The natal dispersal of tree swallows in a continuous mainland environment

Authors


David W. Winkler, Department of Ecology and Evolutionary Biology, Corson Hall, Cornell University, Ithaca, NY 14853, USA. Tel: 607 254 4216; Fax: 607 255 8088; E-mail: dww4@cornell.edu

Summary

  • 1To minimize study-area artefacts in a study of natal dispersal distances in tree swallows, we intensively banded nestlings and adults in nest-boxes in upstate New York and recruited and trained 70 volunteer banders within a 400 km radius of Ithaca. We banded 26 567 nestlings in the years 1985 through 1998, and captured 4774 adults at the nest, 630 of which had been banded as nestlings and were recaptured the year after fledging.
  • 2To correct for spatial variation in capture intensity, we resampled the distribution of all boxes where an adult capture was made under uniform, exponential and Cauchy null hypotheses. Compared to the null distributions, the frequency of observed dispersal events was significantly higher at 0–10 km and lower for all larger dispersal distances. The Cauchy distribution came closest to approximating the observed dispersal distance distribution.
  • 3Dispersal distances were sensitive to the distribution of nesting sites, and all measurements of dispersal distance distributions must be seen as being habitat-specific. We could detect no effect of dispersal distance on the subsequent timing of breeding and no effect of the timing of fledging on dispersal distance. The sexual differences in raw mean dispersal distances (8·38 km for females, 2·44 km for males) are similar to those reported for other species.
  • 4While it is tempting to conclude that studies in smaller areas have not missed a great deal, even a study area of 10 km extent (substantially larger than most) would have missed 11% of the dispersing birds detected and 95% of the range of distances recorded. Despite this sizeable component that would have been missed with a smaller study area, the relatively low frequencies of long-distance dispersal overall reinforce the conclusion that tree swallows, and probably most other migratory passerines, generally disperse much less far from their natal sites than the distances of their annual migrations might lead one to expect.

Introduction

Dispersal, the movement from a natal or breeding site to a new breeding site, is probably the most important and least understood life history trait (Clobert et al. 2001). It is of core importance to the ecological understanding of landscapes, populations and organisms, and it connects ecology to evolution through life history theory, biogeography and population genetics. Population geneticists have long understood that dispersal can act both as a source of genetic variation for evolutionary change (e.g. Wright 1982; Bohonak 1999) as well as a limit to local adaptation (e.g. Dhondt et al. 1990; Hendry et al. 2001; Lenormand 2002). Population ecologists have also recently rediscovered the importance of dispersal in metapopulation and range dynamics (Andrewartha & Birch 1954, cf. Durrett & Levin 1994; Tilman & Kareiva 1997; King & With 2002): dispersal is the glue that binds together the components of a metapopulation, and it effects source–sink dynamics and the demographic interconnection that is essential to metapopulation dynamics. Without dispersal, the dramatic range extensions that we see today (Shaw 1995; Veit & Lewis 1996), and that we infer in the recent past (Mila et al. 2000), would not have been possible. Finally, demographers know dispersal as the principal confounding factor in estimating survival rates (e.g. Lebreton et al. 1992; Lindberg et al. 2001; Blums et al. 2002), as a bird that disappears from a marked open population can only be known with certainty to have dispersed or died if it is found again after leaving.

Because birds are so vagile, avian dispersal has been studied most effectively to date where potential dispersal destinations are constrained to discrete ‘islands’: in the ocean (Pärt 1994, 1995, 1996; Spear et al. 1998; Wheelwright & Mauck 1998; Young 1998), in seas of unfavourable habitat (e.g. Matthysen et al. 1995; Stith et al. 1996) or where social behaviour dictates either a very clumped distribution of breeders (e.g. colonial breeders; Brown & Brown 1992; Pradel 1996; Negro et al. 1997; Hafner et al. 1998; Lindberg et al. 1998; Schjørring 2001) or limited dispersal options (e.g. cooperative breeders; Walters et al. 1988; Zack 1990; Young 1998; Koenig et al. 2000). For all the tractability that these choices of systems provide, the majority of avian species occupy more continuous habitat, where the distribution of dispersal distances is expected to be continuous; or the movements of dispersers, while impacted by interactions with competitors, are not constrained by habitat availability per se. By ‘continuous habitat’ we do not imply that there are not patches of unsuitable habitat, simply that these patches are not arrayed so as to bound dispersal distributions.

There have been few studies of passerines in continuous habitat (Plissner & Gowaty 1996; Verhulst et al. 1997), and especially of obligate migrants (Payne 1990, 1991; Shutler & Clark 2003). The biggest problem in studying dispersal empirically in these habitats is that distributions of dispersal distances are confounded by the unequal probabilities of detecting dispersal movements of differing length, i.e. that the dispersal distances actually observed are dictated largely by the dispersal distances that could be observed (e.g. Porter & Dooley 1993; van Noordwijk 1995; Koenig et al. 1996). There is also the fundamental problem of distinguishing between mortality and dispersal to a breeding site outside the study area. These and other biases inherent in most estimates of dispersal are now widely acknowledged, and there has been considerable recent interest in developing computational methods that quantify dispersal distances and survival more accurately (e.g. Barrowclough 1978; Manly & Chatterjee 1993; Baker et al. 1995; Pradel 1996; Thomson et al. 2003). These methods still involve extremely simplifying assumptions that may not apply in most systems. Regardless, no matter how sophisticated our corrections for bias may be, we cannot measure dispersal accurately until almost all the potential dispersal distances are sampled (Baker et al. 1995; Koenig et al. 1996).

Despite the availability of these and other methods, in thinking about dispersal most ornithologists are still guided by generalizations that arise from early spatially constrained studies of dispersal (e.g. Murray 1967; Greenwood 1980; Greenwood & Harvey 1982; Clarke et al. 1997): that female birds disperse further than males, that dispersal frequency declines geometrically with distance from the natal site, etc. We examine here the validity of these generalizations in a large-scale study of dispersal in a continuous mainland environment. We explore the natal dispersal distance distributions (DDDs) of a widely distributed Neotropical migrant bird, the tree swallow (Tachycineta bicolor, Vieillot 1808). This paper focuses on natal dispersal, the movement from the natal site to the first breeding site. Such movements are of larger scale than ‘breeding dispersal’ movements among successive breeding sites, both in general (Greenwood & Harvey 1982) and in tree swallows (Winkler et al. 2004), and they encompass the largest component of the spatial ecology of these birds. We compare the observed DDDs to those that could have been observed in our study area on the basis of the distribution of recapture effort. These comparisons help to weigh the ‘true’ DDD free of constraining study area boundaries. We explore further the potential costs and benefits of dispersal decisions by investigating the relationship between dispersal distances and the density of breeding opportunities, sex of the disperser and the relative timing of the disperser's fledging and its first breeding attempt.

study system and methods

Like most other North American passerines, tree swallows are Neotropical migrants. They fly every year between breeding grounds throughout North America to wintering areas in the Gulf Coast of North America, the Caribbean and Central America (Robertson et al. 1992). Their dispersal distances are thus not constrained by any limitation of movement. Tree swallows are secondary cavity nesters that rely on woodpeckers (or humans) to create the tree holes (or nest-boxes) that they require for nesting.

Our studies of tree swallows around Ithaca were begun in ‘UNIT’ study areas with the erection of 105 nestboxes in 1985 at Cornell University's Experimental Ponds Unit 1. Boxes were established at Experimental Ponds Unit 2 (128 boxes) in 1989, and on Cornell farm land at the top of Mt Pleasant (Unit 4: 60 boxes) in 1991 and along Hanshaw Road. (Unit 5: 22 boxes) in 1993 (see map in Winkler et al. 2004). Boxes at each of these UNIT sites are 20 m from the nearest neighbouring box. In the late 1980s we began monitoring variable numbers of boxes erected by others on private property surrounding our intensive study areas on the UNITs. Searching the roads of Tompkins County and creating a database for the locations, conditions, occupants, owners and permissions to visit each of these boxes, by 1993 we built a network (dubbed TOCO) for exploring the dispersal of swallows all around Tompkins County.

We extended the reach of our recapture efforts further by recruiting participants to a dispersal study that was part of the Cornell Nest Box Network (CNBN). Through this network we recruited and trained a subset of CNBN participants in New York and surrounding states to band bluebirds and swallows. In addition to these subpermitees on our master banding permits, several independent banders were also recruited to participate in the study. Although CNBN transformed into the Birdhouse Network (birds.cornell.edu/birdhouse/) in the late 1990s, the Swallow/Bluebird Dispersal Study (SDS) continued to function from Winkler's laboratory at Cornell through the 2003 breeding season. CNBN began reaching cooperators outside Tompkins County in 1994, with eight banders trained, and increased through the next 2 years to between 66 and 73 active banders state-wide from 1997 to 1999. The inclusion of banders throughout New York and surrounding states allowed us to conceive of our dispersal study area as a circle of 400 km radius around our original study site at Unit 1 (Fig. 1).

Figure 1.

The extended area for this study. The circle of 400-km radius around Ithaca encompasses all the CNBN participants who were trained as banders (filled dots) as well as some independent banders recruited to CNBN (large filled stars).

The central preoccupation of studying dispersal is gathering a collection of line segments, each of which represents the connection between a bird's natal site from which it fledged and its first breeding site. To ensure the accuracy of these line segments, we took great pains to assure that both capture locations, for chicks when they were banded in the nest and for adults when they were recaptured as breeders elsewhere, were recorded as accurately as possible. Box locations were mapped to an accuracy of less than 100 m using USGS 7·5 minute topographic quads. The other critical information available from our database is the distribution of nestboxes in which an adult swallow was captured. This distribution offers an integrated estimate of the ‘eyes’ of our project, as it incorporates not only the distribution of boxes, but also the distribution of adult-capturing effort.

It is the nature of birds that nest in continuous habitat that a check of all possible nesting sites is impossible. Thus, raw recapture rates cannot be taken as estimates of survival rates, and our study aimed not to capture every bird dispersing but rather to sample the dispersal distances across as wide a range of distances as possible. We conducted randomization tests with s-plus (2002) to evaluate the deviations of the observed DDD from those expected under various null distributions. Taking the natal box of each dispersal event as a starting-point, we calculated the distance from the natal box to every other box (henceforth ‘capture-boxes’) in the study area at which we captured an adult the following year. (The same qualitative results were obtained if we evaluated capture-boxes in the year of fledging.) Then, to judge the extent to which the observed DDD was dictated by the distribution of dispersal events that could have been observed, we conducted randomization tests on the distribution of all capture-boxes to see whether the observed DDD represented a significantly different distribution.

The first randomization test was based on a uniform null distribution, with an equal probability of a fledgling settling to breed in any capture-box. One draw was taken from the distribution of capture-boxes for each of the natal nests that was the origin of a dispersal line segment. This process was repeated 1000 times to produce an estimate of the median and range of the expected DDD for all dispersal events. The uniform null model assumed that returning birds were equally aware of all the nesting opportunities in our entire 400 km-radius study circle. One alternative to this null model is that the birds search for available nesting sites starting at their natal site and working outwards from there until they find an unoccupied site. Random-walk local searches produce a geometric decline in frequency with distance (e.g. Murray 1967; Waser 1985), and we created a similar exponential null distribution by regressing the overall observed log probabilities of capture on distance and using the slope and intercept of this regression to parameterize the null distribution. Note that in this paper, in the interest of comparability, we use a one-parameter exponential model, with a steeper drop-off in probability of settlement with distance than in the two-parameter exponential null used in Winkler et al. (2004). Finally, we used a very similar procedure to generate a half-Cauchy distribution with its shape parameter derived from the observed data by non-linear regression. The Cauchy distribution is the distribution resulting from the ratio of two independent normal distributions, and it has the heavy tails that characterize what we know of other empirical DDDs (Sutherland et al. 2000; Paradis et al. 2002).

In analyses relating dispersal distance to breeding phenology, we standardized for annual variations in laydates by subtracting the mean laydate for all nests (not just those that produced dispersal recaptures) in each year from each laydate. We then added either this standardized natal lay date for each disperser or a three-state (early, mid, late) laydate code for each to the mixed-model analyses. With the same methods, we also standardized the laydates of dispersing females in their first breeding year.

We tested for nest density effects in the sample of known first breeders from natal years 1993 onwards, as it is only in these latter years that large numbers of birds were being recaptured from all three networks (Table 1). Within this sample, we divided the area around each natal site into a series of concentric bands of increasing radius. We then related the observed dispersal distances to the numbers of capture-boxes in these bands, tallied for the year of breeding. In a mixed-model analysis with year as a random effect we also included the effect of sex and destination network, along with interactions of all these with each other and the distance rings, as fixed effects.

Table 1.  Summary of the numbers of all dispersers (both known first-year breeders and not) captured in each network for each year
Natal YearCNBNTOCOUNITsTotal
1985  0  0  1  1
1986  0  0  9  9
1987  0  0 15 15
1988  1  0  3  4
1989  0  0 16 16
1990  1  1 26 28
1991  2  1 46 49
1992  0  0 26 26
1993  7  0 64 71
1994  9 10 90109
1995 20 27 99146
1996 33 31 61125
1997 30 33 74137
1998 22  9 24 55
Total125112554791

Mixed-model analyses were conducted using the mixed procedure in SAS statistical software version 8·2 (Littell et al. 1996). Model selection proceeded from a fully parameterized model, with interaction terms eliminated, weakest first, that had P > 0·25. Natal year was included as a random effect in all mixed models. To moderate the effect of rare very long-distance dispersal events, we conducted the analyses with loge-transformed distance. In interpreting the fixed effect coefficients, one cannot merely take the antilog of the coefficient to estimate the mean effect of a change in the predictor on the distance dispersed. It is more direct to think about the median, because the log(median distance) = median of log(distance), which is not true of the mean. To focus, for example, on only the effect of sex on dispersal distance, the median of log(distance) = intercept + beta × sex and, taking antilogs: median distance = exp(intercept + beta × sex) = exp(intercept)exp(beta × sex). Because we coded sex as 1 for females and 0 for males, the expressions for the sex-effects are exp(intercept)exp(beta) and exp(intercept), respectively. The ratio of female to male distance is thus: median(females)/median(males) = exp(beta).

Results

known dispersal events

From records of the 26 567 banded tree swallows that fledged from 1985 to 1998, we detected a total of 791 events involving a nestling tree swallow banded and recaptured as a breeder. Of these, 630 (80%) were captured 1 year after fledging and were thus known cases of natal dispersal. The remaining 161 cases were first recaptured 2 (80%) or more years after fledging. These were excluded from the analyses because they are a heterogeneous sample of birds that could include unknown proportions of individuals that (1) we failed to trap in their first season of breeding, (2) delayed breeding for 1 or more years or (3) bred elsewhere and dispersed subsequently to the capture locality. The sample of 630 certain natal dispersal events included more females (n = 355) than males (n = 257, 18 were not sexed) because females are much easier to capture at the nest than are males. About half of all captures in this study (Table 2) came from CNBN and half came from the detailed studies in Ithaca (UNITs and TOCO combined). However, because of the limited spatial scale of swallow natal dispersal (below), more than 73% of the natal dispersal events detected came from the detailed studies in Ithaca (Table 3). The map of known dispersal events (Fig. 2) shows this preponderance of records from central New York; however, long-distance dispersal events across the state were observed, and shorter-distance dispersal within other regions of the state was also detected.

Table 2.  Summary of captures of adult and nestling tree swallows, by network, between 1985 and 1998. Each cell contains the raw number together with (in parentheses) the percentage of the column and row (in italics) totals comprised by the cell. For further details on the UNITs, TOCO and CNBN capture networks, see text
 UNITsTOCOCNBNTotals
Adults 1 233 (10, 26) 625 (20, 13) 2 916 (18, 61) 4 774 (15)
Nestlings10 752 (90, 40)2569 (80, 10)13 246 (82, 50)26 567 (85)
Totals11 985 (38)3194 (10)16 162 (52)31 341
Table 3.  Summary of natal and first-breeding sites for each of the 630 known natal dispersal events. Each cell contains the raw number together with (in parentheses) the percentage of the column and row (in italics) totals comprised by the cell. For further information on the UNITs, TOCO and CNBN capture networks, see text
Fledged fromFirst found breeding in
CNBNTOCOUNITsTotals
CNBN50 (57, 52)23 (28, 24) 23 (5, 24) 96 (16)
TOCO11 (13, 27)17 (21, 42) 12 (3, 30) 40 (6)
UNITs27 (31, 5)42 (51, 8)425 (92, 86)494 (78)
Totals88 (14)82 (13)460 (73)630
Figure 2.

The observed line segments for all 630 natal dispersal events observed between natal years 1985 and 1998 in this study. Note the preponderance of records in the Ithaca region south of Cayuga Lake, a consequence of more intense banding and nest-boxing activity in that area.

estimating and evaluating the ddd

Converting this map of dispersal events into a DDD (Fig. 3) reveals the steep drop-off in the frequency of dispersal with distance from the natal site that other biologists have observed (Greenwood 1980; Greenwood & Harvey 1982; Payne 1991), but with a much ‘heavier’ tail than one might expect (see also Sutherland et al. 2000; Paradis et al. 2002).

Figure 3.

The relation between the observed (open circles) dispersal distance distribution and that expected under (a) uniform, (b) exponential and (c) truncated Cauchy null hypotheses. For each null distribution, the median frequency of dispersers in a given distance band from one thousand draws from the detectable distance distribution is indicated by a filled square, and the range of all 1000 draws is indicated by a vertical bar.

The disparity between observed and uniform null DDDs (Fig. 3a) indicates that tree swallows dispersed more often than expected in the distance band less than 10 km from the natal site. At all larger distances, the observed frequencies of dispersing birds were substantially smaller than those expected under the uniform null. Even the single-parameter exponential null (Fig. 3b) did not produce expected numbers of birds staying closer than 10 km from the natal site, and it produced higher numbers than expected at distances between 10 and 40 km from the natal site. At distances greater than 60 km, the drop-off in settlement probability was sufficiently strong that no recaptures were predicted. Between these two extreme alternatives, the half-Cauchy null model (Fig. 3c) produced intermediate results: it performed almost as well as the exponential and better than the uniform at short distances, and much better than either in the longer distances, although the predicted dispersal frequencies at longer distances are still generally over-estimated. In sum, the swallows dispersed to sites closer than 10 km from their natal site more often than expected under any null and generally less often at all longer distances.

Thus, the enlarged area covered in this study has allowed us to measure the DDD in a study area that is large enough that its shape is not merely an artefact of the distribution of recaptures that we could have observed. However, there is another factor that can potentially cause variability in estimates of the DDD: available nest density. If fledglings leave the nest with an innate search algorithm, then their patterns of movement are likely to be affected by the availability of nesting habitat. On the contrary, if the DDD is guided by a distance template that is under strong genetic influences, it might be expected that the observed DDD should not be affected strongly by variations in local nest density.

effects of nest density

The random effect of year on log dispersal distance was not significant (P = 0·20), and the solutions for fixed effects after model simplification (Table 4) revealed strong effects of sex, destination network and numbers of capture-boxes in the distance rings on dispersal distance. Given that, for the sex effect, beta = 0·75 (Table 4, also see Methods), the median dispersal distance for females is 2·12 times that of the males, a result that reinforces the results of simple univariate analyses (Fig. 4).

Table 4.  Solutions and tests for fixed effects in a mixed model analysis of potential predictors of the log of dispersal distance (ln meters) for 486 cases of known-sex natal dispersal (breeding year after 1993). The coefficient estimates for the categorical variable sex are for females relative to a value of zero for males. For destination, UNITs is assumed to have a coefficient of 0, with separate values for the other destination networks. Denominator degrees of freedom are calculated with the Kenward Roger method in SAS Proc Mixed, and the coefficients and their standard errors and t-tests come from the Solutions for Fixed Effects. All the remaining statistics for all effects other than the intercept come from the Type 3 Tests of Fixed Effects. AIC = 1881 and P for random natal year effect = 0·20. For interpretation of coefficients, see text
EffectGroupCoefficientSEd.f.tP > t
Intercept 10·160·734643·85  0·0003
SexF0·750·2346414·27  0·0002
Boxes 0–2 km −0·0250·002846437·77< 0·0001
Boxes 2–3 km −0·00470·002646010·10  0·0016
Boxes 3–10 km −0·00510·00171071·04  0·31
Boxes 10–20 km −0·000820·0016 68·50·02  0·89
Boxes 20–400 km 0·000480·00048  9·70·98  0·35
Destination   45414·56< 0·0001
CNBN−4·310·92454−4·66< 0·0001
TOCO−0·890·99465−0·89  0·37
Destination × sex   4632·85  0·059
CNBN1·160·494652·39  0·017
TOCO0·0800·454620·17  0·86
Destination × 0–2 km   4648·84  0·0002
CNBN−0·00210·0078466−0·27  0·79
TOCO0·0220·00554624·07< 0·0001
Destination × 2–3 km   46412·18< 0·0001
CNBN0·0370·00754664·93< 0·0001
TOCO0·00370·00454630·83  0·41
Destination × 3–10 km   46414·26< 0·0001
CNBN0·0110·00234664·75< 0·0001
TOCO0·00140·00234630·59  0·56
Destination × 10–20 km   4633·47  0·032
CNBN0·00520·00204552·57  0·011
TOCO0·00150·00204660·73  0·46
Sex × 10–20 kmF−0·00260·00144603·30  0·070
Figure 4.

Uncorrected dispersal distance distributions for male and female tree swallows, summed in 10 km bands from the natal nest. The mean distance for the 355 females (8·38 km, SD = 23·08, light grey bars) is larger than the mean for the 257 males (2·44 km, SD = 3·85, dark grey bars; unequal variances t381 = 4·76, P < 0·0001). Note the much larger dispersion of the DDD for females, with eight captures in bands from 50 to 210 km.

All but one of the coefficients relating dispersal distance to capture-box availability are negative (Table 4), indicating that birds disperse for shorter distances when the availability of boxes is high. The strength of this effect generally declines as box availability further from the natal site is considered; and the coefficients for these effects are generally quite small and only significant up to 3 km from the natal box.

The negative coefficients and P-values for the individual destination effects indicate that CNBN recaptures alone are driving the significant overall main effect (Table 4), and the surprisingly large negative coefficient (−4·31) indicates that the median dispersal distance of birds recaptured in CNBN is only 1·3% that of birds recaptured in UNIT nests. This much smaller median distance in CNBN probably arises from the fact that, despite several very long-distance recaptures in CNBN nests (Fig. 2), the vast majority of CNBN recaptures come from nests close by, very often captures by the same CNBN collaborator. By contrast, the intensive capturing effort represented by the UNITs brings in recaptures from all sources, and the majority of captures there are birds originating elsewhere within the UNIT network at small to moderate distances from the natal site.

Although the interactions between destinations and the availability bands are significant and should be retained in the model, they all have small coefficients (Table 4) indicating that the effects of destination on the relationship between dispersal distance and box availability are nowhere more than about 4% more than the availability effect for UNIT nests.

effects between dispersal and phenology

To test for these effects we simply added natal or first laydate to the model from Table 4, and neither continuous nor categorical effects for natal laydate had a significant effect on the log of dispersal distance (P = 0·30 and 0·19, respectively). Thus, there is no evidence in this data set that later-reared chicks disperse different distances than their earlier cohort members. The analysis of the effects of dispersal distance on standardized laydate in the first breeding season is restricted to a sample of 294 females from the UNITs for which we had reliable estimates of laydate. This smaller data set revealed even less suggestion of an effect (P = 0·88 and 0·95, respectively) of dispersal distance on laydate in the first breeding season. Thus, we can find no evidence that birds that disperse further from their natal site incur a later first laying date as a result.

Discussion

Although the general form of the DDD is reminiscent of the pattern seen in DDDs from more limited study areas, this is the first study to extend a local study of passerine dispersal to two orders of magnitude beyond the limit of 23 median dispersal distance units that characterized earlier studies (Sutherland et al. 2000). The median dispersal distance in our study was 2·282 km and, given that our study could detect dispersal distances of over 300 times this distance, it seems fair to say that this is the first site-based study to explore the details of the long-distance dispersal of a passerine migrant across continuous habitat. The results of our randomization tests indicate that the observed DDD is not merely a consequence of the distribution of recaptures that could have been made. Tree swallows return to within 10 km of their natal site much more often, and at distances between 10 and 40 km much less often, than expected under any of the null DDDs. At distances of 40 km and longer, the uniform null still predicted more recaptures than observed, and the observed distribution was generally not as heavy-tailed as the Cauchy predicted.

There have been several attempts to estimate the true DDD free of the artefacts imposed by study areas of limited extent (e.g. Baker et al. 1995; Thomson et al. 2003). The approach we have taken in comparing the observed to null DDDs is most similar to that proposed by Thomson et al. (2003). However, we rely on the actual distributions of nests where adults were captured to generate our null distributions, and we analyse the fates of birds from multiple scattered release locations as opposed to one or a few central ones. While we are working actively on methods to use more effectively all available information in a maximum-likelihood estimation of the ‘true’ DDD (Hiebeler et al. in preparation), we have restricted ourselves here to simply testing three alternative hypothetical forms of bird dispersal pattern and beginning to explore the biological causes and consequences of the patterns observed.

The three null DDDs tested reflect different models of nest discovery. If, as a result of their relatively large-scale movements, swallows always know of a great number of potential sites, then something approaching the uniform DDD would be the most appropriate null model for movement. If, on the other hand, swallows require a considerable investment in time to locate potential nest sites, then the exponential null distance distribution is most appropriate. Finally if, as Winkler (2005) suggested, swallows are balancing largely local information with that obtained by more distant flights, then the half-Cauchy distribution may be more appropriate. The fact (Table 4) that these birds are responsive to variation in the availability of nesting opportunities within 3 km of the natal site indicates that they are not merely dispersing according to some set genetic distance template and that they are actually gathering information before selecting a nesting site. It is a little surprising that the significant effects of nest density are localized to only 3 km (Table 4), for birds recently fledged could obtain information on availability of nesting sites at distances up to about 20 km during foraging and exploratory flights immediately post-fledging and at even greater distances once they begin visiting large nocturnal roost sites which, in the case of Ithaca, would probably take the birds 60 km NW (Burney 2002). The fact that the Cauchy distribution seemed to perform best in approximating the DDD of these birds suggests that this longer-distance information is indeed being brought to bear on the dispersal distribution for at least some of the dispersing birds. It must be borne in mind, however, that none of the null hypotheses tested fitted the observed data very well and all were spatially static: they were fitted from the raw data without corrections for spatial heterogeneity in nest-site availability, and they did not take into account the fact that local competition for sites might shift substantially the distributions under the null. A more satisfying ESS model at the behavioural level would require the inclusion of competition over nest-boxes, the timing of settlement, etc. as drivers of the pattern, and patterns modified by these behavioural details may look very different from those generated by these simple null models.

measures of nest-site availability need to be improved

Counting nest-boxes in which adults were actually captured yields a good estimate of recapture effort, and it thus serves well as the basis for randomization tests comparing the observed to expected DDDs. As an indicator of the true availability of nest-sites and a clue to the mechanisms behind dispersal movements, however, it is much less satisfying. We have tried unsuccessfully to obtain reliable independent estimates of nest-site availability. Many schemes of categorizing boxes as being available or not for swallow nesting, based on habitat, hole-size, etc. have had to be abandoned, as gradations are slight and our understanding of swallow nest-site selection is sufficiently poor to make convincing categorization impossible. There is also the complication that the availability of natural nesting sites and their use varies considerably across our extended study area. It appears likely that tree swallows will often prefer boxes to natural sites (Robertson & Rendell 1990), but it would be very difficult to obtain accurate estimates of the proportion of birds using natural sites in Tompkins County and next to impossible to do so for larger areas. Despite all the problems with the lack of an independent measure of nest-site availability it is our firm belief, from experience in various parts of our extended study area, that nest-site availability is correlated with capture effort. Although much further work refining estimates of nest-site availability would be helpful, it is perhaps surprising that the relationship between variation in the numbers of capture nests and dispersal distances is as strong as it is (Table 4).

ddds are flexible and responsive

As important as it is to know the DDD free of the artefacts imposed by the geography of study areas and recapture effort, it is valuable to bear in mind that the DDD is very much a function of the environment in which it is measured. The effects of the density of capture boxes on dispersal distance (Table 4) suggest a situation analogous to the estimation of heritability: no measure of a DDD can be taken as independent of the environment in which it is measured. However, just as geneticists can seek better understanding of the genetic variance component that is the numerator of the heritability ratio, behavioural ecologists must seek an ever-better understanding of the behavioural processes that drive dispersal.

One way to probe these processes is to try to uncover the costs and benefits that might accrue to variation in dispersal distance. We could detect no temporal cost or benefit associated with reproductive timing: birds that dispersed further did not lay later in their first year of breeding, and birds that fledged later did not have to disperse further to find a breeding place. These results should not be taken as definitive until larger numbers of dispersing birds can be studied in detail, but a complementary study (Shutler & Clark 2003) could find no effects of previous breeding experience on tree swallow dispersal distances.

It is clear that the dispersal distances of tree swallows are not largely at variance to those of other migratory passerines, although they may disperse slightly further than others (cf. uncorrected estimates in Payne 1991; Plissner & Gowaty 1996; Sutherland et al. 2000). Regardless, given the large-scale movements of these birds as they migrate to and from wintering areas around the Gulf of Mexico and Caribbean and their considerable movements to foraging areas during, and roosting areas after, the breeding season, one of the largest unanswered questions is why these birds do not travel further when they disperse (Weatherhead & Forbes 1994). It is difficult to scale the costs of movement but, because swallows are already flying most of every day to forage, their incremental costs of dispersal must be smaller than for the majority of passerine dispersers, who must segregate large movements from foraging. Thus, dispersing larger distances clearly poses a relatively small cost, and the reasons for these relatively restricted movements must be sought in other causes (cf. Pärt 1990; Hansson et al. 2002).

Patterns of movement must entail an integration of different sources of information (Schjørring 2002) and a balancing of risks. Our understanding of what those risks may be and their relative importance is still rudimentary, but the primary risks associated with staying close to the natal site would appear to be those associated with mating or competing with kin. Increasing dispersal distance presumably comes at the cost of greater risk of mortality from the increased vulnerability to predation and increased uncertainty of adequate foraging areas that comes with movement to an unfamiliar environment. Birds may have been selected not to leave behind local adaptations (to parasites, mate compatibility, etc.) or knowledge of familiar predators, competitors and feeding and breeding sites (Bélichon et al. 1996; Winkler 2005). No matter how exactly they value and balance these risks, the limited spatial scale of these birds’ movements indicates that the risks involved in staying relatively close to the natal site are low compared to the risks of dispersing further away.

what of longer distances?

Consideration of the scale of this study inevitably leads to questioning how much of the long-distance tail of the DDD we left unsampled. The relatively little that is known about the geographical structuring of genetic variation in tree swallows (Stenzler 2001) suggests that some very long-distance dispersal has occurred some time in the recent evolutionary past, and recent analyses (Hosner & Winkler, in preparation) of the continent-wide database from the US Bird Banding Laboratory indicates that dispersal events longer than 400 km do occur rarely. Just how important these rare longer-distance dispersal events will be depends a great deal on the question being asked. Population ecologists interested in the dynamics of subpopulations and their interactions over a large landscape are likely to be more interested in the mean and modal distances that animals disperse, whereas a population or evolutionary geneticist or an epidemiologist is more likely to focus on the movements of those rare individuals that settle at distances well beyond the rest of their cohort. All of these types of movement are summarized in the DDD and, although DDDs in different systems will always be better estimated in some distances than in others, different workers will always be emphasizing different moments of the DDD in their work.

Conclusions

Although tree swallows may disperse further on average than other passerines, most of which are more limited in the range of their information gathering by breeding territoriality, it is clear that the general shapes of dispersal distance distributions reported for other passerines, although truncated to short distances, are not due solely to the artefacts of limited study areas. Most migratory songbirds, after making long migrations to distant wintering areas, may return to breed within relatively short distances of where they fledged. It appears likely that further progress in the study of natal dispersal will arise from the refinement of technologies that increase the probabilities of rediscovery at breeding sites of individuals that were measured and marked as nestlings.

Acknowledgements

Many thanks to John Fitzpatrick, André Dhondt and Rick Bonney, co-Principal Investigators on the CNBN grant (NSF ESI-9627280), which was the start of that distinctive source of recaptures. CNBN volunteers included J. Abrams, T. Anstey, C. Anstey, C. Banta, B. Best, R. Biss, K. and E. Boehm, J. Bowe, E. Brooks, J. Buckland, D. Buerk, L. Buttel, L. Carpenter, C. Cassady, B. Cicognani, D. Clark, C. Cliffel, P. Coates, E. Colling, L. Connor, S. Cooper, N. Davis-Ricci, C. Delong, J. Deobil, A. Dhondt, G. and W. Dingman, K. Engel, D. Fancher, A. Finney, W. Fisher, M. Forness, K. Fox, C. and D. Fryer, K. Geiger, B. Giles, G. Graziano, A. Greenwood, T. Greg, F. Gricius, D. Griffin, M. Gunning, B. Habalou, D. and I. Hall, S. Hartwig, S. Harwood, D. Hauber, D. Hofer, B. Hohmann, J. Holman, R. Jensen, T. Kast, E. Kellogg, D. Kelly, B. and S. Kozlowski, K. Luther, S. McCutcheon, D. McDermitt, K. McMahon, D. McNaughton, D. Edsinger, D. Moore, G. Nesslage, N. Newcomb, Ontario Bird Banding Association, E. and J. Ormondroyd, L. Owens, S. Podulka, C. Redinger, D. Robie, J. Rogers, B. Ross, N. Rowe, N. Scalzo, G. Schmieder, S. Schwemmer, J. Sedlacek, P. Senesac, W. Shuart, L. Siemion, Jr, D. Slavin, D., P., K., L. and J. Smith, L. Sommers, D. Sosnowski, L. Stenzler, A. Terninko, B. Toner, B. Treiber, J. Van Niel, K. Waters, R. Wells and family, L. Wheeler, D. Whittaker, J. Zarudsky, C. Zenger and M. Zettel. TOCO field assistants included R. Bakelaar, S. Bonilla, T. Bruce, B. Carter, B. Christman, C. Clews, E. Goetze, N. Hamm, J. Henkel, P. Hosner, A. Krakauer, M. Medler, S. Molnich, A. Romero, B. Taft and L. Yang. UNIT crew especially involved in dispersal captures included F. Adler, D. Ardia, S. Barker, T. Bruce, M. Bowlin, B. Carter, M. Clark, C. Burney, B. Christman, K. Dorsey, V. Ferretti, L. Gallant, E. Goetze, S. Kelly, P. Kleinman, E. Klopfer, A. Krakauer, P. Llambías, S. Malone, J. McCarty, M. Medler, L. Merrill, S. Mitra, S. Molnich, R. Moore, R. Morantz, D. O’Neill, M. Quiroga, K. Roux, C. Schiefflin, E. Siemann, B. Stuart, B. Taft, E. VanderWerf, L. Yang and J. Zamon. Dorothy Buerk, Jim Holman and Susan McCutcheon have given many hours to keeping our dispersal study running on less than a shoestring. Kevin Omland and Anders Møller asked thought-provoking questions to earlier presentations of these results, and Charles McCulloch provided essential advice on analysing mixed models in sas. Some analyses in this paper were supported by a Cooperative Agreement (no. 829374) between Winkler and the National Center for Environmental Assessment, Environmental Protection Agency, Washington, DC and by an NSF LTREB grant to Winkler (IBN-013437).

Ancillary