The status and nomenclature of Testudo (graeca) terrestris FORSKAL 1775: with additional notes on some little known land tortoises of Israel including Testudo floweri BODENHEIMER 1935
A C Highfield
THIS ARTICLE IS PUBLISHED ON THE INTERNET FOR SCHOLASTIC USE ONLY. IT IS ©1998 BY THE AUTHOR.
Summary: The taxonomic and nomenclatural status of the alleged middle-eastern sub-species of Testudo graeca, Testudo graeca terrestris is discussed with reference to a recent re-examination of an original copy of 'Descriptiones Animalium' by Petrus Forskal published in 1775. The confusion between Testudo floweri BODENHEIMER 1935 and the 'terrestris' form is also considered.
According to the IUCN publication "The Conservation Biology of Tortoises" (eds. Swingland and Klemens, 1989), which may be considered representative of currently accepted taxonomic opinion with regard to Testudo, there occurs in Asiatic Turkey, northern and western Syria, Lebanon, western Jordan and northern and central Israel a sub specific form of T. graeca named Testudo graeca terrestris which possesses as its principal diagnostic characters "a high and domed carapace, with a yellow spot on the superior and lateral sides of the head". The authority for this description is cited as Wermuth and Mertens (1961). This tortoise is also said to be "smaller and paler than T. g. ibera", and Stanley Flower (1933) is given as the authority for this observation. A total of 6 sub specific forms of T. graeca are listed in "The Conservation Biology of Tortoises", including T. graeca floweri. No reference is made by the author of the 'Conservation Biology of Tortoises' paper on the status of T. graeca , David Stubbs, to the original publication of this name by Bodenheimer in 1935 either in the relevant chapter itself or in the extensive accompanying bibliography. Little of a descriptive nature is said of this tortoise other than that it is "a small form with a pale yellow carapace and black central spots on its scutes". Adding to the confusion is Bodenheimer's own single sentence 'type description' of 1935 for Testudo floweri, which is also devoid of any meaningful scientific description or of any illustration.
This raises the matter of the inclusion of both T. g. floweri and T. g. terrestris thereby suggesting that these are two distinct forms; the fact is that Heinz Wermuth who 'resurrected' the nomenclature 'T. g. terrestris' in 1958 clearly believed that Forskal's 'Testudo terrestris' and Bodenheimer and Mertens' 'floweri' were the same taxonomic entity. That opinion may be mistaken, but as he is the author of the 'revived' name, and the first to apply it in modern times, his view that 'terrestris' and 'floweri' are identical, the latter being a mere junior synonym of the former, cannot simply be ignored.
The author of the 'Conservation Biology of Tortoises' paper on Testudo graeca (Stubbs, 1989) does not provide any references to support the suggestion that the two forms are different, nor does he list any authorities for dispensing arbitrarily with Wermuth's fundamental claim that 'terrestris' occurs in Libya - and yet in the provided table of distributions for 'terrestris' Libya is absent. This although Wermuth (1958) specifically identified his illustrated (plate 20) specimen (Senck. Museum reg. no. 36127) from Derna in Libya as a topotype of Testudo graeca terrestris and which - in addition - he firmly believed also represented Bodenheimer's Testudo floweri i.e. Wermuth states that Testudo graeca terrestris occurs in Libya, that it is exactly the same tortoise as described by Bodenheimer as Testudo floweri from Gaza (rendering that a junior synonym), and that this same tortoise also occurs in the rest of Israel, Syria and the Lebanon. Wermuth also states that the Testudo floweri form as named by Bodenheimer in 1935 had earlier been described and named by Petrus Forskal in 1775, and that therefore this earlier name is the correct one by reason of nomenclatural priority.
As to the presently accepted distribution of 'Testudo graeca terrestris', this alleged sub species is widely reported from Israel (Buskirk, 1966, Pritchard 1979, Arbel, 1984). However, these Israeli tortoises bear little resemblance to Wermuth's illustrated specimen from Derna, Libya (SM 36127) and equally are in no way similar to Bodenheimer's Testudo floweri (described in detail not by Bodenheimer, but by Stanley Flower in 1933).
The geographical attributions of T. g. terrestris to Turkey cited in the 'Conservation Biology of Tortoises' follow a later alleged discovery of the 'terrestris' race in that region by Herrn (1966) and subsequently also described by other authors (Freyhof, 1983, Eiselt-Spitzenberger, 1967). These accounts are fortunately illustrated and clearly the tortoises their authors are referring to bear no resemblance to Wermuth's specimen from Libya, although they closely match his specimen from Syria. Most significantly, nor do they bear any resemblance to Testudo floweri as described by Flower (1933) and named by Bodenheimer (1935).
It is also important to note that in his (1946) account of the alleged sub specific distributions and relationships of 'Testudo graeca' Robert Mertens refers not to Testudo graeca terrestris but to Testudo graeca floweri; "The S. Palestinian race of Testudo graeca, which the author unfortunately does not know from personal experience, is remarkable for its modest size and bright colouring and thereby very noticeably resembles Testudo leithii (syn = T. kleinmanni)"
It was not until 1958, when Heinz Wermuth revised and modified Robert Mertens' earlier (1946) account (with Mertens' agreement), that 'terrestris' effectively replaced 'floweri' as one of the 4 T. graeca sub species on the presumption that 'floweri' was a junior synonym of 'terrestris'.
There is one curious but very significant reference in Wermuth's (1958) account. He refers to the distribution of Testudo graeca terrestris as "SW Asia (Syria and Israel) and NW Africa (from N. Egypt westwards to Libya)" The author gives no details of any specimen derived from Egypt other than one "purchased by his secretary from a street vendor". It seems that this purchase is the sole basis for the claim that 'T. graeca terrestris' occurs "from N. Egypt....". This does not seem to represent a very satisfactory basis for a valid scientific record of distribution, Egyptian street traders being well known for their inventiveness concerning the antiquity or geographical origins of the goods offered to tourists.
I have recently examined colour slides of this specimen (CFM 82661), and it is rather typical of the general population of tortoises inhabiting Cyranaica, Libya where tortoises are found in some abundance (Zavattari, 1930). It does not conform in any way to the diagnosis of T. floweri as given by Flower (1933) and is different again from the typical tortoises of Israel as described by Buskirk (1966) and Arbel (1984). Whilst the true origins of CFM 82661 must remain a mystery, it should not be attributed to Egypt on merely on the basis of having been purchased there.
Finally, to further illustrate the depths of nomenclatural and taxonomic confusion surrounding this tortoise, it is worth pointing out that in Mertens' (1946) account of the alleged sub specific forms of T. graeca the selfsame specimen later used by Wermuth to resurrect T. graeca terrestris (SM 36127) features as one of the specimens used to differentiate T. graeca graeca from T. (g). ibera. There is no evidence that at any time Mertens examined specimens of T. graeca derived from the type locality of Oran, Algeria during his comparative 'T. graeca' studies. His overall sample was in any event extremely small, consisting of a mere 16 tortoises of north African origin (various localities from Morocco to Libya) and 21 specimens of T. ibera.
In the case of 'Testudo terrestris' we therefore have a name, but lack any adequate description of what it should be applied to, and have no type specimen or even any illustration of one contemporary with the claimed original description (Forskal, 1775). Very much the same applies to Testudo floweri, although in that case we can fortunately refer to a very detailed (earlier) description by S. S. Flower (1933) which the author of the name, Bodenheimer (1935), cites as his principal reference.
Identity of populations
The first problem to address is whether the tortoises from Palestine described by Flower in 1933 (and named by Bodenheimer in 1935) are the same as Wermuth later identified as 'terrestris' from Libya and Syria. This latter group is typified by two specimens that he used to illustrate his (1958) account (including SM 36127 which was collected by W. Sturmer on the 29th of September 1941). There really is not much difficulty in resolving this. Flower's discussion is revealing and is quoted in full:
'"Canon H. B. Tristram (1888,p. 156) noted that two forms of Land-Tortoise occurred in Palestine. Of "Testudo ibera" (i.e., T. graeca) he wrote:- "This is the common tortoise of the Holy Land, and is found in every part of the country, quite irrespective of the nature of the soil, till we reach Hebron. The hill country of Judea appears to be its southern limit, south of which and of the Dead Sea it does not occur." Of "Testudo kleinmanni" that it "is the Tortoise of the region between Hebron and Bersheeba, and of the Arabah, south of the Dead Sea."
Actually these two forms are both races of Testudo graeca, and, as far as my experience goes, are two extreme forms, Syrian specimens attaining larger dimensions and darker colour than the typical Moroccan, or Mauritanian, race, whilst specimens from southern Palestine are smaller and lighter coloured than any other tortoises of the species in all the countries of its wide distribution.
Tristram's second form is not T. kleinmanni (which is a synonym of T. leithii), but is the smallest known geographical race of T. graeca, which has superficial resemblance to T. leithii in its small size and yellow colour, but from which it is easily distinguished at close quarters. Among the points of difference are:-
Major M. Portal, D.S.O., very kindly gave me four adult specimens that he collected himself in 1917 and 1918 in the coastal plain between the neighbourhood of Gaza and Bir Salem, near Jaffa; these were in length of carapace in a straight line in median line respectively 112, 123, 128 and 132mm. This largest one was a male."'
- 1. Scales on front of arm, number of longitudinal series: T. graeca, 4 to 6; T. leithii, 3.
- 2. Spur on back of thigh: T. graeca, always present; T. leithii, none.
- 3. In the south Palestine form of T. graeca each vertebral and costal shield has a conspicuous black spot near the centre of the shield; in T. leithii there is no trace of these centre spots.
The dimensions cited bear out Flower's comparison to T. kleinmanni and it should further be noted that in the published description by Bodenheimer the size of T. floweri is said to be "tiny".
I am grateful to James Buskirk for providing me with a series of slides of the floweri form in the wild. These slides tend to establish the accuracy of Stanley Flower's description quoted above. I was also very fortunate in being able to examine and photograph the sole surviving museum specimen of Major Portal's s examples, donated to the British Museum by Stanley Flower. This specimen is not labelled 'Testudo floweri', but instead carries two paper tags in the same hand. I am also grateful to the Senckenberg Museum for allowing me unrestricted access to Wermuth's specimens and for permission to photograph them. A side-by-side comparison of the two series leaves little room for doubt that these are two very different tortoises. By no stretch of the imagination could the Senckenberg specimens be said to be "tiny" or even small. They are of average size for adult T. graeca (Libya) and T. g. ibera (Syria). The Frankfurt specimen SM 36127 for example, a male, measures 149.2mm. Other Israeli specimens of 'Testudo graeca terrestris' have been recorded as measuring up to 250mm Straight Carapace Length (Arbel, 1984). Buskirk (1966), in confirmation, states of Israeli 'Testudo g. terrestris' "This is a medium sized tortoise, reaching a maximum of twenty-five centimetres but averaging about fifteen centimetres". In no way could these tortoises be confused with, or compared to, Testudo kleinmanni.
This dimensional evidence alone casts serious doubt therefore, upon Wermuth's conclusion that the Flower-Bodenheimer tortoises and his own series were the same thing.
A brief chronology of the name Testudo graeca terrestris FORSKAL 1775
A fundamental issue concerns the 'resurrection' of Petrus Forskal's nomenclature Testudo terrestris by Wermuth in 1958. This point is an important one, as it lies at the base of all the subsequent confusion. A chronology of the usage is illuminating:-
- 1765 - Firmin publishes a description of what appears to be the Matamata, Chelus fimbriatus (Schneider 1783) from Surinam under the name Testudo terrestris.
- 1775 - Forskal refers to the presence 'Testudo terrestris' from the Lebanon and Syria in his Descriptiones Animalium. No illustrations are provided. There is no descriptive paragraph.
- 1933 - Flower describes a miniature race of 'graeca' from the Gaza strip region, roughly comparable in size to Testudo kleinmanni, but does not name it.
- 1935 - Bodenheimer names this tortoise, but does not describe it (other than to comment that it is "tiny"), after Flower calling it Testudo floweri.
- 1946 - Robert Mertens concluded, whilst admitting that he has no personal knowledge of the tortoise, that this must be a subspecies of Testudo graeca. He then 'relegates' it as a subspecies listing it as Testudo graeca floweri along with T. g. graeca, T. g. ibera and T. g. zarudnyi (this latter of which he also knew only as a single preserved specimen in the Vienna Museum). Before Mertens' intervention all the above 'subspecies' had been generally considered to be full species.
- 1958 - H. Wermuth concludes that T. g. floweri was previously described by Forskal in 1775, and that 'floweri' is therefore merely a junior synonym of Forskal's earlier designation. This should therefore take priority as 'T. g. terrestris'.
- 1966 - Testudo graeca terrestris further claimed from Turkey (Herrn, 1966).
- 1989 - T. g. terrestris continues to be claimed from Libya (Schleich, 1989).
Validity of nomenclatural acts relating to 'Testudo terrestris' and Testudo floweri
For a resolution to this problem, the first point of reference should be to Petrus Forskal's 'Descriptiones Animalium' of 1775 of which I have fortunately been able to examine a rare original copy. This is an important text as it forms the basis of Wermuth's resurrection of the name Testudo terrestris as a specific scientific name. The results of this examination were surprising.
The 'type description' of a tortoise that should exist on pages 12 and 13 of the book does not in fact exist. Instead there is a compound paragraph of miscellaneous observations relating to land tortoises generally under the heading 'Testudo terrestris' and also - very significantly - to marine turtles under the heading 'Testudo marina'. The content of these miscellaneous observations is primarily directed toward the medicinal values of the eggs of tortoises and turtles, which are said to be excellent for colic and fever. Forskal records observing land tortoises at Aleppo, in Syria and also in the Lebanon. There is nothing in these pages which could conceivably constitute a type description as required by the normal rules of zoological nomenclature. There is also no illustration of 'Testudo terrestris' or 'Testudo marina' and no reference to the collection or designation of any specimen or type.
That this is the case is further confirmed by a careful study of the rest of Forskal's book for where he intended to describe and name an animal or plant there is invariably a paragraph with the heading 'DESCR.', this is always quite detailed, and follows the normally accepted format of presenting data on size, shape and colour, etc. This is then followed by a second paragraph headed 'OBS.' for general observations. These are typically quaint in nature, often featuring local folklore.
In the case of 'Testudo terrestris' there is no paragraph headed 'DESCR.', and therefore it seems probable that Forskal was in this instance not attempting a description. Only the paragraph headed 'OBS.' is present and this confines itself to a very general discussion as outlined above. It must also be noted that in Latin 'Testudo terrestris' simply means 'land tortoise' and the fact that it is used in the same paragraph as 'Testudo marina' or 'water tortoise' strongly suggests that here it is being employed in a generic and not in a specific sense. The same pattern is followed throughout the text.
There are several well-known precedents for this. Dumeril & Bibron (1835) in the Erpetologie General (Tome 2), for example, provide an extensive list of synonyms and miscellaneous names pre-dating Linnaeus that have been applied to various land tortoises from the time of Pliny. No fewer than five separate citations refer to 'Testudo terrestris'. The term again being employed generically rather than specifically. Marine turtles or 'Water tortoises' are similarly referred to as 'Testudo marina', as it appears, was Forskal's intent.
Latin is the commonplace language of Forskal's entire book. Whenever he refers to land or aquatic testudines he invariably refers to 'Testudo terrestris' or 'Testudo marina'. To interpret this as a nomenclatural act, particularly in the absence of any physical description is curious.
There is therefore a strong case for taking the view that as Forskal's publication of 1775 does not appear to constitute a valid type description according to the accepted rules of scientific zoological nomenclature, and noting that before Wermuth in 1958 no-one had ever claimed that it did, there is no justification for applying it (citing Forskal's authority) to any living tortoise either as a specific or sub specific scientific name.
The diagnostic criterion for the alleged subspecies 'Testudo graeca terrestris' is presently highly confused, and the name is currently being applied to a number of physically disparate tortoises from Libya, Israel, Syria, Jordan and S. Turkey. Few of these tortoises - if any - conform to the most important criteria specified by S. Flower in 1933 (and followed by Mertens in his 1946 paper on the alleged '4 subspecies' of T. graeca), that they must be of approximately similar dimensions as Testudo kleinmanni. In this author's personal experience numerous tortoises from North Africa have bright yellow head markings and light coloured carapaces (Highfield, 1990), as do some in Spain (Highfield, 1993) and Southern Turkey (Highfield, in press). This feature may be advantageous in certain bioclimatic conditions , for example, in very hot climates. All such tortoises have, in recent years, tended to be ascribed to 'Testudo graeca terrestris' without further reference to Flower's original descriptive publication (1933) and often ignoring any inconsistency with the dramatic dimensional criterion specified therein.
As to the validity of the name 'Testudo terrestris', if this was (as now appears likely) being used in the generic sense by Forskal for 'land tortoise', there would seem to be a good case, in view of the subsequent confusion, for withdrawing it from usage as a specific or sub specific scientific name. This would also encourage a reassessment of all those tortoises currently burdened with this uncertain attribution. The population of tortoises identified by Flower in 1933 require the most urgent reassessment as recent reports from Israel indicate that their habitat is undergoing drastic degradation and development.
- Arbel, Avraham. (1984) Plants and Animals of the Land of Israel: an illustrated Encyclopedia, Vol. 5. Tel Aviv. (Trans. J. R. Buskirk).
- Bodenheimer, F. S. (1935) Animal Life in Palestine. Jerusalem.
- Buskirk, James. (1966) Turtles of the Holy Land. Int. Turtle Soc. J. 1:20-23.
- Buskirk, James (1986) The endangered Egyptian tortoise Testudo kleinmanni: Status in Egypt and Israel.Proc. 9th International Herpet. Symposium on captive husbandry and propagation. (eds. McKeown, Caporaso & Peterson). San Diego, CA.
- Dumeril, A. M. C. and Bibron, G. (1835) Erpetologie General ou Histoire Naturelle complete des Reptiles. Tome.2. Paris.
- Eiselt, J. and Spitzenberger, F. (1967). Ergbinisse zoologischer Sammelreisen in der Turkei: Testudines. Ann. Nat. Hist. Mus. Wein. 70: 357-378.
- Esterbauer, H. (1985) Okologische und verhaltensbiologische Beobachtungen an der Maurischen Landschildkrote, Testudo graeca terrestris Forskal, 1775, in Sudwestsyrien. Aquar. Terr. Mon. Ornith. Vivarium. 32(1):389-392.
- Flower, S.S. INCLUDE
- Forskal, P. (1775) Descriptiones Animalium; Avium, Amphibiorum, Piscium, Insectorum, Vermium. Heineck and Faber.
- Freyhof, J. (1983) Beobachtungen zur Trachtigkeitsdauer bei der Maurischen Landschildkroten, Testudo graeca terrestris. Das Aquarium (173):605-608.
- Herrn, C. P. (1966) Testudo graeca terrestris Forskal neu fur die Turkei. Stuttgarter Beitrage zur Naturkunde (155):1-2.
- Pritchard, P. C. H. (1979) Encyclopedia of Turtles. T.F.H. Publications, New York.
- Schleich, H. H. (1989) Merkmalsausbildungen an Landschildkroten in Nord-ost-Libyen (Testudines: Testudinidae). Herpetozoa 1(3/4):97-108.
- Stubbs, D. (1989) in The Conservation Biology of Tortoises.IUCN.
- Swingland, I. R. and Klemens, M. (eds.), (1989) The Conservation Biology of Tortoises. IUCN.
- Tristram, H. B. (1885). The Flora and Fauna of Palestine. Palestine Exploration Fund, London.
- Wermuth, H. (1956) Versuch der Deutung einiger bisher ubersehener Schilkkroten-Namen. Zool. Beitr., Berl. (2):399-423.
- Wermuth, H. (1958) Status und nomenklatur der maurischen landschildkrote, Testudo graeca, in S. W. Asien und N. O. Africa. Senck. Biol. 39:149-153.
- Wermuth, H and Mertens, R. (1961). Schildkroten, Krokodile, Bruckenechsen. Gustav Fischer Verlag: Jena
- Zavattari, E. (1930). Erpetologia della Cirenaica. Arch. Zool. Ital. 14(2):263