Contents

3 Are we doing the right type of conservation science? 8 Training two captive Sumatran orang-utans for ultrasound scanning 11 As dead as a dodo – when do we declare a species Extinct? 13 Conservation and data management 17 Morbidity of Alaotran gentle lemurs at Durrell Wildlife Park 22 Effects of diet on the health of captive Geoffroy’s marmosets 25 Biodiversity loss in Gerald Durrell’s early works 27 Assessing muscle condition in captive Livingstone’s fruit bats 30 Conservation update on the ploughshare tortoise 33 Seed dispersal by ecological replacement Aldabra giant tortoises

Welcome to Solitaire 27. This issue starts with a very thought-provoking essay by Professor Richard Griffiths that highlights some of the problems both field and zoo-based conservationists face not only in doing high quality research, but in communicating it to the rest of the conservation world. These are problems Durrell’s staff and students face regularly, and they are problems for which Solitaire is trying to provide some solutions. Although we are aiming at the highest standards, we are often working with small numbers of individuals, and in situa- tions that cannot be easily, if ever, repeated. Solitaire is an outlet for such studies, making our results available to a wider audience and, hopefully, stimulating more research and encouraging collaboration.

This year’s Solitaire demonstrates perfectly what we are trying to do to improve evidence-based conserva- tion and the management of threatened species. Articles range from descriptions of techniques that have the potential to be of benefit to a wide variety of species – for example, habituating our female Sumatran orangutans to ultrasound equipment so that they can be monitored during pregnancy without stress – through updates on vitally important conservation initiatives in the field, to basic research on biology, behaviour and health, providing information that forms the essential underpinning of any conservation action.

I hope you enjoy this issue and that it gives you an insight into the huge range of work that Durrell is doing around the world to achieve its mission of saving species from extinction.

Very best wishes Eluned Price Wildlife Park Research Coordinator [email protected]

Cover photo: Alaotran gentle lemur (James Morgan) Solitaire No. 27 2 Guest essay Are we doing the right type of conservation science?

Richard Griffiths

Durrell Institute of Conservation and Ecology, University of Kent

Although he had no formal scientific training, Gerald how do we go about collecting, analysing and disseminating Durrell recognised the importance of a systematic, scientific the results? approach to animal husbandry and conservation very early on in his career. Few other zoos embraced this philosophy at Hard and fuzzy data the time, which meant that the Jersey Zoological Park was The terms ‘science’, ‘research’ and ‘evidence’ are often used founded using scientific principles that have continued to synonymously, but actually mean slightly different things. underpin the organisation through to the present day. Indeed, When we talk about ‘evidence’ we refer to factual informa- many zoos – and many of the broader conservation organi- tion to support conservation needs. This information may take sations – have been playing catch-up, as ‘evidence-based a variety of forms – published papers and reports, quantita- decision making’ has become increasingly recognised as an tive data, qualitative data, or even information residing in the essential component of good management practice. Howev- heads of experts that is extracted during a workshop. Howev- er, the ‘evidence’ that is needed to do effective conservation er, scientists are a sceptical lot that need a lot of convincing. is often complex, costly, and time-consuming to accumulate. The scientific method is rooted in rigorous hypothesis testing In addition, there are often competing pressures on research- using designs that incorporate observational, experimental ers that dictate the type of research that they do. So what or theoretical methods (or preferably a combination of all type of data do we need to inform sound conservation and three). This is fine if you are working in a physics lab where

It has been difficult to publish field research on the impacts of predation on the Mallorcan midwife toad because of small sample sizes (Photo: Richard A. Griffiths).

Solitaire No. 27 3 Guest essay

extraneous variables can be neatly controlled, or even if you fuzzy data that might cause a journal editor to wince, but it still are working on fruitflies where multiple generations can be may be valuable. For some endangered species it is not unusual bred in a couple of months in a converted broom-cupboard. for there to be practically no data at all, and in such cases you For larger scale ecological and conservation questions, ap- have to rely on opinions from experts or from local people plying the scientific method is enormously challenging. who are familiar with the species or habitats. Fortunately, a In addition to the problem of scale, you may find you are range of methods are available to obtain such data, and these working on a rare and cryptic species that confounds the re- are being increasingly embraced within conservation. The quirements of sample size and research design. We recently Population and Habitat Viability Analysis workshop process investigated how much survey effort would be has been invaluable in extracting information needed to reliably detect a population change from experts and using it to build population in a threatened amphibian species. Our “Within conservation, models of a whole range of threatened statistical models revealed that using tra- species (e.g. Conservation Breeding ditional methods we would actually need scientific rigour needs to Specialist Group 2016). As individual to monitor more sites than were actually experts are not always right (Austen at in existence to reliably detect a decline! be laced with a healthy al. 2016), there are also various tools On another occasion, we were doing field- available for consensus building among work in a remote mountain range on a very dose of pragmatism” experts and consolidating information in rare endemic frog – the Mallorcan midwife a systematic and balanced way (e.g. Pullin toad – that was threatened by introduced preda- et al. 2004; MacMillan & Marshall 2006; tors. We stumbled upon a small pond where it was uniquely Sutherland et al. 2011). In China, carrying out standard possible to do a predator removal experiment. We took some surveys of the giant salamander in streams and rivers has counts and measurements, and then duly removed the preda- proved to be logistically challenging. However, a survey of tors and monitored the impacts on the ecology, behaviour local communities has revealed extremely useful information and morphology of our rare frog over the coming weeks – a about past and present status and distribution (Pan et al. 2015). simple ‘before’ and ‘after’ field experiment using a single Scientists are starting to appreciate that within conservation, site. The results were quite compelling and showed signifi- scientific rigour needs to be laced with a healthy dose of cant predator impact, but we have never been able to publish pragmatism, and that the timescales needed for conservation the results. This is because referees of the paper deemed it action are usually substantially shorter than those needed to do insufficiently replicated. To do it properly we would have the relevant conservation research. Indeed, analytical methods needed to find at least another 10 sites with predators and an- are now emerging that combine available quantitative data other 10 without predators to act as controls. Unfortunately, with expert opinions (Kuhnert et al. 2010), and these have such sites were just not available. considerable potential for informing conservation practice. So what do we do when we just cannot get hard data? The simple answer is you have to work with what you have got When do we stop doing research and start doing and make an informed decision. This may include unpublished conservation? So when do we decide that we have accumulated enough evidence to make an informed decision about a conservation intervention? This problem was neatly summarised by Mc- Coy (1994) when considering what we do about amphibian declines: ‘…do ecologists wear their conservationist hats and muster their expertise in defense of life, or do they wear their scientist hats and muster their expertise in defense of “truth”?’. At one end of the spectrum, if conservationists have insufficient evidence to support a management inter- vention there is a risk they make an expensive mistake. At the other end of the spectrum, if they are too cautious and de- lay action until all necessary evidence has been assembled, the species may be extinct. Amphibian conservation provides examples of both extremes. In the early days of natterjack toad conservation in the UK, there was concern that ponds were drying up too early and killing all the tadpoles. The solution at the time seemed quite simple – deepen the ponds to make them more permanent. Unfortunately, this made the ponds much more attractive to predators and competitors that were unable to survive in temporary ponds, with the result that the natterjacks were actually worse off. Longer term research revealed that breeding in temporary ponds and occasional loss of natterjack Conservation management of the natterjack toad has tadpoles to desiccation is actually quite normal and part of been altered in light of long-term data on its requirements a wider metapopulation process. These are long-lived toads (Photo: Bernard Dupont, Creative Commons Attribution- and only need one or two seasons of reproduction in their Share Alike 2.0 license). lifetime to keep the population ticking over. As a result of

Solitaire No. 27 4 Doing conservation science

editorial aspirations of enlightened scientific journals, and the assessment criteria of grant panels. However, the cross- discipline push in conservation has been largely spearheaded from within one discipline – natural sciences. This is perhaps not surprising, as natural scientists are generally a highly col- laborative lot that relish engagement with those working in other fields and learning new methodologies. Equally, there are rather a lot of natural scientists being churned out by universities (I count myself as one of the crowd). Perhaps it is also not surprising, then, that the heads of many conserva- tion organisations have received a fairly traditional training in biology. However, I see far fewer researchers from the so- cial sciences and humanities moving into cross-disciplinary conservation research. There are a number of related issues that emerge from this trend. Firstly, what is – and what is not – good cross-disciplinary research depends on the discipline you come from. An ecologist may wax lyrical about how the application of an economic model to an ecological system The golden toad became extinct despite being well-stud- may be highly revealing, while an economist may regard ied in the wild. this as both misleading and pedestrian. A social anthropolo- gist may elegantly reveal how cultural practices influence attitudes to conservation using qualitative methodologies, but a biologist may dismiss the findings because they are not based on quantitative data that have been analysed sta- tistically. Academic institutions are still organised around this research the management programme was adjusted disciplinary structures that were formulated in the distant accordingly (Beebee 1993). The extinction of the golden past, and forging effective collaboration between researchers toad of Costa Rica has gone down as one of the classic with different world views and philosophies may be difficult stories of amphibian declines. The toad was once abundant (Evely et al. 2010). in the Monte Verde rainforest and was quite well-studied, but There are certainly good examples of ecologists, declined to extinction within a couple of years (Crump et al. geneticists, veterinarians, spatial modellers, economists, 1992). Conservationists subsequently criticised scientists for planners, lawyers and social anthropologists all doing their letting this happen, and suggested that there was a missed bit in a research project for the greater good. However, these opportunity to set up a safety-net population in captivity collaborations often fail to integrate methodologies at any (Harding 1993). Whether this was a sensible depth or embrace the public or conservation option at the time remains debateable, but practitioners (Evely et al. 2010). Because these two cases illustrate the dilemma “Many fundamental of the predominance of natural that conservationists face when trying scientists within conservation, it to assess the evidence to support conservation problems require also means that cross-disciplinary conservation management. research often leans towards the The sad reality is that we rarely small data that are basic yet research philosophies associated with have the luxury of being able to decide that discipline. Despite the upbeat when to – or when not to – intervene still challenging to collect.” talk coming from funding agencies on the basis of available evidence. The and journals, with research quality decision is often made for us by extraneous circumstances, benchmarks firmly rooted within traditional disciplines it is such as political agendas, availability of funding and difficult to give novel cross-disciplinary research the exposure, timeliness. In such circumstances, all we can do is synthesise support and uptake it needs. If conservation is to truly benefit what is available at the time from all sources of hard and fuzzy from cross-disciplinary research then two challenges must data, and make an informed decision on the basis of that. be surmounted: (1) getting a better balance of researchers involved across the relevant disciplines; and (2) developing Cross-disciplinary research an infrastructure for integrating the contributions from the That conservation requires an approach that straddles differ- different disciplines more effectively. ent disciplines to solve all its problems has become something of a mantra in recent years. For example, ecologists may be Big data and small data needed to assess populations and habitats, veterinarians to Globalisation, new technologies and scientific advances do health checks, social scientists to examine the role of lo- have meant that we can now tackle problems with millions cal communities, lawyers and policy makers to work out of data points that would have been completely intractable a how any plan fits with the political agenda and economists few years ago. The laptop that I am now writing this article to work out the cost-effectiveness. Ideally, some overarch- on has more computer power than the truck-sized mainframe ing recommendations should emerge from all this research. computer that analysed my PhD data some years ago. The This discipline-diverse approach is reflected in the mission availability of such technology is therefore driving scien- statements and staff profiles of many organisations, in the tists to address bigger and bigger problems with larger and

Solitaire No. 27 5 Guest essay

more complex data sets. Of course, this is all very worth- The way forward? while. If we are ever to fully understand the impacts of Despite the shortcomings in research meeting conserva- climate change, for example, we must continually build and tion needs, this is not an issue that has gone ignored by test increasingly sophisticated models that stretch available the conservation and research community. Indeed, explor- computer power. However, many fundamental conservation ing the science-implementation gap is becoming something problems require small data that are rather basic yet still very of a topical research discipline in itself (e.g. Arlettaz et al. challenging to collect. Species conservation action plans are 2010). At the institutional level, some of the issues are be- often hamstrung by the absence of a basic understanding of ing addressed. My own institution – DICE – was founded such things as survival rates, number of offspring and the back in 1989 specifically to bridge the gap between social factors regulating population size. Obtaining such small and natural sciences and between research and implementa- data on rare – and often cryptic – species may take sever- tion. That mission remains as relevant today as it was over al years of challenging fieldwork. This may be beyond the a quarter of a century ago, and over this period DICE has timeframe of a conventional research grant, and beyond the equipped over 900 graduates from nearly 100 countries with patience of a smart young scientist keen to progress quickly the cross-disciplinary tools needed to tackle conservation through the academic ranks. When the hard-won data are problems. I actually feel quite proud that the majority of eventually written-up, it is quite likely that the editor of a our graduates have gone on to forge careers in conservation high-impact journal will regard the work too limited in taxo- practice rather than join the treadmill of academia. Indeed, nomic and geographical scope to be of general interest. This many have gone on to become leaders in their respective creates further disincentives to pursue small data projects fields and pass on their skills and knowledge to a new gen- and motivates ambitious researchers to engage ever-more eration of practitioners. The conservation evidence initiative enthusiastically with the big data agenda. Is it therefore sur- based at Cambridge University is another project that aims prising that conservation practitioners bemoan the fact that to identify the key questions that are most important to prac- scientists are doing work that manifestly fails to address their titioners and develop mechanisms to make research results on-the-ground needs? and other forms of evidence more accessible to practitioners (http://www.conservationevidence.com/). This has result- The scientific journal problem ed in several syntheses of conservation evidence focusing Unfortunately, publishing conservation research in scientific on specific problems or taxa. In addition, the open access journals has become something of a high-stakes game in re- journal Conservation Evidence has been developed specifi- cent years. Within academia, careers, promotions and grant cally to publish case studies of conservation interventions by funding all depend on not so much practitioners. These are exactly the sort of what you are publishing but where case studies that would have difficul- you are publishing it. Getting your “Journals increasingly shun ty in getting published in mainstream work into a broad, mainstream journals, and would end up buried in journal – such as Nature or Sci- studies based on single inaccessible reports or possibly not ence – will indeed be a cause to written-up at all. In addition to being pop open the champagne. As journals species or single ecological completely open access to all, Conser- strive towards broad readership, they in- vation Evidence is also unusual in that creasingly shun studies based on single species systems” it levies no page charges. The Centre for Evidence or single ecological systems unless they can shed Based Conservation at Bangor University (http://www. light on problems of wider interest. However, most practi- cebc.bangor.ac.uk/) promotes evidence-based conservation tioners are more concerned about how to best manage their practice through the development of systematic evidence species or site rather than the broader scientific relevance of reviews, and also produces an open access journal, Environ- doing so. So this results in several trends. Firstly, conserva- mental Evidence. In Australia, the Centre of Excellence for tion journals publish a lot of papers on broad, generic topics Environmental Decisions (CEED: http://ceed.edu.au/) was that are of little relevance to conservation because practi- set up in 2011 to carry out research to solve environmental tioners have not asked the question that is being addressed. problems and evaluate the outcomes. From the initiative has Secondly, most conservation research is carried out in the emerged an online magazine, Decision Point, which carries developed world on species and systems that may not be top lively articles and viewpoints on a wide variety of environ- conservation priorities (Milner-Gulland et al. 2009; Griffiths mental decision-making and the research that informs it. & Dos Santos 2012). Thirdly, scientists are getting quite Scientific journals – the vehicles by which knowledge clever at dressing up their research in grandiose narratives is disseminated – are also having to adapt to a rapidly that make it sound more widely relevant than it actually is. changing research landscape. Governmental drivers towards On top of these issues, is the fact that in true Orwellian fash- research transparency and making publically-funded research ion, some animals are more equal than others. For example, accessible to all means that research findings no longer need Bonnet et al. (2000) neatly showed that if you are working to be buried in a journal only available to those paying a on ‘more popular’ mammals and birds, you can get away subscription: they can be found online with a few few clicks with less conceptual/theoretical introductory material in a of a mouse. Much cross-disciplinary research falls between research article than if you are working on fishes, amphib- the stools of traditional journals, and open access journals that ians and reptiles. So if you want to publish your desert lizard publish on technical merit – rather than prevailing fashions, paper in a top journal, try and slip in the species and habitat agendas and perceptions of what is of ‘broad interest’ – are names later on in the paper, after you have framed it in some providing a welcome home for much research that crosses grander, more theoretical problem of global importance. the divides.

Solitaire No. 27 6 Doing conservation science

Despite the drive towards big data over small data, there is Griffiths, RA and Dos Santos, M (2012) Trends in conservation also a move towards increasing employability of graduates and biology: Progress or procrastination in a new millennium? improving the impact of research. A mechanism by which this Biological Conservation 153: 153-158. can be achieved is through partnerships between Universities Harding, KA (1993) Conservation and the case of the golden and research-users, such as conservation agencies and zoos. toad. British Herpetological Society Bulletin 44: 31-34. Most conservation agencies do not have the resources they Kuhnert, PM, Martin, TG and Griffiths, SP (2010). A guide would like to pursue all of their goals, and student interns to eliciting and using expert knowledge in Bayesian can be an extremely cost-effective way of dealing with this. ecological models. Ecology Letters 13: 900–914. Indeed, it can be a ‘win-win’ situation – the student gains MacMillan, DC and Marshall, K (2006). The Delphi process – valuable work experience that will enhance the CV while an expert-based approach to ecological modelling in the organisation gains an extra pair of hands in the office or data-poor environments. Animal Conservation 9: 11-19. field at little extra cost. Furthermore, students carrying out Milner-Gulland, EJ, Fisher, M, Browne, S, Redford, KH, Spencer, dissertation projects within conservation organisations can M and Sutherland, WJ (2009) Do we need to develop a often collect badly needed small data that would otherwise more relevant conservation literature? Oryx 44: 1–2. not be considered a priority for a University supervisor. So McCoy, E (1994). “Amphibian decline”: a scientific dilemma conservation science seems to be moving in the right direction, in more ways than one. Herpetologica 50: 98-103. but perhaps not at the speed that conservation needs it to. Pan, Y, Wei, G, Cunningham, AA, Li, S, Chen, S, Milner- Gulland, EJ and Turvey ST (2015). Using local ecological References knowledge to assess the status of the Chinese giant Arlettaz, R, Schaub, M, Fournier, J, Reichlin, TS, Sierro, A, salamander Andrias davidianus in Guizhou Province, Watson, JEM and Braunisch, V (2010). From publications China. Oryx 50: 257-264. to public actions: when conservation biologists bridge the Pullin, AS, Knight, TM, Stone, DA and Charman, K (2004). gap between research and implementation. Bioscience Do conservation managers use scientific evidence to 60: 835-842. support their decision-making? Biological Conservation Austen, GA. Bindemann, M, Griffiths, RA and Roberts, DL 119: 245–252. (2016). Species identification by experts and non-experts: Sutherland, WJ, Fleishman, E, Mascia, MB, Pretty, J and comparing images from field guides. Scientific Reports Rudd, MA (2011). Methods for collaboratively identifying 6: 33634. DOI: 10.1038/srep33634. research priorities and emerging issues in science and Beebee, TJC (1993). Conservation of the golden toad. British policy. Methods in Ecology and Evolution 2: 238-247. Herpetological Society Bulletin 46: 28. Bonnet, X, Shine, R and Lourdais, O (2000). Taxonomic chauvinism. Trends in Ecology and Evolution 17: 1-3. Professor Richard Griffiths is Pro- Conservation Breeding Specialist Group (2016). http://www. fessor of Biological Conservation cbsg.org/our-approach/workshop-processes/phva- at the Durrell Institute of Conser- workshop-process (accessed 16 October 2016). vation and Ecology, University of Crump, ML, Hensley, FR and Clark, K (1992). Apparent decline Kent. He has had a long associa- of the golden toad: underground or extinct? Copeia tion with Durrell and in particular 1992: 413-420. with Durrell Conservation Acad- Evely, AC, Fazey, I, Lambin, X, Lambert, E, Allen, S and emy. Pinard, M (2010). Defining and evaluating the impact of cross-disciplinary conservation research. Environmental E-mail: [email protected] Conservation 37: 442-450.

Keep in touch

Durrell Conservation Academy Like our page, share our information and help spread the word so we can achieve our mission – saving species from extinction

Durrell Conservation Learning Network Keep in touch with other graduates, access useful information, comment, post & share links by joining this Alumni Group

Solitaire No. 27 7 Technical report Operant conditioning two captive Sumatran orang-utans to present for ultrasound scanning

Sarah Fowkes and Melissa Frost

Durrell Wildlife Conservation Trust

Abstract Using operant conditioning through positive reinforcement to train captive apes to participate in veterinary procedures is a low stress alternative to inspection via anaesthesia. Two pregnant female orang-utans at Durrell Wildlife Conservation Trust were trained using positive reinforcement techniques to present for voluntary ultrasound scanning to monitor foetal development. The two females differed in their response to training, but both were successfully monitored during pregnancy and delivered healthy infants. Key words: Pongo abelii, positive reinforcement, pregnancy, sonograph, training

Introduction Methods There is a growing trend in zoological, veterinary and labo- Subjects ratory communities to recognise the value of using operant Operant conditioning was used to train two pregnant female conditioning as an animal care and management tool. The Sumatran orang-utans housed at Durrell Wildlife Conserva- use of operant conditioning to train non-human primates to tion Trust, Dana (25 years old) and Annette (30 years old) to co-operate with routine scientific, husbandry, and veteri- present their abdomens for ultra sound scanning. nary procedures is recommended by many legislative and professional guidelines (Home Office 1989; International Training area Primatological Society 1989). Training is recommended as Training was conducted in the exit tunnels of the exhibi- it has been shown to reduce fear, anxiety and distress caused tion areas of the orang-utan building as the mesh was large by traditional captive management procedures such as anaes- enough to pass an ultrasound probe through and subjects thesia (Reinhardt et al. 1995). could be easily isolated (see Fig. 1). In positive reinforcement based training, animals are rewarded with something they like for responding to a caregiver’s cue or command (Laule et al. 2003). The positive reinforcement approach to non-human primate captive management is now used in zoological communities to increase primate welfare by successfully supporting the cooperation of captive primates in animal transfers (Bloomsmith et al. 1998) and husbandry procedures, including veterinary interventions, while improving the rapport between keeper and animal (Savastano et al. 2003). Colahan and Breeder (2003) trained captive primate populations at Disney Animal Kingdom to display a wide range of behaviour using positive reinforcement to improve animal care, veterinary inspection and infant care. For example, positive reinforcement was used to train pregnant females for separation from groups for inspection, breast manipulation for lactation assessment and infant pick-up, present and holding behaviours. This paper discusses how two pregnant female orang-utans at Durrell Wildlife Conservation Trust were trained to present voluntarily for ultrasound scanning to monitor foetal development using positive reinforcement techniques without the need for anaesthesia. Figure1. Annette at training station in exit tunnel.

Solitaire No. 27 8 Orangutan ultrasound training

Positive reinforcement A training plan was first devised to outline the steps nec- essary to condition the subjects to present for ultrasound scanning. A positive reinforcement technique was then used to complete these steps. Conditioned commands, for exam- ple “touch”, were given by the keeper trainer. If this was followed by a conditioned response by the orang-utans, such as allowing the trainer to touch the abdomen, then a reward of a favoured food item was given. Once the conditioned responses became habituated, using the bridge phrase “good girl” was enough to sustain the behaviour and a reward was given only when the behaviour had been completed (see Fig. 2). Once one command had been learnt the next was intro- duced, resulting in the building of behaviours necessary for ultrasound scanning. All training sessions were recorded in daily reports.

Materials A jackpot reward of three grapes was given when a new be- haviour was first displayed. A reward of a single grape was then given when the behaviour was repeated. Grapes were used as a reward as they are a favoured food item that is not part of the orang-utans’ daily diet. A peanut was used ini- tially but it was too highly prized a reward, resulting in the orang-utan becoming impatient and in turn frustrated during training sessions. The grapes were originally put in a yo- ghurt pot in sight of the orang-utan but this caused Dana and Annette to become quickly distracted. Therefore grapes were instead placed in a pouch attached to the trainer’s belt so that they were out of view, and this resulted in much higher levels of concentration. A fake ultrasound machine, constructed from card board boxes (see Fig. 3), was used to desensitise Dana to novel stimuli, while latex gloves and a damp cloth were used to desensitise her to novel textures. Annette did not have the same suspicion of novel stimuli so these items were not necessary for her training. Portable Sonosite Micromax© and Sonosite 180 Plus© scanners (see Fig. 3) were used to conduct the sonographs. Clear Image© medical ultrasound gel was used as a conductive medium between the skin and the ultrasound probes. Dana did not like the sensation of the gel but preferred medical gel to the Sonogel© that is typically used for veterinary scanning. As Dana was adverse to the gel, a trans-vaginal probe was initially used externally to conduct ultrasound scans with both subjects, as it required only a small amount of gel. A larger probe head with a depth of 22 cm was used later, once Dana was desensitised to the gel, as the trans-vaginal probe’s depth of only 10cm was not sufficient to gain clear images of Dana’s developing foetus. Non-invasive scanning enabled us to monitor the pregnancies carefully, and was particularly beneficial in Dana’s case, as she was considered at high risk because of a previous haemorrhage following a stillbirth (Maclachlan et al. in press).

CC → CR → BRIDGE → R “touch”→ present stomach→ “good girl”→ grape

Figure 2. Positive reinforcement used to train orang-tans to Figure 3. (Above) Fake ultrasound machine, and (below) present for ultrasound training. Sono 180 plus© portable ultrasound machine.

Solitaire No. 27 9 Technical report

Table 1. Training steps for each female orang-utan. Dana Annette Time taken Time taken to complete to complete Step step step 1: Present 1 session If Dana put her stomach close to the 3 sessions If Annette put her stomach close to stomach mesh and allowed the keeper to touch it the mesh and allowed the keeper to with her hand the command ‘touch’ was touch it with her hand the command given, followed by a reward. ‘touch’ given, followed by a reward. 2: Tolerate a 1 session If Dana tolerated the fake probe being Annette had no fear of novel stimuli fake probe pressed on her stomach when she pre- so this step was skipped. sented it the ‘touch’ command was given, followed by a reward. 3: Desensi- 1 session A latex glove was placed over the fake Annette had no fear of novel textures tise to novel ultrasound probe with ultrasound gel so this step was skipped. textures underneath to desensitise Dana to novel textures. If Dana tolerated the fake probe with the glove on the end of it being pressed on her stomach when she pre- sented it the ‘touch’ command was given, followed by a reward. 4: Hold and 4 sessions Dana given touch command to present 5 sessions Annette given touch command bridge her stomach and allow the fake probe to to present her stomach and al- be pressed against her stomach. Probe low the fake probe to be pressed held on her stomach and Dana given a against her stomach. Probe held on ‘hold’ command. If Dana held her stom- her stomach and Annette given a ach in position for < 45 seconds without ‘hold’ command. If Annette held her grabbing the probe then she was given stomach in position for< 45 seconds a reward. The bridge word ‘good girl was without grabbing the probe then she used to sustain the hold behaviour until the was given a jackpot reward. The reward was given bridge word ‘good girl was used to sustain the hold behaviour until the reward was given 5: Desensi- 1 session Dana rewarded for tolerating the presence 1 session Annette rewarded for tolerating the tise to novel of veterinary staff at training sessions presence of veterinary staff at train- people ing sessions 6: Desensi- 2 sessions Fake ultrasound probe covered with a Annette had no fear of novel tem- tise to novel warm damp cloth to mimic the tempera- peratures so this step was skipped. temps ture of ultrasound gel and Dana rewarded for presenting and holding her stomach on command and allowing keeper to touch her stomach with the cloth covered probe. This was then repeated with gel behind the cloth. 7: Introduce 5 sessions Dana rewarded for presenting and holding 1 session Annette rewarded for presenting and probe on command for a novel trans vagianl holding on command for a novel probe trans vaginal probe 8: Introduce 5 sessions Ultrasound gel put on end of probe. Dana 1 session Ultrasound gel put on end of probe. ultrasound rewarded for presenting and holding her Annette rewarded for presenting and gel stomach on command and allowing the holding her stomach on command keeper to touch her stomach with the gel and allowing the keeper to touch her covered probe for a successful ultrasound stomach with the gel covered probe scan for a successful ultrasound scan 9: Larger 1 session Dana rewarded for presenting and holding Larger probe not necessary for An- probe intro- her stomach on command and allowing nette duced the keeper to touch her stomach with the larger gel covered probe for a successful ultrasound scan.

Conclusion voluntary ultrasound scan, while Annette took only 11 ses- Operant conditioning was successfully used to train two sions. Similar individual differences have been found at pregnant female orang-utans for ultrasound scanning. The Durrell using a similar approach with tamarins (Brayshaw time investment devoted to train naïve animals was minimal et al. 2015). Annette has a very trusting nature and enjoyed but was affected by individual differences in the orang-utan’s the interaction with staff members, and so she did not have temperament. Dana took 20 training session to complete a to be desensitised to novel stimuli and participated in training

Solitaire No. 27 10 Orangutan ultrasound training

sessions with enthusiasm. In comparison Dana has a history References of medical interventions by keepers and veterinary staff, re- Bloomsmith, MA, Stone, AM and Laule, GE (1998). Positive sulting in suspicion of novel stimuli and people. Dana also reinforcement training to enhance the voluntary did not like the sensation of the ultrasound gel. Therefore movement of group-housed chimpanzees within their Dana’s training involved more steps to desensitise her to enclosures. Zoo Biology 17: 33–341. these stimuli. Brayshaw, SK, Lopez FJ and Wormell, D (2015). Ultrasound We believe that the long-term welfare gains of monitoring training for pregnancy monitoring in tamarins. Solitaire the orang-utans’ pregnancies and the ability to conduct 26: 30–33. necessary ultrasound scanning in a low stress manner, out- Colahan, H., and Breder, C. (2003). Primate training at weigh the initial costs in time to train the subjects. The Disney’s Animal Kingdom. Journal of Applied Animal long-term welfare advantages of increased ease of veterinary Welfare Science 6: 235–246. assessment and treatment and the more subtle benefits of Home Office (1989). Code of Practice for the Housing and voluntary participation in management procedures and the Care of Animals Used in Scientific Procedures. London: building of trust between animals and care givers are vast. Her Majesty’s Stationery Office. It should be noted that a key factor in the success of the International Primatological Society (1989). IPS International training was consistency. A primary keeper trainer taught all guidelines for the acquisition, care and breeding of the behaviours necessary for ultrasound scanning. A training nonhuman primates. Primate Report 25: 3–27. plan was designed with each step outlined prior to interaction Laule, GE, Bloomsmith, MA and Schapiro, SJ (2003). The use with the animals. Furthermore all training session were of positive reinforcement training techniques to enhance recorded in daily reports to keep track and to avoid missing and care, management, and welfare of primates in the or skipping training steps. These precautions were taken to laboratory. Journal of Applied Animal Welfare Science reduce confusion and frustration for the orang-utans and to 6: 163–173. ensure training sessions were always a positive experience. Maclachlan, N, Hunt, G, Fowkes, S, Frost, M, Miller, J, Purcell- The training steps could be used and are recommending Jones, G, Sullivan, P, Barbon, A, Routh A., López FJ and for ultrasound scanning future pregnant orang-utans and other Price EC (in press). Successful treatment of infertility in a species. However using a clicker as a bridge instead of the female Sumatran orangutan Pongo abelii. Zoo Biology. ‘good girl’ vocal cue would be more effective as it would be Reinhardt, V, Liss, C and Stevens, C (1995). Restraint methods consistent across keepers. More training for the keepers on of laboratory nonhuman primates: A critical review. how to position and manipulate the scanning probe to gain Animal Welfare 4: 221–238. the best ultrasound images would also be of an advantage. Savastano, G, Hanson, A and McCann, C (2003). The In conclusion, operant conditioning, through positive development of an operant conditioning training reinforcement, can be used to train orang-utans to present program for New World primates at the Bronx Zoo. Journal for ultrasound scanning, increasing the wellbeing by of Applied Animal Welfare Science 6: 247–261. reducing stress induced by veterinary procedures. Potential disadvantages of training are the lack of consistency Sarah Fowkes formerly worked at Durrell Wildlife between trainers and individual differences in subjects. The Park, specialising in the care of apes and Old World predominant cost of training is the time investment required to monkeys. She is now at Bristol Zoo. Mel Frost was a change behavioural repertoires. However the welfare benefits veterinary nurse at Durrell Wildlife Park at the time of of positive reinforcement far outweigh these costs. this study. E-mail: [email protected] As dead as a dodo – when do we declare a species Extinct?

Rosalind Kennerley

Durrell Wildlife Conservation Trust; IUCN SSC Small Mammal Specialist Group (SMSG)

There is no doubt that some species, such as the Dodo, sadly cialist Group (SMSG)’s latest Red List reassessments that no longer exist. However, in some cases it is not always sim- demonstrate some of the challenges involved. By December ple to come to a decision about whether the last individual of 2016 the SMSG had reassessed over two thousand species a species has truly disappeared or not. This article considers and assessed new species for the Red List. This updating of how the IUCN Red List of Threatened Species defines and the Red List provides regular checks for species, which are applies the Extinct category, using a number of interesting crucial to our understanding of how the world’s wildlife is examples that have come out of the Small Mammal Spe- changing and for setting conservation priorities.

Solitaire No. 27 11 Essay

The IUCN Standards and Petitions Subcommittee (2016) first stage of producing this strategy will be to take stock of acknowledges that extinction is very difficult to determine. the current situation to identify the species which are already In the guidance provided by the IUCN, a species should be in captivity, and the purpose they fulfil within the collections. considered Extinct when there is no reasonable doubt that The second phase will then be to conduct a systematic needs the last individual has gone. In order to presume extinction assessment for ex-situ and in-situ captive breeding for SMSG there should be evidence from exhaustive surveys in suitable species, using a series of questions and criteria to identify habitat, which should be undertaken at appropriate times goals, threats and conservation options. The SMSG intends (diurnal, seasonal, annual), and throughout its historic range. to have these prepared in 2017, with the aim of identifying Additionally, surveys should take place over a time frame species at most risk in the wild which are in need of immediate appropriate to its life cycles and life form. rescue to avoid extinction. So, what constitutes sufficient survey effort? In the There are inevitably a few situations where we have following example it appeared to be a clear cut case. Most confirmation that the Extinct category was incorrectly people will not have heard of the Bramble Cay melomys assigned. An unusual case arose recently, which involved the (Melomys rubicola), which became the first mammal likely Machu Picchu arboreal chinchilla rat (Cuscomys oblativa). to have been wiped out by severe storms and rising sea level, Since the beginning of the Red List in 1982, this species has events which were probably exacerbated by human-induced been considered Extinct, with the justification that it was climate change. It was the only mammal species endemic to described from just two skulls collected at an Inca burial site Australia’s Great Barrier Reef and resident on one tiny island (Pacheco et al. 2008). These specimens were estimated to be with an area of just five hectares. The severe weather and around 400 years old and of unknown origin (Pacheco et al. rising sea level caused dramatic habitat loss, making the cay 2008). This particular species featured very low down (number gradually less inhabitable. After sightings became increasingly 77) on a list of 90 animals with fewer than five total sightings rare, researchers set out to capture any remaining individuals ranked in order of their probability of rediscovery (Fisher and start a captive breeding programme. The rescue mission and Blomberg 2012). Remarkably, in 2009 C. oblativa was to such a remote location required significant planning, and photographed and rediscovered. Finding living examples of upon reaching the island five months later the team found no taxa that were previously considered to be extinct represents trace of the species despite targeted surveys. Given the small a special case of the “Lazarus effect” and this has only size of the island, the fact that almost no vegetation remained, been noted a handful of times among mammals and other and the level of surveying, it was a straightforward decision vertebrates (Dawson et al. 2006). This term is usually used to reclassify the Bramble Cay melomys from Critically to describe the reappearance of taxa after a long absence in Endangered in 2008 (Leary et al. 2008) to Extinct (Woinarski the fossil record (Fara 2001; Jablonski 1986). Interestingly, and Burbidge 2016b) in the most recent Red List reassessment. concerns had been raised previously that there was doubt over However, questions were raised about this decision in the whether or not C. oblativa could still be found there, with discussions in Gynther et al. (2016), who proposed that the Emmons (1999) suggesting that there was no actual evidence species or a close relative may still exist in the Fly River delta that it had become extinct since 1500 and that it was likely to region on the island of New Guinea. Whilst it is true that be still extant. Despite these ideas, Pacheco et al. (2008) stated New Guinea warrants further comprehensive small mammal that the species was not detected during various surveys of the inventories, because it is only a hypothesis that this particular region. The current situation is that very little is known about species may have originated from there, this is not sufficient the species and as a result it is now listed as Data Deficient reason to prevent the use of the Extinct category according (Roach 2016). to the IUCN definition. Minimising subjectivity in Red List accounts is vital, as This unfortunate example of the Bramble Cay melomys there will always be some differences of opinion about how extinction, which could have been avoidable, provides the to apply the guidelines. Indeed, this is reflected by the fact specialist group with a timely reminder about the urgency that guidelines for interpretation of the standards of evidence to have a captive breeding strategy in place for species supportive of Extinct and Critically Endangered (Possibly represented by the group and also that the threat of climate Extinct) classifications are currently under development change needs to be taken into account within the strategy. The (IUCN Standards and Petitions Subcommittee 2016). Such guidelines would limit inconsistencies. In the example of the lesser stick-nest rat (Leporillus apicalis), there has been a lack of consensus across the years between assessors. It was listed as Extinct in all previous assessments, except in 2008, when it was moved to Critically Endangered by Robinson and Burbidge (2008). It was then declared Extinct again in the most recent reassessment (Woinarski and Burbidge 2016a). The switches between categories result from differences in the interpretation of the guidelines and also a paucity of data. The species is considered to have severely declined, predominantly due to predation by feral cats (Felis catus), though other threats may have contributed to its decline, for example predation by red foxes (Vulpes vulpes) and habitat degradation caused by The Bramble Cay melomys, now extinct as a result of rising introduced herbivores, in particular during severe droughts sea levels (Photo: Queensland Government, Creative Com- (Woinarski and Burbidge 2016a). The last two specimens of mons Attribution 3.0 Australia (CC BY) licence. lesser stick nest rat were collected near Mt Crombie, south

Solitaire No. 27 12 Defining extinction

of the Musgrave Ranges in north-western South Australia in of the Extinction of the Bramble Cay melomys Melomys 1933. Since then, there have been a couple of unconfirmed rubicola on Bramble Cay, Torres Strait: Results and records, such as one from 1970 and occasional reports of fresh Conclusions from a Comprehensive Survey in August– vegetation being added to old stick-nests. It seems likely that September 2014. Unpublished report. Brisbane: a species which builds such obvious nests would have been Department of Environment and Heritage Protection, found, if still extant, but it is worth noting that much of this Queensland Government. species’ range is in remote portions of central Australia that IUCN Standards and Petitions Subcommittee (2016). have not been fully surveyed. Guidelines for Using the IUCN Red List Categories and In conclusion, there will always be a degree of subjectivity Criteria. Version 12. Prepared by the Standards and about how and when to declare a species officially Extinct Petitions Subcommittee. Available at: http://www. on the Red List. There are likely to be knock-on effects for iucnredlist.org/documents/RedListGuidelines.pdf conservation of declaring a species Extinct. The “Romeo Leary, T, Singadan, R, Menzies, J, Wright, D and Thomson, B Error” (Collar 1998) was a term coined to describe when (2008). Melomys rubicola. The IUCN Red List of Threatened funding and protective measures are removed from threatened Species 2008: e.T13132A3412635. species in the mistaken belief that they are already extinct. Pacheco, V, Zeballos, Z, Vivar E and Dunnum, J (2008). An evidence-based approach to classifying extinctions is Cuscomys oblativa. The IUCN Red List of Threatened required to encourage continued conservation work until Species 2008: e.T136658A4324252. there is no reasonable doubt that the last individual of a Roach, N (2016). Cuscomys oblativa. The IUCN Red List of species has died. On the other hand, if assessments are too Threatened Species 2016: e.T136658A22182152. evidentiary, then extinction rates based on the Red List risk Robinson, T and Burbidge, A (2008). Leporillus apicalis. The IUCN being underestimates. Red List of Threatened Species 2008: e.T11633A3298296 Woinarski, J and Burbidge, AA (2016a). Leporillus apicalis. References The IUCN Red List of Threatened Species 2016: Collar, NJ (1998). Extinction by assumption: or, the Romeo e.T11633A22457421. Error on Cebu. Oryx 32: 239-244. Woinarski, J and Burbidge, AA (2016b). Melomys rubicola. Emmons, LH (1999). A new genus and species of abrocomid The IUCN Red List of Threatened Species 2016: rodent from Peru (Rodentia, Abrocomidae). American e.T13132A97448475. Museum Novitates 3279: 14. Dawson, MR, Marivaux, L, Li CK, Beard, KC and Métais, G (2006). Laonastes and the” Lazarus effect” in recent Ros Kennerley has been involved in mammals. Science 311:1456-1458. a number of threatened bird and Fara, E (2001). What are Lazarus taxa? Geographical Journal mammal programmes around 36: 291-303. the world. She began her role as Fisher, DO and Blomberg, SP (2012). Inferring extinction of Programme Office for the SMSG mammals from sighting records, threats, and biological based at Durrell in April 2015. Prior traits. Conservation Biology 26: 57-67. to this she completed her PhD on Jablonski, D (1986). Causes and consequences of mass the conservation of two endemic extinctions: a comparative approach. In: Elliott, DK (ed.), mammals in the Dominican Re- Dynamics of Extinction. New York: Wiley-Interscience: public, as part of the Durrell-led 183–229. Darwin Initiative funded Last Survivors Project. Gynther, I, Waller, N and Leung, LK-P (2016). Confirmation E-mail: [email protected] Conservation and data management

Finella Blair

Durrell Wildlife Conservation Trust

A major component of running effective programmes for saving endangered species, in or ex situ, is the and the Caribbean, grasslands in India, offices in Bath monitoring of individuals and populations using and at the zoo in Jersey. These data start life mostly evidence-based scientific research to inform actions in notebooks, but also on camera storage cards, and evaluate the impact of programmes. This is spreadsheets and in the heads of the researchers, something Durrell is striving to improve both in our zoo managers and field workers. and overseas research programmes. Why is data management important? These efforts result in an abundance of data generated Data are valuable; in conservation organisations they in programmes all over the world, on islands in Mauritius are often the result of long hours in the field in difficult

Solitaire No. 27 13 Technical report

Excel is an excellent tool for many tasks and is often thought of as a “database”. It is intuitive and is the software package of choice for many researchers as it has some analysis functionality. Although it is great for things like pivot tables and charting, it can be very difficult to sort and report on relational data, particularly when data are spread between sheets and different files. Over time, inconsistency of data entry, particularly when more than one person is entering the data, and sheer volume makes querying or reporting the data more and more time consuming. At this point, a relational database such as MS Access is often a much Figure 1. Illustration of the natural loss of information about more appropriate tool. datasets (metadata) with time (from Michener et al., 1987). Part of building an effective relational database is understanding how the data are collected and recorded, the logistical difficulties of moving it to conditions and, in the case of critically endangered electronic format (adverse weather, lack of internet species, in situations that cannot be repeated if data connection or electricity), how best to collate and store are lost. Good data management is often common it and make sure that it can be reported effectively. As sense but can be given low priority when there are so a result, I spent part of my time in Mauritius visiting Ile many other demands on everyone’s time. aux Fouquets, Ile aux Aigrettes and Round Island with MWF staff, to take part in and understand as many type Good data management practice, when incorporated of surveys as possible. with technological solutions such as data entry forms on tablets and phones, can also save time, something As part of the reptile project I field tested the use of data anybody who has had to re-enter all their data because collection forms on smartphone and tablets for different they didn’t have a back-up will identify with. types of surveys. The forms are created using MS Excel and Open Data Kit (ODK), free open source software What counts as data? that can be used with a Google app (see Figures 2 and Data can be many things, including notebooks full 3). Correctly set up, forms can be context sensitive, data of surveys, spreadsheets of information, photograph consistency can be improved by selecting answers from collections (digital and analogue), sound recordings drop-down boxes and data is saved immediately and and information about data, known as ‘metadata’. automatically sent to a central internet repository as A lot of incredibly valuable data is often just stored soon as an internet connection becomes available. in people’s heads and only missed when they leave! The data can then be downloaded and imported into Without putting too fine a point on it, if you were a central database or further analysed in programs (unfortunately) hit by a bus tomorrow, would anybody such as ‘R’. Directly entering data onto a tablet in the tasked with carrying on your work where you left off know field cuts out the need to transfer data from notebook where your data was, how it was captured and what into an Excel spreadsheet at a later date, a process it represents? The ‘Information Entropy’ diagram (Fig. that can take a long time, opens up opportunities for 1) illustrates the potential hazards of not documenting typographic errors and creates delays in analysis. The your data properly. time from field to central database can be reduced from days or weeks to hours or even minutes. Good data management will ensure that valuable and hard-earned data will be preserved for future use to aid Field testing is in its early stages in Mauritius. The use the endangered species that we care about so much. of ODK forms instead of notebook and pencil for It should also ensure that researchers get the credit for Capture-Mark-Recapture surveys of Bojer’s skinks has the data they collect. Increasingly, scientific journals the potential to speed up the data handling process as are beginning to insist that authors provide access to data are completed and sent while still on the island, the datasets on which their papers are based, generally rather than waiting for access to a computer and MS by assigning Digital Object Identifiers (DOI). Anybody Excel back at the office. Feedback from researchers in asking to reuse the data must then ask permission and the field indicates that further modification and testing correctly cite the dataset and its originators in any is required. papers they publish. Ideally ODK forms are used during the survey, replacing Managing data in Mauritius notebook and pencil, but in some circumstances, for I was very fortunate to spend two months in Mauritius example in distance sampling where recordings are in 2016, helping Dr Nik Cole review a large volume of rapid, the screen refreshing required by ODK forms may historical reptile survey data, which was stored in MS slow down data entry, so ODK forms can be used for Excel. data entry immediately after the survey is complete.

Solitaire No. 27 14 Managing conservation data

Back on the mainland, I designed and built a database structure in MS Access to incorporate the reptile survey data collected on both the new forms and the historical data. It’s important to ensure that the structure can accommodate all sorts of reporting requests and one of the skills of a good developer is to try and anticipate these. The reptile database will enable faster reporting of reptile survey data across species, islands and survey methods. MS Access has powerful querying and sorting capabilities and data reported can be exported if necessary, to documents, spreadsheets or ‘R’ for further analysis.

Back in the Bath office we are building several databases, some of which will support the development of the Durrell Index. We now have a publication database of Durrell staff-authored research papers, books and book chapters, academic theses, action plans and grey literature. A species database, for any animal species associated with Durrell, past or present, has been set up and recently a plant section has been added to support work on endemic and invasive plants in Mauritius. Work has just started to further develop an Academy database which will link Academy course Figure 2. ODK Main menu on smartphone. attendance with the Alumni Network. A new intern at the Bath office, Rachael Gerrie, is working on the data that Durrell need to collect to assess the impact of Durrell Conservation Academy for the Durrell Index and this will be linked to the Academy database in due course.

The data management team in Bath, headed by Lianne Concannon, will be producing some information on guidelines for good data management over the next few months. In the meantime, I have included a few pointers below. Most are common sense and are probably already part of your routine; there are also some really good resources on the internet, the best of which are shown below.

Figure 3. Example of ODK data collection form showing drop-down list. Figure 4. MS Access database design.

Solitaire No. 27 15 Technical report

If you can identify with any of the following situations, Sorting Be careful using sort in Excel. Sorting can go then following some straightforward data management very wrong, for example when there is a blank row guidelines could help you. somewhere within your data.

Writing a report: all the data you need is in different Online resources formats or spreadsheets or locations (or a combination Data management: https://www.dataone.org/sites/all/ of all three)? documents/L01_DataManagement_Handout_FINAL. pdf Field notebooks: the precious field notebook with all its data has been lost or damaged? Data sharing: https://www.dataone.org/sites/all/ documents/L02_DataSharing_Handout_FINAL.pdf Laptop or hard drive: a component fails just as an important report is required? Best practices for spreadsheets and data files: https://www.dataone.org/sites/all/documents/L04_ Excel spreadsheets: Everything was set up beautifully DataEntryManipulation_Handout_FINAL.pdf at the beginning of the surveys but years have passed and the dataset is huge and extracting the Data quality control: https://www.dataone.org/sites/ information needed for reporting is either difficult or all/documents/L05_DataQualityControlAssurance_ impossible? Handout_FINAL.pdf

Backups Backups and archives: https://www.dataone.org/sites/ All data should be backed up, but the more of a disaster all/documents/L06_DataProtectionBackups_Handout_ it would be if you lost it, the more important it is that you FINAL.pdf back it up, preferably to two different locations. Metadata – documentation of your data: https:// Field notebooks www.dataone.org/sites/all/documents/L07_Metadata_ Notebooks are very vulnerable to loss and damage. If Handout_FINAL.pdf you have field notebooks full of data that haven’t been transcribed to MS Excel yet, then scan or photograph each page. It might take half an hour for each Reference notebook but it is then safe. Use a tablet or smartphone to take a photo of each notebook page at the end of Michener, W.K., Feller, R.J. and Edwards, D.G. (1987). each survey then email it to yourself for a permanent Development, management, and analysis of a long- record. term ecological research information base: example for marine macrobenthos. In Boyle, T.P. (ed.), New Approaches Laptops to Monitoring Aquatic Ecosystems. Philadelphia: ASTM Back up your important files to an external drive or thumb International, 173–188. drive or a central facility like Sharepoint or Dropbox if you have it. Make sure that the backups are labelled with the date of backup and are sensibly named.

While Dropbox can be extremely frustrating where connection speeds are poor, having important files in Finella Blair has been working part- Dropbox (preferably not shared to prevent anyone else time for Durrell since October 2015, inadvertently deleting it) at least ensures that you can after completing her Masters in Con- retrieve previous versions of files. servation Biology at DICE, University of Kent. Before that, she spent many File names years building relational databases It is really easy to end up with multiple copies of files and managing data, something she where it is difficult to tell which is most up to date, so also does part time for a NERC-funded incorporate version numbers and dates when making tropical forest programme at the University of Oxford. back-up copies. E.g. reptile_database_v1.2_220716 As Data Management Intern for Durrell, she is working with Lianne Concannon at Durrell’s Conservation Sci- MS Excel files ence hub in Bath to facilitate the flow of data from Master files If you are working constantly on a file (e.g. the point of data collection in the field up to the cleaning data) take a copy of the file and label it with global headline indicators presented on the Durrell the date at the end of each work session. Then if you Index, Durrell’s organisational monitoring and evalua- find that you have made a mistake (e.g. on a sort) you tion framework, and has set up databases to support can track back through your copies to see where it reporting, data analysis and decision making. happened and you don’t have to go back to the last backup which may be many sessions out of date. E-mail: [email protected]

Solitaire No. 27 16 Health of Hapalemur alaotrensis Morbidity of Alaotran gentle lemurs (Hapalemur alaotrensis) at Durrell Wildlife Park (1990–2014)

Alberto R. Barbon, Zoavina Randriana and Gale Glendewar

Durrell Wildlife Conservation Trust

Abstract This study presents a preliminary overview of medical problems encountered in captive Alaotran gentle lemurs at Durrell Wildlife Park over a period of 24 years. Individual medical records for 65 lemurs (36 males and 23 females) were examined. Overall, gastrointestinal and integumentary problems were the most prevalent issues, but age and sex were not good predictors of the occurrence of medical problems. Additional research on the nutrition and behaviour of captive specimens may help to understand and reduce the incidence of some of these pathologies. Key words: diet, disease, medical problems, veterinary care

Introduction 1, while the current diet is shown in Table 2. Variations in the The Alaotran gentle lemur (AGL) or bandro (Hapalemur diet provided in different enclosure have been established alaotrensis) is one of the five bamboo lemur species that given the differing availability of forage in the outside areas. exist in Madagascar (Mittermeier et al. 2010). It is a small bodied lemur (1240 ±140 g), characterised by its highly spe- Results cialised feeding patterns (Mutschler et al. 1998). The AGL is Medical records from 65 AGL (36 males and 23 females) found in the marsh habitat surrounding Lac Alaotra, the larg- were evaluated. Seven different clinicians were involved in est lake in central Madagascar (Mutschler & Feistner, 1995) the description of the medical problems. (Figure 1). Its limited geographical range, combined with Six individuals were stillborn and only post mortem increasing anthropogenic pressure, has led to a population information was available. Of the remaining 59 individuals, 12 decline of more than 80% over a period of 24 years, and it is currently classified as critically endangered with a declin- ing trend (Andriaholinirina et al. 2014) The remaining wild population is estimated to comprise around 2500 individuals (Ratsimabazafy, pers. comm., 2013). A small population is managed in zoological collections. Ninety-nine individuals (40.44.15) are maintained in 24 institutions worldwide (T. Wright and G. Glendewar, unpublished data). From a founding population of 10 animals, in the period to 2014, Durrell Wildlife Park (DWP) in Jersey had held 65 individuals. The aim of this study was to establish the most common medical problems found in the AGL housed at the DWP over a period of 24 years (1990 to 2014). Methods AGL at DWP have been housed in a wide range of enclosures. (Figure 2). At the time of this study three different types of enclosures were being used. The oldest enclosures consisted of an outside planted area (38.5 m2) and an indoor heated area with vertical and diagonal bamboo poles. Another enclosure built in 1998 included a very large outside area of around 800 m2, with planted bamboo and willow trees and a heated shed. The newest enclosure consists of a heated shed and planted outside area with additional vertical bamboo poles. AGL have usually been housed in pairs and families in single species enclosures. Diet has changed over time; a description of the diet in 1995 (Courts 1995) is given in Table Figure 1. Alaotran gentle lemurs in their natural habitat.

Solitaire No. 27 17 Research report

Table 1. Diet described in 1995 for captive gentle lemurs (Courts 1995). 0900 1 small (approx. 5 g) piece of: apple, banana, 2 other fruits eg. pear, plum; 2 small pieces of carrot, 2 soaked Old World Monkey pellets (Special Diet Services) 1200 Forage: bamboo stem (preferably with shoots), reeds, grass or willow 1600 Fresh vegetables/salad, eg.: 2 small (2–4 g) pieces of lettuce heart, celery, cucumber, tomato, carrot plus half a hard-boiled egg (no shell) twice per week Supplements 0.5 g Vionate (Ciba-Geigy Animal Health). Diluted condensed milk given to lactating females

either died naturally or were euthanased for medical reasons, and 39 were exported to other institutions. Eight individuals were still housed at DWP at the time of the study. In total, 78 individual medical problems were identified (Table 3). Figures 3 to 7 illustrate some of the pathologies observed. Prevalence of medical problems within the eight body systems by sex is presented in Table 4. Approximately half of the AGL population (49.1%) housed at Durrell from 1990 to 2014 had some medical problems, and the prevalence Figure 2. Current enclosures for AGL at DWP. was slightly higher in females (52%) than in males (47.2%), but this difference was not statistically significant (F=0.78; p>0.05). Gastrointestinal and integumentary pathologies showed the highest prevalence, with diarrhoea and wounds being the most common problems across the whole population. Prevalence of medical problems within body systems by age is presented in Table 5. Most of the medical problems were diagnosed in the adult group. The juvenile group showed

Table 2. Diet from 2002 to present. Daily extras are no longer offered in the current diet iteration, but they were offered in the past.

0900 Approx. 15 g dry leaf-eater pellet, 2 pieces (approx. 5 g) carrot. Added Vionate supple- Figure 3. Digit injury exposing distal phalanx. ment. Forage Available through the day. Bamboo (Phyl- feed lostachys sp), willow (Salix sp), perennial grass (Miscanthus sinensis), sedges (Carex hachi- joensis, Carex pendula) Chicory (Cichorium sp). Bamboo is offered through the year; not all the plants listed are available in every enclosure, which has led to diet variations depending on the enclosure. Afternoon Chicory, ¼ tomato, 3 small pieces celery, 2 feed small pieces cucumber, 2 small pieces carrot. Fruit (including apple, pear, banana, grapes, nectarine, plum, pomegranate, melon, rasp- berry) only occasionally. Supple- 1 ml calcium vitamin D supplement (Collo-cal ments D) administered in bread for 6 months to nurs- ing mothers Daily Bread (Monday), ¼ boiled egg (Tuesday), ¼ extras boiled egg (Thursday), small piece cooked Figure 4. Severe dental disease, showing gingivitis, tartar sweet potato (Friday), small piece cooked sweet potato (Sunday). accumulation, and gingival recession.

Solitaire No. 27 18 Health of Hapalemur alaotrensis

Figure 5. Axillary glands in gentle lemurs exhibiting normal secretions; these glands are bilateral and present a nodu- lar appearance, not to be mistaken with subcutaneous abscesses.

Figure 7. Localised alopecia in the base of the tail, suspect- ed to be caused by prolonged pressure in this area due to Figure 6. Subcutaneous liposarcoma. specimens resting position next to heaters in indoors areas.

Table 3. Individual medical problems diagnosed in AGL, classified by system, showing the total number of cases and the percentage of the total number of medical problems identified.

Body system (number of cases) Medical problems Number of cases Percentage of total Dental (9) Dental abscess 6 7.7 Gingivitis 1 1.3 Maxillary localised swelling 2 2.5 Gastrointestinal (18) Constipation 1 1.3 Diarrhoea 14 18 Vomiting 3 3.8 Integumentary (30) Abscess 2 2.5 Cyst 1 1.3 Liposarcoma 2 2.5 Localised alopecia 4 5.1 Localised swelling 4 5.1 Wound 17 21.8 Metabolic (6) Obesity 6 7.7 Musculoskeletal (5) Lameness 3 3.8 Fracture 1 1.3 Vertebral spondylosis 1 1.3 Reproductive (2) Miscarriage 1 1.3 Leydig interstitial cell tumour 1 1.3 Sensory (2) Cataract 1 1.3 Corneal ulcer 1 1.3 Urinary (6) Renal cysts 1 1.3 Renal failure 5 6.4

Solitaire No. 27 19 Research report

Table 4. Prevalence of medical problems (%) within Table 5. Prevalence of medical problems (%) within body systems by sex of captive Alaotran gentle lemurs body system by age (years) of captive Alaotran gentle (Hapalemur alaotrensis) housed at Durrell Wildlife Park, lemurs (Hapalemur alaotrensis) housed at the Durrell Jersey (1990–2014). Wildlife Park, Jersey (1990–2014).

Body system Male Female Durrell population Body system Infant Juvenile Adult Geriatric (n=36) (n=23) (n=59) (0–0.5) (>0.5–2) (>2–16) (> 16) Dental 16.6 13 15.2 Dental 0 1.7 11.9 1.7 Gastrointestinal 19.4 47.8 30.5 Gastrointestinal 0 0 23.7 6.7 Integumentary 61.1 34.8 50.8 Integumentary 1.7 6.7 37.2 5.1 Metabolic 11.1 8.7 10.2 Metabolic 0 5.1 5.1 0 Musculoskeletal 8.3 8.7 8.5 Musculoskeletal 0 1.7 6.7 0 Reproductive 2.8 4.3 3.4 Reproductive 0 0 3.4 0 Sensory 0 8.7 3.4 Sensory 0 0 3.4 0 Urinary 11.1 8.7 10.2 Urinary 0 6.7 1.7 1.7

the highest prevalence of renal pathology. However, age was To our knowledge this is the first review of morbidity in not statistically a significant predictor of medical problems captive AGLs; however, our study has significant limitations. occurrence in captive AGL (F=0.69, p>0.05). Apart from the relatively small sample size for a morbidity Faecal parasites (see Table 6), specificallyEntomoeba sp., study, multiple clinicians were involved in the description of were linked to diarrhoea only in one instance; the remaining medical problems, introducing a potential bias factor in the positive results were incidental findings, not associated with diagnosis. Additionally, the diagnosis on multiple occasions clinical signs. The following bacterial species and genera have was limited to clinical signs rather than a specific pathology. been isolated as part of routine or diagnostic faecal screening: Aetiology was rarely identified and in some cases insufficient Acaligenes xylosalidans, Campylobacter sp., Campylobacter evidence was provided to associate the aetiology and the jejuni, Escherichia coli, Enterobacter agglomerans, pathology identified. Klebsiella pneumoniae, Manheimia haemolytica, Pasteurella Despite these shortcomings, some trends emerge from pneumotropica, Pseudomonas sp, Rahnella aquatilis, this morbidity review that may help to focus future research Staphylococcus epidermis and Vibrio metschnikovi. None of efforts on this species’ pathology in captivity. the isolates were associated with diarrhoea or other clinical Diet is one of the main aetiologies, or at least a contributory signs. factor, to consider for over half of the problems observed, Most wounds were the result of conspecific aggression including gastrointestinal, renal and dental pathologies and towards males. A detailed breakdown of wound aetiology, if obesity. Dietary studies in their natural habitat have shown known, is given in Table 7. that Alaotran gentle lemurs are highly specialised feeders, relying for most of their diet on only four plant species Discussion from the Cyperaceae and Poaceae families, resulting in a Alaotran gentle lemurs are underrepresented in biomedical low-energy (17.3±1.7 MJ/kg dry matter), low protein (crude research literature on prosimians, despite extensive reviews protein 131.6 ± 56.7 g/kg dry matter) and moderate fibre diet of prosimian medicine (Benirschke et al. 1985; Junge 1999; (crude fibre 221.1±72.5 g/kg dry matter) (Mutschler 1998). Junge 2003; Williams 2015). The only published report re- This is in strong contrast to diets offered in captivity during lates to the death of two animals at DWP of renal failure certain periods of time in terms of food variety and nutritional following the ingestion of Russian vine (Polygonum balds- composition. Dietary analysis performed in 1996 revealed chuanicum) (Robert and Allchurch 1995). relatively high levels of protein in some plant parts selected by AGLs during certain times of the year (between 339 and 373 g/kg dry matter) (Fidgett et al, 1996). Differences in dietary requirements between the highly specialised Hapalemur and other lemur species are also reflected in gastrointestinal passage time, but despite a longer passage time in captive Table 6. Faecal parasitology findings in Alaotran gentle gentle lemurs it is suspected that this is shorter in captivity in lemurs at DWP from 1990 to 2014. comparison with wild specimens. (Cabre-Vert and Feistner Result Number of Percentage 1995). samples of samples The distribution of food intake through the day may play No parasites observed 266 86% a role in some of the medical problems observed; in the wild Entomoeba sp. 8 3% this species appears to be cathemeral, feeding in periods Hymenolepis sp. 8 3% throughout the whole day. Providing the food over a shorter Cryptosporidium sp. 4 1% period of time during the day (8 hours). instead of over a Giardia sp. 8 3% 24-hour period as in the wild, may contribute to obesity as Unidentified cestode ova 11 4% energy density and intake may change, as shown in humans Unidentified nematode larvae 4 1% (Ello-Martin et al. 2005).

Solitaire No. 27 20 Health of Hapalemur alaotrensis

Table 7. Wound aetiology. Courts SE (1995). The Jersey Wildlife Preservation Trust Dietary Number of Manual. Jersey: Jersey Wildlife Preservation Trust. Aetiology cases Percentage Male/female Denk D, Boufana B, Masters NJ and Stidworthy MF (2016). Fatal Conspecific 10 62.5% 9/1 echinococcosis in three lemurs in the United Kingdom – a aggression case series. Veterinary Parasitology 218:10–14. Unknown 5 31% 2/3 Ello-Martin JE, Ledikwe JH and Rolls BJ (2005). The influence Capture 1 6% 0/1 of food portion size and energy density on energy intake: Enclosure 1 6% 0/1 implications for weight management. American Journal of Clinical Nutrition 82 (suppl): 236–241. Fidgett AL, Feistner ATC and Galbraith H (1996). Dietary intake, Given variations in the diet over time and the limited food composition and nutrient intake in captive Alaotran sample size, it is difficult to establish the role of diet in the gentle lemurs Hapalemur griseus alaotrensis. Dodo, aetiology of many pathological processes. A cross institutional Journal of the Jersey Wildlife Preservation Trust 32:44–62. study with standardised diets would be helpful in increasing Isbell, LA (1991). Contest and scramble competition: patterns our understanding of the role of diet in morbidty and mortality of female aggression and ranging behavior among in this species. primates. Behavioral Ecology 2:143–55. Cryptosporidium infection has been associated with severe Junge, RE (1999). Diseases of prosimians. In Fowler, ME diarrhoea in some lemur species such as sifakas, while lemurs and Miller, RE (eds), Zoo and Wild Animal Medicine, 4th are reportedly asymptomatic carriers of Giardia (Williams edition. Philadelphia: Saunders, 365–368. 2015). Despite the detection of these parasites in AGLs, no Junge RE (2003). Prosimians. In Fowler, ME and Miller, RE (eds), clinical signs were associated with their presence. Zoo and Wild Animal Medicine, 5th edition. Philadelphia: Despite the relatively high frequency of observation Saunders, 334–346. of cestodes in faecal parasitology, other than infection in Mittermeier, RA, Louis Jr, EE, Richardson, M, Schwitzer, C, prosimians as an intermediate aberrant host by Taenia and Langrand, O, Rylands, AB, Hawkins, F, Rajaobelina, Equinococcus (Denk et al. 2016), no morbidity was associated S, Ratsimbazafy, J, Rasoloarison, RM and Roos, C with the detection of cestodes in AGLs. (2010). Lemurs of Madagascar, 3rd edn. Arlington, VA: Wounds are the most prevalent problem, with males Conservation International. receiving the majority of the injuries, Field studies have Mutschler, T (1998). Folivory in a small-bodied lemur. The revealed that females are more dominant and more aggressive nutrition of the Alaotran gentle lemur (Hapalemur griseus than males, with intersexual conflicts arising mostly during alaotrensis). In Rakotosamimanana B, Rasamimanana feeding times (Waeber and Hemelrijk 2003). In primates H, Ganzhorn JU and Goodman SM (eds), New Directions intragroup competition has been linked to food distribution in Lemur Studies. Kluwer Academics/Plenum Publishers, (Isbell 1991), and so evaluating the behaviour of gentle lemurs 221–240. in captivity under different food presentation regimes may help Mutschler, T and Feistner, ATC (1995). Conservation status to reduce aggression and therefore rates of injury. and distribution of the Alaotran gentle lemur Hapalemur griseus alaotrensis. Oryx 29: 267–274. References Mutschler, T, Feistner, ATC and Nievergelt CM (1998). Andriaholinirina N, Baden A, Blanco M, Chikhi L, Cooke A, Preliminary field data on group size, diet and activity in Davies N, Dolch R, Donati G, Ganzhorn J, Golden C, the Alaotran gentle lemur Hapalemur griseus alaotrensis. Groeneveld LF, Hapke A, Irwin M, Johnson S, Kappeler P, Folia Primatologica 69: 325–330. King T, Lewis R, Louis EE, Markolf M, Mass V, Mittermeier Robert N and Allchurch A (1995). Poisoning of Alaotran RA, Nichols R, Patel E, Rabarivola CJ, Raharivololona gentle lemur (Hapalemur griseus alaotrensis) by Russian B, Rajaobelina S, Rakotoarisoa G, Rakotomanga B, vine (Polygonum baldschuanicum) at the Jersey Rakotonanahary J, Rakotondrainibe H, Rakotondratsimba Wildlife Preservation Trust. Verhandlungsberichte des G, Rakotondratsimba M, Rakotonirina L, Ralainasolo FB, Internationalen Symposiums über die Erkrankungen der Ralison J, Ramahaleo T, Ranaivoarisoa JF, Randrianahaleo Zootiere 37: 207–210. SI, Randrianambinina B, Randrianarimanana L, Waeber PO and Hemelrijk CK (2003). Female dominance Randrianasolo H, Randriatahina G, Rasamimananana and social structure in Alaotran gentle lemurs. Behaviour H, Rasolofoharivelo T, Rasoloharijaona S, Ratelolahy 140: 1235–1246. F, Ratsimbazafy J, Ratsimbazafy N, Razafindraibe H, Williams CV (2015). Prosimians. In Fowler, ME and Miller, Razafindramanana J, Rowe N, Salmona J, Seiler M, RE (eds), Zoo and Wild Animal Medicine, 8th edition. Volampeno S, Wright P, Youssouf J, Zaonarivelo J and Philadelphia: Elsevier-Saunders, 291–301. Zaramody A (2014). Hapalemur alaotrensis. The IUCN Red List of Threatened Species 2014 (downloaded on 2 Alberto Barbon is a veterinary surgeon at Durrell Wild- November 2016). life Park and has carried out research into the health Benirschke K, Miller C, Ippen R and Heldstab A (1985). The of a number of key species. Zoavina Randriana is pathology of prosimians, especially lemurs. Advances in a Malagasy veterinarian and a Durrell Conservation Veterinary Science and Comparative Medicine 30:167– Academy graduate, with a history of working with 208. small animals and wildlife conservation in Madagas- Cabre-Vert N and Feistner ATC (1995). Comparative gut car. Gale Glendewar is a senior mammal keeper at passage time in captive lemurs. Dodo, Journal of the Durrell. Jersey Wildlife Preservation Trust 31:76–81. E-mail: [email protected]

Solitaire No. 27 21 Research report Effects of dietary changes on the health of captive Geoffroy’s marmosets Callithrix geoffroyi

Dominic Wormell and Eluned Price

Durrell Wildlife Conservation Trust

Abstract This study used analysis of medical records at Durrell Wildlife Conservation Trust to investigate the impact of dietary changes, namely the addition of acacia gum and an increase in insects, on the health of Geoffroy’s marmosets Callithrix geoffroyi. Rates of death, illness and individual symptoms were calculated per animal per year for two periods: the eight years before the changes, and seven years following the changes. Rates of death, bouts of illnesses and occurrences of specific symptoms per animal-year were all lower, often markedly so, following the changes to the diet. This study emphasises the importance of using information on diets in the wild to manage the nutrition of captive animals. Key words: Callitrichidae, gum, insects, morbidity, mortality, nutrition

Introduction had had some health problems (Price 1992b); severe weight It is now clear from the growing number of field studies of loss, diarrhoea, poor coat condition and lethargy were the callitrichids that their various ecological adaptations are as- commonest symptoms. Broken limbs in some monkeys may sociated with considerable differences in the composition have been related to calcium deficiency. of the diet in the wild (e.g. Ferrari and Lopes Ferrari 1989; The species’ diet was therefore investigated to see if some Ferrari 1993; Rylands 1996), yet until relatively recently, light could be shed on these problems. The results (Price most institutions holding callitrichids in captivity have 1992a) suggested that the diet at the time of the study may provided largely the same diet for all species. Health and have been low in protein, in some vitamins and minerals, and breeding problems that have been experienced with cap- possibly also in energy, particularly perhaps during pregnancy tive callitrichids, such as stillbirths and abortions, marmoset and lactation. Consequently, the diet was modified, with an “wasting syndrome”, and susceptibility to infection (e.g. increase in insects and the daily provision of gum arabic. In Shimwell et al. 1979; Barnard et al. 1988; Crook 1989; Price this paper we attempt to assess whether or not the health of 1992a; Wormell et al. 1996) may be due at least in part to the Geoffroy’s marmosets at Durrell improved following the a diet which is not suitable for the species concerned. An introduction of gum to the diet. understanding of natural diets is therefore vital for captive management. The most significant feature of the feeding habits of marmosets (genera Callithrix, Mico and Cebuella) is their use of plant exudates (Martins and Setz 2000; Ferrari and Rylands 1994; Ferrari and Digby 1996; Digby and Barreto 1996; Rylands 1989), and they possess adaptations in dentition and gut morphology to enable them to exploit these (e.g. Coimbra-Filho and Mittermeier 1976; Ferrari 1993; Power and Oftedal 1996). Geoffroy’s marmoset Callithrix geoffroyi (Figure 1) is endemic to the Atlantic coastal forests of Brazil and, like other marmosets, its diet in the wild is heavily dependent on plant exudates and insects (Passamani 1998; Passamani and Rylands 2000). An investigation of the nutrition of Geoffroy’s marmosets at Durrell Wildlife Conservation Trust was initiated in 1991 in an attempt to tackle frequent adult health problems and high infant mortality that was not due to incompetent parental behaviour (Price 1992a,b). There was also a worrying loss of young adults one to two years old, and Figure 1. Geoffroy’s marmoset (Photo: Hans Hillewaert; Yersinia pseudotuberculosis was a common cause of death. Creative Commons Attribution-Share Alike 3.0 Unported Up to December 1991, 84% of the animals in the collection licence).

Solitaire No. 27 22 Marmoset diet and health

Methods Analysis To obtain information on disease and mortality in Durrell As the number of marmosets in the collection varied over the Wildlife Conservation Trust’s Geoffroy’s marmosets, colony study period, the length of time each marmoset that survived records were examined for the period between the arrival past the age of 30 days was present in the collection was of the first marmosets in Jersey on 1 December 1984, until calculated. These values were summed for each phase to the end of December 1999 (from 2000 onwards very few give the number of animal-years. Frequencies of illnesses, marmosets remained in the collection). As gum, in the form deaths and symptoms per animal-year in each phase were of gum arabic diluted with water (Herron et al. 2001) was then calculated and compared. Statistical analyses were not introduced as a daily component of the diet at the end of performed as the data were not independent: some marmo- November 1992, this period was split into two phases: (1) sets were present for both phases, others for only one. 1 December 1984 - 31 December 1992, and (2) 1 January 1993 - 31 December 1999. The following information was Results extracted: (1) dates and causes of death (if known) of all non- Mortality infant marmosets that had died while at the zoo; (2) details Other than premature births and infants less than 1 month of illnesses that had not resulted in death. For the latter, both old, for whom accurate causes of death were not available, 21 separate bouts of illness (which could include several symp- Geoffroy’s marmosets died during the study period, 17 from toms), and each symptom within an episode of illness, were the end of 1984 to the end of 1992, and 4 from 1993 to 1999. tabulated. A bout of illness was considered terminated when Of deaths before the end of 1992, seven (46.7%) died of con- it was followed by a record of improved health. No account firmed or suspected pseudotuberculosis, three (20%) of other was taken of the duration of bouts. The number of different infections, and one (6.7%) of septicaemic shock following a individuals affected by non-fatal illness was also tabulated. fight with conspecifics. Cause of death was not firmly estab- As the main focus of the study was health in relation to diet, lished in the other six cases, two of which were infants less injuries caused by fighting were excluded from the analysis. than six weeks old. Of the four subsequent deaths, two were Symptoms were categorised as follows: due to pseudotuberculosis, and the remaining two were sud- Weight loss: A record of a drop in weight from a previous den, one occurring during an examination. Rates of death per known weight. animal-year were lower following the introduction of gum to Diarrhoea: Loose faeces, consistency varying from wa- the diet (Figure 2). tery to more solid but unformed stools. Poor pelage condition: Matted coat, hair loss. Other illness Lethargy: Reluctance to move, sometimes associated From 1984 to 1992, of 31 animals held at Jersey who survived with seeking heat and blinking, but able to move approxi- to the age of more than 30 days, 22 (71%) were affected by mately normally if necessary. non-fatal illnesses. An additional four animals who had not Weakness/lack of mobility: Inability to move properly. had any previous record of illness died, resulting in a total Broken bones: Usually presenting as limping or lack of of 26 individuals (84%) who at one time or another suffered use of a limb and confirmed on x-ray. some degree of ill-health. The severity of symptoms ranged Abnormal urine: Any abnormality detected using Multi- from isolated and mild cases of diarrhoea or vomiting, to stix urine test. fatal infections. Other: Rare symptoms such as vomiting shown only by A total of 95 instances of symptoms of poor health were one or two marmosets on isolated occasions. recorded. Most common were observations of weight loss

Deaths

Bouts of illness

Weight loss

Diarrhoea

Poor pelage condition After gum Weakness/lack of mobility Before gum Broken bones

Lethargy

Abnormal urine

Other

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Figure 2. Occurrence of deaths, non-fatal bouts of illness and individual symptoms within bouts, calculated per marmoset per year, before and after the introduction of gum to the diet.

Solitaire No. 27 23 Research report

(20%), diarrhoea (17%), poor coat condition (14%) and Coimbra-Filho, AF and Mittermeier, RA (1976). Exudate- lethargy or listlessness (14%). There were nine records of eating and tree-gouging in marmosets. Nature 262: 630. injury, but six of these were broken limbs associated with other Crook, G (1989). A nutritional reversal of marmoset wasting illness, especially suspected calcium deficiency in one female. syndrome. Australian Primatology 4:21. From 1993 to 1999, 21 marmosets were held at Jersey, Digby, LJ and Barreto, CE (1996). Activity and ranging excluding infants which did not survive to more than 30 patterns in common marmosets (Callithrix jacchus). days. Of these, 9-10 (the exact number is not known as some Implications for reproductive strategies. In: Norconk, animals in one group were not individually identifiable) were MA, Rosenberger, AL and Garber, PA (eds). Adaptive affected by non-fatal illness. Another died suddenly without Radiations of Neotropical Primates. Plenum Press: New having shown any previous signs of illness. Thus, after York, 173-185. the introduction of gum, 43-48% of the marmosets in the Ferrari, SF (1993). Ecological differentiation in the collection suffered some degree of ill-health, just over half the Callitrichidae. In: Rylands, AB (ed.). Marmosets and percentage observed in the period before gum was introduced. Tamarins. Systematics, Behaviour, and Ecology. Oxford: In all, 34 instances of separate symptoms were noted. Oxford University Press, 314-328. Bouts of illnesses and occurrences of individual symptoms Ferrari, SF and Digby, LJ (1996). Wild Callithrix groups: stable per animal per year are presented in Figure 2. In all cases there extended families? American Journal of Primatology was a reduction in the rate of occurrence after the introduction 38:19-27. of gum to the diet, and this decrease was often marked. Ferrari, SF and Lopes Ferrari, MA (1989). A re-evaluation of the social organisation of the Callitrichidae, with reference Discussion to the ecological differences between genera. Folia Finding the root causes of health problems in captive ani- Primatologica 52:132-147. mals, and developing effective means of combatting them, Ferrari, SF and Rylands, AB (1994). Activity budgets and is a complex task. Improved management techniques, such differential visibility in field studies of three marmosets as minimally disruptive husbandry routines and larger cages (Callithrix spp.). Folia Primatologica 63:78-83. with more complex furnishings, have in some cases dramati- Herron, S, Price, E and Wormell, D (2001). Feeding gum cally improved the breeding success of captive callitrichids arabic to New World monkeys: species differences and (e.g. Snowdon et al. 1985). Many of these measures have palatability. Animal Welfare 10: 249-256. been adopted at Durrell (e.g. Wormell et al. 1996). However, Martins, MM and Setz, EZF (2000). Diet of buffy tufted-eared one of the most important factors is likely to be diet. Prob- marmosets (Callithrix aurita) in a forest fragment in lems such as osteomalacia in breeding females, for example, southeastern Brazil. International Journal of Primatology may be a result of calcium deficiency. 21: 467-476. Marmosets are obligate gummivores, with anatomical and McGrew, WC, Brennan, JA and Russell, J. (1986). An artificial physiological adaptations to feeding on exudates (Coimbra- “gum-tree” for marmosets (Callithrix j. jacchus). Zoo Filho and Mittermeier 1976; Ferrari 1993; Power and Oftedal Biology 5:45-50. 1996). Gum is an important source of calcium and energy Passamani, M (1998). Activity budget of Geoffroy’s marmoset (Passamani and Rylands 2000) and in the wild, gum forms (Callithrix geoffroyi) in an Atlantic forest in southeastern nearly 70% of the diet of Callithrix geoffroyi (Passamani Brazil. American Journal of Primatology 46: 333-340. 1998; Passamani and Rylands 2000). In the past, gum was Passamani, M and Rylands, AB (2000). Feeding behavior of typically offered to captive marmosets only as a means Geoffroy’s marmoset (Callithrix geoffroyi) in an Atlantic of environmental enrichment (e.g. McGrew et al. 1986). forest fragment of south-eastern Brazil. Primates 41: 27- Providing gum arabic on a daily basis for all Callithrix in 38. Durrell’s collection was one of the first steps taken to combat Power, ML and Oftedal, OT (1996). Differences among health problems in this species, and the results of this analysis captive callitrichids in the digestive responses to dietary suggest it has had considerable health benefits. Since we gum. American Journal of Primatology 40:131-144. began our investigations of health and diet in marmosets, Price, EC (1992a). The nutrition of Geoffroy’s marmoset gum has become a standard component of the diet for captive Callithrix geoffroyi at the Jersey Wildlife Preservation Trust. marmosets in many institutions and similar benefits have been Dodo, Journal of the Wildlife Preservation Trusts 28:58-69. reported (S. Muir, pers. comm.). Price, EC (1992b). The Health and Nutrition of Geoffroy’s Of course, management of captive animals should be under Marmoset (Callithrix geoffroyi) at the Jersey Wildlife constant review, and diet is only one element of a suite of Preservation Trust. Unpublished report. Jersey: Jersey changes that have been made to the husbandry of callitrichids Wildlife Preservation Trust. at Durrell to address problems in the health and reproduction of Rylands, AB (1989). Sympatric Brazilian callitrichids: the black a number of species (e.g. Wormell et al. 1996), but we believe tufted-ear marmoset, Callithrix kuhli, and the golden- that our experience emphasises the importance of taxon- headed lion tamarin, Leontopithecus chrysomelas. specific diets to the success of captive breeding programmes Journal of Human Evolution 18: 679-695. for endangered species. Rylands, AB (1996). Habitat and the evolution of social and reproductive behavior in Callitrichidae. American Journal References of Primatology 38: 5-18. Barnard, D, Knapka, J and Renquist, D (1988). The apparent Shimwell, M, Warrington, BF and Fowler, JSL (1979). Dietary reversal of a wasting syndrome by nutritional intervention habits relating to ‘wasting marmoset syndrome’ (WMS). in Saguinus mystax. Laboratory Animal Science 38:282- Lab Animals 13:139-142. 288. Snowdon, CT, Savage, A and McConnell, PB (1985). A

Solitaire No. 27 24 Marmoset diet and health

breeding colony of cotton-top tamarins (Saguinus Dominic Wormell is Head of the Mammal Department oedipus). Laboratory Animal Science 35: 477-480. at Durrell Wildlife Conservation Trust, and Eluned Price Wormell, D, Brayshaw, M, Price, E and Herron, S (1996). is Zoo Research Coordinator. Both have conduct- Pied tamarins Saguinus bicolor bicolor at the Jersey ed a wide range of research into the conservation, Wildlife Preservation Trust: Management, behaviour and behaviour, health and management of marmosets reproduction. Dodo, Journal of the Wildlife Preservation and tamarins. Trusts 32: 6-97. E-mail: [email protected]

Biodiversity loss in Gerald Durrell’s early works

Gregory Byrnes

Between 1947 and 1950, Gerald Durrell made three I The Overloaded Ark animal collecting expeditions. A year or so later, he 1. Page 15 began his professional writing career with accounts of 2. Pages 19–21 these expeditions which were published in 1953 and 3. Pages 167–8 1954. 4. Page 229

Superficially, Durrell’s first three books belong to the genre of light travel literature. Their style, for the most part informal and humorous, may have disguised the fact that they also have a serious purpose. Attentive reading, in fact, identifies numerous passages which highlight biodiversity loss.

Biodiversity, an abbreviation of biological diversity, had not been coined in the period examined here, but Durrell was concerned about loss of biodiversity avant la lettre, as will be shown below.

His first three expeditions were to the British Cameroons in 1947–8 and again in 1949, and to British Guiana in 1950. Sources for studying them include diaries and letters, but this essay confines itself to Durrell’s published accounts: The Overloaded Ark (1953), Three Singles to Adventure (1954) and The Bafut Beagles (also 1954).

Reading them with an awareness of rhetorical devices such as repetition, sarcasm and contrast, one notices a sustained concern about the destruction of forests for large-scale cultivation and the commercial, industrial and residential structures that follow in its wake. Comparison of Durrell’s personal reminiscences with independently established facts and dates in standard reference works confirms that his observations are reliable.

For convenience, some passages in which particular attention is drawn to biodiversity issues are set out in abbreviated form below: Gerald Durrell (Photo: Durrell Wildlife Conservation Trust).

Solitaire No. 27 25 Essay

II Three Singles to Adventure The effect of plantations on the people of Cameroon 1. Page 4 is suggested by an understated contrast which speaks 2. Pages 67– 70 for itself: “…a clearing containing a white bungalow, in 3. Pages 265– 6 which lived a European overseer; or a row of horrible corrugated iron sheds, in which lived the banana III The Bafut Beagles pickers…” (I, 4). 1. Page 104 These passages have not always been appreciated: “In One word in particular, “civilisation”, stands out by the fact there is not a glimmer to be found in the book [The way it is used repeatedly and always in a sarcastic Overloaded Ark] – or, for that matter, in its immediate tone: “a landscape denuded of its flora and fauna by sequels…– of Gerald’s views on…animal conservation the beneficial influences of civilization” (I, 1); “This was a in general…Yet before the mid-fifties there is no written bit of civilized Africa, and I shuddered to look at it…” (I, evidence at all of any particular interest in animal 4); “…with cinemas, snack bars, neon lights, and other conservation or in environmental matters as a whole.” doubtful privileges of civilization.” (II, 1); “the blessings (Botting 1999, pp. 206–7). That author dates Durrell’s of civilization” (II, 2). “first public protest at the role of man in the destruction of nature” to a radio talk on 29 August 1956.( Botting Another word much repeated and always negative is 1999, p. 239). The opinion that “little of this way of “plantation”: “apparently endless palm plantations… thinking [i.e. conservation] could be found in his [Durrell’s] the endless rows of palm trees about us, serried ranks…” books” is repeated on page 241. (I, 2 ); “the palm plantations …around Tiko…” (I, 4); and regarding banana plantations, “mile upon mile That interpretation is contradicted by the quotations of nothing but banana trees in a great characterless above. They show that Durrell expressed his concern sheet, arranged in neat rows like a green chessboard. about loss of biodiversity in print in 1953 and 1954, Hideous regimentation, a thousand million banana trees based on experiences which date back to 1948. standing in serried ranks…” (I, 4). Clearly, it is worthwhile revisiting these early books. Further Small-scale farming around Bafut, however, is shown in study of them, and of archival material not included a positive light: “Below us lay a mosaic of small fields, here, could throw more light on the development, green and silver and fawn, broken up by minute palm chronological and philosophical, of Durrell’s thinking thickets and an occasional patch of rust red where the about conservation. earth of a field had been recently hoed.” (III, 1). Galagos were living in the forest around those farms. Acknowledgements I thank the State Library of New South Wales, the Sydney Then there was “the real forest country”, the “beauty and City Library and the University of New South Wales Library. colour” of which is described in lyrical detail (I, 2, and, for Guiana, compare II, 3: “a feeling of awe”). References Botting, D (1999). Gerald Durrell: The Authorised Biography. Durrell, from a lookout, saw a vast forest and London: Collins. contemplated how it “ stretched, almost unbroken, Durrell, GM, (1953). The Overloaded Ark. London: Faber. right across Africa, until it merged into the savannah Durrell, GM (1954a). Three Singles to Adventure. London: land of the east.” (I,3). He attempted to estimate the Rupert Hart-Davis [Pagination referred to here is that of number of animals living there but the mental arithmetic the Ulverscroft Large Print Edition]. made him dizzy; the unstated implication is that every Durrell, GM (1954b). The Bafut Beagles. London: Rupert Hart- plantation meant so much biodiversity lost. Davis. Fa, JE, Funk, SM and O’Connell, D (2011). Zoo Conservation These texts also show concern for indigenous people. Biology. Cambridge: Cambridge University Press. This is relevant because, as mentioned by Fa et al. (2011, p. 5), “ [h]uman cultural diversity could also be considered part of biodiversity since human cultures Gregory Byrnes, MPhil is from Sydney, Australia. He has represent ‘solutions’ to the problems of survival in a background in the humanities and currently teach- particular environments.” es English to international students. He enjoys natural history and is a member of the Australian Wildlife Soci- The destruction of indigenous communities in South ety and a Taronga Zoo Friend. At present he is writing America is presented in a way that hints at the parallel a book about Gerald Durrell’s little-known second ex- confinement of wildlife in protected areas: “…the pedition around Australia in 1969-1970. Amerindians have had their country wrested away from them and are forced to live in reserves…” (II, 2). Email: [email protected]

Solitaire No. 27 26 Measuring muscle in bats Assessing muscle condition in captive Livingstone’s fruit bats

Dominic Wormell1, Scot Ramsay2, Will Masefield1, David Houston3 and Eluned Price1

1Durrell Wildlife Conservation Trust, Jersey 2James Hutton Institute, Aberdeen, UK 3Institute of Biodiversity, Animal Health and Comparative Medicine, University of Glasgow, UK

Abstract It can be difficult to provide captive fruit bats (Pteropodidae) with adequate space for normal flight, and this can be a factor in the development of obesity and related medical conditions. This is often a particular problem in dominant males, who defend feeding territories and are therefore largely sedentary. Lack of muscle use can lead to atrophy, thus exacerbating the problem. A muscle moulding technique previously used in birds was tested in six adult males from a colony of Livingstone’s fruit bats (Pteropus livingstonii) before and after a move to a larger enclosure. The technique enabled us to compare muscle mass between individuals, and within individuals over time, and has great potential in evaluating methods of alleviating health problems associated with captivity. Key words: alginate, captive management, muscle moulding, Pteropodidae, Pteropus livingstonii

Introduction changes can have a substantial number of health consequences Fruit bats (flying foxes) are keystone species in many tropi- (Wilmore and Costill 2004; McArdle et al. 2000). cal forests. They play important roles in both pollination Following their arrival, obesity as a result of lack of and seed dispersal of forest plants (McConkey and Drake exercise and year-round good quality food became a problem 2006; Reiter et al. 2006). Many species are now threatened in the Livingstone’s bats at Durrell. The mean weight of the because of habitat loss and persecution, and captive breed- ten adult wild-caught male bats at capture was 657g; after ing programmes have been established for a number of such about 1.5–2 years in captivity, their weight had increased by species (Carroll et al. 1995; Masefield 2003). Among these an average of 16%. is the Critically Endangered Livingstone’s fruit bat Pteropus Although some of the bats had been moved to an enclosure livingstonii from the Comoro Islands in the Indian Ocean with an outdoor aviary in 1995 (Courts 1996), the space (Mickleburgh et al. 2008; see Figure 1). In 1992–1995, available was still relatively restricted and flying activity 17 wild-caught Livingstone’s bats were brought to Durrell Wildlife Conservation Trust’s headquarters in Jersey, and the captive population now numbers over 60. Wild fruit bats typically roost in groups of varying sizes (Smith and Leslie 2006; Kerth 2008), flying at night to feeding sites that may be several kilometres or more away (Richter and Cumming 2006; Tidemann and Nelson 2003). When fruit bats are brought into captivity it is impossible to provide comparable flying conditions to those in the wild, and as a consequence animals may suffer from muscle atrophy and become obese (Courts 1996, 1999). Dominance hierarchies, particularly amongst males, are also established in these colonial animals, and dominant bats tend to defend small feeding territories (Kunz and Lumsden 2003). In the wild, the bats must fly to and from these territories, but in captivity territory-holding males remain relatively sedentary and as a result often becoming overweight, leading to difficulty flying (Wilson 1988; LeBlanc 2009; Courts 1999). It is well known from domestic animals and human health studies that these Figure 1. Livingstone’s fruit bat (Photo: James Morgan).

Solitaire No. 27 27 Research report

Figure 2. Exterior view of the flight tunnel at the time of the Figure 3. Interior view of the flight tunnel at the time of the study. study.

was extremely limited. In 2003, as part of the programme a discrete muscle block (Yalden and Morris 1975), and is to improve the health status of the captive group, animals covered in dense fur. However, the muscle mass on the wings began moving into a modified horticultural polythene tunnel forms a discrete, lightly furred, and easily moulded mass. (Wormell 2012). This provided a massive increase in the We used alginate moulding to record the cross-sectional area space available for flight, providing a straight flight path of of the whole wing muscle mass overlying the humerus bone. 38 m and a total volume of approximately 1200 m3 (Figures The anaesthetised bats were laid on their backs, the wings 2 and 3). As well as providing excellent opportunities for extended, the membrane dusted with talcum powder, and flying, the enclosure also supplied a large area for climbing. dental alginate poured over the muscle mass overlying the Measuring muscle mass would be a valuable indicator of humerus bone; we also included some of the radius, to give a fitness and a means of assessing the success of such changes fixed reference point for later sectioning. After the alginate had in management. The present study tested whether a body- set, it was removed from the wing, and the wing membrane moulding technique that has been successfully used in birds lightly wiped with a damp cloth to remove excess talcum. to predict the mass of the pectoral muscles (e.g. Veasey et al. The cavity in the mould was then filled with a liquid plaster 2000; Barboutis et al. 2011) could also be applied in fruit bats. of Paris, which was left to set. The plaster “cast” of the muscle mass was then removed from the mould, and sectioned in a Methods band saw at a set distance of 1 cm from a clearly identifiable Subjects bony protuberance at the proximal end of the radius. A Six captive-born male bats were studied (see Table 1). All photocopy image was obtained of each cross-section (both were adult (pteropodid bats reach maturity at 12–24 months; sides) by placing the cast face down on the photocopier plate, Pierson and Rainey 1992; Welbergen 2010). Weights were together with a scale bar 50 mm long. The paper copy was recorded and muscle mouldings taken when they were subsequently scanned using a digital scanner to produce a moved to the tunnel in August 2003, and again in October digital image of the cross sections together with a scale bar. 2004. This was done as part of routine health screening of To standardise the area of muscle measured, we identified the animals, under anaesthetic. two places on each cross section where the wing membrane

Moulding We moulded muscle shape using dental alginate, CA37, manufactured by Cavex Holland, P.O. Box 852, 2003 RW Haarlem, Netherlands. This technique has been used in birds to record the cross-sectional area of the pectoral muscle re- gion, a measure which has been shown to correlate well with the actual weight of the pectoral muscle mass (Selman and Houston 1996). Alginate, when mixed with water, remains in a cream-like condition for about 90 seconds, before so- lidifying within a few seconds. Once hardened, it remains rigid enough to retain the shape of the moulded object in great detail. It does not adhere to skin or fur, and can easily be removed from the bat’s wing after treatment. In bats the wing muscle system on the torso is not suitable for moulding Figure 4. Cross-sectional area of cast showing cut-off line because it is more complex than that in birds, does not form drawn between wing membranes.

Solitaire No. 27 28 Measuring muscle in bats

Table 1. Details of bats studied.

Mean muscle area Mean muscle area Weight at 1st Weight at 2nd at 1st moulding at 2nd moulding Durrell ID Name Date of birth moulding (g) moulding (g) (mm2) (mm2) M2244 Ghost 6 May 1996 1064 961 303 277 M2277 Iggy 24 August 1996 1233 1046 348 321 M2504 Ruggiero 5 March 1999 1001 1043 269 287 M2618 Lysander 17 May 2000 937 1013 318 342 M2682 Balthazar 7 May 2001 890 918 275 324 M2694 Melchior 26 June 2001 793 943 231 293

Mean 986.3 987.3 290.7 307.2 CV 14.1 5.0 12.9 7.5

meets the arm, and used this as the cut-off point: a horizontal whilst in 2004 between-bat variability accounted for 70.4% line was drawn between these two points (see Figure 4), and of total observed variation. This indicates that the technique the muscle area above this line was measured using ‘Scion is sufficiently reliable and accurate to be useful for our needs. Image’ (Scion Image Beta 4.0.2.; Scion Corporation). As the positioning of this line required some judgement, we repeated Discussion each measure and used the mean of the two measures. Captivity has effects not only on the behaviour of animals, but also on their morphology (O’Regan and Kitchener 2005). Results The roost and feeding sites of wild bats are usually a signifi- Mean determined muscle areas for all bats are presented in cant distance apart: satellite telemetry studies of wild Eidolon Table 1. We were not able to establish whether the cross helvum and Pteropus poliocephalus suggest that flying foxes sectional muscle area we measured was correlated with ac- tend to forage 15–50 km from their day roosts each night tual muscle mass, because this would have involved killing (Richter and Cumming 2006; Tidemann and Nelson 2003). animals: such a relationship, however, has been established Captive fruit bats, on the other hand, typically spend only a for this technique in birds (Selman and Houston 1996). very small proportion of their time flying (e.g. Carroll 1979). One non-destructive way to indirectly examine the Bats brought into captivity are therefore almost certain accuracy of the method, however, is to examine “within-bat” to lose muscle condition, while gaining fat. In this study, variation (between left and right wings) in comparison to we have demonstrated that muscle mass in fruit bats can be “between-bat” variation (variation in mean wing muscle area measured by a technique originally developed for use in birds. between different bats). Observed within-bat variation would This method should help in evaluating the extent to which result from both actual natural variation in muscle size between modifications to the housing and management of captive fruit left and right wings and any inherent inaccuracies or errors in bats are successful in increasing activity, especially flight, the moulding technique. Between-bat variation, on the other and thus health. hand, should reflect actual differences in gross muscle mass between individuals plus, again, any inherent inaccuracies Acknowledgements or errors in the moulding technique. As one would expect We are grateful to all the mammal and veterinary staff at Dur- flight muscles to be fairly symmetrical, the actual difference rell Wildlife Park for their support. between left and right wing muscles within individuals should be small relative to the differences in muscle mass References between individuals. Also, the level of error in the technique Barboutis, C, Mylonas, M and Fransson, T (2011). Breast should be roughly similar between wings and between bats. muscle variation before and after crossing large Therefore, if observed “between-bat” variation proved to be ecological barriers in a small migratory passerine (Sylvia significantly greater than observed “within-bat” variation, this borin, Boddaert 1783). Journal of Biological Research – would indicate that the technique should be reliable enough Thessaloniki 16: 159–165. to detect reasonable changes in body muscle size (at least Carroll, JB (1979). The general behavioural repertoire of the changes of a magnitude similar to mean differences in muscle Rodrigues fruit bat Pteropus rodricensis in captivity at the size seen between different individuals). From Table 1 it can Jersey Wildlife Preservation Trust. Dodo, Journal of the be seen that in both 2003 and 2004, between-bat variation Jersey Wildlife Preservation Trust 16: 51–59. was significantly greater than within-bat variation (i.e. the Carroll, JB, Gilmour, L and Courts, S (1995). Rodrigues Fruit variation between left and right wings). In 2003 between-bat Bat Pteropus rodricensis International Studbook. Volume variability accounted for 87.3% of total observed variation, 1. Jersey: Jersey Wildlife Preservation Trust.

Solitaire No. 27 29 Research report

Courts, SE (1996). An ethogram of captive Livingstone’s fruit Tracking bat-dispersed seeds using fluorescent pigment. bats Pteropus livingstonii in a new enclosure at Jersey Biotropica 38: 64–68. Wildlife Preservation Trust. Dodo, Journal of the Wildlife Richter, HV and Cumming, GS (2006). Food availability and Preservation Trusts 32: 15–37. annual migration of the straw-coloured fruit bat (Eidolon Courts, SE (1999). Dietary studies of Livingstone’s fruit bat helvum). Journal of Zoology 268: 35–44. Pteropus livingstonii: feeding behaviour, diet evaluation Selman, R and Houston, DC (1996). A method for estimating and modification. Dodo 35: 26–47. the muscle condition of small, live passerine birds. Ibis Kerth G (2008). Causes and consequences of sociality in 138: 348–350. bats. BioScience 58: 737–746. Smith, SJ and Leslie, DM Jr (2006). Pteropus livingstonii. Kunz, TH and Lumsden, LF (2003). The ecology of cavity and Mammalian Species 792, 1–5. foliage roosting bats. In: Kunz, TH and Fenton, MB (eds). Tidemann, CR and Nelson, JE (2003). Long-distance Bat Ecology. Chicago: University of Chicago Press, 3–89. movements of the grey-headed flying fox (Pteropus LeBlanc, D (2009). Environmental enrichment for long-term poliocephalus). Journal of Zoology 263: 141–146. captive bats. In: Barnard SM (ed.). Bats in Captivity, Veasey, JS, Houston, DC and Metcalfe, NB (2000). Flight Volume 3: Diet and Feeding. Washington DC: Logos muscle atrophy and predation risk in breeding birds. Press, 281–305. Functional Ecology 14: 115–121. Masefield, W (2003). European Studbook for Livingstone’s fruit Welbergen, JA (2010). Growth, bimaturation, and sexual size Bat (Pteropus livingstonii), 1st edn. Jersey: Durrell Wildlife dimorphism in wild gray-headed flying foxes (Pteropus Conservation Trust. poliocephalus). Journal of Mammalogy 91: 38–47. McArdle, WD, Katch, FI and Katch, VL (2000). Essentials of Wilmore, JH and Costill, DL (2004). Physiology of Sports Exercise Physiology, Vol 3. Baltimore: Limmincott, Williams Science, 3rd edn. Champaign: Human Kinetics. and Wilkins Wilson, DE (1988). Maintaining bats for captive studies. In: McConkey, KR and Drake, DR (2006). Flying foxes cease to Kunz, TH (ed.). Ecological and Behavioral Methods for function as seed dispersers long before they become the Study of Bats. London: Smithsonian Institution Press, rare. Ecology 87: 271–176. 247–263. Mickleburgh, S, Hutson, AM and Bergmans, W (2008). Pteropus Wormell, D (2012). The recycled roost. Zooquaria 80: 22–24. livingstonii. The IUCN Red List of Threatened Species. Yalden, BW and Morris, PA (1975). The Lives of Bats. Newton Version 2014.1. www.iucnredlist.org (downloaded on 11 Abbot: David and Charles. July 2014). O’Regan, HJ and Kitchener, AC (2005). The effects of captivity Dominic Wormell is Head of the Mammal Depart- on the morphology of captive, domesticated and feral ment at Durrell. Scot Ramsay is at the James Hutton mammals. Mammal Review 35: 215–230. Institute in Aberdeen. Eluned Price is Zoo Research Pierson, ED and Rainey, WE (1992). The biology of flying Coordinator, and Will Masefield was formerly a senior foxes of the genus Pteropus: a review. In: Wilson, DE and mammal keeper at Durrell. Professor David Houston Graham, GL (eds). Pacific Island Flying Foxes: Proceedings is Honorary Senior Research Fellow at the Institute of of an International Conservation Conference. US Fish and Biodiversity Animal Health and Comparative Medi- Wildlife Service Biological Report 90: 1–17. cine, University of Glasgow. Reiter, J, Curio, E, Tacud, B, Urbina, H and Geronimo, F (2006). E-mail: [email protected] Conservation update on the ploughshare tortoise

Angelo Ramy Mandimbihasina

Durrell Wildlife Conservation Trust

The ploughshare tortoise, Astrochelys yniphora, locally named “angonoka”, is one of Madagascar’s endemic land tortoises (Figure 1). Its distribution is very limited, in the vicinity of the Baly Bay area, north-western Madagascar. Its habitat consists of bamboo scrub and scrub shrub of a total of 160 km2. Wild populations of A. yniphora are very fragmented and subject to many threats such as bush fires, use of its habitat by zebus belonging to local people, and poaching. Combined with the slow growth Figure 1. Ploughshare tortoise (Photo: Matt Goetz).

Solitaire No. 27 30 Ploughshare tortoise conservation

of this species – it takes at least 20 years for an individual in 2008. It is listed in Appendix 1 of CITES, which means to be able to lay eggs – these threats are causing its that exporting ploughshare tortoises or any parts derived decline. The population size of wild A. yniphora has been from the species from Madagascar is forbidden as is estimated to be less than a thousand since the late importing it into a member country of CITES. 1990s. Because of its low numbers, the ongoing decline and its limited distribution, A. yniphora has been listed Ongoing conservation activities as Critically Endangered in the Red List of threatened To reverse the situation of wild ploughshare tortoise species (Leuteritz and Pedrono 2008). populations, Durrell Wildlife Conservation Trust has been investing in many activities, in order to protect the Historical overview of ploughshare conservation remaining ploughshare tortoise population, of course The ploughshare tortoise was described in 1885 by with help from donors. A key action has been to continue Léon Vaillant, a French zoologist, from a specimen to increase public awareness with supports from local obtained by sailors. Before the mid 1980s, captive authorities, by means of big celebrations, village breeding was attempted in Ivoloina Zoological Park meetings, and collaboration with schools. Another but reproduction was unsuccessful. In the mid 1980s, important action is lobbying all higher authorities in Durrell, then known as Jersey Wildlife Preservation Trust order to have transparency in actions and to prove to (JWPT), was appointed by the IUCN to save the species. people that we need to protect this species. A new captive breeding centre was established in 1986 by Durrell in Ampijoroa, not far from its natural Technology has improved our fire survey system. Our range. Nesting and the first hatching success occurred field team is now receiving fire alerts by e-mail from the following year, and has continued since then. NASA’s Fire Information for Resource Management Research in the field has continued to study the habitat System (FIRMS) and Conservation International’s requirements, distribution and biology of the species. “Firecast” Systems in near real time (NRT). The purpose of this is to improve rapid local actions to fight fires. Conservation of the ploughshare tortoise with the local community started in the early 1990s with public Since 2010, village patrols, known as para-rangers, awareness campaigns through festivals. As a result, a have been put in place to undertake daily patrols National Park was created at the end of 1997, and all inside the habitat of the ploughshare tortoise, and known habitats of this tortoise have been included in ensure a permanent presence in the field (Kiester et al. this park. Since the creation of the Park, local people 2013). Local villagers have been living closely with their have participated in the conservation of this species by forests and they know the park very well, so they know creating and maintaining fire breaks around the Park the poachers’ access routes. They also inform the Park boundary and around the limit of core zones. Manager by phone if they see a fire or poachers. This system has led to a reduction in the damage caused A trial release of captive-born ploughshare tortoises by fire as it enables a more rapid response; it has also took place in 1998 into an abandoned habitat and has led to the arrest of poachers. shown success. In 2005, this reintroduction continued with a set of 20 tortoises every year. Over 100 tortoises Since 2014, village patrols have been improved with have been reintroduced, and they have produced the use of GPS and GPS cameras in order to easily more than 20 juveniles. locate reported signs of incursion or other threats to the species. In addition, patrol data have been recorded Despite the success of reintroducing captive-raised into SMART® in order to show transparency and facilitate tortoises back into the wild, recent research has decision making. SMART imports all GPS tracks recorded highlighted the rapid and worsening decline of the wild ploughshare tortoise population as a result of worsening poaching activities for the international illegal wildlife trade.

Since the 1960s, the species has been protected by Malagasy law 60–128 which prohibits all harvesting, eating and trade. That law was updated in the mid 2000s and the species is still in category 1, class 1, which means that it is protected, no harvesting or poaching is allowed, and eating the tortoises is prohibited. All habitats of the ploughshare tortoise have been declared part of the core zone of the National Park, which gives more protection to the species as entry in these zones is forbidden without permits from the Ministry of the Environment and Forest, and are granted only for education, research, or conservation purposes. The Figure 2. Fire breaks created and maintained by villag- species was listed as Endangered (EN) in the 1990s in the ers, following the park boundary in Andranomatavy (Photo: IUCN Red List, and uplisted to Critically Endangered (CR) Angelo Ramy Mandimbihasina).

Solitaire No. 27 31 Conservation report

analysis, reporting and facilitate decision-making. Baly Bay National Park began investing in SMART at the end of 2014 and it was successfully in use a year later despite a lack of equipment. As ploughshare habitats are very remote, no electricity generation facilities are in place yet, and it is still a big challenge to charge batteries in the field for the use of GPS. Investing in efficient solar panels and batteries is an option that could help our field team. In addition, the use of smartphones with power banks and/or small solar panels should help. The use of smartphones for in-field mobile data entry can greatly reduce time-consuming and error-prone and we will be trialling “Cybertracker” over the coming months. Figure 3. Photo taken by a camera trap on 08 April 2016 at 09:03 pm inside a ploughshare habitat. Fires have been seen inside Baly Bay National Parks even when firebreaks were made properly following park boundaries and limits of core zones (Figure 2). Fires inside ploughshare habitats have been noticed during patrols; this can show the patrol effort in terms both by fire alerts and patrollers. During 2016, we have of time spent and distance walked. Patrollers are noted an increased number of fires and burnt areas supervised by a responsible person from the Durrell compared to 2015. team and also by Park Rangers. Information gathered by the village patrols has led to As Baly Bay is a National Park, there are eight Park arrests. Some of the people that were arrested were Rangers undertaking patrols there, including within successfully prosecuted, went to jail and were fined for ploughshare habitats. Their patrol data are also put entering the core zone of the National Parks, even if they into SMART. They undertake supervision of the villagers had not taken any tortoises, which is very encouraging! doing patrols and also ensure that villagers are helping to maintain fire breaks around the habitats and park Camera-traps have shown good results and people boundaries. They also mark trees with permanent have been arrested directly because of the images. enamel paint in order to let people know where park Figure 3 shows two poachers captured by a camera boundaries and core zones are. trap in one of the ploughshare habitats in April 2016, at night. The third type of patrol is done with the military. During the ploughshare’s active period, three groups of armed Conclusion people from the military along with Park rangers, The information and data arising from these actions representatives of the Ministry of Ecology, Environment show that poaching of A. yniphora exists and continues. and Forest, and a team member from Durrell also Poachers may not come in when military personnel carry out patrols. Patrol data from these teams are also are present on site, so a permanent presence and entered into SMART in order to measure their effort and patrols by armed guards may be necessary in order to to better understand the effectiveness of these patrols ensure protection of this species in the wild. However, all conservation actions should be continued and To enhance security, camera traps have been installed reinforced with awareness-raising activities in order to inside the ploughshare habitats since 2012 in order to get good results. Efforts should be put into the use of take pictures of poachers and trespassers. GPS and adoption of and training in technology so that patrol effort can be evaluated and results can flow more Conservation impact rapidly, ensuring that appropriate responses to fires and Patrols are one of the ways in which we can reduce incursions by poachers can be taken more quickly. poaching activities, but the known trails inside the park and other ploughshare habitats only enable us Angelo Ramy Mandimbihasi- to access about 7000 ha, which is about 45% of the na started working with Durrell entire habitat. Creating new trails would mean creating in 2000 as a student studying more access into the habitats, which might backfire by the population genetics of the giving improved access for poachers, so it is important ploughshare tortoise. He joined that we find ways to conduct patrols without following us as Baly Bay Project Manager defined trails and increase the reach of the existing in 2005, and received training patrols. Even with this coverage, poaching signs have in Jersey in 2009. Since 2010 been seen and reported by patrollers. Angelo has focused on coor- dinating research activities for The use of SMART to monitor patrols, threats and the ploughshare tortoise, in- biodiversity is a new conservation tool to enhance data cluding his own PhD. E-mail: [email protected]

Solitaire No. 27 32 Tortoise seed dispersal Reviving lost interactions: seed dispersal by ecological replacement Aldabra giant tortoises on Ile aux Aigrettes, Mauritius Martin Kastner1, Ysabella Montaño1 and Nicolas Zuël2

1Durrell Wildlife Conservation Trust 2Mauritian Wildlife Foundation

Abstract Giant tortoises are keystone herbivores and seed dispersers on many oceanic islands. Their eradication throughout much of their former range has led to the dispersal limitation of many large-seeded tree species, putting those species at increased risk of extinction. The introduction of ecological replacement tortoises could restore seed dispersal of endangered tree species, which may be crucial to the recovery of their populations. We quantified seed dispersal distances of the critically endangered ebony Diospyros egrettarum by ecological replacement Aldabra giant tortoises (Aldabrachelys gigantea) on Ile aux Aigrettes, Mauritius. Average seed dispersal distance was 85.0 m (range 0–350.3 m). Dispersal by male tortoises was higher than that of females, but the difference was not statistically significant. We recorded a large proportion of long-distance dispersal events, with 40% of recorded dispersals over 100 m, and 6% over 200 m. Our results suggest that dispersal by the ecological replacement tortoises may be depositing most ebony seeds beyond the range of density-dependent mortality effects, potentially encouraging germination and establishment. Long distance dispersal could promote gene flow within the population as well as colonisation of suitable habitat. Our study contributes to the important and ongoing process of documenting the results of ecological replacement experiments. Key words: Diospyros egrettarum, ebony, habitat restoration, islands

Introduction solution to restore a large number of interactions within an Until relatively recently, tortoises were the dominant herbi- ecosystem (Kaiser-Bunbury et al. 2010). Indeed, Gibbs et vores and seed dispersers on many oceanic islands (Hansen et al. (2008) documented the positive effect a giant tortoise al. 2010), but they have been driven to extinction throughout reintroduction had on the recruitment of an endangered most of their former range. Large and giant tortoises play a keystone cactus species in the Galápagos. However, high rates number of important ecological roles within the ecosystems of endemicity on oceanic islands mean that in many cases, they inhabit, including the maintenance of habitat openness local extirpation is synonymous with global extinction (Hansen and heterogeneity through browsing, grazing and trampling 2010). In that case, practical options for management include (Gibbs et al. 2010), the promotion of nutrient cycling (Hunter letting the ecosystem reach a new equilibrium (Corlett 2013), et al. 2013), as well as seed dispersal (Kaiser-Bunbury et al. managing the ecosystem with a human workforce (Griffiths et 2010). al. 2013) and introducing an ecological replacement species Mauritius was historically inhabited by two endemic (Griffiths et al. 2013; Hansen 2010; Kaiser-Bunbury et al. species of giant tortoise of the genus Cylindraspis, itself 2010; Parker et al. 2010). unique to the Mascarene Islands. It appears that both species Results from experiments using non-native tortoise species were particularly concentrated in the lowland palm-dominated or subspecies as ecological replacements on offshore islets in forests, which were likely to have been kept relatively open by Mauritius (Griffiths et al. 2010, 2011, 2013) and Pinta Island their grazing pressure. The tortoises were eradicated from the in the Galápagos (Hunter et al. 2013) have shown positive mainland of Mauritius by the early 18th century, though some impacts on the restoration of ecosystem function and potential individuals persisted on offshore islands into the 19th century. for application elsewhere (Hansen et al. 2010). Their extinction was caused by a combination of intense There is a small but growing literature on the role of hunting and predation by introduced pigs and cats (Cheke tortoises as seed dispersers (Blake et al. 2012; Carlson et al. and Hume 2008). The loss of the Cylindraspis tortoises and 2003; Gibbs et al. 2008; Guzmán and Stevenson 2008; Griffiths other endemic frugivores has led to the dispersal limitation of et al. 2011; Jerozolimski et al. 2009; Strong and Fragoso a suite of tree species in Mauritius (Hansen et al. 2008). On 2006). Tortoises often live at high population densities, spread Ile aux Aigrettes, an offshore islet, Griffiths (2010; Griffiths et relatively evenly over the landscape (Gibson and Hamilton al. 2011) confirmed the absence of recruitment of Diospyros 1984; Strong and Fragoso 2006), eat large quantities of egrettarum, a critically endangered endemic ebony, prior to seeds of many species (Blake et al. 2012; Jerozolimski et al. the introduction of ecological replacement tortoises. 2009), move over relatively long distances (Blake et al. 2012; Reintroducing a recently extirpated generalist seed Guzmán and Stevenson 2008) and rest both in covered and disperser could, in theory, be a relatively straightforward open areas for thermoregulation (Griffiths 2010; Strong and

Solitaire No. 27 33 Research report

Fragoso 2006). Their gut passage may be important for the contains some of the last remnants of ebony-rich lowland germination success of native species (Griffiths et al. 2011; forest in the country (Griffiths et al. 2011; Varnham et al. Rick and Bowman 1961) and can have detrimental impacts 2002). As a nature reserve operated by the Mauritian Wildlife on that of non-natives (Carlson et al. 2003; Rick and Bowman Foundation since 1987, it has been the site of intensive con- 1961; but see Waibel et al. 2013). servation and restoration work, including the eradication of The Janzen-Connell model predicts that seed and seedling black rats – important seed predators – weeding of invasive mortality has an inverse relationship with dispersal distance. plants and the reintroduction of native and endemic plant and Therefore, dispersal away from the parent tree may be crucial animal species (Varnham et al. 2002). Ile aux Aigrettes is for seedling establishment (e.g. Hansen et al. 2008). Griffiths demarcated by a permanent 12.5 m x 12.5 m survey grid. et al. (2011) demonstrated that the introduction of ecological Diospyros egrettarum Richardson (Ebenaceae) is a replacement tortoises on Ile aux Aigrettes was successful critically endangered ebony species endemic to Mauritius in breaking the longstanding recruitment limitation of D. (Page 1998; see Figure 1a). It was once a dominant coastal egrettarum, through the effects of their gut passage as well as hardwood, but logging has reduced it to fewer than 10 dispersal beyond the seed shadow of parent trees. individuals on the mainland. The only viable population is on The only estimates of dispersal distance for tortoises Ile aux Aigrettes, where it has benefitted from the eradication published to date (Blake et al. 2012; Guzmán and Stevenson of rats and weeding of exotic plants (Page 1998), as well as the 2008; Jerozolimski et al. 2009; Strong and Fragoso 2006) introduction of Aldabra giant tortoises (Griffiths et al. 2011). have been derived through inference as opposed to direct Its fruits are large and roughly spherical, each containing 8 ± observation. Our study aimed to fill this gap. Our principal 2 large seeds (see Figure 1b, c), encased in a sticky pulp and aim was to measure the seed dispersal distances of ecological a brittle exocarp (see Griffiths 2010). The species is generally replacement Aldabra giant tortoises on Ile aux Aigrettes. dioecious. Its main historical seed dispersers would have been Dispersal distance depends on a variety of factors including Cylindraspis tortoises and, possibly, Leiolopisma mauritiana, gut passage time, which in turn is affected by fruit pulp, and a giant skink (Griffiths 2010). seed shape and size (C.J. Griffiths, pers. comm.). In order to Aldabra giant tortoises (Aldabrachelys gigantea; see increase the reliability of the results, we focused the study on Figure 1d) were first introduced to Ile aux Aigrettes in 2000 the dispersal of a single species, the fleshy-fruited ebony D. as ecological replacements for the extinct Cylindraspis egrettarum, which is commonly consumed by the tortoises species. They are considered suitable “proxies” given their (see Griffiths 2010; Griffiths et al. 2011). phylogenetic and ecological similarities to the endemic Our primary research question was: how far are Aldabra species (Griffiths et al. 2010). All animals were born and giant tortoises dispersing ebony seeds on Ile aux Aigrettes? raised in captivity, but they adapted well to the natural Hansen et al. (2008) suggest that “even a tortoise will likely setting; they are not fed but are provided water for welfare move a greater distance than 25 m within 1–3 weeks.” reasons, as there is no standing water on the island. They While we did not specifically measure gut passage time, were first held in enclosures, and were then released into our observations in the field indicated that it nearly always the wild when it was deemed that they were not harming the fell within the range of 10–21 days over the course of our native vegetation (Griffiths et al. 2012). There have been as experiment. We therefore predicted that the tortoises would many as 26 non-juvenile tortoises on the island (Griffiths et disperse Diospyros egrettarum seeds significantly further than al. 2012), but at present there are 21 individuals, 10 male 25 metres. Secondly, in their native habitat, male and female and 11 female (including one subadult). All the tortoises on Aldabra giant tortoises have differing spatial distribution the island are individually identifiable, and some of their life patterns (Gibson and Hamilton 1983). We therefore expected history is known (see Griffiths et al. 2012). The individuals ebony seed dispersal distances to be different for male and included in the study were all adults, with a sex ratio of 10 female Aldabra giant tortoises on Ile aux Aigrettes. females to 9 males.

Methods Dispersal distance study Study site and species Our dispersal distance study was conducted between June Ile aux Aigrettes (57o73’05”E, 20o42’03”S) is a 25 ha coral- and August 2014 (see Table 1). Ripe ebony fruit was collect- line islet located 700 m off the southeast coast of Mauritius, ed opportunistically over the island. Only fruits found on the reaching a maximum elevation of 12 m above sea level. It ground – likely candidates for tortoise consumption – were

Table 1. Numbers and ratios of Aldabra giant tortoises fed on Ile aux Aigrettes, by sex and feeding event, and corresponding faecal deposit recovery rates.

Females Total fed (% Males fed tortoises Faeces recovered – Faeces recovered Faeces recovered Dates fed total) (% total) fed females (% fed) – males (% fed) – total (% fed) June 13–14, 2014 8 (47) 9 (53) 17 8 (100) 8 (89) 16 (94) June 26–27, 2014 9 (50) 9 (50) 18 7 (78) 9 (100) 16 (89) July 10–11, 2014 10 (52.5) 9 (47.5) 19 9 (90) 8 (89) 17 (89.5) Overall 27 (50) 27 (50) 54 24 (89) 25 (92.5) 49 (91)

Solitaire No. 27 34 Tortoise seed dispersal

Figure 1. (a) Diospyros egrettarum, showing signs of dispersal limitation; (b) D. egrettarum fruit; (c) D. egrettarum seeds; (d) Al- dabrachelys gigantean; (e) Feeding A. gigantean; (f) faeces containing coloured pellets and D. egrettarum seeds. All photos: M. Kastner 2014.

collected. Fruit that appeared overripe, desiccated or dam- fed twice and another only once, because those individuals aged was avoided, and gathered fruit was never kept longer could not be located at the time. than three days. In order to feed the tortoises, we located them and Two tortoises were excluded from the study: one, a large presented them with 12–13 ebony fruits, along with a unique male, because he appeared to be lethargic, and the other, a combination of coloured, non-toxic plastic pellets (1 tbsp subadult female, in order to maintain consistency in age class. per colour; 1–5 mm in diameter; Albert GmbH and Co. KG, The remaining tortoises (n=19; 10 female and 9 male) were Bünde, Nordrhein-Westfalen), in a shallow metal plate (see each fed a total of three times, with two exceptions. One was Figure 1e). The ebony fruits were fed to the tortoises to

Figure 2. Maps of Ile aux Aigrettes showing (a) Aldabra giant tortoise feeding locations and (b) faecal deposits. Note the greater number of deposits than feeding locations.

Solitaire No. 27 35 Research report

Figure 3. Dispersal distances by Aldabra giant tortoises on Ile aux Aigrettes by (a) sex and (b) feeding event.

encourage corresponding “natural” gut passage times, and Results hence representative dispersal distances. The pellets were Dispersal distances encrusted in a banana to encourage consumption (Waibel A total of 143 faecal deposits were found (µTOTAL = 2.6 et al. 2013). The tortoises always ate the banana and almost deposits/feed; µMale = 2.5, µFemale = 2.8) for 57 feeding always ate all (and never less than half) of the ebony fruits. events (see Figure 2). Faecal deposits were found for 91% of The coordinates of each feeding location were recorded using feeding events (Feed 1: 94%, Feed 2: 89%, Feed 3: 89.5%; a global positioning system (GPS) device (eTrex 20; Garmin, see Table 1). Schaffhausen). On average, tortoises dispersed seeds 85.0±62.1 m away Faecal deposits were located systematically as well as from the site of ingestion. Male dispersals (96.2±62.6m) opportunistically. Systematic searches were conducted over were longer than those of females (75.1±60.4m; see Figure the entire area of known tortoise distribution, plus a one grid- 3a), but not significantly so (F = 1.883; p = 0.193). Average square buffer surrounding that area. Any tortoise faeces found female dispersal distance was skewed by two exceptionally during searches were examined for coloured pellets (see Figure long dispersal events, of 276.5 and 350.3m (the next-longest 1f) and broken apart if necessary – it was recorded if at least female dispersal was of 172.5m). Dispersal distances were one identifiable pellet of each colour was found within the higher for the third feeding event (105.6±77.1m; see Figure faecal matter. For each deposit, we recorded GPS coordinates, 3b) than for the first (76.4±53.8m) or the second (71.3±45.1m), pellet colours, faeces appearance (fresh or not fresh), whether but not significantly so (F = 3.052; p = 0.0517). or not it contained ebony seeds, and any other noteworthy Overall, the vast majority of recorded dispersals (88.11%) contents. A deposit was considered unique if it was separated were over 25 m (males: 89.55%; females: 86.84%), and the by a minimum of 1m from its nearest neighbour. square-root transformed data (n = 143; µ= 8.56; SD = 3.4)

Table 2. Dispersal distances by Aldabra giant tortoises on Ile aux Aigrettes in distance classes <25m, 25–100m, 100–200m, and >200m, by sex and feeding event.

Dispersal distance <25m 25–100m 100–200m >200m Total

Dates fed F M Total F M Total F M Total F M Total F M Total

June 13–14, 2014 4 2 6 11 7 18 9 7 16 0 0 0 24 16 40 June 26–27, 2014 2 3 5 21 11 32 3 11 14 0 1 1 26 26 52

July 10–11, 2014 4 2 6 9 11 20 11 7 18 2 5 7 26 25 51

Overall 10 7 17 41 29 70 23 25 48 2 6 8 76 67 143 (% total) (13) (10) (11) (54) (43) (49) (30) (37) (34) (3) (9) (6) (100) (100) (100)

Solitaire No. 27 36 Tortoise seed dispersal

Table 3. Generalised linear mixed effects models of their respective sex (see Figure 4a). Males and females assessing dispersal distance by tortoise sex (fixed were therefore not clustered by sex when we located them effect), and different combinations of random effects. for feeding (F = 0.264; p = 0.615). K: number of parameters, LL: log-likelihood, AIC: Male tortoises tended to be located further (µMales Akaike’s Information Criterion, ΔAIC: differences in AIC. = 73.6±45.1m), on average, than females (µFemales = 59.2±30.1m), from the nearest water point (see Figure 4b), although the difference was not significant (F = 0.727; p = Random effects K LL AIC ΔAIC 0.406). ID + FEED 4 -1987 3981 0 ID + FEED + RAINFALL 5 -1944 3987 6 Discussion Coloured plastic pellets have been used with tortoises to ID + RAINFALL 4 -1997 4003 22 measure gut passage times (Blake et al. 2012) and to mark ID 3 -2143 4291 310 individuals in a germination experiment (Waibel et al. 2013). Our study is novel in using plastic pellets to measure seed dispersal distances in situ. We recovered pellets for 91% of individuals fed (see Table 1). This result suggests that the method is viable for collecting robust dispersal distance datasets from free-ranging tortoises. Their tendency to avoid was significantly greater (t = 9.136; p < 2.2 x 10-16) than dense woody vegetation (Gibbs et al. 2008) and to rest in a simulated normal distribution of equal sample size and clearings (Strong and Fragoso 2006) may help in achieving variance. Nearly half (46.27%) of the males’ dispersals were high recovery rates. The method is nevertheless work-inten- over 100m (females: 32.89%), and 8.96% were over 200m sive, and benefited, in our case, from a relatively small and (females: 2.63%; see Table 2). high-density population. Recovery is most efficient when tortoises are visited daily over the expected period of gut Mixed-effects model passage. A generalised linear mixed-effects model for dispersal dis- Strong and Fragoso (2006) estimated a minimum tance, with SEX as a fixed effect and ID and FEED as random dispersal distance of 91.2 m for Geochelone carbonaria and effects, was selected based on its relative AIC as compared G. denticulata in the Brazilian Amazon, by multiplying the to other iterations of the model with different combinations mean daily dispersal by the minimum gut passage time they of random effects. AIC dropped substantially when FEED recorded. This method may be problematic, since it assumes was added as a random effect. RAINFALL did help model fit linearity in movement over consecutive days. Guzmán and when added to ID, though not as much as FEED; including Stevenson (2008) recorded a similar figure (89.6 ± 9.2 m) it alongside ID and FEED only complicated the model (see for G. deniculata in Peru, whereas Jerozolmiski et al. (2009) Table 3). SEX was not a significant predictor of dispersal estimated mean dispersal distances of 174.1 m in the rainy distance (p = 0.479), but it did account for 53.43% of the season and 276.7 m in the dry season for the same species. variation with the standard error of the model’s fixed effects. The latter study overlaid the range of gut passage times they recorded in captivity with movement data measured in the Nearest-neighbour analyses field to produce potential “seed shadows”. Blake et al. (2012) On average, male tortoises were located 45.8±32.9m, and used similar methodology on Chelonoides nigra on Santa Cruz female tortoises 53.9±43.9m, away from the nearest member Island in the Galápagos to produce an average seed dispersal

Figure 4. Distance of each Aldabra giant tortoise fed on Ile aux Aigrettes to the nearest (a) member of the same sex, an indicator of clustering, and (b) water point.

Solitaire No. 27 37 Research report

distance of 394 m. We recorded a mean of 85.0 m, with a that tortoises are effective seed dispersers, in that, beyond minimum of 0 m and a maximum of 350.3 m. Given the range dispersing over relatively long distances, they also consume of species, habitats and methods used to produce the estimates, large quantities of a wide variety of fruit, their digestion is it is perhaps less remarkable that estimates differ substantially relatively benign and may sometimes help in germination, than that they are all within the same order of magnitude. their populations are often dense but scattered over the The vast majority of recorded dispersals (over 88%; see landscape, and they occupy a wide range of microhabitats Table 2), were longer than 25 m, and mean dispersal distance (Jerozolimski et al. 2009; Strong and Fragoso 2006). We was significantly higher than 25 m. We can therefore reject observed dispersal by Aldabra giant tortoises of a variety of the null hypothesis that Aldabra giant tortoise ebony seed large-seeded native and non-native species, including Eugenia dispersal distance on Ile aux Aigrettes is less than or equal on a regular basis, as well as Dracaena, Hibiscus, Scaevola, to 25 m, and confirm the accuracy of Hansen et al.’s (2008) Morinda and Acacia species. Griffiths et al. (2011) have prediction that tortoise seed dispersal tends to be longer than demonstrated the beneficial effects of tortoise gut passage that distance. This suggests that dispersal by the ecological for D. egrettarum. Those benefits may extend to other fleshy- replacement tortoises is generally far enough to allow ebony fruited species potentially prone to fungal infection, such as seeds to escape most density-dependent mortality factors Eugenia. Tortoises on Ile aux Aigrettes certainly do deposit (Hansen et al. 2008), although some density-dependent seeds over a full gradient of shaded to open microhabitats effects are certainly manifested within tortoise-dispersed (pers. obs.). It seems reasonable to hypothesise, therefore, that seed clusters (Griffiths 2010). This interpretation is based on Aldabra giant tortoises are acting as effective seed dispersers the assumption that the Janzen-Connell model applies to this on Ile aux Aigrettes. On the other hand, their impact will tree species. The 25 m threshold may be conservative, since logically be restricted to their area of occupancy, and it does the diameter of an average D. egrettarum canopy on Ile aux appear that their range is restricted on the island (see Figure Aigrettes, the seed shadow within which most non-dispersed 2). Their potential habitat may be confined by access to seeds are confined, is probably smaller than that distance. water, locally dense vegetation or barriers of coralline rock; Our findings support the evidence compiled by Griffiths alternatively, they may not be exploiting all available habitats et al. (2011) demonstrating that dispersal by Aldabra giant if the population is below its carrying capacity. tortoises is helping ebony on Ile aux Aigrettes to overcome Sex was not a significant predictor of dispersal distance. long-standing recruitment limitation caused by the extinction However, it does account for a substantial proportion of the of their historical dispersal agents. It should be noted that the variation within the model, and male tortoises did disperse removal of pulp during gut passage, and deposition within seeds farther, on average, than females. Figure 2 clearly shows nutrient-rich and moisture-retaining faecal matter, are other differential distribution of tortoises by sex on the island, with potentially important factors contributing to the germination females especially concentrated near buildings by the western and eventual survival of seedlings (Griffiths 2010). coast and males more common inland. Cain et al. (2000) stress the critical importance of long- Although the result was not statistically significant, male distance seed dispersal for plant population dynamics, and the tortoises were generally encountered further from water points scarcity of studies capturing the tails of seed dispersal curves. than females. Fresh water appears to be the major motivating Nearly 40% of the dispersal distances we recorded were longer factor for long distance movements by tortoises on Ile aux than 100m, and 6% were longer than 200m: such dispersals can Aigrettes (Z. Ahamud and N. Zuël, pers. comm.), and it is be considered as “long distance” (Cain et al. 2000) and “very possible that male tortoises are travelling relatively long long distance” (Blake et al. 2012), respectively. Long distance distances on a regular basis to drink water, with important seed dispersals contribute significantly to dynamics and gene implications for seed dispersal. Gibson and Hamilton (1983) flow of tree populations (Cain et al. 2000; Jerozolimski et al. noted differential habitat use by sex in Aldabra giant tortoises 2009). They may be especially important for dioecious species in their native range. While the species is relatively gregarious, such as D. egrettarum, which need to compensate for the loss some level of home range fidelity by the tortoises on Ile aux of one sex role relative to their hermaphroditic competitors Aigrettes in apparent. Differential home range size (Diemer (Griffiths 2010). Long distance dispersals may also contribute 1992) and movement patterns (McRae et al. 1981) by sex to maintaining species diversity within a forest (Janzen and have been documented for Gopherus polyphemus in Florida. Martin 1982; Terborgh et al. 2008), and to the restoration of Furthering our understanding of tortoise behaviour, including disturbed landscapes (Wunderle 1997). D. egrettarum was sex-dependent variations therein, will help us refine seed logged on some portions of Ile aux Aigrettes (Griffiths et al. dispersal models in the future (Russo et al. 2006). Skewing 2011) and tortoise-mediated dispersal may help the species to the sex ratio of populations is common practice in animal re-establish on suitable habitat. It will only be possible to truly translocations (Armstrong and Reynolds 2012), especially for assess the effect of tortoise seed dispersal on such ecosystem the purpose of maximising breeding rates. If the goal of the processes in the very long term, as ebony trees are particularly translocation is the restoration of ecological processes, as is the slow-growing, and have long generational times. Nevertheless, case in rewilding projects (Seddon et al. 2014), then managers our results support the body of evidence demonstrating that should consider the possibility of sex-related differences in the introduction of ecological replacement tortoises on Ile aux contribution to ecosystem function within a species. Aigrettes is reviving important plant-animal interactions on It would be useful to expand this study to cover seasonal the island (Hansen et al. 2010; Kaiser-Bunbury et al. 2010). and annual variations in environmental conditions and ebony There are a number of components to seed dispersal fruit availability. It is also difficult to assess the applicability effectiveness beyond dispersal distance, such number of of the method to other locations. The small size of the seeds dispersed, effects of gut passage and faecal matter, and tortoise population and restricted island area may have been deposition site (Schupp 1993). Several authors have suggested advantageous in achieving a high recovery rate for pellets.

Solitaire No. 27 38 Tortoise seed dispersal

Difficulties may be encountered, however, in areas with rapid Gibson, CWD and Hamilton, J (1983). Feeding ecology and decomposition rates and/or difficult access. Nevertheless, seasonal movements of giant tortoises on Aldabra atoll. our study is the first to measure precise point-to-point seed Oecologia 56: 84–92. dispersal distances for tortoises in a natural setting, and our Gibson, CWD and Hamilton, J (1984). Population processes results contribute to the existing evidence suggesting that in a large herbivorous reptile: the giant tortoise of Aldabra the ecological replacement tortoises are acting as effective atoll. Oecologia 61: 230–240. seed dispersers on the island. They appear to be filling at Griffiths, CJ (2010). Conservation and Restoration of Mauritian least part of the ecological niche left vacant by the extinction Plant Communities using Taxon Substitutes. PhD thesis. of the native Cylindraspis tortoises, and contributing to the Bristol: University of Bristol. restoration of critical plant- animal mutualisms. However, the Griffiths, CJ, Jones, CG, Hansen, DM, Puttoo, M, Tatayah, RV, success of this ecological replacement experiment, and others Müller, CB and Harris, S (2010). The use of extant non- that may be inspired by it, will only be confirmed in the long indigenous tortoises as a restoration tool to replace extinct term, and the perceived benefits should always be weighed ecosystem engineers. Restoration Ecology 18: 1–7. against any potential or actual negative impacts. Griffiths, CJ, Hansen, DM, Jones, CG, Zuël, N and Harris, S (2011). Resurrecting extinct interactions with extant Acknowledgements substitutes. Current Biology 21: 762–765. Many thanks to Christine Griffiths for feedback. Thank you Griffiths, CJ, Zuël, N, Tatayah, V, Jones, CG, Griffiths, O and Zairabee Ahamud for teaching us all about the tortoises on Harris, S (2012). The welfare implications of using exotic Round Island. Thanks to Mark O’Connell and Rosemary tortoises as ecological replacements. PloS One 7: Moorhouse-Gann (and Nicolas Zuël of course) for advice on e39395. statistics. Jamie Copsey, thank you for facilitating the whole Griffiths, CJ, Zuel, N, Jones, CG, Ahamud, Z and Harris, S process. MK’s participation in the 2014 Durrell Postgradu- (2013). Assessing the potential to restore historic grazing ate Diploma in Endangered Species Recovery course was ecosystems with tortoise ecological replacements. supported by a New Noah scholarship from Wildlife Preser- Conservation Biology 27: 690–700. vation Canada. Guzmán, A and Stevenson, PR (2008). Seed dispersal, habitat selection and movement patterns in the Amazonian References tortoise, Geochelone denticulata. Amphibia–Reptilia 29: Armstrong, DP and Reynolds, MH (2012). Modelling 463–472. reintroduced populations: the state of the art and future Hansen, DM (2010). On the use of taxon substitutes in rewilding directions. In Ewen, JG, Armstrong, DP and Parker, KA projects on islands. In: Pérez-Mellado, V and Ramon, C (eds). Reintroduction Biology: Integrating Science and (eds). Islands and Evolution. Menorca: Institut Menorquí Management. Chichester: John Wiley and Sons, 165– d’Estudis, 111–146. 222. Hansen, DM, Donlan, CJ, Griffiths, CJ and Campbell, KJ Bates, D and Maechler, M (2009). lme4: Linear mixed-effects (2010). Ecological history and latent conservation models using S4 classes. R package version 0.999375- potential: large and giant tortoises as a model for taxon 31. Available at: http://lme4.r-forge.r-project.org/. substitutions. Ecography 33: 272–284. Blake, S, Wikelski, M, Cabrera, F, Guezou, A, Silva, M, Hansen, DM, Kaiser, CN and Müller, CB (2008). Seed dispersal Sadeghayobi, E, Yackulic, CB and Jaramillo, P (2012). and establishment of endangered plants on oceanic Seed dispersal by Galapagos tortoises. Journal of islands: the Janzen-Connell model, and the use of Biogeography 39: 1961–1972. ecological analogues. PLoS One 3: e2111. Cain, ML, Milligan, BG and Strand, AE (2000). Long-distance Hunter, EA, Gibbs, JP, Cayot, LJ, and Tapia, W (2013). seed dispersal in plant populations. American Journal of Equivalency of Galápagos giant tortoises used as Botany 87: 1217–1227. ecological replacement species to restore ecosystem Carlson, JE, Menges, ES and Marks, PL (2003). Seed dispersal functions. Conservation Biology 27: 701–709. by Gopherus polyphemus at Archbold Biological Station, Janzen, DH and Martin, PS (1982). Neotropical anachronisms: Florida. Florida Scientist 66: 147–154. the fruits the gomphotheres ate. Science 215: 19–27. Cheke, A and Hume, JP (2008). The Lost Land of the Dodo: An Jerozolimski, A, Ribeiro, MBN and Martins, M (2009). Are Ecological History of Mauritius, Réunion and Rodrigues. tortoises important seed dispersers in Amazonian forests? New Haven: Yale University Press. Oecologia 161: 517–528. Corlett, RT (2013). The shifted baseline: Prehistoric defaunation Kaiser-Bunbury, CN, Traveset, A and Hansen, DM (2010). in the tropics and its consequences for biodiversity Conservation and restoration of plant–animal mutualisms conservation. Biological Conservation 163: 13–21. on oceanic islands. Perspectives in Plant Ecology, Diemer, JE (1992). Home range and movements of the Evolution and Systematics 12: 131–143. tortoise Gopherus polyphemus in northern Florida. Journal McRae, WA, Landers, JL and Garner, JA (1981). Movement of Herpetology 26: 158–165. patterns and home range of the gopher tortoise. Gibbs, JP, Marquez, C and Sterling, EJ (2008). The role American Midland Naturalist 106: 165–179. of endangered species reintroduction in ecosystem Page, W (1998). Diospyros egrettarum. The IUCN Red List of restoration: tortoise–cactus interactions on Española Threatened Species. Version 2014.2. www.iucnredlist.org Island, Galápagos.Restoration Ecology 16: 88–93. (downloaded on 17 September 2014). Gibbs, JP, Sterling, EJ and Zabala, FJ (2010). Giant tortoises Parker, KA, Seabrook-Davison, M and Ewen, JG (2010). as ecological engineers: a long-term quasi-experiment Opportunities for nonnative ecological replacements in in the Galápagos Islands. Biotropica 42: 208–214. ecosystem restoration. Restoration Ecology 18: 269–273.

Solitaire No. 27 39 Research report

Rick, CM and Bowman, RI (1961). Galapagos tomatoes and Invasive Species Specialist Group, 342–349. tortoises. Evolution 15: 407–417. Waibel, A, Griffiths, CJ, Zuël, N, Schmid, B, and Albrecht, M Russo, SE, Portnoy, S and Augspurger, CK (2006). Incorporating (2013). Does a giant tortoise taxon substitute enhance animal behavior into seed dispersal models: implications seed germination of exotic fleshy-fruited plants? Journal for seed shadows. Ecology 87: 3160–3174. of Plant Ecology 6: 57-63. Schupp, EW (1993). Quantity, quality and the effectiveness of Wunderle Jr, JM (1997). The role of animal seed dispersal seed dispersal by animals. In: Fleming, TH and Estrada, in accelerating native forest regeneration on degraded A (eds). Frugivory and Seed Dispersal: Ecological and tropical lands. Forest Ecology and Management 99: 223- Evolutionary Aspects. Netherlands: Springer, 15–29. 235. Seddon, PJ, Griffiths, CJ, Soorae, PS and Armstrong, DP (2014). Reversing defaunation: Restoring species in a changing Martin Kastner and Ysabella Montaño carried out world. Science 345: 406–412. this study as part of their PG Dip course in Endan- Strong, JN and Fragoso, J (2006). Seed dispersal by gered Species Recovery with Durrell Conservation Geochelone carbonaria and Geochelone denticulata Academy, Mauritius in 2014. Martin participated in in northwestern Brazil. Biotropica 38: 683–686. the course on a New Noah scholarship from Wildlife Terborgh, J, Nuñez-Iturri, G, Pitman, NC, Valverde, FHC, Preservation Canada. He is currently based in Guam Alvarez, P, Swamy, V, Pringle, EG and Paine, CT (2008). Tree working as an avian restoration project manager for recruitment in an empty forest. Ecology 89: 1757–1768. Colorado State University. Ysabella is completing her Varnham, KJ, Roy, SS, Seymour, A, Mauremootoo, J, Jones, MSc in Environmental Biology at Suranaree University CG and Harris, S (2002). Eradicating Indian musk shrews of Technology, Thailand. Nicolas Zuël is Fauna Man- (Suncus murinus, Soricidae) from Mauritian offshore islands. ager at the Mauritian Wildlife Foundation. In: Veitch, CR and Clout, MN (eds). Turning the Tide: The Email: [email protected] Eradication of Invasive Species. Gland, Switzerland: IUCN Learn from the conservation experts

We are world leaders in saving species from extinction and empowering others to run conservation projects worldwide, operating right at the interface between animals in the wild and those in captivity.

Our courses We cover a wide variety of topics and cross-disciplinary skills relevant to successful conservation and wildlife management. From 3–5 day courses to a 6-month Post Graduate Diploma, the Academy provides career-long training services in Jersey, Mauritius, UK and online.

Endangered species recovery techniques Technical training in topics such as GIS, education, husbandry, conservation management and leadership

Specific information, course outlines and dates can be found online at www.durrell.org To apply, please email [email protected]

Solitaire is edited by Eluned Price E-mail: [email protected]

Contact us: Durrell Wildlife Conservation Trust, Les Augres Manor, Trinity, Tel: +44 (0) 1534 860000 Jersey, Channel Islands, Fax: +44 (0) 1534 860001 JE3 5BP www.durrell.org