A wolf in the hand is worth two in the bush: a response to Ciucci et al. (2007)

The use of carcasses in demographic studies of mammals is not a new issue (e.g. Englund, 1980, for the red fox Vulpes vulpes; Sidorovich et al., 2007, for the grey wolf Canis lupus; Gonzalez & Crampe, 2001, for the Pyrenean chamois Rupicapra pyrenaica), but the reliability of these studies varies with the origin and quality of the samples. Ciucci et al. (2007) maintain that ‘opportunistic or convenience sampling is not acceptable and should not be encouraged’ in demographic studies. They generalize from the data reported in a recent paper of ours (Lovari et al., 2007) on mortality parameters of the Apennine wolf Canis lupus italicus Altobello (1921), in Central-eastern Italy, based on 154 wolf carcasses, over an 11-year period. We agree with Ciucci et al. (2007) that sound data on the live population, with adequate sampling, are definitely preferable to data from carcasses. However, demographic data can be quite hard to collect from the live population of elusive species, for example the wolf. The first peerreviewed paper on the biology of the wolf in Italy was published 32 years ago (Zimen & Boitani, 1975). Since then, studies on wolves in Italy have been carried out by several research teams across most of the wolf distribution range (e.g. Meriggi et al., 1991; Boitani, 1992; Meriggi et al., 1996; Ciucci et al., 1997; Ciucci & Boitani, 1999; Gazzola et al., 2002; Apollonio et al., 2004). In spite of these research efforts, only food habits (for a review: Meriggi & Lovari, 1996), some aspects of genetics (Randi, Lucchini & Francisci, 1993; Lorenzini & Fico, 1995; Randi et al., 2000; Scandura, Apollonio & Mattioli, 2001; Scandura, 2005; Scandura et al., 2006; Fabbri et al., 2007) and several predator/prey interactions (Meriggi et al., 1996; Mattioli et al., 2004; Gazzola et al., 2005) have been relatively well assessed. In the absence of information on demography, data from large samples of carcasses can be useful as a source of information to help formulate conservation models, but with appropriate cautionary interpretation. We also agree that ‘proper methodology and acknowledgement of potential sources of bias should be common practice’ (Ciucci et al., 2007). This was why we repeatedly emphasized the limits of our sample of dead wolves in our paper (Lovari et al., 2007). Ciucci et al. (2007) also mentioned the theoretical qualities that a sample of dead animals should have to reflect the live population structure. Because these qualities are quite hard to meet in field conditions, their conclusion was that the use of dead animals found in the wild should not be encouraged. If data on the live population are missing (as it is very often the case for larger Mammals), when demographic information is required (e.g. for population modelling), educated guesses become the most likely alternative. The question is whether such guesses are better than using information from carcasses even though this may contain biases. We think that, when data are needed for elusive or rare species, the soundest approach is to test demographic models on real data from the wild, even from carcasses, if information on the live population is lacking or very hard to obtain. Variation in carcass recovery effort is a potentially important source of bias but, in our study, the ‘recovery effort’ was approximately constant throughout the study period. Carcasses were recovered mainly (no o98%) by rangers of the State Forestry Service, whose units were evenly distributed in our study area, that is the historic core of the wolf distribution in Italy. By law (Ministry of Agriculture and Forestry, doc. n. 30668, 24th October 1991), all dead wolves recovered had to be delivered to a single agency, where one of us (R. F.) performed necropsies, thus ensuring a consistent evaluation (Lovari et al., 2007).