Rise and Fall of Aldabran Giant Tortoise Bourn Et Al

8/3/2019 Rise and Fall of Aldabran Giant Tortoise Bourn Et Al

http://slidepdf.com/reader/full/rise-and-fall-of-aldabran-giant-tortoise-bourn-et-al 1/11

The Rise and Fall of the Aldabran Giant Tortoise PopulationAuthor(s): David Bourn, Charlie Gibson, Dave Augeri, Cathleen J. Wilson, Julia Church, SimonI. HaySource: Proceedings: Biological Sciences, Vol. 266, No. 1424 (Jun. 7, 1999), pp. 1091-1100Published by: The Royal SocietyStable URL: http://www.jstor.org/stable/51351 .

Accessed: 07/09/2011 11:40

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

The Royal Society is collaborating with JSTOR to digitize, preserve and extend access to Proceedings:

Biological Sciences.

http://www.jstor.org

http://www.jstor.org/action/showPublisher?publisherCode=rsl

http://www.jstor.org/stable/51351?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

http://www.jstor.org/page/info/about/policies/terms.jsp

http://www.jstor.org/stable/51351?origin=JSTOR-pdf

http://www.jstor.org/action/showPublisher?publisherCode=rsl



[ -THE ROYALE94 SOCI ETY

T h e r i s e a n d f a l l o f t h e Aldabran g i a n t tortoise

population

David Bourn', Charlie Gibson2, Dave Augeril, Cathleen J. Wilson', Julia Churchl

and Simon I. Hay3

'EnvironmentalResearchGroupOxfordLimited, PO Box 346, OxfordOX] 3QE, UK (bournnordergo@,compuserve.com)2BioscanLimited,StandingfordHouse,CaveStreet,OxfordOX4iBA, UK

3 TALA ResearchGroup,DepartmentofZoology, Universityof Oxford,South Parks Road, OxfordOX] 3PS, UK

At the end of the 19th century, after prolonged and extensive harvesting, indigenous giant tortoises had

been eliminated from all islands in the Indian Ocean, except Aldabra atoll, where only a few survived.

With greatly reduced levels of exploitation during the 20th century, the population recovered to a

revised estimated total of 129 000 in 1973-1974, when the first sample census was conducted. A repeat

census in 1997 revealed a highly significant reduction in numbers over the past 24 years to an esti-

mated total of 100 000. The great majority of tortoises are still found at relatively high density in

south-eastern Grande Terre, where the number of animals has declined by more than one-third. In

contrast, low-density subpopulations on Malabar and Picard have almost doubled in size, but they

represent less than 5% of the total population. Corroborative evidence for the crash in the Grande

Terre subpopulation comes from two independent observations: a significant increase in tortoise

mortality; and a significant decline in tortoise counts on long-term population monitoring transects.

These population changes are attributed to natural population regulatory mechanisms, exacerbated by

low rainfall years in the period 1980-1997, including two consecutive years of below average rainfall

in 1995-1996 and 1996-1997.

Keywords: Aldabra; giant tortoise Geocheloneigantea Schweigger; population census; monitoring;

Seychelles; Indian Ocean

1. INTRODUCTION

Scientific interest in the giant tortoises (Geochelone igantea

Schweigger) of Aldabra atoll in the western Indian Ocean

(figures 1 and 2) and concern for their survival date back

more than a century (Gunther 1877). Fears of excessive

harvesting and possible extinction of the species prompted

eminent scientists of the day, including Hooker, Owen,

Darwin, Newton and Gunther, to petition the Governor

of Mauritius and the Seychelles in April 1874 to ensure

their protection (Stoddart & Peake 1979).

By the end of the 19th century, there is little doubt that

Aldabra's tortoise population had been severely depleted

by repeated collection (figure 3a), removal and consump-

tion by man (Rothschild 1915). Wharton (1879) stated

that 'the reptiles are now very scarce' and that a party of

his sailors was only able to find one specimen'after much

trouble and search'. In 1892 Spurs, the resident lessee,

reintroduced tortoises to Ile Picard from the 'principal

section' (Grande Terre?) and claimed that there 'were

more than 1000 tortoises' on the atoll (Riseley-Griffith

1892). However, Abbott (1893), after a stay of four

months, believed this to be a 'considerable overestimate'

Voeltzkow (1895) found six tortoises near Dune d'Messe,and only three others during a four-day search of the

south-east of the atoll. Roberts (1905), after a four-month

stay, commented that most tortoises were to be found on

Grande Terre, but that he had not seen many. Fryer's

diary of a six-month sojourn on Aldabra in 1908 makes

little mention of tortoises in areas where they are now

common (Stoddart 1971). Nicoll (1908) claimed that

tortoises were restricted to the northern side of the atoll.

Fryer (1911, p. 420) reported that the tortoise 'still occurs

in fair numbers on the extreme east of the Main Island

[Grande Terre] and is scattered in small numbers in the

rest of the atoll. It is also found on Malabar and Picard

islands but is stated to have been introduced .., only two

were found on Malabar, both young specimens'. Eighteen

years later, Dupont (1929, p. 17) reported that he saw 'a

great many tortoises all over the place' and that tortoises

had 'selected the Cinq Cases area for their breeding

ground almost to the exclusion of other parts of the atoll'.

Public concern was reawakened in the mid-1960s,

when plans for development of Aldabra as a military

base became known (Gaymer 1968; Stoddart 1968a,b;

Beamish 1970). An international campaign to 'save'

Aldabra was launched and, following a strategic review

of Britain's defence policy east of Suez and devaluation

of the pound sterling, plans for military development of

the atoll were cancelled in 1967. After an initial

expeditionary phase in 1967-1969, the Royal Societyestablished a research station on the atoll in 1971, and

continued to manage a programme of studies to

investigate the terrestrial ecology of Aldabra until 1980.

Proc. R. Soc. Lond. B (1999) 266, 1091-1100

Received 1December 1998 Accepted 8 January 19991091 C)1999 The Royal Society



1092 D. Bournand others Reassessing ldabra'sgiantortoiseopulation

40E 50E

N\/ / 0 250 500

kilometres

SOMALI

-0 0 -!~~~~~~~~~ N

KENYA / // / A

iL / / *

0Pemba Mahe

< QZanzibarZ

t

Z |?Mafia N SEYCHELLESNNAldabra

'IU.l. N9Q A

-ios~~~~~~I

os-

m COMOROS

J \ ~~MADAGASCAe

20S /MAURITIUSO20S

-

/ / ~~~REUNIONt40E | 50E

Figure1. LocationmapofAldabraatollin thewestern ndianOcean.Dashed ineindicates heextentof theSeychelles xclusive

economicmaritime one.

Operations were financed by a ParliamentaryGrant-in-Aid, with additional support from the US NationalAcademy of Sciences, the Smithsonian Institute and the

World Wildlife Fund (now the World Wide Fund for

Nature).A key component of the researchprogramme relatedto

the ecology and dynamicsof the giant tortoisepopulation.After initial investigations by Gaymer (1968, 1973) andGrubb (1971), long-termtortoise-monitoringprogrammecommenced with a sample census in 1973-1974, and the

semi-permanent marking of some 7882 individuals withnumberedtitanium discs (Bourn & Coe 1978).A series ofinterrelated studies followed, focusing on interactions

between tortoises and vegetation (Merton et al. 1976;Hnatiuk et al. 1976);reproduction(Bourn 1977;Swingland& Coe 1978);mortality (Bourn & Coe 1979);recruitment(Swingland & Coe 1979); biomass, production and

carrying capacity (Coe et al. 1979);feeding and seasonal

movement (Hamilton & Coe 1982; Gibson & Hamilton1983); and concluding with a synthesis of populationprocesses (Gibson& Hamilton 1984).

Since 1980, Aldabra has been managed by theSeychelles Islands Foundation(SIF), as a Special Reserveunder the Seychelles National Parks and Conservancy

Act, and the tortoise population has been assessed inter-mittently using strip transects. In recognition of itsunique ecological and geological attributes,Aldabrawas

designatedas a World Heritage Site by UNESCO in 1982(Seatonet al. 1991;Amin etal. 1995).

This paper presents the findings of a repeat samplecensus of the Aldabran giant tortoise population during

July and August 1997 (early dry season) undertakenonbehalf of SIF by the Environmental Research GroupOxford Limited (ERGO 1997) as part of the SeychellesBiodiversityConservationand Marine Abatement Project

(GEF 1992).

Proc.R. Soc.Lond.B (1999)



ReassessingAldabra'sgiant ortoise opulation D. Bourn and others 1093

PICARD POLYMNIE MALABAR

a~~~~~~~~~ ~

8 *17/4 and 997 157

_ago on C(15

kilometrespredominan vegetati nq

GRANDEERRE mang sCases

t.0t' - ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~akmk

............ Southmiedscrb,grsslndandoter

Figure... ......ortoise2censuschectares.......::::::::.:::::. ...1973/74 and 1997 (157)

< O ~ ~ ~ 8 a1973/74 only (135)

kilometres predominant vegetationGRANDETERRE m3mangroves

:1pemphis scrub

o mixed scrub, grassland and others

Figure 2. Tortoise census hectares.

2. CENSUS METHOD

(a) Hectare sample plots

The first census of giant tortoises on Aldabra was carried out

over an 18-month period from April 1973 to September 1974,in

conjunction with a large-scale marking programme, and was

based on a 5% stratified random sample of 292 ha shown in

figure 2. Each sample hectare was located by means of aerial

photographstaken in 1960, demarcatedby a perimeter rope laid

out on compass bearings, and searched thoroughly by three

people. All live tortoisesand carapace remains found were indi-

vidually measuredand recorded (Bourn & Coe 1978).

Essentially the same method was used in July and August

1997.The only differences were that because of time constraints:

circular, rather than square, hectare sample plots were used;

two teams of three people, rather than one, were employed as

search parties; and sample intensity in the south-east and south

coast regions of GrandeTerrewas reducedfrom 5% to 2.5%.

The geographical coordinatesof the original sample hectares

were selected randomly and numberedindividually from 1-292.

To avoid bias in subsample selection, only odd-numbered

hectareson GrandeTerrewere resampled, therebyestablishinga

repeated-measures sample of the original random selection.

Eighty-four out of the original 168 ha in the south-east and 46out of the original 92 ha along the south coast of Grande Terre

were re-examined. All 32 of the original hectares on Malabar

were resampled. Thus, in total, 162 out of the original 292 ha

were includedin the 1997 census.

Copies of the original aerial photographswere used to relo-

cate sample hectares. Quite unexpectedly, it was found that

most of the original hectare site markers, consisting of rock

cairns, thin mangrove poles and engraved plastic tags held in

place with stainless steel wire, had survived 24 years of tropical

exposure (figure 3b). It was possible, therefore, to positively

identify the majorityof re-examined hectares.

(b) Information collectedThe Universal TransverseMercator (UTM) coordinates of

sample hectareswere determined by means of hand-held Global

Positioning System (GPS) receivers (Garmin 45XL) for future

relocation. Having identified the centre of a sample hectare, the

area within a radius of 56.4 m of that point was systematically

searchedby a team of three using a measuring tape and ropes to

define incremental sectors. All the live tortoises and carapace

remains (figure 3c) found were examined and marked tempora-

rily to avoid multiple recording.The curved width of the third

dorsal scute was measured to the nearestmillimetre, and the sex

of all individuals with a third scute width > 20 cm was assessed

by inspectionof tail length, plastrondepth and hind claw length.

3. ESTIMATION F POPULATION IZEStatistically robust estimates of the number of tortoises

on Aldabra and objective assessments of change must take

account of variations inherent within the sample data,

the uneven distribution of tortoises around the atoll,

seasonal movements between habitats, and general prefer-

ences for some vegetation types over others.

(a) Data transformation

The wide variation in tortoise numbers between sample

hectares requires allowance to be made for the frequency

distribution of census data. Bourn & Coe's (1978) original

analyses were carried out on untransformed data,

assuming a normal distribution, whereas a Poisson, ormore skewed distribution, was likely. Examination of both

the 1973-1974 and 1997 census data showed that

frequency distributions were skewed away from normal

and approximated Poisson distributions. Square-root

transformations of the data produced frequency distribu-

tions that approximated normal, so all subsequent

analyses were carried out on square-root transformed

data, but are presented as back-transformed figures.

Hectares in both census periods are paired observations,

so paired t-tests on the transformed data were used to test

for the significance of change.

(b) SubpopulationsThe giant tortoise population of Aldabra is divided into

various subpopulations by geographical and habitat

barriers (see figure 2). Within subpopulations, circulation

and movement is likely, but between them it is unlikely.

Proc. R. Soc. Lond. B (1999)



. ourn an o ers eassess ng a ra sg an or o se opu a on

(a) (b)

_ ~ ~~~- _h_

|_j ~~~~~~~~~~~~~~~(d)

ENK~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~V

(e) ()

A..

Figure 3. (a) Tortoise collection pens in the vicinity of Cinq Cases creels,probably dating from the early l9th century (Stoddart& Peake 1979). (b) The 1997 tortoise census team at the rock cairn with a mangrove pole and ID tag, marking the site of one of

the 1973-1974 census hectares near Cinq Case in south-east Grande Terre. (c) Carapace remains of tortoise 05654 found near

BasinFregate in 1997, marked 23 years earlier near Anse Takamaka, a straight-line distance of 6 km away. (d) Collection of

carapace remains at various stages of disintegration found in the immediate vicinity of Cinq Cases shade trees, south-cast Grande

Terre. (e) The sedge, Cyperusigularis,as spread extensively inland from Cinq Cases since the mid-1I970s.Note the numerousdeadtrees n thebackground,mostlyGuettardapeciosa.n Congregationf tortoisesn pool,south-eastGrandeTerre.

The strength of the barriers ranges from near completeisolation of subpopulationson different islands, to gradualbut slow movement between units separated only byhabitat barriers (Gibson & Hamilton (1983), based ontheir own data and that derived from Bourn & Coe(1978) and Swingland& Coe (1979)).

Bourn & Coe (1978) identified two subpopulationsinthe eastern region and along the south coast of GrandeTerre, and two other subpopulationson Ile Malabar andIle Picard.Allowances were also made for tortoises occu-pying dense Pemphis cidula crub and other suitable areas,including minor islands within the lagoon. Analysis ofrecaptured marked tortoises by Gibson & Hamilton

(1984), revealed a further subpopulationdivide within theeastern region of GrandeTerre.They identified a partialhabitat barrier formed by pool, mangrove and shade-freeridge vegetation separating a 'Takamaka'subpopulationfrom a'Cinq Cases' subpopulation.

(c) Seasonal movementSeasonal movements of tortoises between habitats in

the Cinq Cases region (figures3a-f) have been reportedsince the earliest observations on Aldabra (Fryer 1911).Vegetation and primary production studies by Gibson &Phillipson (1983a,b) provideda sound basis for reanalysisof the original tortoise census data (Bourn & Coe 1978)




Reassessingldabraigiantortoisepopulation D. Bournando3thers 1095

CB | ,, | X w~~~~coc I- u- u- | | o b

b20~~~~~~~~~0{O CO CO 0)03r$

nCOco zco 0 I d

a iS~~~~~~~~~~~~Coco n O = c cor ?

G~~~~~~~~~~~)~It co c co It < U1)

S~~~~~~~~~~0) cn "I r- co C) L =

o C S ~~~~~~~o) (C + I I I I

CS

| ts | 2 |

~~n - On?| |r

x | | o ]~~~~~~~~~~0n 1 co co CcC

u I c I _ ;HZ~~o4 c

> O cocorcor=2~~c

< C7- C"f O tD Ng? ,)o | ! |L Nc

+~~ ~ ~~~~~~~ ~ ~~~~- co) C) 00 |n O)

v = n u^)~~~~~~~~~~~On 01- 0 Q -0LnO CR N > N t

1::~~~~~~~~~~~~~Lc - - n

r H UE 5 , < ,: < < tc; = ; UEC

Proc. R:. Sc. Lond.B (\1999) ~ ~ ~ -4 r



1096 D. Bourn and others ReassessingAldabra$sgianttortoisepopulation

and quantitative assessment of seasonal movements

(Gibson & Hamilton 1983).

Results of seasonal movement studies showed that

during the dry season giant tortoises are relatively evenly

dispersed between the different habitat types, although

P acidula and mangrove vegetation is little used, and

there is a twofold difference in density between the closed

mixed scrub types and more open areas. At the first

significant rain, there is an immediate concentration of

tortoises on the open Sporobolusvirginicus swards of the

coastal fringe, followed by concentrations in open mixed

scrub and associated coastal mixed scrubs, as the 'tortoise

turf' begins to grow. In consequence, there can be as

much as a five- or even tenfold difference in density

between different habitat types during the wet season

(Gibson & Hamilton 1983).

To minimize the effects of this density variability

between habitats, the 1997 tortoise census was carried out

during the dry season. The 1973-1974 census, however,

was conducted as part of a time-consuming, baseline,

marking and measuring programme, which extended

over both dry and wet seasons.

Bourn & Coe's (1978) original counts were adjusted for

the seasonal changes in density in the five different

habitat types (S. virginicusturf, coastal mixed scrub, open

mixed scrub, mixed scrub and pool vegetation) as

estimated by Gibson & Hamilton (1983). These adjust-

ments are based on one or at most two years' estimates of

density for each month and are inevitably approximate,

whereas the true figure is likely to be affected by a

particular year's rainfall and hence productivity patterns.

Habitats in Cinq Cases and Takamaka are spatially

arranged in a mosaic of large blocks, with seasonal move-

ments of tortoises between them. Elsewhere, habitats tend

to be in narrow bands and small patches, accessible by

short-term daily movements (Gibson & Hamilton 1983).

Seasonal movements of subpopulations in these linear

habitats are, therefore, less likely to be a source of bias in

population density estimates.

(d) Habitat stratification

Having corrected for seasonal movements between

habitats, an overall residual difference in tortoise density

between habitat types remained. This difference,

however, was not useful in further stratification, as

increased precision was outweighed by decreasedaccuracy of smaller sample sizes. Confidence limits, as a

proportion of the mean for individual habitats, were

similar to, or higher than, the 11-13% obtained from the

two censuses, using overall data without habitat stratifica-

tion. No further correction was therefore justified.

(e) Overall population estimates

Revised tortoise population estimates for 1973-1974

and new figures for 1997 are presented in table 1 and

figure 4. Estimates are based on square-root transformed

data from the 162ha examined on both occasions, and

have been adjusted for seasonal variation in tortoise

distribution. Mean density and population values aregiven, together with 95% confidence limits. The

probabilities of chance occurrence are also indicated.

Data for Picard in both censuses were based on an

unreplicated sample, so only a single density figure is

40

coCinqCaseso ~~~~~~~~~~~~~~~~~-p:! 30

south coast

a) 20 - Malabar nkchange akaX ~ Pi d =-+p-= 0.0001 -p= 0.022

o

subpopulation and year

Figure 4. Density changes in tortoise subpopulations.

given for each occasion. Data for 'south coast other',

'Malabar Pemphis', Picard Pempliis' and 'other areas' are

based on non-quantitative estimates, and are included so

that a nominal approximation to the whole Aldabra

population can be given. Bourn & Coe's (1978) original

figures for these areas in 1973-1974 are given. The 1997

figures have been discounted at the same rate of overall

population change estimated for census areas with quanti-

tative and replicated data.

The total population of giant tortoises on Aldabra atoll

in 1997 was estimated to be 100 473, down from and a

revised estimated total of 129 415 in 1973-1974. This

revised figure is considered to be more realistic than

Bourn & Coe's (1978) original estimate of 150 466, which

was based on analysis of untransformed data.

Tortoise numbers on Grande Terre have declined signif-

icantly since 1973-1974, with reductions of 33% in theCinq Cases area (p= <0.001) and 41% in the Takamaka

area (p =0.02).

The south coast population of Grande Terre has also

declined, but the reduction of 8% is not significant.

Tortoise numbers on Malabar and Picard have

increased significantly since 1973-1974, with a doubling of

the population on Malabar (p-0.0001).

The magnitude and direction of population changes

appear to be density dependent, as indicated in figure 4.

The lowest density populations on Malabar and Picard

have increased; the intermediate density population on

the south coast has remained approximately steady, and

the high-density linked subpopulations of Cinq Cases andTakamaka have declined.

4. CORROBORATIVEVIDENCE

Corroborative evidence for the tortoise population

crash in south-eastern Grande Terre comes from two

independent observations: a signlificant increase in

tortoise mortality in recent years; and a significant

decline in tortoise counts on population monitoring

transects.

A OA

(a) Tortoise remains

In the absence of large scavengers on Aldabra, tortoisecarapaces may persist for two to three years, depending

on size and exposure (Bourn & Coe 1979). All carapaces

found in sample hectares during both censuses were

examined and recorded. Back-transformed mean densities




ReassessingAldabra'sgiant ortoise opulation D. Bourn and others 1097

Table 2. Density of tortoiseremains

location density range1973-1974 pairedt-test density range1997

probabilityof chance95% low mean 95% high occurrence 95% low mean 95% high % increase

Grande Terresouth-east 0.47 0.70 0.96 0.0001 1.10 1.47 1.92 110south coast 0.19 0.38 0.61 0.0015 0.59 0.87 1.20 129

Ile Malabar 0.03 0.17 0.33 0.356 0.12 0.27 0.45 59

2.5-South Eastp= 0.0001

,2-

south coastp 0.0015

Q 1.5-

ZC

E 1-Malabarp= 0.356

*O0-5

0

1973-74 1997 1973-74 1997 1973-74 1997

subpopulation and year

Figure 5. Density changes in tortoise remains.

are compared in table 2 and figure 5. A significant

increase in mortality is apparent.

Tortoise remains on Grande Terre were more than

twice as common in 1997 than in 1973-1974, with highest

concentrations in the south-east. From the observed

density of tortoise remains, it would appear that around

7950 tortoises died between 1995 and 1997. A major

contributory factor to this mass mortality is likely to have

been two consecutive years of well below average rainfall

in 1995-1996 and 1996-1997, as recorded at the meteo-

logical station on Picard (SIF, unpublished data).

(b) Transect counts

Three tortoise population monitoring transects in

south-east Grande Terre-Coco, Groves and Southern-have been surveyed intermittently over the past 20 years

and provide an index of population abundance. Quarterly

mean tortoise transect counts are summarized in figure 6,

derived from Aldabra Research Station files with addi-

tional information provided by C. W. D. Gibson and

J. Hamilton (personal communication). Given the

seasonal redistribution of tortoises and the variety of

transect recorders, the high variability of tortoise counts

is hardly surprising. Nevertheless, visual inspection indi-

cates that counts during the 1990s have been appreciably

lower than in the late 1970s and 1980s, especially on Coco

and Southern transects.

This impression is strengthened when all three transectcounts are summed (figure 7), which reduces some of the

overall variation, as an increase on one transect may be

offset by a decrease on another, to reveal a highly signifi-

cant decline in tortoise numbers since 1981 (p <0.001).

The decline in numbers is most clearly demonstrated by

the end of wet season counts, when tortoises are most

widely distributed (figure 8). Over the 16 years from

February 1981 to February-March 1997, end of wet

season transect counts indicate that there has been a 25%

decline in relative abundance of tortoises in south-eastern

Grande Terre, from 320 to 240. This is consistent with the

findings of the 1997 hectare sample census, which indicate

a 33-41% decline since 1974.

5. DISCUSSIONAND CONCLUSIONS

(a) Changing population levels

This study has substantiated earlier indications of the

non-equilibrium status of the Aldabra giant tortoise popu-

lations (Bourn & Coe 1978,1979; Swingland & Coe 1979;

Gibson & Hamilton 1984). The directions and rates of

change are clearly associated with population density

(figure 4). The low-density population on Malabar

increased rapidly from 1974 to 1997, while the highestdensity subpopulations, all on Grande Terre, decreased.

Although not formally testable, the lowest density popula-

tion, on Ile Picard, also increased rapidly from an esti-

mated five to eight animals per hectare. The patterns of

change appear to be related to density in a relatively

simple manner, despite the various settings of the different

populations. For instance, populations on Picard, Malabar

and Grande Terre are relatively isolated, while the different

subpopulations on Grande Terre have higher rates of

emigration and immigration (Gibson & Hamilton 1983).

At 100 000 individuals, the Aldabra giant tortoise

population is still far more numerous than its endangered

relative on the Galapagos, where fewer than 15 000remain (Stolzenburg 1996). Densities attained by the

Grande Terre tortoise subpopulations are higher than any

other surviving giant tortoise population. However, they

do reflect anecdotal reports of typical densities encoun-

tered before human exploitation or other disturbance of

natural populations. Descriptions of natural densities on

the Galapagos, Seychelles, Mascarenes and Aldabra itself

(Stoddart & Peake 1979; Desmond & Moore 1992)

resemble south-east Grande Terre in the late 20th century,

not the current low densities of surviving populations.

Nevertheless, the Aldabra populations have all been

disturbed (as outlined in ?1 of this paper). We cannot

place a precise date on the end of significant disturbance,but the Grande Terre populations appear to have started

their recovery shortly before the beginning of the 20th

century. Significant interference with the more accessible

Malabar population continued much later, probably




1098 D. Bourn and others Reassessing ldabra'sgiantortoiseopulation

350

300-

o 250-0

(' 200-

(D10o

0

E 5

0

4~~~~~~~~~?tbtt Wi11i

0~~~~~~~~~~~~~~~~~~~~

500

00

3N0 ';'

FigureC7.otal tortoise t cuts in south trande Te

cot uii TN;TN4i ttEuEtNttV0 i;0A0itX0Sf00000it;V0ityitt t!tti te ;t0XV0utRtCit0t tE00t 0

ouu- SkS:00:tl;S tH

;000~~~400. ti1 tlS;0 y ;: ; 4 m~~~~~~~~~~~~~~Int;S;;;;SS,gtSEC t00 t00 t;tC tt;t i tEE t0~~ 0f<::;tiS t t.01 i

0 ~~ ~~~~~~~~~X00A))Q00E0XSX)0j0l0l\0.\A)Aj0)TXi j 0 E0l:0t2000X0a0i4) :A00X)0X; ;:

co 400-

:~~~~~~~~~~~~~~~~000, 9~ - 6 5w p C 5 5

1<* ' *<v <eQ ZQ1z 4v ~1

Fgr7. Toaoros00 asctcut in sot-eatGad er - -.23+9.,R .83

until the late 1950s, while on Picard human removal of

tortoises is likely to have continued until the research

station was established in 1968. The densities found in

the initial census in 1973-1974 reflect these differences in

the period of time that each population has had to

recover.

The significance of these changes depends on their

cause (s), which warrant further investigation and con-

tinued tortoise population monitoring. Whatever the

cause, they are important scientifically because they are

the only example where large reptile populations can bestudied at, or near, their natural densities. Of particular

interest is the significant decrease, or 'crash', in high-

density populations, which clearly demonstrate that a

stable, steady state was not achieved.

(b) Correlates of change and implication for

conservation

Leaving aside changes in initially low-density popula-

tions on Malabar and Picard, that are clearly associated

with population recovery, the decline in high-density

populations on Grande Terre is of obvious potential

concern for species conservation. The most parsimonious

explanation for the decline is that it is a consequence of

the inherent dynamics of giant tortoise populations. For

example, Grande Terre populations might be in a state of

equilibrium that is not constant, but is cyclic or chaotic;or they might be fluctuating as part of a dynamic

recovery process from repeated collections during the

18th and 19th centuries that has yet to run its full course.

If either of these explanations were true, there would be

Proc. 1.So. Tond. B (1999)



Reassessing ldabra'sgiantortoiseopulation D. Bourn and others 1099

500 -

o = 400-

300,

m 200-

+5~ 100

0

Figure 8. Total late wet season tortoise transect counts in

south-east Grand Terre.y = -0.4423x + 753.83, R2= 0.6185.

important consequences for these animals and others with

similar life-history traits. At the least extreme, time,

patience and perhaps direct intervention would be needed

to ensure their long-term survival.

Alternatively, it is known that the Grande Terre popula-

tions are very close to their carrying capacity for food

supply and, in turn, that rainfall drives both the produc-

tivity of key vegetation and reproduction and growth in

these tortoises (Gibson & Hamilton 1983, 1984). In conse-

quence, population levels, as in African elephants (Laws

et al. 1975; Phillipson 1975; Caughley 1976; Eltringham

1979) could be highly sensitive to fluctuations in annual

rainfall. Following relatively high rainfall years in the

early to mid-1970s, there were several years of drought

between then and the early 1990s (SIF, unpublished

data). If these low rainfall years alone were responsible

for the decline, there is little import for the animals'

conservation, unless there has been a permanent change

in the rainfall regime.

Feral animals have had less impact on Aldabra than

the Galapagos, where they have contributed to the

general decline of tortoises and their extinction from

some islands, through competition or predation on nests

and young animals (MacFarland et al. 1974a,b). Feral

goats have, however, been present on Aldabra at least

since the 1870s. Until the 1970s, these appeared to have

remained at relatively low levels, maintained perhaps by

human predation, but after establishment of the research

station, they increased sufficiently to have marked local

effects on vegetation cover (Hambler 1984; Newing et al.1984; Scoones et al. 1988; Burke 1988a,b). Subsequent

control measures have once again reduced goats to low

levels (Coblentz et al. 1990; Rainbolt 1995). During the

period 1987-1997 a total of some 1800 goats were killed,

with an estimated 30-50 remaining in June 1997

(M. Bergeson, personal communication).

Goats, which are mainly browsers, have two potential

effects on tortoises through food competition and through

removal of shade, which tortoises need to avoid over-

heating (Hnatiuk et al. 1976; Merton et al. 1976). Overall,

tortoises browse little and are mainly grazers, but rely on

browsing for short but critical periods during the dry

season (Gibson & Hamilton 1983). Likewise, mosthabitats used by tortoises on Aldabra have plenty of

shade, but shade is critical in others, such as the coastal

S. virginicus swards, that many tortoises rely on for short

periods of the year. Goats, therefore, may have contrib-

uted marginally to the observed decline through impactson bottlenecks in the tortoises' annual food and shaderequirements.If so, this provides further justification forthe continued control, or elimination, of feral goats overand above that of their damage to native vegetation.

The evidence presented in this paper has identifiedmajor changes in the Aldabra tortoise populations and anumber of possible causes that correlate especially withthe observed declines in the high-density populations.These changes demonstrate the value of long-term moni-toring of this last remaining set of long-lived, terrestrialreptiles living under near-natural conditions.The resultsreveal a dynamic situation, not a static one, and areinstructive for the better understanding of the ecologyand conservation of tortoisesand a wide range of animalsthat share some of their life-historytraits.

The management implications of our findings dependon what conservationobjectivesand prioritiesare consid-ered appropriate,and the financial and human resourcesavailable.Aldabra'sremote, rugged and relatively inhospi-table environmentimposes severe constraints on realisticand sustainableobjectives.The atoll is managed as aWorldHeritage Site by the SIF, under non-specific AldabraSpecial Reserve Regulations, 1981 (Seaton et al. 1991).Whether or not conservation management objectivesshould maintain the present status quo of minimal inter-ference,or take a more proactiveinterventionistapproachto the management of tortoisesubpopulations,remains anopen question.

Fieldworkor thisstudywas carriedout on behalfof the SIFandfunded by the Global EnvironmentFacility,as part of the

Seychelles BiodiversityConservationand Marine PollutionAbatement roject.Wewould iketo thankall SIFandMinistryof Tourism taff on Aldabraand on Mahe, especiallyFrancisAlcindor,GeorgeAuguist,Roselle Chapman,LindsayChong-Seng, Dave Morel,Jean-ClaudNicette,MarkNicette,WalterRoselie (deceased) nd GordonSmith, or their hardworkandenthusiasm.Many hanksalso to MalcolmCoe,Julie Hamilton,Len Mole, Jeanne Mortimer,MarianneNord, David Rogers,RichardSouthwood,David Stoddart, onToomerand WilliamWint,for theirsupportandencouragementehind the scenes.SIHis supported y a grantfromthe Departmentor Interna-tionalDevelopment DFID),LivestockProduction rogramme(ZC0012).However,he DFIDtakesno responsibilityor infor-mationprovided r viewsexpressed.

REFERENCES

Abbott, W. L. 1893 Notes on the natural history of Aldabra,Assumption and Glorioso Islands, Indian Ocean. Proc.US NatlMus. 16, 759-764.

Amin, M., Willets, D. & Skerrett, A. (eds) 1995 Aldabra:worldheritage ite. Nairobi: Camerapix Publishers International forSeychelles Islands Foundation.

Beamish, T. 1970Aldabra lone.London: George Allen & Unwin.Bourn, D. 1977Reproductive study of giant tortoiseson Aldabra.

_>Zool.Lond.182, 27-38.Bourn, D. & Coe, M. 1978 The size, structure and distribution

of the giant tortoise population on Aldabra. Phil.Trans.R. Soc.Lond.B282, 139-175.

Bourn, D. & Coe, M. J. 1979 Features of tortoise mortality and

decomposition on Aldabra. Phil. Trans.R. Soc. Lond. B 286,188-193.

Burke, M. G. 1988a Status, impact and conservation implica-tions of feral goats on Aldabra atoll. Bull. Biol. Soc.Wash.8,129-138.




1100 D. Bourn and others Reassessing ldabra'sgiantortoiseopulation

Burke, M. G. 1988b The feral goats of Aldabra: ecology and

population dynamics.NatlGeogr. es.4, 272-279.Caughley 1976 The elephant problem an alternative hypoth-

esis. E. Afr.Wildl. _ 14, 265-283.

Coblentz, B. E., Van Vuren, D. & Main, M. B. 1990 Control of

feral goats on Aldabra atoll. AtollRes.Bull 337, 1-10.

Coe, M. J., Bourn, D. & Swingland, I. R. 1979 The biomass,production and carrying capacity of giant tortoises on

Aldabra.Phil.Trans.R. Soc.Lond.B 286, 163-176.Desmond, A. & Moore, J. 1992 Darwin. London: Penguin.

Dupont, P. R. 1929 Colonyf Seychelles;isitto theoutlyingslandsbytheGovernor.ictoria, Mahe: Government PrintingOffice.

Eltringham, S. K. 1979 Theecologynd conservationf largeAfricanmammals. ondon: Macmillan.

ERGO 1997 Aldabra revisited: final report on the 1997Aldabran giant tortoise population and vegetation study.

Report by the Seychelles Islands Foundation,Victoria, Mahe.

Environmental Research Group Oxford Limited. (Held at

the Zoology Library of the University of Oxford.)

Fryer,J. C. F. 1911The structureand formation of Aldabra and

neighbouring islands-with notes on their flora and fauna.Trans. inn.Soc.Lond.Zool.2 14, 397-442.

Gaymer, R. 1968 The Indian Ocean giant tortoise Testudo

gigantea n Aldabra. 7.Zool.Lond.154, 341-363.

Gaymer, R. 1973A marking method for giant tortoises and field

trials on Aldabra. _i.Zool.Lond.169, 393-401.GEF 1992 Seychelles biodiversity conservation and marine

pollution abatement project document. Washington: GlobalEnvironment Facility,The World Bank.

Gibson, C. W. D. & Hamilton, J. 1983 Feeding ecology and

seasonal movements of giant tortoises on Aldabra atoll.

Oecologia6, 84-92.Gibson, C. W. D. & Hamilton, J.1984 Population processesin a

large herbivorousreptile: the giant tortoise of Aldabra atoll.

Oecologia1, 230-240.

Gibson, C. W. D. & Phillipson, J. 1983a The vegetation ofAldabra atoll: preliminary analysis and explanation of the

vegetation map.Phil. Trans. . Soc.Lond.B 302, 201-235.Gibson, C.W. D. & Phillipson,J. 1983bThe primary production

of Aldabra atoll, with reference to habitats used by giant

tortoises.Phil. Trans. . Soc.Lond.B 302, 167-199.

Grubb,P. 1971The growth, ecology and population structuresof

giant tortoises on Aldabra. Phil. Trans.R. Soc.Lond.B 260,

327-372.Gunther,A. C. L. G. 1877Thegiganticand ortoisesliving nddead)n

the ollectionftheBritishMuseum. ondon:Taylor& Francis.

Hambler, C. 1984 Goat threat. NewSci.104, 46.

Hamilton,J. & Coe, M. J. 1982 Feeding, digestion and assimila-

tion of a population of giant tortoiseson Aldabra atoll. I.Arid

Env.5, 127-144.Hnatiuk, R. J., Woodell, S. R. J. & Bourn, D. M. 1976 Giant

tortoise and vegetation interactions on Aldabra atoll. II.

Coastal. Biol.Conserv., 305-316.

Laws, R. M., Parker, I. S. C. & Johnstone, R. C. B. 1975

Elephantsnd heirhabitats: heecology felephantsnNorthBunyoro,Uganda.Oxford: Clarendon Press.

MacFarland, C. G., Villa, J. & Toro, B. 1974aThe Galapagosgiant tortoises (Geochelone lephantopus). . Status of the

surviving populations.Biol.Conserv., 118-133.

MacFarland, C. G., Villa, J. & Toro, B. 1974bThe Galapagosgiant tortoises (Geochelone elephantopus). II. Conservation

methods. Biol. Conserv., 198-212.Merton, L. F. H., Bourn, D. M. & Hnatiuk, R. J. 1976 Giant

tortoise and vegetation interactions on Aldabra atoll. I.Inland. Biol. Conserv., 293-304.

Newing, T. J., Daly, K. H. & Hambler, K. 1984 Feral goats onAldabra. Final report, Southampton University to Aldabra1982. University of Southampton.

Nicoll, M. J. 1908 Three oyagesfa naturalist.London: Witherbyand Co.

Phillipson, J. 1975 Rainfall, primary productivityand 'carryingcapacity' of Tsavo National Park (East), Kenya. E. Afr Wildl.

_7.3, 171-201.Rainbolt, R. E. 1995 Final report on the Aldabra feral goat

control project. SeychellesIslands Foundation.Riseley-Griffith, T. 1892 Report on Aldabra to R. Risely-

Griffith, Esq. Administrator of Seychelles Islands. Port

Victoria, Seychelles, July 19, 1891. Colonial report no. 40,Mauritius (Seychellesand Rodriguez).

Roberts, B. 0. 1905 Letter to d'Emmerez, dated 6 September1905. Seychelles Archives, PortVictoria, Mahe.Rothschild, W. 1915 On the gigantic land-tortoises of the

Seychelles and Aldabra-Madagascar group, with some noteson certain forms of the Mascarene group. Novit. Zool. 22,418-442.

Scoones, T., Hambler, C., Woodroffe, R., Linfield, M. &

Spillett, D. 1989 Aldabra '88. Final report of the OxfordUniversity expedition to Aldabra 25 June-2 October 1988.

Zoology Department, Universityof Oxford.Seaton, A. J., Beaver, K. & Afif, M. (eds) 1991 A focus on

Aldabra. Conserving the Environment, vol. 3. Victoria,Seychelles: Ministry of Education,Governmentof Seychelles.

Stoddart, D. R. 1968a The Aldabra affair. Biol. Conserv. ,63-69.

Stoddart, D. R. 1968bThe conservationof Aldabra. Geog. . 134,471-485; discussion 485-486.

Stoddart, D. R. 1971Settlement, development and conservationof Aldabra. Phil.Trans. . Soc.Lond.B 260, 611-628.

Stoddart, D. R. & Peake,J. F. 1979Historical records of Indian

Ocean giant tortoise populations. Phil. Trans.R. Soc. Lond.B 286, 147-161.

Stolzenburg,W. 1996 Collision at the Galapagos: slow road to

evolution meets the fast track of extinction. Nat. Conserv.Mag.46. (http: www.tnc.org/news/magazine/nor_dec/galap.htm.)

Swingland, I. R. & Coe, M. J. 1978 The natural regulation of

giant tortoise populations on Aldabra atoll: reproduction.9.Zool.Lond. 186, 285-309.

Swingland, I. R. & Coe, M. J. 1979 The natural regulation of

giant tortoise populations on Aldabra atoll: recruitment.Phil.Tra,ns.. Soc.Lond.B 286, 177-188.

Voeltzkow, A. 1895 Einleitung: Madagaskar, Juan de Nova,Aldabra. Abh. enckenb.aturf.Ges.21, 1-76.

Wharton,W.J. H. 1879Letter on Aldabra. Ann.Mag.Nat. Hist.

3, 165-166.

As this paper exceeds the maximum length normally permitted,

the authors have agreed to contribute to production costs.

Proc. R. Soc. Lond. B (1999)

Rise and Fall of Aldabran Giant Tortoise Bourn Et Al

Documents

Transcript of Rise and Fall of Aldabran Giant Tortoise Bourn Et Al