Information about activities. How to join the TT. Discussion list, for members and others! Species-specific Care sheets Articles from our Newsletter. Links to important sites. Links to members' sites. Search our site! What has been updated, and when.

A revision of the Testudines of North Africa, Asia and Europe - Genus: Testudo

A C Highfield & J. Martin

Abstract
This paper is the first in a series which presents a complete review, analysis and revision of the taxonomic and zoogeographic status of all identifiable extant members of the genus Testudo which is generally accepted by the present authors as defined by Williams in Loveridge and Williams (1957).
This first part includes a report of an important morphological diagnostic character which indicates that the present status of Testudo (graeca) zarudnyi NIKOLSKI 1896 (Terra typica = Birjand province, Eastern Iran), Testudo (graeca) ibera PALLAS 1814 (Terra typica = Kura valley, Caucasus) as sub-species of Testudo graeca LINNAEUS 1758 (Terra typica = Oran, Algeria) is not an accurate reflection of their true taxonomic status.

KEYWORDS; Taxonomy - Testudo - Africa - Asia - Europe - Testudines -Iran - Testudo graeca - Testudo zarudnyi - Holotypes - Nikolski -Distribution.

Introduction
It was in an attempt to clarify the rather confused current diagnostic situation relating to Testudo that the present work commenced. Never has the need for reliable data on the taxonomy of Testudo been greater. Throughout their range these animals are threatened, principally by habitat destruction and human incursion. It is generally agreed that conservation captive breeding projects can play a major role in preventing their total extinction, yet in order for such projects to function reliable information on identification and inter-species relationships are essential. The authors objectives therefore are to provide an accurate and relatively accessible guide for use by field herpetologists and those involved in captive breeding projects in which all extant members of the genus are catalogued and described, to identify cases where errors of classification have occurred, and to advance a tentative theory of zoogeographic distribution based upon some new observations and previously unremarked characters.
A complicating factor in many published descriptions of terrestrial tortoises is the often imprecise and subjective nature of many character observations e.g, carapaces are frequently described as ''low vaulted'', ''high domed'' or ''broad''. Colouration is notoriously difficult to express with any degree of accuracy, e.g ''greenish-olive'' or ''yellow ground with central brown-black dot''. Such terminology also creates considerable difficulty for translators. Even attempts at using a standardised system of colour nomenclature fail to satisfy the requirement for a usable field reference system (Ridgway, 1912 , Eiselt & Spitzenberger, 1967). For this reason photographs and line drawings offer many advantages and although we shall provide a written description of characters, it is the authors opinion that the illustrations in combination with dimensional statistics provide the most reliable key to practical identification of living specimens in the field.
During the course of this work several specimens were encountered which could not be referred to any of the currently recognised taxa. These will be described in detail in a subsequent part and our tentative suggestions as to their derivation and possible attribution discussed.
Synonymies on Testudo are notoriously difficult to compile with any degree of accuracy. The status of species referred have undergone a great many changes, each change introducing an additional level of complexity and making bibliographic research on the taxa extremely difficult. Most early and not a few later checklists contain a very high proportion of entirely spurious entries, and a considerable number of described species are now considered invalid -either because they are homonyms, non-binomial or for some other reason. It is also necessary for those undertaking taxonomic or bibliographic research on this genus to recognise that authors have at various times employed entirely different conventions of nomenclature from that accepted today. Thus, for many authors ''T. graeca L.'' was the name habitually used to describe T. hermanni GMELIN and vice versa e.g see Boulenger (1889) pp.177 and Lortet, (1887) pp.5, tab.4. Fortunately a table of the various authors methodology is fairly easy to compile and this certainly eases the task of the bibliographer. If there is sufficient demand, the table compiled by the present authors will be published at a later date. The synonyms provided here should make our interpretations of the identifications clear. To facilitate this, whenever possible we have listed them in the following manner, e.g;
Testudo marginata SCHOEPFF 1792 As Testudo campanulata; Strauch 1862a,b, 1890 etc. As Testudo tabulata var. campanulata; Walbaum 1782 As Testudo nemoralis; Schreiber 1875 etc. As Testudo marginata; Lortet 1887,Pritchard 1979 etc.
In cases where the list is particularly extensive and confused as the is the case with Testudo graeca L. 1758, we have elected to publish a condensed and edited listing. A full listing, even if it were possible to compile would run to many pages and would achieve little of practical value. All important and significant references are recorded . Type localities are also often very confused in early works. Where this is relevant we have noted the fact, e.g;
Gervais, P. (1836) Enumeration de quelques especes de Reptiles provenant de Barbarie. Ann. Sci. Nat. (2) vi. pp.308-313; Testudo marginata SCHOEPFF 1792 (omit N. African locality references).
or:
Strauch, A. (1862b) Essai D'une Erpetologie de L'Algerie. Mem. L'Acad. Imp. Sci. St. Petersbourg. Tome IV, No.7. p.14 . Testudo campanulata WALBAUM 1782 (omit Egyptian locality reference).

Materials and methods
Wherever possible original rather than secondary sources have been consulted. For this the authors are grateful to a number of museums and libraries which will receive full acknowledgement in a subsequent part of this paper and also to our translators who have coped splendidly with some often very difficult transcriptions. We are also grateful to numerous correspondents who supplied us with field and other data, and to owners of captive specimens who permitted us (sometimes at great length) to photograph, measure and catalogue their collections.
We have also endeavoured to examine types and paratypes where these were available. In addition, the authors are fortunate in having at their disposal an extensive collection of living specimens which includes representatives from all of the currently recognised species (and sub-species) of the genus. We have also had access to post-mortem material derived from a wide variety of sources following the natural demise of specimens in various other collections. This has proved to be of very considerable value. In osteology we have mainly followed Gaffney (1972 & 1975), Romer (1956) and Siebenrock (1897). For definitions of general herpetological terminology employed see Peters (1964) and the International Code of Zoological Nomenclature. In the course of this study, several hundred living tortoises were weighed, measured and carefully photographed both in colour and black and white. For maximum definition and quality, medium format 120 roll-film cameras were usually employed. 35mm cameras were used for most of the extreme close-up photographs. A 55mm macro lens was invariably used in such instances. Black and white photographs were taken using ISO 50 and ISO 200 materials developed to a high degree of acutance and finished to archival standards of permanence. A second series of 120 format colour transparencies for reprographic use was taken at the same time using Kodak Ektachrome professional or Fuji RD100 film. 35mm colour transparencies were also taken for use in lectures and for inclusion in a photographic library. A series of colour transparencies and prints to accompany this series will be made available at a later date.
Several views were usually taken of each tortoise;- a view from above to show overall body outline and carapace markings, a plastral view, a side view to show carapace height and profile and a head-on view which clearly demonstrates carapace curvature. Where necessary, these views are supplemented by extreme close-ups of notable characters. Comparative side-by-side views were also taken to illustrate sexual dimorphism or the differences between various species and sub-species.
Weight and measurement data was entered onto a computer system and the results made available in a series of charts and graphs which were then integrated with a relational database to produce the final revised taxonomic data tables. For maximum weighing accuracy electronic digital scales were used, and all measurements were executed using a specially designed 'Tortometer' capable of obtaining carapace length, height and width. More detailed measurements of scutes, eggs and external cranial characteristics were obtained using calipers and micrometers.
The numbers of specimens available to us varied very considerably depending upon the rarity of the species involved. In some cases we were able to examine 400 - 500 living animals of identical type together with numerous museum paratypes, in another instance it proved possible to locate only one living representative. Some of the races are exceedingly rare, and to the best of our knowledge are currently not the subject of any active conservation or captive breeding efforts. Indeed, we know almost nothing of the present population status of some races. It is the hope of both authors that this work will encourage further interest in their ecology and conservation and that it may not yet be too late to institute measures which will lead to their preservation in the wild.

Current status and definitions
The most widely accepted and used classification of the land tortoises of the genus Testudo to date is as follows. This is broadly based upon the work of Robert Mertens (1946) and Mertens and Wermuth (1961). The conclusions reached by these authors have found general support from Obst and Meusel (1963 and 1978), and Pritchard (1979). This is the system adopted by most current works of reference and hence it is important that it is included here. Excluding any disputes at species level and below, various schemes have been suggested which attempt to classify chelonians (including fossil examples) at Order and Family level. The most notable of these are the suggestions of Chkhkvadze (1970), Mlynarski (1969), Romer (1956), Gaffney (1975) and Williams (1950).

Suborder:- Cryptodira
Family:- Testudinidae
Genus:- Testudo
Species:- Testudo graeca
Sub-species:-
Testudo graeca graeca LINNAEUS 1758
Testudo graeca ibera PALLAS 1814
Testudo graeca terrestris FORSKAL 1775
Testudo graeca zarudnyi NIKOLSKI 1896
Species:- Testudo marginata SCHOEPFF 1792
Species:- Testudo hermanni
Sub-species:-
Testudo hermanni hermanni GMELIN 1789
Testudo hermanni robertmertensi WERMUTH 1952

The present authors accept the revision of Khozatsky and Mlynarski (1966) which removed the Afghan or Horsfields tortoise from the genus Testudo and placed it in the newly designated genus Agrionemys as Agrionemys horsfieldii (GRAY 1844) on the basis of its unique carapace and plastron morphology and other distinctive features (see also Mlynarski, 1966). However, although Agrionemys horsfieldii is therefore in a strict sense outside the scope of the present work which is restricted to members of the genus Testudo, it features a number of extremely interesting characters which closely tie in with the asiatic grouping of this genus, and thus will be referred to from time to time for the purposes of comparison. A similar situation applies to Psuedotestudo kleinmanni (LORTET ). This latter species is of considerable zoogeographic and evolutionary interest and its status will be discussed in some detail.
The definition adopted by the present authors for the genus Testudo follows that of Williams in Loveridge and Williams (1957). We do have a number of reservations about some specific characters but these are comparatively minor in nature and for the most part we regard the definition of Loveridge and Williams as reasonably satisfactory;

''Skull with triturating surface of maxilla moderately or weakly ridged or without ridging; median premaxillary ridge absent; maxillary not entering roof of palate ; anterior palantine foramina small, concealed or large and conspicuous; prootic typically concealed dorsally and anteriorly by parietal; quadrate enclosing stapes or not; neck with third or fourth centrum biconvex; carapace never hinged; typically the anterior neurals alternately octagonal and quadrilateral; outer side of third costal scute about as long as, or longer than, that of the fourth; submarginal scute absent; frequently a single suprapygal, if two, they are typically separated by a straight transverse suture; plastron with posterior lobe more or less hinged in one or both sexes; gular region but little thickened or produced; gulars paired, longer than broad.''

Any attempt or describe and classify the Testudinidae is made considerably more difficult by reason of the very wide range and variation of diagnostic characters to be found within even closely related animals. This has had the effect of a) masking the precise delineations which exist between genuine species and sub-species groups and b) causing numerous cases of mistaken identity. Regional clines are common, and by no means all of these have yet been catalogued or adequately described. Truly anomalous animals are also relatively frequently encountered; for example, we have in our own collection a very large (264mm, 3,420g) female Testudo hermanni which conforms to type perfectly in every respect save for an additional vertebral scute and a perfectly formed set of 4 claws both anteriorly and posteriorly instead of 5 anteriorly and 4 posteriorly (Highfield, 1988). Excluding specimens where obvious congenital deformities, accidental damage or carapace deformities consequent upon identifiable nutritional deficiency effects exist (see Highfield, 1989), a surprising number of anomalous animals are encountered. The external characters most often affected include extra or missing claws, extra or missing scutes, scute markings or dimensions which are inconsistent with the species to which the specimen is referred, unusually exaggerated dimensions either greater or lesser than normal for the species and certainly in the case of T.graeca, missing or twinned thigh tubercles. It is necessary to differentiate between those characters found in an individual specimen which are truly anomalous or eccentric and those which may be considered either diagnostic or normal for a species or sub-species, even though in some cases they may only be found in a minority of the population - e.g divided supracaudal scutes in Testudo graeca L.
The present authors have taken the view that where a character is present in less than 0.5% of a representative population sample we regard the character as anomalous or eccentric. Where it occurs at a level of between 0.5% and 5% we regard it as probably anomalous. Where unusual and anomalous features have been observed, these are recorded under the relevant species heading together with estimates of their probable incidence.
At one time it was normal for species to be defined entirely on the basis of morphological characters, often based on preserved specimens alone. This method is still of prime importance, but it has gradually been recognised that other factors should also be taken into consideration and that careful observation of the living animal is equally as important as the study of museum specimens. These additional characters may include behavioural differences between two groups, in the case of reptiles differences in preferred optimum temperatures, habitat niches occupied, and even different pathological patterns. The temperatures at which eggs incubate or produce males or females according to Environmental Sex Determination criteria may also provide useful evidence of speciation. Although open to subjective interpretation, it cannot have escaped those charged with caring for captive tortoises on a day-to-day basis that there are also clear behavioural differences between the various species, with some noticeably more aggressive than others. This too can provide a number of useful clues. It is notable for example that male-female leg biting is much more frequent in T. hermanni than in T. graeca, but that male T. hermanni do not indulge to anything like the same extent in the ''shell butting'' or ramming courtship behaviour pattern so typical of T. graeca and T. ibera.These behavioural characteristics, whilst not necessarily diagnostic in isolation can provide useful indicators when considered in combination with morphological factors.
The question of actual or alleged hybridisation between various species and sub-species of this genus poses special problems. Anecdotal evidence of such occurrences are not unusual, but such allegations are rarely adequately supported.
Most cases of alleged hybridisation reported to us were based upon simple mis-identification of the parents, others upon pre-fertilisation of a female by a male of identical type but subsequently mated by male of a different species which was then assumed to be a parent. The sometimes very extended gestation periods of tortoises between conception and laying frequently leads to this type of mistake. Some hybridisation may indeed occur naturally between some geographically adjacent races, and it may be possible to induce it between normally widely separated races when both are held in captivity. We conclude however that any such instances must be regarded as exceptional and certainly that the physical construction of some species, the size of their eggs and unique biotypic adaptions would suggest that any such reports must be treated with skepticism until carefully verified.
Some rare published references to alleged inter-species breeding include Obst and Meusel (1978) [ T.marginata + T.graeca ibera], Basoglu (1977) refers to specimens observed in Turkey possessing intermediate characters between T.g.ibera and T.g.terrestris and the same observation is reported by Eiselt and Spitzenberger (1967). The authors would welcome further reports and references.
It is obviously extremely important that those engaged in taxonomic studies recognise clinical symptoms of osteological deformity whenever present in a specimen as failure to take this factor into account can lead to gross errors of attribution. Specimens exhibiting carapacial deformity should therefore be excluded from morphological comparative data. This is clearly of particular importance when dealing with long-term captives or with captive-bred specimens where such deformities are unfortunately all too common (e.g see Jackson and Trotter, 1976). Similarly, growth and size data derived from entirely captive bred specimens may not always reflect the natural circumstances of the species and this must be considered whenever data from such specimens is included in studies.
From the checklist of species given earlier, it will be noted that both Testudo zarudnyi and Testudo ibera are both currently regarded as subspecies of the African race Testudo graeca. This position was defined by Mertens in 1946 on the basis of some shared characteristics with the African graeca, e.g thigh tubercles and alleged ''minor differences in shell proportions'' between Testudo graeca L. 1758 from North Africa and Testudo ibera PALLAS 1814 from trans-caucasia. That both Testudo zarudnyi and Testudo ibera have features in common with each other and also with Testudo graeca L. is not disputed. We would argue however that they have at least as much if not more in common with Testudo marginata and Testudo hermanni than with any N.African race. In the matter of carapace proportions the question is less open to subjective interpretation (measurements based on a randomly selected series of 50 specimens of each race);

Average carapace length female T. graeca L. = 175.80mm
Average carapace length male T. graeca L. = 152.40mm
Average carapace length female T. ibera = 201.08mm
Average carapace length male T. ibera = 180.23mm

Similar divergence is found in the ratios of carapace height to length and width and also in overall body mass. A comprehensive set of morphometric data will be presented later in this series. It is our opinion that these differences should not be regarded as minor, however. Additional points of divergence include skull structure, colouration, skin colour and texture, scutellation and egg morphology. Given such a range of divergent characters, it is not easy to reconcile these to the claim that one race is merely a sub-species of another. Divergence of such a range and of such scale surely requires a re-assessment of Mertens's conclusion. One of the first key morphological characters to be identified by us which led us to believe that members of the Asiatic and European races are not subspecies of the North African Testudo graeca L. was the remarkable consistency of form exhibited by the 1st frontal vertebral scute of all the Asiatic-European specimens. This shape of scute does not occur in the N.African Testudo above the level of anomaly. The N.African Testudo consistently exhibits an entirely different (rounded) frontal vertebral scute construction. Again, we have not found the N.African form of scute in Euro-Asiatic specimens above the level of abnormality. The (straight-edged, angular) form of scute found in Testudo ibera PALLAS is common also to Testudo zarudnyi, Testudo hermanni and Agrionemys horsfieldii (see plates 1 to 6). This form of scute does not occur in animals conforming to the type of Testudo graeca L. 1758. We have observed the rounded form frontal vertebral scute only twice in Testudo ibera PALLAS out of over 450 living animals of this race examined and catalogued to date. In each case there was a definite thickening of keratin and our diagnosis was that there was clear evidence of abnormal carapace development consequent upon nutritional factors. Both were long-term captives and had been in our opinion subject during an early growth phase to a diet excessively high in protein resulting in an abnormal thickness of keratin (Highfield, 1989). We have never observed this shape of scute in this species in the absence of additional osteological pathogeny.
We suggest that the distinct forms of frontal vertebral scute are diagnostic characters of some very considerable importance and indicative of membership of a) the Euro-Asiatic racial group or b) the African racial group. There is also a third group, the Middle-eastern racial group. This presents certain very difficult and interesting problems of its own and will be examined in some detail in a later part of this survey (with particular reference to Testudo terrestris FORSKAL 1775).
It is our view that the African race Testudo graeca L. should be zoogeographically restricted to N. Africa and that no convincing evidence exists to support the theory that the races T. ibera PALLAS (Type locality; Caucasia) and T. terrestris FORSKAL (Type locality;Lebanon) are sub-races of the African graeca. This matter will be discussed in some detail when in the next part of this revision we describe and illustrate the various different races of tortoise which currently form the miscellaneous grouping alleged to comprise the single race Testudo (graeca) terrestris FORSKAL 1775.

TESTUDO ZARUDNYI (NIKOLSKI 1896)

Important publications
As Testudo ibera: Blandford 1876; Werner 1895. As Testudo zarudnyi: Nikolski 1896, 1897, & 1899; Zarudny 1903; Werner 1929, 1936 & 1938; Forcart 1950; Anderson 1963. As Testudo graeca zarudnyi: Mertens 1946 & 1956; Wermuth and Mertens 1961; Pritchard 1966 & 1979; Obst 1986; Obst and Meusal 1974; Anderson 1974 and 1979.

DISTRIBUTION
Endemic to the Central Plateau of Iran. Confirmed localities include Zahedan in Baluchistan, Khash, Askhalabad, Arusan, Kerman (central to range), steppe between Nabid and Bam, Bendun and Gurmukh. One record from Ashkhabad, Turkmen, USSR (Gabriel collection, Werner, 1938).

HABITAT
Reported altitude range from 1,065 m to 2,500 m. Hillsides, uncultivated rocky slopes and plains. This tortoise inhabits xeric and inaccessible areas far from human habitation and contact. Ashkhabad specimen recorded at 1800m (Werner, 1938).

NOMENCLATURE
This tortoise was named by Aleksandr Mikhailovich Nikolski after Nikolai Aleksyevich Zarudny who made extensive collections of reptiles, amphibians and fishes in Iran circa 1895-1904. The first full description of the species was that of Nikolski in the Annals of the Museum of Zoology of the Academy of Imperial Science of St.Petersburg in 1896. This paper also bears an excellent illustration.

DIAGNOSIS
Maxillary with weak ridge; quadrate enclosing stapes; forelimb with 5 claws; one claw noticeably smaller than the others; large subconical tubercles on posterior aspect of thighs; supracaudal shield normally undivided; 5th vertebral shield not broader than 3rd; carapace elongate in outline, with upturned strongly emarginate posterior border; carapace colouration very dark brown-olive with little pattern visible; edges of scutes may be very slightly horn coloured ; edges of posterior and frontal marginals may feature transluscent-yellow horn tips; scales of fore-limbs large, black and triturated, pointed scales on heels of rear limbs; noticeable diagonal reticulations on costal and vertebral scutes; eyes usually elongate-almond in shape rather than round (as seen in T. ibera PALLAS 1814); anterior ridge of marginals extended; intergular very broad; nuchal long and narrow, sword shaped; plastron mostly dark with few horn coloured markings; skin light olive-brown; anterior limbs flattened and adapted for digging.

DIMENSIONS
The living (female) specimen in our collection has a straight-line carapace length of 282mm, a maximum width of 201mm and a carapace height measured at the 3rd vertebral of 130mm. Weight was recorded at 3,700g. This is considerably larger than most T. ibera of Turkish, Iranian or Syrian origin. Werner (1938) provides the following dimensional statistics for two specimens from Ashkalabad (1800m) and Kharistan (2,500m). The smaller example of which Werner says ''The colouring is a bit like T. marginata'' is a juvenile, and the larger a female;

  • Carapace, length (with string over curve) 275mm & 175mm
  • Carapace, length (direct) 230mm & 135mm
  • Carapace, width (string) 260mm & 180mm
  • Carapace, width (direct) 135mm & 98mm
  • Carapace, height 95mm & 70mm
  • Plastron, length 180mm & 117mm
  • Nuchal 14 X 8mm & 8 X 4mm
  • Ratio of length of individual plastron plates to median seam; 25 + 30 + 8 + 70 + + 14 + 38 (I) 18 + 17 + 8 + 48 + 6 + 31 (II)

Of these two specimens Werner additionally notes the large scales on the heels and the reticulations of the carapace shields.

CRANIAL OSTEOLOGY
Due to the extreme rarity of this species it has not proved possible to obtain any sufficiently reliable material for osteological study. Nor has it proved possible to trace any previously published illustrations detailing cranial or carapace morphology. The authors are however continuing to survey museum collections and it is hoped that eventually sufficient material will be located to permit an appendix to this paper to be produced describing these characters at a later date.

REPRODUCTION
Few details on the reproductive behaviour of this species are available. Anderson (1979) observed a female excavate a nest in a west facing rocky slope between two Artemisia bushes. The nest was 11cm deep by 15cm long and 12cm wide. The female deposited 4 eggs before filling the nest. Unfortunately no figures are given for the dimensions and weight of the eggs. A subsequent attempt to incubate these eggs artificially was unsuccessful.

ILLUSTRATIONS
Few illustrations of this species have been published. The most notable are those of Nikolski (1896) Pl.17, and Anderson (1979) Figure 10. Werner (1938) also includes two small plates of a specimen originating from Ashkhabad. The illustration on p.563 of Pritchard (1979) appears to be a large Testudo ibera PALLAS 1814 rather than a Testudo (graeca) zarudnyi as labeled. The marginals of the specimen depicted do not conform to type, and the carapace reticulations so typical of zarudnyi are absent. We are also extremely doubtful of an illustration of an alleged Testudo zarudnyi accompanying a report of the USNM Nat.Hist reptile collection in Iran (Tuck, 1971). This specimen also does not conform to type and in addition was photographed in Najafabad, well outside the range of the species as defined by Anderson (1979). The tortoise illustrated has characters typical of Testudo ibera PALLAS 1814 to which we believe it should be referred.
The range of Testudo zarudnyi NIKOLSKI 1896 can be defined as exclusively to the east of a line drawn from Babol on the Caspian sea to Niza inland of the Gulf (Fig.1). Six tortoises located by Pritchard (1966) and referred by him to zarudnyi were found aestivating under rocky ledges beside a stream in an otherwise exceptionally arid habitat 15 miles east of Moorcheh Khort which is 33 miles north of Isfahan - a considerable distance to the west of the limit of this species accepted range, but well within the range of Testudo ibera PALLAS.
Some idea of the low population densities of terrestrial chelonia in Iran can be gained from Pritchards observation that in 10 weeks of surveying, covering up to 8 miles per day though relatively open ground, only 17 living tortoises were encountered. This should be contrasted with reports from workers in Greece, Turkey and Africa where it is not uncommon for the count to reach double figures in a matter of a few hours searching - even in difficult terrain offering good concealment potential. Descriptions of some methods useful for surveying wild tortoise populations can be found in Stubbs, Hailey , Pulford and Tyler (1981 and 1984) and in Lambert (1982).
The problem of confusion between Testudo ibera PALLAS and Testudo zarudnyi NIKOLSKI is one which has dogged many researchers, collectors and taxonomists. The fact is that Testudo zarudnyi are very rare, and are consequently infrequently available for direct comparison. The provenance of many museum reference specimens is also consequently in doubt. The Testudo ibera which inhabit the eastern central Zagros mountains of Iran are also fairly melanistic, a trait which extends to those inhabiting Iraq and the extreme Eastern region of Turkey. Many specimens of Testudo ibera also feature flared posterior marginals, although in our experience never to the extent found in Testudo zarudnyi or Testudo marginata. The projection of the anterior marginals found in Testudo zarudnyi in combination with the gross projection and serration of the posterior marginals is quite remarkable - the only other extant tortoises we have observed which bear a marked similarity are Testudo marginata. Even here, the anterior marginals are reduced by comparison in marginata. Although Testudo ibera may possess in some cases quite a flared posterior marginal, the anterior marginal is never in our experience as pronounced as that found in Testudo zarudnyi and the posterior marginal is never as serrated. The presence of a flared or extended posterior margin in T.ibera has on a number of occasions given rise to the claim that some individuals demonstrate intermediate characters with T. zarudnyi. We believe this view to be mistaken and based upon a misinterpretation of the range of normal variation to be found within ibera. We have examined and recorded many hundreds of specimens of T.ibera collected from various locations and can state with confidence that no specimen seen by us has possessed significant diagnostic characters which in our view could be justifiably interpreted in this way.
Anderson (1979) makes reference (p.518) to a paratype of Testudo zarudnyi held in the British Museum collection. He suggests that this specimen is the same as that examined by Bird (1936) and Pritchard (1966), both of whom voiced doubts as to the validity of Testudo zarudnyi, even at sub-specific level. These doubts are not new and indeed Boulenger (1920) comments '' I am not certain whether T. zarudnyi deserves specific-recognition; at any rate the characters pointed out by Siebenrock (1909) are worthless. A specimen from Zirkuck, E. Persia, received from the Petrograd museum in 1899 as T. zarudnyi has the first vertebral shield a little broader in front than behind, the third vertebral not broader than he third costal, and the posterior margin of the carapace not more strongly serrated than in some individuals of T. ibera''. Unfortunately neither Bird nor Pritchard provide any identification numbers for the specimen they saw, but the senior author has carefully examined what is believed to be the same one (numbered BM 1947.3.5.17) and which once formed part of the St.Petersburg Museum collection. This appears to be the only alleged zarudnyi in the British Museum, so it is likely that Anderson (1979) is correct. This carapace is definitely not a Testudo zarudnyi NIKOLSKI but is that of a medium-sized female Testudo ibera PALLAS. It is somewhat discoloured with age, but clearly has a distinctly amber coloured plastron and carapace and scute markings typical of Testudo ibera, only moderately flared marginals and in every respect conforms to the type description not of zarudnyi but of ibera. The senior author has seen and photographed several tortoises from Turkey absolutely identical to BM 1947.3.5.17 in size, colour and carapace morphology. Unfortunately only the carapace of the St. Petersburg tortoise is preserved. This is perfectly adequate to identify it however, and even to estimate the age of the animal which we would place at circa 40 years. On the basis of this specimen it is not surprising that doubts were expressed regarding the status of Testudo zarudnyi. What is curious is that BM 1847.3.5.17 does not even appear to conform to the T. ibera PALLAS of the Zagros mountains. It is much lighter in colouration than a typical Zagros T.ibera, and is more typical of a north-eastern Turkish rather than an Iranian specimen. We would regard the general provenance of this specimen as doubtful.
It is understandable that a Testudo ibera with flared posterior marginals could be mistaken for a Testudo zarudnyi if the observer was not acquainted with the limits of the natural range of variation to be found in both species. Age can also affect a tortoises colouration and markings further complicating identification; in some cases this results in an increase in visible yellow-orange and in others an apparent loss of yellow-orange (neither change occurs in Testudo zarudnyi, but these changes do occur in Testudo ibera where some very elderly specimens can lose almost all distinguishing marks). Clinal and other variations in the morphology and range of carapace markings of Turkish, Syrian and Iranian Testudo ibera will be discussed in a later part of this series (e.g see Siebenrock ,1913, Nikolski, 1915, and Wettstein, 1951).
Finally, brief mention should be made of two alleged species to which references may be found in older accounts of Testudo zarudnyi. These concern Testudo buxtoni (Boulenger, 1920), type locality Manjil, and Testudo baluchiorum (Annadale, 1906) type locality Baluchistan. The former is undoubtedly a synonym of Testudo ibera PALLAS 1814, and the latter a straight synonym of Agrionemys horsfieldii (GRAY 1844). Neither should be confused with Testudo zarudnyi.
The elongate carapace of Testudo zarudnyi and extensively flared and serrated marginals also have far more in common with Testudo marginata than with Testudo graeca L. This same comparison is drawn by Mertens (1946) who goes on to say that T. zarudnyi (which he knew from a specimen in the Vienna Museum collection and which at the time of writing the present authors have not examined) ''appears to be connected through intermediate specimens with ibera, so that it is most probably correct to regard zarudnyi as only the most eastern race of graeca''. The opinion of the present authors however is that the two races are entirely separate and fully deserving of independent species status.

CONCLUSIONS
Analysis of the morphological characters of Testudo zarudnyi NIKOLSKI 1896 indicate that it is sufficiently diverse from the African race Testudo graeca L. 1758 to be regarded as an autonomous species. It clearly belongs to the trans-caucasian grouping of the genus which in our view comprises;

  • Testudo zarudnyi NIKOLSKI 1896
  • Testudo ibera PALLAS 1814
  • Testudo marginata SCHOEPFF 1792
The relationships between these races pose a number of very interesting and perplexing problems. These will be explored as this series progresses. The status of Testudo hermanni GMELIN 1789 will be also be discussed in relation to the above races.

To be continued:-

REFERENCES

  • Anderson, S. C. (1963) Amphibians and Reptiles from Iran. Proc. Calif. Acad. Sci. 4th series, 31:417-498.
  • Anderson, S. C. (1979) Synopsis of the Turtles, Crocodiles, and Amphisbaenians of Iran. Proceedings of the California Academy of Sciences, 4th Series. Vol XL1, No.22, pp. 501-528.
  • Annadale, N. (1906) Testudo baluchiorum, a new species. Journal of the Asiatic Society of Bengal, Vol.11, No.3, pp. 75-76.
  • Basoglu, M. (1977) Turkiye Surungenleri. Ege Universitesi Fen Fakultesi Kitaplar Serisi No. 76. pp.191-223.
  • Bird, C. G. (1936) The Distribution of Reptiles and Amphibians in Asiatic Turkey. Ann. Mag. Nat. Hist. Series 10, 18: 257-281.
  • Blandford, W. T. (1876) Eastern Persia, an account of the journeys of the Persian Boundary Commission, 1870-1872. Vol.2, The Zoology and Geology. London, 516pp.
  • Boulenger, G. A. (1889) Cat. of the Chelonians, Rhynchocephalians, and Crocodiles in the British Museum. London.
  • Boulenger, G. A. (1920) Description of a new land tortoise from northern Persia. Journ. Bombay Nat. Hist. Soc. 27:251-252.
  • Chkhkvadze, V. M. (1970) On the origin of the modern Palaearctic land tortoises. Bull. Acad. Sci. Georgian S. S. r. 57(1):245-247
  • Eiselt, J. and Spitzenberger, F. (1967) Ergebnisse zoologischer Sammelreisen in der Turkei: Testudines. Ann. Naturhistor. Mus. Wien. 70:357-378.
  • Forcart, L. (1950) Amphibien und Reptilien von Iran. Verh. Naturforsch. Ges. Basel. 61:141-156.
  • Forskal, P. (1775) Descriptiones Animalium; Avium, Amphibiorum, Piscium, Insectorum, Vermium. Heineck and Faber.
  • Gaffney, E. S. (1972) An Illustrated Glossary of Turtle Skull Nomenclature. Amer. Mus. Noviates, No. 2486:1-33.
  • Gaffney, E. S. (1975) A phylogeny and classification of the higher categories of turtles. Bull. Amer. Mus. Nat. Hist., 155(5):387-436.
  • Gervais, P. (1836) Enumeration de quelques especes de Reptiles provenant de Barbarie. Ann. Sci. Nat. (2) vi. pp.308-313.
  • Gmelin, J. F. (1789) Systema Naturae. Ed. 13. (3). Amphibia et Pisces. Leipzig.
  • Gray, J. E. (1844) Catalogue of the tortoises, crocodiles and amphisbaenians in the collection of the British Museum. London, 289pp.
  • Highfield, A. C. (1988) New size record for T. hermanni GMELIN 1789. The Rephiberary. Newsletter of ASRA (U.K) No.132: 5-6.
  • Highfield, A. C. (1989) Notes on dietary constituents for Herbivorous terrestrial chelonia and their effects on Growth and Development. ASRA (U.K) Journal, Vol.3 (3):7-20.
  • Jackson, C. G. and Trotter,J. (1976) Accelerated growth rate and early maturity in Gopherus agassizi (Reptilia: Testudines). Herpetologica 32:139-145.
  • Khozatsky, L.I and Mlynarski, M. (1966) Agrionemys - nouveau genre de tortues terrestres (Testudinidae). Bull. Ac. Polon. Sci. (2):123-125.
  • Lambert, M. R. K. (1982) Studies on the growth, structure and abundance of the Mediterranean spur-thighed tortoise, Testudo graeca in field populations. J. Zool. Soc. London, 196:165-189.
  • Linnaei, C. (1758) Systema Naturae. Ed. decima.
  • Lortet, L. (1887) Observations sur les Tortues Terrestres et Paludines du Bassin de la Mediterranee. Arc. Mus. Nat. Hist. Lyon, Tom. IV, pp.1-26.
  • Loveridge, A. and Williams, E.E. (1957) Revision of the African tortoises and turtles of the suborder Cryptodira. Bull. Mus. Comp. Zool. Harv. Coll. 115:163-557.
  • Mertens, R. (1946) Uber einige mediterrane Schildkrotenrassen. Senckenbergiana Biol. 27:111-118.
  • Mertens, R. (1956) Amphibien aus Reptilien aus s. o. Iran, 1954. Jahresh. Ver. Vaterl. Naturkd. Wurttemb, 111:90-97.
  • Mlynarski, M. (1966) Morphology of the shell of Agrionemys horsfieldii (Gray 1844) (Testudines, Reptilia). Acta. Biol. Cracovia IX:219-223.
  • Mlynarski, M. (1969) Fossile Schildkroten. A. Zeimsen Verlag. pp. 1-128.
  • Nikolski, A. M. (1896) Diagnosis reptilium et amphibiorum novorum in Persia orientali a N. Zarudny collectorum. Ann. Mus. Zool. Acad. Sci. St. Petersbourg, 4:369-372.
  • Nikolski, A. M. (1897) Les Reptiles, amphibiens, et poissons recueillis par M. N. Zarudny dans la Perse orientale. Ann. Mus. Zool. Acad. Imp. Sci. St. Petersbourg, 2:306-348.
  • Nikolski, A. M. (1899) Reptiles, amphibiens, et poissons, recueillis pendant le voyage de M. N. A. Zarudny en 1898 dans la Perse. Ann. Mus. Zool. Acad. Imp. Sci. St. Petersbourg, 4:375-417.
  • Nikolksi, A. M. (1915) Fauna of Russia and adjacent Countries; Reptiles, Vol.1. Chelonia and Sauria. Translation; Israel Prog. Sci. Trans. Jerusalem 1963, 352pp.
  • Obst, F. J. and Meusel, W. (1963 and 1978) Die Landschildkroten Europas. A. Ziemsen Verlag, Wittenberg Lutherstadt.
  • Obst, F. J. (1986) Turtles, Tortoises and Terrapins. Macmillan of Australia. 231pp.
  • Pritchard, P. C. H. (1966) Notes on Persian Turtles. Brit. Journ. Herpetology. 3:271-275.
  • Pallas, P. S. (1814) Zoographia rosso-asiatica, sistens omnium Animalium in extenso Imperio rossico et adjacentibus Maribus observatorum Recensionem, Domicilia, Mores et Descriptiones, Anatomem atque Icones plurimorum. 3 Vols.
  • Peters, J. A. (1964) A Dictionary of Herpetology. Hafner,N. Y., 392pp.
  • Pritchard, P. C. H (1979) An Encylopedia of Turtles. TFH Publications Inc., N. J.
  • Ridgway (1912) Color standards and color Nomenclature. Washington.
  • Romer, A. S. (1956) The Osteology of Reptiles. Univ. of Chicago Press.
  • Schoepff, J. D. (1792) Historia Testudinum iconibus illustrata. Erlangen.i-xii.
  • Siebenrock, F. (1897) Das Kopfskelet der Schildkroten. Sitz.ber. Akad. Wiss. Wien, Mathem. naturwiss. Kl.; 106:245-328.
  • Siebenrock, F. (1909) Synopsis der rezenten Schildkroten, mit Berucksichtigung der in historischer Zeit ausgestorbenen Arten. Zool. Jb. Suppl. 10:427-618.
  • Siebenrock, F. (1913) Schildkroten aus Syrien und Mesopotamien. Ann. Nat. Mus. Wien, 27:171-225.
  • Strauch, A. (1862a) Chelonologische Studien, mit besonderer beziehung auf die Schildkrotensammlung. Mem. L'Academie Imp. des Sci. St. Petersbourg. Tome V, No.7.
  • Strauch, A. (1862b) Essai D'une Erpetologie de L'Algerie. Mem. L'Acad. Imp. Sci. St. Petersbourg. Tome IV, No.7.
  • Stubbs, D., Hailey, A., Tyler, W. and Pulford, E. (1981) Expedition to Greece; A Report. University of Lond. Union Nat. Hist. Soc.
  • Stubbs,D. et. al. (1984) Population ecology of European Tortoises: review of field techniques. Amphibia-Reptilia (5):57-68.
  • Tuck, R. G. (1971) Amphibians and Reptiles from Iran in the United States National Museum Collection. Bull. Maryland Herpet. Soc. Vol 7, No.3:48-86.
  • Walbaum, J. J. (1782) Chelonographia oder Beschreibung einiger Schildkroten nach naturlichen Urbildern.
  • Wermuth, H. (1952) Testudo hermanni robertmertensi n. subsp. und ihr Vorkommen in Spanien. Senckenbergiana, 33:157-164.
  • Wermuth, H. and Mertens, R. (1961) Schildkroten, Krokodile, Bruckenechsen. Jena. xxxvi, 422pp.
  • Werner, F. (1895) Ueber eine Sammlung von Reptilien aus Persien, Mesopotamien und Arabien. Verh. Zool. Bot. Ges. Wien, 45:14.
  • Werner, F. (1929) Beitrage zur Kenntnis der Fauna von Syrien und Persien. Zool. Anz. 81:238-242.
  • Werner, F. (1936) Reptilien und Gliedertiere aus Persien. Festschrift zum 60 Geburtstage von Prof. Dr. Embrick Strand, 2:193-202.
  • Werner, F. (1938) Reptilien aus Iran und Belutschistan. Zool. Anz. 121:265-271.
  • Wettstein, O. (1951) Ergebnisse der Ostereichischen Iran-Expedition 1949/1950, Amphibien und Reptilien. Versuch einer tiergeographischen Gliederung Iran auf Grund des Reptilien-verbreitung. Sitzungsber. Osterr. Acad. Wiss. Math.-Naturwiss. Kl. 160:427-428.
  • Williams, E. E. (1950) Testudo cubensis and the evolution of Western Hemisphere tortoises. Bull. Amer. Mus. Nat. Hist. 94 (9) :505-562.
  • Zarudny, N. A. (1903) Les reptiles, amphibiens, et poissons de la Perse orientale. Zap. Imp. Russ. Geog. Obs. Geog. 36(3):1-42.