G. M. Dunnet’s research while affiliated with University of Aberdeen and other places

This short paper describes the background, initiation and development of what has turned out to be a long-term (41 years to date) study of Fulmars on Eynhallow in Orkney. Most of the scientific results have been published (see Appendix 1) so that only the major findings will be referred to here, to illustrate how the study has evolved as knowledge increased, new techniques became available and objectives changed.

.1. A study of individually marked Fulmars during three seasons has shown that both sexes can breed in successive years. The evidence for the theory of intermittent breeding is examined and definitions are given of the various categories of birds which compose the population during the breeding season. It is suggested that the supposed non-breeding adults which are seen, often in moult, in July are failed-breeders which have lost their egg or nestling. 2. No marked individual changed its nest-site between years, although the number and identity of sites occupied each July varied considerably. 3. No changes of mate occurred among the six marked pairs which nested in two consecutive years. 4. There is some evidence that both parents remain near the colony during the week before and after hatching. 5. Analysis of the 90 known sites of a colony subject to avian but not wild mammalian predation has revealed a wide range of habitat selection and a strong tendency toward close concentration of nests. 6. There is some evidence that the breeding season, as measured by date of hatching, varies from year to year.

. 1 The relation between a population of Starlings and its food supply during the breeding season was investigated at Craibstone, near Aberdeen, during 1950, 1951 and 1952.2 In a square mile, including Craibstone, the number of breeding females increased from 43 in 1950 to 52–54 in 1951, and to 78 in 1952.3 During the breeding season the Starlings obtained most of their food in the upper layer of the soil in pasture fields. Leatherjackets (Tipula sp. larvae) were the most important, and the seasonal variation in their biomass was measured. The period of availability is long, but that of great biomass is short. Earthworms were abundant in the soil, but formed a small proportion of the food taken.4 Synchrony of laying the first clutches is marked; in second clutches (common at Craibstone in each year) it is less pronounced and is derived from that of the earlier layin.5 Starlings fed from a limited number of feeding areas.6 The food brought to the nestlings was studied in part by high-speed Hash photography.(a) On the average leatherjackets formed over 80% of the food brought to the nestlings and earthworms 16%.(b) The weight of, and variation in, individual feeds were calculated.(c) There is a marked peak of feeding activity in the early morning, often with a lesser peak in the afternoon. Earthworms tend to be brought early in the morning.(d) The proportions of different items do not vary seasonally.(e) The relation between feeding rate and the amount of food brought is not constant.(f) Owing to individual differences in the feeding activity of parents and the response of the young, the amount of food required to rear a brood cannot be estimated.(g) The proportion of available leatherjackets consumed by nestlings was about 1·9% in 1951 and 7% in 1952.7 The mean clutch-size was similar in each year, and second clutches were smaller than first.8 Nestling weights did not differ significantly between broods of different sizes or in relation to the amount of food available.9 It was concluded that:(a) The synchrony of first clutches may be due to the advantage of having successful second broods, and the evidence suggests that these are ultimately controlled by the date the leatherjackets emerge. The synchrony may also have survival value in providing the opportunity for gregarious feeding by the adults and, later, the juveniles.(b) The occurrence of second broods appears to be determined by how early breeding starts and when the leather-jackets disappear from the soil.(c) Breeding-rate and success were not closely related to the available food supply.(d) There is much evidence to suggest that the available food-supply is the ultimate factor timing the end of the breeding season, but it probably does not determine the onset of breeding.(e) The density of the breeding population is not controlled by the food available during the breeding season.

Colour-ringed breeding Fulmars were studied on the island of Eynhallow, Orkney, from 1950 onwards. During the 28 years, there was an average increase in population size of about 4% (9 birds) per annum, standard deviation 23%, The mean population size was 161 birds, with about 22 new breeders and 26 new nest sites per annum. There were marked fluctuations in numbers of birds and occupied sites per annum, causing deviations from the underlying trend. Most nest sites were on the coast. The distribution of occupied nests was variable, the most favoured areas changing over the years. A mean of 46 chicks fledged per annum. The proportion of breeding pairs which fledged a chick varied from 16 to 52% (even though there was no known change in conditions). Success in small areas sometimes differed considerably from that of the island as a whole. Although most fledglings were ringed, only 3% have returned and bred on Eynhallow. Between 89 and 94% of those which survived to breeding age must have bred elsewhere. It was concluded that the population size and breeding success of small areas could differ quite markedly from that on a wider scale and could vary quite markedly over the years, facts that have important implications for monitoring of seabird colonies.

The food taken by Aberdeenshire Starlings during autumn and winter was investigated from stomach contents of samples shot regularly at roosts. The food available was known approximately. During severe weather feeding habits of the Starlings change so that they become almost entirely dependent on man. The gross weight and the fat‐deposits of the birds increased from autumn to mid‐winter and were not greatly affected by two snowy periods. The period of most acute food‐shortage for the Starling may be the early autumn, after the leatherjackets have disappeared and before oats become available. At this time the total weight and the fat‐weight of adults and juveniles are low and the birds are moulting. Density‐governed mortality, which contributes importantly to stability of the breeding‐population density, may occur at this time.

Five cliff-nesting species were identified as being particularly vulnerable, namely fulmar (Fulmarus glacialis), shag (Phalacrocorax aristotelis), kittiwake (Rissa tridactyla), guillemot (Uria aalge) and razorbill (Alca torda). Eider (Somateria mollissima) was also surveyed. On many cliffs all five of these species nest together and the choice of colonies took this into account for logistic reasons. The data are presented in several ways: numbers, actual numbers of items counted; annual means, whereby at any colony a count will be made in each sample plot on a number of different days, usually during June of each year, for each data the plot counts being summed to provide a "colony count'; and an index in which in order to consider trends of population change over Shetland as a whole the data for each colony have been converted to an "Index'. -after Authors