We considered rarity which is a component that is both extensively used in conservation strategy, and is of potential interest in ecosystem functioning (e.g. Walker et al., 1999; Lyons et al. 2005; Ellingsen et al., 2007). This concept is usually defined on the basis of the level of species local abundance or the species range size (occurrence). But whatever the approach used, definitions of rarity are necessarily arbitrary (Gaston, 1994). Because species with limited occurrence are usually the most vulnerable to environmental change (Thomas and Mallorie, 1985), the occurrence criterion has been the most widely used for conservation purposes. Thus, in our study we used the criterion of occurrence (species range size) rather than local abundance to define rarity.
The definition of a threshold (rarity cutoff) to identify which species are rare (and which are not) is a necessary but subjective stage. Here, instead of a priori defining the rarity cutoff in a fully subjective way (as is often the case in the literature), we have defined it on the basis of the empirical distribution of frequency of species occurrence (i.e. the number of species for different percentages of occurrence). The threshold of percentage of occurrence we adopted to define the list of rare species corresponded to the drop in frequency point in this distribution (i.e. where the number of species by percentage of occurrence sharply decreases when increasing this percentage). While not perfect, this threshold definition is less subjective than definitions currently available. For instance, Gaston (1994) and more recently Ellingsen et al. (2007), who noted no consensus in the literature to define a threshold, proposed another subjective threshold, called the “quartile definition” (defining as rare every species with a lower abundance or occurrence below 25%). In contrast, in our study, using the above-mentioned selection method enabled us to define a posteriori the percentage of occurrence used as the rarity cutoff.
In this paper, the rarity index we used corresponds to the number of rare species in each sample (haul), with rare species occurring in less than 5 % of the 1454 hauls for the whole MEDITS zone (Table 1). 26 species, including 12 Chondrichthyes, were below this threshold and thus considered as rare species in our work (see Table S1).