The taxonomical position of Cuora amboinensis is one that is argued about. It is placed in the family of Emydidae by some (Dijk, pers. comm.), whereas others place it in the Family of Bataguridae. In recent times it is most commonly accepted to belong to the family of Bataguridae (Dijk, pers. comm.) therefore this is the position we accept in this report. Table 1.1. describes the place of Cuora amboinensis in the animal kingdom.

Table 1.1 Taxonomy of Cuora amboinensis (Daudin, 1802).

(Colijn, 20-04-2001; Dijk, 2002)

Adult Cuora amboinensis attain a length of up to 20 cm with a high arched carapace (Ernst & Barbour, 1989). While researching more than 200 preserved specimens, Rummler and Fritz (1991) found the largest specimen to measure 217,5 mm (Rummler & Fritz, 1991). The carapace is uniformly dark brown to black, the plastron is yellow to light brown (Ernst & Barbour, 1989).

The species is described to occur from the Nicabar Islands, Bangladesh and Assam south through Burma, Thailand, Kampuchea, Vietnam, and Malaya, and east in Indonesia to Sulawesi and Ambinia. It also reaches the Philippines and Celebes (Ernst & Barbour, 1989).
Figure 1.2 shows the distribution of Cuora amboinensis and each of its subspecies.

Currently three subspecies are described:

Cuora amboinensis amboinensis (Daudin, 1802), (Terra typica: Island of Ambon, Moluccas)
The nominate form occurs on the Philippines, Mollucas, and Sulawesi. It is characterised by a flat, broad carapace with a distinct margin, and a plastral pattern with larger dark blotches than in the other subspecies. The carapace ratios for the subspecies C. a. amboinensis are shown in table 1.2 (Rummler & Fritz, 1991).

The populations of the Philippine Islands differ from other C. a. amboinensis through extremely narrow light head stripes, more extensive dark plastral blotches and different most common plastral formulae (Rummler & Fritz, 1991).

Cuora amboinensis kamaroma (Terra typica: 50 km north of Bangkok, Thailand)
This subspecies lives on the South-East-Asian continent and Borneo. It has a highly domed and narrower carapace than the other subspecies without a well-developed margin. The carapace ratio for the subspecies C. a. kamaroma is shown in table 1.2 (Rummler & Fritz, 1991).


Cuora amboinensis couro (Schweigger, 1812), (Terra typica: Java)
This subspecies occurs on Sumatra and Java. This subspecies is though to be derived from a former intergrade population between C.a. amboinensis and C.a. kamaroma having intermediate characteristics. The shell is of intermediate shape with or without an indistinct margin. The carapace ratio for the subspecies C. a. couro is shown in table 1.2 (Rummler & Fritz, 1991).

Table 1.2 Carapace ratio of Cuora amboinensis amboinensis, Cuora amboinensis kamaroma and Cuora amboinensis couro.

(Rummler & Fritz, 1991).
Table 1.2 indicates physiological differences between the three subspecies. The numbers above suggest the shape of the animals of each subspecies, especially the matter in which the carapace is domed. This difference in carapace shape is thoroughly described in an article by Filella (2000). A personal view of the authors is that the matter in which a carapace is domed could give an indication about the niche each subspecies fills. A more domed carapace limits the speed of a swimming individual but is stronger in withstanding predators or trampling hoofs of land animals. As a result of this the shape of the carapace could indicate the in-situ lifestyle of its bearer, possibly also indicating differences in diet.

Cuora amboinensis is a semi-aquatic turtle that inhabits lowland water bodies with soft bottoms and slow current, such as marshes, swamps, ponds, pools, streams and manmade flooded rice paddies. Although highly aquatic it is often found far from water. Ernst and Barbour (1989) refer to an article by Taylor (1920) stating that juveniles are entirely aquatic (Ernst & Barbour, 1989).

In 2000 this species was listed vulnerable on the Red List (Asian Turtle Trade Working Group) it was listed a CITES Appendix II species 19/07/00. Despite this, large numbers were still exported from Indonesia (18,000) and Malaysia (50,000) in 2001 (CITES, 01-04-2002)
Vietnam exported a total of 13440 animals between 1994 and 1999. Animals exported within the Asian region are most likely to be destined for consumption. Animals exported to America and Europe are often sold as pets (Dijk, Stuart & Rhodin, 2000).

The species Heosemys spinosa is placed in the family Emydidae, this position seems to be commonly accepted, in contrast to that of Cuora amboinensis.
Table 1.3. describes the place of Heosemys spinosa in the animal kingdom.

Table 1.3 Taxonomy of Heosemys spinosa (Gray, 1831).

(Sellmann, 01-01-1999).

This species can attain a carapace length of up to 22 cm (Ernst and Barbour, 1989). Herman (1993) describes two approximately 19-year-old specimens that both measured 203 mm, which is confirm with the mentioned maximum (Herman, 1993). The plastron of young animals is marked with a distinct ‘ray-pattern’, this is most distinct in half grown animals that are 12-15 cm in length and disappears with older age leaving a uniform blackish brown colour in fully grown animals (Mertens, 1971). On the outside of the carapace this species has spiny edges. In adult animals spines become less distinct and often disappear due to growth and wearing down, making them much smoother than juveniles. This species ranges from Tenasserim, Burma and southern Thailand southward through Malaysia to Sumatra and Borneo (Ernst and Barbour, 1989). Figure 1.4 shows the distribution of Heosemys spinosa.

No subspecies have been recognised in the species Heosemys spinosa, although great morphological differences seem to occur (Kerlen, pers. comm.). A carapace ratio for this species does not seem to be available.

Herman (1993) mentions this species lives primarily in rainforest streams at altitudes of 170 up to 1000 m, it is thought to be mostly diurnal and sluggish. This species is often reported to be both aquatic and terrestrial (Herman, 1993). Mertens (1971) quotes an article by MÒller (1906) stating that young animals are aquatic and adults terrestrial (Mertens, 1971).

In 2000 this species was listed Vulnerable on the Red List of IUCN. Despite high demand on the food market the number of animals traded in Indonesia has declined by 50% over a very short period of time, indicating rapid decline of the natural population. Its high demand in the food market, its scattered small populations and a low reproductive output are reason for a proposed endangered status (Dijk, Stuart & Rhodin, 2000). Information on capture- or export numbers is not available, this species does not appear in CITES legislation and therefore is not protected by international law, no numbers are available on capture or export. (CITES, 01-04-2002).

In this study data about the captive management two confiscated Asian turtle species were collected:
Cuora amboinensis; Malaysian Box Turtle
Heosemys spinosa; Spiny Turtle

A study was done to find the nutritional requirements of these two species, making it possible to give a nutritional advice for their captive management. The two species have been selected from the entire group of species because of expected dietary similarity. In literature both are said to be primarily herbivorous (Ernst & Barbour, 1989; Mertens, 1971). A literature review was carried out to compare in-situ data with collected ex-situ data about these species.

In order to give an advice about the nutritional requirements and the correct diet of Cuora amboinensis and Heosemys spinosa several questions needed to be answered. This paragraph describes the process used to answer these questions.

• What is the natural habitat of Cuora amboinensis and Heosemys spinosa?
• What is known about the digestive system of Cuora amboinensis and Heosemys spinosa?
• What is known about the natural diet of Cuora amboinensis and Heosemys spinosa?

These questions were answered by a literature review and by contacting herpetological specialists. A part of the literature review was conducted at the library of the natural historic museum 'Naturalis'. Throughout the review references were found in books, articles and on the Internet. Relevant publications were ordered through the library of the Van Hall Instituut and processed in different parts of the report.

To gain insight in the captive diets of Cuora amboinensis and Heosemys spinosa the following questions were composed:

• What is the composition and nutritional value of the diets currently fed to Cuora amboinensis and Heosemys spinosa in captivity?
• Which diets are fed at institutions that have a good reproduction success?
• What nutrition related diseases and deficiencies occur amongst Cuora amboinensis and Heosemys spinosa in captivity?

These questions were answered using data gathered by a questionnaire. The set-up of this questionnaire is explained in the next paragraph. It was sent out to zoos keeping the species and to the ESF studbooks that are being kept for the species Cuora amboinensis and Heosemys spinosa.
57 Zoos were contacted; of which 18 keep animals from the Hong Kong impoundment. ESF studbook keepers have not given out data on the number of participants.

In order to give the nutritional advice that was set out as the main goal of this project two questions were to be answered:

• How can knowledge about the composition of the in-situ diet be applied in an ex-situ situation?
• What is the most successful ex-situ diet?

These questions were answered by comparing the natural, in-situ, situation with data on ex-situ keeping of these animals and successes or failures in ex-situ keeping.
Comparison was made on a dry matter basis as is often done in zoo nutrition. This data was then processed further to make it possible to compare on energy basis. All diets were screened and compared on the following nutritional qualities:

Gross Energy
The intake of food is determined by the energy content. This means an animal will eat to satisfy its energy requirements, all other nutritional requirements need to be balanced with energy. The most accurate picture is given by looking at Metabolisable Energy (ME). However, since no indications were found as to how these turtles utilise proteins, fats, carbohydrates and crude fibre, the Gross Energy (GE) content of the diets is used in calculations and comparisons. Gross Energy values were used as given in literature, where needed it was calculated from levels of Crude Protein, Crude Fats, Carbohydrates and Crude Fibre as shown in Appendix VIII.

Crude Protein
Protein is important for the growth and repair and is also a source of energy (M. Knight, 1976). In turtles protein is also required for the growth of the carapaces scutes and the horny covering of the beak (Zwart, 2000). No indications have been found on the amino acid requirements of turtles, or reptiles in general.

Crude Fats
Fats have a high energy density, something that often is not wanted in turtle diets (Dennert, 2001). However fats do serve as a carrier for several vitamins (such as vitamin A and D, discussed below) and essential fatty acids (Palika, 1997). Data on essential fatty acids in turtle nutrition does not seem to be available.

Carbohydrates
Carbohydrates are the most important suppliers of energy, furthermore they can be converted to bodily fat, useful when a diet contains limited amounts of fat.

Crude Fibre
Herbivorous diets largely consist of crude fibre. Fibre is needed to activate the colon and is therefore crucial in the digestive process. Apart from this fibre can also supply energy to herbivorous animals that utilise hindgut fermentation.

These nutritional values are researched in every nutritional study and are of major importance for the maintenance of the animals.

Calcium
Reptiles very often suffer from calcium deficiencies due to malnutrition or wrong maintenance. (See also: Phosphorous and Vitamin D3) A lack of calcium is compensated by resorption from calcified body parts, resulting in bone-, and (in case of turtles) shell deformities (Zwart, 2000).

Phosphorous
Phosphorous plays an important role in the same bodily functions as mentioned under calcium. Apart from this it also is important in many enzyme systems. Excess of phosphorous may lead to symptoms similar to that of calcium deficiencies (Zwart, 2000).

Calcium : Phosphorous
For optimal usage of both calcium and phosphorous the ratio of the two minerals is very important. A wrong ratio can have the same results as a lack, or overdose of either of the two (Zwart, 2000).

Vitamin A
It is an essential vitamin for growth but the best known function of vitamin A is its influence on the functioning of the vision. Terrapins are often seen suffering from vitamin A deficiency due to insufficient nutrition (Zwart, 2000). Basically vitamin A is available in two main forms; as Retinol, which is the actual vitamin, and as b-carotene that can be utilised in an animal's body in the production of vitamin A (Retinol).

Vitamin B1
Vitamin B1 aids in different bodily functions such as enzymatic digestion and the working of the nervous system (Burger, 1995).

Vitamin D3
Vitamin D3 is essential in the absorption of previously mentioned calcium and phosphorous. Apart from utilising it in its original form as present in the diet, reptiles can also convert cholesterol into vitamin D3 by means of UV-B radiation coming from natural sunlight, or special light sources (Zwart, 1990).

Vitamin E
Vitamin E is an important factor for reproduction success and growth. This vitamin is especially important in the forming of eggs and the strength and shape of the young that is to grow in it. (Dennert, 2001)

Requirements of mentioned nutrients are researched in chapter 3, with additional calculation methods in Appendix VIII.

In-situ data and part of ex-situ data was gained by a literature review. The literature was collected at the libraries of the van Hall Institute, the Library ‘Naturalis’ in Leiden and by searching on the Internet. Table 2.1 shows the search engines and keywords that were used in the Internet search.

Table 2.1 Internet search engines, keywords used and dates of reference.

With searching criteria as listed above several written sources were found, in their turn supplying more useful references for the literature review.

In order to study the captive management of the five turtle species private keepers and zoos needed to be questioned about the data on their population, enclosure and the diets of the animals kept. A questionnaire was designed to collect the information of concerning institutions.

The questionnaire used to get these and other data was based on a questionnaire made by Schils and Smeets (2001) who researched the dietary needs of Aldabra Giant Tortoises (Dipsochelys dussumieri) (Schils & Smeets, 2001). Because the study by Schils and Smeets had a good response the questionnaire for this project was made in a similar style. The original questionnaire about Aldabra Giant Tortoises was altered to match the five turtle species. Institutions were asked to fill in the questionnaire separately for each species.
Some questions in the original questionnaire by Schils and Smeets proved to be difficult to answer by participating zoos (Huisman pers. comm.) In an attempt to overcome some of the problems encountered in the Aldabra Giant Tortoise study the new questionnaire was more simplified. Questions asking for food amounts now ask specifically for grams, thus limiting the chance of getting answers that are too diverse to be compared in analyses.

In this study three groups of keepers of Cuora amboinensis and Heosemys spinosa were recognised; those that do not keep animals from the impoundment (keepers of non-Hong Kong animals), those that do keep Hong Kong animals (keepers of Hong Kong animals) and private owners of Cuora amboinensis and Heosemys spinosa participating in ESF studbooks. Different questionnaires were made for these three groups, making it possible to ask for specific data on the animals kept.

1. Keepers of non-Hong Kong animals
The questionnaire made for keepers of animals that are not from the Hong Kong impoundment consisted of:

• Introduction: This part was made to get general information about the animals. Information about their sex, age and weight is asked to get insight in the demographics of the population.
• Housing conditions: These were asked to get insight in enclosure size, temperature and UV-radiation, this is done to reflect back to successful or unsuccessful keeping of the species.
• Feeding information: This was asked to get insight in feeding regime, amounts and ingredients of diets fed.
• Reliability: In order to estimate the reliability of given information questions were included about using nutritional programs and measuring/weighing of the diet and supplementations.
• Diet success: In an attempt to judge the successfulness of a diet questions are asked about reproduction success and overall health of the animals kept. Reproduction success is measured by clutch sizes, numbers of eggs fertilised, number of young, and first month/first year mortality of the young produced.

The institutions keeping one or more of the species were looked up at the ISIS abstracts homepage, reptile section (ISIS, 02-05-2002). The e-mail addresses of these institutions were provided by Rotterdam Zoo.

2. Keepers of Hong Kong animals
The keepers of animals coming from the Hong Kong impoundment were also asked the questions described above. However, to get insight in the acceptance of nutritional advice such as was given at the distribution of these confiscated turtles an additional set of questions was included in this questionnaire. This part of the questionnaire dealt with the following topics related to the information-flow within the organisation:

• Distribution advice: A number of questions were asked to get insight in the difference between the diets fed at the institutions and the advice that was given by EZNC at distribution.
• Organisation: A number of questions are included in order to see which people in zoo organisations receive the given nutritional advices, who processes them and who should receive such advices.
• Advice: Institutions are asked for their opinion about the format in which they would like to see the nutritional advices given to them.

The addresses for keepers of Hong Kong-animals were provided by Mr. Henk Zwartepoorte.

3. Private keepers
A number of reptiles kept in private collections are managed in studbooks. These studbooks are co-ordinated by the European Studbook Foundation (ESF). ESF Studbook keepers for the species Cuora amboinensis and Heosemys spinosa were asked to participate in this study in order to get insight in the situation private keepers maintain. ESF studbook members have been approached through these studbook keepers only.
Even though several members of these studbooks have received animals from the impoundment, the questionnaires used were similar to those sent to zoos not keeping Hong Kong animals. This difference was made because the specific questionnaires designed for keepers of Hong Kong animals dealt with the information-flow within the organisation. Private keepers would not have been able to answer most of these questions, therefore the choice was made to eliminate this part for private keepers all together.

All questionnaires were sent by e-mail to eliminate the cost of postal services and to reduce the time lost by shipping and handling of regular mail.
The questionnaires were accompanied by an introducing letter, making up the message part of the e-mail. This part explained the need of the study and the way it was conducted. This letter was signed by Dr. Ir. Walter Jansen and Joeke Nijboer, B.Sc. who assigned the project, and by Mr. Henk Zwartepoorte, closely involved with the implementation of the project. These names were included to emphasise the importance and value of the project.
The questionnaires also contained an example document, describing the situation of Cuora amboinensis at Rotterdam Zoo. This was done to clarify the way the questionnaire was to be filled out.
For the complete questionnaire and the introducing letter see Appendix III.

The data that was gathered by the questionnaires was stored and processed in Excel tables. In these tables all given data was processed and nutritional compositions of individual ingredients were used to calculate the composition of the total diet. To define the ingredient composition of the food items fed Nevo (2001), Souci (1986), Zootrition ì, Dietary Management Software (2001), Dierenfeld et al. (1996), Mader (1996), Markwell & Hurley (2001) and other publications were used. When the nutritional composition of ingredients could not be found comparisons with similar food items were made. The same was done when a few of the wanted nutrient values were missing in otherwise useable data. A few calculations had to be made to fill in missing nutritional values. Appendix VIII shows the assumptions and calculations made.

The ingredients and nutritional compounds of the diet fed by each zoo and private keeper are discussed in chapter 4 and completely given in Appendix IV.

Data on nutritional requirements of the animals was simplified to one easily comparable list of data. This simplifying was done by calculating all demands for an average representative of the particular species into demand per MJ GE of food. For these calculations average food intake and energy demand were withdrawn from the questionnaires and a standard bodyweight of 1000 grams was assumed.

Demands for nutrients are given on different levels by different authors. Several calculations had to be made to come to the same level (per MJ GE). Examples of different calculations made are given next:

• Calculating the amount of an energy-supplying nutrient with existing demand as a percentage of total energy supply, assuming ME = GE.

Crude protein = 15%GE/ (average energy density) = demand CP (g)= (15% * 1,00 MJ)/ (0,02412 MJ/ g) = 6,22 g

• Calculating phosphorous demand /MJ GE from known calcium/MJ demand and Ca/P ratio:

Phosphorus = 0,714 (max Ca)/ 1,25 = 0,571 g /MJ GE

• Calculating vitamin demand/ MJ GE from known mg/kg level:

Demand mg/MJ GE = Demand mg/kg * weight food (kg)/ energy content food (MJ GE)

Vitamin B1 = 4,4 mg/kg x 0,09266/ 0,367 = 1,11

Very little is known about the digestive system of Cuora amboinensis. Being herbivorous it is very likely to utilise hindgut fermentation and thus have large intestine with a relatively large capacity. (Dennert, 2001). No specific information on the digestive system of Cuora amboinensis could be found.

Cuora amboinensis often is referred to as being largely herbivorous in the wild (Ernst & Babour, 1989; Mudde, 1981, 1982). Liat & Das (1999) specifically mention wild animals to eat aquatic insects, molluscs, crustaceans, fungi and worms, next to the bulk that is made up by plant matter. However no indication is made as to what part of the diet is made up of this animal matter (Liat & Das, 1999). The first faeces that were produced by Hong Kong impounded Cuora amboinensis consisted of compact plant mass. This mass contained a lot of long fibres, indicating the presence of grasses and/or reed in the natural diet at time of capture (Kerlen pers. comm.).

According to Liat and Das (1999) wild Cuora amboinensis are semi-aquatic, feeding both on land and in water. It is also said to be a good swimmer suggesting that it is capable of catching insects and crustaceans that Liat and Das also mention as part of the diet of Cuora amboinensis (Liat & Das, 1999).

No literature seems to be available on the digestive system of Heosemys spinosa. It is likely Heosemys spinosa utilises hindgut fermentation for the digestion of vegetable matter.

Smith (1931) who is cited in Mertens (1971) found indications of the natural diet of Heosemys spinosa. The species is described to feed entirely upon vegetable matter (Mertens, 1971). Kerlen, who received five Heosemys spinosa from the Hong Kong impoundment, reported the presence of seeds in the first faeces that were produced in captivity (Kerlen pers. comm.). The seeds involved have not been identified, they are shown in figure 3.1. The presence of these seeds indicates that the natural diet of these animals at time of capture (November/December 2001) at least partly consisted of seed bearing plants.

Liat and Das (1999) report this species to be a poor swimmer. As a result of this it is unlikely this species relies heavily on deep-water plant species as a part of its natural diet (Liat & Das, 1999).

A limited amount of literature was found on the species Cuora amboinensis and Heosemys spinosa. To extend the amount of data the search was expanded to similar species.
The choice of this study to collect data of similar species was based on the assumption that the diets of herbivorous freshwater turtles are basically the same. Freshwater areas in similar climate zones throughout the world are likely to contain similar plant species, thus providing similar ingredients to the turtle’s diets.

According to Mader (1996) energy requirements of reptiles are strongly depended on the temperature of their surroundings and the related body temperature (Mader, 1996). Figure 3.2 already showed the intake of aquatic turtles at different temperatures. Food intake in basic is determined by the amount of energy needed (Burger, 1993). As a result of this figure 3.2 thus shows an increasing need for energy when temperature rises. Mader (1996) determines daily energy requirements by estimation from research data, extrapolation from domestic species and by practical experience (Mader, 1996).

A Standard Metabolic Rate (SMR) describes the energy needed to meet the cost of essential work done by the cells and the organs of the individual (Burger, 1993).
The SMR of turtles in MJ/day is estimated by Mader (1996) to be 0,1344*(W ). Where W is bodyweight in kg. The outcome of this formula must be multiplied by a number between 1,1 and 2,5 according to activity (Mader, 1996). Although Energy requirements of reptiles also are heavily influenced by temperature (Parmenter and Avery, 1990) and, as in every animal, by physiological status (Schmidt-Nielsen, 1997) this is not included in the equation.
This equation is determined in animals at 30°C, a temperature that, according to Mc Arthur is within the preferred temperature range of Asiatic Box turtles (27-30 °C) (Mc Arthur, 1996).
Dennert (2001) mentions the Metabolic energy of reptiles to be; ME = 0,042*(W )
This value also must be corrected for level of activity. Dennert describes those as follows:

Animals at rest/sleeping: x 1,25
Adults with little activity: x 1,50
Adults active: x 2
Growing animals, little activity: x 2
Egg carrying animals: x 2
Animals under stress: x 2-2,5.
(Dennert, 2001)

Protein
Protein levels advised by Mader (1996) are shown in table 4.1. This table clearly shows the difference in protein amounts advised for herbivores and carnivores (Mader, 1996).
Zwart (2000) reported that the recommendations concerning the protein content of the food differ between 5% in the Californian desert tortoise (Gopherus agassizii) and 13% in snapping turtles (Chelydra spp. and Macroclemmys spp.). However, mentioned in the same article is an extensive study on red-eared terrapins (Trachemys scripta elegans) fed a diet containing 24-27% protein (Zwart, 2000). No indications were given whether these levels are for an ‘as fed’ diet or on Dry Matter basis. Comparing levels to those advised by Mader (1996) listed below it is likely to deal with levels in the diet ‘as fed’.

Fats
Mader (1996) describes the fat requirements for reptiles as a part of the Metabolisable Energy (ME) (Mader, 1996). The values given by Mader (1996) are shown in table 4.1. Again there is a large difference between the demands of herbivorous and carnivorous animals.

Carbohydrates
Carbohydrate levels in reptile diets as advised by Mader (1996) are given in table 4.1 (Mader, 1996

Table 4.1 Estimated nutritional energy requirements
for captive reptiles as percentage of ME.

Fibre
Not much is mentioned about fibre requirements of turtles. Zwart (2000) does mention the need of roughage because of its influence on the motility of the intestinal tract (Zwart, 2000). This is a quality that fibre is known for in many different animal species.
Mader (1996) also emphasises the importance of fibre for the gut motility. Herbivorous turtles are said to need at least 12% of DM to consist of fibre, where 20-30% would be closer to optimal (Mader, 1996).

Calcium
Calcium is of major importance in many metabolic processes in the body. Calcium requirements are said to be somewhere between 428 and 714 mg/MJ. These values are based on experiences with domestic mammals and birds, exact requirements are not available (Mader 1996). Zwart (2000) mentions a level of 1% Calcium in dry matter (Zwart, 2000).

Phosphorous
Calculating back from Calcium/ Phosphorous ratio’s as given by Zwart (2000) results in a phosphorous requirement of 0,7-1,7% in dry matter. Something that Zwart (2000) also backs up by stating that a 0,8% level of phosphorous in dry matter is adequate (Zwart, 2000).
Mader (1996) states reptiles’ diet should contain 0.6-1% Phosphorus (DM), a level that is in line with the range given by Zwart (2000).

Calcium/ Phosphorous ratio
Regardless of the actual levels of calcium and phosphorous separately, the ratio between the two is of great importance. Exact requirements of this ratio are not know in full grown animals, but from a great number of observations and extrapolations of other animals it is concluded that a calcium contents of 1,0% in combination with a phosphorus content of 0,8% (in dry matter) are adequate, meaning a Ca/P ratio of 1,25:1 (Zwart, 2000).
Also stated by Zwart (2000) is a diet containing 1,4 to 2,0% calcium in the dry matter at a Ca:P ratio of 1,2:1 or even 2:1 (Zwart, 2000). The Ca/P ratio is something that is rather disputable, ranging to as high as 4:1 or even 6:1 (McArthur, 1996) Although the higher values are likely to apply to egg producing animals, something that is also seen in birds (Huisman, pers. comm.).

Vitamin A
The daily requirement of vitamin A of chelonians is estimated to be 400 IU/kg bodyweight being 2.800 IU/kg bodyweight per week (Zwart, 2000). Dennert (2001) gives a vitamin A requirement of 1.500 IU/kg bodyweight per week (Dennert, 2001). Vitamin A is Retinol, b-carotene often is mentioned as being vitamin A. However it is actually pro-vitamin A that chelonians utilise in the production of vitamin A (Retinol). Hypervitaminoses can occur in the intake of Retinol, b-carotene is not toxic and does not pose a problem if given in high amounts.
If one is to look at amounts of b-carotene it is important to realise that 6 IU of b-carotene are needed to produce 1 IU of Retinol. (See also Appendix VIII).

Vitamin B1
Zwart (1990) proposes a vitamin B1 concentration ranging between 4.4 and 11 mg/kg food (Zwart, 1990).

Vitamin D3
The nutrient requirement in vitamin D3 is related to the Ca:P ratio in the food. There is still no exact data available on the vitamin D3 requirement. If the data from higher vertebrates and other reptiles is extrapolated the vitamin D3 requirement varies between 10 and 100 IU per kg bodyweight per day, being 70-700 IU per kg BW per week. The higher estimation is said likely to be somewhat too high (Zwart, 2000). Dennert (2001) advices 150 IU / kg bodyweight every week (Dennert, 2001). Mader (1996) gives a level of 200-1.000 IU/ kg dry matter in the diet (Mader, 1996).Vitamin D3 can also be produced by the reptiles’ own body under influence of UV-B radiation coming from unfiltered sunlight or special UV-B lighting (Bruins, 1999). In this process cholesterol is transformed into vitamin D3 (Zwart, 1990).

Vitamin E
Dierenfeld (1989) advices a vitamin E level between 1142 and 5709 mg/MJ (Dierenfeld, 1989).

A summarisation of all parameters mentioned above is given in table 4.2.

Table 4.2 Summarised requirements of general turtle diets.

The requirements summarized in table 4.2 can be calculated for both species when a ‘standard’ animal is chosen to fill in missing data. In table 4.3 the requirements of Cuora amboinensis are given assuming the animal weighs 1000 grams, weekly diet fed is 92,66 g, of which 22,93 g is dry matter, and the amount of Gross Energy is 0,367 MJ. Is set to 1000 grams for easy calculating and comparison, other values are averages derived from questionnaire diets given by zoos and private keepers (see Appendix IV and table 5.8).

Table 4.3 Dietary requirements for Cuora amboinensis, assuming bodyweight is 1000 grams, diet as fed is 92,66 grams, dry matter is 22,93 grams and Gross Energy content is 0,367 MJ.

Assumptions for Heosemys spinosa are based on a bodyweight of 1000 grams, and ‘as fed’ weight of the diet of 507,71 grams with a dry matter of 71,96 grams and a GE content of 1,024 MJ. Weight is set to 1000 grams for easy processing and comparison, other values are derived from questionnaire diets given by zoos and private keepers (See Appendix IV and table 5.10).

Table 4.4 Requirements for the diet of Heosemys spinosa assuming bodyweight is 1000 grams, diet as fed is 507,71 grams, dry matter is 71,96 grams and Gross Energy content is 1,024 MJ.

In case of malnutrition animals are likely to eventually show this by a changing physiological condition. This paragraph explains the disorders most commonly associated with poorly fed turtles, and reptiles in general.

Protein
A deficiency in proteins may occur when an inadequate diet is fed in captivity that consists exclusively of the outer leaves of salad and fruits. Periods of protein deficiency may lead to disturbance in growth of scutes of the carapace and the horny covering of the beak as well as to failure in reproduction (Zwart, 2000).
Protein over dosage can lead to growth problems such as pyramiding shell scutes, unproportionated limbs and shell size and overweight (Mc Arthur, 1996) gout (Zwart, 2000) and according to Palika (1997) cited in Schils and Smeets (2001) it can even cause kidney failure.
When the diet composition suddenly changes an overfeeding with proteins could occur. This overfeeding may lead to disturbance of the intestinal flora, with production of abnormal products, diarrhoea and large amounts of gas (Zwart, 2000).

Fats
Disturbance in the fat metabolism does occur. Yellow fat disease and in addition also a few cases of focal degeneration of fat tissues have been described in red-eared terrapins (Trachemys scripta elegans) (Zwart, 2000).

Hypovitaminosis A (vitamin A deficiency) in terrapins
A deficiency in vitamin A is most common in omnivorous and carnivorous terrapins. If older animals that are caught in the wild are transferred to a deficient diet in captivity they may develop clinical signs after a period of several months to one year, depending on the reserves in vitamin A. Vitamin A deficiency may lead to gastrointestinal disturbance such as anorexia, diarrhoea or even excessive salivation. Bulging and swelling of the eyelids are clinical signs of hypovitaminosis A. Reptiles that suffer from hypovitaminosis are likely to be deficient in other vitamins and minerals as well (Zwart, 2000).

Hypovitaminosis B1 (vitamin B1 deficiency)
Only a few indications of hypovitaminosis B1 occurring in tortoises and terrapins are know. A deficiency in vitamin B1 may lead to anorexia, poor growth and chronic loss of weight in spite of sufficient uptake of food. A clinical sign is that the eye is retracted in the orbita (Zwart, 2000).

Hypovitaminosis D3 (vitamin D3 deficiency)
When a vitamin D3 deficiency occurs a change in the diet is essential. The vitamin D3 deficiency was solved in reacted in red-eared terrapins (Trachemys scripta elegans) by feeding a product (Carmix, Hope farms, Woerden, the Netherlands) containing 11,65% Ca, 0,4% P and 20.000 IU vitamin D3. This was applied by rubbing the product intensely into pieces of meat (Zwart, 2000). Vitamin D3 is produced under the influence of UV-B radiation. Correct lighting in captivity or the presence of unfiltered sunlight overcomes the need of supplementation with vitamin D3 (Bruins, 1999; Mc Arthur, 1996).

Calcium deficiency
A deficiency in calcium leads to nutritional osteodystrophia fibrosa. The clinical signs of osteodystrophia fibrosa are a breakdown of the original bone and the production of a fibrocollagenous connective tissue. Osteodystrophia fibrosa can be prevented by feeding selected greens and fruits. Greens, which have a Ca:P ratio higher than, or close to 1, are preferable (Zwart, 2000). Examples of such greens are: dandelion (2,64), endive (1,86) and pak-choi (2,00) (Mader, 1996; Souci, 1986; Nevo, 2001).

Problems with Calcium, Phosphorous and vitamin D3 are popularly known as rachitis (Göltenboth & Klös, 1995) or MBD (Metabolic Bone Disease) (Davies & Davies, 1997) both are terms often used to describe deficiency diseases associated with a shortage in calcium or vitamin D3 or an unequal ratio of calcium and phosphorous. This disorder is associated with weak and soft bones, often surrounded by connective tissue. Animals suffering an advanced stage of MBD often have deformed body parts, turtle sufferers are easy recognisable by soft shells (Davies & Davies, 1997).

Litophagy
In several species of tortoises and terrapins litophagy, the uptake of stones occurs. Regular uptake of larger masses of stones leads to intestinal obstruction (Zwart, 2000).

Geophagy
Geophagy is the uptake of sand and is to some degree physiological in tortoises and contributes to a firmer substance of the faeces. But an exceptionally high uptake of sand may lead to intestinal impaction and atony and eventually causing death of an individual (Zwart, 2000).

Despite the fact that wild Cuora amboinensis are mostly referred to as herbivorous, this species often is fed a carnivorous diet in captivity. An article by Hofstra (1994) describes a diet used for captive bred young C. amboinensis. These animals are fed on a diet of mealworms, heart, beef (whole pieces as well as minced and mixed with Gistocal), earthworms, dry cat food, cod filet and large quantities of pulverised chicken egg scales. The author feeds his adult animals almost exclusively on dry cat food (Hofstra, 1994).
Mudde (1981) also is aware of the fact that literature describes this species as herbivorous. According to his experience the animals are more carnivorous and occasional plant eaters, which is an experience, he has with most aquatic turtles. His animals were fed shrimp, earthworms, slugs and snails (both terrestrial and aquatic). Less popular were meat (no specification) and fish, where only tiger barbs (Barbus tetrazona) were accepted. The author does note the fish have to be largely immobilised by cold temperatures or a low water level for the turtles to catch them. Guppies (Poecilia reticulata) were never accepted, neither live or dead, even though a population of this species constantly lived in with the animals. Plant matter was accepted by his group of C. amboinensis and the author mentions apple, banana, and cucumber as well as floating plants kept in the enclosure (Mudde, 1981).In a 1982 article by the same author he describes all things ever accepted in the diet of this species, no quantities or frequencies are mentioned though;
Animal matter: Shrimps, commercial dry turtle food, soaked dry cat food, canned cat food, pieces of day-old chicks, meat (no mentioned of which animal), pike (Esox lucius), cod (Gadus morhua), smelt (Osmerus eperlanus), coalfish (Pollachius virens), worms, mealworms, crickets, wax worms, mosquito larvae, Daphnia, Mysis, snails, slugs, Gammarus, woodlice (Porcellio scaber).
Plant matter:banana, cucumber, several water plants, boiled rice, beans, peas, soaked corn and even shrimp crackers, wholebran biscuits and popped corn were accepted (Mudde, 1982).

Limited information is published about the captive diet of this species. A report by Herman (1993) describes the diet of three captive Heosemys spinosain the Zoo of Atlanta (USA) and the offspring they produced. The animals had a preference for tomatoes (Lycopersicon esculentum). Tomatoes were the only things that were accepted in the first year of captivity. After this year other fruits and vegetables were also accepted. Finally mixed salads were given, consisting of several fruits, vegetables and collard greens (no species were mentioned). Animals also accepted chopped, skinned mice that were presented occasionally. Why these mice were skinned is not mentioned. Eventually the author mentions the quantity of mice in the diet was drastically cut back because of the animals becoming obese (Herman, 1993).An article by MÒller (1906) describes a captive youngster that feeds on fruits, salads and aquatic plants (including duckweed). Apart from this the turtle also accepted animal matter such as earthworms, mealworms and red meat (Muller, 1906).

This paragraph describes the results from the questionnaires that were sent to zoos and private keepers of Cuora amboinensis and Heosemys spinosa. Zoos and private keepers that sent back the questionnaire and data on the overall response of the questionnaire are given in Appendix V.

This paragraph will answer the sub-question:

• What is the composition and nutritional value of the diets currently fed to Cuora amboinensis and Heosemys spinosa in captivity?