CSIRO PUBLISHING

www.publish.csiro.au/journals/wr Wildlife Research, 2005, 32, 715–722

Mortality and survivorship of the quokka (Setonix brachyurus)

(Macropodidae:Marsupialia) in the northern jarrah forest
of Western Australia

Matt W. HaywardA,B,C,D,E, Paul J. de ToresB, Michael L. AugeeA and Peter B. BanksA

A
School of Biological, Earth and Environmental Science, University of New South Wales,
Sydney, NSW 2052, Australia.
B
Department of Conservation and Land Management, Wildlife Research Centre,
PO Box 51, Wanneroo, WA 6946, Australia.
C
Department of Conservation and Land Management, Dwellingup Research Centre,
Banksiadale Road, Dwellingup, WA 6213, Australia.
D
Present address: Terrestrial Ecology Research Unit, Department of Zoology,
Nelson Mandela Metropolitan University, PO Box 77000, Port Elizabeth 6031, Eastern Cape Province, South Africa.
E
Corresponding author. Email: [email protected]

Abstract. The potential for the quokka (Setonix brachyurus (Quoy & Gaimard, 1830)), a threatened macropodid
marsupial, to increase in abundance following the initiation of predator control was investigated by determining the
cause of deaths of radio-collared individuals. Predation was identified as a major cause of death followed by road
kills. The non-parametric Kaplan–Meier method modified for staggered entry of individuals was used to estimate
survivorship. Although males and females were affected differently by each cause of mortality, their overall
survivorship did not differ significantly. Individuals alive at the beginning of the 25-month study had a 61% chance
of surviving to the end. This represented an 81% chance of surviving for 1 year. There was no significant difference
in survivorship between adults and juveniles. Current rates of adult and juvenile survivorship should allow
population recovery, although none has been evident. Pouch young mortality is hypothesised to have inhibited the
anticipated quokka population increase since the initiation of predator control. The observed expulsion of pouch
young by females when threatened may be a primary predator avoidance strategy.

Introduction summer (Main 1959), which is attributable to protein starva-

For species that exist as a metapopulation, the extinction of tion from feeding on nutrient-deficient succulents in
local populations is a natural process and of little concern attempts to derive sufficient moisture to survive through the
provided that colonisation and extinction rates are in equi- dry season (Storr 1964). The ecology of the mainland popu-
librium (Harrison 1991). When a metapopulation is in a state lation is very different from that of the Rottnest Island popu-
of collapse, however, the extinction of each local population lation, being in a more mesic environment and subject to
is of much greater bearing. predation pressure from the red fox. This paper estimates
The quokka (Setonix brachyurus) is a threatened, macro- survivorship for quokkas from sites within the northern
podid marsupial that is endemic to south-western Australia jarrah forest metapopulation and assesses this in light of the
and that once existed naturally as a classic metapopulation recovery of the species.
(Hayward 2002; Hayward et al. 2003, 2004, 2005). The
species suffered a dramatic decline in the 1930s, primarily Methods
from predation by the introduced European red fox (Vulpes General procedures
vulpes), and secondarily from habitat alteration and possibly Quokkas were trapped seasonally at five sites in the northern jarrah
disease (White 1952; Cook 1960; Hayward 2002). This led forest of Western Australia (Chandler, Rosella Road, Kesners, Hadfield
to the collapse of the metapopulation structure into a non- and Victor Road) from spring 1998 until summer 2000–01 using two
equilibrium state (as described by Harrison 1991), such that types of wire mesh cage traps baited with apples or universal bait and
the conservation status of the quokka is now listed as vul- set for eight consecutive days (methods and results of this trapping are
described in detail in Hayward et al. 2003). The northern jarrah forest
nerable according to IUCN criteria (Hayward 2002). comprises the northern half of the Jarrah Forest bioregion (Thackway
The quokka has been extensively studied on Rottnest and Cresswell 1995) and encompasses the Darling Plateau from east of
Island where the species experiences seasonal mortality over Perth to Collie in the south. All but the Victor Road site (a control site)

© CSIRO 2005 10.1071/WR04111 1035-3712/05/080715

716 Wildlife Research M. W. Hayward et al.

were baited monthly with dried meat baits injected with 4.5 mg of captured individuals were collared (48 of 48 trapped adults and 10 of 23
sodium monofluoroacetate (1080) poison to control introduced preda- juveniles were collared: Hayward et al. 2004), these individuals were
tors. No estimates of fox or cat density were conducted around the considered a random sample of the population. Other assumptions sat-
quokka sites, although we assume that their density was very low given isfied were that survival of each individual was independent and that
that baiting seasonally in surrounding areas reduced introduced preda- censoring was random (Pollock et al. 1989a, 1989b). Any newly col-
tor impacts significantly, especially when coupled with perimeter lared animals were assumed to have the same survival rate as those pre-
baiting every two months (de Tores 1999). viously collared (Pollock et al. 1989a).
Captured quokkas were sedated and a radio-transmitter was fitted to Survival estimates were derived from 58 animals from the five
64 (Hayward et al. 2004) of an estimated adult population of 84 populations (Hadfield, where 17 of 24 individuals were collared;
(Hayward et al. 2003). Radio-collars (Biotrack, Institute of Terrestrial Kesners, 26 of 26; Chandler, 4 of 7; Victor Road, 10 of 13; and Rosella
Ecology, Wareham, UK) were configured to maximise signal strength Road, 1 of 1) and 10 of these were juveniles. An additional eight indi-
and battery life while minimising weight. The collars incorporated viduals were excluded from the analyses as their collars ceased trans-
movement-sensitive circuitry with a 2.5-h immobile period required to mitting within the recommended seven-day conditioning period
trigger the mortality pulse frequency of ~110 pulses min–1, compared to (Pollock et al. 1989a). Two deaths associated with trapping were
55 pulses min–1 in ‘live’ mode. excluded from the analyses as these would negatively bias the survival
Breakaway collars were used on juveniles and these employed a per- estimates (Pollock et al. 1989a). We don’t know whether this condi-
ishable rubber collar that stretched with the animal’s growth (Hayward tioning period was of appropriate length or not, but the first death in the
et al. 2004). Juveniles weighed 0.55 to 1.6 kg for females and 2.5 kg for collared individuals occurred two months after collaring. Animals
males (whereupon they were considered adult, after Hayward et al. whose collar ceased transmitting and whose fate was unknown were
2003) and were fitted with the expandable collar while they were in the censored. Considerable effort was made to locate censored individuals
presence of their mother and still suckling (Hayward et al. 2004). in order to reduce the confidence intervals associated with the survivor-
Animals were regularly monitored between November 1998 and ship estimates. Although many censored animals were later recaptured
November 2000, with a particularly intensive period between March (27 recaptured out of 46 censored from 450 individuals at risk) without
1999 and May 2000 so as to allow direct comparison between the sur- their collars, they were still classified as censored in the analyses to
vivorship of animals at each site. This comparative monitoring period avoid bias (Pollock et al. 1989b).
occurred when all sites had collared animals and ceased when collar Survival estimates from the Kaplan–Meier method were compared
removal began. The location of each collared individual was determined using the log-rank (χ2) test (Pollock et al. 1989a). This test assumes that
at least monthly and generally more frequently (up to four times per censoring is random but may be violated by a predator killing an animal
week). This meant that all but two of the dead individuals were found and simultaneously destroying the transmitter (Pollock et al. 1989a).
within two days of death. Consequently, the most conservative modification of the log-rank test
The cause of death was determined by examining characteristic (with a modified variance of the number of deaths) (Pollock et al.
markings on the carcass (Augee et al. 1996). Predation was attributed 1989a) was used in this study. The K-sample Mann–Whitney test was
to the red fox if there was minimal marking on the collar or, if marked, used to compare survivorship between sites (Zar 1996). Comparisons
it showed evidence of large canid teeth, if the gut had been eaten or between survival estimates at the time of the final sample were con-
cached in an excavated hole, and if the fur and uneaten remains were ducted using the normal test statistic equation (Pollock et al. 1989a).
cached (Augee et al. 1996; de Tores 1999). Predation was attributed to The log-rank test compared survival functions in their entirety while the
feral cats (Felis catus) if there was evidence of bite marks to the back of approximation to the normal (Z) test compared the survival curves at
the skull, which often included the removal of the brain, if the gut the end of the 15-month comparative period. Individual comparisons
remained near the carcass, and if there was no evidence of caching or were not conducted on the Rosella Road site as the only collared animal
simple covering of the carcass with debris (Augee et al. 1996; de Tores was censored after 8 months.
1999). The western quoll or chuditch (Dasyurus geoffroii), although
present at each site, was not considered a potential predator of quokkas
considering its largely insectivorous diet and it being known to prey Results
only on mammals significantly smaller in size than quokkas Overall survivorship
(Soderquist and Serena 1994). Dingoes (Canis lupus dingo) were not
considered as potential predators of the quokka because of their scarcity The 58 collared individuals were monitored over the 25-
in the northern jarrah forest (MWH, personal observation; PdeT, month study period, during which eight deaths were recorded
unpublished data). Raptor predation was characterised by carcasses (Fig. 1). Of the known causes of death, three deaths were
being located below trees, when skulls were intact and/or skin peeled
attributed to predation and two were road kills (Table 1).
back and soft tissue removed (Augee et al. 1996). The regular moni-
toring meant that predation by carpet pythons (Morelia spilota) would Only females were depredated, whereas only males at the
have been identified by the presence of the collared animal still within Kesners site were killed after being struck by motor vehicles.
the snake (Augee et al. 1996). Other evidence of python predation Despite the existing predator-control program (de Tores
would have been the collar found within a regurgitated pellet contain- 1994, 1999), foxes and cats still preyed on quokkas.
ing crushed or fractured bones. Road-killed individuals were located on
The Kaplan–Meier survivorship estimate for all quokkas
road verges, had obvious massive trauma and showed no evidence of
predation. Carcasses that could not be satisfactorily placed into one of for the entire 25-month study was 0.61 (95% confidence inter-
these categories were classed as unknown mortality events. vals (CI) = 0.34–0.87) (Fig. 1). This equates to an annual sur-
vivorship of 0.81. This means that an individual alive at the
Survival estimate
beginning of the study had a 61% chance of surviving for 25
The Kaplan–Meier method (product limit estimator) with staggered months or an 81% chance of surviving for a year. There was no
entry was used to assess survivorship, because the assumptions of the
alternative parametric estimators were not always met in this study
significant difference between the survival functions between
(Pollock et al. 1989a; Kendall and Pollock 1992). As trapping was years for all sites combined (November 1998 until October
assumed to be unbiased (Hayward et al. 2003) and a high proportion of 1999 and November 1999 until October 2000) (Table 2).
Mortality and survivorship of the quokka Wildlife Research 717

Kaplan–Meier survivorship (S[t])

1.0

0.8

0.6
Fig. 1. Kaplan-Meier survivorship plot of quokkas
from all sites combined over 25 months beginning
0.4 8 November 1998. Confidence limits (95 percentile)
are shown as dashed lines.
0.2

0.0
Apr-99
May-99

Jul-99

Jul-00
Nov-98

Aug-99
Sep-99
Oct-99

Apr-00
May-00
Nov-99

Aug-00
Sep-00
Oct-00
Nov-00
Jan-99
Feb-99
Mar-99

Jun-99

Jan-00
Feb-00
Mar-00

Jun-00
Dec-98

Dec-99

Comparative period of survivorship (March 1999–May 2000) Discussion

Quokkas at each site were intensively monitored over a Survivorship
15-month comparative period from March 1999 to May Overall, quokkas at the study sites had an 81% chance of sur-
2000. There was no difference between the survival curves of viving for 1 year. The similar survivorship values for each sex
all sites combined (Kruskal–Wallis K-sample test) or for the found in this study is unusual for a sexually dimorphic species
final survival estimate (Z statistic) (Table 2, Fig. 2). like the quokka. Such species generally exhibit lower male sur-
There was no significant difference between the survival vivorship owing to the costs associated with competition for
curves or estimates for males and females (Table 2, Fig. 3). mates, dispersal and increased food demands (see discussion in
Males had a final survival estimate of 0.82 (95% CI = Owen-Smith 1993). The absence of dispersal in quokkas and
0.61–1.02), whereas that for females was 0.75 (95% CI = similar home-range sizes between the sexes (Hayward et al.
0.53–0.98). There was also no significant difference between 2004) may explain the similar survival of the sexes.
the survival curves and final estimates of adults and juve- The similarity in the survival curves between the sexes
niles (Fig. 4, Table 2). Adult survivorship was 0.77 (95% and for all sites combined should be interpreted cautiously
CI = 0.59–0.94) during the comparative 15 months, and because of the wide confidence intervals arising from the
during that period no collared juveniles died. number of censored animals (Fig. 2). Comparisons between

Table 1. Causes of death of the 58 radio-collared quokkas in the northern jarrah forest
Study site and approximate date of death are also shown

Fate Males Females Study sites (approximate date of death)

Predation by fox 0 2 Hadfield (20.v.1999); Kesners (26.viii.1999)
Predation by cat 0 1 Kesners (16.xii.1999)
Road kill 2 0 Kesners (29.viii.1999, 15.iii.2000)
Unknown 2 1 Hadfield (2.iv.1999); Kesners (1.x.2000); Chandler (27.i.1999)

Table 2. Results of log-rank and normal tests of survivorship

Total number of animals collared in each comparison is shown in parentheses. The Bonferroni correction was applied to the comparisons between
sites yielding a significant probability of P = 0.008

Comparison Log-rank test for pairwise comparison/Kruskal–Wallis Z statistic

test for all sites
χ2 /H d.f. Probability Z d.f. Probability
All sites 1.8711 4 0.90 > P > 0.75 2.46 4 0.75 > P > 0.50
Males (33) versus females (25) 0.1781 1 0.75 > P > 0.50 1.14 1 0.50 > P > 0.25
Adults (48) versus immatures (10) 0.9699 1 0.50 > P > 0.25 2.61 1 0.25 > P > 0.10
Nov. 1998 – Oct. 1999 versus Nov. 1999 – Oct. 2000 0.2849 1 0.75 > P > 0.50
718 Wildlife Research M. W. Hayward et al.

1.0
Kaplan–Meier survivorship

0.8

Fig. 2. Survival functions of quokkas at the five

0.6
sites in the northern jarrah forest between March
1999 and May 2000. The number of collared
0.4 individuals at each is shown in parentheses followed
by the number of censored individuals
(collared/censored). The Rosella and Chandler sites
0.2
were conservatively assumed to have a survival
function of 0 and a variance of 1 after all collared
0.0 animals were censored.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Month
Hadfield (17/8) Victor Road (10/5) Kesners (26/16)
Chandler (4/4) Rosella (1/1)

adults and juveniles suffered similar problems and these mortem examination of the least decomposed (an adult
were compounded by the short lifespan of breakaway collars female) of the three deaths from this study not attributed to
that were fitted to juveniles (Hayward et al. 2004). predation or road kill revealed fat reserves, a full stomach and
bladder and an absence of obvious internal parasitic cysts,
Mortality events and the causes of the quokka’s decline suggesting that disease was not the cause. Although this death
Natural deaths for the mainland quokka populations in the occurred in summer, when quokkas may die of dehydration
past derived from disease, predation, starvation and the owing to their high water requirements (Main and Yadav
direct and indirect effects of fire (Hayward 2002; Hayward 1971), this seems unlikely because water was plentiful at the
et al. 2003). Although deaths associated with starvation and Chandler site throughout the year owing to the presence of a
fire were not recorded in this study, various forms of preda- dam and the animal had a full bladder. The other two quokkas
tion, and possibly disease, were. whose cause of death could not be categorised may have suc-
Wild quokkas have been found to be susceptible to infec- cumbed to disease. A recent re-evaluation of the deaths that
tion by Toxoplasma parasites (Gibb et al. 1966) and Cook (1960) attributed to disease indicates that surplus
Salmonella bacteria (Iveson and Hart 1983; Hart et al. 1986). killing by foxes may have been the cause (Short et al. 2002).
Captive quokkas are known to be parasitised by the genera Although it is unlikely that surplus killing by foxes caused the
Progamotaenia (cestode tapeworms); Austrostrongylus, deaths of these three quokkas, it may have been important in
Breinlia, Cloacina, Dipetalonema, Filaria, Microfilaria the initial decline of the quokka in the 1930s as we find it
(nematode flatworms); and Entamoeba and Toxoplasma unlikely that disease was a major factor (Hayward 2002).
protozoans (Collins 1973). The herpes virus has also caused The droppings and regurgitated pellets of nocturnal birds
mortality in captive quokka colonies (Burnet 1970). Post- of prey, such as the masked owl (Tyto novaeholliandiae) or

1.0

0.8
Kaplan–Meier survivorship

0.6

Fig. 3. Survival functions of male and female quokkas

0.4
in the northern jarrah forest between March 1999 and
May 2000.

0.2
Males Females

0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Month
Mortality and survivorship of the quokka Wildlife Research 719

1.0
Kaplan–Meier survivorship

0.8

0.6
Fig. 4. Survival functions of adult and immature
0.4
quokkas in the northern jarrah forest between March
1999 and May 2000.
0.2 Adults Sub-adults

0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Month

barking owl (Ninox connivens), contain quokka bones common sources of quokka mortality (Table 1). Roads and
(Archer and Baynes 1972). Quokka bones were found below tracks traverse the jarrah forest, delineating forest blocks as
a wedge-tail eagle (Aquila audax) eyrie on Bald Island (Storr management units. Where these cross the preferred habitat
1965a) and ospreys (Pandion haliaetus) prey on quokkas on of quokkas, in the upper reaches of creek systems (Hayward
Rottnest Island (Storr 1965b). Although this suggests that et al. 2005), the potential for road kills exists. Most of these
birds of prey are threats to the quokka, there was no evidence tracks are closed to the general public and are poorly main-
of raptor predation during this study. tained, which limits vehicle speed, such that they are a
The Aboriginal, and later European, inhabitants of the minimal threat. The high traffic volumes of surfaced roads
south-west regularly preyed upon quokkas (Gould 1863; and the speed that vehicles travel on them provides a more
Stewart 1936; Whittell 1954; Gardner 1957). The removal of serious threat to wildlife, including quokkas. The Kesners
the Noongar people of south-western Australia from a tradi- site has a 200-m tributary of the main swamp bisected by a
tional lifestyle has meant that such predation events no bitumen road, and two individuals were killed crossing to this
longer occur. Despite recent records of humans killing isolated section of swamp, probably to access the adult
quokkas on Rottnest Island (AAP 1998; Squires 2003), no females known to inhabit the area (Hayward 2002). Although
predation event on quokkas attributable to humans was quokkas in the Kesners swamp are most at risk of road kill,
recorded in this study. the lower reaches of the Hadfield swamp is also bisected by
Hunting pressure has been replaced by predation pressure a large, but unsurfaced, road that carries a high volume of
from other introduced placental predators, such as the dingo, traffic travelling at high speed. Given the impact of road kills
European red fox and feral cat. The dingo occurs at such low on quokka populations, it would be beneficial to incorporate
numbers in the jarrah forest that it is not considered a signif- traffic-calming structures and wildlife underpasses into the
icant threat to quokkas. The fox and, to a lesser extent, the cat design of roads that potentially affect quokka swamps.
are significant threats to the survival of quokkas (Hayward
2002; Hayward et al. 2003) and the only predation events Reasons for the lack of response to introduced
recorded in this study were attributed to these species. predator control
Quokkas may be most susceptible to predation in the Female quokkas can produce 1.8–2.0 young per year
wetter months when swamps become inundated and the core (Sharman 1955a, 1955b; Shield 1964) and thus have the
home range shifts towards their periphery (Hayward et al. potential to produce 17 young over a lifetime (Hayward et al.
2004). Higher rainfall also results in toxins in the dried meat 2003). This high potential fecundity did not result in sub-
baits used to control introduced predators leaching out more stantial population increases in the study sites despite the
rapidly, minimising their long-term effectiveness (Fleming relaxation of predation pressure (Hayward et al. 2003).
and Parker 1991; Saunders et al. 2000; Twigg et al. 2000). Evidence of the lack of population increases include a
Two predation events occurred in this period and both were paucity of recent museum records (Western Australian
attributed to foxes (Table 1). Although these small sample Museum, personal communication, in Hayward et al. 2003),
sizes are inconclusive, hunting by foxes may be hindered by reports of quokka population declines elsewhere (Maxwell
dense vegetation in swamps, in contrast to the cat, which may et al. 1996), and trapping in the 1970s and early 1990s at
be equally as successful inside the swamps as outside. several of our study sites that indicated similar, and generally
In addition to recently arrived predators, other causes of larger, populations previously (Hayward 2002) when com-
mortality have appeared since European arrival. Today, road pared to now (Hayward et al. 2003). There is evidence that
kills are one of the most common ways of identifying new the decline is due to the low juvenile recruitment rate
quokka sites throughout the mainland and one of the most because fewer than one in two pouch young survived to inde-
720 Wildlife Research M. W. Hayward et al.

pendence (Hayward et al. 2003). The collaring of juveniles northern jarrah forest, except possibly that at the Rosella
in this study shortly before their permanent departure from Road site, which consists of a lone male (Hayward et al.
the pouch ensured that it was not an early juvenile death that 2003). On-going predator control appears crucial because,
we missed by fitting collars to juveniles after the time they even with reduced predation intensity, introduced carnivores
are most at risk (i.e. initial independence). still cause mortality in already small quokka populations.
It seems unlikely that nutrition is the cause of the low Additionally, the extent and causes of mortality of juveniles
recruitment rate, given that populations are considered to be and pouch young are unknown but are likely to explain why
well below carrying capacity (Hayward et al. 2003) and the expected population increase has not been observed
nutrition is not considered a limiting factor at these sites (Hayward et al. 2003). Where roads that carry high traffic
(Hayward 2005). Similarly, genetic analysis of quokkas volume cut swamps containing extant quokka populations, it
trapped at each of the five sites revealed that they are not would be beneficial to construct traffic-calming devices to
inbred or lacking genetic diversity (Alacs 2001). slow vehicle speeds, and possibly also to construct wildlife
The behaviour of female quokkas within traps offers a underpasses to facilitate movement between isolated patches
possible explanation, however. When trapped animals but bearing in mind that these can become predator traps
became aware of the researcher approaching they generally (Foster and Humphrey 1995; Clevenger and Waltho 2000).
attempted to escape, during which larger young were often On the mainland the quokka cannot be managed as one
expelled or released from the pouch of females. The pouch biological entity and it is fundamental that conservation
young would flounder on the ground and ‘hiss’ loudly. managers utilise its metapopulation structure (Hayward et al.
Whether the young were actually physically expelled by the 2003, 2004, 2005) during management activities. Local
mother or became dislodged during these violent, evasive extinction probability has a greater influence on meta-
movements is unknown but considering the muscular control population persistence time than the rate of colonisation
that female quokkas have over the pouch opening (high- (Etienne and Heesterbeek 2001). Given this, and the minimal
lighted when the muscle becomes relaxed during sedation) it likelihood of increasing the colonisation rate between local
seems likely that this is a behavioural response rather than quokka populations because of the large distances separating
accidental. Pouch young expelled in traps were taped back them (Hayward et al. 2003), we advocate reinstating
into the pouch and the 100% survivorship of radio- metapopulation function by continuing control of introduced
monitored individuals suggests that trapping was unlikely to predators. We also recommend that management be directed
have inhibited the population recovery. at increasing the possibility of colonisation from the existing
If the pouch release is found to be a physical action by the isolated populations to adjacent, threat-free habitat.
mother then it can be viewed as a useful predator-avoidance Available habitat should be maximised by means of fine-
strategy. When a female quokka becomes aware of a life- scale, delicate control burns at existing sites and slightly
threatening interaction with a predator she may expel her off- larger burns at sites surrounding extant populations
spring whose subsequent noise attracts the attention of the
(Hayward et al. 2005). In the absence of these management
approaching predator. It seems likely that the predator would
actions, we predict further localised, and eventually regional,
take the young in such circumstances. Comparable
extinction of the quokka.
behaviour has been observed in other marsupials: grey
kangaroos expel their pouch young when chased by foxes
Acknowledgments
(Banks 1997), as do swamp wallabies (Wallabia bicolor)
when chased by dingoes (Robertshaw and Harden 1986). This project was funded by the Western Australian
The high proportion of young swamp wallabies in the diet of Department of Conservation and Land Management and
the dingo were thought to have been consumed after being additional funds for fox baiting were provided by Alcoa
ejected by the mother (Robertshaw and Harden 1985, 1986). World Alumina Australia Ltd. Approval for the project was
provided by CALMScience Ethics Committee Approval
Management implications Number CAEC 1/97 and trapping permits SF002928 and
Although almost 8% of quokkas at Kesners were killed by subsequent renewals. MWH thanks the Museum of Western
introduced predators, overall the threat to the quokka from Australia for providing access to their collection records, as
predation in the northern jarrah forest appears to have been well as numerous friends, family and volunteers who assisted
alleviated. Adult survivorship following fox control is at in the field, notably Phil, Joan, Miffy and Mary Hayward,
levels that would allow population increases assuming that Doug Kimber, Anna Sherriff, Beth MacArthur, Kathy
pouch young are successfully weaned. That this noticeable Himbeck, Marika Maxwell, Nick Valentine, Wazza Saunders
population increase has not been apparent to date highlights and Darren Stephen. This paper has been improved by the
the likelihood that mortality of pouch young is limiting this reviews of Joan Hayward, Jenny Taylor, Elizabeth Jefferys,
increase (Hayward et al. 2003). Conservation efforts should Barry Fox, Ian Abbott, Nicky Marlow, Gina Dawson and two
continue to focus on all of the extant populations in the anonymous reviewers.
Mortality and survivorship of the quokka Wildlife Research 721

References Hayward, M. W. (2005). Diet of the quokka (Setonix brachyurus)

(Macropodidae: Marsupialia) in the northern jarrah forest of
AAP (1998). Youth charged for ‘quokka soccer’. The Newcastle
Western Australia. Wildlife Research 32, 15–22. doi:10.1071/
Herald. p. 3, 08/12/1998.
WR03051
Alacs, E. (2001). Conservation genetics of the quokka, Setonix
Hayward, M. W., de Tores, P. J., Dillon, M. J., and Fox, B. J. (2003).
brachyurus. Ph.D. Thesis, Murdoch University.
Local population structure of a naturally-occurring metapopulation
Archer, M., and Baynes, A. (1972). Prehistoric mammal faunas from
of the quokka (Setonix brachyurus Macropodidae:Marsupialia).
two small caves in the extreme south-west of Western Australia.
Biological Conservation 110, 343–355. doi:10.1016/S0006-
Journal of the Royal Society of Western Australia 55, 80–90.
3207(02)00240-9
Augee, M. L., Smith, B., and Rose, S. (1996). Survival of wild and
Hayward, M. W., de Tores, P. J., Augee, M. L., Fox, B. J., and
hand-reared ringtail possums (Pseudocheirus peregrinus) in bush-
Banks, P. B. (2004). Home range and movements of the quokka
land near Sydney. Wildlife Research 23, 99–108.
Setonix brachyurus (Macropodidae: Marsupialia), and its impact on
Banks, P. B. (1997). Predator–prey interactions between foxes, rabbits the viability of the metapopulation on the Australian mainland.
and native mammals of the Australian Alps. Ph.D. Thesis, Journal of Zoology 263, 219–228. doi:10.1017/S0952836904
University of Sydney. 005060
Burnet, S. M. (1970). ‘Changing Patterns: an Atypical Autobiography.’ Hayward, M. W., de Tores, P. J., and Banks, P. B. (2005). Habitat use of the
(Sun Books: Melbourne.) quokka Setonix brachyurus (Macropodidae: Marsupialia) in the north-
Clevenger, A. P., and Waltho, N. (2000). Factors influencing the effec- ern jarrah forest of Australia. Journal of Mammalogy 86, 683–688.
tiveness of wildlife underpasses in Banff National Park, Alberta, doi:10.1644/1545-1542(2005)086[0683:HUOTQS]2.0.CO;2
Canada. Conservation Biology 14, 47–56. doi:10.1046/j.1523- Iveson, J. B., and Hart, R. P. (1983). Salmonella on Rottnest Island:
1739.2000.00099-085.x implications for public health and wildlife management. Journal of
Collins, L. R. (1973). ‘Monotremes and Marsupials – a Reference for the Royal Society of Western Australia 66, 15–20.
Zoological Institutions.’ (Smithsonian Institution Press: Kendall, W. L., and Pollock, K. H. (1992). The robust design in
Washington, DC.) capture–recapture studies: a review and evaluation by Monte Carlo
Cook, D. L. (1960). Some mammal remains found in caves near simulation. In ‘Wildlife 2001: Populations’. (Eds C. McCulloch and
Margaret River. Western Australian Naturalist 7, 107–108. R. H. Barret.) pp. 31–72. (Noyes Publishing: New Jersey.)
de Tores, P. J. (1994). Fox control manual – operational guidelines for Main, A. R. (1959). Rottnest Island: the Rottnest Biological Station and
control of the red fox, Vulpes vulpes, through the use of sodium recent scientific research. Rottnest Island as a location for biologi-
monofluoroacetate or ‘1080’ on CALM managed estate and in other cal studies. Journal of the Royal Society of Western Australia 42,
CALM programs. Western Australian Department of Conservation 66–67.
and Land Management, Perth. Main, A. R., and Yadav, M. (1971). Conservation of macropods in
de Tores, P. J. (1999). Control and ecology of the red fox in Western reserves in Western Australia. Biological Conservation 3, 123–133.
Australia – prey response to 1080 baiting over large areas. doi:10.1016/0006-3207(71)90014-0
CALMScience, Wildlife Research Centre, Department of Maxwell, S., Burbidge, A. A., and Morris, K. D. (1996). ‘The Action
Conservation and Land Management, Wanneroo, Western Australia. Plan for Australian Monotremes and Marsupials.’ (Wildlife
Etienne, R. S., and Heesterbeek, J. A. P. (2001). Rules of thumb for con- Australia: Canberra.)
servation of metapopulations based on a stochastic winking-patch Owen-Smith, N. (1993). Comparative mortality rates of male and
model. American Naturalist 158, 389–407. doi:10.1086/321986 female kudus: the costs of sexual size dimorphism. Journal of
Fleming, P. J. S., and Parker, R. W. (1991). Temporal decline of 1080 Animal Ecology 62, 428–440.
within meat baits used for control of wild dogs in New South Wales. Pollock, K. H., Winterstein, S. R., Bunck, C. M., and Curtis, P. D.
Wildlife Research 18, 729–740. doi:10.1071/WR9910729 (1989a). Survival analysis in telemetry studies: the staggered entry
Foster, M. L., and Humphrey, S. R. (1995). Use of highway underpasses design. Journal of Wildlife Management 53, 7–15.
by Florida panthers and other wildlife. Wildlife Society Bulletin 23, Pollock, K. H., Winterstein, S. R., and Conroy, M. J. (1989b).
95–100. Estimation and analysis of survival distributions for radio-tagged
Gardner, C. A. (1957). The fire factor in relation to the vegetation of animals. Biometrics 45, 99–109.
Western Australia. Western Australian Naturalist 5, 166–173. Robertshaw, J. D., and Harden, R. H. (1985). The ecology of the dingo
Gibb, D. G. A., Kakulas, B. A., Perrett, D. H., and Jenkyn, D. J. (1966). in north-eastern New South Wales. II. Diet. Australian Wildlife
Toxoplasmosis in the Rottnest quokka (Setonix brachyurus). Research 12, 39–50. doi:10.1071/WR9850039
Australian Journal of Experimental Biology and Medical Science Robertshaw, J. D., and Harden, R. H. (1986). The ecology of the dingo
44, 665–672. in north-eastern New South Wales. IV. Prey selection by dingoes,
Gould, J. (1863). ‘The Mammals of Australia.’ (Privately published: and its effect on the major prey species, the swamp wallaby,
London.) Wallabia bicolor (Desmarest). Australian Wildlife Research 13,
Harrison, S. (1991). Local extinction in a metapopulation context: an 141–163. doi:10.1071/WR9860141
empirical evaluation. Biological Journal of the Linnean Society 42, Saunders, G., McLeod, S., and Kay, B. (2000). Degradation of sodium
73–88. monofluoroacetate (1080) in buried fox baits. Wildlife Research 27,
Hart, R. P., Bradshaw, S. D., and Iveson, J. B. (1986). Salmonella infec- 129–135. doi:10.1071/WR99031
tions and animal conditions in the mainland and Bald Island popu- Sharman, G. B. (1955a). Studies on marsupial reproduction. II. The
lations of the quokka (Setonix brachyurus: Marsupialia). Journal of oestrous cycle of Setonix brachyurus. Australian Journal of Zoology
the Royal Society of Western Australia 69, 7–11. 3, 44–55. doi:10.1071/ZO9550044
Hayward, M. W. (2002). The ecology of the quokka (Setonix brachyu- Sharman, G. B. (1955b). Studies on marsupial reproduction. III.
rus) (Macropodidae: Marsupialia) in the northern jarrah forest of Normal and delayed pregnancy in Setonix brachyurus. Australian
Australia. Ph.D. Thesis, Biological, Earth and Environmental Journal of Zoology 3, 56–70. doi:10.1071/ZO9550056
Science, University of New South Wales, Sydney. Shield, J. W. (1964). A breeding season difference in two populations of
http://www.library.unsw.edu.au/~thesis/adt-NUN/public/adt- the Australian macropod marsupial (Setonix brachyurus). Journal of
NUN20030325.154836/index.html. Mammalogy 45, 616–625.
722 Wildlife Research M. W. Hayward et al.

Short, J., Kinnear, J. E., and Robley, A. (2002). Surplus killing by intro- Thackway, R. M., and Cresswell, I. D. (1995). ‘An Interim
duced predators in Australia – evidence for ineffective anti-predator Biogeographic Regionalisation for Australia: a Framework for
adaptations in native prey species? Biological Conservation 103, Setting Priorities in the National Reserves System Cooperative
283–301. doi:10.1016/S0006-3207(01)00139-2 Program.’ (Australian Nature Conservation Agency: Canberra.)
Soderquist, T. R., and Serena, M. (1994). Dietary niche of the western Twigg, L. E., Eldridge, S. R., Edwards, G. P., Shakeshaft, B. J.,
quoll, Dasyurus geoffroii, in the jarrah forest of Western Australia. dePreu, N. D., and Adams, N. (2000). The longevity and efficacy of
Australian Mammalogy 17, 133–136. 1080 meat baits used for dingo control in central Australia. Wildlife
Squires, N. (2003). Rare marsupials kicked to death in ‘quokka soccer’. Research 27, 473–481.
The Telegraph. p. 3, 15/01/2003. White, S. R. (1952). The occurrence of the quokka in the south-west.
Stewart, D. W. R. (1936). Notes on marsupial damage in pine planta- Western Australian Naturalist 3, 101–103.
tions. Australian Journal of Forestry 1, 41–44. Whittell, H. M. M. (1954). John Gilbert’s notebook on marsupials.
Storr, G. M. (1964). Studies on marsupial nutrition – IV. Diet of the Western Australian Naturalist 4, 105–114.
quokka, Setonix brachyurus (Quoy & Gaimard), on Rottnest Island, Zar, J. H. (1996). ‘Biostatistical Analysis.’ (Prentice-Hall Inc.: Upper
Western Australia. Australian Journal of Biological Sciences 17, Saddle River, NJ.)
469–481.
Storr, G. M. (1965a). Notes of Bald Island and the adjacent mainland.
Western Australian Naturalist 9, 187–196.
Storr, G. M. (1965b). The avifauna of Rottnest Island, Western
Australia. I. Marine birds. Emu 64, 48–60. Manuscript received 8 November 2004, accepted 12 October 2005

http://www.publish.csiro.au/journals/wr