Dr Richard W. G. Caldow, Institute of Terrestrial Ecology, Furzebrook Research Station, Furzebrook Road, Wareham, Dorset BH20 5AS, UK. Tel. 01929 551518; Fax 01929 551087. E-mail: R.Caldow@ite.ac.uk
1. Individual variation in the competitive ability of foraging animals arises from variation in their intrinsic foraging efficiency and in their susceptibility to interference from competitors. Empirical and theoretical studies have concentrated on quantifying the latter and examining its role in determining the distribution and dynamics of animal populations, but have seldom considered the role of variation in foraging efficiency. Using the frequency of occurrence of oystercatchers in supplementary feeding habitats as an index of their competitive ability, we assessed the relative importance of foraging efficiency and susceptibility to interference in determining the overall competitive ability of an individual.
2. Individual mussel-feeding oystercatchers varied in their tendency to supplement their low-tide intake by feeding on other prey on upshore tidal flats and in fields. Individuals that opened mussels by stabbing occurred on upshore flats more often than birds which hammered mussels, whereas young birds were more likely to visit the fields than older birds. This difference between habitats emphasized the need to understand the individual differences underlying these class effects.
3. Foraging efficiency increased with age and differed between feeding methods. Dominance also increased with age, but did not differ between feeding methods. An individual's foraging efficiency was not related to its dominance.
4. Individual variation in the usage of either upshore flats or fields, and of each habitat separately, was related to individual variation in foraging efficiency, but not to variation in dominance. Individuals of poor intrinsic foraging ability made greater use of supplementary feeding habitats than did more efficient foragers.
5. Our results show that, even in an interference-prone system and across a wide range of circumstances, individual variation in foraging efficiency is the major determinant of overall competitive ability. We believe, therefore, that this source of individual variation is of greater importance in determining variation in mortality risk within a population than the effort invested in its study hitherto would suggest. We suggest that a modelling approach is necessary to establish the competitive conditions under which susceptibility to interference might become the more important determinant of competitive ability and, hence, whether such conditions are ever likely to occur in natural populations. We argue that greater emphasis needs to be placed on identifying the determinants of foraging efficiency and its variation between individuals.
The importance of differences in competitive ability between individual animals to understanding population ecology was first emphasized by Łomnicki (1978, 1980, 1988). In recent years, there has been an increasing awareness of the need to quantify individual variation in the critical demographic variables, survival chances and reproductive success, and to identify those aspects of individual physiology and behaviour which are associated with such variation (Clutton-Brock 1988; Newton 1989; Goss-Custard 1996; Sutherland 1996). Individual variation in the competitive ability of foraging animals arises primarily from variation in two characteristics. An individual's foraging efficiency – its intrinsic ability to forage in the absence of competitors – determines how well it can cope with depletion of its food resources, which will be particularly severe at large population sizes. An individual's susceptibility to interference – the rate at which its intake rate changes with an increasing density of competitors – determines how well an individual can cope with the presence of other competitors around it (Goss-Custard & Durell 1987a; Sutherland & Parker 1992; Goss-Custard & Sutherland 1997). Many empirical studies have focused on measuring differences in susceptibility to interference between individuals (Ens & Goss-Custard 1984; Goss-Custard, Clarke & Durell 1984; Goss-Custard & Durell 1987a,b,c, 1988). Numerous theoretical studies have used models to explore the significance to population dynamics of the overall strength of interference and of the variation in susceptibility to interference between individuals (Sutherland 1992; Sutherland & Parker 1992; Sutherland & Dolman 1994). Others have explored the appropriateness of different models of interference (van der Meer & Ens 1997; Stillman, Goss-Custard & Caldow 1997). In contrast, few studies have even measured variation in foraging efficiency between individuals (Partridge 1976; Goss-Custard & Durell 1988), let alone addressed the relative importance of foraging efficiency and susceptibility to interference to individual's competitive ability or to population dynamics (although see Sutherland & Parker 1992 and Illius et al. 1995 for exceptions). The aim of this paper is to quantify the relative importance of individual variation in foraging efficiency and susceptibility to interference, as defined above, in determining variation in the overall competitive ability of a foraging shorebird.
Shorebirds typically feed on inter-tidal patches during low tide and rest at communal roosting sites at other times when the preferred inter-tidal patches are not available. However, in response to deterioration in the inter-tidal feeding conditions, shorebirds feed increasingly on upshore flats on the receding and advancing tides, and in fields at high-tide (Goss-Custard et al. 1996a; Hulscher 1996). This entails prolonged periods of foraging, and the use of areas where the risks of predation, parasitism and of accidents are likely to be elevated (Goater 1993; Cresswell 1994; Goater, Goss-Custard & Kennedy 1995; Goss-Custard et al. 1996a). The circumstances in which such feeding occurs and these potentially serious drawbacks to ‘supplementary’ feeding, suggest that it is a good indicator of the difficulty that shorebirds face in meeting their requirements.
Studies of oystercatchers (Haematopus ostralegus L.) have consistently revealed differences in the frequency of occurrence in supplementary feeding areas between birds of different ages and diets (see Goss-Custard et al. 1996a for review). Moreover, birds that feed on mussels (Mytilus edulis L.), but which open their prey in different ways, also differ in their usage of supplementary feeding areas (Goss-Custard & Durell 1988). Young birds are generally less efficient foragers and are more susceptible to interference than older birds (Goss-Custard & Durell 1987a,c). Birds that stab mussels are, on average, poorer foragers than those that hammer them (Goss-Custard & Durell 1988). These findings lead to the hypothesis that differences in the usage of supplementary feeding areas between birds of different classes arise from the underlying differences in the difficulty they have in meeting their energy requirements at low tide on the principal feeding grounds (Goss-Custard et al. 1996a), i.e. their competitive ability as foragers.
Goss-Custard & Durell (1988) showed that, in a sample of 12 adult stabbers, individual differences in foraging efficiency were of greater importance in determining variation in the frequency with which birds supplemented their intake in fields at high tide than were differences in dominance, a reliable index of their susceptibility to interference (Goss-Custard & Durell 1988; Ens & Cayford 1996). This finding was contrary to expectations (Goss-Custard & Durell 1988), but can be explained on the basis that, in an interference-prone system, the relative importance of foraging efficiency and susceptibility to interference in determining overall competitive ability is likely to depend upon competitor density. Goss-Custard & Durell (1988) noted that most of their searches for birds feeding in the fields were made on spring tides, when the density of oystercatchers on the mussel beds at low tide would have resulted in relatively little interference (Stillman et al. 1996). Moreover, adult stabbers, with which Goss-Custard & Durell were concerned are, among mussel-feeding oystercatchers, the least susceptible to interference (Goss-Custard & Durell 1987a,c, 1988). It is perhaps not surprising therefore that in this unique, empirical study, Goss-Custard & Durell (1988) found foraging efficiency to be the major determinant of competitive ability, as measured by field use.
A behaviour-based model of foraging oystercatchers has, however, predicted that, even under conditions of intense competition, an individual's foraging efficiency is more likely to be important in determining its chances of survival than its dominance (Goss-Custard et al. 1996b; Goss-Custard & Sutherland 1997). Results of Birds of Estuaries Enquiry (BoEE) counts indicate that the mean over-winter population of oystercatchers on the Exe estuary doubled between 1983 and 1984 (mean over-winter count of 1618), when the study of Goss-Custard & Durell (1988) was conducted, and the period of the current study (mean over-winter count 1989–90 of 3624 and 1990–91 of 3696) (S.E.A. le V. dit Durell, unpublished data). In this paper we examine the variation in the usage of supplementary feeding areas by individual oystercatchers to test whether, at this higher population density, individual variation in foraging efficiency, rather than in susceptibility to interference, remained the more important component of competitive ability and hence of greater influence in determining variation in over-winter mortality risk within the population.
Materials and methods
This study was conducted on individually ringed, mussel-feeding oystercatchers on the Exe estuary in England over two winters (September 1989–March 1990 and October 1990–March 1991). Birds ringed previously on the Exe estuary provided most of the adults in our study. Juvenile and immature birds were marked between 1989 and 1991. Individuals’ intake rates, dominances and feeding methods were measured on their low-tide feeding grounds. Observations were made in sequences of consecutive 5-min periods during each of which, records were made of the numbers and lengths of all mussels consumed and of the number and outcome of all encounters with other birds.
The intake rate achieved during each 5-min period was calculated from the numbers and lengths of mussels consumed, using established procedures (Goss-Custard & Durell 1987a; Goss-Custard et al. 1987). First, a sample of 50 mussels spanning the entire size range eaten by oystercatchers was collected. These mussels were used in trials under field-conditions to derive the relationship, for each observer, between estimated and actual mussel length. All estimated mussel lengths were then converted to actual lengths. Secondly, further samples of mussels (n = 50–100) were collected from each mussel-bed where observations were made. These were used to derive, for each mussel bed, the relationship between the actual length of a mussel and the ash-free dry mass of its flesh. The intake (mg ash-free dry mass) per 5-min period was then calculated by conversion of the actual length of each mussel consumed to its flesh content, according to the equation specific to the mussel-bed where the observation was made, and summation of the mass of flesh consumed across all the mussels eaten in that interval.
Interference among oystercatchers only occurs above a threshold density of competitors (Stillman et al. 1996). Intake rates achieved at competitor densities below this threshold represent, by definition, an individual's interference-free intake rate, which we assume indicates their intrinsic foraging efficiency. Oystercatchers occur at higher densities on the mussel beds of the Exe estuary at the beginning and end of each exposure period than at low-tide (Goss-Custard & Durell 1988). Records of intake rate collected during the first and last quarter of the exposure period were therefore excluded from calculations of birds’ foraging efficiencies. Following Caldow & Goss-Custard (1996), a bird's dominance was defined as the proportion of all encounters with other oystercatchers that was won by the focal bird. A focal bird was classified as the winner of an encounter when it initiated an attack and the recipient retreated, and when, as a recipient, it rebuffed another's attack. Preliminary analyses revealed no tidal effect on birds’ dominance scores and so encounters from all stages of the exposure period were included in the calculation of a bird's dominance, the index of its susceptibility to interference.
On the advancing and receding tide, birds can feed on two areas of intertidal, upshore flats. In the winter of 1989-1990 regular searches for ringed birds were made on one of these areas. In 1990-1991, both areas were searched equally often throughout the winter. Almost without exception, a ringed bird would use only one or other area. The tendency of a ringed individual to feed on upshore flats (pu) was calculated as:
where ru = the number of sightings of a bird on its principal upshore feeding area and nu is the number of searches of that area.
The use of fields was assessed by regular high-tide searches of all fields known to be used by oystercatchers. Individual fields were grouped into geographical blocks that were searched between six and 12 times each winter. Almost without exception, a ringed bird would only use fields within one particular block. Thus, the tendency of a ringed individual to feed in fields at high tide (pf) was calculated as:
where rf = the number of sightings of a bird on its principal block of fields and nf is the number of searches of those fields.
In order to exclude estimates of dominance and foraging efficiency that were likely to be particularly inaccurate, records of dominance based on fewer than 10 encounters and records of foraging efficiency based on less than half an hour of observations were excluded from the analyses. The resulting dataset comprised 96 records of the use of supplementary feeding areas by 1st (n = 6), 2nd (n = 21), 3rd (n = 18), 4th year (n = 6) and adult birds (n = 45) of which 83%, 76%, 33%, 17% and 16%, respectively, were records of stabbers. Preliminary analyses revealed no significant differences between ventral or dorsal hammerers (Hulscher 1996) in the usage of either upshore flats or fields. Accordingly, these feeding methods were merged and classified simply as hammerers in all analyses.
The GLM procedure of SAS (SAS 1989) was used to construct models relating the use of: (a) fields at high tide; (b) upshore flats on the advancing and receding tide; and (c) both habitats at either stage of the tidal cycle to (i) a bird's age and feeding method, and (ii) its dominance and foraging efficiency. Since the use of each supplementary feeding area was expressed as the proportion of searches on which a bird was seen, an empirical logistic transformation was applied to these binomial values as follows:
where r and n are as before, and parameters of the resulting logistic model fitted by weighted least squares regression (see McCullagh & Nelder 1989, p107). The optimum weight for each record is the inverse of its variance (i.e. 1/Var(Z)), where the variance for each individual record of field (Zf) or upshore (Zu) usage is estimated as:
(McCullagh & Nelder 1989, p. 107). This weighting depends on both the number of searches (n) and the frequency of sightings (r), and is expected to give more precise estimates of model parameters than when simply weighting by n.
For analyses of the usage of either habitat a simple un-weighted average of the logit-transformed proportions of usage of the two separate habitats was calculated as:
and the variance calculated as:
All models initially included a variable to denote the year of observation, and models of the usage of upshore flats or of fields each included a variable that characterized birds’ usage of the other habitat. Preliminary analyses indicated that these variables did not improve the fit of any of the models and so these were not included in the further analyses presented in this paper.
Variation in the use of supplementary feeding habitats
Approximately one-third of mussel-feeders were never seen in supplementary feeding habitats (i.e. either on fields or on upshore flats), and no individual was seen on more than 50% of occasions (Fig. 1a). Nonetheless, there was statistically significant variation between birds in the extent to which they visited supplementary feeding areas (Fig. 1a, χ2 = 415·5, d.f. = 95, P < 0·001). This pattern was also apparent in the usage of both fields and of upshore flats when these were examined separately (Fig. 1b). These two habitats appeared to have been used to the same extent by the population, the frequency distributions of pf and pu being statistically indistinguishable (Kolmogorov–Smirnov Test; D = 0·104; n = 96,96; P > 0·5). There was, however, no significant association between the tendency of individuals to visit fields and upshore flats (Fisher's exact test P = 0·872).
Relationship between age and feeding method and use of supplementary feeding habitats
The frequency with which birds were seen supplementing their low tide intake in either of the two available habitats was significantly related to both their age and to their feeding method (Table 1, Fig. 2a). Supplementary habitats were used more frequently by stabbers, irrespective of their age, and by young birds, irrespective of their feeding method. However, there was a clear difference in the relative importance of a bird's age and feeding method in determining its usage of the two alternative habitats (Table 1, Fig. 2b,c). Controlling for the effect of age, a bird's feeding method had a significant effect on its usage of upshore flats. The opposite was not the case. Stabbers occurred on upshore flats significantly more often than hammerers did (Fig. 2b). In contrast, a bird's usage of fields at high tide was independent of its feeding method, but strongly dependent upon its age. Younger birds were much more likely to visit the fields at high tide than older birds, irrespective of their feeding method (Fig. 2c).
Table 1. The effect of feeding method and age on the frequency with which individual oystercatchers supplemented their low tide intake by feeding on either supplementary feeding habitat, on upshore flats and in fields. Three weighted, linear regression models were constructed to show the effect of feeding method and age on logit-transformed frequency of supplementary feeding. For each feeding method and age, the values show the estimated coefficient, standard error and significance level. The coefficients for each feeding method are compared against each other while the coefficient for age is compared against zero. See Materials and methods for the details of the transformation and the weighting factors used in the models
n = 96; *P < 0·05; **P < 0·01; ***P < 0·001.
Interrelationships between age, feeding method, dominance and foraging efficiency
A bird's age and feeding method had significant partial effects on its foraging efficiency, which increased significantly with age in both feeding methods, and was greater among hammerers than stabbers across all ages (Table 2, Fig. 3a). Dominance increased with age but did not differ between feeding methods (Table 2, Fig. 3b). No relationship was found between a bird's foraging efficiency and its dominance (Pearson correlation coefficient; r = 0·082; n = 96; P > 0·1).
Table 2. The effect of feeding method and age on the foraging efficiency and dominance of individual oystercatchers. Age was log-transformed in the analysis of birds’ foraging efficiencies. Values given are the estimated coefficient, its standard error and significance level. The coefficients for each feeding method are compared against each other, while the coefficient for age is compared against zero
Relationship between foraging efficiency and dominance and the use of supplementary habitats
The relative importance of a bird's foraging efficiency and dominance in determining its reliance on supplementary habitats was consistent across models of the use of either habitat, upshore flats and fields (Table 3, Fig. 4a–c). Birds with low foraging efficiencies made greater use of supplementary feeding habitats than more efficient foragers (Table 3, Fig. 4a) by visiting either upshore flats or fields more frequently (Table 3, Fig.4b,c). In contrast, a bird's dominance, controlling for its foraging efficiency, had no significant effect on its usage of supplementary feeding habitats (Table 3, Fig. 4a–c).
Table 3. The effect of foraging efficiency and dominance on the frequency with which individual oystercatchers supplemented their low tide intake by feeding on either supplementary feeding habitat, on upshore flats or in fields. Three weighted, linear regression models were constructed to show the effect of foraging efficiency and dominance on logit-transformed frequency of supplementary feeding. For each variable the values show the estimated coefficient, standard error and significance level. See Materials and methods for the details of the transformation and the weighting factors used in the models
Our understanding of the factors that influence the tendency of oystercatchers to occur in supplementary feeding habitats has, until now, been based largely on simple analyses of differences between birds of different diets, ages and feeding methods (Goss-Custard et al. 1983, 1988; Goss-Custard et al. 1996a). These analyses have also focused solely on the usage of fields at high tide. Our analyses indicate that upshore flats are used as frequently by the population as a whole, as are fields. However, the overall effect of a bird's age and feeding method on its tendency to supplement its low-tide intake at other times and places does not reflect an equal importance of each of these characteristics in determining usage of fields and upshore flats. A bird's feeding method influenced its usage of upshore flats, whereas its age influenced its usage of fields. Moreover, there was no association between the frequencies with which individuals visited the two habitats. These unexpected findings emphasize the shortcomings of previous studies in focusing solely on the usage of fields at high-tide, and the need to identify the underlying causes of the association between a bird's age and feeding method and its usage of supplementary habitats.
In accord with previous studies (Goss-Custard & Durell 1987a,b, 1988; Stillman et al. 1996), birds of different ages differed in both their foraging efficiency and dominance, whereas birds of different feeding methods differed only in their foraging efficiency, and there was no association between a bird's foraging efficiency and its dominance. The results of the current study thus confirm our understanding of the ways in which birds of different age and feeding method differ in the two major characteristics that independently determine competitive ability.
Statistical models of the frequency of occurrence of birds on either fields or upshore flats, or on each habitat separately clearly showed that, across ages and feeding methods, an individual's foraging efficiency was of considerably more importance in determining its usage of these areas than was its dominance. Assuming that usage of supplementary feeding habitats is indicative of a mussel-feeding bird's ability to meet its requirements on mussel beds at low tide, i.e. its competitive ability on the principal feeding grounds, our results suggest that, even in this interference-prone system (Ens & Cayford 1996; Sutherland 1996; Stillman et al. 1997), a bird's foraging efficiency is, across a wide range of competitor densities, the principal determinant of competitive ability. Given the elevated risks of predation, parasitism and of accidents that are likely to be associated with feeding in supplementary habitats, and the fact that food in these areas becomes unavailable under freezing conditions (Goss-Custard et al. 1996a), a bird's foraging efficiency is likely to be the principal determinant of its over-winter mortality risk.
Our findings verify the preliminary results of Goss-Custard & Durell (1988). However, a number of features of our dataset, in comparison with that of Goss-Custard & Durell (1988), make the lack of a significant effect of susceptibility to interference more surprising. First, the mean over-winter population of oystercatchers in 1989–91 was more than twice that during the previous study (S.E.A. le V. dit Durell, unpublished data). Moreover, 75% of all searches of supplementary feeding areas between 1989 and 1991 were made on days of neap tides when the density of oystercatchers at low tide and, hence, the intensity of interference among mussel-feeding birds would have been at its greatest (Goss-Custard & Durell 1988). In addition, over half of the records in the current dataset were of juvenile and immature birds that, given their low dominance status, would have been particularly prone to interference (Goss-Custard & Durell 1987a,c).
Stillman et al. (1996) showed that, in oystercatchers, interference only occurs above a threshold density of competitors. This finding leads to the hypothesis that at competitor densities lower than this threshold, foraging efficiency will be the sole determinant of competitive ability and that as densities rise, its importance relative to that of susceptibility to interference will diminish. The precise shape of the function describing the relationship between the relative importance of foraging efficiency and susceptibility to interference in determining overall competitive ability and competitor density has yet to be determined. In support of the predictions generated by the behaviour-based models of foraging oystercatchers (Goss-Custard et al. 1996b; Goss-Custard & Sutherland 1997), our results indicate that, even when the behaviour on neap tides of the birds most susceptible to interference is considered, the densities at which oystercatchers forage on the Exe are insufficient for susceptibility to interference to become the principal determinant of competitive ability. A modelling approach will be required in order to establish the competitor densities at which this will occur and, hence, indicate whether such conditions are ever likely to be sustained in natural populations. We are currently exploring this approach.
Although a low competitive ability, arising from a poor foraging efficiency, requires birds to compensate by supplementing their intake, the strategy which they adopt in order to do so differs between age classes and feeding methods. Birds whose efficiency when feeding on mussels is poor primarily because of their age, are more likely to supplement their diet by visiting the fields than are birds whose efficiency is poor primarily because of their feeding method. These birds are more likely to supplement their diet by visiting the upshore flats rather than the fields. One possible explanation for this is that all prey items consumed in fields are soft-bodied invertebrates, mainly earthworms (Family Lumbricidae; Hulscher 1996), whereas on upshore flats, cockles (Cerastoderma edule L.) are a major prey item. Juvenile oystercatchers are as successful as adults when feeding on worms, at least on intertidal flats (Triplet 1989; Durell, Goss-Custard & Perez-Hurtada 1996). Thus, field-feeding may be a more profitable strategy for juveniles of low competitive ability, than feeding on upshore areas on the advancing and receding tides. In contrast, if cockles, like the bivalve Macoma balthica L., are particularly vulnerable to being stabbed on the advancing and receding tides when their shells are agape (Hulscher 1996), feeding on upshore flats may be a more profitable strategy for stabbers of low competitive ability.
Feeding in fields almost certainly exposes oystercatchers to greater risks of accidental death due to collisions with cars, trains or overhead wires en route to and from fields than if they were to restrict their foraging to inter-tidal areas. Small shorebirds have been found to be most at risk from predation by raptors when they feed near to the high-water mark and thus close to cover from which raptors can launch surprise attacks (Whitfield, Evans & Whitfield 1988; Cresswell & Whitfield 1994). It is likely that feeding in hedge- and tree-lined fields may pose an even greater risk of predation than feeding on upshore flats. Oystercatchers are exposed to the infective stages of parasitic worms through their consumption of molluscs and earthworms (Goater et al. 1995; Quénec’hdu 1998). The trematode Psilostomum brevicolle (Creplin) is one of the two most common intestinal helminths of oystercatchers on the Exe estuary, and is acquired through feeding on mussels and cockles (Goater et al. 1995). Trematodes of the family Psilostomatidae are reputed to be of low pathogenicity (Euzeby 1966) and their high prevalence within the oystercatcher population of the Exe estuary lends support to this idea (Quénec’hdu 1998). In contrast, parasitic worms such as the nematodes Capillaria spp., which are less commonly found in oystercatchers, have been known to bring about heavy mortalities among captive fowl (Brugère-Picoux & Silim 1992). Capillaria spp. have lumbricid worms as their intermediate hosts, and oystercatchers that feed in the fields on earthworms are therefore at greater risk of infection by a highly pathogenic parasite than are those that feed solely on the inter-tidal zone (Quénec’hdu 1998). Altogether, therefore, it is likely that field-feeding represents the more risky of the two alternative supplementary feeding options available to oystercatchers of low foraging efficiency. Experience of these elevated risks acquired by older birds or disproportionate mortality of young individuals that feed in the fields, may also go some way to explaining the lower tendency of older birds to feed in the fields at high tide. Further work is required to test these ideas.
Animals generally compete for resources on an individual basis and the outcome of direct competitive encounters are generally clear-cut, with a winner and loser (Drews 1993). These characteristics of competitive interactions have made dominance an ideal subject for both empiricists and theoreticians. Numerous empirical studies have measured variation in dominance, and examined its correlates and its fitness consequences (Pusey & Packer 1997). Studies of competing individuals across a range of taxa have related the intake rates of foraging animals to their dominance (Baker et al. 1981; Forrester 1991), whereas others have described interference functions and estimated susceptibility to interference (Dolman 1995; Cresswell 1997, 1998). Only in the case of the oystercatcher has variation in susceptibility to interference between individuals been related to their dominance (Ens & Goss-Custard 1984; Goss-Custard et al. 1984; Goss-Custard & Durell 1988; Stillman et al. 1996). Theoretical studies of the significance of individual variation in competitive ability have examined the effect of changes in the magnitude of the interference constant, and of its variation between individuals on population level functional and numerical responses and mortality rates (Sutherland 1992; Sutherland & Parker 1992; Sutherland & Dolman 1994). Comparatively few studies have quantified individual variation in the other major component of competitive ability, i.e. foraging efficiency, and how this might influence mortality risk and, hence, population dynamics. A notable exception is the study of Illius et al. (1995) in which individual differences in the jaw structure of Soay sheep (Ovis aries L.) had very clear fitness consequences during a period of high population density in relation to resources. Individuals with wide incisor arcades survived, while those with narrow arcades did not. Since incisor breadth was a primary determinant of foraging efficiency under the conditions of the study (Illius et al. 1995), it would seem that in this system, individual variation in foraging efficiency is a major determinant of variation in mortality risk between individuals. Illius et al. (1995) suggested that intense selection on foraging efficiency is likely to be a widespread phenomenon. The findings of the current study suggest that this is likely to be true, even in a highly interference-prone system. These studies support the assertion of Goss-Custard & Sutherland (1997) that, if behavioural ecologists are to contribute all they can to predicting population responses to habitat loss and change, greater effort is needed to identify the determinants of foraging efficiency and its variation between individuals.
We are grateful to members of the Dawlish Warren Nature Reserve Management Committee, for granting permission to ring birds, and to members of the Devon Wader Ringing Group for assistance in doing so. We would also like to thank Mr Ralph Clarke, for advice on statistical analyses, and two anonymous referees for their helpful comments.