloacae of females are larger than those of males due to the egg passage, but this technique is limited to the period during egg laying for breeding pairs (Boersma & Davies 1987; O’Dwyer et al. 2006). Differences in vocalisation patterns have been used to determine sex in some shearwater species (Cure et al. 2009), but this too relies on birds calling and cannot be used at times of the year when birds are silent. As a result, morphological differences assessed by discriminant function analysis (DFA) have been widely used to determine sex for a variety of bird species (Bertelloti et al. 2002; Mallory & Forbes 2005; Einoder et al. 2008; Liordos & Goutner 2008; Landers et al. 2011). This analytical technique identifies those morphological characters which best discriminate males and females, and the resulting canonical classification function may be used to classify sex (McGarigal et al. 2000). Results from the DFA allows one to estimate the proportion of correctly identified individuals in a sample of birds using the classification function (DechaumeMoncharmont et al. 2011).