Page 2 - THE_ISLAND_RULE_2012

1732 P. RAIA AND S. MEIRI

The Theoretical Framework the risk of intraguild predation. Moreover, prey size is of
crucial importance to carnivores, because carnivores need to
We predict that ecological interactions affect carnivores be large enough to subdue their prey. Therefore, even in the
and herbivores differently, with dwarfing being more com- absence of competitors, size reduction may not be adaptive,
mon and more drastic in herbivores, because of dietary dif- unless large prey is absent (Jessop et al. 2006). Furthermore,
ferences and the peculiarity of intraguild killing in carnivores. islands often offer extremely abundant resources in the form
Size reduction in mammals often entails an increase in life-
of carcasses, fishes, and marine bird nests (Case and Schwa-
time offspring production (Calder 1996, p. 248). Offspring
ner 1993; Goltsman et al. 2005). Thus, resource abundance
number, however, is negatively correlated with average vi-
should be much more frequently important to carnivores, and
ability, with large litters facing increased mortality (Prom-
size would evolve accordingly.
islow and Harvey 1990). Thus, life history remains a matter We therefore predict different evolutionary trajectories for
of strategy (Roff 2002), and no size confers an a priori ad- carnivores and herbivores with the most remarkable cases of
vantage (Blackburn and Gaston 1996, Kozłowski and Ter- size reduction occurring in the latter. We predict herbivore
iokhin 1999). Raia et al. (2003) suggested that size reduction size will be related to the prevailing competition and pre-
of large insular mammals depends on selection for greater dation pressures. We expect carnivores to evolve to be small-
reproductive investment under condition of reduced external er or larger on islands primarily in relation to the relative
mortality. Raia and colleagues argued that at the individual size and abundance of their prey. We do not expect ancestral
level organ growth and repair compete with reproductive ef- body size to be a major predictor of size change, except
fort (Holliday 1989; Mangel and Stamps 2001). As a con- through its influence on the above factors. Four predictions
sequence, increased energy allocation toward reproduction arise from our theory.
occurs at the expense of growth. Impressively high numbers The biological interaction effect hypothesis (fossil ungu-
of juveniles seem to point to intense reproductive effort in lates). If competition and predation drive the extent of size
insular dwarf fossil mammals both in the absence (e.g., Mal- decrease, we would expect different degrees of size reduction
atesta 1980; Raia et al. 2003) and presence (Abbazzi et al. for the same species depending on the presence and the size
2004) of predators. Unfortunately, data are generally too of competitors and on the presence of predators. Because
scarce to be tested statistically, but it seems that herbivores smaller competitors can prevent size decrease (Dayan et al.
that actively defend themselves and their calves from pred- 1989), we expect the degree of dwarfism to be higher in the
ators (e.g., elephants) retain their ancestral size if carnivores absence of smaller competitors.
are present. Competition similarly affects offspring mortality Overdispersion hypothesis (fossil ungulates). If compe-
because competitors reduce resource availability. Poulakakis tition is important, body size should be overdispersed in mul-
et al. (2002) and Raia et al. (2003) suggested that different tispecies assemblages (i.e., size ratios between species should
levels of dwarfism in Pleistocene elephants on Mediterranean be more equal than expected by chance; Dayan and Simber-
islands depended on the presence of smaller species with loff 2005). Late colonizing/new species should remain large
overlapping diets. Raia et al. (2003) specifically advocated if smaller competitors are present because of niche incum-
the incumbency of smaller competitors in the same niche bency (Rosenzweig and McCord 1991). In our data there is
(Rosenzweig and McCord 1991) and predation to explain only one case where this prediction is testable (Crete fossil
differential dwarfism in the straight-tusked elephant Elephas deer).
antiquus in fossil assemblages on Sicily. In carnivores, in- Carnivore resource-competition hypothesis (extant carni-
traguild predation and interference competition could se- vores). We expect body size change in insular carnivores
verely counterbalance any fitness advantage accrued by size to depend on the presence of potential predators, smaller and
reduction. Large carnivores often kill smaller ones (see, e.g., larger competitors, dietary preferences, and the nature of the
Palomares and Caro 1999; Creel and Creel 2002). Often this resource base. Size is predicted to evolve toward that of miss-
results in smaller carnivores living in peripheral sink popu- ing competitors, decrease in the absence of predators, and
lations that are relatively predator free (Mills and Gorman correlate positively with the size and abundance of available
1997; Creel and Creel 2002). Intraguild predation may be prey.
even more influential on islands, where predator-free space Sexual size dimorphism hypothesis (extant carnivores only
is scarce and sink populations are not viable (Rosenzweig because fossils cannot be accurately sexed). A greater in-
vestment in reproduction should be paralleled by increased
1995).
sexual size dimorphism (SSD). This could occur because
Size differences have been claimed to promote coexistence
males gain less than females from size reduction. Large size
in herbivores (Bell 1971; McNaughton and Georgiadis 1986).
is often an effective mean of displacing rival males at mating,
Large herbivores can access low-quality food because at least
up to about rhino size, gut retention time scales positively and reproduction is less costly for males. We therefore predict
females of insular dwarves will decrease more in size relative
with body size (Demment and Van Soest 1985; Illius and to males of the same populations, thereby increasing SSD on
Gordon 1992; Clauss and Hummel 2005). Thus, in classical islands.
‘‘grazing chains’’ (Bell 1971), larger species feed on the low
quality forage, permitting smaller species to exploit the high-
MATERIALS AND METHODS
er quality parts of the plants. This coexistence mechanism
Pleistocene Ungulates of the Mediterranean Islands
should accrue to island herbivores as well.
In carnivores, the presence of similar-sized or larger com- We gathered data on large mammals of Plio-Pleistocene
petitors could impede the onset of dwarfism by increasing assemblages of different Mediterranean islands from the lit-