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.

Notes on Dietary Constituents for Herbivorous Terrestrial Chelonians and their effect on Growth and Development

A C Highfield

This is the original article, based upon early Tortoise Trust research, that literally revolutionised the way people thought about feeding tortoises in captivity. When this was published, in the mid-late 1980's, almost all books, veterinary texts and articles were recommending diets that were high in protein and promoted fast growth. The vast majority of captive bred tortoises at this time developed shell deformities, and died within a few years. This article changed perceptions, and made a major contribution to advancing the whole concept of dietary management for captive chelonians.


Dietary related disorders represent a major cause of mortality in captive bred hatchlings and juveniles (Highfield 1986, Zwart 1987). Lambert (1986, 1988) reports that the median survivorship of UK captive bred Testudo graeca is 1.5 years, that of T. hermanni 1.75 years and T. marginata 2.3 years but does not provide any clinical reason for such mortalities other than to indicate that growth rates are more rapid than those attained in the wild, and that carapace deformities are characteristic of such animals. These median survivorship figures agree with those recorded by the present author from submissions made by members of the public and other inexperienced keepers. Detailed examination of the affected living and deceased specimens reveals that a combination of recognised clinical factors are invariably present, principally acute secondary nutritional osteodystrophy resulting from and inadequate dietary Ca:P ratio (Wallach and Hoessele, 1968) frequently combined with hepatic dysfunction and renal dysfunction urea and deposits of uric acid within the renal tubules. The latter condition is most often observed in cases where artificially high protein diets have been imposed by keepers upon herbivorous species which in their native habitat would experience much lower levels of dietary protein. High incidences of renal and hepatic disease have been noted for some time both in veterinary reports and necropsy surveys of long-term captive chelonians. Keymers survey of 1978 which details the result of 144 pathological examinations reports that G.I. tract, hepatic and renal disease accounted respectively for 27%, 1 1% and 9.7% of mortalities with miscellaneous nutritional disorders contributing a further 22.2% of the total. 77.8% of the study group had been housed in the London Zoo, the remainder being drawn from animals belonging to members of the public (Keymer, 1978).
A second survey by Rosskopf drew exclusively upon animals which had been maintained as pets. In this survey the findings were that 72.6% of mortalities exhibited severe hepatic disorders, 53% respiratory disease (principally pneumonia), 50.7% G.1 tract disorders, 40.6% renal disease and 34.3% cardiac dysfunction (Rosskopf, 1981). It is notable that in both surveys gastrointestinal disorders are particularly significant (with parasites accounting for some, but not all of these), and that the incidence of hepatic and renal disease is much higher in animals which have been maintained principally as pets. A currently unpublished survey of tortoises originating from pet keepers and subsequently acquired by the Tortoise Trust also reveals a high incidence of renal and hepatic disease. It is interesting to note that captive herbivorous iguanas often display almost identical nutritional disorders (Wallach and Hoessele, 1968, 1966).

Nutritional Disorders

The disorders may be usefully divided into two main groups consisting of:

a) diseases of excess

b) diseases of deficiency.

Clinically, combinations of the two groups are common, e.g., acute carapace distortion due to lack of dietary calcium and concurrent hepatic and renal dysfunction resulting from dangerously high consumption of saturated fats, proteins and nitrates.
In practice, it is possible to achieve a combination of normal growth and satisfactory osteological carapace development without incurring the mortality noted by Lambert provided adequate attention is given to the specific dietary requirements of the species concerned and that sufficient levels of all essential trace elements are present (Highfield, 1988). The dietary factor is of considerably more significance to growth and development that climatological influences, although the latter is frequently of great importance with regard to initiation of the breeding cycle (Reiter and Follett, 1980; Vivien-Roels, Arendt and Bradtke, 1979).

Dietary Calcium

The calcium requirement of chelonians is variable and increases during growth phases and in the case of females during egg formation. Given that a 100% growth-weight increase frequently occurs during the first eight weeks of life, it is clear that this period is especially critical. The calcium requirement is also dependent upon external factors such as the quantity of dietary phosphorous. Vitamin D3 availability is also of relevance. In the wild, tortoises from desert habitats typically experience a calcium to phosphorous ratio of between 5: 1 to 8: 1. The greater proportion of plants consumed are considerably richer in calcium than they are in phosphorous, are comparatively low in protein and are also very high in fibre. Furthermore, additional trace elements are consumed as soil, sand and grit particles. Zwart (1987) notes that a critical Ca:P deficiency ratio of 1.2:1 is typical and that at this level osteoporosis and osteomalacia (softshell syndrome) together with other deformities of the scutes recorded by Lambert (1983) manifest. Supplementation with Vionate (Ciba-Geigy) is frequently inadequate, as this product itself contains both calcium and phosphorous in the ratio of 2:1, and excess phosphorous in the principal diet rapidly reduces the ratio actually attained to critical deficiency levels. Certain food items used by some captive breeders are particularly disruptive of the calcium balance, especially legumes, sprouting seeds and processed cat or dog food of animal origin. This latter frequently has a negative Ca:P ratio by as much as 1:44 or more (Collins, 1971), whilst lOOg of peas typically contains 42mg Ca to 127mg P, broad beans 27mg Ca to 160mg P and mung sprouts 19mg Ca to 64mg P.
The high phytic acid content of legumes which mobilizes Ca is also cause for concern. Each of these items has in the past been recommended as a suitable substitute item for inclusion in the diet of hatching chelonians. Lettuce, which is frequently condemned as a dietary constituent, is in fact relatively neutral in the Ca:P sense, ranging from 20mg Ca to 22mgP/lOOg in the case of Iceberg varieties, to 68mg Ca to 25mg P/lOOg for Romaine varieties. Although inadequate alone, it is useful neutral base for further controlled artificial supplementation. Another common tortoise food item, dandelion, is of extremely high quality comprising 187mg Ca to 66mgP/lOOg and combining this with 14,000 i.u- Vitamin-A, 1. 6g fibre, 0 - 19mg thiamine and 2.7g protein. Banana fruit and leaves which are frequently utilized in tropical captive breeding projects as a readily obtainable staple dietary item are strongly Ca:P negative and liable to induce a relative Ca deficiency in rapidly growing animals (typically 8mg Ca to 30mgP/100g fresh, 32mg Ca to 104mgP/l00g dried).
Some food items which appear initially to represent good sources of dietary calcium are not as attractive on closer examination. Beet greens, kale, spinach and members of the Goosefoot family
Chenopodiaceae contain relatively high Ca levels but this is bound up with oxalic acid which reacts with calcium to form insoluble calcium oxalate. Opuntia cacti, a typical food item of arid habitat tortoises including Gopherus agassizzii and Geochelone elephantopus offers by comparison 1.89% calcium to 0.02% phosphorous (Rosskopf, 1982). Other self-selected native food items of Gopherus agassizzii reveal similarly positive Ca:P ratios.
The dietary preferences of
T. hermanni and T. graeca in the wild have seldom been recorded, notable exceptions being the work of Swingland (1984) and Stubbs who noted that 25% of the diet of T. hermanni consisted of Rubiaceae (Bedstraw family), 22% of Leguminosae (Peaflower family), 10% of Compositae (Daisy family) and 8% Ranunculaceae (Buttercup family). Comparative figures for T. graeca include 30% Plantago (Plantain family), 26% Compositae and 10% Rubiaceae. An average Ca:P ratio for the above is 3.5:1, and a typical (wet basis) protein content is 2.75% (Highfield, unpublished notes).
A practical solution is to botanically screen all dietary items carefully and exclude items which are strongly Ca:P negative from regular consumption. The remaining neutral or positive Ca:P constituents can then be supplemented with a multi-vitamin and mineral powder boosted if necessary by the addition of raw calcium carbonate until a true ratio of at least 5:1 is achieved. In practice, hatchling
T. graeca, Geochelone pardalis and Geochelone carbonaria raised on such a regime do not display the deformed carapaces and raised vertebral shields so typical of nutritional secondary osteodystrophy fibrosa which were remarked upon by Lambert (1986) and which results from an excess growth of keratin combined with the sub-optimum development of the underlying bony plates.

Vitamin D and Ultraviolet Light

Animals in their natural habitat are extremely unlikely to suffer hypovitaminosis-D3. Deficiencies are possible in captive animals which are deprived of access to sunlight or a suitable artificial U.V. source of sufficient intensity, e.g., True-lite @ (Durolite Corporation) or blacklight. Symptoms of deficiency include poor locomotion, osteomalacia and osteoporosis. Plant foods contain nil vitamin D. The skin of tortoises is however, extremely rich in oils containing sterols which react with U.V. to produce the vitamin and provided adequate U.V. exposure is attained oral supplementation is not necessary (Kauffield, 1969; Wagner, 1977). It is common for herpetologists to over estimate D3 demand and to grossly overdose orally. One possible consequence of this practice is metastatic mineralisation of the soft tissues (Barten, 1982; Wallach and Hoessle, 1966). Vitamin D3 is highly toxic and extreme caution should be exercised whenever oral supplementation is employed (Finlayson and Woods, 1977). If calcium and phosphorous are provided in suitable ratios and sufficient quantity and quality of U.V. are available, hypovitaminosis D3 is not at all likely. Human demand is for 10 micrograms (400iu) per day which can be obtained from as little as 3 hours exposure to sunlight. The requirements of tortoises are not known in detail, although Zwart (1987) suggest that 10-20,000 iu of D3 per Kg of general vitamin-mineral supplement dosed routinely at 2% food volume is an effective prophylactic measure where exposure to U.V. is inadequate.


B complex deficiencies have been recorded by the author in both hatchling and adult T. graeca and T. kleinmanni where animals have been maintained by owners on what amount to lettuce-only diets. The B group includes thiamine riboflavin, pyridoxine, nicotinic acid, pantothenic acid, biotin, folic acid and cobalamin. Clinical signs of deficiency include lack of neuromuscular coordination and pernicious anaemia. In the wild, it is probable that tortoises are able to obtain adequate levels of B12 to interact with dietary folic acid by gut microflora activity involving trace dietary cobalt. B12 is absent in vegetation, but essential to life. Although it has been recommended by some herpetologists (Reid, 1982; Engberg, 1980) raw fish is categorically not suitable dietary constituent for terrestrial chelonians due to its thiaminase content which prevents synthesis and absorption of B-group vitamins. Deficiency is particularly common following severe colitis and malabsorption syndrome resulting from pathogenic flagellate infection of the G.I. tract.

Dietary Iodine

Fibrous goitre or hypothyroidism is commonplace among captive herbivorous chelonians, particularly Geochelone elephantopus and Megalochelys gigantea maintained in zoological collections and which have been fed diets rich in vegetables containing high levels of anionic goitrogens (e.g. glucosinolates and thiocyanates) such as cabbage and kale. The condition has also been observed in hatchling T. graeca and T. marginata subjected to an identical dietary regime. The consumption of the responsible class of vegetables needs to be limited and in addition a multi-mineral supplement containing traces of iodine at a suggested dose rate of 6-10mg per Kg of supplement should by routinely provided with every meal. Vegetation from mountainous areas, or from soil rich in limestone may be lower than average in dietary iodine.

Dietary Fat

A disease of excess, steatitis or fatty infiltration of the liver is encountered frequently in captive chelonians and is an established cause of mortality in both hatchlings and in adults (Will, 1975; Rosskopf, 1981). Herbivorous chelonians are poorly equipped to metabolize saturated fats (Tammar, 1974), and when subjected to high fat content diets develop serious hepatosis resulting in jaundice and the inability to retain vitamin-A; hence hypovitaminosis-A is often concurrent clinically with steatitis. In the wild, virtually no saturated fats whatsoever are consumed, yet in captivity many keepers habitually provide dietary sources which are extremely rich in these substances. Tinned cat and dog foods are undoubtedly the worst offenders in this respect. Taking into account the normally reduced metabolic rate of most captive animals compared with wild specimens due to reduced temperatures and photo-periods, it is therefore not surprising that many mortalities are found post mortem to be suffering from obesity and gross fatty lesions of the liver. Botanical analysis of the native diet of T. graeca and G. agassizzii which may be assumed typical for species inhabiting similar biotypes, indicate that an average level of polyunsaturated fat consumption is 0.35g per lOOg of raw vegetation. Existing cases of steatitis may respond to veterinary treatment with thyroxine and vitamin-E.

Dietary Protein

Captive chelonians are frequently placed on diets richer in useable protein by a factor of several magnitudes greater than they would possibly be able to attain in the wild, and this, combined with a lack of dietary calcium is a major direct cause of mortalities in the median of 1.5 to 1.75 age group. The growth of such specimens is greatly accelerated, and full sexual maturity has been observed by the author in one captive bred male T. ibera at just 19 months of age. This animal weighed 565g and had a carapace length of 148mm. There was marked carapace deformity and overgrowth of keratin, the beak being extremely overgrown to the extent of interfering with normal feeding. Similar observations have been made in respect of Gopherus agassizzii hatchlings captive-bred in the USA, when calcium deficient protein-rich diets of vegetable origin were employed to rear hatchlings by amateur herpetologists instead of a diet of designed to replicate as closely as possible the naturally occurring diet of the species (Jackson and Trotter, 1976; Hansen, Johnson, Van Devender 1976). The photographs illustrating the first reference show grossly deformed carapaces typical of the most acute form of nutritional osteodystrophy fibrosa.
In addition to stimulating excessive growth, both generally and of the keratin in particular thus increasing the calcium demand yet further, the intake of high levels of protein has two further effects a) there is a direct effect upon calcium absorption ability (Margen, 1974) and b) high protein results in high levels of blood urea and consequently increases the amount of nitrogenous waste to be processed via the renal system. Unfortunately, reliable data on blood urea nitrogen and creatinine levels of wild populations is not readily available and most studies so far published have relied upon data extracted from captive specimens, mostly pet animals belonging to members of the public (e.g. see Rosskopf, 1982). Although useful, this data is unlikely to represent a true picture of the normal blood chemistry of specimens in the native habitat. It is not at all unusual for herbivorous chelonians which have been maintained on unnaturally high protein diets to exhibit symptoms of renal dysfunction. Dehydrated animals are obviously most at risk, as the accumulated uric acid becomes deposited not only in the renal tubules but invasively throughout the pericardium, liver and other organs (Frye, 1974; Wallach, 1971). Specific conditions recorded include interstitial nephrosis and glomerulosclerosis (Rosskopf, 1981). Once again, processed tinned pet foods of animal origin are particularly dangerous as not only are they high in protein (typically 17%) and saturated fats, rich in phosphorous and relatively low in calcium but most are also very rich in other nitrates. Cheese, another item which has seriously been suggested as an appropriate dietary component for
T. graeca is similarly adverse in effect (protein content typically 25%). Legumes are also spectacularly high in protein for a vegetable source (typically 10%), and should be avoided for the same reasons.
The protein requirement for most reptiles have not been studied in sufficient detail and no specific figures have been established for herbivorous terrestrial chelonians. Analysis of the native diet of
Gopherus agassizzii, which in many respects is typical of and habitat chelonian herbivores, suggests that the protein content of the food intake ranges from 1% (Opuntia sp. ) with grasses at a median content of 5% constituting a major part of the dietary intake (Rosskopf, 1982; Hansen et al, 1976). A safe upper protein limit for items which are regularly included in the diet would seem to be 7% as this is about as high as is ever attained in the wild by most species, even during peak periods of food availability. An average intake level of 4% would represent a close approximation of that experienced in the natural habitat.
Despite the lack of detailed information on protein demand, it is certain that the figure is very much lower Kg for Kg than mammals where 0.5g of usable protein per Kg would be a typical daily requirement. It seems probable that the daily requirement of a growing tortoise is in the approximate region of 0.20g of usable protein per Kg, although this may well vary considerably according to species and metabolic rate. Against this it should be noted that even such a low quality food item as lettuce contains an average of Ig per lOOg and most legumes contain well in excess of 7g/100g.
In the wild tortoises consume not only plant leaf material, but also seeds, fruits, flowers, roots, bark and grasses. See Samour, Spratt, Hart, Savage and Hawkey (1987) for an excellent survey which conveys the scale of the dietary eclecticism of
Megalochelys (Geochelone) gigantea, for example. The principal importance of this varied intake is not merely that it provides a relatively wide range of vitamins, minerals and fibre, but that by combining leaf, bark grass and seed sources an improved range of essential amino-acids is made available in a complementary process. This has the effect of increasing the potential Net Protein Utilization (NPU) factor of the diet. Where tortoises are provided with a suitable intake of essential amino-acids derived from the correct balance of vegetable matter they can exist perfectly satisfactorily on diets which are surprisingly low (in percentage terms) of protein content. Artificial amino-acid supplementation is not necessary provided a reasonable range of suitable food items are available although this has sometimes been adopted (bac,-)n, 1980). In most captive situations pathology results consistently indicate that it is protein excess rather than deficiency which is the principal danger. Meat based products have an extremely high NPU as they represent a complete amino acid source. lOOg of dog food containing 15% protein is, therefore, utilized at a greater rate than a vegetable containing an equivalent percentage of raw protein. In some circumstances and with some reptiles this may be of benefit, in the case of herbivores it is entirely inappropriate. The diet of a wild tortoise typically contains between 2%-6% plant protein (wet basis) which is utilized at an approximate rate of 55%. The high protein regimes adopted by some herpetologists for herbivorous chelonians include up to 20% protein which, because it is derived from amino-acid complete sources is utilized at a typical rate of 70%.
It is necessary to comment upon the claim often made that terrestrial chelonians from and habitats receive significant additional protein in the wild as a result of consuming carrion, arthropods and other insects.
This is not supported by faecal pellet analysis (Dearden, Hansen and Steinhorst, 1974) which indicates that such intake is so low as to be of virtually nil dietary significance. The levels paralleled that of other miscellaneous detritus also consumed, including small rocks, sand, bird feathers, lizard skin casts and mammal hairs (Hansen et al, 1976). Most tortoises will consume anything which is presented to them whether palatable and nutritious or not. Until separated from the public by barriers, giant tortoises at San Diego zoo were known to consume popcorn, balloons, yogurt, film wrappers, chewing gum and red-painted toenails (Bacon, 1980). Habitat analysis suggests that animals from and areas would not frequently encounter available sources of animal protein, but that animals from more humid habitats may have greater opportunities in this respect. Certainly, some T. hermanni may occasionally display an inclination to consume a passing slug or worm, but they make no effort to actively seek out such delicacies. In experiments conducted by the author, T. graeca,T. ibera,T. marginata and T. kleinmanni showed no interest whatsoever in slugs, worms or insects presented and indeed demonstrated active avoidance behaviour in many instances.

The Role of Symbiotic Digestive Microflora

The intestinal microflora of herbivorous chelonians is geared to processing relatively large quantities of fibrous, carbohydrate-rich cellulose matter. There is reason to suppose that protozoans and ciliate organisms play a role in this process together with the more unusual bacterial agents (See Fenchel, McRoy, Ogden, Parker and Rainey, 1979). Non-specific enteritis is common in captive collections (Keymer, 1978; Hunt, 1957) ac counting for up to 40% of total mortalities in some cases. Deficiencies of dietary fibre are certainly one factor (Shaw, 1961; Throp, 1969; Bacon, 1980) and an adequate intake of dietary fibre may also be of importance in controlling populations of potentially pathogenic parasites. An additional factor responsible for the high incidence of gastrointestinal disease noted may be inclusion of food items of animal origin to which the slow fermentation process of the herbivores digestive system appears ill suited (Bellairs, 1969; Holt, 1978; Dandifrosse, 1974; Sokol, 1967; Skoczylas, 1978).


The dietary requirements of captive herbivorous chelonians are far more complex than has previously been assumed by many keepers. The interaction and interdependence between the mineral, vitamin and protein metabolisms have rarely been studied in sufficient detail and much work remains to be done in this field. The role of the endocrine system in relation to the mineral metabolism is yet another area where currently much awaits discovery. It is apparent that the simplistic approach of providing a high quality diet in mammalian terms is totally inadequate to meet the real needs of chelonians which have a completely different set of requirements. Indeed, that which may represent a high quality diet for a mammal or carnivorous reptile may have entirely negative consequences when presented to a chelonian herbivore.

1997 The Tortoise Trust/A. C. Highfield