Observations on the incidence and taxonomic significance of divided supracaudal scutes in Mediterranean tortoises genus Testudo.
A C Highfield & J. R. Bayley
Key Words: Testudo graeca; Testudo hermanni;Taxonomy; Diagnostic characters; Field populations.
Introduction
The presence or absence of a divided supracaudal scute is regularly cited in many scientific and popular works of reference as a factor useful in taxonomic diagnoses within the genus Testudo. It is usually stated that whilst Testudo hermanni possesses this character, Testudo graeca does not (Ernst and Barbour, 1989; Pritchard, 1979). Stubbs (1989) states of T. graeca that "reliable distinguishing features are an undivided supracaudal (and) absence of horny terminal tail scale...". Alderton (1988) is more absolute in his diagnostic criteria, stating of T. graeca that "the supracaudal is never divided".
Materials and methods
The following data, from field observations in Turkey and in North Africa, supplements an earlier report which demonstrated that captive T. graeca which possess a divided supracaudal are occasionally encountered and that among captive-bred siblings it is not unusual to find some individuals with a supracaudal division and others without (Highfield, 1989).
Discussion
It is evident that the statement that Testudo graeca complex tortoises invariably feature an undivided supracaudal scute is not supported by this data. Supracaudal division occurs in a small but significant minority of T. graeca complex specimens in both North Africa and Turkey. These results should be compared with data from Testudo hermanni where a divided supracaudal is considered normal for the species. Out of 300 wild Testudo hermanni hermanni Gmelin 1789 recorded by the authors in France, Spain and Italy, all featured a divided supracaudal. The only recorded instance of an undivided supracaudal in this race is believed to be that observed in a specimen on Mallorca (Bruekers, 1995 and pers. comm.). However, in the case of Testudo hermanni boettgeri Mojsisovics 1889, the situation is more complex. From a total sample of 438 specimens examined by the authors (including captive specimens in European collections and wild specimens in Greece), 78 individuals (18%) possessed an undivided supracaudal. Petzold (1966) reported that between 8-10% of eastern Adriatic populations featured an undivided supracaudal whilst the supracaudals of the rest were divided. Meek and Inskeep (1981) also found that 12% of a population of Montenegran T. hermanni possessed a supracaudal which, when viewed laterally, appeared to be undivided. These authors stated, however, that a division was visible when the supracaudal was subsequently viewed ventrally. This latter observation is not consistent with the condition of the supracaudal scute of any Testudo hermanni boettgeri examined by the present authors and may possibly represent a further regional variation. A much earlier report, frequently overlooked, is that of Siebenrock (1906) who noted that the supracaudal of T. hermanni is "highly variable, and can frequently be undivided". In this report Siebenrock refers specifically to 3 museum specimens from Herzegovina and 2 from Albania with undivided supracaudals.
Conclusion
Although numerous works of reference continue to cite the condition of the supracaudal as a reliable diagnostic character for determining Testudo species, these results indicate that the character is much more variable than is generally believed and that earlier results from captive-bred specimens (Highfield, op. cit.) are closely paralleled in wild populations. Whilst a divided supracaudal is consistently present within T. h. hermanni, it is highly variable in T. h. boettgeri and is also found in a minority of T. graeca complex specimens, thereby limiting the value of this character in taxonomic diagnosis.
LITERATURE CITED
A total of 112 T. (g.) ibera Pallas 1814 were recorded in Turkey during a 1 week survey in June 1993. Of this total, 3 individuals possessed a divided supracaudal scute (2.68%). A further 23 specimens were recorded in September 1993 in the region between Antalya and Aspendos. Of these, 2 individuals possessed a divided supracaudal (8.70%). Of an additional 6 specimens studied in the vicinity of Antakya (Antioch) in Southern Turkey, 1 individual possessed a divided supracaudal (16.66%). The overall incidence of supracaudal division in the tortoises surveyed in Turkey was therefore 4.26%.
A similar survey was subsequently undertaken in southern Morocco during December 1994 (tortoises in this region do not hibernate and are very active during the mild winter). Field densities in Morocco are considerably lower than those in Turkey (Lambert, 1982) and the available sample for T. g. graeca L. 1758 is therefore lower than for Turkish T. (g.) ibera. A total of 39 tortoises were located, measured, weighed and photographed. Of the specimens studied, 3 exhibited a divided supracaudal, forming 7.70% of the total sample.