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.

Preliminary report on the Taxonomic, Biotypic and Conservation status of the land tortoises of Tunisia

A C Highfield

Field research assistant: L. E. King

Introduction
Little detailed work has been published on the terrestrial tortoises of Tunisia which have generally been assumed by most authors to be purely standard specimens of Testudo graeca graeca LINNAEUS 1758 demonstrating little or no regional variation from those of type locality (Oran, Algeria) or from those of Morocco. The most important papers published to date although providing some useful data on distribution and conservation do not attempt to re-examine the taxonomic status of these tortoises, nor do they provide any morphometric data or illustrations. Field work recently undertaken by the author suggests that a re-examination is overdue as the tortoises of Tunisia are remarkably different in several important respects from those of the neighboring countries of Algeria or Libya and may represent several unique and taxonomically separate species or sub-species requiring a major revision of currently accepted nomenclature.

Distribution & published records
The most important data on the distribution of tortoises within Tunisia is that provided by Blanc (1978) which lists by region many important centres of population. The majority of observations are recorded in the north of the country within the Tunisian Atlas (Khroumirie) mountain range principally in the Teboursouk, Beja, Jendouba, Testour, Le Kef and Ain Draham areas which are typified by a high annual rainfall and are thickly vegetated with holm and cork oak forestation. Further observations of coastal and steppe habitat populations are noted in the Grombalia, Nabeul and Kairouan areas. Additional recent data on distribution is given by Lambert (1982 and personal communications) who lists previously unpublished observations in the Tunis and Hammamet regions and in the oases of Gabes.
Other useful papers on distribution and habitat include those by Mayet (1903) who provides good general data and an extremely important observation on egg morphology until now overlooked, and Mosauer (1934) who lists observations in the Rades, Sidi bou Ali and Kairouan regions.
Historically important records of chelonians in Tunisia are also provided by Gadeau de Kerville (1908) who records T. graeca, Emys orbicularis and Mauremys caspica leprosa in the Khroumirie, by M. Blanc (1935) who notes records of T. graeca in the Khroumirie, Cap Bon, Kelibia, Sousse, Kairouan and in the region of Tunis with additional records of M. c. leprosa and E. orbicularis including Ain Draham, Tabarca, Zaghouan and Grombalia. Other interesting (but less detailed) records are provided by Boulenger (1891) who cites T. graeca in the ruins of Zarzis, by Konig (1892), Olivier (1896), Anderson (1892) and by Seurat (1922 and 1927). Useful accounts of some sympatric reptile species are also provided by Mayet (1903), Mosauer (1934), Chpakowski and Chneour (1953), Domergue (1956-1966 and 1959), Escherich (1896), Lanza and Bruzzone (1960), Laurens (1975), Mocquard (1895) and Thilenius (1897) whilst Mertens (1929) describes in some detail the herpetofauna of Belvedere Park in Tunis and that of the oases of Gabes.

General nomenclature and taxonomy of north African terrestrial tortoises
For many years it was assumed that all terrestrial tortoises in north Africa with the exception of Testudo kleinmanni LORTET 1887 which occurs in Egypt and Libya belonged to the nominate subspecies Testudo graeca graeca LINNAEUS 1758 (terra typica = Oran, Algeria). Common synonyms include Testudo mauritanica DUMERIL & BIBRON 1835 and Testudo ibera PALLAS 1814 (actual type locality = Georgia, U.S.S.R and which does not occur in north Africa). For a brief analysis of the morphology and history of the holotype of T. g. graeca see Highfield (1990a). Recently the present author has also demonstrated that two other full species also exist in the region; Testudo whitei BENNETT 1836 (Highfield & Martin, 1989) and Testudo flavominimaralis HIGHFIELD & MARTIN 1989. The former is the largest species of mediterranean land tortoise so far known to science, routinely reaching an adult body mass of over 3kg and a straight-line carapace length of 250mm by 25 years of age with the largest individual recorded measuring 292mm and weighing 4.550kg at 30 years of age (Highfield, in press). Testudo flavominimaralis is by comparison a dwarf or miniature species, which even at a great age rarely exceeds 135mm or 500g. It should be noted that in both cases field workers observing these animals have almost invariably assumed that specimens of Testudo whitei BENNETT 1836 were actually very old specimens of Testudo graeca graeca L. 1758 (or earlier, specimens of Testudo marginata SCHOEPFF 1792; e.g see Gray, 1870 and Gervais, 1836) and that adult Testudo flavominimaralis were merely juvenile T. graeca L. 1758. In fact, both T. whitei BENNETT 1836 and T. flavominimaralis HIGHFIELD & MARTIN 1989 possess a number of absolutely unique cranial and osteological features including neural and suprapygal constructions which conclusively separate them from the Testudo graeca of Linnaeus (Das & Highfield, in preparation).
Mention should be made in this context of the alleged sub-species Testudo graeca terrestris FORSKAL 1775 which according to Wermuth (1958) is found in Libya. Recent taxonomic studies have demonstrated this attribution to be entirely erroneous and that it definitely does not occur anywhere in north Africa (Highfield & Martin, in press). There are in any event no alleged records of this tortoise for Tunisia.

The current taxonomic status of north African land tortoises is therefore cited as follows;

  • Testudo graeca graeca LINNAEUS 1758
  • Testudo whitei BENNETT 1836
  • Testudo flavominimaralis HIGHFIELD & MARTIN 1989
Notes on measurements and photographic records
During field work it is important to record as many details as possible of individual specimens examined. Some of these details are of peripheral taxonomic value (although they may prove of considerable value in other study areas such as growth and development), but others are of primary taxonomic importance; the following listed measurements are of special diagnostic value:

  1. Straight-line carapace length (L)
  2. Carapace length over curve (LoC)
  3. Median body width (Mw)
  4. Maximum width at marginals (MwM)
  5. Body width at M1 - M2 (front) marginal suture (Fw)
  6. Maximum carapace height at V2-V3 (Ch)
  7. Weight (W)
  8. Plastral suture lengths P1, P2, P3, P4, P5 and P6.
  9. Post anal gap (PaG); (of prime importance in female specimens)
A more complete set of measurements will also include length x width of V1-V5, M1-M11, supracaudal, nuchal, C1 - C4 and on the plastron dimensions of the intergular, gular, humeral, pectoral, abdominal, femoral and anal plates.

It is absolutely essential that all specimens are photographed both in colour and black and white; the following views will be found most useful;

  1. Directly from above
  2. Plastron
  3. Head-on frontal view showing V1
  4. Rear view towards showing V5 and supracaudal
  5. Side view showing marginals and costals
  6. Close-ups of head from above and sides
  7. Frontal close-up of head showing beak
It is also good practice to photograph each animal exactly as found in situ giving general views of the environment and location. If possible a shade and open temperature reading should be taken together with data on relative humidity.
Where eggs are to be recorded a length, width and weight should be taken; these should also be photographed against a 25mm graduated scale.
Finally, it is vitally important to note that great care must be taken not to assume that all very small tortoises are juveniles or that all large examples are of great age - by 15 years of age a Testudo whitei BENNETT 1836 will often surpass a 50 year old specimen of Testudo graeca graeca LINNAEUS 1758 by over 200% in both length and weight. With practice, it becomes relatively easy to recognise genuine juveniles. Until this skill is acquired however, mistakes often occur.
It is interesting to note also that according to Flowers records (1944) a juvenile T. whitei BENNETT 1836 exceeds the dimensions of even a fully grown adult Tunisian tortoise of maximum attainable size from the Sidi Kalifa region by 9 years of age. These statistics are confirmed by independent records in our possession (Highfield, in press).

The value of comparative data
Each population must be examined and recorded as a separate entity; data collected from diverse populations, if integrated at random, often leads to inaccuracy and confusion; a good example is the inclusion of data from T. whitei in many reports on Algerian T. graeca graeca. This problem can be avoided by providing each individual specimen or record with a unique identification code thereby allowing every set of data to be sorted or re-sorted later according to revised criteria and with a permanent record of exactly where the specimen referred to was found. Without location data records are of extremely limited value. By keeping accurate records of where specimens were located it becomes much easier to draw valid conclusions as to the presence of unique localised characters, regional clines, or make decisions which may ultimately affect the genetic integrity of a population, e.g repopulation by captive breeding or the release of transported or previously captive animals.
The value of accurate records from clearly identifiable sites cannot be overestimated. It should be noted that the practical uses of such data are considerable; for example it becomes possible to identify with a high degree of accuracy the precise area from which a tortoise originated by comparison of its characters with data from others on file - this has obvious applications in conservation enforcement and wildlife protection, making the detection and prevention of illegal trading or exportation very much easier and more effective.
We are confident that the taxonomic characters and indicators of individual populations within Tunisia are sufficiently diverse to permit correlation of translocated specimens to specific sites with a very high degree of accuracy.

Observations of tortoises 25/12/89 to 30/12/89 in Tunisia

Locality records

  1. Between Nabeul (7km) and Grombalia (21km) (25/12/89)
    adult females; 118mm (L) X 104mm (Mw) and 130mm (L) X 107mm (Mw)
    12.45 to 1.20 pm both feeding on sun exposed slopes. Soil temperature 27.6 oC, shade temperature 23.2 oC. Relative humidity circa 55%.
  2. Sidi bou Ali (26/12/89)
    1 adult male; 113mm (L) X 79mm (Mw)
    This tortoise was examined in a house in the village where it was being kept for utilisation as a medication. The tortoise had been purchased by the inhabitants from a mobile door to door salesman who had recently passed through the area.
  3. Enfidaville (26/12/89)
    1 adult male; 112mm (L) X 75.5mm (Mw)
    This specimen was being kept as a pet in a large walled garden. Morphologically identical to Sidi bou Ali example.
  4. Enfidaville (28/12/89)
    1 adult female; 134mm (L) X 99.5 (Mw)
    1 adult male; 122mm (L) X 83mm (Mw)
    Pair of pet tortoises kept in garden. History of successful captive breeding. Both specimens clearly of identical race, but different from previous specimens.
  5. Forest of Sidi Kalifa (28/12/89)
    1 adult male; 105mm (L) X 82mm (Mw)
    Observed grazing at 11.00 am. Ground temperature 14oC, shade temperature 18.5oC. Relative humidity 60%. Altitude approximately 400m.
    This specimen morphologically identical to specimens 2 and 3 (above). Clearly these also were collected from this locality.
  6. M'hadba (29/12/89)
    2 adult females; 138mm (L) X 101.5 (Mw) and 134mm (L) X 95mm (Mw)
    2 juvenile females ; 96mm (L) X 74mm (Mw) and 67mm (L) X 87mm (Mw)
    Observed in olive grove. Smallest juvenile estimated age 18 months - 2 years; larger juvenile estimated age 3 years. All 4 specimens were identically marked and were very different in general appearance from the previous specimens observed.
  7. Nabeul (30/12/89)
    1 small juvenile female; photographed but not measured.
    On display and being offered for sale (10 TD) from street sellers stall alongside mounted scorpions and other natural history objects. When questioned, stallholder claimed that many were available. When asked who purchased them, he became evasive but eventually stated that tourists were the intended customers and that they were being purchased as pets for illegal export to Europe. At 10 TD only a foreigner would be likely to purchase as this price is very excessive by Tunisian standards. When this was pointed out the stallholder explained that tortoises were in demand by foreign tourists as 'Les tortues grecque' were no longer available legally in Europe.
  8. Hammamet (30/12/89)
    4 adults; photographed but not measured.
    Observed being offered for sale from a small crate outside a shop in the medina. Price asked = 20 TD each. Shop owner questioned and replied that he was aware that sale and export from Tunisia was illegal, but that he could make a large profit from selling tortoises to tourists. English tourists nearby observed us examining the tortoises and assumed that we were interested in making a purchase; they then approached us to inform us that they believed export or sale of these animals in Tunisia was illegal as they had seen the educational posters issued by the Department of Forests which are displayed in some hotels.
  9. Hammamet (30/12/89)
    28 very small juveniles (dead); 55mm - 80mm (estimated).
    On sale in shop called 'Le Souk D'Art Artisanal D'Hammamet'. Preserved and coated with varnish. Price asked = 6.500 TD each. This shop accepts Access, Visa, American Express and Diners Club credit cards. Other native Tunisian fauna also on sale including scorpions. Estimated age of specimens approximately 1 to 6 years of age. All apparently collected locally and killed specifically to supply the illegal 'souvenir' export trade.
Note on exploitation
Despite our observations of the encouragement of illegal export of tortoises by a minority of traders, it is clear that the Tunisian authorities are promoting effective educational efforts directed at discouraging such illegal and destructive activities. The poster dealing with the export of prohibited flora and fauna issued by the Direction General des Forets does illustrate tortoises and other reptiles in addition to birds and prohibited plants . This is obviously read and understood by many tourists as confirmed by our encounter in the medina. All of the dealers we spoke to were also fully aware that they were engaged in an illegal and prohibited trade which was contrary to the wishes of the Tunisian government and Department des Forets. Regarding the local utilisation of individual tortoises by villagers for medicinal purposes (the flesh is used for stomach illnesses and the blood is also sometimes used as a cure for warts), we feel that this is less of a threat than the commercial illegal export trade which in our opinion does pose a serious danger to some already depleted low density populations. The souvenir trade in dead juveniles has a particularly destructive ecological potential and should definitely be suppressed.
We have also received reports of identical trade in other areas of the country; live juvenile specimens were on sale in the summer of 1989 in the medina of Sousse and preserved specimens similar to those observed by us in Hammamet were on sale in El Djem (J. Bruekers, Netherlands. Personal communication).
Ministry of Agriculture directive (Art.6 alinea 5) of 2nd August 1973 prohibits the collection, buying, door to door trading, selling and export of Tunisian tortoises.

Note; Blanc (1978) quotes then current prices of tortoises in Tunisia as between 50 and 120 millimes - compare to the illegal dealer in Hammamet medina now asking 20 Dinars.

Morphometric data on recorded specimens
For the purpose of this preliminary report specimens of only one of the 4 distinct and readily identifiable populations observed will be described for the purpose of comparison with Testudo graeca graeca L. 1758 and T. whitei BENNETT 1836. In order to make clear the comparison male specimens only are described.

This particular group was selected for special study for the following reasons;

  1. We already had computer data on file for specimens conforming exactly to this groups characters - this data was obtained from tortoises exported from Tunisia via the bulk pet trade before the present conservation ban on export was adopted.
  2. This data is invaluable in that it provides records of lengths and weights over a period as long as 30 years for some individual specimens thus permitting an unparalleled study of growth and development for the population type.
  3. We have several living specimens of this tortoise in our reference collection in England which we have carefully monitored for some years; these animals are all survivors of the bulk pet trade which were donated to us by pet keepers no longer wishing to retain them.
  4. We have an exact locality in which tortoises of this type can be found in their natural habitat (Forest of Sidi Kalifa).
  5. e) Our museum reference collection includes osteological material for this group.
Morphological observations
Specimens 1 and 2 measured in Tunisia. The measurements for T. g. graeca LINNAEUS 1758 and T. whitei BENNETT 1836 are averages obtained from a sample of over 50 specimens recorded on our computer. Specimen number 3 is a tortoise identical in appearance to the Sidi Kalifa specimens but which has been in captivity in England for 20 years. This specimen is believed to be circa 45 years of age and visible growth has now ceased. Specimens 1 and 2 both displayed some slight evidence of continuing growth but were definitely mature adults with an estimated age of circa 15-18 years.

Our records contain further data on several other examples which are identical to specimens (1) and (2) observed in Tunisia in all important diagnostic characters. Of these, one is of special interest as the tortoise is obviously of very extreme age and has a confirmed history in captivity in England of precisely 31 years. Comparison of scute surface deterioration with other examples of Testudo spp. of confirmed age suggests an estimated age of between 70 - 80 years for this specimen.

The primary measurements of this very ancient tortoise are as follows;

L = 120mm, Ch = 55mm, Mw = 75mm, MwM = 80mm. The body mass is 360g.

There has been no measurable growth in any parameter since the tortoise was obtained by the present keeper 27 years ago. Comparison with specimens (1), (2) and (3) suggests that we can therefore conclude that 120-125mm straight line carapace length probably represents the maximum dimensions attained by males of this race even as very mature adults. This compares with a measured absolute maximum recorded of L = 188mm, Ch = 102mm, MwM = 118mm for Moroccan-Algerian T. graeca graeca and average carapace lengths of circa L = 155mm for the same species (Highfield & Martin, 1989; Lambert 1982 & 1983).
In comparative dimensional terms, the tortoises of Sidi Kalifa are slightly larger than Testudo kleinmanni LORTET 1887 (Flower, 1933), but significantly smaller than Testudo flavominimaralis HIGHFIELD & MARTIN 1989. As such they represent one of the worlds smallest Testudo spp.

Carapace scute measurements
Further dimensional statistics obtained from the Sidi Kalifa tortoises are as follows;

Supracaudal; 30.5 X 18.5 & 29 X 18.5
The supracaudal shape is not consistent with that of T. g. graeca from the type locality of Oran, Algeria but is significantly more projected beyond the marginals in a manner which somewhat resembles that found in T. kleinmanni LORTET 1887. Frontal vertebral scute
The f.v.s is also not consistent with a diagnosis of T. g. graeca (Highfield, in press), being significantly less round and more angular.

Beak construction
Bi-cuspid rather than tri-cuspid as T. g. graeca (Loveridge & Williams, 1957).

Adult Skull dimensions

Skeletal and cranial analysis of this tortoise is at an early stage of research, but some useful observations already made include;

  • Basal length = 25mm
  • Quadrate width = 18mm
  • Supraoccipital - Basioccipital separation = 5mm
  • Frontal width = 7mm
Full details of comparative cranial osteology will be published subsequently.

Suprapygal osteology
The suprapygal construction of the Tunisian tortoises differs completely from that defined as in accordance with T. g. graeca LINNAEUS 1758 (Loveridge & Williams, 1957).
In the Tunisian specimens there are two suprapygals, the anterior being the largest and enclosing the posterior between two symmetrical posteriorly directed rami; the posterior suprapygal bone semi-oval on its anterior face.
In T. g. graeca L. 1758 there is a fused suprapygal sometimes laterally divided by a straight transverse suture. The rear facing horns (rami) are absent. This suprapygal pattern is considered general for Testudo as a whole, including T. ibera PALLAS 1814 from eastern Europe.
This result suggests that not only are the Tunisian specimens not T. g. graeca LINNAEUS 1758 but that they do not even conform to currently accepted standards of diagnosis for the genus Testudo (Bour, 1989). That the suprapygal constructions of the nominate species T. graeca LINNAEUS 1758 and the Sidi Kalifa population are entirely different is a finding of some very considerable taxonomic and conservation significance.

General notes on osteology of the carapace
Typically alternate octagonal and quadrilateral anterior neurals; xiphiplastra hinged; two suprapygals as described above.

Plastral markings
Bour (1987) has demonstrated the importance of plastral markings in taxonomic diagnosis; however, in the case of T. graeca it has usually been assumed that great individual variation occurs. This view is not shared by us. In all cases we have found that variation within separate populations or species is very slight. In the case of the Tunisian tortoises there are major differences between the Grombalia (Cap Bon) population and the Sidi Kalifa population (see plates). More specimens from both locations need to be studied, but it is likely that these differences will prove consistent.

Parasitology
Microscopic examination of a faeces sample taken from one of the Sidi Kalifa tortoises revealed a zero nematode count; a very low flagellate and ciliate count and no evidence of additional intestinal parasites. This is in contrast to French T. h. hermanni studied by Bruekers (1986) where nematodes were very common. The sample also revealed a relatively high intake of grity silicate material and a high dietary fibre intake. No external parasites (ticks, etc.) were noted on any specimen examined.

Carapace condition
One specimen was observed with severe crushing damage to the marginals and costals (M' Hadba, adult female). These injuries appeared most likely to have been caused by dropping or by vehicle impact. The tortoise was in otherwise good health and feeding normally. Another specimen (Grombalia) had suffered carapacial damage consistent with having survived a fire. There was no evidence of carapace malformation due to congenital or dietary causes on any specimen examined.

Egg morphology
It has often been assumed that the eggs of the various races and species comprising the homogenous grouping generally referred to as 'Testudo graeca' are of identical construction. Where differences have been noted these have usually been attributed to individual variation. In fact each species group within what has generally been regarded as the 'graeca' complex produces a very characteristic size and shape of egg with a similarly characteristic clutch size. Although a wide variety of egg forms are encountered the range of individual variation within each species is surprisingly small - the eggs may therefore be usefully employed in taxonomic diagnosis.
The size, shape and weight of tortoise eggs from specified species have rarely been studied in detail. In particular, studies relating to Testudo graeca have tended to include data from races other than the true Testudo graeca described by Linnaeus in 1758 and based upon the holotype of Edwards, 1748. i.e, most studies have randomly grouped egg data from what were previously regarded as sub-species of Testudo graeca or from unrecognised forms, thus obscuring the very real differences which actually exist between the various species in their 'pure' forms (e.g see Cooper, 1983 and Smith, 1985).
One of the most significant and important findings to emerge in this present study was the alleged minute dimensions of the eggs of some Tunisian tortoises; we were repeatedly shown drawings of eggs which were very much smaller than anything we have previously encountered in any land tortoise. Unfortunately, due to the time of year of our visit, no actual egg material was available for examination.
We find, however confirmation of these truly remarkable egg dimensions in V. Mayets (1903) paper on Tunisian reptiles. Mayet reports that the eggs measure a mere 15mm X 13mm.
For comparison, the following dimensions should be noted; this data clearly demonstrates the tremendous difference between the alleged egg dimensions of some Tunisian populations and anything else currently recognised.

Average egg dimensions of some species of tortoise (From; Highfield, in preparation).

All lengths in millimeters, all weights in grammes.

In comparison to these, the alleged sizes of Tunisian tortoise eggs at only 15mm long X 13mm wide are of extraordinary significance.
Even the diminutive Testudo kleinmanni LORTET 1887 lays eggs which by comparison to the reported egg dimensions of the Tunisian tortoises are enormous - T. kleinmanni eggs typically measure approximately 23mm X 29mm (Flower, 1933, Kammerer, 1906).
There is no doubt that Mayet (1903) refers to land tortoises in this context as he specifically mentions their nesting behaviour and hard calcified egg shells. This matter is of major taxonomic importance, and suggests that the tortoises referred to are undoubtedly very unique animals. We are pleased to recognise at last this important scientific observation by M. Mayet on the tortoises of Tunisia which has been overlooked for 87 years.

Hybridisation & genetic interaction between populations
More data is required on movement of individuals, degree of sympatricity and environmental limitations on transit between the various populations before any conclusions can be reached on this subject. However, initial impressions are that the populations experience little if any transit of individuals for various environmental reasons which again require further research. Individual tagging of specimens could prove a useful technique for researching this question.
Neo (1978) reports that heterozygous values tend to be higher in reptiles than in birds or mammals, however island populations are more generally homogenous (Frankel and Soule', 1981). Because of their highly specific biotypic requirements and generally isolated habitats with minimal transit of individuals between populations a high degree of genetic convergence is often found within groups of reptiles. Where a greater exchange of individuals occurs, in more easily traversed and larger habitat areas then a state of balanced polymorphism may occur between homozygous, dominant homozygous and recessive individuals (Croudace, 1989). This occurs due to the phenomena of ''heterozygous advantage'' described by Strickenberger (1968) where the heterozygous individual has greater reproductive fitness.
The effects of isolation, species evolution and genetic inbreeding suppression in Tunisian tortoises are at present unknown but require further study (Chesser, 1983).

Habitat and biotype

It is obvious that Tunisia offers an excellent range of habitats which are very well suited to the requirements of terrestrial tortoises and which are capable of supporting high density populations; furthermore, most of these habitats are well managed and many are under the direct control of the Direction General des Forets.
The main threat from development and industrialisation concerns the coastal populations which are also under threat from tourist exploitation. These populations should therefore be considered as at highest overall risk. Most inland sites appear reasonably secure from habitat loss due to development and there is definitely less of a threat from collecting for the tourist souvenir trade. There are obviously some localised exceptions to this general observation however.

General biotypes of Tunisian tortoises
Tortoises are most often found in grazing and pasture land, on sunny but well vegetated undulating slopes, among rocky forest verges and to some lesser extent in olive groves and traditionally managed agricultural zones. Tortoises are much less common in areas of intensive agriculture.
Tortoises are not common in arid zones or found at all in desert regions, other than in oasis microclimates. Preferred humidity range would appear to be between 50-70%. Lambert (1983) suggests that limiting factors in the distribution of north African land tortoises include minimum mean annual rainfall and vegetation. In Morocco Lambert notes tortoise records in areas where the mean annual rainfall (P) = 1062mm - 1112mm. In Tunisia, especially high density populations are noted in the Khroumirie mountains (Ain Draham, Teboursouk etc.), one of north Africas areas of highest annual precipitation.
In very hot weather or in cold winters the tortoises retreat under rocks or hide in thick thorny bushes. In mild winters the tortoises of the coastal zones do not hibernate. The egg laying and mating periods of the mountain tortoises are later than those of the coastal plain and hill populations.
Altitude range is 0-800m (Blanc, 1978). In December we found some tortoises at 400m which were still active and feeding, but further inland tortoises at 500m were not to be found. According to local observers, all of the tortoises had retreated under the rocks for the past 6 weeks and were expected to emerge again in February, March and April. In these locations we recorded ground and shade temperatures of 12oC and 14oC respectively at 12.00 a.m. Vegetation in these same areas was more sparse, mainly Opuntia, Mandragora officinarum and similar species.
In coastal sites and in the Cap Bon tortoises were found on gentle rocky slopes among very lush vegetation; some sympatric plant species included Opuntia, Leontice leontopetalum, Lupinus spp., Genista spp., Cistus spp., Cyclamen spp., Convolvulus spp., Plantago spp., Arisarum vulgare (plentiful), Tulipa spp., Taraxacum spp., Tamaricaceae spp., etc.
A more detailed botanical checklist of tortoise habitats will be published separately.
In every location we found tortoises the ground was light in texture but not completely sandy; the soil was invariably populated with a thick carpet of plants many of which are recognised food plants of the tortoise.

RECOMMENDATIONS & CONCLUSIONS

Exploitation and illegal export
Educational efforts directed at preventing tourists from purchasing tortoises or objects derived from them for illegal export need to be continued and increased; it would be advantageous if all hotels displayed posters and had publications available listing prohibited items and explaining the need for conservation of Tunisias unique flora and fauna. The assistance of the tour operators could also be enlisted in distributing such material. More information at Airports would be especially useful.
The activities of street sellers and souvenir shops in tourist areas need to be subject to continual monitoring to prevent violations. Enforcement action against offenders would have a useful deterrent effect.

Habitat
The impact of land development and agricultural usage on tortoise populations needs to be further studied.

Genetic integrity of populations
It is clear that there are massive differences between the genetic make-up of certain individual populations - even among those which may be in relatively close proximity to each other. This feature urgently requires additional research.
In the meantime, until more is understood of the genetic relationships of the various populations the artificial movement of tortoises between different sites should be strongly discouraged. No captive breeding releases should be made for the same reason unless their genetic descent can be traced with certainty to the proposed release site.
This finding also indicates that the protection of individual sites and populations may require a much higher conservation priority than is currently accorded.

Taxonomy and nomenclature
The tortoises of Tunisia are not identical to the tortoises of Morocco or those of the type locality of Testudo graeca graeca LINNAEUS 1758 in Algeria.

The reported egg dimensions, confirmed by Mayet (1903) indicate that at least some populations may belong to an as yet unclassified and genetically separate species which is unique to Tunisia.

The carapace osteology is also very different to that of T. g. graeca LINNAEUS 1758 and indeed does not even conform to currently accepted standards permitting a diagnosis of Testudo at genus level.

The conservation implications of these findings are profound and may require extensive revisions of currently accepted nomenclature and/or diagnostic criteria at both species and genus levels.

FURTHER RESEARCH PROPOSED

Further work is required to establish the following;

  1. Average comparative clutch sizes and egg dimensions from each population group.
  2. Hatchling size and weight, plus data on comparative early phase development.
  3. Skeletal variations between populations; we would stress that this work should be done without recourse to taking specimens for killing in order to obtain data. Adequate biological material is available by other means, e.g utilisation of seized or confiscated 'souvenir' specimens etc.
  4. Seasonal activity pattern recording. Much more data is required on comparative mating, nesting and hatching periods for example in mountainous, coastal and steppe habitats.
  5. Population stability. Long term monitoring required.
  6. Food chain interaction with other species, e.g carnivorous mammals and avian predators. Nest site damage and hatchling predation rates.
  7. Environmental data. e.g how activity is affected by weather, temperature and humidity etc. Are there differences between individual population types?. External morphology suggests that evolutionary environmental adaptions may have occurred at certain locations.
  8. Average longevity and dimensional statistics. Much more field data is required.
  9. More data is required on hibernation activity and behaviour in the wild of the various populations - especially comparative data between the southern and northern populations and between coastal and mountain region populations.
  10. Transit of individuals between populations; more data required. A tracking system might be worth investigating. Combined with systematic data recording and marking this should provide a good deal of useful information.
  11. Dietary intake; more information required to establish precise dietary preferences and food value averages.