INVASIVE RESEARCH, REDUCED
LONGEVITY AND MORTALITY IN AN ENDANGERED SPECIES,
|
"…. the resolution of one
question remains imperfectly ragged, despite exhaustive attention, and that
is whether handling or vaccinating wild dogs had inadvertently contributed
to their demise in the Serengeti-Mara ecosystem"
David Macdonald, Chair of the Canid Specialist Group, in Preface to: - IUCN,
Species Survival Commission, Canid Specialist Group (CSG), Status Survey
& Conservation Action Plan, 'The African Wild Dog' (Woodroffe
et.al.1997).
INTRODUCTION
The African Wild Dog (Lycaon pictus) an African endemic species is highly
endangered with now perhaps less than 5,000 individuals. Between 1985-1991
the entire wild dog study population comprising 14 packs containing approximately
200 individuals died or disappeared from two study areas in the Serengeti-Mara
ecosystem of Tanzania (Serengeti) and Kenya (Mara), East Africa where the
species had been considered a 'flagship species' for conservation.
The ecosystem population did not become extinct in 1991 a non study population
persisted within and around the ecosystem throughout the study period and
persists to date.
All study packs were the subject of conservation research by scientists who
routinely used invasive research techniques (known as 'handling'); including
immobilisation by anaesthetisation for radio-collaring and blood sampling
of all new packs or groups located. Also in the Mara study population from
1987 onwards, sporadic experimental rabies vaccinations were carried out
despite the lack of any serological evidence at that time that the packs
in the Mara had been exposed to rabies virus. In 1990 a mass vaccination
was carried out in the Serengeti study population which had been naturally
exposed to rabies virus pre vaccination. Of the 13 Serengeti-Mara study packs
known to have become extinct 77% survived less than 4 months after
anaesthetisation and radio collaring of one or more individuals in the
pack.
Rabies was confirmed in all cases where tissue samples suitable for analysis
were obtained in both study areas and is the most probable cause of extinction
of both unvaccinated and vaccinated study packs post 1985.
The claimed major aim of such invasive research is conservation by facilitating
the rapid detection of changes in wild dog demography, particularly the extent
and causes of mortality and providing early warning of major adverse changes
in populations over time. This it is claimed will enable wildlife managers
to counter any identified short or long term problems particularly in endangered
species. These claims remain unsubstantiated, as do claims that the benefits
of invasive research outweigh the risks; the converse seems to apply.
The two Serengeti-Mara study populations one in 'Serengeti' (Tanzania) and
the other approximately 175 km to the north in the 'Mara' (Kenya), together
with a non study population formed part of a breeding population of at least
200 individuals with interchange between widely spaced packs via long range
dispersal (often >100 km) of yearlings. Population Viability Analysis
models suggest that large (>100) wild dog populations occupying large
(>=5,000 km2) areas are unlikely to become extinct due to a chance event,
such as a disease outbreak, and suggest that some other factor(s) were involved
in the extirpation of the Serengeti (and Mara) study packs between 1985-91.
The anaesthetisation of wild dogs in captivity is known to be followed by
a significant rapid increase in the levels of stress related corticosteroid
hormones which are associated with immunosuppression. The efficacy of the
inactivated rabies vaccines used, in producing an effective immune response
in Lycaon, is unknown.
Analysis of the Serengeti data from 1985-August 1990 showed a statistically
significant reduction in longevity of both handled individual wild dogs and
their packs and suggested a causal relationship existed between handling
and the rabies related extinction of the study packs within a few months
of a handling event.
A 'handling-stress' hypothesis (also referred to as the
'handling-stress-immunosuppression' hypothesis) was proposed to explain
this.
Recently, significant population declines have taken place in other intensively
handled wild dog study populations in Kruger National Park (South Africa)
and Moremi Game Reserve (Botswana). Until 2000 the Kruger population, the
only viable free living population in South Africa, was considered to be
stable, with no significant disease related adult mortality since pre 1990.
However, in 2001 it became apparent that between 1995-2000, there had been
a significant, but apparently unnoticed, decline of approximately 60% in
number of individuals and a 30% reduction in the number of packs in this
very intensively handled and managed population. There was no evidence of
disease. Kruger scientists attribute this population crash to changes in
rainfall in 1996-97 which they associate with reduced reproductive success
of the packs.
However, just prior to the crash 2 adults in each of the packs (36 in 1995)
were anaesthetised and blood sampled with the alpha pair in each pack, if
known, being selected. Pack disruption is known to follow the loss of either
of the alpha pair in a pack. The alpha pair are the main and usually the
only breeding pair in a wild dog pack in any one year. Also in Kruger, between
1990-93, 34 % of all radio collared individuals survived less than 12
months.
In Moremi Game Reserve (Botswana), disease related adult mortality in wild
dogs was unknown until 1996 when an unidentified disease, thought to be rabies
or canine distemper, caused the extinction of 5 (38%) of the study packs
during a three week period. In the on-going Moremi study, as in Kruger, one
of the alpha pair in each study pack is usually selected for handling, this
time for radio-collaring, with sometimes both alpha individuals being collared.
As radios expire within 19 months the Moremi dogs are re-anaesthetised to
replace them. Unfortunately, despite the presence of radio-collared individuals
in each study pack no tissue samples were obtained from the 5 Moremi packs
lost in 1996. In Serengeti of the 5 handled packs which died in 1991, 4
containing radio-collared individuals but, as in Moremi, no tissue samples
were taken.
No data are available on the longevity of the radio-collared alpha individuals
in Moremi although it is known that 25% of all radio collared dogs in this
study between 1989-93 died within 12 months of radio-collaring, but whether
this was after the first collaring or re-collaring is not reported.
In South Africa and Namibia death from rabies of individuals vaccinated against
rabies in captivity prior to release in reintroduced packs and groups emerged
as a 'new' post 1988 phenomenon.
"Despite the potential impact that such activities [handling] may have,
it is frequently tacitly assumed ,or subjectively assessed, that these techniques
have little effect on individuals in the long term" (White & Garrot,
1990).
"It is likely…. that in many studies any adverse effects [of handling]
are either unnoticed, perhaps because they are rare, or, more likely, because
they are not reported" (Laurenson 1992)
Wild dog pack extinction (other than by shooting) was not known in Serengeti
between 1965-85 i.e. before the introduction of intensive invasive research
techniques
In contrast to the confirmed population declines in Kruger and Moremi, despite
many published opinions expressed to the contrary, there is no evidence for
a significant decline in the 'Serengeti' wild dog population (either in number
of adults or number of resident packs) between 1965-78, or that the Serengeti
plains study population was subject to a series of disease related population
declines in any one year between 1965-85 i.e. before routine invasive research
began (see
Appendices
1 &
3).
Likewise there is no evidence that the significant increase, between the
mid 1960s and 1986, in the number of spotted hyaena commuting to the Serengeti
plains in the wet seasons when the dogs denned, had any significant effect
on resident wild dog pack's nutrition, reproductive success or mortality.
The hyaena increase had no observable effect, either pre 1985 when wild dog
reproductive success was very low and immigrants very rare, or post 1985
when reproductive success was high, the population expanding and immigrants
common. Due to frequent immigration of wild dogs to the Serengeti plains
post 1979, the small study population was not, as is often claimed, genetically
isolated.
Unfortunately any anecdotal hint of a decline in the Serengeti wild dog
population before routine handling was introduced is considered by some
commentators to be firm evidence of a series of catastrophic disease related
population collapses in Serengeti
(see
Appendices 3a.-d.). By contrast, in Kruger where a 60% loss is
known to have occurred following an invasive research exercise, this is claimed
to be a natural population fluctuation to which wild dog packs are prone;
and little comment has been made on the loss of the 5 study packs in
Moremi.
Extinction of Wild dog packs from disease : A new phenomenon in 'Serengeti'
post 1985
Prior to 1986 no pack extinction in the Serengeti was known other than those
shot by rangers/game wardens as vermin, a practice which fortunately ceased
in 1973. In 1986 the first disease related pack extinction (with pack behaviour
suggesting rabies) in the ecosystem occurred within three months of the
anaesthetisation and radio-collaring, of one subordinate adult dog, in one
of the two Serengeti study packs.
The rate of disease related study pack extinction in Serengeti increased
paralleling the increasing use of routine intensive invasive research techniques
post 1985. Despite the extinction of Serengeti study packs in 1986, 1988,
1989 and 1990 (2) the Serengeti and Mara population was expanding with new
packs replacing those lost (Appendix 8 and Table 3). This resulted from good
reproductive success of the remaining study packs and the arrival of immigrant
groups of both sexes.
Additional stressors to the study packs resulted from monthly or more frequent
aerial monitoring between 1986-90 and the presence of tourist vehicles which
followed hunting packs and the exposure to a variety of pathogens including
a new virus, canid parvovirus. In September/October 1990 more than 80% of
the study population in Serengeti was handled during mass vaccination against
rabies, all had died with empty radio-collars being found from some individuals
in the packs the rest simply disappeared by June 1991 (see
Appendix
8 &
Tables
4 &
5).
An unhandled non study resident denning population persisted within and around
the ecosystem throughout the period 1985-91 and persists to date. In both
sectors of the ecosystem unhandled packs were still present after
all study packs had died (Burrows et.al. 1995, Kat 1992).
The pattern of mortality in Serengeti changed significantly post
vaccination. Between June 1986 - August 1990 (pre vaccination) 5 widely spaced
study packs died, with rabies confirmed in one, all within 4 months of the
radio-collaring of one or more individuals in the pack; other study and non
study packs, some with home ranges overlapping that of the diseased pack,
survived each study pack death. There was no evidence of inter-pack transmission
of pathogens or of domestic dog vectors.
Post vaccination all 5 remaining study packs disappeared, 2 packs
died within 3 months of the next anaesthetisation event in the pack
which involved the radio collaring of a previously vaccinated
individual.
The effect of repeat handling of individual wild dogs is currently unknown
(see
Table 4).
After the Serengeti vaccinations, blood samples taken up to 2 years before
vaccination, were finally screened and 46% were then found to be rabies
seropositive. The population had been exposed to rabies in the environment
pre vaccination and some survived. This suggests that some natural immunity
existed pre but not post vaccination.
Prior to 1989, rabies had not been confirmed in any free-living population
of wild dogs
In 1987 the first experimental vaccination of free living wildlife in Africa
against rabies began in the Mara with the sporadic vaccinations of a few
wild dogs in the only large study pack - the Aitong. In August 1989 older
adults in the Aitong pack began to die within 2 months of the anaesthetisation
and blood sampling of 8 individuals and the radio-collaring and rabies
vaccination of 3 pups (10 months old).
The first free living pack in which some individuals had been vaccinated
against rabies became the first free living pack in which rabies was confirmed.
Between December 1990 and January 1991 the other two Mara study packs died
again with rabies confirmed in the only pack from which tissue samples were
taken. One pack had been completely and the second partially vaccinated against
rabies.
(Appendix
8 &
Table
4).
As all serum samples taken from Mara wild dogs prior to August 1989 (unlike
those taken in the Serengeti) were claimed to be sero-negative for rabies.
The rationale for the initial sporadic Mara vaccination of a few individuals
in the Aitong pack in 1987 and subsequently, using inactivated vaccines developed
for domestic dogs, in the absence of any confirmed threat of rabies is therefore
unclear.
Similar strains of a locally common canid rabies virus were isolated from
the Aitong Pack in the Mara and from an unvaccinated Serengeti wild dog study
pack which in August 1990 died from rabies within 2 months of the radio-collaring
of two older adults one a subordinate female with pups.
In southern Africa there is no serological evidence of natural exposure of
free living wild dogs to rabies. However, in Etosha NP, Namibia (1989
&1990) and Madikwe GR, South Africa (1997 & 2000) in packs in which
some individuals had been vaccinated against rabies in captivity both vaccinated
and unvaccinated individuals died from rabies after release. In 1993 rabies
was for the first time confirmed in a free living Lycaon pack in Zimbabwe
in the intensively handled population in Hwange N.P.
Rabies in wild dogs in southern Africa, where disease related mortality in
wild dogs is rare, is a new phenomenon unreported prior to the introduction
of intensive invasive handling including vaccination. There is no evidence
that currently used rabies or canine distemper vaccines protect free living
or captive wild dogs from challenge (East & Burrows 2001 and
Appendix
1 &
8
and Tables
4-5).
The Handling-Stress Hypothesis
The Serengeti data from 1985-August
1990 suggested that a direct causal relationship existed between 'handling'
and reduced individual and pack longevity. A handling-stress hypothesis (also
later referred to as the 'handling-stress-immunosuppression' hypothesis)
was proposed to explain this. (Burrows 1992).
The hypothesis is based on the basic premise that all forms of handling
of wild dog is stressful to the individual(s) and their pack and results
in the elevation of stress related corticosteroid hormone levels.
The data suggest that anaesthetisation and radio-collaring of older (mainly
post 'dispersers' i.e those that have left their natal pack) subordinate
adult individuals has the greatest adverse effect on individual and their
pack's longevity with little evidence of such effects when yearlings are
handled for radio-collaring.
The hypothesis attempts to explain the phenomenon of pack extinction from
disease between 1985-90 and why there was a sudden change in the pattern
of disease related study pack extinction in the Serengeti in 1991:-
Between 1985 and August 1990 sporadic pack deaths occurred with rabies confirmed
or suspected within a few months of a handling event for radio-collaring
and blood sampling mainly of individual older dogs. Most of the individuals
in these packs were 'unhandled'. In 1991 however, within a few months of
the mass vaccination (of more than 80% of the study pack individuals) against
rabies all the study packs died with rabies confirmed in two packs in the
Mara sector of the ecosystem. Non study unvaccinated packs and groups survived
with no evidence of a general rabies epidemic (Burrows et.al. 1995, Kat
1992).
Why did rabies emerge post 1985 as a new problem in 'handled' wild dog
packs?
To explain why free living packs, some known to have been naturally exposed
to rabies virus and with some sero-positive individuals, should die from
rabies post handling, a possible mechanism was suggested:
If an asymptomatic wild dog 'carrier' of rabies in which the virus is 'latent'
is handled, the stress induced rapid rise in corticosteroid hormone levels
may result in immunosuppression allowing activation of the virus leading
to clinical rabies and pack death. The anaesthetisation by dart of wild dogs
in captivity in South Africa is known to be followed by a significant rapid
increase in the stress related corticosteroid hormones associated with
immunosuppression.
The handling-stress hypothesis can explain the sporadic loss of 5 Serengeti
radio-collared packs pre vaccination, the sudden extinction of the remaining
5 Serengeti study packs in 1991 following the mass vaccinations of August
/September 1990, and also the loss of the 3 Mara study packs from rabies
between 1989 and 1990. Pre vaccination i.e. pre September 1990, the chance
of targeting a 'carrier' in a pack when a single dog was selected for
radio-collaring was much less than that when most individuals in the packs
were handled for vaccination in 1990.
An alternative mechanism, also consistent with the basic premise of
the hypothesis, is that handling stress induced immunosuppression, particularly
following the Serengeti mass vaccination against rabies of packs already
naturally exposed to rabies virus, may lead to greater susceptibility
of some stress-immunosuppressed individuals to rabies virus (or any other
pathogen) in the environment when the pathogen is next encountered. This
again leading to clinical rabies and transmission to the rest of the pack.
This alternative mechanism could explain not only the outbreaks of
rabies in Serengeti- Mara but also those in Namibia and South Africa.
It has been suggested that the rabies virus which killed the Serengeti study
packs was contracted, in some unspecified way, from domestic dogs in villages
to the west of Serengeti National Park where rabies may have been endemic
since the last domestic dog epidemic in 1977 and possibly maintained
in this low-density population by seropositive infectious 'carriers' which
show no clinical symptoms of the disease. The existence of asymptomatic rabies
'carrier' domestic dogs has been confirmed in Nigeria and Ethiopia. Recently
Serengeti Spotted hyaena have been found to be 'carriers' of a different
rabies strain from the one isolated from the Serengeti and Mara
Lycaon, but with no associated mortality. Both Bat-eared fox and
White-tailed mongoose in Serengeti are known to suffer periodic limited
mortality from rabies (Burrows 1995).
The possibility that some seropositive wild dog or other species may act
as asymptomatic potentially infectious 'carriers' of rabies has not been
investigated.
IUCN Canid Specialist Group (CSG) response to pack extinction and the
handling-stress hypothesis
Initially the extinction of all 5 Serengeti study packs from rabies in 1991
was questioned by IUCN Canid Specialist Group (CSG) claiming that despite
the retrieval of empty radio collars and some skeletal material the disappearance
of all study packs did not prove high mortality, and that failure to locate
the packs was probably due to lack of monitoring and some emigration. Rabies
had been confirmed in all three study packs in the ecosystem from which suitable
tissue samples were obtained, including those in 1989 and 1990 from vaccinated
individuals in the Mara. Despite this in 1992 CSG claimed that as the Serengeti
study packs had been vaccinated against rabies they probably did not die
from rabies. It was suggested that some other pathogen such as Canine Distemper
may be involved, despite there being no serological evidence of exposure
of any wild dogs in the ecosystem to this pathogen.
[The claim made in a recent book that in 1990 the Tanzanian and Kenya study
packs were decimated by rabies is correct (McNutt & Boggs 1996). The
claim by the same authors that in 1991 when the rest of the study population
died out that, " ..evidence indicated that a canine distemper virus,
originating from the domestic dog population in the area, was responsible
for the second and final epidemic". is incorrect, there is no such
evidence.
However, in 1992 CSG proposed an immediate moratorium on handling of wild
dog in Serengeti and initiated an analysis of data (to be carried out by
Dr. J.Ginsberg) from other handled study populations, including the Mara.
Data from the Mara was crucial as this was the only other open plains dwelling
Lycaon population, all the remaining study populations in Africa being
in woodlands. The Mara was also the only other study population with high
disease related mortality and, crucially, in which free living packs had
been exposed to, vaccinated against and died from rabies (East 1996).
Unfortunately any possible adverse effects of the first experimental vaccination
of free living wild dogs in Africa, which began in the Mara in 1987, were
not considered by CSG in their data analysis (Ginsberg et al 1995a.). Dart
vaccination was not considered a form of 'handling' and individuals so treated,
at least 16 individuals including all 9 in one of the three study packs (cf.
just 2 in one pack as stated by Ginsberg 1996), were included in CSG's analysis
in the 'unhandled' category.
The results of the data analysis were circulated in August 1993. At that
date, and up to 1999, CSG reported the existence of an unhandled
population in the ecosystem. CSG recommended that only half of the packs,
which 'migrate' back into the Serengeti, should be handled, in order to permit
a comparison between handled and non handled packs. Packs do not
migrate, single sex groups and individuals emigrate from existing
packs.
The claim that "… the data presented by Burrows (Burrows et al 1994,
Burrows 1995) show clearly that the decline in the Serengeti Lycaon population
began in the early 1970s, nearly two decades before intensive handling
began ". (Ginsberg et al 1995b, p. 673) is misleading (see
Appendix
3a.-d. &
4.).
The implied significant decline (1970-73) and pack losses was based on incorrect
data for 1970 and the actual decline post 1970 was not statistically significant.
It was due to a reduction in the mean numbers of individuals in the study
packs due to poor reproductive success and lack of immigrants and not due
to a fall in the number of packs. There was no sudden pack extinction, other
than from shooting, until 1986 (Burrows 1995).
The loss of whole packs from disease was a post 1985 phenomenon coincident
with the introduction of intensive invasive handling of the packs in
both sectors of the ecosystem (see
Appendices
1,
2
&
3
and
Tables
3, 4
&
5).
As a number of authors, including Ginsberg, have pointed out, the pack rather
than the individual must be considered the basic unit of the population (e.g.
Woodroffe & Ginsberg 1999).
When in 1997 the death of all the study population was finally acknowledged
by CSG it was accepted that rabies was probably the cause of study pack losses
in all cases. CSG also accepted that a statistically significant association
between handling and reduced longevity of both individuals and packs had
been demonstrated in the Serengeti data, but in their reviewer/spokesperson's
(Dr.R.Woodroffe) opinion it is not necessary to invoke a causal relationship.
It is claimed that the Serengeti data can be explained as due to a chance
post vaccination epidemic/disease outbreak from which rabies vaccinations
failed to protect some packs (Woodroffe 1997). The chair of CSG is not so
certain (see quotation at top of page).
CSG's alternative 'disease outbreak' hypothesis cannot explain sporadic
study pack extinctions 1986-90 and persistence of an unhandled non study
population pre and post 1991
This hypothesis claims that the Serengeti data can be simply explained not
only by a hypothetical disease outbreak but by its timing which, based on
the claim that vaccinated dogs had less time to live before the epidemic
than unvaccinated dogs (Woodroffe 1997), the putative epidemic must have
been post vaccination. The first Serengeti pack to disappear post vaccination
(the Ndoha) was last seen alive in January 1991 (Woodroffe 1997 Table A1.1)
so any post vaccination 'epidemic' in Serengeti must have begun in 1991 not
vaguely in 1990-91 as Woodroffe claims.
Further :- " ..it seems unlikely that a highly pathogenic strain of rabies
was responsible for the disappearance of wild dogs from the Serengeti -Mara
" ( Woodroffe 1997).
An epidemic of rabies of a non highly pathogenic strain in 1991 cannot explain
the sporadic deaths of 5 unvaccinated Serengeti study packs between 1986-August
1990 one with rabies confirmed and suspected in the others, which was the
very basis for the stress/rabies hypothesis, nor can it explain the extinction
between 1989 and late 1990 of all the Mara study packs, again with rabies
confirmed.
Why should a non highly pathogenic strain of rabies virus cause limited sporadic
death of 5 Serengeti wild dog packs within a few months of a radio-collaring
event in each pack pre vaccination but suddenly in 1991, post vaccination
against rabies with an inactivated vaccine, lead to the elimination of all
study packs?
The evidence suggests that the same common 'canid' strain of rabies virus
identified in Serengeti in the rabid pack of 1990 (pre vaccination) was most
likely also the cause of the sporadic study pack deaths in Serengeti from
1986-90 and also caused the deaths of the 3 packs in the Mara between 1989-90
and the 5 pack deaths in Serengeti in 1991 post vaccination.
No evidence for a general rabies epidemic in the ecosystem in 1991
Rabies virus of at least two strains
(one a common canid strain and the other unique to hyaena) is known to be
endemic in some Serengeti wildlife but there is no evidence of epidemic
rabies in wildlife or domestic dogs in, or around, the ecosystem post 1977.
Despite the loss of all study packs by mid 1991 a non-study population of
wild dogs persisted and persists to date. There is no evidence for a general
rabies epidemic in 1991 which would have been expected not only to affect
wild dogs but other susceptible wildlife and local domestic dogs in and around
Serengeti. No such mortality was reported by researchers, rangers or the
many tourists which closely observe this area throughout the year.
An unhandled non study Wild dog population exsisted pre 1991 and persists
to date
In 1999, mindful of the problem presented by a putative 'epidemic' that
apparently selectively afflicted only handled wild dog study packs, the impact
of CSG's hypothetical 1991 'disease outbreak' was suddenly expanded to a
claim that the entire ecosystem wild dog population and with it a unique
genotype became extinct (Woodroffe & Ginsberg 1999).
It had earlier been claimed that the many confirmed wild dog sightings in
the ecosystem post 1991 were all 'immigrants'. This is neither consistent
with the claim that the Serengeti population was isolated, nor CSG's distribution
maps in their 1997 Action Plan which show wild dogs as 'vagrants' or absent
in the areas adjacent to the former study areas in Tanzania and Kenya. (Woodroffe
et.al. 1997).
The claimed extinction of the entire Serengeti-Mara ecosystem wild dog population
in 1991 is a myth: an unhandled resident breeding population existed in
the ecosystem pre 1991 and persists to date (2003).
Woodroffe's alternative 1991 'disease outbreak' hypothesis and claimed extinction
of the entire Serengeti -Mara ecosystem wild dog population can be dismissed
as it neither explains the pattern of Serengeti study pack extinction
beginning in 1986, nor can it explain the Mara data or why a non study
population survived.
Population Viability Analysis models
Doubts were recently expressed
concerning the ability of Population Viability Analysis models to help elucidate
the cause of the loss of the Serengeti packs:-
"…. from our results and PVA models presented previously (Burrows
et al 1995 ,Ginsberg et al 1995b), it is difficult to determine whether the
extinction of the Serengeti wild dog population is more likely to be due
to chance alone (Ginsberg et al 1995b) or to invasive research methods (Burrows
1992)." Cross and Beissinger (2001).
The Ginsberg et al's use of demographic data for 1975-76 on which their PVA
model is largely based is fatally flawed (see
Appendix
3d.). Unfortunately, Cross and Beissinger would not have been aware
of this when drawing their important conclusion.
CSG accept a statistically significant association between handling and
reduced longevity in Serengeti data but claim that handling is not causally
associated with mortality and reduced longevity in any ecosystem.
This claim is based on a flawed 'survival' analysis of data from originally
5 (Ginsberg et al 1995a) now reduced to 4 other 'ecosystems' with the Mara
data now excluded (East 1996, Woodroffe et al 1997) and a Population Viability
Assessment model (Ginsberg et al 1995b) using incorrect data (see
Appendix
3d.).
The validity of the Mara data in CSG's 'survival' analysis was questioned,
following which an attempt, that unfortunately failed, was made to obtain
the full Mara data from the researchers concerned, particularly those data
relating to the actual numbers of individuals vaccinationed (Woodroffe 1997).
Therefore in 1997 the Mara data supplied to CSG in 1993, the only data
of relevance to those from the Serengeti, was discounted as 'incomplete'.
Yet it was mainly on the basis of this 'incomplete' and inaccurate data from
the Mara (Burrows et al 1995), the northern sector of the same ecosystem
in Ginsberg's analysis that, in 1993, CSG recommended the lifting of the
moratorium on handling in Serengeti. If as Woodroffe & Ginsberg (1998)
claim the entire ecosystem Lycaon population had become extinct there
would have been little point in lifting the moratorium on handling.
The claim that handling had no effect on the survival of handled wild dogs
in any of the five ecosystems studied between 1987-93 (Ginsberg et al 1995a)
has now been reduced to four: Kruger NP (South Africa), Moremi GR (Botswana),
Selous GR (Tanzania) and Hwange NP (Zimbabwe). All are in woodland habitats
where, up to 1996, the populations were considered stable, or increasing,
and disease related adult mortality was rare with no evidence of exposure
to rabies. Data from such populations was therefore largely irrelevant as
a test of the handling-stress hypothesis which related specifically to a
wild dog population in Serengeti known to be exposed to, vaccinated against
and died from rabies (East 1996).
Is there evidence for handling-stress related problems in other wild dog study populations ?
Although the handling-stress hypothesis
was originally based on Serengeti data, other published data suggests handling
problems in other populations, whatever the precise mechanism involved may
be. Handling-stress may explain population declines and rabies deaths in
other wild dog study populations. For example data from the Mara study area
in Kenya (1987-90) show that 33% radio-collared dogs and 70% of those blood
sampled survived less than 12 months following anaesthetisation. In South
Africa and Namibia unexpected outbreaks of rabies in vaccinated translocated
or reintroduced packs and groups emerged as a 'new' post 1988 phenomenon.
In Kruger (1990--93) 53% of individuals implanted with radio-transmitters
and 40% of those fitted with radio-collars survived less than 12 months post
handling (Ginsberg et.al. 1995a: also see this
Appendix
7).
The comparative data for survival of unhandled individuals in the
Ginsberg et.al. study is based on the false assumption that dispersal equals
death. The confirmed high survival of dispersing unhandled individuals (and,
in Serengeti, radio collared yearlings) post emigration is ignored (Burrows
1995, Burrows et.al. 1995).
More recently in Kruger NP, prior to a dramatic decline in reproductive success
in 1996 and 1997, the alpha pair if known, or if not another adult pair
in each pack (36 packs in 1995), were anaesthetised and blood sampled
(Girman et al 2001). Similarly in Moremi at least one of the alpha pair was
usually selected for anaesthetisation this time for radio-collaring and blood
sampling (McNutt 1996). Any adverse effect on subsequent reproductive success,
or on disease related or accidental death, following the deliberate selection
of breeding individuals for anaesthetisation to provide data for such research
projects was not, it appears, considered to be important. The death of
either of the alpha pair will result in pack dissolution (Burrows:
http://www.africanconservation.org/wilddogs)
These worrying developments suggest that a causal association between handling
and rabies in wild dog packs exists in other ecosystems but is being ignored
. A population decline in Kruger following the anaesthetisation of so many
adult dogs in the National Park would be predicted from the handling-stress
hypothesis whatever the precise causal mechanism involved.
Uncritical use of raw and inappropriate data.
Most reviewers of the Serengeti
wild dog demographic data for the 1960s and 1970s uncritically accept raw
data showing inter-year fluctuations in the number of known and unknown
individuals seen in a study area in a year by different observers as
an accurate reflection of the health of the population.
The Serengeti plains packs studied from 1964 onwards was NOT the entire
population of 'Serengeti', the National Park, the 'ecological unit' or the
ecosystem as is implied by some commentators (e.g Estes 1991, Creel 1992)
it was geographically and logistically a convenient sample to study.
Anecdotal accounts and opinions expressed in popular books and some research
publications of claimed disease related population crashes in Serengeti wild
dog study packs between 1967-77 have been uncritically accepted as facts.
Such 'facts' are the basis of a claimed 1971-3 population crash, continued
decline and further crash in 1975-6 which would, due to chance events alone,
such as an 'epidemic', lead to the inevitable extinction of the small Serengeti
ecosystem population of wild dogs (e.g. Estes 1991, Ginsberg et al 1995b,
Ginsberg 1996, Creel 1992).
Therefore, it is conveniently concluded, that there is no need to look for
any other cause. Any possible adverse effects of handling can therefore be
ignored and routine handling for data collection continue unhindered. However,
this convenient self-fulfilling prophecy is contradicted by all the available
evidence (Burrows 1994, Anon 1994; and see this
Appendix
8 and Tables 2 &
3).
Based on published data the same number of packs (5) were present
on the Serengeti plains in early 1991 as in 1970 (Burrows 1993, and
see
Appendix
4 & Table 1. There is no evidence for either a series of population
crashes or a dramatic decline in the Serengeti population of wild dogs before
the introduction of routine invasive handling i.e. pre 1985.
Prior to 1991 there is no evidence for a significant change in the number
of wild dog packs resident on the Serengeti plains from 1964-91 (Burrows
et al 1994, & this Table 1).
The decline in the Serengeti wild dog study population in the 1970s was in
the mean number of adults in the study packs not in the number
of packs the basic unit of the wild dog population ( Woodroffe &
Ginsberg 1999) and was caused by a prolonged (8 year), but temporary significant
decline in reproductive success and a lack of immigrants
(Appendix
4.).
The lack of immigrants suggests that whatever the cause of poor reproductive
success it affected wild dog packs not only in Serengeti but over a very
wide area (see
Appendix
7 for a possible explanation). Post 1978 the wild dog population
in both Serengeti and the Mara was expanding up to 1990, despite the
sporadic extinction of whole study packs, due to good reproductive success
of the remaining packs and immigration of unknown groups (Fuller et al 1992,
Burrows 1995).
By contrast in Kruger just 2 years of poor reproductive success and
pup survival is claimed to have resulted in a significant decline in numbers
of individuals which might be expected. What this would not explain is the
dramatic fall in the number of packs which would not be expected.
The loss of reproductive success alone cannot explain the Kruger data from
1995-2000, which suggests that this population decline, unlike that in
Serengeti in the 1970s, involved high adult mortality.
No data analysis of the effects of the extensive handling post 1993 has
been published
To date, since their limited survival analyses based on data collected
up to and including 1993 (Ginsberg et. al. 1995a), no further data from
any of the ecosystems where handling of wild dog continued with greatly
increasing frequency for up to a further nine years post 1993, has been published
to support CSG's contention that handling has no effect on longevity or survival
in any of the ecosystems, or is essential for the conservation of wild dogs.
Currently there are no unhandled study packs in Africa from which
base line data on individual and pack longevity is being collected by non
invasive techniques. The only such data comes from Serengeti where both the
longevity of unhandled individuals and their packs was significantly
greater than that following the introduction of routine invasive handling
in that ecosystem in 1985 (Burrows et al 1994).
In the opinion of Woodroffe, the sole author of CSG Action Plan Appendix
1, the association, post 1985, between intensive handling and reduced longevity
in the Serengeti data set can be explained by a hypothetical disease
outbreak/epidemic without assuming a causal relationship (Woodroffe 1997).
There is no evidence for such an epidemic at such a time. The chair of CSG
is obviously equally unsure (see above).
However, on the basis of a flawed PVA model, an irrelevant and a flawed
'survival' analysis (Ginsberg et.al. a&b) and the timing of a hypothetical
epidemic for which there is no evidence (Woodroffe 1997), Woodroffe claims
to be able to reject the handling-stress hypothesis as " not the best
explanation" for the loss of the Serengeti -Mara study packs between
1995-91 (Woodroffe 1997).
It may not be the 'best' but it is currently the only explanation that fits
the data for the reduced longevity of handled individuals and their packs
post 1985 and the associated pattern of mortality pre and post vaccination
in Serengeti and Mara wild dogs whatever the precise mechanism involved
may be (Burrows et al 1995 & 1995).
The claimed benefits of invasive research lack supporting data
"Despite the potential impact that such activities [handling] may have,
it is frequently tacitly assumed ,or subjectively assessed, that these techniques
have little effect on individuals in the long term" (White & Garrot,
1990).
"The tacit assumption that immunisation can only benefit a population
is a dangerous one, particularly when dealing with free-living
animals."(Hall & Harwood 1990).
"It is likely…. that in many studies any adverse effects [of
handling] are either unnoticed, perhaps because they are rare, or, more
likely, because they are not reported" (Laurenson 1992)
The current routine use of radio collaring in wild dog research is based
on an assumption that any inevitable risks involved are outweighed by the
benefits and a claim that it is essential in facilitating data collection.
Unfortunately the plethora of research papers on wild dogs post 1990, most
based on use of data collected using invasive techniques, have failed to
provide information of any practical help to wildlife managers such as early
warning of disease problems or major fluctuations in population and their
causes. Apart from Serengeti, nothing is published from other studies relating
to the longevity of radio-collared individuals compared with that
of unhandled members of the same population, or the cause of death of so
many radio collared individuals, most vaguely attributed to 'natural causes'
(Ginsberg et al 1995a see
Appendix
5).
Is routine invasive handling of free living wild dog essential for the welfare and conservation of the species?
To date invasive research on wild
dogs has failed to provide any clear welfare or conservation benefits to
any population of this highly endangered species
(Appendix
1 &
5
and Tables
3 - 5).
Analysis of data from wild dogs obtained by invasive research showing that
such techniques enhance, rather than decrease, the survival of free living
or translocated populations has yet to be produced. Analysis of Serengeti
and Mara data suggest adverse effects of handling. This begs the question:
why continue to use such techniques when non invasive techniques are available?
(Appendix
2.).
"To minimize risks, however, wild dogs should only be immobilized or radio
collared when …. there is a clear conservation benefit to the study
…." (Woodroffe 2001).
Unfortunately to date, the 'risks' involved in any ecosystem where such risks
need to be minimized have not been assessed. Woodroffe's conclusions concerning
the risks associated with immobilization/anaesthetisation and radio-collaring
are based on the 'incomplete' and very limited analysis of data from 4 ecosystems
up to 1993 on survival of individuals up to 12 months post handling for
radio-collaring and blood sampling only (Ginsberg et al 1995a) and a
statistically flawed analysis (Creel et al 1997) of the effect of radio collaring
on stress hormones mainly in males selected for their vigour (East et al
1997).
Until the risks involved in all forms of handling are properly assessed,
taking into account the published data on handling collected from packs
throughout Africa pre 1994 and from those that have continued to be intensively
handled since 1993 which remain unpublished; any benefits of handling
to welfare and conservation of wild dogs in the wild, or in translocated
packs remain to be demonstrated.
It is essential that all handling data including details of age, sex and
social position of the individuals handled and the historic vaccination data
from the Mara should be made available.
The costs in terms of death of wild dog packs and groups following all
vaccinations and within less than 4 months of radio collaring in Serengeti-Mara
and unexplained population declines in Kruger and Moremi following
anaesthetisation of alpha individuals appear to be very high.
The Serengeti analysis, associating handling with reduced longevity of
individuals and packs is recognised (Anon 1994, Woodroffe 1997). It now appears
that handling induced stress may also contribute to other population declines
and the loss of handled free living and managed reintroduced wild dog study
populations from rabies in southern Africa, whatever the mechanism involved
may be.
The ability of some researchers to ignore the obvious risks involved in
anaesthetising any animal and their dismissal of any suggestion that handling
may be detrimental to wild dog individuals and their packs is clearly displayed
in a recent publication:-
In Kruger National Park which has
the only viable wild dog population in South Africa :- ..................
"To test
..[an]…hypothesis, the alpha male and female if known (Girman
et al 1997), or one adult male and one female from each pack in the Kruger
population were selected" . (Girman et al 2001).
The selected individuals were anaesthetised and blood sampled just prior
to the dramatic 60% population crash of 1995-2000.
The Kruger research clearly seems to conflict with recent statements made
by IUCN CSG spokesperson on wild dogs:-
"It is vital that future plans for wild dog conservation be based upon
realistic assessments of the costs and benefits of intervention. "
and :- "To minimize risks, however, wild dogs should only be immobilized
or radio collared when …. there is a clear conservation benefit to the
study …." (Woodroffe 2001).
Given the known risks of immobilising/anaesthetising individuals, the practise
of radio collaring and re-collaring previously collared alpha individuals
as in Moremi GR would appear to be particularly dangerous and is perhaps
one of the reasons for the reported low pack longevity in Moremi with most
packs not lasting more than 2 years (McNutt 2000) and the catastrophic decline
in the Kruger population between 1995-2000.
Despite the accepted statistically significant association between handling
and reduced longevity in the Serengeti study population and the confirmation
of rabies as the only known cause of study pack deaths in the Serengeti and
the Mara and the comments of the IUCN CSG (see above and page 1) an extraordinary
claim and recommendation is made :-
"Causes of mortality in the wild dog population are still unknown. To
rescue this endangered species an intensive radiotelemetry study is
required" (Sincair 1995).
If, as is claimed, the entire Serengeti-Mara population became extinct in
1991 (Woodroffe & Ginsberg 1999) Sinclair's recommendation would be the
equivalent to attempting to rescue the Dodo. It is also extremely ill advised
given the uncertainty of the extent to which handling precipitated the extinction
of the Serengeti-Mara study packs in 1991 and the association of all forms
of handling both free living and captive populations of Lycaon with
high mortality (Burrows et.al. 1994 & 1995, Anon 1994, East 1996, East
et.al. 1997, Woodroffe 1997, East & Burrows 2001, van de Bildt 2002 ).
However, as the entire Serengeti-Mara population did not become extinct in
1991, then by ignoring all the available scientific data, Sinclair's
recommendation for an intensive radio telemetry study is more likely to hasten
the decline of this endangered species than to rescue it.
A former warden in Serengeti from 1956-72 wrote:- "One thing is sure :
it [research] was a great confidence trick, and virtually nothing has ever
come out of it to help the hard-pressed animals of East Africa (Turner
1987,163).
CONCLUSIONS AND RECOMMENDATIONS
Based on the Serengeti and Mara data derived from study packs all of which were radio-collared it can be concluded that:-
1. The risks involved in the Serengeti
of immobilization/anaesthetisation of individuals exposed to rabies and in
particular of older adults not in their natal pack's home range appear to
be very high with death of the individual and their pack from rabies all
within 4 months (n=6, average 2.5 months).
2. The survival of individuals of any age (and their packs) which
have been/are exposed to rabies virus, dart vaccinated and subsequently radio
collared, i.e. handled at least twice, is low (n=2, average 2 months). The
majority (70%) of dart vaccinated individuals which died were in these 2
packs (Salei & Ndoha) which were handled in this way in early 1991. This
accounts for the fact that dart-vaccinated individuals which were not immobilized
had the shortest mean survival after handling.
3. The average survival post immobilization/anaesthetisation of unvaccinated
yearling pre dispersers is 25 months cf 2.5 months for older individuals
handled as post-dispersers (1. above).
4. The vaccination of free living wild dogs and those in captivity prior
to release with inactivated vaccines developed for domestic dogs is associated
with high mortality from rabies.
5. The death of either of the alpha pair in Serengeti and other ecosystems
is usually followed by dissolution of the pack with the pack splitting into
single sex groups.
6. Genetic and other studies can now be undertaken using non invasive methods.
7. The claim that routine invasive research is 'safe', with any risks outweighing
the inevitable risks involved, and is essential for the conservation of
Lycaon has no data support from any ecosystem .
From the conclusions above practical recommendations follow:-
1. If for very good conservation reasons it is considered essential
to radio-collar a free living wild dog, it should be a yearling still in
its natal pack's home range.
2. Due to the inevitability of pack dissolution or other disruption following
the accidental death of an alpha dog in a pack, no individual of alpha status
should be immobilised/anaesthetised or 'bounce dart' tissue sampled.
3. It is crucial that the role of cell mediated immunity in protecting wild
dogs from naturally occurring pathogens in their environment be understood.
Currently the effect of stress on cell mediated immunity in Lycaon is unknown.
The available vaccination data (from both free living and captive groups)
strongly suggests that, the immune system, particularly of older
Lycaon individuals is unusual.
4. Free living wild dog packs and individuals should not be subjected to
experimental invasive research.
5. The remaining viable and self sustaining populations such as those in
Ruaha National Park and Selous Game Reserve / Mikumi National Park in Tanzania
should be invasive research free areas and, by concentrating on the conservation
of habitat and education, conserve all their diverse inhabitants including
their vitally important populations of wild dogs.
6. If research on this endangered species is to achieve its claimed aims
of improving the welfare and survival of free living and captive wild dogs
all relevant data and the sometimes contentious and inconvenient conclusions
drawn from them must be fully published and taken into account.
These recommendations, if adopted, are likely to have very positive effects
on the future welfare and conservation of the remaining free living wild
dogs in Africa.
"History can offer perspectives that we would not otherwise have, and
case histories of endangered species programs have the potential to serve
as a corrective for past mistakes. This tool is diminished if accounts are
not chronicled accurately."
Ken Alvarez ,'Twilight of the Panther: Biology, Bureaucracy and Failure in
an Endangered Species Program'
"If such a mess can be made of efforts to save a creature as attractive
as the black-footed ferret in a country as well organized and prosperous
as the United States, prospects for conservation in other parts of the world
are indeed bleak." Robert May Nature 1986.
All Tables and Appendices - Bibliography & References
To
Browse
 |
To
Search
|
|
