Ronald E. Willemsen

Ronald Willemsen has studied Mediterranean tortoises since 1975, at first in Greece, and since 1990 mostly in Italy.

Ronald E. Willemsen
MonteCassinostraat 35
7002 ER Doetinchem
The Netherlands

Some differences in morphology between Greek, Italian and Corsican tortoise populations

During his visit to the Galapagos Islands in 1835 Charles Darwin was told by the Ecuadorian governor of the islands that giant tortoises, Geochelone elephantopus, of different islands had different sizes and shapes (Darwin, 1859). Darwin's conclusion was that these differences were the results of natural selection under different environmental circumstances.

In the Balkans Testudo hermanni is the most common tortoise species. It is found all over Greece from sea level to up 1500 metres in the mountains of the Peloponnese (Willemsen & Hailey, 1989). Greece is divided into many regions, each with different climatic conditions and habitats, by mountain ridges which are impassable (or only with difficulty) by tortoises. The Greek geographical situation is comparable with that of the Galapagos island Isabela (Albemarle) where each of its five volcanos has a different form of tortoise (Thornton, 1971). It is thus possible that similar morphologcal differences would exist in Greek T. hermanni and perhaps in other Greek tortoise species as well. Differences in body size and plastral pigmentation between western T. hermanni (T. h. hermanni, formerly T. h. robermertensi) and the eastern T. hermanni from the Balkans (T. h. boettgeri, formerly T. h. hermanni) were already described by Wermuth (1952) and Stemmler (1963, 1968). In T. h. hermanni two bands of dark pigment are found on the plastron (Tuscany, above) and the body size is smaller than in T. h. boettgeri. In T. h. boettgeri spots of often diffuse pigment (Meteora, right) occur on the plastron, but the pigmentation pattern in this subspecies is more variable than the western subspecies where the two-banded pattern is found in almost all individuals.

In 1975 the research in Greece began as a trial near Olympia. During that time it became clear that the marking and recapture method was working well and that differences in plastral pigmentation existed in that population. In the following years work was done at more sites in Greece (Willemsen & Hailey, 1989). A correlation was found between latitude and altitude and body size, with the largest animals found in northern Greece or at high altitude; in fact in the coolest environments (Willemsen & Hailey, 1999a). The quantity of plastral pigmentation decreased from the south to the north, but did not vary with altitude (Willemsen & Hailey, 1999b). No significant differences in carapace pigmentation were found between populations. There was often, but not always, a correlation between the quantity of plastral and carapace pigmentation; animals with a dark carapace can have little pigmentation on the plastron and vice versa. More information about body size and ecological differences can be found at:, which also describes conservation studies on Greek tortoise populations. The original research site at Olympia was abandoned in the early 1980s after tortoises were killed by pesticides (Willemsen & Hailey, 2001). The main research site was then Meteora in central Greece. Parts of this site have also been destroyed, by fire and mechanical clearance of the vegetation. The area shown at the right had a high population density of T. hermanni till 1984. The habitat was gradually degraded, and in in the last two years was completely destroyed by flattening and house building. A fragment of the original habitat seen in the right of the picture still had many tortoises in 2001 but was being filled with rubbish.

In addition to differences in the quantity of pigment on the plastron, there are also significant differences in pigmentation patterns between populations, which could be useful for indentifying the origin of a sample of tortoises. On the Peloponnese, southern Greece, T. hermanni is small-sized and often with a heavily pigmented plastron (Willemsen & Hailey, 1999a,b). There are large dark areas on the plastron; two, often wide, bands of dense black pigment are found in most animals (Sparta, left). A totally black (melanistic) plastron is frequently seen in all Peloponnese T. hermanni populations. Diffuse plastral pigment, which is common in northern mainland populations, is also found in all Peloponnese populations, but the pigmented area is in almost all cases larger and darker than in northern populations.

Along the west coast of the Greek mainland T. hermanni is medium-sized (Willemsen & Hailey, 1999a,b). Separated spots of black pigmentation are found on the plastron in most animals (Parga, right). On the east coast, animals are also medium-sized with the spotted pattern of the west coast, but the two-banded pattern is here more common than on the west coast. In most animals the pigmented area is smaller than in tortoises from the Peloponnese, for both black spots and bands. In both western and eastern coastal populations a substantial number of animals were found with diffuse pigmentation. Medium-sized T. hermanni are found in lower altitude areas of central Greece, and most animals have spots of diffuse pigmentation on the plastron. Animals with little or no pigmentation are common (see photo of 5 animals at Meteora - the second tortoise figure above). In many animals the gular and anal scutes lack pigmentation, and even animals with little or no pigment on the humeral and femoral scutes occur. These animals, if they have more than some small spots of diffuse pigmentation, have a small and short band of diffuse or dark pigment. The pigment on the abdominal scute is often radiating to the head, a pattern which is common in all populations in central and northern Greece, but is extremely rare in southern Greece. In all populations individuals with a coastal and even a southern pigmentation pattern were also found. Large-bodied T. hermanni occur in northern Greece, as described at Deskati, Agios Dimitrios and Kastoria, but large animals are also common along the Albanian border between Neapolis and Eptachorion. This type of tortoise was also found in the region around lake Ochrid and up to Skopje in the former Yugoslavia. The patterns are like those in the medium-sized T. hermanni from central Greece, but the short-band pattern is more common here (Deskati, left).

The T. hermanni populations in north-east Greece (Mikri Volvi and Keramoti) are also large-sized (Willemsen & Hailey, 1999a,b), although those from Mikri Volvi are rather smaller than the largest animals. In these populations most animals have spots of diffuse pigment and unpigmented animals are rather common (Mikri Volvi, right). T. hermanni from the island of Euboea seem to be different from those of the Greek mainland. Only a small sample was found on that island, so it is difficult to say if these results are conclusive for that island. The males were small-sized and with a dark plastron like those from the Peloponnese, while females were larger than expected and less pigmented than those on the Peloponnese (Kymi, left). The body size and plastral pigmentation patterns will make it possible to approximately determine the location of a sample of Greek T. hermanni and those from the south-western part of the former Yugoslavia. For individuals this is only possible in a very crude way, for all types of pigmentation are found in each population, although often very rare. But in adults the combination of body size and pigmentation pattern will give an indication of the origin even for one individual. Determination will only be possible in adult tortoises, for juveniles from all populations are much the same.

The morphology of T. marginata in Greece is puzzling. It is difficult to collect data from a large sample of different populations, and numbers in a population are often small. It seems that there could be a similar body size gradient as in T. hermanni in Greece, but with a smaller range. In the south-western part of the Peloponnese dwarf T. marginata are found (Bour, 1996; Artner, 1996). The real status of these dwarf T. marginata is still unclear, but it is not excluded that they will be a separate subspecies. There are no differences in pigmentation between T. maginata populations. Pigmentation on the plastron does not change during life; carapace pigmentation increases, then decreases again in old age (Bringsøe, Buskirk & Willemsen, 2001). No morphological differences are found between Greek T. graeca populations.

On the mainland of Italy all T. hermanni populations are small and scattered and all belong to the western subspecies T. h. hermanni (Bour,1987; Ballasina, 1995; Willemsen, 1990). Only in Tuscany and Latium are populations with high and medium density found. In Campania, Calabria and Puglia small low density populations are still found. Almost all mainland populations are threatened by development, illegal sampling, forest fires and by predation from increasing populations of introduced central European wild boar (Sus scrofa). It seems that there will be a similar geographical gradient in body size to the Greek populations (Willemsen, 1990). Animals from the lower altitude parts of Tuscany are small-sized, compared to the Greek tortoises, and those from the higher parts are medium-sized. Animals from Calabria are the smallest known T.hermanni (Willemsen, 1990). Adult females and males are often hardly larger than 10 cm, the usual juvenile limit for T. hermanni. However, the scattered and often very small and low density populations surviving today make it impossible to construct a cline as for Greek T. hermanni. Pigmentation is almost uniform in Italy. There are two bands of black pigment on the plastron, similar to but often narrower than in the Peloponnese (Tuscany, above). Few animals have spots of diffuse pigment on the plastron or a totally black plastron. The carapace is more darkly pigmented than in most of the Greek animals, but in some Greek populations dark animals are also rather common. Strangely T. hermanni on Sicily and Sardinia are larger than those of southern mainland Italy. Populations on Sicily have medium-sized animals (Ballasina & Willemsen, unpublished data). In these populations the two-banded pigmentation pattern is predominant (Sicily, right). The situation on Sardinia is puzzling as individuals as large as the largest Greek tortoises are common (Fritz et al., 1996), but medium-sized animals also occur. Plastral pigmentation is like the Sicilian and mainland Italian populations.

The situation is also puzzling on Corsica, where medium-sized and large-sized animals are found. The larger animals have the shape of T. h. boettgeri (Joubert & Cheylan, 1989) but according to plastron morphology they all belong to the western subspecies. The pigmentation pattern is mixed. Some animals have the two-banded pattern of the western subspecies, others have spots of often diffuse pigment like central Greek T. h. boettgeri populations (Corsica, left). Along the Italian west coast scattered small populations of T. marginata are found. It is likely that this species was introduced by the Etruscans, which is also likely to be responsible for the populations on Sardinia. The Italian and Sardinian T. marginata are often larger than the Greek specimens (Ballasina & Willemsen, unpublished data). The status of the T. graeca which occur on Sardinia (Ballasina, 1996; Buskirk,1996; Fritz et al., 1996) is unclear. It is most likely that these populations have been introduced by man, but they could have different sources; North Africa, Greece or Asia Minor.

The classical taxonomy of tortoises is based on differences in shape, body size, structure of the shell and skull, and coloration . Many of these characteristics could have environmental causes, so taxonomy will always be questionable, especially at the level of subspecies and below. The new techniques of gene sequencing have opened the possibilty to study differences in DNA and RNA, and comparing taxa at all levels, even individuals. The costs of these analyses make it impossible to use large samples, which limit the results in most research. Recently a study was made of the phylogenetic relationships among the species of the genus Testudo using mitochondrial rRNA (van der Kuyl et al., 2002). This study confirms most, but not all, of the results of classical taxonomy.

It is generally assumed, though not by herpetologists, that recent tortoises are ancient animals, remainders of the era of the dinosaurs. But fossil records and the RNA analysis of all living Testudo sp., the large African Geochelone sp. and the two South American forest Geochelone sp. show that these are recent species on a geological time scale. The evolution of the genus Testudo is monophyletic, as shown in the tree cluster diagram. The earliest division of the Testudo group did not occur before the middle of the miocene, about 20 million years ago, if a slow mutation rate of mitichondrial RNA is assumed; if a fast rate is assumed then the division occurred in the villafranchian period at the begining of the pleistocene, 2.6 million years ago. The miocene option looks the most likely, for during that period the climate became drier and caused open landscapes with steppe vegetiation, which is favourable to tortoises. These changes in climate and vegetation caused an explosion of mammal species adapted to the new habitats. At that time there were also routes passable by tortoises from Europe to Africa and the American continents. The fossil tortoise record of Europe, Asia and America shows a rich variation of species in the miocene and pliocene, often of very large size, which disappeared rapidly when the ancestors of man appeared.

A second division occurred with separation of the ancestors of T. hermanni and T. horsfieldii at the beginning of the pliocene about 11 million years ago; later if the mutation rate is fast, about 1.4 million years ago in the pleistocene at the beginning of the ice ages. The division of T. graeca was not before the pleistocene, 2.6 million years ago, and possibly as late as 0.3 million years ago between two ice ages. The separation of T. marginata and T. kleinmanni occurred in the pliocene, at the earliest 9 million years ago, or at the latest in the early pleistocene, 1.1 million years ago. It is likely that during the last ice age T. hermanni in western Europe had only one small refugium in the very south of Italy. From that refugium the species crept to the north and colonized the coastal area of Italy and France. There are no real barriers for this species along the Italian and French coast. A small ancestral stock of T. h. hermanni can explain why this subspecies is so uniform; no population can be older than 10,000 years. It is likely that there were more refugia in the Balkans during the last ice age, so this can explain the larger variation in the Balkan subspecies, T. h. boettgeri. It is also likely that all tortoise populations on islands such as Corsica, Sardinia and Mallorca are human introductions. No morphological differences were found between T. h. hermanni from Mallorca and Tuscany (unpublished data), but large differences were found between T. hermanni from Sardinia and Corsica. Their body sizes are like the large animals of the Balkans, but all their other morphology is like the western subspecies. If these populations were introduced by man then it happened long ago and these animals had a rather fast evolution. More samples and sequencing of different genes could give more detailed data about relationships and migration of Mediterranean tortoises.


Artner, H. (1996). Beobachtungen and der Zwerg-Breitrandschildkröte Testudo weissingeri in Messenien/Griechenland und Diskussion über die Validität ihres Artstatus. Emys 3, 2-5.

Ballasina, D. (1995). Distribuzione e situazione delle tartarughe terresstri in Italia. In BALLASINA, D. (Hrsg.): Red data book on Mediterranean Chelonians. Bolognia (Ediagricole), pp 147-168.

Ballasina, D. (1996). Conservation and reproduction techniques at the Carapax Center, Italy. In: Proceedings of the International Congress of Chelonian Conservation , Gonfaron, 6th July 1995. Gonfaron (SOPTOM), pp 210-213.

Bour, R. (1987). L'identité des tortues terrestres européennes: spécimens-types et localités-types. Rev. Fr. Aquariol. 13, 111-122.

Bour, R. (1996). Une nouvelle espèce de tortue terrestre dans le Peloponnèse (grèce). Dumerilia 2, 23-54.

Bringsøe, H., Buskirk, J. & Willemsen, R. E. (2001). Handbuch der Reptilien and Amphibien Europas. Testudo marginata Schoepff, 1792 - Breitrandschildkröte. Herausgegeben von Uwe Fritz. AULA-Verlag Wiebelheim.

Buskirk, J. (1996). On the absence of spur-tighted tortoises, Testudo graeca, from Egypt. Chelon. Conserv. Biol. 2, 118-120.

Darwin, C. (1859). The origin of species. John Murray, London.

Fritz, U., Petters, G. W. & Matzanke, M. (1996). Zur Schildkrötenfauna Nordsardiniens, Teil 2 Herpetofauna 18, 5-14.

Joubert, L. & Cheylan, M. (1989). La tortue d'Hermann de Corse. Trav. Sci. parc. Nat. Reg., Rés. Nat. de Corse 22, 1-54.

Stemmler, O. (1968). Zur Kenntnis von Testudo hermanni in tyrrhenischen Gebiet. Aquar. Terrar. 7, 49-52.

Stemmler-Gyger, O. (1963). Ein Beitrag zur Brutbiologie der mediterranen Landschildkröten. Aquar. Terrar. Z. 16, 181-183.

Thornton, I. (1971). Darwin's Islands: A natural History of the Galapágos. The Natural History Press. Garden City, New York.

van der Kuyl, A. C., Ballasina, D. L. P., Dekker, J. T., Maas, J., Willemsen, R. E. & Goudsmit, J. (2002). Phylogenetic relationships among the species of the genus Testudo (Testudines: Testudinidae) inferred from mitochondrial 12S rRNA gene sequences. Mol. Phylog. Evol. 22, 174-183. abstract

Wermuth, H. (1952). Testudo hermanni robertmertensi n. subsp. und ihr Vorkommen in Spanien.Senckenbergiana 33, 157-164.

Willemsen, R. E. (1990). The status of Testudo hermanni hermanni in southern Tuscany (Province Grosseteo and Livorno) WWF of Tuscany. Firenze.

Willemsen, R. E. & Hailey, A. (1989). Status and conservation of tortoises in Greece. Herpetol. J. 1, 315-330. abstract

Willemsen, R. E. & Hailey, A. (1999a). Variation of adult body size of the tortoise Testudo hermanni in Greece: proximate and ultimate causes. J. Zool. Lond. 248, 379-396. abstract

Willemsen, R. E. & Hailey, A. (1999b). A latitudinal cline of dark plastral pigmentation in the tortoise Testudo hermanni in Greece. Herpetol. J. 9, 125-132. abstract

Willemsen, R. E. & Hailey, A. (2001). Effects of spraying the herbicides 2,4-D and 2,4,5-T on a population of the tortoise Testudo hermanni in southern Greece. Environ. Pollut. 113: 71-78. abstract

With Geochelone sulcata at the Centro Carapax, Italy. Other interests include model ship building. The steel four-masted barque Herzogin Cecilie was built in Germany in 1902 and wrecked at Ham Stone, Devon in 1936. The model is now in a museum at Åland, Finland.

Other models:

© Copyright R. E. Willemsen 2002-. Webmaster A. Hailey