Inconstancy in predator/prey ratios in Quaternary large mammal communities of Italy, with an appraisal of mechanisms Pasquale Raia a,b, ⁎ , Carlo Meloro b , Carmela Barbera b a Dipartimento STAT Università degli Studi del Molise, Via Mazzini 10, 86170 Isernia, Italy b Dipartimento di Scienze della Terra, L.go San Marcellino 10, 80138 Napoli, Italy Received 15 February 2006 Available online 10 January 2007 Abstract Constancy in predator/prey ratio (PPR) is a controversial issue in ecological research. Published reports support both constancy and inconstancy of the ratio in animal communities. Only a few studies, however, specifically address its course through time. Here we study the course of predator/prey ratio in communities of large Plio-Pleistocene mammals in Italy. After controlling for taphonomic biases, we find strong support for PPR inconstancy through time. Extinction, dispersal events, and differences in body size trends between predators and their prey were found to affect the ratio, which was distributed almost bimodally. We suggest that this stepwise dynamic in PPR indicates changes in ecosystem functioning. Prey richness was controlled by predation when PPR was high and by resources when PPR was low. © 2006 University of Washington. All rights reserved. Keywords: Predator/prey ratio; Quaternary; Large mammals; Paleocommunities; Megaherbivores; Alternative community states; Predation Introduction The question of constancy in predator/prey ratio (PPR) in animal communities is an old one in ecology that is still almost unresolved (Rosenzweig, 1995; Spencer et al., 1999; Croft, 2006). Arnold (1972) suggested that the number of predators follows the number of prey. Thus, the ratio should be constant. Some studies have supported constant PPRs across commu- nities (Cohen, 1977; Jeffries and Lawton, 1985; Sugihara et al., 1989; Shoenly et al., 1991; Gaston et al., 1992; Warren and Gaston, 1992). Mithen and Lawton (1986) performed a computer-based experiment indicating PPRs are discrete, meaning that not all ratios are equally feasible. Inconsistent support for a constant ratio was reported by Simberloff (1976), and Valentine et al. (2002). More importantly, Wilson (1996) demonstrated that a constant ratio is expected by chance as one compares communities with different species diversities. Studies of PPR in extinct mammal communities are much less common (but see Croft, 2006; Fortelius et al., 1996; Van Valkenburgh and Janis, 1993). Van Valkenburgh and Janis (1993) analyzed patterns in PPR in Cenozoic large mammals from North America. They found evidence for strong variation in PPR through time. More specifically, Van Valkenburgh and Janis found that carnivore diversity did not increase at the same rate as herbivore diversity. They suggested that carnivores are less likely to share niche space and that this, in turn, would put a ceiling on their increase in number. In Van Valkenburgh and Janis' study, both the methods and the interpretation of results were based on the assumption that PPR reflects the effects of ecological interactions on species diversity in different trophic groups. It is worth mentioning that other authors have interpreted PPR differently. For instance, Gaston et al. (1992) noted that in most studies, the prey category often includes quarries for no predator. Consequently, they suggested using the term “non-predator” instead of “prey”. Similar arguments persuaded Valentine et al. (2002) to use the expression “carnivorous to non-carnivorous” for their marine gastropods. The spirit of our study is the same as that of Van Valkenburgh and Janis; we analyze the course of PPR through time in Plio- Pleistocene Italian large mammal communities. We present this study in two related parts. First, we show the results for the computation of PPRs and its validation, taking into account the influence of possible sampling biases. We then seek explanatory Quaternary Research 67 (2007) 255 – 263 www.elsevier.com/locate/yqres ⁎ Corresponding author. Fax: +39 081 552 09 71. E-mail address: pasquale.raia@libero.it (P. Raia). 0033-5894/$ - see front matter © 2006 University of Washington. All rights reserved. doi:10.1016/j.yqres.2006.10.005