To this point, my account of the balanced growth equation has centered on energy flow, but ecological fluxes involve more than the transfer of energy. When an animal eats or breathes, defecates or grows, a quantity of material with a definite chemical composition accompanies each energy flux. One could produce budgets for each compound and element in animal tissues and so repeat the foregoing chapters considering balanced growth equations for phosphorus (Peters and Rigler 1973), nitrogen (Nagy & Shoemaker 1975), nickel (Hall 1978), and so forth. These flows are the autecological expression of nutrient turnover, mineral utilization, and toxicant accumulation, areas of prime environmental concern that are not addressed by ecological energetics.

Whatever advantages such material budgets might offer, they have been far less extensively analyzed than energy budgets, and the treatment here will be correspondingly brief. The prime reason for this asymmetrical development is theoretical. Beginning with Lindeman's (1942) seminal paper and culminating in the International Biological Programme (1964–1974), ecologists hoped that patterns of energy flow would provide the basis for ecological generality. This theoretical attraction gave impetus to the development of techniques and traditions that emphasize energy transfers. Finally, the number of possible compounds that could be studied ecologically is so vast that biologists must hope that the basic proportionalities of energy flow will also apply to mass fluxes.

Locomotion is the most obvious and most characteristic of animal activities. Like any activity, movement requires energy and, therefore, increases an animal's metabolic rate. From our own experience with walking, running, or swimming, we know that the power demands for movement can be very large. The demands of locomotion may, therefore, be a large component of respiration in the balanced energy equation and a significant energetic cost for the moving animal. This raises a quantitative question: How important are locomotive costs?

This chapter examines the interrelations between the metabolic costs of movement, body size, velocity, distance traveled, time spent traveling, and mode of locomotion – flying, swimming, or running. It provides the basic information required to estimate the metabolic rate of a moving animal. This is achieved by first considering the results of empirical studies relating metabolism to body mass and velocity for terrestrial, aquatic, and aerial locomotion. Because such data are largely derived in the laboratory, equations describing these data have limited ecological relevance when considered in isolation. To use them, we also require some estimate of an organism's speed in nature. Available allometric descriptions of average velocity are, therefore, introduced and then used to calculate the metabolic rates of moving animals. Whenever possible, the ecological importance of these relations is considered by comparing speeds and costs with the demands of existence.

Almost all relations in this book are derived from and applied to individual animals. The ecological implications of biological scaling are, therefore, primarily autecological. Such information is valuable and important, but many processes depend on other members of the population or the community. For example, excretion, ingestion, growth, and reproduction are each physiological processes that scale to body size. These autecological rates are given new significance when multiplied by the number of individuals in the population or community. Excretion becomes nutrient regeneration, ingestion might represent prey mortality, and growth and reproduction become production. Each of these population and community processes, and many others, can be calculated from allometric equations if we know population size and community size structure (i.e., the number of organisms of each size, regardless of species, in the community). Such composite values may even be better than individual predictions, because they average over many different members of the population or over many different species and individuals.

Ideally, figures for abundance would be measured accurately and precisely, but such estimates are often unavailable for financial, practical, and biological reasons. One then searches the literature for approximations, values from similar sites and similar organisms. This book is dedicated to the premise that size is a major criterion of similarity, consequently this chapter summarizes the literature relating animal abundance to individual size as a necessary step toward the application of allometry to higher ecological levels.

Models of preferential mating

In this chapter, I bring together and analyse the different lines of evidence on the mechanism of sexual selection in the Arctic Skua. The models to be fitted to the data have already been described in sections 9.2 and 9.4. They are all models of preferential mating, not male competition. As we proved in section 9.3, male competition cannot maintain a polymorphism without heterozygous advantage. Heterozygotes suffer a slight disadvantage compared to dark homozygotes (section 8.5). Since the polymorphism of the Arctic Skua is stable, models of male competition cannot plausibly be fitted to the data.

Males who are the objects of female preference will tend to be chosen as mates earlier in the breeding season. We may postulate that a certain proportion of females will prefer to mate with particular male phenotypes. If the males have already arrived on their territories or joined a ‘club’ of unmated males competing for the females' attention, then as groups of females come into breeding condition they will be able to choose males from among those who are still unmated. If we are to simulate this process of mate selection as it goes on through the breeding season, we must allow for the intervals of time which are required for courtship and mate selection. On average, as we saw in section 8.4.2, new pairs take an extra 7–8 days to breed compared to old, established pairs.

In this book I have described the results of my study of the population ecology and genetics of the Arctic Skuas on Fair Isle. When a population ecologist studies an organism, he asks questions, such as the following, about its population size and numbers:

1. (i) Are its populations stable, increasing, or decreasing?

2. (ii) Can future population changes be predicted?

3. (iii) How are population numbers regulated – for example, is regulation density-dependent?

Seeing a polymorphism, a population geneticist asks such questions as:

1. (i) What are the genetics of the polymorphism?

2. (ii) Is the polymorphism stable, or is it still evolving?

3. (iii) How are the gene frequencies spatially distributed – is their distribution uniform, or are there clines from one area to another?

4. (iv) What selective forces are acting on the phenotypes – natural selection, sexual selection, or both – and are they sufficient to ‘protect’ the polymorphism against extinction of alleles?

I hoped I might answer such questions when I set out to study the Arctic Skua and its striking polymorphism in plumage. In previous chapters – chapter 2 on population ecology, chapter 5 on genetics and chapters 6, 8 and 10 on natural and sexual selection, I have given some answers. Some answers are tentative: some – on the population changes and the natural and sexual selection of the population – are bold and decisive. Now, in this last chapter, I can omit the details of analysis, summarize the results, and draw general conclusions about the evolutionary forces that maintain the polymorphism.

This book describes the results of research spanning a period of 30 years. The late Kenneth Williamson began studying the Arctic Skuas of Fair Isle in 1948. Only 15 pairs were then nesting on the island. In 1957, Peter Davis took up the study, having succeeded Williamson as Warden of the Fair Isle Bird Observatory. As part of research for the Ph.D. degree of the University of Cambridge, I spent three years from 1958 to 1961 studying the genetics of the Arctic Skua. I was supported by a Nature Conservancy Research Studentship. This initial period of research on the Arctic Skuas ended in 1962 when Peter Davis left Fair Isle. By this time, 71 pairs were nesting on the island.

In 1970, R. J. Berry and Peter Davis published a paper analysing the breeding dates of the different phenotypes of the Arctic Skuas (Berry & Davis, 1970). They found that pale males, breeding for the first time, bred several days later on average than first-time, intermediate and dark males. They interpreted this as a behavioural adaptation of pale birds to their more northerly distribution where later breeding might be advantageous. At that time, I was working on models of Darwin's theory of sexual selection. The later breeding of certain male phenotypes in new pairs is exactly what the models predict, whereas the adaptation postulated by Berry and Davis should have been shown by all birds.

Female preferences and male competition in monogamous birds

Let us suppose that some males possess a genetic characteristic that can increase female arousal. Some females, responding more rapidly to these males, thus mate with them preferentially. Other females, however, show no preferences: they just choose any male at random. To construct a theoretical model of the sequence of matings, I assume that groups of females come into breeding condition during successive intervals of the breeding season. They choose their mates from the available pool of unmated males. This pool decreases, of course, as each successive group of females choose their mates. Later females may not be able to mate with the males they prefer because none may be left in the pool.

In monogamous species, the sequence of matings is clearly important in determining what choices can be made. This applies to the matings taking place within a particular group of individuals. Females with particular preferences might choose their mates before, at the same time as, or after the females who mate at random. If preferential matings come first, more females will be able to mate with the males they prefer. But if the random matings come first, some preferred males will already have been chosen when the preferential matings take place: females will then have less opportunity to express their preferences.

The behavioural mechanism of mate selection will presumably determine the sequence of preferential and random matings.

Darwin's theory of sexual selection

Sexual selection occurs when some individuals are more successful in finding mates than others. Like other characteristics of animals, sexual behaviour varies. Some males will defend their territories more fiercely than others, or court the females more vigorously, or display more attractive plumage. These characteristics will be selected if they increase the chances of mating. This is sexual selection as Darwin defined it in The Origin of Species:

And Darwin thought that sexual selection would be the cause of the evolution of many of the differences in structure and coloration of the sexes:

Genetic analysis of matings between phenotypes

The first step in studying any polymorphism is to determine the genetics of the different phenotypes. At first sight, the polymorphism of the Arctic Skua seems to depend on a simple genetic mechanism. As adults, the pale, non-melanic birds are completely distinct from the dark and intermediate melanics. Dark and intermediate birds show a continuous range of phenotypic expression with no clearly definable difference separating darks from intermediates. Matings of pale × pale birds have only produced pale offspring. We may thus infer: (i) that pale birds are homozygous; and (ii) that melanism is semi-dominant to pale. If only two alleles determine the phenotypes, pales must be homozygous for the pale allele, while darks and intermediates are either homozygous or heterozygous for the alternative, melanic allele. Presumably the melanic homozygotes would tend to be darker than the heterozygotes.

Chicks produced by matings

A difficulty is encountered when classifying the chicks produced by matings of different phenotypes. The phenotypes of the chicks show no absolutely clear differences. All gradations can be observed, from pale chicks with a large patch of white feathers on their belly, to those with just a few white feathers, then to those with some darkening at the tips of the mainly white feathers, through intermediates with bands of dark pigment, and so to darks with completely dark belly feathers. Although there is this continuous gradation from pale to dark, the great majority of birds are easily classified by phenotype.

The breeding season

The Arctic Skuas start returning to their breeding grounds on Fair Isle in the third week of April. In the years 1973–75, the first Arctic Skua was seen on the island on 16 April. In 1976 the first bird was seen a day later. Three or four birds have usually arrived by 18 April. Then numbers increase rapidly:

1. (i) in 1973, 12 birds were seen on 20 April, 80+ on 8 May;

2. (ii) in 1974, 59 birds were seen on 20 April, 120 on 8 May;

3. (iii) in 1975, 63 birds were seen on 2 May, 200+ on 15 May.

In 1974, when the most careful and detailed counts of arriving Arctic Skuas were made, a total of 116 pairs eventually bred on Fair Isle. About 25 per cent of the breeding birds had arrived by 30 April, 50 per cent by 8 May. The first arrivals are breeding birds that were colour-ringed in previous seasons. Non-breeding birds arrive much later, forming ‘clubs’ in late June. As we have already seen (table 2.5, section 2.3, chapter 2), the Bonxies arrive about two or three weeks before the Arctic Skuas. The Bonxies thus gain a great advantage in the competition for territories: a Bonxie may already have established himself on an Arctic Skua's territory before the Arctic Skua returns. The Arctic Skua is forced to seek an unoccupied territory elsewhere.

Territoriality, pair formation and mating behaviour

Observation of a few pairs of Arctic Skuas may reveal only very little of their courtship and mating behaviour. Pairs often copulate with no obvious preliminary courtship; they may have bred together for many years. Birds that know each other well enough perhaps have no need for elaborate courtship.

Perdeck (1963) observed the pairing of a few single birds and the mating behaviour of some established pairs in a colony of Arctic Skuas on the Faeroe island of Mykines. Since the birds had not been ringed, pairs from previous years could not have been distinguished from newly formed pairs. Single birds in the act of pairing may simply have been rejoining their former mates. If they had been seeking new mates, their behaviour might have been different. A detailed analysis of pairing and mating behaviour would have been possible on Fair Isle where all birds from previous years are known by their colour rings. But my own study of the population biology of the Fair Isle Arctic Skuas never allowed time for systematic observations of behaviour. Of course, I observed many aspects of their agonistic, pairing and mating behaviour. My own observations of their calls and postures agree closely with those of Perdeck, whose terminology I have used in the following account.

Agonistic behaviour in defence of territory

Perdeck (1963) observed the agonistic behaviour of Arctic Skuas defending their territories.