Atlantic Yellow-legged Gull (atlantis)

(last update: October 12, 2011)

Rui Caratão (Portugal)
Mars Muusse (Netherlands)

Atlantic YLG 4cy August

Mesopelagic fish eaten by Yellow-legged Herring Gulls
Larus argentatus atlantis in the Azores.
Published in: SEABIRD 16: 30-33, 1994.

K. C. Hamer, D.R. Thompson, A.J. Rundle, S.A. Lewis and F.M. Stewart

Herring Gulls Larus argentatus are highly flexible and opportunistic feeders; their diets have been recorded to include, among other items, marine fish (including discards from fishing boats), marine invertebrates, seabird eggs and chicks, earthworms and domestic refuse (Furness & Monaghan 1987). Yellow-legged Herring Gulls L argentatus atIantis breed throughout the Mediterranean and the south-east region of the North Atlantic Ocean, and have been increasing in number for several decades, at least in the Azores archipelago (Le Grande et al. 1984). Despite this, their diets are poorly known, although Mediterranean populations have been recorded feeding upon marine fish (Wilt 1974, Borg & Zamrnit 1987) and those in the Selvagens Islands on seabird eggs (Zino et al. 1987) and terrestrial snails (K.C. Hamer unpublished data). This paper provides information concerning the diets of Yellow-legged Herring Gulls in the Azores.

Field-work was conducted at Ilheu Topo, an island of about 20ha, 0.5km from the eastern tip of Sao Jorge in the centre of the archipelago, where less than 100 pairs of Yellow-legged Herring Gulls breed each year. Regurgitated pellets containing the undigested elements of the gulls' food were collected from the colony between 28 and 30 August 1989, shortly after the end of that year's breeding season. The contents of all pellets were identified in the field, and any sagittal otoliths present were retained and later identified by A.J.R., using keys in Nolf (1985) augmented by otoliths dissected from material at the British Museum of Natural History.

A total of 510 pellets were collected, of which 89% were fish and 8% were goose-barnacle Lepas sp. (Table I). All other food items were present in less than 1% of pellets, and all pellets contained remains of a single food-type. Of 280 complete otoliths, representing 21 species of fish (Table II), 44% were identified to species with certainty. A further 48% were identified to genus, but a lack of described otoliths from Azorean fish prevented further identification. The remaining 8% of otoliths were not identifiable to genus, because of erosion by stomach acid before regurgitation. However, they were separable into ten species (Table II).
Whole otoliths were obtained from 36% of fish pellets. A further 32% (omitted from Table II)
contained fragmented otoliths, but apparently included no new species. The remainder contained no otoliths, but mostly had a similar appearance to pellets that contained otoliths, and probably represented the same species. However, 31 pellets (6% of the total), all lacking otoliths, consisted of very fine bones, probably of horse mackerel Trachurus picturatus, a species which has very small otoliths that are likely to be dissolved by stomach acid before regurgitation (Jobling & Breiby 1986).
Three species contributed 83% of otoliths; 42% were Diretmus argenteus, 33% were a single
species of Symbolophorus and 8% a single species of Myctophum. These are all mesopelagic fish occurring at depths below 100m (Whitehead et al. 1984; Table II). All other fish identified to genus also occur at depths below 100m, and three (Coelorhynchus, Trachichthodes and Beryx) are benthopelagic genera, found mainly on the sea-bed away from the coast (Table II), while the Batrachoididae (toadfish) are characterized by a coastal benthic habitat (Nelson 1984).

Analysis of regurgitated pellets generally underestimates the importance of items that produce little indigestible material (Johnstone et al. 1990), while infrequent collection underestimates the importance of foods that result in unstable pellets, since these tend to be broken up and dispersed during periods of bad weather (Furness & Hislop 1981). Although the very low summer rainfall on Ilheu Topo probably reduced the latter source of error, we do not claim that the data in Tables I and II allow a quantitative assessment of the diet of Yellow-legged Herring Gulls at this colony in 1989. However, they do allow an assessment of whether different items were important dietary components.
Almost 90% of pellets were composed of fish (Table I), indicating that they were an important resource for adults at this colony. The true proportion of fish in the gulls' diet may in fact have exceeded this figure, since pellets containing crustacean and molluscan remains were comparatively large and easily visible, while those containing black rat were very durable, and so may have been over-represented.
Horse mackerel, or other fine-boned species, may have been under-represented if their pellets were less stable than those produced from other species, although this did not appear to be the case in the pellets that we obtained. Even allowing for the uncertainties inherent in analysis of pellet samples, it was obvious that mesopelagic fish made a significant contribution to the diets of Yellow-legged Herring Gulls at this colony.
Meso- and benthopelagic fish species may be made available to surface predators as discards
from fishing vessels (Hudson & Fumess 1988). However, the fishery in the central Azores is very small. Fish could have been eaten by vertically-migrating piscine predators, and entered surface-feeding seabirds when the predatory fish were themselves eaten. However, this should have resulted in pellets containing otoliths of more than one species, which was not the case in this study. Fish in the genera Symbolophorus and Myctophum - from which 41% of otoliths were derived, are nyctoepipelagic (found in surface waters at night) and so could have been caught by Yellow-legged Herring Gulls feeding nocturnally, as they do in the Mediterranean (Borg & Zamrnit 1987). However, the other species are not nyctoepipelagic. Otoliths of D. argenteus represented 42% of the total; this species is mesopelagic when adult and, although juveniles are found near the surface by day (Whitehead et al. 1984), these are less than 1cm long and therefore too small to have produced the otoliths in this sample.
Meso- and benthopelagic species may be driven to the surface by marine predators, particularly tuna (Au & Pitman 1988). Periodic associations between surface-schooling tuna, cetaceans and various seabird species have long been recognized and exploited by fishermen (Evans 1982) but gulls are considered to feed largely independently of fish schools in pelagic waters and to rarely occur with either tuna or cetaceans (Au & Pitman 1988). However, Martin (1986) recorded Yellow-legged Herring Gulls in the Azores feeding in association with Atlantic spotted dolphin Stenella frontalis and Cory's Shearwater Calonectris diomedea, although on all occasions these three species were feeding upon horse mackerel rather than upon mesopelagic species. Cetaceans and seabirds were reported by fishermen in the Azores to reliably pinpoint feeding tuna, although these were not present in any of the small number of associations observed by Martin (1986). Hence, the high frequency of mesopelagic fish in the diets of Yellow-legged Herring Gulls at Ilheu Topo suggested that they regularly feed in association with surface-schooling tuna.

We thank the Camara Municipal de Calheta, Sao Jorge for logistical support and permission to work on llheu Topo, and Mike Imber for identifying squid beaks. This work formed part of an expedition funded by Glasgow University Exploration Society, the Linnean Society, the British Ornithologists' Union, the Biological Council & the British Ecological Society. We thank Mike Harris and Tony Martin for helpful comments on an earlier manuscript.

Over 70% of pellets regurgitated by Yellow-legged Herring Gulls in the Azores contained the remains of diurnally meso- or benthopelagic fish species. A third of pellets were of nyctoepipelagic species, which may have been caught at the surface at night. However most other pellets, including the most abundant type, were of species not normally found at depths less than 100m. These were probably made available to Herring Gulls by the activities of tuna, which drive mesopelagic fish to the surface.

See PDF:



Pellet type: Number of pellets: Percentage of pellets:
Fish 447 89.4
Goose-barnacle 41 8.2
Black rat Rattus rattus 4 0.8
Squid 4 0.8
Gastropod mollusc 3 0.6
Unidentified crustacea 1 0.2

Note: Squid were identified from beaks as one species of Chiroteuthis, one species of Taonius and two
individuals in the family Ommastrephidae. Black rats were identified from mandibles, molar teeth and long bones.



# of otoliths

% of otoliths Habitat and depth


- Opisthoproctus grimaldii



mesopelagic; 200-600m

- Myctophum sp. 21 7.5

mesopelagic; 100-1000m

- Notoscopelus sp. 1 0.4 mesopelagic; 100-1000m
- Protomyctophum sp. 3 1.1 mesopelagic; 100-1000m
- Symbolophorus sp. 93 33.2 mesopelagic; 100-1000m
- Undetermined 2 0.7  


- Coelorhynchus sp. 11 3.9 benthopelagic; 140-2000m
- Nezumia aequalis 2 0.7 benthopelagic; 140-2000m
- Undetermined 1 0.4 coastal, benthic
- Trachichthodes sp. 4 1.4 benthopelagic; deep water
- Diretmus argenteus 117 41.8 mesopelagic; 500-700m
- Beryx sp. 1 0.4 benthopelagic; 400-600m
- Caprosaper 1 0.4 mesopelagic; 100-600m
- Undetermined; 2 species 2 0.7  
Undetermined Family; 4 species 15 5.4  
Undetermined Order; 2 species 2 0.7  

Note: Data concerning habitats and depth ranges are from Whitehead et al. (1984) and Nelson (1984). Where fish have been identified to genus, the depth range given is the maximum covering all species recorded in the North-east Atlantic.