Skip to main content Accessibility help
×
Home
Hostname: page-component-99c86f546-z5d2w Total loading time: 0.754 Render date: 2021-12-06T06:57:53.023Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "metricsAbstractViews": false, "figures": true, "newCiteModal": false, "newCitedByModal": true, "newEcommerce": true, "newUsageEvents": true }

1 - Hunting Strategies of Predators as Revealed in Field Studies of Great Tits

Published online by Cambridge University Press:  01 April 2021

T. Royama
Affiliation:
Canadian Forest Service

Summary

This chapter develops the concept of ‘hunting by profitability’ as a basic principle underlying the hunting strategy of predators. When the birds look for prey (insects) at various hunting sites (niches) in the habitat, they allocate their time among the potential sites according to the profit they can gain at each site (the biomass of prey they can collect per unit effort of hunting). The main objective of this chapter is not to describe the behavioural ecology of the tits as observed, but to show how I made observations, interpreted the results, and inferred the concept of profitability as the basis for the strategy of predators hunting their prey. The concept plays an important role in the analyses of the multi-species interaction processes discussed in Chapters 8 and 9.

Type
Chapter
Information
Animal Population Ecology
An Analytical Approach
, pp. 7 - 25
Publisher: Cambridge University Press
Print publication year: 2021

1.1 Preamble

Towards the end of my undergraduate years at the University of Tokyo, I decided to go to the graduate school in forest zoology where I could study birds in the field. Although called ‘forest zoology’, in practice it was an entomology laboratory. However, it was the only place where I could possibly study birds in a natural environment. At that time in Japan, field ornithology was not an academic subject in any university.

I proposed to study the breeding behaviour and ecology of the great tit, Parus major minor (now classified as the independent species P. minor). At the admission interview, one professor (then the head of the University Research Forests) became interested in my proposal and suggested that I use the Fuji Research Forest by the scenic Lake Yamanaka at the foot of Mount Fuji, about 80 km south-west of Tokyo. He also provided me with financial support and with all sorts of facilities at the research forest during my three field seasons from 1958 to 1960.

However, I could not find anybody to guide my study, as there were only a few academic ornithologists in Japan then, and they were mostly taxonomists studying dead specimens. In contrast, there were quite a few works on the tits in Europe, especially in the Netherlands and Britain, which I used as trailblazers. As my study progressed, I recognized certain aspects that had not been well understood even in the European studies. The topic I talk about here is one of the few aspects that I think I understood in depth: the strategy of the great tits hunting their prey (insects). Here I describe how I made observations, interpreted the results, and conceived a general principle underlying the strategy with which predators hunt their prey. But, before getting into the main theme, I review a major theory from the European studies.

1.2 Search Image Theory by Lukas Tinbergen

In the final year of my graduate study in Tokyo, I happened to see an article by Lukas (Luuk) Tinbergen (Reference Tinbergen1960) of the University of Groningen (henceforth just Luuk). In this article, Luuk proposed a theory based on his intensive study of the great tit (P. m. major, the subspecies of Parus major on continental Europe) feeding their nestlings with various insects in the Dutch pinewoods. In particular, he studied the frequency of a given insect brought to the nest to feed the young in relation to its abundance in the woods.

As a typical example, he selected the case for pine beauty (Panolis flammea, Lepidoptera: Noctuidae) caterpillars. What he found was summarized in his figure 14, which contained two curves. One was the frequency (%) occurrence of the Panolis caterpillars in the tits’ diet and the other the relative (%) abundance of the caterpillars in the woods, both being plotted against their density in the woods. He postulated the following: (i) if a parent tit searched the woods at random, it should have encountered the caterpillars in proportion to their relative abundance in the woods; (ii) then, the % frequency of the caterpillars in the diet should be expected to be proportional to their (%) abundance in the woods. In other words, the two curves in the graph should have matched in shape. Luuk found no match but the following tendency: when the (%) abundance of the insect in the woods was low, its frequency in the nestling diet was below that expected; when the abundance was at a moderately high level, the frequency was substantially above that expected; but at a much higher level in abundance, the frequency did not increase as expected but levelled off below that expected. Luuk interpreted this tendency in the following way.

When a given insect species was sufficiently abundant, the tits encountered it so frequently that they would have been conditioned to its image when searching for prey, i.e. the formation of a search image. This enabled the birds to see the insect more readily among the leaves and twigs that would have otherwise obscured the presence of the well-camouflaged insect. Conversely, when the insect was scarce, the birds did not encounter it frequently enough to form a search image and did not recognize the insect as readily. But, then, Luuk had a dilemma: the search-image hypothesis would not explain why the caterpillars were not brought to the nest proportionately more frequently when they became very abundant. He got around the dilemma by postulating that the birds must have avoided the diet becoming too monotonous with a very abundant prey species. I was not comfortable with Luuk’s inferences.

First of all, I was not happy with his postulation of ‘random search’ as a criterion with which he compared the observed frequency of the prey species in the tits’ diet. As is well known, forest insects occur on their own distinct places or niches, such as trees, shrubs, or ground vegetation. Even within each type of place, they tend to occur on different host plants, or even in different parts of the same plant (micro niches). Second, I knew through my own observation that a parent tit as a rule brought home only one prey item on each hunting trip. Under these circumstances, the random search criterion necessarily implies that the tit would visit one of these niches at random in each hunting trip, regardless of the places it had visited on the previous trips. The crucial point here is: would the tits actually do that? It is crucial because, if they did not (and I knew they didn’t), Luuk’s ‘expected’ frequencies in the diet would no longer be an appropriate basis on which the formation of search images was deduced. Besides, I was not comfortable with the duality in Luuk’s explanation: the formation of search images and the avoidance of monotonous diet.

By the time this paper of Luuk’s came out (in fact published posthumously), I had already finished my three-year field study at the research forest in Yamanaka. My routines in the field included mapping the foraging sites of the birds and recording the insects fed to the nestlings; I did the recording at the nests by direct observation as well as with an automatic camera I designed and built. So, I knew that my tits did not search the woods haphazardly for prey. Instead, each tit tended to make a series of trips to a particular site (a tree, a clump of shrubs, or a patch of ground vegetation) and brought home a particular kind of prey in succession. You might say: ‘That strongly suggests that the bird formed a search image of the insect after a successful catch’. Well, that is just using the term for conveniently referring to a complex process of learning and memorizing the whereabouts of the things of interest. The formation of a search image that Luuk envisaged was altogether different from the above process of learning. As he described it explicitly, it was a specific (albeit hypothetical) process of recognizing the insect by the bird through its eye among noisy surroundings. It would serve like a template, as it were, of the image of the object formed in the bird’s brain by repeated encounters. An image received through the eye would then be recognized by the brain if it matched the template.

In contrast, to understand the bird returning to the same spot repeatedly to look for a particular kind of prey, as I observed, would not have required the assumption of a specific mechanism of visually detecting (recognizing) the prey at the hunting site. All we need to assume is simply that the bird learned where the prey could be readily found. I do not mean that the formation of search images as Luuk envisaged is improbable. All I need to say is: the postulation of the specific mechanism is unnecessary to understand the cause of the trend that Luuk observed, i.e. the pattern of deviation from the ‘expected’ in the random encounter hypothesis. So, what is an alternative view? Before I delve into it, I would like to note one thing.

I often wondered why Luuk did not consider any alternative to the random-encounter hypothesis before he decided on it. He had first-hand knowledge of tits’ searching behaviour in the woods, yet he did not make use of his knowledge. Why? I do not believe that he was satisfied with his own interpretation: it is hard to believe that a researcher of his calibre would have been. He was the youngest of the remarkable Tinbergen brothers: the eldest, Jan, was an economist, a pioneer in econometrics, and second brother Nikolaas (Niko) was, as well known among biologists, a founder of comparative ethology. Both were Nobel laureates in their own fields. Luuk was a professor in zoology, whose earlier works had been highly regarded by his colleagues, but died young at the age of 39. He was very close to his brother Niko, and was greatly influenced by him. Niko was an extremely rigorous thinker: in fact, in later years, I was a regular in his evening student discussion group in Oxford. He never allowed us students to leave even a trace of ambiguity in reasoning; he never allowed us to make an assumption without a logically valid basis. Luuk must also have been influenced by Jan in rigorous statistical inferences. So, it is rather surprising that Luuk did not put more thought into his last work. I think I know why: he took his own life, I am sad to say, and he must have been going through a tough time dealing with the emotional ordeal. So, he could not have put as much thought into the work as he would have otherwise. He left drafts of his work which were published posthumously in two parts: Professor Herman Klomp (a close friend of Luuk’s and involved in the publication of the work by Luuk) told me the story when I visited him at the Agricultural University of Wageningen in the early 1960s. I wish I could have met Luuk; I would have had a great time chatting with him.

In the following, I talk about an alternative theory that I conceived during my graduate work in two parts: one at the Fuji Research Forest in Yamanaka, Japan (henceforth the Yamanaka study) and the other at Wytham Woods, near Oxford, UK (the Wytham study).

1.3 Alternative Theory: Hunting by Profitability

1.3.1 The Yamanaka Study

A major objective of my field study in Yamanaka in the summers of 1958–1960 was to find what the tits fed their young with. I recorded 3300 prey items in a total of 210 hours of direct observations at a few nests, and 12,000 photographs on 53 nestling-days at several other nests. But I had a serious problem: I could not identify most of the insects by name, as they were mostly in their larval stages. There were also quite a few adult moths but, before being taken to the nests, their wings had been removed by the parent tits. However, I could distinguish most of them by their characteristic appearance and colour. So, I recorded them by the codes I invented for my own use. I needed an illustrated identification guide of caterpillars, but none was available then. Detailed descriptions of morphological keys for identification in most taxonomic books did not help, as I could not examine each specimen closely in my hand. So, I did not know for sure where these caterpillars occurred in the woods. Another problem was that, by the time Luuk Tinbergen’s paper came to my attention in 1961, I had already finished my field work and did not have an opportunity to test his theory. I was only sure that what I had observed in Yamanaka did not agree with Luuk’s interpretation.

1.3.2 Study in Wytham Woods

In the late summer of 1962, I obtained a British Council scholarship for further study at the Edward Grey Institute of Field Ornithology (EGI), Oxford, to join its long-term project on the genus Parus under the directorship of Dr David E. Lack. There, as David suggested, I concentrated on the hunting behaviour of the great tit (Parus major newtoni, the British subspecies of Parus major) with a supplementary observation of the blue tit (Parus caeruleus, now moved to the genus Cyanistes).

The EGI tits studies were being conducted mostly in Wytham Woods, a biological reserve of the University. Also, Professor George C. Varley of the Hope Department of Entomology was conducting his population studies of the oak-feeding insects with his right-hand man, George R. Gradwell. They were interested in my proposed study and allowed me to set up my recorder camera at a corner of their study plot. Also, they provided me with all sorts of information on the insects in Wytham, including their famous work on the winter moth, Operophtera brumata, which was in fact one of the most important food sources for the tits. George Gradwell loaned me the Victorian book The larvae of the British Lepidoptera and their food plants by Owen S. Wilson with beautifully hand-drawn illustrations by Eleonora Wilson. After a series of on-site tutorials George Gradwell gave me in the woods, I could identify most of the caterpillars. He also provided me with his original population data of the oak-feeding insects. So, I was all set. In Wytham, I took 29,000 photographs on 97 nestling-days over three seasons (1963–1965) at seven nests. Most aspects of the tits’ hunting behaviour that I had already observed in my Yamanaka study were confirmed in greater detail in the Wytham study.

1.3.2.1 Key Features in Tits’ Hunting Behaviour

I found several key features of the tits’ behaviour that provided bases of my alternative theory. First, all individual great tits (both in Yamanaka and Wytham) brought home almost invariably only one prey item on each hunting trip: the closely related blue and coal tits often carried several items. However, except for the first few days after all chicks hatched, each prey item brought to the nest was large; the majority of the caterpillars looked fully mature.

Second, as already mentioned, an individual tit tended to make a number of trips consecutively to the same hunting site (e.g. a particular tree) and brought home the same prey species (a hunting bout, say); but after a while switched to a different prey species in another bout; and so on to repeat the cycle. This tendency is clearly seen in the series of food brought home by a parent tit plotted on a chart over the course of a day (cf. figure 9, p. 647, in Royama, Reference Royama1970). You would see a run of a particular prey species for a certain length of time, usually not much more than half an hour, only sporadically interrupted by odd individuals of different species, followed by a run of another kind of prey for another time period.

Third, as already pointed out, different caterpillars occurred on distinct plant species. For example, one occurred mainly on oak but another on hawthorn, etc. However, even if occurring on the same tree, the caterpillars therein tended to be resting during the daytime in distinct micro-niches, e.g. leaves, branches, or the trunk of the tree. This suggests that changes in the runs of prey species from one bout to another indicate changes in micro-niches for hunting. Furthermore, each run of a given species often began abruptly, a tendency that happened more often than a run starting with sporadic appearances of a particular species mixed with others. This implies that, at the beginning of each hunting bout, the tits had already decided exactly where to hunt for what prey.

These features convinced me that the tits were very efficient hunters. They must have been able to assess how profitable a given prey species, or a given hunting site, could be. They must have been able to assess it (call it ‘profitability’) by the amount (e.g. most practically, biomass) of food collected in a given hunting bout. Then, the tits must have frequently been sampling many potential hunting sites to evaluate which site was more profitable than other sites and accordingly to allot their time among those sites to gain most profit. This implies that sampling different sites must be a regularly and frequently conducted routine because of the ever-changing levels of profitability among the potential sites.

I now describe in detail what I saw and examine the factors and conditions that most likely determined the profitability of a hunting site. [Note: I invented the term ‘profitability’. As English is not my native tongue, I was not sure if it was appropriate. So, I consulted some colleagues at the Zoology Department. I was encouraged to stick with it: even the Journal of Animal Ecology accepted the term without fuss.]

1.3.2.2 Chronological Changes in Profitability and Successions of Prey Composition in Tits’ Diet

To describe what I observed in Wytham (where I could identify most of the prey), I conveniently divide the nestling period into three major subperiods by the timing of the occurrence of given prey species in the diet. [Note: For the first few days after hatching, the nestlings were fed with small pieces of food, e.g. spiders or a fragment of a large piece of whatever it was. The fragments were difficult to identify on the film. So, I did not operate the camera until a few days later when the chicks were fed mostly with mature larvae. You have to see it for yourself to believe how big a piece a young chick can swallow. It can swallow an item as big as its head. Incredible!]

Early period (the last week in May). For the first 3 days of operating the camera at one nest in 1963 (cf. appendix table 2, p. 660, in Royama, Reference Royama1970), the mature larvae of the two similar-sized geometrids, Operophtera brumata (winter moth) and Oporinia dilutata (November moth; syn.: Epirrita dilutata), constituted almost 75% (in number) of all prey items brought home by the parent tits. Although comparatively small in size (20–25 mm long), O. brumata and E. dilutata combined were by far the most abundant prey species in the woods in this early period.

Of the other items, about 14% consisted of miscellaneous prey, including: tortricoid larvae, mainly Tortrix vilidana (green totrix), some spiders, and adult dipterans, including tipulids (crane flies), bibionids (March flies), empidids (dagger flies), and syrphids (hoverflies). The individual sizes of these miscellaneous prey were mostly no larger than that of a brumata/dilutata larva. The remaining 9%, however, were the larvae of the two similar-sized geometrids, Colotois pennaria (feathered thorn) and Phigaria pilosaria (pale brindled beauty); these are combined in one category, the pennaria/pilosaria larvae. Each of them was twice the body length, and at least 4 or 5 times in volume, as a brumata/dilutata larva. Few other caterpillars as large as or larger than a pennaria/pilosaria larva were around in the woods during this period of the season.

Now, the number of brumata/dilutata larvae brought home per day steadily declined towards the beginning of June: from 441 on 23 May down to 323 on 27 May; sharply down to 170 on 28 May; further down to 89 on 31 May; only 17 in total during the first week in June; and that was it. I considered two possible causes of the decline: one was certain and the other very probable. First, these larvae began to pupate. Like many other caterpillars, they descended to the ground to pupate just under the duff, meaning they were no longer available to the tits. However, George Gradwell’s trap catches showed that their descent for pupation did not begin until 29 May; the number descending peaked 3 days later; and by 10 June, few larvae remained on the trees. In other words, until 28 May, these larvae must have been as available to the tits as they were on the first 3 days. So, the substantial decline in the number brought home from 23 to 28 May could not be attributable to pupation.

The other and probable cause of the decline was the larvae of the other geometrids, C. pennaria and P. pilosaria, rapidly becoming fully mature. Although much less abundant than brumata/dilutata larvae in number, they were individually at least 4 times larger in volume. So one hunting trip by the tit that yielded a single pennaria/pilosaria larva would have been worth as much as 4 or 5 trips each with one brumata/dilutata larva; remember that, somehow, an individual great tit (be it Japanese or British) brought home only one prey item per trip regardless of its size. As the pennaria/pilosaria larvae were rapidly maturing, they became more profitable, and the tits steadily lost interest in the smaller prey, even though they were still as abundant as before. Thus, while the number of the large geometrids brought home increased from 49 on 23 May to 163 on 28 May, the number of small prey (O. brumata etc.) declined from more than 500 to less than 200.

Intermediate period. Following the first week of observation, in addition to the pennaria/pilosaria larvae, the parent tits began to regularly bring home the mature larvae of several noctuids, including most frequently Allophyes oxyacanthae (green-brindled crescent) and less frequently Orthosia populeti (lead-coloured drab), Amphipyra pyramidea (copper underwing), Brachionycha sphinx (sprawler), and Griposia aprilina (Merveille-du-jour; syn.: Dichonia aprilina). The successions of the prey species in the tits’ diet are most certainly attributable to the fact that the larvae of different kinds were becoming as profitable (as they mature) as the preceding pennaria/pilosaria larvae. I will now elaborate a little more on this issue.

The first major prey of the season, brumata/dilutata larvae, is reported to feed on many varieties of plants. There were at least a dozen eligible trees and shrubs in Wytham, but the oak (Quercus robur) was by far the most preferred. The larger geometrid larvae, pennaria/pilosaria, are reported to feed on a variety of plants, as do brumata/dilutata larvae. But in Wytham, according to George Gradwell, they were fairly rare on oak; they occurred more frequently on other trees or shrubs, such as ash, hazel, or hawthorn. Thus, in the tits’ diet, the transition from the brumata/dilutata to the pennaria/pilosaria larvae means that the tits were actually switching their main hunting sites from oak to non-oak plants.

This switching of main hunting sites by the tits was manifested more clearly in the appearance of the noctuid larvae in the nestling diet after the brumata/dilutata larvae were rarely brought home: Allophyes oxyacanthae fed mostly on hawthorn (Crataegus monogyna) and blackthorn (Prunus spinosa); Orthosia populeti fed exclusively on aspen (Populus tremula); and Griposia aprilina fed exclusively on oak. [Note: the A. pyramidea and B. sphinx larvae reportedly feed on a variety of trees, including oak, hawthorn, and aspen, but their frequency occurrences in the diet closely coincided with the rise and fall of the O. populeti as well as the A. oxyacanthae larvae. This indicates that A. pyramidea and B. sphinx were likely collected mainly on aspen and hawthorn.] The simultaneous appearance of certain key prey species in the daily menu suggests that the tits were hunting on a number of different tree or shrub species, switching frequently from one site to another during the day.

Furthermore, during the daytime, some larvae stayed on the leaves (e.g. all these green caterpillars: Op. brumata, Orth. populeti, A. pyramidea, and B. sphinx), while others hid elsewhere. For example, A. oxyacanthae larvae, which perfectly resemble a piece of bark of the hawthorn, most often rested on the trunk or thick branches, whereas the pennaria/pilosaria larvae stayed on small branches with a posture to mimic a twig. G. aprilina larvae, although feeding exclusively on oak leaves, were dark-coloured and hid in crevices in the tree trunk during the daytime. In other words, the tits not only made successive trips to a given tree species for a while before switching to a different one, but they also knew what to search for in a specific micro-niche (leaves, branches, or trunk) on each tree.

Late period. After the first week of June, the pennaria/pilosaria larvae and aforementioned noctuids in turn began to decline –they were still regularly brought to the nest, but in a reduced frequency. The beginning of their decline coincided more or less with the time when the tits started bringing home a large number of tortricoid pupae, mostly Tortrix viridana (green tortrix), together with the larvae of a dozen species, including several Orthosia species – O. incerta (clouded drab), O. gothica (Hebrew character), O. cruda (small quaker), but mostly O. stabilis (common quaker; syn.: O. cerasi) – as well as the geometrids Erannis leucophaearia (spring usher; now in the genus Agriopi) and Eupithecia irriguata (marbled pug), and even the hairy lymantrid caterpillars Lymantria monacha (black arches), although in single-digit numbers.

The significance of the above transition in species composition in the tits’ diet is that all of these late-occurring prey species fed either exclusively or preferentially on oak; moreover, most of them, as they were green-coloured, stayed on leaves during the daytime. In the meantime, Griposia aprilina, even though it was an oak specialist, dramatically declined in the diet: as already mentioned, this caterpillar, being dark-coloured, hid itself in crevices in the oak tree trunk during the daytime. Clearly, the transition of the species composition in the nestling diet was due to a shift in the main hunting sites from largely non-oak (during the preceding period) to oak, especially its foliage. This interpretation can also be extended to explain the following fact.

Incidental catches. As already mentioned, earlier in the season when the tits were collecting a large number of brumata/dilutata larvae, they also collected some dipteran adults quite regularly, if only in small numbers. However, when the tits switched their attention from the brumata/dilutata larvae on the oak foliage to the non-oak-feeding larvae, the dipteran adults were brought to nest only sporadically. However, during the late period when the tits switched their hunting sites back to the oak foliage, some of the dipterans, mostly the tipulids, reappeared in the diet. This indicates that the tits must have collected those dipterans as they came across them while searching in oak foliage for the caterpillars as the main source of food for the nestlings. In other words, the dipterans were likely incidental catches because they were obviously a trivial source of food compared with the fat and juicy noctuid larvae. The parent tits most probably ate those small items as they came across them while searching for the main foods to bring home.

Although the foregoing events were fairly easy to comprehend, there was one thing which puzzled me for quite a while.

1.3.2.3 A Puzzle to Solve

The miscellaneous tortricoid pupae, collected by the tits in exceedingly large numbers in the late period, were mostly Tortrix viridana. They were extremely abundant at an outbreak level in the years I was working in Wytham; its host tree, almost exclusively oak, was heavily defoliated. As a tortricid (leaf roller), the larva folds back the tip of a lobe of the oak leaf over itself, like a Mexican taco wrap. When disturbed, it swiftly wriggles out of the wrap (forwards or backwards) and drops on a silk, a typical predator-avoidance action of many tortricid larvae. So, they are not particularly easy prey for the great tit, which is not as adept at catching these larvae as its cousin the blue tit.

Now, the viridana larvae on the trees had been fully mature at the very beginning of the observation (when brumata/dilutata larvae were still being brought home in large numbers) and were more abundant (in my study years) than brumata/dilutata larvae with a big margin. Besides, viridana larvae shared the oak leaves with brumata/dilutata larvae. Nonetheless, during that period of time, the tits brought home viridana larvae in only small numbers compared with brumata/dilutata larvae. Initially, I thought that it was simply because a viridana larva was not an easy prey because of its swift reaction to predators, and that the tits took only those in a prepupal stage when they must have been easier to catch.

Eventually, I realized that the problems were not that simple. The viridana larvae began to pupate in the last week of May, at more or less the same time as the brumata/dilutata larvae. However, unlike the brumata/dilutata, the viridana larvae did not descend to the ground but stayed on the leaves (within the ‘taco’ wraps) to pupate. Therefore, they were still available in large numbers as easy prey.

By the end of May, the majority had pupated. Nonetheless, the tits paid them little attention, but collected them only sporadically and in single-digit numbers per day; they appeared to be no more than incidental catches. All of a sudden, however, the (year 1963) pair of tits began to collect the pupae in large numbers: 35 on 6 June, which jumped to 272 on 8 June, and then a whopping 571 a few days later. It took me a while before I began to comprehend. What I realized was the following. There are two distinct issues to recognize: the profitability of the pupae as an individual species and the profitability of the site (oak foliage) of which the viridana pupae were a constituent. Let me begin with the first issue.

Each tortricoid pupa was so small that 7 or 8 of them could be worth just about a single large, mature noctuid larva (e.g. Allophyes oxyacanthae). Translating this into 571 pupae, and allowing for my crude estimate, they were worth, on average, only 70–80 large larvae. On the other hand, the tits routinely brought home well over 100 of these large larvae a day. To compensate for the small size of the pupae, the tits had to make many more trips between the nest and the hunting site. Was it really worthwhile allotting any amount of time to collecting the super-abundant but individually small prey? [Remember again that, somehow, the great tits brought home only one item at a time regardless of size.] To investigate this issue, I needed to assess what the difference in profitability could actually have been between the pupae and the regularly occurring larger larvae. But how could I do it? After a while, I found a way: crude, but good enough a trick for the purpose.

I looked at a daily chart of the prey items brought to the nest by one female tit; I picked the female simply because the chart I happened to use was that of the female. Further, I arbitrarily picked a long, uninterrupted run of A. oxyacanthae larvae and that of the viridana pupae, each with a very short average interval between consecutive trips. I picked those runs for the following reason: during such a run, the tit must have minimized the time spent in activities other than those necessary for collecting an individual prey, i.e. travelling from the nest to the hunting site, finding a prey (very quickly), picking it up, bringing it to the nest, feeding it to a nestling, going back to the hunting site, repeating the cycle, and little else. Then, I divided the number of the prey individuals by the length (in minutes) of each run to estimate the minimum average rate at which a single prey individual could be collected and brought home by an individual tit per unit time.

In the particular runs I examined, 16 oxyacanthae larvae were brought home without interruption over 28 minutes (or 5.7 larva/10 min) and 25 viridana pupae over 13 minutes (or 19.2 pupae/10 min). Now, a full-grown oxyacanthae larva would be 7–8 times larger in volume than a viridana pupa, so that 19 pupae would be worth only about 2.5 or so larvae, although more nutritional values could have been packed in a pupa. In other words, hunting oxyacanthae larvae for a given length of time (given their abundance on that day) could yield a profit at least twice as much as hunting viridana pupae. As crude as it may be, the above calculation suggests that hunting the oxyacanthae larvae must have been substantially more profitable than hunting the viridana pupae.

With the above result, I further investigated how the tits actually allotted their time in hunting between the two types of prey. I looked into a day (8 June) on which the tits (male and female combined) collected 272 viridana pupae. On the same day, they brought home 202 oxyacanthae larvae. Applying the rate of hunting per unit time (for a single parent bird), I figured that it could have taken 68 minutes for the pair of tits to collect the 272 pupae, as opposed to 177 minutes to collect the 202 oxyacanthae larvae.

Setting aside how to interpret the above estimates in absolute value, the point of interest here is the relative values: about 2.6 times more minutes were allotted to the more profitable larvae than to the less profitable pupae on that day. But, why did the tits bother allotting any of their time to the less profitable viridana pupae? Why didn’t they concentrate solely on the more profitable oxyacanthae larvae? You might say: to avoid a monotonous menu, as Luuk Tinbergen suggested. I doubt it, as I have a more plausible explanation.

1.3.2.4 Allotment of Hunting Time Among Prey Species vs. Among Hunting Sites

An obvious and crucial point to consider is the profitability of a site as a whole, e.g. oak foliage where the viridana pupae were just a constituent, albeit a major one. This suggests that the tits must have allotted their time primarily among the productive sites, rather than among the individual prey species: the hawthorn trunk–branch site vs. oak foliage, rather than oxyacanthae larvae vs. viridana pupae. I had this idea when I recognized the following.

On the hawthorn trunks were mainly oxyacanthae larvae and few else; similarly, among the twigs there were pennaria/pilosaria larvae and few else. In contrast, by the end of the first week of June, as many as a dozen large species had become available among the oak foliage simultaneously. In particular, in the oak leaves I sampled, I quite often found, beside numerous viridana pupae, large noctuid caterpillars. This prompted me to look closely into the daily charts of prey brought home by the tits (cf. figure 3a and 9 in Royama, Reference Royama1970). Indeed, I found that a single run of viridana pupae was often mixed with large noctuid larvae of oak feeding species. It was particularly noticeable that these large caterpillars tended to occur individually in the run of the pupae but seldom occurred in a series. At first, I thought that these caterpillars were merely incidental finds in the oak foliage where the tits were looking primarily for the pupae. However, this does not make good sense because, as already argued, viridana pupae on their own could not have been as profitable a source as oxyacanthae larvae. The following is a more plausible interpretation.

Recall again that a parent tit brought home only one food item at a time. Thus, at each trip to the oak foliage, the first thing that the tit encountered would probably have been a viridana pupa, but every now and then it could have been a large caterpillar. Because of the sheer number of pupae on the foliage, the odds are that the tits found and collected them in succession, whereas encountering a big caterpillar was more sporadic. Nonetheless, the tit could count on big, juicy caterpillars that occur quite regularly in the foliage. In other words, the tits must have assessed the oak foliage to be as profitable a hunting site as the hawthorn trunk–twigs site where oxyacanthae/pennaria/pilosaria larvae were collected. In short, it is likely that the tit became attached to a profitable site and took whatever it came across first, rather than looking for a particular prey species with its image in mind. The crucial question is: how do the tits assess the profitability of a given site?

1.3.2.5 How Do Tits Assess the Profitability of a Hunting Site?

In theory, the profitability can be measured by the amount (biomass) of food the tit can collect there to bring home per unit time. But, pragmatically, how could the tit do it without a stopwatch or weighing scale? The answer is by watching the chicks’ state of hunger or satiation.

In the Yamanaka study, I spent 210 hours (mainly in the first year before I built the automatic camera) directly watching what was going on inside the nest boxes. I saw that each time a parent tit came home with food, every chick stretched its neck upwards as high as possible and with the gape wide open; it did this no matter how hungry it was. The parent placed the food item in one of the gapes, apparently at random. If the chick was hungry, it immediately grabbed the food and began to swallow it. The parent watched it until the chick finished swallowing the prey and then went out to collect more. On the other hand, if not hungry, the chick did not swallow the food as quickly; often, it even kept the gape wide open with the piece of food already placed in it. Then, the parent pulled it out of the gape and put it in another gape, and so on, until a hungry chick was found. This had already been noted by the EGI alumna Monica Betts, whose pioneering work on the tits’ feeding habit guided me in field observations.

If the chicks were quick to react, the parents kept collecting more food. But when all the chicks became reluctant to swallow the food quickly, the parents relaxed their hunting activities. Because the parents tended to stick to a particular site for a while, they must have been able to assess whether the site was profitable by the intensity of the chicks’ demand: hunting at a good site would satisfy the chicks more quickly than hunting at a lesser site. However, the parents must also spend time regularly sampling other potential sites to assess ever-changing levels of profitability in the woods. They would probably do this at a slack time when feeding themselves.

1.4 Profitability Curve

Nobody, I hope, would disagree that the profitability of a prey species or a hunting site must be positively correlated with the abundance of the prey therein. But what sort of relationship is likely expected? An idealization (or oversimplification) would help us to make an inquiry into the underlying principle. So, let us assume that the tit hunts at a given site with no structure (a single niche, e.g. leaves or trunk of a tree) within which prey individuals of just one kind are distributed at random.

Now, we plot the abundance (density) of the prey at the site on the horizontal axis and, on the vertical axis, we plot the average number of prey individuals (mean total biomass) that the tit could collect and bring home in a unit period of time, i.e. the measure of profitability. Let us call the resultant curve ‘the profitability curve’, as illustrated in Figure 1.1. In particular, the curve starts at the origin (zero density, zero profitability) and increases as prey density increases, but with an ever-decreasing rate of increase. More particularly, as prey density increases, the curve increases monotonically and smoothly to ever more closely (i.e. asymptotically) approach a horizontal line which is the maximum profitability that the tit can gain at the site. The reason for this shape is as follows.

Figure 1.1 Idealized example of profitability curve.

Each hunting trip consumes a certain amount of time required for two major categories of activity: searching for prey within the site and dealing with the prey found. The latter category would typically involve: capturing (chasing if necessary) prey; preparing the catch (killing and removing appendages, if any, e.g. wings); flying back to the nest with it and feeding it to a chick; tending the chicks for a while (e.g. removal of faeces); and flying out to the site to begin searching for another prey.

[Digression: Let me talk about an (extreme) example of how a tit prepared a prey it caught before bringing it home. For some time, I was wondering about a couple of mysterious things in the woods. One was the shrivelled bodies of large hairy caterpillars, like Lymantria monacha (black arches), hanging on branches. I just thought they were the carcasses of those that had died of a viral disease. Meanwhile, every now and then, I saw a tit bringing home a blob of dark, shapeless material, looking soft, fresh and juicy; whatever it was, the chicks seemed to love it. Then, one day I saw a tit pick up a fully mature, very hairy L. monacha caterpillar in the oak foliage, carry it over to a branch, grab its body with one foot, and repeatedly peck hard at the joint between the head and the first thoracic segment. As the tit decapitated it and pulled the head off the body, out came the whole length of innards. The tit rolled them up neatly into a blob around its beak (just like those I saw fed to the chicks in the nest) and flew away, leaving the empty hairy caterpillar skin hanging on the branch. It made my day.]

Now, just think of a tit searching for prey along a random path, forgetting about capturing the prey for now. That is, just think of the number of prey individuals that the tit would encounter per unit time, given the prey density, each time it searched the site. Because we are assuming a random occurrence of prey individuals at a site with no structure, the tit would be expected to encounter the prey in proportion to prey density: here, I am plagiarizing Luuk Tinbergen’s random search. The number encountered per unit time at the site is represented by the straight (slanted) line in the graph, starting at the origin (0 density, 0 encounter) and at a certain angle with the (horizontal) density axis. The slope of the line is the proportionality factor, determined by the level of ease (or difficulty) in finding the prey within the site: the easier it is to find, the steeper the slope. This serves as a reference line for drawing a profitability curve.

Now that the profitability (i.e. the number of prey found, captured, and dealt with per unit time) could never exceed the number found (encountered), the profitability curve should always be below the slanted reference line, except at the origin: no prey, no encounter, no profit. When the prey was present but very scarce, only a very few individuals would be encountered per unit time. Then, the proportion of total time spent in hunting prey would be used up mostly in searching: or conversely, little time would be spent in dealing with the prey captured. Thus, the steepness of the profitability curve would initially be very close to, albeit never steeper than, the reference (encounter) line.

As the prey abundance increased a little further, accordingly a few more individuals would be encountered and captured and, hence, a slightly higher portion of time would be spent dealing with the catches. So, the number found, captured, and brought home per unit time (i.e. the profitability), albeit increasing with prey abundance, deviates from the slanted (encounter) line a little further downwards, and so on. For a further increase in prey density, the profitability curve should sooner or later approach an asymptotic level. This is because, at a high level of prey abundance, comparatively little time would be needed to find prey, whereas dealing with the catches would occupy most of the total time spent hunting. In other words, the profitability curve (the number of prey brought home per unit time) would practically reach its ceiling and would stay there no matter how much more abundant the prey became from then on. [The profitability curve is well described by the popular ‘disk equation’ of C. S. Holling (Reference Holling1959): details are in Royama (Reference Royama1970), pp. 642–3.]

1.5 Allotment of Hunting Time among Different Sites

I now look into how the tit would allot its time hunting among potential sites. The foregoing profitability curve provides the basis for this inquiry in the following idealization. [Here, I generalize ‘profitability’ in terms of the total biomass of food fed to the chicks per unit effort (energy expenditure) of hunting.] For simplicity, consider how the tit allots its time between two sites, A and B, and let PA and PB designate the level of profitability at A and B, respectively. Further, let PA vary, while PB is fixed at its maximum level. Let us consider the three situations: PA is more or less equal to PB; PA is very low; and PA is in between.

In the first situation, there would be no advantage (no reason) for the tit to select A over B, or vice versa. Then, the allocation of time between the two sites would become a matter of arbitrary choice and, hence, would be subject purely to a chance variation. That is, we would expect a fifty–fifty allocation of time between A and B. Second, if PA is very low, we would expect that the tit would allot little time to site A. In the third situation, in which PA is less than PB but not very low, what we would expect to happen is not readily conceivable.

Theoretically, there would be no point for the tit to allot any portion of time to site A when PA was less than PB. But it is not realistic to expect that the tit spends little time in A until PA becomes close to PB and that, at that moment, there occurs a sudden jump in time spent at site A. What we would expect is the following. The tit spends a certain amount of time at A to monitor changes in PA so as not to run the risk of prematurely accepting or rejecting the site as profitable or unprofitable. On the other hand, spending too much time would be a waste. It follows that the tits must optimize the allocation of time. This depends on their ability to assess PA. There could even be considerable individual variations among the tits under varying environmental conditions. [Note: You might wonder if the tits have this sort of cognitive ability to assess the profitability. Surely it is unlikely that the tits solve the problem intellectually. However, I am not surprised if their instinctive behaviour, evolved through natural selection, would guide them to the solution.]

1.5.1 Consequence of Hunting by Profitability

In these circumstances, what consequence do we expect in terms of the amount of prey (from site A) fed to the chicks in the nest?

Suppose that we made repeated observations to measure the proportion of the total time the tit allotted to site A, and plot it on the vertical (percentage) axis against prey abundance at the site on the horizontal axis. It would most certainly form a scattergram along a sigmoidal curve as an average trend: starting with a slow rate of increase, followed by a rapid increase, but slowing down again to level off. At what level of prey abundance and how rapidly the curve (average trend) starts increasing should depend on the efficacy of the tits’ instinctive assessment.

Now, multiplying the profitability PA of site A by the time allotted to the site, we would expect in principle that the amount of prey (from A) fed to the chicks per unit time, as plotted against the prey abundance (in A), is likely to be sigmoidal. This explains the trend that Luuk Tinbergen actually observed in his study with respect to Panolis flammea larvae: few of the larvae were fed to the chicks when scarce in the woods; the proportion of the larvae in the diet increased substantially when moderately abundant in the woods; but, for a higher level of abundance, the percentage in the diet did not keep increasing but levelled off.

1.6 Hunting by Profitability as Principle

My conception of hunting by profitability is not just a means of comprehending the tits’ behaviour. As I see it, the survival (and fitness) of an animal would in principle depend on how to allot optimal amounts of time (effort, energy) among its feeding sites (niches) according to the profitability that each site would provide. I will come back to this issue in Chapters 8 and 9 in which I talk about the evolution of niche selection by animals in general.

Figure 0

Figure 1.1 Idealized example of profitability curve.

You have Access

Send book to Kindle

To send this book to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle.

Note you can select to send to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Available formats
×

Send book to Dropbox

To send content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about sending content to Dropbox.

Available formats
×

Send book to Google Drive

To send content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about sending content to Google Drive.

Available formats
×