Page 3 - THE_ISLAND_RULE_2012

THE ISLAND RULE IN LARGE MAMMALS 1733

TABLE 1. Fossil ungulate guilds on islands with more than one faunal complex included in this study.
Island Herbivores Carnivores Age
Crete Cervus dorothoensis absent late Pleistocene
Cervus major
Cervus rethymnensis
Elephas creutzburgi
Hippopothamus creutzburgi
Praemegaceros cretensis
Praemegaceros ropalophorus
Crete Mammuthus creticus absent early Pleistocene
Hippopothamus creutzburgi
Sardinia Sus sondaari Chasmaportetes melei, Cynotherium sp. late Pliocene—early Pleistocene
Nesogoral sp.
Caprinae indeterminate
Sardinia Praemegaceros sp. Cynotherium sp. middle Pleistocene
Caprinae indeterminate
Sardinia Mammuthus lamarmorae Cynotherium sardous late Pleistocene
Praemegaceros cazioti
Sicily Bison priscus sicliliae Crocuta crocuta, Ursus arctos, Panthera leo, Canis lupus middle to late Pleistocene
Bos primigenius siciliae
Cervus elaphus siciliae
Dama carburangelensis
Elephas mnaidriensis
Equus hydruntinus
Hippopotamus pentlandi
Sus scrofa
Sicily Elephas falconeri absent early Pleistocene
Cyprus Elephas cypriotes absent middle Pleistocene
Phanourios minor
Corse Praemegaceros cazioti absent late Pleistocene
Corse Praemegaceros cazioti Cynotherium sardous, Canis sp. middle Pleistocene
Cervus elaphus rossii

erature, supplemented by measurement of Sicilian species mainland population was used as a reference to calculate size
and Cervus elaphus thyrrenicus taken by P. Raia (Table 1; reduction indices (SR; size on island/size on the mainland)
Appendix 1 available online only at http://dx.doi.org/10. in all insular descendants (Appendix 2, available online only
1554/05-664.1.s1). We excluded small mammals because of at http://dx.doi.org/10.1554/05-664.1.s2; and Appendix 3,
their patchy fossil record (Damuth 1982). We did not test for available online only at http://dx.doi.org/10.1554/05-664.1.
the effects of island area on fossil ungulate size as Pleistocene s3).
island areas are not reliable. We include island area as a Some islands had more than one faunal complex (i.e., chro-
predictor when analyzing carnivore sizes. We used the length nologically distinct fossil assemblages): Sardinia, Sicily, and
of the third lower molar (M 3 ) as an index of size, because Crete had four each. In these islands there were cases of in
this measurement is well correlated with body size in many situ speciation (deer in Crete, a caprine in Sardinia). Some
mammalian families (Damuth and McFadden 1990). When colonization events were recurrent: dwarfism of the straight-
this was not reported we used the length of the lower molar tusked elephant E. antiquus occurred on Tilos, Crete, Sicily
row or the third upper molar. A separate size index was (twice), Malta (twice), Rhodos, Cyprus, Naxos, Delos, and
derived from long bone measurements: we used metatarsal Favignana. Similarly, the megacerine deer Praemegaceros
lengths, and when these were unavailable either tibiae or verticornis colonized Sardinia, Corsica, Crete, Kasos, and
metacarpal lengths. Teeth and bone measurements other than Karpathos. Red deer (Cervus elaphus) inhabited Crete, Cor-
third molars and metapodials lengths can admittedly be better sica, Capri, Malta, and Sicily. Hippos inhabited Cyprus,
predictors of body size (Janis 1990; Scott 1990). Yet, M 3 is Crete, Malta, and Sicily. Large carnivores include the running
the most morphologically distinctive tooth and could be eas- hyena, Chasmaportetes melei, and the canid, Cynotherium
ily recognized (and measured) in both ruminants and pro- sardous, on Sardinia and cave hyena Crocuta crocuta, wolf
boscideans (the third molar of elephants has distinctive shape (Canis lupus), brown bear (Ursus arctos), and lion (Panthera
because, being the last to erupt, it is not pushed forward by leo) on Sicily. We followed the biostratigraphic accounts of
any other tooth; thus, its rear part is elongated). Similarly, Dermitzakis and Vos (1987; Crete), Kotsakis (1990; other
metapodials are easily recognized and often distinctive East Mediterranean islands), Abbazzi et al. (2004; Sardinia),
among ruminants and are more abundant in fossil samples and Marra (2005; Sicily). The species we included date from
than any other long bone. The choice of these measurements the earliest Pleistocene (oldest faunal complex in Sardinia
therefore maximizes both the number of species included and with C. melei) to the latest Pleistocene (Crete largest deer in
sample size for each species. Measurements of mainland spe- Simonelli Cave, 32,500 20% years before present [BP];
cies were preferentially taken from Mediterranean popula- Karpathos deer, 14,320 20% years BP; Reese et al. 1996)
tions (Italy, Greece, Spain, and southern France). The same and mid-Holocene (one record, pigmy Tilos elephant, 4390