Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
VOLUME 21 (2008)
<strong>Seabird</strong> publishes papers and short<br />
communications on any aspect of<br />
seabird biology, conservation, identification,<br />
and status. <strong>The</strong> geographical<br />
focus of the journal is the Atlantic<br />
Ocean and adjacent seas, but contributions<br />
are also welcome from other<br />
parts of the world provided they are of<br />
general interest. Detailed guidelines<br />
for contributors are available at<br />
www.seabirdgroup.org.uk.<br />
Editor:<br />
Martin Heubeck, Sumburgh<br />
Lighthouse, Virkie, Shetland ZE3 9JN.<br />
Email: martinheubeck@btinternet.com<br />
Publishing Editor:<br />
Linda Wilson, JNCC, Dunnet House,<br />
7 Thistle Place, Aberdeen, AB10 1UZ.<br />
Email: linda.wilson@jncc.gov.uk<br />
Editorial Board:<br />
Tycho Anker-Nilssen (Norway), Josep<br />
Arcos (Spain), Kees Camphuysen (<strong>The</strong><br />
Netherlands), John Chardine (Canada),<br />
Stefan Garthe (Germany), Morten<br />
Frederiksen (Denmark), Norman<br />
Ratcliffe (UK), Jim Reid (UK), Sarah<br />
Wanless (UK)<br />
Photographic Consultant:<br />
Hugh Harrop (www.hughharrop.com)<br />
Design & Typesetting:<br />
Harry Scott (Pica Design)<br />
Print & Production:<br />
F. Crowe & Son Ltd<br />
(using FSC paper)<br />
Front cover photo:<br />
Leach’s Storm-petrel © Sue Tranter<br />
<strong>Seabird</strong> 21 (2008) is the first re-launched<br />
volume of the <strong>Seabird</strong> <strong>Group</strong>’s former<br />
journal <strong>Seabird</strong>, which discontinued in 1998<br />
(Volume 20). Between 1998 and 2008, the<br />
<strong>Seabird</strong> <strong>Group</strong> published the journal<br />
Atlantic <strong>Seabird</strong>s (Vols 1–8) jointly with the<br />
Dutch <strong>Seabird</strong> <strong>Group</strong> (Nederlandse<br />
Zeevogelgroep, NZG). Back issues of these<br />
journals will shortly be available online at:<br />
www.seabirdgroup.org.uk.<br />
Registered Charity 260907<br />
SEABIRD<br />
Volume 21 (2008)<br />
CONTENTS<br />
PAPERS<br />
1 Identifying giant petrels, Macronectes giganteus and M. halli,<br />
in the field and in the hand CARLOS, C. J., & VOISIN, J.-F.<br />
16 Vagrancy of Brünnich’s Guillemot Uria lomvia in Europe<br />
VAN BEMMELEN, R., & WIELSTRA, B.<br />
32 A survey of Leach’s Oceanodroma leucorhoa and European<br />
Storm-petrel Hydrobates pelagicus populations on North<br />
Rona and Sula Sgeir, Western Isles, Scotland MURRAY, S.,<br />
MONEY, S., GRIFFIN, A. & MITCHELL, P. I.<br />
44 <strong>The</strong> diet of European Shag Phalacrocorax aristotelis, Blacklegged<br />
Kittiwake Rissa tridactyla and Common Guillemot<br />
Uria aalge on Canna during the chick-rearing period<br />
1981–2007 SWANN, R. L., HARRIS, M. P. & AITON, D. G.<br />
55 Colony habitat selection by Little Terns Sternula albifrons in<br />
East Anglia: implications for coastal management RATCLIFFE,<br />
N., SCHMITT, S., MAYO, A., TRATALOS, J. AND DREWITT, A.<br />
64 Descriptive anatomy of the subcutaneous air diverticula in<br />
the Northern Gannet Morus bassanus DAOUST, P.-Y., DOBBIN,<br />
G. V., RIDLINGTON ABBOTT, R. C. F. & DAWSON, S. D.<br />
77 Population decline of Leach’s Storm-petrel Oceanodroma<br />
leucorhoa within the largest colony in Britain and Ireland<br />
NEWSON, S. E., MITCHELL, P. I., PARSONS, M., O’BRIEN, S. H., AUSTIN,<br />
G. E., BENN S., BLACK J., BLACKBURN, J., BRODIE, B., HUMPHREYS, E.,<br />
LEECH, D., PRIOR, M. & WEBSTER, M.<br />
NOTES<br />
85 Rafting behaviour of Manx Shearwaters Puffinus puffinus<br />
WILSON, L. J., MCSORLEY, C. A., GRAY, C. M., DEAN, B. J., DUNN, T. E.,<br />
WEBB, A. & REID, J. B.<br />
93 Late breeding by Great Cormorants Phalacrocorax carbo<br />
CRAIK J. C. A. & BREGNBALLE T.<br />
98 A pilot study of the phenology and breeding success of<br />
Leach’s Storm-petrel Oceanodroma leucorhoa on St Kilda,<br />
Western Isles MONEY, S., SÖHLE, I. & PARSONS, M.<br />
102 Use of gulls rather than terns to evaluate American Mink<br />
Mustela vison control CRAIK, J. C. A.<br />
104 Use of gulls rather than terns to evaluate American Mink<br />
Mustela vison control. A response to Craik (2008) RATCLIFFE, N.<br />
105 Fish brought to young Atlantic Puffins Fratercula arctica on<br />
Burhou, Channel Islands SANDERS, J. G.<br />
108 REVIEWS
Identification of giant petrels<br />
Identifying giant petrels, Macronectes<br />
giganteus and M. halli, in the field<br />
and in the hand<br />
Carlos, C. J. 1,2 *, & Voisin, J.-F. 3<br />
*Correspondence author. Email: cjcarlos@bol.com.br or macronectes1@yahoo.co.uk<br />
1 Laboratório de Elasmobrânquios e Aves Marinhas, Departamento de Oceanografia,<br />
Fundação Universidade Federal do Rio Grande, CP 474, 96201-900, Rio Grande, RS,<br />
Brazil; 2 Current address: Rua Mário Damiani Panatta 680, Cinqüentenário, 95013-290,<br />
Caxias do Sul, RS, Brazil; 3 USM 305, Département Ecologie et Gestion de la Biodiversité,<br />
Muséum National d’Histoire Naturelle, CP 51, 57 rue Cuvier, F-75 005 Paris, France.<br />
Abstract<br />
<strong>The</strong> two similar-looking species of giant petrels, the Northern Giant Petrel<br />
Macronectes halli and the Southern Giant Petrel M. giganteus, are renowned for being<br />
difficult to identify. In this paper we review and offer new guidelines on identification<br />
of these birds at sea, on land, and as dead specimens. Criteria for identifying giant<br />
petrels are available in the scientific literature, especially regarding bill-tip coloration<br />
which readily differ from one species to another. Plumage characters, although useful<br />
to discriminate species, are not adequately covered at present. Thus, for each species<br />
we describe in detail and illustrate distinctive age-related plumage stages, or types,<br />
from juveniles through to adult breeders.We also comment on giant petrel biometrics,<br />
body weight, and some aspects of their behaviour, in order to help ornithologists and<br />
birdwatchers separate males and females, and eventually specimens from South<br />
America–Gough Island, Antarctic and sub-Antarctic regions.<br />
Introduction<br />
Giant petrels Macronectes are large Procellariiformes with a wingspan usually in excess<br />
of two meters in males, slightly less in females, and are similar in size to Thalassarche<br />
mollymawks. <strong>The</strong>y display a wide range of plumages, from dull-black to brownish-grey<br />
and even white, according to age and species. <strong>The</strong> discovery of two distinct<br />
populations breeding at Macquarie Island (Australian Antarctic Territory; Figure 1),<br />
laying about six weeks apart (Warham 1962), led Bourne & Warham (1966) to restrict<br />
the name Macronectes giganteus (Gmelin, 1789) to those birds breeding on Antarctic<br />
islands, the Antarctic Peninsula and continent, and on islands of the South Atlantic, and<br />
to resurrect the name Macronectes halli Mathews, 1912 for birds breeding on sub-<br />
Antarctic islands, from South Georgia through the Indian Ocean to the New Zealand<br />
area. Voisin & Bester (1981) and Voisin (1982b) showed later that the form breeding<br />
on the Falkland and Gough Islands (South Atlantic), as well as on islands off the<br />
Argentine coast, represented a well-marked subspecies: Macronectes giganteus<br />
solanderi Mathews, 1912, for which Carlos et al. (2005) proposed the vernacular name<br />
South Atlantic Giant Petrel. Hereafter, we follow current usage in referring to M. halli<br />
as Northern Giant Petrel and M. giganteus as Southern Giant Petrel.<br />
SEABIRD 21 (2008): 1–15<br />
1
2<br />
Identification of giant petrels<br />
Ringing recoveries and subsequent satellite-tracking studies have shown that both<br />
species of giant petrel roam the southern ocean widely, especially immatures, but also<br />
adults in the breeding and non-breeding seasons (Hunter 1984; Voisin 1990; Parmalee<br />
1992;Trivelpiece & Trivelpiece 1998; González-Solís et al. 2000a, b; Patterson & Hunter<br />
2000; Patterson & Fraser 2000, 2003; BirdLife International 2004; González-Solís et al.<br />
2008). Antarctic populations of Southern Giant Petrel are especially wide-ranging, and<br />
one should not assume that, for example, a Southern Giant Petrel seen off the<br />
Argentinean–Brazilian coast at any time of the year is necessarily M. g. solanderi.<br />
Recently, Penhallurick & Wink (2004) proposed to re-group M. halli as a subspecies of M.<br />
giganteus on the basis of low mitochondrial cytochrome-b gene divergence. Such a<br />
position is untenable since both species have long bred sympatrically, and often in close<br />
proximity, on South Georgia, Marion, Crozet and Macquarie Islands (Figure 1) without<br />
regular interbreeding. A few cases of inter-specific breeding involving male M. giganteus<br />
paired to female M. halli have been reported from Marion (Burger 1978; Cooper et al.<br />
2001) and Macquarie Islands (Johnstone 1978), although the resulting eggs did not<br />
hatch. Only at South Georgia has occasional hybridisation been reported (Hunter 1982,<br />
1987). Given these circumstances, we can reasonably assume that gene flow between<br />
the two species is extremely limited. In birds, many closely related species are known to<br />
hybridise regularly, e.g. Anas dabbling ducks (Carboneras 1992), and Common<br />
Phylloscopus collybita and Iberian P. ibericus Chiffchaffs (Salomon et al. 2003), without<br />
merging into one single species. Even if ‘their apparent failure to interbreed is not quite<br />
as straightforward as if they bred at the same time without interbreeding’ (Penhallurick<br />
& Wink 2004), M. giganteus and M. halli do not genetically mix.<br />
Several authors have dealt with giant petrel identification, especially of birds at sea<br />
(Johnstone 1971, 1974; Conroy et al. 1975; Voisin 1982a; Hunter 1983; Voisin &<br />
Teixeira 1998; Jiguet 2000; Shirihai & Jarrett 2002). Here we review, and offer new<br />
guidelines on, giant petrel identification at sea, on land, and as dead specimens.<br />
Separating giant petrels from other Procellariiformes<br />
At sea, and in poor light conditions, inexperienced observers may confuse giant petrels<br />
with Diomedea albatrosses, Thalassarche mollymawks or Phoebetria sooty albatrosses.<br />
When flying, giant petrels adopt a typically hump-backed posture which, when viewed<br />
from the side, allows easy separation from albatrosses and mollymawks. Also, the<br />
massive, pale bill of giant petrels is quite different to the slender one of mollymawks<br />
and sooty albatrosses (Figure 2), and contributes to give them a clumsy appearance.<br />
Due to a larger wing-loading (Pennycuick 1982; Obst & Nagy 1992), the flight of giant<br />
petrels is more laboured (usually 4–5 flaps followed by a stiff-winged glide) than that<br />
of mollymawks and sooty albatrosses, which are of similar size but are decidedly more<br />
graceful. Macronectes halli and dark form M. giganteus can also be distinguished from<br />
juvenile Wandering Albatrosses (sensu lato) by their dark, not white, underwing and<br />
from sooty albatrosses by their short and round, not long and pointed, tail (Figure 2).<br />
<strong>The</strong> white form of M. giganteus (see Plumage characters) is distinguished from<br />
Diomedea albatrosses and mollymawks by its white, not dark remiges, and most often<br />
by its diagnostic random dark brown flecking.<br />
SEABIRD 21 (2008): 1–15
Identification of giant petrels<br />
Figure 1. At-sea and breeding distribution of Northern Macronectes halli (Ng) and Southern Giant Petrels M.<br />
giganteus (Sg), and locations where the two species are known to breed sympatrically. (1) South Georgia, (2) Marion<br />
Island, (3) Crozet Islands, (4) Kerguelen Island, and (5) Macquarie Island.<br />
On land, separating giant petrels from albatrosses and<br />
mollymawks is straightforward for adults, because of<br />
their characteristic bill shape, smaller and flattened, not<br />
rounded, heads and plumage coloration. Although the<br />
bill shape of giant petrel chicks is somewhat different<br />
from that of adults, it is distinctive enough to allow<br />
separation from mollymawk chicks. Giant petrels nest in<br />
a variety of situations, from tight to loose colonies and<br />
even in solitary nests. Like all Procellariiformes, giant<br />
petrels are rather smelly birds, but their odour is stronger<br />
and less ‘sweet’ than that of albatrosses and<br />
mollymawks, and persists for a very long time on old<br />
nest material, feathers, and museum specimens, and<br />
once experienced is easy to recognise.<br />
Separating M. halli and M. giganteus<br />
Bill coloration: <strong>The</strong> best criterion for identifying giant<br />
petrels is bill coloration. In breeding M. halli, the basic<br />
colour of the bill is pinkish-horn to reddish-brown,<br />
turning brick-red on the maxillary and mandibulary<br />
ungues, or nails, and sometimes also on the latericorns<br />
Figure 2. Silhouettes, drawn to about the<br />
same scale, of (1) a Thalassarche mollymawk,<br />
(2) a Phoebetria sooty albatross, and (3) a<br />
giant petrel Macronectes. <strong>The</strong> birds are similar<br />
in size, but giant petrels have massive bills<br />
compared to the slender ones of mollymawks<br />
and sooty albatrosses. Giant petrels also have<br />
a shorter, round tail, which is longer and more<br />
pointed in sooty albatrosses.<br />
SEABIRD 21 (2008): 1–15<br />
3
4<br />
Identification of giant petrels<br />
Figure 3. Northern Macronectes halli (plumage type H5; left) and Southern Giant Petrels M. giganteus (plumage type<br />
G7; right), Abattoir Outlet, Falkland Islands, 2006 © Steve Copsey.<br />
Figure 4. White form Southern Giant Petrel M. giganteus, South Georgia, 1998 © Ronald Saldino.<br />
(horny plates running along the maxilla or upper-mandible; Figure 3). A few bluishblack<br />
markings can also be found on both nails. <strong>The</strong> number of birds displaying such<br />
markings varies between populations, being frequent in some and rare in others.When<br />
seen from a distance and in good light conditions, the bill of M. halli may look ‘bicoloured’,<br />
with an obvious darker tip.<br />
In young M. halli chicks, the bill is light brown with a more or less pronounced reddish<br />
hue on the nails, and becomes coloured like that of an adult when the mesoptile<br />
down (the second set of down) is acquired. <strong>The</strong> reddish tinge on the bill of fledglings<br />
may fade somewhat and some immature birds have entirely yellowish-brown bills,<br />
with just a slight red colour on the nails. A few birds in adult plumage encountered at<br />
sea may also have uniformly coloured bills, and may be birds taking a ‘sabbatical<br />
leave’ from breeding (Voisin 1988).<br />
SEABIRD 21 (2008): 1–15
Identification of giant petrels<br />
Figure 5. Heads of newly hatched Northern Macronectes halli (left) and Southern Giant Petrels<br />
M. giganteus (right) chicks (redrawn from Voisin 1968).<br />
In M. giganteus, the bill is yellowish with green nails, the tinge of which can vary from<br />
bluish to apple-green (Figure 3). In birds indulging in sexual displays, this green often<br />
extends to the latericorns, and mainly so on their lower part. This is especially so in M.<br />
g. solanderi, and birds breeding at Gough Island in particular may develop entirely<br />
green bills (Voisin & Bester 1981). When seen from a distance, the bill of M. giganteus<br />
may also appear ‘bi-coloured’, but the pale green nails contrast only slightly with the<br />
remaining yellowish bill.<br />
In young M. giganteus chicks, the bill is pinkish-cream with a bluish-green tip, and<br />
evolves to a pattern similar to that of adults, but usually somewhat paler. Even though<br />
bill colours fade rather in juvenile and immature giant petrels, they remain sufficiently<br />
characteristic for specific identification when seen in good light.<br />
Plumage characters: <strong>The</strong> plumage of M. giganteus is dimorphic, with a grey-brown,<br />
pale-headed dark form and a white form. According to Shaughnessy (1970), this<br />
dimorphism is controlled by two autosomal alleles with white dominant to dark.White<br />
form birds (up to 15% in some populations; Shaughnessy 1971) are entirely white<br />
from hatching to adulthood, except for a variable amount of dark brown feathers<br />
irregularly scattered throughout the body (Figure 4). Bill colour of the white form is<br />
identical to that of the dark form. Dark form birds have dark grey legs and feet, while<br />
white form ones have legs that vary from bluish-grey to pink-grey. Totally white-<br />
Figure 6. Southern Giant Petrels Macronectes giganteus: plumage type G2 (left) and G4 (right), Abattoir Outlet,<br />
Falkland Islands, 2006 © Steve Copsey.<br />
SEABIRD 21 (2008): 1–15<br />
5
6<br />
Identification of giant petrels<br />
plumaged birds (leucistic; Conroy et al. 1975), lacking the dark brown mottling of the<br />
white form, with a uniform horn-coloured tip to a pinkish beak, and pink legs and feet,<br />
are regarded to be homozygous white (Shaughnessy & Conroy 1977). Identification of<br />
white form M. giganteus is straightforward at all ages, as M. halli has no white form.<br />
Chicks: Newly hatched dark form M. giganteus chicks are covered with pearl-grey<br />
down, sometimes with a slightly darker head. At Gough Island, however, some South<br />
Atlantic Giant Petrel chicks appear similar to M. halli chicks (Voisin & Bester 1981). At<br />
some locations, such as the Crozet and Kerguelen Islands, almost all newly hatched M.<br />
halli are whitish below and ash-grey on the back, darkest on the crown and nape,<br />
where it forms a cap contrasting with a paler forehead, face and sides of head (Voisin<br />
1968; Figure 5). At other sites they are simply darker grey than M. giganteus chicks.<br />
Chicks of both species in mesoptile down are of the same ash-grey colour.<br />
Juveniles and immatures: Fledgling Northern and dark form Southern Giant Petrels have<br />
the same uniform dull glossy blackish-brown to black plumage.According to locality, some<br />
individuals may also display a number of white feathers on their forehead and cheeks, as<br />
is frequently the case with M. halli chicks from the Crozet and Kerguelen Islands.<br />
After one year, these black juveniles moult into uniformly dull brown immatures<br />
(Figure 6). <strong>The</strong>se birds remain mostly at sea for up to ten years or more, rarely coming<br />
to colonies (Conroy 1972; Hunter 1984; Voisin 1988), and their plumage changes are<br />
poorly understood. After each moult, their plumage acquires a more brownish tinge,<br />
and the area around the base of the bill becomes mottled and paler (Figure 6).<br />
Plumage differences between immature Northern and Southern Giant Petrels, if any,<br />
are very slight and of no help in specific identification.<br />
Adults: <strong>The</strong> plumage of adults, like that of immatures, becomes paler at each moult,<br />
but these changes do not progress in the same way in nominate M. g. giganteus, in M.<br />
g. solanderi and M. halli.<br />
Identification of white form and ‘totally white’ Southern Giant Petrels is straightforward.<br />
In the dark form of the nominate subspecies, a very few birds breed while still in dark<br />
plumage, having the area around the base of the bill paler, mottled yellowish and brown.<br />
<strong>The</strong> plumage then becomes more greyish-brown and the mottling around the base of<br />
the bill extends progressively, encompassing the face, sides of the head and throat. <strong>The</strong><br />
head and neck then gradually become entirely whitish-grey, with a few feathers on the<br />
middle of the crown and along the dorsal side of the neck remaining darker brown. Old<br />
birds have almost entirely white heads and necks. Even in these, the chest, belly, and<br />
undertail are as dark as the upperparts, giving them a characteristic dark-bodied and<br />
white-necked appearance. In a few cases, the underparts may be slightly paler than the<br />
upperparts, but never whitish.<strong>The</strong> transition between the whitish colour of the head and<br />
neck and the dark colour of the rest of the body is narrow, and from a distance appears<br />
as a clear-cut demarcation line. Voisin (1982b) distinguished eight plumage stages in<br />
Southern Giant Petrels (Table 1, Figure 7). Many old birds have a pale leading edge to the<br />
wing, but this is not a diagnostic feature, as some old M. halli at South Georgia also show<br />
this character (Hunter 1983; Voisin & Teixeira 1998).<br />
SEABIRD 21 (2008): 1–15
Identification of giant petrels<br />
Figure 7. Sketches of ‘old’ adult Southern Giant Petrels Macronectes giganteus, plumages types G5–G8 (based on<br />
Voisin 1982b). See Table 1 for descriptions and comments.<br />
Table 1. Plumage types of dark form Southern Giant Petrels Macaronectes giganteus (modified from Voisin<br />
1982b). White form birds are white throughout their lives.<br />
Type Description Comments<br />
G1 Entirely black birds. Juveniles (1st calendar year).<br />
G2 Entirely brown birds, the area around the base of the bill Nominate giganteus: immatures;<br />
not distinctly paler. M. g. solanderi: some breeders.<br />
G3 Entirely brown birds, the area around the base of the bill Nominate: older immatures and a<br />
light yellowish-brown. few breeders; M. g. solanderi: breeders.<br />
G4 Brown birds with a pale face and a pale area around the Both subspecies: breeders.<br />
base of the bill.<br />
G5 Brown birds with face, lower forehead and throat all pale; Nominate: breeders; M. g. solanderi:<br />
underparts as dark, or almost as dark as upperparts (Figure 7). breeders (most).<br />
G6 Birds with a pale face, sides of the head, throat and upper Nominate: breeders (most); M. g.<br />
foreneck; underparts as G5 (Figure 7). solanderi: breeders.<br />
G7 Birds with a pale head and neck, still having a brown cap on Both subspecies: breeders.<br />
the head and some dark feathering along the dorsal side of<br />
the neck; underparts as G6 (Figure 7).<br />
G8 Birds with entirely whitish heads and necks, with sometimes Both subspecies: breeders.<br />
a few dark feathers on the nape; underparts as G7 (Figure 7).<br />
Notes: Birds belonging to types G7 and G8 show a fairly clear-cut demarcation line between the whitish colour<br />
of their heads and necks and the dark colour of the rest of their bodies (Figure 7). An exception to this is M. g.<br />
solanderi at Gough Island, which shows a more diffuse transition between both colours. Exceptionally, dark form<br />
M. giganteus may have their underparts slightly paler than their upperparts, but never light grey or whitish. It is<br />
not known whether this difference is permanent or just transitory.<br />
SEABIRD 21 (2008): 1–15<br />
7
8<br />
Identification of giant petrels<br />
Figure 8. Sketches of ‘old’ adult Northern Giant Petrels Macronectes halli, plumages types H5–H8. See Table 2 for<br />
descriptions and comments.<br />
Table 2. Plumage types of Northern Giant Petrels Macronectes halli.<br />
Type Description Comments<br />
H1 Entirely black birds. Juveniles (1st calendar year).<br />
H2 Entirely brown birds, the area around the base of the bill not Immatures.<br />
distinctly paler.<br />
H3 Entirely brown birds, the area around the base of the bill light Immatures.<br />
yellowish-brown.<br />
H4 Brown birds with a pale face and a pale area around the base May breed. <strong>The</strong> pale area on upper<br />
of the bill and on the upperthroat. throat on average a little larger than<br />
in M. giganteus.<br />
H5 Brown birds with face, lower forehead and throat all pale Breeders.<br />
(Figure 8).<br />
H6 Birds with a pale face, sides of the head to ear coverts, foreneck Breeders.<br />
paler than hindneck, underparts a little paler than upperparts<br />
(Figure 8).<br />
H7 Pale areas on the head larger, nape dark, foreneck much paler Breeders.<br />
than hindneck, underparts much paler than upperparts (Figure 8).<br />
H8 Like H7, but dark areas on the nape and hindneck reduced; Breeders.<br />
chest and belly whitish (Figure 8).<br />
Notes: Types H1–H4 are virtually indistinguishable from types G1–G4 of M. giganteus. Type H5 can only be<br />
reliably determined in the field when seen in good conditions (see Figures 7 & 8).<br />
SEABIRD 21 (2008): 1–15
Identification of giant petrels<br />
<strong>The</strong> South Atlantic Giant Petrel differs from the nominate subspecies in having a<br />
generally darker plumage, with a more intense grey tone. It also acquires a paler head<br />
and neck, but more slowly and at an older age. In the Falkland Islands and Argentinean<br />
populations, the border between the pale head and neck and darker body is as clearcut<br />
as in nominate giganteus, whereas it is more diffuse in the population breeding at<br />
Gough Island (Voisin & Bester 1981).<br />
In Northern Giant Petrels, the plumage of young breeders is quite similar to that of<br />
young breeding M. giganteus, and the two species cannot be separated on that<br />
character alone. As birds get older, the face, sides of the head and foreneck become<br />
Table 3. Measurements (mm) and body mass (kg) of male and female Southern Giant Petrels Macronectes<br />
giganteus from several breeding places. Data are: mean, (± standard deviation; range; sample size). (1) Falkland<br />
Islands (Voisin 1982b); (2) Chubut, Argentina (Copello et al. 2006); (3) Gough Island; (4) Crozet Islands; (5)<br />
Antarctic Peninsula (Voisin & Bester 1981); (6) South Orkney Islands (Conroy 1972); (7) Frazier Island,<br />
Antarctica; and (8) Macquarie Island (G. W. Johnstone pers. comm.).<br />
Males Females<br />
Wing (1) 510.8 (9.58; 500–525; 5) 485.5 (9.06; 470–500; 10)<br />
(2) 517 (12; 507–530; 3) 491 (0.9; 478–508; 8)<br />
(3) 507.5 (485–535; 4) 484.5 (478, 491; 2)<br />
(4) 533.2 (14.1; 492–555; 15) 494.9 (23.85; 460–522; 7)<br />
(5) 510 (25.17; 465–552; 8) 499.8 (13.44; 477–515; 8)<br />
(6) 553.5 (10.86; 534–571; 13) 518.9 (15.21; 498–541; 13)<br />
(7) 542 (8.54; 530–550; 7) 512 (7.18; 500–518; 5)<br />
(8) 552 (10.66; 534–577; 20) 526.8 (8.52; 513–540; 21)<br />
Culmen (1) 97.9 (3.21; 94.5–102; 5) 84.4 (2.55; 80–89; 10)<br />
(2) 95.5 (2.6; 91.7–101.2; 18) 82.9 (2.1; 80.12–87.3; 22)<br />
(3) 95.2 (2.3; 91.5–98.5; 17) 83.3 (2.9; 79.5–88; 12)<br />
(4) 104.7 (3.97; 97–111; 16) 89.1 (3.28; 84–94; 8)<br />
(5) 98.5 (2.02; 96–102; 9) 89.1 (3.97; 85.5–95 ; 8)<br />
(6) 101.4 (2.45; 97.4–108.2; 66) 87.4 (3; 82–97; 73)<br />
(7) 98.9 (2.29; 96–103.1; 7) 88 (2.09; 85.8–90.4; 5)<br />
(8) 101.9 (2.15; 95.8–105; 20) 89.7 (2.41; 84.6–93.5; 21)<br />
Tarsus (1) 93 (1.46; 91.5–95; 5) 84 (2.2; 79.5–86; 10)<br />
(2) 92.2 (2.1; 87.7–96.9; 15) 84.8 (3.6; 80.5–99.1; 22)<br />
(3) 95.3 (2.02; 92–99; 17) 88.3 (3.11; 85–95; 12)<br />
(4) 102 (2.64; 97–107; 14) 95.1 (6.83; 88–111; 8)<br />
(5) 94.7 (5.12; 85–101; 9) 89 (3.29; 84–93; 6)<br />
(6) 96.3 (1.69; 94–99; 13) 87.1 (5.45; 82–95; 13)<br />
(7) 99.3 (2.34; 95.8–102; 7) 90.4 (2.19; 87.8–93.8; 5)<br />
(8) 102.7 (1.71; 99.5–105.3; 20) 94.2 (1.88; 90.1–98; 21)<br />
Body mass (2) 3.5 (0.3; 15) 2.5 (0.2; 21)<br />
(3) 3.77 (0.36; 2.3–4.55; 11) 3.17 (0.11; 2.7–3.9; 11)<br />
(4) 4.93 (0.34 ; 4.2–5.5; 15) 3.95 (0.17; 4.7–3.3; 8)<br />
(6) 4.94 (0.41; 4.1–5.8; 37) 3.85 (0.37; 3–4.8; 37)<br />
(8) 5.14 (0.42; 4.3–5.6; 20) 4.2 (0.44; 3.3–5.2; 21)<br />
SEABIRD 21 (2008): 1–15<br />
9
10<br />
Identification of giant petrels<br />
Table 4. Measurements (mm) and body mass (kg) of male and female Northern Giant Petrels Macronectes<br />
halli from several breeding places. Data are: mean, (± SD; range; sample size). (1) South Georgia (González-<br />
Solís et al. 2000b, González-Solís 2004); (2) Crozet Islands (J-FV, unpubl. data); (3) Macquarie Island (G. W.<br />
Johnstone, pers. comm.); (4) Chatham Islands (G. W. Johnstone & C. J. Robertson, pers. comm.).<br />
Males Females<br />
Wing (1) 526 (507–550; 71) 500.6 (479–518; 77)<br />
(2) 519.7 (2.86; 503–535; 14) 474 (8.76; 435–502; 8)<br />
(3) 532.7 (2.4; 509–545; 17) 497.2 (1.8; 483–513; 22)<br />
(4) 525.9 (1.67; 506–538; 20) 496 (2.11; 475–515; 21)<br />
Culmen (1) 103.7 (98.4–109.4; 71) 89.3 (84.3–94.4; 77)<br />
(2) 103.3 (0.82; 96–110; 17) 88.3 (1.06; 84–92; 8)<br />
(3) 103.9 (0.8; 95.1–109.7; 18) 89.8 (0.37; 86.3–92.4; 23)<br />
(4) 97.8 (0.63; 93.9–104.7; 20) 85.4 (0.42; 81.8–89.5; 21)<br />
Tarsus (1) 100.8 (95.5–106.9; 71) 90.7 (86.7–94.7; 77)<br />
(2) 105.7 (2.24; 91–123; 13) 90.8 (1.45; 84–95.5; 8)<br />
(3) 102.27 (0.71; 94.5–107.1; 18) 92 (0.39; 88.1–95.6; 23)<br />
(4) 94.4 (0.51; 89.1–98.5; 20) 83.4 (0.49; 80.8–89; 21)<br />
Mass (1) 4.6 (0.2; 33) 3.7 (0.2; 35)<br />
(2) 4.47 (0.42; 3.6–5.35; 14) 4.31 (0.62; 3.3–5.35; 7)<br />
(3) 4.49 (0.37; 3.85–5.4; 16) 3.38 (0.24; 3–3.85; 22)<br />
(4) 3.66 (0.32; 3.15–4.45; 19) 2.83 (0.17; 2.5–3.25; 21)<br />
more mottled and then gradually whiter, while the crown, nape and back of the neck<br />
remain mottled brown (Table 2, Figure 8). With age, the pale grey or whitish colour of<br />
the foreneck then extends progressively down the whole chest, belly and undertail<br />
coverts. At the same time, the upperparts become greyish-brown (but never as pale as<br />
in old M. giganteus). In this way, old birds acquire a dark-above, pale-below appearance,<br />
quite different from that of M. giganteus.<br />
In both species, it is possible that the plumage of males pales more rapidly with age<br />
than that of females. This has never been examined systematically, but one published<br />
photograph (Anonymous 2002) shows a female M. halli ringed as an adult in South<br />
Georgia in 1963 and recovered in the Falkland Islands in 2002, and therefore over 40<br />
years old, but still in stage H6 plumage.<br />
Sexual differences: No differences in plumage have yet been demonstrated between<br />
male and female giant petrels, except for the possibility (above) that male Northern<br />
Giant Petrels become paler sooner with age than females. On the other hand, there is a<br />
marked sexual dimorphism in size in both species, males being much larger and about<br />
20% heavier than females (Hunter 1983; González-Solís et al. 2000b; González-Solís<br />
2004; J-FV; Tables 3 & 4), and size differences may help distinguish the sexes of a pair at<br />
the nest or displaying. However, one should be extremely cautious in using body mass<br />
to sex birds, as it may change considerably during the year, with average variations of<br />
over 20% having been recorded for each sex in both species on the Crozet Islands (J-FV).<br />
SEABIRD 21 (2008): 1–15
Identification of giant petrels<br />
Biometrics: Both giant petrel species are of similar size and mass where they breed<br />
sympatrically. Large samples measured at South Georgia (Hunter 1984), and more<br />
limited data from Crozet and Macquarie Islands (Tables 3 & 4) both suggest that M.<br />
giganteus is, on average, larger than M. halli although there is much overlap and the<br />
characteristics cannot be used in most individuals to confirm the species. Northern and<br />
Southern Giant Petrels breeding at localities where the other species does not occur<br />
(the Argentinean coast, Falkland and Gough Islands, and Terre Adélie for M. giganteus;<br />
Chatham Islands for M. halli) are smaller in size, and sometimes significantly smaller<br />
than their counterparts from other sub-Antarctic islands (Voisin & Bester 1981; Voisin<br />
1982b; Carlos et al. 2005; Tables 3 & 4). <strong>The</strong>se differences may be slight but, for<br />
instance, a nominate M. giganteus and/or a M. halli seen amongst M. g. solanderi is<br />
easily detected, as it is clearly larger (Figure 9).<br />
Voice: <strong>The</strong> vocalisations of the two species are very similar, and consist of croaking,<br />
hissing sounds uttered during quarrelling and displays, as well as a kind of neighing, often<br />
emitted when the birds fly low over the observer’s head. <strong>The</strong> calls of M. giganteus are<br />
higher-pitched and uttered in a faster rhythm than those of M. halli, a difference best<br />
noticed when birds fly overhead (Voisin 1968, 1978; Shirihai & Jarrett 2002; C. Chappuis<br />
pers. comm.). At sea, both species are usually silent, except when quarrelling over food.<br />
Nest location: Giant petrels exhibit a varying degree of coloniality, from lone nests and<br />
small groups to large loose or tight colonies. However, lone nests or small groups may<br />
have become isolated because reproduction failed at other nests around them. M.<br />
giganteus reputedly nests more colonially than M. halli, but the difference is mainly a<br />
matter of statistics, although tight colonies, recalling gull colonies, seem to be<br />
encountered regularly in the South Atlantic Giant Petrel (Voisin 1982b). M. giganteus<br />
generally breeds in more open situations than M. halli, but the overlap is large and of<br />
little use in field identification, except for the few nests that are more or less hidden in<br />
cavities, which are M. halli (Voisin 1976, 1986). However, the South Atlantic Giant Petrel<br />
at Gough Island may breed in very sheltered sites, with the nest sometimes hidden<br />
completely among tussock grass clusters and tree ferns (Johnstone et al. 1976;<br />
Shaughnessy & Fairall 1976; Voisin & Bester 1981).<br />
Breeding periods: Where both species breed together, M. halli commences breeding<br />
about six weeks earlier than M. giganteus (Table 5). In South Georgia, where the climate<br />
is harsher, both species breed about six weeks later than at other sympatric locations<br />
(Hunter 1984, 1987). Similarly, M. giganteus lays from late October to early November<br />
in Terre Adélie (Mougin 1975). In contrast, South Atlantic Giant Petrels from Gough<br />
Island lay between late August and early September (Voisin & Bester 1981), whereas<br />
in the Falklands (Woods 1975) and Argentina (Humphrey & Livezey 1983; Quintana et<br />
al. 2005) they lay in late October.<br />
Dead specimens: Museum specimens can be identified by plumage if they are of<br />
‘old’ breeders. <strong>The</strong> distinctive coloration of giant petrels’ bills fades rapidly after<br />
death if the birds are not kept in good conditions. This is especially so for birds that<br />
have been kept for a certain time in fluids, some of which (e.g. formalin) can be very<br />
SEABIRD 21 (2008): 1–15<br />
11
12<br />
Identification of giant petrels<br />
detrimental in this respect. Submergence in seawater and/or other fluids can result<br />
in an overall olive-green ground colour, very different from the reddish hue found in<br />
live M. halli, or the bluish-green in M. giganteus. This is often the case in museum<br />
specimens, and while enough of the original bill colours may remain to allow, or<br />
assist, specific identification in some specimens, others remain unidentified (Bourne<br />
& Warham 1966). Decomposition and exposure to sunlight can quickly alter bill<br />
colours in dead giant petrels, especially at low latitudes. In the case of freshly dead,<br />
stranded specimens, it is advisable to look at the side of the bill lying on the ground,<br />
which may have kept its coloration better (Figure 10). Post-mortem alteration is<br />
certainly the reason for some puzzling bill colour descriptions found on museum<br />
labels. Moreover, describing colours can be rather subjective, and such descriptions<br />
should be interpreted with caution.<br />
Figure 10. Freshly dead, stranded Northern Macronectes halli (left) and Southern Giant Petrels M. giganteus<br />
(right), State of Rio Grande do Sul in south Brazil, 2003 © Fernanda I. Colabuono. <strong>The</strong> side of the bill lying on the<br />
ground kept enough of the original colour to allow specific identification.<br />
SEABIRD 21 (2008): 1–15<br />
Conclusion<br />
Giant petrels have a reputation for being<br />
difficult to identify, but fortunately this is<br />
only partly true. If seen at close range (e.g. at<br />
sympatric breeding sites), even chicks and<br />
juveniles may be specifically identified using<br />
bill coloration alone, quite apart from<br />
plumage. With the exception of white form<br />
M. giganteus, bill colour, if seen under good<br />
light conditions, can still be the diagnostic<br />
Figure 9. A Northern Macronectes halli<br />
(background, plumage type H8) and two South<br />
Atlantic Giant Petrels M.g. solanderi (plumage<br />
type G1), Abattoir Outlet, Falkland Islands, 2006<br />
© Steve Copsey.
Identification of giant petrels<br />
Table 5. Laying dates of Northern Macronectes halli and Southern M. giganteus Giant Petrels at locations<br />
where the two species breed sympatrically.<br />
Species Locality Range of laying dates Reference<br />
M. halli South Georgia 19 September–10 October Hunter (1984, 1987)<br />
Marion Island 4 August–1 September Burger (1978);<br />
Cooper et al. (2001)<br />
Crozet Islands 16 August–5 September Voisin (1968)<br />
Macquarie Island 11 August–6 September Johnstone (1978)<br />
M. giganteus South Georgia 30 October–24 November Hunter (1984, 1987)<br />
Marion Island 12 September–25 October Burger (1978);<br />
Cooper et al. (2001)<br />
Crozet Islands 26 September–17 October Voisin (1968)<br />
Macquarie Island 27 September–19 October Johnstone (1978)<br />
criterion for birds observed far away in the field or at sea, but the observer has mostly<br />
to rely on plumage characteristics of ‘old’ adults (i.e. plumage types H6–H8 for M. halli<br />
and G6–G8 for M. giganteus; see Figures 7 & 8), which means that a number of birds<br />
may remain unidentified.<br />
Acknowledgements<br />
We are grateful to Steve Copsey, Fernanda I. Colabuono and Ronald Saldino for<br />
allowing us to use their photographs. Renata F. Cunha kindly prepared Figures 7 &<br />
8. CJC received support from Fundação de Aperfeiçoamento de Pessoal de Nível<br />
Superior (CAPES), Brazil. <strong>The</strong> typescript benefited from the input of Bernie<br />
Zonfrillo, Stephen Hunter, Bill Fraser, an anonymous referee and the editorial<br />
advice of Martin Heubeck.<br />
References<br />
Anonymous, 2002. <strong>The</strong> oldest stinker in town? Falkland Conservation Newsletter (Sep. 2002): 4.<br />
BirdLife International, 2004. Tracking ocean wanderers: the global distribution of albatrosses<br />
and petrels. Results from the Global Procellariiform Tracking Workshop; 2003 September 1-5;<br />
Gordon’s Bay, South Africa. BirdLife International, Cambridge, UK.<br />
Bourne, W. R. P. & Warham, J. 1966. Geographical variation in the giant petrels of the genus<br />
Macaronectes. Ardea 54: 45–67.<br />
Burger, A. E. 1978. Interspecific breeding attempts by Macronectes giganteus and M. halli. Emu<br />
78: 234–235.<br />
Carboneras, C. 1992. Family Anatidae (Ducks, Geese and Swans). In: del Hoyo J., Elliott A. &<br />
Sargatal J. (eds.), Handbook of the Birds of the World. Vol. 1: 536–628. Lynx Edicions, Barcelona.<br />
Carlos, C. J., Voisin, J.-F. & Vooren, C. M. 2005. Records of Southern Giant Petrel Macronectes<br />
giganteus solanderi and Northern Giant Petrel M. halli off southern Brazil. Bulletin of the British<br />
Ornithologists Club 125: 288–292.<br />
Conroy, J. W. H. 1972. Ecological aspects of the biology of the Giant Petrel Macronectes<br />
giganteus (Gmelin) in the Maritime Antarctic. British Antarctic Survey Scientific Report 75: 1–74.<br />
Conroy, J. W. H., Bruce, G. & Furse, J. R. 1975. A guide to the plumage and iris colours of the<br />
giant petrels. Ardea 63: 87–92.<br />
SEABIRD 21 (2008): 1–15<br />
13
14<br />
Identification of giant petrels<br />
Cooper, J., Brooke, M. L., Burger, A., Crawford, R. J. M., Hunter, S. & Williams, T. A. J. 2001.<br />
Aspects of the breeding biology of the Northern Giant Petrel (Macronectes halli) and the<br />
Southern Giant Petrel (Macronectes giganteus) at sub-Antarctic Marion Island. International<br />
Journal of Ornithology 4: 53–68.<br />
Copello, S., Quintana, F. & Somoza, G. 2006. Sex determination and sexual size-dimorphism in<br />
Southern Giant Petrels (Macronectes giganteus) from Patagonia, Argentina. Emu 106: 141–146.<br />
González-Solís, J. 2004. Sexual size dimorphism in Northern Giant Petrels: ecological correlates<br />
and scaling. Oikos 105: 247–254.<br />
González-Solís, J., Croxall, J. P. & Wood A. G. 2000a. Foraging partitioning between giant<br />
petrels Macronectes spp. and its relationship with breeding population changes at Bird Island,<br />
South Georgia. Marine Ecology Progress Series 204: 279–288.<br />
González-Solís, J., Croxall, J. P. & Wood, A. G. 2000b. Sexual dimorphism and sexual<br />
segregation in foraging strategies of Northern Giant Petrels, Macronectes halli, during<br />
incubation. Oikos 90: 390–398.<br />
González-Solís, J., Croxall, J. P. & Afanasyev, V. 2008. Offshore spatial segregation in giant<br />
petrels Macronectes spp.: differences between species, sexes and seasons. Aquatic<br />
Conservation: Marine and Freshwater Ecosystems 17: S22–S36.<br />
Humphrey, P. S. & Livezey, B. C. 1983. Giant Petrels (Macronectes giganteus) nesting in Chubut,<br />
Argentina. Gerfaut 73: 3–8.<br />
Hunter, S. 1982. Interspecific breeding in giant petrels at South Georgia. Emu 82 (suppl.): 312–314.<br />
Hunter, S. 1983. Identification of giant petrels Macronectes spp. Sea Swallow 32: 72–77.<br />
Hunter, S. 1984. Breeding biology and population dynamics of giant petrels Macronectes at<br />
South Georgia (Aves: Procellariiformes). Journal of Zoology, London 203: 441–460.<br />
Hunter, S. 1987. Species and sexual isolating mechanisms in sibling species of giant petrels<br />
Macronectes. Polar Biology 7: 295–301.<br />
Jiguet, F. 2000. <strong>The</strong> two giant petrels. Birding World 13: 108–116.<br />
Johnstone, G. W. 1971. Bird in the hand–giant petrels. <strong>The</strong> Australian Bird Bander 9: 86–87.<br />
Johnstone, G. W. 1974. Field characteristics and behaviour at sea of giant petrels in relation to<br />
their oceanic distribution. Emu 74: 209–218.<br />
Johnstone, G. W. 1978. Interbreeding by Macronectes halli and M. giganteus at Macquarie<br />
Island. Emu 78: 235.<br />
Johnstone, G. W. Shaughnessy, P. D. & Conroy, J. H. W. 1976. Giant petrels in the South<br />
Atlantic: new data from Gough Island. South African Journal of Antarctic Research 6: 19–22.<br />
Mougin, J.-L. 1975. Ecologie comparée des Procellariidae antarctiques et subantarctiques.<br />
Comité National Français dês Recherches Antarctiques 36: 1–195.<br />
Obst, B. S. & Nagy, K. A. 1992. Field energy expenditures of the Southern Giant Petrel. Condor<br />
94: 801–810.<br />
Parmelee, D. F. 1992. Antarctic Birds: Ecological and Behavioral Approaches. University of<br />
Minnesota Press, Minneapolis.<br />
Patterson, D. L. & Hunter, S. 2000. Giant petrel Macronectes spp. band recovery analysis from<br />
the International Giant Petrel Banding Project, 1988/89. Marine Ornithology 28: 69–74.<br />
Patterson, D. L. & Fraser, W. 2000. Foraging movements of Southern Giant Petrels on the<br />
Antarctic Peninsula: preliminary findings of a satellite tracking study during the breeding<br />
season. In: Flint, E. & Swift, K. (eds.) 2nd International Conference on the Biology and<br />
Conservation of Albatrosses and Petrels, Honolulu, Hawaii, USA, 8-12 May 2000. Marine<br />
Ornithology 28: 125–152.<br />
SEABIRD 21 (2008): 1–15
Identification of giant petrels<br />
Patterson, D. L. & Fraser, W. 2003. Satellite tracking Southern Giant Petrels at Palmer Station,<br />
Antarctica. Feature Articles, Microwave Telemetry, Inc. 8: 3–4.<br />
Penhallurick, J. & Wink, M. 2004. Analysis of the taxonomy and nomenclature of the<br />
Procellariiformes based on complete nucleotide sequences of the mitochondrial cytochrome b<br />
gene. Emu 104: 125–147.<br />
Pennycuick, C. J. 1982. <strong>The</strong> flight of petrels and albatrosses (Procellariiformes) observed in<br />
South Georgia and its vicinity. Philosophical Transactions of the Royal Society of London Series B<br />
300: 75–106.<br />
Quintana, F., Schiavini A. & Copello, S. 2005. Estado poblacional, ecología y conservación del<br />
Petrel Gigante del Sur (Macronectes giganteus) en Argentina. Hornero 20: 25–34.<br />
Salomon, M., Voisin, J.-F. & Bried, J. 2003. On the taxonomic status and denomination of the<br />
Iberian Chiffchaff. Ibis 145: 87–97.<br />
Shaughnessy, P. D. 1970. <strong>The</strong> genetics of plumage phase dimorphism of the Southern Giant<br />
Petrel Macronectes giganteus. Heredity 25: 501–506.<br />
Shaughnessy, P. D. 1971. Frequency of the white phase of the Southern Giant Petrel,<br />
Macronectes giganteus. Australian Journal of Zoology 19: 77–83.<br />
Shaughnessy, P. D. & Conroy, J.W. D. 1977. Further data on the inheritance of plumage phases<br />
of the Southern Giant Petrel Macronectes giganteus. British Antarctic Survey Bulletin 45: 25–28.<br />
Shaughnessy, P. D. & Fairall, N. 1976. Notes on seabirds at Gough Island. South African Journal<br />
of Antarctic Research 6: 23–25.<br />
Shirihai, H. & Jarrett, B. 2002. <strong>The</strong> Complete Guide to Antarctic Wildlife: Birds and Marine<br />
Mammals of the Antarctic Continent and the Southern Ocean. Princeton University Press,<br />
Princeton and Oxford.<br />
Trivelpiece, S. G. & Trivelpiece, W. Z. 1998. Postfledging dispersal of Southern Giant Petrels<br />
Macronectes giganteus banded at Admiralty Bay, King George Island, Antarctica. Marine<br />
Ornithology 26: 63–68.<br />
Voisin, J.-F. 1968. Les pétrels géants (Macronectes halli et M. giganteus) de I’lle de la Possession.<br />
L’Oiseau et la Revue Française d’Ornithologie (Special No.) 38: 95–122.<br />
Voisin, J.-F. 1976. Observations sur les pétrels géants de l’île aux Cochons (archipel Crozet).<br />
Alauda 44: 411–429.<br />
Voisin, J.-F. 1982a. Sur la détermination des pétrels géants Macronectes giganteus et M. halli de<br />
l’Océan indien sud. Info Nature, Ile de la Réunion 19: 157–165.<br />
Voisin, J.-F. 1982b. Observations on the Falkland Islands Giant Petrels Macronectes giganteus<br />
solanderi. Gerfaut 72: 367–380.<br />
Voisin, J.-F. 1988. Breeding biology of the Northern Giant Petrel Macronectes halli and the<br />
Southern Giant Petrel M. giganteus at île de la Possession, îles Crozet, 1966–1980. Cormorant<br />
16: 65–97.<br />
Voisin, J.-F. 1990. Movements of giant petrels Macronectes spp. banded as chicks at Iles Crozet<br />
and Kerguelen. Marine Ornithology 18: 27–36.<br />
Voisin, J.-F. & Bester, M. N. 1981. <strong>The</strong> specific status of giant petrels Macronectes at Gough<br />
Island. In: Cooper, J. (ed.) Proceedings of the Symposium on Birds of the Sea and Shore: 215–222.<br />
African <strong>Seabird</strong> <strong>Group</strong>, Cape Town.<br />
Voisin, J.-F. & Teixeira, D. M. 1998. <strong>The</strong> identification of giant petrels (Aves, Procellariidae) in the<br />
South Atlantic. Boletim da Fundação Brasileira para a Conservação da Natureza 25: 129–134.<br />
Warham, J. 1962. <strong>The</strong> biology of the Giant Petrel Macronectes giganteus. Auk 79: 139–160.<br />
Woods, R. W. 1975. <strong>The</strong> Birds of Falkland Islands. Anthony Nelson, Oswestry, England.<br />
SEABIRD 21 (2008): 1–15<br />
15
16<br />
Vagrancy of Brünnich’s Guillemot<br />
Vagrancy of Brünnich’s Guillemot Uria<br />
lomvia in Europe<br />
Van Bemmelen, R. 1 *, & Wielstra, B. 2<br />
*Correspondence author. Email: rvanbemmelen@gmail.nl<br />
1 Stavangerweg 535, 1013 AX Amsterdam, <strong>The</strong> Netherlands; 2 Dr. Benthemstraat 10–91,<br />
7514 CM Enschede, <strong>The</strong> Netherlands.<br />
Abstract<br />
We review the occurrence of vagrant Brünnich’s Guillemots Uria lomvia in Europe. <strong>The</strong><br />
104 records of 109 individual birds that could be traced showed a distinct seasonal<br />
pattern. <strong>The</strong>re were no September records, but a small autumn peak was apparent in<br />
late October and early November. Numbers increased again in early December, peaked<br />
in late January and early February, and declined through spring, with only seven<br />
records in summer. Autumn records were mostly of first-winter birds, whereas<br />
relatively more adults were recorded in winter, in line with expectations based on<br />
timing of migration for these different age classes. We speculate that vagrant birds to<br />
western Europe have strayed from the wintering grounds and migration routes south<br />
of Iceland and along the Norwegian coast, while overland movements from the<br />
Barents Sea may explain inland records from northern Scandinavia, and some from the<br />
Baltic Sea. Two spatial clusters of records were evident, one in Scotland and the other<br />
in the Skagerrak, Kattegat and southern Baltic Sea. Between 1975/76 and 2005/06, the<br />
number of records declined in the former region but increased in the latter, which may<br />
represent a real decrease in occurrence, and increased observer effort, respectively.<br />
Introduction<br />
Brünnich’s Guillemot Uria lomvia is a circumpolar species, breeding in both the high<br />
and low (sub) Arctic (Figure 1a) (Nettleship & Evans 1985). <strong>The</strong> Atlantic population<br />
numbers c. 6.5 million birds, breeding at colonies in eastern Canada, west and east<br />
Greenland, Iceland, Jan Mayen, Svalbard, northern Norway, and the western Russian<br />
Arctic (CAFF 2004). Ringing recoveries have shown that the breeding populations of<br />
Canada and Greenland winter in the northwest Atlantic, where they are joined by a<br />
substantial (but unknown) number of birds from colonies in Iceland, Svalbard, Norway<br />
and Russia, although many of the latter remain in the Norwegian and Barents Seas<br />
throughout the year (Kampp 1988; Nikolaeva et al. 1996; CAFF 2004; Bakken &<br />
Mehlum 2005). Thus, a large number of Brünnich’s Guillemots migrate in a<br />
southwesterly direction across the North Atlantic each autumn. Occasionally,<br />
Brünnich’s Guillemots are recorded further south in Europe than usual (Figure 1b). We<br />
review the occurrence of such vagrants, analyse their temporal and geographic distribution,<br />
and examine differences between age classes. We attempt to relate the<br />
patterns found to the normal seasonal distribution of the species in the Atlantic and<br />
speculate on possible origins and causes of these extra-limital occurrences. Finally, we<br />
discuss whether Brünnich’s Guillemots found in western Europe are truly lost, or<br />
whether they winter there in small numbers.<br />
SEABIRD 21 (2008): 16–31
Vagrancy of Brünnich’s Guillemot<br />
Figure 1. <strong>The</strong> normal range of Brünnich’s Guillemot Uria lomvia in the North Atlantic (A), and the distribution of records<br />
of vagrants in Europe from 1900–2006 (B). Breeding areas are after Kampp (1988), boundaries of winter distribution are<br />
after Cramp (1985) and Gaston & Jones (1998), and extreme distribution limits are drawn from Gaston & Jones (1998).<br />
SEABIRD 21 (2008): 16–31<br />
17
18<br />
Vagrancy of Brünnich’s Guillemot<br />
Methods<br />
A database was compiled of records of Brünnich’s Guillemot in Europe that have been<br />
accepted by national rarity committees, for which location (at least province) and date (at<br />
least month) were known. A cut-off date of 31 August 2006 was used. <strong>The</strong> species is<br />
considered by all European national rarities committees, except those of Iceland, Norway<br />
and Russia, which were therefore left out of the analysis. <strong>The</strong> German rarities committee<br />
has not yet examined records of Brünnich’s Guillemot, so we included only those German<br />
records for which museum specimens were available.We took a neutral stand on decisions<br />
made by rarities committees, but tried to gather as much additional detail as possible on<br />
age, sex and plumage by studying photos, drawings, and museum specimens.<br />
Records were not analysed by calendar year, but for intervals from September to<br />
August, hereafter referred to as years. Seasons were classed as autumn (September to<br />
November), winter (December to February), spring (March to May) and summer (June<br />
to August). Months were divided into three periods of approximately 10 days: early<br />
(1–10), mid (11–20) and late (21–28/30/31). Fisher’s Exact Tests were performed to<br />
test whether years where a season had records had been preceded more frequently by<br />
a season holding records compared to years in which the season of interest did not<br />
have records. A potential shift over time in the proportion of dead versus live birds was<br />
tested for using a binomial linear regression analysis. <strong>The</strong> software package R was used<br />
for statistical analyses (R Development Core Team 2005).<br />
Results<br />
We traced 104 records of European vagrants, distributed as follows: UK and Ireland<br />
(39), the southern North Sea and English Channel (16), the Skagerrak, Kattegat and<br />
Baltic Sea (41), Lapland (6), and inland Western Europe (2) (Figure 1, Appendix 1). Five<br />
records involved two birds together, making a total of 109 individuals. Brünnich’s<br />
Guillemot has only been recorded regularly since the mid 1950s, with just nine coastal<br />
records (ten individuals) and one inland record (two individuals) prior to 1950/51.<br />
During 1954/1955–2005/2006, an average of 1.9 individuals were recorded in Europe<br />
per year, with records in 41 (79%) of the 52 years (Figure 2). Since 1975, the number<br />
of individuals increased in the Skagerrak/Kattegat/Baltic Sea area (t [1.4] = 3.624, P =<br />
0.02), but declined in the UK/Ireland (t [1.4] = -3.292, P = 0.03) (Figure 3). A negative<br />
trend was also suggested for the southern North Sea and the English Channel, but this<br />
was not statistically significant (t [1.4] = -1.207, P = 0.294).<br />
<strong>The</strong> occurrence of Brünnich’s Guillemot showed a distinct seasonal pattern, with 19%<br />
of records in autumn, 56% in winter, 18% in spring, and 7% in summer (Figure 4).<br />
<strong>The</strong>re were no records from mid August to late September.<br />
Data for 1954/55–2005/06 showed that for years in which Brünnich’s Guillemots<br />
were observed in a particular season, the preceding season did not hold records more<br />
frequently than those years that did not hold records in that season (Fisher’s Exact Test<br />
for: summer and the preceding spring (P = 0.589); summer and the preceding winter<br />
(P = 1.000); spring and the preceding winter (P = 0.237); and winter and the preceding<br />
autumn (P = 0.703, n = 52 for all comparisons)).<br />
SEABIRD 21 (2008): 16–31
Vagrancy of Brünnich’s Guillemot<br />
Figure 2. <strong>The</strong> number of records of vagrant Brünnich’s Guillemots Uria lomvia per year in Europe during<br />
1950/1951–2005/2006.<br />
Of the 109 individual Brünnich’s Guillemots, 43 (39%) were found dead. <strong>The</strong> remaining<br />
66 (61%) were alive when discovered, but at least nine of these died soon after<br />
(including three in Sweden and one in Denmark that were shot). Seven birds were<br />
reported as oiled, five of which were dead when found while the other two died soon<br />
after being found. From 1975 onwards, the proportion of dead birds relative to live birds<br />
declined from 80% in 1975/76–1979/80 to 17% in 2000/01–2004/05 (z = 4.224, 4 df,<br />
P < 0.01). However, this did not hold for all regions. Whereas the proportion of dead<br />
individuals declined in UK and Ireland (z = 2.521, P = 0.01, 3 df), it remained stable in<br />
the Skagerrak/Kattegat/Baltic Sea area (z = 1.081, P = 0.279, 3 df) (Figure 3). Eight<br />
records have been accepted of birds seen only in flight, seven from Sweden and one from<br />
Denmark, occurring in October, November (2), January, February (2), April and May.<br />
Only 38 birds were reportedly aged, 25 as adult, one as a first-summer, and 12 as firstwinter.<br />
Three of the adults were found in autumn, ten were in winter, eight in spring<br />
and four in summer.<strong>The</strong> first-summer bird was in July, while six of the first-winter birds<br />
were in autumn and six were in winter. Apart from the 12 aged as first-winter (which<br />
by definition show a winter-plumaged head pattern), plumage details were noted for<br />
61 birds. Fourteen were in breeding plumage, in October, January (2), March, April (2),<br />
May (2), June (2), July (2), and August (2). <strong>The</strong>se were all reported as adults, except for<br />
four Swedish records from April (2), May and June, and a British record for June, for<br />
which no age was given. <strong>The</strong> 47 birds noted as being in non-breeding plumage were<br />
recorded in all months from October until April. Eleven of these were aged as adults,<br />
in October, November, December (2), January (3), February, March (2), and April. No<br />
details on age were provided for the remaining 36, recorded in October (3), November<br />
(7), December (10), January (6), February (9), and March. (Figure 4, Appendix 1).<br />
Discussion<br />
Seasonal occurrence: Records of vagrant Brünnich’s Guillemots in Europe show a<br />
distinct seasonal pattern, and our results corroborated those from an earlier review of<br />
records from northwest Europe spanning the period 1963–1992 (Rønnest 1994).<strong>The</strong>re<br />
SEABIRD 21 (2008): 16–31<br />
19
20<br />
Vagrancy of Brünnich’s Guillemot<br />
Figure 3. <strong>The</strong> number of live (upper bar parts) and dead (lower bar parts) vagrant Brünnich’s Guillemot Uria lomvia<br />
individuals per 5-year period for three areas: the UK and Ireland, the Skagerrak, Kattegat and Baltic Sea, and the<br />
southern North Sea and English Channel.<br />
Figure 4. <strong>The</strong> seasonal occurrence of vagrant Brünnich’s Guillemot Uria lomvia individuals in Europe per 10-day<br />
period (n=105), divided into age and plumage category. Three Swedish records of four individuals could not be<br />
assigned to 10-day periods (two unaged birds in non-breeding plumage in January 1925 and another in February<br />
1962, and one with no plumage details in May 1955), and are therefore not shown.<br />
were no September records, but a small autumn peak from mid October to mid<br />
November coincides with the onset of migration of birds from Barents Sea colonies<br />
(Gaston & Jones 1998; Bakken & Mehlum 2005). Records increased again from early<br />
December to a peak in late January and early February, perhaps reflecting prolonged<br />
migration of Brünnich’s Guillemots, with many still arriving in wintering areas in<br />
December and January (Gaston & Jones 1998; Kampp 1988; Nikolaeva et al. 1996;<br />
SEABIRD 21 (2008): 16–31
Vagrancy of Brünnich’s Guillemot<br />
Bakken & Mehlum 2005). Some individuals found in winter may have arrived in the<br />
area in autumn but taken time to weaken, die and beach, but our data suggested that<br />
winters with records were not more likely to follow autumns with records, and it seems<br />
more probable that the winter peak was mainly of newly arrived birds.<br />
Brünnich’s Guillemots arrive back at their colonies in March–May, depending on latitude<br />
and ice conditions (Cramp 1985; Isaksen & Bakken 1995; Gaston & Jones 1998). <strong>The</strong><br />
declining number of records from mid February, with only a few from May to August,<br />
suggests most vagrants had either returned to their normal range or died by then.<br />
Whether summer records refer to new arrivals from the north, or to lingering but<br />
previously undetected birds is unclear, but summers with records did not more often<br />
follow springs or winters with records. Some summer vagrants were clearly very disorientated,<br />
such as the adults found inland in Germany in August 1987, and in Belgium in<br />
August 2006, the only inland records apart from those in northern Scandinavia. In<br />
contrast, there are two UK records of Brünnich’s Guillemots near Common Guillemot<br />
Uria aalge colonies (Farne Islands, July 1977; St Kilda, May/June 1992), and one of a bird<br />
present in a Common Guillemot colony for nearly a month (Shetland, June/July 1989).<br />
Seasonal distribution of age classes: First-winter birds tend to arrive at wintering<br />
areas in the northwest Atlantic in October and November, whereas adults mainly<br />
arrive during or after December (Gaston & Jones 1998; Bakken & Mehlum 2005). <strong>The</strong><br />
later migration of adults (or its slower progression) might result in a higher proportion<br />
of first-winter birds among autumn records, and a higher proportion of adults in<br />
winter. Indeed, of the nine aged records in the small autumn peak of European<br />
vagrants, six were identified as first-winters and three as adults (Figure 2).<br />
Furthermore, the three birds in mid October (and possibly the three in early<br />
November) in non-breeding plumage that were not aged may also have been firstwinters,<br />
since adults are probably still in transitional moult at that time (Gaston &<br />
Jones 1998), but note the two records of adults in non-breeding plumage in late<br />
October and early November (Figure 4). Of the 16 aged individuals in winter, ten were<br />
reported as adults compared to six as first-winters.<br />
Ageing Brünnich’s Guillemots in the field is difficult, especially in winter (Cramp 1985;<br />
Blomdahl et al. 2003). Only nine of the live birds were reportedly aged in the field<br />
(eight adults and one first-winter). Separating first-winter Brünnich’s Guillemots from<br />
older birds in winter is not particularly difficult in the hand (Gaston 1984; Gaston &<br />
Jones 1998), so although criteria used for ageing corpses were not documented for<br />
most records, ages given for museum specimens and weakened birds that were caught<br />
were probably correct. Separating immature birds (non-adults after their first winter)<br />
from adults is much harder (Cramp 1985), and the number of immatures among the<br />
records listed may have been underestimated. Since immature birds generally do not<br />
return to the breeding areas (Cramp 1985; Bakken & Mehlum 2005), summer records<br />
may include more immatures than is realised. <strong>The</strong> only record aged as immature was<br />
in July (1978, St. Cyrus, Scotland), and among the four summer records of adults, Grant<br />
(1981) suggested that the live bird at the Farne Islands in July 1977 (Ribbands 1977)<br />
may have been a first-summer bird rather than an adult.<br />
SEABIRD 21 (2008): 16–31<br />
21
22<br />
Vagrancy of Brünnich’s Guillemot<br />
Spatial distribution, origin, and possible causes of vagrancy: <strong>The</strong>re are two obvious<br />
geographic clusters in the records of vagrant Brünnich’s Guillemot in Europe. Of the 82<br />
records during 1975/76 to 2004/05, 34 (41%) were from the UK and Ireland, mostly<br />
in northern Scotland, and 35 (43%) were from the Skagerrak, Kattegat and the<br />
entrance to the Baltic Sea (Figure 1b). <strong>The</strong> cluster of records in Scotland is not<br />
surprising, since Scotland is closest to the species’ normal migratory and wintering<br />
range around Iceland and the coast of northern Norway (Cramp 1985; Kampp 1988;<br />
Nikolaeva et al. 1996; Bakken & Mehlum 2005; Fraser et al. 2007). Brünnich’s<br />
Guillemots also stray further south along the Norwegian coast than their normal range<br />
(Engebretsen & Pettersen 2001), and such birds are probably the source of many of<br />
those recorded in the eastern Skagerrak and the Kattegat.<br />
<strong>The</strong> six inland records in northern Sweden and northern Finland, five in winter and one in<br />
spring, suggest that at least some birds recorded in the Baltic Sea may have flown<br />
overland across Lapland. That Brünnich’s Guillemots can move inland from the Barents<br />
Sea was illustrated by a large influx covering much of Finland from mid November 1902<br />
to mid January 1903 (Mela 1903).<strong>The</strong>se records were not dealt with by the Finnish rarities<br />
committee so were not included in our analysis, but can be seen at<br />
http://koti.netplaza.fi/~pply/havikset/lajisto/support/urilom_1902.htm. In the same<br />
winter, two birds were shot inland in Norrbotten, northern Sweden. Mela (1903)<br />
attributed the invasion to an early onset of severe weather conditions in the Barents and<br />
White Seas. Such an influx, which must have involved many hundreds of birds, has never<br />
recurred, and the most recent inland records from the area were of single birds in northern<br />
Finland in December 1986 and northern Sweden in January 1987. In contrast, inland<br />
movements of Brünnich’s Guillemots in the Great Lakes region of Canada and northeast<br />
USA, mainly during 1890–1910, were thought to have been related to changes in prey<br />
abundance (Gaston 1988). <strong>The</strong> 1902/03 Finnish influx also involved Atlantic Puffins<br />
Fratercula arctica, but no other alcids were recorded in the Great Lakes movements.<br />
TEXTBOX: Little Auk invasions and Brünnich’s Guillemot vagrancy<br />
Four recent autumn reports of Brünnich’s Guillemot around the North Sea in 2005–2007, which have<br />
not yet been accepted by rarity committees, coincided with (or just preceded) large coastal<br />
movements (invasions) of Little Auks Alle alle, and it is tempting to speculate whether the two<br />
phenomena were related. Breeding and wintering areas of both species overlap to a large extent<br />
(Cramp 1985) and Camphuysen & Leopold (1996) suggested that Little Auks wintering in the North<br />
Sea originate from the Barents Sea, as we do for Brünnich’s Guillemots that turn up in the North Sea.<br />
Invasions of Little Auks generally coincide with their arrival on the wintering grounds in late autumn,<br />
and Camphuysen & Leopold (1996) suggested that such movements are triggered by local food<br />
shortages, with wind conditions playing only a secondary role in making the displacement visible to<br />
sea-watchers. Gaston (1988) noted a lack of synchrony between invasions of the two species in<br />
Canada and the northeast USA and attributed this to differences in diet. However, there may be<br />
some dietary overlap in winter in large zooplankton and small pelagic fish (Blake 1983; Skov et al.<br />
1989; Moody & Hobson 2007; Stempniewicz 2001), so food shortage may in certain circumstances<br />
affect both species simultaneously. However, as Gaston (1988), we found no support for<br />
synchronized occurrence. An analysis of years with high numbers of Little Auks (taken from<br />
SEABIRD 21 (2008): 16–31
Vagrancy of Brünnich’s Guillemot<br />
Camphuysen & Leopold (1996)) and records of Brünnich’s Guillemots during autumn and winter,<br />
both only in countries directly surrounding the North Sea, did not find more records of Brünnich’s<br />
Guillemots in years with Little Auk invasions than in years without (analysed by year for 1975/1976<br />
- 1996/1997: Fisher’s Exact Test: n = 52, P = 0.380). However the few records of Brünnich’s<br />
Guillemots might make it impossible to detect any potential correlation.<br />
For Brünnich’s Guillemot sightings, see Van Bemmelen et al. 2005 (Figure 9); Birding World 18: 425,<br />
430 (photo), 19: 30; 19: 442, 447; 20: 447; British Birds 99: 658; 101: 54. For details on Little Auk<br />
numbers, see Van Bemmelen & Wielstra 2005; Birding World 18: 405; 19: 447; 20: 447, 450; British<br />
Birds 99: 658; 101: 54.<br />
Brünnich’s Guillemots breeding in the northeast Atlantic average slightly larger than<br />
those breeding in the western Atlantic, but there is considerable overlap in<br />
measurements (Cramp 1985; Gaston & Jones 1998), and biometrics are unlikely to<br />
help identify the geographic origin of a small sample of vagrant birds (but see Gaston<br />
(1988) for an attempt). Although stable-isotope analysis has been successfully applied<br />
to trace bird movements (Rubenstein & Hobson 2004; Fox et al. 2007; Fox & Bearhop<br />
2008), strong longitudinal gradients in isotope ratios are absent in the Atlantic<br />
(Kroopnick 1980; Rubenstein & Hobson 2004) and as the latitudinal distribution of<br />
Brünnich’s Guillemot colonies is relatively narrow (at least in the northeast Atlantic),<br />
assignment of individuals to specific geographic regions by this method seems<br />
problematic. It proved impossible to unambiguously segregate birds from different<br />
populations based on mitochondrial DNA sequence data (Birt-Friesen et al. 1992).<br />
Similar difficulties are to be expected for the nuclear genome, but no attempts have<br />
yet been made as far as we are aware.<br />
Given the above, the recovery of a Brünnich’s Guillemot ringed at a breeding colony<br />
seems the most likely way of identifying its origin. However, while extensive ringing<br />
has been carried out in the past (Kampp 1988; Nikolaeva et al. 1996; Bakken &<br />
Mehlum 2005), and expanded programmes have been proposed for the future (CAFF<br />
2004), the relatively small number of European vagrants means the chance of finding<br />
a ringed, extra-limital Brünnich’s Guillemot is slim, making the only known ringing<br />
recovery all the more remarkable, a bird ringed in August 1948 at Novaya Zemlya,<br />
Russia and caught alive in April 1964 in Poland (Tomialojć 1990; Rønnest 1994).<br />
As with other vagrant seabirds, one likely explanation for the occurrence of Brünnich’s<br />
Guillemots in western Europe is displacement by severe weather events. A number of<br />
live birds have been found moribund or sheltering in harbours after severe gales, and<br />
some temporal clusters of records coincided with or followed periods of strong<br />
northerly winds over the seas around and to the east of Iceland. This was the case in<br />
January 1981, when six birds were found after two spells of northerly gales, and again<br />
in January 1995, after an intense anti-cyclone in the mid Atlantic and low pressure<br />
over Scandinavia at the beginning of the year brought Arctic northerly gales across the<br />
northern North Sea; the five records that month were the only ones in 1994/95.<br />
However, records in other years (such as the seven in 1986/87) were more spread<br />
SEABIRD 21 (2008): 16–31<br />
23
24<br />
Vagrancy of Brünnich’s Guillemot<br />
throughout the year and there was little obvious to connect them to any specific<br />
weather event. In some circumstances, wind may simply help land-based observers<br />
detect vagrants already in the area instead of being the cause of their displacement.<br />
Environmental factors other than brief weather events, and operating over a larger<br />
sea area and a longer timeframe, may also influence the occurrence of Brünnich’s<br />
Guillemots in Europe. Winter movements of Brünnich’s Guillemots are generally<br />
contained within cold currents of Arctic water temperature (Cramp 1985), and a<br />
reduction in sea temperature could cause a shift in the southerly limit of their normal<br />
range, decreasing the distance to the North Sea. Irons et al. (2008) found that<br />
population fluctuations of Brünnich’s Guillemots were associated with decadal<br />
changes in winter (January to March) sea surface temperature (SST) in the vicinity of<br />
breeding colonies, and plotted a 3-year running mean of deviation of SST in the<br />
northeast Atlantic from the 50-year average. Interestingly, the lowest two points on<br />
this graph (i.e. when SSTs were coldest) were in 1981 and 1987, coincident with<br />
above average number of vagrant Brünnich’s Guillemots (Figure 2), although a third<br />
peak in records (the seven in 1997/98, six of which were in the Skaggerak/Kattegat)<br />
occurred when SSTs had returned to above average levels.<br />
Trends and status: Vagrant Brünnich’s Guillemots have been only recorded regularly in<br />
Europe since the mid 1950s, with three records or more in 14 of the 29 years from<br />
1977/78. However, the balance has shifted between the two geographic clusters of<br />
records, with fewer in the UK and Ireland (Scotland, essentially) and more in the<br />
Skagerrak/Kattegat (Figure 3). <strong>The</strong>ir seasonal occurrence also differed considerably, with<br />
more autumn records in the latter region (14) than the former (2). <strong>The</strong>re is no obvious<br />
explanation for these disparities in both annual and seasonal occurrence. Given the<br />
increase in observer activity and identification skills in recent decades, and the increased<br />
proportion of live birds found (Figure 3), the decrease in the number of records in the UK<br />
and Ireland must reflect (and under-estimate) a decreased occurrence of the species<br />
there. Whether this is due to changes in population size, migratory routes or wintering<br />
areas, or disruptive climatic events or regimes is unknown. <strong>The</strong> contrasting increase in<br />
southern Scandinavia, however, where virtually all recent records have been of live birds<br />
(Figure 3), is more likely to reflect an increased sea-watching effort, a better<br />
understanding of alcid identification, higher quality optical equipment, and perhaps a<br />
greater willingness to submit and accept records of birds seen only in flight (since<br />
1991/1992, records of nine ‘migrating’ individuals have been accepted in Denmark and<br />
Sweden). Brünnich’s Guillemots may also have been ranging further south along the<br />
Norwegian coast than in the past, in which case they would be more likely to be found<br />
in southern Scandinavia than in the western North Sea, but there is no systematic<br />
evidence for this (G. Mobakken, T. Anker-Nilssen, R. Barrett pers. comm.).<br />
In the UK, the accumulation of Brünnich’s Guillemot records has led some to suggest<br />
the species may winter regularly in small numbers in the northern North Sea (e.g.<br />
Rogers et al. 1978), but others consider this unlikely (Cramp 1985; Fraser et al. 2007;<br />
McGeehan 1991). In Shetland, just three dead Brünnich’s Guillemots were found on<br />
systematic beached bird surveys from 1979 to 2004 compared to 15,107 Common<br />
SEABIRD 21 (2008): 16–31
Figure 7. Brünnich’s Guillemot Uria lomvia, adult summer,<br />
Lille, Antwerp, Belgium, 5 August 2006 © Tom Goossens.<br />
Vagrancy of Brünnich’s Guillemot<br />
Figure 8. Brünnich’s Guillemot Uria lomvia, adult<br />
summer, Yell, Shetland, UK, 4 May 2006 © Mick Mellor.<br />
SEABIRD 21 (2008): 16–31<br />
Figure 5 (above). Brünnich’s<br />
Guillemot Uria lomvia, probable<br />
first-winter, Lerwick, Shetland,<br />
UK, December 2005 © Hugh<br />
Harrop.<br />
Figure 6 (left). Brünnich’s<br />
Guillemot Uria lomvia, probably<br />
2nd calendar year, Scousburgh,<br />
Shetland, UK, 25 March 2007 ©<br />
Roger Riddington. This record fell<br />
outside the period under consideration<br />
here, but has been<br />
accepted by the British Birds<br />
Rarities Committee (Hudson et<br />
al. 2008).<br />
25
26<br />
Vagrancy of Brünnich’s Guillemot<br />
Guillemots (Heubeck 2006), and Brünnich’s Guillemot has never been seen during 31<br />
years of ship-based inshore surveys of wintering seabirds there (M. Heubeck pers.<br />
comm.). <strong>The</strong> lack of any English record south of Northumberland (Figure 1), and totals<br />
of 45,357 Common Guillemots but no Brünnich’s Guillemots found on systematic<br />
beached bird surveys during 1965–2007 in <strong>The</strong> Netherlands (C. J. Camphuysen pers.<br />
comm.) also suggest no regular presence in the western or southern North Sea.<br />
Despite this, observers should always consider the possibility of encountering the<br />
species during beached bird surveys or sea-watches. Field identification of Brünnich’s<br />
Guillemot may be challenging, but is – even in flight – not insurmountable (McGeehan<br />
1991; Ullman 1998; Blomdahl et al. 2003).<br />
Acknowledgements<br />
Many people helped us prepare this article and we thank them all. Earlier drafts were<br />
improved by comments from Martin Heubeck, who also provided information on some<br />
Shetland records. Mardik Leopold encouraged us and helped gather records and<br />
forward requests for information. Museum specimens were checked by Bob McGowan<br />
(National Museums of Scotland), Joan McLaren (Montrose Museum, UK), Mike Nicoll<br />
(Dundee Museum, UK) and Richard Sutcliffe (Glasgow Museum, UK). <strong>The</strong> following<br />
sent lists of national records: Jochen Dierschke (Germany), Roger Riddington (UK),<br />
Marnix Vandegehuchte (Belgium) and Georges Olioso (France), while Killian Mullarney<br />
commented on the Irish record. John Chardine and Eeva-Liisa Alanen provided useful<br />
references, Arnoud van den Berg helped at the Dutch Birding Association library, and<br />
we thank the EuroBirdNet community for their help. Finally, we thank two anonymous<br />
referees for their comments.<br />
Figure 9. Alcid identified in the field as a Brünnich's Guillemot Uria lomvia, Schiermonnikoog,<br />
Friesland, <strong>The</strong> Netherlands, 23 October 2005 © Martijn Renders. This 'fly-by' sighting was<br />
rejected by the Dutch rarities committee, but this decision is to be reviewed.<br />
SEABIRD 21 (2008): 16–31
Vagrancy of Brünnich’s Guillemot<br />
References<br />
Bakken, V. & Mehlum, F. 2005. Wintering areas and recovery rates of Brünnich’s Guillemots<br />
Uria lomvia ringed in the Svalbard Archipelago. Arctic 58: 268–275.<br />
Birt-Friesen,V. L., Montevecchi,W.A., Gaston,A. J. & Davidson,W. S. 1992. Genetic structure<br />
of Thick-Billed Murre (Uria lomvia) populations examined using direct sequence analysis of<br />
amplified DNA. Evolution 46: 267–272.<br />
Blake, B. F. 1983. A comparative study of the diet of auks killed during an oil incident in the<br />
Skagerrak in January 1981. Journal of Zoology, London 201: 1–12.<br />
Blomdahl, A., Breife, B. & Holmström, N. 2003. Flight identification of European seabirds.<br />
Christopher Helm, London.<br />
Breife, B., Hirschfeld, E., Kjellén, N. & Magnus, U. 1990. Sällsynta fåglar i Sverige. Sveriges<br />
Ornitologiska Förening, Lund.<br />
CAFF. 2004. Distribution of murres outside the breeding season. “Circumpolar Murre Banding<br />
Program – North Atlantic Region”. CAFF Technical Report No. 13. CAFF International<br />
Secretariat, Akureyri, Iceland.<br />
Camphuysen, C. J. & Leopold, M. F. 1996. Invasies van de Kleine Alk Alle alle: voorkomen en<br />
achtergronden. Sula 10: 169–182.<br />
Cramp, S. (ed.) 1985. <strong>The</strong> Birds of the Western Palearctic.Vol. IV. Oxford University Press, Oxford.<br />
Dubois, P. J. & Yésou, P. 1991. Les oiseaux rares en france. Chabaud, Bayonne Cedex.<br />
Engebretsen, S. & Pettersen, M. 2001. Østfold neste stopp? – Polarlomvi Uria lomvia. Natur I<br />
Østfold 20: 79-84.<br />
Fox, T. A. D., Christensen, T. K., Bearhop, S & Newton, J. 2007. Using stable isotope analysis of<br />
multiple feather tracts to identify moulting provenance of vagrant birds: a case study of Baikal<br />
Teal Anas formosa in Denmark. Ibis 149: 622–625.<br />
Fox,T.A. D. & Bearhop, S. 2008. <strong>The</strong> use of stable-isotope ratios in ornithology. British Birds 101:<br />
112–130.<br />
Fraser, P.A., Rogers, M. J. & the Rarities Committee 2007. Report on rare birds in Great Britain<br />
in 2005. Part I: non-passerines. British Birds 100: 16–61.<br />
Gaston, A. J. 1984. How to distinguish first-year murres, Uria spp., from older birds in winter.<br />
Canadian Field-Naturalist 98: 52–55.<br />
Gaston, A. J. 1988. <strong>The</strong> mystery of the murres: Thick-billed Murres, Uria lomvia, in the Great<br />
Lakes region, 1890–1986. Canadian Field-Naturalist 102: 705–711.<br />
Gaston A. J. & Jones, I. L. 1998. <strong>The</strong> Auks. Oxford University Press, Oxford.<br />
Goethe, F. & Ringleben, H. 1964. Ein neuerer Nachweis von Uria lomvia (L.) für Deutschland.<br />
Journal of Ornithology 105: 54–56.<br />
Goossens, T. 2006. Kortbekzeekoet in Belgisch binnenland. Dutch Birding 28: 340–341.<br />
Grant, P. J. 1981. Mystery photographs 51. British Birds 74: 144–145.<br />
Heubeck, M. 2006. <strong>The</strong> Shetland beached bird survey 1979–2004. Marine Ornithology 34: 123–127.<br />
Irons, D. B., Anker-Nilssen, T. A., Gaston, A. J., Byrd, G. V., Falk, K., Gilchrist, G., Hario, M.,<br />
Hjernquist, M., Krasnov,Y.V., Mosbeck, A., Olsen, B., Petersen, A., Reid, J. B., Robertson, G. J.,<br />
Strøm, H. & Wohl, K. D. 2008. Fluctuations in circumpolar seabird populations linked to climate<br />
oscillations. Global Change Biology 14: 1455–1463.<br />
Hudson, N. & the Rarities Committee. 2008. Report on rare birds in Great Britain in 2007.<br />
British Birds 101: 516–577.<br />
Isaksen, K. & Bakken, V. 1995. Breeding populations of seabirds in Svalbard. Pp 11–15 In Isaksen,<br />
K. & Bakken, V. (eds.) <strong>Seabird</strong> populations in the northern Barents Sea – source data for the<br />
impact assessment of the effects of oil drilling activity. Norsk Polarinstitutt Meddelelser 135.<br />
SEABIRD 21 (2008): 16–31<br />
27
28<br />
Vagrancy of Brünnich’s Guillemot<br />
Jaeschke, G. & Schulz, U. 1968. Dickschnabellumme (Uria lomvia) 1966 an der Mecklenburgischen<br />
Ostseküste. Journal of Ornithology 109: 131.<br />
Kampp, K. 1988. Migration and winter ranges of Brünnich’s Guillemots Uria lomvia breeding<br />
or occurring in Greenland. Dansk Ornitologisk Forenings Tidsskrift 82: 117–130.<br />
Kroopnick, P. 1980. <strong>The</strong> distribution of 13C in the Atlantic Ocean. Earth and Planetary Science<br />
Letters 49: 469–484.<br />
McGeehan, A. 1991. Brünnich’s Guillemot: rare or overlooked? Irish Birding News 1: 105–112.<br />
Mela, A. J. 1903. Pohjankiisla. Luonnon Ystävä 7: 37–45.<br />
Meyer, M. & Thorwarth, H. J. 1987. Erstnachweis der Dickschnabellumme (Uria lomvia) für das<br />
Binnenland der DDR. Beiträge zur Vogelkunde 34: 313–314.<br />
Moody, A. T. & Hobson, K. A. 2007. Alcid winter diet in the northwest Atlantic determined by<br />
stable isotope analysis. Marine Ornithology 35: 39–46<br />
Mullarney, K. 1988. Brünnich’s Guillemot in County Wexford – an addition to the Irish list. Irish<br />
Birds 3: 601–605<br />
Nettleship, D. N. & Evans, P. G. H. 1985. Distribution and status of the Atlantic Alcidae. In: Nettleship,<br />
D.N. & Birkhead T.R. (eds.). <strong>The</strong> Atlantic Alcidae. <strong>The</strong> evolution, distribution and biology of the auks<br />
inhabiting the Atlantic Ocean and adjacent water areas: 53–154. Academic Press, London.<br />
Nikolaeva, N. G., Krasnov, Y. V. & Barrett, R. T. 1996. Movements of Common Uria aalge and<br />
Brünnich’s Murre U. lomvia breeding in the southern Barents Sea. Fauna norvegica Series C,<br />
Cinclus 19: 9–20.<br />
Olsen, K. M. 1992. Danmarks fugle – en oversigt. Dansk Ornitologisk Forening, København.<br />
R Development Core Team 2005. R: A language and environment for statistical computing. R<br />
Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org.<br />
Ribbands, J. B. 1977. Brünnich’s Guillemot in Northumberland. British Birds 73: 225–226.<br />
Risberg, L. 1990. Sveriges fåglar. Aktuell översikt över deras utbredning, numerär och flyttning<br />
samt något om svensk ornitologi. Sveriges Ornitologiska Förening, Stockholm.<br />
Rogers, M. J. & the Rarities Committee. 1978. Report on rare birds in Great Britain in 1977.<br />
British Birds 71: 481–532.<br />
Rønnest S. 1994. Sjældne fugle i Danmark. En oversigt over forekomsten af sjældne fugle i<br />
Danmark og Nordvesteuropa 1963–1992. Forlaget Pinus, Skjern.<br />
Rubenstein, D. R. & Hobson, K. A. 2004. From birds to butterflies: animal movement patterns<br />
and stable isotopes. Trends in Ecology and Evolution 19:256–263.<br />
Skov, H., Durinck, J. & Danielsen, F. 1989. Sammenfaldende forekomst mellem Søkonger Alle<br />
alle og Glaskutling Aphya minuta i Flekkefjord, Sydnorge, November 1988. Pelagicus 4: 22–24.<br />
Stempniewicz, L. 2001. Little Auk Alle alle. BWP Update. <strong>The</strong> Journal of Birds of the Western<br />
Palearctic 3: 145–201.<br />
Tomialojć, L. 1990. Ptaki Polski. Rozmieszczenie i liczebno´sć. Państwowe Wydawnictwo<br />
Naukowe, Warszawa.<br />
Ullman, M. 1998. Fältbestämning av spetsbergsgrissla. Vår Fågelvärld 8: 28–33.<br />
Van Bemmelen, R. S. A., Wielstra, B., Renders, M. & Hendriks, K. 2005. Black-and-white<br />
weekend: Kleine Alken en Kortbekzeekoet langs Schiermonnikoog. Dutch Birding 27: 448–449.<br />
Van Bemmelen, R. S. A. & Wielstra, B. M. 2005. Invasie van kleine alken op 23 oktober 2005:<br />
kort maar krachtig. NZG Nieuwsbrief 7: 1–4.<br />
Van Gompel, J. 1981. De massale zeevogelsterfte aan de Belgische kust tijdens de voorbije<br />
winter. Wielewaal 47: 137–142.<br />
Van Gompel, J. 1982. First record in Belgium of a living specimen of Brünnich’s Guillemot (Uria<br />
lomvia). Wielewaal 48: 169–170.<br />
SEABIRD 21 (2008): 16–31
Vagrancy of Brünnich’s Guillemot<br />
Appendix 1.<br />
List of vagrant Brünnich’s Guillemot records in Europe, arranged by country and date. Details given are date,<br />
location, age, plumage, sex, condition, and the fate of the specimen, as far this information could be traced.<br />
Abbreviations used are. Age: Ad = adult, 1st w = first-winter, 1st s = first-summer; Plumage: B = breeding, NB<br />
= non-breeding; Sex: F = female, M = male. Museums: NMS = National Museums of Scotland; NNM =<br />
Nationaal Natuurhistorisch Museum, Leiden; ZMA = Zoologisch Museum Amsterdam. References to rarity<br />
reports and reports of rarities committees are provided in a simplified form, providing the name of the journal,<br />
page numbers and information on published photographs. Journal abbreviations: IB = Irish Birds, BB = British<br />
Birds, SB = Scottish Birds, DB = Dutch Birding, DOFT = Dansk Ornitologisk Forenings Tidsskrift, SF = Sveriges<br />
Fåglar, FoFl = Fauna och Flora, VF = Vår Fågelvärld, Fåg = Fågelåret.<br />
Ireland (1 record, 1 individual)<br />
24 Dec. 1986 Kilmore Quay, Ballyteigue Bay, Wexford. NB, alive. IB 3: 478; BB 80: 547; Mullarney 1988.<br />
United Kingdom (38 records, 39 individuals)<br />
10 Dec. 1908 Craigielaw Point, Lothian. 1st w, F, dead, specimen at NMS. BB 2: 425.<br />
15 Apr. 1960 Middleton Sands, near Morecambe, Lancashire & North Merseyside. Ad, dead.<br />
BB 54: 188, 284–286.<br />
20 Mar. 1968 Norwick, Unst, Shetland. NB, F, dead, specimen at NMS. BB 62: 473; SB 5: 272<br />
(plate 19b), 285–287.<br />
11 Oct. 1969 Knapdale, Loch Caolisport, Argyll. Dead. BB 63: 281; SB 6: 334–335.<br />
31 Jan. 1976 Reay, Thurso, Caithness. Ad, NB, dead, specimen at NMS. BB 71: 509–510.<br />
13 Jul. 1977 Farne Islands, Northumberland. Ad, B, alive, at sea near a Common Guillemot colony.<br />
BB 72: 528; 73: 225–226.<br />
18 Dec. 1977 Sumburgh, Shetland. Dead. BB 71: 509.<br />
14 Jul. 1978 St Cyrus, northeast Scotland. 1st s, B, M, dead, specimen at Montrose Museum. BB 72: 528.<br />
25 Feb. 1979 Rattray Head, northeast Scotland. Dead. BB 73: 514.<br />
9 Feb. 1980 Kilspindie Beach, Lothian. Ad, NB, dead, specimen at Glasgow Museum. BB 74: 478.<br />
9 Feb. 1980 Ferry Ness, Lothian. Dead, specimen in L. Simmen private collection. BB 74: 478.<br />
24 Feb. 1980 Burrafirth, Unst, Shetland. NB, dead. BB 81: 569.<br />
16–17 Oct. 1980 Fair Isle, Shetland. Ad, B, alive but presumed to have died on 17 Oct.. BB 74: 478<br />
(plate 279); 81: 569.<br />
26 Dec. 1980 At Sea, Brent Oilfield, Norwegian Sea, 61°03’N 01°43’E. Alive. BB 75: 511.<br />
25 Jan. 1981 Johnshaven, Kincardine, northeast Scotland. Dead, specimen at Dundee Museum. BB 75: 511.<br />
29 Dec. 1981 Bay of Ireland, Stenness, Orkney. Dead. BB 75: 511.<br />
3 Feb. 1982 Brora, Sutherland, Highland. Dead. BB 76: 502.<br />
3 Apr. 1982 Stromness, Orkney. Dead. BB 77: 537.<br />
24 Dec. 1982 Golspie, Highland. 1st w, dead, specimen at NMS. BB 100: 55-56.<br />
30 Oct. 1983 Banna Minn, West Burra, Shetland. 1st w, F, dead, specimen at NMS. BB 78: 561.<br />
20 Mar. 1984 Birsay, Orkney. Dead. BB 78: 561.<br />
9 Jan. 1985 Scapa Bay, Orkney. NB, dead. BB 79: 558.<br />
3–7 Feb.1987 Off Hamnavoe, West Burra, Shetland. Alive, NB & 7 Feb 1987, Off Hamnavoe, West Burra,<br />
Shetland. NB, dead, (different individual, live bird still present), specimen in D. Coutts<br />
private collection. BB 81: 569.<br />
9 Mar. 1988 Dunnet Bay, Caithness. Ad, M, dead, specimen at NMS. BB 82: 533.<br />
16 Jun.–12 Jul. 1989 Sumburgh Head, Shetland. Alive, B, site holding in a Common Guillemot colony. BB 83: 469.<br />
25 Jan. 1991 Sule Skerry, Orkney. Alive. BB 85: 531.<br />
26 May–8 Jun. 1992 Hirta, St Kilda, Outer Hebrides. Alive. BB 86: 496.<br />
27 Mar. 1993 Musselburgh, Lothian. Ad, B, alive. BB 87: 536.<br />
6 Feb. 1994 Seafield, Lothian. Alive. BB 88: 523; 89: 509.<br />
12 Feb. 1994 Wadbister Voe, Shetland. 1st w, dead, oiled, specimen at NMS. BB 88: 523.<br />
SEABIRD 21 (2008): 16–31<br />
29
30<br />
Vagrancy of Brünnich’s Guillemot<br />
4 Jan. 1995 Gulberwick, Shetland. Alive, weak, taken into care, released at Wadbister Voe 1 February,<br />
last seen there 2 February. BB 89: 509.<br />
23 Jan. 1995 At Sea, north of Fair Isle, 62°41’N 01°34’W. Alive. BB 89: 509.<br />
27 Mar. 1996 Kilchoan Bay, Ardnamurchan, Argyll. Alive. BB 90: 488.<br />
25–30 Dec. 1997 Fetlar, Shetland. Alive. BB 91: 496.<br />
21 Dec. 2000 Scapa Flow, Orkney. Ad, NB, dead, specimen at NMS. BB 94: 480.<br />
29 Jan. 2001 Orkney North Ronaldsay, Orkney. 1st w, dead, specimen at NMS. BB 95: 501.<br />
30 Nov.–20 Dec. 2005 Lerwick & Bressay, Shetland. NB, alive. BB 100: 55–56 (photo); 100: 55 (photo).<br />
4 May 2006 Southladie Voe, Yell, Shetland. Ad, B, dead, specimen at NMS. BB 100: 694–754.<br />
France (3 records, 3 individuals)<br />
21 Apr. 1978 near Plougerneau, Finistère. Dead, oiled. Dubois & Yésou 1991.<br />
21 Jan. 1981 near Santec, Finistère. Dead, oiled. Dubois & Yésou 1991.<br />
3 Feb. 2003 Audinghen, Pas-de-Calais. Alive. Ornithos 12: 24.<br />
Belgium (5 records, 5 individuals)<br />
4 Jan. 1981 Klemskerke/De Haan,West-vlaanderen. 1st w, dead, oiled. Oriolus 52: 73;Van Gompel 1981.<br />
18 Jan. 1981 Wenduine, West-vlaanderen. Ad, dead, oiled. Oriolus 52: 73.<br />
7 Dec. 1981 Blankenberge, West-vlaanderen. 1st w, alive, oiled, died same day. Oriolus 52: 73; Van<br />
Gompel 1982.<br />
21 Jan. 1995 Bredene, West-vlaanderen. Ad, NB, dead. Aves 34: 195-223; Oriolus 63: 81-82; DB 17: 79,<br />
86 (photo).<br />
4 Aug. 2006 Visbeek Wechelderzande, Lille, Antwerp. Ad, B, alive. Goossens 2006; BW 19: 320 (photo).<br />
<strong>The</strong> Netherlands (7 records, 7 individuals)<br />
24 Dec. 1919 Noordwijk aan Zee, Noordwijk, Zuid-Holland. Ad, F, NB, dead, specimen at NNM.<br />
Ardea 9: 32–33.<br />
28 Dec. 1924 Noordwijk aan Zee, Noordwijk, Zuid-Holland. M, NB, dead, specimen at NNM.<br />
19 Feb. 1969 Texel, Noord-Holland. Dead, specimen at NNM.<br />
10 Mar. 1974 Oostkapelle, Domburg, Zeeland. M, NB, dead, specimen at ZMA.<br />
4–10 Feb. 1979 Brouwersdam, Goedereede, Zeeland. M, NB, alive, oiled, found dead on 10 February,<br />
specimen at ZMA. DB 1: 109–111 (photos); 2:20 (photo).<br />
10 Jan. 1981 Monster, Zuid-Holland. Ad, B, dead, speciman in private collection. DB 3: 99 (photo).<br />
18 Apr. 1992 Texel, Noord-Holland. Ad, NB, dead, specimen at ZMA. DB 15: 43.<br />
Germany (3 records, 3 individuals)<br />
1 Nov. 1959 Voslapp, north of Wilhemshaven, Niedersachsen. NB, dead, specimen at Museum<br />
Heineanum Halberstadt and Vogelwarte Helgoland, Wilhelmshaven. Goethe &<br />
Ringleben 1964.<br />
8 Mar. 1966 Warnemunde, Mecklenburg-Vorpommern. Ad, NB, F, dead, specimen in Meereskundlichen<br />
Museums Stralsund. Jaeschke & Schulz 1968.<br />
7 Aug. 1987 Leipzig, Sachsen. Ad, B, alive, weak, died next day, specimen in Naturkunde-museum<br />
Leipzig. Meyer & Thorwarth 1987 (photo).<br />
Denmark (10 records, 10 individuals)<br />
14 May 1886 Storsømmen, Sjælland. Dead. Olsen 1992.<br />
2 Oct. 1905 Kalveboderne, København, Sjælland. Alive, shot. Olsen 1992.<br />
8 Nov. 1925 Mesinge ved Kerteminde, Fyn. Dead. Olsen 1992.<br />
29 Oct. 1974 Blåvand Strand, Ribe. 1st w, dead. DOFT 74: 134–135 (photo); Olsen 1992.<br />
10 Dec. 1989 Hundested Havn, Sjælland. NB, alive. DOFT 85: 27 (photo); Olsen 1992.<br />
14 Jan. 1991 At Sea between Hanstholm & Kristianssand, Nordjylland. Alive. DOFT 87: 236.<br />
17 Jan. 1995 Hirtshals Havn, Nordjylland. Alive. DOFT 91: 142.<br />
21 Jan. 1998 Læsø Rende, Nordjylland. Ad, B, alive. DOFT 93: 133<br />
SEABIRD 21 (2008): 16–31
Vagrancy of Brünnich’s Guillemot<br />
23 Jan. 1998 Strandby Havn, Nordjylland. Ad, alive. From 24 January to 3 February 1998 at<br />
Frederikshavn Havn, Nordjylland. DOFT 93: 133.<br />
28 Feb. 1998 Kikhavn, Hundested, Sjælland. Alive, migrating. DOFT 93: 133.<br />
Sweden (34 records, 38 individuals)<br />
10 Dec. 1874 Kosterfjorden, Bohuslän. NB, dead, specimen at Strömstads Museum. Risberg 1990.<br />
5 Feb. 1875 Strömstad, Bohuslän. NB, alive. FoFl 39: 95.<br />
1 Dec. 1902 Parish of Pajala, Norrbotten. NB, 2 individuals, alive, shot. FoFl 2: 216–217.<br />
Jan. 1925 Grebbestad, Bohuslän. NB, 2 individuals, alive. SF 4: 64.<br />
May 1955 Karesuando, Lappland. Alive. SF 1978: 130.<br />
17 Dec. 1955 Kälvudden, Överkalix, Norrbotten. NB, alive, ringed and released. VF 15: 280.<br />
Feb. 1962 Killingholmen, Boden, Norrbotten. NB, alive. VF 32: 307.<br />
9 Feb. 1964 Båstad, Skåne. NB, dead. Anser 5: 9.<br />
6 Jan. 1981 Fjällbacka, Bohuslän. NB, alive, shot. VF 46: 327.<br />
19 Apr. 1982 Lilla Karlsö, Gotland. B, alive. VF 42: 401.<br />
5 Nov. 1983 Busör, Halland. NB, alive. VF 43: 544; Breife et al. 1990 (photo).<br />
11 Jan. 1987 Klöverträsk - Älvsbyn, Norrbotten. NB, dead, found on a road. VF 47: 456.<br />
21 Jul. 1987 Sotenäs, Bohuslän. Ad, B, alive. VF 47: 456.<br />
28 Nov.–6 Dec. 1987 Varbergs hamn, Halland. 1st w, alive, eventually “collected”. VF 47: 456.<br />
3 Jan. 1991 St. Amundön, SW Askim, Västergötland. Ad, NB, alive. VF 51 7: 24.<br />
21 Jan. 1993 Sote huvud, Bohuslän. NB, alive, migrating. Fåg 1993: 112.<br />
14 Nov. 1993 Sebybadet, Öland, NB, alive, migrating. Fåg 1993: 112.<br />
18–19 Oct. 1995 Hovs hallar,Skåne, 2 individuals, one also seen at Kullen, Skåne, NB, alive, migrating.<br />
Fåg 1995: 134.<br />
8 Dec. 1995 Segerstads fyr, Öland. NB, alive. Fåg 1995: 134.<br />
16–17 Nov. 1996 Soten, Sotenäs, Bohuslän. Two individuals, NB, alive. Fåg 1996: 140.<br />
17 Oct. 1997 Busör, Halland. NB, alive. Fåg 1997: 160.<br />
28–29 Oct. 1997 Välen - Askimsviken, Västergötland. 1st w, F, alive, found dead on 30 October, specimen<br />
at Göteborg Natural History Museum. Fåg 1997: 160.<br />
30 Dec. 1997 Kullen, Skåne. NB, alive. Fåg 1997:160.<br />
28 Feb. 1998 Hovs hallar, Skåne, NB, alive, migrating. Fåg 1998: 145.<br />
20–21 May 2000 Lilla Karlsö, Gotland. B, alive, migrating. Fåg 2001: 136.<br />
10 Jun. 2001 Revsudden, Kalmarsund, Småland. B, alive. Fåg 2001:1 36.<br />
30 Oct. 2001 Tussebo, Skepplanda, Västergötland. Ad, M, NB, alive, died on 1 November, specimen at<br />
Göteborg Natural History Museum. Fåg 2001: 136.<br />
3 Nov. 2001 Busör, Halland. NB, alive, migrating. Fåg 2001: 136.<br />
7 Feb. 2002 Valnäsudden, Nordkoster, Bohuslän. NB, alive. Fåg 2002: 139.<br />
7 Apr. 2002 Hermanö huvud, Bohuslän. B, alive, migrating. Fåg 2002: 139.<br />
17 Jan. 2003 Bua udde, Väröbacka, Halland. Alive. Fåg 2003: 184.<br />
15 Nov. 2003 Busör, Halland. 1st w, alive. Fåg 2003: 184.<br />
3 Dec. 2003 Segelskär, Tanum, Bohuslän. NB, alive. Fåg 2003: 184<br />
5–19 Nov. 2004 Guleskärskajen, Kungshamn, Bohuslän. Ad, NB, alive. Fåg 2004: 145.<br />
Poland (1 record, 1 individual)<br />
10 Apr. 1964 Gdynia, Pomorskie. Ad, alive, trapped, ringed at Novaya Zemlya in August 1948 (number<br />
D-117833 Moskva). Tomialojć, 1990; Rønnest 1994.<br />
Finland (2 records, 2 individuals)<br />
14 Dec. 1986 Inari, Lappland. Alive, brought to Korkeasaari Zoo, Helsinki, where it died on 26 January<br />
1987. Lintumies 23: 199.<br />
22 Oct. 1988 Nauvo Sandö, Varsinais-Suomi. 1st w, dead, drowned in fishnet. Lintumies 23: 278.<br />
SEABIRD 21 (2008): 16–31<br />
31
32<br />
Leach’s and European Storm-petrels<br />
A survey of Leach’s Oceanodroma leucorhoa<br />
and European Storm-petrel Hydrobates<br />
pelagicus populations on North Rona and<br />
Sula Sgeir, Western Isles, Scotland<br />
Murray, S. 1 *, Money, S. 2 , Griffin, A. 3 & Mitchell, P. I. 4<br />
*Correspondence author. Email: murraysurvey@yahoo.co.uk<br />
1 Craigie Dhu, Cardney, Dunkeld, Perthshire PH8 0EY, UK; 2 Raintree House, Church Lane,<br />
Drayton St Leonard, Oxfordshire OX10 7AU, UK; 3 15 Horologie Hill, Arbroath, Angus<br />
DD11 5AE, UK; 4 Joint Nature Conservation Committee, Dunnet House, 7 Thistle Place,<br />
Aberdeen AB10 1UZ, UK.<br />
Abstract<br />
Leach’s Storm-petrel Oceanodroma leucorhoa was first recorded breeding on North<br />
Rona in 1883 and on Sula Sgeir in 1939. European Storm-petrel Hydrobates pelagicus<br />
was first recorded on North Rona in 1885 and on Sula Sgeir in 1958. Since then, there<br />
have been attempts to estimate the population size of both species on North Rona but<br />
there is little information about their current status on Sula Sgeir. In 2001, systematic<br />
surveys of both species using tape playback were conducted for the first time on both<br />
islands. North Rona held 1,133 Apparently Occupied Sites (AOS) of Leach’s Stormpetrel<br />
but only 371 AOS of European Storm-petrel; numbers on Sula Sgeir were five<br />
and eight AOS respectively. <strong>The</strong> combined population of both North Rona and Sula<br />
Sgeir of Leach’s Storm-petrel and European Storm-petrel, comprise 2.3% and 1.4%<br />
respectively, of the total number of each species breeding in Great Britain.<br />
Introduction<br />
North Rona (area = 128 ha, highest point = 108 m), uninhabited since 1844, lies about<br />
70 km north of the Butt of Lewis in the Western Isles at 59°08’N 5°50’W. Leach’s Stormpetrel<br />
Oceanodroma leucorhoa and European Storm-petrel Hydrobates pelagicus were<br />
first discovered there in the 1880s (Swinburne 1885; Harvie-Brown 1888), but there has<br />
never been an accurate census of either species. Both are difficult to survey, since they<br />
nest in burrows or in rock crevices and are nocturnal. However, the recently developed<br />
tape playback technique (Ratcliffe et al. 1998) enabled this study to obtain an accurate<br />
count of Apparently Occupied Sites (AOS) of both species for the first time.<br />
Sula Sgeir (20 ha, 70 m) lies about 17 km west of North Rona, at 59°06’N 6°09’W. It<br />
is no more than 1 km in length and just 200 m across at its widest point. It is sparsely<br />
vegetated due to lack of soil, and is subject to heavy erosive pressure from breeding<br />
seabirds and sea-spray. It has rarely been visited by ornithologists and ‘stormy petrels’,<br />
species uncertain, were first confirmed present in 1930 (Dougal 1937). North Rona<br />
holds extensive evidence of past human occupation, most prominently a ruined<br />
village, graveyard and chapel enclosed by pronounced cultivation ridges. Sula Sgeir,<br />
SEABIRD 21 (2008): 32–43
Leach’s and European Storm-petrels<br />
although never permanently occupied has been visited annually for centuries by the<br />
men of Ness in Lewis, who have built dry stone bothies on the rock. <strong>The</strong> structures on<br />
both islands are an important component of the available breeding habitat for both<br />
petrel species. North Rona and Sula Sgeir are designated as a Special Protection Area<br />
(SPA) under Article 4.1 of the EC Birds Directive (79/409/EEC) for supporting more<br />
than 1% of the Great Britain breeding populations of Leach’s and European Stormpetrel,<br />
which are both listed in Annex 1 of the Directive (Stroud et al. 2001). <strong>The</strong><br />
islands also qualify for SPA designation under Article 4.2 of the Directive by supporting<br />
more than 1% of the relevant bio-geographic breeding populations of Northern<br />
Gannet Morus bassanus and Common Guillemot Uria aalge (Stroud et al. 2001).<br />
Leach’s and European Storm-petrel were cited as qualifying species for the North Rona<br />
and Sula Sgeir SPA based on estimates of colony size and Great Britain population size<br />
given in Lloyd et al. (1991). <strong>The</strong>se estimates were derived before tape playback was<br />
developed and most were expressed as orders of magnitude of the number of birds<br />
present at colonies during the night. <strong>The</strong> aim of the present study was to use tape<br />
playback to accurately census both species of storm-petrel and, combined with the<br />
latest Great Britain population estimates (also derived using tape playback) (Mitchell<br />
2004; Mitchell & Newton 2004), assess their conservation status and validate the<br />
islands’ SPA designation with respect to these species.<br />
Methods<br />
Colony census methods: <strong>The</strong> playback method entails playing recordings of the chatter<br />
call of a male Leach’s Storm-petrel and the purr call of European Storm-petrel in<br />
suitable habitat during the incubation period, in order to elicit a reply from an<br />
incubating adult within a burrow. A hand-held dictaphone with integral speakers was<br />
used.All accessible areas on both islands, including the highest sea cliffs on North Rona,<br />
were systematically surveyed using the tape playback technique (Gilbert et al. 1998)<br />
and responses were counted and mapped. <strong>The</strong> main drawback with the tape playback<br />
method is that not all individuals will respond to the taped calls (Ratcliffe et al. 1998),<br />
so a count of responses will underestimate the total number of AOS at a colony.<br />
Furthermore, Leach’s Storm-petrel will only respond to taped chatter calls of the same<br />
sex (Taoka et al. 1989), therefore it is necessary to measure what proportion of birds<br />
present in burrows are responding to the taped calls. This was achieved by setting up a<br />
calibration plot for each species, which entails repeatedly visiting a delimited section of<br />
the colony on successive days and each time marking new responding AOS. Calibration<br />
plots were set up on the North Rona storm beach for European Storm-petrel and along<br />
the village graveyard wall for Leach’s Storm-petrel. On the first visit to each plot, 20<br />
AOS were located and their positions marked with flagged canes. Both plots were<br />
visited on a total of six days between 26 June and 1 July 2001.<strong>The</strong> total number of AOS<br />
on each island was then estimated by multiplying counts of responding birds by the<br />
response rate derived from the calibration plots (Equation 1).<br />
Equation 1: Number of AOS = no. responses x (1 / response rate)<br />
where response rate is estimated as shown in Equation 2.<br />
SEABIRD 21 (2008): 32–43<br />
33
34<br />
Leach’s and European Storm-petrels<br />
North Rona: <strong>The</strong> island was subdivided into 16 sections (A to P, Figure 1), using clearly<br />
defined natural or man-made boundaries, and each was searched systematically for<br />
both species. To ensure full coverage of each section only one species was searched for<br />
at a time by three surveyors. This ensured that each section received the same level of<br />
effort. In some areas, mainly Fianuis (section L) and the storm beach (section M), ropes<br />
and canes were used to subdivide the ground into strips several metres wide in order<br />
to aid coverage. Island coverage was close to 100% and the only site possibly holding<br />
breeding petrels that was not surveyed was the lower, inaccessible half of Geo Mairi,<br />
between sections O and P.<br />
Sula Sgeir<br />
Figure 1. North Rona showing survey sections A to P, and place names given in the text. Sula Sgeir, showing the<br />
survey area and location of the bothies.<br />
SEABIRD 21 (2008): 32–43<br />
North Rona
Figure 2. Aerial photograph of North Rona © J. A. Love.<br />
Leach’s and European Storm-petrels<br />
<strong>The</strong> Toa Rona cliffs (section P) were difficult to access, held large numbers of Northern<br />
Fulmars Fulmarus glacialis and Atlantic Puffins Fratercula arctica and vegetation was<br />
unstable over shallow soil. <strong>The</strong>refore, to lessen risks to surveyors and to minimise<br />
disturbance to the colonies, the area was surveyed by a single surveyor only. Overall,<br />
the slope was delimited by steep rock walls, narrow gullies and sharply defined<br />
vegetation boundaries, which simplified surveying.<br />
European Storm-petrels were surveyed between 26 June and 2 July and Leach’s Stormpetrels<br />
between 2 and 8 July, (except for the village ruins (section E), which were<br />
surveyed on 29 June 2001).<br />
Sula Sgeir: Sula Sgeir was surveyed on 24 June 2001. To minimise disturbance to the<br />
dense assemblage of breeding seabirds on the flat top of the rock, only the five bothies,<br />
set in a small area of eroded soil edged by boulders, were surveyed (Figure 1).<br />
Estimating response rate: <strong>The</strong>re are two methods for estimating the response rate of<br />
a storm-petrel population from calibration plot data:<br />
1. Simple arithmetic (Equation 2)<br />
Equation 2: response rate = total number of responses / (total AOS x number of visits).<br />
SEABIRD 21 (2008): 32–43<br />
35
36<br />
Leach’s and European Storm-petrels<br />
Confidence limits were calculated using a Generalised Linear Model, with the number<br />
of responses from each AOS as the dependent variable and the number of visits made<br />
to each AOS as the independent variable. However, this method is sensitive to birds<br />
becoming habituated to the taped calls, so that progressively fewer previously located<br />
AOS responded on successive days. Using this method in a habituated plot will tend<br />
to underestimate the response rate of a population not yet exposed to the taped calls.<br />
Furthermore, if not all AOS in the plot had been found by the time of last visit,<br />
response rate will be overestimated.<br />
2. Iterative regression. <strong>The</strong> advantage of this method over the simple arithmetic<br />
method is that not all AOS in a plot need to be found; this is useful if a surveyor can<br />
only make a limited number of visits to the plot, which is often the case on remote<br />
islands that are difficult and costly to visit. For each calibration plot, an asymptotic<br />
regression model was fitted to the cumulative number of AOS that had been found<br />
following each visit (see Mayhew et al. 2000). <strong>The</strong> model took the form of Equation 3<br />
and the parameters a and b were predicted using the iterative regression function of<br />
S-Plus® 2000 (Mathsoft Inc., Seattle, Washington).<br />
Equation 3: where y = number of AOS detected on a given visit (x); b = the<br />
exponential (e) proportional rate of increase to the asymptote (a).<br />
Thus, the coefficient a is an estimate of the total number of AOS present in the study<br />
plot. <strong>The</strong> response rate was calculated by substituting the values of the coefficients for<br />
b into Equation 4.<br />
Equation 4: response rate =<br />
<strong>The</strong> upper and lower 95% confidence limits of the response rate were determined<br />
from the Equations 5 and 6 respectively.<br />
Equation 5: 95% UCL response rate = 1 - e-b-(se x 1.96)<br />
Equation 6: 95% LCL response rate = 1 - e-b+(se x 1.96) where se is the standard error<br />
of the estimate of the coefficient b.<br />
Results<br />
North Rona: European Storm-petrel response rate; <strong>The</strong> results from the European<br />
Storm-petrel calibration plot are shown in Table 1 and the cumulative number of AOS<br />
found following on each successive visit are plotted in Figure 3. Table 1 shows a clear<br />
reduction in total responses following the first visit, with only six birds responding on<br />
day six. This would suggest that some degree of habituation to the taped calls was<br />
taking place. Hence, the use of the simple arithmetic technique to calculate response<br />
rate was inadvisable in this case. Instead an iterative regression was applied to the plot<br />
Table 1. European Storm-petrel Hydrobates pelagicus calibration plot results.<br />
Visit no. 1 2 3 4 5 6<br />
No. of new burrows found per visit 20 9 7 3 4 2<br />
Cumulative total of burrows 20 29 36 39 43 45<br />
No. of responses per visit 20 13 11 13 16 6<br />
SEABIRD 21 (2008): 32–43
Cumulative no. of AOS<br />
Leach’s and European Storm-petrels<br />
Table 2. Parameters of the iterative regression of the cumulative number of European Storm-petrel<br />
Hydrobates pelagicus Apparently Occupied Sites (AOS) on successive visits to the calibration plot. <strong>The</strong><br />
resultant correction factor was multiplied by the total counts of responses to estimate the total number of<br />
AOS. Figures in parentheses are 95% CLs.<br />
60<br />
55<br />
50<br />
45<br />
40<br />
35<br />
30<br />
25<br />
20<br />
a b Response rate Correction factor<br />
(1/response rate)<br />
46.3 0.509 0.400 2.5<br />
(44.1–48.6) (0.035 s.e.) (0.356–0.439) (2.28–2.81)<br />
t 5 = 40.329 t 5 = 14.385<br />
European Storm-petrel<br />
Leach's Storm-petrel<br />
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15<br />
Days after first visit<br />
Figure 3. Change in cumulative number of new<br />
AOS found using diurnal playback on successive<br />
visits to the calibration plots for European Stormpetrel<br />
Hydrobates pelagicus and Leach’s Stormpetrel<br />
Oceanodroma leucorhoa on North Rona.<br />
<strong>The</strong> curve was fitted to the European Storm-petrel<br />
data using iterative regression, dotted lines<br />
indicate 95% confidence limits. where a = 46.3<br />
(44.1–48.6 s.e.) and b = 0.509 (0.035 s.e.)<br />
and the results are shown in Table 2. <strong>The</strong> extrapolation of the response curve (Figure<br />
3) estimated that 46 AOS were present in the plot, just one more than the number<br />
actually found after six visits. <strong>The</strong> response rate predicted from the slope of the curve<br />
was 0.40, giving a correction factor of 2.5 (i.e. 1 / response rate, see Equation 1).<br />
European Storm-petrel population estimates; One hundred and forty-seven birds<br />
responded to the taped calls at 18 sites, in 11 out of 16 colony sections (Tables 3 & 4,<br />
Figure 1), representing an estimated 371 AOS (95% CL 335– 413). <strong>The</strong> largest concentration<br />
of 203 AOS was found in the storm beach (section M). <strong>The</strong> remainder were<br />
thinly scattered across the island, with man-made structures, e.g. the chapel ruins,<br />
Fianuis bothies, walls, cairns and enclosures forming an important component of the<br />
breeding habitat, accounting for 25% of all AOS (Table 3). None were found in the<br />
large, apparently suitable area of rock scree on the south side of Geodha Lèis (section<br />
K ), but there were five AOS bordering the area. On the Toa Rona cliffs (section P), eight<br />
AOS were found along the cliff top, but none elsewhere. Both areas were densely<br />
occupied by breeding Atlantic Puffins and Geodha Lèis by Razorbills Alca torda also.<br />
Leach’s Storm-petrel response rates; <strong>The</strong>re was no evidence of habituation by Leach’s<br />
Storm-petrel to the taped calls, and the number of responses from the plot actually<br />
increased over the last three visits (Table 5). However, the iterative regression was not<br />
used to estimate response rate because, with hindsight, too few visits were made to<br />
SEABIRD 21 (2008): 32–43<br />
37
38<br />
Leach’s and European Storm-petrels<br />
Table 3. Habitat selection and estimated AOS of European Storm-petrels Hydrobates pelagicus on North Rona.<br />
Habitat type No. of sites No. of responses % of total responses AOS (95% CL)<br />
Storm beach 1 81 55 203 (185–228)<br />
Other natural sites 8 29 20 74 (66–81)<br />
Man-made structures 9 37 25 94 (84–104)<br />
Total 18 147 100 371 (335–413)<br />
Table 4. Number of AOS of Leach’s Storm-petrel Oceanodroma leucorhoa and European Storm-petrel<br />
Hydrobates pelagicus on North Rona. See Figure 1 for map of sections A to P.<br />
Leach’s Storm European Leach’s Storm European<br />
Section -petrel Storm-petrel Section -petrel Storm-petrel<br />
A 0 3 I 85 10<br />
B 65 23 J 97 8<br />
C 0 0 K 159 33<br />
D 12 0 L 185 30<br />
E 328 20 M 28 203<br />
F 5 0 N 51 0<br />
G 23 0 O 21 30<br />
H 30 3 P 44 8<br />
Total AOS 1,133 371<br />
Table 5. Leach’s Storm-petrel Oceanodroma leucorhoa calibration plot results.<br />
Visit no. 1 2 3 4 5 6 Total<br />
No. of new burrows found per visit 20 9 8 13 6 2 58<br />
Cumulative total of burrows 20 29 37 50 56 58 58<br />
No. of responses per visit 20 16 19 34 35 27 151<br />
Table 6. Habitat selection and estimated AOS of Leach’s Storm-petrels Oceanodroma leucorhoa on North Rona.<br />
Habitat type No. of responses % of total responses AOS (95% CL)<br />
Storm beach 12 2% 28 (26–29)<br />
Other natural sites 306 62% 708 (673–745)<br />
Man made structures 172 35% 397 (374–422)<br />
Total 490 100% 1,133 (1,065–1,202)<br />
the plot and consequently, produced too limited a range of x-values to enable<br />
Equation 3 to be meaningfully applied (see Figure 3). <strong>The</strong> simple arithmetic technique<br />
(Equation 2) was applied instead. <strong>The</strong> mean response rate across the six visits was<br />
0.434, with 95% CL of 0.408–0.460 (GLM, t 56 = 17.82, P < 0.001). <strong>The</strong> correction<br />
factor for Leach’s Storm-petrel was therefore, 1/0.434 = 2.31 (95% CL 2.173–2.453).<br />
SEABIRD 21 (2008): 32–43
Leach’s and European Storm-petrels<br />
Leach’s Storm-petrel population estimates; Four hundred and ninety birds responded to<br />
the taped calls, representing 1,133 AOS (95% CL 1065–1202) ( Table 4 & 6, Figure 1).<br />
<strong>The</strong> largest and densest subcolony on the island was found in the village ruins (section<br />
E), which held 328 AOS, 29% of the island total. A further 69 AOS (6%) were found in<br />
other man-made structures, including the low turf dyke demarcating section J on Toa<br />
Rona, which held 97 AOS in the second densest subcolony on the island.<br />
Overall, 708 AOS (62%) were found in natural sites, widely distributed across the<br />
island, with the exception of sections A and C where none were found. Nests were<br />
situated in stone piles, under embedded boulders, along cliff edges and in well drained<br />
turf, but not in the old cultivation ridges of section F, or in the deep, waterlogged soils<br />
of sections D and H. None were found in the largest Atlantic Puffin colony on Toa Rona<br />
(section P), although they were present in small numbers along the cliff tops. Similarly,<br />
at other, smaller puffin subcolonies, e.g. Geodha Lèis and Geodha Blatha Mor (both in<br />
section K), Leach’s Storm-petrel nests were found in abandoned puffin burrows on the<br />
edge of the colonies, but none were found in areas of densely occupied puffin burrows.<br />
Sula Sgeir: Leach’s and European Storm-petrels were found only in the walls of the<br />
bothies, but not in any of the cairns or other man-made structures that were surveyed.<br />
Around the bothies the ground was badly eroded and occupied by a high density of<br />
nesting Northern Fulmars. Potentially suitable storm-petrel habitat in boulders around<br />
the margins of the survey area (Figure 1) held nesting Atlantic Puffins and Common<br />
Guillemots, and probably as a result, did not appear to be occupied by either stormpetrel<br />
species. In total, only two responses from Leach’s Storm-petrel and three from<br />
European Storm-petrel were elicited during the survey of Sula Sgeir. Using the<br />
response rates estimated on North Rona (Tables 2 & 5), these responses equated to<br />
five AOS and eight AOS respectively.<br />
Discussion<br />
European Storm-petrel on North Rona: <strong>The</strong> estimate of 371 AOS of European Stormpetrel<br />
is the lowest made of the North Rona population to date. However, the distribution<br />
of AOS found during the current study is broadly similar to earlier accounts. For<br />
example, the storm beach (section M) was considered by Bagenal & Baird (1959) to<br />
hold the largest concentration of burrows on the island and Love (1978) considered<br />
the storm beach to hold many more pairs than the village. In 2001, the storm beach<br />
also held the largest subcolony, with 55% of the total number of AOS on the island.<br />
By contrast, low numbers have consistently been noted in the village ruins (section E),<br />
with nests found only in the chapel and in a nearby boulder pile. In 2001, 20 AOS were<br />
found there, which was identical to the number of pairs estimated to be breeding there<br />
in 1936 by Ainslie & Atkinson (1937a). Despite conducting additional searches of the<br />
village during the night in 2001, none were found breeding in any structure other than<br />
the chapel. Likewise, Robson (1968) found a few nests in the chapel, but none<br />
elsewhere, and Love (1978) noted that 72% of his mist net captures of European<br />
Storm-petrels in the village were made close to the chapel. Elsewhere on the island,<br />
AOS were thinly distributed and in low numbers.<br />
SEABIRD 21 (2008): 32–43<br />
39
40<br />
Leach’s and European Storm-petrels<br />
<strong>The</strong> mean response rate of European Storm-petrels on North Rona was very similar to<br />
other colonies in northwest Scotland (Mitchell & Newton 2004). <strong>The</strong> calibration and<br />
simultaneous census of North Rona were conducted early in the incubation period<br />
expected of European Storm-petrels at Scottish colonies (i.e. late June and most of July),<br />
so response rate may have been expected to increase if not all of the birds had<br />
commenced incubation by the start of the study (Ratcliffe et al. 1998). However, daily<br />
response rate in the calibration plot at least stayed fairly constant apart from on the day<br />
of the last visit when it was considerably lower. Furthermore, detailed daily searches of<br />
AOS in the village did not find any more than were there on the first day of searching.<br />
<strong>The</strong>refore, the mean response rate estimated in the plot appeared to have accurately<br />
reflected the response rate across the rest of the colony when the census was conducted.<br />
Leach’s Storm-petrel on North Rona: On North Rona it is difficult to make any<br />
assessment of change prior to this study, since no systematic survey of Leach’s Stormpetrel<br />
numbers, using comparable methods, has been conducted previously. Our count<br />
of 328 AOS in the village ruins is almost identical to the 327 occupied burrows found<br />
there in 1936 by Ainslie & Atkinson (1937b).<strong>The</strong>y estimated only 50 burrows for the rest<br />
of the island, giving a total population of 380 pairs. In 1958, a study using ringing and<br />
recapture, estimated the village population to be around ten times larger, i.e. 2–3,000<br />
pairs, with a total island population of about 5,000 pairs (Bagenal & Baird 1959).<br />
However, estimates of population size derived from mist-netting and mark-recapture<br />
include non-breeding birds as well as breeders, so will be greater than concurrent<br />
estimates derived from counts of occupied nest sites. Bagenal & Baird’s (1959) estimate<br />
was also far greater than subsequent assessments of colony size, with both Robson<br />
(1968) and Love (1978) considering the village population to be little different in size<br />
compared to Ainslie & Atkinson’s (1937b) estimate in 1936. Outside the village, calling<br />
birds have been heard at many different sites, but only in small numbers. P. G. H. Evans<br />
in 1972, cited in Lloyd et al. (1991) estimated the entire population, including that of the<br />
village, at 500 pairs. Our estimate in 2001 of 1,133 AOS of Leach’s Storm-petrel on<br />
North Rona makes this the third largest colony, after St Kilda and the Flannan Isles, of<br />
only eight known colonies in Britain and Ireland (Mitchell 2004).<br />
<strong>The</strong> response rate measured in the calibration plot on North Rona was probably a<br />
slight overestimate because not all the AOS in the plot were found, though visual<br />
inspection of the calibration plot data in Figure 3 would suggest that only 2–3 AOS<br />
were missed. <strong>The</strong>refore, the 1,133 AOS is probably only a slight underestimate of the<br />
size of the colony on North Rona. A further inaccuracy may have resulted if the birds<br />
had responded differently during the census on 2–8 July, than they had done a week<br />
earlier when the response calibration was conducted (Ellis et al. 1998). <strong>The</strong> greatest<br />
determinant of response rate is the level of diurnal occupancy of burrows by the<br />
adults. At least one adult should be present in the burrow during the day throughout<br />
incubation and up to five days post-hatching. <strong>The</strong>re are no published data on the<br />
phenology of Leach’s Storm-petrels on North Rona, but recent observations on St<br />
Kilda have shown hatching to start during the second and third week in July (Money<br />
et al. 2008). It is reasonable to assume that breeding phenology is similar in both<br />
colonies, since they most probably use the same feeding grounds (Mitchell 2004).<br />
SEABIRD 21 (2008): 32–43
Leach’s and European Storm-petrels<br />
<strong>The</strong>refore it is unlikely that the Leach’s Storm-petrel on North Rona were responding<br />
to the taped calls differently during the calibration and the census. Furthermore, the<br />
response rate estimated in the current study was very similar to that recorded during<br />
4–10 July 2003 on St Kilda, using the same method (Newson et al. 2008).<br />
European Storm-petrel on Sula Sgeir: European Storm-petrels have been rarely<br />
recorded breeding on Sula Sgeir. Although they may have been present in 1930<br />
(Dougal 1937), they were not noted by either Stewart (1932), Atkinson & Ainslie<br />
(1940) or McGeoch (1954a,b). Bagenal & Baird (1959) caught one and heard others<br />
calling in June 1958 and this appears to be the first confirmed record. None were heard<br />
in 1980 but they were noted in 1986 (Benn et al. 1989) and the eight AOS in 2001<br />
were all found within bothy walls.<br />
Leach’s Storm-petrel on Sula Sgeir: On Sula Sgeir Leach’s Storm-petrels were<br />
probably the ‘stormy petrels’ observed in 1930 by Dougal (1937) and in 1932 by<br />
Stewart (1934), as being present in bothy walls. Breeding was proven in 1939 by<br />
Atkinson & Ainslie (1940) who found both young and incubating adults. <strong>The</strong>y carried<br />
out an overnight survey of the rock, starting from the bothies and surveying<br />
southwards to the summit cairn. <strong>The</strong>y found birds calling from underground near the<br />
bothies and along the length of the summit ridge (Figure 3).<strong>The</strong>y considered that they<br />
were at least as numerous as the population on Rona and estimated that 400 pairs<br />
were breeding there. <strong>The</strong> subsequent loss of soil and vegetation around the bothies,<br />
caused largely by the increase in nesting Northern Fulmars, and on the plateau by the<br />
expansion of the gannetry, probably reduced the available breeding habitat. As a<br />
Figure 4. Leach's Storm-petrel Oceanodroma leucorhoa, Merseyside, 23 September 2004 © Sue Tranter.<br />
SEABIRD 21 (2008): 32–43<br />
41
42<br />
Leach’s and European Storm-petrels<br />
consequence, in 1954, McGeoch (1954a,b) found only a dozen birds, although he<br />
searched for several nights up to the edge of the gannetry. Bagenal & Baird (1959)<br />
caught 18 in mist nets set overnight between the bothies in June 1958, and a few were<br />
heard in bothy walls, but nowhere else, during an overnight stay in June 1980 (S.<br />
Murray pers. obs.). Since then there have been further increases in both Northern<br />
Gannet and Northern Fulmar numbers (Mitchell et al. 2004) and in 2001, only five<br />
AOS of Leach’s Storm-petrel were found, all in the walls of bothies.<br />
Conclusions<br />
Although the European Storm-petrel appears never to have bred in large numbers on<br />
Sula Sgeir, numbers of the Leach’s Storm-petrel have declined substantially since 1939,<br />
probably caused by habitat loss due to soil erosion and through competition for space<br />
from the larger and more aggressive Northern Gannets and Northern Fulmars. In 2001,<br />
three fresh and unmarked European Storm-petrel corpses were found lining a<br />
Northern Fulmar nest.We speculated that all had been killed by the incubating fulmar,<br />
as they made their way in or out of the adjacent bothy wall. Whatever the case, it<br />
illustrates the vulnerability of small petrels at this site.<br />
It is unclear from previous records whether or not numbers of Leach’s and European<br />
Storm-petrel breeding on North Rona have changed significantly during the last century.<br />
However, future surveys of either species on the island should use the results of the 2001<br />
survey (shown in Tables 3, 4, & 6; Figure 1) as a baseline for assessing changes in breeding<br />
numbers and distribution. <strong>The</strong> precision of the 2001 estimates of colony size (see 95%<br />
CL in Tables 3 & 6) should enable future surveys to detect changes in breeding numbers<br />
of both species. Changes will have to be greater than 9% for European Storm-petrel and<br />
6% for Leach’s Storm-petrel in order to be statistically significant.<br />
<strong>The</strong> combined population of both North Rona and Sula Sgeir of Leach’s Storm-petrel<br />
and European Storm-petrel, comprise 2.3% and 1.4% respectively, of the total number<br />
of each species breeding in Great Britain (Mitchell 2004). <strong>The</strong>refore, North Rona and<br />
Sula Sgeir qualify as an SPA under the EC Birds Directive for their importance as a<br />
breeding site for both Annex 1 listed species of storm-petrels.<br />
Acknowledgements<br />
<strong>The</strong> work outlined in this report was funded by the European Research and<br />
Development fund under the Atlantic Area Programme of the Interreg II C Initiative, as<br />
part of Project No. 414, awarded to the Joint Nature Conservation Committee in<br />
partnership with BirdWatch Ireland. We are grateful to Norman Ratcliffe and Adrian<br />
Plant for their comments on an earlier draft; to Paddy Pomeroy for the use of the Sea<br />
Mammal Research Unit hut and equipment; and especially to our boat skippers, Murdo<br />
MacDonald and Andy Tibbits.<br />
References<br />
Ainslie, J.A. & Atkinson, R. 1937a. Summer bird notes from North Rona. Scottish Naturalist 49: 7–13.<br />
Ainslie, J. A. & Atkinson, R. 1937b. On the breeding habits of Leach’s Fork-tailed Petrel. British<br />
Birds 30: 234–248.<br />
SEABIRD 21 (2008): 32–43
Leach’s and European Storm-petrels<br />
Atkinson, R. & Ainslie, J.A. 1940. <strong>The</strong> British breeding status of Leach’s Fork-tailed Petrel. British<br />
Birds 34: 50–55.<br />
Bagenal, T. B. & Baird, D. E. 1959. <strong>The</strong> birds of North Rona in 1958, with notes on Sula Sgeir.<br />
Bird Study 6: 153–174.<br />
Benn, S., Murray, S. & Tasker, M. L. 1989. <strong>The</strong> birds of North Rona and Sula Sgeir. NCC,<br />
Peterborough, UK.<br />
Dougal, J. W. 1937. Island Memories. Edinburgh, Moray Press, Edinburgh.<br />
Ellis, P., Ratcliffe, N. & Suddaby, D. 1998. Seasonal variation in diurnal attendance and<br />
response to playback by Leach’s Petrel Oceanodroma leucorhoa, on Gruney, Shetland. Ibis 140:<br />
336–339.<br />
Gilbert, G., Gibbons, D. W. & Evans, J. 1999. Bird Monitoring Methods, a Manual of Techniques<br />
for Key UK Species. Royal Society for the Protection of Birds, UK.<br />
Harvie-Brown, J. A . & Buckley, T. E. 1888. A Vertebrate Fauna of the Outer Hebrides. David<br />
Douglas, Edinburgh.<br />
Lloyd,C.,Tasker,M.L.& Partridge,K.1991. <strong>The</strong> Status of <strong>Seabird</strong>s in Britain and Ireland. Poyser, London.<br />
Love,J.A.1978.Leach’s and Storm-petrels on North Rona 1971–74. Ringing & Migration 2: 15–19.<br />
Mayhew, P., Chisholm, K., Insley, H. & Ratcliffe, N. 2000. A survey of Storm Petrels on Priest<br />
Island. Scottish Birds 21: 78–84.<br />
McGeoch, J. 1954a. August migrants at Sula Sgeir. Fair Isle Bird Observatory Bulletin 25: 201–208.<br />
McGeoch, J. 1954b. ‘Birds of Sula Sgeir, August 1954.’ Unpubl. report to SNH Western Isles.<br />
Mitchell, P. I. 2004. Leach’s Storm-petrel Oceanodroma leucorhoa. In: Mitchell, P. I., Newton,<br />
S. F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. <strong>Seabird</strong> Populations of Britain and Ireland: 101–114.<br />
Poyser, London.<br />
Mitchell, P. I., Newton, S. F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. <strong>Seabird</strong> Populations of<br />
Britain and Ireland. Poyser, London.<br />
Mitchell, P. I. & Newton, S. F. 2004. European Storm-petrel Hydrobates pelagicus. In: Mitchell,<br />
P. I., Newton, S. F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. <strong>Seabird</strong> Populations of Britain and<br />
Ireland: 81–100. Poyser, London.<br />
Money, S., Söhle, I. & Parsons, M. 2008. A pilot study of phenology and breeding success of<br />
Leach’s Storm-petrel Oceanodroma leucorhoa on St Kilda, Western Isles. <strong>Seabird</strong> 21: 98–101.<br />
Newson, S. E., Mitchell P. I., Parsons, M., O’Brien, S. H., Austin, G. E., Benn, S., Black, J.,<br />
Blackburn, J., Brodie, B., Humphreys, E., Leech, D. I., Prior, M. & Webster, M. 2008.<br />
Population decline of Leach’s Storm-petrel Oceanodroma leucorhoa within the largest colony<br />
in Britain and Ireland. <strong>Seabird</strong> 21: 77–84.<br />
Ratcliffe, N., Vaughan, D., Whyte, C. & Shepherd, M. 1998. Development of playback census<br />
methods for Storm-petrels Hydrobates pelagicus. Bird Study 45: 302–312.<br />
Robson, M. J. H. 1968. <strong>The</strong> breeding birds of North Rona. Scottish Birds 5: 126–156.<br />
Stewart, M. 1934. <strong>The</strong> status of petrels in certain remote Scottish islands. Scottish Naturalist<br />
46: 95–98.<br />
Stroud, D. A., Chambers, D., Cook, S., Buxton, N., Fraser, B., Clement, P., Lewis, P., McLean, I.,<br />
Baker, H. & Whitehead, S. 2001. <strong>The</strong> UK SPA network: its scope and content. JNCC,<br />
Peterborough, U.K.<br />
Swinburne, J. 1885. Notes on the islands of Sula Sgeir or North Barra and North Rona, with a<br />
list of birds inhabiting them. Proceedings of the Royal Physical Society of Edinburgh 8: 51–67.<br />
Taoka, M., Sato, T., Kamada, T. & Okumura, H. 1989. Sexual dimorphism of chatter calls and<br />
vocal sex recognition in Leach’s Storm-petrels. Auk 106: 498–501.<br />
SEABIRD 21 (2008): 32–43<br />
43
44<br />
<strong>Seabird</strong> diet on Canna<br />
<strong>The</strong> diet of European Shag Phalacrocorax<br />
aristotelis, Black-legged Kittiwake Rissa<br />
tridactyla and Common Guillemot Uria<br />
aalge on Canna during the chick-rearing<br />
period 1981–2007<br />
Swann, R. L. 1 *, Harris, M. P. 2 & Aiton, D. G. 3<br />
*Correspondence author. Email: robert.swann@homecall.co.uk<br />
1 14 St Vincent Road, Tain, Ross-shire IV19 1JR, UK; 2 Centre for Ecology and Hydrology,<br />
Hill of Brathens, Banchory, Aberdeenshire AB31 4BY, UK (current address: CEH Bush Estate,<br />
Penicuik, Midlothian EH26 0QB, UK); 3 14 Buckstone Howe, Edinburgh EH10 6XF, UK.<br />
Abstract<br />
Chick diet of European Shags Phalacrocorax aristotelis, Black-legged Kittiwakes Rissa<br />
tridactyla and Common Guillemots Uria aalge at Canna was investigated over a 27year<br />
period. <strong>The</strong> diet was mainly composed of Sprats Sprattus sprattus, Lesser Sandeels<br />
Ammodytes marinus and members of the Gadidae (a variety of species but mainly<br />
Trisopterus spp. and Whiting Merlangius merlangus). Other groups (ten families of fish,<br />
crustaceans, cephalopod molluscs and polychaete worms) were of minimal<br />
importance. Lesser Sandeels dominated the diet of young Black-legged Kittiwakes and<br />
European Shags, Sprats the diet of young Common Guillemots, whereas gadid otoliths<br />
were by far the commonest items found in pellets regurgitated by older European<br />
Shags. <strong>The</strong>re were few significant temporal changes in species composition or the size<br />
of prey taken over the 27 years and the results confirm earlier findings that gadids are<br />
a normal and important part of the diet of seabirds at this colony.<br />
Introduction<br />
Over the last 40 years seabird populations in Britain have undergone major changes.<br />
Up to the late 1980s numbers were tending to increase, but since then major declines<br />
have been observed (Mavor et al. 2006). For some species, e.g. Black-legged Kittiwake<br />
Rissa tridactyla, changes in numbers are thought to be linked to changes in food supply,<br />
particularly sandeels (Ammodytidae, mainly Lesser Sandeels Ammodytes marinus)<br />
(Frederiksen et al. 2004). Several studies on seabird diet undertaken at colonies in the<br />
North Sea and Shetland have shown the importance of sandeels in the diets of<br />
seabirds in these areas (Pearson 1968; Monaghan 1992; Daunt et al. 2008), but<br />
relatively little information is available on the food of seabirds at colonies in the west<br />
of Britain. Swann et al. (1991) summarised information on the food of seabirds on<br />
Canna during the chick-rearing periods 1981–90, and showed that although sandeels<br />
were the main species taken by European Shag Phalacrocorax aristotelis, Black-legged<br />
Kittiwake and Common Guillemot Uria aalge, in contrast to some other places gadids<br />
(Gadidae, cod-fishes) and clupeids (Clupeidae, Herring Clupea harengus and Sprats<br />
SEABIRD 21 (2008): 44–54
<strong>Seabird</strong> diet on Canna<br />
Sprattus sprattus) also featured highly in the diet. This paper updates information on<br />
the diet of these seabirds on Canna up to 2007 and compares the findings with<br />
information from other locations in northern Britain.<br />
Methods<br />
Canna (57°03’N, 6°32’W) is situated in the southern Minch in western Scotland and<br />
annual visits were made to the colony to ring and monitor seabirds in the period<br />
1981–2007.Three visits were made each year, one for five days in late May, one for seven<br />
days in late June/early July and one for five days in late July/early August. Regurgitations<br />
were collected from young and adults feeding chicks of Black-legged Kittiwake (hereafter<br />
‘Kittiwake’) and European Shag (‘Shag’). Fish being carried back to chicks by adult<br />
Common Guillemots (‘Guillemot’) were also collected. Adult Guillemots returning with<br />
fish to the colony were targeted and captured and the fish they dropped were collected<br />
(Swann et al. 1991). Kittiwake and Guillemot samples were mainly collected between 28<br />
June and 10 July, whilst Shag samples were collected, if available, on all visits. <strong>The</strong><br />
contents of regurgitates and pellets were broken up in warm water or were digested in a<br />
warm solution of biological washing powder (biotex), fish present were identified, using<br />
otoliths where necessary (Härkönen 1986) and any intact specimens were measured.<br />
Sandeels, mostly if not all Lesser Sandeels, were classified as 0-group (young of the year)<br />
or older, based on the absence or presence of annual growth rings in otoliths (ICES 1995).<br />
Fish from Guillemots were identified, using otoliths where necessary, and the majority<br />
were measured (tip of snout to tip of tail) and weighed (to 0.1 g), although we are aware<br />
that these fish will have lost weight due to desiccation while being carried back to the<br />
colony held in the adult’s beak (Montevecchi & Piatt 1987). Prior to 1990 gadids were<br />
measured and weighed but measurements were not always matched with species.<br />
Unless otherwise stated, all Clupeidae appeared to be Sprat, although the identification<br />
of very small specimens was problematic, so the name sprat is used for that family.<br />
Pellets regurgitated by fully-grown Shags of unknown breeding status were collected<br />
between May and September and were digested in biotex until the mucous was<br />
dissolved. Otoliths were identified to family and when they were not too worn, and<br />
time allowed, to species level. Other items (crustacean fragments, polychaete jaws,<br />
cephalopod beaks) were assigned to the lowest possible taxa. Otoliths from sandeels<br />
fell obviously into two sizes. ‘Large’ were from older sandeels, but ‘small’ were a<br />
mixture of those from 0-group and those from older individuals that had their outer<br />
parts (and so growth rings) eroded away by the acid of the stomach (Johnstone et al.<br />
1990) so no attempt was made to age these otoliths.<br />
Results are expressed as frequency of occurrence in samples and, for otoliths and other<br />
hardparts, the actual numbers present. No attempt was made to pair up otoliths or<br />
mandibles of polychaete worms but fragments of crustacea and pairs of cephalopod<br />
beaks were treated as single items. A few regurgitates and many pellets contained more<br />
than one category of food so percentages of occurrence can total more than 100%.<br />
Some of the earlier data are described in Swann et al. (1991). Samples were collected<br />
on only a few days so we could not look for within-year variation in diet and we assume<br />
that what we report are representative of the chick-rearing season as a whole.<br />
SEABIRD 21 (2008): 44–54<br />
45
46<br />
<strong>Seabird</strong> diet on Canna<br />
Results<br />
Shag: Regurgitations were dominated by sandeels that were found in 67% of the<br />
134 samples (Table 1). <strong>The</strong>re was no significant trend in the proportion of samples<br />
with sandeels during the period (linear regression on arcsin transformed data: n = 24<br />
years weighted by sample size, R 2 = 13.8%, P = 0.14). Of 40 regurgitations where<br />
sandeel size was assessed, 19 (48%) had small sandeels (0-group) and 24 (60%)<br />
large (older) sandeels. Gadids occurred in 36% of samples; 17 fish were identified to<br />
genus or species – Trisopterus spp. (8), Whiting Merlangius merlangus (7), Cod Gadus<br />
morhua (1) and rockling Ciliata/Gaidropsarus sp. (1). Two samples contained Sprat<br />
(1), two unidentified wrasse (Labridae), whilst single samples contained Herring,<br />
Butterfish Pholis gunnellus (Pholidae), Bull-rout Myoxocephalus scorpius (Cottidae),<br />
Viviparous Blenny Zoarces viviparus (Zoarcidae), pipefish (presumably Snake Pipefish<br />
Entelurus aequoreus (Syngnathidae)), unidentified fish, Sea Mouse Aphrodite<br />
aculeate (Polychaete worm) and small crab (Crustacea).<br />
Figure 1. Annual frequency of occurrence of<br />
otoliths and other remains of sandeels<br />
(Ammodytidae), cod-fishes (Gadidae) and other<br />
prey groups in pellets of European Shag<br />
Phalacrocorax aristotelis on Canna.<br />
% occurrence<br />
Table 1. Contents of regurgitations from European Shags Phalacrocorax aristotelis on Canna. See text for<br />
details of other items.<br />
Sample % % % Other<br />
Year size Sandeels Gadidae items<br />
1981 6 50 33 17<br />
1982 2 100 0 0<br />
1983 2 100 0 0<br />
1984 5 80 20 0<br />
1987 6 100 0 0<br />
1988 3 100 0 0<br />
1989 5 100 20 0<br />
1990 8 50 50 0<br />
1991 4 100 25 25<br />
1992 6 50 67 17<br />
1993 2 0 50 50<br />
1994 7 86 14 0<br />
1995 9 44 44 22<br />
SEABIRD 21 (2008): 44–54<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
1986 1989 1992 1995 1998 2001 2004<br />
Sandeels Gadids Other<br />
Sample % % % Other<br />
Year size Sandeels Gadidae items<br />
1996 3 33 67 0<br />
1997 14 85 14 7<br />
1998 11 82 36 0<br />
1999 4 75 50 0<br />
2000 1 100 0 0<br />
2001 8 25 75 25<br />
2002 5 100 0 0<br />
2003 11 55 45 0<br />
2004 3 100 0 0<br />
2005 0<br />
2006 6 17 83 17<br />
2007 3 33 100 33<br />
Total 134 67 36 8
<strong>Seabird</strong> diet on Canna<br />
Table 2. Frequency of occurrence and total numbers of otoliths and other items in pellets regurgitated by<br />
European Shags Phalacrocorax aristotelis on Canna. See text for details of other items.<br />
% of pellets containing % of otoliths containing<br />
Period No. of Pellets Sandeels Gadidae Other No. of Otoliths Sandeels Gadidae Other<br />
1986 early August 6 50 100 0 231 5 95 0<br />
1989 early August 10 30 100 0 4195 5 95 0<br />
1993 early July 5 0 100 40 547 0 98 2<br />
1993 early August 8 0 100 0 3314 0 100 0<br />
1993 mid Sept 3 0 100 33 475 0 100
48<br />
<strong>Seabird</strong> diet on Canna<br />
No systematic attempts were made to identify gadid otoliths, but many species were<br />
present including Trisopterus spp. (25% of 3,621 that were assigned to genus), Saithe<br />
Pollachius virens, rockling,Whiting, Haddock Melanogrammus aeglefinus, Cod and Ling<br />
Molva molva. Most of these otoliths would have come from fish 50–150 mm long. Of<br />
the sandeel otoliths, 626 (67%) were ‘small’ (probably mainly 0-group but including<br />
some eroded otoliths from older fish) and 310 (33%) were ‘large’.<br />
Kittiwake: Sandeels were the commonest species found in regurgitations in 14 of the<br />
20 years, and occurred in 60% of the 227 samples (Table 3), gadids were the<br />
commonest food in four years, sprat in one, and in one year the two families were<br />
equally represented. Gadids and sprat were present in 22% of samples and very large<br />
numbers of very small pelagic crustaceans were found in 5%. <strong>The</strong> ‘other’ category was<br />
made up of four Lumpsuckers Cyclopterus lumpus (Cyclopteridae), two pipefish (both<br />
in 2007), one Three-spined Stickleback Gasterosteus aculeatus (Gasterosteidae) and a<br />
wrasse. Of the 62 samples where sandeels were aged, 58 (94%) were 0-group. Due to<br />
the lack of a colony-specific otolith:fish length relationship we could not calculate the<br />
lengths of these fish, but the bulk of relatively intact 0-group were 60–80 mm. Other<br />
fish were mostly very digested but most gadids appeared to be 60–100 mm long,<br />
whereas sprats were smaller at 50–60 mm.<br />
Table 3. Frequency of occurrence of fish families, crustacea and other items in regurgitations of Black-legged<br />
Kittiwakes Rissa tridactyla on Canna. See text for details of other items.<br />
% regurgitations with<br />
Year No. of Samples Sprats Sandeels Gadidae Crustacea Others<br />
1987 7 0 14 86 0 0<br />
1988 6 0 0 83 17 17<br />
1989 17 0 94 12 0 0<br />
1990 6 0 100 0 0 0<br />
1991 3 0 100 33 0 0<br />
1992 8 25 25 63 13 0<br />
1993 3 0 67 33 0 0<br />
1994 6 0 100 0 17 0<br />
1995 5 0 40 40 40 0<br />
1996 9 33 67 11 0 0<br />
1997 20 5 70 15 0 15<br />
1998 4 50 25 50 50 0<br />
1999 22 36 59 18 0 0<br />
2000 32 16 81 0 13 3<br />
2001 12 0 67 42 0 0<br />
2002 22 27 73 0 0 0<br />
2003 11 9 82 27 0 0<br />
2004 21 95 5 5 0 0<br />
2005 9 11 22 67 11 11<br />
2006 0<br />
2007 4 0 75 50 0 50<br />
Total 227 22 60 22 5 4<br />
SEABIRD 21 (2008): 44–54
<strong>Seabird</strong> diet on Canna<br />
Guillemot: Of the 1,562 fish collected, 753 (48%) were Clupeidae (three Herring and<br />
the rest Sprat), and 416 (27%) were Gadidae including 314 Whiting, 52 Trisopterus spp.<br />
(mostly unidentified but including Poor Cod T. minutus, Bib T. luscus and Norway Pout<br />
T. esmarkii), 25 Haddock, 10 Saithe, two Blue Whiting Micromesistius poutassou and<br />
two Cod; the rest of the gadids were not identified to species level. Sandeels (24%)<br />
made up the remainder (Table 4). <strong>The</strong>re were large variations in the fish taken from<br />
year to year (Figure 2), but generally sprats were less important in the 1980s and<br />
2000s than during the 1990s (x2 = 57.5, 2 df, P < 0.001). On an annual basis, sprats<br />
were the most important, making up on average 47 ± 4% of the fish. Sandeels and<br />
Gadidae contributed 26 ± 3% and 27 ± 3%, respectively.<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
1982 1986 1990 1994 1998 2002 2006<br />
Ammodytidae Gadidae Clupeidae<br />
Figure 2. Fish family composition (% by number) of prey taken by Common Guillemots Uria aalge on Canna.<br />
Excludes 1981 when sample size < 10.<br />
<strong>The</strong> length of sprats ranged from 68–155 mm, with a peak at 110–14 mm (Figure 3a).<br />
Most of these fish will have been 1-group or older. <strong>The</strong>re was a significant decline in<br />
mean length over the period (linear regression: Length (mm) = 765 – 0.328 year, n =<br />
717, P < 0.001); however since the overall decrease was only 5 mm, and year explained<br />
only 3% of the variation, this decline was unlikely to be biologically meaningful. <strong>The</strong><br />
mean length of gadids varied from 42–147 mm with a peak at 80–84 mm (Figure 3b).<br />
<strong>The</strong> length of Whiting, the commonest gadid, showed no significant change in size over<br />
the study period (linear regression: R 2 = 0.0%, n = 274, P = 0.903), though there were<br />
marked annual variations in the size of fish taken (Appendix 1). Sandeels varied in length<br />
from 66–231 mm, with peaks around 90, 130 and 160 mm (Figure 3c). Fish were not<br />
aged but, given that most 0-group sandeels in samples from Kittiwakes were 60–80 mm<br />
long, it is unlikely that many of these from Guillemots were 0-group. <strong>The</strong>re was no<br />
significant change in the size of sandeels taken over the period (linear regression: n =<br />
315, R 2 = 0.1%, P = 0.66). Overall the mean weights of individual sprats, gadids, and<br />
sandeels were 10.7 ± 0.1 g (n = 669), 6.7 ± 0.2 g (339) and 7.8 ± 0.3 g (280), respectively;<br />
assuming that each of these fish had lost equal weight due to desiccation, then<br />
59%, 20% and 21% of the biomass of food fed to chicks came from these fish families.<br />
SEABIRD 21 (2008): 44–54<br />
49
50<br />
<strong>Seabird</strong> diet on Canna<br />
Table 4. Fish collected from Common Guillemots Uria aalge on Canna.<br />
Gadidae Gadidae Gadidae<br />
Year Number % Sandeels % Sprats % Whiting % other species % unknown<br />
1981 5 20 40 0 40 0<br />
1982 27 56 26 0 15 0<br />
1983 23 35 48 9 4 4<br />
1984 34 15 62 6 12 6<br />
1985 24 33 50 4 13 0<br />
1986 72 69 17 4 8 0<br />
1987 22 18 73 9 0 0<br />
1988 64 48 28 13 9 2<br />
1989 74 16 35 31 18 0<br />
1990 72 32 31 28 8 1<br />
1991 106 32 35 20 12 1<br />
1992 77 1 86 10 3 0<br />
1993 76 16 49 32 4 0<br />
1994 66 18 74 6 2 0<br />
1995 87 10 71 17 1 0<br />
1996 78 14 71 10 4 1<br />
1997 27 7 85 7 0 0<br />
1998 81 20 36 40 5 0<br />
1999 76 13 70 16 1 0<br />
2000 89 21 49 26 1 1<br />
2001 113 27 52 12 7 3<br />
2002 24 8 67 25 0 0<br />
2003 56 9 50 39 2 0<br />
2004 119 19 35 45 1 0<br />
2005 11 64 18 18 0 0<br />
2006 40 40 8 13 38 3<br />
2007 19 47 5 16 16 16<br />
Total 1562 24 48 20 7 1<br />
Discussion<br />
<strong>The</strong> finding that sandeels were the main constituent (62%) of regurgitations from young<br />
Shags and adults feeding chicks accords well with numerous studies in Britain and<br />
Europe, although most reported higher frequencies (Pearson 1968; Harris & Riddiford<br />
1989; Barrett et al. 1990; Guyot 1990; Harris & Wanless 1991; Harris & Wanless 1993).<br />
However, the frequency of gadids (37%) was unexpectedly high when compared with<br />
none in 57 and 141 regurgitates from Fair Isle, Shetland in 1986-1989 and the Isle of<br />
May, Firth of Forth in 1985–1990, respectively (Harris & Riddiford 1989; Harris &<br />
Wanless 1991). <strong>The</strong>re are problems in describing the diet of Shags using otoliths, arising<br />
from the differential rates of digestion of otoliths of different sizes and from different<br />
families. For instance, otoliths from gadids are much more robust than those from<br />
sandeels and particularly sprats, whereas Butterfish and gobies have notably small<br />
otoliths for fish of their size (Härkönen 1986; Johnstone et al. 1990). However, gadids<br />
were represented in more than twice as many pellets as were sandeels and the total<br />
numbers recovered were over seven-times greater, so it seems likely that the diet of<br />
these individuals, full-grown but of unknown breeding status, that produced the pellets<br />
SEABIRD 21 (2008): 44–54
200<br />
180<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
80<br />
60<br />
40<br />
20<br />
0<br />
60<br />
40<br />
20<br />
0<br />
a) Clupeidae (n = 717)<br />
b) Gadidae (n = 384)<br />
c) Sandeels (n = 315)<br />
<strong>Seabird</strong> diet on Canna<br />
40- 50- 60- 70- 80- 90- 100- 110- 120- 130- 140- 150- 160- 170- 180- 190- 200- 210- 220+<br />
44 54 64 74 84 94 104 114 124 134 144 154 164 174 184 194 204 214<br />
Length (mm)<br />
Figure 3. Distribution of lengths (mm) of fish from Common Guillemots Uria aalge on Canna.<br />
was substantially different to that of the chicks. Such a difference in the diet of chicks<br />
and full-grown Shags has been recorded elsewhere (Harris & Wanless 1993). <strong>The</strong><br />
significant lower proportion of pellets later in the season that had sandeels present was<br />
possibly the result of sandeels leaving the water column and burying themselves in the<br />
sand, where they spend most of the rest of the summer and winter (Macer 1966).<br />
Although sandeels occurred in 60% of regurgitations from Kittiwakes and was the<br />
commonest family in 14 of the 20 years, these figures were low compared to many<br />
other British colonies. Sandeels made up an average of 83% of the biomass fed to<br />
chicks on the Isle of May over 18 years, were present in all 66 regurgitates on Fair Isle<br />
1986–88, and in 98 of 100 regurgitates on Farne Islands, Northumberland 1998–2000<br />
(Harris & Riddiford 1989; Lewis et al. 2001; Bull et al. 2004; Wanless et al. 2007).<br />
Gadids have been recorded at several other colonies, e.g. Farne Islands, Inchkeith,<br />
Inchcolm and Isle of May off northeast Britain (Bull et al. 2004), but not at such a high<br />
rate as on Canna. In contrast Sprats are regular in the diet and sometimes made up<br />
the bulk of the food at Inchkeith and Inchcolm.<br />
SEABIRD 21 (2008): 44–54<br />
51
52<br />
<strong>Seabird</strong> diet on Canna<br />
Figure 1. Common Guillemots Uria aalge with gadids, Canna, July 2006 © Kenny Graham.<br />
By number, approximately half the diet of young Guillemots was sprat, a quarter<br />
sandeels and a quarter gadids, and in biomass terms sprats contributed 59%.<strong>The</strong>se are<br />
the normal constituents of the diet of young Guillemots in northern Britain. However,<br />
normally sandeels make up the bulk of the diet with sprats the usual alternative. On<br />
the Isle of May between 1981 and 2004 the relative mean annual percentages of these<br />
two families (by number) were 56% and 42% with the remainder being made up by<br />
a variety of other families (Wanless et al. 2005). Among 303 fish brought to Fair Isle<br />
in 1986-89, only four (two Sprat and two gadids) were not sandeels (Harvey et al.<br />
1990). Gadids have been recorded from the stomachs of Guillemots shot during the<br />
summer in the 1980s in Shetland (11/83 birds with food remains), the Summer Isles<br />
(5/27), the Sound of Jura (1/5), the Clyde Front (1/5) and St Kilda (11/19) (Blake et al.<br />
1985; Halley et al. 1995). However, it is difficult to compare these with fish brought<br />
to chicks since there are differences in the diet of parent birds and their chicks even at<br />
the same time at the same colony (Wilson et al. 2004).<br />
Canna seabirds fed their chicks on a wide variety of fish species but the bulk of the<br />
diet was made up of sprats, sandeels and gadids (a variety of species but mainly<br />
Trisopterus spp. and Whiting). Other groups were of minimal importance. <strong>The</strong>re are no<br />
data available on the numbers of small fish in the waters around Canna to allow a<br />
meaningful discussion as to whether the birds were selecting these species or were<br />
just eating the main species available to them. Over the period of the study there were<br />
few significant changes in species composition or the size of prey taken by Canna<br />
seabirds and the results confirm earlier findings that gadids are a normal and<br />
important part of the diet of seabirds at this colony.<br />
SEABIRD 21 (2008): 44–54
<strong>Seabird</strong> diet on Canna<br />
Acknowledgements<br />
We thank the many people who have helped collect the field data over many years, in<br />
particular Andrew Call, Simon Foster, Alan Graham, Kenny Graham, Kathryn<br />
Mackinnon, Andrew Ramsay and Alastair Young. John Hislop gave invaluable help with<br />
fish and otolith identification, and JNCC provided financial support through their<br />
seabird monitoring programme. Rob Barrett and Kees Camphuysen improved the<br />
manuscript with their criticisms.<br />
References<br />
Barrett, R.T., Røv, N., Loen, J. & Montevecchi,W.A. 1990. Diets of Shags Phalacrocorax aristotelis<br />
and Cormorants P. carbo in Norway and possible implications for gadoid stock recruitment.<br />
Marine Ecology Progress Series 66: 205–218.<br />
Blake, B. F., Dixon, T. J., Jones, P. H. & Tasker, M. L. 1985. Seasonal changes in the feeding<br />
ecology of Guillemots (Uria aalge) off north and east Scotland. Estuarine, Coastal and Shelf<br />
Science 20: 559–568.<br />
Bull, J., Wanless, S., Elston, D. A., Daunt, F., Lewis, S. & Harris, M. P. 2004. Local-scale<br />
variability in the diet of Black-legged Kittiwakes Rissa tridactyla. Ardea 92: 43–82.<br />
Daunt, F., Wanless, S., Greenstreet, S. P. R., Jensen, H., Hamer, K. C. & Harris, M. P. 2008. <strong>The</strong><br />
impact of the sandeel fishery closure in the northwestern North Sea on seabird food consumption,<br />
distribution and productivity. Canadian Journal of Fisheries and Aquatic Sciences 65: 362–391.<br />
Frederiksen, M., Wanless, S., Harris, M. P., Rothery, P. & Wilson, L. J. 2004. <strong>The</strong> role of the<br />
industrial fishery and climate change in the decline of North Sea Black-legged Kittiwakes.<br />
Journal of Applied Ecology 41: 1129–1139.<br />
Guyot, I. 1990. Le Cormoran Huppé en Corse: Biologie et interactions avec la pêche profesionnelle.<br />
Travaux Scientifiques du Parc Naturel Regional et des Réserves Naturelles de Corse 28: 1–40.<br />
Halley, D. J., Harrison, N., Webb, A. & Thompson, D. R. 1995. Seasonal and geographical<br />
variations in the diet of Common Guillemots Uria aalge off western Scotland. <strong>Seabird</strong> 17: 12–20.<br />
Härkönen, T. 1986. Guide to the Otoliths of Bony Fishes of the Northeast Atlantic. Danbiu ApS,<br />
Hellerup, Denmark.<br />
Harris, M. P. & Riddiford, N. J. 1989. <strong>The</strong> food of some young seabirds on Fair Isle in 1986–88.<br />
Scottish Birds 15: 119–125.<br />
Harris, M. P. & Wanless, S. 1991. <strong>The</strong> importance of the Lesser Sandeel Ammodytes marinus in<br />
the diet of the Shag Phalacrocorax aristotelis. Ornis Scandinavica 22: 375–882.<br />
Harris, M. P. & Wanless, S. 1993. <strong>The</strong> diet of Shags Phalacrocorax aristotelis during the chickrearing<br />
period assessed by three methods. Bird Study 40: 135–139.<br />
Harvey, P. V., Harris, M. P., Osborn, K., Riddiford, N. & Silcocks, A. F. 1990. <strong>The</strong> breeding success<br />
and diet of Fair Isle’s seabirds in 1986–1989. Fair Isle Bird Observatory Report 42: 47–54.<br />
ICES 1995. Report of the ICES workshop on sandeel otolith analysis: Review of sandeel biology.<br />
ICES, CM 1995/G:4.<br />
Johnstone, I. G., Harris, M. P., Wanless, S. & Graves, J. A. 1990. <strong>The</strong> usefulness of pellets for<br />
assessing the diet of adult Shags Phalacrocorax aristotelis. Bird Study 37: 5–11.<br />
Lewis, S.,Wanless, S.,Wright, P. J., Harris, M. P., Bull, J. & Elston, D. A. 2001. Diet and breeding<br />
performance of Black-legged Kittiwakes Rissa tridactyla at a North Sea colony. Marine Ecology<br />
Progress Series 221: 277–284.<br />
Macer, C. T. 1966. Sandeels (Ammodytidae) in the southwestern North Sea; their biology and<br />
fishery. MAFF Fishery Investigations 2. 24: 1–51.<br />
SEABIRD 21 (2008): 44–54<br />
53
54<br />
<strong>Seabird</strong> diet on Canna<br />
Mavor, R. A., Parsons, M., Heubeck, M. & Schmitt, S. 2006. <strong>Seabird</strong> numbers and breeding<br />
success in Britain and Ireland, 2005. Joint Nature Conservation Committee, Peterborough. (UK<br />
Nature Conservation No. 30.)<br />
Monaghan, P. 1992. <strong>Seabird</strong>s and sandeels: the conflict between exploitation and conservation<br />
in the northern North Sea. Biodiversity and Conservation 1: 98–111.<br />
Montevecchi, W. A. & Piatt, J. F. 1987. Dehydration of seabird prey during transport to the<br />
colony: effects on wet weight energy densities. Canadian Journal of Zoology 65: 2822–2824.<br />
Pearson, T. H. 1968. <strong>The</strong> feeding biology of seabird species breeding on the Farne Islands,<br />
Northumberland. Journal of Animal Ecology 37: 521–552.<br />
Swann, R. L., Harris, M. P. & Aiton, D. G. 1991. <strong>The</strong> diet of some young seabirds on Canna,<br />
1981-90. <strong>Seabird</strong> 13: 54–58.<br />
Wanless, S., Frederiksen, M., Daunt, F., Scott, B. E. & Harris, M. P. 2007. Black-legged<br />
Kittiwakes as indicators of environmental change in the North Sea: Evidence from long-term<br />
studies. Progress in Oceanography 72: 30–38.<br />
Wanless, S., Harris, M. P., Redman, P. & Speakman, J. 2005. Low fish quality as a probable cause<br />
of a major seabird breeding failure in the North Sea. Marine Ecology Progress Series 294: 1–8.<br />
Wilson, L. J., Daunt, F. & Wanless, S. 2004. Self-feeding and chick provisioning diets differ in<br />
the Common Guillemot Uria aalge. Ardea 92: 197–208.<br />
Appendix 1. <strong>The</strong> mean lengths (and their standard error, SE) of fish from collected from Common Guillemots<br />
Uria aalge on Canna.<br />
Sandeels Sprats Whiting<br />
Year Number Mean (mm) SE Number Mean (mm) SE Number Mean (mm) SE<br />
1983 10 112.2 0.8<br />
1984 21 114.1 1.7<br />
1985 7 147.6 8.7 11 111.1 1.5<br />
1986 48 112.7 4.3 12 126.0 3.6<br />
1987 16 106.4 1.7<br />
1988 20 149.2 6.5 16 111.6 2.3<br />
1989 12 154.0 9.1 26 112.2 1.8<br />
1990 23 125.6 7.5 22 114.4 1.0<br />
1991 34 134.8 6.4 36 112.3 1.2 21 86.0 1.8<br />
1992 1 143.0 64 113.1 1.3 8 84.8 3.3<br />
1993 12 157.9 8.5 36 113.5 1.2 24 83.0 2.0<br />
1994 12 141.3 11.3 49 110.6 1.5 4 89.0 6.5<br />
1995 8 162.1 7.2 62 111.2 0.8 15 92.7 2.2<br />
1996 11 132.5 13.7 55 111.2 0.7 8 94.4 3.4<br />
1997 2 162.5 37.5 23 111.8 2.3 2 92.0 12.0<br />
1998 16 126.6 8.3 29 116.8 1.5 32 90.2 3.7<br />
1999 10 97.2 8.5 53 105.6 1.2 12 86.8 5.1<br />
2000 19 126.8 11.4 42 112.8 1.5 22 103.1 2.5<br />
2001 23 103.3 6.7 47 106.5 1.4 13 93.4 2.9<br />
2002 2 105.0 12.0 16 108.7 1.6 6 100.2 5.0<br />
2003 5 143.6 27.9 22 115.4 2.4 22 92.1 2.1<br />
2004 23 127.8 6.5 42 102.4 1.6 52 88.7 1.8<br />
2005 7 148.9 10.9 2 110.5 0.5 2 82.0 11.0<br />
2006 16 153.9 6.7 3 110.7 5.3 5 75.6 9.2<br />
2007 4 151.8 9.4 1 119.0 3 66.0 8.1<br />
SEABIRD 21 (2008): 44–54
Colony habitat selection by Little Terns<br />
Colony habitat selection by Little Terns<br />
Sternula albifrons in East Anglia:<br />
implications for coastal management<br />
Ratcliffe, N. 1 *, Schmitt, S. 2 ,Mayo,A. 2 , Tratalos, J. 3 and Drewitt, A. 4<br />
*Correspondence author. Email: notc@bas.ac.uk<br />
1 RSPB, East Scotland Regional Office, 10 Albyn Terrace, Aberdeen AB10 1YP, UK (Current<br />
address: British Antarctic Survey, NERC, High Cross, Cambridge CB3 0ET, UK); 2 RSPB, <strong>The</strong><br />
Lodge, Sandy, Bedfordshire SG19 2DL, UK; 3 University of East Anglia, Norwich NR4 7TJ, UK;<br />
4 English Nature, Northminster House, Peterborough PE1 1UA, UK.<br />
Abstract<br />
Little Terns Sternula albifrons are unusual among UK seabirds in that a large proportion<br />
of the population breed on mainland beaches in East Anglia. Relative sea-level rise<br />
means that such habitats are under threat in this region, and so we quantified colony<br />
habitat selection of beach nesting Little Terns in order to inform habitat restoration<br />
and creation initiatives. Random 1 km sections of beach were selected and the<br />
presence or absence of a Little Tern colony within each was related to physical<br />
(substrate type, height and width), biotic (vegetation cover, predator activity) and<br />
anthropogenic (disturbance) characteristics using logistic regression models. Little<br />
Terns positively selected for beaches with vegetation cover and negatively for those<br />
with high disturbance levels. <strong>The</strong>y showed no selection according to width and height<br />
or Red Fox Vulpes vulpes presence, even though these are likely to affect flood and<br />
predation risk respectively. Red Foxes were found to be widespread on beaches<br />
irrespective of tern colony presence, and so movement of tern colonies will not always<br />
result in predator avoidance. Little Tern habitat creation needs to be integrated into<br />
coastal management plans in order to safeguard their population from the combined<br />
threats of relative sea-level rise, predation and disturbance.<br />
Introduction<br />
<strong>The</strong> population size and range of Little Terns Sternula albifrons in the UK have declined<br />
by 24% since the mid 1980s (Pickerell 2004), and qualifies it for inclusion on the Amber<br />
list based on a moderate rate of population decline (Gregory et al. 2002). It is also listed<br />
on Annex I of the Birds Directive (EC Directive on the conservation of wild birds<br />
79/409/EEC) and 67% of the GB population lies within Special Protection Areas (Stroud<br />
et al. 2001). Little Tern distribution through the UK has been broadly stable since the<br />
1960s, being wide and patchy with a stronghold in East Anglia where 69% of the<br />
population currently breed (Pickerell 2004).<strong>The</strong>re is an absence of sandy offshore barrier<br />
islands in the UK, and so Little Terns nest mainly on low-lying shingle beaches, spits and<br />
estuarine islets, and here they face threats from disturbance by coastal tourism,<br />
predation and inundation by high tides or storms (Ratcliffe 2004). <strong>The</strong>se threats are<br />
mitigated to varying extents by intensive colony protection schemes (Smart 2004) that<br />
currently cost conservation organisations tens of thousands of pounds per annum, and<br />
SEABIRD 21 (2008): 55–63<br />
55
56<br />
Colony habitat selection by Little Terns<br />
need to be run in perpetuity in order to maintain Little Tern populations. Furthermore,<br />
the colony habitat of Little Terns is under threat from relative sea level rise and increased<br />
storminess (Norris & Buisson 1994; Gill 2004), such that protection of existing colonies<br />
from disturbance and predation may be inadequate to conserve them in the future.<br />
Rather than continuing to focus on colony site protection, a strategic approach to the<br />
management of Little Tern distribution relative to threats may result in more effective<br />
conservation. Little Terns require small areas of habitat that can be restored or created<br />
quickly and inexpensively, and these can be sited in areas where threats to breeding<br />
success are lower. For example, creation of dredge spoil islands offshore could reduce<br />
predation and disturbance without need for fencing and wardening (Parnell et al. 1986;<br />
Burgess & Hirons 1992; Erwin et al. 2001), and sediment recharge of eroded narrow<br />
beaches could mitigate loss of habitat or colony flooding risks (Charlton & Allcorn<br />
2004). <strong>The</strong> characteristics of Little Tern colony sites need to be quantified and means<br />
of creation or restoration have to be tested in order to achieve this (Gill 2004).<br />
Shoreline management plans are currently under review for East Anglia and decisions are<br />
being made that will affect coastal habitats in future decades. <strong>The</strong>se decisions are mainly<br />
influenced by engineering and economic considerations at present, but it is important<br />
that conservation requirements are also included (Gill 2004). A strategic review of Little<br />
Tern habitat requirements would be timely in order to influence coastal planning decisions<br />
in a manner that is sympathetic towards Little Terns and to exploit opportunities that may<br />
arise as a by-product of managed realignment or coastal defence (Gill 2004).<br />
This project examines colony habitat selection by Little Terns in relation to beach<br />
characteristics, including width, height, shoreward habitat type, substrate type,<br />
disturbance and predation risk.<strong>The</strong> implications of the results for habitat management<br />
and creation are discussed and recommendations for further research are made.<br />
Study sites and methods<br />
<strong>The</strong> entire length of beaches in Norfolk and Suffolk (east England) were divided into 1<br />
km lengths using GIS, and these were used as the sampling unit for further data<br />
collection and analyses.<br />
Beach habitat characteristics: Beach characteristics were studied in 68 randomly<br />
selected 1 km sample sections of beach in Norfolk and Suffolk, representing 33% of<br />
the entire beach length within these counties. Beach width from the landward limit of<br />
the beach to mean high water mark was extracted every 200 m from Ordnance Survey<br />
maps using GIS. Data on maximum beach height was extracted from Environment<br />
Agency beach profile databases. Substrate was quantified within 30 2.5 x 2.5 m<br />
quadrats placed randomly along the beach.<strong>The</strong> percentage cover of sand, shingle, rock,<br />
debris and vegetation was recorded in each. <strong>The</strong> values were averaged within each of<br />
the 1 km sample units. Fieldwork was conducted in April 2003 prior to the return of<br />
Little Terns to their colonies to avoid disturbance. <strong>The</strong> presence or absence of breeding<br />
Little Terns in each sample was determined by walking and scanning each section of<br />
coast three times during May and June 2003 when they would be incubating eggs.<br />
SEABIRD 21 (2008): 55–63
Colony habitat selection by Little Terns<br />
Disturbance: Observations of disturbance within a subsample of 45 of the previous 68<br />
beach sections were conducted, during May and June 2003, by walking each 1 km<br />
section of beach and counting the number of people and dogs. Each section was<br />
visited three times, with samples being taken at both weekdays and weekends to<br />
account for variation in disturbance owing to day of the week. Samples were only<br />
taken on fine days that were conducive to beach recreation to reduce variation owing<br />
to weather.<strong>The</strong>se counts were averaged to provide an estimate of disturbance pressure<br />
within the sampled 1 km section.<br />
Dummy nests were used to assess relative trampling risk in 40 sections of beach during<br />
late May and early June. Twenty Common Quail Coturnix coturnix eggs were each<br />
placed in separate artificial scrapes along the length of the beach section above the<br />
high tide line and were marked with a numbered peg. <strong>The</strong>se were visited every 3–5<br />
days over a period of 23 days to determine whether they were stepped upon by<br />
humans. Daily trampling rates for each section were obtained using the Mayfield<br />
method (Mayfield 1975).<br />
Predation: Predation risk to Little Terns was determined during May and June 2003 by<br />
using sand traps and dummy nests. Sand traps were deployed at 30 sections of beach<br />
sampled from the previous 45, with five traps being set within each. <strong>The</strong>se comprised<br />
2.5 m 2 of smoothed-over sand with the perimeter marked with lines of pebbles to<br />
enable it to be relocated. A spoonful of cat food was placed in the centre of this. <strong>The</strong><br />
traps were set in the evening and visited early in the morning and any footprints<br />
present were identified by reference to Bang & Dahlstrom (2001). Five traps were set<br />
on each of three nights in each sample section, with Red Foxes Vulpes vulpes being<br />
classed as present if footprints were found in a trap on one or more occasions.<br />
Daily nest predation rates were assessed using the dummy egg experiment described<br />
above, except that the numbers of nests predated were analysed as opposed to<br />
numbers trampled.<br />
Analysis: <strong>The</strong> presence or absence of Little Terns in a sample stretch of beach was<br />
related to the explanatory variables using logistic regression. We used presence or<br />
absence rather than number of breeding pairs as an index of suitability because we<br />
were interested in correlates of colony site selection rather than colony size. Moreover,<br />
abundance of Little Terns is more annually variable than site occupancy, and<br />
abundance in colonial birds is prone to non-independence (i.e. pairs selecting a site<br />
because of the presence of conspecifics, rather than owing to habitat characteristics).<br />
A forward stepwise procedure was used to select the minimal adequate model, using a x2<br />
alpha of 0.05 as the threshold for inclusion of an explanatory variable in the model. Chisquare<br />
statistics for each explanatory variable in the results are taken from the minimal<br />
adequate model. It was not possible to build a global model of all variables of interest<br />
since owing to the uneven sampling effort among parameters caused by manpower<br />
limitations (see sections above), and so separate models were fitted for each. Observed<br />
frequencies and model fits were plotted graphically according to Smart et al. (2004).<br />
SEABIRD 21 (2008): 55–63<br />
57
Likelihood of Occupancy<br />
0.8<br />
0.7<br />
0.6<br />
0.5<br />
0.4<br />
0.3<br />
0.2<br />
58<br />
Colony habitat selection by Little Terns<br />
Results<br />
Beach characteristics: Nineteen of the 68 samples were occupied by Little Tern<br />
colonies. <strong>The</strong> likelihood of a beach being used by breeding Little Terns increased significantly<br />
with percentage vegetation cover (x2 1 = 3.77, P = 0.05). Occupancy likelihood<br />
was 23% in the absence of vegetation, but rose to 70% when coverage was 20% (Figure<br />
1). Fitting the square of vegetation cover to the model did not significantly reduce the<br />
residual deviance, suggesting that there was no decline as vegetation became more<br />
dense within the range of cover observed. Fitting linear and quadratic models that<br />
included the variables beach height, beach width, the percentage cover of sand, gravel,<br />
rock or debris did not significantly reduce the residual deviance (all P > 0.05).<br />
Disturbance: Occupancy likelihood declined significantly with the average number of<br />
people observed per km of beach (x2 1 = 4.5, P < 0.05), but not with the number of<br />
dogs (x2 1 = 1.30, P > 0.2). Occupancy likelihood was 47% in the absence of people but<br />
fell to zero when the average number of people exceeded 25 (Figure 2).<br />
Daily nest trampling rates of dummy clutches were 1.33%, equating to a 26% survival<br />
rate over a 23-day incubation period, in the absence of other forms of loss. <strong>The</strong><br />
likelihood of a site being occupied by Little Terns was not significantly affected by nest<br />
trampling rates of dummy clutches within sample sites (x2 1 = 2.6, P > 0.1).<br />
Predation risk: Sand traps revealed that foxes were present on 62% of beaches, but<br />
the likelihood of a beach being occupied by breeding Little Terns was not significantly<br />
affected by the presence of foxes (x2 1 = 0.03, P > 0.8). Daily nest predation rates on<br />
dummy clutches were 1.08%, equating to a survival over a 23-day incubation period<br />
of 22%. Site occupancy was not explained by predation rates experienced by dummy<br />
clutches within the sampled stretches of beach (x2 1 = 0.00, P > 0.9).<br />
0.1<br />
0 0.0<br />
0<br />
0 5 10 15 20 25 30<br />
0 20 40 60 80 100<br />
Percentage Vegetation Cover<br />
Average Number of People<br />
Figure 1 (left). <strong>The</strong> effect of percentage vegetation cover on the likelihood of a site being occupied by breeding<br />
Little Terns Sternula albifrons. <strong>The</strong> line represents the best fit line from the logistic regression model plotted on the<br />
left y axis , while the histograms show observed frequencies of presence (on the top x axis, which is inverted) and<br />
absence (on the bottom x axis) plotted against the right hand y axis (see Smart et al. 2004). Figure 2 (right). <strong>The</strong><br />
effect of human disturbance pressure on the likelihood of a site being occupied by breeding Little Terns Sternula<br />
albifrons. See the legend of Figure 1 for details.<br />
SEABIRD 21 (2008): 55–63<br />
0<br />
10<br />
20<br />
20<br />
10<br />
Frequency<br />
Likelihood of Occupancy<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0<br />
5<br />
Frequency<br />
5
Colony habitat selection by Little Terns<br />
Discussion<br />
While numbers of breeding pairs and breeding success within Little Tern colonies in<br />
East Anglia varies markedly among years, the locations of occupied sites has remained<br />
remarkably consistent since regular monitoring began in the 1960s (Pickerell 2004;<br />
Ratcliffe 2004). This suggests colony sites are selected consistently according to<br />
temporally stable habitat characteristics rather than selected at random or in response<br />
to previous breeding success. This is the first study to attempt to identify the physical<br />
factors driving colony site selection by Little Terns at a regional scale within the UK.<br />
Of the physical beach habitat characteristics measured, only vegetation cover had a<br />
significant positive effect on colony site selection. Similarly, Goutner (1990) found<br />
Little Terns nested near low vegetation in Greece, and the closely related and ecologically<br />
similar Least Terns S. antillarum also selected sites with sparse vegetation cover<br />
(Kotliar & Burger 1986). However, thick vegetation cover greater than the 30%<br />
maximum found in this study might result in avoidance: Burger & Gochfeld (1990)<br />
found that Least Tern nests were closer to vegetation than random points where cover<br />
was sparse and further away where cover was dense.<br />
Studies of Least Tern colony site selection in the USA (Thomas & Slack 1982; Gochfeld<br />
1983; Kotliar & Burger 1986; Burger & Gochfeld 1990) have shown a preference for<br />
mixed sand and gravel or shell substrates (which aids camouflage of nests) and high,<br />
wide beaches (which reduce the risk of tidal inundation). However, in this study,<br />
substrate and beach height/width variables had no effect on colony site selection by<br />
Little Terns. This may have been because these factors are not selected for by Little<br />
Terns in East Anglia or that power to detect such effects was insufficient owing to<br />
sample size or co-linearity of explanatory variables. Alternatively, selection for beach<br />
characters may have been obscured by another factor that was not measured, with the<br />
most likely of these being food availability.<br />
Little Terns have very short foraging ranges compared to most seabirds, with most food<br />
generally being obtained from within 5 km of the colony and 1 km of the shore (Davies<br />
1981; Fasola & Bogliani 1990; Allcorn et al. 2003; Bertolero et al. 2005; Perrow et al.<br />
2006). Consequently, Little Terns have to nest close to areas of high prey availability so<br />
that they can provision their chicks with sufficient food to maintain their growth and<br />
survival (Holloway 1993; Bertolero et al. 2005). Little Terns in the UK generally feed on<br />
small (30–40 mm) clupeids and sandeel (Davies 1981; Cramp 1985; Perrow et al. 2004),<br />
and on the east Norfolk coast selected sites where the availability of such prey was<br />
greatest (Perrow et al. 2004).As such, significant colony site selection according to beach<br />
characteristics might only be evident when controlling for offshore food availability.<br />
Collecting data on prey availability at the broad scales required to understand regional<br />
colony site selection is prohibitively expensive, but there may be potential to infer this<br />
from marine habitat variables such as turbidity (Perrow et al. 2004).<br />
Little Terns showed a significant aversion to human disturbance, as noted elsewhere in<br />
Europe (Scarton et al. 1994; Catry et al. 2003), and for Least Terns in the USA (Thomas<br />
& Slack 1982; Gochfeld 1983; Kotliar & Burger 1986). Disturbance by humans and<br />
SEABIRD 21 (2008): 55–63<br />
59
60<br />
Colony habitat selection by Little Terns<br />
dogs has the potential to cause trampling or predation of eggs or chicks, or to disrupt<br />
courtship, incubation or brooding by parent birds, causing reduced breeding success<br />
and site abandonment (Thomas & Slack 1982; Gochfeld 1983; Kotliar & Burger 1986).<br />
Indeed, our dummy nest experiment showed that trampling on East Anglian beaches<br />
would destroy 74% of clutches over a 23-day incubation period in the absence of<br />
other sources of loss. Humans are active on beaches during both the day and the<br />
colony-prospecting period of Little Terns, and so birds are able to perceive this pressure<br />
and select sites accordingly. Continuing to maintain protective cordons around Little<br />
Tern colonies is therefore essential to prevent their extirpation, but there may also be<br />
potential to reduce disturbance in other areas that are deemed suitable to encourage<br />
colonisation. However, excluding tourists from popular areas of beach will be difficult<br />
and politically contentious, especially in the context of UK Government’s objective to<br />
increase public access to the coast (Defra 2004).<br />
Predation is among the most important causes of breeding failure at Little Tern<br />
colonies, with Red Foxes being the predator most frequently recorded to cause<br />
significant losses of both eggs and chicks (Ratcliffe 2004). Little Terns did not select<br />
colony sites according to the occurrence of Red Foxes or the rates of predation failure<br />
of dummy clutches, and so their ability to perceive, or scope to avoid, predation risks<br />
appeared to be poor.This is probably because the presence of nocturnal predators only<br />
becomes apparent to terns during the incubation or chick rearing stage when they<br />
have committed to breeding at a site.<br />
Red Foxes were present on a large proportion of the beaches despite the fact that<br />
trapping effort was relatively low. Red Foxes have increased their range in East Anglia<br />
owing to reduced gamekeeping effort (Tapper 1992) and this has probably resulted in<br />
the reduced availability of predator-free colony sites for ground-nesting birds.<br />
Similarly, declines of beach-nesting terns in Virginia were associated with an increased<br />
occupancy of barrier islands by Red Foxes and Racoons Procyon lotor and a consequent<br />
reduction in available predator-free nesting sites (Erwin et al. 2001). Red Fox presence<br />
was not related to Little Tern presence or absence, suggesting that they are not<br />
venturing onto beaches solely to exploit tern prey. It is therefore likely that Red Foxes<br />
forage on most East Anglian beaches to feed on tide-line carrion and scraps dropped<br />
by tourists, and will opportunistically prey on tern colonies if they encounter them.<br />
This discredits the widely held belief that high predation rates are an indirect<br />
consequence of disturbance, which reduces habitat availability and hence the ability<br />
of birds to move to avoid predators. This assumes that Red Foxes only forage on<br />
beaches following a chance encounter with a tern colony, whereas the data presented<br />
in this study suggest they already occur along most stretches of beach and will<br />
opportunistically predate nests or chicks in any new colonies forming there.<br />
<strong>The</strong> combined effects of disturbance avoidance and predation losses mean that<br />
intensive protection of mainland beach sites will be required in perpetuity if Little Tern<br />
populations are to be maintained on existing habitat. However, an alternative in the<br />
longer term is to create islets using dredge spoil which would enable Little Terns to<br />
nest in sites that are free from predators and people (Burgess & Hirons 1992; Krogh &<br />
SEABIRD 21 (2008): 55–63
Colony habitat selection by Little Terns<br />
Schweitzer 1999). This method has been successfully employed the USA and now a<br />
large proportion of some regional tern populations nest on artificial islands (Parnell et<br />
al. 1986; Visser & Peterson 1994). In southern England, Little Terns typically breed on<br />
natural islets in saline lagoons or estuaries rather than beaches (Pickerell 2004), but<br />
such habitats are unavailable throughout much of East Anglia. Further research into<br />
colony habitat selection for islets is required to ensure new sites are designed<br />
appropriately. In particular, size is likely to be an important consideration, since larger<br />
islands are attractive to other species of tern and gulls, which can result in competitive<br />
exclusion of Little Terns (Fasola & Canova 1992; Scarton et al. 1994).<br />
Correct location of the islets will be essential to ensure that they are colonised and to<br />
maximise breeding success. Positioning them within foraging range of concentrations<br />
of available prey will be essential, and studies of turbidity combined with surveys of<br />
prey availability using a fine-mesh, epipelagic trawl net (Perrow et al. 2004) would be<br />
required to achieve this. However, other factors have to be taken into account for<br />
practical and legal reasons. For example, construction of islets in some locations might<br />
be unfeasible as they might cause a hazard to navigation, suffer excessive erosion or<br />
contravene legislation protecting designated nature conservation sites (Harrison &<br />
Allcorn 2004). Conversely, offshore coastal defence structures or remnants of breached<br />
sea walls may present opportunities for site creation (Charlton & Allcorn 2004). Once<br />
islands of a suitable design are created in appropriate locations, Little Terns can be<br />
attracted to them using decoys and recordings of courtship calls (Kotliar & Burger<br />
1984; Burger 1988; Jeffries & Brunton 2001; Colombé & Allcorn 2004), which will<br />
increase the likelihood of them being colonised.<br />
Further work is required to develop an improved understanding of Little Tern habitat<br />
requirements and to assess how a network of colony sites can be delivered within the<br />
constraints and opportunities presented by wider coastal planning initiatives. While<br />
this work is in progress, it is essential that protection at existing sites is established and<br />
maintained in order to prevent their extirpation and further population declines.<br />
Acknowledgements<br />
This work was funded by a partnership between RSPB and English Nature. We are<br />
grateful to the wardens of tern colonies in Norfolk and Suffolk for granting permission<br />
to work on their sites and for assisting with data collection. <strong>The</strong> Environment Agency<br />
kindly provided access to the beach profile data. We thank the two anonymous<br />
referees for constructive criticism of an earlier draft.<br />
References<br />
Allcorn, R. I., Eaton, M. A. Cranswick, P. A., Perrow, M., Hall, C., Smith, L., Reid, J., Webb, A.,<br />
Smith, K.W. S., Langston, R. H.W. & Ratcliffe, N. 2003. A pilot study of breeding tern foraging<br />
ranges in NW England and East Anglia in relation to potential development areas for offshore<br />
windfarms. RSPB/WWT/JNCC report to DTI, Sandy, UK.<br />
Bang, P. & Dahlstrom, P. 2001. Animal Tracks and Signs. Oxford University Press, Oxford.<br />
Bertolero,A., Oro, D.,Vilalta,A. M. & López, M. À. 2005. Selection of foraging habitats by Little<br />
Terns Sterna albifrons at the Ebro Delta, NE Spain. Revista Catalana d’Ornitologia 21: 37–42.<br />
SEABIRD 21 (2008): 55–63<br />
61
62<br />
Colony habitat selection by Little Terns<br />
Burger, J. 1988. Social attraction in nesting Least Terns: effects of numbers, spacing and pair<br />
bonds. Condor 90: 575–582.<br />
Burger, J. & Gochfeld, M. 1990. Nest site selection in Least Terns Sterna antillarum in New<br />
Jersey and New York. Colonial Waterbirds 13: 31–40.<br />
Burgess, N. D. & Hirons, J. M. 1992. Creation and management of artificial nesting sites for<br />
wetland birds. Journal of Environmental Management 34: 285–295.<br />
Catry, T., Ramos, J. A., Catry, I., Allen-Revez, M. & Grade, N. 2003. Are salinas a suitable<br />
alternative breeding habitat for Little Terns Sterna albifrons? Ibis 145: 258–268.<br />
Charlton, P., & Allcorn, R. I. 2004. Habitat creation for Little Terns. In: Allcorn, R. I. (ed.).<br />
Proceedings of a Symposium on Little Terns Sterna albifrons: 32–38. RSPB Research Report No.<br />
8. Sandy, UK.<br />
Colombé, S. & Allcorn, R. I. 2004. An overview of seabird social facilitation projects in the UK,<br />
US and elsewhere with particular reference to the Sterna terns. In: Allcorn, R.I. (ed.). Proceedings<br />
of a Symposium on Little Terns Sterna albifrons: 60–65. RSPB Research Report No. 8. Sandy, UK.<br />
Cramp, S. (ed.) 1985. <strong>The</strong> Birds of the Western Palearctic. Vol. IV. Oxford University Press, Oxford.<br />
Davies, S. 1981. Development and behaviour of Little Tern chicks. British Birds 74: 291–298.<br />
Defra 2004. Delivering the essentials of life: Defra’s five year strategy. Department of<br />
Environment, Food and Rural Affairs report to Government. HMSO, Norwich.<br />
Erwin, R. M., Truitt, B. R. & Jimenez, J. E. 2001. Ground-nesting waterbirds and mammalian<br />
carnivores in the Virginia Barrier Islands region: running out of options. Journal of Coastal<br />
Research 17: 292–296.<br />
Fasola, M. & Bogliani, G. 1990. Foraging ranges of an assemblage of Mediterranean seabirds.<br />
Colonial Waterbirds 13: 72–74.<br />
Fasola, M. & Canova, L. 1992. Nest habitat selection by eight syntopic species of<br />
Mediterranean gulls and terns. Colonial Waterbirds 15: 169–178.<br />
Gill, J. 2004. Potential impacts of climate change on Little Terns in East Anglia. In: Allcorn, R. I.<br />
(ed.). Proceedings of a Symposium on Little Terns Sterna albifrons: 29–31. RSPB Research Report<br />
No. 8. Sandy, UK.<br />
Gochfeld, M. 1983. Colony site selection by Little Terns: physical attributes of sites. Colonial<br />
Waterbirds 6: 205–213.<br />
Goutner, V. 1990. Habitat selection of Little Terns in the Evros Delta, Greece. Colonial<br />
Waterbirds 13: 108–114.<br />
Gregory, R. D., Wilkinson, N. I., Noble, D. G., Robinson, J. A., Brown, A. F., Hughes, J., Proctor,<br />
D. A., Gibbons, D. W. & Galbraith, C. A. 2002. <strong>The</strong> population status of birds in the United<br />
Kingdom, Channel Islands and the Isle of Man: an analysis of conservation concern 2002–2007.<br />
British Birds 95: 410–450.<br />
Harrison, D. & Allcorn, R. I. 2004. Legislative framework affecting habitat creation in the UK.<br />
In: Allcorn, R. I. (ed.). Proceedings of a Symposium on Little Terns Sterna albifrons: 66–72. RSPB<br />
Research Report No. 8. Sandy, UK.<br />
Holloway, M. 1993. <strong>The</strong> variable breeding success of the Little Tern Sterna albifrons in South-<br />
East India and protective measured needed for its conservation. Biological Conservation 65: 1–8.<br />
Jeffries, D. S. & Brunton, D. H. 2001. Attracting endangered species to “safe” habitats: responses<br />
of Fairy Terns to decoys. Animal Conservation 4: 301–305.<br />
Kotliar, N. B. & Burger, J. 1984. <strong>The</strong> use of decoys to attract Least Terns (Sterna antillarum) to<br />
abandoned colony sites in New Jersey. Colonial Waterbirds 7: 134–138.<br />
SEABIRD 21 (2008): 55–63
Colony habitat selection by Little Terns<br />
Kotliar, N. B. & Burger, J. 1986. Colony site selection and abandonment by Least Terns Sterna<br />
antillarum in New Jersey, USA. Biological Conservation 37: 1–21.<br />
Krogh, M. G. & Schweitzer, S. H. 1999. Least Terns nesting on natural and artificial habitats in<br />
Georgia, USA. Waterbirds 22: 290–296.<br />
Mayfield, H. 1975. Suggestions for calculating nest success. Wilson Bulletin 87: 456–466.<br />
Norris, K. J. & Buisson, R. 1994. Sea-level rise and its impact upon coastal breeding birds in the<br />
UK. RSPB Conservation Review 8: 63–71.<br />
Parnell, J. F., Needham, R. N., Soots, R. F. Jr., Fussel, J. O. III, McCrimmon, D. A. Jr., Bork, R. D.<br />
& Shields, M. A. 1986. Use of dredged-material deposition sites by birds in coastal North<br />
Carolina, USA. Colonial Waterbirds 9: 210–217.<br />
Perrow, M. R., Tomlinson, M. L., Lines, P., Benham, K., Howe, R. & Skeate, E. R. 2004. Is food<br />
supply behind Little Tern Sterna albifrons colony location? <strong>The</strong> case of the largest colony in the<br />
UK at the North Denes/Winterton SPA in Norfolk. In: Allcorn, R. I. (ed.). Proceedings of a<br />
Symposium on Little Terns Sterna albifrons: 39–59. RSPB Research Report No. 8. Sandy, UK.<br />
Perrow, M. R., Skeate, E. R., Lines, P., Brown, D. & Tomlinson, M. L. 2006. Radio telemetry as a<br />
tool for impact assessment of wind farms: the case of Little Terns Sterna albifrons at Scroby<br />
Sands, Norfolk, UK. Ibis 148: S1 57-75.<br />
Pickerell, G. 2004. Little Tern Sterna albifrons. In: Mitchell, P. I., Newton, S., Ratcliffe, N. and<br />
Dunn, T.E. (eds.). <strong>Seabird</strong> populations of Britain and Ireland: 339–349. Poyser, London.<br />
Ratcliffe, N. 2004. Little Terns in Britain and Ireland: estimation and diagnosis of population<br />
trends. In: Allcorn, R. I. (ed.). Proceedings of a Symposium on Little Terns Sterna albifrons: 4–18.<br />
RSPB Research Report No. 8. Sandy, UK.<br />
Scarton, F., Valle, R. & Borella, S. 1994. Some comparative aspects of the breeding biology of<br />
Black-headed Gull (Larus ridibundus), Common Tern (Sterna hirundo) and Little Tern (Sterna<br />
albifrons) in the Lagoon of Venice, NE Italy. Avocetta 18: 199–123.<br />
Smart, J. 2004. Managing colonies for Little Terns. In: Allcorn, R. I. (ed.). Proceedings of a<br />
Symposium on Little Terns Sterna albifrons: 19–28 RSPB Research Report No. 8. Sandy, UK.<br />
Smart, J., Sutherland, W. J., Watkinson, A. R. & Gill, J. A. 2004. A new means of presenting the<br />
results of logistic regression. Bulletin of the Ecological Society of America 85: 100–102.<br />
Stroud, D. A, Chambers, D., Cook, S., Buxton, N., Fraser, B., Clement, P., Lewis, P., McLean, I.,<br />
Baker, H. & Whitehead, S. 2001. <strong>The</strong> UK SPA network: its scope and content. Volume 2: species<br />
accounts. Joint Nature Conservation Committee, Peterborough, UK.<br />
Tapper, S. C. 1992. Game heritage: an ecological review from shooting & gamekeeping records.<br />
<strong>The</strong> Game Conservancy, Fordingbridge, UK.<br />
Thomas, B. C. and Slack, D. 1982. Physical aspects of colony selection by Least Terns on the<br />
Texas coast. Colonial Waterbirds 5: 161–168.<br />
Visser, J. M. & Peterson, G. W. 1994. Breeding populations & colony site dynamics of seabirds<br />
nesting in Louisiana. Colonial Waterbirds 17: 146–152.<br />
SEABIRD 21 (2008): 55–63<br />
63
64<br />
Subcutaneous air diverticula of Northern Gannet<br />
Descriptive anatomy of the subcutaneous<br />
air diverticula in the Northern Gannet<br />
Morus bassanus<br />
Daoust, P.-Y. 1 *, Dobbin, G. V. 1 , Ridlington Abbott, R. C. F. 2 & Dawson, S. D. 3<br />
*Correspondence author. Email: daoust@upei.ca<br />
1 Department of Pathology & Microbiology, Atlantic Veterinary College, University of<br />
Prince Edward Island, 550 University Avenue, Charlottetown, PE C1A 4P3, Canada<br />
(Current address for G. V. D: Department of Laboratory Medicine, Queen Elizabeth<br />
Hospital, P.O. Box 6600, Charlottetown, PE C1A 8T5, Canada); 2 Class of 2005,<br />
Department of Biology, Faculty of Science, University of Prince Edward Island, 550<br />
University Avenue, Charlottetown, PE C1A 4P3, Canada (Current address: 772 Osborne<br />
Street, Summerside, PE C1N 4N5, Canada); 3 Department of Biomedical Sciences,<br />
Atlantic Veterinary College, University of Prince Edward Island, 550 University Avenue,<br />
Charlottetown, PE C1A 4P3, Canada.<br />
Abstract<br />
Northern Gannets Morus bassanus typically forage by diving from high above the<br />
water surface. <strong>The</strong>ir subcutaneous (s-c) tissues are invested by an elaborate system of<br />
air diverticula that presumably function in cushioning the impact of their entry into<br />
the water. <strong>The</strong> anatomical details of this system were studied by dissection and latex<br />
injection in 15 carcasses of these birds. <strong>The</strong> s-c air diverticula consist mainly of two<br />
independent systems of intercommunicating compartments that are bilaterally<br />
symmetrical, cover the ventral and lateral regions of the trunk and the proximal<br />
portions of the wings and legs, and communicate with the ipsilateral region of the<br />
clavicular respiratory air sac. This communication, which opens into the axillary region,<br />
is through a narrow gap between the subcoracoideus and coracobrachialis caudalis<br />
muscles. Two other, smaller, independent systems of s-c air diverticula, also bilaterally<br />
symmetrical, may contribute to cushioning the Northern Gannet’s body during its<br />
dives: one at the thoracic inlet, which communicates with the corresponding side of<br />
the clavicular air sac, and the other along the neck, which communicates with the<br />
nasal cavities and the choanal opening. Further work is required to define more<br />
precisely the function of these extensive air diverticula and air circulation within them.<br />
Introduction<br />
<strong>The</strong> avian respiratory system is the most efficient among those of all air-breathing<br />
vertebrates and is unique in its basic structure (King & McLelland 1984). Its extensive<br />
system of air sacs allows a near-continuous flow of fresh air through the pulmonary air<br />
capillaries at countercurrent to the blood circulation and throughout the respiratory<br />
Footnote: Definition of terms used in the text for anatomical orientation: cranial, toward the head; caudal,<br />
toward the tail; ventral, toward the front of the body; dorsal, toward the back of the body; proximal, closer<br />
to the centre of the body; distal, farther from the centre of the body; medial, closer to the body’s midline;<br />
lateral, farther from the body’s midline; rostral, toward the beak or tip of the beak.<br />
SEABIRD 21 (2008): 64–76
Subcutaneous air diverticula of Northern Gannet<br />
cycle. Most avian species have four paired air sacs (cervical, cranial thoracic, caudal<br />
thoracic, abdominal) and one unpaired air sac (clavicular). Depending on the species,<br />
some of these air sacs can project complex systems of diverticula between muscles and<br />
into the subcutis and pneumatic bones of the trunk, pectoral and pelvic girdles, and limbs<br />
(McLelland 1989; O’Connor 2004). Some members of the order Pelecaniformes have an<br />
elaborate and extensive system of subcutaneous (s-c) air diverticula. Northern Gannets<br />
Morus bassanus, which typically forage by diving from heights of up to 30 m above water<br />
and reaching speeds of up to 100 km/h on impact with water, are thought to use these<br />
s-c air diverticula as a means of cushioning this impact (Montagu 1813; Gurney 1913;<br />
Nelson 1978). It is not known, however, whether these diverticula are inflated voluntarily<br />
prior to diving or whether air is simply prevented from exiting them as the bird hits the<br />
water. Regardless, an efficient communication is likely needed between the respiratory<br />
tract and the system of s-c air diverticula and among the various compartments of this<br />
system. Subcutaneous air diverticula were described, albeit only partially, many years ago<br />
in the Northern Gannet (Montagu 1813; Gurney 1913) and in the Brown Pelican<br />
Pelecanus occidentalis (Richardson 1939). According to the study of the Brown Pelican<br />
by Richardson (1939), the communication between the respiratory system, specifically<br />
the clavicular air sac, and the s-c air diverticula is located caudolaterally to the head of<br />
the coracoid bone and below the head of the humerus, ‘primarily between the M<br />
[muscle] coracobrachialis posterior and the M subcorachoideus’. Similarly, in his study of<br />
the Northern Gannet, Gurney (1913), quoting C. B. Ticehurst, states that the s-c air<br />
diverticula communicate with the respiratory system by way of a passage just outside<br />
the coracoid bone and close to the tendon of the ‘pectoralis minor muscle’ (‘M coracobrachialis<br />
posterior’, according to Richardson (1939)). <strong>The</strong>se authors also briefly describe<br />
the distribution of the s-c air diverticula along the ventral region of the trunk and down<br />
the thighs and wings and the separation of these diverticula between left and right sides<br />
of the body. <strong>The</strong> description of s-c air diverticula that they offer is, however, insufficient<br />
to fully understand the exact pattern of air flow among their various compartments.<br />
<strong>The</strong> objective of this study was to provide a more detailed description, complemented<br />
by photographs, of the anatomy of the s-c air diverticula in the Northern Gannet than<br />
is currently available in the literature. More specifically, we describe the distribution of<br />
s-c air diverticula along the body and the communication between the system of s-c<br />
air diverticula and the respiratory system in this species. We also hypothesise that the<br />
wings’ position may alter this communication as it changes from extended away from<br />
the body while flying and soaring to flexed against the body when diving. More specifically,<br />
we predict that wing flexion against the body closes the communication<br />
between the two systems, thus preventing air from escaping the s-c air diverticula,<br />
thus ensuring a firm cushion on impact.<br />
Materials and Methods<br />
Fifteen carcasses of Northern Gannets in a good state of preservation (ten adult, one<br />
immature and three full-grown hatch-year based on their plumage, one of<br />
undetermined age; six male, four female, five of undetermined sex) were dissected in<br />
the course of this study. <strong>The</strong> carcasses were of wild birds that had drowned in fishing<br />
nets, had died of emaciation/starvation, or had been euthanized because of a broken<br />
SEABIRD 21 (2008): 64–76<br />
65
66<br />
Subcutaneous air diverticula of Northern Gannet<br />
limb. Two carcasses were refrigerated until used a few days later, whereas the 13 other<br />
carcasses were frozen at minus 20°C for a period varying between three weeks and 19<br />
months (average, six months) prior to use. In 13 carcasses, a solution of either red or<br />
blue latex (Carolina Biological Supply, Burlington, NC, USA) was injected in various<br />
locations in order to make casts of the different cavities under study, i.e. respiratory air<br />
sacs and/or s-c air diverticula. Injection sites included: intrachoanal; intratracheal (via<br />
a small incision of the skin and tracheal wall in the mid-cervical region); and<br />
perihumeral, axillary, ventral, subclavicular and cervical regions of s-c air diverticula<br />
(via small skin incisions). In five instances, intratracheal injection was carried out while<br />
one wing was extended away from the body and the other flexed (folded) against the<br />
body. In each of these 13 cases, the carcass, if frozen, was completely thawed (over a<br />
period of 48h); latex was injected into the selected location, being allowed to settle<br />
strictly by gravity; and the carcass was frozen again for c.48h, in order to promote<br />
polymerisation of the latex solution (Tompsett 1970), and subsequently thawed for<br />
dissection. <strong>The</strong> amount of latex injected varied among the locations selected and was<br />
largest when latex was injected intratracheally, in which case up to 550 ml were used<br />
in order to adequately fill some of the respiratory air sacs. Within 48h following the<br />
start of dissection, the carcass (minus abdominal viscera) was immersed in a solution<br />
of 10% formalin in order to prevent decomposition. It was possible to make casts into<br />
two different locations in the same bird by using a latex solution of one colour,<br />
freezing the carcass for c.48h and thawing it in order to inject latex of the other colour.<br />
Two carcasses were dissected without prior latex injection in order to better examine<br />
the sites of origin and insertion of muscle masses particularly relevant to the anatomy<br />
of the s-c air diverticula. George & Berger (1966), Vanden Berge (1975), and Nickel et<br />
al. (1977) were consulted for muscle identification and terminology.<br />
Results<br />
Anatomical observations on the 15 Northern Gannets used in this study were<br />
consistent among all birds, with one exception pertaining to the possible role of wing<br />
flexion in air circulation (see below). According to these observations, the s-c air<br />
diverticula of this species consist mainly of two independent systems of intercommunicating<br />
compartments that extend from the respiratory tract, are bilaterally<br />
symmetrical along the cranio-caudal midline of the trunk and cover its ventral and<br />
lateral regions. Latex injected into the trachea easily fills most of the volume of the<br />
respiratory air sacs and their diverticula within the thoracic cavity. Latex further flows<br />
from what are interpreted as the left and right ventrolateral regions of the clavicular<br />
air sac into the axillary regions of the ipsilateral s-c air diverticula through a narrow<br />
gap between the subcoracoideus and coracobrachialis caudalis muscles (Figures 1 & 2).<br />
Both muscles are located immediately caudal to the coracoid bone. <strong>The</strong> subcoracoideus<br />
muscle originates on the inner (ventral) surface of the cranial region of the<br />
scapula and on the medial surface of the cranial region of the coracoid bone and<br />
inserts on the medial surface of the proximal humerus. <strong>The</strong> coracobrachialis caudalis<br />
muscle originates on the caudal region of the coracoid bone, primarily its lateral<br />
surface but with some fibres originating on its dorsal and ventral surfaces, and it<br />
inserts on the medial surface of the proximal humerus, proximal to the pneumatic<br />
foramen and distal to the insertion of the subcoracoideus muscle.<br />
SEABIRD 21 (2008): 64–76
Subcutaneous air diverticula of Northern Gannet<br />
Figure 1. Right lateral view of the thoracic cavity<br />
of a Northern Gannet Morus bassanus. <strong>The</strong> head<br />
is directed to the right side of the Figure. Red<br />
latex injected via the trachea filled the clavicular<br />
air sac (CAS) and extended (a) into the<br />
pneumatic foramen of the right humerus (H)<br />
and (b) between the coracobrachialis caudalis<br />
muscle (CBC) and subcoracoideus muscle (SC)<br />
into the axillary region of the right subcutaneous<br />
(s-c) air diverticulum. (Blue latex had previously<br />
been injected into both axillary regions, and<br />
small amounts of it had reached the clavicular<br />
air sac.) Co, caudal end of coracoid bone<br />
detached from its insertion onto the sternum.<br />
Latex emerges from the clavicular air sac into the left and right axillary regions, located<br />
laterally between the sternum and corresponding pectoralis muscle (Figure 2). In three of<br />
five instances in which intratracheal injection was carried out while one wing was flexed<br />
and the other extended, flow of latex into the axillary region occurred on the extended<br />
side, but not on the flexed side. Latex also flows from the clavicular air sac into the<br />
pneumatic foramen of the right and left humeri through a separate communication<br />
located immediately caudal to the corresponding coracobrachialis caudalis muscle (Figure<br />
1). This flow of latex into the humerus occurred whether or not the wing was flexed.<br />
From the axillary region, the air diverticulum extends ventrally between the sternum<br />
and the pectoralis muscle and caudally along the lateral side of the trunk and along<br />
the medial side of the leg down to the distal region of the tibia (Figure 2). No<br />
communication was found between the region of the s-c air diverticulum along the leg<br />
and the ipsilateral abdominal air sac.<br />
Immediately caudal to the axilla, the diverticulum originating from the axillary region<br />
also extends dorsally and then cranially, dorsal to the scapulohumeralis muscle and<br />
ventral to (underneath) the latissimus dorsi muscle (Figure 2). <strong>The</strong> diverticulum<br />
emerges subcutaneously dorsal to the shoulder, curves around the cranial region of the<br />
shoulder, and opens caudally into a large compartment that covers the whole ventral<br />
surface of the pectoralis muscle and extends slightly caudal to it (Figure 3). <strong>The</strong> lateral<br />
region of this ventral compartment is subdivided into approximately six pockets that<br />
communicate widely with each other ventrally and are formed by thin transparent<br />
membranous partitions extending about 3–4 cm from the lateral wall of the<br />
compartment and attached to the skin along their outer border and to the pectoralis<br />
muscle along their inner border. <strong>The</strong> lateral wall of this ventral compartment, also<br />
consisting of a thin transparent membrane, separates it from the portion of the air<br />
diverticulum originating from the axillary region; this lateral wall extends along the<br />
SEABIRD 21 (2008): 64–76<br />
67
68<br />
Subcutaneous air diverticula of Northern Gannet<br />
lateral side of the trunk and down the medial side of the leg. <strong>The</strong> ventral compartment<br />
is separated from the contralateral side by a thin transparent membranous partition<br />
which is continuous along the ventral midline from the cranial extremity of the keel<br />
to the vent and is attached to the skin along its outer border and to the keel and, more<br />
caudally, the abdominal muscle wall along its inner border. Thin bands of fibrous tissue<br />
and small blood vessels and nerves course through these various membranous<br />
partitions and may thus reinforce them (Figure 3).<br />
Figure 2. Right lateral view of the trunk of the skinned carcass of a Northern Gannet Morus bassanus. <strong>The</strong> head is<br />
directed to the right side of the Figure. <strong>The</strong> arrows show the communication among compartments of the right s-c<br />
air diverticulum, starting in the axillary region with its origin from the (intrathoracic) clavicular air sac between the<br />
subcoracoid muscle (SC) and coracobrachialis caudalis muscle (CBC) and spreading between the sternum (St) and<br />
ribs medially and the pectoralis muscle (P, reflected away from the sternum) laterally (1).<strong>The</strong> diverticulum continues<br />
caudally along the trunk and the medial side of the leg. Caudal to the axillary region, it passes dorsally and then<br />
cranially between the scapulohumeralis muscle (SH) and latissimus dorsi muscle (LD) (2) to emerge subcutaneously<br />
at the level of the shoulder (3). From underneath the pectoralis muscle, the diverticulum also passes cranially and<br />
then dorsally over the proximal end of the coracoid bone (Co) (4), emerges underneath the tensor propatagialis<br />
muscle (TP) to extend along muscles of the proximal portion of the wing (5), and also joins the compartment that<br />
emerges subcutaneously from underneath the latissimus dorsi muscle (6) (see Figures 4 & 5). <strong>The</strong> diverticulum then<br />
proceeds ventrally and caudally between the pectoralis muscle and skin (7) (see Figure 3). In addition, the<br />
diverticulum extends from the axillary region along the humerus between the muscle biceps brachii (BB) ventrally<br />
and the muscle triceps brachii dorsally (8) (see Figure 6). A third compartment of the diverticulum extending along<br />
the wing, besides (5) and (8), originates from the portion of the diverticulum as it emerges from underneath the<br />
latissimus dorsi muscle (see Figure 5) (hidden from view in this Figure). Cl, clavicle; F, femur; H, humerus.<br />
SEABIRD 21 (2008): 64–76
Subcutaneous air diverticula of Northern Gannet<br />
Figure 3. Ventral view of the trunk of a Northern Gannet Morus bassanus. <strong>The</strong> head is directed to the right side<br />
of the Figure. <strong>The</strong> skin has been partly reflected, revealing the large ventral compartment of the right s-c air<br />
diverticulum. <strong>The</strong> lateral region of this compartment is partly subdivided into individual pockets by thin<br />
transparent membranous partitions. Its lateral wall (LatW), also consisting of thin transparent tissue, separates it<br />
from the axillary region of the diverticulum, whereas a median wall (MedW) separates it from the ventral<br />
compartment of the left diverticulum. Thin bands of fibrous tissue and small blood vessels and nerves course<br />
through these various membranous partitions (thick arrows). <strong>The</strong> long arrow on the right shows the wide<br />
communication over the shoulder between this ventral compartment of the diverticulum and its dorso-lateral<br />
region (see Figure 1, point 7). P, pectoralis muscle.<br />
From the axillary region, and in addition to extending ventrally and caudally between<br />
the sternum and pectoralis muscle, the diverticulum also extends through a narrow<br />
canal dorsally over the cranial end of the coracoid bone and spreads distally<br />
underneath the tensor propatagialis muscle (the origin of which is partly on the cranial<br />
end of the coracoid bone) toward the distal end of the humerus (Figures 4 & 5). From<br />
underneath the tensor propatagialis muscle, this portion of the diverticulum is<br />
continuous via a narrow opening with the portion of the diverticulum that emerges<br />
subcutaneously from underneath the latissimus dorsi muscle (Figure 5).<br />
Air diverticula extending along muscles of the proximal portion of the wing are<br />
supplied from at least three sources. One, just described, originates from underneath<br />
the tensor propatagialis muscle (Figures 4 & 5). Another, which proceeds distally from<br />
underneath the deltoideus major muscle, originates from the portion of the<br />
diverticulum that travels beneath, and emerges subcutaneously cranial to, the<br />
latissimus dorsi muscle (Figure 5). A third source originates directly from the axillary<br />
SEABIRD 21 (2008): 64–76<br />
69
70<br />
Subcutaneous air diverticula of Northern Gannet<br />
Figure 4. Right shoulder of a Northern Gannet Morus bassanus in lateral view. <strong>The</strong> head is directed to the right side<br />
of the Figure. Blue latex injected subcutaneously along the distal region of the right humerus spread underneath<br />
the tensor propatagialis muscle (TP) and, through a narrow canal running dorsally around the cranial end of the<br />
coracoid bone (Co), reached the axillary region of the s-c air diverticulum, located between sternum and pectoralis<br />
muscle (removed). Cl, clavicle; H, proximal end of humerus.<br />
SEABIRD 21 (2008): 64–76
Subcutaneous air diverticula of Northern Gannet<br />
Figure 5. Right shoulder of a Northern Gannet Morus bassanus in lateral view. <strong>The</strong> head is directed to the right side<br />
of the Figure. Blue latex was injected into the axillary region, between sternum and pectoralis muscle (P). After<br />
having travelled dorsally around the cranial end of the coracoid bone, the latex solution extended underneath the<br />
tensor propatagialis muscle (TP) and further distally along the humerus (H). It also partly filled, through a narrow<br />
opening (a), the portion of the s-c air diverticulum that would normally emerge subcutaneously from underneath<br />
the latissimus dorsi muscle (LD, cranial and caudal heads). Some of the latex in the latter portion also extended<br />
underneath the deltoideus major muscle (DM) (b) distally along muscles of the proximal region of the wing.<br />
region and proceeds between the biceps brachii muscle ventrally and the triceps<br />
brachii muscle dorsally (Figure 6).<br />
In addition to the large systems of air diverticula covering the ventro-lateral region of the<br />
trunk, two other, small, bilaterally symmetrical, independent systems may contribute to<br />
cushioning the gannet’s body during dives: one at the thoracic inlet and another along<br />
the neck. Air diverticula located at the thoracic inlet on either side of the cranio-caudal<br />
midline lie cranial to the main compartment of the respiratory clavicular air sac and<br />
internal and cranial to the corresponding clavicle (Figure 7). Each of these subclavicular<br />
air diverticula communicates with the corresponding side of the clavicular air sac via a<br />
tubular channel along the lateral wall of the thoracic cavity. Large s-c air diverticula, each<br />
possibly composed of a number of interconnecting compartments, lie ventrally along<br />
either side of the neck (Figure 8). Each of these two cervical s-c air diverticula<br />
communicates with the nasal cavities and the choanal opening through a small aperture<br />
along the roof of the pharynx. No communication was found between these cervical sc<br />
air diverticula and the ipsilateral cervical respiratory air sacs.<br />
SEABIRD 21 (2008): 64–76<br />
71
72<br />
Subcutaneous air diverticula of Northern Gannet<br />
Figure 6. Ventral view of the right axilla and<br />
humerus of a Northern Gannet Morus bassanus.<br />
<strong>The</strong> head is directed to the right side of the<br />
Figure. Blue latex injected via the trachea<br />
extended into the axillary region and, from there,<br />
along the humerus between the biceps brachii<br />
muscle (BB) ventrally and the triceps brachii<br />
muscle dorsally. St, sternum, from which the<br />
pectoral muscle has been removed.<br />
Figure 7. Ventral view of the thoracic girdle of<br />
a Northern Gannet Morus bassanus. <strong>The</strong> head is<br />
directed to the right side of the Figure. Red<br />
latex injected via the trachea (T) has filled the<br />
left and right subclavicular air diverticula (stars)<br />
via their communication with the clavicular air<br />
sac. (Blue latex had previously been injected<br />
into both axillary regions, and small amounts of<br />
it had reached the clavicular air sac and,<br />
subsequently, both subclavicular air<br />
diverticula.) Cl, clavicle; K, keel of the sternum;<br />
P, left and right pectoralis muscles.<br />
SEABIRD 21 (2008): 64–76
Subcutaneous air diverticula of Northern Gannet<br />
Figure 8. Ventral view of the neck and head of a Northern Gannet Morus bassanus. <strong>The</strong> head is to the right; the<br />
rostral portion of the upper beak was cut off. <strong>The</strong> mandible and caudal region of the palate (P) were removed in<br />
order to exposed the nasal cavities. Red latex injected directly into the right cervical s-c air diverticulum extends<br />
through a small opening (arrow) into the right nasal cavity. Q, quadrate bone, which articulates with the mandible.<br />
Discussion<br />
<strong>The</strong> results of this study confirm and expand upon earlier observations of the anatomy<br />
of s-c air diverticula in the Northern Gannet.Assuming that air flow among the various<br />
compartments of these diverticula can be inferred from that of latex, their distribution,<br />
more specifically their voluminous size along the ventral surface of the neck and trunk<br />
and their paucity along the back, supports their putative function in cushioning the<br />
impact of entry into the water as the bird dives from high above the water surface.We<br />
had originally hypothesised that flexion of the wings against the body at the start of<br />
a dive could close the communication between respiratory air sacs and s-c air<br />
diverticula of the trunk, thus allowing the air trapped in these diverticula to provide an<br />
efficient cushion against the impact of entry into the water. This was demonstrated in<br />
three of five instances. It is possible that, in the two birds in which latex could flow<br />
from the clavicular air sac into the axillary region of the s-c air diverticulum on the<br />
side of the flexed wing, incomplete flexion of the wing against the body and/or<br />
postmortem stiffness of the tissues could have prevented complete closure of the<br />
communication. Alternatively, during a dive, a gannet lays its wings back, flat against<br />
the body but fully extended rather than flexed (Elphick et al. 2001), and this position<br />
might better close the communication between respiratory air sacs and s-c air<br />
diverticula of the trunk. <strong>The</strong> hypothesis proposed above therefore requires further<br />
testing. More cranially, escape of air from inflated cervical air diverticula (via the nasal<br />
cavities) and inflated subclavicular air diverticula (via the clavicular air sac and into the<br />
lungs and trachea) could be prevented by complete closure of the Northern Gannet’s<br />
mouth, as the nostrils are permanently closed by an overgrowth of epithelial cells in<br />
this species (King & McLelland 1984).<br />
SEABIRD 21 (2008): 64–76<br />
73
74<br />
Subcutaneous air diverticula of Northern Gannet<br />
Our localisation of the communication between respiratory air sacs and s-c air<br />
diverticula of the trunk is as described in general terms in the Northern Gannet by<br />
Gurney (1913) and more precisely in the Brown Pelican by Richardson (1939), namely<br />
between the subcoracoideus and coracobrachialis caudalis muscles. Richardson (1939)<br />
assumed, based on the work of others, that this communication comes from the<br />
clavicular air sac and ‘probably corresponds to the axillary diverticulum’ of this air sac.<br />
We make a comparable assumption, i.e. that the s-c air diverticulum along the trunk<br />
on each side of the cranio-caudal midline represents a massive extension of an<br />
extrathoracic diverticulum originating from the ipsilateral (paired) lateral chamber of<br />
the clavicular air sac as described by King (1975) and McLelland (1989).<strong>The</strong> only other<br />
possible alternative would be a communication between the s-c air diverticula and the<br />
cranial thoracic air sacs. <strong>The</strong>se air sacs were not visualised in this study, as it was not<br />
our intent to provide a detailed description of the system of respiratory air sacs in the<br />
Northern Gannet. However, whereas the clavicular air sac typically has several intraand<br />
extra-thoracic diverticula, including a humeral diverticulum aerating the humerus<br />
in some species (as was found in our birds), the cranial thoracic air sacs are not known<br />
to have diverticula in any species (King 1975; Nickel et al. 1977; McLelland 1989),<br />
although they may aerate some bones (sternal ribs, sternum) in some species (e.g.<br />
Psittaciforms) (Evans 1996). King (1975) describes three diverticula of the lateral<br />
chamber of the clavicular air sac in the chicken (pectoral, humeral, and axillary),<br />
whereas McLelland (1989), citing Groebbels (1932), describes in general five such<br />
diverticula (subscapular, axillary, subpectoral, suprahumeral, and a fifth diverticulum<br />
under the latissimus dorsi muscle) but adds that ‘considerable interspecific variation in<br />
development of the diverticula appears to exist’. We did not attempt to ascribe the sc<br />
air diverticula along the trunk of the Northern Gannet to an extension of any one or<br />
more of the diverticula mentioned above. Bezuidenhout et al. (1999) describe<br />
diverticula originating from the abdominal air sacs and extending among muscles and<br />
under the subcutis of the legs in the Ostrich Struthio camelus. No such communication<br />
was found between the region of the s-c air diverticula along the legs of Northern<br />
Gannets and the ipsilateral abdominal air sacs.<br />
<strong>The</strong> location of what we describe as the subclavicular air diverticula in the Northern<br />
Gannet corresponds to that of the craniolateral diverticulum of the median chamber of<br />
the clavicular air sac as described in the chicken by King (1975) and in general by<br />
McLelland (1989) and of the left and right cranial parts of the clavicular air sac as<br />
described in the chicken by Nickel et al. (1977). What we describe as cervical s-c air<br />
diverticula represent a greatly expanded cervicocephalic system of air sacs, as there was<br />
a clear communication between these diverticula and the nasal cavities. According to<br />
Richardson (1939), there is in pelicans ‘sometimes a connection between the air cavities<br />
of the pharyngonasal system of the head region and the pulmonary cavities of the neck’,<br />
but no further description is provided. Walsh & Mays (1984) described in psittacine birds<br />
a s-c cervicocephalic air sac with a cephalic portion situated caudodorsally to the skull and<br />
a cervical portion extending bilaterally dorsolaterally along the neck; this sac<br />
communicates with the infraorbital sinus cranially, but not with the respiratory air sacs.<br />
According to McLelland (1989), the cervical respiratory air sacs make ‘an especially large<br />
contribution to the subcutaneous diverticula [along the ventral surface of the neck] in the<br />
SEABIRD 21 (2008): 64–76
Subcutaneous air diverticula of Northern Gannet<br />
Gannet’. This author added, however, that ‘extensions of the vertebral diverticula [of the<br />
cervical respiratory air sacs] penetrate between the neck diverticula of the cervicocephalic<br />
system of air sacs’, suggesting that two independent systems of cervical s-c air diverticula<br />
are present in this species.We were unable in this study to find a distinct communication<br />
between the cervical s-c air diverticula and the cervical respiratory air sacs.<br />
Several unanswered questions remain regarding the function of s-c air diverticula and air<br />
circulation within them. Within the six families of the order Pelecaniformes, all sulids<br />
(boobies and gannets) forage by plunge-diving, but among the other five families, only<br />
Brown Pelicans and Neotropic Cormorants Phalacrocorax olivaceous habitually do so (del<br />
Hoyo et al. 1992). We are not aware of a detailed description of s-c air diverticula in<br />
species of the order Pelecaniformes that do not plunge-dive. Such information could help<br />
to elucidate the evolution of s-c air diverticula in plunge-divers and might offer<br />
alternative explanations for their function. Bignon (1889) suggested a number of<br />
possible functions for the cervicocephalic air diverticula in birds besides that of cushions,<br />
including heat retention, buoyancy control, and head support during flight. Control<br />
mechanisms for inflation and deflation of the different s-c air diverticula are also<br />
incompletely understood. Although s-c air diverticula of the trunk and the subclavicular<br />
s-c air diverticula can be inflated voluntarily via the respiratory tract, it is less clear<br />
whether or how a bird can voluntarily inflate the cervical s-c air diverticula via the<br />
choanal opening. Owen (1866) suggested that birds (pelicans and gannets) may<br />
voluntarily expulse air from s-c air diverticula ‘when the bird is about to descend, in order<br />
to increase its specific gravity, and enable it to dart with rapidity upon a living prey’; he<br />
added that this can be done through the action of those muscles that are connected to<br />
the skin by membranous partitions and bands of fibrous tissue, blood vessels and nerves.<br />
In conclusion, the subcutis of the ventral region of the trunk and neck of the Northern<br />
Gannet is invested by an elaborate air mattress that may be related to the particular<br />
foraging behaviour of this species. Further work is required, however, to describe in<br />
more detail the numerous peripheral extensions of this system of s-c air diverticula<br />
and the potential anatomic variation in these extensions among individual birds and,<br />
particularly, to define more precisely its functional adaptation.<br />
Figure 9. Northern Gannet Morus bassanus, Quendale Bay, Shetland, 2 September 2007 © Hugh Harrop.<br />
SEABIRD 21 (2008): 64–76<br />
75
76<br />
Subcutaneous air diverticula of Northern Gannet<br />
Acknowledgements<br />
We thank Shelley Ebbett, Mike Needham, and Fiep de Bie for photography of the<br />
specimens and preparation of the figures. We also thank three anonymous reviewers<br />
for their excellent comments.<br />
References<br />
Bezuidenhout,A. J., Groenewald, H. B., & Soley, J.T. 1999. An anatomical study of the respiratory<br />
air sacs in ostriches. Onderstepoort Journal of Veterinary Research 66: 317–325.<br />
Bignon, F. 1889. Contribution to the study of pneumacity of birds. <strong>The</strong> cervicocephalic air cells<br />
of the birds and their relation with the bones of the head. Mémoires Société zoologique de<br />
France 2: 260–318.<br />
del Hoyo, J., Elliott, A. & Sargatal, J. (eds.). 1992. Handbook of the Birds of the World. Vol. 1.<br />
Lynx Edicions, Barcelona.<br />
Elphick, C., Dunning Jr, J. B. & Sibley, D. A. 2001. <strong>The</strong> Sibley Guide to Bird Life & Behavior.<br />
National Audubon Society. Alfred A. Knopf Inc., New York.<br />
Evans, H. E. 1996. Anatomy of the budgerigar and other birds. In: Rosskopf Jr, W. J. & Woerpel,<br />
R. W. (eds.). Diseases of Cage and Aviary Birds: 79–162. Williams & Wilkins, Baltimore. 3rd edn.<br />
George, J. C. & Berger, A. J. 1966. Avian Myology. Academic Press, New York.<br />
Groebbels, F. 1932. Der Vögel. Bau, Funktion, Lebenserscheinung, Einpassung.Vol. 1. Borntraeger, Berlin.<br />
Gurney, J. H. 1913. <strong>The</strong> Gannet, a bird with a history. Witherby, London.<br />
King, A. S. 1975. Aves Respiratory System. In: Getty, R. (ed.). Sisson and Grossman’s <strong>The</strong> Anatomy<br />
of the Domestic Animals. Vol. 2: 1883–1918. W. B. Saunders Co., Philadelphia. 5th edn.<br />
King, A. S. & McLelland, J. 1984. Birds. <strong>The</strong>ir structure and function: 110–144. Baillière Tindall,<br />
Toronto. 2nd edn.<br />
McLelland, J. 1989. Anatomy of the lungs and air sacs. In: King, A. S. & McLelland, J. (eds.). Form<br />
and Function in Birds. Vol. 4: 221–279. Academic Press, Toronto.<br />
Montagu, G. 1813. First description of the gannet’s subcutaneous air sacs. Supplement to the<br />
Ornithological Dictionary.<br />
Nelson, B. 1978. <strong>The</strong> Gannet. Buteo Books, Vermilion.<br />
Nickel, R., Schummer, A. & Seiferie, E. 1977. <strong>The</strong> Anatomy of the Domestic Birds. Translated by<br />
W. G. Siller & P. A. L. Wight. Springer-Verlag, New York.<br />
O’Connor, P. M. 2004. Pulmonary pneumaticity in the postcranial skeleton of extant Aves: a<br />
case study examining Anseriformes. Journal of Morphology 261: 141–161.<br />
Owen, R. 1866. Anatomy of Vertebrates.Vol. II. Birds and Mammals. Longmans, Green and Co., London.<br />
Richardson, F. 1939. Functional aspects of the pneumatic system of the California Brown<br />
Pelican. <strong>The</strong> Condor 41: 13–17.<br />
Tompsett, D. H. 1970. Anatomical Techniques: 259–265. E. & S. Livingstone, London. 2nd edn.<br />
Vanden Berge, J. C. 1975. Aves Myology. In: Getty, R. (ed.). Sisson and Grossman’s <strong>The</strong> Anatomy<br />
of the Domestic Animals. Vol. 2: 1802–1848. W. B. Saunders Co., Philadelphia. 5th edn.<br />
Walsh, M. T. & Mays, M. C. 1984. Clinical manifestations of cervicocephalic air sacs of<br />
psittacines. Compendium of Continuing Education for the Practicing Veterinarian 6: 783–789.<br />
SEABIRD 21 (2008): 64–76
Population decline of Leach’s Storm-petrel on St Kilda<br />
Population decline of Leach’s Storm-petrel<br />
Oceanodroma leucorhoa within the largest<br />
colony in Britain and Ireland<br />
Newson, S. E. 1 *, Mitchell, P. I. 2 , Parsons, M. 2 , O’Brien, S. H. 2 , Austin, G. E. 1 , Benn S. 3 ,<br />
Black J. 4 , Blackburn, J. 1 , Brodie, B. 3 , Humphreys, E. 5 , Leech, D. 1 , Prior, M. 6 & Webster, M. 7<br />
*Correspondence author. Email: stuart.newson@bto.org<br />
1 British Trust for Ornithology, <strong>The</strong> Nunnery, <strong>The</strong>tford, Norfolk IP24 2PU, UK; 2 Joint<br />
Nature Conservation Committee, Dunnet House, 7 Thistle Place, Aberdeen AB10 1UZ,<br />
UK; 3 19 Culloden Court, Inverness IV2 7DX, UK; 4 3 Newgate Street, Bingham,<br />
Nottingham NG13 8FD, UK; 5 BTO Scotland, School of Biological and Environmental<br />
Sciences, Cottrell Building, University of Stirling FK9 4LA, UK; 6 1 Rother Close,<br />
Greenmeadow, Swindon, Wiltshire SN25 3PZ, UK; 7 16 Grice Close, Sheringham, Norfolk<br />
NR26 8UG, UK.<br />
Abstract<br />
This study used diurnal playback of vocalisations to examine the abundance of<br />
breeding Leach’s Storm-petrel Oceanodroma leucorhoa on Dun, St Kilda in 2003 and<br />
2006 in relation to the only previous survey conducted using similar methodology in<br />
1999. <strong>The</strong> number of Apparently Occupied Sites in 2006 was 12,770, not significantly<br />
different to the 14,490 found in 2003, but significantly lower than the 27,811 found<br />
in 1999, by 54%. <strong>The</strong> magnitude and rate of the decline are of major conservation<br />
concern. Great Skua Stercorarius skua predation is thought the most likely cause but<br />
other factors such as poor food supply cannot be ruled out. <strong>The</strong> importance of<br />
continued monitoring of Leach’s Storm-petrel and Great Skua is discussed.<br />
Introduction<br />
<strong>The</strong> Leach’s Storm-petrel Oceanodroma leucorhoa has a highly localised distribution<br />
in the east Atlantic, with breeding confirmed on only about 13 remote islands and<br />
archipelagos off the coasts of the Republic of Ireland, Scotland, Faeroes, Iceland and<br />
Norway (Mitchell 2004). One of the largest of these is on the St Kilda archipelago off<br />
northwest Scotland, probably second only in size to the colony on the Westmann<br />
Islands off southern Iceland that is estimated to hold 80,000–150,000 pairs (Icelandic<br />
Institute of Natural History 2000). During 1999, the <strong>Seabird</strong> 2000 survey (Mitchell<br />
2004) estimated that St Kilda held about 94% of the British and Irish population<br />
(45,433 Apparently Occupied Sites, AOS) of which 27,811 were in a single subcolony<br />
on the island of Dun. <strong>The</strong> lack of accurate population estimates for this species in<br />
Britain and Ireland before this time makes it impossible to assess whether there has<br />
been a significant long-term change in the status of this species.<br />
On Dun and other islands of St Kilda, predation by Great Skuas Stercorarius skua is<br />
thought to pose a serious threat to Leach’s Storm-petrel (Phillips et al. 1997, 1999a;<br />
Votier et al. 2006). Great Skuas have increased dramatically on St Kilda, with the<br />
SEABIRD 21 (2008): 77–84<br />
77
78<br />
Population decline of Leach’s Storm-petrel on St Kilda<br />
population rising from about 42 pairs in 1986 to a peak of 240 pairs in 2000 (Furness &<br />
Ratcliffe 2004). Based on their numbers in 1996, it has been estimated that Great Skuas<br />
predate approximately 14,850 Leach’s Storm-petrel on St Kilda each year (Phillips et al.<br />
1999b). Although there has been a small decline in the proportion of Leach’s Stormpetrel<br />
in the diet of Great Skua since 1996, the level of predation appears to have<br />
remained high (Votier et al. 2006). With declines in the availability of offal and discards<br />
from commercial fishing boats predicted due to changes in fisheries management, Great<br />
Skuas may increase their reliance on seabird prey (Votier et al. 2004). Because of the<br />
importance of Dun for Leach’s Storm-petrel in Britain and Ireland, there was a need to<br />
provide reliable monitoring data for this population. Here we report on two subsequent<br />
surveys of Leach’s Storm-petrel on Dun conducted in 2003 and 2006, and examine<br />
whether the size of the population has changed since the first survey in 1999.<br />
Methods<br />
Survey methods: A playback survey was carried out on Dun, St Kilda (57°47’N,<br />
08°33’W; Figure 1) on 4–5 and 7–10 July 2003 and 17–20 June 2006 using the<br />
methods developed by Ratcliffe et al. (1998) that were used previously on Dun in 1999<br />
and throughout the rest of Britain and Ireland during <strong>Seabird</strong> 2000 (Mitchell 2004).<br />
Both surveys were conducted between 09.00–18.30 BST, since response rate is known<br />
to increase significantly towards dusk (Mitchell 2004). Leach’s Storm-petrel burrows on<br />
Dun are virtually invisible, hidden in a thick sward on unconsolidated ground, which is<br />
broken by boulders and by Atlantic Puffin Fratercula arctica burrows. Transect lines<br />
running the width of the island (southwest to northeast; Figure 1) were positioned at<br />
10 m intervals along the northwest side of the island and at larger 25 m intervals on<br />
the southeast side to minimise disturbance of breeding Atlantic Puffins in this part of<br />
the island. A Global Positioning System (GPS) was used to ensure that the positions of<br />
transects were similar (± 5 m) during the 2003 and 2006 surveys. At randomly selected<br />
points along these transects, a total of 331 and 303 5 m x 5 m quadrats were laid out<br />
in 2003 and 2006 respectively, which for each year is about 5% of the total area of Dun<br />
(147,396 m 2 ). Each quadrat was then surveyed by a single observer playing a recorded<br />
call on a dictaphone at approximately 1 m intervals and the number of responses was<br />
recorded. <strong>The</strong> tape playback used a male Leach’s Storm-petrel chatter call recorded on<br />
St Kilda. <strong>The</strong> same recording was used in 1999, 2003 and 2006.<br />
Data analysis: As only a proportion of birds will respond to tape playback, it is<br />
necessary to apply a response rate correction to any estimate. Such a correction factor<br />
is normally calculated by repeatedly sampling a calibration plot (Ratcliffe et al. 1998;<br />
Mayhew et al. 2000). On the first day of the survey in 2003, a calibration plot was<br />
established on the northwest end of Dun (chosen to avoid disturbance to the dense<br />
Atlantic Puffin colony on the southeast end). <strong>The</strong> plot yielded 25 responses from<br />
Leach’s Storm-petrel on the first visit. <strong>The</strong> plot was subsequently visited on each day<br />
of the survey (except on 9 July, when deteriorating sea conditions dictated a<br />
premature departure from the island). Tape playback in the calibration plot was made<br />
at various times of day, but always within the same time period as the survey.<br />
Unfortunately, during the 2006 survey deteriorating sea conditions again necessitated<br />
a premature departure from the island and calibration data for 2006 were too limited<br />
SEABIRD 21 (2008): 77–84
Population decline of Leach’s Storm-petrel on St Kilda<br />
Figure 1. Map showing the location of Dun, St Kilda and surveyed quadrats.<br />
to determine a correction. <strong>The</strong>refore, we applied the same correction factor estimated<br />
from the calibration plot in 2003 to the survey data from 2006.<br />
<strong>The</strong> population size (number of Apparently Occupied Sites, AOS t), of Leach’s Stormpetrel<br />
on Dun for each year was estimated as:<br />
AÔS t =∑S AOS s + S AOS s<br />
x (A t – A q x N)<br />
(N x A q)<br />
Where AOSs is the number of birds counted in each surveyed quadrat, N is the number<br />
of quadrats surveyed, At is the total area of Dun and Aq is the area of a quadrat.<br />
To produce confidence intervals we used a bootstrap resampling procedure of 10,000<br />
iterations (Crowley 1992). For each iteration, AOS s were randomly resampled with<br />
replacement, the sample drawn equal to the number of quadrats surveyed. <strong>The</strong><br />
response rate correction factor was included within the above analyses, by dividing<br />
SAOS s across quadrats from each bootstrap iteration by the mean response rate. <strong>The</strong><br />
errors around the response rate were incorporated by specifying a mean and standard<br />
deviation for each value and randomly selecting response rate values from a binomial<br />
SEABIRD 21 (2008): 77–84<br />
79
Cumulative no. responses<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
80<br />
Population decline of Leach’s Storm-petrel on St Kilda<br />
distribution defined by these values and constrained to remain between 0 and 1. By<br />
doing this 10,000 estimates of population size were produced. <strong>The</strong> 250th and 9,750th<br />
ordered bootstrap values across iterations were taken to give the lower and upper 95%<br />
confidence limits of the estimates respectively. Bootstrap t-tests were used to test for<br />
a significant difference in population estimates between years (Manly 1998).<br />
Results<br />
Calibration: <strong>The</strong> results of tape playback at the calibration plot on Dun in 2003 gave<br />
a mean response rate of 0.386 (Table 1 & Figure 2), very similar to that obtained in<br />
2000 on Boreray, St Kilda (0.382, 95% CL 0.338–9.422) and applied to the 1999<br />
census of Dun (Mitchell 2004).<br />
Population size: As in 1999, there was no significant difference in 2003 between the<br />
mean number of responses per quadrat obtained on the northwest side of the island<br />
(0.945, 95% CL 0.776–1.132) compared to the southeast side (0.955, 95% CL<br />
0<br />
0<br />
0 2 4 6 8 10 12 14 16 18 20<br />
1999 2000 2001 2002 2003 2004 2005 2006<br />
Days after first visit<br />
Year<br />
Figure 2 (left). Plot of cumulative number of responses to tape playback against days after first visit at the<br />
calibration plot on Dun, 4–10 July 2003. <strong>The</strong> solid line represents line of best fit from the observed responses<br />
(diamonds) and the broken lines represent the lower and upper 95% confidence limits of the line. Figure 3 (right).<br />
Estimated population size (AOS with 95% confidence intervals) of Leach’s Storm-petrel Oceanodroma leucorhoa<br />
on Dun, St Kilda in 2003 and 2006 in relation to a previous survey carried out in 1999.<br />
Table 1. Response rate of Leach’s Storm-petrel Oceanodroma leucorhoa to tape playback, estimated from the<br />
calibration plot on Dun, 4–10 July 2003.<br />
Date in July 4 5 6 7 8 10<br />
No. responses 25 25 29 25 25 17<br />
New responses 25 9 8 10 6 1<br />
Cumulative no. responses 25 34 42 52 58 59<br />
Response rate 1 0.397 0.397 0.460 0.397 0.397 0.270 0.386 MEAN<br />
0.336 LCL<br />
0.436 UCL<br />
1 <strong>The</strong> plot of the cumulative number of responses against days after first visit (Figure 2) gave an asymptote of<br />
63, which was taken to be the total number of birds present in the calibration plot.<br />
SEABIRD 21 (2008): 77–84<br />
AOS<br />
40000<br />
35000<br />
30000<br />
25000<br />
20000<br />
15000<br />
10000<br />
5000
Population decline of Leach’s Storm-petrel on St Kilda<br />
0.754–1.159, t 329 = 0.04, n.s.). Thus the data from both sides of the island were pooled<br />
in order to calculate a mean density of responses across the whole island. However in<br />
2006, there were a significantly greater number of responses per quadrat on the<br />
northwest side of the island (1.08, CL 0.842–1.34) than on the southeast side (0.611;<br />
CL 0.45–0.77, t 301 = 3.03, P < 0.01). For this reason, data were resampled separately<br />
for each side of the island and treated as if they were two separate populations to<br />
produce a separate population estimate for each. Population estimates produced from<br />
each iteration for each side of the island were then summed to produce 10,000<br />
estimates for the whole of Dun and 95% confidence intervals taken as before.<br />
With the application of the response rate, the total population of Leach’s Storm-petrel<br />
on Dun was estimated as 14,490 AOS (95% CL 12,110–17,439) in 2003 and 12,770<br />
AOS (95% CL 10,046–17,086) in 2006. This represents a significant decrease of 48%<br />
between 1999 and 2003 (t 435 = 2.95, P < 0.01) but a non-significant 12% decrease<br />
from 2003 to 2006 (t 632 = 0.19, n.s.; Figure 3).<br />
Discussion<br />
<strong>The</strong> estimate of 12,770 AOS in 2006 represents a decrease of approximately 15,000<br />
AOS since 1999 and 1,700 AOS since 2003. <strong>The</strong> scale of decline since 1999 is a major<br />
concern, particularly if the other subcolonies within the archipelago have also<br />
decreased by a similar proportion (i.e. 54%), which would have reduced total numbers<br />
breeding on St Kilda from 45,433 AOS in 1999 to 20,899 AOS in 2006, a loss of about<br />
49,000 breeding birds. Because estimates of population size of Leach’s Storm-petrel on<br />
Dun prior to 1999 are qualitative, it is not possible to interpret this finding in relation<br />
to longer-term trends for this species.<br />
Before exploring possible ecological or environmental causes of the decline, it is<br />
important to rule out the possibility that it was an artefact of the playback method<br />
used during each of the three surveys. <strong>The</strong> largest error in estimating the number of<br />
AOS would arise from using an inaccurate estimate of response rate to adjust the<br />
number of responses obtained during the playback survey. In theory, the decline in<br />
numbers between 1999 and 2003 could have arisen if: a) in 2003, the response rate<br />
used was too high and therefore underestimated the number of AOS; and/or b) in<br />
1999, the response rate was too low and therefore overestimated the number of AOS.<br />
It is unlikely that the response rate used in 2003 was inaccurate, since this was<br />
measured on Dun at exactly the same time as the playback survey was carried out.<br />
Furthermore the response rate measured in 2003 on Dun was very similar to that<br />
measured at other subcolonies within St Kilda and at other colonies in the British Isles<br />
(Mitchell 2004). In contrast, in 1999, no simultaneous measurement of response rate<br />
was conducted on Dun during the survey and instead, a response rate was used that<br />
had been measured the following year on the neighbouring island of Boreray. This<br />
response rate was almost identical to that recorded on Dun in 2003. If the number of<br />
AOS in 1999 were in fact similar to that in 2003, such that there was no decline, the<br />
number of responses obtained in 1999 – 10,574 – would need to be corrected by a<br />
factor of 1.36 rather than 2.62, which was based on a response rate of 0.38 (correction<br />
factor = 1 / response rate). However, a correction factor of 1.36 would only have been<br />
SEABIRD 21 (2008): 77–84<br />
81
82<br />
Population decline of Leach’s Storm-petrel on St Kilda<br />
derived if the response rate in 1999 was 0.74. Response rates to a male chatter call of<br />
more than 0.52 have yet to be recorded at a Leach’s Storm-petrel colony (Mitchell<br />
2004) and are unlikely to be achieved since only the incubating males are thought to<br />
respond to recorded male chatter calls (Taoka et al. 1989). <strong>The</strong>re is no reason to<br />
suggest that a higher proportion of males than females should be incubating at any<br />
one time, with sexes sharing incubation equally with a 3-day changeover time (Wilbur<br />
1969; Watanuki 1985). On Boreray in 2000, the response rate of Leach’s Storm-petrel<br />
did increase significantly towards dusk from a mean of 0.38 (95% CL 0.34–0.42)<br />
between 07.10 and 17.15 to 0.52 (95% CL 0.45–0.58) between 20.00 and 22.00, but<br />
the survey in 1999 was conducted no later than 17.00.<br />
<strong>The</strong> response rate to playback by Leach’s Storm-petrel will vary during the breeding<br />
season, with the highest and most consistent response rates obtained during the peak<br />
of incubation when the highest number of burrows will be occupied by an adult (Ellis<br />
et al. 1998). In the UK, peak incubation occurs around mid June to early July and each<br />
egg is incubated for 41–42 days (Cramp & Simmons 1977). As chicks start to hatch,<br />
daytime occupancy rates and hence response rates will decrease as adults spend days<br />
away from the burrow foraging, leaving young unattended (Ellis et al. 1998).<strong>The</strong> survey<br />
in 1999 was conducted on 30 June and 1 July, and on 4–10 July in 2003, but around<br />
two weeks earlier in 2006 on 17–20 June. <strong>The</strong> surveys in 1999 and 2003 were<br />
conducted towards the very end of incubation. However, of nine burrows examined on<br />
neighbouring Hirta in 2003 at the same time as the survey of Dun, the first chick<br />
hatched on 9 July, the day before the survey on Dun was completed. Furthermore, there<br />
is no evidence of a decrease in response rate during the course of the survey (Table 1),<br />
suggesting that the majority of birds during both the 1999 and 2003 surveys were still<br />
incubating or brooding small chicks, but nevertheless still occupying burrows during the<br />
day (Leach’s Storm-petrel chicks are brooded continuously for five days after hatching<br />
(Cramp & Simmons 1977)). It is thus unlikely that the decline in AOS that occurred<br />
between 1999 and 2003 and 2006 was an artefact of the difference in timing of the<br />
three surveys, and is more likely to represent a real decline in the population on Dun.<br />
One potential cause of the observed decline in Leach’s Storm-petrels is predation by<br />
Great Skuas, which had increased in number exponentially on St Kilda prior to the 1999<br />
survey of Leach’s Storm-petrel (see above).A study on the neighbouring island of Hirta<br />
found that the diet of Great Skuas was dominated by Leach’s Storm-petrel, and that<br />
they hunted during the night when Leach’s Storm-petrels were returning to their<br />
colonies (Votier et al. 2006). Phillips et al. (1999b) estimated that the total Great Skua<br />
population on St Kilda in 1996 consumed approximately 15,000 Leach’s Storm-petrels<br />
per year. Given that between 1996 and 2004, the proportion of Leach’s Storm-petrels<br />
in the diet of Great Skuas remained relatively constant (Votier et al. 2006), around<br />
60,000 Leach’s Storm-petrels would have been killed on the archipelago during the<br />
period between the surveys in 1999 and 2003, assuming that the model of Phillips et<br />
al. (1999b) is correct and that the Great Skua population has not changed substantially<br />
since 1998. Whilst this is more than the 45,000 breeding Leach’s Storm-petrels<br />
we estimate may have been lost from the archipelago, it is likely that recruitment will<br />
offset some of the losses due to predation mortality. In addition it is also likely that<br />
SEABIRD 21 (2008): 77–84
Population decline of Leach’s Storm-petrel on St Kilda<br />
non-breeding adult Leach’s Storm-petrels that also visit colonies at night are being<br />
predated by Great Skuas, but the age or breeding status of Leach’s Storm-petrel<br />
remains could not be determined from skua pellet analysis. Furthermore, it is possible<br />
that there may have been greater declines in the Leach’s Storm-petrel populations on<br />
the other islands than there were on Dun.<br />
Whilst the Leach’s Storm-petrel is not rare in an international context, with an estimated<br />
9,000,000–10,600,000 pairs worldwide (Mitchell 2004), the Great Skua, with just 16,000<br />
pairs is scarce globally (Furness & Ratcliffe 2004). As St Kilda is a Special Protection Area<br />
(SPA) under the European Commission’s Birds Directive, with both Leach’s Storm-petrel<br />
and Great Skua listed as qualifying species, if it were thought necessary to devise a<br />
management solution, it would not be straightforward (Phillips et al. 1999a).<br />
<strong>The</strong> question as to whether Great Skuas are the main cause of the decline remains<br />
unproven. Unfortunately there are no other demographic data available for Leach’s<br />
Storm-petrel on St Kilda or any other colony in the northeast Atlantic that could be used<br />
to identify other causes. In fact St Kilda is the only colony of Leach’s Storm-petrel in the<br />
northeast Atlantic where trends in population size have been accurately measured.<br />
Recent steps have been taken on St Kilda to initiate monitoring of productivity and<br />
adult survival of Leach’s Storm-petrel. In 2004, JNCC installed 25 nest boxes on Hirta<br />
and set up a constant effort mist-netting site for ringing. By attracting Leach’s Stormpetrels<br />
to nest in the boxes it is hoped that monitoring of productivity (including<br />
hatching rates, chick growth rates and fledging success) could be carried out more easily<br />
than in natural nest cavities that have proved extremely difficult to investigate. While<br />
low productivity, perhaps related to poor food supply, may have contributed to the<br />
decline in the breeding population, it was most unlikely to have been the sole cause of<br />
such a steep decline. However, monitoring productivity could give an indication of the<br />
state of feeding conditions during the breeding season and the potential for the<br />
population to recover.<strong>The</strong> main aim of ringing Leach’s Storm-petrel is to gain estimates<br />
of annual survival rates of adults to determine whether high mortality is indeed the<br />
main driver of the decline in the breeding population. However, simultaneous<br />
monitoring of predation rates by skuas is required to determine if observed levels of<br />
mortality are likely to have resulted from predation at the colony or from some other<br />
factor, e.g. in the non-breeding areas in the southeast Atlantic. Monitoring changes in<br />
the numbers of Leach’s Storm-petrel breeding at other colonies in the British Isles and<br />
elsewhere in the northeast Atlantic would be useful to determine whether the decline<br />
is restricted to St Kilda or represents a more widespread phenomenon.<br />
Acknowledgements<br />
Thanks to Maggie Robinson for help with survey work in 2003; to QinetiQ for use of<br />
their boat in 2003; to Neil Mitchell, Sarah Money and Susan Bain of the National<br />
Trust for Scotland and the QinetiQ staff for their support on Hirta. <strong>The</strong> survey in<br />
2006 was part funded through a grant from the <strong>Seabird</strong> <strong>Group</strong>. Thanks also to<br />
Norman Ratcliffe and Steve Newton who provided extremely useful comments on<br />
an earlier version of this manuscript.<br />
SEABIRD 21 (2008): 77–84<br />
83
84<br />
Population decline of Leach’s Storm-petrel on St Kilda<br />
References<br />
Cramp, S. & Simmons, K. E. L. (eds.) 1977. <strong>The</strong> Birds of the Western Palearctic. Vol. I. Oxford<br />
University Press, Oxford.<br />
Crowley, P. H. 1992. Resampling methods for computation-intensive data analysis in ecology<br />
and evolution. Annual Review of Ecology & Systematics 23: 405–447.<br />
Ellis, P., Ratcliffe, N. & Suddaby, D. 1998. Seasonal variation in diurnal attendance and<br />
response to playback by Leach’s Petrels Oceanodroma leucorhoa on Gruney, Shetland. Ibis 140:<br />
336–339.<br />
Furness, R. W. & Ratcliffe, N. 2004. Great Skua Stercorarius skua. In: Mitchell, P. I., Newton, S.<br />
F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. <strong>Seabird</strong> Populations of Britain and Ireland: 173–186.<br />
Poyser, London.<br />
Icelandic Institute of Natural History 2000. Red List of Threatened Species in Iceland. Vol. 2.<br />
Náttúrufræ∂istofnun Islands, Reykjavik.<br />
Manly, B. F. J. 1998. Randomization, Bootstrap and Monte Carlo Methods in Biology. Chapman &<br />
Hall, London.<br />
Mayhew, P., Chisholm, K., Insley, H. & Ratcliffe, N. 2000. A survey of Storm Petrels on Priest<br />
Island. Scottish Birds 21: 78–84.<br />
Mitchell, P. I. 2004. Leach’s Storm-petrel Oceanodroma leucorhoa. In: Mitchell, P. I., Newton, S.<br />
F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. <strong>Seabird</strong> Populations of Britain and Ireland: 101–114.<br />
Poyser, London.<br />
Phillips, R. A., Catry, P., Thompson, D. R., Hamer, K. C. & Furness, R. W. 1997. Inter-colony<br />
variation in diet and reproductive performace of Great Skuas Catharacta skua. Marine Ecology<br />
Progress Series 152: 285–293.<br />
Phillips, R. A., Bearhop, S., Hamer, K. C. & Thompson, D. R. 1999a. Rapid population growth<br />
of Great Skuas Catharacta skua at St Kilda: implications for management and conservation.<br />
Bird Study 46: 174–183.<br />
Phillips, R. A., Thompson, D. R. & Hamer, K. C. 1999b. <strong>The</strong> impact of Great Skua predation on<br />
seabird populations on St Kilda: a bioenergetics model. Journal of Applied Ecology 36: 218–232.<br />
Ratcliffe, N., Vaughan, D., Whyte, C. & Shepherd, M. 1998. Development of playback census<br />
methods for Storm Petrels Hydrobates pelagicus. Bird Study 45: 302–312.<br />
Taoka, M., Sato, T., Kamada, T. & Okumura, H. 1989. Heterosexual response to playback calls<br />
of the Leach’s Storm-petrel Oceanodroma leucorhoa. Journal of Yamashina Institute of<br />
Ornithology 21: 84–89.<br />
Votier, S. C., Furness, R. W., Bearhop, S., Crane, J. E., Caldow, R. W. G., Catry, P., Ensor, K.,<br />
Hamer, K. C., Hudson, A. V., Kalmbach, E., Klomp, N. I., Pfeiffer, S., Phillips, R. A., Prieto, I. &<br />
Thompson, D. R. 2004. Changes in fisheries discard rates and seabird communities. Nature<br />
427: 727–730.<br />
Votier, S. C., Crane, J. E., Bearhop, S., de León, A., McSorley, C. A., Mínguez, E., Mitchell, P.<br />
I., Parsons, M., Phillips, R. A. & Furness, R. W. 2006. Nocturnal foraging by Great Skuas<br />
Stercorarius skua: implications for conservation of storm-petrel populations. Journal of<br />
Ornithology 147: 405–413.<br />
Watanuki, Y. 1985. Breeding biology of Leach’s Storm Petrels Oceanodroma leucorhoa on<br />
Daikoku Island, Hokkaido, Japan. Journal of Yamashina Institute of Ornithology 17: 9–22.<br />
Wilbur, H. M. 1969. <strong>The</strong> breeding biology of the Leach’s Petrel Oceanodroma leucorhoa. Auk 86:<br />
433–442.<br />
SEABIRD 21 (2008): 77–84
Rafting behaviour of<br />
Manx Shearwaters<br />
Puffinus puffinus<br />
Wilson, L. J. 1 *, McSorley, C. A. 1a ,Gray,C.M. 2 ,<br />
Dean, B. J. 1 , Dunn,T. E. 1 ,Webb,A. 1 & Reid, J. B. 1<br />
*Correspondence author.<br />
Email: linda.wilson@jncc.gov.uk<br />
1 Joint Nature Conservation Committee, 7<br />
Thistle Place, Aberdeen AB10 1UZ, UK<br />
( a Current address: Scottish Natural Heritage, 1<br />
Kilmory Industrial Estate, Lochgilphead, Argyll<br />
PA31 8RR, UK); 2 Countryside Council for Wales,<br />
Maes Y Ffynnon, Penrhosgarnedd, Bangor LL57<br />
2DW, UK (Current address: Peak District<br />
National Park Authority, Aldern House, Baslow<br />
Road, Bakewell, Derbyshire DE45 1AE, UK).<br />
Abstract<br />
Radio-telemetry data were collected on<br />
rafting Manx Shearwaters Puffinus puffinus at<br />
Skomer (southwest Wales), Rum (northwest<br />
Scotland) and Bardsey (northwest Wales)<br />
between 2003 and 2005. <strong>The</strong>se were used to<br />
investigate whether Manx Shearwaters tend<br />
to raft adjacent to their breeding areas and<br />
whether rafts move closer towards shore as<br />
the evening progresses. On Skomer and<br />
Bardsey, there was a tendency for birds to raft<br />
in an area roughly adjacent to where they<br />
bred, although they did not raft exclusively<br />
opposite their breeding site. On Rum, birds<br />
breeding at two different locations appeared<br />
to show different preferences for rafting areas.<br />
However, it was difficult to draw conclusions<br />
from this, as signal coverage around the island<br />
was poor, and the breeding locations were<br />
close together. At all three islands, there was<br />
strong evidence that birds tended to move<br />
closer inshore as the evening progressed.<br />
Introduction<br />
Many nocturnal Procellariids form rafts, or<br />
dense flocks of birds, on the sea adjacent to<br />
their breeding colonies from late afternoon<br />
onwards, before coming ashore at nightfall<br />
(Brooke 2004). Species of shearwater known<br />
to engage in such rafting behaviour include<br />
Cory’s Shearwater Calonectris diomedea,<br />
Short Notes<br />
Streaked Shearwater C. leucomelas, Manx<br />
Shearwater Puffinus puffinus, Flesh-footed<br />
Shearwater P. carneipes, Great Shearwater P.<br />
gravis, Sooty Shearwater P. griseus and Shorttailed<br />
Shearwater P. tenuirostris (Brooke 2004).<br />
Nocturnal attendance at the colony is thought<br />
to be a predator avoidance strategy (Brooke &<br />
Prince 1991; Mougeot & Bretagnolle 2000;<br />
Keitt et al. 2004), but the function of rafting is<br />
unclear. It is possible that the birds have<br />
difficulty in timing their return to their colony<br />
from their distant foraging grounds to<br />
precisely coincide with nightfall, so they<br />
assemble in rafts until it is safe to come<br />
ashore (Warham 1996; Brooke 2004). Most<br />
Procellariids in nearshore rafts do not feed, or<br />
only do so if schooling fish are present<br />
(Lockley 1942; Warham 1990), but rafts may<br />
provide an arena for courtship behaviour and<br />
other social interactions, as well as<br />
maintenance behaviour such as preening and<br />
resting (Brooke 1990; Warham 1996).<br />
Rafts can comprise thousands of birds and it is<br />
assumed that they include both breeding and<br />
non-breeding birds, as both visit the colony<br />
after nightfall (Furness et al. 2000). Anecdotal<br />
observations indicate that rafting birds are<br />
generally more restless during windy<br />
conditions, frequently flying and re-alighting<br />
to maintain position (Brooke 2004),<br />
suggesting that raft position is of importance<br />
to participating birds. It is thought that Manx<br />
Shearwaters may raft adjacent to where they<br />
come ashore and their rafts might tend to<br />
approach the shore, once darkness falls<br />
(Brooke 1990). Based on visual observations of<br />
birds coming ashore close to where they had<br />
been rafting, Furness et al. (2000) also<br />
assumed Cory’s Shearwaters tended to raft<br />
opposite their breeding site.<br />
Most observations of rafting have been<br />
anecdotal, and to our knowledge, there have<br />
been no detailed studies of rafting behaviour<br />
itself. Recently, we used radio-telemetry to<br />
determine how far Manx Shearwater rafts<br />
extend offshore, as part of a project to inform<br />
the issue of possible extensions to breeding<br />
colony Special Protection Areas into adjacent<br />
marine areas (McSorley et al. 2008; Wilson et<br />
SEABIRD 21 (2008)<br />
85
86<br />
Short Notes<br />
al. in prep.). In this paper, we use those data to<br />
further investigate whether Manx Shearwaters<br />
tend to raft adjacent to their breeding areas<br />
and whether rafts move closer towards shore<br />
as the evening progresses.<br />
Methods<br />
Detailed methods of the fieldwork can be found<br />
in McSorley et al. (2008) and Wilson et al. (in<br />
prep.). <strong>The</strong> main points are summarised here.<br />
Study colonies: <strong>The</strong> study took place from<br />
May to August on the islands of Skomer<br />
(southwest Wales), Rum (northwest Scotland)<br />
and Bardsey (northwest Wales) (Figure 1a–c),<br />
during 2003, 2004 and 2005 respectively.<br />
<strong>The</strong>se islands host the world’s three largest<br />
Manx Shearwater breeding colonies (Skomer:<br />
101,800 pairs, Rum: 120,000 pairs, Bardsey:<br />
16,183 pairs (Newton et al. 2004)). On each<br />
island, we fitted radio-tags to breeding adults<br />
at a number of geographically distinct breeding<br />
areas, allowing us to test whether the location<br />
of the breeding area influenced raft location<br />
around the island. Four study breeding areas<br />
were chosen on Skomer, two on Rum and five<br />
on Bardsey (Table 1, Figure 1a–c). It was not<br />
possible to get a wide geographical spread of<br />
study sites on Rum, as the colony is confined to<br />
the high mountains, where access is difficult.<br />
SEABIRD 21 (2008)<br />
An initial visit to each colony in May (during<br />
the incubation period) allowed the marking of<br />
potential study burrows, with adults from each<br />
being sexed using cloacal inspection where<br />
possible (Gray & Hamer 2001).<br />
Radio-tag attachment and radio-tracking:<br />
Radio-tag attachment was carried out during<br />
a second visit in July (during the chick rearing<br />
period). A VHF radio transmitter (supplied by<br />
Biotrack Ltd) was attached to the two central<br />
tail feathers of one adult from each accessible,<br />
occupied study burrow. To reduce any adverse<br />
effects of the tag and the possibility of chick<br />
desertion, only the heaviest individuals, with<br />
chicks older than five days were tagged. Tags<br />
weighed 8.9 g on Skomer and 4.4 g on Rum<br />
and Bardsey, less than 2.5% (Skomer) or 1.3%<br />
(Rum and Bardsey) of average adult body<br />
weight. Tags were attached using either selfamalgamating<br />
tape (Skomer), or Tesa® tape<br />
(Rum and Bardsey). Adults were returned to<br />
their burrows immediately following the<br />
tagging procedure, which took less than ten<br />
minutes. At the end of the study, birds were<br />
recaptured and their tags removed. Some birds<br />
could not be recaptured and it was assumed<br />
that these would lose their tags during their<br />
subsequent post-breeding moult.<br />
Table 1. <strong>The</strong> number of marked burrows and birds that were tagged and subsequently located rafting, for each<br />
Manx Shearwater Puffinus puffinus breeding site on each island.<br />
Total no. of No. of birds No. of birds<br />
Island Breeding Site burrows marked tagged located rafting<br />
Skomer Pigstone 14 7 6<br />
<strong>The</strong> Wick 35 8 4<br />
Behind house 23 8 3<br />
<strong>The</strong> Neck 26 7 6<br />
Total 98 30 19<br />
Rum Hallival 45 18 11<br />
Askival 75 10 9<br />
Total 120 28 20<br />
Bardsey Cristin 9 2 2<br />
Nant 32 11 11<br />
Pen Cristin 35 9 9<br />
NW Fields 21 5 5<br />
South-end 7 3 3<br />
Total 104 30 30
Radio-tracking, using Sika receivers and fivebar<br />
rigid Yagi antennas (Biotrack Ltd), began on<br />
the first evening of tag attachment and took<br />
place on: 14 evenings between 15 and 29 July<br />
2003 (Skomer); 15 evenings between 15 July<br />
and 6 August 2004 (Rum); and 18 evenings<br />
between 31 July and 19 August 2005<br />
(Bardsey). Radio-tracking was conducted from<br />
three locations on Skomer and mainland<br />
Pembrokeshire (50–80 m above sea level<br />
(a.s.l.)), six locations on Rum (200–250 m<br />
a.s.l.), and five locations on Bardsey and<br />
mainland Gwynedd (100–160 m a.s.l.),<br />
although at any one time only two or, usually,<br />
three locations were used. Radio-tracking<br />
generally commenced 17.00–19.00 (GMT)<br />
and continued until the last detectable tagged<br />
bird returned to the colony (usually by 24.00).<br />
During this time simultaneous compass<br />
bearings for each detectable signal (bird) in<br />
the area were taken by two or three observers,<br />
following a coordinated tracking schedule<br />
(one bird every three minutes).<br />
Analysis: Analyses were performed only on<br />
those birds that were thought to be rafting;<br />
data were checked prior to analyses and any<br />
bearings that were clearly incorrect, or were for<br />
birds that were travelling or foraging, were<br />
removed. This was determined by signal<br />
strength, notes taken by the trackers and<br />
comparison of signal direction between<br />
tracking locations. Compass bearings were<br />
corrected for magnetic north using the<br />
appropriate adjustment for each area. Bird<br />
locations were estimated by triangulation of<br />
corrected bearing data with LOAS TM 3.0.2<br />
(Location Of A Signal) software © 1998–2004<br />
(Ecological Software Solutions TM). A maximum<br />
likelihood estimator was used for triangulation.<br />
To test whether raft location around each<br />
island was influenced by birds’ breeding<br />
locations, a quadrant was centred on the<br />
‘central study burrow’ and rafting birds from<br />
each breeding area were classified according<br />
to which quadrant they fell within (i.e.<br />
whether they were located in rafts northeast,<br />
northwest, southeast or southwest of the<br />
island, or in the case for Rum, east or west).<br />
Observed and expected frequencies were<br />
compared using a Chi-squared test. <strong>The</strong><br />
Short Notes<br />
‘central study burrow’ was defined as the<br />
geographical mean centre of all study burrows<br />
containing a tagged bird, using the central<br />
feature tool within Esri® Arcmap TM 9.2.<br />
A Spearman rank correlation of raft distance<br />
against time was used to investigate whether<br />
rafts moved closer to the colony over time.<br />
<strong>The</strong> distance between each estimated raft<br />
location and the nearest point of land on the<br />
island (‘raft distance’) was calculated using<br />
the spatial join tool within Esri® Arcmap TM 9.2<br />
and based on a polygon of the high water<br />
mark supplied by Ordnance Survey under<br />
licence [JNCC][100017995][2008].<br />
Figure 1a. Rafting locations of Manx Shearwaters Puffinus<br />
puffinus around Skomer, according to their breeding area. <strong>The</strong><br />
position of each quadrant is shown, centred on the mean centre<br />
of all study burrows.<br />
SEABIRD 21 (2008)<br />
87
88<br />
Short Notes<br />
Results and Discussion<br />
Thirty birds on Skomer, 28 on Rum and 30 on<br />
Bardsey, were fitted with a radio-tag, and of<br />
these it was possible to estimate raft locations<br />
for 19, 20 and 30 birds respectively (Table 1).<br />
Within Skomer and Rum, the sample of birds<br />
for which we were unable to estimate raft<br />
locations included individuals breeding at each<br />
of the study areas, making it unlikely that the<br />
lack of data from these birds biased our results.<br />
Six birds from Skomer and three from Rum<br />
were not detected at all after tag attachment,<br />
and of these, four (all from Skomer) were<br />
recaptured of which all had lost their tag.<br />
<strong>The</strong>refore, it seems likely that our failure to<br />
detect some birds was due to tag loss, rather<br />
than because they were not in the vicinity of<br />
the colony. Tag loss was reduced during the<br />
Rum and Bardsey studies by having a more<br />
secure attachment using Tesa® tape and lighter<br />
tags. In the sample of birds for which we<br />
obtained raft locations, the sex ratio did not<br />
significantly differ from 50:50 (Skomer: 9m:9f,<br />
x2 = 0.11; Rum: 11m:7f, x2 = 0.56; Bardsey:<br />
8m:7f, x2 = 0.17; P > 0.05, 1 df for all).<br />
SEABIRD 21 (2008)<br />
Of those birds located rafting, 58% (Skomer),<br />
40% (Rum) and 93% (Bardsey) were located<br />
in rafts at least twice during the study period.<br />
Guilford et al. (2008) suggested that most<br />
breeding Manx Shearwaters on Skomer may<br />
not join rafts. This was based on data from<br />
birds fitted with global positioning devices<br />
which showed that some birds were some<br />
distance from the colony at nightfall. We were<br />
unable to estimate the number of evenings<br />
that birds visited the colony without rafting,<br />
as we were much more likely to detect birds<br />
that were rafting than birds which simply flew<br />
straight from their foraging grounds to their<br />
burrow (where the radio signal was not<br />
detectable except at close range). However,<br />
our data does not exclude the possibility that<br />
some birds attended their burrow without<br />
joining rafts beforehand, although we feel this<br />
would be a minority.<br />
Figure 1(a–c), shows the rafting locations<br />
estimated from our analysis around each island,<br />
colour coded according to breeding location.<br />
Figure 1b. Rafting locations of Manx Shearwaters Puffinus puffinus around Rum, according to<br />
their breeding area. <strong>The</strong> position of the line dividing the island into west and east is shown,<br />
centred on the mean centre of all study burrows.
Skomer (Figure 1a): No raft location data were<br />
obtained to the east or west of Skomer and<br />
sample sizes were small to the south. This was<br />
largely due to the island’s topography<br />
interfering with signals from tagged birds and<br />
the locations of trackers (at the highest points<br />
on the island, falling in an east–west orientated<br />
line) making it difficult to triangulate to the<br />
east or west. Casual observations before<br />
nightfall, together with signals coming from<br />
these directions (but which could not be<br />
triangulated), indicated that birds were rafting<br />
in the areas for which we lacked data.<br />
<strong>The</strong>re was a significant difference in the<br />
observed and expected frequencies of raft<br />
locations in each quadrant for Skomer (x2 =<br />
33.53, 9 df, P < 0.001, Figure 2a). Most raft<br />
locations were to the northeast (53%) and the<br />
northwest (37%). Birds breeding at Pigstone<br />
(in the west of the island) rafted most often to<br />
the northwest while birds breeding at the<br />
Neck (in the east of the island) were located<br />
more often to the northeast. Observations<br />
from birds breeding at the House (in the east<br />
Short Notes<br />
of the island) were equally divided between<br />
the northeast and northwest. Almost all of the<br />
observations in the two southern quadrants<br />
were from birds breeding at the Wick (in the<br />
south of the island), although most of<br />
observations from the Wick were in the<br />
northeast quadrant.<br />
As there were few observations to the south<br />
of Skomer, the Chi-squared test was repeated<br />
excluding raft locations to the southeast and<br />
southwest; there was still a significant<br />
difference (x2 = 11.29, 3 df, P < 0.05).<br />
Thus, there was evidence that the location of<br />
the breeding area had some influence on raft<br />
location, with most of birds breeding in the<br />
east of the island, rafting to the northeast,<br />
most birds breeding in the west of the island,<br />
rafting to the northwest, and almost all of the<br />
raft observations to the south, being from<br />
birds breeding in the south.<br />
Rum (Figure 1b): No raft location data were<br />
obtained to the north or south of Rum. On<br />
Figure 1c. Rafting locations of Manx Shearwaters Puffinus puffinus around Bardsey,<br />
according to their breeding area. <strong>The</strong> position of each quadrant is shown, centred on the<br />
mean centre of all study burrows.<br />
SEABIRD 21 (2008)<br />
89
Percentage of rafting observations<br />
in each quadrant<br />
100<br />
90<br />
80<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
90<br />
Short Notes<br />
5<br />
5<br />
south<br />
-east<br />
south<br />
-west<br />
north<br />
-east<br />
north<br />
-west<br />
Rum, the position of trackers was limited due<br />
to access difficulties, and the surrounding<br />
topography allowed adequate signal coverage<br />
only to the west and the east. However, birds<br />
have been observed rafting to the north (S.<br />
Morris, Scottish Natural Heritage, pers.<br />
comm.) and our tracking indicated that birds<br />
were probably also rafting to the south.<br />
Most (81%) raft locations were to the east of<br />
Rum, reflecting greater observer effort there.<br />
<strong>The</strong>re was a significant difference in the<br />
SEABIRD 21 (2008)<br />
40<br />
12<br />
2<br />
1<br />
House (east) Neck (east) Pigstone (west) Wick (south)<br />
Breeding area<br />
Figure 2a. <strong>The</strong> percentage of Manx Shearwater Puffinus puffinus raft locations in each of four quadrants around<br />
Skomer, in relation to breeding area (Pigstone, House, Neck and Wick). Quadrants were centred on the mean centre<br />
of all the study burrows. Bar numbers are frequency numbers.<br />
100<br />
90<br />
80<br />
70<br />
60<br />
50<br />
29<br />
West<br />
East<br />
22<br />
40<br />
30<br />
20<br />
10<br />
80<br />
133<br />
0<br />
Askival<br />
Breeding area<br />
Hallival<br />
Figure 2b. <strong>The</strong> percentage of Manx Shearwater Puffinus puffinus<br />
raft locations in each of two areas around Rum, in relation to<br />
breeding area (Askival and Hallival). <strong>The</strong> dividing line between the<br />
two areas was centred on the mean centre of all the study<br />
burrows. Bar numbers are frequency numbers.<br />
Percentage of rafting observations<br />
in each sub-area<br />
22<br />
26<br />
observed and expected frequencies of raft<br />
locations between each breeding area, within<br />
the east and west areas (x2 = 19.14, 1 df, P <<br />
0.01, Figure 2b). This is perhaps surprising<br />
considering the close proximity of the two<br />
breeding areas. Most (57%) of the westerly<br />
locations were from birds that bred at Askival<br />
and most (62%) of the easterly locations were<br />
from birds that bred at Hallival, although<br />
individuals from both breeding areas rafted<br />
both to the east and west of the island.<br />
Birds rafting to the west of Rum may position<br />
themselves there so that they can reach their<br />
burrow by flying directly east up the Glen<br />
Harris valley system (Figure 1b) across land<br />
and around the back of the site. However,<br />
anecdotal observations suggested that some<br />
birds that rafted to the west gradually moved<br />
southwards and around to the east side of the<br />
island, presumably to access their burrow from<br />
the east. Thus, it seems that on Rum,<br />
topography and access routes to the burrow,<br />
as well as burrow location itself, may influence<br />
raft locations, particularly just prior to birds<br />
returning to their colony.<br />
Bardsey (Figure 1c): <strong>The</strong> highest sample size<br />
of both rafting observations and tagged birds,<br />
and the most complete signal coverage<br />
achieved, was on Bardsey. <strong>The</strong>re was a<br />
significant difference in the observed and<br />
3<br />
11<br />
25<br />
22
expected frequencies between each quadrant<br />
for Bardsey (x2 = 57.78, 12 df, P < 0.001). On<br />
Bardsey, most (56%) raft locations were to the<br />
southeast of the island, while the other<br />
quadrants each contained only 14–15% of<br />
raft locations (Figure 2c). Birds from all<br />
subcolonies (except South-end) appeared to<br />
show a preference for rafting to the southeast<br />
of the island, possibly because this provides<br />
the most sheltered area from wind and/or<br />
strong tidal currents. Once the southeast<br />
locations are accounted for, there was<br />
evidence that birds tended to raft adjacent to<br />
Percentage of rafting observations<br />
in each quadrant<br />
100<br />
90<br />
80<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
80<br />
13<br />
32<br />
13<br />
Nant<br />
(northeast)<br />
south<br />
-east<br />
south<br />
-west<br />
north<br />
-east<br />
north<br />
-west<br />
18<br />
5<br />
8<br />
14<br />
NW fields<br />
(northwest)<br />
Short Notes<br />
their breeding area, with birds breeding in the<br />
northeast (Nant) tending to raft to the<br />
northeast, birds breeding in the northwest<br />
(Northwest fields) tending to raft to the<br />
northwest, and birds breeding in the<br />
southwest (South-end) tending to raft to the<br />
southwest. In addition, most of rafting<br />
observations in the preferred southeast<br />
quadrant were from birds breeding in the<br />
southeast (Pen Cristin).<br />
Do rafts move closer to shore over time?:<br />
<strong>The</strong>re were no locations generated earlier than<br />
95<br />
31<br />
10<br />
27<br />
Pen Cristin<br />
(southeast)<br />
Breeding area<br />
7<br />
8<br />
5<br />
South-end<br />
(southwest)<br />
SEABIRD 21 (2008)<br />
91<br />
17<br />
1<br />
1<br />
Cristin<br />
(middle)<br />
Figure 2c. <strong>The</strong> percentage of Manx Shearwater Puffinus puffinus raft locations in each of four quadrants around<br />
Bardsey, in relation to breeding area (Nant, NW fields, Pen Cristin, South end and Cristin). Quadrants were centred<br />
on the mean centre of all the study burrows. Bar numbers are frequency numbers.<br />
Mean (+SD) raft distance (km)<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
Bardsey<br />
0<br />
1645-1900 1901-2000 2001-2100 2101-2200 2201-2300 2301-2400<br />
Time interval (GMT)<br />
Figure 3. <strong>The</strong> mean and standard deviation (SD) of rafting distance (km) of Manx Shearwaters Puffinus puffinus<br />
from their breeding islands, at different time intervals over the evening.<br />
Rum<br />
Skomer
92<br />
Short Notes<br />
19.00 (GMT) on Skomer and Rum, as birds did<br />
not start rafting within range of the trackers<br />
until at least 19.00 on Skomer and 19.30 on<br />
Rum. On Bardsey, the earliest locations<br />
generated were much earlier, at 16.45,<br />
reflecting the earlier onset of nightfall later on<br />
in the season, when the Bardsey study took<br />
place. On all three islands, there was evidence<br />
that rafting birds moved significantly closer to<br />
the colony as the evening progressed<br />
(Spearman Rank correlation: r s = -0.19, n =<br />
174, P < 0.01, Skomer; r s = -0.13, n = 264, P <<br />
0.05, Rum; r s = -0.32, n = 385, P < 0.001,<br />
Bardsey) (Figure 3). Positioning closer to shore<br />
as darkness falls may allow a better<br />
assessment of the approach to the breeding<br />
burrow, which could be associated with<br />
assessing predation threat and local light<br />
levels. By reducing flight time to the burrow,<br />
temporary periods of reduced light levels, such<br />
as when moonlight is obscured by clouds, may<br />
be taken advantage of more readily.<br />
Conclusion<br />
This study highlights how radio-telemetry<br />
can be a useful tool to examine rafting<br />
behaviour. <strong>The</strong>re was clear evidence from all<br />
three colonies that rafting birds tended to<br />
approach closer to shore as the evening<br />
progressed. <strong>The</strong> data also indicated that raft<br />
location around an island may be influenced<br />
by the location of the breeding area, although<br />
the situation was not clear-cut and data were<br />
confined to one season for each island. For<br />
Skomer and Rum, data were limited in some<br />
areas by poor signal coverage, making it more<br />
difficult to assess the true influence of<br />
breeding location on raft location. <strong>The</strong> rafts<br />
around Bardsey included a few birds which<br />
were feeding and although these birds were<br />
removed from the analysis, the presence of<br />
feeding birds might attract other birds and so<br />
influence the location of rafts. It is likely that<br />
other factors also influence raft location, such<br />
as weather conditions, local oceanographic<br />
variables, and access routes to the colony, and<br />
that the combined effect of such factors will<br />
vary between colonies. Around fairly small<br />
islands, such as the ones in this study, the<br />
choice of raft site may not be as crucial as<br />
around larger islands, as differences in travel<br />
SEABIRD 21 (2008)<br />
time to the burrow from different raft<br />
locations around the island could be<br />
insignificant. Thus, a stronger influence of<br />
breeding location on rafting location might<br />
be more evident where colonies are on larger<br />
islands, such as with Cory’s Shearwaters in<br />
the Azores archipelago.<br />
Acknowledgements<br />
All work was carried out under licence from<br />
the Countryside Council for Wales, Scottish<br />
Natural Heritage and the British Trust for<br />
Ornithology, and was funded by the Joint<br />
Nature Conservation Committee. Thanks to<br />
Skomer and Skokholm Islands Management<br />
Committee, Scottish Natural Heritage, Ynys<br />
Enlli Trust and Bardsey Bird and Field<br />
Observatory for permission to work on the<br />
islands, and to the wardens for their<br />
assistance. <strong>The</strong> fieldwork involved many<br />
individuals, to whom we are very grateful.<br />
Advice on radio-tracking techniques was<br />
gratefully received from Peter Smith and Brian<br />
Cresswell (Biotrack Ltd), and Robert Kenward,<br />
while the manuscript was improved by<br />
comments from Chris Perrins.<br />
References<br />
Brooke, M. 1990. <strong>The</strong> Manx Shearwater. Poyser, London.<br />
Brooke, M. 2004. Albatrosses and Petrels across the<br />
World. Oxford University Press, Oxford.<br />
Brooke, M. de L. & Prince, P. A. 1991. Nocturnality<br />
in <strong>Seabird</strong>s. Proceedings of the International<br />
Ornithological Congress 20: 1113–1121.<br />
Furness, R. W., Hilton, G. & Monteiro, L. R. 2000.<br />
Influences of coastal habitat characteristics on the<br />
distribution of Cory’s Shearwaters Calonectris<br />
diomedea in the Azores archipelago. Bird Study 47:<br />
257–265.<br />
Gray, C. M. & Hamer, K. C. 2001. Food-provisioning<br />
behaviour of male and female Manx Shearwaters,<br />
Puffinus puffinus. Animal Behaviour 62: 117–121.<br />
Guilford, T. C., Meade, J., Freeman, R., Biro, D.,<br />
Evans, T., Bonadonna, F., Boyle, D., Roberts, S.,<br />
Perrins, C. M. 2008. GPS tracking of the foraging<br />
movements of Manx Shearwaters Puffinus puffinus<br />
breeding on Skomer Island,Wales. Ibis 150: 462–73.<br />
Keitt, B. S., Tershy, B. R. & Croll, D. A. 2004.<br />
Nocturnal behavior reduces predation pressure on<br />
Black-vented Shearwaters Puffinus opisthomelas.<br />
Marine Ornithology 32: 173–178.
Lockley, R. M. 1942. Shearwaters. Dent, London.<br />
McSorley, C. A., Wilson, L. J., Dunn, T. E., Gray, C.,<br />
Dean, B. J., Webb, A. & Reid, J. B. 2008. Manx<br />
Shearwater Puffinus puffinus evening rafting<br />
behaviour around colonies on Skomer, Rum and<br />
Bardsey: its spatial extent and implications for<br />
recommending seaward boundary extensions to<br />
existing colony Special Protection Areas in the UK.<br />
JNCC Report 406. Peterborough, UK.<br />
Mougeot, F. & Bretagnolle, V. 2000. Predation risk<br />
and moonlight avoidance in nocturnal seabirds.<br />
Journal of Avian Biology 31: 376–386.<br />
Newton, S. F., Thompson, K. & Mitchell, P. I. 2004.<br />
Manx Shearwater Puffinus puffinus. In: Mitchell, P.<br />
I., Newton, S. F., Ratcliffe, N. & Dunn, T. E. (eds.)<br />
<strong>Seabird</strong> Populations of Britain and Ireland: 63–80.<br />
Poyser, London<br />
Late breeding by<br />
Great Cormorants<br />
Phalacrocorax carbo<br />
Craik J. C. A. 1 * and Bregnballe T. 2<br />
*Correspondence author.<br />
Email: clive.craik@sams.ac.uk<br />
1 Scottish Association for Marine Science,<br />
Dunstaffnage, Oban, Argyll PA37 1QA, UK; 2<br />
National Environmental Research Institute,<br />
University of Aarhus, Department of Wildlife<br />
Ecology and Biodiversity, Kalø, Grenåvej 14,<br />
DK-8410 Rønde, Denmark.<br />
Abstract<br />
Two distinct waves of synchronised breeding<br />
occurred at a colony of Great Cormorants<br />
Phalacrocorax carbo in west Scotland in 2007.<br />
<strong>The</strong> second wave led to young fledging in<br />
September, exceptionally late for this species<br />
locally. Evidence from elsewhere suggests that<br />
this second wave was not double-brooding<br />
(raising of two broods in one year by the same<br />
pair) but breeding by newly-arrived birds that<br />
had failed at a nearby colony earlier in the year.<br />
Introduction<br />
At a small colony of Great Cormorants<br />
Phalacrocorax carbo (hereafter ‘Cormorants’)<br />
in Scotland in 2007, there were two distinct<br />
Short Notes<br />
Warham, J. 1990. <strong>The</strong> Petrels. <strong>The</strong>ir Ecology and<br />
Breeding Systems. Academic Press, London.<br />
Warham, J. 1996. <strong>The</strong> Behaviour, Population Biology<br />
and Physiology of the Petrels. Academic Press, London.<br />
Wilson L. J., McSorley, C. A., Gray, C. M., Dean, B.<br />
J., Dunn, T. E., Webb, A. & Reid, J. B. In prep.<br />
Radio-telemetry as a tool to define protected<br />
areas for seabirds in the marine environment.<br />
waves of successful breeding separated by 2.5<br />
to 3 months. In June, the normal fledging<br />
season, 47 young fledged from 18 nests. <strong>The</strong>n,<br />
in September, 28 young fledged from 14 nests<br />
in the same small area (details in Appendix 1).<br />
A well-synchronised pulse of many young<br />
fledging together so late in the year is unusual,<br />
at least in this part of Scotland where such an<br />
occurrence seems not to have been reported<br />
before.<strong>The</strong> first purpose of this short note is to<br />
place this event on record.<br />
Our second objective is to consider whether<br />
these two waves were caused by the same<br />
pairs breeding for a second time (successful<br />
‘double-brooding’), or by incoming birds<br />
nesting in the same small colony area, or<br />
possibly a mixture of these. <strong>The</strong>re were no<br />
ringing data to resolve this question directly.<br />
We therefore approach it indirectly by<br />
examining other records of late breeding in<br />
this and a closely related species. If it was<br />
genuine double-brooding, 14 of the original 18<br />
pairs raised successful second broods. This<br />
percentage (78%) will be compared with<br />
similar measures from elsewhere.<br />
Late breeding<br />
Late breeding has been recognised as an aspect<br />
of Cormorant breeding behaviour at, at least,<br />
two colonies in continental Europe. In 1992 and<br />
1993, the colony at Val Campotto in the Po<br />
SEABIRD 21 (2008)<br />
93
94<br />
Short Notes<br />
delta, Italy, was studied by Grieco (1994).<br />
Breeding occurred from February to September<br />
with two peaks of laying, the main one in spring<br />
and a smaller one in summer. In 1992 when the<br />
colony had 250 nests, there were 13 late<br />
breeders in a sample of 75 nests (17%). In 1993<br />
when the colony had 270 nests, this proportion<br />
was 19/115 (17%). Most of the later breeders<br />
used nests built by other birds earlier in the<br />
year but, in both 1992 and 1993, Grieco<br />
suggested that some cases of reuse of a nest<br />
after successful fledging occurred because the<br />
same parents initiated a new clutch. He based<br />
this suggestion on the behaviour of the adults<br />
concerned; thus, for example, the adults<br />
sometimes fed the large or fledged young and<br />
did not drive them away. In 1992 Grieco<br />
recorded four such nests with presumed second<br />
breeding, in one of which a second brood<br />
fledged (three young in August). Thus, in his<br />
1992 sample of 75 nests, the proportion with<br />
successful double broods was 1/75 (1.3%).<br />
Late breeding has been described at other<br />
colonies in Europe, including the large colony<br />
at Vorsø in Denmark. This grew from 1,000<br />
nests in 1980 to 5,000 nests in 1991. Each<br />
year a proportion of breeding attempts failed,<br />
mostly during incubation but some at the<br />
chick stage. In some of these nests, a second<br />
clutch was laid, either by the same pair or by<br />
another pair. Most second clutches occurred<br />
after the loss of the first eggs or young or,<br />
more rarely, after young had fledged. <strong>The</strong><br />
proportion of second clutches varied from<br />
year to year: 9–19% of nests in 1980–84,<br />
0–2% in 1985–88, and 0–3.2% in 1990 and<br />
1993–95 (based on 97–348 nests with eggs<br />
followed throughout each season; Bregnballe<br />
1996). In 1980–1983, the years in which<br />
second clutches were most frequent, the<br />
annual percentages of new clutches after<br />
fledging of a first brood were in the range<br />
3.3–9.6%, and successful second broods (nests<br />
where chicks fledged both in the first and in<br />
the second attempt) ranged from 2.3 to 4.6%.<br />
Both these quantities are expressed as<br />
percentages of the number of nests where<br />
eggs were incubated in the first attempt (214<br />
nests in 1980, 348 in 1981, 293 in 1982 and<br />
280 in 1983) (Bregnballe & Gregersen<br />
unpubl.). In some of these nests, chicks of the<br />
SEABIRD 21 (2008)<br />
first brood were tolerated by the adults if they<br />
returned to a nest where a second clutch was<br />
being incubated, suggesting that, in at least<br />
some cases, one or both of the adults was<br />
breeding for a second time after fledging a<br />
brood. <strong>The</strong> range of 2.3 to 4.6%, calculated for<br />
the ‘best’ of the study years, thus places an<br />
upper limit of about 5% on same-pair<br />
successful second broods.<br />
Second-brood behaviour has also been seen at a<br />
tree colony in France (Demongin 1993) but, like<br />
the above Danish and Italian data, these reports<br />
were not based on marked birds. Unequivocal<br />
identification of double-brooding requires<br />
observation of individually marked birds.<br />
Double-brooding by marked adults<br />
At Vorsø, 23–318 breeders (on average 156)<br />
with individually marked colour-rings were<br />
followed from egg-laying onwards annually<br />
from 1981 to 2004. During this period, there<br />
were four cases of marked birds with successful<br />
first broods followed by a second clutch that<br />
was incubated (Bregnballe & Gregersen<br />
unpubl.). Of these, the outcome of one (in<br />
1987) was unknown. Of the other three, all in<br />
1984, one fledged no young, while two fledged<br />
second broods. Both involved marked females<br />
that had each laid first clutches in February,<br />
fledged a first brood in May, laid second<br />
clutches starting 30 May and 1 June, and<br />
respectively fledged one and two young. <strong>The</strong><br />
number of colour-ringed birds known to have<br />
had at least one clutch in 1984 was 186, so the<br />
proportion of known successful second broods<br />
in that year was 2/186 or 1.1%. It was zero<br />
among the ringed parents studied in the other<br />
23 years of the study.<br />
Data from similar species<br />
At a colony of European Shags P. aristotelis<br />
(hereafter ‘Shags’) in Brittany, France, Cadiou<br />
(1994) described a ringed 12-year-old female<br />
which, in 1993, began its first clutch on 7–9<br />
February, fledged three young in early May,<br />
laid a second clutch early in June and fledged<br />
two young early in September. Cadiou cited<br />
reports of similar cases from three sites in<br />
Britain (one at each), noting that at one the<br />
adults had not been marked so two pairs<br />
might have been involved. He considered the
frequency of such cases to be very low - one<br />
in several thousand nests.<br />
Wanless & Harris (1997) made a detailed<br />
study of Shags nesting at the Isle of May,<br />
Scotland during 1985–1996. Breeding<br />
numbers varied between a maximum of<br />
1,916 nests in 1987 and a minimum of 501<br />
nests in 1995 after a major population crash<br />
in 1994. About 40% of the birds had been<br />
ringed and, when possible, identities were<br />
checked of birds involved in two attempts at<br />
breeding in the same year. Second breeding<br />
attempts were noted only in 1987 and<br />
1995, years in which the median laying<br />
dates were the earliest in the study. In 1987,<br />
three nests from which young had fledged in<br />
June continued to be occupied in July; two<br />
of these pairs then laid and one clutch<br />
hatched, but no young fledged. In 1995 a<br />
remarkable 27 nests continued to be<br />
occupied, eggs were laid in 20, five clutches<br />
hatched and all five fledged two young each.<br />
<strong>The</strong> authors did not specify whether any of<br />
these five successful pairs were individually<br />
identifiable but they noted that, of the 27<br />
cases, two involved the same pair breeding<br />
Short Notes<br />
twice while, in four, one adult was known to<br />
have remained the same. <strong>The</strong>re was no<br />
evidence in any of the 27 that change in<br />
site-holders had taken place. <strong>The</strong> authors<br />
suggested that the unusually high number of<br />
second breeders in 1995 was a densitydependent<br />
response to the crash of 1994.<br />
<strong>The</strong> five successful second broods<br />
represented 1% of the 1995 breeding<br />
population of 501 nests, but no such cases<br />
were recorded in the other 11 years of the<br />
12-year study. Wanless & Harris cited other<br />
evidence that genuine double-brooding in<br />
this and related species is uncommon.<br />
Conclusion<br />
<strong>The</strong>se studies of both Cormorants and Shags<br />
show that successful double-brooding is a<br />
rare event. While second clutches are not<br />
unusual after failed clutches, they are less<br />
common after failed broods and rare after<br />
successful broods. <strong>The</strong> two most detailed<br />
studies with individually marked birds, those<br />
of Bregnballe & Gregersen (unpubl.) with<br />
Cormorants and Wanless & Harris (1997) with<br />
Shags, concur in several important findings.<br />
Both found that in most years there was no<br />
Figure 1. Great Cormorant Phalacrocorax carbo chicks, Glas Eilean, Loch Fyne, 27 August 2007 © Tom Callan.<br />
SEABIRD 21 (2008)<br />
95
96<br />
Short Notes<br />
double-brooding, but that there were<br />
occasional good years when, after successful<br />
first broods, second clutches and broods<br />
tended to occur. <strong>The</strong> few available records of<br />
identifiable individuals with successful second<br />
broods showed that all laid their first clutches<br />
early in the year (February in the case of<br />
Cormorants in continental Europe, March-<br />
April in Shags in Scotland) – perhaps not<br />
surprisingly, in view of the time needed to<br />
raise two successive broods. In the occasional<br />
years when they occurred, successful second<br />
broods formed very low proportions of the<br />
number of nests at a colony, about 1% in<br />
each study. Data from unmarked Cormorants<br />
at Vorsø (Denmark) and from Val Campotto<br />
(Italy) suggested upper limits to this quantity<br />
of about 5% and 1% respectively.<br />
<strong>The</strong>se percentages are very much lower than<br />
the 78% observed at Glas Eilean in 2007. (<strong>The</strong><br />
difference is so large that we are justified in<br />
ignoring the relatively minor differences in<br />
methods and interpretation between the<br />
projects considered above.) Thus it is<br />
reasonable to infer that the event at Glas<br />
Eilean was not genuine double-brooding and<br />
that most or all of the 14 late-breeding pairs<br />
were not among the original 18 pairs. Where<br />
might these new birds have come from?<br />
<strong>The</strong>re are three other annually-occupied<br />
colonies of Cormorants within 30 km of Glas<br />
Eilean. <strong>The</strong> nearest is at Eilean Buidhe, 16 km<br />
to the south. In 2007, 22 pairs of Cormorants<br />
nested there but almost all failed. Human<br />
interference was suspected. <strong>The</strong> incubating<br />
adults were counted from the sea on 21 May<br />
and again on 11 June; no young were visible<br />
on either date. On 2 July counting ashore<br />
found 14 empty nests and three nests with<br />
small or medium-small young. On 23 July<br />
there were three large young.<br />
<strong>The</strong>se facts strongly suggest that Eilean<br />
Buidhe was the source of the second wave of<br />
birds that bred at Glas Eilean. Not only was<br />
Eilean Buidhe the nearest Cormorant colony,<br />
but the numbers of birds involved and the<br />
timing of events were both consistent with<br />
this explanation (c. 20 pairs left Eilean Buidhe<br />
some time between 11 June and 2 July; 14<br />
SEABIRD 21 (2008)<br />
pairs started laying at Glas Eilean from late<br />
June to early July, estimated from clutches and<br />
small young there on 5 August). It remains<br />
unclear how these birds were able to<br />
reproduce successfully so late in the year, at a<br />
time when Cormorants in this area have<br />
normally stopped breeding and environmental<br />
cues are presumably unfavourable.<br />
Acknowledgements<br />
We are grateful to Tom Callan for first alerting<br />
JCAC to the Cormorants breeding at Glas<br />
Eilean, for keeping him informed of their<br />
progress, and for assistance with ringing. We<br />
thank Jens Gregersen for his dedicated effort<br />
in observing breeding Cormorants at the Vorsø<br />
colony. We thank Robin Sellers, Martin<br />
Heubeck, Sarah Wanless and Mike Harris for<br />
considerable help and advice during writing.<br />
References<br />
Bregnballe, T. 1996. ‘Reproductive performance in<br />
Great Cormorants during colony expansion and<br />
stagnation’. PhD <strong>The</strong>sis, National Environmental<br />
Research Institute, University of Aarhus, Denmark.<br />
Cadiou, B. 1994. Un évènement rarissime: l’élevage<br />
de deux nichées avec succès par un couple de<br />
cormorans huppés Phalacrocorax aristotelis.<br />
Alauda 62: 134–135.<br />
Demongin, L. 1993. Premières nidification du Grand<br />
Cormorant (Phalacrocorax carbo) sur la resérve de<br />
la Grande-Noé (Eure). Le Cormoran 8: 303–306.<br />
Gibson, J. A. 1958a. Notes on the breeding birds of<br />
the Lochgilphead Islands. Glasgow Bird Bulletin 7:<br />
67–68.<br />
Gibson, J. A. 1958b. <strong>The</strong> breeding birds of the small<br />
Clyde islands. Glasgow Bird Bulletin 7: 99–116.<br />
Gibson, J. A. 1969. Population studies of Clyde<br />
seabirds, part 1. Transactions of the Buteshire<br />
Natural History Society 17: 79–95.<br />
Gibson, J. A. 1985. Population Studies of Clyde<br />
seabirds, part 4. Transactions of the Buteshire<br />
Natural History Society 22: 85–105.<br />
Gibson, J. A. 1990. Population studies of Clyde<br />
seabirds, part 5. Transactions of the Buteshire<br />
Natural History Society 23: 81–107.<br />
Grieco, F. 1994. Fledging rate in the Cormorant<br />
Phalacrocorax carbo at the colony of Val Campotto<br />
(Po Delta, north-east Italy). Avocetta 18: 57–61.<br />
Wanless, S. & Harris, M. P. 1997. Successful<br />
double-brooding in European Shags. Colonial<br />
Waterbirds 20: 291–294.
Appendix 1.<br />
Description of Glas Eilean and events there<br />
in 2007: Glas Eilean (grid reference<br />
NR912857; 56°01’N 5°21’W) consists of two<br />
small islets about 350 m from the mainland in<br />
Loch Fyne, a long sea loch in southwest<br />
Scotland. Records since about 1950 show that,<br />
while large gulls have bred there for decades,<br />
Cormorants were not recorded breeding until<br />
2007. Shags bred for the first time in 2004<br />
(one or two pairs) and 2005 (two pairs) but<br />
none were recorded in 2006 (Gibson 1958a,<br />
1958b, 1985; Craik unpublished annual records<br />
1987–2007).<br />
Several seabird species bred on Glas Eilean in<br />
2007, mainly Herring Gull Larus argentatus (c.<br />
75 pairs) and Great Black-backed Gull L.<br />
marinus (c. ten pairs) together with smaller<br />
numbers of Shag (five pairs), Black Guillemot<br />
Cepphus grylle and Common Eider Somateria<br />
mollissima (at least two pairs of each, probably<br />
considerably more). <strong>The</strong> Cormorant colony<br />
was on the larger, southern islet, which<br />
measures c. 150 x 100 m. Observations at the<br />
Cormorant colony in 2007 were as follows.<br />
27 May: During the annual count ashore of<br />
gull clutches and nests, the new colony of<br />
Cormorants was not approached closer than c.<br />
100 m in a line of vision, as experience<br />
elsewhere has shown that large gulls quickly<br />
learn to prey on exposed eggs and small young<br />
of Shags and Cormorants. On this occasion,<br />
the sitting Cormorants did not leave their<br />
nests and an incomplete count of ten<br />
incubating adults was made from the sea.<br />
11 June: <strong>The</strong> Cormorant colony was entered.<br />
<strong>The</strong>re was one disused nest. <strong>The</strong> 18 active<br />
nests all held large, mobile young, and<br />
(unusually) there were no eggs or small young.<br />
Thirty-seven young were ringed and c. ten<br />
were too large to catch. Thus c. 47 young are<br />
estimated to have fledged from 18–19 pairs in<br />
this first laying.<br />
21 July: On a routine visit to count large<br />
young gulls, the ringed young Cormorants<br />
were flying in the area, settling on tidally<br />
exposed rocks and at the colony edge (as<br />
expected); at least 19 were counted.<br />
Short Notes<br />
Unexpectedly, however, at least 12 adult<br />
Cormorants were seen incubating on nests in<br />
the same nesting area. <strong>The</strong> colony was again<br />
not approached, for the above reason.<br />
5 August: <strong>The</strong> colony was entered and nest<br />
contents were recorded. <strong>The</strong>re were 14 nests<br />
(two with eggs and 12 with small young, none<br />
large enough to ring). Continuous observation<br />
after departure revealed no predation by gulls<br />
before the return of the adult Cormorants.<br />
14 August: Seven nests held 15 young that<br />
were large enough to ring.<strong>The</strong> other seven held<br />
15 young too small to ring, including two runts.<br />
27 August: Thirteen more chicks were ringed.<br />
Two of the 15 (the two runts?) were not found<br />
and had probably died, although a search for<br />
remains was not made. Thus 15 + 13 = 28<br />
young are believed to have fledged in early to<br />
mid September from the 14 late-nesting pairs.<br />
JCAC has ringed Cormorant chicks in this part<br />
of west Scotland for over 20 years and, later<br />
each summer, often visited many of the same<br />
sites to count and ring chicks of other species.<br />
After the normal midsummer fledging of most<br />
Cormorant chicks, it is not unusual to find a<br />
few late-nesting Cormorant pairs with eggs or<br />
small young in early July. However, there are<br />
no local records of an entire colony with many<br />
young fledging together in September. In<br />
2007, JCAC monitored six other Cormorant<br />
colonies in the area, all on small islands along<br />
the mainland coast between Kintyre and<br />
Mallaig, but none of these behaved in this way.<br />
One failed completely as a result of mink<br />
predation of eggs; the colony at Eilean Buidhe<br />
raised almost no young, apparently because of<br />
human interference and/or disturbance<br />
allowing predation of eggs and young by large<br />
gulls; and four colonies reared single broods<br />
that fledged in midsummer in the normal way.<br />
That the breeding pattern at Glas Eilean in<br />
2007 was atypical was confirmed by Robin<br />
Sellers, National Organiser of the UK<br />
Cormorant Breeding Colony Survey (R. M.<br />
Sellers pers. comm.).<br />
SEABIRD 21 (2008)<br />
97
98<br />
Short Notes<br />
A pilot study of the<br />
phenology and breeding<br />
success of Leach’s Stormpetrel<br />
Oceanodroma<br />
leucorhoa on St Kilda,<br />
Western Isles<br />
Money, S., 1 *, Söhle, I. 2 & Parsons, M 2 .<br />
*Correspondence author.<br />
Email: catharacta7@yahoo.co.nz<br />
1 Raintree House, Church Lane, Drayton St<br />
Leonard, Oxfordshire OX10 7AU, UK; 2 Joint<br />
Nature Conservation Committee, 7 Thistle<br />
Place, Dunnet House, Aberdeen AB10 1UZ, UK.<br />
Abstract<br />
From June to September 2007, 27 occupied<br />
Leach’s Storm-petrel Oceanodroma leucorhoa<br />
burrows on St Kilda were investigated by<br />
endoscope to estimate phenology and<br />
breeding success. From 17 burrows that could<br />
be followed to chick development, a minimum<br />
productivity of 0.59 young per egg laid was<br />
estimated, similar to other published figures of<br />
breeding success from the northwest Atlantic.<br />
Similar life cycle timings to those of Leach’s<br />
Storm-petrels in the northwest Atlantic were<br />
revealed, with egg laying in late May to mid<br />
June, hatching in mid to late July and fledging<br />
from mid September onwards. Implications for<br />
conservation and monitoring are discussed,<br />
and guidance for future studies suggested.<br />
Introduction<br />
<strong>The</strong> St Kilda archipelago (57°49’N, 08°35’W),<br />
66 km west of the Western Isles, holds a<br />
population of around 45,000 pairs of Leach’s<br />
Storm-petrels Oceanodroma leucorhoa<br />
which represents 94% of the British and Irish<br />
population (Mitchell 2004). Recent work in<br />
the northwest Atlantic by Huntington et al.<br />
(1996) and A. Hedd (unpubl.) has been carried<br />
out on the breeding biology of Leach’s<br />
Storm-petrel, but little is known about their<br />
breeding habits within Britain, and little has<br />
been published since that of Ainslie &<br />
Atkinson (1937) on the Flannan Isles and<br />
North Rona, apart from some exploratory<br />
SEABIRD 21 (2008)<br />
work on Gruney, Shetland in the 1990s (Ellis<br />
et al. 1998). In addition, it has been suggested<br />
that the Leach’s Storm-petrel breeding<br />
population on St Kilda may have declined<br />
significantly since the last full census in<br />
1999/2000 (Mitchell 2004; Newson et al.<br />
2008), and studies on breeding success may<br />
therefore give insights as to the reasons for<br />
this decline.<br />
Methods<br />
In 2007, 27 Leach’s Storm-petrel burrows<br />
which were occupied (i.e. a response was<br />
obtained from playing a tape of a male ‘St<br />
Kildan’ chatter call in mid June and nest<br />
material was visible) and which were<br />
accessible by endoscope (Everest VIT PXL<br />
Videoprobe) were located and marked with<br />
two short bamboo canes with a tape flag.<br />
Burrows were situated on grassy slopes, under<br />
rocks and in cleits (man-made stone<br />
structures unique to St. Kilda), and were<br />
monitored by endoscope a total of 12 times<br />
throughout the breeding season. Burrow<br />
occupancy was determined on seven days<br />
during peak incubation time between 19 and<br />
27 June, on three days during hatching and<br />
early chick stage (24 and 26 July, and 8<br />
August) and twice during the late chick stage<br />
(on 10 and 26 September). Of the 27 initially<br />
occupied study burrows, eight (30%) did not<br />
contain an egg and therefore were presumably<br />
occupied by immature or prospecting birds.<br />
Eggs were laid in 19 burrows, two of which<br />
could not be relocated due to marker loss,<br />
leaving 17 burrows that were used in<br />
estimates of breeding success (Table 1).<br />
Results<br />
Phenology: <strong>The</strong> presence of unattended small<br />
chicks and incubating adults on 24 and 26 July,<br />
but only unattended chicks on 8 August (Table<br />
1), suggested that hatching took place<br />
between mid and late July (assuming that<br />
chicks are brooded for an average of five days<br />
(Cramp & Simmons 1977)). Given an<br />
incubation period of six weeks and a fledging<br />
period of 9–10 weeks (Cramp & Simmons<br />
1977) this would suggest a laying period from<br />
around late May to mid June and fledging<br />
from mid September. Indeed, our observations
showed that all study birds had laid their eggs<br />
by mid June and the majority of the chicks<br />
fledged between 10 and 26 September, with<br />
the remaining chicks presumably fledging<br />
after 26 September (Table 1). In<br />
Newfoundland, Canada, egg laying of Leach’s<br />
Storm-petrel begins in late May, the mean<br />
hatching time is late July and fledging begins<br />
in mid September (Huntington et al. 1996; A.<br />
Hedd, unpubl.), suggesting that the phenology<br />
in these two geographically distinct<br />
populations is very similar. A study of nine<br />
burrows investigated by endoscopy on St Kilda<br />
in 2003 (O’Brien et al. 2005) found the first<br />
Short Notes<br />
hatching to be on 9 July and a further three<br />
had hatched by 14 July, which indicates a<br />
similar (perhaps slightly earlier) timing to that<br />
presented here, and work carried out on<br />
Gruney in Shetland (Ellis et al. 1998) also<br />
found chicks on a slightly earlier date (6 July)<br />
than in the current study.<br />
Breeding success: Out of 17 burrows in which<br />
an egg was recorded, ten were confirmed to<br />
have chicks by 10 or 26 September (Table 1).<br />
Of the remaining seven burrows, four failed at<br />
the egg stage, two stone cavity nesters<br />
(burrows 19 and 20) had ambiguous results<br />
Table 1. Leach’s Storm-petrel Oceanodroma leucorhoa burrow contents as determined by endoscopy.<br />
19 20 22 23 24 25 27 24, 26 8 10 26<br />
Jun Jun Jun Jun Jun Jun Jun Egg Jul Aug Sep Sep Summary<br />
1 0 0 0 0 0 0 0 0 No egg laid<br />
2 R R R 1,E 1 1,E R Y A,E SC LC 0 Presumed fledged<br />
3 1 1 R R 1,E 1,E 1 Y SC MC 0 Presumed fledged<br />
4 R 1 1 0,E 0,E 1,E 0,E Y 0 E Abandoned egg<br />
5 1 R R R 0 1,E R Y SC LC 0 Presumed fledged<br />
6 R 0 0 1 0 0 0 0 No egg laid<br />
7 R R 1,E R R 1,E R Y SC MC 0 Presumed fledged<br />
8 R R R R R 1,E R Y 0 0 Failed to hatch<br />
9 1,E R R R 1 1,E 1 Y A,BE,SC? SC LC LC Presumed fledged<br />
10 0 0 0 0 0 0 0 0 No egg laid<br />
11 0 0 R R R 1,E 1,E Y A E Abandoned egg<br />
12 0 R R R R 1,E 1 Y SC 0 0 Chick died/fledged early?<br />
13 1 R R R 1 1,E 1 Y ? Burrow not relocated<br />
14 R R R 0,E 0,E 0,E R Y 0 Failed to hatch<br />
15 R R R 1,E 1,E 1,E R Y SC LC LC Presumed fledged<br />
16 R R 0,E 1,E 1 1,E R Y A,SC LC 0 Presumed fledged<br />
17 R R 0 R 0 0 0 ? 0 No egg laid<br />
18 1 1 0,E R R 1,E 1 Y SC MC LC Presumed fledged<br />
19 R R R R 1 1,E 1 Y A BE 0 0 Failed to hatch?<br />
20 1 0 1 1 R 1,E R Y A BE BE BE Failed to hatch?<br />
21 1 0 1 1,E 1 1,E 1 Y BE BE MC 0 Presumed fledged<br />
22 1,E R 0,E 1,E 1 1,E R Y ? Burrow not relocated<br />
23 R 1 0 R R 1 R ? Burrow not relocated<br />
24 0 R 0 0 R 1 R 0 No egg laid<br />
25 0 0 0 0 0 0 0 0 No egg laid<br />
26 0 0 0 0 0 0 0 0 No egg laid<br />
27 0,E R R 0,E 1 1,E R Y A SC MC LC Presumed fledged<br />
R = Response, no endoscope Y = Yes BE = Broken eggshell<br />
0 = No response, no adult 0 = Empty burrow SC = Small unattended chick<br />
1 = Adult present A = Adult in burrow MC = Medium chick<br />
E = Egg seen A,E = Adult sitting on egg LC = Large, well-feathered chick<br />
SEABIRD 21 (2008)<br />
99
100<br />
Short Notes<br />
Figure 1. Sarah Money using endoscope on Hirta, summer 2006.<br />
Figure 2. Leach’s Storm-petrel Oceanodroma leucorhoa, Hirta, St<br />
Kilda, September 2007 © Will Miles.<br />
SEABIRD 21 (2008)<br />
due to the possible retreat of the chick out of<br />
view from the endoscope (which happened<br />
once in the case of one of the successful<br />
burrows) and one (burrow 12) held a chick<br />
which may have fledged before the first check<br />
on 10 September (fledged Leach’s Stormpetrels<br />
have been observed on St Kilda by this<br />
date in previous years: National Trust for<br />
Scotland, unpubl.). Thus, the minimum<br />
estimate of productivity was 0.59 young per<br />
egg laid, comparable with published figures of<br />
breeding success of 0.48–0.73 from the<br />
northwest Atlantic (Huntington et al. 1996),<br />
and suggesting that Leach’s Storm-petrels on<br />
St Kilda might not be currently experiencing<br />
major breeding problems. Of interest was the<br />
fact that during incubation a number of<br />
burrows were observed to hold unattended<br />
eggs for periods of up to three days. Three out<br />
of these six burrows (those where the eggs<br />
were left unattended for single days at a time)<br />
went on to produce chicks, which suggests that<br />
leaving the egg unattended for a single day<br />
does not necessarily prevent hatching. Indeed,<br />
this phenomenon has been observed for a wide<br />
range of petrel species (Warham 1990).<br />
Discussion<br />
This pilot study gives a useful insight into the<br />
breeding success and timing of breeding of<br />
Leach’s Storm-petrel on St Kilda in 2007. It<br />
should be emphasised that this work was of an<br />
exploratory nature, rather than a detailed<br />
study of productivity and phenology, but can<br />
nonetheless be used for comparison with<br />
future studies both on St Kilda and elsewhere<br />
in Britain. Such work on Leach’s Storm-petrels<br />
is of necessity always going to be difficult due<br />
to the remote locations involved and<br />
limitations on the use of endoscopes imposed<br />
by the weather. However, lessons to be<br />
learned which would refine monitoring<br />
techniques for future studies include:<br />
� burrow marker loss could be prevented<br />
through the use of GPS and improved<br />
marking;<br />
� increasing the sample size of burrows and<br />
the spread and frequency of monitoring<br />
dates would help determine breeding cycle<br />
dates and success more accurately, and;
� only grassy slope burrows instead of stone<br />
cavities should be monitored to prevent<br />
ambiguous outcomes from nesting chambers.<br />
Information on the phenology of breeding is<br />
critical to the correct application of methods<br />
for monitoring breeding numbers of stormpetrels.<br />
<strong>The</strong> favoured technique of tape<br />
playback (Ratcliffe et al. 1998) should be<br />
undertaken during the peak period of<br />
incubation to maximise the likelihood of<br />
burrow occupancy at the time of survey;<br />
therefore, an understanding of phenology and<br />
its year-to-year variability has important<br />
practical applications.<br />
Currently there are concerns about the conservation<br />
status of Leach’s Storm-petrel on St<br />
Kilda. Surveys in 1999 and 2003 suggested a<br />
48% decline in the population on Dun, an island<br />
within the St Kilda archipelago (Mitchell 2004),<br />
although a further survey in 2006 suggested<br />
that this rapid decrease may have ceased<br />
(Newson et al. 2008). Predation by Great Skuas<br />
Stercorarius skua is suggested as a potential<br />
explanation for the decline (Phillips et al. 1999;<br />
Mitchell 2004; Votier et al. 2005). However,<br />
alternative factors, such as food shortages,<br />
which currently affect a wide range of Britain’s<br />
seabird species, may also have played a role.<br />
Our pilot study suggests that shortage of<br />
suitable food did not affect Leach’s Stormpetrel<br />
chicks on Hirta, St Kilda during the 2007<br />
breeding season, as breeding success was quite<br />
high, but caution should be applied in<br />
interpreting these results as our sample size was<br />
small. However, studies of breeding success<br />
(combined with those of adult survival) could<br />
contribute to an early warning of potential<br />
population change, and it would therefore be<br />
appropriate to establish a long-term productivity<br />
study for this species on St Kilda.<br />
Acknowledgements<br />
This work was supported by the National Trust<br />
for Scotland and the Joint Nature Conservation<br />
Committee, Aberdeen. We also thank Stuart<br />
Murray for encouraging this publication, April<br />
Hedd for allowing access to unpublished data,<br />
and two anonymous referees for improving an<br />
earlier draft of the manuscript.<br />
Short Notes<br />
References<br />
Ainslie, J. A. & Atkinson, R. 1937. On the breeding<br />
habits of Leach’s Fork-tailed Petrel. British Birds 30:<br />
234–48.<br />
Cramp, S. & Simmons, K. E. L. (eds.) 1977. <strong>The</strong> Birds<br />
of the Western Palearctic. Vol. I. Oxford University<br />
Press, Oxford.<br />
Ellis, P., Ratcliffe, N. & Suddaby, D. 1998. Seasonal<br />
variation in diurnal attendance and response to<br />
playback by Leach’s Petrels Oceanodroma<br />
leucorhoa on Gruney, Shetland. Ibis 140: 336–339.<br />
Huntington, C . E., Butler, R. G. & Mauck, R. A.<br />
1996. Leach’s Storm-petrel (Oceanodroma<br />
leucorhoa). In: Poole, A. & Gill, F. (eds.) <strong>The</strong> Birds of<br />
North America: 233. Birds of North America Inc.,<br />
Philadelphia, PA & American Ornithologists’ Union,<br />
Washington DC.<br />
Mitchell, P. I. 2004. Leach’s Storm-petrel Oceanodroma<br />
leucorhoa. In: Mitchell, P. I., Newton, S., Ratcliffe, N.<br />
& Dunn,T. E. (eds.) <strong>Seabird</strong> Populations of Britain and<br />
Ireland: 101–114. Poyser, London.<br />
Newson, S. E., Mitchell, P. I., Parsons, M., O’Brien,<br />
S. H., Austin, G. E., Benn, S., Black, J., Blackburn,<br />
J., Brodie, B., Humphreys, E., Leech, D. I., Prior, M.<br />
& Webster, M. 2008. Population decline of<br />
Leach’s Storm-petrel Oceanodroma leucorhoa<br />
within the largest colony in Britain and Ireland.<br />
<strong>Seabird</strong> 21: 77–84.<br />
O’Brien, S. H., Mitchell, P. I., Parsons, M. and<br />
Mavor, R. A. 2005. ‘<strong>Seabird</strong> Monitoring on St Kilda<br />
in 2003’. Unpublished JNCC Report, Aberdeen.<br />
Phillips, R. A., Thompson, D. R., & Hamer, K. C.<br />
1999. <strong>The</strong> impact of Great Skua predation on<br />
seabird populations at St Kilda: a bioenergetics<br />
model. Journal of Applied Ecology 36: 218–232.<br />
Ratcliffe, N., Vaughan, D., Whyte, C. & Shepherd,<br />
M. 1998. Development of playback census<br />
methods for Storm Petrels Hydrobates pelagicus.<br />
Bird Study 45: 302–312.<br />
Votier, S. C., Crane, J. E., Bearhop, S., De Leon, A.,<br />
McSorley, C. A., Minguez, E., Mitchell, P. I.,<br />
Parsons M., Phillips, R. A. & Furness, R. W. 2005.<br />
Nocturnal foraging by Great Skuas Stercorarius<br />
skua: implications for conservation of storm-petrel<br />
populations. Journal of Ornithology 147: 405–413.<br />
Warham, J. 1990. <strong>The</strong> Petrels: <strong>The</strong>ir Ecology and<br />
Breeding Systems. Academic Press, London.<br />
SEABIRD 21 (2008)<br />
101
102<br />
Short Notes<br />
Use of gulls rather<br />
than terns to evaluate<br />
American Mink<br />
Mustela vison control<br />
Craik, J. C. A.<br />
Email: clive.craik@sams.ac.uk<br />
Scottish Association for Marine Science,<br />
Dunstaffnage Marine Laboratory, Oban, Argyll<br />
PA37 1QA, UK.<br />
Ratcliffe et al. (2006) compared the breeding<br />
biology of terns Sterna spp. on the Uists,<br />
Western Isles, where American Mink Mustela<br />
vison (hereafter ‘mink’) had been removed,<br />
and on nearby Lewis, where there had been no<br />
mink control. <strong>The</strong>y showed that hatching<br />
success was significantly higher on the Uists,<br />
indicating that mink removal improved this<br />
aspect of breeding, but found no significant<br />
difference in colony productivity between the<br />
two areas. <strong>The</strong> years in which the comparisons<br />
were made, 1993 and 2005, were years of low<br />
tern productivity, probably caused by poor<br />
weather and food shortage, respectively. <strong>The</strong><br />
authors considered that such factors overrode<br />
any effect of mink by killing young terns<br />
before mink could take them, and suggested<br />
SEABIRD 21 (2008)<br />
that any effects of mink would be more<br />
detectable in years of high tern productivity<br />
when more young terns survived.<br />
This is a problem I have encountered in<br />
another part of Scotland – on small islands in<br />
the sealochs and sounds of mainland Argyll,<br />
including Mull and the adjacent part of<br />
Highland Region (hereafter ‘Argyll’). For<br />
almost 20 years, we have been removing mink<br />
from near tern colonies and comparing<br />
productivity with similar colonies in<br />
unprotected areas, using broadly the same<br />
field methods as Ratcliffe et al. (2006). <strong>The</strong>re<br />
have been years when mink removal led to<br />
marked improvement of tern productivity, but<br />
there were also years when other factors<br />
severely reduced tern productivity and mink<br />
removal had no detectable effect. Ratcliffe et<br />
al. (2006) and my own work demonstrate that<br />
tern breeding performance is so variable that<br />
it is of limited value as an indicator of the<br />
success of mink control. If mink removal is to<br />
be justified, we must be able to show that it<br />
has immediate measurable benefits.<br />
Showing that mink are absent in trapped areas<br />
by searching for signs such as scats and dens<br />
can be difficult, particularly since any mink are<br />
likely to be present at low density. Moreover,<br />
Table 1. Common Gull Larus canus productivity in areas with and without American Mink Mustela vison<br />
control in Argyll and Lochaber.<br />
Areas with mink control Areas without mink control<br />
No.of No. of Productivity No. of No. of Productivity % that mink<br />
pairs (No. of fledged (chicks/pair), pairs (No. fledged (chicks/pair), reduced<br />
Year colonies) young ‘a’ of colonies) young ‘b’ productivity<br />
by = (a-b)/a<br />
1996 477 (6) 385 0.807 460 (24) 84 0.182 77<br />
1997 547 (10) 378 0.691 577 (20) 230 0.399 42<br />
1998 700 (10) 612 0.874 357 (9) 134 0.375 57<br />
1999 828 (18) 510 0.616 506 (26) 132 0.261 58<br />
2000 682 (13) 490 0.718 764 (36) 281 0.368 49<br />
2001 899 (19) 760 0.845 471 (31) 182 0.386 54<br />
2002 760 (13) 508 0.668 375 (14) 89 0.237 65<br />
2003 652 (18) 383 0.587 584 (36) 220 0.377 36<br />
2004 761 (15) 504 0.662 285 (10) 138 0.484 27<br />
2005 636 (8) 703 1.105 493 (16) 138 0.28 75<br />
2006 715 (18) 704 0.985 544 (31) 134 0.246 75<br />
2007 552 (9) 258 0.467 455 (17) 118 0.259 45
Short Notes<br />
Table 2. Herring Gull Larus argentatus productivity in areas with and without American Mink Mustela vison<br />
control in Argyll and Lochaber.<br />
Areas with mink control Areas without mink control<br />
No.of No. of Productivity No. of No. of Productivity % that mink<br />
pairs (No. of fledged (chicks/pair), pairs (No. fledged (chicks/pair), reduced<br />
Year colonies) young ‘a’ of colonies) young ‘b’ productivity<br />
by = (a-b)/a<br />
1997 698 (6) 703 1.01 1709 (14) 809 0.473 53<br />
1998 754 (6) 805 1.07 2507 (25) 860 0.343 68<br />
1999 1700 (7) 1373 0.808 6492 (42) 2465 0.38 53<br />
2000 1637 (8) 1978 1.21 6905 (48) 5427 0.786 35<br />
2001 1386 (7) 1145 0.826 6791 (46) 4266 0.628 24<br />
2002 1294 (9) 1198 0.926 5571 (38) 3639 0.653 29<br />
2003 1521 (10) 1174 0.77 5675 (50) 3298 0.581 25<br />
2004 1650 (10) 1473 0.893 2183 (11) 1159 0.531 41<br />
2005 1031 (9) 887 0.86 2929 (20) 1557 0.5316 38<br />
2006 2198 (11) 2299 1.05 3066 (23) 2064 0.67 36<br />
2007 1073 (11) 798 0.744 3337 (28) 2477 0.742 0.3<br />
at any time, more mink may arrive from<br />
untrapped areas nearby (Birks 1986). Arguably,<br />
the only practical and effective way to<br />
demonstrate that local mink control has been<br />
successful is to show that it has prevented the<br />
damage that mink were causing.<br />
Mink can cause widespread breeding failure of a<br />
variety of ground-nesting birds, as well as terns.<br />
Since tern breeding success is a poor indicator of<br />
the success of mink control, breeding success of<br />
other species may serve the purpose better. In<br />
Argyll, we have measured the annual breeding<br />
success of Common Gull Larus canus and<br />
Herring Gull L. argentatus colonies since 1996<br />
and 1997, respectively, both in areas where<br />
mink have been removed (usually to protect<br />
terns) and in areas with no mink control. New<br />
methods have been developed for this purpose,<br />
allowing simple and rapid determinations of gull<br />
productivity (Craik 2000) and more efficient<br />
control of mink (Craik 2008). <strong>The</strong> study area<br />
includes most colonies along the mainland<br />
coast between Mallaig in the north and West<br />
Loch Tarbert in the south. In areas without mink<br />
control, in any year there were some island<br />
colonies with good productivity that were<br />
naturally free of mink for various reasons<br />
(chiefly distance from the mainland shore), and<br />
some colonies with zero or low productivity,<br />
some of which were affected by mink; this<br />
applied to both gull species.<br />
In each of the years 1996–2007, the productivity<br />
(total number of fledged young /total<br />
number of pairs in a group of colonies) of<br />
Common Gulls was considerably higher in the<br />
mink-free group than in the mink-affected<br />
group (Table 1). This was also true for Herring<br />
Gulls in each of the years 1997–2006. <strong>The</strong><br />
exception was in 2007 when productivity was<br />
almost identical in the two groups of Herring<br />
Gulls (Table 2). This seems to have been<br />
because, in 2007, some Herring Gull colonies<br />
lost many or most eggs and young during a<br />
storm-tide in mid May, while some others<br />
were more than usually affected by Brown<br />
Rats Rattus norvegicus.<br />
Hence, gull productivity provides a much<br />
more reliable indicator of the success of mink<br />
control than the breeding success of the terns,<br />
even when the mink control may be directed<br />
primarily at conserving tern colonies.<br />
Footnote: Ratcliffe et al. (2006) state in their Discussion: ‘For example, in south-west Scotland between 1990 and 2006, 58% of<br />
unprotected tern colonies were not mink-affected (J. C. A. Craik, unpubl.)’, ‘unprotected’ here meaning those with no mink control.<br />
While 42% of unprotected colonies showed definite signs of mink predation, the other 58% included some colonies which had low<br />
or zero productivity for unknown reasons, and where the possibility of predation by mink could not be excluded.Thus their statement<br />
should be rephrased: ‘At least 42% of unprotected colonies are known to have been attacked by mink.’<br />
SEABIRD 21 (2008)<br />
103
104<br />
Short Notes<br />
Acknowledgements<br />
I am most grateful to Norman Ratcliffe, Ian<br />
Mitchell and Martin Heubeck for comments and<br />
corrections to earlier drafts of this note, and to<br />
the many people in the Argyll area who have<br />
helped with mink control and seabird counting<br />
over the years. I particularly thank Rob, Audrey<br />
and Niall Lightfoot and the late Ian Hynd.<br />
Use of gulls rather than<br />
terns to evaluate American<br />
Mink Mustela vison control.<br />
A response to Craik (2008)<br />
Ratcliffe, N.<br />
Email: notc@bas.ac.uk<br />
British Antarctic Survey, Natural Environmental<br />
Research Council, High Cross, Madingley Road,<br />
Cambridge CB3 0ET, UK.<br />
I thank Clive Craik for the opportunity to<br />
clarify points made in the 2006 paper. Clive<br />
presents a convincing argument that gull<br />
rather than tern productivity was a better<br />
indicator of the benefits of mink control<br />
during his study. However, I maintain that I<br />
was justified in restricting the analysis to terns<br />
in my study for three reasons:<br />
<strong>The</strong> Hebridean Mink Project was designed to<br />
produce conservation benefits for a range of<br />
taxa that included terns, but not gulls.<br />
<strong>The</strong>refore I was obliged to test the efficacy of<br />
the project in terms of changes in the<br />
breeding success of terns, since that of gulls<br />
was irrelevant to the project’s objectives.<br />
In contrast to Clive’s study area, gulls and terns<br />
on the Western Isles occupy discrete habitats<br />
that differ with respect to mink predation risk.<br />
Gulls nest inland on moors where mink are rare,<br />
while terns nest on the coast where mink are<br />
more common.As such, gull productivity on the<br />
SEABIRD 21 (2008)<br />
References<br />
Birks, J. 1986. Mink. <strong>The</strong> Mammal Society, Anthony<br />
Nelson, Oswestry.<br />
Craik, J. C. A. 2000. A simple and rapid method of<br />
estimating gull productivity. Bird Study 47: 113–116.<br />
Craik, J. C.A. 2008. Sex ratio in catches of American<br />
Mink – how to catch the females. Journal for<br />
Nature Conservation 16: 56–60.<br />
Ratcliffe, N., Houghton, D., Mayo, A., Smith, T. &<br />
Scott, M. 2006. <strong>The</strong> breeding biology of terns on<br />
the Western Isles in relation to mink eradication.<br />
Atlantic <strong>Seabird</strong>s 8: 127–135 [published April 2008].<br />
Western Isles will be less sensitive to removal of<br />
mink than that of terns.<br />
An analysis of Clive’s own data demonstrated<br />
that tern productivity during 1998–2006 at<br />
colonies protected from mink was on average<br />
253% higher than that at unprotected ones<br />
(Ratcliffe et al. 2008).As such, tern productivity<br />
clearly has a greater value as an indicator of the<br />
effects of mink removal than Clive’s<br />
commentary suggests.<br />
I concede the point in the Footnote<br />
concerning the detectability of mink<br />
predation being less than one. Quantifying the<br />
likelihood of an unprotected colony escaping<br />
mink predation is therefore difficult, but data<br />
certainly show that unprotected colonies can,<br />
on occasion, escape predation and experience<br />
high productivity. Hence, detecting the effects<br />
of mink control statistically requires sampling<br />
at a large number of colonies, and certainly<br />
more than the two sampled on Lewis in 1992.<br />
Reference<br />
Ratcliffe, N., Craik, C., Helyar, A., Roy, S. & Scott,<br />
M. 2008. Modelling the benefits of mink<br />
management options for terns in West Scotland.<br />
Ibis 150 (Suppl. 1): 114–121.
Fish brought to<br />
young Atlantic Puffins<br />
Fratercula arctica on<br />
Burhou, Channel Islands<br />
Sanders, J. G.<br />
Email: jerry.sanders@cwgsy.net<br />
<strong>The</strong> Alderney Ornithological <strong>Group</strong>, PO Box 24,<br />
Alderney GY9 3AP, UK.<br />
Introduction<br />
During the last 100 years breeding numbers of<br />
Atlantic Puffins Fratercula arctica (hereafter<br />
‘Puffins’) in the Channel Islands and Brittany,<br />
the most southern colonies in the eastern<br />
Atlantic, have declined dramatically and in<br />
2000 totalled less than 600 pairs (Harris &<br />
Wanless 2004). Most colonies now have trivial<br />
numbers and that on Burhou, Alderney,<br />
Channel Islands is one of only two remaining<br />
probably viable breeding populations south of<br />
the English Channel, and is therefore regionally<br />
of considerable conservation importance.<br />
Short Notes<br />
occupied in three subcolonies in 2005 (L.<br />
Soanes pers. comm.), and a maximum of c.250<br />
individuals was present in early July 2006 and<br />
2007 (pers. obs.).<br />
Between 1969 (Operation Seafarer) and 2005,<br />
breeding Lesser Black-backed Gulls Larus fuscus<br />
on Burhou increased from 70 to 1,102 pairs and<br />
Herring Gulls Larus argentatus from 87 to 202<br />
pairs, while c.15 pairs of Great Black-backed<br />
Gulls Larus marinus also presently breed (L.<br />
Soanes pers. comm.). Interactions with Puffins<br />
include gulls loafing near Puffin burrows,<br />
kleptoparasitism by Lesser Black-backed and<br />
Herring Gulls, and predation by Great Blackbacked<br />
Gulls, which became acute in the 1970s<br />
and 1980s but diminished after 1996 (Lockley<br />
1953; Sanders 2007).<br />
Nothing appears to have been known of the diet<br />
of Puffin chicks at Burhou and this note presents<br />
the results of an attempt to remedy this.<br />
Burhou (49o44’ N, 2o16 W) in the Alderney<br />
group is the most northerly Channel Island,<br />
about 760 m long by 300 m wide, 18 hectares<br />
in area, and 21 m above sea level at its highest<br />
point. Rabbits Oryctolagus cuniculus are<br />
common, having been introduced by the 14th<br />
century (Coysh 1985), the vegetation is short,<br />
but no other mammals are present. <strong>The</strong> island<br />
is uninhabited and now closed to the public<br />
from mid March to late July. Puffin numbers<br />
were estimated at 50,000 pairs in 1948–1949<br />
(Lockley 1953), making it by far the largest<br />
colony on the southern side of the English<br />
Channel at that time. <strong>The</strong> population fell to<br />
c.350 individuals (but less than 100 breeding<br />
pairs) by 1980, through a combination of<br />
chronic oil pollution and major tanker sinkings,<br />
burrow trampling by unrestricted day-trippers,<br />
and changes in vegetation due to soil<br />
enrichment by gull faeces (Danchin &<br />
Cordonnier 1980; Hill 1989; Sanders 2007,<br />
2005). <strong>The</strong> population almost stabilised after<br />
access to the island became restricted in the<br />
Figure 1. Puffin Fratercula arctica with Gadidae, Burhou, July 2006.<br />
early 1980s (Sanders 2007), 120 burrows were Figure 2. Puffin Fratercula arctica with sandeels, Burhou, July 2007.<br />
SEABIRD 21 (2008)<br />
105
106<br />
Short Notes<br />
Table 1. Fish carried by Atlantic Puffins Fratercula arctica on<br />
Burhou, 2006 & 2007.<br />
Date in Time of No. fish Species<br />
2006 photo in load of fish<br />
01 July 12.22 5 Gadidae sp.<br />
01 July 17.17 3 Gadidae sp.<br />
02 July 07.58 12 Small sandeels<br />
02 July 08.00 5 Gadidae sp.<br />
02 July 08.04 4 Gadidae sp.<br />
02 July 08.16 7 Gadidae sp.<br />
02 July 08.21 3 Gadidae sp.<br />
02 July 08.35 6 Gadidae sp.<br />
02 July 08.57 4 Gadidae sp.<br />
02 July 09.13 3 Gadidae sp.<br />
02 July 09.17 4 Gadidae sp.<br />
02 July 11.42 11 Small sandeels<br />
02 July 11.47 4 2 small sandeels, 2 Gadidae sp<br />
02 July 11.51 2 Gadidae sp.<br />
02 July 12.12 2 Gadidae sp.<br />
02 July 12.21 2 Gadidae sp.<br />
02 July 12.31 4 Gadidae sp.<br />
Total 81, of which: Small sandeels 25<br />
Gadidae sp. 56<br />
Date in Time of No. fish Species<br />
2007 photo in load of fish<br />
09 July 13.13 6 Large sandeels<br />
09 July 13.13 6 Mixed-size sandeels<br />
09 July 13.43 8 7 Large sandeels, 1 Gadidae sp.<br />
09 July 13.44 7 Large sandeels<br />
09 July 13.50 7 Small sandeels<br />
09 July 13.56 8 Large sandeels<br />
09 July 15.28 2 Large sandeels<br />
11 July 09.56 6 Small sandeels<br />
11 July 09.58 3 2 Gadidae sp,, 1 small sandeel<br />
11 July 10.02 5 Large sandeels<br />
11 July 10.02 6 Small sandeels<br />
11 July 10.04 8 Large sandeels<br />
11 July 12.25 10 Large sandeels<br />
Total 82, of which: Small sandeels 20<br />
Mixed-size sandeels 6<br />
Large sandeels 53<br />
Gadidae sp. 3<br />
SEABIRD 21 (2008)<br />
Methods<br />
<strong>The</strong> usual method of assessing Puffin chick diet<br />
is to catch adults carrying fish as they come<br />
ashore. However, the Burhou population is very<br />
small and since mist-netting in the 1980s<br />
appeared to cause undue disturbance to the<br />
main colony (pers. obs.; M. G. Hill pers. comm.),<br />
it was decided instead to photograph Puffins<br />
carrying fish, a somewhat novel method.<br />
On arrival at the main colony Puffins carrying<br />
fish usually circled a number of times partly<br />
over the burrow area and partly over the sea,<br />
and if there were loafing gulls near the<br />
burrows, as was often the case, they<br />
frequently landed on the water in the bay<br />
adjacent to the colony, from which they could<br />
see the location of their burrow. <strong>The</strong>se rafting<br />
Puffins were photographed on 1 & 2 July 2006<br />
and 9 & 11 July 2007 using a digital Nikon<br />
Coolpix 995 camera with a Leica APO 77<br />
spotting telescope as lens. Taking the photos<br />
may in some cases have kept the Puffins away<br />
from their burrows for a slightly longer period<br />
than usual, but at other times it was possible<br />
to photograph without observer disturbance.<br />
Of many hundreds of photographs taken, only<br />
those clearly showing the fish carried by<br />
individual Puffins were selected for this study.<br />
To prevent double-recording, identification of<br />
individual Puffins in the photographs was<br />
ensured by comparing the number, size,<br />
position and distribution in the bill of the fish<br />
carried by each Puffin. Fish were identified to<br />
species by M. P. Harris, and their size and<br />
number were noted. This was helped by there<br />
usually being several photos of the same bird.<br />
Results<br />
Photographs taken in 2006 showed 17 Puffins<br />
with a load of fish. Two were carrying only<br />
small sandeels Ammodytes sp. (probably 0group,<br />
young of the year), 14 carried fish of<br />
the cod family Gadidae, most likely Whiting<br />
Merlangius merlangus, and one load contained<br />
some of each.Average load was 4.8 fish, and of<br />
the 81 fish in total, 25 (31%) were sandeels,<br />
and 56 (69%) were gadids. (Table 1).
In the 2007 photographs, three and seven of<br />
the 13 Puffins were carrying only small and<br />
large sandeels, respectively. One Puffin had<br />
seven large sandeels and one probable gadid,<br />
one carried a mixture of small and large<br />
sandeels (mixed-size sandeels), and one<br />
carried a small sandeel and two gadids. <strong>The</strong><br />
average load (6.3 fish) was not significantly<br />
greater than in 2006 (t = -1.617, P = 0.117).<br />
Of the total of 82 fish, 20 (24%) were small<br />
sandeels, 53 (64%) were large sandeels, six<br />
(7%) were mixed-size sandeels, and three<br />
(4%) were gadids. On 23 July, all 25 loads seen<br />
clearly through a telescope appeared to<br />
contain only sandeels, some of which were<br />
large (M. P. Harris & S. Wanless pers comm.).<br />
Discussion<br />
Puffins returning to the colony with fish for<br />
their chick normally enter their burrow as<br />
quickly as possible to prevent the catch being<br />
stolen (Harris 1984). Where they are prevented<br />
from doing so by gulls loafing near the burrow<br />
entrance and have to stand around for a while,<br />
or in the case of Burhou raft on the sea, digital<br />
photography can be a useful and non-invasive<br />
tool for assessing chick diet, given that the fish<br />
are large and reasonably easy to identify. It<br />
does, however, have its limitations on Burhou in<br />
that calm conditions are needed to get good<br />
pictures. Of the three days that photographs<br />
were taken in 2007, one day produced no<br />
useable photographs due to the movement of<br />
the birds in rough seas.<br />
All Puffins photographed appeared to be<br />
carrying as many fish as they could cope with.<br />
All the fish were of good size, and the overall<br />
size of the loads was at least as good as is<br />
currently seen at most northern British<br />
colonies (M. P. Harris pers. comm.). Generally,<br />
gadids are not considered as good food for<br />
young seabirds since they are generally of<br />
lower calorific value than other more oil-rich<br />
species such as sandeels, and Sprats Sprattus<br />
sprattus and Herring Clupea harengus (Harris<br />
& Hislop 1978). However, the gadids<br />
photographed in 2006 were substantially<br />
larger than those normally fed to Puffin chicks<br />
and this will at least in part have compensated<br />
for their low energy density.<br />
Short Notes<br />
Chick diet at a colony can change dramatically<br />
within a season and between years (Harris<br />
1984), and the difference between 2006<br />
(predominantly gadids) and 2007 (predominantly<br />
sandeels) at Burhou was striking.Around<br />
Guernsey, 28 km from Burhou, where there is a<br />
small-scale fishery for sandeels, fishermen<br />
reported the sandeel cycle of 2006 to be late,<br />
and abundance the lowest for a long time, with<br />
none caught on some traditionally productive<br />
grounds (M. Roger pers. comm.). This would<br />
suggest that sandeel abundance around Burhou<br />
in 2006 was also lower than in most years, and<br />
the 2007 observations give rise to some<br />
optimism on the state of the recent summer<br />
food supply for Puffins on Burhou.<br />
Acknowledgements<br />
Thanks are due to the Burhou Warden, John<br />
Dupont for arranging the visits to Burhou, to<br />
Pierre Dupont for transport, to Mike Roger for<br />
permission to use the photographs, to Mike<br />
Harris for help with the manuscript and for<br />
identifying the fish in the photographs, to<br />
Louise Soanes for Puffin counts, and to Mark<br />
Atkinson, Alderney County Bird Recorder, for<br />
other assistance.<br />
References<br />
Coysh, V. 1985. Channel Islets. <strong>The</strong> Guernsey Press,<br />
Guernsey.<br />
Danchin E. & Cordonnier P. 1980. ‘Rapport<br />
d’Activité,Aurigny (Alderney) Iles Anglonormandes.’<br />
Unpublished Report, C.R.B.P.O., Museum de Paris.<br />
Harris, M.P. 1984. <strong>The</strong> Puffin. Poyser, Calton.<br />
Harris, M. P. & Hislop, J. R. G. 1978. <strong>The</strong> food of<br />
young Puffins Fratercula arctica. Journal of Zoology,<br />
London. 185: 213-236.<br />
Harris, M. P. & Wanless, S. 2004. <strong>The</strong> Atlantic Puffin<br />
Fratercula arctica. In: Mitchell, I. P., Newton, S. F.,<br />
Ratcliffe, N. and Dunn,T. E. (eds.) <strong>Seabird</strong> populations<br />
in Britain and Ireland: 392-406. Poyser, London.<br />
Hill, M. G. 1989. <strong>The</strong> Spread of Bracken Pteridium<br />
aquilinium in Burhou Island, Alderney from 1973 to<br />
1978. Report and Transactions of La Société<br />
Guernesiaise. 22: 628-632.<br />
Lockley, R. M. 1953. Puffins. Dent, London.<br />
Sanders, J. G. 2005. Alderney’s Breeding <strong>Seabird</strong>s.<br />
Report and Transactions of La Société Guernesiaise.<br />
25: 702-728.<br />
Sanders, J. G. 2007. <strong>The</strong> Birds of Alderney.<strong>The</strong> Press at<br />
St. Anne, Alderney.<br />
SEABIRD 21 (2008)<br />
107
108<br />
Reviews<br />
REVIEWS<br />
Petrels – night & day. A<br />
Sound Approach Guide By<br />
Magnus Robb, Killian<br />
Mullarney & <strong>The</strong> Sound<br />
Approach. <strong>The</strong> Sound<br />
Approach, Poole. 2008. ISBN<br />
978-90-810933-2-3. 300<br />
pages, numerous photos,<br />
maps, artwork and<br />
sonagrams, 2 CDs. Hardback,<br />
£34.95 plus p&p outside UK.<br />
Although I had commented on<br />
one chapter at the draft stage, I<br />
had little idea what to expect<br />
when the review copy of Petrels<br />
– night & day dropped on my<br />
desk. Certainly I had no inkling<br />
that, between the book’s covers,<br />
I would find an extraordinary<br />
combination of ripping yarns,<br />
beautiful artwork, two evocative<br />
CDs and scientific speculation<br />
that was well informed and bang<br />
up-to-date.<br />
How is this pulled off? Well, the<br />
bare bones are that, one by one,<br />
the 22 procellariiform species<br />
breeding in the northeast<br />
Atlantic, from Cape Verdes<br />
northwards and including the<br />
SEABIRD 21 (2008)<br />
Mediterranean, are given similar<br />
treatment; that is, an account of<br />
the authors’ visits to the species’<br />
colonies, a quick resumé of<br />
breeding biology, a more detailed<br />
description of calls illustrated by<br />
CD tracks, and a discussion of<br />
whether the vocalisations<br />
suggest taxonomic revision<br />
might be called for.<br />
<strong>The</strong> sheer enthusiasm in the face<br />
of the misadventures of island<br />
hopping is a delight.‘We had seen<br />
photographs of boats used by<br />
birders before, and they looked<br />
big enough to carry a football<br />
team. Our boat was not only very<br />
small, but also appeared to be<br />
falling apart’, describes a trip to<br />
Raso to search out Cape Verde<br />
Shearwaters Calonectris<br />
edwardsii. Having made the same<br />
journey, possibly in the same<br />
boat, to study Raso Larks Alauda<br />
razae, I know the feeling. Or again<br />
in the Cape Verdes, but this time<br />
in pursuit of White-faced Stormpetrels<br />
Pelagodroma marina, an<br />
outboard could only be provoked<br />
into temporary life by the<br />
insertion of tubing. When the<br />
boatman and suspect motor<br />
abandoned Magnus Robb, taking<br />
his recording kit, he pleaded with<br />
fate. ‘<strong>The</strong> optimist in me hoped<br />
that [the boatman] was just<br />
going back to get another<br />
mouthful of tubing. I politely<br />
asked the pessimist in me not to<br />
think at all.’<br />
Within this mesh of excitement<br />
emerge extraordinary facts that I<br />
did not know. For example the<br />
shearwaters of Menorca appear<br />
to be ill-understood hybrids of<br />
Puffinus mauretanicus and P.<br />
yelkouan. Quite amazingly, the<br />
Atlantic’s very first Swinhoe’s<br />
Storm-petrel Oceanodroma<br />
monorhis, ringed on the Salvages<br />
in 1983, was re-caught in August<br />
2007. Incidentally, the fact that<br />
that news is included in the text<br />
of this book, published early in<br />
2008, illustrates a speed of<br />
publication that shames the<br />
larger, established publishers.<br />
<strong>The</strong> photographs are of a consistently<br />
high standard. Some are<br />
outstanding: a White-faced<br />
Storm-petrel dancing off the sea<br />
surface, a Leach’s Storm-petrel O.<br />
leucorhoa sliding down the face<br />
of a spume-strewn yellow-brown<br />
wave breaking off the English<br />
coast. No less pleasing are Killian<br />
Mullarney’s plates, essentially one<br />
per species.While the plates show<br />
features crucial to identification,<br />
and the text captions are clearly<br />
aimed at the hardcore birder, the<br />
jizz of the birds is superbly<br />
captured and the plates just look<br />
aesthetically right.<br />
Dipping into the CDs was a treat.<br />
I reminded myself of the<br />
difference between Madeiran O.<br />
castro and Leach’s Storm-petrels.<br />
I learnt the sound of Bulwer’s<br />
Petrel Bulweria bulwerii, a species<br />
I have never heard. Perhaps the<br />
ideal listening arrangement<br />
would be a foul winter’s night,<br />
the wind lashing on the<br />
windows, the log fire subsiding,<br />
and the chance to let gruff<br />
female and higher-pitched male<br />
Cory’s Shearwater C. diomedea
calls drag me on a magic<br />
acoustic carpet to a balmy<br />
Macaronesian evening.<br />
Reviews such as this almost<br />
routinely carry a ‘but’. My but is<br />
small (so to speak). <strong>The</strong> authors<br />
place great weight on the use of<br />
calls as an aid to taxonomic<br />
decisions. Not for a moment<br />
would I deny their utility but, for<br />
me, a difference in the calls of<br />
two populations does not,<br />
without corroborating evidence<br />
from other features, provide<br />
sufficient justication for a<br />
Antarctica: Exploring a<br />
Fragile Eden By Jonathan<br />
and Angela Scott. Collins,<br />
London, UK. 2007. ISBN 10<br />
0-00-718345-3. 255 pages,<br />
many colour photographs,<br />
line drawings and maps.<br />
Hardback, £25.<br />
<strong>The</strong> Scotts are well known for<br />
their work in the Masai Mara and<br />
on the BBC with Big Cat Diary<br />
and other series. <strong>The</strong>ir passion<br />
for Antarctica is much in<br />
evidence throughout this book.<br />
It is a very personal narrative,<br />
filled with observed cameos of<br />
their numerous visits to the<br />
region. Reading it I almost feel I<br />
am travelling on a cruise ship<br />
with them, leaning over the rails<br />
watching Cape Petrels Daption<br />
capense, fighting past fur seals<br />
on a beach, or sitting in the<br />
ship’s lecture theatre learning<br />
about penguin biology. All I am<br />
missing is the vast quantities of<br />
luxurious food!<br />
Like many cruises to the continent,<br />
we first visit the Falkland Islands<br />
and then South Georgia before<br />
arriving at the Antarctic Peninsula.<br />
Interspersed are chapters on the<br />
discovery of Antarctica,<br />
albatrosses, Shackleton, whales,<br />
the Pole, and pack ice. Despite the<br />
many clichés (albatrosses are<br />
‘ocean wanderers’, whales are<br />
taxonomic split, at any level. For<br />
example, the splitting of the<br />
Mediterranean Storm-petrel as a<br />
full species Hydrobates melitensis<br />
might be a step too far for some.<br />
I could cite other, similar<br />
examples which generate the 22<br />
species mentioned above. In<br />
their defence, the authors<br />
explain their reasoning<br />
thoroughly and do not propose<br />
any scientific names other than<br />
those already existing.<br />
This is a terrific book, a record of<br />
a love affair with the birds and<br />
‘great leviathans’), and some<br />
rather odd metaphors (the<br />
southern ocean is ‘a giant watery<br />
glove’), these chapters do impart a<br />
fair bit of knowledge and are very<br />
readable. <strong>The</strong> fur seals’ demise at<br />
man’s hand and their subsequent<br />
recovery, and the Falkland’s squid<br />
fishery and its impact on<br />
Rockhopper Penguin Eudyptes<br />
chrysocome and Black-browed<br />
Albatross Thalassarche<br />
melanophrys populations there,<br />
are just two of many stories well<br />
told. <strong>The</strong>ir retelling of Shackleton’s<br />
epic journey, and other accounts of<br />
astonishing hardship and bravery<br />
by early polar explorers, greatly<br />
enhance this book.<br />
Inevitably a few mistakes have<br />
crept in, such as the authors<br />
getting confused by the identity<br />
of oystercatchers (describing a<br />
Blackish Haematopus ater yet<br />
calling it Magellanic H.<br />
leucopodus) and steamerducks<br />
(they call the Falkland<br />
Steamerduck Tachyeres<br />
brachypterus a Magellanic T.<br />
pteneres). Birders will be more<br />
concerned by their uncertain<br />
approach to albatross taxonomy.<br />
Despite stating that there are 24<br />
species, all forms are thereafter<br />
lumped in the conservative<br />
manner. A more consistent<br />
approach would have been<br />
better. <strong>The</strong>se are, however, rather<br />
Reviews<br />
their islands. It is also a<br />
convincing riposte to anyone<br />
who thinks passion and scientific<br />
curiosity are uncomfortable<br />
bedfellows. Buy it.<br />
M. de L. Brooke<br />
SPECIAL READER’S OFFER<br />
Petrels – night & day. £30 per<br />
book (normally £34.95)<br />
To order, phone 01202 676622 or<br />
email info@soundapproach.co.uk,<br />
quoting ‘<strong>Seabird</strong> <strong>Group</strong>’.<br />
minor mishaps that do not<br />
detract from the Antarctic<br />
experiences, which the Scotts<br />
generally convey well.<br />
As one would expect from two<br />
award-winning photographers,<br />
the photos are all good and<br />
occasionally sensational. Some<br />
can also however seem<br />
somewhat clichéd, views through<br />
ice caves having been taken on<br />
another Scott’s journey a<br />
century ago. You also feel you<br />
have seen some of the wildlife<br />
photos before, but I suppose<br />
there are only so many angles<br />
from which to view a King<br />
Penguin Aptenodytes<br />
SEABIRD 21 (2008)<br />
109
110<br />
Reviews<br />
patagonicus colony, or an Adelie<br />
Penguin Pygoscelis adeliae on ice.<br />
Birders will be disappointed by<br />
the range of species depicted. Of<br />
the 71 photos where birds<br />
feature, 54 depict eight species<br />
of penguin, including 17 Adelie,<br />
14 Emperor A. forsteri and 10<br />
King. Nine albatross photos<br />
feature four each of Blackbrowed<br />
and Wandering<br />
Diomedea exulans, and one of<br />
Royal, while Procellariidae are<br />
<strong>The</strong> Birds of Alderney By<br />
Jeremy G. Sanders. <strong>The</strong> Press<br />
at St Anne, Alderney. 2007.<br />
ISBN 978-0-946760-61-9.<br />
320 pages, line drawings<br />
throughout, three maps.<br />
Hardback, £25, available (incl.<br />
p&p) from J. G. Sanders, PO<br />
Box 24, Alderney, GY9 3AP,<br />
Channel Islands, UK.<br />
Alderney is the most northerly<br />
of the British Channel Islands<br />
lying just 16 km due west of the<br />
Cotentin peninsula of France. It<br />
is tiny, at 6 km long at its<br />
greatest length and 2.4 km wide<br />
at its greatest width. Some will<br />
SEABIRD 21 (2008)<br />
covered by just two giant petrels<br />
and a Cape Petrel and there are<br />
none of Hydrobatidae. <strong>The</strong> text is<br />
similarly biased, leaving birders<br />
the impression that this book is<br />
for general tourists, not them.<br />
<strong>The</strong> Scotts would no doubt make<br />
very good and knowledgeable<br />
companions on a visit to<br />
Antarctica, and to be fair it was<br />
their intention to tell the<br />
continent’s story as if on a cruise<br />
argue that there is little scope<br />
for a volume about the birds of<br />
this small place compared to<br />
one that deals with birds of the<br />
Channel Islands as a whole. I<br />
have a different opinion. For<br />
starters, an assemblage of all<br />
the information about the birds<br />
of Alderney is a precursor to a<br />
volume with a wider scope.<br />
Furthermore, Alderney is unique<br />
and some of its uniqueness<br />
would be lost in a volume<br />
covering all Channel Islands.<br />
Thus, I welcome Sanders’ book.<br />
<strong>The</strong> book covers more than 130<br />
years of recording, drawing on<br />
published and unpublished<br />
sources, including assessment of<br />
thousands of records from over a<br />
hundred observers. However, the<br />
contents by Sanders’ admission<br />
offer a personal view built around<br />
his own observations over a<br />
period of 25 years. <strong>The</strong> text is<br />
lightly peppered with line<br />
drawings by Carmen Watson,<br />
some of which are quite pleasing.<br />
<strong>The</strong> book comprises three<br />
introductory chapters, the<br />
systematic list (the bulk of the<br />
book), and a somewhat<br />
redundant section on sketches of<br />
Alderney birds that repeats<br />
content from earlier in the book.<br />
<strong>The</strong> first introductory chapter<br />
provides a general description of<br />
Alderney. We are briefly<br />
introduced to the climate and to<br />
prehistoric times, followed by<br />
ship, which is how the vast<br />
majority of people lucky enough<br />
to visit Antarctica will experience<br />
it. Many points of interest are<br />
well enough covered to give a<br />
good insight into the whole<br />
Antarctic experience. It is<br />
certainly worth a read before a<br />
visit, and makes a good souvenir<br />
of an Antarctic cruise.<br />
Richard Schofield<br />
more extensive accounts of the<br />
various habitats including town,<br />
farmland, and coastal regions. <strong>The</strong><br />
second provides an overview of<br />
the birds in categories including<br />
seabirds, coastal wading birds,<br />
inland breeding birds, migrants,<br />
and vagrants. <strong>The</strong> third<br />
introductory chapter reviews<br />
ornithology on Alderney. An<br />
important element of this for the<br />
systematic list is the reliability of<br />
the records. <strong>The</strong>re will always be<br />
issues in this regard for a volume<br />
that covers over a century of<br />
records, especially with the<br />
earliest records. For Alderney,<br />
issues of validation of records<br />
persist in modern times since<br />
there is no single formal records<br />
panel. Indeed, unavoidably the<br />
author more-or-less is the sole<br />
adjudicator, but is transparent in<br />
this regard.<br />
<strong>The</strong> systematic list follows the<br />
species order in use by the British<br />
Ornithologists Union in 2006.<br />
Each account starts with the<br />
species’ English name and<br />
scientific name, briefly states the<br />
status on Alderney, and then<br />
delivers a succinct yet comprehensive<br />
account of its<br />
occurrence on Alderney from the<br />
earliest to most recent records. I<br />
would like to have seen a<br />
checklist of the birds of Alderney<br />
following the systematic list.<br />
Alderney is probably most<br />
famous for its important
Northern Gannet Morus<br />
bassanus colonies and for this<br />
reason Northern Gannet<br />
correctly receives extra<br />
attention in the systematic list,<br />
in a span of ten pages, while the<br />
demise of Atlantic Puffin<br />
Fratercula arctica from 100,000<br />
birds in the 1940s to the current<br />
250 pairs is afforded 17 pages.<br />
<strong>The</strong>re are many species that<br />
have rarely been recorded on<br />
Alderney, including a few<br />
surprises like Gadwall Anas<br />
strepera on only two occasions,<br />
Wildlife of Seychelles By<br />
John Bowler. WILDGuides,<br />
Old Basing. 2006. ISBN 1-<br />
903657-06-7. 192 pages, 51<br />
photographic plates.<br />
Hardback, £17.95 (but see<br />
reader offer).<br />
<strong>The</strong> opening of Seychelles<br />
International Airport in July 1971<br />
led to a surge in tourism and also<br />
in biodiversity research. Apart<br />
from the Smithsonian Preliminary<br />
field guide to birds of the Indian<br />
Ocean, published in 1963, not<br />
widely available, and including<br />
only rough monochrome<br />
sketches, the early air travellers<br />
had little help in identifying what<br />
they saw. Since that time there<br />
has, of course, been a huge proliferation<br />
of field guides worldwide,<br />
including two on Seychelles’ birds.<br />
John Bowler’s Wildlife of<br />
Seychelles aims to facilitate<br />
identification of most of the<br />
more common animals. A brief<br />
introductory section describes<br />
the islands’ geography and<br />
climate, habitats, and conservation.<br />
In this section the scope<br />
of the book is defined as the<br />
granitic islands, along with the<br />
two nearby coralline islands, Bird<br />
and Denis. This is on the basis<br />
that the more remote southern<br />
island groups are ‘relatively<br />
depauperate in terms of the<br />
terrestrial habitats they support<br />
putting it on par with Greenish<br />
Warbler Phylloscopus<br />
trochiloides! Such seemingly<br />
odd statistics are so typical of<br />
small island avifauna. Inclusion<br />
of Levantine (Yelkouan)<br />
Shearwater Puffinus yelkouan<br />
might raise a few eyebrows<br />
given issues surrounding safe atsea<br />
identification. Sacred Ibis<br />
Threskiornis aethiopicus is<br />
included as a bona fide visitor.<br />
Alderney boasts a fine list of<br />
continental vagrants. <strong>The</strong> only<br />
Nearctic vagrants are Blue-<br />
and the number of terrestrial<br />
species present’ and are also ‘….<br />
very difficult or expensive for the<br />
average person to visit’. While<br />
the latter may be true, it is a pity<br />
that the outer islands are<br />
dismissed in this way without<br />
further mention, since some<br />
harbour massive and internationally<br />
important breeding<br />
colonies of seabirds, along with<br />
the wintering of most of the<br />
world population of the Indian<br />
Ocean endemic Crab Plover<br />
Dromas ardeola; Aldabra and<br />
Cosmoledo also host indigenous<br />
and endemic land birds!<br />
<strong>The</strong> bulk of the book comprises<br />
brief species accounts,<br />
accompanied by photographs, of<br />
many of the islands’ commoner<br />
animals (birds, mammals, reptiles,<br />
amphibians, insects, arachnids,<br />
crustaceans, millipedes and<br />
centipedes, and molluscs), a<br />
significant proportion of which<br />
are endemic and many have not,<br />
to my knowledge, been illustrated<br />
before. Most photographs are<br />
excellent, while the dragonfly<br />
section includes a key<br />
accompanied by a helpful set of<br />
drawings. <strong>The</strong> texts for the birds<br />
include identification, voice,<br />
behaviour and breeding, whereas<br />
other taxa have a variety of<br />
headings, most including both<br />
identification and description; the<br />
contents of these two sections<br />
Reviews<br />
winged Teal Anas discors,<br />
Killdeer Charadrius vociferus,<br />
and Wilson’s Phalarope<br />
Phalaropus tricolor. Overall, the<br />
systematic list makes for<br />
interesting reading.<br />
This book is informative and<br />
must be read if visiting Alderney<br />
or studying the avifauna of the<br />
Channel Islands. It is well worth a<br />
read for general interest on a wet<br />
day or winter’s evening.<br />
Robert L. Flood<br />
are often repetitive. Checklists at<br />
the back of the book cover<br />
regularly occurring animals<br />
(including birds) and a more<br />
comprehensive one of birds<br />
recorded up to December 2005;<br />
in these, information on birds is<br />
duplicated, but using different<br />
symbols to designate status.<br />
All of the breeding seabirds are<br />
adequately described and<br />
illustrated, as are some of the<br />
regular migrants. However,<br />
describing the Black-naped Tern<br />
Sterna sumatrana as ‘possibly an<br />
annual vagrant’ ignores its<br />
breeding in the neighbouring<br />
Amirantes group, where<br />
important but unmentioned<br />
colonies of Roseate Terns S.<br />
dougallii arideensis also breed.<br />
Gygis alba is allocated the<br />
English name ‘Fairy Tern’, for<br />
SEABIRD 21 (2008)<br />
111
112<br />
Reviews<br />
which my preference is the more<br />
widely accepted ‘White Tern’.<br />
Notwithstanding these shortfalls,<br />
this book is undoubtedly a<br />
valuable addition to Seychelles<br />
guides, allowing visitors to gain a<br />
greater appreciation of the<br />
uniqueness of Seychelles’ fauna<br />
but, more importantly,<br />
Birds and Mammals of the<br />
Falkland Islands By Robin W.<br />
& Anne Woods. WILDGuides,<br />
Old Basing. 2006. ISBN 1-<br />
903657-10-5. 144 pages,<br />
over 100 colour photographs,<br />
43 plates. Hardback, £17.95<br />
(but see reader offer).<br />
As soon as you pick up this book,<br />
you get the feeling it is well<br />
bound and going to last the<br />
course. <strong>The</strong> front cover is<br />
appealing and just inside is a map<br />
of the islands, while opening the<br />
back cover reveals the grid codes<br />
used in the Breeding Birds Survey.<br />
<strong>The</strong> Introduction takes us<br />
through the islands’ topography,<br />
climate, vegetation, human<br />
presence and influence on the<br />
environment, predation on birds,<br />
and changes in the natural<br />
vegetation, before a succinct<br />
chapter on ‘Conservation:<br />
positive action for wildlife’. Here<br />
we learn among other things<br />
about the local WATCH group,<br />
the only overseas branch of this<br />
junior section of the UK Wildlife<br />
Trusts, so important for the<br />
future wildlife of the islands.<br />
<strong>The</strong>n come the birds – a brief<br />
introduction to them followed by<br />
a checklist; after which help in<br />
knowing ‘How to use this book’.<br />
Next we come to a brief<br />
summary of each bird family, with<br />
associated tables. Between this<br />
and the species accounts is an<br />
illustration of bird topography. If<br />
the book is to be used to its full<br />
potential, I would have found it<br />
SEABIRD 21 (2008)<br />
encouraging the Seychellois to<br />
learn about and value their<br />
heritage. For me, its greatest<br />
utility lies in its coverage of taxa<br />
not treated in other guides; keen<br />
birders will need more comprehensive<br />
field guides.<br />
(Another photographic guide<br />
with the same title, by Mike Hill<br />
easier to have the section on bird<br />
families after the bird topography<br />
section, integrated in some way<br />
with the species accounts –<br />
otherwise you are constantly<br />
looking back and forward to<br />
marry the two together.<br />
Each species’ text is succinct,<br />
containing a lot of information in<br />
a small space, including a box<br />
giving status details. <strong>The</strong> species<br />
are all illustrated with<br />
photographs opposite respective<br />
texts, over a hundred of which<br />
are by Alan Henry. Most are of<br />
high quality, although one or two,<br />
for example, Magellanic Diving<br />
Petrel Pelecanoides magellani<br />
and Grey-backed Storm-Petrel<br />
Garrodia nereis are not quite so<br />
sharp. Some, such as the Dolphin<br />
Gull Leucophaeus scoresbi and<br />
particularly the Variable Hawk<br />
Buteo polyosoma, show the range<br />
of different plumages that can be<br />
a minefield for the unwary.<br />
However, I found some of the<br />
annotations, explained in the bird<br />
plates’ introduction, a little<br />
difficult to read at times.<strong>The</strong> only<br />
drawback to my mind with<br />
photographs is that what you see<br />
is what you get: illustrations not<br />
necessarily comparable in<br />
plumage, size or jizz for similar<br />
and difficult species.<br />
But, as the title tells us, birds<br />
aren’t the only creatures covered<br />
here. So, following them, there is a<br />
good section on marine<br />
mammals and non-domesticated<br />
terrestrial mammals. Of the<br />
latter, the only native terrestrial<br />
and David Currie, was published<br />
by Collins in 2007)<br />
Chris Feare<br />
WILDGuides will be supporting the<br />
work of <strong>The</strong> Island Conservation<br />
Society and <strong>The</strong> Nature Protection<br />
Trust of Seychelles through<br />
proceeds of this book.<br />
mammal, the Falkland Islands Fox<br />
Dusicyon antarcticus, known by<br />
early settlers as the Warrah, was<br />
exterminated by 1876. All other<br />
species were introduced. Again,<br />
good photographs lie opposite<br />
respective texts.<br />
I’m rather a pedant regarding<br />
typographical errors and I’m<br />
pleased to say could find none.<br />
Altogether I liked this<br />
approachable book a lot, which is<br />
nothing less than expected with<br />
authors so synonymous with the<br />
birds and other wildlife of the<br />
Falkland Islands. I feel it will have<br />
a much wider audience than its<br />
stated aim of appealing to<br />
passengers on cruise ships and<br />
land-based visitors.<br />
Wendy Dickson<br />
A contribution will be made by<br />
WILDGuides Ltd. to the work of<br />
Falklands Conservation for every<br />
book sold.<br />
SPECIAL READER’S OFFER<br />
Wildlife of Seychelles and Birds<br />
and Mammals of the Falkland<br />
Islands. £15 per book (normally<br />
£17.95) or £27.50 if both ordered<br />
at the same time (p&p free).<br />
Orders, mentioning <strong>Seabird</strong> to be<br />
sent to WILDGuides, PO Box 680,<br />
Maidenhead, Berks SL6 9ST, UK<br />
with cheques made out to<br />
‘Wildguides’, or by phone for<br />
credit card orders 01628-529297
<strong>The</strong> <strong>Seabird</strong> <strong>Group</strong><br />
<strong>The</strong> <strong>Seabird</strong> <strong>Group</strong> is a registered charity, founded in<br />
1966 to promote and help co-ordinate the study and<br />
conservation of seabirds. <strong>The</strong> <strong>Group</strong> organises regular<br />
international conferences. Members receive, and can<br />
contribute to, regular newsletters, and the journal<br />
<strong>Seabird</strong>, published annually. Members are encouraged<br />
to participate in surveys of seabirds, monitoring<br />
programmes and other seabird research work. As part<br />
of this, members can apply for small grants towards<br />
the cost of research or survey projects.<br />
Membership<br />
Membership of the <strong>Seabird</strong> <strong>Group</strong> is open to all with an<br />
interest in seabirds, whether amateur enthusiasts, students or<br />
professionals.<br />
Membership Secretary: David Sowter, <strong>The</strong> <strong>Seabird</strong> <strong>Group</strong>,<br />
5 <strong>The</strong> Grove, Penwortham, Preston, PR1 0UU, UK.<br />
Email: davidsowter@freenet.co.uk<br />
Annual Membership Rates:<br />
Standard £ 10.00<br />
Standing Order £ 9.00<br />
Students £ 5.00<br />
the<br />
<strong>Seabird</strong><br />
<strong>Group</strong><br />
Current Executive Committee<br />
Norman Ratcliffe (Chairman), Alan Leitch (Secretary), Andrew<br />
Ramsay (Treasurer), David Sowter (Membership Secretary),<br />
Martin Heubeck (<strong>Seabird</strong> Editor), Linda Wilson (<strong>Seabird</strong><br />
Publishing Editor), Liz Humphreys (Newsletter Editor), Claire<br />
Smith (Assistant Newsletter Editor), Juan Brown, Simon Foster.<br />
General correspondence<br />
<strong>The</strong> <strong>Seabird</strong> <strong>Group</strong>, c/o BTO, <strong>The</strong> Nunnery, <strong>The</strong>tford, Norfolk<br />
IP24 2PU, UK. Email: seabird@bto.org<br />
Further information about the group, including grant<br />
guidelines and membership application can be found at:<br />
www.seabirdgroup.org.uk
Leach’s Storm-Petrel Oceanodroma leucorhoa, Merseyside, UK, 20th September 2004 © S. Young.<br />
ISSN 1757-5842