Captive Breeding and Marketing of Turtles

Home , Chelodina burrungandjii, Chelodina canni, Chelodina expansa, Chelodina steindachneri, Chinese box turtle, Elseya dentata

by Grahame J.W. Webb, S. Charlie Manolis and Michelle Gray

February 2008

RIRDC Publication No 08/012 RIRDC Project No WMI-3A

ISBN 1 74151 601 3 ISSN 1440-6845

Captive Breeding and Marketing of Turtles Publication No. 08/012 Project No. WMI-3A

The information contained in this publication is intended for general use to assist public knowledge and discussion and to help improve the development of sustainable regions. You must not rely on any information contained in this publication without taking specialist advice relevant to your particular circumstances.

While reasonable care has been taken in preparing this publication to ensure that information is true and correct, the Commonwealth of Australia gives no assurance as to the accuracy of any information in this publication.

The Commonwealth of Australia, the Rural Industries Research and Development Corporation (RIRDC), the authors or contributors expressly disclaim, to the maximum extent permitted by law, all responsibility and liability to any person, arising directly or indirectly from any act or omission, or for any consequences of any such act or omission, made in reliance on the contents of this publication, whether or not caused by any negligence on the part of the Commonwealth of Australia, RIRDC, the authors or contributors.

The Commonwealth of Australia does not necessarily endorse the views in this publication.

This publication is copyright. Apart from any use as permitted under the Copyright Act 1968, all other rights are reserved. However, wide dissemination is encouraged. Requests and inquiries concerning reproduction and rights should be addressed to the RIRDC Publications Manager on phone 02 6271 4165

Researcher Contact Details Grahame Webb PO Box 530, Sanderson, NT 0813 Wildlife Management International Pty Limited (WMI) Phone: (08) 89224500 Fax: (08) 89470678 Email: [email protected]

In submitting this report, the researcher has agreed to RIRDC publishing this material in its edited form.

RIRDC Contact Details Rural Industries Research and Development Corporation Level 2, 15 National Circuit BARTON ACT 2600 PO Box 4776 KINGSTON ACT 2604

Phone: 02 6271 4100 Fax: 02 6271 4199 Email: [email protected]. Web: http://www.rirdc.gov.au

Published in February 2008 Printed by Canprint

ii Foreword

The farming of Hawksbill turtles in northern Australia is a potentially viable wildlife industry. Like crocodile farming, overcoming some technological obstacles associated with captive breeding may be a pre-requisite before international trade, currently prohibited by the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES), takes place. The demand for Hawksbill turtle shell is high, and captive breeding provides the quickest route through CITES for legal international trade. Preliminary research indicated high growth rates of animals under captive conditions, and the current work confirmed early attainment of maturity. Research only spanned one breeding season, but indicated that social, behavioural, physiological and environmental factors may be implicated in successful nesting in captivity. The prototype breeding pen provided basic conditions for growth and reproduction. With successful nesting, small numbers of females would be required to produce sufficient numbers of hatchlings for farming purposes. An industry based on Hawksbill turtles could provide tangible economic benefits for indigenous landowners in regional areas, and provide positive commercial incentives for conservation of sea turtles and their habitats.

The demand for freshwater turtles for food, particularly in China, provides potential opportunities for Australian species. Australian species are sought after in the international pet trade, but current Federal legislation prohibiting export of live animals for commercial purposes is an impediment to accessing and/or testing these market opportunities. Domestic consumption (as food) appears to be a viable option for turtles in the short-term, and perhaps in Traditional Chinese Medicine, a rapidly expanding industry in Australia.

This project was funded from RIRDC core funds which are provided by the Australian Government.

This report is an addition to RIRDC’s diverse range of over 1700 research publications. It forms part of our New Animal Products R&D sub-program that aims to accelerate the development of viable new animal industries.

Most of our publications are available for viewing, downloading or purchasing online through our website:

• downloads at www.rirdc.gov.au/fullreports/index.html • purchases at www.rirdc.gov.au/eshop

Peter O’Brien Managing Director Rural Industries Research and Development Corporation

iii Acknowledgments

We are especially grateful to Boyd Simpson and Jessie Rutter for their assistance in raising the Hawksbill turtles leading up to this study, and to Cathi Shilton (NT Department Primary Industry, Fisheries and Mines) who performed most autopsies and provided interpretation of histological and pathology results. Patty Richards (Darwin) carried out ultrasonography on the female Hawksbill turtles.

We are also thankful to Wan Ziming, (Beijing), Wan Quihong (Hanzhou City) and Mac Min Jiang Li (Panyu City, Guangzhou) who provided information on freshwater turtles in China, and to Erin O'Brien who assisted in the collation of information on Australian freshwater turtles.

iv Abbreviations

CITES Convention on International Trade in Endangered Species of Wild Fauna and Flora

IUCN IUCN-World Conservation Union

SCL straight carapace length

CCL curved carapace length

TCM Traditional Chinese Medicine

v Contents

Foreword...... iii Acknowledgments...... iv Abbreviations...... v Executive Summary- ...... x 1. Introduction and Background...... 1 2. Objectives...... 6 3. Methodology ...... 7 3.1 Captive Breeding of Hawksbill Turtles...... 7 3.1.1 Breeding Pen ...... 7 3.1.2 Water Circulation ...... 9 3.1.3 Water Parameters ...... 9 3.1.4 Cleaning ...... 10 3.1.5 Temperature...... 10 3.1.6 Nesting Banks ...... 11 3.1.7 Running Costs ...... 11 3.1.8 Turtles...... 12 3.1.9 Inventories...... 12 3.1.10 Foods and Feeding ...... 13 3.1.11 Illness and Mortalities ...... 14 3.1.12 Ultrasonography ...... 14 3.2 Marketing of Turtles...... 14 4. RESULTS...... 15 4.1 Captive Breeding of Hawksbill Turtles...... 15 4.1.1 Breeding Pen ...... 15 4.1.2 Temperature...... 15 4.1.3 Water Parameters ...... 18 4.1.4 Turtles...... 18 4.1.5 Sex Ratios...... 20 4.1.6 Mortalities ...... 21 4.1.7 Maturity...... 22 4.1.8 Mating Behaviour...... 25 4.1.9 Nesting ...... 25 4.1.10 Growth Rates...... 26 4.1.11 Foods and Feeding ...... 27 4.1.12 Running Costs ...... 28 4.2 Marketing of Turtles...... 29 4.2.1 Asian Trade in Freshwater Turtles ...... 29 4.2.2 Trade...... 29 4.2.3 Food...... 31 4.2.4 Traditional Chinese Medicine ...... 32 4.2.5 Pet Trade and Religious Ceremonies ...... 32 4.2.6 Turtle Farming...... 33 4.3 Australian Turtles...... 34 4.3.1 Pet trade...... 36 4.3.2 Food...... 37 4.3.3 Traditional Chinese Medicine ...... 38 4.3.4 Turtle Shell...... 38

vi 5. Discussion...... 39 5.1 Captive Breeding of Hawksbill Turtles...... 39 5.1.1 General ...... 39 5.1.2 Breeding Pen ...... 39 5.1.3 Reproduction ...... 40 5.1.4 Growth Rates...... 42 5.1.5 Mortalities ...... 42 5.1.6 Foods and Feeding ...... 42 5.2 Marketing of Turtles...... 43 6. Implications...... 44 7. Recommendations ...... 45 8. References ...... 46

vii List of Tables

Table 1. Mean monthly water temperatures measured in the morning (am = 0800-0900 h) and afternoon (pm= 1600-1700 h) in the breeding pen, and the difference between the two mean readings.

Table 2. Morphometric measurements of 19 ranched and one wild-caught E. imbricata prior to introduction into the breeding pen. SCL= straight carapace length; SCW= straight carapace width; CCL= curved carapace length; CCW= curved carapace width; HL= head length; BWt= bodyweight. *= wild-caught individual.

Table 3. Shell grading of 20 adult E. imbricata on 21 October 2003. Turtles are listed from the "least desirable" colouration to the "most desirable".

Table 4. Initial housing of adult E. imbricata in the breeding pen on 27 October 2003. SCLi= initial straight carapace length. See Table 3 for details of "Shell Grade".

Table 5. Numbers and mean diameter (md; in mm) of three size classes of ovarian follicles recorded in 6 female E. imbricata. L= left ovary; R= right ovary.

Table 6. Female E. imbricata listed in order of size, with assessment of reproductive status. Immature? = no sign of follicular development by ultrasound; * = reproductive status ascertained by ultrasonography.

Table 7. Growth rates (GR; mm SCL/d) of adult E. imbricata for the 341 day period prior to introduction into the breeding pen , and after different amounts of time in the breeding pen. Mean 1= mean of calculated growth rates; Mean 2 = mean of growth rates assuming negative growth rates were zero. * = turtles that died.

Table 8. Summary of main running costs associated with operation of the breeding pen, from 27 October 2003 to 30 April 2005.

Table 9. Species of freshwater turtles found in Asian food markets (Sources: NCMA 2002; Compton 2000; Lau and Haitao 2000; Chen et al. 2000; CITES 2002, 2003). F= food, TCM= traditional Chinese medicine, PT= pet trade, RR= religious release. *= US species.

Table 10. Freshwater and marine turtle species in Australia (sources: Cogger 1992; Greer 2003; Cann 1998). * = most common species. In brackets; PNG= Papua New Guinea, IND= Indonesia. Only distribution in Australia is shown for marine turtles.

viii List of Figures

Figure 1. Stylised aerial and lateral views of the breeding pen with associated pumps, filters, etc.

Figure 2. View of breeding pen showing shade cloth roof.

Figure 3. View of breeding pen showing netting partitions, nesting banks (right) and one of two the 35 kl storage tanks (background).

Figure 4. Sand bank showing tracks made by a female E. imbricata the previous night (19 June 2005).

Figure 5. Adult, captive-bred E. imbricata; female (left), male (right).

Figure 6. Left C1 dorsal shell plate. The left flipper appears at top left, and the head (not shown) is at top right.

Figure 7. Mean monthly water temperatures in the breeding pen, measured in the morning (closed circles) and afternoon (open circles) (see Table 1).

Figure 8. Water and air temperatures in the breeding pen recorded every hour with data loggers, in December 2003. High and low spikes with water temperature reflect when the pen was drained and refilled respectively.

Figure 9. Water and air temperatures in the breeding pen recorded every hour with data loggers, in April 2004. High and low spikes with water temperature reflect when the pen was drained and refilled respectively.

Figure 10. Enlarged ovarian follicles from ranched, 69.0 cm SCL female, E. imbricata (February 2004).

Figure 11. Hole (nest chamber) dug by a ranched, 76.2 cm SCL female E. imbricata. The ruler is 30 cm long.

ix Executive Summary

What the report is about and who the report is targeted at

This project assessed the suitability of a prototype breeding pen for Hawksbill sea turtles, and subsequent reproductive performance of captive-raised animals. The data collected will assist organizations interested in pursuing captive breeding of sea turtles, as has occurred with the crocodile industry. Information was also gathered on the potential for Australian freshwater turtle species for national and international markets. This information will assist people already working on captive breeding of turtles, particularly indigenous communities that are involved or interested in being involved in this wildlife industry.

Background

Wild sea turtles have a long history of being harvested for meat, eggs, leather and other products. But harvesting has often been excessive and unsustainable, causing declines in wild stocks, and generating conservation problems. Experimentation with sea turtle farming was initiated at the same time as crocodilian farming, but commercial incentives to develop these technologies to their full potential with sea turtles were constrained, largely by the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES). The harvest of wild eggs (ranching) is considered a safe form of wild harvest because it relies on, life stages that have little chance of survival to adulthood, and it also links gains by industry to the maintenance of the natural resource. Captive breeding is perceived as being the safest form of use because its impact on wild populations is negligible, and as such, once the techniques and protocols for captive breeding have been developed and proved, it is the easiest pathway through CITES for international trade. Unlike other sea turtle species, there is a high international demand for the valuable shell of Hawksbill turtles, and meat and other by-products could also be marketed domestically and internationally.

The high demand for freshwater turtles in Asia for food, particularly in China, has placed great pressure on wild populations. To meet the demand, captive breeding of turtles has expanded greatly over the last 10-15 years, in China and other Southeast Asian countries (eg Vietnam, Malaysia, Thailand, Cambodia, Taiwan). The reasons why certain species are sought over others are unclear, although soft-shelled turtle species dominate the food market. The distinction between food and medicine in Chinese culture is not clear, and certain foods, like freshwater turtles, shark fin and bird's nest, have medicinal properties attributed to them, and are considered "good for your health".

Aims/Objectives

The aim was to advance our understanding of the captive husbandry of Hawksbill turtles. The primary objective was to develop and test a prototype pen for captive breeding of farm-raised E. imbricata. A secondary objective was to advance our understanding of potential market opportunities for Australian turtle meat and other products (including freshwater turtle species), and to assess ways in which access to those markets may be able to be realised.

Methods used

A breeding pen incorporating water and sandy “beach” areas was built in Darwin, NT. Captive-raised Hawksbill turtles (hatched March 1996) were placed in different densities and sex ratios, over one breeding season. Water parameters, growth rates and reproductive behaviours were monitored over the period of study. Information on the demand for Australian turtles was sought through literature searches and communication with wildlife authorities, non-government organisations and private individuals (eg restaurants, Traditional Chinese Medicine practitioners), and attendance at meetings (eg CITES 2002, 2004) and sea turtle symposia (2003, 2004, 2005), and visits to countries in Asia (eg China, Singapore, Malaysia, Japan, Cambodia, Indonesia, Vietnam and Thailand). In particular, information on potential differences in quality between different species was sought.

x Results/Key findings

Hawksbill turtles are amenable to captive raising, display high growth rates and early attainment of maturity (males as early as 1.5 years, females <6 years). The study spanned one breeding season, and unforeseen problems occurred when turtles were placed together. Nonetheless, the development of ovarian follicles by females and mating behaviour indicated that captive breeding is possible, as demonstrated with Green turtles (Chelonia mydas) in the Caribbean. Suggested modifications to the prototype breeding pen include provision of areas of warm water during the cool season, and mechanisms (eg evaporative cooling) to cool water during the warmer times of the year.

Domestic trade is the most promising option for a turtle industry in Australia in the short-term, for food and perhaps other products (eg shells) for the growing Traditional Chinese Medicine industry. The Environment Protection Biodiversity Conservation 1999 Act is an impediment to international trade from Australia, as it prohibits live exports of native species for commercial purposes. A strong market appears to exist for Australian freshwater turtles as pets in Europe and the USA, but their suitability for the Asian market remains unknown. It is unlikely that processed (frozen) turtles from Australia could compete with live turtles that can be obtained readily from farms and other sources in the Asian region. However, with the demand on wild turtle populations in Asia and the increased regulation of trade through CITES, this situation may well change over time.

Implications for relevant stakeholders

Hawksbill turtles are highly amenable to ranching and subsequent raising in captivity, and early maturation in captivity supports their potential for captive breeding. An industry based on Hawksbill turtles could involve indigenous landowners, who have shown an interest in this type of wildlife industry over many years. In Australia, most nesting areas for E. imbricata lie on indigenous land. Captive breeding provides a shorter route for approval of international trade through CITES, and would be the most logical approach to take in the first instance, rather than ranching. A domestic market for freshwater turtles as food could be developed in Australia, and assessment of other derivatives such as the shell for Traditional Chinese Medicine could be undertaken at the same time. Ranching of freshwater turtles presents an opportunity for economic development of indigenous landowners, and has already been initiated by one community in Arnhem Land. The sustainable use of wildlife is now recognised as a legitimate tool that can contribute to conservation, thereby indirectly benefiting the wider community as a whole.

Recommendations

Successful nesting of Hawksbill turtles involves a complex interaction between a suite of social, behavioural, physiological and environmental factors prior to and during the breeding season. Further research over a longer period of time would help quantify optimum conditions for captive breeding of the species. A more detailed analysis of potential markets for freshwater turtles, with particular emphasis on species that that are amenable to ranching and/or captive breeding, would allow cost- benefits for domestic trade to be estimated more realistically.

1. Introduction and Background

Wild sea turtles and crocodilians have a long history of being harvested for meat, eggs, leather and other products. But harvesting was often excessive and unsustainable, causing declines in wild stocks, and thus generating series of conservation problems. The declining supply of wild crocodiles stimulated the historical efforts to farm crocodilians through closed or semi-closed farming systems, for commercial production and in some cases for conservation (restocking programs). The technologies developed have been remarkably successful, and the majority of skins in world trade now come from crocodile farms. Experimentation with sea turtle farming technologies was initiated at the same time, for the same reasons, in a similar limited number of countries. However, commercial incentives to develop these technologies to their full potential were constrained by comparison, largely by the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES).

In 1975, CITES was enacted, controls and restrictions on international trade in wild-caught sea turtles and crocodilians were implemented. The aim of CITES was to ensure international trade was legal, verifiable and sustainable, and that wild harvesting for international trade did not contribute to the further endangerment of wild populations.

The immediate effect of CITES was often an international ban on trade in wild-caught specimens, until status could be clarified and improved, and management protocols for sustainable use could be developed. For species listed on Appendix II of CITES, trade could occur if sustainable management programs were implemented that met CITES criteria requiring the ongoing demonstration that trade was not detrimental to the wild population. For species listed on Appendix I of CITES, where trade was considered a serious present threat to the survival of species, the only option for international trade was through captive breeding. Again, over time reasonably strict CITES criteria were derived for the commercial production through captive breeding of Appendix-I species.

With crocodilians, the disparity between supply and demand in the early 1970s had stimulated investment and research into captive breeding and ranching (collecting wild eggs or hatchlings and raising them in captivity for commercial sale), which continued after CITES was implemented. The Parties to CITES began to sanction trade in captive-bred Appendix-I crocodilians, by registering commercial captive breeding facilities. In the 1980's, they began transferring crocodilian species from Appendix I to Appendix II, so that controlled trade involving wild crocodilian species could occur. Confidence in the ability to trade, if CITES criteria were met, created confidence in the ability to gain a commercial return on investment in research and development, which proceeded with crocodilians around the world. Thus CITES was extremely effective in encouraging nations to improve conservation and management, and legal trade sanctioned by CITES largely replaced illegal trade (Hutton and Webb 2002; MacGregor 2002).

With sea turtles, by comparison, early efforts by the Parties to CITES to encourage sustainable use in some countries through ranching (eg Suriname) were not acted upon, criteria for captive breeding were introduced that required the demonstration of second generation breeding, which meant long delays between investment and return, and the development of criteria for ranching programs (CITES Resolution Conf. 9.20, in 1994), were protracted. The ability to sustain commercial efforts to derive technologies for the captive breeding and production of sea turtles, which was with crocodilians could provide an alternative to the unsustainable use of wild populations, was constrained. Nevertheless, considerable progress was made.

Largely to improve the survival prospects of wild eggs and hatchlings, a number of programs pursued the relocation and/or artificial incubation of wild eggs, and the raising of hatchlings for release at a larger size (head-starting), when survival prospects were considered greater. The investment was usually minor and the results mixed (Donnelly 1994). At least one major program with Kemp's Ridley (Lepidochelys kempi) in Texas, achieved considerable success with raising technology and the stock

1 they released have now returned and started nesting in areas where they were essentially extinct (Caillouett et al. 1995; Fontaine and Shaver 2005; Shaver et al. 2005).

In Australia, the potential for commercial captive production of sea turtles was recognised by the Federal Government in the early 1970s. It embarked on a turtle farming research and development program in northern Australia, concentrated in the Torres Strait. The program was designed to provide economic benefits for indigenous people, many of whom have limited access to employment in their homelands, and was based on ranching Green Turtles (Chelonia mydas). Wild eggs were collected and incubated and the hatchlings were raised on farms primarily for meat. After a series of initial problems, particularly with husbandry (Garnett 1983), the turtles were raised for release to the wild rather than for slaughter (Kowarsky 1977). Given the pioneering nature of this project, the biological problems that needed to be addressed, the logistics of working in very remote areas, and the ambitious social and economic goals, it is perhaps not surprising that this program failed to meet its expectations and was eventually abandoned. However, it did leave a lasting expectation in the Torres Strait in particular, that turtle farming was a real potential commercial activity for the future, and one that was culturally and logistically suitable.

A more significant pioneering turtle farming project with C. mydas, which proved remarkably successful despite trade restrictions, started in the Cayman Islands, in the Caribbean, in 1968. The Cayman Island Turtle Farm (CITF) has now been operating continuously for 37 years, and has overcome many of the technical problems that plagued the early Australian effort.

The CITF maintains a captive population of C. mydas adults, that produce fertile eggs annually, which are then incubated under controlled conditions on the farm. The hatchlings are raised in captivity for around 3 years before being killed and processed for meat (Fosdick and Fosdick 1994; Simon et al. 1975). Because of the time taken for CITES to agree on a pathway forward with sea turtles, and the requirement to establish that the technologies were able to proceed through to second generation breeding (which has now been done), neither the meat nor any of the by-products could be exported and sold. So the scope of the operation had to be dramatically reduced, and tourism introduced as an income stream. As the Cayman Islands have a long tradition of harvesting and eating wild sea turtles, and local demand for the meat is strong, all the turtle meat produced by CITF is sold and consumed domestically.

In 2002 the UK, on behalf of the Government of the Cayman Islands, proposed that CITES register The Cayman Island Turtle Farm as a commercial captive breeding operation for an Appendix-I species. The only product they wished to export was the carapaces of C. mydas, which when polished and lacquered were a popular curio with tourists (CITES 2002a). Allegations that some of the original founder stock, from 1960s had not been obtained with the correct legal permits clouded and confused the debate over the Cayman Island proposal, even though these concerns were later shown to be unfounded and spurious. So the approvals sought by the Cayman Island were not granted. .

The emphasis on C. mydas in farming research in both Australia and the Cayman Islands was explicable in terms of turtle meat, which was a product with a long commercial history. However, is difficult to evaluate the commercial viability of these operations as a whole, because the international trade in meat and other by-products (skin, oil, shell plates, whole carapaces) was constrained by CITES, and thus both products and their value could only be determined by local markets.

When considering the potential of producing Hawksbill turtles (Eretmochelys imbricata) through captive breeding and/or ranching, some notable differences with C. mydas need to be noted. . Both species produce meat and other by-products such as oil and fat, but the keratinous "shell" plates on the carapace of E. imbricata are significantly more valuable than those of C. mydas. Indeed, like the skin of crocodiles, they would almost certainly be the primary commercial product produced from E. imbricata farming.

2 The shell plates of E. imbricata (see Figs. 5 and 6) are imbricate (overlapping), thick and uniquely coloured with the original "tortoiseshell" pattern, made up of patches of black to red pigment on white to yellow substrates (ROC 1998). The plastic-like consistency of the plates and the ease with which they can be moulded, shaped and carved, meant that they have been used for thousands of years to make items of practical, decorative and traditional significance. They were essentially a natural renewable plastic, available well before plastics came into being.

Throughout most of the range of E. imbricata coastal peoples still fashion the shell plates into bangles, earrings and other items. Even in countries where the species is protected, the shell plates of animals caught incidentally is still often used rather than discarded. In Palau the shell plates are used as a form of traditional money (Ishmael and Antonio 2005). In Japan, E. imbricata shell has been used as a raw material for making objects (called "Bekko") of art and culture for some 500 years. The bekko tradition survives today using a stockpile of shell plates accumulated before trade ceased in 1993. During the 1970s, countries importing E. imbricata shell plates, and using them for various forms of traditional craft, included: Holland, United Kingdom, Hong Kong, France, United States of America, Italy, Japan, Germany, Belgium and Austria (Carrillo et al. 1998).

The majority of countries ceased importing E. imbricata shell when the species was listed on Appendix I of CITES in 1977. In France, the tradition was maintained using shell imported from French colonies. Japan, which developed what is arguably the highest level of perfection in its traditional craft with E. imbricata shell, lodged a reservation on E. imbricata when it joined CITES. This meant Japan could continue to import E. imbricata shell from non-Party States or from nations (eg Cuba) that also held a reservation on this species.

In 1992 Japan lifted its reservation and could no longer import E. imbricata shell from any country. They had to rely on stockpiles of shell held by various traders in Japan, which was then estimated to be about 100 tonnes in total. Since 1992, the Japanese bekko industry has been trying to encourage Parties to implement the types of sustainable use programs now common with crocodilians, so that a legal, verifiable and sustainable supply of shell could be produced in accordance with CITES requirements (eg JBA 1994). The demand is estimated as 5-10 tonnes per year at an import price of around $AUD1000/kg. A wild E. imbricata of 70-80 cm straight carapace length has around 1.5 kg of shell; more in captive-raised individuals.

Cuba maintains a traditional and demonstrably sustainable harvest of E. imbricata, and has twice gone before the Parties of CITES with proposals to trade (ROC 1998, 2000). However, CITES is also a political forum and opposition by the USA, in particular, which unilaterally imposed and maintains an economic trade ban on Cuba, has been a significant factor thwarting their efforts. There are no real biological or conservation problems, only political problems. If successful, Cuba could provide Japan with around 600 kg per year from their scaled-down traditional wild harvest (Cuban Ministry of Fishing Industries, unpubl. data).

The current Australian draft Action Plan for Marine Turtles (Australian Commonwealth Government) recognises the potential future role of sustainable, commercial use of marine turtles in Australia (EA 2002), and specifically with E. imbricata in the Northern Territory, where the species is abundant and quite possibly are carrying capacity. To satisfy CITES, there are really two "safe" options for production in Australia that are not mutually exclusive: closed-cycle captive breeding and/or ranching.

Ranching is considered the safest form of wild harvest because it relies on the harvest of wild eggs, which have little chance of survival to adulthood. It also links any gains from industry to the maintenance of the natural resource - particularly the adults that lay the eggs. There are currently CITES guidelines for ranching of marine turtles (Resolution Conf. 9.20), but no ranching programs have been submitted for approval.

3 The other option is captive breeding. It is often perceived as being the safest form of use because its impact on wild populations is negligible, and as such, once the techniques and protocols for captive breeding have been developed and proved, it is the easiest pathway through CITES. There are as yet no commercial captive breeding facilities for any sea turtle species that are registered with CITES, but only the Cayman Islands has tried.

In 1995 WMI researchers were contracted to assist Cuba develop its proposals to CITES (in 1997, 2000). It soon became apparent to WMI that relative to crocodilians, there was a paucity of information available on the captive husbandry of E. imbricata. Assumptions about potential constraints were untested and in most cases misguided and spurious: for example, long periods of time to reach maturity, high levels of embryo mortality when eggs are moved, poor hatching success, high mortality rates among hatchlings and high susceptibility to disease. In addition, some critical issues associated with farming reptiles, such as the effects of water temperature on growth, were being ignored. Similar potential and real problems were proposed as constraints on crocodilian farming in the 1970s, but they were overcome through targeted research and development during the 1980s and 1990s.

In 1996 in conjunction with Cuban scientists, WMI initiated research with E. imbricata on egg collection, incubation, captive husbandry of hatchlings, hatchling nutrition, effects of salinity and water quality and the effects of water temperature (Nodarse et al. 1998). This was followed by raising trials aimed at quantifying the size-age relationship in captivity over a longer time period. As in crocodilians, growth rates in captivity were appreciably higher than those reported in the wild. The first male E. imbricata with secondary sexual characteristics and active spermatogenesis in the gonads was 1.7 years of age (41 cm SCL), which can be compared with an estimated 65-68 cm SCL and 27- 30 years of age for wild males to reach maturity in the Indian-Pacific region (Limpus 1992).

Additional WMI research was oriented at the shell plates themselves, in both wild and captive-raised E. imbricata. Changes in shell plate size and shape with increasing animal size and age were quantified, and various investigations made into the structure of shell plates, the deposition of colour, the relationship between growth rate and colour deposition, and clutch-specific differences in colour. The value of E. imbricata shell plates is partly dependent on colour, with yellow (amber) being more valuable than red or red/black, which are more valuable than black. Genetic and environmental factors both influence shell colour, so through captive breeding and raising within controlled environments there is the potential to control some aspects of shell quality.

Commercial production through captive breeding or ranching can benefit conservation, just as it has done for crocodilians (Hutton and Webb 2002). At one extreme, if wild populations are depleted the potential exists to restock them, which has worked successfully with L. kempii in Texas and Mexico (Shaver 2005). Perhaps more relevant, is the linking of a commercial industry, through ranching, to the maintenance of wild populations: it directly links the generation of wealth to the pursuit of conservation. People are motivated to take action to conserve wildlife because they value the wildlife (Webb 2002). The more reasons people have to value wildlife, and the more people who value wildlife as a consequence, the greater will be the public and political will to ensure wild populations are conserved and used sustainably.

The current situation with E. imbricata in northern Australia is similar to that which existed for crocodiles in Australia in the early 1980s. At that time Saltwater crocodiles (Crocodylus porosus) were on Appendix I of CITES, which prevented trade (except through captive breeding). Yet the crocodile industry was expanding with a strong commitment to R&D aimed at overcoming various problems. The results of that R&D demonstrated to the Parties to CITES that crocodile farming was biologically and economically feasible and viable, which has now been established unequivocally.

In the case of Saltwater crocodiles, the first skin exports from the Northern Territory occurred in 1987, after approval by CITES in 1985. This was 5-7 years after the farms were initially established. The situation with Hawksbill turtles is likely to be the same, in the sense that serious commitment needs to

4 be made before approval by CITES can be sought. Crocodile skins are the main product of farming, which are perishable and need to be treated cautiously in storage and transport. By comparison, the shell plates of E. imbricata are dried and can be stored indefinitely. The meat of both is a valuable by- product.

In addition to sea turtles and crocodiles, there is increasing evidence that freshwater turtles have potential for commercial production. They have been consumed by local people throughout the world for centuries, and are part of the traditional Aboriginal diet in northern Australia today.

Over the last 15-20 years, improved economic conditions in many Asian countries, particularly China, has stimulated significant international trade in freshwater turtles. The main demand is for food (meat and eggs), which is also valued as a Traditional Chinese Medicine (TCM). A much lesser demand exists for the pet trade, and in some countries local species are purchased and released during religious ceremonies (Compton 2000): but these are minor relative to the demand for food.

Trade in freshwater turtles for food is generally non-species specific, and involves the harvesting and export of all species from South and Southeast Asia, from India to Papua New Guinea. However, within the retail market place in China market place some species of freshwater turtle are highly valued relative to others. At the national level in many countries, controls on exports and imports are now starting to be implemented. More and more freshwater turtle species are being listed on the Appendices of CITES in order to improve controls on international trade and prevent trade contributing to the endangerment of wild populations (eg CITES 2000, 2002a, 2003a, 2004; SMS 2004; Rhodin 2001).

Production of some freshwater turtle species through captive breeding is now commonplace in Asia, although relatively few ranching programs for freshwater turtles appear to have been implemented anywhere. The biological and economic potential of Australian freshwater turtle species in the international trade for food is largely unknown.

5 2. Objectives

The main focus of this study is to advance our understanding of on the captive husbandry of turtles. The primary objective was to develop and test unitised pens for the captive breeding of farm-raised E. imbricata. Successful captive breeding would allow production independent of wild populations, which would satisfy some conservation concerns. It would also potentially allow more controls to be exerted over the genetic aspects of shell quality, which has obvious commercial significance. Production through captive breeding, as in the Cayman Islands, is a highly conservative and very safe avenue through which the Parties to CITES could approve international trade. A second objective of this study was to advance our understanding of potential market opportunities for Australian turtle meat and other products, and to assess ways in which access to those markets may be able to be realised.

6 3. Methodology

3.1 Captive Breeding of Hawksbill Turtles

3.1.1 Breeding Pen

In designing a prototype breeding pen for captive adult E. imbricata, that could be operated at the WMI research facility which does not have access to salt water, particular attention was paid to the physical aspects that would maintain water quality (circulation, filtration), allow ease of cleaning (eg drainage, pen surface),) and allow the pen to be subdivided into enclosures of different sizes for research. Drawing on previous experience (since 1996) with rearing of E. imbricata in both controlled- environment raising tanks and larger outdoor ponds, and similar experience with the captive breeding of crocodilians, a pen design was refined (Fig. 1).

Figure 1. Stylised aerial and lateral views of breeding pen with associated pumps, filters, etc.

Figure 2. View of breeding pen showing shadecloth roof.

Construction (Figs. 2 and 3) began with the excavation of an area slightly greater than the intended dimensions of the pen, and installation of drainage points and pipework (Fig. 1). The sides and bottom were then lined with reinforced concrete (50-100 mm thick). After curing, the concrete was covered with fibreglass to provide a smooth surface. Pumps, filters, protein skimmers, storage tanks and other associated pipework were installed, and the system tested.

Dimensions of the pool are 30 m x 5 m x 0-0.8 m deepest point (Fig. 1). Water depth varies from 0.8 m along one side, to 0 m at the other side, where a gentle incline allows turtles to crawl out onto the land area. Total water volume was around 70 kl. The land area along one side of the pen consisted of a 1.7 m wide area filled with sand (70 cm deep) for nesting (Figs. 1 and 3). On the other side of the pen consists of a bitumen pathway (Fig. 2).

Within the first three days of operation (28-30 October 2003), temporary shade was established over the water surface of the pen (about 1 m from the water surface). This was removed 5 months later (March 2004) when a permanent shade structure (90% shade cloth) was erected over the entire pen (Fig. 2).

Initially, plastic mesh (3 mm diameter; 5 x 5 cm mesh) was used to subdivide the pen into separate enclosures, but this material was later replaced with more durable polyethylene prawn trawl netting (3 mm, 5 x 5 cm mesh; 1.5 mm, 3 x 3 cm mesh; 3 mm, 2.5 x 2.5 cm mesh). The netting sits 20-30 cm above the waterline, which proved sufficient to stop turtles climbing into adjacent enclosures.

Figure 3. View of breeding pen showing netting partitions, nesting banks (right) and one of two the 35 kl storage tanks (background).

3.1.2 Water Circulation

A 1.1 kW water pump at one end of the pen (Fig. 1) took water from a sump connected to the drainage line under the pen. The water was pumped through a conventional pool sand filter and back into the pen at the opposite end. A second 1.1 kW water pump took water from the same sump and pumped it independently through two protein skimmers, from which it flowed back into opposite end of the pen (Fig. 1). There were thus two pumps taking water from one end of the pool, one going through the protein skimmer and another through a sand filter, with both draining into the opposite end of the pool.

From October 2003 to June 2004, water was circulated through the pen for an average of 20 h each day. The pumps were stopped for feeding and cleaning (see below). From July 2004 to May 2005 pumps were only run during the day, for an average of 8.6 h per day.

Two interconnected 35 kl plastic water tanks (Fig. 1), connected to the filter pump, were used as temporary storage tanks for the saline water.

3.1.3 Water Parameters

For most of the period for which the pen was operated (October 2003 to May 2005), it was filled with saline water of 20-25 ppt, made up with fresh water and pool salt. Fresh water or salt were added periodically to compensate for evaporation and minor water losses incurred during cleaning. In previous raising trials 20-25 ppt salinity proved adequate for large, captive-raised E. imbricata.

On 15 October 2004, the pen was filled with sea water, collected from Darwin Harbour by road tanker. Evaporation and cleaning losses were replaced with freshwater, until salinity reached around 20 ppt, at

9 which time 20-35 kl of "new" sea water was used to top up the pen and raise the salinity. The use of sea water was abandoned after 5 months (in March 2005), and the previous protocol (see above) adopted again.

All water in the pen was discarded when it was considered unsuitable for further use. This occurred three times: December 2003, March 2004 and October 2004 (when sea water was used).

Chlorine was added either on feed days or if the water appeared slightly discoloured, to attain levels of <1 ppm. Liquid chlorine (13% hypochlorite) was poured directly into the sump (see Fig. 1) or directly into the pen, and 5-10 chlorine tablets (200g; trichloroisocyanuric acid) were placed in a dispenser within the sump (Fig. 1). Attempts were made to maintain cyanuric acid levels between 50 and 70 ppm by the addition of dissolved cyanuric acid into the sump and through a dispenser (0.5 kg capacity) within the sump, as required. The chlorine tablets also contributed to cyanuric acid levels (see above). Sodium bicarbonate was used as required to increase pH to around 7.2-7.5.

Chlorine, cyanuric acid and ammonia levels were monitored using a "Palintest Photometer 5000," and pH using a sodium thiosulphate-phenol red test kit. Salinity was measured with a hydrometer initially (to 7 May 2004), but later with an Aqua-TDS/pH meter (TPS Pty. Ltd.).

3.1.4 Cleaning

During feeding, the saline water was pumped into the storage tanks, and shallow freshwater introduced to the pens that could later be taken to waste, which in this case was through a recycling system. Leftover food and faeces were easily hosed down the drains after feeding. Until November 2004, light algal growth was scrubbed off the surface of the pen using a hard bristled broom. In December 2004, when sea water was being used, an algae that stained the pen surface black appeared, and different methods were tried to retard its rapid growth. Initially, hydrated lime was added to fresh water overnight to increase pH to around 10.5, and after the water was drained high water pressure applied to the affected surfaces. Copper sulphate was then added to achieve a concentration of 4.3 ppm. This was repeated after 3 weeks, and again later as required to prevent algal growth.

3.1.5 Temperature

Water temperature was monitored regularly, with morning and afternoon readings taken with a calibrated thermometer. Between December 2003 and May 2004, calibrated Tidbit data loggers ("Onset") were placed in the water, approximately 10 cm beneath the water surface, and enclosed in a short piece of PVC to prevent turtles biting it. Air temperature was measured approximately 10 cm above the water surface, also within a piece of PVC. Readings were taken every hour.

In April 2004, shade cloth covers were attached at the ends of the pen to reduce direct sunlight hitting the water. At the same time, some of the water being circulated back into the pen after filtration (see above) was diverted through sprinklers placed along the length of the pen to create a fine spray over the water surface. During the day, three upright oscillating fans were directed at the spray to facilitate evaporative cooling. By early June 2004 this was not considered necessary, and it was stopped. It was started in early December 2004, but with only one fan this time, until mid-March 2005, when ambient temperatures were lower.

Two methods of heating the pen water during the coolest time of year (July) were tried. The first consisted of hot water being pumped to a coil of PVC pipe anchored at the bottom of one of the enclosures (20 June 2004). This was continued for 4 days and was not effective in creating a "warm water" enclave within the pen.

More successful was a fibreglass chamber (185 x 80 x 77 cm), open at one end (80 x 77 cm), with a free-moving clear plastic flap allowing entrance or exit to the chamber. A single 2400 W coil heater

10 was fixed within the chamber, at the bottom, at the end furthest from the door flap. The top of the chamber was just out of the water. To reduce heat loss to the air, a clear plastic "ceiling" was placed on the chamber (50-150 mm above the water surface) and a second plastic flap placed 50 cm from the outer flap. Heated water was thus trapped between three walls, a ceiling and two "doors", and once a turtle swam into the chamber it was itself located in a warm water bath.

Within each chamber, the coil heaters were left on continuously for two months. On 20 August 2004 two coil heaters were placed in one chamber to provide additional heat to a sick turtle, until 13 September 2004 when both chambers were removed completely.

Water temperature inside the chambers and in the surrounding water was monitored.

3.1.6 Nesting Banks

The sand area available to the turtles for nesting was inspected daily, and if necessary was raked smooth so that any tracks could be easily seen the following day (Fig. 4). During the wet season, when the sand was sometimes compacted by rain, it was loosened with a pitchfork and then raked flat using a sand rake.

3.1.7 Running Costs

Running costs associated with operation of the breeding pen were either recorded precisely (eg labour, consumables) or estimated for equipment using power ratings and time of operation (eg electricity).

Figure 4. Sand bank showing tracks made by a female E. imbricata the previous night (19 June 2005).

11 3.1.8 Turtles

Nineteen (8 males, 11 females) of the 20 E. imbricata used in this study were hatched from ranched eggs in March 1996, and were thus 7.6 years of age in October 2003 (Fig. 5). One female E. imbricata was a wild-caught individual that was brought into captivity on 10 September 1997, measuring 31.6 cm straight carapace length (SCL). It is likely that this animal was at least 10 years of age by October 2003.

For the first 18 months of life, all ranched turtles were housed in an indoor complex of environmentally controlled tanks (29oC water temperature). On 14 November 1997, an outdoor fibreglass pond was commissioned, and some turtles were moved into it. Over the next 6 years turtles spent varying amounts of time in the indoor and/or outdoor facilities. The outdoor pond measured 9 x 7 x 0.3-0.9 m depth, and there was no control over water temperature.

Turtles were involved in various experiments on food preferences, density, etc., from the time of hatching.

3.1.9 Inventories

All turtles were inventoried on 21 October 2003, one week before being placed in the breeding pen. The standard measurements taken were straight carapace length (SCL) to the tip and the notch; straight carapace width (SCW); curved carapace length (CCL) to the tip and the notch; curved carapace width (CCW); and, head length (HL). They were individually weighed and photographed. Using the C1 dorsal shell plate (Fig. 6) as the index of colouration, the shell of each turtle was assessed on the basis of: extent of yellow colouration; colour of the dark colouration (eg red to black); and, extent of secondary yellow colouration underlying the top of the plate. Each of these characteristics was ranked from 1 to 10 (1= little yellow; 10= very yellow), and an aggregate of these three scores used as the Shell Grade.

Figure 5. Adult, captive-bred E. imbricata; female (left), male (right).

Figure 6. Left C1 dorsal shell plate. The left flipper appears at top left, and the head (not shown) is at top right.

On the basis of sex, size (SCL) and shell grade, turtles were assigned to different enclosures within the breeding pen. With mortalities over the first 6-7 months of operation, these initial combinations became somewhat redundant. The remaining male (No. 1295113) was moved between the enclosures of the remaining females every 2-3 weeks. One female was so aggressive towards the male that he could not be placed with her.

3.1.10 Foods and Feeding

The feeding regime was varied throughout the period of study, sometimes because of the availability of different foods, but mostly to refine the procedures. All turtles had been fed a diet consisting mainly of marine fish (pilchards, Tommy Ruff) and barramundi pellets (Ridleys Aquafeed) from the time they were about 6 months of age. The previous regime of feeding in the morning, three times per week, was maintained for the first two weeks of operation in the breeding pen. This was changed to feeding two times per week, with shark meat and pellets offered on alternate feeds for three weeks (to 2 December 2003), and then shark/pellets and fish/pellets on alternate feeds for two weeks (to 15 December 2003). The fish consisted of Tommy Ruff, except for the wild-caught turtle that would only accept pilchards.

On 16 December 2003 the feeding regime was altered to determine if, like crocodiles, they would feed readily on chicken heads (a source of calcium), with squid and pellets once per week, and fish/pellets once per week. This regime was continued until 16 March 2004, when only fish and pellets were offered. From 28 December 2004 onwards pilchards replaced Tommy Ruff as the fish component of the diet, and from 17 February 2005 squid was used in lieu of fish for one feed in four. Between 18 October 2004 and 21 December 2004, feeding was carried out in the afternoon (see below).

Squid was used to add variety in the diet, but also because it was an effective way of delivering vitamin supplements. Beginning in December 2003, each turtle was fed a “Pluravit” (Bayer Australia Pty. Ltd.) multivitamin capsule every two weeks, except during periods of cool weather when food intake decreased. Capsules were enclosed in the food, with the best results being achieved with chicken heads and squid.

Until 29 June 2004, the procedure associated with feeding was as follows. Prior to feeding the salt water was pumped into two 35 kl PVC storage tanks. The pen was filled with about 20 kl of fresh water (depth= 20-40 cm), and the turtles were allowed to feed for at least 1 h, in the morning. After feeding, all water was drained to waste, and leftover food and faeces hosed into the 10 drainage points on the bottom of the pen (Fig. 1). After cleaning, the salt water held in the storage tanks was released back into the pen and recirculation started again.

Due to the reduced number of turtles, from 30 June to 12 October 2004 it was decided to feed the turtles in the salt water. Food was offered in the afternoon (twice per week), and left in for about 20 minutes, after which time any leftover food was scooped out. It was still necessary to pump all water into the storage tanks and hose the pen clean once per week.

In an attempt to reduce the contamination created in the water during feeding, from 13 October to 21 December 2004 the amount of food offered per feed was reduced, but it was offered 5 times per week. Instead of being offered fish and pellets together, turtles were offered one or the other on a feed day.

From 22 December 2004 the feeding/cleaning procedures reverted back to feeding in fresh water, twice per week.

3.1.11 Illness and Mortalities

Turtles showing any visible external signs that they were ill were kept under close observation with veterinary assistance. Some became lethargic to the point that they could not lift their heads, and they were removed from the breeding pen and placed in plastic isolation tanks. Turtles were treated with antibiotics (eg Tetracycline, Baytril, Trivetrin) and multivitamin injections as prescribed by a qualified veterinarian. Most turtles that died were autopsied by a NT Department of Primary Industry, Fisheries and Mines veterinarian.

3.1.12 Ultrasonography

On 27 June 2005, female E. imbricata were subjected to an ultrasound scan, with standard equipment used for humans. The ultrasonic probe was placed on the inguinal area between the plastron and the rear legs, as described by Owen (1999), to report on follicle size in the ovaries.

3.2 Marketing of Turtles

Information on the demand and type of processing required by the marketplace for Australian turtles was sought through literature searches and communication with wildlife authorities, non-government organisations (eg TRAFFIC) and private individuals. In particular, information on potential differences in quality between different species was sought.

WMI staff participated in the 12th and 13th Conferences of the Parties to CITES (November 2002 and October 2004, respectively) and the 22nd, 23rd and 24th Annual Sea Turtle Symposia (March 2003, February 2004 and January 2005 respectively), where they were able to communicate with different people from around the world. Additional information was obtained during visits by WMI staff to China, Singapore, Malaysia, Japan, Cambodia, Indonesia, Vietnam and Thailand.

Chinese restaurants in Sydney, Melbourne and Darwin were also contacted, as were practitioners dealing with TCM in Australia.

14 4. RESULTS

4.1 Captive Breeding of Hawksbill Turtles

4.1.1 Breeding Pen

The design of the breeding pen allowed for efficient feeding and cleaning. The most time-consuming aspect was pumping salt water into the two storage tanks, but this was addressed through the installation of timers (pumping started well before staff arrived) and a mercury switch (to avoid pump accidentally running dry and burning the motor out).

The total volume of water held in the pen (70 kl) did not allow for cost-effective regulation of water temperature (heating or cooling), although modifications tested in the cool and hot times of the year appeared to be effective to some degree (see below).

The slope out of the water up to the land allowed turtles to crawl out onto the land. In addition, the shallower water on the slope was sometimes used by turtles, particular any that were sick, to warm up.

4.1.2 Temperature

Water temperatures varied throughout the year, and were relatively high at the beginning of the study, before permanent shade was constructed over the pen (Table 1; Fig. 7). Water temperature increased during the day, with greater increases during summer than in winter (Table 1), due to higher ambient temperatures.

Temperature (C) 26

22 0 ODNJFMAJJ2 4 6 MONDFMAM8 10 A 12S 14 16 J18 20 22 2005 2003 2004 Figure 7. Mean monthly water temperatures in the breeding pen, measured in the morning (closed circles) and afternoon (open circles) (see Table 1).

15 Table 1. Mean monthly water temperatures measured in the morning (am = 0800- 0900 h) and afternoon (pm= 1600-1700 h) in the breeding pen, and the difference between the two mean readings.

Month/Year am pm Difference (oC) (oC) (oC)

October 2003 31.5 34.4 2.9 November 2003 29.7 31.8 2.1 December 2003 28.5 30.8 2.3 January 2004 28.9 30.7 1.8 February 2004 28.2 29.8 1.6 March 2004 28.5 30.2 1.7 April 2004 26.9 28.3 1.4 May 2004 25.5 26.5 1.0 June 2004 22.8 23.2 0.4 July 2004 22.7 23.5 0.8 August 2004 24.2 25.2 1.0 September 2004 24.9 26.1 1.2 October 2004 28.0 29.2 1.2 November 2004 28.8 29.9 1.1 December 2004 29.0 29.9 0.9 January 2005 28.3 29.4 1.1 February 2005 28.5 30.0 1.5 March 2005 28.4 29.4 1.0 April 2005 27.4 28.8 1.4 May 2005 26.3 27.4 1.1

Fluctuations of water and air temperature in December 2003 (hot season) and April 2004 (cool season) are in Figures 8 and 9. These trends reflect readings taken in the morning and afternoon (Table 1):the water cools at night and heats during the day. Discrete decreases in temperature over a few days, as indicated in December 2003 (Fig. 8), reflect extended periods of rainfall.

40 300 - Air 38 735 - Water 36

28 Temperature (C) 26

20 1st 3rd 4th 6th 7th 8th 2nd 21st 10th 11th 12th 13th 15th 16th 17th 19th 20th 24th 25th 26th 28th 29th 30th 22nd December 2003 Figure 8. Water and air temperatures in the breeding pen recorded every hour with data loggers, in December 2003. High and low spikes with water temperature reflect when the pen was drained and refilled respectively.

40 300 - Air 38 735 - Water 36

28 Temperature (C) 26

20 1st 3rd 4th 6th 7th 8th 9th 2nd 21st 23rd 11th 12th 13th 14th 16th 17th 18th 19th 24th 26th 27th 28th 29th 22nd April 2004 Figure 9. Water and air temperatures in the breeding pen recorded every hour with data loggers, in April 2004. High and low spikes with water temperature reflect when the pen was drained and refilled respectively.

17 The sprinkler system installed along the length of the pen in April 2004, working in conjunction with oscillating fans, allowed evaporative cooling to reduce water temperature by 1-2oC.

The provision of heated chambers was more successful than simply placing a source of heat in the water. Heat from the coil simply dissipated too quickly into the surrounding water to provide an effective area of warm water, and it was disbanded after 4 days. On the other hand, the enclosed heating chambers allowed a reservoir of warm water to be maintained in the ponds. Within three days, the turtles would swim through the door flap and sit within the chambers for most of the day, only leaving the chamber to feed.

For the chamber without a top, the mean difference between outside water temperature (mean= 23.0oC) and that in the heating chamber was 2.5oC (range 0.6 to 4.7). The addition of a top and a second flap increased the gradient to a mean of 4.2oC (range 1.2 to 8.9).

4.1.3 Water Parameters

Mean salinity over the entire study period was 25.8 ppt (SD= 6.1; N= 192). During the 5 months when sea water was used (mean= 26.0 ppt; SD= 5.4; N= 33) it was not significantly different to when reconstituted water was used (mean= 25.8 ppt; SD= 6.2; N= 159). Bringing sea water by road to fill the pen and compensate for evaporation and other losses proved impractical relative to adding salt to freshwater.

Ammonia concentrations varied between 0 and 5.0 ppm (mean= 0.91; SD= 1.17; N= 68). Cyanuric acid levels were difficult to maintain over the first 6 months (mean= 11.6 ppm; SD= 12.56; N= 16), but this changed once the numbers of turtles had been reduced over the last 12 months (mean= 48.1 ppm; SD= 18.81; N= 47). Mean chlorine concentration was 0.35 ppm (SD= 0.67; N= 65), and mean pH was 7.35 (SD= 0.22; range 6.8-8.2).

4.1.4 Turtles

Female E. imbricata introduced into the breeding pen were on average longer (65.8 to 81.1 cm SCL) and heavier (41.1-76.2 kg) than the males of the same age (64.1 to 72.2 cm SCL; 30.1-47.5 kg) (Table 2).

Shell grades varied between individual turtles (Table 3). One animal (No. 5867632) had little primary yellow colouration or secondary colouration, and was thus graded very low (4) - the least desirable. On the other hand, two individuals (No. 5606378, 1822625) scored high on primary yellow colouration and secondary colouration (Table 3), and were considered to have the highest quality shell (25, 27).

Due to the low growth rates and problems with mortality, there was little shell growth with most individuals. Where it was evident at the anterior margins of the dorsal shell plates, colouration was dark.

18 Table 2. Morphometric measurements of 19 ranched and one wild-caught E. imbricata prior to introduction into the breeding pen. SCL= straight carapace length; SCW= straight carapace width; CCL= curved carapace length; CCW= curved carapace width; HL= head length; BWt= bodyweight. *= wild-caught individual.

Turtle No. Sex SCL SCW CCL CCW HL BWt (mm) (mm) (mm) (mm) (mm) (kg)

1567258 Female 658.1 515.5 695.0 622.0 152.9 41.1 1371552 Female 659.1 483.0 691.0 594.0 148.4 36.3 1891103 Female 677.3 513.1 707.0 582.0 148.0 34.6 2036114 * Female 680.5 524.9 753.0 600.0 157.7 44.0 1614310 Female 722.0 549.3 761.0 642.0 155.8 50.2 1369539 Female 723.2 590.3 766.0 670.0 162.3 53.5 1563770 Female 736.7 529.1 775.0 633.0 160.8 42.2 1592063 Female 737.1 573.9 771.0 650.0 156.0 53.4 1605864 Female 737.5 543.5 769.0 659.0 166.2 49.3 1329823 Female 750.6 551.9 810.0 639.0 159.6 53.0 6051363 Female 767.8 608.8 816.0 684.0 163.7 58.3 2101860 Female 811.2 637.6 870.0 711.0 167.6 76.2 Mean 721.8 49.3

2364815 Male 641.0 475.7 658.0 525.0 153.8 30.1 1887127 Male 649.5 540.0 663.0 574.0 152.7 36.3 5606378 Male 685.0 534.7 703.0 575.0 156.5 47.3 5867632 Male 692.3 523.7 725.0 614.0 156.3 41.6 1295113 Male 712.9 565.3 758.0 612.0 155.9 44.6 1822625 Male 713.4 534.7 741.0 601.0 150.4 42.5 9999999 Male 720.6 555.1 747.0 614.0 163.0 43.2 1561610 Male 722.4 560.6 755.0 635.0 157.0 47.5 Mean 692.1 41.6

19 4.1.5 Sex Ratios

The effectiveness of different combinations of males and females assigned on the basis of sex, size and shell colouration (Table 4) could not be determined, because there were few males and sample sizes were reduced by mortality (see below).

However, during the first two months all but one of the mature males, regardless of the number of females with them, became very active and aggressive towards the females that appeared to be courting behaviour. The males attempted to mate with the females as soon as they were placed together in the pen. These behaviours decreased in frequency after the first 1-2 weeks, but never stopped.

Table 3. Shell grading of 20 adult E. imbricata on 21 October 2003. Turtles are listed from the "least desirable" colouration to the "most desirable".

No. Red:Black Yellow:Dark Secondary Shell in dark Colour on Shell Plate Colouration Grade

5867632 2 0 2 4 1614310 5 2 5 12 1295113 7 6 3 16 1605864 7 2 7 16 9999999 3 8 6 17 1592063 5 7 5 17 2364815 9 7 3 19 2036114 6 8 5 19 1371552 7 5 7 19 2101860 9 2 9 20 6051363 4 10 6 20 1567258 7 3 10 20 1369539 4 10 7 21 1329823 6 7 8 21 1561610 7 10 4 21 1887127 9 4 9 22 1891103 6 8 10 24 1563770 10 10 4 24 5606378 10 5 10 25 1822625 10 9 8 27

20 Table 4. Initial housing of adult E. imbricata in the breeding pen on 27 October 2003. SCLi= initial straight carapace length.

Turtle No. Sex Pen M:F Shell SCLi No. Grade (mm)

2101860 Female 2 1:1 20 811.2 1887127 Male 22 649.5

5606378 Male 3 1:1 25 685.0 1891103 Female 24 677.3

9999999 Male 4 3:5 17 720.6 1369539 Female 21 723.2 1592063 Female 17 737.1 1605864 Female 16 737.5 1614310 Female 12 722.0 5867632 Male 4 692.3 2364815 Male 19 641.0 6051363 Female 20 767.8

1329823 Female 5 1:2 21 750.6 1561610 Male 21 722.4 2036114 Female 19 680.5

1567258 Female 6 1:2 20 658.1 1371552 Female 19 659.1 1295113 Male 16 712.9

1822625 Male 7 1:1 27 713.4 1563770 Female 24 736.7

4.1.6 Mortalities

The planned experiments were seriously compromised by a sharp increase in mortality rate. One female died when it became trapped in a hole in the sand area. However the exact causes of mortality in the others are unclear.

Turtles started showing external, visible signs of illness after about 1.5 months in the pen. With the benefit of hindsight, the most likely cause of health being severely compromised in the first instance was the extreme activity undertaken by the males, which caused the females to be very active as well. We now believe this increased activity caused metabolic heat production, that resulted in body temperatures exceeding limits that caused problems, that later led to other complications. Body temperature was not measured at the time. In some of the later mortalities, we consider it possible that the experimental feeding with chicken may have introduced Salmonella to which most turtles and crocodiles are resistant, but perhaps sea turtles are not. With shade and cooling improved, and chicken excluded from the diet, survival and health of the survivors has been excellent. But the original planned experiment was severely compromised by the early deaths.

Three males exhibited signs of illness after 41-47 days (December-January), two after 95-99 days (January-February) and two after 148-152 days (March). Amongst the females, two showed signs of illness 89-91 days (January), 3 after 125-134 days (March) and three after 167 (April), 293 (August) and 539 days (April) respectively.

All animals that died showed signs of lethargy and began floating high in the water, often leaning to one side. Despite intense treatment from veterinarians, the majority of turtles showing these signes

21 ultimately died and were autopsied. Twelve of 15 had some degree of skeletal muscle degeneration histologically, which could have been related to excessive body temperatures when first introduced into the breeding pens. Lesions varied from acute (degeneration with no inflammatory or fibrotic response) to subacute (some changes but with macrophages coming in to clean up degenerate fiber) to chronic (all of these changes plus fibrosis). Most dead turtles also exhibited signs of renal complications to some degree. Isolation of a number of environmental organisms (eg Edwardsiella tarda, Salmonella spp., Fusarium spp.), suggested some type of immunosuppression may have been in place.

4.1.7 Maturity

Seven of the 8 males in the study show clear signs of sexual dimorphism, with an elongated tail. This characteristic was clearly evident by the time these individuals were 3 years of age. Previous histological examination of the testes of a two-year-old sibling turtle indicated active spermatogensis, so there was no reason to suspect that they were not mature.

Examination of 7 autopsied males indicated that 6 of them had enlarged mature testes (mean= 140 x 29 x 9 mm), and the one (No. 2364815) without an elongated tail had small testes (60 x 15 x 3 mm) and was still immature. One male also showed signs of sperm in the cloaca. One 15-month-old E. imbricata (38.3 cm SCL) currently held by WMI has a greatly enlarged tail extending well past the margin of the carapace, and is possibly already mature.

The reproductive status of the captive-raised females was not well known when this study began. Examination of younger E. imbricata in the past, had not indicated any ovarian follicular development, but the turtles had never been housed in conditions that were conducive to extensive interaction between the sexes.

The autopsy of 8 females during the course of this study revealed that 7 (78%) of them were clearly mature. Six of these 7 females had enlarged, developing ovarian follicles (Fig. 10), which could be assigned to three broad size classes - large, medium and small (Table 5). Mean diameter of the largest follicles ranged from 17.9 to 22.1 mm, from 12.5 to 15.5 mm for medium-sized follicles, and 4.5 to 9.3 mm for the smallest. The seventh mature female was just showing the beginnings of follicular development, with the largest follicles at around 5 mm diameter - it is likely that this female was just beginning her first reproductive cycle.

Figure 10. Enlarged ovarian follicles from a ranched, 69.0 cm SCL, female E. imbricata (February 2004).

For the 6 mature females with large follicles, the numbers of large follicles present in each ovary were similar. One female had a low number of large follicles (N= 43) relative to the other females (range 144 to 270). Five of the 6 females also had similar numbers of medium-sized follicles in each ovary, and numbers of small follicles in each ovary varied greatly for all females (Table 5). Two females (22%) examined appeared to be immature, having small, undeveloped ovaries, and no sign of previous follicular development.

Ultrasonography of three live females in June 2005 indicated that two had ovarian follicles in at least two size classes (estimated diameters of 16.5 and 23.6 mm; and 18.3 and 21.4 mm, respectively), and the third showed no follicular development at all (see Table 6). There was no clear trend between maturity and size, with one of the smallest females (68.2 cm SCL) and one of the largest females (73.8 cm SCL) being immature (Table 6).

23 Table 5. Numbers and mean diameter (md; in mm) of three size classes of ovarian follicles recorded in 6 female E. imbricata. (L)= left ovary; (R)= right ovary.

Turtle No. 1369539 (L) 1369539 (R) 1329823 (L) 1329823 (R)

Large (md) 126 (21.9) 144 (22.1) 107 (19.9) 108 (19.9) Medium (md) 68 (15.5) 67 (15.0) 138 (14.2) 148 (14.4) Small (md) 473 (6.0) 156 (7.6) 286 (6.5) 381 (7.7) Total 667 367 531 637

Turtle No. 1371552 (L) 1371552 (R) 2036114 (L) 2036114 (R)

Large (md) 75 (18.1) 69 (17.9) 140 (21.2) 151 (20.1) Medium (md) 27 (13.7) 60 (13.1) 120 (14.7) 106 (13.2) Small (md) 71 (8.9) 225 (6.9) 300 (9.3) 240 (7.4) Total 173 354 560 497

Turtle No. 6051363 (L) 6051363 (R) 2101860 (L) 2101860 (R)

Large (md) 26 (19.3) 17 (19.0) 113 (19.5) 112 (20.4) Medium (md) 32 (14.5) 36 (13.4) 39 (13.5) 34 (12.5) Small (md) 87 (6.8) 38 (7.5) 322 (4.5) 416 (4.8) Total 145 91 474 562

Reproductive Straight Carapace Status Length (cm)

Mature 66.2 Mature 66.5 Immature 68.2 Mature 69.0 Mature 72.4 Mature * 73.0 Immature? * 75.2 Mature * 76.2 Immature 73.8 Mature 75.0 Mature 78.9 Mature 81.1

Mean - Mature 73.1 Mean - Immature 71.7

24 4.1.8 Mating Behaviour

As soon as turtles were introduced to each other in the breeding pens on 27 October 2003, 6 of the 8 males began aggressively chasing the female/s, biting them on the flippers and neck, and attempting to mate with them. This vigorous activity continued for about 2 hours. Over the next week, the aggression and attempts to mate by 5 of the males were less frequent, but occurred on a daily basis. They were observed much less frequently after this time, about 1-2 times every 1-2 weeks. One male, in a 1:2 enclosure, continued to actively chase the females for about two weeks before settling down.

The remaining male attempts to mate with females with which he is placed, with the latest attempts being made at the time of writing of this report. The immature male (as revealed by autopsy) and one of the mature males were never observed attempting to mate. They seemed to ignore the other turtles with which they were housed.

Males attempting to mate were able to clasp the females and curl their tail under the females, thus bringing their cloacas together. It was not possible to ascertain whether intromission actually occurred.

The plastrons of 6 mature males were softened and took on the concave appearance typical of wild male E. imbricata during mating activities (Wibbels et al. 1991; van Dam et al. 2004); one of these males was never seen mating with the females. Of the two males where the plastron remained hard, one was immature, and one was a mature animal that had been mating. Repeated attempts to mount the females also resulted in the keratin layer of the males' plastrons being eroded to some degree.

4.1.9 Nesting

No egg-laying occurred during the single "breeding season" spanning the period of study, although the reproductive status of the majority of females suggests that this may have occurred had the females survived.

It was difficult to establish which females from communal pens came up onto the nesting banks, and crawled around on the sand. Assuming that the males did not leave the water, at least 7 of the 12 females crawled up onto the sand banks at some time. This activity always took place at night, with the exception of one female that was observed still on the land in the morning on one occasion (see below).

Two females dug distinct holes during the study. These were either: wide and shallow (30 cm diameter x 10-20 cm deep); wide and deep (30 cm diameter x 30-40 cm deep); or, distinct chambers (15-30 cm diameter, 13-30 cm deep) typical of a turtle nest chamber (Fig. 11).

Distinct holes were recorded at similar times in 2004 and 2005. Two females dug holes (one each) in July 2004, and in June 2005, by the time of writing of this report one of these females had dug 9 distinct chambers, one of which was partially re-filled with sand. This particular female had been sighted digging a hole during the day in 2004, at which time she had not appeared to be disturbed by activities of staff around her. With the exception of this case, all other holes were dug at night.

Figure 11. Hole (nest chamber) dug by a ranched, 76.2 cm SCL female E. imbricata. The ruler is 30 cm long.

4.1.10 Growth Rates

When released into the breeding pen, turtles grew slower than they had over the previous 11 months (Table 7). Growth rates of animals that died may have been affected by the illness, but even turtles that did not appear to be affected by illness did not grow as fast as they had previously [mean GR= 0.077 mm/d (0.038-0.100) versus mean GR= 0.030 mm/d (0.011-0.054) in the breeding pen]. The negative growth rates calculated for some turtles (Table 7) suggest changes in body shape of, perhaps the result of loss of bodyweight. In the case of the males the softening and concavity of the plastron may have contributed to "shrinkage" between inventories.

All turtles lost weight during the study period, and even the surviving turtles now weigh less than they did before introduction to the breeding pen: they lost a mean of 10.9% (range 8.3-12.7%) of their original bodyweight after 447 days, although prior to being introduced into the breeding pen they were exceedingly fat.

There was no significant relationship between growth rate and initial size (r2= 0.0002; p= 0.948; N= 20), and this did not change when negative growth rates were assumed to be zero (r2= 0.001; p= 0.883; N= 20). The inclusion of the number of days spent in the breeding pen in the analyses did not result in initial size or time in the pen being significant.

Turtle No. Sex Enclosure Previous GR Days Initial SCL Final GR (mm SCL/d) in Pen (mm) (mm SCL/d)

1567258* Female Indoor 0.1364 550 658.1 0.0125 1371552* Female Indoor 0.0713 151 659.1 0.0185 1891103* Female Indoor 0.0806 129 677.3 0.0380 2036114* Female Outdoor -0.0244 123 680.5 0.0748 1614310 Female Indoor 0.0978 447 722.0 0.0188 1369539* Female Indoor 0.0133 171 723.2 0.0018 1563770 Female Outdoor 0.0714 447 736.7 0.0340 1592063* Female Indoor 0.0136 302 737.1 0.0030 1605864 Female Indoor 0.0998 447 737.5 0.0544 1329823* Female Indoor 0.0104 186 750.6 -0.0022 6051363* Female Indoor 0.0946 229 767.8 0.0939 2101860* Female Indoor 0.0891 224 811.2 -0.0013 Mean 1 0.0628 0.0288 Mean 2 0.0649 0.0291

2364815* Male Outdoor 0.0490 55 641.0 0.0527 1887127* Male Outdoor 0.0340 71 649.5 0.0056 5606378* Male Indoor 0.0311 143 685.0 -0.0007 5867632* Male Indoor 0.0211 62 692.3 -0.0403 1295113 Male Indoor 0.0376 447 712.9 0.0112 1822625* Male Outdoor 0.0178 160 713.4 -0.0025 9999999* Male Indoor 0.0143 151 720.6 -0.0132 1561610* Male Indoor 0.0401 173 722.4 0.0069 Mean 1 0.0306 0.0025 Mean 2 0.0306 0.0096

4.1.11 Foods and Feeding

Of the foods offered, squid was eaten in preference to fish (pilchards before Tommy Ruff) or barramundi pellets. However, squid is expensive and deficient in calcium when fed alone. Pellets are reduced to a powder after more than 1-2 hours in the water, or after being bitten by the turtles. Nonetheless, pellets require less time for preparation and storage is easier.

The system of feeding the turtles in fresh water was effective. It was more preferable than feeding directly in the salt water, which became dirty quite quickly. Even feeding with smaller amounts of food offered more frequently (5 days/week) led to the water becoming dirty.

Under the current feeding procedures (ie fed in fresh water), wastage (leftover food) varied between the different types of foods (pellets 13-15%, marine fish 3-5%, chicken heads 15%, squid 20%). When turtles were fed in the afternoon in the salt water, there was increased wastage (pellets 31%, fish 9%). Smaller amounts of food, offered more frequently (5 days/week), resulted in very little leftover food (<5%).

27 Feed was offered at a mean rate of 0.11 kg/turtle/day (dry weight). This equates to 0.096 kg/turtle/day of pellets (wet weight) and 0.074 kg/turtle/day of fish and other feeds (wet weight) (Table 8).

4.1.12 Running Costs

The costs of running the breeding for 18 months are in Table 8. The main cost was labour (41%), associated with food preparation, feeding of turtles, cleaning pens, record-keeping and monitoring water parameters.

The relatively low cost of feed is due to the low number of turtles held over most of the study period, and do not reflect the likely costs had all animals survived. Based on number of turtles and time spent in the pen, the mean cost of food per turtle per day was calculated at $0.55.

Transporting sea water to the facility was expensive, and overall costs would have been lower if it had not been used. Costs of chemicals to maintain water quality would have been the same regardless of the type of water used.

Table 8. Summary of main running costs associated with operation of the breeding pen, from 27 October 2003 to 30 April 2005.

Chemicals Salt 20.2 t $3929 Lime 6 kg $4 Copper sulphate 1.8 kg $8 Cyanuric acid 9.5 kg $86 Liquid chlorine 917 l $812 Chlorine tablets 303 $1333 Sodium bicarbonate 24 kg $40 Subtotal $6211 17.4%

Water Fresh 3000 kl $2025 Sea 151 kl $3514 Subtotal $5539 15.5%

Electricity Pumps (2) $2236 Fans (3) $46 Heaters $827 Subtotal $3109 8.7%

Vehicle $3449 9.7%

Feed Pellets 450 kg $1349 Fish (incl. squid) 345 kg $1347 Supplements $20 Subtotal $2716 7.6%

Labour (feeding, cleaning, etc.) $14,638 41.1%

Total $35,662 Mean costs per month $1,981

28 4.2 Marketing of Turtles

4.2.1 Asian Trade in Freshwater Turtles

Freshwater turtles have been used as source of food and TCM for thousands of years. The explosive economic development in China over the last decade has resulted in a greatly increased demand for freshwater turtles, mainly for food. This trade has subsequently led to concerns about the sustainability of unregulated harvesting of many species from the wild.

Many of the species in trade are now listed on CITES Appendix I (no trade), or on Appendix II (regulated trade) and are protected from international trade in their countries of origin (CITES 2000). Concerns about the status of wild turtle populations are exemplified by the increasing numbers of freshwater turtle species that have been proposed for listing on the CITES Appendices, to assist the regulation of international trade (eg CITES 2000, 2002, 2003a, 2004; SMS 2004; Rhodin 2001).

The bulk of the trade in freshwater turtles is for food (Lau and Hitao 2000), but they are also used for TCM, the pet trade, curio market and for religious releases (Compton 2000).

4.2.2 Trade

As an indication of the increased demand for turtles, official Hong Kong trade statistics indicated over 13,500 tons of live turtle (equivalent to about 9 million individuals) were imported in 1998 alone - a 28 fold increase over 1992 imports. Eighty-four species were recorded in Hong Kong markets in 1998, 26 (31%) more than were encountered 5 years earlier (Lau et al. 2000). Over 99% of these turtles were for the food trade.

Taiwan is a major producer of non-endemic soft-shell turtles (1998 production = 2237 tons). Over a 3- year period (1996-1998), Taiwan exported 3308 tons of live soft-shell turtles, 0.5 million live hatchings and 95,000 turtle eggs, mainly to countries in Southeast Asia (eg Hong Kong/China imported 23% of live turtles, 33% of hatchlings, 5% of eggs). Over the same period, 0.4 tons of live hard-shell turtles were reported to be exported (Chen et al. 2000).

The Asian trade in freshwater turtles is non-species specific, with the majority of species in the region being involved in trade (Table 9) - the majority are harvested from the wild. With the exception of farmed Pelodiscus sinensis and Trachemys scripta, the 10 most traded species in Hong Kong during the period May 1998 to May 1999 were: pet trade, Chinemys reevesii, Chelydra serpentina, Ocadia sinensis, Graptemys pseudogeographica; food/pet trade, Mauremys mutica, Cuora amboinensis, Cuora flavomarginata, Cuora galbinifrons, Cuora trifasciata; food, Notochelys platynota (Lau et al. 2000). Eight of these 10 species are of Asian origin. At least 10 North American species have been reported from Asian markets, with Trachemys scripta and Chrysemys picta being the commonly traded ones (Table 9).

As the main consumer nation, China has taken significant steps in recent years to address the extensive trade in freshwater turtles. This includes suspension of exports for commercial purposes for all species except Pelodiscus sinensis and Chinemys reevesii, importation only from countries with annual quotas, suspension of imports from some countries, and size limits on alien species. Malaysia established national export quotas for all endemic species in 2003, and suspended exports of wild Notochelys platynota, Malayemys subtrijuga and Cyclemys dentata (CITES 2004).

29 Table 9. Species of freshwater turtles found in Asian food markets (Sources: NCMA 2002; Compton 2000; Lau and Haitao 2000; Chen et al. 2000; CITES 2002, 2003). F= food, TCM= traditional Chinese medicine, PT= pet trade, RR= religious release. *= US species.

Species Common Name Use

Common in Trade Amyda cartilaginea Asian Softshell Turtle F Chinemys reevesii Reeves' turtle F, TCM Cuora amboinensis Asian Box Turtles TCM Indotestudo forstenii Travancore Tortoise Heosemys grandis Giant Asian Pond Turtle TCM Heosemys spinosa Spiny Turtle PT Hieremys annandalii Yellow-headed Temple Turtle RR, TCM Indotestudo elongata Elongated tortoise PT, F, TCM Mauremys mutica Chinese Freshwater Turtle TCM Manouria impressa Impressed Tortoise F, TCM Manouria emys PT, F, TCM Platysternon megacephalum Big-headed turtle F, TCM Pyxidea mouhotii Keeled Box Turtle F, TCM Ocadia sinensis Chinese stripe-necked turtle F, TCM Orlitia borneensis Malaysian Giant Turtle F Palea steindachneri Wattle-necked Soft-shell Turtle F, TCM Pelodiscus sinensis Chinese Softshell Turtle F Notochelys platynota Malayan Flat-shelled Turtle Siebenrockiella crassicollis Black Marsh Turtle TCM Chelydra serpentina * Common Snapping Turtle PT, F Trachemys scripta * Red-eared Slider PT, F

Occassionally in Trade Annamemys annamensis Annam Leaf Turtle Chelus fimbriata Matamata Turtle Chinemys megalocephala Chinese Broad-headed Pond Turtle F Chinemys nigricans Chinese Red-necked Pond Turtle F Chrysemys picta * Painted Turtle Cuora aurocapitata Golden-headed Box Turtle F, PT Cuora flavomarginata Yellow-margined Box Turtle F, TCM Cuora galbinifrons Indo-Chinese Box Turtle F Cuora mccordi McCord's Box Turtle F, PT Cyclemys dentata Asian Leaf Turtle TCM Cyclemys tcheponensis Stripe-necked Leaf Turtle Emydura subglobosa Red-bellied Short-necked Turtle Geochelone carbonaria South American Red-footed Tortoise Geochelone elegans Indian Star Tortoise PT, TCM Geochelone pardalis Leopard Tortoise Geochelone platynota Burmese Star Tortoise Geochelone sulcata African Spurred Tortoise Geoemyda spengleri Black-breasted Leaf Turtle F, PT, TCM Graptemys pseudogeographica * Mississippi Map Turtle Hardella thurjii Crowned River Turtle TCM Kachuga tecta Indian Roofed Terrapin TCM Lissemys punctata Indian Flapshell turtle Macroclemys temminckii * Alligator Snapping Turtle Malayemys subtrijuga Malayan Snail-eating Turtle TCM Melanochelys trijuga Indian Black Turtle

30 Table 9 continued. Species of freshwater turtles found in Asian food markets (Sources: NCMA 2002; Compton 2000; Lau and Haitao 2000; Chen et al. 2000; CITES 2002, 2003). F= food, TCM= traditional Chinese medicine, PT= pet trade, RR= religious release. *= US species.

Species Common Name Use

Morenia petersi Indian eyed turtle Pelochelys cantorii Giant Soft-shell Turtle F Phrynops hilarii Hilaire's Spotted-bellied Side-necked Turtle Sacalia bealei Eye-spotted Turtle F Sacalia psuedocellata False Eye-spotted Turtle PT, F, TCM Sacalia quadriocellata Four Eye-spotted Turtle F Sternotherus odoratus * Common Musk Turtle Testudo hermanni Hermann's Tortoise

Minor/Occassional Trade Aspideretes gangeticus Indian Soft-shell Turtle Aspideretes hurum Indian Peacock Soft-shell Turtle Batagur baska Batagur Terrapin Callagur borneoensis Painted Terrapin Carettochelys insculpta Pig-nosed Turtle PT Chelodina siebenrocki Cuora trifasciata Golden Coin Turtle TCM, PT, F Geoclemys hamiltonii Spotted Pond Turtle Lissemys scutata Malaclemys terrapin * Diamondback Terrapin Morenia ocellata Burmese Eyed Turtle Ocadia glyphistoma Guangxi Stripe-necked Turtle Ocadia philippeni Philippen's Stripe-necked Turtle PT, F Pseudemys nelsoni * Florida Red-bellied Turtle Terrapene carolina * Terrapene ornata * Ornate Box Turtle Testudo horsfieldii Central Asian Tortoise F

Rarely Traded Cuora pani Pan's Box Turtle PT, F Cuora zhoui Zhou's Box Turtle PT, F

4.2.3 Food

Various species of soft-shell turtles are highly desired for food, including Pelochelys cantorii, Pelodiscus sinensis, Palea steindachneri, and Amyda cartilaginea. Although the market of turtles for human consumption favours the soft-shelled species, it appears that the majority of turtle species are consumed for food. Even odorous species such as Sternotherus odoratus from the USA are frequently found in Chinese markets (CNMA 2002).

The market has changed slightly over time with regard to species composition. For example, species common in the markets a few years ago are now being less frequently sighted (eg Geoemyda yuwonoi, Morenia petersi, Lissemys punctata, Chinemys reevesii, Mauremys mutica) (Ades et al. 2000). The reasons for these changes include decreased supply, but also increased availability of other species.

31 Informants in China (Beijing, Hangzhou, Chiangxing, Guangzhou) and Singapore were unable to provide specific information as to why certain species were favoured over others for eating purposes, other than turtles were considered a delicacy. A common way of cooking freshwater turtle is to double boil them in a soup with herbs.

The serving of freshwater turtle (and other wildlife species) during meals to business associates, guests, etc., reflects social status, and is a sign of affluence and a mark of respect towards the former. During two visits to China, both hard-shelled and soft-shelled turtles were served as courses during formal banquets. All markets and large restaurants visited had turtles for sale.

The distinction between food and medicine in Chinese culture is not clear. This is because certain foods have medicinal properties attributed to them. Like shark fin and bird's nest soup, eating turtle is considered "good for your health". There were no indications that eating turtles has an aphrodisiac effect.

The reasons why some turtle species command higher prices than others were also not clear. At Pudu Market in Kuala Lumpur hard-shelled turtles (Trachemys scripta) were being traded at $10 each and soft-shelled turtles (Pelodiscus sinensis) at $16 each ($20/kg and $40/kg respectively), the difference in price being attributed to the soft-shelled turtles being considered better for eating. When cooked, the soft-shell takes on a gelatinous texture and is eaten.

Turtles are typically transported live to the marketplace. Together with other live foods (seafood, crocodiles, snakes, frogs, insects, etc.), they are held in restaurants, usually where patrons can see them and select them for eating. It is important that food like turtles is fresh (live).

4.2.4 Traditional Chinese Medicine

After the food trade, the TCM market is the most significant, with both hard-shelled and soft-shelled turtles being used for medicinal purposes. The shell is believed to treat a variety of maladies, including dizziness, profuse sweating and impotence (Momphard 2004). It is also considered good for the kidneys, and is used to treat sore feet and hips, for strengthening bones and muscles, and treating diarrhoea, pancreatic infections and tumours. Chinese practitioners of Chinese medicine classify turtles into five or six general types, based on locale and medicinal effect - these groupings do not coincide with western scientific taxonomy.

At least 20 species are commonly traded for TCM purposes, including Cuora amboinensis, Malayemys subtrijuga and Siebenrockiella crassicollis (Chen et al. 2000). The Golden Coin Turtle (Cuora trifasciata) is highly desired in TCM because of its reputed cancer-curing qualities, and can fetch very high prices (up to $US1000/kg; Shi and Parnham 2001). Species imported from North America, such as Trachemys scripta, are farmed extensively in Southeast Asia, yet the shell is not readily accepted for TCM, as its clinical qualities are not known.

The shell used in TCM is usually a by-product of the food industry. Taiwan is a large importer of turtle shell for TCM from China and other parts of Southeast Asia, as it is unable to produce sufficient quantities itself. Between 1992 and 1998 1000 tons of hard-shelled turtles and 228 tons of shell were imported into Taiwan (Chen et al. 2000).

4.2.5 Pet Trade and Religious Ceremonies

Turtles are regarded as a symbol of long life and creatures with supernatural powers, and for these reasons and because of their reputation as being undemanding, they are commonly kept as pets in Asia. Although a small-scale international pet trade has been occurring for decades, a more substantial commercial international trade has developed in recent years.

32 Species that have charismatic qualities or are unusual are desired in the pet trade, and are exported from Asia to Europe and the USA (eg Manouria emys, Indotestudo elongata, Heosemys spinosa, Cuora trifasciata; Compton 2000). On the other hand, the USA exports many species of its endemic turtles and terrapins to the Asian market (NCMA 2002; Table 9). Species such as the Red-eared Slider Turtles (Trachemys scripta), a prolific breeder, are raised in large quantities on turtle farms in the USA, and are now being farmed in Asia. Other species in the pet trade include; the Pig-nose Turtle (Carettochelys insculpta), the Southeast Asian Box Turtle (Cuora amboinensis), the Asiatic Soft-shell Turtle (Amyda cartilaginea), the Asian Brown Tortoise (Manouria emys), Alligator Snapping Turtle (Macroclemys temminckii), Indian Star Tortoise (Geochelone elegans) and Siebenrock’s Snake-necked Turtle (Chelodina siebenrocki). Japan is a major pet market for freshwater turtles.

The pet trade in some parts of Asia is often associated with temples, and provides turtles for religious ceremonies. Releasing of turtles is a popular activity, as Buddhists believe that showing mercy by releasing live animals they will be rewarded with good fortune (Chen et al. 2000). In Taiwan, species sold for this purpose include Ocadia sinensis, Trachemys scripta and sometimes Mauremys mutica. Of the 24 species reported in the pet trade in Taiwan, three species (Ocadia sinensis, Trachemys scripta, Pelodiscus sinensis) represented 61% of the number of turtles reported in the survey being for religious releases (Chen et al. 2000). Species such as Hieremys annandalii are specifically sought after as temple turtles.

4.2.6 Turtle Farming

Farming of the Chinese Soft-shell Turtle Pelodiscus sinensis is known from Japan in the 19th Century (CITES 2003), and has been undertaken in Taiwan since at least the early 1900s (Chen et al. 2000). The demand and relatively high prices offered for freshwater turtles in the 1980s and 1990s spurred the development of turtle farming in countries such as Thailand, Malaysia, Taiwan and China. In recent years large numbers of hatchlings were exported from Taiwan to various countries, where they were used for farming (eg Thailand, Malaysia).

China has emerged as a significant source of farmed turtles. The full extent of turtle farming in China was not clear until the Chinese Government undertook a review of farming operations (CITES 2002c). The results indicated that farms were operating in 14 Provinces, and that significant numbers of turtles were being held. Of the 309 million turtles reported at the time of review, the most common species were Pelodiscus sinensis (303 million), Chinemys reevesii (2.8 million) Ocadia sinensis (1.5 million) and Chrysemys scripta (0.9 million). In order of numbers held, the following species were also recorded; Palea steindachneri, Maremys mutica, Pyxidea mouhotii, Cuora trifasciata, Chelydra serpentina, Geoemyda spengleri, Platysternon megalocephalum, Indotestudo elongata, Macroclemys temmincki, Sacalia quadriocellata and Cuora flavimarginata. Output from Chinese turtle farms in 2002 was estimated to reach 68,000 tons.

Pelodiscus sinensis is the most common species being farmed in China and other parts of Asia, and millions of individuals can be seen in markets in southern China at any one time (Artner and Hofer 2001). This species is favoured for commercial production because of high growth rates, high reproductive potential, widespread consumer acceptance and the understanding of conditions for farming (CITES 2003b). Due to the massive scale and successful breeding, the establishment of new farms in China in the 1990s, led to the production of Pelodiscus sinensis exceeding demand, and many farmers began to rear hard-shelled species such as Cuora trifasciata and Mauremys mutica. Some farmers ran into financial difficulties because of decreased prices (CITES 2003b). Farming of hard- shelled turtles is not considered as efficient as it is for soft-shelled species, as the former have slower growth rates and lower reproductive potential (CITES 2003b). However, species such as Cuora trifaciata are farmed in modest quantities due to their high individual market value (perceived cancer curing qualities), or for niche markets such as TCM, religious release and the pet trade (CITES 2000).

33 There is little published information available on turtle farming in Asia, with many farmers being guarded and secretive about their farming methods. Shi and Parham (2001) reported on a farm in Hainan Island (China) that began operations with Cuora trifasciata, Mauremys mutica and Ocadia sinensis in 1983, but now has some 50 species of turtle, including the soft-shelled Pelodiscus sinensis. Numbers of turtles of each species on this farm largely reflect the different market value of each species. Most turtles are housed in an outdoor (8 ha), vegetated enclosure, and in "dozens" of indoor, concrete breeding pools (connected to sand-filled nesting rooms). Eggs are collected and incubated at ambient temperature, and hatchlings reared in plastic buckets until they are 8-12 cm long, at which time they are transferred to the outdoor enclosure. Cuora trifasciata take 3 years to reach 1 kg, and 5 years to reach 2 kg, with high survivorship at each stage (95%).

Two species that are farmed extensively in Southeast Asia, Pelodiscus sinensis and Trachemys scripta, are not endemic there, but were originally imported from Taiwan and the USA respectively. In China, Chrysemys scripta has been farmed since 1986, and Chelydra serpentina and Macroclemmys temmincki were imported more recently from North America; the latter because of muscle content, high growth rates and high reproductive potential (CITES 2002a). Macroclemys temminckii were recorded at restaurants in Hangzhou, and were apparently farmed in the area (Zheijiang Province). CITES (2003b) indicates that Trachemys scripta is also extensively farmed in China.

4.3 Australian Turtles

Australia has at least 28 species of freshwater turtle, divided into two taxonomic families across 6 genera (Table 10), and inhabiting a range of freshwater habitats in all Australian States and Territories except Tasmania. Six of the world's seven species of marine turtle are also found in Australian waters, and nest on the mainland and various islands.

At the State/Territory level, the responsibility for wildlife lies with each appropriate wildlife agency. Where export out of Australia is involved, the trade must comply with Federal legislation, specifically the Environment Protection and Biodiversity Conservation Act 1999 (EPBC 1999 Act). Under this Act, commercial export is allowed where the specimens are derived from an approved source, including: approved captive breeding, aquaculture, and wildlife trade operations; approved wildlife trade management plans or an accredited wildlife trade management plan. Imports into Australia must also comply with the EPBC 1999 Act.

The Australian Government Department of Environment and Heritage is the designated CITES Management Authority. Under CITES, countries are able to adopt "stricter domestic measures" with regard to international trade in CITES-listed species, and Australia does so under the EPBC 1999 Act. This is exemplified by the prohibition on exports on live native mammals, birds, reptiles and amphibians for commercial purposes (Steensby 2004).

In assessing the potential for marketing of Australian turtles, a distinction was made between domestic and international trade, and the possible impediments to the development of each.

34 Table 10. Freshwater and marine turtle species in Australia (sources: Cogger 1992; Greer 2003; Cann 1998). * = most common species. In brackets; PNG= Papua New Guinea, IND= Indonesia. Only distribution in Australia is shown for marine turtles.

Species Common Name Distribution

Freshwater Turtles Carettochelys insculpta Pig-nosed Turtle NT (PNG, IND) Chelodina burrungandjii NT, WA Chelodina canni * Cann's Long-neck Turtle NT, QLD Chelodina expansa Broad-Shelled River Turtle NSW, VIC, SA, QLD Chelodina kuchlingi WA Chelodina longicollis * Eastern Snake-necked Turtle NSW, VIC, SA, QLD, ACT Chelodina novaeguineae New Guinea Snake-necked Turtle QLD (PNG, IND) Chelodina oblonga * Oblong Turtle WA Chelodina rugosa * Northern Snake-necked Turtle NT, WA, QLD Chelodina steindachneri * Flat-Shelled Turtle WA Elseya bellii Namoi River Snapping Turtle NSW Elseya dentata * Northern Snapping Turtle NT, QLD, WA Elseya georgesi George's Turtle NSW Elseya irwini Irwin's Turtle QLD Elseya latisternum Saw-Shelled Turtle NSW, QLD Elseya lavarackorum Gulf Snapping Turtle QLD Elseya purvisi Manning River Turtle NSW Elusor macrurus Mary River Turtle QLD Emydura australis Nort-west Red-faced Turtle WA Emydura krefftii * Krefft's River Turtle QLD, NT? Emydura macquarii * Murray Turtle NSW Emydura subglobosa Painted Short-necked Turtle QLD Emydura victoriae Northern Red-faced Turtle WA, NT, QLD Emydura signata Brisbane River Turtle NSW, QLD Emydura tanybaraga Northern Yellow-faced Turtle NT, QLD Emydura worrelli Worrell's Turtle NT Pseudemydura umbrina Western Swamp Turtle WA Rheodytes leukops Fitzroy Turtle QLD

Marine Turtles Natator depressus Flatback Turtle NT, WA, QLD Carreta carreta Loggerhead Turtle NT, WA, QLD Eretmochelys imbricata Hawksbill Turtle NT, WA, QLD Chelonia mydas Green Turtle NT, WA, QLD Lepidochelys oliviciea Olive Ridley Turtle NT, QLD Dermochelys coriacea Leatherback Turtle WA, QLD, NSW, NT?

35 4.3.1 Pet trade

Domestic: The current market for freshwater turtles within Australia is largely limited to the pet trade. Species that may be kept as pets within each State/Territory varies according to State/Territory legislation on the status and suitability of species within their jurisdiction. For example, in Western Australia the keeping of turtles as pets is limited to two species, Chelodina oblonga and C. steindachner, and species such as Emydura macquarii that are commonly kept as pets in other States are not permitted, due to their pest potential if they become established in the wild.

The unique appearance of many species maintains the demand for them in the pet trade. These include the Pig-nosed Turtle (Carettochelys insculpta), the long-necked and snake-necked turtles (Chelodina spp.), the Northern Snapping Turtle (Elseya dentata) and the Saw-shelled Turtle (E. latisternum). Most animals in the pet trade are derived from captive breeding, but a limited trade now occurs with Northern Long-necked turtle (Chelodina rugosa) hatchlings derived from wild eggs (ie ranching). This has been undertaken since 2000 by the Bawinanga Aboriginal Community (Maningrida, NT) as a commercial venture involving the harvest of eggs (induced from mature females) and captive raising of Chelodina rugosa to provide economic benefits to indigenous landowners. This program has been assisted with research support from the Applied Research Group at the University of Canberra and the Key Centre for Tropical Wildlife Management at Charles Darwin University (R. Hall, pers. comm.).

Some species are not considered for the pet trade due to their population status. For example, the Western Swamp Turtle (Pseudemydura umbrina) has a highly restricted distribution and is listed as Critically Endangered, and Mary River Turtle (Elusor macrurus) is listed as Endangered, at a Federal level.

The export and import of turtles between States/Territories requires permits from the relevant wildlife authorities.

International: Carettochelys insculpta is available in large numbers in the pet trade in Asia and elsewhere, and hatchlings were sighted by the authors in markets in Kuala Lumpur, Malaysia. In Indonesia (Papua Province) and Papua New Guinea, eggs are collected and artificially incubated, and the resulting hatchlings exported primarily to Asian markets, where they can be sold for up to $27 each (WCT 2005). From Asia, animals can make their way to the USA and Europe, where prices are much higher. As Carettochelys insculpta was listed on Appendix II in late 2004, it may become more difficult for collectors in Europe and the USA to acquire (Lunsford 2005). In Australia, the species has some characterstics suited to ranching: nests are located on sandy banks during the dry season (August-September), and eggs can be transported and incubated easily (Webb et al. 1986).

The price for freshwater turtles depends in part on their rarity in trade, and on unique appearance or characteristics. For this reason, Australian Chelonids (long-necks) are sought after in the pet trade in Europe and the USA. Lunsford (2005) indicates prices of $US50 for hatchling Elseya novaeguineae and $US100-200 for adults. Salzberg (2001) listed the price for a juvenile Elseya novaeguineae as $US65 in the USA. Hatchling Chelodina longicollis could fetch up to $US200 each, and adults a "few hundred dollars" (Lunsford 2005).

The low numbers of Australian species in the market is largely due to the EPBC 1999 Act) which does not allow for the export of live wildlife for commercial purposes. As many Australian turtles in the market were illegally acquired, it was difficult to gauge the extent of trade that may be occurring now. Nonetheless, the demand for Australian species is high.

In the USA alone, around 5.7 million turtles (142 taxa) were imported between 1989 and 1997, and like other reptiles they are becoming increasingly popular as pets (Telecky 2000). Australian species could certainly fill a niche in this market.

36 4.3.2 Food

Domestic: There is currently no established commercial market for turtles as food in Australia. However, indigenous people have harvested turtles for food for thousands of years, and they are a favoured bush tucker item (Vardon et al. 1999). The Aboriginal people of northern Australia still traditionally harvest freshwater and marine turtles as a source of protein.

To coastal indigenous people, access to marine environments allows the hunting of marine turtles and the collection of their eggs. Vardon et al. (1997) reported that on the basis of what coastal Aboriginal people in their survey were "willing to pay", green turtle (C. mydas) ranked much higher than magpie geese, wallaby, possum and flying fox. Hawksbill turtle meat is not as readily eaten as Green turtle, because the meat from wild turtles can sometimes be toxic. Human fatalities have been recorded in Australia, Oceania and Asia, and to a much lesser degree in the Caribbean (see Márquez 1990). Cuba had an extensive wild harvest of Hawksbill turtles for decades, with meat being the main product of the harvest (ROC 1998); illness as a result of eating E. imbricata meat has not been reported (E. Carrillo, pers. comm.). It is likely that the toxicity of the meat from different localities reflects differences in the composition of the diet in the wild.

Australia is a diverse multicultural society, and many cultures within it consider turtles as food (eg the Chinese community). There is a demand for freshwater turtles as food, mostly in Asian communities in southern States. Several Chinese restaurants in Sydney and Melbourne indicated that they would include freshwater turtles on their menus if they were available, and there is some pressure from some operators to pursue this sooner than later (R. Hall, pers. comm.). The preferred option would be to have turtles delivered live, so they could be processed prior to cooking. However, this may pose some problems with exporting between States, particularly where the species is not distributed in the importing State. Restaurants also indicated that sea turtle meat could be utilised.

The Chinese food industry is an immediate market identified for turtles, but as has occurred with crocodiles and other wildlife meats in Australia, attitudes of other cultures (eg Anglo-Saxon) can quickly change, and result in a more diverse market.

The procedures for the processing of crocodiles for human consumption were developed in the late 1980s (see Manolis and Webb 1990). The issues of concern that were addressed at that time with crocodiles will be the same ones for turtles, with Salmonella being the main one to address (see Manolis et al. 1991) - turtles, like crocodiles, harbour the bacteria naturally. Procedures for processing turtles, within approved facilities, at the point of export, will follow similar guidelines as have been established for crocodiles. Should turtles be processed at the restaurant level, procedures will need to be established in collaboration with the appropriate authorities.

Following its success with production of turtles for the pet trade, the Bawinanga Aboriginal Corporation is planning to expand into commercial production of turtles for food, based on the sustainable use of the resource. It is now preparing a management plan. Whether wild turtles are taken for food or as pets, current legislation in States/Territories prohibits the taking of wild freshwater turtles for commercial purposes, unless there is an approved management plan that outlines the sustainability of the proposed operation and ensures that the potential impacts on the wild populations are monitored (with the exception of New South Wales, where only kangaroos are subject to wild harvest).

Those species that are common and widely distributed in Australia and/or can be reproduced easily in captivity, would be most suited if markets can be created. Smaller, more exclusive markets could also be created for rarer species through careful management.

International: In Asia, soft-shell turtles are generally favoured for eating over hard-shelled species, and they are farmed in large numbers (see above). In Australia, there are no true soft-shelled turtles. The

37 Pig-nose turtle is a relatively large species, with mature individuals reaching 0.5 m length and weighing 20 kg. Unlike other freshwater turtles, it is characterised by flippers similar to marine turtles (Cogger 1992). It would be the most likely candidate for the food trade if production in captivity in large numbers were possible. Other Australian species may also be suitable, but there does not appear to be any particular characteristic that would make them more appealing than the Asian species currently being farmed or harvested from the wild (Table 9).

The trade in freshwater turtles for food involves live animals. As the EPCB 1999 Act does not allow the export of live animals for commercial purposes. Any exports, at least in the first instance, would need to be either whole dead or processed turtles. This would probably impact negatively on the price that could be obtained for Australian turtles.

4.3.3 Traditional Chinese Medicine

Domestic: TCM is becoming increasingly popular in Australia, (MacLennan et al. 2002; B. May, pers. comm.) which is reflected in a four-fold increase in imports of Chinese herbal medicines between 1992 and 1996 (DHS 1996). Opportunities exist for the utilisation of freshwater turtle shell into the TCM industry.

With the listing of many freshwater turtles on the Appendices of CITES importation of prepared medicines containing turtle shell into Australia is becoming more. The use of local shell in TCM would allow for quality assurance in the manufacture of medicines (Li 2003), and address issues of quality involved with imported products (Ernst 2002). Trials would need to be undertaken to test the TCM properties of Australian turtle shell before production on a larger scale could begin (B. May, pers. comm.).

Commercial production of freshwater turtles only for shell is not considered to be a feasible option, but as the shell would be a by-product of a potential food industry, its inclusion into TCM would add additional value for each turtle used.

International: Australian freshwater turtles do not currently feature in TCM practiced in Asia, and so they are unlikely to be accepted into that market unless medicinal properties can be confirmed through testing.

4.3.4 Turtle Shell

The decoration of shell carapaces (eg C. rugosa) by indigenous people is a tradition that continues to this day. The potential for the commercial use of decorated carapaces has been investigated (Kennett 1994), but it is unclear why it has not expanded to any degree. Through an approved management program such items could be traded on domestic and international markets. The shell of Hawksbill turtles is in high demand, particularly in Japan (see above).

38 5. Discussion

5.1 Captive Breeding of Hawksbill Turtles

5.1.1 General

The Hawksbill turtle is a more suitable species for ranching and farming than other marine turtles (eg C. mydas), as in addition to meat, the shell has a high value. The technology of raising E. imbricata in captivity is not complicated, and high growth rates have been achieved (Nodarse et al. 1998; WMI, unpubl. data). The international demand for shell from the artisanal Japanese Bekko industry is high, and will continue to be so as its stockpiles continue to decease in the absence of any legal international trade.

As all sea turtles are listed under Appendix I of CITES, there is currently no legal international trade in sea turtles of products derived from them. Transfer of E. imbricata to CITES Appendix II, which would allow for regulated international trade, could be on the basis of ranching or captive breeding. The pathway through CITES for ranching would be complicated, and would be need a more long-term view. However, captive breeding offers a shorter pathway through CITES.

Although successful nesting with captive raised individuals did not occur during the short period of this study, the study is continuing and the current adult females have enlarged follicles (determined with ultrasound), mate regularly, regularly excavate nest holes, and are expected to nest. Turtles were able to reach sizes for maturity at much younger ages than in the wild (see later).

Captive breeding as part of an aquaculture industry has many benefits. A small number of females can potentially produce a large number of eggs, but perhaps more importantly, genetic effects on shell colour could be better controlled.. The value of shell depends largely on its colour, with yellow (amber) being more valuable than other colours (red, red/black, black). The manipulation of raising conditions is another way in which shell colour can be influenced.

The potential benefits of an industry based on E. imbricata are wide-ranging and significant. For example, with most E. imbricata rookeries in northern Australia on Aboriginal lands, indigenous landowners could benefit in social (employment) and economic terms from an industry. Many coastal indigenous communities have traditional links with sea turtles that are not only used as a food source but are also of cultural significance. Interest has been already shown in turtle farming/ranching by a number of indigenous communities in eastern and western Arnhemland and Queensland (eg Maningrida, Nhulunbuy and Torres Strait).

Captive rearing of sea turtles would also have potential conservation benefits. It could provide a source of turtles for release back into the wild should that be considered necessary, and create incentives for conservation of habitats on which turtles rely, as has occurred with crocodiles in northern Australia.

5.1.2 Breeding Pen

The prototype breeding pen was designed to provide conditions conducive to reproduction in E. imbricata, and also to provide a suitable set-up for researching other aspects of captive raising of marine turtles. For example, the netting is effective in dividing the pen into different-sized enclosures, without compromising water circulation. Whether it proves suitable for breeding can only be established if the turtles nest, which they have not done so far.

Aspects of the pen design that merit further investigation are mainly related to maintaining water temperatures within acceptable levels throughout the year. At the end of the dry season, ambient

39 temperatures may be high (<35oC), resulting in water temperatures that may be unacceptably high. Costs to cool the entire volume of water (70 kl) in the pen by conventional means, even by 1-2oC, would be prohibitive. The modifications tested in this study, to cool some of the water by evaporative means, were cost-effective. They were easily integrated into the existing recirculation system, and utilised cheap, readily available fans. A disadvantage is that the fine water spray created by the sprinklers has the potential to cause rust in steel structures around the pen.

Like cooling, the costs involved in heating the large amount of water in the pen during the cool times of the year may be prohibitive. The provision of discrete enclaves of warm water were effective, and turtles quickly learned to enter the heating chamber to take advantage of the warmer water. The two turtles that utilised the heated chambers were also the first animals to begin feeding readily at the end of the cool season.

Doody et al. (2001) reported on Pig-nosed turtles (Carettochelys insculpta) in the Daly River (NT) sitting on thermal springs in the river when water temperatures were lower than that of the springs (ie cool, dry season). Nesting in areas with thermal springs preceded that in areas without springs, suggesting that these animals had accelerated follicular development. Maintaining higher body temperatures clearly assists reptiles to digest food and remain more active. This was evidenced by the two turtles provided with heated chambers beginning to feed before animals that had not been provided with heating.

There are no doubt improvements that can be made to this type of heating system. For example, insulation around the chambers would have made heating the water more cost-effective, and result in higher water temperatures in the chamber. Other options to provide heating could be the provision of enclosed, shallow water areas adjacent to the main pen, and that can be heated effectively. As reported here, the turtles appear to learn quickly, and take advantage of the heated areas.

5.1.3 Reproduction

In the Cayman Islands, captive male and female C. mydas were initially kept apart until just before the nesting season (Wood and Wood 1980), but since 1996 they have been housed together (J. Parsons, pers. comm.). Chelonia mydas have a well-defined nesting period, whereas E. imbricata nest throughout the year, with a 2-3 month peak of nesting (eg Moncada et al. 1998). In northern Australia, the peak of E. imbricata nesting occurs in the latter half of the year, from July to September (Chatto 1998). Thus, separation of the sexes until just before the nesting season is probably not a viable option, in the absence of data on when females are most likely to nest.

Captive C. mydas, females show a distinct 2-4 day "heat" period when they are receptive to the males - prior to and after this period males show no attention towards the females (Wood and Wood 19809). In the case of E. imbricata in this study, it was unclear whether females were "receptive" to the males, although after the first few days together they allowed males to mount them. Specific interest by females towards males was reported three times, in each case by the same female. In two of these cases regular mating was observed for a few days after, and hole-digging followed about two months later.

Mating in the wild takes place near the nesting sites, and is believed to involve multiple insemination prior to the nesting season (Witzell 1983; Miller 1994). Wood and Wood (1980) reported relationships between the proportion of captive C. mydas nesting and time elapsed between nesting after mating, and the amount of time spent mating. The low fertility reported for captive C. mydas has been the basis of a theory that the mating system of sea turtles requires an excess of reproductively active males, and that competition between males is necessary for maximum fertility (Owens 2000).

The mean ages taken to reach maturity for the captive-raised E. imbricata, 3 years for males and 7.6 years for females, are much less than those for wild E. imbricata. For our females it is likely that maturity, as evidenced by follicular development (Limpus and Reed 1985) could have been reached at

40 younger ages, as they had not been placed in "breeding" conditions. Fast growth rates have clearly allowed the turtles to reach sizes at which maturity could be expected. In Mexico and Cuba, where there are extensive areas of shallow, warm waters, E. imbricata grow relatively fast (MIP, unpubl. data; Garduño and Marquez 1994; CITES 2002), and attain maturity at much younger ages (10-15 years; Prieto et al. 2001) than has been estimated for E. imbricata growing slowly in cooler waters (>30 years; Limpus 1992).

Mean size of nesting E. imbricata varies between different locations (Witzell 1990; Márquez 1990), and females as small as 58.5 cm SCL have been recorded laying eggs in Cuba (Moncada et al. 1999), 54.9 cm in Malaysia (predicted from CCL; Pilcher and Ali 1999), 60.1 cm in the northern Great Barrier Reef (Dobbs et al. 1999), and 50.5 cm in the Arabian Gulf (Pilcher 1999). The smallest mature female in this study was 66.2 cm SCL, and the mean mature female was 73.1 cm SCL (Table 6). This mean size is smaller than that reported in the Caribbean [Cuba 83.5 cm; Puerto Rico 82.8 cm; Costa Rica 83.8 cm; Barbados 85.6 cm; Guyana 85.5 cm; Mexico 70.9-109.0 cm; Antigua 72.6-98.0 cm; (see CITES 2002)], but similar to that reported in the northern Great Barrier Reef [75.0 cm SCL; Loop et al. 1995; Dobbs et al. 1999)], Torres Strait [76.4 cm SCL (Limpus et al. 1983)] and southeast Asia [70.6-80.5 cm SCL; (predicted from CCL; Chan and Liew 1999); 69.9 cm SCL (predicted from CCL; Pilcher and Ali 1999)].

The use of ultrasound to monitor folicular development in adult females (Owen 1999; Rostal 2005), is perhaps better than laparoscopy (Limpus and Reed 1985) which is invasive, and in captive turtles more likely to lead to infection (Owen 1999).

Most captive-raised males were mature at around 50 cm SCL. The first mature male was histologically confirmed at 1.7 years (41 cm SCL), and the first signs of maturity as evidence by secondary sexual characteristics, in another group of E. imbricata being raised, were at 38 cm SCL. Based on growth rates in Cuban water, and assuming maturity at 68 cm SCL, Prieto et al. (2001) estimated that mean age would be 7 years at maturity. The data here indicate that maturity can be reached at much younger ages/sizes. Unfortunately research has tended to focus on nesting females, and little information is available on the male portion of the population.

The cues used by individual female E. imbricata to place their eggs on a particular portion of beach, including distance from the water, are unknown. Kamel and Mrosovsky (2004) reported that nest site choice for female E. imbricata in Antigua differed significantly among individuals, indicating distinct nesting preferences. Repeatability of nest location was significant, indicating nest-site choice is a consistent behaviour. However, a similar study in Guadeloupe suggested that nest site selection (distance from the water), was random for E. imbricata, and not a heritable trait (Mason et al. 2005).

The turtles in this study left the water and crawled around on the sand, often without digging, a behaviour that is also displayed by wild sea turtles. It is possible that the sand areas, in terms of thermal characteristics, consistency of sand, distance from the water, etc., were not considered appropriate by the females. Unfortunately mortality disrupted potential nesting before ovulation and egg formation (shell) occurred. Manipulation of the shape of the sand banks, perhaps looking at more slope, as they would appear in the wild situation, may merit further investigation.

41 5.1.4 Growth Rates

The decrease in growth rates was not surprising, particularly in view of the various factors that could have affected them. Growth rates of E. imbricata approaching maturity slow, and on animals reaching maturity they may approach zero (ie no growth). It is likely that the advanced follicular development of the mature females (Table 5) occurred during their time in the breeding pen, and may have impacted on growth. However, the transfer from small indoor controlled environment tanks and small outdoor pond is considered to have had a greater effect on growth rate of all animals, due to greatly increased swimming activity. For example, at 20 months of age, the growth rates of turtles moved to the outdoor pen were less than half those of turtles in the indoor facility, largely reflecting increased swimming (WMI, unpubl. data).

Relative to wild E. imbricata, the captive turtles were much heavier for their length when they were introduced into the pen. Even with the loss in weight over the study period, the remaining turtles were still more than 10-15% heavier than their wild counterparts (MIP, unpubl. data).

5.1.5 Mortalities

The vigorous activity displayed by the turtles after being put together, coupled with relatively high water temperatures at that time, (Fig. 7) are considered the most likely cause of the skeletal muscle myopathy associated with most turtles that died. Unlike many other reptiles, sea turtles may produce significant amounts of metabolic heat, particularly during swimming (Hirth 1962; Mrosovsky and Pritchard 1971). The difference between the body temperature of swimming sea turtles and the surrounding water may be much greater than the difference for turtles that are inactive (Standora et al. 1982; Spotila and Standora 1985). Hirth (1962) reported mean cloacal temperatures for nesting E. imbricata at 29.9oC.

The ability to produce significant amounts of metabolic heat may have caused critical levels of overheating to occur in this case. Nutritional myopathy caused by nutritional deficiencies may appear the same as exertional myopathy exacerbated by hyperthermia, but as the animals were fed a similar diet for over 7 years, it is unlikely to be implicated. The most obvious behavioural change that occurred immediately after the turtles were placed in the breeding pen was extreme exertion and interaction (chasing) with other turtles. Even turtles previously held in an outdoor pond had never reacted in this manner towards each other when put together.

Hypothermia could also be a major immunosuppression factor contributing to other secondary bacterial and fungal infections, and may have contributed to the kidney and heart lesions through increased oxygen demand due to hyperthermia (C. Shilton, pers. comm.). The initial symptoms of floating, often at an angle to the water surface, and difficulty diving, are commonly reported in sea turtles suffering from infectious and non-infectious diseases (Jacobson 1998). Like crocodilians, the effects of "stress" may remain for long periods of time, even after causative factors are removed or treated.

5.1.6 Foods and Feeding

Food conversion rates (FCR) could not be calculated for animals in this study due to loss of weight, but food conversion data for these animals at younger ages are available. From 2 to 3 years of age, with E. imbricata in indoor facilities, feeding pilchards resulted in 27% of the wet weight of food eaten being converted to bodyweight (6.7% dry weight). This was much higher than that obtained with a different type of fish (dollar fish; 3.9% dry weight, 15.5% wet weight) or pellets (15.5% wet weight, 16.7% dry weight). This confirms that pellets are a more cost-effective source of food.

There is currently no pellet that meets the specific needs of E. imbricata. A pellet has been developed for C. mydas farmed in the Cayman Islands, but this species is herbivorous. The pellets used for

42 raising these and other E. imbricata in Darwin appear to be suitable for attaining high growth rates, but there is still a need to elucidate in more detail the nutritional requirements of this species.

Although turtles appear to prefer marine fish and squid to pellets, from a cost point of view pellets are a cheaper source of protein. The rate at which food is converted to body weight is also an important variable to consider in evaluating different foods.

The FCR for a mixture of fish and pellets was higher for turtles in the indoor tanks (11.9% wet weight, 22.5% dry weight) than was recorded for turtles in the outdoor pond (4.5% wet weight, 11.9% dry weight). This difference reflects the additional energy expended by turtles actively swimming in the outdoor pond.

5.2 Marketing of Turtles

There is no doubt that Australian species would be suitable for the food trade. The main challenges to developing an aquaculture industry for freshwater turtles will be to identify those species with the greatest potential for captive raising, and to develop cost-effective raising technologies to produce a competitively-priced product, for either domestic or international trade. Many aspects of farming freshwater turtles are known, and there is considerable information available from the USA where turtle farming has been established for many years. In northern Australia, the experiences of the crocodile industry are a valuable asset that would assist commercial-scale turtle farming operations.

The EPBC 1999 Act is a clear impediment to international trade from Australia, as it prohibits live exports of native turtles. A strong market appears to exist for Australian freshwater turtles in Europe and the USA, but their suitability for the Asian market is unknown. Given the current levels and nature of the turtle trade for food in Asia, it is unlikely that processed (frozen) turtles could compete with live turtles that can be obtained readily from farms and other sources in the region. However, with the pressure on wild turtle populations in Asia and the increased regulation of trade through CITES, this situation may well change over time. Even so, live turtles for food will probably continue to be the preferred option in Asia, and efforts may be required to review the EPBC 1999 Act to facilitate live exports for commercial purposes.

Domestic trade offers the most promising option for a turtle industry, at least in the first instance. Turtles for the food trade would be the main product from such an industry, with shells perhaps being marketed to the growing TCM industry in Australia. A legal source of turtles would be welcomed by the restaurant trade in southern States, particularly larger restaurants, and could replace the limited illegal trade rumoured to be operating now.

Potential concerns about legal trade encouraging illegal harvesting of wild turtles could be addressed through a marking system for turtles, and by limiting distribution and retail access in the first instance. This same approach was taken by the Northern Territory with crocodile meat, and it allowed effective monitoring of trade whilst the industry became more established. In addition, methods to distinguish wild and captive-raised turtle meat and other products could be developed [eg see Moncada et al. (1998) for wild versus captive Hawksbill turtles].

Commercial enterprises based on ranching and captive breeding of turtles for meat and other products could benefit both Aboriginal and non-Aboriginal landowners in Australia. Conservation benefits would also accrue, by creating incentives for conservation of habitats on which turtles rely, as has occurred with crocodiles in northern Australia. The Bawinanga Aboriginal Corporation has initiated the development of a ranching operation with Chelodina rugosa, linked to a discrete research program to obtain the basic data required to develop a management program. They would utilise the traditional knowledge and resources of the area, and provide tangible economic benefits to the traditional landowners.

43 6. Implications

Hawksbill turtles are highly amenable to ranching and subsequent raising in captivity, and early maturation in captivity supports their potential for captive breeding. Like the crocodile industry, an industry based on Hawksbill turtles could involve indigenous landowners, who have shown an interest in this type of wildlife industry over many years. Most nesting areas for E. imbricata lie on indigenous land. Captive breeding is considered a safer option than ranching, and provides a shorter route for approval of international trade through CITES - it is the most logical approach to take in the first instance.

A domestic market for freshwater turtles as food could be developed in Australia, and assessment of other derivatives such as the shell for Traditional Chinese Medicine could be undertaken at the same time. Ranching of freshwater turtles presents an opportunity for economic development of indigenous landowners, and has already been initiated by one community in Arnhem Land. The sustainable use of wildlife is now recognised as a legitimate tool that can contribute to conservation.

44 7. Recommendations

Successful breeding (nesting) of Hawksbill turtles involves a complex interaction between a suite of social, behavioural, physiological and environmental factors prior to and during the breeding season. In light of the results obtained during this study, additional research over a longer period of time could help quantify the optimum conditions for captive breeding of Hawksbill turtles. A more detailed analysis of potential markets for freshwater turtles, with particular emphasis on species that would be amenable to ranching and/or captive breeding, would allow cost-benefits for domestic trade to be estimated more realistically.

45 8. References

Ades, G., Banks, C.B., Buhlmann, K.A., Chan, B., Chang, H.C., Chen, T.H., Crow, P., Haupt, H., Kan, R., Lai, J.Y., Lau, M., Lin, H.C. and Haitao, S. (2000). Turtle trade in northeast Asia: regional summary (China, Hong Kong, and Taiwan). Pp. 52-54 in Asian Turtle Trade: Proceedings of a Workshop on Conservation and Trade of Freshwater Turtles and Tortoises in Asia, ed. by P.P. van Dijk, B.L. Stuart and A.G.J. Rhodin. Chelonian Research Foundation: Massachusetts. (Chelonian Research Monographs 2: 52-54).

Artner, H. and Hofer, A. (2001). Observations in the Qing Ping Free Market, Guangzhou, China, November 2000. Turtle and Tortoise Newsletter 3: 14.

Cann, J. (1998). Australian Freshwater Turtles. Beaumont Publishing Pty. Ltd.: Singapore.

Caillouet, C.W., Fontaine, C.T., Manzella-Tirpak, S.A. and Shaver, D..J. (1995). Survival of head- started Kemp’s ridley sea turtles (Lepidochelys kempii) released into the Gulf of Mexico or adjacent bays. Chel. Conserv. Biol. 1: 285-292.

Carrillo, E., Moncada, F., Elizalde, S., Nodarse, S., Pérez, C. and Rodríguez, A.M. (1998). Annex 4. Historical harvest, trade and sampling data. Rev. Cubana Invest. Pesq. 22(1): 82-88.

CCMA (1998). Clarification and update: Cuba's proposal on Hawksbill turtles (Eretmochelys imbricata). Rev. Cubana Invest. Pesq. 22(1): 196-205.

Chan, E.H. and Liew, H.C. (1999). Hawksbill turtles, Eretmochelys imbricata, nesting on Redang Island, Terengganu, Malaysia, from 1993 to 1997. Chel. Conserv. Biol. 3(2): 326-329.

Chatto, R. (1998). A preliminary overview of the locations of marine turtle nesting in the Northern Territory. In Marine Turtle Conservation and Management in Northern Australia, ed. by R. Kennett, A. Webb, G. Duff, M. Guinea and G. Hill. CINCRM and CTWN, Northern Territory University: Darwin.

Chen, T.H., Lin, H.C. and Chang, H.C. (2000). Current status of chelonians in Taiwan. Pp. 45-51 in Asian Turtle Trade: Proceedings of a Workshop on Conservation and Trade of Freshwater Turtles and Tortoises in Asia, ed. by P.P. van Dijk, B.L. Stuart and A.G.J. Rhodin. Chelonian Research Foundation: Massachusetts. (Chelonian Research Monographs 2: 45-51).

CITES (2000). Trade in Freshwater Turtles and Tortoises to and in Southeast Asia. Doc.11.35, COP11, Gilgiri, Kenya, 10-20 April 2000. (Available at www.cites.org).

CITES (2002a). Application to register a captive breeding operation involving Chelonia mydas on Grand Cayman, Cayman Islands. Submitted by UK of Britain and Northern Island to COP12, Santiago, Chile, October 2002.

CITES (2002b). Hawksbill turtles in the Caribbean region: basic biological characteristics and population status. Information Document to CITES Regional Dialogue Meeting on Hawksbill Turtles. CITES: Geneva.

CITES (2002c). Report on the captive breeding of tortoises and freshwater turtles in China. Inf. 8, COP12, Santiago, Chile, 3-15 November 2002. (Available at www.cites.org).

CITES (2003a). Conservation and trade of turtles and tortoises: updated review of status and threats with recommendations for CITES listings. AC19 Doc. 15.1. 19th meeting of the Animals Committee, Geneva, Switzerland, 18-21 August 2003. (Available at www.cites.org).

CITES (2003b). Trade in tortoises and freshwater turtles: development of mid- and long-term conservation measures for tortoises and freshwater turtles. AC19 Doc.15.2 Rev. 1, 19th meeting of the Animals Committee, Geneva, Switzerland, 18-21 August 2003. (Available at www.cites.org).

CNMA (CITES Management Authority of China) (2002). Identification Manual for Common Turtles and Tortoises. China Forestry Publishing House: Beijing.

Cogger, H.G. (1992). Reptiles and Amphibians of Australia. Fifth edition. Reed Books: Chatswood, NSW.

Compton, J. (2000). An overview of Asian turtle trade. Pp. 24-29 in Asian Turtle Trade: Proceedings of a Workshop on Conservation and Trade of Freshwater Turtles and Tortoises in Asia, ed. by P.P. van Dijk, B.L. Stuart and A.G.J. Rhodin. Chelonian Research Foundation: Massachusetts. (Chelonian Research Monographs 2: 24-29).

Davis, P.M.C. (1981). Anatomy and physiology. Pp. 9-73 in Diseases of the Reptilia, ed. by J.E. Cooper and O.F. Jackson. Academic Press: London.

DHS (1996). dhs.vic.gov.au/pdpd/chinese/report/contents.html.

Dobbs, K.A., Miller, J.D., Limpus, C.J. and Landrey, Jr., A.M. (1999). Hawksbill turtle, Eretmochelys imbricata, nesting at Milman Island, northern Great Barrier Reef, Australia. Chel. Conserv. Biol. 3(2): 344-361.

Donnelly, M. (1994). Sea Turtle Mariculture. A Review of Relevant Information for Conservation and Commerce. Centre for Marine Conservation: Washington.

Doody, J.S., Sims, R.A. and Georges, A. (2001). Use of thermal springs for aquatic basking by the Pig-nosed turtle, Carettochelys insculpta. Chel. Conserv. Biol. 41(1): 81-87.

EA (Environment Australia) (2002). Draft Recovery Plan for Marine Turtles in Australia. Environment Australia: Canberra.

Ernst E. (2002). Toxic heavy metals and undeclared drugs in Asian herbal medicines. Trends Pharmacol. Sci. 23:136-139.

Fontaine, C. and Shaver, D. (2005). Head-starting the Kemp's Ridley sea turtle, Lepidochelys kempii, at the NMFS Galveston Laboratory, 1978-1992: a review. Chel. Conserv. Biol. 4(4): 838-845.

Fosdick, P. and Fosdick, S. (1994). Last Chance Lost? Can and farming save the green sea turtle? The story of Mariculture Ltd. - Cayman Turtle Farm. Irvin S. Naylor: York, Pennsylvania.

Garduño, M.A. and Márquez, R. (1994). Tagging and returns of hawksbill sea turtle in Las Coloradas, Yucatan, Mexico. In Proceedings of the International Workshop on the Management of Marine Turtles ‘94. 28-30 March 1994, Tokyo.

Garnett, S. (1983). Nutrition and Farm Husbandry of the Green Turtle (Chelonia mydas) and the Estuarine Crocodile (Crocodylus porosus). Unpublished PhD Thesis, James Cook University, Townsville.

Green, D. (1998). basking in Galapagos green turtles. Pp. 60-62 in Proceedings of the 17th Annual Sea Turtle Symposium. Orlando, USA, 4-8 March 1997. US Dep. Commerc. NOAA Tech. Memo. NMFS-SEFSC-415.

47 Greer, A.E. (2003). Encyclopedia of Australian Reptiles. Australian Museum Online, http://www.amonline.net.au/herpetology/research/encyclopedia.pdf Version date: 25 May 2004.

Hirth, H.F. (1962). Cloacal temperatures of green and hawksbill sea turtles. 1962: 647-648.

Hutton, J. and Webb, G.J.W. (2002). Crocodiles: legal trade snaps back. Pp. 108-120 in Trade in Wildlife: Regulation for Conservation, ed. by S. Oldfield. Earthscan Publications: London.

Ishmael, D. and Antonio, D.P. (2005). Merir Island, Republic of Palau Turtle Monitoring Project. In Proceedings of the 25th Annual Sea Turtle Symposium. Savannah, Georgia, USA. (in press).

Jacobson, E.R. (1998). Buoyancy problems in sea turtles: causes and diagnosis. Pp. 64 in Proceedings of the 17th Annual Sea Turtle Symposium. Orlando, USA, 4-8 March 1997. US Dep. Commerc. NOAA Tech. Memo. NMFS-SEFSC-415.

JBA (Japan Bekko Association) (1994). Proceedings of the The International Workshop on the Management of Marine Turtles in Tokyo '94. 28-30 March 1994. JBA: Tokyo.

Kamel, S. and Mrosovsky, N. (2004). Repeatability of nesting preferences and their fitness consequences in hawksbill sea turtles. In Proceedings of the 24th Annual Sea Turtle Symposium. San Jose, Costa Rica, (in press).

Kennett, R. (1994). A proposed commercial trade in turtle shell artefacts by Aboriginal people in the Northern Territory. Conservation Commission of the Northern Territory: Darwin.

Kennett, R. (1996). Growth models for two species of freshwater turtle, Chelodina rugosa and Elseya dentata, from the wet-dry tropics of northern Australia. Herpetologica 52: 383-395.

Kowarsky, J. (1977). Culture of young green turtles, Chelonia mydas, in Torres Strait, northern Australia. Aquaculture 11: 197-215.

Lau, M. and Haitao, S. (2000). Conservation and trade of terrestrial; and freshwater turtles and tortoises in the People's Republic of China. Pp. 30-38 in Asian Turtle Trade: Proceedings of a Workshop on Conservation and Trade of Freshwater Turtles and Tortoises in Asia, ed. by P.P. van Dijk, B.L. Stuart and A.G.J. Rhodin. Chelonian Research Foundation: Massachusetts. (Chelonian Research Monographs 2: 30-38).

Lau, M., Chan, B., Crow, P. and Ades, G. (2000). Trade and conservation of turtles and tortoises in the Hong Kong Special Administrative Region, People's Republic of China. Pp. 39-44 in Asian Turtle Trade: Proceedings of a Workshop on Conservation and Trade of Freshwater Turtles and Tortoises in Asia, ed. by P.P. van Dijk, B.L. Stuart and A.G.J. Rhodin. Chelonian Research Foundation: Massachusetts. (Chelonian Research Monographs 2: 39-44).

Li, G.Q. (2003). The quality and safety of traditional Chinese medicines. Aust. Prescr. 26: 128-130.

Limpus, C.J. (1992). The Hawksbill turtle, Eretmochelys imbricata, in Queensland: population structure within a southern Great Barrier Reef feeding ground. Wildl. Res. 19: 489-506.

Limpus, C.J. (1998). Definition of "adult" for marine turtle growth models. Pp. 67 in Proceedings of the 17th Annual Sea Turtle Symposium. Orlando, USA, 4-8 March 1997. US Dep. Commerc. NOAA Tech. Memo. NMFS-SEFSC-415.

48 Limpus, C.J., Miller, J.D., Baker, V. and McLachlan, E. (1983). The Hawksbill turtle, Eretmochelys imbricata (L.), in north-eastern Australia: the Campbell Island rookery. Aust. Wildl. Res. 10: 185-197.

Limpus, C.J., Miller, J.D. and Preece, N. (1994). The basking greens of Bountiful Island - Kay's turtles revisited. Pp. 76-77 in Proceedings of the 14th Annual Symposium on Sea Turtle Biology and Conservation. NOAA Tech. Memo. NMFS-SEFSC-351.

Limpus, C.J. and Reed, P.C. (1985). The green turtle, Chelonia mydas, in Queensland: a preliminary description of the population structure in a coral reef feeding ground. Pp. 47-52 in Biology of Australasian Frogs and Reptiles, ed. by G. Grigg, R. Shine and H. Ehmann. Surrey Beatty & Sons: Chipping Norton.

Loop, K.A., Miller, J.D. and Limpus, C.J. (1995). Nesting by the Hawksbill Turtle (Eretmochelys imbricata) on Milman Island, Great Barrier Reef, Australia. Wildl. Res. 22: 241-252.

Lunsford, R. (2005). Concise guide to choosing exotic turtles as pets. (Available at www.austinsturtlepage.com/Articles).

MacGregor, J. (2002). International trade in crocodilian skins: review and analysis of the trade and industry dynamics for market-driven conservation. Pp. 12-18 in Crocodiles: proceedings of the 16th Working Meeting of the IUCN-SSC Crocodile Specialist Group. Gainesville, Florida, USA, 7-10 October 2002. IUCN: Gland, Switzerland.

MacLennan AH, Wilson DH, Taylor AW. (2002). The escalating cost and prevalence of alternative medicine. Prev. Med. 35:166-173.

Manolis, S.C. and Webb, G.J.W. (1990). Crocodile management and research in the Northern Territory: 1987-88. Pp. 38-68 in Crocodiles. Proceedings of the 9th Working Meeting of the IUCN-SSC Crocodile Specialist Group, Lae, Papua New Guinea, 19-21 October 1988. IUCN: Gland, Switzerland.

Manolis, S.C., Webb, G.J.W., Pinch, D., Melville, L. and Hollis, G. (1991). Salmonella in captive crocodiles (Crocodylus johnstoni and C. porosus). Aust. Vet. J. 68: 102-105.

Márquez, R. (1990). FAO Species Catalogue Volume II. Sea Turtles of the World. FAO Fisheries Synopsis No. 125. FAO: Rome.

Mason, P., Valentine, A. and Stapleton, S. (2005). Nest site selection in Hawksbill turtles (Ertemochelys imbricata), within and among seasons, Antigua, West Indies. In Proceedings of the 25th Annual Sea Turtle Symposium. Savannah, Georgia, USA. (in press).

Miller, J.D. (1994). The hawksbill turtle, Eretmochelys imbricata: a perspective on the species. Pp. 25- 38 in Proceedings of the Australian Marine Turtle Conservation Workshop. ANCA: Canberra.

Moncada, F., Carrillo, E., Saenz, A. and Nodarse, G. (1999). Reproduction and nesting of the Hawksbill turtle (Eretmochelys imbricata) in the Cuban Archipelago. Chel. Conserv. Biol. 3(2): 257-263.

Moncada, F., Pérez, C., Nodarse, G., Elizalde, S., Rodríguez, A.M. and Meneses, A. (1998). Reproduction and nesting of E. imbricata in Cuba. Rev. Invest. Pesq. 22(1): 109-116.

Mrosovsky, N. and Pritchard, P.C.H. (1971). Body temperatures of Dermochelys coriacea and other sea turtles. Copeia 1971: 624-631.

49 Nodarse, G., Meneses, A., Manolis, S.C., Webb, G.J.W., Carrillo, E. and Pelegrin, E. (1998). Management program and procedures - ranching program. Rev. Invest. Pesq. 22(1): 162-165.

Owens, D.W. (1999). Reproductive cycles and endocrinology. Pp. 119-123 in Research and Management Techniques for the Conservation of Sea Turtles, ed. by K.L. Eckert, K.A. Bjorndal, F.A. Abreu-Grobois and M. Donnelly. IUCN Marine Turtle Specialist Group Publication No. 4.

Owens, D.W. (2000). Reproductive problems in captive and wild sea turtles. Pp. 2 in Proceedings of the 18th International Sea Turtle Symposium. Mazatlan, Mexico, 3-7 March 1998. US Dep. Commerc. NOAA Tech. Memo. NMFS-SEFSC-436.

Pilcher, N.J. (1999). The Hawksbill turtle, Eretmochelys imbricata, in the Arabian Gulf. Chel. Conserv. Biol. 3(2): 312-317.

Pilcher, N.J. and Ali, L. (1999). Reproductive biology of the Hawksbill turtle, Eretmochelys imbricata, in Sabah, Malaysia. Chel. Conserv, Biol. 3(2): 330-336.

Prieto, A., Moncada, F., Nodarse, G., Puga, R., De León, M.E, Díaz-Fernández, R., Espinosa, G., Castillo, D., Hernández, M., Pelegrin, E., de Arazoza, M., Salabarria, D., Morales, E., Webb, G., Manolis, C. and Gómez, R. (2001). Aspectos biologicos y ecologicos relacionados con la poblacion de tortuga carey en aguas cubanas. Informe de la República de Cuba. First CITES Dialogue Meeting on Hawksbill Turtles in the Greater Caribbean. Mexico City, 15-17 May 2001.

Rhodin, A.G.J. (2001). Making progress in freshwater turtle and tortoise conservation. Turtle and Tortoise Newsletter 3: 2-3.

Rice, M.R., Balasz, G.H., Hallacher, L., Dudley, W., Watson, G., Krussel, K. and Larson, B. (2000). Diving, basking, and foraging patterns of a sub-adult green turtle at Punalu'u, Hawaii. Pp. 229- 231 in Proceedings of the 18th International Sea Turtle Symposium. Mazatlan, Mexico, 3-7 March 1998. US Dep. Commerc. NOAA Tech. Memo. NMFS-SEFSC-436.

ROC (Republic of Cuba) (1998). An annotated transfer of the Cuban population of Hawksbill turtles (Eretmochelys imbricata) from Appendix I to Appendix II of CITES. Rev. Invest. Pesq. 22(1): 32-58.

ROC (Republic of Cuba) (2000). An annotated transfer of that part of the Caribbean population of Hawksbill turtles (Eretmochelys imbricata) inhabiting Cuban waters from Appendix I to Appendix II. Proposal 11.40, COP11, Nairobi, Kenya, April 2000.

Rostal, D.C. (2005). Seasonal reproductive biology of Kemp's Ridley sea turtle (Lepidochelys kempii): comparison of captive and wild populations. Chel. Conserv. Biol. 4(4): 788-800.

Salzberg, A. (20011). Turtles for sale. Turtle and Tortoise Newsletter 3: 22.

Shaver, D.J., Amos, A.F., Higgins, B. and Mays, J. (2005). Record 42 Kemp's Ridley nests found in Texas in 2004. Marine Turtle Newsletter 108: 1-3.

Shi, H. and Parham, J.F. (2001). Preliminary observations of a large turtle farm in Hainan Province, People's Republic of China. Turtle and Tortoise Newsletter 3: 4-6.

Simon, M. H., Ulrick, G.F. and Parks, A.S. (1975). The Green Sea Turtle (Chelonia mydas); mating, nesting and hatching on a farm. J. Zool. Lond.177; 411-423.

SMS (Species Management Specialists) (2004). A Guide to the Working Documents and Amendment Proposals. Uniprint NT: Darwin.

Spencer, R. (2002). Growth patterns of two widely distributed freshwater turtles and a comparison of common methods used to estimate age. Aust. J. Zool. 50: 477-490.

Spotila, J.R. and Standora, E.A. (1985). Environmental constraints on the thermal energetics of sea turtles. Copeia 1985(3): 694-702.

Standora, E.A., Spotila, J.R. and Foley, R.E. (1982). Regional endothermy in the sea turtle, Chelonia mydas. Journal of Thermal Biology 7(3): 159-165.

Steensby, C. (2004). International trade and CITES - tags, permits and stricter domestic measures. Pp. 102-106 in Crocodiles. Proceedings of the 17th Working Meeting of the IUCN-SSC Crocodile Specialist Group. Darwin, NT, 24-29 May 2004. IUCN: Gland, Switzerland.

Telecky, T. (2000). United States import and export of live turtles and tortoises. Turtle and Tortoise Newsletter 4: 81-3.

Van Dam, R., Diez, C. and Colon, L. (2004). Close and often: migratory and breeding behaviour of male hawksbill turtles. In Proceedings of the 24th Annual Sea Turtle Symposium. San Jose, Costa Rica. (in press).

Vardon, M.J., Gaston, S.M., Niddrie, J. and Webb, G.J.W. (1999). Wildlife use at Momega, north- central Arnhem Land. Australian Biologist 12(1): 15-22.

Vardon, M.J., Missi, C., Cleary, M. And Webb, G.J.W. (1997). Aboriginal use and conservation of wildlife: a cultural necessity. Pp. 241-245 in Conservation Outside Nature Reserves, ed. by P. Hale and D. Lamb. University of Queensland: Brisbane.

WCT (World Chelonian Trust) (2005). Carettochelys insculpta - Pig-nosed Turtle. (Available at www.chelonia.org/Articles/China/china29cinsculpta.htm).

Webb, G.J.W. (2002). Conservation and sustainable use of wildlife - an evolving concept. Pacific Conserv. Biol. 8(1): 12-26.

Webb, G.J.W., Choquenot, D. and Whitehead, P.J. (1986). Nests, eggs and embryonic development of Carettochelys insculpta (Chelonia: Carettochelidae) from northern Australia. J. Zool. Lond. (B) 1986(1): 521-550.

Wibbels, T., Owens, D.W. and Rostal, D.R. (1991). Soft plastra of adult male sea turtles: an apparent secondary sexual characteristic. Herp. Rev. 22: 47-49.

Witzell, W.N. (1983). Synopsis of Biological Data on the Hawksbill Turtle Eretmochelys imbricata (Linnaeus, 1766). FAO Fisheries Synopsis No. 137. FAO: Rome.

Wood, F. and Wood, J. (1980). Reproductive biology of captive green sea turtles Chelonia mydas. Amer. Zool. 20(3): 499-506.