Modest increases in densities of burrow-nesting petrels following the removal of cats (Felis catus) from Marion Island

Abstract

Introduced predators are one of the main threats facing seabirds breeding on oceanic islands. Cats (Felis catus) were introduced to subantarctic Marion Island (290 km2) in 1949, and by the 1970s some 2000 cats were killing about 450,000 seabirds per year, greatly reducing burrowing petrel populations. Cats were eradicated by 1991, but house mice (Mus musculus) remain. The densities of utilised petrel burrows were estimated in 2013 by systematically searching for their burrows in 741 10 × 10 m sample quadrats in the north-eastern sector of Marion Island, repeating the sampling design and methods used by Schramm in 1979. The mean burrow densities and 95 % CIs were compared between surveys by species for the different habitat and vegetation types, with non-overlapping CIs considered indicative of an increase in burrow density. With cats eradicated and the potential for immigration from nearby Prince Edward Island (free of introduced mammals), we could expect a multi-fold increase in petrel numbers over the last two decades; however, burrow densities at Marion have increased by only 56 % since 1979. White-chinned petrels (Procellaria aequinoctialis) showed the greatest increase, despite being listed as vulnerable due to incidental mortality on fishing gear at sea. The recovery of other summer-breeding species decreased with decreasing body size, and winter-breeding species showed even smaller recoveries, similar to patterns of breeding success at Gough Island, where mice are major predators of petrel chicks and eggs. Predation by mice is the most likely explanation for the limited recovery of Marion’s petrel populations.

Full text

1 23
Polar Biology
ISSN 0722-4060
Polar Biol
DOI 10.1007/s00300-016-1985-z
Modest increases in densities of burrow-
nesting petrels following the removal of cats
(Felis catus) from Marion Island
Ben J.Dilley, Michael Schramm & Peter
G.Ryan

1 23
Your article is protected by copyright and
all rights are held exclusively by Springer-
Verlag Berlin Heidelberg. This e-offprint is
for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

ORIGINAL PAPER
Modest increases in densities of burrow-nesting petrels following
the removal of cats (Felis catus) from Marion Island
Ben J. Dilley
1
•Michael Schramm
1,2
•Peter G. Ryan
1
Received: 7 June 2015 / Revised: 24 May 2016 / Accepted: 25 May 2016
ÓSpringer-Verlag Berlin Heidelberg 2016
Abstract Introduced predators are one of the main threats
facing seabirds breeding on oceanic islands. Cats (Felis
catus) were introduced to subantarctic Marion Island
(290 km
2
) in 1949, and by the 1970s some 2000 cats were
killing about 450,000 seabirds per year, greatly reducing
burrowing petrel populations. Cats were eradicated by
1991, but house mice (Mus musculus) remain. The densi-
ties of utilised petrel burrows were estimated in 2013 by
systematically searching for their burrows in 741
10 910 m sample quadrats in the north-eastern sector of
Marion Island, repeating the sampling design and methods
used by Schramm in 1979. The mean burrow densities and
95 % CIs were compared between surveys by species for
the different habitat and vegetation types, with non-over-
lapping CIs considered indicative of an increase in burrow
density. With cats eradicated and the potential for immi-
gration from nearby Prince Edward Island (free of intro-
duced mammals), we could expect a multi-fold increase in
petrel numbers over the last two decades; however, burrow
densities at Marion have increased by only 56 % since
1979. White-chinned petrels (Procellaria aequinoctialis)
showed the greatest increase, despite being listed as vul-
nerable due to incidental mortality on fishing gear at sea.
The recovery of other summer-breeding species decreased
with decreasing body size, and winter-breeding species
showed even smaller recoveries, similar to patterns of
breeding success at Gough Island, where mice are major
predators of petrel chicks and eggs. Predation by mice is
the most likely explanation for the limited recovery of
Marion’s petrel populations.
Keywords Seabird conservation Predation Island
restoration (Mus musculus)(Felis catus)
Introduction
Many seabird species are threatened with extinction, and
one of the major threats, particularly for oceanic species, is
the introduction of mammalian predators onto their
breeding islands (Croxall et al. 2012). The subantarctic
Prince Edward Islands (46°540S, 37°450E) in the south-west
Indian Ocean provide a sobering example of the conse-
quences of such introductions, which extend beyond the
impacts on seabird populations to affect the structure and
functioning of the islands’ terrestrial ecosystems (Chown
and Smith 1993; Smith et al. 2002). The Prince Edward
Island group comprises two islands: Marion (290 km
2
) and
Prince Edward Island (44 km
2
). They support 28 species of
breeding seabirds (Ryan and Bester 2008), but Marion
Island has a much larger complement of introduced spe-
cies, following the establishment of a weather station on
the island in 1948 (Chown and Froneman 2008). House
mice (Mus musculus) were introduced accidentally to
Marion Island, most likely by sealers or shipwrecks in the
early nineteenth century (Watkins and Cooper 1986).
Domestic cats (Felis catus) were taken to the island’s
weather station in 1949 to control mice (van Aarde 1977),
Electronic supplementary material The online version of this
article (doi:10.1007/s00300-016-1985-z) contains supplementary
material, which is available to authorized users.
&Ben J. Dilley
dilleyben@gmail.com
1
Percy FitzPatrick Institute of African Ornithology, DST/NRF
Centre of Excellence, University of Cape Town,
Rondebosch 7701, South Africa
2
NISC (Pty) Ltd, PO Box 377, Grahamstown 6140, South
Africa
123
Polar Biol
DOI 10.1007/s00300-016-1985-z
Author's personal copy

but they soon turned feral and started killing the island’s
seabirds (Rand 1954). By the mid-1970s, an estimated
2000 cats were killing some 450,000 birds per year, most
of which were burrow-nesting petrels (van Aarde 1980).
Population densities were reduced more than 20-fold
(Schramm 1986), and some petrels were apparently extir-
pated (van Aarde 1980; Ryan and Bester 2008). By com-
parison, Prince Edward Island has not had any introduced
mammals (Gleeson and van Rensburg 1982).
The population sizes of burrow-nesting petrels (Procel-
lariidae, Pelecanoididae and Hydrobatidae) at the Prince
Edward Islands are poorly known relative to surface-nest-
ing species (Diomedeidae and Spheniscidae; Crawford
et al. 2009; Ryan et al. 2009a). At least nine species of
burrowing petrels breed on Marion Island, but two small
species probably were extirpated [black-bellied storm pet-
rel (Fregetta tropica) and common diving petrel (Pele-
canoides urinatrix)], and populations of other species were
greatly depressed by cat predation (Ryan and Bester 2008).
A multifaceted cat eradication programme was started in
the late 1970s, and by 1991 the last cat had been killed
(Bester et al. 2002), allowing burrow-nesting petrel num-
bers to recover. There have only been two detailed esti-
mates of burrow-nesting petrel population sizes at the
Prince Edward Islands based on burrow densities
(Schramm 1986; Ryan et al. 2012). There has been little
effort to assess how petrel populations at Marion Island
responded following the eradication of cats. Initial indi-
cations were positive; following the removal of cats, there
were marked increases in the breeding success of bur-
rowing petrels, especially the great-winged petrel (Ptero-
droma macroptera), which breeds in winter when cat
predation pressure was most severe (Cooper and Fourie
1991; Cooper et al. 1995). However, recent evidence from
subantarctic skuas (Catharacta antarctica lo¨nnbergi),
which are major predators of burrowing petrels, was less
encouraging (Cerfonteyn and Ryan 2016). There is a sig-
nificant relationship between food availability and repro-
ductive success in skuas (Phillips et al. 1996), but numbers
of skuas have decreased steadily at Marion Island since
cats were eradicated, whereas their numbers at Prince
Edward Island have remained stable (Ryan et al. 2009b).
At other islands, petrel populations have recovered
rapidly following the eradication of terrestrial predators. For
example, numbers of spectacled petrel (Procellaria con-
spicillata), a species closely related to white-chinned petrels,
have increased at roughly 7 % per year following the dis-
appearance of pigs from Inaccessible Island (Ryan et al.
2006; Ryan and Ronconi 2011). If similar recoveries have
occurred at Marion Island since cats were eradicated, we
expect burrowing petrel populations to have increased
threefold to fivefold by 2013 (based on 5–7 % per annum
growth), and potentially even more given immigration from
nearby Prince Edward Island (22 km to the NE), creating
easily detectable signals of population recovery (Ryan et al.
2006). Such increases are not unrealistic, because densities
of burrowing petrel nests on Marion Island at the height of
the cat era were approximately 25 times lower (200 ha
-1
)
than those on Prince Edward Island (5000 ha
-1
; Schramm
1986). Schramm (1986) estimated the density of burrowing
petrel nests in the north-east sector of Marion Island in 1979.
We repeated Schramm’s survey in 2013, more than 20 years
after cats were removed from the island, to assess the extent
to which petrel populations at Marion Island have recovered
since cats were eradicated.
Study area and methods
Repeat survey
To estimate petrel breeding densities, BD systematically
searched for their burrows in 741 10 910 m sample
quadrats in the north-eastern sector of Marion Island,
repeating the sampling design and methods used by
Schramm (1986) over the austral summer of 1979–1980.
Burrows were sampled at 13 sites from altitudes of
12–373 m, covering an area of approximately 1040 ha
(Fig. 1) with five habitat types: steep vegetated slopes,
coastal lowland, vegetated lava hummocks, partly vege-
tated lava hummocks and cinder slopes. At 11 of these
sites, quadrats were arranged in blocks of three 10 910 m
(n=215) and laid out at 25-m intervals perpendicular to
transect lines 200 m apart. At the remaining two sites,
random transect lines (1200 m) were chosen and
10 910 m quadrats (n=96) sampled at 50-m intervals
for more extensive coverage in the vegetated and partly
vegetated lava hummocks. The location of Schramm’s
(1986) sample quadrats was only crudely mapped in 1979.
To accurately repeat the 1979 sampling protocol, MS
returned to Marion Island in April–May 2012 to identify
the location of the original transects (Fig. 1). Sites were
matched to the same habitats sampled in 1979 and marked
with poles for future monitoring.
The actual surveys to estimate burrow densities for all
species were repeated in summer (29 January 2013–22
March 2013), when most burrowing petrels breed. At the
time of the surveys, Salvin’s prion (Pachyptila salvini) and
blue (Halobaena caerulea), soft-plumaged (Pterodroma
mollis), white-chinned (Procellaria aequinoctialis), South
Georgian diving (Pelecanoides georgicus) and common
diving petrels were still breeding, and Kerguelen petrel
(Aphrodroma brevirostris) chicks had recently fledged
(Fig. 2). During the repeat survey, the two winter-breeding
species [great-winged and grey petrels (Procellaria
cinerea)] were not active as the timing of this repeat survey
Polar Biol
123
Author's personal copy

fell outside their breeding cycle. Therefore, we intensively
surveyed a *300 ha area around base for utilised nests of
these two species during the winter of 2009 and 2012. A
proportion of these burrows were monitored to record the
birds breeding phenology for additional research projects.
These studies provided a reasonable baseline for identify-
ing recently active burrows of these two winter-breeding
species during the actual survey.
Fig. 1 Study area in the north-east corner of Marion Island, showing the locations and arrangement of the 741 (10 910 m) sample plots. The
inset shows the location of the Prince Edward Islands, with Prince Edward Island 22 km to the north-east of Marion
Fig. 2 Breeding months of burrowing petrels at Marion Island.
Vertical bars indicate the hatching periods, and the central vertical
bars give the average hatching dates. Arrows indicate the timing of
the 2013 repeat survey of 741 (10 910 m) quadrats. Data sources:
1
FitzPatrick Institute unpubl. data, Marion Island 2009–2014;
2
Schramm (1983);
3
assumed breeding period for diving petrels
Pelecanoides spp., Payne and Prince (1979).
Polar Biol
123
Author's personal copy

For each 10 910 m sample quadrat, we recorded slope
aspect (using a compass), soil depth (using a graduated 1 m
metal rod), slope angle (using a clinometer), altitude (using
a Garmin GPSMAP 62 s), the dominant vegetation type
(based on Huntley 1971; Smith 1976) and the habitat type
based on the five habitat types described in detail by
Schramm (1986) based on the classification by Verwoerd
(1971). In both surveys, these parameters were recorded
once for each quadrat and assumed to be representative for
all burrows within the quadrat [following methods descri-
bed in Schramm (1986)].
Burrow identification
In both surveys, all the active burrows (with a bird present)
and recently active burrows were identified and counted.
Recently active burrows were identified by signs of fresh
excavations, freshly cropped vegetation (used as nest lin-
ing), fragments of new egg shell (indicative of a failed
nest), feathers, droppings, fresh footprints or disturbed
moat water (clear, settled moat water signifies birds have
not recently been through the entrance). Burrows which
were overgrown or collapsed were not counted. During his
return to Marion Island in April–May 2012, MS demon-
strated to BD how he used the relative shape and size of the
burrow entrance (Fig. 3) and the physical burrow charac-
teristics (Online Resource 1) to infer the species occupying
a burrow. Further confirmation by probing the burrow with
a stick to elicit a response Ryan et al. (2012) or using
playback recordings or vocal impersonations was a reliable
method for identifying white-chinned petrels (Berrow
2000; Ryan et al. 2012), blue petrels (Crawford 1952;
Fugler et al. 1987), grey petrels (Barbraud et al. 2009; pers.
obs. 2009, 2012, 2014) and Kerguelen petrels (pers. obs.
2012), whereas great-winged and soft-plumaged petrels
tended to be less responsive. However, even among highly
responsive species, some individuals are less likely to call
back than others.
Data on vocalisations of petrels and on physical char-
acteristics of their burrows (Online Resource 1) were col-
lected prior to the survey at study burrows in study colonies
of grey, great-winged, white-chinned and blue petrels
which BD conducted at Marion in 2009–2010 and in
2012–2013. We did not collect further data on the physical
characteristics during this repeat survey. For Kerguelen and
soft-plumaged petrels, we used data on burrow dimensions
from Schramm (1983). Data on burrow dimensions of grey
petrels were from a study colony on Gough Island con-
ducted by BD in 2014, since grey petrels predominantly
nest in caves on Marion, but also use burrows. In addition
to these methods, we used a custom-made burrow scope
with a high-resolution conical pinhole camera, LED torch
and an 18 921 cm colour monitor, which provided a clear
image of the inside of the burrow. Although the burrow
scope allowed for low-impact inspection of burrow con-
tents, some burrows were too complex or deep to see the
nest chamber from the burrow entrance. Burrows with two
entrances that connected to one passage were counted as
one, and burrows on the edge of a quadrat were included if
the burrow entrance was completely within the square.
Using a combination of these burrow identification meth-
ods, BD allocated a species to each utilised burrow in each
quadrat. The resultant burrow densities in both surveys are
a measure of how many utilised burrows were found, but
since these were once-off burrow checks, the results do not
accurately quantify how many burrows contained breeding
pairs or non-breeders, nor account for inter-annual vari-
ability in burrow occupancy.
Data analysis
Schramm (1986) reported burrow densities (no ha
-1
)of
eight petrel species at 13 sites (total of 741 quadrats) as
mean ±one standard deviation (SD) with the associated
number of sample quadrats (0.01 ha) per habitat type and
sample site (Online Resource 2). From these data, the
standard errors (SE = SD/
ffiffiffi
n
p) and 95 % confidence inter-
vals (CI =mean ±2 SE) were calculated from the mean
burrow densities of each species for the five habitat types
and seven vegetation types. Where 2013 data produced a
negative CI, these data were bootstrapped (analysis was run
using library boot (Canty and Ripley 2012) in R (R Core
Team 2014) with 5000 iterations). These mean burrow
densities and 95 % CIs were compared between surveys by
species for the different habitat and vegetation types, with
Fig. 3 Average burrow entrance dimensions of Salvin’s prion
(n=17)
1
, blue petrel (n=30)
1
, soft-plumaged petrel (n=16)
2
,
Kerguelen petrel (n=15)
2
, great-winged petrel (n=50)
1
, white-
chinned petrel (n=50)
1
and grey petrel (n=51)
3
.Error bars
represent ±1 SD Data sources:
1
This study;
2
Schramm (1983);
3
BJD
unpubl. data, Gough Island.
Polar Biol
123
Author's personal copy

non-overlapping CIs considered indicative of an increase in
burrow density. Means are presented ±SD unless other-
wise indicated.
Mean body mass of each species (data from: Payne and
Prince 1979; Schramm 1983; Berruti and Hunter 1986;
Fugler et al. 1987; FitzPatrick Institute unpubl. data) was
log (ln)-transformed to interpret the relationship between
body mass and the apparent increase in burrow density
between surveys, estimated as the density in 2013/density
in 1979. Following methods described in Schramm (1986),
the 2013 burrow density data were extrapolated to the
larger study area (1041 ha) by habitat type to estimate the
number of burrows for each species. The percentage
increase in the number of burrows between surveys was
calculated as the difference in number of burrows between
surveys/number of burrows in 1979.
Inter-specific differences in mean soil depth (mm) and
slope angle (degrees) were tested using Kruskal–Wallis and
post hoc Tukey tests. We calculated the average direction
burrow entrances faced for each species using the package
‘circular’ (Agostinelli and Lund 2013) in R 3.1 (R Core
Development Team 2014). We calculated 95 % CIs from
aspect data [bootstrapped with 1000 iterations, using
library boot (Canty and Ripley 2012) in R (R Core Team
2014)]. Species with non-overlapping 95 % CIs were
considered to be significantly different, and statistical tests
were two-tailed with p\0.05 as the cut-off for
significance.
Results
Of the six petrel species recorded in both studies, only the
summer-breeding white-chinned petrels and Salvin’s pri-
ons showed marked increases in densities from 1979 to
2013. Recovery of other summer-breeding species
decreased with decreasing body size (Fig. 4), and winter-
breeding species showed even smaller recoveries, with
numbers of grey petrels apparently having decreased. Blue
petrels, soft-plumaged petrels, Kerguelen petrels and great-
winged petrels showed marginal increases in densities.
Although overall burrow numbers in the study area
(1041 ha) increased by 56 % between the two surveys, if
we exclude white-chinned petrels, the increase is 51 %.
Changes in burrow densities and distribution
White-chinned petrels
White-chinned petrels showed the greatest increase in the
number of burrows (3.3 times the number of burrows in
1979, Table 1, Online Resource 3). MS perceived a visible
increase in the number of burrows in 2012 since he was last
on the island in 1980. Burrow densities increased most in
the steep vegetated slopes (14.6–56.1 burrows ha
-1
) and
coastal lowland habitats (24.6–85.6 burrows ha
-1
, Fig. 5
and Online Resource 2), with the highest mean density of
113 burrows ha
-1
recorded in the Van den Boogaard
sample quadrats. Previously (Poa cookii) tussock supported
the highest densities of white-chinned petrel burrows (30
burrows ha
-1
), but in 2013 both open fernbrake (61 bur-
rows ha
-1
) and (Acaena magellanica) herbfield (50 bur-
rows ha
-1
) supported higher densities (Online Resource 1).
Salvin’s prions
Salvin’s prions also showed an overall increase in burrow
densities (1.4 times the number of burrows in 1979,
Table 1and Online Resource 2), with the most notable in-
creases in the steep vegetated slopes (64–100 bur-
rows ha
-1
) and the partly vegetated lava hummocks in
Hoppie’s Hell (279–393 burrows ha
-1
). The latter area has
become more extensively vegetated since 1979 and now
represents more of a ‘vegetated lava hummocks’ habitat
with deeper soils and more vegetation cover for burrows
(Niek Gremmen and Valdon Smith, pers. comm. 2013).
Salvin’s prions showed a strong preference across all study
sites for burrowing in Acaena (increase from 100 to 211
burrows ha
-1
, Table 2). There was also an increase in
Salvin’s prion burrows on the cinder slopes (Table 1),
particularly the south-facing and east-facing (coastal)
slopes of Hendrik Vister Kop, where prions utilised lower
slopes dominated by Acaena and (Azorella selago)
(Table 2).
Fig. 4 Relationship between body mass (g) and the increase in
number of burrows in north-east Marion Island from 1979 to 2013.
Data from Kerguelen and soft-plumaged petrels are pooled due to the
low number of burrows recorded and their similar burrow size.
Winter breeders represented with open circles
Polar Biol
123
Author's personal copy

Kerguelen petrels
Kerguelen petrels showed little change in their burrow
densities. Breeding in early summer, these small- to
medium-sized gadfly petrels prefer steep vegetated slopes
and, to a lesser extent, coastal lowlands, usually choosing a
steep slope (mean 35°) for easy take off. In both surveys,
the highest densities were recorded on the steep, upper
ridges of Blue Petrel Bay. Although their chicks had
fledged at time of the survey, the distinctive nature of the
burrow, especially the neat and extended moat (Online
Resource 1), confirmed their presence along this coastal
slope. The transect counts revealed little change in the
densities for Kerguelen petrels, but there is a general per-
ception that Kerguelen petrels have become less common
over the last few decades. For example, MS monitored 49
study burrows along the slopes above Gentoo Lake in
1979, but only two active burrows were found in this area
in 2012. As a further example of their scarcity, during 26
nights spotlighting near the Base in October–November
2012 (peak incubation period for this species), only two
Kerguelen petrels were sighted in flight.
Soft-plumaged petrels
Although the results from this repeat survey show the
number of soft-plumaged petrel burrows has increased 3.2
times since 1979 (Table 1), these data are from a small
base. This large factor increase is being driven by burrows
found in areas, which have become more vegetated since
the 1979 survey, for example the vegetated lava hummocks
on the side of Junior’s Kop and the partly vegetated lava
hummocks of Hoppie’s Hell. Moderate increases in the
number of burrows in the steep vegetated slopes along the
coastal ridges at Albatross Lakes, Blue Petrel Bay and
Skua Ridge, and in the steep slopes below the east-facing
cliffs of Piew Crags all contributed to the overall increase
since 1979 (Online Resource 2). The steep coastal vege-
tated slopes of Macaroni Bay used to support numerous
soft-plumaged petrel burrows, but these slopes are now
dominated by dense patches of invasive (Agrostis castel-
lana) and (A. stolonifera) (Gremmen 1997) forming an
impenetrable mat of roots and grass unsuitable for bur-
rowing birds. In contrast to the Kerguelen petrels, soft-
plumaged petrels are regularly seen and heard at night on
Marion, which does lend support to this apparent increase
in burrow numbers. However, the overall density of soft-
plumaged petrels is still relatively low.
Great-winged petrels
Great-winged petrels had marginal increases in burrow
densities (1.8 times the number of burrows in 1979,
Table 1 Estimated numbers of petrel burrows (for eight species) within the six habitat types of the 1041 ha study area at Marion Island
Habitat type (area) Salvin’s prion Blue petrel Great-winged
petrel
Kerguelen
petrel
Soft-plumaged
petrel
White-chinned
petrel
Other petrels All petrels
Year 1979 2013 1979 2013 1979 2013 1979 2013 1979 2013 1979 2013 1979 2013 1979 2013
Steep vegetated slopes (71 ha) 4587 7157 5488 8797 1243 1328 518 838 1271 1697 1037 3983 0 56
b
14,144 23,899
Fjaeldmark and mire plateaux (147 ha) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Coastal lowland (38 ha) 908 1733 0 0 84 137 137 110 0 0 935 3196 0 0 2064 5176
Vegetated lava hummocks (397 ha) 17,826 28,345 596 1630 14,967 28,008 0 0 0 1630 1747 4927 1151
a
530
a
36,287 63,710
Partly vegetated lava hummocks (367 ha) 102,466 144,488 0 0 0 0 0 0 0 771 0 0 770
a
3046
b
103,236 148,305
Cinder slopes (21 ha) 351 920 0 53 80 200 0 0 0 0 8 17 44
b
16
b
483 1208
Total (1041 ha) 126,138 182,643 6084 10479 16,374 29,672 655 948 1271 4097 3727 12,123 1965 3650 156,214 243,613
% composition 80.7 74.8 3.9 4.3 10.5 12.2 0.4 0.4 0.8 1.7 2.4 5.0 1.3 1.5 100.0 100.0
Factor increase in number of burrows 1.44 1.72 1.81 1.45 3.22 3.25 1.86 1.56
Areas calculated by Schramm (1986)
a
Grey petrel,
b
Diving petrel
Polar Biol
123
Author's personal copy

Table 1and Online Resource 2) and still favour the shel-
tered slopes of vegetated lava hummocks (e.g. Nellie
Humps and the sides of Junior’s Kop). Burrows were also
recorded in dense Acaena on the northern lower slopes of
Junior’s Kop, with an overall increase from 2 to 39 bur-
rows ha
-1
in areas dominated by Acaena (Online Resource
2). Schramm (1986) found the highest densities (45 bur-
rows ha
-1
)inPoa tussock along Skua Ridge. This site is
still favoured (103 burrows ha
-1
), but is now dominated by
Acaena and the low fern (Blechnum penna-marina). The
steep vegetated slopes around Albatross Lakes were good
burrowing petrel habitat in 1979, especially for great-
winged petrels, with deep soils suited to large burrows.
However, this habitat has since been invaded by the alien
grass (Agrostis stolonifera) (Gremmen et al. 1998), which
forms a dense, impenetrable mat where few burrows were
recorded in 2013.
Blue petrels
Blue petrels also had marginal increases in burrow densi-
ties (1.7 times the number of burrows in 1979, Table 1and
Online Resource 2); however, these results should be
viewed with some caution because blue petrels are colonial
breeders, and thus, the results are very sensitive to shifts in
colony boundaries. Transects intersected colonies at Mac-
aroni Bay, where birds still favoured the steep coastal
slopes dominated by dense Acaena patches, and at Blue
Petrel Bay, where they favoured the lower coastal slopes
dominated by Poa tussock (104–527 burrows ha
-1
, Online
Resource 2). Although this was the highest burrow density
recorded in this survey, it is low compared to blue petrel
densities at Prince Edward Island, where densities in Poa
tussock were 2600–8300 burrows ha
-1
in 1979 (Schramm
1986). The steep vegetated slopes above Ship’s Cove
continue to be heavily invaded by (A. stolonifera) (Grem-
men et al. 1998); however, some blue petrels were found in
a few isolated Acaena patches on these slopes. The upper
slopes of the cove are largely free of A. stolonifera and are
well utilised by burrowing petrels, including blue petrels,
resulting in an overall increase in burrow densities for
Ship’s Cove.
Diving petrels
Two species of diving petrels Pelecanoides spp. breed at
the Prince Edward Islands: South Georgian diving petrels
are largely confined to cinder slopes, whereas common
Fig. 5 Increases in estimated densities (±95 % CI) of white-chinned
petrel burrow densities in two habitat types in north-east Marion
Island from 1979 and 2013
Table 2 Mean burrow densities (burrows ha
-1
)±SD of six petrel species in eight vegetation types at Marion Island in 2013
Vegetation
type
Salvin’s prion Blue petrel Great-
winged
petrel
Kerguelen
petrel
Soft-
plumaged
petrel
White-
chinned petrel
Other
petrels
All petrels
Cotula
herbfield
0 800.0 ±707.1 0 16.7 ±40.8 33.3 ±51.6 183.3 ±240.1 0 1033.3 ±981.2
Poa tussock 75.7 ±222.9 527.0 ±929.1 0 8.1 ±27.7 16.2 ±44.2 45.9 ±114.5 2.7 ±16.4
b
675.7 ±906.0
Closed
fernbrake
113.1 ±204.1 4.9 ±28.3 41.0 ±88.8 8.2 ±30.4 13.1 ±56.0 20.5 ±52.9 2.5 ±15.6
b
203.3 ±234.9
Open
fernbrake
98.4 ±170.5 4.1 ±34.8 15.9 ±55.3 4.9 ±23.4 7.7 ±37.0 61.4 ±118.2 0.4 ±6.4
a
192.7 ±237.9
Acaena
herbfield
211.0 ±236.9 60.4 ±255.1 39.6 ±95.3 7.7 ±34.1 22.0 ±57.4 50.5 ±100.4 2.2 ±14.7
b
393.4 ±343.8
Agrostis
mire
18.6 ±62.7 0 0 0 0 34.9 ±113.1 0 53.5 ±162.3
Azorella
fjaeldmark
42.9 ±105.6 0 0 0 0 0 1.8 ±13.6
b
44.8 ±105.7
Unvegetated 0 0 0 0 0 0 0 0
a
Grey petrel,
b
Diving petrel
Polar Biol
123
Author's personal copy

diving petrels nest on well-vegetated coastal slopes. Both
species are scarce at Marion Island; only eight active div-
ing petrel burrows were found in 2013. On the scoria
cones, 19 burrows were found, but only two had fresh
feathers and faeces and were assumed to be South Geor-
gian diving petrel burrows. This represents a modest
decrease from 1979 (3.2–1.3 burrows ha
-1
, Table 1), but
the sample size is too small to make any firm conclusions
about trends. Four common diving petrel burrows were
found in well-established Blechnum slopes in Hoppie’s
Hell. One contained an adult common diving petrel, with
no sign of a chick or egg, and the other three burrows had
signs of recent activity. Although breeding was not con-
firmed, this is the first record of the species ashore in a
burrow since it was assumed to have been extirpated by
cats (Ryan and Bester 2008). Four skua nests in Hoppie’s
Hell had numerous diving petrel wings, further indicating
their presence in the area (Ryan et al. 2009b; although
some birds flying to the interior may be caught by skuas,
Schramm 1986). Active diving petrel burrows at Piew
Crags (n=1) and Blue Petrel Bay (n=1) could not be
identified to species. Diving petrel wings were found in
skua middens south-east of Blue Petrel Bay, but none were
found in two middens below Piew Crags.
Grey petrels
Only 15 grey petrel nests were found during the intensive
survey of a &300 ha area: one in an earth burrow and 14 in
rock caves, giving a rough density of 0.05 burrows ha
-1
for
this area. Using the transect results, burrow densities
appear to have decreased (3.1–1.0 burrows ha
-1
), albeit
from a very low base.
Changes in nest site selection
Overall, the estimated densities of burrowing petrels
increased across all five habitat types, although overlapping
95 % confidence intervals suggest that the increases are not
significant in either of the lava hummock habitats (Fig. 6).
Partly vegetated lava hummocks had the highest density of
burrows, due to the predominance of Salvin’s prions, but
the increase in burrow density (44 %) was lowest in this
habitat (Fig. 6, Online Resource 2). The numbers of bur-
rows in the vegetated lava hummocks increased 69 %
overall, attributable to the site at the base of Junior’s Kop,
which increased from 81 to 214 burrows ha
-1
, mainly due
to increases in Salvin’s prions, which are able to utilise
shallow soils and natural cavities, as well as species
requiring deeper soils such as white-chinned and great-
winged petrels (Online Resource 2). The steep vegetated
slopes had the highest burrow densities for blue, great-
winged, Kerguelen and soft-plumaged petrels and showed
a significant increase in petrel densities (69 %, Fig. 6). The
large increase in burrow densities in the coastal lowlands
(251 %) was primarily a result of the increase in white-
chinned petrels at Van den Boogaard and Trypot. The
cinder slopes also showed a large increase in overall bur-
row densities (250 %), but remained less than half the
density in any other habitat type sampled (Fig. 6).
The 2013 data confirmed that larger petrel species are
associated with sites with deeper soils for their larger
burrows (Schramm 1986; Online Resource 4). Burrow
entrances of most species faced mainly to the east, away
from the prevailing westerly winds, but also facing
downslope towards the sea (Fig. 1, Online Resource 4).
Great-winged petrels typically faced more south-east, away
from the north-westerly winter winds, differing signifi-
cantly from the other species surveyed (Online Resource
4).
Discussion
Burrow densities on Marion Island are still relatively low
when compared to neighbouring Prince Edward Island and
other Southern Indian Ocean Island groups (Crozet and
Kerguelen Islands). Burrowing petrel populations are dif-
ficult to census precisely (Brooke 2004), making it hard to
conduct comparable repeat counts, especially over long
periods. We were fortunate to have continuity in approach
between the two studies as MS, who made the 1979 counts,
was present to ensure that the areas sampled were matched
as closely as possible and to advise on burrow identifica-
tion and the criteria for discriminating active burrows. The
repeat survey was conducted to determine the extent of any
recovery of the petrel community following the removal of
cats more than two decades ago. Comparison with Prince
Edward Island, 22 km north-east of Marion Island,
emphasises the impact cats had on Marion’s burrowing
petrel populations. In 1979, Schramm (1986) estimated that
petrel densities on Prince Edward Island were roughly 25
times greater than those on Marion Island and attributed the
difference to predation by the large population of feral cats
on Marion ([2000 cats, Bester et al. 2002). If we assume
that adult and subadult petrels from both islands share
similar challenges at sea (e.g. changes in prey availability
and distribution [Cherel and Hobson 2007) or fisheries
bycatch (Barnes et al. 1997)], then the reasons for the huge
differences in petrel densities between the two islands are
almost certainly due to factors on the islands which affect
petrel breeding success and survival. There are no esti-
mates of petrel survival rates, but their breeding success
increased immediately following the removal of cats
(Cooper et al. 1995; Ryan and Bester 2008) and remains at
Polar Biol
123
Author's personal copy

moderate levels (FitzPatrick Institute unpubl. data), sug-
gesting that petrel populations have the potential to
recover.
When seabirds recolonise an island after extirpation, the
growth of a new colony is usually slow and limited, in part,
by delayed natal recruitment (Warham 1990). However,
with the exception of the common diving petrel (van Aarde
1980), Marion did not experience species extirpations
during the cat era and the remaining populations should
have had the potential to grow rapidly. At the time of our
survey, cats had been absent from Marion Island for
22 years, and their numbers greatly reduced for several
years prior to 1990 (Bester et al. 2002). Given recovery
rates at other islands where introduced predators have been
extirpated of around 5–7 % per year from endogenous
growth alone (e.g. Ryan et al. 2006), we would expect
petrel numbers to have increased at least threefold to
fivefold if cats were the sole factor depressing petrel pop-
ulations. White-chinned and soft-plumaged petrels were
the only species to attain these levels of growth. Consid-
ering that white-chinned petrels are one of the widest
ranging of seabirds when breeding (Weimerskirch et al.
1999) and are the seabird most often killed on longlines in
the Southern Ocean (Petersen et al. 2009; Delord et al.
2010), their strong recovery relative to other species less
prone to fishing mortality is unexpected. The increase in
soft-plumaged petrels was driven mainly by their apparent
colonisation of vegetated lava hummocks, a habitat from
which they were not recorded in 1979 (Table 1). Other
petrel populations had more modest growth rates over the
last few decades, less than doubling their numbers. Thus, it
appears that while removing the cats was a crucial step
towards the recovery of Marion Island’s burrowing petrels,
other factors may continue to suppress their populations.
In addition to endogenous growth, petrel numbers on
islands can also be boosted by immigration. Situated just
22 km to the NE of Marion, Prince Edward Island is an
ideal source for petrel immigration to Marion Island. One
example is at Aorangi Island where Buller’s Shearwater
(Puffinus bulleri) increased at 20 % per year with growth
enhanced by immigration from nearby (440 m) Tawhiti
Rahi Island (Harper 1983). Although the burrowing petrel
species at the Prince Edward Islands are probably less
aggressive colonisers than Buller’s shearwaters and the
immigration source is further away, this study shows that
immigration allows petrel populations to grow rapidly
following the removal of an introduced predator. Under
such a scenario, the Marion populations could be expected
to have increased more than 40-fold by 2013, to levels
similar to those recorded at Prince Edward Island. How-
ever, it is possible that the reduced petrel population on
Marion Island allowed for an increase in petrel densities on
Prince Edward Island through, for example, reduced
competition for food resources. The continued slow
recovery of the burrowing petrels on Marion could be a
result of this density-dependent control. Unless Prince
Edward Island becomes saturated with birds, to the point
where competition for burrow space is limiting population
growth, then juveniles and subadults are likely to return to
their natal breeding grounds (Warham 1990) and not
expand to neighbouring Marion Island.
Other seabird colonies have shown marked responses
following the eradication of feral cat populations (e.g.
Natividad and San Roque Islands off Mexico and Raoul
Island off New Zealand, Jones et al. 2011). Grey petrels on
subantarctic Macquarie Island increased sevenfold (8–59
burrows in which nesting was confirmed, Schulz et al.
2005) within 3 years of cats being eradicated (year 2000).
Fig. 6 Changes in density
estimates (±95 % CI) of all
petrel burrows (Salvin’s prions,
blue petrels, soft-plumaged
petrels, Kerguelen petrels,
great-winged petrels, white-
chinned petrels, grey petrels and
South Georgian diving petrels)
in five habitat types in north-
east Marion Island from 1979 to
2013
Polar Biol
123
Author's personal copy

Since 2003, researchers have seen substantial increases in
grey petrels and blue petrels on mainland Macquarie
Island, although rates of increase have still to be quantified
(Rachael Alderman, pers. comm.). Ascension Island
(97 km
2
), in the tropical South Atlantic Ocean, once hosted
huge seabird colonies, but the introduction of cats in 1815
proved catastrophic for local seabird populations which
eventually were restricted to cat-free cliff edges, stacks and
islets where their population sizes were limited by nest site
availability (Ashmole et al. 1994). The eradication of cats
in 2003 increased adult survival of sooty terns (Ony-
choprion fuscatus), and four species of seabird have reco-
lonised the main island from adjacent relict colonies,
despite the presence of black rats (Rattus rattus) and house
mice (Ratcliffe et al. 2010). On Jarvis Island (5 km
2
) in the
central Pacific Ocean, cats were introduced in 1936, dras-
tically reducing local seabird populations including extir-
pating numerous smaller species (Rauzon et al. 2011). Cats
also extirpated both species of introduced rats, but mice
survived. By the 1980s, cats were eradicated and most of
the extirpated seabird species began to recolonise the
island. Petrels were much slower to recover than the sur-
face-nesting boobies, frigatebirds, noddies and terns, but by
1996 seabird diversity and abundance were returning to
historically recorded levels (cf. Rauzon et al. 2011). Why
haven’t Marion’s petrel populations recovered?
Mice are the likely culprit to explain the slow recovery
of burrowing petrels at Marion Island. For 30 years, the
petrel populations were impacted by cats (top-predators)
and perhaps mice (mesopredators). While mice probably
target eggs and chicks (Fugler et al. 1987), reducing
reproductive success, cat predation was far more detri-
mental because they killed chicks and adults, affecting both
reproduction and adult survival (Le Corre 2008). Burrow-
ing petrels have long lifespans and low reproductive rates,
making their populations very sensitive to changes in adult
survival (Warham 1990). Removal of the top-predator
benefited adult survival, but may have triggered a ‘meso-
predator release effect’ (Zavaleta et al. 2001; Le Corre
2008), whereby rodent numbers expand, increasing their
impact on petrel populations (Rayner et al. 2007). How-
ever, mice were not an important prey item for cats (van
Aarde 1980), so the mouse population may not have been
limited by cat predation (van Aarde et al. 1996). Mouse
densities on Marion Island are thought to be regulated by
bottom-up processes; McClelland (2013) reported mouse
densities have increased 145 % over the past decades due
to the local effects of global climate change (warmer, drier
and less extreme climate). Peak mouse densities in
2008–2011 were 237 mice ha
-1
in mire habitats
(McClelland 2013), similar to peak densities on Gough
Island (266 mice ha
-1
; Cuthbert et al. 2016), central South
Atlantic, where mouse predation has dramatically reduced
chick survival rates of burrowing petrels and other birds
(Wanless et al. 2012; Cuthbert et al. 2013; Davies et al.
2015; Dilley et al. 2015a).
Mouse injured albatross chicks were first recorded on
Marion in 2003 and thereafter attacks continued at a low
level affecting \1 % of the albatross population (Jones
and Ryan 2010). In 2015, there was a sudden increase in
mice attacks, which were widespread across the island
affecting 9 % of large, well-feathered albatross chicks
(Dilley et al. 2015b). In 2016, the frequency and spread of
mice attacks were similar to 2015 (FitzPatrick Institute
unpubl. data). Burrow cameras installed from 2012 to 2015
revealed that mice frequently enter nest chambers and
harass chicks, with three fatal attacks recorded on film in
the winter of 2015 (one grey petrel chick and two great-
winged petrel chicks, FitzPatrick Institute unpubl. data). It
is probable that mice frequently kill burrowing petrel
chicks, which would account for the lesser recovery of
small petrels and winter-breeding petrels at Marion Island
(Fig. 4), consistent with the patterns detected at Gough
Island where mice are significant predators of seabirds
(Cuthbert et al. 2013; Davies et al. 2015; Dilley et al.
2015a). Mice on Gough are significantly larger than those
on Marion (Cuthbert et al. 2016), and although this might
confer an advantage in subduing smaller petrel chicks
(Dilley et al. 2015a), recent observations on Marion Island
show that large body mass is not necessarily a prerequisite
for mice attacking large albatross chicks (Dilley et al.
2015b).
Shortly after the introduction of cats (1951–1952),
common diving petrels were regarded by Rand (1954)as
being ‘common’ on Marion. However, during 1965–1966
Van Zinderen Bakker (1971) found no nests and common
diving petrels were thought to have been extirpated from
Marion (van Aarde 1980). This survey showed a decrease
in diving petrel burrow density on Junior’s Kop since 1979
(3.2–1.3 burrows ha
-1
), but an expansion in their distri-
bution (Table 1) with burrows now found in the more
vegetated slopes of Hoppie’s Hell, the coastal slopes of
Blue Petrel Bay and the vegetated slopes below the eastern
cliffs of Piew Crags. No breeding diving petrels were found
during this survey; however, in 2015 diving petrels were
recorded incubating in burrows on the Poa slopes of Good
Hope Bay (pers. comm. Stefan Schoombie).
Grey and white-chinned petrels have similar habitat
requirements for their burrows, preferring deep soils for
their large nesting chambers, but grey petrels breed in
winter, whereas white-chinned petrels breed in summer.
Grey petrels are scarce on Marion Island; in 1952, Rand
(1954) described them as ‘not common’, so perhaps their
numbers were low even before cats arrived (but they are
locally common on nearby Prince Edward Island; Ryan and
Bester 2008). Mouse predation may be more regular on this
Polar Biol
123
Author's personal copy

species’ chicks, because they hatch in late winter when
mice have few other food sources (Gleeson and van
Rensburg 1982; Smith et al. 2002). By contrast, the white-
chinned petrel population has increased more than three-
fold since 1979. Its large size coupled with its mid-sum-
mer-breeding season probably protects its chicks from
mouse predation. The population increase at Marion Island
contrasts with nearby Ile de la Possession in the Crozet
archipelago, where white-chinned petrels are decreasing
(Barbraud et al. 2008). The wandering albatross (Diomedea
exulans) is another species susceptible to longline mortal-
ity, but its population is stable at both Marion (Nel et al.
2002) and Ile de la Possession (Inchausti and Weimerskirch
2002), but decreasing at South Georgia (Poncet et al.
2006). The contrasting fate of the white-chinned petrel
population on Marion may reflect the presence of black rats
on Ile de la Possession, which are more aggressive preda-
tors of petrel chicks than mice.
In summary, densities of burrowing petrels remain low
on Marion Island compared to neighbouring Prince Edward
Island and other islands in the south-west Indian Ocean
which lack introduced mammalian predators. We recom-
mend regular monitoring of burrowing petrels to assess the
long-term changes in population size. Annual assessment
of breeding success in study colonies of selected species
would be valuable to assess the severity of mouse preda-
tion. Chicks should be carefully inspected for mouse
wounds. Banding adults and fledglings within study colo-
nies should be conducted to assess natal recruitment and
adult survival, provided the disturbance does not cause
undue emigration from study colonies. An expedition to
Prince Edward Island is vital to assess current burrow
densities to compare with those estimated in 1979
(Schramm 1986). Eradicating mice from Marion Island
would benefit not only the burrowing petrel populations but
also help to restore the original structure and functioning of
the island’s terrestrial ecosystems. Benefits would be both
direct (e.g. recovery of native invertebrate populations,
reduced seed predation) and indirect by promoting key
ecological processes driven by burrowing petrels (e.g. soil
disturbance and marine nutrient imports, especially to
inland sites; Caut et al. 2012).
Acknowledgments We thank Delia Davies for support and assis-
tance in the field and the 69th Marion Expedition Team members for
their support. We commend Martha
`n Bester and all the cat hunters for
their perseverance. Susan Cunningham and Jessica Shaw assisted with
data analyses. Rachael Alderman (Wildlife Management Branch,
DPIPWE, Tasmania) provided information on Macquarie Island.
Stefan Schoombie found breeding diving petrels on Marion in 2015.
The South African Department of Environmental Affairs provided
logistical support. Financial support was received from the National
Research Foundation (South African National Antarctic Programme).
We thank Graham Parker, Jeroen Creuwels and one anonymous
reviewer for their helpful comments.
References
Agostinelli C, Lund U (2013) R package ‘circular’: circular statistics
(version 0.4-7). https://r-forge.r-project.org/projects/circular/
Ashmole NP, Ashmole MJ, Simmons KEL (1994) Seabird conser-
vation on ascension island. In: Nettleship DN, Burger J,
Gochfeld M (eds) Seabirds on Islands: threats, case studies and
action plans. BirdLife International, Cambridge, pp 94–121
Barbraud C, Marteau C, Ridoux V, Delord K, Weimerskirch H (2008)
Demographic response of a population of white-chinned petrels
Procellaria aequinoctialis to climate and longline fishery
bycatch. J Appl Ecol 45:1460–1467
Barbraud C, Delord K, Marteau C, Weimerskirch H (2009) Estimates
of population size of white-chinned petrels and grey petrels at
Kerguelen islands and sensitivity to fisheries. Anim Conserv
12:258–265
Barnes KN, Ryan PG, Boix-Hinzen C (1997) The impact of the hake
Merluccius spp. longline fishery off South Africa on procellari-
iform seabirds. Biol Conserv 82:227–234
Berrow SD (2000) The use of acoustics to monitor burrow-nesting
white-chinned petrels Procellaria aequinoctialis at Bird Island,
South Georgia. Polar Biol 23:575–579
Berruti A, Hunter S (1986) Some aspects of the breeding biology of
Salvin’s prion Pachyptila vittata salvini at Marion Island.
Cormor 13:98–106
Bester MN, Bloomer JP, van Aarde RJ, Erasmus BH, van Rensburg
PJJ, Skinner JD, Howell PG, Naude TW (2002) A review of the
successful eradication of feral cats from sub-Antarctic Marion
Island, southern Indian Ocean. S Afr J Wild Res 32:65–73
Brooke M (2004) Albatrosses and petrels across the world. Oxford
University Press, Oxford
Canty A, Ripley B (2012) boot: Bootstrap R (S-Plus) functions. R
package version 1.3-4. R Core Team. 2013. R: A language and
environment for statistical computing (version 3.0.1). Founda-
tion for Statistical Computing, Vienna
Caut S, Angulo E, Pisanu B, Ruffino L, Faulquier L, Lorvelec O,
Chapuis J-L, Pascal M, Vidal E, Courchamp F (2012) Seabird
modulations of isotopic nitrogen on islands. PLoS One 7:e39125
Cerfonteyn M, Ryan PG (2016) Have burrowing petrels recovered on
Marion Island two decades after cats were eradicated? Evidence
from sub-Antarctic skua prey remains. Antarct Sci 28:51–57
Cherel Y, Hobson KA (2007) Geographical variation in carbon
stable isotope signatures of marine predators: a tool to investi-
gate their foraging areas in the Southern Ocean. Mar Ecol-Prog
Ser 329:281–287
Chown SL, Froneman PW (eds) (2008) The Prince Edward Islands:
land–sea interactions in a changing ecosystem. SunMedia,
Stellenbosch, pp 121–164
Chown SL, Smith VR (1993) Climate change and the short term
impact of feral house mice at the sub-Antarctic Prince Edward
Islands. Oecol 96:508–516
Cooper J, Fourie A (1991) Improved breeding success of great-
winged petrels Pterodroma macroptera following control of
feral cats Felis catus at sub-Antarctic Marion Island. Bird
Conserv Int 1:171–175
Cooper J, Marais AVN, Bloomer JP, Bester MN (1995) A success
story: breeding of burrowing petrels (Procellariidae) before and
after the extinction of feral cats Felis catus at sub-Antarctic
Marion Island. Mar Ornithol 23:33–37
Crawford AB (1952) The birds of Marion Island, South Indian Ocean.
Emu 52:73–85
Crawford RJM, Whittington PA, Upfold L, Ryan PG, Petersen SL,
Dyer BM, Cooper J (2009) Recent trends in numbers of four
species of penguins at the Prince Edward Islands. Afr J Mar Sci
31:419–426
Polar Biol
123
Author's personal copy

Croxall JP, Butchart SH, Lascelles B, Stattersfield AJ, Sullivan B,
Symes A, Taylor P (2012) Seabird conservation status, threats
and priority actions: a global assessment. Bird Conserv Int
22:1–34
Cuthbert RJ, Louw H, Lurling J, Parker G, Rexer-Huber K, Sommer
E, Visser P, Ryan PG (2013) Low burrow occupancy and
breeding success of burrowing petrels at Gough Island: a
consequence of mouse predation. Bird Conserv Int 23:113–124
Cuthbert RJ, Wanless RM, Angel A, Burle
´M-H, Hilton GM, Louw
H, Visser P, Wilson J, Ryan PG (2016) drivers of predatory
behaviour and extreme body size in house mice Mus musculus on
Gough Island. J Mammal 97:533–544
Davies D, Dilley BJ, Bond AL, Cuthbert RJ, Ryan PG (2015) Trends
and tactics of mouse predation on Tristan Albatross Diomedea
dabbenena chicks at Gough Island, South Atlantic Ocean. Avian
Conserv Ecol 10:5
Delord K, Cotte
´C, Pe
´ron C, Marteau C, Pruvost P, Gasco N,
Duhamel G, Cherel Y, Weimerskirch H (2010) At-sea distribu-
tion and diet of an endangered top predator: relationship between
white-chinned petrels and commercial longline fisheries. Endang
Species Res 13:1–16
Dilley BJ, Davies D, Bond AL, Ryan PG (2015a) Effects of mouse
predation on burrowing petrel chicks at Gough Island. Antarct
Sci 27:543–553
Dilley BJ, Schoombie S, Schoombie J, Ryan PG (2015b) ‘Scalping’
of albatross fledglings by introduced mice spreads rapidly at
Marion Island. Antarct Sci 28:73–80
Fugler SR, Hunter S, Newton IP, Steele WK (1987) Breeding biology
of blue petrels Halobaena caerulea at the Prince Edward Islands.
Emu 87:103–110
Gleeson JP, van Rensburg PJJ (1982) Feeding ecology of the house
mouse Mus musculus on Marion Island. S Afr J Antarct Res
12:34
Gremmen NJ (1997) 59 Changes in the vegetation of sub-Antarctic
Marion Island resulting from introduced vascular plants. Antarct
Commun Species Struct Surviv 417
Gremmen NJM, Chown SL, Marshall DJ (1998) Impact of the
introduced grass Agrostis stolonifera on vegetation and soil
fauna communities at Marion Island, sub-Antarctic. Biol Con-
serv 85:223–231
Harper PC (1983) Biology of the Buller’s shearwaters (Puffinus
bulleri) at the Poor Knights Islands, New Zealand. Notornis
30:299–318
Huntley BJ (1971) Vegetation. In: Van Zinderen Bakker EM,
Winterbottom JM, Dyer RA (eds) Marion and Prince Edward
Islands: Report on the South African Biological and Geological
Expedition 1965-1966. AA Balkema, Cape Town, pp 98–160
Inchausti P, Weimerskirch H (2002) Dispersal and metapopulation
dynamics of an oceanic seabird, the wandering albatross, and its
consequences for its response to long-line fisheries. J Anim Ecol
71:765–770
Jones MGW, Ryan PG (2010) Evidence of mouse attacks on albatross
chicks on sub-Antarctic Marion Island. Antarct Sci 22:39–42
Jones HP, Towns DR, Bodey T, Miskelly C, Ellis JC, Rouzon M,
Kress S, McKnown M (2011) Eradication of invasive predators
on seabird islands. In: Mulder CPH, Anderson WB, Towns DR,
Bellingham PJ (eds) Seabird Islands: ecology, invasion and
restoration. Oxford University Press, Chicago, pp 324–325
Le Corre M (2008) Conservation biology: cats, rats and seabirds.
Nature 451:134–135
McClelland GT (2013) Ecology of the black-faced sheathbill on
Marion Island. Dissertation, Stellenbosch University
Nel DC, Ryan PG, Watkins BP (2002) Seabird mortality in the
Patagonian toothfish longline fishery around the Prince Edward
Islands, 1996-2000. Antarct Sci 14:151–161
Payne MR, Prince PA (1979) Identification and breeding biology of
the diving petrels Pelecanoides georgicus and P. urinatrix exsul
at South Georgia. New Zeal J Zool 6:299–318
Petersen SL, Honig MB, Ryan PG, Underhill LG (2009) Seabird
bycatch in the pelagic longline fishery off southern Africa. Afr J
Mar Sci 31:191–204
Phillips RA, Caldow R, Furness R (1996) The influence of food
availability on the breeding effort and reproductive success of
Arctic skuas Stercorarius parasiticus. Ibis 138:410–419
Poncet S, Robertson G, Phillips RA, Lawton K, Phalan B, Trathan
PN, Croxall JP (2006) Status and distribution of wandering,
black-browed and grey-headed albatrosses breeding at South
Georgia. Polar Biol 29:772–781
R Development Core Team (2014) R: a language and environment for
statistical computing. Version 3.1 http://www.r-project.org.
R Foundation for Statistical Computing, Vienna, Austria
Rand RW (1954) Notes on the birds of Marion Island. Ibis
96:173–206
Ratcliffe N, Bell M, Pelembe T, Boyle D, Benjamin R, White R,
Godley B, Stevenson J, Sanders S (2010) The eradication of feral
cats from Ascension Island and its subsequent recolonization by
seabirds. Oryx 44:20–29
Rauzon MJ, Forsell DJ, Flint EN, Gove JM (2011) Islands 25 years
after cat eradication: the recovery of seabirds in a biogeograph-
ical context. In: Veitch CR, Clout MN, Towns DR (eds) Island
invasives: eradication and management. IUCN, Gland,
pp 345–349
Rayner MJ, Hauber ME, Imber MJ, Stamp RK, Clout MN (2007)
Spatial heterogeneity of mesopredator release within an oceanic
island system. Proc Natl Acad Sci 104:20862–20865
Ryan PG, Bester MN (2008) Pelagic predators. In: Chown SN,
Froneman W (eds) The Prince Edward Islands: land-sea
interactions in a changing ecosystem. African Sun Media,
Stellenbosch, pp 121–164
Ryan PG, Ronconi RA (2011) Continued increase in numbers of
spectacled petrels Procellaria conspicillata. Antarct Sci
23:332–336
Ryan PG, Dorse C, Hilton GM (2006) The conservation status of the
Spectacled Petrel Procellaria conspicillata. Biol Conserv
131:575–583
Ryan PG, Jones MGW, Dyer BM, Upfold L, Crawford RJM (2009a)
Recent population estimates and trends in numbers of albatrosses
and giant petrels breeding at the sub-Antarctic Prince Edward
Islands. Afr J Mar Sci 31:409–417
Ryan PG, Whittington PA, Crawford RJM (2009b) A tale of two
islands: contrasting fortunes for Subantarctic skuas at the Prince
Edward Islands. Afr J Mar Sci 31:431–437
Ryan PG, Dilley BJ, Jones MGW (2012) The distribution and
abundance of white-chinned petrels (Procellaria aequinoctialis)
breeding at the sub-Antarctic Prince Edward Islands. Polar Biol
35:1851–1859
Schramm M (1983) The breeding biologies of the petrels Pterodroma
macroptera, P. brevirostis and P. mollis at Marion Island. Emu
83:75–81
Schramm M (1986) Burrow densities and nest site preferences of
petrels Procellariidae at the Prince Edward Islands. Polar Biol
6:63–70
Schulz M, Robinson S, Gales R (2005) Breeding of the Grey Petrel
(Procellaria cinerea) on Macquarie Island: population size and
nesting habitat. Emu 105:323–329
Smith VR (1976) Standing crop and nutrient status of Marion Island
(sub-Antarctic) vegetation. J S Afr Bot 42:231–263
Smith VR, Avenant NL, Chown SL (2002) The diet and impact of
house mice on a sub-Antarctic island. Polar Biol 25:703–715
Polar Biol
123
Author's personal copy

Van Aarde RJ (1977) Voeding, habitatsvoorkeur en voortplanting van
die wildehuiskat (Felis catus Linnaeus, 1758) op Marioneiland.
MSc thesis, University of Pretoria, Pretoria
Van Aarde RJ (1980) The diet and feeding behaviour of feral cats,
Felis catus at Marion Island. S Afr J Wildl Res 10:123–128
Van Aarde R, Ferreira S, Wassenaar T, Erasmus DG (1996) With the
cats away the mice may play. S Afr J Sci 92:357–358
Van Zinderen Bakker EM (1971) Comparative avian ecology. In:
Winterbottom JM, Dyer RA, Van Zinderen Bakker EM Sr (eds)
Marion and Prince Edward Islands: Report on the South African
Biological and Geological Expedition, 1965–1966. Balkema,
Cape Town, pp 161–172
Verwoerd WJ (1971) Geology. In: Winterbottom JM, Dyer RA, Van
Zinderen Bakker EM Sr (eds) Marion and Prince Edward
Islands. Rep South Afr Biol Geol Exp 1965–1966. Balkema,
Cape Town, pp 40–62
Wanless RM, Ratcliffe N, Angel A, Bowie BC, Cita K, Hilton GM,
Kritzinger P, Ryan PG, Slabber M (2012) Predation of Atlantic
petrel chicks by house mice on Gough Island. Animal Conserv
15:472–479
Warham J (1990) The petrels—their ecology and breeding systems.
Academic Press, London
Watkins BP, Cooper J (1986) Introduction, present status and control
of alien species at the Prince Edward Islands, sub-Antarctic.
S Afr J Antarct Res 16:86–94
Weimerskirch H, Catard A, Prince PA, Cherel Y, Croxall JP (1999)
Foraging white-chinned petrels Procellaria aequinoctialis at
risk: from the tropics to Antarctica. Biol Conserv 87:273–275
Zavaleta ES, Hobbs RJ, Mooney HA (2001) Viewing invasive species
removal in a whole-ecosystem context. Trends Ecol Evol
16:454–459
Polar Biol
123
Author's personal copy