COMPLETE.pdf - The Seabird Group

VOLUME 21 (2008)

Seabird publishes papers and short 

communications on any aspect of 

seabird biology, conservation, identification, 

and status. The geographical 

focus of the journal is the Atlantic 

Ocean and adjacent seas, but contributions 

are also welcome from other 

parts of the world provided they are of 

general interest. Detailed guidelines 

for contributors are available at 

www.seabirdgroup.org.uk. 

Editor: 

Martin Heubeck, Sumburgh 

Lighthouse, Virkie, Shetland ZE3 9JN. 

Email: martinheubeck@btinternet.com 

Publishing Editor: 

Linda Wilson, JNCC, Dunnet House, 

7 Thistle Place, Aberdeen, AB10 1UZ. 

Email: linda.wilson@jncc.gov.uk 

Editorial Board: 

Tycho Anker-Nilssen (Norway), Josep 

Arcos (Spain), Kees Camphuysen (The 

Netherlands), John Chardine (Canada), 

Stefan Garthe (Germany), Morten 

Frederiksen (Denmark), Norman 

Ratcliffe (UK), Jim Reid (UK), Sarah 

Wanless (UK) 

Photographic Consultant: 

Hugh Harrop (www.hughharrop.com) 

Design & Typesetting: 

Harry Scott (Pica Design) 

Print & Production: 

F. Crowe & Son Ltd 

(using FSC paper) 

Front cover photo: 

Leach’s Storm-petrel © Sue Tranter 

Seabird 21 (2008) is the first re-launched 

volume of the Seabird Group’s former 

journal Seabird, which discontinued in 1998 

(Volume 20). Between 1998 and 2008, the 

Seabird Group published the journal 

Atlantic Seabirds (Vols 1–8) jointly with the 

Dutch Seabird Group (Nederlandse 

Zeevogelgroep, NZG). Back issues of these 

journals will shortly be available online at: 

www.seabirdgroup.org.uk. 

Registered Charity 260907 

SEABIRD 

Volume 21 (2008) 

CONTENTS 

PAPERS 

1 Identifying giant petrels, Macronectes giganteus and M. halli, 

in the field and in the hand CARLOS, C. J., & VOISIN, J.-F. 

16 Vagrancy of Brünnich’s Guillemot Uria lomvia in Europe 

VAN BEMMELEN, R., & WIELSTRA, B. 

32 A survey of Leach’s Oceanodroma leucorhoa and European 

Storm-petrel Hydrobates pelagicus populations on North 

Rona and Sula Sgeir, Western Isles, Scotland MURRAY, S., 

MONEY, S., GRIFFIN, A. & MITCHELL, P. I. 

44 The diet of European Shag Phalacrocorax aristotelis, Blacklegged 

Kittiwake Rissa tridactyla and Common Guillemot 

Uria aalge on Canna during the chick-rearing period 

1981–2007 SWANN, R. L., HARRIS, M. P. & AITON, D. G. 

55 Colony habitat selection by Little Terns Sternula albifrons in 

East Anglia: implications for coastal management RATCLIFFE, 

N., SCHMITT, S., MAYO, A., TRATALOS, J. AND DREWITT, A. 

64 Descriptive anatomy of the subcutaneous air diverticula in 

the Northern Gannet Morus bassanus DAOUST, P.-Y., DOBBIN, 

G. V., RIDLINGTON ABBOTT, R. C. F. & DAWSON, S. D. 

77 Population decline of Leach’s Storm-petrel Oceanodroma 

leucorhoa within the largest colony in Britain and Ireland 

NEWSON, S. E., MITCHELL, P. I., PARSONS, M., O’BRIEN, S. H., AUSTIN, 

G. E., BENN S., BLACK J., BLACKBURN, J., BRODIE, B., HUMPHREYS, E., 

LEECH, D., PRIOR, M. & WEBSTER, M. 

NOTES 

85 Rafting behaviour of Manx Shearwaters Puffinus puffinus 

WILSON, L. J., MCSORLEY, C. A., GRAY, C. M., DEAN, B. J., DUNN, T. E., 

WEBB, A. & REID, J. B. 

93 Late breeding by Great Cormorants Phalacrocorax carbo 

CRAIK J. C. A. & BREGNBALLE T. 

98 A pilot study of the phenology and breeding success of 

Leach’s Storm-petrel Oceanodroma leucorhoa on St Kilda, 

Western Isles MONEY, S., SÖHLE, I. & PARSONS, M. 

102 Use of gulls rather than terns to evaluate American Mink 

Mustela vison control CRAIK, J. C. A. 

104 Use of gulls rather than terns to evaluate American Mink 

Mustela vison control. A response to Craik (2008) RATCLIFFE, N. 

105 Fish brought to young Atlantic Puffins Fratercula arctica on 

Burhou, Channel Islands SANDERS, J. G. 

108 REVIEWS

Identification of giant petrels 

Identifying giant petrels, Macronectes 

giganteus and M. halli, in the field 

and in the hand 

Carlos, C. J. 1,2 *, & Voisin, J.-F. 3 

*Correspondence author. Email: cjcarlos@bol.com.br or macronectes1@yahoo.co.uk 

1 Laboratório de Elasmobrânquios e Aves Marinhas, Departamento de Oceanografia, 

Fundação Universidade Federal do Rio Grande, CP 474, 96201-900, Rio Grande, RS, 

Brazil; 2 Current address: Rua Mário Damiani Panatta 680, Cinqüentenário, 95013-290, 

Caxias do Sul, RS, Brazil; 3 USM 305, Département Ecologie et Gestion de la Biodiversité, 

Muséum National d’Histoire Naturelle, CP 51, 57 rue Cuvier, F-75 005 Paris, France. 

Abstract 

The two similar-looking species of giant petrels, the Northern Giant Petrel 

Macronectes halli and the Southern Giant Petrel M. giganteus, are renowned for being 

difficult to identify. In this paper we review and offer new guidelines on identification 

of these birds at sea, on land, and as dead specimens. Criteria for identifying giant 

petrels are available in the scientific literature, especially regarding bill-tip coloration 

which readily differ from one species to another. Plumage characters, although useful 

to discriminate species, are not adequately covered at present. Thus, for each species 

we describe in detail and illustrate distinctive age-related plumage stages, or types, 

from juveniles through to adult breeders.We also comment on giant petrel biometrics, 

body weight, and some aspects of their behaviour, in order to help ornithologists and 

birdwatchers separate males and females, and eventually specimens from South 

America–Gough Island, Antarctic and sub-Antarctic regions. 

Introduction 

Giant petrels Macronectes are large Procellariiformes with a wingspan usually in excess 

of two meters in males, slightly less in females, and are similar in size to Thalassarche 

mollymawks. They display a wide range of plumages, from dull-black to brownish-grey 

and even white, according to age and species. The discovery of two distinct 

populations breeding at Macquarie Island (Australian Antarctic Territory; Figure 1), 

laying about six weeks apart (Warham 1962), led Bourne & Warham (1966) to restrict 

the name Macronectes giganteus (Gmelin, 1789) to those birds breeding on Antarctic 

islands, the Antarctic Peninsula and continent, and on islands of the South Atlantic, and 

to resurrect the name Macronectes halli Mathews, 1912 for birds breeding on sub- 

Antarctic islands, from South Georgia through the Indian Ocean to the New Zealand 

area. Voisin & Bester (1981) and Voisin (1982b) showed later that the form breeding 

on the Falkland and Gough Islands (South Atlantic), as well as on islands off the 

Argentine coast, represented a well-marked subspecies: Macronectes giganteus 

solanderi Mathews, 1912, for which Carlos et al. (2005) proposed the vernacular name 

South Atlantic Giant Petrel. Hereafter, we follow current usage in referring to M. halli 

as Northern Giant Petrel and M. giganteus as Southern Giant Petrel. 

SEABIRD 21 (2008): 1–15 

1

2 


Ringing recoveries and subsequent satellite-tracking studies have shown that both 

species of giant petrel roam the southern ocean widely, especially immatures, but also 

adults in the breeding and non-breeding seasons (Hunter 1984; Voisin 1990; Parmalee 

1992;Trivelpiece & Trivelpiece 1998; González-Solís et al. 2000a, b; Patterson & Hunter 

2000; Patterson & Fraser 2000, 2003; BirdLife International 2004; González-Solís et al. 

2008). Antarctic populations of Southern Giant Petrel are especially wide-ranging, and 

one should not assume that, for example, a Southern Giant Petrel seen off the 

Argentinean–Brazilian coast at any time of the year is necessarily M. g. solanderi. 

Recently, Penhallurick & Wink (2004) proposed to re-group M. halli as a subspecies of M. 

giganteus on the basis of low mitochondrial cytochrome-b gene divergence. Such a 

position is untenable since both species have long bred sympatrically, and often in close 

proximity, on South Georgia, Marion, Crozet and Macquarie Islands (Figure 1) without 

regular interbreeding. A few cases of inter-specific breeding involving male M. giganteus 

paired to female M. halli have been reported from Marion (Burger 1978; Cooper et al. 

2001) and Macquarie Islands (Johnstone 1978), although the resulting eggs did not 

hatch. Only at South Georgia has occasional hybridisation been reported (Hunter 1982, 

1987). Given these circumstances, we can reasonably assume that gene flow between 

the two species is extremely limited. In birds, many closely related species are known to 

hybridise regularly, e.g. Anas dabbling ducks (Carboneras 1992), and Common 

Phylloscopus collybita and Iberian P. ibericus Chiffchaffs (Salomon et al. 2003), without 

merging into one single species. Even if ‘their apparent failure to interbreed is not quite 

as straightforward as if they bred at the same time without interbreeding’ (Penhallurick 

& Wink 2004), M. giganteus and M. halli do not genetically mix. 

Several authors have dealt with giant petrel identification, especially of birds at sea 

(Johnstone 1971, 1974; Conroy et al. 1975; Voisin 1982a; Hunter 1983; Voisin & 

Teixeira 1998; Jiguet 2000; Shirihai & Jarrett 2002). Here we review, and offer new 

guidelines on, giant petrel identification at sea, on land, and as dead specimens. 

Separating giant petrels from other Procellariiformes 

At sea, and in poor light conditions, inexperienced observers may confuse giant petrels 

with Diomedea albatrosses, Thalassarche mollymawks or Phoebetria sooty albatrosses. 

When flying, giant petrels adopt a typically hump-backed posture which, when viewed 

from the side, allows easy separation from albatrosses and mollymawks. Also, the 

massive, pale bill of giant petrels is quite different to the slender one of mollymawks 

and sooty albatrosses (Figure 2), and contributes to give them a clumsy appearance. 

Due to a larger wing-loading (Pennycuick 1982; Obst & Nagy 1992), the flight of giant 

petrels is more laboured (usually 4–5 flaps followed by a stiff-winged glide) than that 

of mollymawks and sooty albatrosses, which are of similar size but are decidedly more 

graceful. Macronectes halli and dark form M. giganteus can also be distinguished from 

juvenile Wandering Albatrosses (sensu lato) by their dark, not white, underwing and 

from sooty albatrosses by their short and round, not long and pointed, tail (Figure 2). 

The white form of M. giganteus (see Plumage characters) is distinguished from 

Diomedea albatrosses and mollymawks by its white, not dark remiges, and most often 

by its diagnostic random dark brown flecking. 

SEABIRD 21 (2008): 1–15


Figure 1. At-sea and breeding distribution of Northern Macronectes halli (Ng) and Southern Giant Petrels M. 

giganteus (Sg), and locations where the two species are known to breed sympatrically. (1) South Georgia, (2) Marion 

Island, (3) Crozet Islands, (4) Kerguelen Island, and (5) Macquarie Island. 

On land, separating giant petrels from albatrosses and 

mollymawks is straightforward for adults, because of 

their characteristic bill shape, smaller and flattened, not 

rounded, heads and plumage coloration. Although the 

bill shape of giant petrel chicks is somewhat different 

from that of adults, it is distinctive enough to allow 

separation from mollymawk chicks. Giant petrels nest in 

a variety of situations, from tight to loose colonies and 

even in solitary nests. Like all Procellariiformes, giant 

petrels are rather smelly birds, but their odour is stronger 

and less ‘sweet’ than that of albatrosses and 

mollymawks, and persists for a very long time on old 

nest material, feathers, and museum specimens, and 

once experienced is easy to recognise. 

Separating M. halli and M. giganteus 

Bill coloration: The best criterion for identifying giant 

petrels is bill coloration. In breeding M. halli, the basic 

colour of the bill is pinkish-horn to reddish-brown, 

turning brick-red on the maxillary and mandibulary 

ungues, or nails, and sometimes also on the latericorns 

Figure 2. Silhouettes, drawn to about the 

same scale, of (1) a Thalassarche mollymawk, 

(2) a Phoebetria sooty albatross, and (3) a 

giant petrel Macronectes. The birds are similar 

in size, but giant petrels have massive bills 

compared to the slender ones of mollymawks 

and sooty albatrosses. Giant petrels also have 

a shorter, round tail, which is longer and more 

pointed in sooty albatrosses. 

SEABIRD 21 (2008): 1–15 

3

4 


Figure 3. Northern Macronectes halli (plumage type H5; left) and Southern Giant Petrels M. giganteus (plumage type 

G7; right), Abattoir Outlet, Falkland Islands, 2006 © Steve Copsey. 

Figure 4. White form Southern Giant Petrel M. giganteus, South Georgia, 1998 © Ronald Saldino. 

(horny plates running along the maxilla or upper-mandible; Figure 3). A few bluishblack 

markings can also be found on both nails. The number of birds displaying such 

markings varies between populations, being frequent in some and rare in others.When 

seen from a distance and in good light conditions, the bill of M. halli may look ‘bicoloured’, 

with an obvious darker tip. 

In young M. halli chicks, the bill is light brown with a more or less pronounced reddish 

hue on the nails, and becomes coloured like that of an adult when the mesoptile 

down (the second set of down) is acquired. The reddish tinge on the bill of fledglings 

may fade somewhat and some immature birds have entirely yellowish-brown bills, 

with just a slight red colour on the nails. A few birds in adult plumage encountered at 

sea may also have uniformly coloured bills, and may be birds taking a ‘sabbatical 

leave’ from breeding (Voisin 1988). 

SEABIRD 21 (2008): 1–15


Figure 5. Heads of newly hatched Northern Macronectes halli (left) and Southern Giant Petrels 

M. giganteus (right) chicks (redrawn from Voisin 1968). 

In M. giganteus, the bill is yellowish with green nails, the tinge of which can vary from 

bluish to apple-green (Figure 3). In birds indulging in sexual displays, this green often 

extends to the latericorns, and mainly so on their lower part. This is especially so in M. 

g. solanderi, and birds breeding at Gough Island in particular may develop entirely 

green bills (Voisin & Bester 1981). When seen from a distance, the bill of M. giganteus 

may also appear ‘bi-coloured’, but the pale green nails contrast only slightly with the 

remaining yellowish bill. 

In young M. giganteus chicks, the bill is pinkish-cream with a bluish-green tip, and 

evolves to a pattern similar to that of adults, but usually somewhat paler. Even though 

bill colours fade rather in juvenile and immature giant petrels, they remain sufficiently 

characteristic for specific identification when seen in good light. 

Plumage characters: The plumage of M. giganteus is dimorphic, with a grey-brown, 

pale-headed dark form and a white form. According to Shaughnessy (1970), this 

dimorphism is controlled by two autosomal alleles with white dominant to dark.White 

form birds (up to 15% in some populations; Shaughnessy 1971) are entirely white 

from hatching to adulthood, except for a variable amount of dark brown feathers 

irregularly scattered throughout the body (Figure 4). Bill colour of the white form is 

identical to that of the dark form. Dark form birds have dark grey legs and feet, while 

white form ones have legs that vary from bluish-grey to pink-grey. Totally white- 

Figure 6. Southern Giant Petrels Macronectes giganteus: plumage type G2 (left) and G4 (right), Abattoir Outlet, 

Falkland Islands, 2006 © Steve Copsey. 

SEABIRD 21 (2008): 1–15 

5

6 


plumaged birds (leucistic; Conroy et al. 1975), lacking the dark brown mottling of the 

white form, with a uniform horn-coloured tip to a pinkish beak, and pink legs and feet, 

are regarded to be homozygous white (Shaughnessy & Conroy 1977). Identification of 

white form M. giganteus is straightforward at all ages, as M. halli has no white form. 

Chicks: Newly hatched dark form M. giganteus chicks are covered with pearl-grey 

down, sometimes with a slightly darker head. At Gough Island, however, some South 

Atlantic Giant Petrel chicks appear similar to M. halli chicks (Voisin & Bester 1981). At 

some locations, such as the Crozet and Kerguelen Islands, almost all newly hatched M. 

halli are whitish below and ash-grey on the back, darkest on the crown and nape, 

where it forms a cap contrasting with a paler forehead, face and sides of head (Voisin 

1968; Figure 5). At other sites they are simply darker grey than M. giganteus chicks. 

Chicks of both species in mesoptile down are of the same ash-grey colour. 

Juveniles and immatures: Fledgling Northern and dark form Southern Giant Petrels have 

the same uniform dull glossy blackish-brown to black plumage.According to locality, some 

individuals may also display a number of white feathers on their forehead and cheeks, as 

is frequently the case with M. halli chicks from the Crozet and Kerguelen Islands. 

After one year, these black juveniles moult into uniformly dull brown immatures 

(Figure 6). These birds remain mostly at sea for up to ten years or more, rarely coming 

to colonies (Conroy 1972; Hunter 1984; Voisin 1988), and their plumage changes are 

poorly understood. After each moult, their plumage acquires a more brownish tinge, 

and the area around the base of the bill becomes mottled and paler (Figure 6). 

Plumage differences between immature Northern and Southern Giant Petrels, if any, 

are very slight and of no help in specific identification. 

Adults: The plumage of adults, like that of immatures, becomes paler at each moult, 

but these changes do not progress in the same way in nominate M. g. giganteus, in M. 

g. solanderi and M. halli. 

Identification of white form and ‘totally white’ Southern Giant Petrels is straightforward. 

In the dark form of the nominate subspecies, a very few birds breed while still in dark 

plumage, having the area around the base of the bill paler, mottled yellowish and brown. 

The plumage then becomes more greyish-brown and the mottling around the base of 

the bill extends progressively, encompassing the face, sides of the head and throat. The 

head and neck then gradually become entirely whitish-grey, with a few feathers on the 

middle of the crown and along the dorsal side of the neck remaining darker brown. Old 

birds have almost entirely white heads and necks. Even in these, the chest, belly, and 

undertail are as dark as the upperparts, giving them a characteristic dark-bodied and 

white-necked appearance. In a few cases, the underparts may be slightly paler than the 

upperparts, but never whitish.The transition between the whitish colour of the head and 

neck and the dark colour of the rest of the body is narrow, and from a distance appears 

as a clear-cut demarcation line. Voisin (1982b) distinguished eight plumage stages in 

Southern Giant Petrels (Table 1, Figure 7). Many old birds have a pale leading edge to the 

wing, but this is not a diagnostic feature, as some old M. halli at South Georgia also show 

this character (Hunter 1983; Voisin & Teixeira 1998). 

SEABIRD 21 (2008): 1–15


Figure 7. Sketches of ‘old’ adult Southern Giant Petrels Macronectes giganteus, plumages types G5–G8 (based on 

Voisin 1982b). See Table 1 for descriptions and comments. 

Table 1. Plumage types of dark form Southern Giant Petrels Macaronectes giganteus (modified from Voisin 

1982b). White form birds are white throughout their lives. 

Type Description Comments 

G1 Entirely black birds. Juveniles (1st calendar year). 

G2 Entirely brown birds, the area around the base of the bill Nominate giganteus: immatures; 

not distinctly paler. M. g. solanderi: some breeders. 

G3 Entirely brown birds, the area around the base of the bill Nominate: older immatures and a 

light yellowish-brown. few breeders; M. g. solanderi: breeders. 

G4 Brown birds with a pale face and a pale area around the Both subspecies: breeders. 

base of the bill. 

G5 Brown birds with face, lower forehead and throat all pale; Nominate: breeders; M. g. solanderi: 

underparts as dark, or almost as dark as upperparts (Figure 7). breeders (most). 

G6 Birds with a pale face, sides of the head, throat and upper Nominate: breeders (most); M. g. 

foreneck; underparts as G5 (Figure 7). solanderi: breeders. 

G7 Birds with a pale head and neck, still having a brown cap on Both subspecies: breeders. 

the head and some dark feathering along the dorsal side of 

the neck; underparts as G6 (Figure 7). 

G8 Birds with entirely whitish heads and necks, with sometimes Both subspecies: breeders. 

a few dark feathers on the nape; underparts as G7 (Figure 7). 

Notes: Birds belonging to types G7 and G8 show a fairly clear-cut demarcation line between the whitish colour 

of their heads and necks and the dark colour of the rest of their bodies (Figure 7). An exception to this is M. g. 

solanderi at Gough Island, which shows a more diffuse transition between both colours. Exceptionally, dark form 

M. giganteus may have their underparts slightly paler than their upperparts, but never light grey or whitish. It is 

not known whether this difference is permanent or just transitory. 

SEABIRD 21 (2008): 1–15 

7

8 


Figure 8. Sketches of ‘old’ adult Northern Giant Petrels Macronectes halli, plumages types H5–H8. See Table 2 for 

descriptions and comments. 

Table 2. Plumage types of Northern Giant Petrels Macronectes halli. 

Type Description Comments 

H1 Entirely black birds. Juveniles (1st calendar year). 

H2 Entirely brown birds, the area around the base of the bill not Immatures. 

distinctly paler. 

H3 Entirely brown birds, the area around the base of the bill light Immatures. 

yellowish-brown. 

H4 Brown birds with a pale face and a pale area around the base May breed. The pale area on upper 

of the bill and on the upperthroat. throat on average a little larger than 

in M. giganteus. 

H5 Brown birds with face, lower forehead and throat all pale Breeders. 

(Figure 8). 

H6 Birds with a pale face, sides of the head to ear coverts, foreneck Breeders. 

paler than hindneck, underparts a little paler than upperparts 

(Figure 8). 

H7 Pale areas on the head larger, nape dark, foreneck much paler Breeders. 

than hindneck, underparts much paler than upperparts (Figure 8). 

H8 Like H7, but dark areas on the nape and hindneck reduced; Breeders. 

chest and belly whitish (Figure 8). 

Notes: Types H1–H4 are virtually indistinguishable from types G1–G4 of M. giganteus. Type H5 can only be 

reliably determined in the field when seen in good conditions (see Figures 7 & 8). 

SEABIRD 21 (2008): 1–15


The South Atlantic Giant Petrel differs from the nominate subspecies in having a 

generally darker plumage, with a more intense grey tone. It also acquires a paler head 

and neck, but more slowly and at an older age. In the Falkland Islands and Argentinean 

populations, the border between the pale head and neck and darker body is as clearcut 

as in nominate giganteus, whereas it is more diffuse in the population breeding at 

Gough Island (Voisin & Bester 1981). 

In Northern Giant Petrels, the plumage of young breeders is quite similar to that of 

young breeding M. giganteus, and the two species cannot be separated on that 

character alone. As birds get older, the face, sides of the head and foreneck become 

Table 3. Measurements (mm) and body mass (kg) of male and female Southern Giant Petrels Macronectes 

giganteus from several breeding places. Data are: mean, (± standard deviation; range; sample size). (1) Falkland 

Islands (Voisin 1982b); (2) Chubut, Argentina (Copello et al. 2006); (3) Gough Island; (4) Crozet Islands; (5) 

Antarctic Peninsula (Voisin & Bester 1981); (6) South Orkney Islands (Conroy 1972); (7) Frazier Island, 

Antarctica; and (8) Macquarie Island (G. W. Johnstone pers. comm.). 

Males Females 

Wing (1) 510.8 (9.58; 500–525; 5) 485.5 (9.06; 470–500; 10) 

(2) 517 (12; 507–530; 3) 491 (0.9; 478–508; 8) 

(3) 507.5 (485–535; 4) 484.5 (478, 491; 2) 

(4) 533.2 (14.1; 492–555; 15) 494.9 (23.85; 460–522; 7) 

(5) 510 (25.17; 465–552; 8) 499.8 (13.44; 477–515; 8) 

(6) 553.5 (10.86; 534–571; 13) 518.9 (15.21; 498–541; 13) 

(7) 542 (8.54; 530–550; 7) 512 (7.18; 500–518; 5) 

(8) 552 (10.66; 534–577; 20) 526.8 (8.52; 513–540; 21) 

Culmen (1) 97.9 (3.21; 94.5–102; 5) 84.4 (2.55; 80–89; 10) 

(2) 95.5 (2.6; 91.7–101.2; 18) 82.9 (2.1; 80.12–87.3; 22) 

(3) 95.2 (2.3; 91.5–98.5; 17) 83.3 (2.9; 79.5–88; 12) 

(4) 104.7 (3.97; 97–111; 16) 89.1 (3.28; 84–94; 8) 

(5) 98.5 (2.02; 96–102; 9) 89.1 (3.97; 85.5–95 ; 8) 

(6) 101.4 (2.45; 97.4–108.2; 66) 87.4 (3; 82–97; 73) 

(7) 98.9 (2.29; 96–103.1; 7) 88 (2.09; 85.8–90.4; 5) 

(8) 101.9 (2.15; 95.8–105; 20) 89.7 (2.41; 84.6–93.5; 21) 

Tarsus (1) 93 (1.46; 91.5–95; 5) 84 (2.2; 79.5–86; 10) 

(2) 92.2 (2.1; 87.7–96.9; 15) 84.8 (3.6; 80.5–99.1; 22) 

(3) 95.3 (2.02; 92–99; 17) 88.3 (3.11; 85–95; 12) 

(4) 102 (2.64; 97–107; 14) 95.1 (6.83; 88–111; 8) 

(5) 94.7 (5.12; 85–101; 9) 89 (3.29; 84–93; 6) 

(6) 96.3 (1.69; 94–99; 13) 87.1 (5.45; 82–95; 13) 

(7) 99.3 (2.34; 95.8–102; 7) 90.4 (2.19; 87.8–93.8; 5) 

(8) 102.7 (1.71; 99.5–105.3; 20) 94.2 (1.88; 90.1–98; 21) 

Body mass (2) 3.5 (0.3; 15) 2.5 (0.2; 21) 

(3) 3.77 (0.36; 2.3–4.55; 11) 3.17 (0.11; 2.7–3.9; 11) 

(4) 4.93 (0.34 ; 4.2–5.5; 15) 3.95 (0.17; 4.7–3.3; 8) 

(6) 4.94 (0.41; 4.1–5.8; 37) 3.85 (0.37; 3–4.8; 37) 

(8) 5.14 (0.42; 4.3–5.6; 20) 4.2 (0.44; 3.3–5.2; 21) 

SEABIRD 21 (2008): 1–15 

9

10 


Table 4. Measurements (mm) and body mass (kg) of male and female Northern Giant Petrels Macronectes 

halli from several breeding places. Data are: mean, (± SD; range; sample size). (1) South Georgia (González- 

Solís et al. 2000b, González-Solís 2004); (2) Crozet Islands (J-FV, unpubl. data); (3) Macquarie Island (G. W. 

Johnstone, pers. comm.); (4) Chatham Islands (G. W. Johnstone & C. J. Robertson, pers. comm.). 

Males Females 

Wing (1) 526 (507–550; 71) 500.6 (479–518; 77) 

(2) 519.7 (2.86; 503–535; 14) 474 (8.76; 435–502; 8) 

(3) 532.7 (2.4; 509–545; 17) 497.2 (1.8; 483–513; 22) 

(4) 525.9 (1.67; 506–538; 20) 496 (2.11; 475–515; 21) 

Culmen (1) 103.7 (98.4–109.4; 71) 89.3 (84.3–94.4; 77) 

(2) 103.3 (0.82; 96–110; 17) 88.3 (1.06; 84–92; 8) 

(3) 103.9 (0.8; 95.1–109.7; 18) 89.8 (0.37; 86.3–92.4; 23) 

(4) 97.8 (0.63; 93.9–104.7; 20) 85.4 (0.42; 81.8–89.5; 21) 

Tarsus (1) 100.8 (95.5–106.9; 71) 90.7 (86.7–94.7; 77) 

(2) 105.7 (2.24; 91–123; 13) 90.8 (1.45; 84–95.5; 8) 

(3) 102.27 (0.71; 94.5–107.1; 18) 92 (0.39; 88.1–95.6; 23) 

(4) 94.4 (0.51; 89.1–98.5; 20) 83.4 (0.49; 80.8–89; 21) 

Mass (1) 4.6 (0.2; 33) 3.7 (0.2; 35) 

(2) 4.47 (0.42; 3.6–5.35; 14) 4.31 (0.62; 3.3–5.35; 7) 

(3) 4.49 (0.37; 3.85–5.4; 16) 3.38 (0.24; 3–3.85; 22) 

(4) 3.66 (0.32; 3.15–4.45; 19) 2.83 (0.17; 2.5–3.25; 21) 

more mottled and then gradually whiter, while the crown, nape and back of the neck 

remain mottled brown (Table 2, Figure 8). With age, the pale grey or whitish colour of 

the foreneck then extends progressively down the whole chest, belly and undertail 

coverts. At the same time, the upperparts become greyish-brown (but never as pale as 

in old M. giganteus). In this way, old birds acquire a dark-above, pale-below appearance, 

quite different from that of M. giganteus. 

In both species, it is possible that the plumage of males pales more rapidly with age 

than that of females. This has never been examined systematically, but one published 

photograph (Anonymous 2002) shows a female M. halli ringed as an adult in South 

Georgia in 1963 and recovered in the Falkland Islands in 2002, and therefore over 40 

years old, but still in stage H6 plumage. 

Sexual differences: No differences in plumage have yet been demonstrated between 

male and female giant petrels, except for the possibility (above) that male Northern 

Giant Petrels become paler sooner with age than females. On the other hand, there is a 

marked sexual dimorphism in size in both species, males being much larger and about 

20% heavier than females (Hunter 1983; González-Solís et al. 2000b; González-Solís 

2004; J-FV; Tables 3 & 4), and size differences may help distinguish the sexes of a pair at 

the nest or displaying. However, one should be extremely cautious in using body mass 

to sex birds, as it may change considerably during the year, with average variations of 

over 20% having been recorded for each sex in both species on the Crozet Islands (J-FV). 

SEABIRD 21 (2008): 1–15


Biometrics: Both giant petrel species are of similar size and mass where they breed 

sympatrically. Large samples measured at South Georgia (Hunter 1984), and more 

limited data from Crozet and Macquarie Islands (Tables 3 & 4) both suggest that M. 

giganteus is, on average, larger than M. halli although there is much overlap and the 

characteristics cannot be used in most individuals to confirm the species. Northern and 

Southern Giant Petrels breeding at localities where the other species does not occur 

(the Argentinean coast, Falkland and Gough Islands, and Terre Adélie for M. giganteus; 

Chatham Islands for M. halli) are smaller in size, and sometimes significantly smaller 

than their counterparts from other sub-Antarctic islands (Voisin & Bester 1981; Voisin 

1982b; Carlos et al. 2005; Tables 3 & 4). These differences may be slight but, for 

instance, a nominate M. giganteus and/or a M. halli seen amongst M. g. solanderi is 

easily detected, as it is clearly larger (Figure 9). 

Voice: The vocalisations of the two species are very similar, and consist of croaking, 

hissing sounds uttered during quarrelling and displays, as well as a kind of neighing, often 

emitted when the birds fly low over the observer’s head. The calls of M. giganteus are 

higher-pitched and uttered in a faster rhythm than those of M. halli, a difference best 

noticed when birds fly overhead (Voisin 1968, 1978; Shirihai & Jarrett 2002; C. Chappuis 

pers. comm.). At sea, both species are usually silent, except when quarrelling over food. 

Nest location: Giant petrels exhibit a varying degree of coloniality, from lone nests and 

small groups to large loose or tight colonies. However, lone nests or small groups may 

have become isolated because reproduction failed at other nests around them. M. 

giganteus reputedly nests more colonially than M. halli, but the difference is mainly a 

matter of statistics, although tight colonies, recalling gull colonies, seem to be 

encountered regularly in the South Atlantic Giant Petrel (Voisin 1982b). M. giganteus 

generally breeds in more open situations than M. halli, but the overlap is large and of 

little use in field identification, except for the few nests that are more or less hidden in 

cavities, which are M. halli (Voisin 1976, 1986). However, the South Atlantic Giant Petrel 

at Gough Island may breed in very sheltered sites, with the nest sometimes hidden 

completely among tussock grass clusters and tree ferns (Johnstone et al. 1976; 

Shaughnessy & Fairall 1976; Voisin & Bester 1981). 

Breeding periods: Where both species breed together, M. halli commences breeding 

about six weeks earlier than M. giganteus (Table 5). In South Georgia, where the climate 

is harsher, both species breed about six weeks later than at other sympatric locations 

(Hunter 1984, 1987). Similarly, M. giganteus lays from late October to early November 

in Terre Adélie (Mougin 1975). In contrast, South Atlantic Giant Petrels from Gough 

Island lay between late August and early September (Voisin & Bester 1981), whereas 

in the Falklands (Woods 1975) and Argentina (Humphrey & Livezey 1983; Quintana et 

al. 2005) they lay in late October. 

Dead specimens: Museum specimens can be identified by plumage if they are of 

‘old’ breeders. The distinctive coloration of giant petrels’ bills fades rapidly after 

death if the birds are not kept in good conditions. This is especially so for birds that 

have been kept for a certain time in fluids, some of which (e.g. formalin) can be very 

SEABIRD 21 (2008): 1–15 

11

12 


detrimental in this respect. Submergence in seawater and/or other fluids can result 

in an overall olive-green ground colour, very different from the reddish hue found in 

live M. halli, or the bluish-green in M. giganteus. This is often the case in museum 

specimens, and while enough of the original bill colours may remain to allow, or 

assist, specific identification in some specimens, others remain unidentified (Bourne 

& Warham 1966). Decomposition and exposure to sunlight can quickly alter bill 

colours in dead giant petrels, especially at low latitudes. In the case of freshly dead, 

stranded specimens, it is advisable to look at the side of the bill lying on the ground, 

which may have kept its coloration better (Figure 10). Post-mortem alteration is 

certainly the reason for some puzzling bill colour descriptions found on museum 

labels. Moreover, describing colours can be rather subjective, and such descriptions 

should be interpreted with caution. 

Figure 10. Freshly dead, stranded Northern Macronectes halli (left) and Southern Giant Petrels M. giganteus 

(right), State of Rio Grande do Sul in south Brazil, 2003 © Fernanda I. Colabuono. The side of the bill lying on the 

ground kept enough of the original colour to allow specific identification. 

SEABIRD 21 (2008): 1–15 

Conclusion 

Giant petrels have a reputation for being 

difficult to identify, but fortunately this is 

only partly true. If seen at close range (e.g. at 

sympatric breeding sites), even chicks and 

juveniles may be specifically identified using 

bill coloration alone, quite apart from 

plumage. With the exception of white form 

M. giganteus, bill colour, if seen under good 

light conditions, can still be the diagnostic 

Figure 9. A Northern Macronectes halli 

(background, plumage type H8) and two South 

Atlantic Giant Petrels M.g. solanderi (plumage 

type G1), Abattoir Outlet, Falkland Islands, 2006 

© Steve Copsey.


Table 5. Laying dates of Northern Macronectes halli and Southern M. giganteus Giant Petrels at locations 

where the two species breed sympatrically. 

Species Locality Range of laying dates Reference 

M. halli South Georgia 19 September–10 October Hunter (1984, 1987) 

Marion Island 4 August–1 September Burger (1978); 

Cooper et al. (2001) 

Crozet Islands 16 August–5 September Voisin (1968) 

Macquarie Island 11 August–6 September Johnstone (1978) 

M. giganteus South Georgia 30 October–24 November Hunter (1984, 1987) 

Marion Island 12 September–25 October Burger (1978); 

Cooper et al. (2001) 

Crozet Islands 26 September–17 October Voisin (1968) 

Macquarie Island 27 September–19 October Johnstone (1978) 

criterion for birds observed far away in the field or at sea, but the observer has mostly 

to rely on plumage characteristics of ‘old’ adults (i.e. plumage types H6–H8 for M. halli 

and G6–G8 for M. giganteus; see Figures 7 & 8), which means that a number of birds 

may remain unidentified. 

Acknowledgements 

We are grateful to Steve Copsey, Fernanda I. Colabuono and Ronald Saldino for 

allowing us to use their photographs. Renata F. Cunha kindly prepared Figures 7 & 

8. CJC received support from Fundação de Aperfeiçoamento de Pessoal de Nível 

Superior (CAPES), Brazil. The typescript benefited from the input of Bernie 

Zonfrillo, Stephen Hunter, Bill Fraser, an anonymous referee and the editorial 

advice of Martin Heubeck. 

References 

Anonymous, 2002. The oldest stinker in town? Falkland Conservation Newsletter (Sep. 2002): 4. 

BirdLife International, 2004. Tracking ocean wanderers: the global distribution of albatrosses 

and petrels. Results from the Global Procellariiform Tracking Workshop; 2003 September 1-5; 

Gordon’s Bay, South Africa. BirdLife International, Cambridge, UK. 

Bourne, W. R. P. & Warham, J. 1966. Geographical variation in the giant petrels of the genus 

Macaronectes. Ardea 54: 45–67. 

Burger, A. E. 1978. Interspecific breeding attempts by Macronectes giganteus and M. halli. Emu 

78: 234–235. 

Carboneras, C. 1992. Family Anatidae (Ducks, Geese and Swans). In: del Hoyo J., Elliott A. & 

Sargatal J. (eds.), Handbook of the Birds of the World. Vol. 1: 536–628. Lynx Edicions, Barcelona. 

Carlos, C. J., Voisin, J.-F. & Vooren, C. M. 2005. Records of Southern Giant Petrel Macronectes 

giganteus solanderi and Northern Giant Petrel M. halli off southern Brazil. Bulletin of the British 

Ornithologists Club 125: 288–292. 

Conroy, J. W. H. 1972. Ecological aspects of the biology of the Giant Petrel Macronectes 

giganteus (Gmelin) in the Maritime Antarctic. British Antarctic Survey Scientific Report 75: 1–74. 

Conroy, J. W. H., Bruce, G. & Furse, J. R. 1975. A guide to the plumage and iris colours of the 

giant petrels. Ardea 63: 87–92. 

SEABIRD 21 (2008): 1–15 

13

14 


Cooper, J., Brooke, M. L., Burger, A., Crawford, R. J. M., Hunter, S. & Williams, T. A. J. 2001. 

Aspects of the breeding biology of the Northern Giant Petrel (Macronectes halli) and the 

Southern Giant Petrel (Macronectes giganteus) at sub-Antarctic Marion Island. International 

Journal of Ornithology 4: 53–68. 

Copello, S., Quintana, F. & Somoza, G. 2006. Sex determination and sexual size-dimorphism in 

Southern Giant Petrels (Macronectes giganteus) from Patagonia, Argentina. Emu 106: 141–146. 

González-Solís, J. 2004. Sexual size dimorphism in Northern Giant Petrels: ecological correlates 

and scaling. Oikos 105: 247–254. 

González-Solís, J., Croxall, J. P. & Wood A. G. 2000a. Foraging partitioning between giant 

petrels Macronectes spp. and its relationship with breeding population changes at Bird Island, 

South Georgia. Marine Ecology Progress Series 204: 279–288. 

González-Solís, J., Croxall, J. P. & Wood, A. G. 2000b. Sexual dimorphism and sexual 

segregation in foraging strategies of Northern Giant Petrels, Macronectes halli, during 

incubation. Oikos 90: 390–398. 

González-Solís, J., Croxall, J. P. & Afanasyev, V. 2008. Offshore spatial segregation in giant 

petrels Macronectes spp.: differences between species, sexes and seasons. Aquatic 

Conservation: Marine and Freshwater Ecosystems 17: S22–S36. 

Humphrey, P. S. & Livezey, B. C. 1983. Giant Petrels (Macronectes giganteus) nesting in Chubut, 

Argentina. Gerfaut 73: 3–8. 

Hunter, S. 1982. Interspecific breeding in giant petrels at South Georgia. Emu 82 (suppl.): 312–314. 

Hunter, S. 1983. Identification of giant petrels Macronectes spp. Sea Swallow 32: 72–77. 

Hunter, S. 1984. Breeding biology and population dynamics of giant petrels Macronectes at 

South Georgia (Aves: Procellariiformes). Journal of Zoology, London 203: 441–460. 

Hunter, S. 1987. Species and sexual isolating mechanisms in sibling species of giant petrels 

Macronectes. Polar Biology 7: 295–301. 

Jiguet, F. 2000. The two giant petrels. Birding World 13: 108–116. 

Johnstone, G. W. 1971. Bird in the hand–giant petrels. The Australian Bird Bander 9: 86–87. 

Johnstone, G. W. 1974. Field characteristics and behaviour at sea of giant petrels in relation to 

their oceanic distribution. Emu 74: 209–218. 

Johnstone, G. W. 1978. Interbreeding by Macronectes halli and M. giganteus at Macquarie 

Island. Emu 78: 235. 

Johnstone, G. W. Shaughnessy, P. D. & Conroy, J. H. W. 1976. Giant petrels in the South 

Atlantic: new data from Gough Island. South African Journal of Antarctic Research 6: 19–22. 

Mougin, J.-L. 1975. Ecologie comparée des Procellariidae antarctiques et subantarctiques. 

Comité National Français dês Recherches Antarctiques 36: 1–195. 

Obst, B. S. & Nagy, K. A. 1992. Field energy expenditures of the Southern Giant Petrel. Condor 

94: 801–810. 

Parmelee, D. F. 1992. Antarctic Birds: Ecological and Behavioral Approaches. University of 

Minnesota Press, Minneapolis. 

Patterson, D. L. & Hunter, S. 2000. Giant petrel Macronectes spp. band recovery analysis from 

the International Giant Petrel Banding Project, 1988/89. Marine Ornithology 28: 69–74. 

Patterson, D. L. & Fraser, W. 2000. Foraging movements of Southern Giant Petrels on the 

Antarctic Peninsula: preliminary findings of a satellite tracking study during the breeding 

season. In: Flint, E. & Swift, K. (eds.) 2nd International Conference on the Biology and 

Conservation of Albatrosses and Petrels, Honolulu, Hawaii, USA, 8-12 May 2000. Marine 

Ornithology 28: 125–152. 

SEABIRD 21 (2008): 1–15


Patterson, D. L. & Fraser, W. 2003. Satellite tracking Southern Giant Petrels at Palmer Station, 

Antarctica. Feature Articles, Microwave Telemetry, Inc. 8: 3–4. 

Penhallurick, J. & Wink, M. 2004. Analysis of the taxonomy and nomenclature of the 

Procellariiformes based on complete nucleotide sequences of the mitochondrial cytochrome b 

gene. Emu 104: 125–147. 

Pennycuick, C. J. 1982. The flight of petrels and albatrosses (Procellariiformes) observed in 

South Georgia and its vicinity. Philosophical Transactions of the Royal Society of London Series B 

300: 75–106. 

Quintana, F., Schiavini A. & Copello, S. 2005. Estado poblacional, ecología y conservación del 

Petrel Gigante del Sur (Macronectes giganteus) en Argentina. Hornero 20: 25–34. 

Salomon, M., Voisin, J.-F. & Bried, J. 2003. On the taxonomic status and denomination of the 

Iberian Chiffchaff. Ibis 145: 87–97. 

Shaughnessy, P. D. 1970. The genetics of plumage phase dimorphism of the Southern Giant 

Petrel Macronectes giganteus. Heredity 25: 501–506. 

Shaughnessy, P. D. 1971. Frequency of the white phase of the Southern Giant Petrel, 

Macronectes giganteus. Australian Journal of Zoology 19: 77–83. 

Shaughnessy, P. D. & Conroy, J.W. D. 1977. Further data on the inheritance of plumage phases 

of the Southern Giant Petrel Macronectes giganteus. British Antarctic Survey Bulletin 45: 25–28. 

Shaughnessy, P. D. & Fairall, N. 1976. Notes on seabirds at Gough Island. South African Journal 

of Antarctic Research 6: 23–25. 

Shirihai, H. & Jarrett, B. 2002. The Complete Guide to Antarctic Wildlife: Birds and Marine 

Mammals of the Antarctic Continent and the Southern Ocean. Princeton University Press, 

Princeton and Oxford. 

Trivelpiece, S. G. & Trivelpiece, W. Z. 1998. Postfledging dispersal of Southern Giant Petrels 

Macronectes giganteus banded at Admiralty Bay, King George Island, Antarctica. Marine 


Voisin, J.-F. 1968. Les pétrels géants (Macronectes halli et M. giganteus) de I’lle de la Possession. 

L’Oiseau et la Revue Française d’Ornithologie (Special No.) 38: 95–122. 

Voisin, J.-F. 1976. Observations sur les pétrels géants de l’île aux Cochons (archipel Crozet). 

Alauda 44: 411–429. 

Voisin, J.-F. 1982a. Sur la détermination des pétrels géants Macronectes giganteus et M. halli de 

l’Océan indien sud. Info Nature, Ile de la Réunion 19: 157–165. 

Voisin, J.-F. 1982b. Observations on the Falkland Islands Giant Petrels Macronectes giganteus 

solanderi. Gerfaut 72: 367–380. 

Voisin, J.-F. 1988. Breeding biology of the Northern Giant Petrel Macronectes halli and the 

Southern Giant Petrel M. giganteus at île de la Possession, îles Crozet, 1966–1980. Cormorant 

16: 65–97. 

Voisin, J.-F. 1990. Movements of giant petrels Macronectes spp. banded as chicks at Iles Crozet 

and Kerguelen. Marine Ornithology 18: 27–36. 

Voisin, J.-F. & Bester, M. N. 1981. The specific status of giant petrels Macronectes at Gough 

Island. In: Cooper, J. (ed.) Proceedings of the Symposium on Birds of the Sea and Shore: 215–222. 

African Seabird Group, Cape Town. 

Voisin, J.-F. & Teixeira, D. M. 1998. The identification of giant petrels (Aves, Procellariidae) in the 

South Atlantic. Boletim da Fundação Brasileira para a Conservação da Natureza 25: 129–134. 

Warham, J. 1962. The biology of the Giant Petrel Macronectes giganteus. Auk 79: 139–160. 

Woods, R. W. 1975. The Birds of Falkland Islands. Anthony Nelson, Oswestry, England. 

SEABIRD 21 (2008): 1–15 

15

16 

Vagrancy of Brünnich’s Guillemot 

Vagrancy of Brünnich’s Guillemot Uria 

lomvia in Europe 

Van Bemmelen, R. 1 *, & Wielstra, B. 2 

*Correspondence author. Email: rvanbemmelen@gmail.nl 

1 Stavangerweg 535, 1013 AX Amsterdam, The Netherlands; 2 Dr. Benthemstraat 10–91, 

7514 CM Enschede, The Netherlands. 

Abstract 

We review the occurrence of vagrant Brünnich’s Guillemots Uria lomvia in Europe. The 

104 records of 109 individual birds that could be traced showed a distinct seasonal 

pattern. There were no September records, but a small autumn peak was apparent in 

late October and early November. Numbers increased again in early December, peaked 

in late January and early February, and declined through spring, with only seven 

records in summer. Autumn records were mostly of first-winter birds, whereas 

relatively more adults were recorded in winter, in line with expectations based on 

timing of migration for these different age classes. We speculate that vagrant birds to 

western Europe have strayed from the wintering grounds and migration routes south 

of Iceland and along the Norwegian coast, while overland movements from the 

Barents Sea may explain inland records from northern Scandinavia, and some from the 

Baltic Sea. Two spatial clusters of records were evident, one in Scotland and the other 

in the Skagerrak, Kattegat and southern Baltic Sea. Between 1975/76 and 2005/06, the 

number of records declined in the former region but increased in the latter, which may 

represent a real decrease in occurrence, and increased observer effort, respectively. 

Introduction 

Brünnich’s Guillemot Uria lomvia is a circumpolar species, breeding in both the high 

and low (sub) Arctic (Figure 1a) (Nettleship & Evans 1985). The Atlantic population 

numbers c. 6.5 million birds, breeding at colonies in eastern Canada, west and east 

Greenland, Iceland, Jan Mayen, Svalbard, northern Norway, and the western Russian 

Arctic (CAFF 2004). Ringing recoveries have shown that the breeding populations of 

Canada and Greenland winter in the northwest Atlantic, where they are joined by a 

substantial (but unknown) number of birds from colonies in Iceland, Svalbard, Norway 

and Russia, although many of the latter remain in the Norwegian and Barents Seas 

throughout the year (Kampp 1988; Nikolaeva et al. 1996; CAFF 2004; Bakken & 

Mehlum 2005). Thus, a large number of Brünnich’s Guillemots migrate in a 

southwesterly direction across the North Atlantic each autumn. Occasionally, 

Brünnich’s Guillemots are recorded further south in Europe than usual (Figure 1b). We 

review the occurrence of such vagrants, analyse their temporal and geographic distribution, 

and examine differences between age classes. We attempt to relate the 

patterns found to the normal seasonal distribution of the species in the Atlantic and 

speculate on possible origins and causes of these extra-limital occurrences. Finally, we 

discuss whether Brünnich’s Guillemots found in western Europe are truly lost, or 

whether they winter there in small numbers. 

SEABIRD 21 (2008): 16–31


Figure 1. The normal range of Brünnich’s Guillemot Uria lomvia in the North Atlantic (A), and the distribution of records 

of vagrants in Europe from 1900–2006 (B). Breeding areas are after Kampp (1988), boundaries of winter distribution are 

after Cramp (1985) and Gaston & Jones (1998), and extreme distribution limits are drawn from Gaston & Jones (1998). 

SEABIRD 21 (2008): 16–31 

17

18 


Methods 

A database was compiled of records of Brünnich’s Guillemot in Europe that have been 

accepted by national rarity committees, for which location (at least province) and date (at 

least month) were known. A cut-off date of 31 August 2006 was used. The species is 

considered by all European national rarities committees, except those of Iceland, Norway 

and Russia, which were therefore left out of the analysis. The German rarities committee 

has not yet examined records of Brünnich’s Guillemot, so we included only those German 

records for which museum specimens were available.We took a neutral stand on decisions 

made by rarities committees, but tried to gather as much additional detail as possible on 

age, sex and plumage by studying photos, drawings, and museum specimens. 

Records were not analysed by calendar year, but for intervals from September to 

August, hereafter referred to as years. Seasons were classed as autumn (September to 

November), winter (December to February), spring (March to May) and summer (June 

to August). Months were divided into three periods of approximately 10 days: early 

(1–10), mid (11–20) and late (21–28/30/31). Fisher’s Exact Tests were performed to 

test whether years where a season had records had been preceded more frequently by 

a season holding records compared to years in which the season of interest did not 

have records. A potential shift over time in the proportion of dead versus live birds was 

tested for using a binomial linear regression analysis. The software package R was used 

for statistical analyses (R Development Core Team 2005). 

Results 

We traced 104 records of European vagrants, distributed as follows: UK and Ireland 

(39), the southern North Sea and English Channel (16), the Skagerrak, Kattegat and 

Baltic Sea (41), Lapland (6), and inland Western Europe (2) (Figure 1, Appendix 1). Five 

records involved two birds together, making a total of 109 individuals. Brünnich’s 

Guillemot has only been recorded regularly since the mid 1950s, with just nine coastal 

records (ten individuals) and one inland record (two individuals) prior to 1950/51. 

During 1954/1955–2005/2006, an average of 1.9 individuals were recorded in Europe 

per year, with records in 41 (79%) of the 52 years (Figure 2). Since 1975, the number 

of individuals increased in the Skagerrak/Kattegat/Baltic Sea area (t [1.4] = 3.624, P = 

0.02), but declined in the UK/Ireland (t [1.4] = -3.292, P = 0.03) (Figure 3). A negative 

trend was also suggested for the southern North Sea and the English Channel, but this 

was not statistically significant (t [1.4] = -1.207, P = 0.294). 

The occurrence of Brünnich’s Guillemot showed a distinct seasonal pattern, with 19% 

of records in autumn, 56% in winter, 18% in spring, and 7% in summer (Figure 4). 

There were no records from mid August to late September. 

Data for 1954/55–2005/06 showed that for years in which Brünnich’s Guillemots 

were observed in a particular season, the preceding season did not hold records more 

frequently than those years that did not hold records in that season (Fisher’s Exact Test 

for: summer and the preceding spring (P = 0.589); summer and the preceding winter 

(P = 1.000); spring and the preceding winter (P = 0.237); and winter and the preceding 

autumn (P = 0.703, n = 52 for all comparisons)). 

SEABIRD 21 (2008): 16–31


Figure 2. The number of records of vagrant Brünnich’s Guillemots Uria lomvia per year in Europe during 

1950/1951–2005/2006. 

Of the 109 individual Brünnich’s Guillemots, 43 (39%) were found dead. The remaining 

66 (61%) were alive when discovered, but at least nine of these died soon after 

(including three in Sweden and one in Denmark that were shot). Seven birds were 

reported as oiled, five of which were dead when found while the other two died soon 

after being found. From 1975 onwards, the proportion of dead birds relative to live birds 

declined from 80% in 1975/76–1979/80 to 17% in 2000/01–2004/05 (z = 4.224, 4 df, 

P < 0.01). However, this did not hold for all regions. Whereas the proportion of dead 

individuals declined in UK and Ireland (z = 2.521, P = 0.01, 3 df), it remained stable in 

the Skagerrak/Kattegat/Baltic Sea area (z = 1.081, P = 0.279, 3 df) (Figure 3). Eight 

records have been accepted of birds seen only in flight, seven from Sweden and one from 

Denmark, occurring in October, November (2), January, February (2), April and May. 

Only 38 birds were reportedly aged, 25 as adult, one as a first-summer, and 12 as firstwinter. 

Three of the adults were found in autumn, ten were in winter, eight in spring 

and four in summer.The first-summer bird was in July, while six of the first-winter birds 

were in autumn and six were in winter. Apart from the 12 aged as first-winter (which 

by definition show a winter-plumaged head pattern), plumage details were noted for 

61 birds. Fourteen were in breeding plumage, in October, January (2), March, April (2), 

May (2), June (2), July (2), and August (2). These were all reported as adults, except for 

four Swedish records from April (2), May and June, and a British record for June, for 

which no age was given. The 47 birds noted as being in non-breeding plumage were 

recorded in all months from October until April. Eleven of these were aged as adults, 

in October, November, December (2), January (3), February, March (2), and April. No 

details on age were provided for the remaining 36, recorded in October (3), November 

(7), December (10), January (6), February (9), and March. (Figure 4, Appendix 1). 

Discussion 

Seasonal occurrence: Records of vagrant Brünnich’s Guillemots in Europe show a 

distinct seasonal pattern, and our results corroborated those from an earlier review of 

records from northwest Europe spanning the period 1963–1992 (Rønnest 1994).There 

SEABIRD 21 (2008): 16–31 

19

20 


Figure 3. The number of live (upper bar parts) and dead (lower bar parts) vagrant Brünnich’s Guillemot Uria lomvia 

individuals per 5-year period for three areas: the UK and Ireland, the Skagerrak, Kattegat and Baltic Sea, and the 

southern North Sea and English Channel. 

Figure 4. The seasonal occurrence of vagrant Brünnich’s Guillemot Uria lomvia individuals in Europe per 10-day 

period (n=105), divided into age and plumage category. Three Swedish records of four individuals could not be 

assigned to 10-day periods (two unaged birds in non-breeding plumage in January 1925 and another in February 

1962, and one with no plumage details in May 1955), and are therefore not shown. 

were no September records, but a small autumn peak from mid October to mid 

November coincides with the onset of migration of birds from Barents Sea colonies 

(Gaston & Jones 1998; Bakken & Mehlum 2005). Records increased again from early 

December to a peak in late January and early February, perhaps reflecting prolonged 

migration of Brünnich’s Guillemots, with many still arriving in wintering areas in 

December and January (Gaston & Jones 1998; Kampp 1988; Nikolaeva et al. 1996; 

SEABIRD 21 (2008): 16–31


Bakken & Mehlum 2005). Some individuals found in winter may have arrived in the 

area in autumn but taken time to weaken, die and beach, but our data suggested that 

winters with records were not more likely to follow autumns with records, and it seems 

more probable that the winter peak was mainly of newly arrived birds. 

Brünnich’s Guillemots arrive back at their colonies in March–May, depending on latitude 

and ice conditions (Cramp 1985; Isaksen & Bakken 1995; Gaston & Jones 1998). The 

declining number of records from mid February, with only a few from May to August, 

suggests most vagrants had either returned to their normal range or died by then. 

Whether summer records refer to new arrivals from the north, or to lingering but 

previously undetected birds is unclear, but summers with records did not more often 

follow springs or winters with records. Some summer vagrants were clearly very disorientated, 

such as the adults found inland in Germany in August 1987, and in Belgium in 

August 2006, the only inland records apart from those in northern Scandinavia. In 

contrast, there are two UK records of Brünnich’s Guillemots near Common Guillemot 

Uria aalge colonies (Farne Islands, July 1977; St Kilda, May/June 1992), and one of a bird 

present in a Common Guillemot colony for nearly a month (Shetland, June/July 1989). 

Seasonal distribution of age classes: First-winter birds tend to arrive at wintering 

areas in the northwest Atlantic in October and November, whereas adults mainly 

arrive during or after December (Gaston & Jones 1998; Bakken & Mehlum 2005). The 

later migration of adults (or its slower progression) might result in a higher proportion 

of first-winter birds among autumn records, and a higher proportion of adults in 

winter. Indeed, of the nine aged records in the small autumn peak of European 

vagrants, six were identified as first-winters and three as adults (Figure 2). 

Furthermore, the three birds in mid October (and possibly the three in early 

November) in non-breeding plumage that were not aged may also have been firstwinters, 

since adults are probably still in transitional moult at that time (Gaston & 

Jones 1998), but note the two records of adults in non-breeding plumage in late 

October and early November (Figure 4). Of the 16 aged individuals in winter, ten were 

reported as adults compared to six as first-winters. 

Ageing Brünnich’s Guillemots in the field is difficult, especially in winter (Cramp 1985; 

Blomdahl et al. 2003). Only nine of the live birds were reportedly aged in the field 

(eight adults and one first-winter). Separating first-winter Brünnich’s Guillemots from 

older birds in winter is not particularly difficult in the hand (Gaston 1984; Gaston & 

Jones 1998), so although criteria used for ageing corpses were not documented for 

most records, ages given for museum specimens and weakened birds that were caught 

were probably correct. Separating immature birds (non-adults after their first winter) 

from adults is much harder (Cramp 1985), and the number of immatures among the 

records listed may have been underestimated. Since immature birds generally do not 

return to the breeding areas (Cramp 1985; Bakken & Mehlum 2005), summer records 

may include more immatures than is realised. The only record aged as immature was 

in July (1978, St. Cyrus, Scotland), and among the four summer records of adults, Grant 

(1981) suggested that the live bird at the Farne Islands in July 1977 (Ribbands 1977) 

may have been a first-summer bird rather than an adult. 

SEABIRD 21 (2008): 16–31 

21

22 


Spatial distribution, origin, and possible causes of vagrancy: There are two obvious 

geographic clusters in the records of vagrant Brünnich’s Guillemot in Europe. Of the 82 

records during 1975/76 to 2004/05, 34 (41%) were from the UK and Ireland, mostly 

in northern Scotland, and 35 (43%) were from the Skagerrak, Kattegat and the 

entrance to the Baltic Sea (Figure 1b). The cluster of records in Scotland is not 

surprising, since Scotland is closest to the species’ normal migratory and wintering 

range around Iceland and the coast of northern Norway (Cramp 1985; Kampp 1988; 

Nikolaeva et al. 1996; Bakken & Mehlum 2005; Fraser et al. 2007). Brünnich’s 

Guillemots also stray further south along the Norwegian coast than their normal range 

(Engebretsen & Pettersen 2001), and such birds are probably the source of many of 

those recorded in the eastern Skagerrak and the Kattegat. 

The six inland records in northern Sweden and northern Finland, five in winter and one in 

spring, suggest that at least some birds recorded in the Baltic Sea may have flown 

overland across Lapland. That Brünnich’s Guillemots can move inland from the Barents 

Sea was illustrated by a large influx covering much of Finland from mid November 1902 

to mid January 1903 (Mela 1903).These records were not dealt with by the Finnish rarities 

committee so were not included in our analysis, but can be seen at 

http://koti.netplaza.fi/~pply/havikset/lajisto/support/urilom_1902.htm. In the same 

winter, two birds were shot inland in Norrbotten, northern Sweden. Mela (1903) 

attributed the invasion to an early onset of severe weather conditions in the Barents and 

White Seas. Such an influx, which must have involved many hundreds of birds, has never 

recurred, and the most recent inland records from the area were of single birds in northern 

Finland in December 1986 and northern Sweden in January 1987. In contrast, inland 

movements of Brünnich’s Guillemots in the Great Lakes region of Canada and northeast 

USA, mainly during 1890–1910, were thought to have been related to changes in prey 

abundance (Gaston 1988). The 1902/03 Finnish influx also involved Atlantic Puffins 

Fratercula arctica, but no other alcids were recorded in the Great Lakes movements. 

TEXTBOX: Little Auk invasions and Brünnich’s Guillemot vagrancy 

Four recent autumn reports of Brünnich’s Guillemot around the North Sea in 2005–2007, which have 

not yet been accepted by rarity committees, coincided with (or just preceded) large coastal 

movements (invasions) of Little Auks Alle alle, and it is tempting to speculate whether the two 

phenomena were related. Breeding and wintering areas of both species overlap to a large extent 

(Cramp 1985) and Camphuysen & Leopold (1996) suggested that Little Auks wintering in the North 

Sea originate from the Barents Sea, as we do for Brünnich’s Guillemots that turn up in the North Sea. 

Invasions of Little Auks generally coincide with their arrival on the wintering grounds in late autumn, 

and Camphuysen & Leopold (1996) suggested that such movements are triggered by local food 

shortages, with wind conditions playing only a secondary role in making the displacement visible to 

sea-watchers. Gaston (1988) noted a lack of synchrony between invasions of the two species in 

Canada and the northeast USA and attributed this to differences in diet. However, there may be 

some dietary overlap in winter in large zooplankton and small pelagic fish (Blake 1983; Skov et al. 

1989; Moody & Hobson 2007; Stempniewicz 2001), so food shortage may in certain circumstances 

affect both species simultaneously. However, as Gaston (1988), we found no support for 

synchronized occurrence. An analysis of years with high numbers of Little Auks (taken from 

SEABIRD 21 (2008): 16–31


Camphuysen & Leopold (1996)) and records of Brünnich’s Guillemots during autumn and winter, 

both only in countries directly surrounding the North Sea, did not find more records of Brünnich’s 

Guillemots in years with Little Auk invasions than in years without (analysed by year for 1975/1976 

- 1996/1997: Fisher’s Exact Test: n = 52, P = 0.380). However the few records of Brünnich’s 

Guillemots might make it impossible to detect any potential correlation. 

For Brünnich’s Guillemot sightings, see Van Bemmelen et al. 2005 (Figure 9); Birding World 18: 425, 

430 (photo), 19: 30; 19: 442, 447; 20: 447; British Birds 99: 658; 101: 54. For details on Little Auk 

numbers, see Van Bemmelen & Wielstra 2005; Birding World 18: 405; 19: 447; 20: 447, 450; British 

Birds 99: 658; 101: 54. 

Brünnich’s Guillemots breeding in the northeast Atlantic average slightly larger than 

those breeding in the western Atlantic, but there is considerable overlap in 

measurements (Cramp 1985; Gaston & Jones 1998), and biometrics are unlikely to 

help identify the geographic origin of a small sample of vagrant birds (but see Gaston 

(1988) for an attempt). Although stable-isotope analysis has been successfully applied 

to trace bird movements (Rubenstein & Hobson 2004; Fox et al. 2007; Fox & Bearhop 

2008), strong longitudinal gradients in isotope ratios are absent in the Atlantic 

(Kroopnick 1980; Rubenstein & Hobson 2004) and as the latitudinal distribution of 

Brünnich’s Guillemot colonies is relatively narrow (at least in the northeast Atlantic), 

assignment of individuals to specific geographic regions by this method seems 

problematic. It proved impossible to unambiguously segregate birds from different 

populations based on mitochondrial DNA sequence data (Birt-Friesen et al. 1992). 

Similar difficulties are to be expected for the nuclear genome, but no attempts have 

yet been made as far as we are aware. 

Given the above, the recovery of a Brünnich’s Guillemot ringed at a breeding colony 

seems the most likely way of identifying its origin. However, while extensive ringing 

has been carried out in the past (Kampp 1988; Nikolaeva et al. 1996; Bakken & 

Mehlum 2005), and expanded programmes have been proposed for the future (CAFF 

2004), the relatively small number of European vagrants means the chance of finding 

a ringed, extra-limital Brünnich’s Guillemot is slim, making the only known ringing 

recovery all the more remarkable, a bird ringed in August 1948 at Novaya Zemlya, 

Russia and caught alive in April 1964 in Poland (Tomialojć 1990; Rønnest 1994). 

As with other vagrant seabirds, one likely explanation for the occurrence of Brünnich’s 

Guillemots in western Europe is displacement by severe weather events. A number of 

live birds have been found moribund or sheltering in harbours after severe gales, and 

some temporal clusters of records coincided with or followed periods of strong 

northerly winds over the seas around and to the east of Iceland. This was the case in 

January 1981, when six birds were found after two spells of northerly gales, and again 

in January 1995, after an intense anti-cyclone in the mid Atlantic and low pressure 

over Scandinavia at the beginning of the year brought Arctic northerly gales across the 

northern North Sea; the five records that month were the only ones in 1994/95. 

However, records in other years (such as the seven in 1986/87) were more spread 

SEABIRD 21 (2008): 16–31 

23

24 


throughout the year and there was little obvious to connect them to any specific 

weather event. In some circumstances, wind may simply help land-based observers 

detect vagrants already in the area instead of being the cause of their displacement. 

Environmental factors other than brief weather events, and operating over a larger 

sea area and a longer timeframe, may also influence the occurrence of Brünnich’s 

Guillemots in Europe. Winter movements of Brünnich’s Guillemots are generally 

contained within cold currents of Arctic water temperature (Cramp 1985), and a 

reduction in sea temperature could cause a shift in the southerly limit of their normal 

range, decreasing the distance to the North Sea. Irons et al. (2008) found that 

population fluctuations of Brünnich’s Guillemots were associated with decadal 

changes in winter (January to March) sea surface temperature (SST) in the vicinity of 

breeding colonies, and plotted a 3-year running mean of deviation of SST in the 

northeast Atlantic from the 50-year average. Interestingly, the lowest two points on 

this graph (i.e. when SSTs were coldest) were in 1981 and 1987, coincident with 

above average number of vagrant Brünnich’s Guillemots (Figure 2), although a third 

peak in records (the seven in 1997/98, six of which were in the Skaggerak/Kattegat) 

occurred when SSTs had returned to above average levels. 

Trends and status: Vagrant Brünnich’s Guillemots have been only recorded regularly in 

Europe since the mid 1950s, with three records or more in 14 of the 29 years from 

1977/78. However, the balance has shifted between the two geographic clusters of 

records, with fewer in the UK and Ireland (Scotland, essentially) and more in the 

Skagerrak/Kattegat (Figure 3). Their seasonal occurrence also differed considerably, with 

more autumn records in the latter region (14) than the former (2). There is no obvious 

explanation for these disparities in both annual and seasonal occurrence. Given the 

increase in observer activity and identification skills in recent decades, and the increased 

proportion of live birds found (Figure 3), the decrease in the number of records in the UK 

and Ireland must reflect (and under-estimate) a decreased occurrence of the species 

there. Whether this is due to changes in population size, migratory routes or wintering 

areas, or disruptive climatic events or regimes is unknown. The contrasting increase in 

southern Scandinavia, however, where virtually all recent records have been of live birds 

(Figure 3), is more likely to reflect an increased sea-watching effort, a better 

understanding of alcid identification, higher quality optical equipment, and perhaps a 

greater willingness to submit and accept records of birds seen only in flight (since 

1991/1992, records of nine ‘migrating’ individuals have been accepted in Denmark and 

Sweden). Brünnich’s Guillemots may also have been ranging further south along the 

Norwegian coast than in the past, in which case they would be more likely to be found 

in southern Scandinavia than in the western North Sea, but there is no systematic 

evidence for this (G. Mobakken, T. Anker-Nilssen, R. Barrett pers. comm.). 

In the UK, the accumulation of Brünnich’s Guillemot records has led some to suggest 

the species may winter regularly in small numbers in the northern North Sea (e.g. 

Rogers et al. 1978), but others consider this unlikely (Cramp 1985; Fraser et al. 2007; 

McGeehan 1991). In Shetland, just three dead Brünnich’s Guillemots were found on 

systematic beached bird surveys from 1979 to 2004 compared to 15,107 Common 

SEABIRD 21 (2008): 16–31

Figure 7. Brünnich’s Guillemot Uria lomvia, adult summer, 

Lille, Antwerp, Belgium, 5 August 2006 © Tom Goossens. 


Figure 8. Brünnich’s Guillemot Uria lomvia, adult 

summer, Yell, Shetland, UK, 4 May 2006 © Mick Mellor. 

SEABIRD 21 (2008): 16–31 

Figure 5 (above). Brünnich’s 

Guillemot Uria lomvia, probable 

first-winter, Lerwick, Shetland, 

UK, December 2005 © Hugh 

Harrop. 

Figure 6 (left). Brünnich’s 

Guillemot Uria lomvia, probably 

2nd calendar year, Scousburgh, 

Shetland, UK, 25 March 2007 © 

Roger Riddington. This record fell 

outside the period under consideration 

here, but has been 

accepted by the British Birds 

Rarities Committee (Hudson et 

al. 2008). 

25

26 


Guillemots (Heubeck 2006), and Brünnich’s Guillemot has never been seen during 31 

years of ship-based inshore surveys of wintering seabirds there (M. Heubeck pers. 

comm.). The lack of any English record south of Northumberland (Figure 1), and totals 

of 45,357 Common Guillemots but no Brünnich’s Guillemots found on systematic 

beached bird surveys during 1965–2007 in The Netherlands (C. J. Camphuysen pers. 

comm.) also suggest no regular presence in the western or southern North Sea. 

Despite this, observers should always consider the possibility of encountering the 

species during beached bird surveys or sea-watches. Field identification of Brünnich’s 

Guillemot may be challenging, but is – even in flight – not insurmountable (McGeehan 

1991; Ullman 1998; Blomdahl et al. 2003). 


Many people helped us prepare this article and we thank them all. Earlier drafts were 

improved by comments from Martin Heubeck, who also provided information on some 

Shetland records. Mardik Leopold encouraged us and helped gather records and 

forward requests for information. Museum specimens were checked by Bob McGowan 

(National Museums of Scotland), Joan McLaren (Montrose Museum, UK), Mike Nicoll 

(Dundee Museum, UK) and Richard Sutcliffe (Glasgow Museum, UK). The following 

sent lists of national records: Jochen Dierschke (Germany), Roger Riddington (UK), 

Marnix Vandegehuchte (Belgium) and Georges Olioso (France), while Killian Mullarney 

commented on the Irish record. John Chardine and Eeva-Liisa Alanen provided useful 

references, Arnoud van den Berg helped at the Dutch Birding Association library, and 

we thank the EuroBirdNet community for their help. Finally, we thank two anonymous 

referees for their comments. 

Figure 9. Alcid identified in the field as a Brünnich's Guillemot Uria lomvia, Schiermonnikoog, 

Friesland, The Netherlands, 23 October 2005 © Martijn Renders. This 'fly-by' sighting was 

rejected by the Dutch rarities committee, but this decision is to be reviewed. 

SEABIRD 21 (2008): 16–31


References 

Bakken, V. & Mehlum, F. 2005. Wintering areas and recovery rates of Brünnich’s Guillemots 

Uria lomvia ringed in the Svalbard Archipelago. Arctic 58: 268–275. 

Birt-Friesen,V. L., Montevecchi,W.A., Gaston,A. J. & Davidson,W. S. 1992. Genetic structure 

of Thick-Billed Murre (Uria lomvia) populations examined using direct sequence analysis of 

amplified DNA. Evolution 46: 267–272. 

Blake, B. F. 1983. A comparative study of the diet of auks killed during an oil incident in the 

Skagerrak in January 1981. Journal of Zoology, London 201: 1–12. 

Blomdahl, A., Breife, B. & Holmström, N. 2003. Flight identification of European seabirds. 

Christopher Helm, London. 

Breife, B., Hirschfeld, E., Kjellén, N. & Magnus, U. 1990. Sällsynta fåglar i Sverige. Sveriges 

Ornitologiska Förening, Lund. 

CAFF. 2004. Distribution of murres outside the breeding season. “Circumpolar Murre Banding 

Program – North Atlantic Region”. CAFF Technical Report No. 13. CAFF International 

Secretariat, Akureyri, Iceland. 

Camphuysen, C. J. & Leopold, M. F. 1996. Invasies van de Kleine Alk Alle alle: voorkomen en 

achtergronden. Sula 10: 169–182. 

Cramp, S. (ed.) 1985. The Birds of the Western Palearctic.Vol. IV. Oxford University Press, Oxford. 

Dubois, P. J. & Yésou, P. 1991. Les oiseaux rares en france. Chabaud, Bayonne Cedex. 

Engebretsen, S. & Pettersen, M. 2001. Østfold neste stopp? – Polarlomvi Uria lomvia. Natur I 

Østfold 20: 79-84. 

Fox, T. A. D., Christensen, T. K., Bearhop, S & Newton, J. 2007. Using stable isotope analysis of 

multiple feather tracts to identify moulting provenance of vagrant birds: a case study of Baikal 

Teal Anas formosa in Denmark. Ibis 149: 622–625. 

Fox,T.A. D. & Bearhop, S. 2008. The use of stable-isotope ratios in ornithology. British Birds 101: 

112–130. 

Fraser, P.A., Rogers, M. J. & the Rarities Committee 2007. Report on rare birds in Great Britain 

in 2005. Part I: non-passerines. British Birds 100: 16–61. 

Gaston, A. J. 1984. How to distinguish first-year murres, Uria spp., from older birds in winter. 

Canadian Field-Naturalist 98: 52–55. 

Gaston, A. J. 1988. The mystery of the murres: Thick-billed Murres, Uria lomvia, in the Great 

Lakes region, 1890–1986. Canadian Field-Naturalist 102: 705–711. 

Gaston A. J. & Jones, I. L. 1998. The Auks. Oxford University Press, Oxford. 

Goethe, F. & Ringleben, H. 1964. Ein neuerer Nachweis von Uria lomvia (L.) für Deutschland. 

Journal of Ornithology 105: 54–56. 

Goossens, T. 2006. Kortbekzeekoet in Belgisch binnenland. Dutch Birding 28: 340–341. 

Grant, P. J. 1981. Mystery photographs 51. British Birds 74: 144–145. 

Heubeck, M. 2006. The Shetland beached bird survey 1979–2004. Marine Ornithology 34: 123–127. 

Irons, D. B., Anker-Nilssen, T. A., Gaston, A. J., Byrd, G. V., Falk, K., Gilchrist, G., Hario, M., 

Hjernquist, M., Krasnov,Y.V., Mosbeck, A., Olsen, B., Petersen, A., Reid, J. B., Robertson, G. J., 

Strøm, H. & Wohl, K. D. 2008. Fluctuations in circumpolar seabird populations linked to climate 

oscillations. Global Change Biology 14: 1455–1463. 

Hudson, N. & the Rarities Committee. 2008. Report on rare birds in Great Britain in 2007. 

British Birds 101: 516–577. 

Isaksen, K. & Bakken, V. 1995. Breeding populations of seabirds in Svalbard. Pp 11–15 In Isaksen, 

K. & Bakken, V. (eds.) Seabird populations in the northern Barents Sea – source data for the 

impact assessment of the effects of oil drilling activity. Norsk Polarinstitutt Meddelelser 135. 

SEABIRD 21 (2008): 16–31 

27

28 


Jaeschke, G. & Schulz, U. 1968. Dickschnabellumme (Uria lomvia) 1966 an der Mecklenburgischen 

Ostseküste. Journal of Ornithology 109: 131. 

Kampp, K. 1988. Migration and winter ranges of Brünnich’s Guillemots Uria lomvia breeding 

or occurring in Greenland. Dansk Ornitologisk Forenings Tidsskrift 82: 117–130. 

Kroopnick, P. 1980. The distribution of 13C in the Atlantic Ocean. Earth and Planetary Science 

Letters 49: 469–484. 

McGeehan, A. 1991. Brünnich’s Guillemot: rare or overlooked? Irish Birding News 1: 105–112. 

Mela, A. J. 1903. Pohjankiisla. Luonnon Ystävä 7: 37–45. 

Meyer, M. & Thorwarth, H. J. 1987. Erstnachweis der Dickschnabellumme (Uria lomvia) für das 

Binnenland der DDR. Beiträge zur Vogelkunde 34: 313–314. 

Moody, A. T. & Hobson, K. A. 2007. Alcid winter diet in the northwest Atlantic determined by 

stable isotope analysis. Marine Ornithology 35: 39–46 

Mullarney, K. 1988. Brünnich’s Guillemot in County Wexford – an addition to the Irish list. Irish 

Birds 3: 601–605 

Nettleship, D. N. & Evans, P. G. H. 1985. Distribution and status of the Atlantic Alcidae. In: Nettleship, 

D.N. & Birkhead T.R. (eds.). The Atlantic Alcidae. The evolution, distribution and biology of the auks 

inhabiting the Atlantic Ocean and adjacent water areas: 53–154. Academic Press, London. 

Nikolaeva, N. G., Krasnov, Y. V. & Barrett, R. T. 1996. Movements of Common Uria aalge and 

Brünnich’s Murre U. lomvia breeding in the southern Barents Sea. Fauna norvegica Series C, 

Cinclus 19: 9–20. 

Olsen, K. M. 1992. Danmarks fugle – en oversigt. Dansk Ornitologisk Forening, København. 

R Development Core Team 2005. R: A language and environment for statistical computing. R 

Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org. 

Ribbands, J. B. 1977. Brünnich’s Guillemot in Northumberland. British Birds 73: 225–226. 

Risberg, L. 1990. Sveriges fåglar. Aktuell översikt över deras utbredning, numerär och flyttning 

samt något om svensk ornitologi. Sveriges Ornitologiska Förening, Stockholm. 

Rogers, M. J. & the Rarities Committee. 1978. Report on rare birds in Great Britain in 1977. 


Rønnest S. 1994. Sjældne fugle i Danmark. En oversigt over forekomsten af sjældne fugle i 

Danmark og Nordvesteuropa 1963–1992. Forlaget Pinus, Skjern. 

Rubenstein, D. R. & Hobson, K. A. 2004. From birds to butterflies: animal movement patterns 

and stable isotopes. Trends in Ecology and Evolution 19:256–263. 

Skov, H., Durinck, J. & Danielsen, F. 1989. Sammenfaldende forekomst mellem Søkonger Alle 

alle og Glaskutling Aphya minuta i Flekkefjord, Sydnorge, November 1988. Pelagicus 4: 22–24. 

Stempniewicz, L. 2001. Little Auk Alle alle. BWP Update. The Journal of Birds of the Western 

Palearctic 3: 145–201. 

Tomialojć, L. 1990. Ptaki Polski. Rozmieszczenie i liczebno´sć. Państwowe Wydawnictwo 

Naukowe, Warszawa. 

Ullman, M. 1998. Fältbestämning av spetsbergsgrissla. Vår Fågelvärld 8: 28–33. 

Van Bemmelen, R. S. A., Wielstra, B., Renders, M. & Hendriks, K. 2005. Black-and-white 

weekend: Kleine Alken en Kortbekzeekoet langs Schiermonnikoog. Dutch Birding 27: 448–449. 

Van Bemmelen, R. S. A. & Wielstra, B. M. 2005. Invasie van kleine alken op 23 oktober 2005: 

kort maar krachtig. NZG Nieuwsbrief 7: 1–4. 

Van Gompel, J. 1981. De massale zeevogelsterfte aan de Belgische kust tijdens de voorbije 

winter. Wielewaal 47: 137–142. 

Van Gompel, J. 1982. First record in Belgium of a living specimen of Brünnich’s Guillemot (Uria 

lomvia). Wielewaal 48: 169–170. 

SEABIRD 21 (2008): 16–31


Appendix 1. 

List of vagrant Brünnich’s Guillemot records in Europe, arranged by country and date. Details given are date, 

location, age, plumage, sex, condition, and the fate of the specimen, as far this information could be traced. 

Abbreviations used are. Age: Ad = adult, 1st w = first-winter, 1st s = first-summer; Plumage: B = breeding, NB 

= non-breeding; Sex: F = female, M = male. Museums: NMS = National Museums of Scotland; NNM = 

Nationaal Natuurhistorisch Museum, Leiden; ZMA = Zoologisch Museum Amsterdam. References to rarity 

reports and reports of rarities committees are provided in a simplified form, providing the name of the journal, 

page numbers and information on published photographs. Journal abbreviations: IB = Irish Birds, BB = British 

Birds, SB = Scottish Birds, DB = Dutch Birding, DOFT = Dansk Ornitologisk Forenings Tidsskrift, SF = Sveriges 

Fåglar, FoFl = Fauna och Flora, VF = Vår Fågelvärld, Fåg = Fågelåret. 

Ireland (1 record, 1 individual) 

24 Dec. 1986 Kilmore Quay, Ballyteigue Bay, Wexford. NB, alive. IB 3: 478; BB 80: 547; Mullarney 1988. 

United Kingdom (38 records, 39 individuals) 

10 Dec. 1908 Craigielaw Point, Lothian. 1st w, F, dead, specimen at NMS. BB 2: 425. 

15 Apr. 1960 Middleton Sands, near Morecambe, Lancashire & North Merseyside. Ad, dead. 

BB 54: 188, 284–286. 

20 Mar. 1968 Norwick, Unst, Shetland. NB, F, dead, specimen at NMS. BB 62: 473; SB 5: 272 

(plate 19b), 285–287. 

11 Oct. 1969 Knapdale, Loch Caolisport, Argyll. Dead. BB 63: 281; SB 6: 334–335. 

31 Jan. 1976 Reay, Thurso, Caithness. Ad, NB, dead, specimen at NMS. BB 71: 509–510. 

13 Jul. 1977 Farne Islands, Northumberland. Ad, B, alive, at sea near a Common Guillemot colony. 

BB 72: 528; 73: 225–226. 

18 Dec. 1977 Sumburgh, Shetland. Dead. BB 71: 509. 

14 Jul. 1978 St Cyrus, northeast Scotland. 1st s, B, M, dead, specimen at Montrose Museum. BB 72: 528. 

25 Feb. 1979 Rattray Head, northeast Scotland. Dead. BB 73: 514. 

9 Feb. 1980 Kilspindie Beach, Lothian. Ad, NB, dead, specimen at Glasgow Museum. BB 74: 478. 

9 Feb. 1980 Ferry Ness, Lothian. Dead, specimen in L. Simmen private collection. BB 74: 478. 

24 Feb. 1980 Burrafirth, Unst, Shetland. NB, dead. BB 81: 569. 

16–17 Oct. 1980 Fair Isle, Shetland. Ad, B, alive but presumed to have died on 17 Oct.. BB 74: 478 

(plate 279); 81: 569. 

26 Dec. 1980 At Sea, Brent Oilfield, Norwegian Sea, 61°03’N 01°43’E. Alive. BB 75: 511. 

25 Jan. 1981 Johnshaven, Kincardine, northeast Scotland. Dead, specimen at Dundee Museum. BB 75: 511. 

29 Dec. 1981 Bay of Ireland, Stenness, Orkney. Dead. BB 75: 511. 

3 Feb. 1982 Brora, Sutherland, Highland. Dead. BB 76: 502. 

3 Apr. 1982 Stromness, Orkney. Dead. BB 77: 537. 

24 Dec. 1982 Golspie, Highland. 1st w, dead, specimen at NMS. BB 100: 55-56. 

30 Oct. 1983 Banna Minn, West Burra, Shetland. 1st w, F, dead, specimen at NMS. BB 78: 561. 

20 Mar. 1984 Birsay, Orkney. Dead. BB 78: 561. 

9 Jan. 1985 Scapa Bay, Orkney. NB, dead. BB 79: 558. 

3–7 Feb.1987 Off Hamnavoe, West Burra, Shetland. Alive, NB & 7 Feb 1987, Off Hamnavoe, West Burra, 

Shetland. NB, dead, (different individual, live bird still present), specimen in D. Coutts 

private collection. BB 81: 569. 

9 Mar. 1988 Dunnet Bay, Caithness. Ad, M, dead, specimen at NMS. BB 82: 533. 

16 Jun.–12 Jul. 1989 Sumburgh Head, Shetland. Alive, B, site holding in a Common Guillemot colony. BB 83: 469. 

25 Jan. 1991 Sule Skerry, Orkney. Alive. BB 85: 531. 

26 May–8 Jun. 1992 Hirta, St Kilda, Outer Hebrides. Alive. BB 86: 496. 

27 Mar. 1993 Musselburgh, Lothian. Ad, B, alive. BB 87: 536. 

6 Feb. 1994 Seafield, Lothian. Alive. BB 88: 523; 89: 509. 

12 Feb. 1994 Wadbister Voe, Shetland. 1st w, dead, oiled, specimen at NMS. BB 88: 523. 

SEABIRD 21 (2008): 16–31 

29

30 


4 Jan. 1995 Gulberwick, Shetland. Alive, weak, taken into care, released at Wadbister Voe 1 February, 

last seen there 2 February. BB 89: 509. 

23 Jan. 1995 At Sea, north of Fair Isle, 62°41’N 01°34’W. Alive. BB 89: 509. 

27 Mar. 1996 Kilchoan Bay, Ardnamurchan, Argyll. Alive. BB 90: 488. 

25–30 Dec. 1997 Fetlar, Shetland. Alive. BB 91: 496. 

21 Dec. 2000 Scapa Flow, Orkney. Ad, NB, dead, specimen at NMS. BB 94: 480. 

29 Jan. 2001 Orkney North Ronaldsay, Orkney. 1st w, dead, specimen at NMS. BB 95: 501. 

30 Nov.–20 Dec. 2005 Lerwick & Bressay, Shetland. NB, alive. BB 100: 55–56 (photo); 100: 55 (photo). 

4 May 2006 Southladie Voe, Yell, Shetland. Ad, B, dead, specimen at NMS. BB 100: 694–754. 

France (3 records, 3 individuals) 

21 Apr. 1978 near Plougerneau, Finistère. Dead, oiled. Dubois & Yésou 1991. 

21 Jan. 1981 near Santec, Finistère. Dead, oiled. Dubois & Yésou 1991. 

3 Feb. 2003 Audinghen, Pas-de-Calais. Alive. Ornithos 12: 24. 

Belgium (5 records, 5 individuals) 

4 Jan. 1981 Klemskerke/De Haan,West-vlaanderen. 1st w, dead, oiled. Oriolus 52: 73;Van Gompel 1981. 

18 Jan. 1981 Wenduine, West-vlaanderen. Ad, dead, oiled. Oriolus 52: 73. 

7 Dec. 1981 Blankenberge, West-vlaanderen. 1st w, alive, oiled, died same day. Oriolus 52: 73; Van 

Gompel 1982. 

21 Jan. 1995 Bredene, West-vlaanderen. Ad, NB, dead. Aves 34: 195-223; Oriolus 63: 81-82; DB 17: 79, 

86 (photo). 

4 Aug. 2006 Visbeek Wechelderzande, Lille, Antwerp. Ad, B, alive. Goossens 2006; BW 19: 320 (photo). 

The Netherlands (7 records, 7 individuals) 

24 Dec. 1919 Noordwijk aan Zee, Noordwijk, Zuid-Holland. Ad, F, NB, dead, specimen at NNM. 

Ardea 9: 32–33. 

28 Dec. 1924 Noordwijk aan Zee, Noordwijk, Zuid-Holland. M, NB, dead, specimen at NNM. 

19 Feb. 1969 Texel, Noord-Holland. Dead, specimen at NNM. 

10 Mar. 1974 Oostkapelle, Domburg, Zeeland. M, NB, dead, specimen at ZMA. 

4–10 Feb. 1979 Brouwersdam, Goedereede, Zeeland. M, NB, alive, oiled, found dead on 10 February, 

specimen at ZMA. DB 1: 109–111 (photos); 2:20 (photo). 

10 Jan. 1981 Monster, Zuid-Holland. Ad, B, dead, speciman in private collection. DB 3: 99 (photo). 

18 Apr. 1992 Texel, Noord-Holland. Ad, NB, dead, specimen at ZMA. DB 15: 43. 

Germany (3 records, 3 individuals) 

1 Nov. 1959 Voslapp, north of Wilhemshaven, Niedersachsen. NB, dead, specimen at Museum 

Heineanum Halberstadt and Vogelwarte Helgoland, Wilhelmshaven. Goethe & 

Ringleben 1964. 

8 Mar. 1966 Warnemunde, Mecklenburg-Vorpommern. Ad, NB, F, dead, specimen in Meereskundlichen 

Museums Stralsund. Jaeschke & Schulz 1968. 

7 Aug. 1987 Leipzig, Sachsen. Ad, B, alive, weak, died next day, specimen in Naturkunde-museum 

Leipzig. Meyer & Thorwarth 1987 (photo). 

Denmark (10 records, 10 individuals) 

14 May 1886 Storsømmen, Sjælland. Dead. Olsen 1992. 

2 Oct. 1905 Kalveboderne, København, Sjælland. Alive, shot. Olsen 1992. 

8 Nov. 1925 Mesinge ved Kerteminde, Fyn. Dead. Olsen 1992. 

29 Oct. 1974 Blåvand Strand, Ribe. 1st w, dead. DOFT 74: 134–135 (photo); Olsen 1992. 

10 Dec. 1989 Hundested Havn, Sjælland. NB, alive. DOFT 85: 27 (photo); Olsen 1992. 

14 Jan. 1991 At Sea between Hanstholm & Kristianssand, Nordjylland. Alive. DOFT 87: 236. 

17 Jan. 1995 Hirtshals Havn, Nordjylland. Alive. DOFT 91: 142. 

21 Jan. 1998 Læsø Rende, Nordjylland. Ad, B, alive. DOFT 93: 133 

SEABIRD 21 (2008): 16–31


23 Jan. 1998 Strandby Havn, Nordjylland. Ad, alive. From 24 January to 3 February 1998 at 

Frederikshavn Havn, Nordjylland. DOFT 93: 133. 

28 Feb. 1998 Kikhavn, Hundested, Sjælland. Alive, migrating. DOFT 93: 133. 

Sweden (34 records, 38 individuals) 

10 Dec. 1874 Kosterfjorden, Bohuslän. NB, dead, specimen at Strömstads Museum. Risberg 1990. 

5 Feb. 1875 Strömstad, Bohuslän. NB, alive. FoFl 39: 95. 

1 Dec. 1902 Parish of Pajala, Norrbotten. NB, 2 individuals, alive, shot. FoFl 2: 216–217. 

Jan. 1925 Grebbestad, Bohuslän. NB, 2 individuals, alive. SF 4: 64. 

May 1955 Karesuando, Lappland. Alive. SF 1978: 130. 

17 Dec. 1955 Kälvudden, Överkalix, Norrbotten. NB, alive, ringed and released. VF 15: 280. 

Feb. 1962 Killingholmen, Boden, Norrbotten. NB, alive. VF 32: 307. 

9 Feb. 1964 Båstad, Skåne. NB, dead. Anser 5: 9. 

6 Jan. 1981 Fjällbacka, Bohuslän. NB, alive, shot. VF 46: 327. 

19 Apr. 1982 Lilla Karlsö, Gotland. B, alive. VF 42: 401. 

5 Nov. 1983 Busör, Halland. NB, alive. VF 43: 544; Breife et al. 1990 (photo). 

11 Jan. 1987 Klöverträsk - Älvsbyn, Norrbotten. NB, dead, found on a road. VF 47: 456. 

21 Jul. 1987 Sotenäs, Bohuslän. Ad, B, alive. VF 47: 456. 

28 Nov.–6 Dec. 1987 Varbergs hamn, Halland. 1st w, alive, eventually “collected”. VF 47: 456. 

3 Jan. 1991 St. Amundön, SW Askim, Västergötland. Ad, NB, alive. VF 51 7: 24. 

21 Jan. 1993 Sote huvud, Bohuslän. NB, alive, migrating. Fåg 1993: 112. 

14 Nov. 1993 Sebybadet, Öland, NB, alive, migrating. Fåg 1993: 112. 

18–19 Oct. 1995 Hovs hallar,Skåne, 2 individuals, one also seen at Kullen, Skåne, NB, alive, migrating. 

Fåg 1995: 134. 

8 Dec. 1995 Segerstads fyr, Öland. NB, alive. Fåg 1995: 134. 

16–17 Nov. 1996 Soten, Sotenäs, Bohuslän. Two individuals, NB, alive. Fåg 1996: 140. 

17 Oct. 1997 Busör, Halland. NB, alive. Fåg 1997: 160. 

28–29 Oct. 1997 Välen - Askimsviken, Västergötland. 1st w, F, alive, found dead on 30 October, specimen 

at Göteborg Natural History Museum. Fåg 1997: 160. 

30 Dec. 1997 Kullen, Skåne. NB, alive. Fåg 1997:160. 

28 Feb. 1998 Hovs hallar, Skåne, NB, alive, migrating. Fåg 1998: 145. 

20–21 May 2000 Lilla Karlsö, Gotland. B, alive, migrating. Fåg 2001: 136. 

10 Jun. 2001 Revsudden, Kalmarsund, Småland. B, alive. Fåg 2001:1 36. 

30 Oct. 2001 Tussebo, Skepplanda, Västergötland. Ad, M, NB, alive, died on 1 November, specimen at 

Göteborg Natural History Museum. Fåg 2001: 136. 

3 Nov. 2001 Busör, Halland. NB, alive, migrating. Fåg 2001: 136. 

7 Feb. 2002 Valnäsudden, Nordkoster, Bohuslän. NB, alive. Fåg 2002: 139. 

7 Apr. 2002 Hermanö huvud, Bohuslän. B, alive, migrating. Fåg 2002: 139. 

17 Jan. 2003 Bua udde, Väröbacka, Halland. Alive. Fåg 2003: 184. 

15 Nov. 2003 Busör, Halland. 1st w, alive. Fåg 2003: 184. 

3 Dec. 2003 Segelskär, Tanum, Bohuslän. NB, alive. Fåg 2003: 184 

5–19 Nov. 2004 Guleskärskajen, Kungshamn, Bohuslän. Ad, NB, alive. Fåg 2004: 145. 

Poland (1 record, 1 individual) 

10 Apr. 1964 Gdynia, Pomorskie. Ad, alive, trapped, ringed at Novaya Zemlya in August 1948 (number 

D-117833 Moskva). Tomialojć, 1990; Rønnest 1994. 

Finland (2 records, 2 individuals) 

14 Dec. 1986 Inari, Lappland. Alive, brought to Korkeasaari Zoo, Helsinki, where it died on 26 January 

1987. Lintumies 23: 199. 

22 Oct. 1988 Nauvo Sandö, Varsinais-Suomi. 1st w, dead, drowned in fishnet. Lintumies 23: 278. 

SEABIRD 21 (2008): 16–31 

31

32 

Leach’s and European Storm-petrels 

A survey of Leach’s Oceanodroma leucorhoa 

and European Storm-petrel Hydrobates 

pelagicus populations on North Rona and 

Sula Sgeir, Western Isles, Scotland 

Murray, S. 1 *, Money, S. 2 , Griffin, A. 3 & Mitchell, P. I. 4 

*Correspondence author. Email: murraysurvey@yahoo.co.uk 

1 Craigie Dhu, Cardney, Dunkeld, Perthshire PH8 0EY, UK; 2 Raintree House, Church Lane, 

Drayton St Leonard, Oxfordshire OX10 7AU, UK; 3 15 Horologie Hill, Arbroath, Angus 

DD11 5AE, UK; 4 Joint Nature Conservation Committee, Dunnet House, 7 Thistle Place, 

Aberdeen AB10 1UZ, UK. 

Abstract 

Leach’s Storm-petrel Oceanodroma leucorhoa was first recorded breeding on North 

Rona in 1883 and on Sula Sgeir in 1939. European Storm-petrel Hydrobates pelagicus 

was first recorded on North Rona in 1885 and on Sula Sgeir in 1958. Since then, there 

have been attempts to estimate the population size of both species on North Rona but 

there is little information about their current status on Sula Sgeir. In 2001, systematic 

surveys of both species using tape playback were conducted for the first time on both 

islands. North Rona held 1,133 Apparently Occupied Sites (AOS) of Leach’s Stormpetrel 

but only 371 AOS of European Storm-petrel; numbers on Sula Sgeir were five 

and eight AOS respectively. The combined population of both North Rona and Sula 

Sgeir of Leach’s Storm-petrel and European Storm-petrel, comprise 2.3% and 1.4% 

respectively, of the total number of each species breeding in Great Britain. 

Introduction 

North Rona (area = 128 ha, highest point = 108 m), uninhabited since 1844, lies about 

70 km north of the Butt of Lewis in the Western Isles at 59°08’N 5°50’W. Leach’s Stormpetrel 

Oceanodroma leucorhoa and European Storm-petrel Hydrobates pelagicus were 

first discovered there in the 1880s (Swinburne 1885; Harvie-Brown 1888), but there has 

never been an accurate census of either species. Both are difficult to survey, since they 

nest in burrows or in rock crevices and are nocturnal. However, the recently developed 

tape playback technique (Ratcliffe et al. 1998) enabled this study to obtain an accurate 

count of Apparently Occupied Sites (AOS) of both species for the first time. 

Sula Sgeir (20 ha, 70 m) lies about 17 km west of North Rona, at 59°06’N 6°09’W. It 

is no more than 1 km in length and just 200 m across at its widest point. It is sparsely 

vegetated due to lack of soil, and is subject to heavy erosive pressure from breeding 

seabirds and sea-spray. It has rarely been visited by ornithologists and ‘stormy petrels’, 

species uncertain, were first confirmed present in 1930 (Dougal 1937). North Rona 

holds extensive evidence of past human occupation, most prominently a ruined 

village, graveyard and chapel enclosed by pronounced cultivation ridges. Sula Sgeir, 

SEABIRD 21 (2008): 32–43


although never permanently occupied has been visited annually for centuries by the 

men of Ness in Lewis, who have built dry stone bothies on the rock. The structures on 

both islands are an important component of the available breeding habitat for both 

petrel species. North Rona and Sula Sgeir are designated as a Special Protection Area 

(SPA) under Article 4.1 of the EC Birds Directive (79/409/EEC) for supporting more 

than 1% of the Great Britain breeding populations of Leach’s and European Stormpetrel, 

which are both listed in Annex 1 of the Directive (Stroud et al. 2001). The 

islands also qualify for SPA designation under Article 4.2 of the Directive by supporting 

more than 1% of the relevant bio-geographic breeding populations of Northern 

Gannet Morus bassanus and Common Guillemot Uria aalge (Stroud et al. 2001). 

Leach’s and European Storm-petrel were cited as qualifying species for the North Rona 

and Sula Sgeir SPA based on estimates of colony size and Great Britain population size 

given in Lloyd et al. (1991). These estimates were derived before tape playback was 

developed and most were expressed as orders of magnitude of the number of birds 

present at colonies during the night. The aim of the present study was to use tape 

playback to accurately census both species of storm-petrel and, combined with the 

latest Great Britain population estimates (also derived using tape playback) (Mitchell 

2004; Mitchell & Newton 2004), assess their conservation status and validate the 

islands’ SPA designation with respect to these species. 

Methods 

Colony census methods: The playback method entails playing recordings of the chatter 

call of a male Leach’s Storm-petrel and the purr call of European Storm-petrel in 

suitable habitat during the incubation period, in order to elicit a reply from an 

incubating adult within a burrow. A hand-held dictaphone with integral speakers was 

used.All accessible areas on both islands, including the highest sea cliffs on North Rona, 

were systematically surveyed using the tape playback technique (Gilbert et al. 1998) 

and responses were counted and mapped. The main drawback with the tape playback 

method is that not all individuals will respond to the taped calls (Ratcliffe et al. 1998), 

so a count of responses will underestimate the total number of AOS at a colony. 

Furthermore, Leach’s Storm-petrel will only respond to taped chatter calls of the same 

sex (Taoka et al. 1989), therefore it is necessary to measure what proportion of birds 

present in burrows are responding to the taped calls. This was achieved by setting up a 

calibration plot for each species, which entails repeatedly visiting a delimited section of 

the colony on successive days and each time marking new responding AOS. Calibration 

plots were set up on the North Rona storm beach for European Storm-petrel and along 

the village graveyard wall for Leach’s Storm-petrel. On the first visit to each plot, 20 

AOS were located and their positions marked with flagged canes. Both plots were 

visited on a total of six days between 26 June and 1 July 2001.The total number of AOS 

on each island was then estimated by multiplying counts of responding birds by the 

response rate derived from the calibration plots (Equation 1). 

Equation 1: Number of AOS = no. responses x (1 / response rate) 

where response rate is estimated as shown in Equation 2. 

SEABIRD 21 (2008): 32–43 

33

34 


North Rona: The island was subdivided into 16 sections (A to P, Figure 1), using clearly 

defined natural or man-made boundaries, and each was searched systematically for 

both species. To ensure full coverage of each section only one species was searched for 

at a time by three surveyors. This ensured that each section received the same level of 

effort. In some areas, mainly Fianuis (section L) and the storm beach (section M), ropes 

and canes were used to subdivide the ground into strips several metres wide in order 

to aid coverage. Island coverage was close to 100% and the only site possibly holding 

breeding petrels that was not surveyed was the lower, inaccessible half of Geo Mairi, 

between sections O and P. 

Sula Sgeir 

Figure 1. North Rona showing survey sections A to P, and place names given in the text. Sula Sgeir, showing the 

survey area and location of the bothies. 

SEABIRD 21 (2008): 32–43 

North Rona

Figure 2. Aerial photograph of North Rona © J. A. Love. 


The Toa Rona cliffs (section P) were difficult to access, held large numbers of Northern 

Fulmars Fulmarus glacialis and Atlantic Puffins Fratercula arctica and vegetation was 

unstable over shallow soil. Therefore, to lessen risks to surveyors and to minimise 

disturbance to the colonies, the area was surveyed by a single surveyor only. Overall, 

the slope was delimited by steep rock walls, narrow gullies and sharply defined 

vegetation boundaries, which simplified surveying. 

European Storm-petrels were surveyed between 26 June and 2 July and Leach’s Stormpetrels 

between 2 and 8 July, (except for the village ruins (section E), which were 

surveyed on 29 June 2001). 

Sula Sgeir: Sula Sgeir was surveyed on 24 June 2001. To minimise disturbance to the 

dense assemblage of breeding seabirds on the flat top of the rock, only the five bothies, 

set in a small area of eroded soil edged by boulders, were surveyed (Figure 1). 

Estimating response rate: There are two methods for estimating the response rate of 

a storm-petrel population from calibration plot data: 

1. Simple arithmetic (Equation 2) 

Equation 2: response rate = total number of responses / (total AOS x number of visits). 

SEABIRD 21 (2008): 32–43 

35

36 


Confidence limits were calculated using a Generalised Linear Model, with the number 

of responses from each AOS as the dependent variable and the number of visits made 

to each AOS as the independent variable. However, this method is sensitive to birds 

becoming habituated to the taped calls, so that progressively fewer previously located 

AOS responded on successive days. Using this method in a habituated plot will tend 

to underestimate the response rate of a population not yet exposed to the taped calls. 

Furthermore, if not all AOS in the plot had been found by the time of last visit, 

response rate will be overestimated. 

2. Iterative regression. The advantage of this method over the simple arithmetic 

method is that not all AOS in a plot need to be found; this is useful if a surveyor can 

only make a limited number of visits to the plot, which is often the case on remote 

islands that are difficult and costly to visit. For each calibration plot, an asymptotic 

regression model was fitted to the cumulative number of AOS that had been found 

following each visit (see Mayhew et al. 2000). The model took the form of Equation 3 

and the parameters a and b were predicted using the iterative regression function of 

S-Plus® 2000 (Mathsoft Inc., Seattle, Washington). 

Equation 3: where y = number of AOS detected on a given visit (x); b = the 

exponential (e) proportional rate of increase to the asymptote (a). 

Thus, the coefficient a is an estimate of the total number of AOS present in the study 

plot. The response rate was calculated by substituting the values of the coefficients for 

b into Equation 4. 

Equation 4: response rate = 

The upper and lower 95% confidence limits of the response rate were determined 

from the Equations 5 and 6 respectively. 

Equation 5: 95% UCL response rate = 1 - e-b-(se x 1.96) 

Equation 6: 95% LCL response rate = 1 - e-b+(se x 1.96) where se is the standard error 

of the estimate of the coefficient b. 

Results 

North Rona: European Storm-petrel response rate; The results from the European 

Storm-petrel calibration plot are shown in Table 1 and the cumulative number of AOS 

found following on each successive visit are plotted in Figure 3. Table 1 shows a clear 

reduction in total responses following the first visit, with only six birds responding on 

day six. This would suggest that some degree of habituation to the taped calls was 

taking place. Hence, the use of the simple arithmetic technique to calculate response 

rate was inadvisable in this case. Instead an iterative regression was applied to the plot 

Table 1. European Storm-petrel Hydrobates pelagicus calibration plot results. 

Visit no. 1 2 3 4 5 6 

No. of new burrows found per visit 20 9 7 3 4 2 

Cumulative total of burrows 20 29 36 39 43 45 

No. of responses per visit 20 13 11 13 16 6 

SEABIRD 21 (2008): 32–43

Cumulative no. of AOS 


Table 2. Parameters of the iterative regression of the cumulative number of European Storm-petrel 

Hydrobates pelagicus Apparently Occupied Sites (AOS) on successive visits to the calibration plot. The 

resultant correction factor was multiplied by the total counts of responses to estimate the total number of 

AOS. Figures in parentheses are 95% CLs. 

60 

55 

50 

45 

40 

35 

30 

25 

20 

a b Response rate Correction factor 

(1/response rate) 

46.3 0.509 0.400 2.5 

(44.1–48.6) (0.035 s.e.) (0.356–0.439) (2.28–2.81) 

t 5 = 40.329 t 5 = 14.385 

European Storm-petrel 

Leach's Storm-petrel 

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 

Days after first visit 

Figure 3. Change in cumulative number of new 

AOS found using diurnal playback on successive 

visits to the calibration plots for European Stormpetrel 

Hydrobates pelagicus and Leach’s Stormpetrel 

Oceanodroma leucorhoa on North Rona. 

The curve was fitted to the European Storm-petrel 

data using iterative regression, dotted lines 

indicate 95% confidence limits. where a = 46.3 

(44.1–48.6 s.e.) and b = 0.509 (0.035 s.e.) 

and the results are shown in Table 2. The extrapolation of the response curve (Figure 

3) estimated that 46 AOS were present in the plot, just one more than the number 

actually found after six visits. The response rate predicted from the slope of the curve 

was 0.40, giving a correction factor of 2.5 (i.e. 1 / response rate, see Equation 1). 

European Storm-petrel population estimates; One hundred and forty-seven birds 

responded to the taped calls at 18 sites, in 11 out of 16 colony sections (Tables 3 & 4, 

Figure 1), representing an estimated 371 AOS (95% CL 335– 413). The largest concentration 

of 203 AOS was found in the storm beach (section M). The remainder were 

thinly scattered across the island, with man-made structures, e.g. the chapel ruins, 

Fianuis bothies, walls, cairns and enclosures forming an important component of the 

breeding habitat, accounting for 25% of all AOS (Table 3). None were found in the 

large, apparently suitable area of rock scree on the south side of Geodha Lèis (section 

K ), but there were five AOS bordering the area. On the Toa Rona cliffs (section P), eight 

AOS were found along the cliff top, but none elsewhere. Both areas were densely 

occupied by breeding Atlantic Puffins and Geodha Lèis by Razorbills Alca torda also. 

Leach’s Storm-petrel response rates; There was no evidence of habituation by Leach’s 

Storm-petrel to the taped calls, and the number of responses from the plot actually 

increased over the last three visits (Table 5). However, the iterative regression was not 

used to estimate response rate because, with hindsight, too few visits were made to 

SEABIRD 21 (2008): 32–43 

37

38 


Table 3. Habitat selection and estimated AOS of European Storm-petrels Hydrobates pelagicus on North Rona. 

Habitat type No. of sites No. of responses % of total responses AOS (95% CL) 

Storm beach 1 81 55 203 (185–228) 

Other natural sites 8 29 20 74 (66–81) 

Man-made structures 9 37 25 94 (84–104) 

Total 18 147 100 371 (335–413) 

Table 4. Number of AOS of Leach’s Storm-petrel Oceanodroma leucorhoa and European Storm-petrel 

Hydrobates pelagicus on North Rona. See Figure 1 for map of sections A to P. 

Leach’s Storm European Leach’s Storm European 

Section -petrel Storm-petrel Section -petrel Storm-petrel 

A 0 3 I 85 10 

B 65 23 J 97 8 

C 0 0 K 159 33 

D 12 0 L 185 30 

E 328 20 M 28 203 

F 5 0 N 51 0 

G 23 0 O 21 30 

H 30 3 P 44 8 

Total AOS 1,133 371 

Table 5. Leach’s Storm-petrel Oceanodroma leucorhoa calibration plot results. 

Visit no. 1 2 3 4 5 6 Total 

No. of new burrows found per visit 20 9 8 13 6 2 58 

Cumulative total of burrows 20 29 37 50 56 58 58 

No. of responses per visit 20 16 19 34 35 27 151 

Table 6. Habitat selection and estimated AOS of Leach’s Storm-petrels Oceanodroma leucorhoa on North Rona. 

Habitat type No. of responses % of total responses AOS (95% CL) 

Storm beach 12 2% 28 (26–29) 

Other natural sites 306 62% 708 (673–745) 

Man made structures 172 35% 397 (374–422) 

Total 490 100% 1,133 (1,065–1,202) 

the plot and consequently, produced too limited a range of x-values to enable 

Equation 3 to be meaningfully applied (see Figure 3). The simple arithmetic technique 

(Equation 2) was applied instead. The mean response rate across the six visits was 

0.434, with 95% CL of 0.408–0.460 (GLM, t 56 = 17.82, P < 0.001). The correction 

factor for Leach’s Storm-petrel was therefore, 1/0.434 = 2.31 (95% CL 2.173–2.453). 

SEABIRD 21 (2008): 32–43


Leach’s Storm-petrel population estimates; Four hundred and ninety birds responded to 

the taped calls, representing 1,133 AOS (95% CL 1065–1202) ( Table 4 & 6, Figure 1). 

The largest and densest subcolony on the island was found in the village ruins (section 

E), which held 328 AOS, 29% of the island total. A further 69 AOS (6%) were found in 

other man-made structures, including the low turf dyke demarcating section J on Toa 

Rona, which held 97 AOS in the second densest subcolony on the island. 

Overall, 708 AOS (62%) were found in natural sites, widely distributed across the 

island, with the exception of sections A and C where none were found. Nests were 

situated in stone piles, under embedded boulders, along cliff edges and in well drained 

turf, but not in the old cultivation ridges of section F, or in the deep, waterlogged soils 

of sections D and H. None were found in the largest Atlantic Puffin colony on Toa Rona 

(section P), although they were present in small numbers along the cliff tops. Similarly, 

at other, smaller puffin subcolonies, e.g. Geodha Lèis and Geodha Blatha Mor (both in 

section K), Leach’s Storm-petrel nests were found in abandoned puffin burrows on the 

edge of the colonies, but none were found in areas of densely occupied puffin burrows. 

Sula Sgeir: Leach’s and European Storm-petrels were found only in the walls of the 

bothies, but not in any of the cairns or other man-made structures that were surveyed. 

Around the bothies the ground was badly eroded and occupied by a high density of 

nesting Northern Fulmars. Potentially suitable storm-petrel habitat in boulders around 

the margins of the survey area (Figure 1) held nesting Atlantic Puffins and Common 

Guillemots, and probably as a result, did not appear to be occupied by either stormpetrel 

species. In total, only two responses from Leach’s Storm-petrel and three from 

European Storm-petrel were elicited during the survey of Sula Sgeir. Using the 

response rates estimated on North Rona (Tables 2 & 5), these responses equated to 

five AOS and eight AOS respectively. 

Discussion 

European Storm-petrel on North Rona: The estimate of 371 AOS of European Stormpetrel 

is the lowest made of the North Rona population to date. However, the distribution 

of AOS found during the current study is broadly similar to earlier accounts. For 

example, the storm beach (section M) was considered by Bagenal & Baird (1959) to 

hold the largest concentration of burrows on the island and Love (1978) considered 

the storm beach to hold many more pairs than the village. In 2001, the storm beach 

also held the largest subcolony, with 55% of the total number of AOS on the island. 

By contrast, low numbers have consistently been noted in the village ruins (section E), 

with nests found only in the chapel and in a nearby boulder pile. In 2001, 20 AOS were 

found there, which was identical to the number of pairs estimated to be breeding there 

in 1936 by Ainslie & Atkinson (1937a). Despite conducting additional searches of the 

village during the night in 2001, none were found breeding in any structure other than 

the chapel. Likewise, Robson (1968) found a few nests in the chapel, but none 

elsewhere, and Love (1978) noted that 72% of his mist net captures of European 

Storm-petrels in the village were made close to the chapel. Elsewhere on the island, 

AOS were thinly distributed and in low numbers. 

SEABIRD 21 (2008): 32–43 

39

40 


The mean response rate of European Storm-petrels on North Rona was very similar to 

other colonies in northwest Scotland (Mitchell & Newton 2004). The calibration and 

simultaneous census of North Rona were conducted early in the incubation period 

expected of European Storm-petrels at Scottish colonies (i.e. late June and most of July), 

so response rate may have been expected to increase if not all of the birds had 

commenced incubation by the start of the study (Ratcliffe et al. 1998). However, daily 

response rate in the calibration plot at least stayed fairly constant apart from on the day 

of the last visit when it was considerably lower. Furthermore, detailed daily searches of 

AOS in the village did not find any more than were there on the first day of searching. 

Therefore, the mean response rate estimated in the plot appeared to have accurately 

reflected the response rate across the rest of the colony when the census was conducted. 

Leach’s Storm-petrel on North Rona: On North Rona it is difficult to make any 

assessment of change prior to this study, since no systematic survey of Leach’s Stormpetrel 

numbers, using comparable methods, has been conducted previously. Our count 

of 328 AOS in the village ruins is almost identical to the 327 occupied burrows found 

there in 1936 by Ainslie & Atkinson (1937b).They estimated only 50 burrows for the rest 

of the island, giving a total population of 380 pairs. In 1958, a study using ringing and 

recapture, estimated the village population to be around ten times larger, i.e. 2–3,000 

pairs, with a total island population of about 5,000 pairs (Bagenal & Baird 1959). 

However, estimates of population size derived from mist-netting and mark-recapture 

include non-breeding birds as well as breeders, so will be greater than concurrent 

estimates derived from counts of occupied nest sites. Bagenal & Baird’s (1959) estimate 

was also far greater than subsequent assessments of colony size, with both Robson 

(1968) and Love (1978) considering the village population to be little different in size 

compared to Ainslie & Atkinson’s (1937b) estimate in 1936. Outside the village, calling 

birds have been heard at many different sites, but only in small numbers. P. G. H. Evans 

in 1972, cited in Lloyd et al. (1991) estimated the entire population, including that of the 

village, at 500 pairs. Our estimate in 2001 of 1,133 AOS of Leach’s Storm-petrel on 

North Rona makes this the third largest colony, after St Kilda and the Flannan Isles, of 

only eight known colonies in Britain and Ireland (Mitchell 2004). 

The response rate measured in the calibration plot on North Rona was probably a 

slight overestimate because not all the AOS in the plot were found, though visual 

inspection of the calibration plot data in Figure 3 would suggest that only 2–3 AOS 

were missed. Therefore, the 1,133 AOS is probably only a slight underestimate of the 

size of the colony on North Rona. A further inaccuracy may have resulted if the birds 

had responded differently during the census on 2–8 July, than they had done a week 

earlier when the response calibration was conducted (Ellis et al. 1998). The greatest 

determinant of response rate is the level of diurnal occupancy of burrows by the 

adults. At least one adult should be present in the burrow during the day throughout 

incubation and up to five days post-hatching. There are no published data on the 

phenology of Leach’s Storm-petrels on North Rona, but recent observations on St 

Kilda have shown hatching to start during the second and third week in July (Money 

et al. 2008). It is reasonable to assume that breeding phenology is similar in both 

colonies, since they most probably use the same feeding grounds (Mitchell 2004). 

SEABIRD 21 (2008): 32–43


Therefore it is unlikely that the Leach’s Storm-petrel on North Rona were responding 

to the taped calls differently during the calibration and the census. Furthermore, the 

response rate estimated in the current study was very similar to that recorded during 

4–10 July 2003 on St Kilda, using the same method (Newson et al. 2008). 

European Storm-petrel on Sula Sgeir: European Storm-petrels have been rarely 

recorded breeding on Sula Sgeir. Although they may have been present in 1930 

(Dougal 1937), they were not noted by either Stewart (1932), Atkinson & Ainslie 

(1940) or McGeoch (1954a,b). Bagenal & Baird (1959) caught one and heard others 

calling in June 1958 and this appears to be the first confirmed record. None were heard 

in 1980 but they were noted in 1986 (Benn et al. 1989) and the eight AOS in 2001 

were all found within bothy walls. 

Leach’s Storm-petrel on Sula Sgeir: On Sula Sgeir Leach’s Storm-petrels were 

probably the ‘stormy petrels’ observed in 1930 by Dougal (1937) and in 1932 by 

Stewart (1934), as being present in bothy walls. Breeding was proven in 1939 by 

Atkinson & Ainslie (1940) who found both young and incubating adults. They carried 

out an overnight survey of the rock, starting from the bothies and surveying 

southwards to the summit cairn. They found birds calling from underground near the 

bothies and along the length of the summit ridge (Figure 3).They considered that they 

were at least as numerous as the population on Rona and estimated that 400 pairs 

were breeding there. The subsequent loss of soil and vegetation around the bothies, 

caused largely by the increase in nesting Northern Fulmars, and on the plateau by the 

expansion of the gannetry, probably reduced the available breeding habitat. As a 

Figure 4. Leach's Storm-petrel Oceanodroma leucorhoa, Merseyside, 23 September 2004 © Sue Tranter. 

SEABIRD 21 (2008): 32–43 

41

42 


consequence, in 1954, McGeoch (1954a,b) found only a dozen birds, although he 

searched for several nights up to the edge of the gannetry. Bagenal & Baird (1959) 

caught 18 in mist nets set overnight between the bothies in June 1958, and a few were 

heard in bothy walls, but nowhere else, during an overnight stay in June 1980 (S. 

Murray pers. obs.). Since then there have been further increases in both Northern 

Gannet and Northern Fulmar numbers (Mitchell et al. 2004) and in 2001, only five 

AOS of Leach’s Storm-petrel were found, all in the walls of bothies. 

Conclusions 

Although the European Storm-petrel appears never to have bred in large numbers on 

Sula Sgeir, numbers of the Leach’s Storm-petrel have declined substantially since 1939, 

probably caused by habitat loss due to soil erosion and through competition for space 

from the larger and more aggressive Northern Gannets and Northern Fulmars. In 2001, 

three fresh and unmarked European Storm-petrel corpses were found lining a 

Northern Fulmar nest.We speculated that all had been killed by the incubating fulmar, 

as they made their way in or out of the adjacent bothy wall. Whatever the case, it 

illustrates the vulnerability of small petrels at this site. 

It is unclear from previous records whether or not numbers of Leach’s and European 

Storm-petrel breeding on North Rona have changed significantly during the last century. 

However, future surveys of either species on the island should use the results of the 2001 

survey (shown in Tables 3, 4, & 6; Figure 1) as a baseline for assessing changes in breeding 

numbers and distribution. The precision of the 2001 estimates of colony size (see 95% 

CL in Tables 3 & 6) should enable future surveys to detect changes in breeding numbers 

of both species. Changes will have to be greater than 9% for European Storm-petrel and 

6% for Leach’s Storm-petrel in order to be statistically significant. 

The combined population of both North Rona and Sula Sgeir of Leach’s Storm-petrel 

and European Storm-petrel, comprise 2.3% and 1.4% respectively, of the total number 

of each species breeding in Great Britain (Mitchell 2004). Therefore, North Rona and 

Sula Sgeir qualify as an SPA under the EC Birds Directive for their importance as a 

breeding site for both Annex 1 listed species of storm-petrels. 


The work outlined in this report was funded by the European Research and 

Development fund under the Atlantic Area Programme of the Interreg II C Initiative, as 

part of Project No. 414, awarded to the Joint Nature Conservation Committee in 

partnership with BirdWatch Ireland. We are grateful to Norman Ratcliffe and Adrian 

Plant for their comments on an earlier draft; to Paddy Pomeroy for the use of the Sea 

Mammal Research Unit hut and equipment; and especially to our boat skippers, Murdo 

MacDonald and Andy Tibbits. 

References 

Ainslie, J.A. & Atkinson, R. 1937a. Summer bird notes from North Rona. Scottish Naturalist 49: 7–13. 

Ainslie, J. A. & Atkinson, R. 1937b. On the breeding habits of Leach’s Fork-tailed Petrel. British 

Birds 30: 234–248. 

SEABIRD 21 (2008): 32–43


Atkinson, R. & Ainslie, J.A. 1940. The British breeding status of Leach’s Fork-tailed Petrel. British 

Birds 34: 50–55. 

Bagenal, T. B. & Baird, D. E. 1959. The birds of North Rona in 1958, with notes on Sula Sgeir. 

Bird Study 6: 153–174. 

Benn, S., Murray, S. & Tasker, M. L. 1989. The birds of North Rona and Sula Sgeir. NCC, 

Peterborough, UK. 

Dougal, J. W. 1937. Island Memories. Edinburgh, Moray Press, Edinburgh. 

Ellis, P., Ratcliffe, N. & Suddaby, D. 1998. Seasonal variation in diurnal attendance and 

response to playback by Leach’s Petrel Oceanodroma leucorhoa, on Gruney, Shetland. Ibis 140: 

336–339. 

Gilbert, G., Gibbons, D. W. & Evans, J. 1999. Bird Monitoring Methods, a Manual of Techniques 

for Key UK Species. Royal Society for the Protection of Birds, UK. 

Harvie-Brown, J. A . & Buckley, T. E. 1888. A Vertebrate Fauna of the Outer Hebrides. David 

Douglas, Edinburgh. 

Lloyd,C.,Tasker,M.L.& Partridge,K.1991. The Status of Seabirds in Britain and Ireland. Poyser, London. 

Love,J.A.1978.Leach’s and Storm-petrels on North Rona 1971–74. Ringing & Migration 2: 15–19. 

Mayhew, P., Chisholm, K., Insley, H. & Ratcliffe, N. 2000. A survey of Storm Petrels on Priest 

Island. Scottish Birds 21: 78–84. 

McGeoch, J. 1954a. August migrants at Sula Sgeir. Fair Isle Bird Observatory Bulletin 25: 201–208. 

McGeoch, J. 1954b. ‘Birds of Sula Sgeir, August 1954.’ Unpubl. report to SNH Western Isles. 

Mitchell, P. I. 2004. Leach’s Storm-petrel Oceanodroma leucorhoa. In: Mitchell, P. I., Newton, 

S. F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. Seabird Populations of Britain and Ireland: 101–114. 

Poyser, London. 

Mitchell, P. I., Newton, S. F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. Seabird Populations of 

Britain and Ireland. Poyser, London. 

Mitchell, P. I. & Newton, S. F. 2004. European Storm-petrel Hydrobates pelagicus. In: Mitchell, 

P. I., Newton, S. F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. Seabird Populations of Britain and 

Ireland: 81–100. Poyser, London. 

Money, S., Söhle, I. & Parsons, M. 2008. A pilot study of phenology and breeding success of 

Leach’s Storm-petrel Oceanodroma leucorhoa on St Kilda, Western Isles. Seabird 21: 98–101. 

Newson, S. E., Mitchell P. I., Parsons, M., O’Brien, S. H., Austin, G. E., Benn, S., Black, J., 

Blackburn, J., Brodie, B., Humphreys, E., Leech, D. I., Prior, M. & Webster, M. 2008. 

Population decline of Leach’s Storm-petrel Oceanodroma leucorhoa within the largest colony 

in Britain and Ireland. Seabird 21: 77–84. 

Ratcliffe, N., Vaughan, D., Whyte, C. & Shepherd, M. 1998. Development of playback census 

methods for Storm-petrels Hydrobates pelagicus. Bird Study 45: 302–312. 

Robson, M. J. H. 1968. The breeding birds of North Rona. Scottish Birds 5: 126–156. 

Stewart, M. 1934. The status of petrels in certain remote Scottish islands. Scottish Naturalist 

46: 95–98. 

Stroud, D. A., Chambers, D., Cook, S., Buxton, N., Fraser, B., Clement, P., Lewis, P., McLean, I., 

Baker, H. & Whitehead, S. 2001. The UK SPA network: its scope and content. JNCC, 

Peterborough, U.K. 

Swinburne, J. 1885. Notes on the islands of Sula Sgeir or North Barra and North Rona, with a 

list of birds inhabiting them. Proceedings of the Royal Physical Society of Edinburgh 8: 51–67. 

Taoka, M., Sato, T., Kamada, T. & Okumura, H. 1989. Sexual dimorphism of chatter calls and 

vocal sex recognition in Leach’s Storm-petrels. Auk 106: 498–501. 

SEABIRD 21 (2008): 32–43 

43

44 

Seabird diet on Canna 

The diet of European Shag Phalacrocorax 

aristotelis, Black-legged Kittiwake Rissa 

tridactyla and Common Guillemot Uria 

aalge on Canna during the chick-rearing 

period 1981–2007 

Swann, R. L. 1 *, Harris, M. P. 2 & Aiton, D. G. 3 

*Correspondence author. Email: robert.swann@homecall.co.uk 

1 14 St Vincent Road, Tain, Ross-shire IV19 1JR, UK; 2 Centre for Ecology and Hydrology, 

Hill of Brathens, Banchory, Aberdeenshire AB31 4BY, UK (current address: CEH Bush Estate, 

Penicuik, Midlothian EH26 0QB, UK); 3 14 Buckstone Howe, Edinburgh EH10 6XF, UK. 

Abstract 

Chick diet of European Shags Phalacrocorax aristotelis, Black-legged Kittiwakes Rissa 

tridactyla and Common Guillemots Uria aalge at Canna was investigated over a 27year 

period. The diet was mainly composed of Sprats Sprattus sprattus, Lesser Sandeels 

Ammodytes marinus and members of the Gadidae (a variety of species but mainly 

Trisopterus spp. and Whiting Merlangius merlangus). Other groups (ten families of fish, 

crustaceans, cephalopod molluscs and polychaete worms) were of minimal 

importance. Lesser Sandeels dominated the diet of young Black-legged Kittiwakes and 

European Shags, Sprats the diet of young Common Guillemots, whereas gadid otoliths 

were by far the commonest items found in pellets regurgitated by older European 

Shags. There were few significant temporal changes in species composition or the size 

of prey taken over the 27 years and the results confirm earlier findings that gadids are 

a normal and important part of the diet of seabirds at this colony. 

Introduction 

Over the last 40 years seabird populations in Britain have undergone major changes. 

Up to the late 1980s numbers were tending to increase, but since then major declines 

have been observed (Mavor et al. 2006). For some species, e.g. Black-legged Kittiwake 

Rissa tridactyla, changes in numbers are thought to be linked to changes in food supply, 

particularly sandeels (Ammodytidae, mainly Lesser Sandeels Ammodytes marinus) 

(Frederiksen et al. 2004). Several studies on seabird diet undertaken at colonies in the 

North Sea and Shetland have shown the importance of sandeels in the diets of 

seabirds in these areas (Pearson 1968; Monaghan 1992; Daunt et al. 2008), but 

relatively little information is available on the food of seabirds at colonies in the west 

of Britain. Swann et al. (1991) summarised information on the food of seabirds on 

Canna during the chick-rearing periods 1981–90, and showed that although sandeels 

were the main species taken by European Shag Phalacrocorax aristotelis, Black-legged 

Kittiwake and Common Guillemot Uria aalge, in contrast to some other places gadids 

(Gadidae, cod-fishes) and clupeids (Clupeidae, Herring Clupea harengus and Sprats 

SEABIRD 21 (2008): 44–54


Sprattus sprattus) also featured highly in the diet. This paper updates information on 

the diet of these seabirds on Canna up to 2007 and compares the findings with 

information from other locations in northern Britain. 

Methods 

Canna (57°03’N, 6°32’W) is situated in the southern Minch in western Scotland and 

annual visits were made to the colony to ring and monitor seabirds in the period 

1981–2007.Three visits were made each year, one for five days in late May, one for seven 

days in late June/early July and one for five days in late July/early August. Regurgitations 

were collected from young and adults feeding chicks of Black-legged Kittiwake (hereafter 

‘Kittiwake’) and European Shag (‘Shag’). Fish being carried back to chicks by adult 

Common Guillemots (‘Guillemot’) were also collected. Adult Guillemots returning with 

fish to the colony were targeted and captured and the fish they dropped were collected 

(Swann et al. 1991). Kittiwake and Guillemot samples were mainly collected between 28 

June and 10 July, whilst Shag samples were collected, if available, on all visits. The 

contents of regurgitates and pellets were broken up in warm water or were digested in a 

warm solution of biological washing powder (biotex), fish present were identified, using 

otoliths where necessary (Härkönen 1986) and any intact specimens were measured. 

Sandeels, mostly if not all Lesser Sandeels, were classified as 0-group (young of the year) 

or older, based on the absence or presence of annual growth rings in otoliths (ICES 1995). 

Fish from Guillemots were identified, using otoliths where necessary, and the majority 

were measured (tip of snout to tip of tail) and weighed (to 0.1 g), although we are aware 

that these fish will have lost weight due to desiccation while being carried back to the 

colony held in the adult’s beak (Montevecchi & Piatt 1987). Prior to 1990 gadids were 

measured and weighed but measurements were not always matched with species. 

Unless otherwise stated, all Clupeidae appeared to be Sprat, although the identification 

of very small specimens was problematic, so the name sprat is used for that family. 

Pellets regurgitated by fully-grown Shags of unknown breeding status were collected 

between May and September and were digested in biotex until the mucous was 

dissolved. Otoliths were identified to family and when they were not too worn, and 

time allowed, to species level. Other items (crustacean fragments, polychaete jaws, 

cephalopod beaks) were assigned to the lowest possible taxa. Otoliths from sandeels 

fell obviously into two sizes. ‘Large’ were from older sandeels, but ‘small’ were a 

mixture of those from 0-group and those from older individuals that had their outer 

parts (and so growth rings) eroded away by the acid of the stomach (Johnstone et al. 

1990) so no attempt was made to age these otoliths. 

Results are expressed as frequency of occurrence in samples and, for otoliths and other 

hardparts, the actual numbers present. No attempt was made to pair up otoliths or 

mandibles of polychaete worms but fragments of crustacea and pairs of cephalopod 

beaks were treated as single items. A few regurgitates and many pellets contained more 

than one category of food so percentages of occurrence can total more than 100%. 

Some of the earlier data are described in Swann et al. (1991). Samples were collected 

on only a few days so we could not look for within-year variation in diet and we assume 

that what we report are representative of the chick-rearing season as a whole. 

SEABIRD 21 (2008): 44–54 

45

46 


Results 

Shag: Regurgitations were dominated by sandeels that were found in 67% of the 

134 samples (Table 1). There was no significant trend in the proportion of samples 

with sandeels during the period (linear regression on arcsin transformed data: n = 24 

years weighted by sample size, R 2 = 13.8%, P = 0.14). Of 40 regurgitations where 

sandeel size was assessed, 19 (48%) had small sandeels (0-group) and 24 (60%) 

large (older) sandeels. Gadids occurred in 36% of samples; 17 fish were identified to 

genus or species – Trisopterus spp. (8), Whiting Merlangius merlangus (7), Cod Gadus 

morhua (1) and rockling Ciliata/Gaidropsarus sp. (1). Two samples contained Sprat 

(1), two unidentified wrasse (Labridae), whilst single samples contained Herring, 

Butterfish Pholis gunnellus (Pholidae), Bull-rout Myoxocephalus scorpius (Cottidae), 

Viviparous Blenny Zoarces viviparus (Zoarcidae), pipefish (presumably Snake Pipefish 

Entelurus aequoreus (Syngnathidae)), unidentified fish, Sea Mouse Aphrodite 

aculeate (Polychaete worm) and small crab (Crustacea). 

Figure 1. Annual frequency of occurrence of 

otoliths and other remains of sandeels 

(Ammodytidae), cod-fishes (Gadidae) and other 

prey groups in pellets of European Shag 

Phalacrocorax aristotelis on Canna. 

% occurrence 

Table 1. Contents of regurgitations from European Shags Phalacrocorax aristotelis on Canna. See text for 

details of other items. 

Sample % % % Other 

Year size Sandeels Gadidae items 

1981 6 50 33 17 

1982 2 100 0 0 

1983 2 100 0 0 

1984 5 80 20 0 

1987 6 100 0 0 

1988 3 100 0 0 

1989 5 100 20 0 

1990 8 50 50 0 

1991 4 100 25 25 

1992 6 50 67 17 

1993 2 0 50 50 

1994 7 86 14 0 

1995 9 44 44 22 

SEABIRD 21 (2008): 44–54 

100 

80 

60 

40 

20 

0 

1986 1989 1992 1995 1998 2001 2004 

Sandeels Gadids Other 

Sample % % % Other 

Year size Sandeels Gadidae items 

1996 3 33 67 0 

1997 14 85 14 7 

1998 11 82 36 0 

1999 4 75 50 0 

2000 1 100 0 0 

2001 8 25 75 25 

2002 5 100 0 0 

2003 11 55 45 0 

2004 3 100 0 0 

2005 0 

2006 6 17 83 17 

2007 3 33 100 33 

Total 134 67 36 8


Table 2. Frequency of occurrence and total numbers of otoliths and other items in pellets regurgitated by 

European Shags Phalacrocorax aristotelis on Canna. See text for details of other items. 

% of pellets containing % of otoliths containing 

Period No. of Pellets Sandeels Gadidae Other No. of Otoliths Sandeels Gadidae Other 

1986 early August 6 50 100 0 231 5 95 0 

1989 early August 10 30 100 0 4195 5 95 0 

1993 early July 5 0 100 40 547 0 98 2 

1993 early August 8 0 100 0 3314 0 100 0 

1993 mid Sept 3 0 100 33 475 0 100

48 


No systematic attempts were made to identify gadid otoliths, but many species were 

present including Trisopterus spp. (25% of 3,621 that were assigned to genus), Saithe 

Pollachius virens, rockling,Whiting, Haddock Melanogrammus aeglefinus, Cod and Ling 

Molva molva. Most of these otoliths would have come from fish 50–150 mm long. Of 

the sandeel otoliths, 626 (67%) were ‘small’ (probably mainly 0-group but including 

some eroded otoliths from older fish) and 310 (33%) were ‘large’. 

Kittiwake: Sandeels were the commonest species found in regurgitations in 14 of the 

20 years, and occurred in 60% of the 227 samples (Table 3), gadids were the 

commonest food in four years, sprat in one, and in one year the two families were 

equally represented. Gadids and sprat were present in 22% of samples and very large 

numbers of very small pelagic crustaceans were found in 5%. The ‘other’ category was 

made up of four Lumpsuckers Cyclopterus lumpus (Cyclopteridae), two pipefish (both 

in 2007), one Three-spined Stickleback Gasterosteus aculeatus (Gasterosteidae) and a 

wrasse. Of the 62 samples where sandeels were aged, 58 (94%) were 0-group. Due to 

the lack of a colony-specific otolith:fish length relationship we could not calculate the 

lengths of these fish, but the bulk of relatively intact 0-group were 60–80 mm. Other 

fish were mostly very digested but most gadids appeared to be 60–100 mm long, 

whereas sprats were smaller at 50–60 mm. 

Table 3. Frequency of occurrence of fish families, crustacea and other items in regurgitations of Black-legged 

Kittiwakes Rissa tridactyla on Canna. See text for details of other items. 

% regurgitations with 

Year No. of Samples Sprats Sandeels Gadidae Crustacea Others 

1987 7 0 14 86 0 0 

1988 6 0 0 83 17 17 

1989 17 0 94 12 0 0 

1990 6 0 100 0 0 0 

1991 3 0 100 33 0 0 

1992 8 25 25 63 13 0 

1993 3 0 67 33 0 0 

1994 6 0 100 0 17 0 

1995 5 0 40 40 40 0 

1996 9 33 67 11 0 0 

1997 20 5 70 15 0 15 

1998 4 50 25 50 50 0 

1999 22 36 59 18 0 0 

2000 32 16 81 0 13 3 

2001 12 0 67 42 0 0 

2002 22 27 73 0 0 0 

2003 11 9 82 27 0 0 

2004 21 95 5 5 0 0 

2005 9 11 22 67 11 11 

2006 0 

2007 4 0 75 50 0 50 

Total 227 22 60 22 5 4 

SEABIRD 21 (2008): 44–54


Guillemot: Of the 1,562 fish collected, 753 (48%) were Clupeidae (three Herring and 

the rest Sprat), and 416 (27%) were Gadidae including 314 Whiting, 52 Trisopterus spp. 

(mostly unidentified but including Poor Cod T. minutus, Bib T. luscus and Norway Pout 

T. esmarkii), 25 Haddock, 10 Saithe, two Blue Whiting Micromesistius poutassou and 

two Cod; the rest of the gadids were not identified to species level. Sandeels (24%) 

made up the remainder (Table 4). There were large variations in the fish taken from 

year to year (Figure 2), but generally sprats were less important in the 1980s and 

2000s than during the 1990s (x2 = 57.5, 2 df, P < 0.001). On an annual basis, sprats 

were the most important, making up on average 47 ± 4% of the fish. Sandeels and 

Gadidae contributed 26 ± 3% and 27 ± 3%, respectively. 

100 

80 

60 

40 

20 

0 

1982 1986 1990 1994 1998 2002 2006 

Ammodytidae Gadidae Clupeidae 

Figure 2. Fish family composition (% by number) of prey taken by Common Guillemots Uria aalge on Canna. 

Excludes 1981 when sample size < 10. 

The length of sprats ranged from 68–155 mm, with a peak at 110–14 mm (Figure 3a). 

Most of these fish will have been 1-group or older. There was a significant decline in 

mean length over the period (linear regression: Length (mm) = 765 – 0.328 year, n = 

717, P < 0.001); however since the overall decrease was only 5 mm, and year explained 

only 3% of the variation, this decline was unlikely to be biologically meaningful. The 

mean length of gadids varied from 42–147 mm with a peak at 80–84 mm (Figure 3b). 

The length of Whiting, the commonest gadid, showed no significant change in size over 

the study period (linear regression: R 2 = 0.0%, n = 274, P = 0.903), though there were 

marked annual variations in the size of fish taken (Appendix 1). Sandeels varied in length 

from 66–231 mm, with peaks around 90, 130 and 160 mm (Figure 3c). Fish were not 

aged but, given that most 0-group sandeels in samples from Kittiwakes were 60–80 mm 

long, it is unlikely that many of these from Guillemots were 0-group. There was no 

significant change in the size of sandeels taken over the period (linear regression: n = 

315, R 2 = 0.1%, P = 0.66). Overall the mean weights of individual sprats, gadids, and 

sandeels were 10.7 ± 0.1 g (n = 669), 6.7 ± 0.2 g (339) and 7.8 ± 0.3 g (280), respectively; 

assuming that each of these fish had lost equal weight due to desiccation, then 

59%, 20% and 21% of the biomass of food fed to chicks came from these fish families. 

SEABIRD 21 (2008): 44–54 

49

50 


Table 4. Fish collected from Common Guillemots Uria aalge on Canna. 

Gadidae Gadidae Gadidae 

Year Number % Sandeels % Sprats % Whiting % other species % unknown 

1981 5 20 40 0 40 0 

1982 27 56 26 0 15 0 

1983 23 35 48 9 4 4 

1984 34 15 62 6 12 6 

1985 24 33 50 4 13 0 

1986 72 69 17 4 8 0 

1987 22 18 73 9 0 0 

1988 64 48 28 13 9 2 

1989 74 16 35 31 18 0 

1990 72 32 31 28 8 1 

1991 106 32 35 20 12 1 

1992 77 1 86 10 3 0 

1993 76 16 49 32 4 0 

1994 66 18 74 6 2 0 

1995 87 10 71 17 1 0 

1996 78 14 71 10 4 1 

1997 27 7 85 7 0 0 

1998 81 20 36 40 5 0 

1999 76 13 70 16 1 0 

2000 89 21 49 26 1 1 

2001 113 27 52 12 7 3 

2002 24 8 67 25 0 0 

2003 56 9 50 39 2 0 

2004 119 19 35 45 1 0 

2005 11 64 18 18 0 0 

2006 40 40 8 13 38 3 

2007 19 47 5 16 16 16 

Total 1562 24 48 20 7 1 

Discussion 

The finding that sandeels were the main constituent (62%) of regurgitations from young 

Shags and adults feeding chicks accords well with numerous studies in Britain and 

Europe, although most reported higher frequencies (Pearson 1968; Harris & Riddiford 

1989; Barrett et al. 1990; Guyot 1990; Harris & Wanless 1991; Harris & Wanless 1993). 

However, the frequency of gadids (37%) was unexpectedly high when compared with 

none in 57 and 141 regurgitates from Fair Isle, Shetland in 1986-1989 and the Isle of 

May, Firth of Forth in 1985–1990, respectively (Harris & Riddiford 1989; Harris & 

Wanless 1991). There are problems in describing the diet of Shags using otoliths, arising 

from the differential rates of digestion of otoliths of different sizes and from different 

families. For instance, otoliths from gadids are much more robust than those from 

sandeels and particularly sprats, whereas Butterfish and gobies have notably small 

otoliths for fish of their size (Härkönen 1986; Johnstone et al. 1990). However, gadids 

were represented in more than twice as many pellets as were sandeels and the total 

numbers recovered were over seven-times greater, so it seems likely that the diet of 

these individuals, full-grown but of unknown breeding status, that produced the pellets 

SEABIRD 21 (2008): 44–54

200 

180 

160 

140 

120 

100 

80 

60 

40 

20 

0 

80 

60 

40 

20 

0 

60 

40 

20 

0 

a) Clupeidae (n = 717) 

b) Gadidae (n = 384) 

c) Sandeels (n = 315) 


40- 50- 60- 70- 80- 90- 100- 110- 120- 130- 140- 150- 160- 170- 180- 190- 200- 210- 220+ 

44 54 64 74 84 94 104 114 124 134 144 154 164 174 184 194 204 214 

Length (mm) 

Figure 3. Distribution of lengths (mm) of fish from Common Guillemots Uria aalge on Canna. 

was substantially different to that of the chicks. Such a difference in the diet of chicks 

and full-grown Shags has been recorded elsewhere (Harris & Wanless 1993). The 

significant lower proportion of pellets later in the season that had sandeels present was 

possibly the result of sandeels leaving the water column and burying themselves in the 

sand, where they spend most of the rest of the summer and winter (Macer 1966). 

Although sandeels occurred in 60% of regurgitations from Kittiwakes and was the 

commonest family in 14 of the 20 years, these figures were low compared to many 

other British colonies. Sandeels made up an average of 83% of the biomass fed to 

chicks on the Isle of May over 18 years, were present in all 66 regurgitates on Fair Isle 

1986–88, and in 98 of 100 regurgitates on Farne Islands, Northumberland 1998–2000 

(Harris & Riddiford 1989; Lewis et al. 2001; Bull et al. 2004; Wanless et al. 2007). 

Gadids have been recorded at several other colonies, e.g. Farne Islands, Inchkeith, 

Inchcolm and Isle of May off northeast Britain (Bull et al. 2004), but not at such a high 

rate as on Canna. In contrast Sprats are regular in the diet and sometimes made up 

the bulk of the food at Inchkeith and Inchcolm. 

SEABIRD 21 (2008): 44–54 

51

52 


Figure 1. Common Guillemots Uria aalge with gadids, Canna, July 2006 © Kenny Graham. 

By number, approximately half the diet of young Guillemots was sprat, a quarter 

sandeels and a quarter gadids, and in biomass terms sprats contributed 59%.These are 

the normal constituents of the diet of young Guillemots in northern Britain. However, 

normally sandeels make up the bulk of the diet with sprats the usual alternative. On 

the Isle of May between 1981 and 2004 the relative mean annual percentages of these 

two families (by number) were 56% and 42% with the remainder being made up by 

a variety of other families (Wanless et al. 2005). Among 303 fish brought to Fair Isle 

in 1986-89, only four (two Sprat and two gadids) were not sandeels (Harvey et al. 

1990). Gadids have been recorded from the stomachs of Guillemots shot during the 

summer in the 1980s in Shetland (11/83 birds with food remains), the Summer Isles 

(5/27), the Sound of Jura (1/5), the Clyde Front (1/5) and St Kilda (11/19) (Blake et al. 

1985; Halley et al. 1995). However, it is difficult to compare these with fish brought 

to chicks since there are differences in the diet of parent birds and their chicks even at 

the same time at the same colony (Wilson et al. 2004). 

Canna seabirds fed their chicks on a wide variety of fish species but the bulk of the 

diet was made up of sprats, sandeels and gadids (a variety of species but mainly 

Trisopterus spp. and Whiting). Other groups were of minimal importance. There are no 

data available on the numbers of small fish in the waters around Canna to allow a 

meaningful discussion as to whether the birds were selecting these species or were 

just eating the main species available to them. Over the period of the study there were 

few significant changes in species composition or the size of prey taken by Canna 

seabirds and the results confirm earlier findings that gadids are a normal and 

important part of the diet of seabirds at this colony. 

SEABIRD 21 (2008): 44–54



We thank the many people who have helped collect the field data over many years, in 

particular Andrew Call, Simon Foster, Alan Graham, Kenny Graham, Kathryn 

Mackinnon, Andrew Ramsay and Alastair Young. John Hislop gave invaluable help with 

fish and otolith identification, and JNCC provided financial support through their 

seabird monitoring programme. Rob Barrett and Kees Camphuysen improved the 

manuscript with their criticisms. 

References 

Barrett, R.T., Røv, N., Loen, J. & Montevecchi,W.A. 1990. Diets of Shags Phalacrocorax aristotelis 

and Cormorants P. carbo in Norway and possible implications for gadoid stock recruitment. 

Marine Ecology Progress Series 66: 205–218. 

Blake, B. F., Dixon, T. J., Jones, P. H. & Tasker, M. L. 1985. Seasonal changes in the feeding 

ecology of Guillemots (Uria aalge) off north and east Scotland. Estuarine, Coastal and Shelf 

Science 20: 559–568. 

Bull, J., Wanless, S., Elston, D. A., Daunt, F., Lewis, S. & Harris, M. P. 2004. Local-scale 

variability in the diet of Black-legged Kittiwakes Rissa tridactyla. Ardea 92: 43–82. 

Daunt, F., Wanless, S., Greenstreet, S. P. R., Jensen, H., Hamer, K. C. & Harris, M. P. 2008. The 

impact of the sandeel fishery closure in the northwestern North Sea on seabird food consumption, 

distribution and productivity. Canadian Journal of Fisheries and Aquatic Sciences 65: 362–391. 

Frederiksen, M., Wanless, S., Harris, M. P., Rothery, P. & Wilson, L. J. 2004. The role of the 

industrial fishery and climate change in the decline of North Sea Black-legged Kittiwakes. 

Journal of Applied Ecology 41: 1129–1139. 

Guyot, I. 1990. Le Cormoran Huppé en Corse: Biologie et interactions avec la pêche profesionnelle. 

Travaux Scientifiques du Parc Naturel Regional et des Réserves Naturelles de Corse 28: 1–40. 

Halley, D. J., Harrison, N., Webb, A. & Thompson, D. R. 1995. Seasonal and geographical 

variations in the diet of Common Guillemots Uria aalge off western Scotland. Seabird 17: 12–20. 

Härkönen, T. 1986. Guide to the Otoliths of Bony Fishes of the Northeast Atlantic. Danbiu ApS, 

Hellerup, Denmark. 

Harris, M. P. & Riddiford, N. J. 1989. The food of some young seabirds on Fair Isle in 1986–88. 

Scottish Birds 15: 119–125. 

Harris, M. P. & Wanless, S. 1991. The importance of the Lesser Sandeel Ammodytes marinus in 

the diet of the Shag Phalacrocorax aristotelis. Ornis Scandinavica 22: 375–882. 

Harris, M. P. & Wanless, S. 1993. The diet of Shags Phalacrocorax aristotelis during the chickrearing 

period assessed by three methods. Bird Study 40: 135–139. 

Harvey, P. V., Harris, M. P., Osborn, K., Riddiford, N. & Silcocks, A. F. 1990. The breeding success 

and diet of Fair Isle’s seabirds in 1986–1989. Fair Isle Bird Observatory Report 42: 47–54. 

ICES 1995. Report of the ICES workshop on sandeel otolith analysis: Review of sandeel biology. 

ICES, CM 1995/G:4. 

Johnstone, I. G., Harris, M. P., Wanless, S. & Graves, J. A. 1990. The usefulness of pellets for 

assessing the diet of adult Shags Phalacrocorax aristotelis. Bird Study 37: 5–11. 

Lewis, S.,Wanless, S.,Wright, P. J., Harris, M. P., Bull, J. & Elston, D. A. 2001. Diet and breeding 

performance of Black-legged Kittiwakes Rissa tridactyla at a North Sea colony. Marine Ecology 

Progress Series 221: 277–284. 

Macer, C. T. 1966. Sandeels (Ammodytidae) in the southwestern North Sea; their biology and 

fishery. MAFF Fishery Investigations 2. 24: 1–51. 

SEABIRD 21 (2008): 44–54 

53

54 


Mavor, R. A., Parsons, M., Heubeck, M. & Schmitt, S. 2006. Seabird numbers and breeding 

success in Britain and Ireland, 2005. Joint Nature Conservation Committee, Peterborough. (UK 

Nature Conservation No. 30.) 

Monaghan, P. 1992. Seabirds and sandeels: the conflict between exploitation and conservation 

in the northern North Sea. Biodiversity and Conservation 1: 98–111. 

Montevecchi, W. A. & Piatt, J. F. 1987. Dehydration of seabird prey during transport to the 

colony: effects on wet weight energy densities. Canadian Journal of Zoology 65: 2822–2824. 

Pearson, T. H. 1968. The feeding biology of seabird species breeding on the Farne Islands, 

Northumberland. Journal of Animal Ecology 37: 521–552. 

Swann, R. L., Harris, M. P. & Aiton, D. G. 1991. The diet of some young seabirds on Canna, 

1981-90. Seabird 13: 54–58. 

Wanless, S., Frederiksen, M., Daunt, F., Scott, B. E. & Harris, M. P. 2007. Black-legged 

Kittiwakes as indicators of environmental change in the North Sea: Evidence from long-term 

studies. Progress in Oceanography 72: 30–38. 

Wanless, S., Harris, M. P., Redman, P. & Speakman, J. 2005. Low fish quality as a probable cause 

of a major seabird breeding failure in the North Sea. Marine Ecology Progress Series 294: 1–8. 

Wilson, L. J., Daunt, F. & Wanless, S. 2004. Self-feeding and chick provisioning diets differ in 

the Common Guillemot Uria aalge. Ardea 92: 197–208. 

Appendix 1. The mean lengths (and their standard error, SE) of fish from collected from Common Guillemots 

Uria aalge on Canna. 

Sandeels Sprats Whiting 

Year Number Mean (mm) SE Number Mean (mm) SE Number Mean (mm) SE 

1983 10 112.2 0.8 

1984 21 114.1 1.7 

1985 7 147.6 8.7 11 111.1 1.5 

1986 48 112.7 4.3 12 126.0 3.6 

1987 16 106.4 1.7 

1988 20 149.2 6.5 16 111.6 2.3 

1989 12 154.0 9.1 26 112.2 1.8 

1990 23 125.6 7.5 22 114.4 1.0 

1991 34 134.8 6.4 36 112.3 1.2 21 86.0 1.8 

1992 1 143.0 64 113.1 1.3 8 84.8 3.3 

1993 12 157.9 8.5 36 113.5 1.2 24 83.0 2.0 

1994 12 141.3 11.3 49 110.6 1.5 4 89.0 6.5 

1995 8 162.1 7.2 62 111.2 0.8 15 92.7 2.2 

1996 11 132.5 13.7 55 111.2 0.7 8 94.4 3.4 

1997 2 162.5 37.5 23 111.8 2.3 2 92.0 12.0 

1998 16 126.6 8.3 29 116.8 1.5 32 90.2 3.7 

1999 10 97.2 8.5 53 105.6 1.2 12 86.8 5.1 

2000 19 126.8 11.4 42 112.8 1.5 22 103.1 2.5 

2001 23 103.3 6.7 47 106.5 1.4 13 93.4 2.9 

2002 2 105.0 12.0 16 108.7 1.6 6 100.2 5.0 

2003 5 143.6 27.9 22 115.4 2.4 22 92.1 2.1 

2004 23 127.8 6.5 42 102.4 1.6 52 88.7 1.8 

2005 7 148.9 10.9 2 110.5 0.5 2 82.0 11.0 

2006 16 153.9 6.7 3 110.7 5.3 5 75.6 9.2 

2007 4 151.8 9.4 1 119.0 3 66.0 8.1 

SEABIRD 21 (2008): 44–54

Colony habitat selection by Little Terns 


Sternula albifrons in East Anglia: 

implications for coastal management 

Ratcliffe, N. 1 *, Schmitt, S. 2 ,Mayo,A. 2 , Tratalos, J. 3 and Drewitt, A. 4 

*Correspondence author. Email: notc@bas.ac.uk 

1 RSPB, East Scotland Regional Office, 10 Albyn Terrace, Aberdeen AB10 1YP, UK (Current 

address: British Antarctic Survey, NERC, High Cross, Cambridge CB3 0ET, UK); 2 RSPB, The 

Lodge, Sandy, Bedfordshire SG19 2DL, UK; 3 University of East Anglia, Norwich NR4 7TJ, UK; 

4 English Nature, Northminster House, Peterborough PE1 1UA, UK. 

Abstract 

Little Terns Sternula albifrons are unusual among UK seabirds in that a large proportion 

of the population breed on mainland beaches in East Anglia. Relative sea-level rise 

means that such habitats are under threat in this region, and so we quantified colony 

habitat selection of beach nesting Little Terns in order to inform habitat restoration 

and creation initiatives. Random 1 km sections of beach were selected and the 

presence or absence of a Little Tern colony within each was related to physical 

(substrate type, height and width), biotic (vegetation cover, predator activity) and 

anthropogenic (disturbance) characteristics using logistic regression models. Little 

Terns positively selected for beaches with vegetation cover and negatively for those 

with high disturbance levels. They showed no selection according to width and height 

or Red Fox Vulpes vulpes presence, even though these are likely to affect flood and 

predation risk respectively. Red Foxes were found to be widespread on beaches 

irrespective of tern colony presence, and so movement of tern colonies will not always 

result in predator avoidance. Little Tern habitat creation needs to be integrated into 

coastal management plans in order to safeguard their population from the combined 

threats of relative sea-level rise, predation and disturbance. 

Introduction 

The population size and range of Little Terns Sternula albifrons in the UK have declined 

by 24% since the mid 1980s (Pickerell 2004), and qualifies it for inclusion on the Amber 

list based on a moderate rate of population decline (Gregory et al. 2002). It is also listed 

on Annex I of the Birds Directive (EC Directive on the conservation of wild birds 

79/409/EEC) and 67% of the GB population lies within Special Protection Areas (Stroud 

et al. 2001). Little Tern distribution through the UK has been broadly stable since the 

1960s, being wide and patchy with a stronghold in East Anglia where 69% of the 

population currently breed (Pickerell 2004).There is an absence of sandy offshore barrier 

islands in the UK, and so Little Terns nest mainly on low-lying shingle beaches, spits and 

estuarine islets, and here they face threats from disturbance by coastal tourism, 

predation and inundation by high tides or storms (Ratcliffe 2004). These threats are 

mitigated to varying extents by intensive colony protection schemes (Smart 2004) that 

currently cost conservation organisations tens of thousands of pounds per annum, and 

SEABIRD 21 (2008): 55–63 

55

56 


need to be run in perpetuity in order to maintain Little Tern populations. Furthermore, 

the colony habitat of Little Terns is under threat from relative sea level rise and increased 

storminess (Norris & Buisson 1994; Gill 2004), such that protection of existing colonies 

from disturbance and predation may be inadequate to conserve them in the future. 

Rather than continuing to focus on colony site protection, a strategic approach to the 

management of Little Tern distribution relative to threats may result in more effective 

conservation. Little Terns require small areas of habitat that can be restored or created 

quickly and inexpensively, and these can be sited in areas where threats to breeding 

success are lower. For example, creation of dredge spoil islands offshore could reduce 

predation and disturbance without need for fencing and wardening (Parnell et al. 1986; 

Burgess & Hirons 1992; Erwin et al. 2001), and sediment recharge of eroded narrow 

beaches could mitigate loss of habitat or colony flooding risks (Charlton & Allcorn 

2004). The characteristics of Little Tern colony sites need to be quantified and means 

of creation or restoration have to be tested in order to achieve this (Gill 2004). 

Shoreline management plans are currently under review for East Anglia and decisions are 

being made that will affect coastal habitats in future decades. These decisions are mainly 

influenced by engineering and economic considerations at present, but it is important 

that conservation requirements are also included (Gill 2004). A strategic review of Little 

Tern habitat requirements would be timely in order to influence coastal planning decisions 

in a manner that is sympathetic towards Little Terns and to exploit opportunities that may 

arise as a by-product of managed realignment or coastal defence (Gill 2004). 

This project examines colony habitat selection by Little Terns in relation to beach 

characteristics, including width, height, shoreward habitat type, substrate type, 

disturbance and predation risk.The implications of the results for habitat management 

and creation are discussed and recommendations for further research are made. 

Study sites and methods 

The entire length of beaches in Norfolk and Suffolk (east England) were divided into 1 

km lengths using GIS, and these were used as the sampling unit for further data 

collection and analyses. 

Beach habitat characteristics: Beach characteristics were studied in 68 randomly 

selected 1 km sample sections of beach in Norfolk and Suffolk, representing 33% of 

the entire beach length within these counties. Beach width from the landward limit of 

the beach to mean high water mark was extracted every 200 m from Ordnance Survey 

maps using GIS. Data on maximum beach height was extracted from Environment 

Agency beach profile databases. Substrate was quantified within 30 2.5 x 2.5 m 

quadrats placed randomly along the beach.The percentage cover of sand, shingle, rock, 

debris and vegetation was recorded in each. The values were averaged within each of 

the 1 km sample units. Fieldwork was conducted in April 2003 prior to the return of 

Little Terns to their colonies to avoid disturbance. The presence or absence of breeding 

Little Terns in each sample was determined by walking and scanning each section of 

coast three times during May and June 2003 when they would be incubating eggs. 

SEABIRD 21 (2008): 55–63


Disturbance: Observations of disturbance within a subsample of 45 of the previous 68 

beach sections were conducted, during May and June 2003, by walking each 1 km 

section of beach and counting the number of people and dogs. Each section was 

visited three times, with samples being taken at both weekdays and weekends to 

account for variation in disturbance owing to day of the week. Samples were only 

taken on fine days that were conducive to beach recreation to reduce variation owing 

to weather.These counts were averaged to provide an estimate of disturbance pressure 

within the sampled 1 km section. 

Dummy nests were used to assess relative trampling risk in 40 sections of beach during 

late May and early June. Twenty Common Quail Coturnix coturnix eggs were each 

placed in separate artificial scrapes along the length of the beach section above the 

high tide line and were marked with a numbered peg. These were visited every 3–5 

days over a period of 23 days to determine whether they were stepped upon by 

humans. Daily trampling rates for each section were obtained using the Mayfield 

method (Mayfield 1975). 

Predation: Predation risk to Little Terns was determined during May and June 2003 by 

using sand traps and dummy nests. Sand traps were deployed at 30 sections of beach 

sampled from the previous 45, with five traps being set within each. These comprised 

2.5 m 2 of smoothed-over sand with the perimeter marked with lines of pebbles to 

enable it to be relocated. A spoonful of cat food was placed in the centre of this. The 

traps were set in the evening and visited early in the morning and any footprints 

present were identified by reference to Bang & Dahlstrom (2001). Five traps were set 

on each of three nights in each sample section, with Red Foxes Vulpes vulpes being 

classed as present if footprints were found in a trap on one or more occasions. 

Daily nest predation rates were assessed using the dummy egg experiment described 

above, except that the numbers of nests predated were analysed as opposed to 

numbers trampled. 

Analysis: The presence or absence of Little Terns in a sample stretch of beach was 

related to the explanatory variables using logistic regression. We used presence or 

absence rather than number of breeding pairs as an index of suitability because we 

were interested in correlates of colony site selection rather than colony size. Moreover, 

abundance of Little Terns is more annually variable than site occupancy, and 

abundance in colonial birds is prone to non-independence (i.e. pairs selecting a site 

because of the presence of conspecifics, rather than owing to habitat characteristics). 

A forward stepwise procedure was used to select the minimal adequate model, using a x2 

alpha of 0.05 as the threshold for inclusion of an explanatory variable in the model. Chisquare 

statistics for each explanatory variable in the results are taken from the minimal 

adequate model. It was not possible to build a global model of all variables of interest 

since owing to the uneven sampling effort among parameters caused by manpower 

limitations (see sections above), and so separate models were fitted for each. Observed 

frequencies and model fits were plotted graphically according to Smart et al. (2004). 

SEABIRD 21 (2008): 55–63 

57

Likelihood of Occupancy 

0.8 

0.7 

0.6 

0.5 

0.4 

0.3 

0.2 

58 


Results 

Beach characteristics: Nineteen of the 68 samples were occupied by Little Tern 

colonies. The likelihood of a beach being used by breeding Little Terns increased significantly 

with percentage vegetation cover (x2 1 = 3.77, P = 0.05). Occupancy likelihood 

was 23% in the absence of vegetation, but rose to 70% when coverage was 20% (Figure 

1). Fitting the square of vegetation cover to the model did not significantly reduce the 

residual deviance, suggesting that there was no decline as vegetation became more 

dense within the range of cover observed. Fitting linear and quadratic models that 

included the variables beach height, beach width, the percentage cover of sand, gravel, 

rock or debris did not significantly reduce the residual deviance (all P > 0.05). 

Disturbance: Occupancy likelihood declined significantly with the average number of 

people observed per km of beach (x2 1 = 4.5, P < 0.05), but not with the number of 

dogs (x2 1 = 1.30, P > 0.2). Occupancy likelihood was 47% in the absence of people but 

fell to zero when the average number of people exceeded 25 (Figure 2). 

Daily nest trampling rates of dummy clutches were 1.33%, equating to a 26% survival 

rate over a 23-day incubation period, in the absence of other forms of loss. The 

likelihood of a site being occupied by Little Terns was not significantly affected by nest 

trampling rates of dummy clutches within sample sites (x2 1 = 2.6, P > 0.1). 

Predation risk: Sand traps revealed that foxes were present on 62% of beaches, but 

the likelihood of a beach being occupied by breeding Little Terns was not significantly 

affected by the presence of foxes (x2 1 = 0.03, P > 0.8). Daily nest predation rates on 

dummy clutches were 1.08%, equating to a survival over a 23-day incubation period 

of 22%. Site occupancy was not explained by predation rates experienced by dummy 

clutches within the sampled stretches of beach (x2 1 = 0.00, P > 0.9). 

0.1 

0 0.0 

0 

0 5 10 15 20 25 30 

0 20 40 60 80 100 

Percentage Vegetation Cover 

Average Number of People 

Figure 1 (left). The effect of percentage vegetation cover on the likelihood of a site being occupied by breeding 

Little Terns Sternula albifrons. The line represents the best fit line from the logistic regression model plotted on the 

left y axis , while the histograms show observed frequencies of presence (on the top x axis, which is inverted) and 

absence (on the bottom x axis) plotted against the right hand y axis (see Smart et al. 2004). Figure 2 (right). The 

effect of human disturbance pressure on the likelihood of a site being occupied by breeding Little Terns Sternula 

albifrons. See the legend of Figure 1 for details. 

SEABIRD 21 (2008): 55–63 

0 

10 

20 

20 

10 

Frequency 

Likelihood of Occupancy 

1.0 

0.8 

0.6 

0.4 

0.2 

0 

5 

Frequency 

5


Discussion 

While numbers of breeding pairs and breeding success within Little Tern colonies in 

East Anglia varies markedly among years, the locations of occupied sites has remained 

remarkably consistent since regular monitoring began in the 1960s (Pickerell 2004; 

Ratcliffe 2004). This suggests colony sites are selected consistently according to 

temporally stable habitat characteristics rather than selected at random or in response 

to previous breeding success. This is the first study to attempt to identify the physical 

factors driving colony site selection by Little Terns at a regional scale within the UK. 

Of the physical beach habitat characteristics measured, only vegetation cover had a 

significant positive effect on colony site selection. Similarly, Goutner (1990) found 

Little Terns nested near low vegetation in Greece, and the closely related and ecologically 

similar Least Terns S. antillarum also selected sites with sparse vegetation cover 

(Kotliar & Burger 1986). However, thick vegetation cover greater than the 30% 

maximum found in this study might result in avoidance: Burger & Gochfeld (1990) 

found that Least Tern nests were closer to vegetation than random points where cover 

was sparse and further away where cover was dense. 

Studies of Least Tern colony site selection in the USA (Thomas & Slack 1982; Gochfeld 

1983; Kotliar & Burger 1986; Burger & Gochfeld 1990) have shown a preference for 

mixed sand and gravel or shell substrates (which aids camouflage of nests) and high, 

wide beaches (which reduce the risk of tidal inundation). However, in this study, 

substrate and beach height/width variables had no effect on colony site selection by 

Little Terns. This may have been because these factors are not selected for by Little 

Terns in East Anglia or that power to detect such effects was insufficient owing to 

sample size or co-linearity of explanatory variables. Alternatively, selection for beach 

characters may have been obscured by another factor that was not measured, with the 

most likely of these being food availability. 

Little Terns have very short foraging ranges compared to most seabirds, with most food 

generally being obtained from within 5 km of the colony and 1 km of the shore (Davies 

1981; Fasola & Bogliani 1990; Allcorn et al. 2003; Bertolero et al. 2005; Perrow et al. 

2006). Consequently, Little Terns have to nest close to areas of high prey availability so 

that they can provision their chicks with sufficient food to maintain their growth and 

survival (Holloway 1993; Bertolero et al. 2005). Little Terns in the UK generally feed on 

small (30–40 mm) clupeids and sandeel (Davies 1981; Cramp 1985; Perrow et al. 2004), 

and on the east Norfolk coast selected sites where the availability of such prey was 

greatest (Perrow et al. 2004).As such, significant colony site selection according to beach 

characteristics might only be evident when controlling for offshore food availability. 

Collecting data on prey availability at the broad scales required to understand regional 

colony site selection is prohibitively expensive, but there may be potential to infer this 

from marine habitat variables such as turbidity (Perrow et al. 2004). 

Little Terns showed a significant aversion to human disturbance, as noted elsewhere in 

Europe (Scarton et al. 1994; Catry et al. 2003), and for Least Terns in the USA (Thomas 

& Slack 1982; Gochfeld 1983; Kotliar & Burger 1986). Disturbance by humans and 

SEABIRD 21 (2008): 55–63 

59

60 


dogs has the potential to cause trampling or predation of eggs or chicks, or to disrupt 

courtship, incubation or brooding by parent birds, causing reduced breeding success 

and site abandonment (Thomas & Slack 1982; Gochfeld 1983; Kotliar & Burger 1986). 

Indeed, our dummy nest experiment showed that trampling on East Anglian beaches 

would destroy 74% of clutches over a 23-day incubation period in the absence of 

other sources of loss. Humans are active on beaches during both the day and the 

colony-prospecting period of Little Terns, and so birds are able to perceive this pressure 

and select sites accordingly. Continuing to maintain protective cordons around Little 

Tern colonies is therefore essential to prevent their extirpation, but there may also be 

potential to reduce disturbance in other areas that are deemed suitable to encourage 

colonisation. However, excluding tourists from popular areas of beach will be difficult 

and politically contentious, especially in the context of UK Government’s objective to 

increase public access to the coast (Defra 2004). 

Predation is among the most important causes of breeding failure at Little Tern 

colonies, with Red Foxes being the predator most frequently recorded to cause 

significant losses of both eggs and chicks (Ratcliffe 2004). Little Terns did not select 

colony sites according to the occurrence of Red Foxes or the rates of predation failure 

of dummy clutches, and so their ability to perceive, or scope to avoid, predation risks 

appeared to be poor.This is probably because the presence of nocturnal predators only 

becomes apparent to terns during the incubation or chick rearing stage when they 

have committed to breeding at a site. 

Red Foxes were present on a large proportion of the beaches despite the fact that 

trapping effort was relatively low. Red Foxes have increased their range in East Anglia 

owing to reduced gamekeeping effort (Tapper 1992) and this has probably resulted in 

the reduced availability of predator-free colony sites for ground-nesting birds. 

Similarly, declines of beach-nesting terns in Virginia were associated with an increased 

occupancy of barrier islands by Red Foxes and Racoons Procyon lotor and a consequent 

reduction in available predator-free nesting sites (Erwin et al. 2001). Red Fox presence 

was not related to Little Tern presence or absence, suggesting that they are not 

venturing onto beaches solely to exploit tern prey. It is therefore likely that Red Foxes 

forage on most East Anglian beaches to feed on tide-line carrion and scraps dropped 

by tourists, and will opportunistically prey on tern colonies if they encounter them. 

This discredits the widely held belief that high predation rates are an indirect 

consequence of disturbance, which reduces habitat availability and hence the ability 

of birds to move to avoid predators. This assumes that Red Foxes only forage on 

beaches following a chance encounter with a tern colony, whereas the data presented 

in this study suggest they already occur along most stretches of beach and will 

opportunistically predate nests or chicks in any new colonies forming there. 

The combined effects of disturbance avoidance and predation losses mean that 

intensive protection of mainland beach sites will be required in perpetuity if Little Tern 

populations are to be maintained on existing habitat. However, an alternative in the 

longer term is to create islets using dredge spoil which would enable Little Terns to 

nest in sites that are free from predators and people (Burgess & Hirons 1992; Krogh & 

SEABIRD 21 (2008): 55–63


Schweitzer 1999). This method has been successfully employed the USA and now a 

large proportion of some regional tern populations nest on artificial islands (Parnell et 

al. 1986; Visser & Peterson 1994). In southern England, Little Terns typically breed on 

natural islets in saline lagoons or estuaries rather than beaches (Pickerell 2004), but 

such habitats are unavailable throughout much of East Anglia. Further research into 

colony habitat selection for islets is required to ensure new sites are designed 

appropriately. In particular, size is likely to be an important consideration, since larger 

islands are attractive to other species of tern and gulls, which can result in competitive 

exclusion of Little Terns (Fasola & Canova 1992; Scarton et al. 1994). 

Correct location of the islets will be essential to ensure that they are colonised and to 

maximise breeding success. Positioning them within foraging range of concentrations 

of available prey will be essential, and studies of turbidity combined with surveys of 

prey availability using a fine-mesh, epipelagic trawl net (Perrow et al. 2004) would be 

required to achieve this. However, other factors have to be taken into account for 

practical and legal reasons. For example, construction of islets in some locations might 

be unfeasible as they might cause a hazard to navigation, suffer excessive erosion or 

contravene legislation protecting designated nature conservation sites (Harrison & 

Allcorn 2004). Conversely, offshore coastal defence structures or remnants of breached 

sea walls may present opportunities for site creation (Charlton & Allcorn 2004). Once 

islands of a suitable design are created in appropriate locations, Little Terns can be 

attracted to them using decoys and recordings of courtship calls (Kotliar & Burger 

1984; Burger 1988; Jeffries & Brunton 2001; Colombé & Allcorn 2004), which will 

increase the likelihood of them being colonised. 

Further work is required to develop an improved understanding of Little Tern habitat 

requirements and to assess how a network of colony sites can be delivered within the 

constraints and opportunities presented by wider coastal planning initiatives. While 

this work is in progress, it is essential that protection at existing sites is established and 

maintained in order to prevent their extirpation and further population declines. 


This work was funded by a partnership between RSPB and English Nature. We are 

grateful to the wardens of tern colonies in Norfolk and Suffolk for granting permission 

to work on their sites and for assisting with data collection. The Environment Agency 

kindly provided access to the beach profile data. We thank the two anonymous 

referees for constructive criticism of an earlier draft. 

References 

Allcorn, R. I., Eaton, M. A. Cranswick, P. A., Perrow, M., Hall, C., Smith, L., Reid, J., Webb, A., 

Smith, K.W. S., Langston, R. H.W. & Ratcliffe, N. 2003. A pilot study of breeding tern foraging 

ranges in NW England and East Anglia in relation to potential development areas for offshore 

windfarms. RSPB/WWT/JNCC report to DTI, Sandy, UK. 

Bang, P. & Dahlstrom, P. 2001. Animal Tracks and Signs. Oxford University Press, Oxford. 

Bertolero,A., Oro, D.,Vilalta,A. M. & López, M. À. 2005. Selection of foraging habitats by Little 

Terns Sterna albifrons at the Ebro Delta, NE Spain. Revista Catalana d’Ornitologia 21: 37–42. 

SEABIRD 21 (2008): 55–63 

61

62 


Burger, J. 1988. Social attraction in nesting Least Terns: effects of numbers, spacing and pair 

bonds. Condor 90: 575–582. 

Burger, J. & Gochfeld, M. 1990. Nest site selection in Least Terns Sterna antillarum in New 

Jersey and New York. Colonial Waterbirds 13: 31–40. 

Burgess, N. D. & Hirons, J. M. 1992. Creation and management of artificial nesting sites for 

wetland birds. Journal of Environmental Management 34: 285–295. 

Catry, T., Ramos, J. A., Catry, I., Allen-Revez, M. & Grade, N. 2003. Are salinas a suitable 

alternative breeding habitat for Little Terns Sterna albifrons? Ibis 145: 258–268. 

Charlton, P., & Allcorn, R. I. 2004. Habitat creation for Little Terns. In: Allcorn, R. I. (ed.). 

Proceedings of a Symposium on Little Terns Sterna albifrons: 32–38. RSPB Research Report No. 

8. Sandy, UK. 

Colombé, S. & Allcorn, R. I. 2004. An overview of seabird social facilitation projects in the UK, 

US and elsewhere with particular reference to the Sterna terns. In: Allcorn, R.I. (ed.). Proceedings 

of a Symposium on Little Terns Sterna albifrons: 60–65. RSPB Research Report No. 8. Sandy, UK. 

Cramp, S. (ed.) 1985. The Birds of the Western Palearctic. Vol. IV. Oxford University Press, Oxford. 

Davies, S. 1981. Development and behaviour of Little Tern chicks. British Birds 74: 291–298. 

Defra 2004. Delivering the essentials of life: Defra’s five year strategy. Department of 

Environment, Food and Rural Affairs report to Government. HMSO, Norwich. 

Erwin, R. M., Truitt, B. R. & Jimenez, J. E. 2001. Ground-nesting waterbirds and mammalian 

carnivores in the Virginia Barrier Islands region: running out of options. Journal of Coastal 

Research 17: 292–296. 

Fasola, M. & Bogliani, G. 1990. Foraging ranges of an assemblage of Mediterranean seabirds. 

Colonial Waterbirds 13: 72–74. 

Fasola, M. & Canova, L. 1992. Nest habitat selection by eight syntopic species of 

Mediterranean gulls and terns. Colonial Waterbirds 15: 169–178. 

Gill, J. 2004. Potential impacts of climate change on Little Terns in East Anglia. In: Allcorn, R. I. 

(ed.). Proceedings of a Symposium on Little Terns Sterna albifrons: 29–31. RSPB Research Report 

No. 8. Sandy, UK. 

Gochfeld, M. 1983. Colony site selection by Little Terns: physical attributes of sites. Colonial 

Waterbirds 6: 205–213. 

Goutner, V. 1990. Habitat selection of Little Terns in the Evros Delta, Greece. Colonial 

Waterbirds 13: 108–114. 

Gregory, R. D., Wilkinson, N. I., Noble, D. G., Robinson, J. A., Brown, A. F., Hughes, J., Proctor, 

D. A., Gibbons, D. W. & Galbraith, C. A. 2002. The population status of birds in the United 

Kingdom, Channel Islands and the Isle of Man: an analysis of conservation concern 2002–2007. 


Harrison, D. & Allcorn, R. I. 2004. Legislative framework affecting habitat creation in the UK. 

In: Allcorn, R. I. (ed.). Proceedings of a Symposium on Little Terns Sterna albifrons: 66–72. RSPB 

Research Report No. 8. Sandy, UK. 

Holloway, M. 1993. The variable breeding success of the Little Tern Sterna albifrons in South- 

East India and protective measured needed for its conservation. Biological Conservation 65: 1–8. 

Jeffries, D. S. & Brunton, D. H. 2001. Attracting endangered species to “safe” habitats: responses 

of Fairy Terns to decoys. Animal Conservation 4: 301–305. 

Kotliar, N. B. & Burger, J. 1984. The use of decoys to attract Least Terns (Sterna antillarum) to 

abandoned colony sites in New Jersey. Colonial Waterbirds 7: 134–138. 

SEABIRD 21 (2008): 55–63


Kotliar, N. B. & Burger, J. 1986. Colony site selection and abandonment by Least Terns Sterna 

antillarum in New Jersey, USA. Biological Conservation 37: 1–21. 

Krogh, M. G. & Schweitzer, S. H. 1999. Least Terns nesting on natural and artificial habitats in 

Georgia, USA. Waterbirds 22: 290–296. 

Mayfield, H. 1975. Suggestions for calculating nest success. Wilson Bulletin 87: 456–466. 

Norris, K. J. & Buisson, R. 1994. Sea-level rise and its impact upon coastal breeding birds in the 

UK. RSPB Conservation Review 8: 63–71. 

Parnell, J. F., Needham, R. N., Soots, R. F. Jr., Fussel, J. O. III, McCrimmon, D. A. Jr., Bork, R. D. 

& Shields, M. A. 1986. Use of dredged-material deposition sites by birds in coastal North 

Carolina, USA. Colonial Waterbirds 9: 210–217. 

Perrow, M. R., Tomlinson, M. L., Lines, P., Benham, K., Howe, R. & Skeate, E. R. 2004. Is food 

supply behind Little Tern Sterna albifrons colony location? The case of the largest colony in the 

UK at the North Denes/Winterton SPA in Norfolk. In: Allcorn, R. I. (ed.). Proceedings of a 

Symposium on Little Terns Sterna albifrons: 39–59. RSPB Research Report No. 8. Sandy, UK. 

Perrow, M. R., Skeate, E. R., Lines, P., Brown, D. & Tomlinson, M. L. 2006. Radio telemetry as a 

tool for impact assessment of wind farms: the case of Little Terns Sterna albifrons at Scroby 

Sands, Norfolk, UK. Ibis 148: S1 57-75. 

Pickerell, G. 2004. Little Tern Sterna albifrons. In: Mitchell, P. I., Newton, S., Ratcliffe, N. and 

Dunn, T.E. (eds.). Seabird populations of Britain and Ireland: 339–349. Poyser, London. 

Ratcliffe, N. 2004. Little Terns in Britain and Ireland: estimation and diagnosis of population 

trends. In: Allcorn, R. I. (ed.). Proceedings of a Symposium on Little Terns Sterna albifrons: 4–18. 

RSPB Research Report No. 8. Sandy, UK. 

Scarton, F., Valle, R. & Borella, S. 1994. Some comparative aspects of the breeding biology of 

Black-headed Gull (Larus ridibundus), Common Tern (Sterna hirundo) and Little Tern (Sterna 

albifrons) in the Lagoon of Venice, NE Italy. Avocetta 18: 199–123. 

Smart, J. 2004. Managing colonies for Little Terns. In: Allcorn, R. I. (ed.). Proceedings of a 

Symposium on Little Terns Sterna albifrons: 19–28 RSPB Research Report No. 8. Sandy, UK. 

Smart, J., Sutherland, W. J., Watkinson, A. R. & Gill, J. A. 2004. A new means of presenting the 

results of logistic regression. Bulletin of the Ecological Society of America 85: 100–102. 

Stroud, D. A, Chambers, D., Cook, S., Buxton, N., Fraser, B., Clement, P., Lewis, P., McLean, I., 

Baker, H. & Whitehead, S. 2001. The UK SPA network: its scope and content. Volume 2: species 

accounts. Joint Nature Conservation Committee, Peterborough, UK. 

Tapper, S. C. 1992. Game heritage: an ecological review from shooting & gamekeeping records. 

The Game Conservancy, Fordingbridge, UK. 

Thomas, B. C. and Slack, D. 1982. Physical aspects of colony selection by Least Terns on the 

Texas coast. Colonial Waterbirds 5: 161–168. 

Visser, J. M. & Peterson, G. W. 1994. Breeding populations & colony site dynamics of seabirds 

nesting in Louisiana. Colonial Waterbirds 17: 146–152. 

SEABIRD 21 (2008): 55–63 

63

64 

Subcutaneous air diverticula of Northern Gannet 

Descriptive anatomy of the subcutaneous 

air diverticula in the Northern Gannet 

Morus bassanus 

Daoust, P.-Y. 1 *, Dobbin, G. V. 1 , Ridlington Abbott, R. C. F. 2 & Dawson, S. D. 3 

*Correspondence author. Email: daoust@upei.ca 

1 Department of Pathology & Microbiology, Atlantic Veterinary College, University of 

Prince Edward Island, 550 University Avenue, Charlottetown, PE C1A 4P3, Canada 

(Current address for G. V. D: Department of Laboratory Medicine, Queen Elizabeth 

Hospital, P.O. Box 6600, Charlottetown, PE C1A 8T5, Canada); 2 Class of 2005, 

Department of Biology, Faculty of Science, University of Prince Edward Island, 550 

University Avenue, Charlottetown, PE C1A 4P3, Canada (Current address: 772 Osborne 

Street, Summerside, PE C1N 4N5, Canada); 3 Department of Biomedical Sciences, 

Atlantic Veterinary College, University of Prince Edward Island, 550 University Avenue, 

Charlottetown, PE C1A 4P3, Canada. 

Abstract 

Northern Gannets Morus bassanus typically forage by diving from high above the 

water surface. Their subcutaneous (s-c) tissues are invested by an elaborate system of 

air diverticula that presumably function in cushioning the impact of their entry into 

the water. The anatomical details of this system were studied by dissection and latex 

injection in 15 carcasses of these birds. The s-c air diverticula consist mainly of two 

independent systems of intercommunicating compartments that are bilaterally 

symmetrical, cover the ventral and lateral regions of the trunk and the proximal 

portions of the wings and legs, and communicate with the ipsilateral region of the 

clavicular respiratory air sac. This communication, which opens into the axillary region, 

is through a narrow gap between the subcoracoideus and coracobrachialis caudalis 

muscles. Two other, smaller, independent systems of s-c air diverticula, also bilaterally 

symmetrical, may contribute to cushioning the Northern Gannet’s body during its 

dives: one at the thoracic inlet, which communicates with the corresponding side of 

the clavicular air sac, and the other along the neck, which communicates with the 

nasal cavities and the choanal opening. Further work is required to define more 

precisely the function of these extensive air diverticula and air circulation within them. 

Introduction 

The avian respiratory system is the most efficient among those of all air-breathing 

vertebrates and is unique in its basic structure (King & McLelland 1984). Its extensive 

system of air sacs allows a near-continuous flow of fresh air through the pulmonary air 

capillaries at countercurrent to the blood circulation and throughout the respiratory 

Footnote: Definition of terms used in the text for anatomical orientation: cranial, toward the head; caudal, 

toward the tail; ventral, toward the front of the body; dorsal, toward the back of the body; proximal, closer 

to the centre of the body; distal, farther from the centre of the body; medial, closer to the body’s midline; 

lateral, farther from the body’s midline; rostral, toward the beak or tip of the beak. 

SEABIRD 21 (2008): 64–76


cycle. Most avian species have four paired air sacs (cervical, cranial thoracic, caudal 

thoracic, abdominal) and one unpaired air sac (clavicular). Depending on the species, 

some of these air sacs can project complex systems of diverticula between muscles and 

into the subcutis and pneumatic bones of the trunk, pectoral and pelvic girdles, and limbs 

(McLelland 1989; O’Connor 2004). Some members of the order Pelecaniformes have an 

elaborate and extensive system of subcutaneous (s-c) air diverticula. Northern Gannets 

Morus bassanus, which typically forage by diving from heights of up to 30 m above water 

and reaching speeds of up to 100 km/h on impact with water, are thought to use these 

s-c air diverticula as a means of cushioning this impact (Montagu 1813; Gurney 1913; 

Nelson 1978). It is not known, however, whether these diverticula are inflated voluntarily 

prior to diving or whether air is simply prevented from exiting them as the bird hits the 

water. Regardless, an efficient communication is likely needed between the respiratory 

tract and the system of s-c air diverticula and among the various compartments of this 

system. Subcutaneous air diverticula were described, albeit only partially, many years ago 

in the Northern Gannet (Montagu 1813; Gurney 1913) and in the Brown Pelican 

Pelecanus occidentalis (Richardson 1939). According to the study of the Brown Pelican 

by Richardson (1939), the communication between the respiratory system, specifically 

the clavicular air sac, and the s-c air diverticula is located caudolaterally to the head of 

the coracoid bone and below the head of the humerus, ‘primarily between the M 

[muscle] coracobrachialis posterior and the M subcorachoideus’. Similarly, in his study of 

the Northern Gannet, Gurney (1913), quoting C. B. Ticehurst, states that the s-c air 

diverticula communicate with the respiratory system by way of a passage just outside 

the coracoid bone and close to the tendon of the ‘pectoralis minor muscle’ (‘M coracobrachialis 

posterior’, according to Richardson (1939)). These authors also briefly describe 

the distribution of the s-c air diverticula along the ventral region of the trunk and down 

the thighs and wings and the separation of these diverticula between left and right sides 

of the body. The description of s-c air diverticula that they offer is, however, insufficient 

to fully understand the exact pattern of air flow among their various compartments. 

The objective of this study was to provide a more detailed description, complemented 

by photographs, of the anatomy of the s-c air diverticula in the Northern Gannet than 

is currently available in the literature. More specifically, we describe the distribution of 

s-c air diverticula along the body and the communication between the system of s-c 

air diverticula and the respiratory system in this species. We also hypothesise that the 

wings’ position may alter this communication as it changes from extended away from 

the body while flying and soaring to flexed against the body when diving. More specifically, 

we predict that wing flexion against the body closes the communication 

between the two systems, thus preventing air from escaping the s-c air diverticula, 

thus ensuring a firm cushion on impact. 

Materials and Methods 

Fifteen carcasses of Northern Gannets in a good state of preservation (ten adult, one 

immature and three full-grown hatch-year based on their plumage, one of 

undetermined age; six male, four female, five of undetermined sex) were dissected in 

the course of this study. The carcasses were of wild birds that had drowned in fishing 

nets, had died of emaciation/starvation, or had been euthanized because of a broken 

SEABIRD 21 (2008): 64–76 

65

66 


limb. Two carcasses were refrigerated until used a few days later, whereas the 13 other 

carcasses were frozen at minus 20°C for a period varying between three weeks and 19 

months (average, six months) prior to use. In 13 carcasses, a solution of either red or 

blue latex (Carolina Biological Supply, Burlington, NC, USA) was injected in various 

locations in order to make casts of the different cavities under study, i.e. respiratory air 

sacs and/or s-c air diverticula. Injection sites included: intrachoanal; intratracheal (via 

a small incision of the skin and tracheal wall in the mid-cervical region); and 

perihumeral, axillary, ventral, subclavicular and cervical regions of s-c air diverticula 

(via small skin incisions). In five instances, intratracheal injection was carried out while 

one wing was extended away from the body and the other flexed (folded) against the 

body. In each of these 13 cases, the carcass, if frozen, was completely thawed (over a 

period of 48h); latex was injected into the selected location, being allowed to settle 

strictly by gravity; and the carcass was frozen again for c.48h, in order to promote 

polymerisation of the latex solution (Tompsett 1970), and subsequently thawed for 

dissection. The amount of latex injected varied among the locations selected and was 

largest when latex was injected intratracheally, in which case up to 550 ml were used 

in order to adequately fill some of the respiratory air sacs. Within 48h following the 

start of dissection, the carcass (minus abdominal viscera) was immersed in a solution 

of 10% formalin in order to prevent decomposition. It was possible to make casts into 

two different locations in the same bird by using a latex solution of one colour, 

freezing the carcass for c.48h and thawing it in order to inject latex of the other colour. 

Two carcasses were dissected without prior latex injection in order to better examine 

the sites of origin and insertion of muscle masses particularly relevant to the anatomy 

of the s-c air diverticula. George & Berger (1966), Vanden Berge (1975), and Nickel et 

al. (1977) were consulted for muscle identification and terminology. 

Results 

Anatomical observations on the 15 Northern Gannets used in this study were 

consistent among all birds, with one exception pertaining to the possible role of wing 

flexion in air circulation (see below). According to these observations, the s-c air 

diverticula of this species consist mainly of two independent systems of intercommunicating 

compartments that extend from the respiratory tract, are bilaterally 

symmetrical along the cranio-caudal midline of the trunk and cover its ventral and 

lateral regions. Latex injected into the trachea easily fills most of the volume of the 

respiratory air sacs and their diverticula within the thoracic cavity. Latex further flows 

from what are interpreted as the left and right ventrolateral regions of the clavicular 

air sac into the axillary regions of the ipsilateral s-c air diverticula through a narrow 

gap between the subcoracoideus and coracobrachialis caudalis muscles (Figures 1 & 2). 

Both muscles are located immediately caudal to the coracoid bone. The subcoracoideus 

muscle originates on the inner (ventral) surface of the cranial region of the 

scapula and on the medial surface of the cranial region of the coracoid bone and 

inserts on the medial surface of the proximal humerus. The coracobrachialis caudalis 

muscle originates on the caudal region of the coracoid bone, primarily its lateral 

surface but with some fibres originating on its dorsal and ventral surfaces, and it 

inserts on the medial surface of the proximal humerus, proximal to the pneumatic 

foramen and distal to the insertion of the subcoracoideus muscle. 

SEABIRD 21 (2008): 64–76


Figure 1. Right lateral view of the thoracic cavity 

of a Northern Gannet Morus bassanus. The head 

is directed to the right side of the Figure. Red 

latex injected via the trachea filled the clavicular 

air sac (CAS) and extended (a) into the 

pneumatic foramen of the right humerus (H) 

and (b) between the coracobrachialis caudalis 

muscle (CBC) and subcoracoideus muscle (SC) 

into the axillary region of the right subcutaneous 

(s-c) air diverticulum. (Blue latex had previously 

been injected into both axillary regions, and 

small amounts of it had reached the clavicular 

air sac.) Co, caudal end of coracoid bone 

detached from its insertion onto the sternum. 

Latex emerges from the clavicular air sac into the left and right axillary regions, located 

laterally between the sternum and corresponding pectoralis muscle (Figure 2). In three of 

five instances in which intratracheal injection was carried out while one wing was flexed 

and the other extended, flow of latex into the axillary region occurred on the extended 

side, but not on the flexed side. Latex also flows from the clavicular air sac into the 

pneumatic foramen of the right and left humeri through a separate communication 

located immediately caudal to the corresponding coracobrachialis caudalis muscle (Figure 

1). This flow of latex into the humerus occurred whether or not the wing was flexed. 

From the axillary region, the air diverticulum extends ventrally between the sternum 

and the pectoralis muscle and caudally along the lateral side of the trunk and along 

the medial side of the leg down to the distal region of the tibia (Figure 2). No 

communication was found between the region of the s-c air diverticulum along the leg 

and the ipsilateral abdominal air sac. 

Immediately caudal to the axilla, the diverticulum originating from the axillary region 

also extends dorsally and then cranially, dorsal to the scapulohumeralis muscle and 

ventral to (underneath) the latissimus dorsi muscle (Figure 2). The diverticulum 

emerges subcutaneously dorsal to the shoulder, curves around the cranial region of the 

shoulder, and opens caudally into a large compartment that covers the whole ventral 

surface of the pectoralis muscle and extends slightly caudal to it (Figure 3). The lateral 

region of this ventral compartment is subdivided into approximately six pockets that 

communicate widely with each other ventrally and are formed by thin transparent 

membranous partitions extending about 3–4 cm from the lateral wall of the 

compartment and attached to the skin along their outer border and to the pectoralis 

muscle along their inner border. The lateral wall of this ventral compartment, also 

consisting of a thin transparent membrane, separates it from the portion of the air 

diverticulum originating from the axillary region; this lateral wall extends along the 

SEABIRD 21 (2008): 64–76 

67

68 


lateral side of the trunk and down the medial side of the leg. The ventral compartment 

is separated from the contralateral side by a thin transparent membranous partition 

which is continuous along the ventral midline from the cranial extremity of the keel 

to the vent and is attached to the skin along its outer border and to the keel and, more 

caudally, the abdominal muscle wall along its inner border. Thin bands of fibrous tissue 

and small blood vessels and nerves course through these various membranous 

partitions and may thus reinforce them (Figure 3). 

Figure 2. Right lateral view of the trunk of the skinned carcass of a Northern Gannet Morus bassanus. The head is 

directed to the right side of the Figure. The arrows show the communication among compartments of the right s-c 

air diverticulum, starting in the axillary region with its origin from the (intrathoracic) clavicular air sac between the 

subcoracoid muscle (SC) and coracobrachialis caudalis muscle (CBC) and spreading between the sternum (St) and 

ribs medially and the pectoralis muscle (P, reflected away from the sternum) laterally (1).The diverticulum continues 

caudally along the trunk and the medial side of the leg. Caudal to the axillary region, it passes dorsally and then 

cranially between the scapulohumeralis muscle (SH) and latissimus dorsi muscle (LD) (2) to emerge subcutaneously 

at the level of the shoulder (3). From underneath the pectoralis muscle, the diverticulum also passes cranially and 

then dorsally over the proximal end of the coracoid bone (Co) (4), emerges underneath the tensor propatagialis 

muscle (TP) to extend along muscles of the proximal portion of the wing (5), and also joins the compartment that 

emerges subcutaneously from underneath the latissimus dorsi muscle (6) (see Figures 4 & 5). The diverticulum then 

proceeds ventrally and caudally between the pectoralis muscle and skin (7) (see Figure 3). In addition, the 

diverticulum extends from the axillary region along the humerus between the muscle biceps brachii (BB) ventrally 

and the muscle triceps brachii dorsally (8) (see Figure 6). A third compartment of the diverticulum extending along 

the wing, besides (5) and (8), originates from the portion of the diverticulum as it emerges from underneath the 

latissimus dorsi muscle (see Figure 5) (hidden from view in this Figure). Cl, clavicle; F, femur; H, humerus. 

SEABIRD 21 (2008): 64–76


Figure 3. Ventral view of the trunk of a Northern Gannet Morus bassanus. The head is directed to the right side 

of the Figure. The skin has been partly reflected, revealing the large ventral compartment of the right s-c air 

diverticulum. The lateral region of this compartment is partly subdivided into individual pockets by thin 

transparent membranous partitions. Its lateral wall (LatW), also consisting of thin transparent tissue, separates it 

from the axillary region of the diverticulum, whereas a median wall (MedW) separates it from the ventral 

compartment of the left diverticulum. Thin bands of fibrous tissue and small blood vessels and nerves course 

through these various membranous partitions (thick arrows). The long arrow on the right shows the wide 

communication over the shoulder between this ventral compartment of the diverticulum and its dorso-lateral 

region (see Figure 1, point 7). P, pectoralis muscle. 

From the axillary region, and in addition to extending ventrally and caudally between 

the sternum and pectoralis muscle, the diverticulum also extends through a narrow 

canal dorsally over the cranial end of the coracoid bone and spreads distally 

underneath the tensor propatagialis muscle (the origin of which is partly on the cranial 

end of the coracoid bone) toward the distal end of the humerus (Figures 4 & 5). From 

underneath the tensor propatagialis muscle, this portion of the diverticulum is 

continuous via a narrow opening with the portion of the diverticulum that emerges 

subcutaneously from underneath the latissimus dorsi muscle (Figure 5). 

Air diverticula extending along muscles of the proximal portion of the wing are 

supplied from at least three sources. One, just described, originates from underneath 

the tensor propatagialis muscle (Figures 4 & 5). Another, which proceeds distally from 

underneath the deltoideus major muscle, originates from the portion of the 

diverticulum that travels beneath, and emerges subcutaneously cranial to, the 

latissimus dorsi muscle (Figure 5). A third source originates directly from the axillary 

SEABIRD 21 (2008): 64–76 

69

70 


Figure 4. Right shoulder of a Northern Gannet Morus bassanus in lateral view. The head is directed to the right side 

of the Figure. Blue latex injected subcutaneously along the distal region of the right humerus spread underneath 

the tensor propatagialis muscle (TP) and, through a narrow canal running dorsally around the cranial end of the 

coracoid bone (Co), reached the axillary region of the s-c air diverticulum, located between sternum and pectoralis 

muscle (removed). Cl, clavicle; H, proximal end of humerus. 

SEABIRD 21 (2008): 64–76


Figure 5. Right shoulder of a Northern Gannet Morus bassanus in lateral view. The head is directed to the right side 

of the Figure. Blue latex was injected into the axillary region, between sternum and pectoralis muscle (P). After 

having travelled dorsally around the cranial end of the coracoid bone, the latex solution extended underneath the 

tensor propatagialis muscle (TP) and further distally along the humerus (H). It also partly filled, through a narrow 

opening (a), the portion of the s-c air diverticulum that would normally emerge subcutaneously from underneath 

the latissimus dorsi muscle (LD, cranial and caudal heads). Some of the latex in the latter portion also extended 

underneath the deltoideus major muscle (DM) (b) distally along muscles of the proximal region of the wing. 

region and proceeds between the biceps brachii muscle ventrally and the triceps 

brachii muscle dorsally (Figure 6). 

In addition to the large systems of air diverticula covering the ventro-lateral region of the 

trunk, two other, small, bilaterally symmetrical, independent systems may contribute to 

cushioning the gannet’s body during dives: one at the thoracic inlet and another along 

the neck. Air diverticula located at the thoracic inlet on either side of the cranio-caudal 

midline lie cranial to the main compartment of the respiratory clavicular air sac and 

internal and cranial to the corresponding clavicle (Figure 7). Each of these subclavicular 

air diverticula communicates with the corresponding side of the clavicular air sac via a 

tubular channel along the lateral wall of the thoracic cavity. Large s-c air diverticula, each 

possibly composed of a number of interconnecting compartments, lie ventrally along 

either side of the neck (Figure 8). Each of these two cervical s-c air diverticula 

communicates with the nasal cavities and the choanal opening through a small aperture 

along the roof of the pharynx. No communication was found between these cervical sc 

air diverticula and the ipsilateral cervical respiratory air sacs. 

SEABIRD 21 (2008): 64–76 

71

72 


Figure 6. Ventral view of the right axilla and 

humerus of a Northern Gannet Morus bassanus. 

The head is directed to the right side of the 

Figure. Blue latex injected via the trachea 

extended into the axillary region and, from there, 

along the humerus between the biceps brachii 

muscle (BB) ventrally and the triceps brachii 

muscle dorsally. St, sternum, from which the 

pectoral muscle has been removed. 

Figure 7. Ventral view of the thoracic girdle of 

a Northern Gannet Morus bassanus. The head is 

directed to the right side of the Figure. Red 

latex injected via the trachea (T) has filled the 

left and right subclavicular air diverticula (stars) 

via their communication with the clavicular air 

sac. (Blue latex had previously been injected 

into both axillary regions, and small amounts of 

it had reached the clavicular air sac and, 

subsequently, both subclavicular air 

diverticula.) Cl, clavicle; K, keel of the sternum; 

P, left and right pectoralis muscles. 

SEABIRD 21 (2008): 64–76


Figure 8. Ventral view of the neck and head of a Northern Gannet Morus bassanus. The head is to the right; the 

rostral portion of the upper beak was cut off. The mandible and caudal region of the palate (P) were removed in 

order to exposed the nasal cavities. Red latex injected directly into the right cervical s-c air diverticulum extends 

through a small opening (arrow) into the right nasal cavity. Q, quadrate bone, which articulates with the mandible. 

Discussion 

The results of this study confirm and expand upon earlier observations of the anatomy 

of s-c air diverticula in the Northern Gannet.Assuming that air flow among the various 

compartments of these diverticula can be inferred from that of latex, their distribution, 

more specifically their voluminous size along the ventral surface of the neck and trunk 

and their paucity along the back, supports their putative function in cushioning the 

impact of entry into the water as the bird dives from high above the water surface.We 

had originally hypothesised that flexion of the wings against the body at the start of 

a dive could close the communication between respiratory air sacs and s-c air 

diverticula of the trunk, thus allowing the air trapped in these diverticula to provide an 

efficient cushion against the impact of entry into the water. This was demonstrated in 

three of five instances. It is possible that, in the two birds in which latex could flow 

from the clavicular air sac into the axillary region of the s-c air diverticulum on the 

side of the flexed wing, incomplete flexion of the wing against the body and/or 

postmortem stiffness of the tissues could have prevented complete closure of the 

communication. Alternatively, during a dive, a gannet lays its wings back, flat against 

the body but fully extended rather than flexed (Elphick et al. 2001), and this position 

might better close the communication between respiratory air sacs and s-c air 

diverticula of the trunk. The hypothesis proposed above therefore requires further 

testing. More cranially, escape of air from inflated cervical air diverticula (via the nasal 

cavities) and inflated subclavicular air diverticula (via the clavicular air sac and into the 

lungs and trachea) could be prevented by complete closure of the Northern Gannet’s 

mouth, as the nostrils are permanently closed by an overgrowth of epithelial cells in 

this species (King & McLelland 1984). 

SEABIRD 21 (2008): 64–76 

73

74 


Our localisation of the communication between respiratory air sacs and s-c air 

diverticula of the trunk is as described in general terms in the Northern Gannet by 

Gurney (1913) and more precisely in the Brown Pelican by Richardson (1939), namely 

between the subcoracoideus and coracobrachialis caudalis muscles. Richardson (1939) 

assumed, based on the work of others, that this communication comes from the 

clavicular air sac and ‘probably corresponds to the axillary diverticulum’ of this air sac. 

We make a comparable assumption, i.e. that the s-c air diverticulum along the trunk 

on each side of the cranio-caudal midline represents a massive extension of an 

extrathoracic diverticulum originating from the ipsilateral (paired) lateral chamber of 

the clavicular air sac as described by King (1975) and McLelland (1989).The only other 

possible alternative would be a communication between the s-c air diverticula and the 

cranial thoracic air sacs. These air sacs were not visualised in this study, as it was not 

our intent to provide a detailed description of the system of respiratory air sacs in the 

Northern Gannet. However, whereas the clavicular air sac typically has several intraand 

extra-thoracic diverticula, including a humeral diverticulum aerating the humerus 

in some species (as was found in our birds), the cranial thoracic air sacs are not known 

to have diverticula in any species (King 1975; Nickel et al. 1977; McLelland 1989), 

although they may aerate some bones (sternal ribs, sternum) in some species (e.g. 

Psittaciforms) (Evans 1996). King (1975) describes three diverticula of the lateral 

chamber of the clavicular air sac in the chicken (pectoral, humeral, and axillary), 

whereas McLelland (1989), citing Groebbels (1932), describes in general five such 

diverticula (subscapular, axillary, subpectoral, suprahumeral, and a fifth diverticulum 

under the latissimus dorsi muscle) but adds that ‘considerable interspecific variation in 

development of the diverticula appears to exist’. We did not attempt to ascribe the sc 

air diverticula along the trunk of the Northern Gannet to an extension of any one or 

more of the diverticula mentioned above. Bezuidenhout et al. (1999) describe 

diverticula originating from the abdominal air sacs and extending among muscles and 

under the subcutis of the legs in the Ostrich Struthio camelus. No such communication 

was found between the region of the s-c air diverticula along the legs of Northern 

Gannets and the ipsilateral abdominal air sacs. 

The location of what we describe as the subclavicular air diverticula in the Northern 

Gannet corresponds to that of the craniolateral diverticulum of the median chamber of 

the clavicular air sac as described in the chicken by King (1975) and in general by 

McLelland (1989) and of the left and right cranial parts of the clavicular air sac as 

described in the chicken by Nickel et al. (1977). What we describe as cervical s-c air 

diverticula represent a greatly expanded cervicocephalic system of air sacs, as there was 

a clear communication between these diverticula and the nasal cavities. According to 

Richardson (1939), there is in pelicans ‘sometimes a connection between the air cavities 

of the pharyngonasal system of the head region and the pulmonary cavities of the neck’, 

but no further description is provided. Walsh & Mays (1984) described in psittacine birds 

a s-c cervicocephalic air sac with a cephalic portion situated caudodorsally to the skull and 

a cervical portion extending bilaterally dorsolaterally along the neck; this sac 

communicates with the infraorbital sinus cranially, but not with the respiratory air sacs. 

According to McLelland (1989), the cervical respiratory air sacs make ‘an especially large 

contribution to the subcutaneous diverticula [along the ventral surface of the neck] in the 

SEABIRD 21 (2008): 64–76


Gannet’. This author added, however, that ‘extensions of the vertebral diverticula [of the 

cervical respiratory air sacs] penetrate between the neck diverticula of the cervicocephalic 

system of air sacs’, suggesting that two independent systems of cervical s-c air diverticula 

are present in this species.We were unable in this study to find a distinct communication 

between the cervical s-c air diverticula and the cervical respiratory air sacs. 

Several unanswered questions remain regarding the function of s-c air diverticula and air 

circulation within them. Within the six families of the order Pelecaniformes, all sulids 

(boobies and gannets) forage by plunge-diving, but among the other five families, only 

Brown Pelicans and Neotropic Cormorants Phalacrocorax olivaceous habitually do so (del 

Hoyo et al. 1992). We are not aware of a detailed description of s-c air diverticula in 

species of the order Pelecaniformes that do not plunge-dive. Such information could help 

to elucidate the evolution of s-c air diverticula in plunge-divers and might offer 

alternative explanations for their function. Bignon (1889) suggested a number of 

possible functions for the cervicocephalic air diverticula in birds besides that of cushions, 

including heat retention, buoyancy control, and head support during flight. Control 

mechanisms for inflation and deflation of the different s-c air diverticula are also 

incompletely understood. Although s-c air diverticula of the trunk and the subclavicular 

s-c air diverticula can be inflated voluntarily via the respiratory tract, it is less clear 

whether or how a bird can voluntarily inflate the cervical s-c air diverticula via the 

choanal opening. Owen (1866) suggested that birds (pelicans and gannets) may 

voluntarily expulse air from s-c air diverticula ‘when the bird is about to descend, in order 

to increase its specific gravity, and enable it to dart with rapidity upon a living prey’; he 

added that this can be done through the action of those muscles that are connected to 

the skin by membranous partitions and bands of fibrous tissue, blood vessels and nerves. 

In conclusion, the subcutis of the ventral region of the trunk and neck of the Northern 

Gannet is invested by an elaborate air mattress that may be related to the particular 

foraging behaviour of this species. Further work is required, however, to describe in 

more detail the numerous peripheral extensions of this system of s-c air diverticula 

and the potential anatomic variation in these extensions among individual birds and, 

particularly, to define more precisely its functional adaptation. 

Figure 9. Northern Gannet Morus bassanus, Quendale Bay, Shetland, 2 September 2007 © Hugh Harrop. 

SEABIRD 21 (2008): 64–76 

75

76 



We thank Shelley Ebbett, Mike Needham, and Fiep de Bie for photography of the 

specimens and preparation of the figures. We also thank three anonymous reviewers 

for their excellent comments. 

References 

Bezuidenhout,A. J., Groenewald, H. B., & Soley, J.T. 1999. An anatomical study of the respiratory 

air sacs in ostriches. Onderstepoort Journal of Veterinary Research 66: 317–325. 

Bignon, F. 1889. Contribution to the study of pneumacity of birds. The cervicocephalic air cells 

of the birds and their relation with the bones of the head. Mémoires Société zoologique de 

France 2: 260–318. 

del Hoyo, J., Elliott, A. & Sargatal, J. (eds.). 1992. Handbook of the Birds of the World. Vol. 1. 

Lynx Edicions, Barcelona. 

Elphick, C., Dunning Jr, J. B. & Sibley, D. A. 2001. The Sibley Guide to Bird Life & Behavior. 

National Audubon Society. Alfred A. Knopf Inc., New York. 

Evans, H. E. 1996. Anatomy of the budgerigar and other birds. In: Rosskopf Jr, W. J. & Woerpel, 

R. W. (eds.). Diseases of Cage and Aviary Birds: 79–162. Williams & Wilkins, Baltimore. 3rd edn. 

George, J. C. & Berger, A. J. 1966. Avian Myology. Academic Press, New York. 

Groebbels, F. 1932. Der Vögel. Bau, Funktion, Lebenserscheinung, Einpassung.Vol. 1. Borntraeger, Berlin. 

Gurney, J. H. 1913. The Gannet, a bird with a history. Witherby, London. 

King, A. S. 1975. Aves Respiratory System. In: Getty, R. (ed.). Sisson and Grossman’s The Anatomy 

of the Domestic Animals. Vol. 2: 1883–1918. W. B. Saunders Co., Philadelphia. 5th edn. 

King, A. S. & McLelland, J. 1984. Birds. Their structure and function: 110–144. Baillière Tindall, 

Toronto. 2nd edn. 

McLelland, J. 1989. Anatomy of the lungs and air sacs. In: King, A. S. & McLelland, J. (eds.). Form 

and Function in Birds. Vol. 4: 221–279. Academic Press, Toronto. 

Montagu, G. 1813. First description of the gannet’s subcutaneous air sacs. Supplement to the 

Ornithological Dictionary. 

Nelson, B. 1978. The Gannet. Buteo Books, Vermilion. 

Nickel, R., Schummer, A. & Seiferie, E. 1977. The Anatomy of the Domestic Birds. Translated by 

W. G. Siller & P. A. L. Wight. Springer-Verlag, New York. 

O’Connor, P. M. 2004. Pulmonary pneumaticity in the postcranial skeleton of extant Aves: a 

case study examining Anseriformes. Journal of Morphology 261: 141–161. 

Owen, R. 1866. Anatomy of Vertebrates.Vol. II. Birds and Mammals. Longmans, Green and Co., London. 

Richardson, F. 1939. Functional aspects of the pneumatic system of the California Brown 

Pelican. The Condor 41: 13–17. 

Tompsett, D. H. 1970. Anatomical Techniques: 259–265. E. & S. Livingstone, London. 2nd edn. 

Vanden Berge, J. C. 1975. Aves Myology. In: Getty, R. (ed.). Sisson and Grossman’s The Anatomy 

of the Domestic Animals. Vol. 2: 1802–1848. W. B. Saunders Co., Philadelphia. 5th edn. 

Walsh, M. T. & Mays, M. C. 1984. Clinical manifestations of cervicocephalic air sacs of 

psittacines. Compendium of Continuing Education for the Practicing Veterinarian 6: 783–789. 

SEABIRD 21 (2008): 64–76

Population decline of Leach’s Storm-petrel on St Kilda 

Population decline of Leach’s Storm-petrel 

Oceanodroma leucorhoa within the largest 

colony in Britain and Ireland 

Newson, S. E. 1 *, Mitchell, P. I. 2 , Parsons, M. 2 , O’Brien, S. H. 2 , Austin, G. E. 1 , Benn S. 3 , 

Black J. 4 , Blackburn, J. 1 , Brodie, B. 3 , Humphreys, E. 5 , Leech, D. 1 , Prior, M. 6 & Webster, M. 7 

*Correspondence author. Email: stuart.newson@bto.org 

1 British Trust for Ornithology, The Nunnery, Thetford, Norfolk IP24 2PU, UK; 2 Joint 

Nature Conservation Committee, Dunnet House, 7 Thistle Place, Aberdeen AB10 1UZ, 

UK; 3 19 Culloden Court, Inverness IV2 7DX, UK; 4 3 Newgate Street, Bingham, 

Nottingham NG13 8FD, UK; 5 BTO Scotland, School of Biological and Environmental 

Sciences, Cottrell Building, University of Stirling FK9 4LA, UK; 6 1 Rother Close, 

Greenmeadow, Swindon, Wiltshire SN25 3PZ, UK; 7 16 Grice Close, Sheringham, Norfolk 

NR26 8UG, UK. 

Abstract 

This study used diurnal playback of vocalisations to examine the abundance of 

breeding Leach’s Storm-petrel Oceanodroma leucorhoa on Dun, St Kilda in 2003 and 

2006 in relation to the only previous survey conducted using similar methodology in 

1999. The number of Apparently Occupied Sites in 2006 was 12,770, not significantly 

different to the 14,490 found in 2003, but significantly lower than the 27,811 found 

in 1999, by 54%. The magnitude and rate of the decline are of major conservation 

concern. Great Skua Stercorarius skua predation is thought the most likely cause but 

other factors such as poor food supply cannot be ruled out. The importance of 

continued monitoring of Leach’s Storm-petrel and Great Skua is discussed. 

Introduction 

The Leach’s Storm-petrel Oceanodroma leucorhoa has a highly localised distribution 

in the east Atlantic, with breeding confirmed on only about 13 remote islands and 

archipelagos off the coasts of the Republic of Ireland, Scotland, Faeroes, Iceland and 

Norway (Mitchell 2004). One of the largest of these is on the St Kilda archipelago off 

northwest Scotland, probably second only in size to the colony on the Westmann 

Islands off southern Iceland that is estimated to hold 80,000–150,000 pairs (Icelandic 

Institute of Natural History 2000). During 1999, the Seabird 2000 survey (Mitchell 

2004) estimated that St Kilda held about 94% of the British and Irish population 

(45,433 Apparently Occupied Sites, AOS) of which 27,811 were in a single subcolony 

on the island of Dun. The lack of accurate population estimates for this species in 

Britain and Ireland before this time makes it impossible to assess whether there has 

been a significant long-term change in the status of this species. 

On Dun and other islands of St Kilda, predation by Great Skuas Stercorarius skua is 

thought to pose a serious threat to Leach’s Storm-petrel (Phillips et al. 1997, 1999a; 

Votier et al. 2006). Great Skuas have increased dramatically on St Kilda, with the 

SEABIRD 21 (2008): 77–84 

77

78 


population rising from about 42 pairs in 1986 to a peak of 240 pairs in 2000 (Furness & 

Ratcliffe 2004). Based on their numbers in 1996, it has been estimated that Great Skuas 

predate approximately 14,850 Leach’s Storm-petrel on St Kilda each year (Phillips et al. 

1999b). Although there has been a small decline in the proportion of Leach’s Stormpetrel 

in the diet of Great Skua since 1996, the level of predation appears to have 

remained high (Votier et al. 2006). With declines in the availability of offal and discards 

from commercial fishing boats predicted due to changes in fisheries management, Great 

Skuas may increase their reliance on seabird prey (Votier et al. 2004). Because of the 

importance of Dun for Leach’s Storm-petrel in Britain and Ireland, there was a need to 

provide reliable monitoring data for this population. Here we report on two subsequent 

surveys of Leach’s Storm-petrel on Dun conducted in 2003 and 2006, and examine 

whether the size of the population has changed since the first survey in 1999. 

Methods 

Survey methods: A playback survey was carried out on Dun, St Kilda (57°47’N, 

08°33’W; Figure 1) on 4–5 and 7–10 July 2003 and 17–20 June 2006 using the 

methods developed by Ratcliffe et al. (1998) that were used previously on Dun in 1999 

and throughout the rest of Britain and Ireland during Seabird 2000 (Mitchell 2004). 

Both surveys were conducted between 09.00–18.30 BST, since response rate is known 

to increase significantly towards dusk (Mitchell 2004). Leach’s Storm-petrel burrows on 

Dun are virtually invisible, hidden in a thick sward on unconsolidated ground, which is 

broken by boulders and by Atlantic Puffin Fratercula arctica burrows. Transect lines 

running the width of the island (southwest to northeast; Figure 1) were positioned at 

10 m intervals along the northwest side of the island and at larger 25 m intervals on 

the southeast side to minimise disturbance of breeding Atlantic Puffins in this part of 

the island. A Global Positioning System (GPS) was used to ensure that the positions of 

transects were similar (± 5 m) during the 2003 and 2006 surveys. At randomly selected 

points along these transects, a total of 331 and 303 5 m x 5 m quadrats were laid out 

in 2003 and 2006 respectively, which for each year is about 5% of the total area of Dun 

(147,396 m 2 ). Each quadrat was then surveyed by a single observer playing a recorded 

call on a dictaphone at approximately 1 m intervals and the number of responses was 

recorded. The tape playback used a male Leach’s Storm-petrel chatter call recorded on 

St Kilda. The same recording was used in 1999, 2003 and 2006. 

Data analysis: As only a proportion of birds will respond to tape playback, it is 

necessary to apply a response rate correction to any estimate. Such a correction factor 

is normally calculated by repeatedly sampling a calibration plot (Ratcliffe et al. 1998; 

Mayhew et al. 2000). On the first day of the survey in 2003, a calibration plot was 

established on the northwest end of Dun (chosen to avoid disturbance to the dense 

Atlantic Puffin colony on the southeast end). The plot yielded 25 responses from 

Leach’s Storm-petrel on the first visit. The plot was subsequently visited on each day 

of the survey (except on 9 July, when deteriorating sea conditions dictated a 

premature departure from the island). Tape playback in the calibration plot was made 

at various times of day, but always within the same time period as the survey. 

Unfortunately, during the 2006 survey deteriorating sea conditions again necessitated 

a premature departure from the island and calibration data for 2006 were too limited 

SEABIRD 21 (2008): 77–84


Figure 1. Map showing the location of Dun, St Kilda and surveyed quadrats. 

to determine a correction. Therefore, we applied the same correction factor estimated 

from the calibration plot in 2003 to the survey data from 2006. 

The population size (number of Apparently Occupied Sites, AOS t), of Leach’s Stormpetrel 

on Dun for each year was estimated as: 

AÔS t =∑S AOS s + S AOS s 

x (A t – A q x N) 

(N x A q) 

Where AOSs is the number of birds counted in each surveyed quadrat, N is the number 

of quadrats surveyed, At is the total area of Dun and Aq is the area of a quadrat. 

To produce confidence intervals we used a bootstrap resampling procedure of 10,000 

iterations (Crowley 1992). For each iteration, AOS s were randomly resampled with 

replacement, the sample drawn equal to the number of quadrats surveyed. The 

response rate correction factor was included within the above analyses, by dividing 

SAOS s across quadrats from each bootstrap iteration by the mean response rate. The 

errors around the response rate were incorporated by specifying a mean and standard 

deviation for each value and randomly selecting response rate values from a binomial 

SEABIRD 21 (2008): 77–84 

79

Cumulative no. responses 

70 

60 

50 

40 

30 

20 

10 

80 


distribution defined by these values and constrained to remain between 0 and 1. By 

doing this 10,000 estimates of population size were produced. The 250th and 9,750th 

ordered bootstrap values across iterations were taken to give the lower and upper 95% 

confidence limits of the estimates respectively. Bootstrap t-tests were used to test for 

a significant difference in population estimates between years (Manly 1998). 

Results 

Calibration: The results of tape playback at the calibration plot on Dun in 2003 gave 

a mean response rate of 0.386 (Table 1 & Figure 2), very similar to that obtained in 

2000 on Boreray, St Kilda (0.382, 95% CL 0.338–9.422) and applied to the 1999 

census of Dun (Mitchell 2004). 

Population size: As in 1999, there was no significant difference in 2003 between the 

mean number of responses per quadrat obtained on the northwest side of the island 

(0.945, 95% CL 0.776–1.132) compared to the southeast side (0.955, 95% CL 

0 

0 

0 2 4 6 8 10 12 14 16 18 20 

1999 2000 2001 2002 2003 2004 2005 2006 

Days after first visit 

Year 

Figure 2 (left). Plot of cumulative number of responses to tape playback against days after first visit at the 

calibration plot on Dun, 4–10 July 2003. The solid line represents line of best fit from the observed responses 

(diamonds) and the broken lines represent the lower and upper 95% confidence limits of the line. Figure 3 (right). 

Estimated population size (AOS with 95% confidence intervals) of Leach’s Storm-petrel Oceanodroma leucorhoa 

on Dun, St Kilda in 2003 and 2006 in relation to a previous survey carried out in 1999. 

Table 1. Response rate of Leach’s Storm-petrel Oceanodroma leucorhoa to tape playback, estimated from the 

calibration plot on Dun, 4–10 July 2003. 

Date in July 4 5 6 7 8 10 

No. responses 25 25 29 25 25 17 

New responses 25 9 8 10 6 1 

Cumulative no. responses 25 34 42 52 58 59 

Response rate 1 0.397 0.397 0.460 0.397 0.397 0.270 0.386 MEAN 

0.336 LCL 

0.436 UCL 

1 The plot of the cumulative number of responses against days after first visit (Figure 2) gave an asymptote of 

63, which was taken to be the total number of birds present in the calibration plot. 

SEABIRD 21 (2008): 77–84 

AOS 

40000 

35000 

30000 

25000 

20000 

15000 

10000 

5000


0.754–1.159, t 329 = 0.04, n.s.). Thus the data from both sides of the island were pooled 

in order to calculate a mean density of responses across the whole island. However in 

2006, there were a significantly greater number of responses per quadrat on the 

northwest side of the island (1.08, CL 0.842–1.34) than on the southeast side (0.611; 

CL 0.45–0.77, t 301 = 3.03, P < 0.01). For this reason, data were resampled separately 

for each side of the island and treated as if they were two separate populations to 

produce a separate population estimate for each. Population estimates produced from 

each iteration for each side of the island were then summed to produce 10,000 

estimates for the whole of Dun and 95% confidence intervals taken as before. 

With the application of the response rate, the total population of Leach’s Storm-petrel 

on Dun was estimated as 14,490 AOS (95% CL 12,110–17,439) in 2003 and 12,770 

AOS (95% CL 10,046–17,086) in 2006. This represents a significant decrease of 48% 

between 1999 and 2003 (t 435 = 2.95, P < 0.01) but a non-significant 12% decrease 

from 2003 to 2006 (t 632 = 0.19, n.s.; Figure 3). 

Discussion 

The estimate of 12,770 AOS in 2006 represents a decrease of approximately 15,000 

AOS since 1999 and 1,700 AOS since 2003. The scale of decline since 1999 is a major 

concern, particularly if the other subcolonies within the archipelago have also 

decreased by a similar proportion (i.e. 54%), which would have reduced total numbers 

breeding on St Kilda from 45,433 AOS in 1999 to 20,899 AOS in 2006, a loss of about 

49,000 breeding birds. Because estimates of population size of Leach’s Storm-petrel on 

Dun prior to 1999 are qualitative, it is not possible to interpret this finding in relation 

to longer-term trends for this species. 

Before exploring possible ecological or environmental causes of the decline, it is 

important to rule out the possibility that it was an artefact of the playback method 

used during each of the three surveys. The largest error in estimating the number of 

AOS would arise from using an inaccurate estimate of response rate to adjust the 

number of responses obtained during the playback survey. In theory, the decline in 

numbers between 1999 and 2003 could have arisen if: a) in 2003, the response rate 

used was too high and therefore underestimated the number of AOS; and/or b) in 

1999, the response rate was too low and therefore overestimated the number of AOS. 

It is unlikely that the response rate used in 2003 was inaccurate, since this was 

measured on Dun at exactly the same time as the playback survey was carried out. 

Furthermore the response rate measured in 2003 on Dun was very similar to that 

measured at other subcolonies within St Kilda and at other colonies in the British Isles 

(Mitchell 2004). In contrast, in 1999, no simultaneous measurement of response rate 

was conducted on Dun during the survey and instead, a response rate was used that 

had been measured the following year on the neighbouring island of Boreray. This 

response rate was almost identical to that recorded on Dun in 2003. If the number of 

AOS in 1999 were in fact similar to that in 2003, such that there was no decline, the 

number of responses obtained in 1999 – 10,574 – would need to be corrected by a 

factor of 1.36 rather than 2.62, which was based on a response rate of 0.38 (correction 

factor = 1 / response rate). However, a correction factor of 1.36 would only have been 

SEABIRD 21 (2008): 77–84 

81

82 


derived if the response rate in 1999 was 0.74. Response rates to a male chatter call of 

more than 0.52 have yet to be recorded at a Leach’s Storm-petrel colony (Mitchell 

2004) and are unlikely to be achieved since only the incubating males are thought to 

respond to recorded male chatter calls (Taoka et al. 1989). There is no reason to 

suggest that a higher proportion of males than females should be incubating at any 

one time, with sexes sharing incubation equally with a 3-day changeover time (Wilbur 

1969; Watanuki 1985). On Boreray in 2000, the response rate of Leach’s Storm-petrel 

did increase significantly towards dusk from a mean of 0.38 (95% CL 0.34–0.42) 

between 07.10 and 17.15 to 0.52 (95% CL 0.45–0.58) between 20.00 and 22.00, but 

the survey in 1999 was conducted no later than 17.00. 

The response rate to playback by Leach’s Storm-petrel will vary during the breeding 

season, with the highest and most consistent response rates obtained during the peak 

of incubation when the highest number of burrows will be occupied by an adult (Ellis 

et al. 1998). In the UK, peak incubation occurs around mid June to early July and each 

egg is incubated for 41–42 days (Cramp & Simmons 1977). As chicks start to hatch, 

daytime occupancy rates and hence response rates will decrease as adults spend days 

away from the burrow foraging, leaving young unattended (Ellis et al. 1998).The survey 

in 1999 was conducted on 30 June and 1 July, and on 4–10 July in 2003, but around 

two weeks earlier in 2006 on 17–20 June. The surveys in 1999 and 2003 were 

conducted towards the very end of incubation. However, of nine burrows examined on 

neighbouring Hirta in 2003 at the same time as the survey of Dun, the first chick 

hatched on 9 July, the day before the survey on Dun was completed. Furthermore, there 

is no evidence of a decrease in response rate during the course of the survey (Table 1), 

suggesting that the majority of birds during both the 1999 and 2003 surveys were still 

incubating or brooding small chicks, but nevertheless still occupying burrows during the 

day (Leach’s Storm-petrel chicks are brooded continuously for five days after hatching 

(Cramp & Simmons 1977)). It is thus unlikely that the decline in AOS that occurred 

between 1999 and 2003 and 2006 was an artefact of the difference in timing of the 

three surveys, and is more likely to represent a real decline in the population on Dun. 

One potential cause of the observed decline in Leach’s Storm-petrels is predation by 

Great Skuas, which had increased in number exponentially on St Kilda prior to the 1999 

survey of Leach’s Storm-petrel (see above).A study on the neighbouring island of Hirta 

found that the diet of Great Skuas was dominated by Leach’s Storm-petrel, and that 

they hunted during the night when Leach’s Storm-petrels were returning to their 

colonies (Votier et al. 2006). Phillips et al. (1999b) estimated that the total Great Skua 

population on St Kilda in 1996 consumed approximately 15,000 Leach’s Storm-petrels 

per year. Given that between 1996 and 2004, the proportion of Leach’s Storm-petrels 

in the diet of Great Skuas remained relatively constant (Votier et al. 2006), around 

60,000 Leach’s Storm-petrels would have been killed on the archipelago during the 

period between the surveys in 1999 and 2003, assuming that the model of Phillips et 

al. (1999b) is correct and that the Great Skua population has not changed substantially 

since 1998. Whilst this is more than the 45,000 breeding Leach’s Storm-petrels 

we estimate may have been lost from the archipelago, it is likely that recruitment will 

offset some of the losses due to predation mortality. In addition it is also likely that 

SEABIRD 21 (2008): 77–84


non-breeding adult Leach’s Storm-petrels that also visit colonies at night are being 

predated by Great Skuas, but the age or breeding status of Leach’s Storm-petrel 

remains could not be determined from skua pellet analysis. Furthermore, it is possible 

that there may have been greater declines in the Leach’s Storm-petrel populations on 

the other islands than there were on Dun. 

Whilst the Leach’s Storm-petrel is not rare in an international context, with an estimated 

9,000,000–10,600,000 pairs worldwide (Mitchell 2004), the Great Skua, with just 16,000 

pairs is scarce globally (Furness & Ratcliffe 2004). As St Kilda is a Special Protection Area 

(SPA) under the European Commission’s Birds Directive, with both Leach’s Storm-petrel 

and Great Skua listed as qualifying species, if it were thought necessary to devise a 

management solution, it would not be straightforward (Phillips et al. 1999a). 

The question as to whether Great Skuas are the main cause of the decline remains 

unproven. Unfortunately there are no other demographic data available for Leach’s 

Storm-petrel on St Kilda or any other colony in the northeast Atlantic that could be used 

to identify other causes. In fact St Kilda is the only colony of Leach’s Storm-petrel in the 

northeast Atlantic where trends in population size have been accurately measured. 

Recent steps have been taken on St Kilda to initiate monitoring of productivity and 

adult survival of Leach’s Storm-petrel. In 2004, JNCC installed 25 nest boxes on Hirta 

and set up a constant effort mist-netting site for ringing. By attracting Leach’s Stormpetrels 

to nest in the boxes it is hoped that monitoring of productivity (including 

hatching rates, chick growth rates and fledging success) could be carried out more easily 

than in natural nest cavities that have proved extremely difficult to investigate. While 

low productivity, perhaps related to poor food supply, may have contributed to the 

decline in the breeding population, it was most unlikely to have been the sole cause of 

such a steep decline. However, monitoring productivity could give an indication of the 

state of feeding conditions during the breeding season and the potential for the 

population to recover.The main aim of ringing Leach’s Storm-petrel is to gain estimates 

of annual survival rates of adults to determine whether high mortality is indeed the 

main driver of the decline in the breeding population. However, simultaneous 

monitoring of predation rates by skuas is required to determine if observed levels of 

mortality are likely to have resulted from predation at the colony or from some other 

factor, e.g. in the non-breeding areas in the southeast Atlantic. Monitoring changes in 

the numbers of Leach’s Storm-petrel breeding at other colonies in the British Isles and 

elsewhere in the northeast Atlantic would be useful to determine whether the decline 

is restricted to St Kilda or represents a more widespread phenomenon. 


Thanks to Maggie Robinson for help with survey work in 2003; to QinetiQ for use of 

their boat in 2003; to Neil Mitchell, Sarah Money and Susan Bain of the National 

Trust for Scotland and the QinetiQ staff for their support on Hirta. The survey in 

2006 was part funded through a grant from the Seabird Group. Thanks also to 

Norman Ratcliffe and Steve Newton who provided extremely useful comments on 

an earlier version of this manuscript. 

SEABIRD 21 (2008): 77–84 

83

84 


References 

Cramp, S. & Simmons, K. E. L. (eds.) 1977. The Birds of the Western Palearctic. Vol. I. Oxford 

University Press, Oxford. 

Crowley, P. H. 1992. Resampling methods for computation-intensive data analysis in ecology 

and evolution. Annual Review of Ecology & Systematics 23: 405–447. 

Ellis, P., Ratcliffe, N. & Suddaby, D. 1998. Seasonal variation in diurnal attendance and 

response to playback by Leach’s Petrels Oceanodroma leucorhoa on Gruney, Shetland. Ibis 140: 

336–339. 

Furness, R. W. & Ratcliffe, N. 2004. Great Skua Stercorarius skua. In: Mitchell, P. I., Newton, S. 

F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. Seabird Populations of Britain and Ireland: 173–186. 


Icelandic Institute of Natural History 2000. Red List of Threatened Species in Iceland. Vol. 2. 

Náttúrufræ∂istofnun Islands, Reykjavik. 

Manly, B. F. J. 1998. Randomization, Bootstrap and Monte Carlo Methods in Biology. Chapman & 

Hall, London. 

Mayhew, P., Chisholm, K., Insley, H. & Ratcliffe, N. 2000. A survey of Storm Petrels on Priest 

Island. Scottish Birds 21: 78–84. 

Mitchell, P. I. 2004. Leach’s Storm-petrel Oceanodroma leucorhoa. In: Mitchell, P. I., Newton, S. 

F., Ratcliffe, N. & Dunn, T. E. (eds.) 2004. Seabird Populations of Britain and Ireland: 101–114. 


Phillips, R. A., Catry, P., Thompson, D. R., Hamer, K. C. & Furness, R. W. 1997. Inter-colony 

variation in diet and reproductive performace of Great Skuas Catharacta skua. Marine Ecology 

Progress Series 152: 285–293. 

Phillips, R. A., Bearhop, S., Hamer, K. C. & Thompson, D. R. 1999a. Rapid population growth 

of Great Skuas Catharacta skua at St Kilda: implications for management and conservation. 

Bird Study 46: 174–183. 

Phillips, R. A., Thompson, D. R. & Hamer, K. C. 1999b. The impact of Great Skua predation on 

seabird populations on St Kilda: a bioenergetics model. Journal of Applied Ecology 36: 218–232. 

Ratcliffe, N., Vaughan, D., Whyte, C. & Shepherd, M. 1998. Development of playback census 

methods for Storm Petrels Hydrobates pelagicus. Bird Study 45: 302–312. 

Taoka, M., Sato, T., Kamada, T. & Okumura, H. 1989. Heterosexual response to playback calls 

of the Leach’s Storm-petrel Oceanodroma leucorhoa. Journal of Yamashina Institute of 


Votier, S. C., Furness, R. W., Bearhop, S., Crane, J. E., Caldow, R. W. G., Catry, P., Ensor, K., 

Hamer, K. C., Hudson, A. V., Kalmbach, E., Klomp, N. I., Pfeiffer, S., Phillips, R. A., Prieto, I. & 

Thompson, D. R. 2004. Changes in fisheries discard rates and seabird communities. Nature 

427: 727–730. 

Votier, S. C., Crane, J. E., Bearhop, S., de León, A., McSorley, C. A., Mínguez, E., Mitchell, P. 

I., Parsons, M., Phillips, R. A. & Furness, R. W. 2006. Nocturnal foraging by Great Skuas 

Stercorarius skua: implications for conservation of storm-petrel populations. Journal of 


Watanuki, Y. 1985. Breeding biology of Leach’s Storm Petrels Oceanodroma leucorhoa on 

Daikoku Island, Hokkaido, Japan. Journal of Yamashina Institute of Ornithology 17: 9–22. 

Wilbur, H. M. 1969. The breeding biology of the Leach’s Petrel Oceanodroma leucorhoa. Auk 86: 

433–442. 

SEABIRD 21 (2008): 77–84

Rafting behaviour of 

Manx Shearwaters 

Puffinus puffinus 

Wilson, L. J. 1 *, McSorley, C. A. 1a ,Gray,C.M. 2 , 

Dean, B. J. 1 , Dunn,T. E. 1 ,Webb,A. 1 & Reid, J. B. 1 

*Correspondence author. 

Email: linda.wilson@jncc.gov.uk 

1 Joint Nature Conservation Committee, 7 

Thistle Place, Aberdeen AB10 1UZ, UK 

( a Current address: Scottish Natural Heritage, 1 

Kilmory Industrial Estate, Lochgilphead, Argyll 

PA31 8RR, UK); 2 Countryside Council for Wales, 

Maes Y Ffynnon, Penrhosgarnedd, Bangor LL57 

2DW, UK (Current address: Peak District 

National Park Authority, Aldern House, Baslow 

Road, Bakewell, Derbyshire DE45 1AE, UK). 

Abstract 

Radio-telemetry data were collected on 

rafting Manx Shearwaters Puffinus puffinus at 

Skomer (southwest Wales), Rum (northwest 

Scotland) and Bardsey (northwest Wales) 

between 2003 and 2005. These were used to 

investigate whether Manx Shearwaters tend 

to raft adjacent to their breeding areas and 

whether rafts move closer towards shore as 

the evening progresses. On Skomer and 

Bardsey, there was a tendency for birds to raft 

in an area roughly adjacent to where they 

bred, although they did not raft exclusively 

opposite their breeding site. On Rum, birds 

breeding at two different locations appeared 

to show different preferences for rafting areas. 

However, it was difficult to draw conclusions 

from this, as signal coverage around the island 

was poor, and the breeding locations were 

close together. At all three islands, there was 

strong evidence that birds tended to move 

closer inshore as the evening progressed. 

Introduction 

Many nocturnal Procellariids form rafts, or 

dense flocks of birds, on the sea adjacent to 

their breeding colonies from late afternoon 

onwards, before coming ashore at nightfall 

(Brooke 2004). Species of shearwater known 

to engage in such rafting behaviour include 

Cory’s Shearwater Calonectris diomedea, 

Short Notes 

Streaked Shearwater C. leucomelas, Manx 

Shearwater Puffinus puffinus, Flesh-footed 

Shearwater P. carneipes, Great Shearwater P. 

gravis, Sooty Shearwater P. griseus and Shorttailed 

Shearwater P. tenuirostris (Brooke 2004). 

Nocturnal attendance at the colony is thought 

to be a predator avoidance strategy (Brooke & 

Prince 1991; Mougeot & Bretagnolle 2000; 

Keitt et al. 2004), but the function of rafting is 

unclear. It is possible that the birds have 

difficulty in timing their return to their colony 

from their distant foraging grounds to 

precisely coincide with nightfall, so they 

assemble in rafts until it is safe to come 

ashore (Warham 1996; Brooke 2004). Most 

Procellariids in nearshore rafts do not feed, or 

only do so if schooling fish are present 

(Lockley 1942; Warham 1990), but rafts may 

provide an arena for courtship behaviour and 

other social interactions, as well as 

maintenance behaviour such as preening and 

resting (Brooke 1990; Warham 1996). 

Rafts can comprise thousands of birds and it is 

assumed that they include both breeding and 

non-breeding birds, as both visit the colony 

after nightfall (Furness et al. 2000). Anecdotal 

observations indicate that rafting birds are 

generally more restless during windy 

conditions, frequently flying and re-alighting 

to maintain position (Brooke 2004), 

suggesting that raft position is of importance 

to participating birds. It is thought that Manx 

Shearwaters may raft adjacent to where they 

come ashore and their rafts might tend to 

approach the shore, once darkness falls 

(Brooke 1990). Based on visual observations of 

birds coming ashore close to where they had 

been rafting, Furness et al. (2000) also 

assumed Cory’s Shearwaters tended to raft 

opposite their breeding site. 

Most observations of rafting have been 

anecdotal, and to our knowledge, there have 

been no detailed studies of rafting behaviour 

itself. Recently, we used radio-telemetry to 

determine how far Manx Shearwater rafts 

extend offshore, as part of a project to inform 

the issue of possible extensions to breeding 

colony Special Protection Areas into adjacent 

marine areas (McSorley et al. 2008; Wilson et 

SEABIRD 21 (2008) 

85

86 

Short Notes 

al. in prep.). In this paper, we use those data to 

further investigate whether Manx Shearwaters 

tend to raft adjacent to their breeding areas 

and whether rafts move closer towards shore 

as the evening progresses. 

Methods 

Detailed methods of the fieldwork can be found 

in McSorley et al. (2008) and Wilson et al. (in 

prep.). The main points are summarised here. 

Study colonies: The study took place from 

May to August on the islands of Skomer 

(southwest Wales), Rum (northwest Scotland) 

and Bardsey (northwest Wales) (Figure 1a–c), 

during 2003, 2004 and 2005 respectively. 

These islands host the world’s three largest 

Manx Shearwater breeding colonies (Skomer: 

101,800 pairs, Rum: 120,000 pairs, Bardsey: 

16,183 pairs (Newton et al. 2004)). On each 

island, we fitted radio-tags to breeding adults 

at a number of geographically distinct breeding 

areas, allowing us to test whether the location 

of the breeding area influenced raft location 

around the island. Four study breeding areas 

were chosen on Skomer, two on Rum and five 

on Bardsey (Table 1, Figure 1a–c). It was not 

possible to get a wide geographical spread of 

study sites on Rum, as the colony is confined to 

the high mountains, where access is difficult. 

SEABIRD 21 (2008) 

An initial visit to each colony in May (during 

the incubation period) allowed the marking of 

potential study burrows, with adults from each 

being sexed using cloacal inspection where 

possible (Gray & Hamer 2001). 

Radio-tag attachment and radio-tracking: 

Radio-tag attachment was carried out during 

a second visit in July (during the chick rearing 

period). A VHF radio transmitter (supplied by 

Biotrack Ltd) was attached to the two central 

tail feathers of one adult from each accessible, 

occupied study burrow. To reduce any adverse 

effects of the tag and the possibility of chick 

desertion, only the heaviest individuals, with 

chicks older than five days were tagged. Tags 

weighed 8.9 g on Skomer and 4.4 g on Rum 

and Bardsey, less than 2.5% (Skomer) or 1.3% 

(Rum and Bardsey) of average adult body 

weight. Tags were attached using either selfamalgamating 

tape (Skomer), or Tesa® tape 

(Rum and Bardsey). Adults were returned to 

their burrows immediately following the 

tagging procedure, which took less than ten 

minutes. At the end of the study, birds were 

recaptured and their tags removed. Some birds 

could not be recaptured and it was assumed 

that these would lose their tags during their 

subsequent post-breeding moult. 

Table 1. The number of marked burrows and birds that were tagged and subsequently located rafting, for each 

Manx Shearwater Puffinus puffinus breeding site on each island. 

Total no. of No. of birds No. of birds 

Island Breeding Site burrows marked tagged located rafting 

Skomer Pigstone 14 7 6 

The Wick 35 8 4 

Behind house 23 8 3 

The Neck 26 7 6 

Total 98 30 19 

Rum Hallival 45 18 11 

Askival 75 10 9 

Total 120 28 20 

Bardsey Cristin 9 2 2 

Nant 32 11 11 

Pen Cristin 35 9 9 

NW Fields 21 5 5 

South-end 7 3 3 

Total 104 30 30

Radio-tracking, using Sika receivers and fivebar 

rigid Yagi antennas (Biotrack Ltd), began on 

the first evening of tag attachment and took 

place on: 14 evenings between 15 and 29 July 

2003 (Skomer); 15 evenings between 15 July 

and 6 August 2004 (Rum); and 18 evenings 

between 31 July and 19 August 2005 

(Bardsey). Radio-tracking was conducted from 

three locations on Skomer and mainland 

Pembrokeshire (50–80 m above sea level 

(a.s.l.)), six locations on Rum (200–250 m 

a.s.l.), and five locations on Bardsey and 

mainland Gwynedd (100–160 m a.s.l.), 

although at any one time only two or, usually, 

three locations were used. Radio-tracking 

generally commenced 17.00–19.00 (GMT) 

and continued until the last detectable tagged 

bird returned to the colony (usually by 24.00). 

During this time simultaneous compass 

bearings for each detectable signal (bird) in 

the area were taken by two or three observers, 

following a coordinated tracking schedule 

(one bird every three minutes). 

Analysis: Analyses were performed only on 

those birds that were thought to be rafting; 

data were checked prior to analyses and any 

bearings that were clearly incorrect, or were for 

birds that were travelling or foraging, were 

removed. This was determined by signal 

strength, notes taken by the trackers and 

comparison of signal direction between 

tracking locations. Compass bearings were 

corrected for magnetic north using the 

appropriate adjustment for each area. Bird 

locations were estimated by triangulation of 

corrected bearing data with LOAS TM 3.0.2 

(Location Of A Signal) software © 1998–2004 

(Ecological Software Solutions TM). A maximum 

likelihood estimator was used for triangulation. 

To test whether raft location around each 

island was influenced by birds’ breeding 

locations, a quadrant was centred on the 

‘central study burrow’ and rafting birds from 

each breeding area were classified according 

to which quadrant they fell within (i.e. 

whether they were located in rafts northeast, 

northwest, southeast or southwest of the 

island, or in the case for Rum, east or west). 

Observed and expected frequencies were 

compared using a Chi-squared test. The 

Short Notes 

‘central study burrow’ was defined as the 

geographical mean centre of all study burrows 

containing a tagged bird, using the central 

feature tool within Esri® Arcmap TM 9.2. 

A Spearman rank correlation of raft distance 

against time was used to investigate whether 

rafts moved closer to the colony over time. 

The distance between each estimated raft 

location and the nearest point of land on the 

island (‘raft distance’) was calculated using 

the spatial join tool within Esri® Arcmap TM 9.2 

and based on a polygon of the high water 

mark supplied by Ordnance Survey under 

licence [JNCC][100017995][2008]. 

Figure 1a. Rafting locations of Manx Shearwaters Puffinus 

puffinus around Skomer, according to their breeding area. The 

position of each quadrant is shown, centred on the mean centre 

of all study burrows. 

SEABIRD 21 (2008) 

87

88 

Short Notes 

Results and Discussion 

Thirty birds on Skomer, 28 on Rum and 30 on 

Bardsey, were fitted with a radio-tag, and of 

these it was possible to estimate raft locations 

for 19, 20 and 30 birds respectively (Table 1). 

Within Skomer and Rum, the sample of birds 

for which we were unable to estimate raft 

locations included individuals breeding at each 

of the study areas, making it unlikely that the 

lack of data from these birds biased our results. 

Six birds from Skomer and three from Rum 

were not detected at all after tag attachment, 

and of these, four (all from Skomer) were 

recaptured of which all had lost their tag. 

Therefore, it seems likely that our failure to 

detect some birds was due to tag loss, rather 

than because they were not in the vicinity of 

the colony. Tag loss was reduced during the 

Rum and Bardsey studies by having a more 

secure attachment using Tesa® tape and lighter 

tags. In the sample of birds for which we 

obtained raft locations, the sex ratio did not 

significantly differ from 50:50 (Skomer: 9m:9f, 

x2 = 0.11; Rum: 11m:7f, x2 = 0.56; Bardsey: 

8m:7f, x2 = 0.17; P > 0.05, 1 df for all). 

SEABIRD 21 (2008) 

Of those birds located rafting, 58% (Skomer), 

40% (Rum) and 93% (Bardsey) were located 

in rafts at least twice during the study period. 

Guilford et al. (2008) suggested that most 

breeding Manx Shearwaters on Skomer may 

not join rafts. This was based on data from 

birds fitted with global positioning devices 

which showed that some birds were some 

distance from the colony at nightfall. We were 

unable to estimate the number of evenings 

that birds visited the colony without rafting, 

as we were much more likely to detect birds 

that were rafting than birds which simply flew 

straight from their foraging grounds to their 

burrow (where the radio signal was not 

detectable except at close range). However, 

our data does not exclude the possibility that 

some birds attended their burrow without 

joining rafts beforehand, although we feel this 

would be a minority. 

Figure 1(a–c), shows the rafting locations 

estimated from our analysis around each island, 

colour coded according to breeding location. 

Figure 1b. Rafting locations of Manx Shearwaters Puffinus puffinus around Rum, according to 

their breeding area. The position of the line dividing the island into west and east is shown, 

centred on the mean centre of all study burrows.

Skomer (Figure 1a): No raft location data were 

obtained to the east or west of Skomer and 

sample sizes were small to the south. This was 

largely due to the island’s topography 

interfering with signals from tagged birds and 

the locations of trackers (at the highest points 

on the island, falling in an east–west orientated 

line) making it difficult to triangulate to the 

east or west. Casual observations before 

nightfall, together with signals coming from 

these directions (but which could not be 

triangulated), indicated that birds were rafting 

in the areas for which we lacked data. 

There was a significant difference in the 

observed and expected frequencies of raft 

locations in each quadrant for Skomer (x2 = 

33.53, 9 df, P < 0.001, Figure 2a). Most raft 

locations were to the northeast (53%) and the 

northwest (37%). Birds breeding at Pigstone 

(in the west of the island) rafted most often to 

the northwest while birds breeding at the 

Neck (in the east of the island) were located 

more often to the northeast. Observations 

from birds breeding at the House (in the east 

Short Notes 

of the island) were equally divided between 

the northeast and northwest. Almost all of the 

observations in the two southern quadrants 

were from birds breeding at the Wick (in the 

south of the island), although most of 

observations from the Wick were in the 

northeast quadrant. 

As there were few observations to the south 

of Skomer, the Chi-squared test was repeated 

excluding raft locations to the southeast and 

southwest; there was still a significant 

difference (x2 = 11.29, 3 df, P < 0.05). 

Thus, there was evidence that the location of 

the breeding area had some influence on raft 

location, with most of birds breeding in the 

east of the island, rafting to the northeast, 

most birds breeding in the west of the island, 

rafting to the northwest, and almost all of the 

raft observations to the south, being from 

birds breeding in the south. 

Rum (Figure 1b): No raft location data were 

obtained to the north or south of Rum. On 

Figure 1c. Rafting locations of Manx Shearwaters Puffinus puffinus around Bardsey, 

according to their breeding area. The position of each quadrant is shown, centred on the 

mean centre of all study burrows. 

SEABIRD 21 (2008) 

89

Percentage of rafting observations 

in each quadrant 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

90 

Short Notes 

5 

5 

south 

-east 

south 

-west 

north 

-east 

north 

-west 

Rum, the position of trackers was limited due 

to access difficulties, and the surrounding 

topography allowed adequate signal coverage 

only to the west and the east. However, birds 

have been observed rafting to the north (S. 

Morris, Scottish Natural Heritage, pers. 

comm.) and our tracking indicated that birds 

were probably also rafting to the south. 

Most (81%) raft locations were to the east of 

Rum, reflecting greater observer effort there. 

There was a significant difference in the 

SEABIRD 21 (2008) 

40 

12 

2 

1 

House (east) Neck (east) Pigstone (west) Wick (south) 

Breeding area 

Figure 2a. The percentage of Manx Shearwater Puffinus puffinus raft locations in each of four quadrants around 

Skomer, in relation to breeding area (Pigstone, House, Neck and Wick). Quadrants were centred on the mean centre 

of all the study burrows. Bar numbers are frequency numbers. 

100 

90 

80 

70 

60 

50 

29 

West 

East 

22 

40 

30 

20 

10 

80 

133 

0 

Askival 

Breeding area 

Hallival 

Figure 2b. The percentage of Manx Shearwater Puffinus puffinus 

raft locations in each of two areas around Rum, in relation to 

breeding area (Askival and Hallival). The dividing line between the 

two areas was centred on the mean centre of all the study 

burrows. Bar numbers are frequency numbers. 


in each sub-area 

22 

26 

observed and expected frequencies of raft 

locations between each breeding area, within 

the east and west areas (x2 = 19.14, 1 df, P < 

0.01, Figure 2b). This is perhaps surprising 

considering the close proximity of the two 

breeding areas. Most (57%) of the westerly 

locations were from birds that bred at Askival 

and most (62%) of the easterly locations were 

from birds that bred at Hallival, although 

individuals from both breeding areas rafted 

both to the east and west of the island. 

Birds rafting to the west of Rum may position 

themselves there so that they can reach their 

burrow by flying directly east up the Glen 

Harris valley system (Figure 1b) across land 

and around the back of the site. However, 

anecdotal observations suggested that some 

birds that rafted to the west gradually moved 

southwards and around to the east side of the 

island, presumably to access their burrow from 

the east. Thus, it seems that on Rum, 

topography and access routes to the burrow, 

as well as burrow location itself, may influence 

raft locations, particularly just prior to birds 

returning to their colony. 

Bardsey (Figure 1c): The highest sample size 

of both rafting observations and tagged birds, 

and the most complete signal coverage 

achieved, was on Bardsey. There was a 

significant difference in the observed and 

3 

11 

25 

22

expected frequencies between each quadrant 

for Bardsey (x2 = 57.78, 12 df, P < 0.001). On 

Bardsey, most (56%) raft locations were to the 

southeast of the island, while the other 

quadrants each contained only 14–15% of 

raft locations (Figure 2c). Birds from all 

subcolonies (except South-end) appeared to 

show a preference for rafting to the southeast 

of the island, possibly because this provides 

the most sheltered area from wind and/or 

strong tidal currents. Once the southeast 

locations are accounted for, there was 

evidence that birds tended to raft adjacent to 


in each quadrant 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

80 

13 

32 

13 

Nant 

(northeast) 

south 

-east 

south 

-west 

north 

-east 

north 

-west 

18 

5 

8 

14 

NW fields 

(northwest) 

Short Notes 

their breeding area, with birds breeding in the 

northeast (Nant) tending to raft to the 

northeast, birds breeding in the northwest 

(Northwest fields) tending to raft to the 

northwest, and birds breeding in the 

southwest (South-end) tending to raft to the 

southwest. In addition, most of rafting 

observations in the preferred southeast 

quadrant were from birds breeding in the 

southeast (Pen Cristin). 

Do rafts move closer to shore over time?: 

There were no locations generated earlier than 

95 

31 

10 

27 

Pen Cristin 

(southeast) 

Breeding area 

7 

8 

5 

South-end 

(southwest) 

SEABIRD 21 (2008) 

91 

17 

1 

1 

Cristin 

(middle) 

Figure 2c. The percentage of Manx Shearwater Puffinus puffinus raft locations in each of four quadrants around 

Bardsey, in relation to breeding area (Nant, NW fields, Pen Cristin, South end and Cristin). Quadrants were centred 

on the mean centre of all the study burrows. Bar numbers are frequency numbers. 

Mean (+SD) raft distance (km) 

12 

10 

8 

6 

4 

2 

Bardsey 

0 

1645-1900 1901-2000 2001-2100 2101-2200 2201-2300 2301-2400 

Time interval (GMT) 

Figure 3. The mean and standard deviation (SD) of rafting distance (km) of Manx Shearwaters Puffinus puffinus 

from their breeding islands, at different time intervals over the evening. 

Rum 

Skomer

92 

Short Notes 

19.00 (GMT) on Skomer and Rum, as birds did 

not start rafting within range of the trackers 

until at least 19.00 on Skomer and 19.30 on 

Rum. On Bardsey, the earliest locations 

generated were much earlier, at 16.45, 

reflecting the earlier onset of nightfall later on 

in the season, when the Bardsey study took 

place. On all three islands, there was evidence 

that rafting birds moved significantly closer to 

the colony as the evening progressed 

(Spearman Rank correlation: r s = -0.19, n = 

174, P < 0.01, Skomer; r s = -0.13, n = 264, P < 

0.05, Rum; r s = -0.32, n = 385, P < 0.001, 

Bardsey) (Figure 3). Positioning closer to shore 

as darkness falls may allow a better 

assessment of the approach to the breeding 

burrow, which could be associated with 

assessing predation threat and local light 

levels. By reducing flight time to the burrow, 

temporary periods of reduced light levels, such 

as when moonlight is obscured by clouds, may 

be taken advantage of more readily. 

Conclusion 

This study highlights how radio-telemetry 

can be a useful tool to examine rafting 

behaviour. There was clear evidence from all 

three colonies that rafting birds tended to 

approach closer to shore as the evening 

progressed. The data also indicated that raft 

location around an island may be influenced 

by the location of the breeding area, although 

the situation was not clear-cut and data were 

confined to one season for each island. For 

Skomer and Rum, data were limited in some 

areas by poor signal coverage, making it more 

difficult to assess the true influence of 

breeding location on raft location. The rafts 

around Bardsey included a few birds which 

were feeding and although these birds were 

removed from the analysis, the presence of 

feeding birds might attract other birds and so 

influence the location of rafts. It is likely that 

other factors also influence raft location, such 

as weather conditions, local oceanographic 

variables, and access routes to the colony, and 

that the combined effect of such factors will 

vary between colonies. Around fairly small 

islands, such as the ones in this study, the 

choice of raft site may not be as crucial as 

around larger islands, as differences in travel 

SEABIRD 21 (2008) 

time to the burrow from different raft 

locations around the island could be 

insignificant. Thus, a stronger influence of 

breeding location on rafting location might 

be more evident where colonies are on larger 

islands, such as with Cory’s Shearwaters in 

the Azores archipelago. 


All work was carried out under licence from 

the Countryside Council for Wales, Scottish 

Natural Heritage and the British Trust for 

Ornithology, and was funded by the Joint 

Nature Conservation Committee. Thanks to 

Skomer and Skokholm Islands Management 

Committee, Scottish Natural Heritage, Ynys 

Enlli Trust and Bardsey Bird and Field 

Observatory for permission to work on the 

islands, and to the wardens for their 

assistance. The fieldwork involved many 

individuals, to whom we are very grateful. 

Advice on radio-tracking techniques was 

gratefully received from Peter Smith and Brian 

Cresswell (Biotrack Ltd), and Robert Kenward, 

while the manuscript was improved by 

comments from Chris Perrins. 

References 

Brooke, M. 1990. The Manx Shearwater. Poyser, London. 

Brooke, M. 2004. Albatrosses and Petrels across the 

World. Oxford University Press, Oxford. 

Brooke, M. de L. & Prince, P. A. 1991. Nocturnality 

in Seabirds. Proceedings of the International 

Ornithological Congress 20: 1113–1121. 

Furness, R. W., Hilton, G. & Monteiro, L. R. 2000. 

Influences of coastal habitat characteristics on the 

distribution of Cory’s Shearwaters Calonectris 

diomedea in the Azores archipelago. Bird Study 47: 

257–265. 

Gray, C. M. & Hamer, K. C. 2001. Food-provisioning 

behaviour of male and female Manx Shearwaters, 

Puffinus puffinus. Animal Behaviour 62: 117–121. 

Guilford, T. C., Meade, J., Freeman, R., Biro, D., 

Evans, T., Bonadonna, F., Boyle, D., Roberts, S., 

Perrins, C. M. 2008. GPS tracking of the foraging 

movements of Manx Shearwaters Puffinus puffinus 

breeding on Skomer Island,Wales. Ibis 150: 462–73. 

Keitt, B. S., Tershy, B. R. & Croll, D. A. 2004. 

Nocturnal behavior reduces predation pressure on 

Black-vented Shearwaters Puffinus opisthomelas. 

Marine Ornithology 32: 173–178.

Lockley, R. M. 1942. Shearwaters. Dent, London. 

McSorley, C. A., Wilson, L. J., Dunn, T. E., Gray, C., 

Dean, B. J., Webb, A. & Reid, J. B. 2008. Manx 

Shearwater Puffinus puffinus evening rafting 

behaviour around colonies on Skomer, Rum and 

Bardsey: its spatial extent and implications for 

recommending seaward boundary extensions to 

existing colony Special Protection Areas in the UK. 

JNCC Report 406. Peterborough, UK. 

Mougeot, F. & Bretagnolle, V. 2000. Predation risk 

and moonlight avoidance in nocturnal seabirds. 

Journal of Avian Biology 31: 376–386. 

Newton, S. F., Thompson, K. & Mitchell, P. I. 2004. 

Manx Shearwater Puffinus puffinus. In: Mitchell, P. 

I., Newton, S. F., Ratcliffe, N. & Dunn, T. E. (eds.) 

Seabird Populations of Britain and Ireland: 63–80. 

Poyser, London 

Late breeding by 

Great Cormorants 

Phalacrocorax carbo 

Craik J. C. A. 1 * and Bregnballe T. 2 


Email: clive.craik@sams.ac.uk 

1 Scottish Association for Marine Science, 

Dunstaffnage, Oban, Argyll PA37 1QA, UK; 2 

National Environmental Research Institute, 

University of Aarhus, Department of Wildlife 

Ecology and Biodiversity, Kalø, Grenåvej 14, 

DK-8410 Rønde, Denmark. 

Abstract 

Two distinct waves of synchronised breeding 

occurred at a colony of Great Cormorants 

Phalacrocorax carbo in west Scotland in 2007. 

The second wave led to young fledging in 

September, exceptionally late for this species 

locally. Evidence from elsewhere suggests that 

this second wave was not double-brooding 

(raising of two broods in one year by the same 

pair) but breeding by newly-arrived birds that 

had failed at a nearby colony earlier in the year. 

Introduction 

At a small colony of Great Cormorants 

Phalacrocorax carbo (hereafter ‘Cormorants’) 

in Scotland in 2007, there were two distinct 

Short Notes 

Warham, J. 1990. The Petrels. Their Ecology and 

Breeding Systems. Academic Press, London. 

Warham, J. 1996. The Behaviour, Population Biology 

and Physiology of the Petrels. Academic Press, London. 

Wilson L. J., McSorley, C. A., Gray, C. M., Dean, B. 

J., Dunn, T. E., Webb, A. & Reid, J. B. In prep. 

Radio-telemetry as a tool to define protected 

areas for seabirds in the marine environment. 

waves of successful breeding separated by 2.5 

to 3 months. In June, the normal fledging 

season, 47 young fledged from 18 nests. Then, 

in September, 28 young fledged from 14 nests 

in the same small area (details in Appendix 1). 

A well-synchronised pulse of many young 

fledging together so late in the year is unusual, 

at least in this part of Scotland where such an 

occurrence seems not to have been reported 

before.The first purpose of this short note is to 

place this event on record. 

Our second objective is to consider whether 

these two waves were caused by the same 

pairs breeding for a second time (successful 

‘double-brooding’), or by incoming birds 

nesting in the same small colony area, or 

possibly a mixture of these. There were no 

ringing data to resolve this question directly. 

We therefore approach it indirectly by 

examining other records of late breeding in 

this and a closely related species. If it was 

genuine double-brooding, 14 of the original 18 

pairs raised successful second broods. This 

percentage (78%) will be compared with 

similar measures from elsewhere. 

Late breeding 

Late breeding has been recognised as an aspect 

of Cormorant breeding behaviour at, at least, 

two colonies in continental Europe. In 1992 and 

1993, the colony at Val Campotto in the Po 

SEABIRD 21 (2008) 

93

94 

Short Notes 

delta, Italy, was studied by Grieco (1994). 

Breeding occurred from February to September 

with two peaks of laying, the main one in spring 

and a smaller one in summer. In 1992 when the 

colony had 250 nests, there were 13 late 

breeders in a sample of 75 nests (17%). In 1993 

when the colony had 270 nests, this proportion 

was 19/115 (17%). Most of the later breeders 

used nests built by other birds earlier in the 

year but, in both 1992 and 1993, Grieco 

suggested that some cases of reuse of a nest 

after successful fledging occurred because the 

same parents initiated a new clutch. He based 

this suggestion on the behaviour of the adults 

concerned; thus, for example, the adults 

sometimes fed the large or fledged young and 

did not drive them away. In 1992 Grieco 

recorded four such nests with presumed second 

breeding, in one of which a second brood 

fledged (three young in August). Thus, in his 

1992 sample of 75 nests, the proportion with 

successful double broods was 1/75 (1.3%). 

Late breeding has been described at other 

colonies in Europe, including the large colony 

at Vorsø in Denmark. This grew from 1,000 

nests in 1980 to 5,000 nests in 1991. Each 

year a proportion of breeding attempts failed, 

mostly during incubation but some at the 

chick stage. In some of these nests, a second 

clutch was laid, either by the same pair or by 

another pair. Most second clutches occurred 

after the loss of the first eggs or young or, 

more rarely, after young had fledged. The 

proportion of second clutches varied from 

year to year: 9–19% of nests in 1980–84, 

0–2% in 1985–88, and 0–3.2% in 1990 and 

1993–95 (based on 97–348 nests with eggs 

followed throughout each season; Bregnballe 

1996). In 1980–1983, the years in which 

second clutches were most frequent, the 

annual percentages of new clutches after 

fledging of a first brood were in the range 

3.3–9.6%, and successful second broods (nests 

where chicks fledged both in the first and in 

the second attempt) ranged from 2.3 to 4.6%. 

Both these quantities are expressed as 

percentages of the number of nests where 

eggs were incubated in the first attempt (214 

nests in 1980, 348 in 1981, 293 in 1982 and 

280 in 1983) (Bregnballe & Gregersen 

unpubl.). In some of these nests, chicks of the 

SEABIRD 21 (2008) 

first brood were tolerated by the adults if they 

returned to a nest where a second clutch was 

being incubated, suggesting that, in at least 

some cases, one or both of the adults was 

breeding for a second time after fledging a 

brood. The range of 2.3 to 4.6%, calculated for 

the ‘best’ of the study years, thus places an 

upper limit of about 5% on same-pair 

successful second broods. 

Second-brood behaviour has also been seen at a 

tree colony in France (Demongin 1993) but, like 

the above Danish and Italian data, these reports 

were not based on marked birds. Unequivocal 

identification of double-brooding requires 

observation of individually marked birds. 

Double-brooding by marked adults 

At Vorsø, 23–318 breeders (on average 156) 

with individually marked colour-rings were 

followed from egg-laying onwards annually 

from 1981 to 2004. During this period, there 

were four cases of marked birds with successful 

first broods followed by a second clutch that 

was incubated (Bregnballe & Gregersen 

unpubl.). Of these, the outcome of one (in 

1987) was unknown. Of the other three, all in 

1984, one fledged no young, while two fledged 

second broods. Both involved marked females 

that had each laid first clutches in February, 

fledged a first brood in May, laid second 

clutches starting 30 May and 1 June, and 

respectively fledged one and two young. The 

number of colour-ringed birds known to have 

had at least one clutch in 1984 was 186, so the 

proportion of known successful second broods 

in that year was 2/186 or 1.1%. It was zero 

among the ringed parents studied in the other 

23 years of the study. 

Data from similar species 

At a colony of European Shags P. aristotelis 

(hereafter ‘Shags’) in Brittany, France, Cadiou 

(1994) described a ringed 12-year-old female 

which, in 1993, began its first clutch on 7–9 

February, fledged three young in early May, 

laid a second clutch early in June and fledged 

two young early in September. Cadiou cited 

reports of similar cases from three sites in 

Britain (one at each), noting that at one the 

adults had not been marked so two pairs 

might have been involved. He considered the

frequency of such cases to be very low - one 

in several thousand nests. 

Wanless & Harris (1997) made a detailed 

study of Shags nesting at the Isle of May, 

Scotland during 1985–1996. Breeding 

numbers varied between a maximum of 

1,916 nests in 1987 and a minimum of 501 

nests in 1995 after a major population crash 

in 1994. About 40% of the birds had been 

ringed and, when possible, identities were 

checked of birds involved in two attempts at 

breeding in the same year. Second breeding 

attempts were noted only in 1987 and 

1995, years in which the median laying 

dates were the earliest in the study. In 1987, 

three nests from which young had fledged in 

June continued to be occupied in July; two 

of these pairs then laid and one clutch 

hatched, but no young fledged. In 1995 a 

remarkable 27 nests continued to be 

occupied, eggs were laid in 20, five clutches 

hatched and all five fledged two young each. 

The authors did not specify whether any of 

these five successful pairs were individually 

identifiable but they noted that, of the 27 

cases, two involved the same pair breeding 

Short Notes 

twice while, in four, one adult was known to 

have remained the same. There was no 

evidence in any of the 27 that change in 

site-holders had taken place. The authors 

suggested that the unusually high number of 

second breeders in 1995 was a densitydependent 

response to the crash of 1994. 

The five successful second broods 

represented 1% of the 1995 breeding 

population of 501 nests, but no such cases 

were recorded in the other 11 years of the 

12-year study. Wanless & Harris cited other 

evidence that genuine double-brooding in 

this and related species is uncommon. 

Conclusion 

These studies of both Cormorants and Shags 

show that successful double-brooding is a 

rare event. While second clutches are not 

unusual after failed clutches, they are less 

common after failed broods and rare after 

successful broods. The two most detailed 

studies with individually marked birds, those 

of Bregnballe & Gregersen (unpubl.) with 

Cormorants and Wanless & Harris (1997) with 

Shags, concur in several important findings. 

Both found that in most years there was no 

Figure 1. Great Cormorant Phalacrocorax carbo chicks, Glas Eilean, Loch Fyne, 27 August 2007 © Tom Callan. 

SEABIRD 21 (2008) 

95

96 

Short Notes 

double-brooding, but that there were 

occasional good years when, after successful 

first broods, second clutches and broods 

tended to occur. The few available records of 

identifiable individuals with successful second 

broods showed that all laid their first clutches 

early in the year (February in the case of 

Cormorants in continental Europe, March- 

April in Shags in Scotland) – perhaps not 

surprisingly, in view of the time needed to 

raise two successive broods. In the occasional 

years when they occurred, successful second 

broods formed very low proportions of the 

number of nests at a colony, about 1% in 

each study. Data from unmarked Cormorants 

at Vorsø (Denmark) and from Val Campotto 

(Italy) suggested upper limits to this quantity 

of about 5% and 1% respectively. 

These percentages are very much lower than 

the 78% observed at Glas Eilean in 2007. (The 

difference is so large that we are justified in 

ignoring the relatively minor differences in 

methods and interpretation between the 

projects considered above.) Thus it is 

reasonable to infer that the event at Glas 

Eilean was not genuine double-brooding and 

that most or all of the 14 late-breeding pairs 

were not among the original 18 pairs. Where 

might these new birds have come from? 

There are three other annually-occupied 

colonies of Cormorants within 30 km of Glas 

Eilean. The nearest is at Eilean Buidhe, 16 km 

to the south. In 2007, 22 pairs of Cormorants 

nested there but almost all failed. Human 

interference was suspected. The incubating 

adults were counted from the sea on 21 May 

and again on 11 June; no young were visible 

on either date. On 2 July counting ashore 

found 14 empty nests and three nests with 

small or medium-small young. On 23 July 

there were three large young. 

These facts strongly suggest that Eilean 

Buidhe was the source of the second wave of 

birds that bred at Glas Eilean. Not only was 

Eilean Buidhe the nearest Cormorant colony, 

but the numbers of birds involved and the 

timing of events were both consistent with 

this explanation (c. 20 pairs left Eilean Buidhe 

some time between 11 June and 2 July; 14 

SEABIRD 21 (2008) 

pairs started laying at Glas Eilean from late 

June to early July, estimated from clutches and 

small young there on 5 August). It remains 

unclear how these birds were able to 

reproduce successfully so late in the year, at a 

time when Cormorants in this area have 

normally stopped breeding and environmental 

cues are presumably unfavourable. 


We are grateful to Tom Callan for first alerting 

JCAC to the Cormorants breeding at Glas 

Eilean, for keeping him informed of their 

progress, and for assistance with ringing. We 

thank Jens Gregersen for his dedicated effort 

in observing breeding Cormorants at the Vorsø 

colony. We thank Robin Sellers, Martin 

Heubeck, Sarah Wanless and Mike Harris for 

considerable help and advice during writing. 

References 

Bregnballe, T. 1996. ‘Reproductive performance in 

Great Cormorants during colony expansion and 

stagnation’. PhD Thesis, National Environmental 

Research Institute, University of Aarhus, Denmark. 

Cadiou, B. 1994. Un évènement rarissime: l’élevage 

de deux nichées avec succès par un couple de 

cormorans huppés Phalacrocorax aristotelis. 

Alauda 62: 134–135. 

Demongin, L. 1993. Premières nidification du Grand 

Cormorant (Phalacrocorax carbo) sur la resérve de 

la Grande-Noé (Eure). Le Cormoran 8: 303–306. 

Gibson, J. A. 1958a. Notes on the breeding birds of 

the Lochgilphead Islands. Glasgow Bird Bulletin 7: 

67–68. 

Gibson, J. A. 1958b. The breeding birds of the small 

Clyde islands. Glasgow Bird Bulletin 7: 99–116. 

Gibson, J. A. 1969. Population studies of Clyde 

seabirds, part 1. Transactions of the Buteshire 

Natural History Society 17: 79–95. 

Gibson, J. A. 1985. Population Studies of Clyde 



Gibson, J. A. 1990. Population studies of Clyde 



Grieco, F. 1994. Fledging rate in the Cormorant 

Phalacrocorax carbo at the colony of Val Campotto 

(Po Delta, north-east Italy). Avocetta 18: 57–61. 

Wanless, S. & Harris, M. P. 1997. Successful 

double-brooding in European Shags. Colonial 

Waterbirds 20: 291–294.

Appendix 1. 

Description of Glas Eilean and events there 

in 2007: Glas Eilean (grid reference 

NR912857; 56°01’N 5°21’W) consists of two 

small islets about 350 m from the mainland in 

Loch Fyne, a long sea loch in southwest 

Scotland. Records since about 1950 show that, 

while large gulls have bred there for decades, 

Cormorants were not recorded breeding until 

2007. Shags bred for the first time in 2004 

(one or two pairs) and 2005 (two pairs) but 

none were recorded in 2006 (Gibson 1958a, 

1958b, 1985; Craik unpublished annual records 

1987–2007). 

Several seabird species bred on Glas Eilean in 

2007, mainly Herring Gull Larus argentatus (c. 

75 pairs) and Great Black-backed Gull L. 

marinus (c. ten pairs) together with smaller 

numbers of Shag (five pairs), Black Guillemot 

Cepphus grylle and Common Eider Somateria 

mollissima (at least two pairs of each, probably 

considerably more). The Cormorant colony 

was on the larger, southern islet, which 

measures c. 150 x 100 m. Observations at the 

Cormorant colony in 2007 were as follows. 

27 May: During the annual count ashore of 

gull clutches and nests, the new colony of 

Cormorants was not approached closer than c. 

100 m in a line of vision, as experience 

elsewhere has shown that large gulls quickly 

learn to prey on exposed eggs and small young 

of Shags and Cormorants. On this occasion, 

the sitting Cormorants did not leave their 

nests and an incomplete count of ten 

incubating adults was made from the sea. 

11 June: The Cormorant colony was entered. 

There was one disused nest. The 18 active 

nests all held large, mobile young, and 

(unusually) there were no eggs or small young. 

Thirty-seven young were ringed and c. ten 

were too large to catch. Thus c. 47 young are 

estimated to have fledged from 18–19 pairs in 

this first laying. 

21 July: On a routine visit to count large 

young gulls, the ringed young Cormorants 

were flying in the area, settling on tidally 

exposed rocks and at the colony edge (as 

expected); at least 19 were counted. 

Short Notes 

Unexpectedly, however, at least 12 adult 

Cormorants were seen incubating on nests in 

the same nesting area. The colony was again 

not approached, for the above reason. 

5 August: The colony was entered and nest 

contents were recorded. There were 14 nests 

(two with eggs and 12 with small young, none 

large enough to ring). Continuous observation 

after departure revealed no predation by gulls 

before the return of the adult Cormorants. 

14 August: Seven nests held 15 young that 

were large enough to ring.The other seven held 

15 young too small to ring, including two runts. 

27 August: Thirteen more chicks were ringed. 

Two of the 15 (the two runts?) were not found 

and had probably died, although a search for 

remains was not made. Thus 15 + 13 = 28 

young are believed to have fledged in early to 

mid September from the 14 late-nesting pairs. 

JCAC has ringed Cormorant chicks in this part 

of west Scotland for over 20 years and, later 

each summer, often visited many of the same 

sites to count and ring chicks of other species. 

After the normal midsummer fledging of most 

Cormorant chicks, it is not unusual to find a 

few late-nesting Cormorant pairs with eggs or 

small young in early July. However, there are 

no local records of an entire colony with many 

young fledging together in September. In 

2007, JCAC monitored six other Cormorant 

colonies in the area, all on small islands along 

the mainland coast between Kintyre and 

Mallaig, but none of these behaved in this way. 

One failed completely as a result of mink 

predation of eggs; the colony at Eilean Buidhe 

raised almost no young, apparently because of 

human interference and/or disturbance 

allowing predation of eggs and young by large 

gulls; and four colonies reared single broods 

that fledged in midsummer in the normal way. 

That the breeding pattern at Glas Eilean in 

2007 was atypical was confirmed by Robin 

Sellers, National Organiser of the UK 

Cormorant Breeding Colony Survey (R. M. 

Sellers pers. comm.). 

SEABIRD 21 (2008) 

97

98 

Short Notes 

A pilot study of the 

phenology and breeding 

success of Leach’s Stormpetrel 

Oceanodroma 

leucorhoa on St Kilda, 

Western Isles 

Money, S., 1 *, Söhle, I. 2 & Parsons, M 2 . 


Email: catharacta7@yahoo.co.nz 

1 Raintree House, Church Lane, Drayton St 

Leonard, Oxfordshire OX10 7AU, UK; 2 Joint 

Nature Conservation Committee, 7 Thistle 

Place, Dunnet House, Aberdeen AB10 1UZ, UK. 

Abstract 

From June to September 2007, 27 occupied 

Leach’s Storm-petrel Oceanodroma leucorhoa 

burrows on St Kilda were investigated by 

endoscope to estimate phenology and 

breeding success. From 17 burrows that could 

be followed to chick development, a minimum 

productivity of 0.59 young per egg laid was 

estimated, similar to other published figures of 

breeding success from the northwest Atlantic. 

Similar life cycle timings to those of Leach’s 

Storm-petrels in the northwest Atlantic were 

revealed, with egg laying in late May to mid 

June, hatching in mid to late July and fledging 

from mid September onwards. Implications for 

conservation and monitoring are discussed, 

and guidance for future studies suggested. 

Introduction 

The St Kilda archipelago (57°49’N, 08°35’W), 

66 km west of the Western Isles, holds a 

population of around 45,000 pairs of Leach’s 

Storm-petrels Oceanodroma leucorhoa 

which represents 94% of the British and Irish 

population (Mitchell 2004). Recent work in 

the northwest Atlantic by Huntington et al. 

(1996) and A. Hedd (unpubl.) has been carried 

out on the breeding biology of Leach’s 

Storm-petrel, but little is known about their 

breeding habits within Britain, and little has 

been published since that of Ainslie & 

Atkinson (1937) on the Flannan Isles and 

North Rona, apart from some exploratory 

SEABIRD 21 (2008) 

work on Gruney, Shetland in the 1990s (Ellis 

et al. 1998). In addition, it has been suggested 

that the Leach’s Storm-petrel breeding 

population on St Kilda may have declined 

significantly since the last full census in 

1999/2000 (Mitchell 2004; Newson et al. 

2008), and studies on breeding success may 

therefore give insights as to the reasons for 

this decline. 

Methods 

In 2007, 27 Leach’s Storm-petrel burrows 

which were occupied (i.e. a response was 

obtained from playing a tape of a male ‘St 

Kildan’ chatter call in mid June and nest 

material was visible) and which were 

accessible by endoscope (Everest VIT PXL 

Videoprobe) were located and marked with 

two short bamboo canes with a tape flag. 

Burrows were situated on grassy slopes, under 

rocks and in cleits (man-made stone 

structures unique to St. Kilda), and were 

monitored by endoscope a total of 12 times 

throughout the breeding season. Burrow 

occupancy was determined on seven days 

during peak incubation time between 19 and 

27 June, on three days during hatching and 

early chick stage (24 and 26 July, and 8 

August) and twice during the late chick stage 

(on 10 and 26 September). Of the 27 initially 

occupied study burrows, eight (30%) did not 

contain an egg and therefore were presumably 

occupied by immature or prospecting birds. 

Eggs were laid in 19 burrows, two of which 

could not be relocated due to marker loss, 

leaving 17 burrows that were used in 

estimates of breeding success (Table 1). 

Results 

Phenology: The presence of unattended small 

chicks and incubating adults on 24 and 26 July, 

but only unattended chicks on 8 August (Table 

1), suggested that hatching took place 

between mid and late July (assuming that 

chicks are brooded for an average of five days 

(Cramp & Simmons 1977)). Given an 

incubation period of six weeks and a fledging 

period of 9–10 weeks (Cramp & Simmons 

1977) this would suggest a laying period from 

around late May to mid June and fledging 

from mid September. Indeed, our observations

showed that all study birds had laid their eggs 

by mid June and the majority of the chicks 

fledged between 10 and 26 September, with 

the remaining chicks presumably fledging 

after 26 September (Table 1). In 

Newfoundland, Canada, egg laying of Leach’s 

Storm-petrel begins in late May, the mean 

hatching time is late July and fledging begins 

in mid September (Huntington et al. 1996; A. 

Hedd, unpubl.), suggesting that the phenology 

in these two geographically distinct 

populations is very similar. A study of nine 

burrows investigated by endoscopy on St Kilda 

in 2003 (O’Brien et al. 2005) found the first 

Short Notes 

hatching to be on 9 July and a further three 

had hatched by 14 July, which indicates a 

similar (perhaps slightly earlier) timing to that 

presented here, and work carried out on 

Gruney in Shetland (Ellis et al. 1998) also 

found chicks on a slightly earlier date (6 July) 

than in the current study. 

Breeding success: Out of 17 burrows in which 

an egg was recorded, ten were confirmed to 

have chicks by 10 or 26 September (Table 1). 

Of the remaining seven burrows, four failed at 

the egg stage, two stone cavity nesters 

(burrows 19 and 20) had ambiguous results 

Table 1. Leach’s Storm-petrel Oceanodroma leucorhoa burrow contents as determined by endoscopy. 

19 20 22 23 24 25 27 24, 26 8 10 26 

Jun Jun Jun Jun Jun Jun Jun Egg Jul Aug Sep Sep Summary 

1 0 0 0 0 0 0 0 0 No egg laid 

2 R R R 1,E 1 1,E R Y A,E SC LC 0 Presumed fledged 

3 1 1 R R 1,E 1,E 1 Y SC MC 0 Presumed fledged 

4 R 1 1 0,E 0,E 1,E 0,E Y 0 E Abandoned egg 

5 1 R R R 0 1,E R Y SC LC 0 Presumed fledged 

6 R 0 0 1 0 0 0 0 No egg laid 

7 R R 1,E R R 1,E R Y SC MC 0 Presumed fledged 

8 R R R R R 1,E R Y 0 0 Failed to hatch 

9 1,E R R R 1 1,E 1 Y A,BE,SC? SC LC LC Presumed fledged 

10 0 0 0 0 0 0 0 0 No egg laid 

11 0 0 R R R 1,E 1,E Y A E Abandoned egg 

12 0 R R R R 1,E 1 Y SC 0 0 Chick died/fledged early? 

13 1 R R R 1 1,E 1 Y ? Burrow not relocated 

14 R R R 0,E 0,E 0,E R Y 0 Failed to hatch 

15 R R R 1,E 1,E 1,E R Y SC LC LC Presumed fledged 

16 R R 0,E 1,E 1 1,E R Y A,SC LC 0 Presumed fledged 

17 R R 0 R 0 0 0 ? 0 No egg laid 

18 1 1 0,E R R 1,E 1 Y SC MC LC Presumed fledged 

19 R R R R 1 1,E 1 Y A BE 0 0 Failed to hatch? 

20 1 0 1 1 R 1,E R Y A BE BE BE Failed to hatch? 

21 1 0 1 1,E 1 1,E 1 Y BE BE MC 0 Presumed fledged 

22 1,E R 0,E 1,E 1 1,E R Y ? Burrow not relocated 

23 R 1 0 R R 1 R ? Burrow not relocated 

24 0 R 0 0 R 1 R 0 No egg laid 

25 0 0 0 0 0 0 0 0 No egg laid 

26 0 0 0 0 0 0 0 0 No egg laid 

27 0,E R R 0,E 1 1,E R Y A SC MC LC Presumed fledged 

R = Response, no endoscope Y = Yes BE = Broken eggshell 

0 = No response, no adult 0 = Empty burrow SC = Small unattended chick 

1 = Adult present A = Adult in burrow MC = Medium chick 

E = Egg seen A,E = Adult sitting on egg LC = Large, well-feathered chick 

SEABIRD 21 (2008) 

99

100 

Short Notes 

Figure 1. Sarah Money using endoscope on Hirta, summer 2006. 

Figure 2. Leach’s Storm-petrel Oceanodroma leucorhoa, Hirta, St 

Kilda, September 2007 © Will Miles. 

SEABIRD 21 (2008) 

due to the possible retreat of the chick out of 

view from the endoscope (which happened 

once in the case of one of the successful 

burrows) and one (burrow 12) held a chick 

which may have fledged before the first check 

on 10 September (fledged Leach’s Stormpetrels 

have been observed on St Kilda by this 

date in previous years: National Trust for 

Scotland, unpubl.). Thus, the minimum 

estimate of productivity was 0.59 young per 

egg laid, comparable with published figures of 

breeding success of 0.48–0.73 from the 

northwest Atlantic (Huntington et al. 1996), 

and suggesting that Leach’s Storm-petrels on 

St Kilda might not be currently experiencing 

major breeding problems. Of interest was the 

fact that during incubation a number of 

burrows were observed to hold unattended 

eggs for periods of up to three days. Three out 

of these six burrows (those where the eggs 

were left unattended for single days at a time) 

went on to produce chicks, which suggests that 

leaving the egg unattended for a single day 

does not necessarily prevent hatching. Indeed, 

this phenomenon has been observed for a wide 

range of petrel species (Warham 1990). 

Discussion 

This pilot study gives a useful insight into the 

breeding success and timing of breeding of 

Leach’s Storm-petrel on St Kilda in 2007. It 

should be emphasised that this work was of an 

exploratory nature, rather than a detailed 

study of productivity and phenology, but can 

nonetheless be used for comparison with 

future studies both on St Kilda and elsewhere 

in Britain. Such work on Leach’s Storm-petrels 

is of necessity always going to be difficult due 

to the remote locations involved and 

limitations on the use of endoscopes imposed 

by the weather. However, lessons to be 

learned which would refine monitoring 

techniques for future studies include: 

� burrow marker loss could be prevented 

through the use of GPS and improved 

marking; 

� increasing the sample size of burrows and 

the spread and frequency of monitoring 

dates would help determine breeding cycle 

dates and success more accurately, and;

� only grassy slope burrows instead of stone 

cavities should be monitored to prevent 

ambiguous outcomes from nesting chambers. 

Information on the phenology of breeding is 

critical to the correct application of methods 

for monitoring breeding numbers of stormpetrels. 

The favoured technique of tape 

playback (Ratcliffe et al. 1998) should be 

undertaken during the peak period of 

incubation to maximise the likelihood of 

burrow occupancy at the time of survey; 

therefore, an understanding of phenology and 

its year-to-year variability has important 

practical applications. 

Currently there are concerns about the conservation 

status of Leach’s Storm-petrel on St 

Kilda. Surveys in 1999 and 2003 suggested a 

48% decline in the population on Dun, an island 

within the St Kilda archipelago (Mitchell 2004), 

although a further survey in 2006 suggested 

that this rapid decrease may have ceased 

(Newson et al. 2008). Predation by Great Skuas 

Stercorarius skua is suggested as a potential 

explanation for the decline (Phillips et al. 1999; 

Mitchell 2004; Votier et al. 2005). However, 

alternative factors, such as food shortages, 

which currently affect a wide range of Britain’s 

seabird species, may also have played a role. 

Our pilot study suggests that shortage of 

suitable food did not affect Leach’s Stormpetrel 

chicks on Hirta, St Kilda during the 2007 

breeding season, as breeding success was quite 

high, but caution should be applied in 

interpreting these results as our sample size was 

small. However, studies of breeding success 

(combined with those of adult survival) could 

contribute to an early warning of potential 

population change, and it would therefore be 

appropriate to establish a long-term productivity 

study for this species on St Kilda. 


This work was supported by the National Trust 

for Scotland and the Joint Nature Conservation 

Committee, Aberdeen. We also thank Stuart 

Murray for encouraging this publication, April 

Hedd for allowing access to unpublished data, 

and two anonymous referees for improving an 

earlier draft of the manuscript. 

Short Notes 

References 

Ainslie, J. A. & Atkinson, R. 1937. On the breeding 

habits of Leach’s Fork-tailed Petrel. British Birds 30: 

234–48. 

Cramp, S. & Simmons, K. E. L. (eds.) 1977. The Birds 

of the Western Palearctic. Vol. I. Oxford University 

Press, Oxford. 

Ellis, P., Ratcliffe, N. & Suddaby, D. 1998. Seasonal 

variation in diurnal attendance and response to 

playback by Leach’s Petrels Oceanodroma 

leucorhoa on Gruney, Shetland. Ibis 140: 336–339. 

Huntington, C . E., Butler, R. G. & Mauck, R. A. 

1996. Leach’s Storm-petrel (Oceanodroma 

leucorhoa). In: Poole, A. & Gill, F. (eds.) The Birds of 

North America: 233. Birds of North America Inc., 

Philadelphia, PA & American Ornithologists’ Union, 

Washington DC. 

Mitchell, P. I. 2004. Leach’s Storm-petrel Oceanodroma 

leucorhoa. In: Mitchell, P. I., Newton, S., Ratcliffe, N. 

& Dunn,T. E. (eds.) Seabird Populations of Britain and 

Ireland: 101–114. Poyser, London. 

Newson, S. E., Mitchell, P. I., Parsons, M., O’Brien, 

S. H., Austin, G. E., Benn, S., Black, J., Blackburn, 

J., Brodie, B., Humphreys, E., Leech, D. I., Prior, M. 

& Webster, M. 2008. Population decline of 

Leach’s Storm-petrel Oceanodroma leucorhoa 

within the largest colony in Britain and Ireland. 

Seabird 21: 77–84. 

O’Brien, S. H., Mitchell, P. I., Parsons, M. and 

Mavor, R. A. 2005. ‘Seabird Monitoring on St Kilda 

in 2003’. Unpublished JNCC Report, Aberdeen. 

Phillips, R. A., Thompson, D. R., & Hamer, K. C. 

1999. The impact of Great Skua predation on 

seabird populations at St Kilda: a bioenergetics 

model. Journal of Applied Ecology 36: 218–232. 

Ratcliffe, N., Vaughan, D., Whyte, C. & Shepherd, 

M. 1998. Development of playback census 

methods for Storm Petrels Hydrobates pelagicus. 

Bird Study 45: 302–312. 

Votier, S. C., Crane, J. E., Bearhop, S., De Leon, A., 

McSorley, C. A., Minguez, E., Mitchell, P. I., 

Parsons M., Phillips, R. A. & Furness, R. W. 2005. 

Nocturnal foraging by Great Skuas Stercorarius 

skua: implications for conservation of storm-petrel 

populations. Journal of Ornithology 147: 405–413. 

Warham, J. 1990. The Petrels: Their Ecology and 

Breeding Systems. Academic Press, London. 

SEABIRD 21 (2008) 

101

102 

Short Notes 

Use of gulls rather 

than terns to evaluate 

American Mink 

Mustela vison control 

Craik, J. C. A. 

Email: clive.craik@sams.ac.uk 

Scottish Association for Marine Science, 

Dunstaffnage Marine Laboratory, Oban, Argyll 

PA37 1QA, UK. 

Ratcliffe et al. (2006) compared the breeding 

biology of terns Sterna spp. on the Uists, 

Western Isles, where American Mink Mustela 

vison (hereafter ‘mink’) had been removed, 

and on nearby Lewis, where there had been no 

mink control. They showed that hatching 

success was significantly higher on the Uists, 

indicating that mink removal improved this 

aspect of breeding, but found no significant 

difference in colony productivity between the 

two areas. The years in which the comparisons 

were made, 1993 and 2005, were years of low 

tern productivity, probably caused by poor 

weather and food shortage, respectively. The 

authors considered that such factors overrode 

any effect of mink by killing young terns 

before mink could take them, and suggested 

SEABIRD 21 (2008) 

that any effects of mink would be more 

detectable in years of high tern productivity 

when more young terns survived. 

This is a problem I have encountered in 

another part of Scotland – on small islands in 

the sealochs and sounds of mainland Argyll, 

including Mull and the adjacent part of 

Highland Region (hereafter ‘Argyll’). For 

almost 20 years, we have been removing mink 

from near tern colonies and comparing 

productivity with similar colonies in 

unprotected areas, using broadly the same 

field methods as Ratcliffe et al. (2006). There 

have been years when mink removal led to 

marked improvement of tern productivity, but 

there were also years when other factors 

severely reduced tern productivity and mink 

removal had no detectable effect. Ratcliffe et 

al. (2006) and my own work demonstrate that 

tern breeding performance is so variable that 

it is of limited value as an indicator of the 

success of mink control. If mink removal is to 

be justified, we must be able to show that it 

has immediate measurable benefits. 

Showing that mink are absent in trapped areas 

by searching for signs such as scats and dens 

can be difficult, particularly since any mink are 

likely to be present at low density. Moreover, 

Table 1. Common Gull Larus canus productivity in areas with and without American Mink Mustela vison 

control in Argyll and Lochaber. 

Areas with mink control Areas without mink control 

No.of No. of Productivity No. of No. of Productivity % that mink 

pairs (No. of fledged (chicks/pair), pairs (No. fledged (chicks/pair), reduced 

Year colonies) young ‘a’ of colonies) young ‘b’ productivity 

by = (a-b)/a 

1996 477 (6) 385 0.807 460 (24) 84 0.182 77 

1997 547 (10) 378 0.691 577 (20) 230 0.399 42 

1998 700 (10) 612 0.874 357 (9) 134 0.375 57 

1999 828 (18) 510 0.616 506 (26) 132 0.261 58 

2000 682 (13) 490 0.718 764 (36) 281 0.368 49 

2001 899 (19) 760 0.845 471 (31) 182 0.386 54 

2002 760 (13) 508 0.668 375 (14) 89 0.237 65 

2003 652 (18) 383 0.587 584 (36) 220 0.377 36 

2004 761 (15) 504 0.662 285 (10) 138 0.484 27 

2005 636 (8) 703 1.105 493 (16) 138 0.28 75 

2006 715 (18) 704 0.985 544 (31) 134 0.246 75 

2007 552 (9) 258 0.467 455 (17) 118 0.259 45

Short Notes 

Table 2. Herring Gull Larus argentatus productivity in areas with and without American Mink Mustela vison 

control in Argyll and Lochaber. 

Areas with mink control Areas without mink control 

No.of No. of Productivity No. of No. of Productivity % that mink 

pairs (No. of fledged (chicks/pair), pairs (No. fledged (chicks/pair), reduced 

Year colonies) young ‘a’ of colonies) young ‘b’ productivity 

by = (a-b)/a 

1997 698 (6) 703 1.01 1709 (14) 809 0.473 53 

1998 754 (6) 805 1.07 2507 (25) 860 0.343 68 

1999 1700 (7) 1373 0.808 6492 (42) 2465 0.38 53 

2000 1637 (8) 1978 1.21 6905 (48) 5427 0.786 35 

2001 1386 (7) 1145 0.826 6791 (46) 4266 0.628 24 

2002 1294 (9) 1198 0.926 5571 (38) 3639 0.653 29 

2003 1521 (10) 1174 0.77 5675 (50) 3298 0.581 25 

2004 1650 (10) 1473 0.893 2183 (11) 1159 0.531 41 

2005 1031 (9) 887 0.86 2929 (20) 1557 0.5316 38 

2006 2198 (11) 2299 1.05 3066 (23) 2064 0.67 36 

2007 1073 (11) 798 0.744 3337 (28) 2477 0.742 0.3 

at any time, more mink may arrive from 

untrapped areas nearby (Birks 1986). Arguably, 

the only practical and effective way to 

demonstrate that local mink control has been 

successful is to show that it has prevented the 

damage that mink were causing. 

Mink can cause widespread breeding failure of a 

variety of ground-nesting birds, as well as terns. 

Since tern breeding success is a poor indicator of 

the success of mink control, breeding success of 

other species may serve the purpose better. In 

Argyll, we have measured the annual breeding 

success of Common Gull Larus canus and 

Herring Gull L. argentatus colonies since 1996 

and 1997, respectively, both in areas where 

mink have been removed (usually to protect 

terns) and in areas with no mink control. New 

methods have been developed for this purpose, 

allowing simple and rapid determinations of gull 

productivity (Craik 2000) and more efficient 

control of mink (Craik 2008). The study area 

includes most colonies along the mainland 

coast between Mallaig in the north and West 

Loch Tarbert in the south. In areas without mink 

control, in any year there were some island 

colonies with good productivity that were 

naturally free of mink for various reasons 

(chiefly distance from the mainland shore), and 

some colonies with zero or low productivity, 

some of which were affected by mink; this 

applied to both gull species. 

In each of the years 1996–2007, the productivity 

(total number of fledged young /total 

number of pairs in a group of colonies) of 

Common Gulls was considerably higher in the 

mink-free group than in the mink-affected 

group (Table 1). This was also true for Herring 

Gulls in each of the years 1997–2006. The 

exception was in 2007 when productivity was 

almost identical in the two groups of Herring 

Gulls (Table 2). This seems to have been 

because, in 2007, some Herring Gull colonies 

lost many or most eggs and young during a 

storm-tide in mid May, while some others 

were more than usually affected by Brown 

Rats Rattus norvegicus. 

Hence, gull productivity provides a much 

more reliable indicator of the success of mink 

control than the breeding success of the terns, 

even when the mink control may be directed 

primarily at conserving tern colonies. 

Footnote: Ratcliffe et al. (2006) state in their Discussion: ‘For example, in south-west Scotland between 1990 and 2006, 58% of 

unprotected tern colonies were not mink-affected (J. C. A. Craik, unpubl.)’, ‘unprotected’ here meaning those with no mink control. 

While 42% of unprotected colonies showed definite signs of mink predation, the other 58% included some colonies which had low 

or zero productivity for unknown reasons, and where the possibility of predation by mink could not be excluded.Thus their statement 

should be rephrased: ‘At least 42% of unprotected colonies are known to have been attacked by mink.’ 

SEABIRD 21 (2008) 

103

104 

Short Notes 


I am most grateful to Norman Ratcliffe, Ian 

Mitchell and Martin Heubeck for comments and 

corrections to earlier drafts of this note, and to 

the many people in the Argyll area who have 

helped with mink control and seabird counting 

over the years. I particularly thank Rob, Audrey 

and Niall Lightfoot and the late Ian Hynd. 

Use of gulls rather than 

terns to evaluate American 

Mink Mustela vison control. 

A response to Craik (2008) 

Ratcliffe, N. 

Email: notc@bas.ac.uk 

British Antarctic Survey, Natural Environmental 

Research Council, High Cross, Madingley Road, 

Cambridge CB3 0ET, UK. 

I thank Clive Craik for the opportunity to 

clarify points made in the 2006 paper. Clive 

presents a convincing argument that gull 

rather than tern productivity was a better 

indicator of the benefits of mink control 

during his study. However, I maintain that I 

was justified in restricting the analysis to terns 

in my study for three reasons: 

The Hebridean Mink Project was designed to 

produce conservation benefits for a range of 

taxa that included terns, but not gulls. 

Therefore I was obliged to test the efficacy of 

the project in terms of changes in the 

breeding success of terns, since that of gulls 

was irrelevant to the project’s objectives. 

In contrast to Clive’s study area, gulls and terns 

on the Western Isles occupy discrete habitats 

that differ with respect to mink predation risk. 

Gulls nest inland on moors where mink are rare, 

while terns nest on the coast where mink are 

more common.As such, gull productivity on the 

SEABIRD 21 (2008) 

References 

Birks, J. 1986. Mink. The Mammal Society, Anthony 

Nelson, Oswestry. 

Craik, J. C. A. 2000. A simple and rapid method of 

estimating gull productivity. Bird Study 47: 113–116. 

Craik, J. C.A. 2008. Sex ratio in catches of American 

Mink – how to catch the females. Journal for 

Nature Conservation 16: 56–60. 

Ratcliffe, N., Houghton, D., Mayo, A., Smith, T. & 

Scott, M. 2006. The breeding biology of terns on 

the Western Isles in relation to mink eradication. 

Atlantic Seabirds 8: 127–135 [published April 2008]. 

Western Isles will be less sensitive to removal of 

mink than that of terns. 

An analysis of Clive’s own data demonstrated 

that tern productivity during 1998–2006 at 

colonies protected from mink was on average 

253% higher than that at unprotected ones 

(Ratcliffe et al. 2008).As such, tern productivity 

clearly has a greater value as an indicator of the 

effects of mink removal than Clive’s 

commentary suggests. 

I concede the point in the Footnote 

concerning the detectability of mink 

predation being less than one. Quantifying the 

likelihood of an unprotected colony escaping 

mink predation is therefore difficult, but data 

certainly show that unprotected colonies can, 

on occasion, escape predation and experience 

high productivity. Hence, detecting the effects 

of mink control statistically requires sampling 

at a large number of colonies, and certainly 

more than the two sampled on Lewis in 1992. 

Reference 

Ratcliffe, N., Craik, C., Helyar, A., Roy, S. & Scott, 

M. 2008. Modelling the benefits of mink 

management options for terns in West Scotland. 

Ibis 150 (Suppl. 1): 114–121.

Fish brought to 

young Atlantic Puffins 

Fratercula arctica on 

Burhou, Channel Islands 

Sanders, J. G. 

Email: jerry.sanders@cwgsy.net 

The Alderney Ornithological Group, PO Box 24, 

Alderney GY9 3AP, UK. 

Introduction 

During the last 100 years breeding numbers of 

Atlantic Puffins Fratercula arctica (hereafter 

‘Puffins’) in the Channel Islands and Brittany, 

the most southern colonies in the eastern 

Atlantic, have declined dramatically and in 

2000 totalled less than 600 pairs (Harris & 

Wanless 2004). Most colonies now have trivial 

numbers and that on Burhou, Alderney, 

Channel Islands is one of only two remaining 

probably viable breeding populations south of 

the English Channel, and is therefore regionally 

of considerable conservation importance. 

Short Notes 

occupied in three subcolonies in 2005 (L. 

Soanes pers. comm.), and a maximum of c.250 

individuals was present in early July 2006 and 

2007 (pers. obs.). 

Between 1969 (Operation Seafarer) and 2005, 

breeding Lesser Black-backed Gulls Larus fuscus 

on Burhou increased from 70 to 1,102 pairs and 

Herring Gulls Larus argentatus from 87 to 202 

pairs, while c.15 pairs of Great Black-backed 

Gulls Larus marinus also presently breed (L. 

Soanes pers. comm.). Interactions with Puffins 

include gulls loafing near Puffin burrows, 

kleptoparasitism by Lesser Black-backed and 

Herring Gulls, and predation by Great Blackbacked 

Gulls, which became acute in the 1970s 

and 1980s but diminished after 1996 (Lockley 

1953; Sanders 2007). 

Nothing appears to have been known of the diet 

of Puffin chicks at Burhou and this note presents 

the results of an attempt to remedy this. 

Burhou (49o44’ N, 2o16 W) in the Alderney 

group is the most northerly Channel Island, 

about 760 m long by 300 m wide, 18 hectares 

in area, and 21 m above sea level at its highest 

point. Rabbits Oryctolagus cuniculus are 

common, having been introduced by the 14th 

century (Coysh 1985), the vegetation is short, 

but no other mammals are present. The island 

is uninhabited and now closed to the public 

from mid March to late July. Puffin numbers 

were estimated at 50,000 pairs in 1948–1949 

(Lockley 1953), making it by far the largest 

colony on the southern side of the English 

Channel at that time. The population fell to 

c.350 individuals (but less than 100 breeding 

pairs) by 1980, through a combination of 

chronic oil pollution and major tanker sinkings, 

burrow trampling by unrestricted day-trippers, 

and changes in vegetation due to soil 

enrichment by gull faeces (Danchin & 

Cordonnier 1980; Hill 1989; Sanders 2007, 

2005). The population almost stabilised after 

access to the island became restricted in the 

Figure 1. Puffin Fratercula arctica with Gadidae, Burhou, July 2006. 

early 1980s (Sanders 2007), 120 burrows were Figure 2. Puffin Fratercula arctica with sandeels, Burhou, July 2007. 

SEABIRD 21 (2008) 

105

106 

Short Notes 

Table 1. Fish carried by Atlantic Puffins Fratercula arctica on 

Burhou, 2006 & 2007. 

Date in Time of No. fish Species 

2006 photo in load of fish 

01 July 12.22 5 Gadidae sp. 


02 July 07.58 12 Small sandeels 










02 July 11.47 4 2 small sandeels, 2 Gadidae sp 





Total 81, of which: Small sandeels 25 

Gadidae sp. 56 

Date in Time of No. fish Species 

2007 photo in load of fish 

09 July 13.13 6 Large sandeels 

09 July 13.13 6 Mixed-size sandeels 

09 July 13.43 8 7 Large sandeels, 1 Gadidae sp. 






11 July 09.58 3 2 Gadidae sp,, 1 small sandeel 





Total 82, of which: Small sandeels 20 

Mixed-size sandeels 6 

Large sandeels 53 

Gadidae sp. 3 

SEABIRD 21 (2008) 

Methods 

The usual method of assessing Puffin chick diet 

is to catch adults carrying fish as they come 

ashore. However, the Burhou population is very 

small and since mist-netting in the 1980s 

appeared to cause undue disturbance to the 

main colony (pers. obs.; M. G. Hill pers. comm.), 

it was decided instead to photograph Puffins 

carrying fish, a somewhat novel method. 

On arrival at the main colony Puffins carrying 

fish usually circled a number of times partly 

over the burrow area and partly over the sea, 

and if there were loafing gulls near the 

burrows, as was often the case, they 

frequently landed on the water in the bay 

adjacent to the colony, from which they could 

see the location of their burrow. These rafting 

Puffins were photographed on 1 & 2 July 2006 

and 9 & 11 July 2007 using a digital Nikon 

Coolpix 995 camera with a Leica APO 77 

spotting telescope as lens. Taking the photos 

may in some cases have kept the Puffins away 

from their burrows for a slightly longer period 

than usual, but at other times it was possible 

to photograph without observer disturbance. 

Of many hundreds of photographs taken, only 

those clearly showing the fish carried by 

individual Puffins were selected for this study. 

To prevent double-recording, identification of 

individual Puffins in the photographs was 

ensured by comparing the number, size, 

position and distribution in the bill of the fish 

carried by each Puffin. Fish were identified to 

species by M. P. Harris, and their size and 

number were noted. This was helped by there 

usually being several photos of the same bird. 

Results 

Photographs taken in 2006 showed 17 Puffins 

with a load of fish. Two were carrying only 

small sandeels Ammodytes sp. (probably 0group, 

young of the year), 14 carried fish of 

the cod family Gadidae, most likely Whiting 

Merlangius merlangus, and one load contained 

some of each.Average load was 4.8 fish, and of 

the 81 fish in total, 25 (31%) were sandeels, 

and 56 (69%) were gadids. (Table 1).

In the 2007 photographs, three and seven of 

the 13 Puffins were carrying only small and 

large sandeels, respectively. One Puffin had 

seven large sandeels and one probable gadid, 

one carried a mixture of small and large 

sandeels (mixed-size sandeels), and one 

carried a small sandeel and two gadids. The 

average load (6.3 fish) was not significantly 

greater than in 2006 (t = -1.617, P = 0.117). 

Of the total of 82 fish, 20 (24%) were small 

sandeels, 53 (64%) were large sandeels, six 

(7%) were mixed-size sandeels, and three 

(4%) were gadids. On 23 July, all 25 loads seen 

clearly through a telescope appeared to 

contain only sandeels, some of which were 

large (M. P. Harris & S. Wanless pers comm.). 

Discussion 

Puffins returning to the colony with fish for 

their chick normally enter their burrow as 

quickly as possible to prevent the catch being 

stolen (Harris 1984). Where they are prevented 

from doing so by gulls loafing near the burrow 

entrance and have to stand around for a while, 

or in the case of Burhou raft on the sea, digital 

photography can be a useful and non-invasive 

tool for assessing chick diet, given that the fish 

are large and reasonably easy to identify. It 

does, however, have its limitations on Burhou in 

that calm conditions are needed to get good 

pictures. Of the three days that photographs 

were taken in 2007, one day produced no 

useable photographs due to the movement of 

the birds in rough seas. 

All Puffins photographed appeared to be 

carrying as many fish as they could cope with. 

All the fish were of good size, and the overall 

size of the loads was at least as good as is 

currently seen at most northern British 

colonies (M. P. Harris pers. comm.). Generally, 

gadids are not considered as good food for 

young seabirds since they are generally of 

lower calorific value than other more oil-rich 

species such as sandeels, and Sprats Sprattus 

sprattus and Herring Clupea harengus (Harris 

& Hislop 1978). However, the gadids 

photographed in 2006 were substantially 

larger than those normally fed to Puffin chicks 

and this will at least in part have compensated 

for their low energy density. 

Short Notes 

Chick diet at a colony can change dramatically 

within a season and between years (Harris 

1984), and the difference between 2006 

(predominantly gadids) and 2007 (predominantly 

sandeels) at Burhou was striking.Around 

Guernsey, 28 km from Burhou, where there is a 

small-scale fishery for sandeels, fishermen 

reported the sandeel cycle of 2006 to be late, 

and abundance the lowest for a long time, with 

none caught on some traditionally productive 

grounds (M. Roger pers. comm.). This would 

suggest that sandeel abundance around Burhou 

in 2006 was also lower than in most years, and 

the 2007 observations give rise to some 

optimism on the state of the recent summer 

food supply for Puffins on Burhou. 


Thanks are due to the Burhou Warden, John 

Dupont for arranging the visits to Burhou, to 

Pierre Dupont for transport, to Mike Roger for 

permission to use the photographs, to Mike 

Harris for help with the manuscript and for 

identifying the fish in the photographs, to 

Louise Soanes for Puffin counts, and to Mark 

Atkinson, Alderney County Bird Recorder, for 

other assistance. 

References 

Coysh, V. 1985. Channel Islets. The Guernsey Press, 

Guernsey. 

Danchin E. & Cordonnier P. 1980. ‘Rapport 

d’Activité,Aurigny (Alderney) Iles Anglonormandes.’ 

Unpublished Report, C.R.B.P.O., Museum de Paris. 

Harris, M.P. 1984. The Puffin. Poyser, Calton. 

Harris, M. P. & Hislop, J. R. G. 1978. The food of 

young Puffins Fratercula arctica. Journal of Zoology, 

London. 185: 213-236. 

Harris, M. P. & Wanless, S. 2004. The Atlantic Puffin 

Fratercula arctica. In: Mitchell, I. P., Newton, S. F., 

Ratcliffe, N. and Dunn,T. E. (eds.) Seabird populations 

in Britain and Ireland: 392-406. Poyser, London. 

Hill, M. G. 1989. The Spread of Bracken Pteridium 

aquilinium in Burhou Island, Alderney from 1973 to 

1978. Report and Transactions of La Société 

Guernesiaise. 22: 628-632. 

Lockley, R. M. 1953. Puffins. Dent, London. 

Sanders, J. G. 2005. Alderney’s Breeding Seabirds. 

Report and Transactions of La Société Guernesiaise. 

25: 702-728. 

Sanders, J. G. 2007. The Birds of Alderney.The Press at 

St. Anne, Alderney. 

SEABIRD 21 (2008) 

107

108 

Reviews 

REVIEWS 

Petrels – night & day. A 

Sound Approach Guide By 

Magnus Robb, Killian 

Mullarney & The Sound 

Approach. The Sound 

Approach, Poole. 2008. ISBN 

978-90-810933-2-3. 300 

pages, numerous photos, 

maps, artwork and 

sonagrams, 2 CDs. Hardback, 

£34.95 plus p&p outside UK. 

Although I had commented on 

one chapter at the draft stage, I 

had little idea what to expect 

when the review copy of Petrels 

– night & day dropped on my 

desk. Certainly I had no inkling 

that, between the book’s covers, 

I would find an extraordinary 

combination of ripping yarns, 

beautiful artwork, two evocative 

CDs and scientific speculation 

that was well informed and bang 

up-to-date. 

How is this pulled off? Well, the 

bare bones are that, one by one, 

the 22 procellariiform species 

breeding in the northeast 

Atlantic, from Cape Verdes 

northwards and including the 

SEABIRD 21 (2008) 

Mediterranean, are given similar 

treatment; that is, an account of 

the authors’ visits to the species’ 

colonies, a quick resumé of 

breeding biology, a more detailed 

description of calls illustrated by 

CD tracks, and a discussion of 

whether the vocalisations 

suggest taxonomic revision 

might be called for. 

The sheer enthusiasm in the face 

of the misadventures of island 

hopping is a delight.‘We had seen 

photographs of boats used by 

birders before, and they looked 

big enough to carry a football 

team. Our boat was not only very 

small, but also appeared to be 

falling apart’, describes a trip to 

Raso to search out Cape Verde 

Shearwaters Calonectris 

edwardsii. Having made the same 

journey, possibly in the same 

boat, to study Raso Larks Alauda 

razae, I know the feeling. Or again 

in the Cape Verdes, but this time 

in pursuit of White-faced Stormpetrels 

Pelagodroma marina, an 

outboard could only be provoked 

into temporary life by the 

insertion of tubing. When the 

boatman and suspect motor 

abandoned Magnus Robb, taking 

his recording kit, he pleaded with 

fate. ‘The optimist in me hoped 

that [the boatman] was just 

going back to get another 

mouthful of tubing. I politely 

asked the pessimist in me not to 

think at all.’ 

Within this mesh of excitement 

emerge extraordinary facts that I 

did not know. For example the 

shearwaters of Menorca appear 

to be ill-understood hybrids of 

Puffinus mauretanicus and P. 

yelkouan. Quite amazingly, the 

Atlantic’s very first Swinhoe’s 

Storm-petrel Oceanodroma 

monorhis, ringed on the Salvages 

in 1983, was re-caught in August 

2007. Incidentally, the fact that 

that news is included in the text 

of this book, published early in 

2008, illustrates a speed of 

publication that shames the 

larger, established publishers. 

The photographs are of a consistently 

high standard. Some are 

outstanding: a White-faced 

Storm-petrel dancing off the sea 

surface, a Leach’s Storm-petrel O. 

leucorhoa sliding down the face 

of a spume-strewn yellow-brown 

wave breaking off the English 

coast. No less pleasing are Killian 

Mullarney’s plates, essentially one 

per species.While the plates show 

features crucial to identification, 

and the text captions are clearly 

aimed at the hardcore birder, the 

jizz of the birds is superbly 

captured and the plates just look 

aesthetically right. 

Dipping into the CDs was a treat. 

I reminded myself of the 

difference between Madeiran O. 

castro and Leach’s Storm-petrels. 

I learnt the sound of Bulwer’s 

Petrel Bulweria bulwerii, a species 

I have never heard. Perhaps the 

ideal listening arrangement 

would be a foul winter’s night, 

the wind lashing on the 

windows, the log fire subsiding, 

and the chance to let gruff 

female and higher-pitched male 

Cory’s Shearwater C. diomedea

calls drag me on a magic 

acoustic carpet to a balmy 

Macaronesian evening. 

Reviews such as this almost 

routinely carry a ‘but’. My but is 

small (so to speak). The authors 

place great weight on the use of 

calls as an aid to taxonomic 

decisions. Not for a moment 

would I deny their utility but, for 

me, a difference in the calls of 

two populations does not, 

without corroborating evidence 

from other features, provide 

sufficient justication for a 

Antarctica: Exploring a 

Fragile Eden By Jonathan 

and Angela Scott. Collins, 

London, UK. 2007. ISBN 10 

0-00-718345-3. 255 pages, 

many colour photographs, 

line drawings and maps. 

Hardback, £25. 

The Scotts are well known for 

their work in the Masai Mara and 

on the BBC with Big Cat Diary 

and other series. Their passion 

for Antarctica is much in 

evidence throughout this book. 

It is a very personal narrative, 

filled with observed cameos of 

their numerous visits to the 

region. Reading it I almost feel I 

am travelling on a cruise ship 

with them, leaning over the rails 

watching Cape Petrels Daption 

capense, fighting past fur seals 

on a beach, or sitting in the 

ship’s lecture theatre learning 

about penguin biology. All I am 

missing is the vast quantities of 

luxurious food! 

Like many cruises to the continent, 

we first visit the Falkland Islands 

and then South Georgia before 

arriving at the Antarctic Peninsula. 

Interspersed are chapters on the 

discovery of Antarctica, 

albatrosses, Shackleton, whales, 

the Pole, and pack ice. Despite the 

many clichés (albatrosses are 

‘ocean wanderers’, whales are 

taxonomic split, at any level. For 

example, the splitting of the 

Mediterranean Storm-petrel as a 

full species Hydrobates melitensis 

might be a step too far for some. 

I could cite other, similar 

examples which generate the 22 

species mentioned above. In 

their defence, the authors 

explain their reasoning 

thoroughly and do not propose 

any scientific names other than 

those already existing. 

This is a terrific book, a record of 

a love affair with the birds and 

‘great leviathans’), and some 

rather odd metaphors (the 

southern ocean is ‘a giant watery 

glove’), these chapters do impart a 

fair bit of knowledge and are very 

readable. The fur seals’ demise at 

man’s hand and their subsequent 

recovery, and the Falkland’s squid 

fishery and its impact on 

Rockhopper Penguin Eudyptes 

chrysocome and Black-browed 

Albatross Thalassarche 

melanophrys populations there, 

are just two of many stories well 

told. Their retelling of Shackleton’s 

epic journey, and other accounts of 

astonishing hardship and bravery 

by early polar explorers, greatly 

enhance this book. 

Inevitably a few mistakes have 

crept in, such as the authors 

getting confused by the identity 

of oystercatchers (describing a 

Blackish Haematopus ater yet 

calling it Magellanic H. 

leucopodus) and steamerducks 

(they call the Falkland 

Steamerduck Tachyeres 

brachypterus a Magellanic T. 

pteneres). Birders will be more 

concerned by their uncertain 

approach to albatross taxonomy. 

Despite stating that there are 24 

species, all forms are thereafter 

lumped in the conservative 

manner. A more consistent 

approach would have been 

better. These are, however, rather 

Reviews 

their islands. It is also a 

convincing riposte to anyone 

who thinks passion and scientific 

curiosity are uncomfortable 

bedfellows. Buy it. 

M. de L. Brooke 

SPECIAL READER’S OFFER 

Petrels – night & day. £30 per 

book (normally £34.95) 

To order, phone 01202 676622 or 

email info@soundapproach.co.uk, 

quoting ‘Seabird Group’. 

minor mishaps that do not 

detract from the Antarctic 

experiences, which the Scotts 

generally convey well. 

As one would expect from two 

award-winning photographers, 

the photos are all good and 

occasionally sensational. Some 

can also however seem 

somewhat clichéd, views through 

ice caves having been taken on 

another Scott’s journey a 

century ago. You also feel you 

have seen some of the wildlife 

photos before, but I suppose 

there are only so many angles 

from which to view a King 

Penguin Aptenodytes 

SEABIRD 21 (2008) 

109

110 

Reviews 

patagonicus colony, or an Adelie 

Penguin Pygoscelis adeliae on ice. 

Birders will be disappointed by 

the range of species depicted. Of 

the 71 photos where birds 

feature, 54 depict eight species 

of penguin, including 17 Adelie, 

14 Emperor A. forsteri and 10 

King. Nine albatross photos 

feature four each of Blackbrowed 

and Wandering 

Diomedea exulans, and one of 

Royal, while Procellariidae are 

The Birds of Alderney By 

Jeremy G. Sanders. The Press 

at St Anne, Alderney. 2007. 

ISBN 978-0-946760-61-9. 

320 pages, line drawings 

throughout, three maps. 

Hardback, £25, available (incl. 

p&p) from J. G. Sanders, PO 

Box 24, Alderney, GY9 3AP, 

Channel Islands, UK. 

Alderney is the most northerly 

of the British Channel Islands 

lying just 16 km due west of the 

Cotentin peninsula of France. It 

is tiny, at 6 km long at its 

greatest length and 2.4 km wide 

at its greatest width. Some will 

SEABIRD 21 (2008) 

covered by just two giant petrels 

and a Cape Petrel and there are 

none of Hydrobatidae. The text is 

similarly biased, leaving birders 

the impression that this book is 

for general tourists, not them. 

The Scotts would no doubt make 

very good and knowledgeable 

companions on a visit to 

Antarctica, and to be fair it was 

their intention to tell the 

continent’s story as if on a cruise 

argue that there is little scope 

for a volume about the birds of 

this small place compared to 

one that deals with birds of the 

Channel Islands as a whole. I 

have a different opinion. For 

starters, an assemblage of all 

the information about the birds 

of Alderney is a precursor to a 

volume with a wider scope. 

Furthermore, Alderney is unique 

and some of its uniqueness 

would be lost in a volume 

covering all Channel Islands. 

Thus, I welcome Sanders’ book. 

The book covers more than 130 

years of recording, drawing on 

published and unpublished 

sources, including assessment of 

thousands of records from over a 

hundred observers. However, the 

contents by Sanders’ admission 

offer a personal view built around 

his own observations over a 

period of 25 years. The text is 

lightly peppered with line 

drawings by Carmen Watson, 

some of which are quite pleasing. 

The book comprises three 

introductory chapters, the 

systematic list (the bulk of the 

book), and a somewhat 

redundant section on sketches of 

Alderney birds that repeats 

content from earlier in the book. 

The first introductory chapter 

provides a general description of 

Alderney. We are briefly 

introduced to the climate and to 

prehistoric times, followed by 

ship, which is how the vast 

majority of people lucky enough 

to visit Antarctica will experience 

it. Many points of interest are 

well enough covered to give a 

good insight into the whole 

Antarctic experience. It is 

certainly worth a read before a 

visit, and makes a good souvenir 

of an Antarctic cruise. 

Richard Schofield 

more extensive accounts of the 

various habitats including town, 

farmland, and coastal regions. The 

second provides an overview of 

the birds in categories including 

seabirds, coastal wading birds, 

inland breeding birds, migrants, 

and vagrants. The third 

introductory chapter reviews 

ornithology on Alderney. An 

important element of this for the 

systematic list is the reliability of 

the records. There will always be 

issues in this regard for a volume 

that covers over a century of 

records, especially with the 

earliest records. For Alderney, 

issues of validation of records 

persist in modern times since 

there is no single formal records 

panel. Indeed, unavoidably the 

author more-or-less is the sole 

adjudicator, but is transparent in 

this regard. 

The systematic list follows the 

species order in use by the British 

Ornithologists Union in 2006. 

Each account starts with the 

species’ English name and 

scientific name, briefly states the 

status on Alderney, and then 

delivers a succinct yet comprehensive 

account of its 

occurrence on Alderney from the 

earliest to most recent records. I 

would like to have seen a 

checklist of the birds of Alderney 

following the systematic list. 

Alderney is probably most 

famous for its important

Northern Gannet Morus 

bassanus colonies and for this 

reason Northern Gannet 

correctly receives extra 

attention in the systematic list, 

in a span of ten pages, while the 

demise of Atlantic Puffin 

Fratercula arctica from 100,000 

birds in the 1940s to the current 

250 pairs is afforded 17 pages. 

There are many species that 

have rarely been recorded on 

Alderney, including a few 

surprises like Gadwall Anas 

strepera on only two occasions, 

Wildlife of Seychelles By 

John Bowler. WILDGuides, 

Old Basing. 2006. ISBN 1- 

903657-06-7. 192 pages, 51 

photographic plates. 

Hardback, £17.95 (but see 

reader offer). 

The opening of Seychelles 

International Airport in July 1971 

led to a surge in tourism and also 

in biodiversity research. Apart 

from the Smithsonian Preliminary 

field guide to birds of the Indian 

Ocean, published in 1963, not 

widely available, and including 

only rough monochrome 

sketches, the early air travellers 

had little help in identifying what 

they saw. Since that time there 

has, of course, been a huge proliferation 

of field guides worldwide, 

including two on Seychelles’ birds. 

John Bowler’s Wildlife of 

Seychelles aims to facilitate 

identification of most of the 

more common animals. A brief 

introductory section describes 

the islands’ geography and 

climate, habitats, and conservation. 

In this section the scope 

of the book is defined as the 

granitic islands, along with the 

two nearby coralline islands, Bird 

and Denis. This is on the basis 

that the more remote southern 

island groups are ‘relatively 

depauperate in terms of the 

terrestrial habitats they support 

putting it on par with Greenish 

Warbler Phylloscopus 

trochiloides! Such seemingly 

odd statistics are so typical of 

small island avifauna. Inclusion 

of Levantine (Yelkouan) 

Shearwater Puffinus yelkouan 

might raise a few eyebrows 

given issues surrounding safe atsea 

identification. Sacred Ibis 

Threskiornis aethiopicus is 

included as a bona fide visitor. 

Alderney boasts a fine list of 

continental vagrants. The only 

Nearctic vagrants are Blue- 

and the number of terrestrial 

species present’ and are also ‘…. 

very difficult or expensive for the 

average person to visit’. While 

the latter may be true, it is a pity 

that the outer islands are 

dismissed in this way without 

further mention, since some 

harbour massive and internationally 

important breeding 

colonies of seabirds, along with 

the wintering of most of the 

world population of the Indian 

Ocean endemic Crab Plover 

Dromas ardeola; Aldabra and 

Cosmoledo also host indigenous 

and endemic land birds! 

The bulk of the book comprises 

brief species accounts, 

accompanied by photographs, of 

many of the islands’ commoner 

animals (birds, mammals, reptiles, 

amphibians, insects, arachnids, 

crustaceans, millipedes and 

centipedes, and molluscs), a 

significant proportion of which 

are endemic and many have not, 

to my knowledge, been illustrated 

before. Most photographs are 

excellent, while the dragonfly 

section includes a key 

accompanied by a helpful set of 

drawings. The texts for the birds 

include identification, voice, 

behaviour and breeding, whereas 

other taxa have a variety of 

headings, most including both 

identification and description; the 

contents of these two sections 

Reviews 

winged Teal Anas discors, 

Killdeer Charadrius vociferus, 

and Wilson’s Phalarope 

Phalaropus tricolor. Overall, the 

systematic list makes for 

interesting reading. 

This book is informative and 

must be read if visiting Alderney 

or studying the avifauna of the 

Channel Islands. It is well worth a 

read for general interest on a wet 

day or winter’s evening. 

Robert L. Flood 

are often repetitive. Checklists at 

the back of the book cover 

regularly occurring animals 

(including birds) and a more 

comprehensive one of birds 

recorded up to December 2005; 

in these, information on birds is 

duplicated, but using different 

symbols to designate status. 

All of the breeding seabirds are 

adequately described and 

illustrated, as are some of the 

regular migrants. However, 

describing the Black-naped Tern 

Sterna sumatrana as ‘possibly an 

annual vagrant’ ignores its 

breeding in the neighbouring 

Amirantes group, where 

important but unmentioned 

colonies of Roseate Terns S. 

dougallii arideensis also breed. 

Gygis alba is allocated the 

English name ‘Fairy Tern’, for 

SEABIRD 21 (2008) 

111

112 

Reviews 

which my preference is the more 

widely accepted ‘White Tern’. 

Notwithstanding these shortfalls, 

this book is undoubtedly a 

valuable addition to Seychelles 

guides, allowing visitors to gain a 

greater appreciation of the 

uniqueness of Seychelles’ fauna 

but, more importantly, 

Birds and Mammals of the 

Falkland Islands By Robin W. 

& Anne Woods. WILDGuides, 

Old Basing. 2006. ISBN 1- 

903657-10-5. 144 pages, 

over 100 colour photographs, 

43 plates. Hardback, £17.95 

(but see reader offer). 

As soon as you pick up this book, 

you get the feeling it is well 

bound and going to last the 

course. The front cover is 

appealing and just inside is a map 

of the islands, while opening the 

back cover reveals the grid codes 

used in the Breeding Birds Survey. 

The Introduction takes us 

through the islands’ topography, 

climate, vegetation, human 

presence and influence on the 

environment, predation on birds, 

and changes in the natural 

vegetation, before a succinct 

chapter on ‘Conservation: 

positive action for wildlife’. Here 

we learn among other things 

about the local WATCH group, 

the only overseas branch of this 

junior section of the UK Wildlife 

Trusts, so important for the 

future wildlife of the islands. 

Then come the birds – a brief 

introduction to them followed by 

a checklist; after which help in 

knowing ‘How to use this book’. 

Next we come to a brief 

summary of each bird family, with 

associated tables. Between this 

and the species accounts is an 

illustration of bird topography. If 

the book is to be used to its full 

potential, I would have found it 

SEABIRD 21 (2008) 

encouraging the Seychellois to 

learn about and value their 

heritage. For me, its greatest 

utility lies in its coverage of taxa 

not treated in other guides; keen 

birders will need more comprehensive 

field guides. 

(Another photographic guide 

with the same title, by Mike Hill 

easier to have the section on bird 

families after the bird topography 

section, integrated in some way 

with the species accounts – 

otherwise you are constantly 

looking back and forward to 

marry the two together. 

Each species’ text is succinct, 

containing a lot of information in 

a small space, including a box 

giving status details. The species 

are all illustrated with 

photographs opposite respective 

texts, over a hundred of which 

are by Alan Henry. Most are of 

high quality, although one or two, 

for example, Magellanic Diving 

Petrel Pelecanoides magellani 

and Grey-backed Storm-Petrel 

Garrodia nereis are not quite so 

sharp. Some, such as the Dolphin 

Gull Leucophaeus scoresbi and 

particularly the Variable Hawk 

Buteo polyosoma, show the range 

of different plumages that can be 

a minefield for the unwary. 

However, I found some of the 

annotations, explained in the bird 

plates’ introduction, a little 

difficult to read at times.The only 

drawback to my mind with 

photographs is that what you see 

is what you get: illustrations not 

necessarily comparable in 

plumage, size or jizz for similar 

and difficult species. 

But, as the title tells us, birds 

aren’t the only creatures covered 

here. So, following them, there is a 

good section on marine 

mammals and non-domesticated 

terrestrial mammals. Of the 

latter, the only native terrestrial 

and David Currie, was published 

by Collins in 2007) 

Chris Feare 

WILDGuides will be supporting the 

work of The Island Conservation 

Society and The Nature Protection 

Trust of Seychelles through 

proceeds of this book. 

mammal, the Falkland Islands Fox 

Dusicyon antarcticus, known by 

early settlers as the Warrah, was 

exterminated by 1876. All other 

species were introduced. Again, 

good photographs lie opposite 

respective texts. 

I’m rather a pedant regarding 

typographical errors and I’m 

pleased to say could find none. 

Altogether I liked this 

approachable book a lot, which is 

nothing less than expected with 

authors so synonymous with the 

birds and other wildlife of the 

Falkland Islands. I feel it will have 

a much wider audience than its 

stated aim of appealing to 

passengers on cruise ships and 

land-based visitors. 

Wendy Dickson 

A contribution will be made by 

WILDGuides Ltd. to the work of 

Falklands Conservation for every 

book sold. 

SPECIAL READER’S OFFER 

Wildlife of Seychelles and Birds 

and Mammals of the Falkland 

Islands. £15 per book (normally 

£17.95) or £27.50 if both ordered 

at the same time (p&p free). 

Orders, mentioning Seabird to be 

sent to WILDGuides, PO Box 680, 

Maidenhead, Berks SL6 9ST, UK 

with cheques made out to 

‘Wildguides’, or by phone for 

credit card orders 01628-529297

The Seabird Group 

The Seabird Group is a registered charity, founded in 

1966 to promote and help co-ordinate the study and 

conservation of seabirds. The Group organises regular 

international conferences. Members receive, and can 

contribute to, regular newsletters, and the journal 

Seabird, published annually. Members are encouraged 

to participate in surveys of seabirds, monitoring 

programmes and other seabird research work. As part 

of this, members can apply for small grants towards 

the cost of research or survey projects. 

Membership 

Membership of the Seabird Group is open to all with an 

interest in seabirds, whether amateur enthusiasts, students or 

professionals. 

Membership Secretary: David Sowter, The Seabird Group, 

5 The Grove, Penwortham, Preston, PR1 0UU, UK. 

Email: davidsowter@freenet.co.uk 

Annual Membership Rates: 

Standard £ 10.00 

Standing Order £ 9.00 

Students £ 5.00 

the 

Seabird 

Group 

Current Executive Committee 

Norman Ratcliffe (Chairman), Alan Leitch (Secretary), Andrew 

Ramsay (Treasurer), David Sowter (Membership Secretary), 

Martin Heubeck (Seabird Editor), Linda Wilson (Seabird 

Publishing Editor), Liz Humphreys (Newsletter Editor), Claire 

Smith (Assistant Newsletter Editor), Juan Brown, Simon Foster. 

General correspondence 

The Seabird Group, c/o BTO, The Nunnery, Thetford, Norfolk 

IP24 2PU, UK. Email: seabird@bto.org 

Further information about the group, including grant 

guidelines and membership application can be found at: 

www.seabirdgroup.org.uk

Leach’s Storm-Petrel Oceanodroma leucorhoa, Merseyside, UK, 20th September 2004 © S. Young. 

ISSN 1757-5842

COMPLETE.pdf - The Seabird Group

Create successful ePaper yourself

Delete template?

Save as template?