Research Paper

INTRODUCTION

Selecting a safe nest site helps birds ensure reproductive success because predation is often the leading cause of nest failure (Martin 1993, Weidinger 2002, Laidlaw et al. 2020). Defense against predation has been cited as an important advantage of colonial breeding, a nesting strategy used by most seabird species (Coulson 2002). Reproductive success for seabirds in large, dense colonies can benefit from predator confusion (Stempniewicz 1995, Davies et al. 2012), synchronous breeding, i.e., dilution effects (Hatchwell 1991, Descamps 2019), and predator mobbing (Clode et al. 2000). Greater reproductive success can thus be afforded to nests occurring where intraspecific nest densities are relatively elevated (Hernández-Matìas et al. 2003, Liljesthröm et al. 2008).

Colonial breeding can nonetheless promote intraspecific competition for nesting habitat (Schreiber and Burger 2002), forcing some individuals to choose nest sites where reproductive success can be relatively low (Hildén 1965, Kokko et al. 2004). The role of competition for resources in shaping patterns of nest-site selection may become increasingly complex when closely related species nest in sympatry, i.e., mixed-species colonies, because one species can outcompete and displace the other(s) when habitat niches completely overlap (Gause 1934, Burger 1979). Accordingly, natural selection is expected to favor species-specific differences in nest-site selection in mixed-species colonies, allowing birds to partition habitat and minimize competition (MacArthur and Levins 1967, Schoener 1974, Buckley and Buckley 2000).

Closely related Arctic Terns (Sterna paradisaea), Common Terns (Sterna hirundo), and Roseate Terns (Sterna dougallii) are three seabird species that nest sympatrically in colonies on small rocky islands in the Northwest and Northeast Atlantic Ocean (Gochfeld and Burger 2020). The three species are characterized by having similar breeding chronologies, placing their nests of similar size on rocky substrate, and defending nesting territories against neighboring terns and avian predators (Arnold et al. 2020, Gochfeld and Burger 2020, Hatch et al. 2020). Mixed colonies of Arctic, Common, and Roseate terns thus provide suitable opportunities to test habitat partitioning theory among closely related species (Bridge et al. 2005). Previous work has shown that each of these three tern species will partition habitat when nesting alongside one other species. For example, Common Terns may select nest sites that are less concealed than those of nearby Roseate Terns, and Roseate Terns will establish sub-colonies within groups of Common Tern nests (Gochfeld and Burger 1987, Burger and Gochfeld 1988a, Ramos and del Nevo 1995). Arctic Terns breeding in colonies with either Common or Roseate terns will establish sub-colonies where cover is sparse compared to nest sites of the other two species (Hawksley 1957, Kress and Hall 2004, Hatch et al. 2020). Still, little is known about how Arctic, Common, and Roseate terns select nest sites where the three species nest sympatrically. The level of niche separation among sympatric species is expected to vary according to the number of interacting species (MacArthur 1958, Northfield et al. 2010, Robertson et al. 2014), thus quantifying differences in nest-site selection among these three tern species will enhance our understanding of the extent to which closely related species can partition fine-scale habitat.

We assessed nest-site selection and habitat partitioning in a colony of Arctic, Common, and Roseate terns on a coastal rocky island in Lobster Bay, southwest Nova Scotia, Canada. Here, the three tern species nested adjacent to a colony of Great Black-backed Gulls (Larus marinus), an important predator of tern eggs and young (Whittam and Leonard 1999, 2000, Donehower et al. 2007). Our study had three objectives. First, we compared habitat characteristics at nest sites, i.e., rocks of different size classes, cover, nearby densities of intra- and interspecific nests, to those at randomly selected sites to elucidate nest-site selection for each species. However, patterns of nest-site selection may not reveal evolutionary processes that shape habitat selection, i.e., correlates of nest success (Pulliam and Danielson 1991, Clark and Shutler 1999). Thus, our second objective was to compare habitat characteristics of successful and unsuccessful nests to assess how selection may shape nest-site selection behavior. Finally, we compared nest-site characteristics among Arctic, Common, and Roseate terns to determine the extent to which these species partition habitat within the colony. We predicted that (i) the three terns would partition aspects of nesting habitat, given that habitat niches among closely related species nesting in sympatry are not expected to completely overlap (MacArthur and Levins 1967); and (ii) nest success would be greatest for nests surrounded by high densities of tern nests, because the efficiency of antipredator tactics can increase with nest densities (Burger and Gochfeld 1988b).

The Roseate Tern is listed as endangered in Canada and in the United States (COSEWIC 2009, U.S. Fish and Wildlife Service 2010). Currently, the Canadian Roseate Tern population is restricted to three islands in Nova Scotia, though ~80% of these birds breed at our study site in Lobster Bay. The annual number of breeding pairs across the three sites (70 to 80 pairs) currently does not meet national recovery goals for the species, i.e., ≥ 150 pairs (Environment Canada 2010). Low productivity and availability of nesting habitat are thought to contribute to limited growth of the Canadian Roseate Tern population (COSEWIC 2009). A key habitat feature for breeding Roseate Terns in Canada is the presence of large numbers of nesting Common and Arctic terns (Rock et al. 2007, Environment Canada 2014). Accordingly, understanding how the three tern species partition nesting habitat will help guide efforts for identifying tern islands, e.g., those with large colonies of Common and Arctic terns, potentially suitable for Roseate Tern nesting, a priority conservation strategy for the species (Environment Canada 2014).

METHODS

Study site

During 2018, we studied a colony of Arctic, Common, and Roseate terns on Gull Island (43°39′40.03″N, 65°54′56.83″W; Figure 1), situated in the upper Gulf of Maine. Gull Island measures 4 ha, is low-lying (~1.5 m above sea level), and consists primarily of rocky substrate with areas of sparse vegetative cover in the form of shrubs and herbaceous plants. The tern colony consisted of approximately 350 nests and was in a ~2000-m² area in the southwest region of the island and adjacent to a small tidal pond. In total, 62 Great Black-backed Gull nests were found on the island, the majority of which were in the northeast portion of the island but within 50 to 200 m of the tern colony. No mammalian nest predators were detected on the island during the study.

Monitoring of this Arctic, Common, and Roseate tern colony in Lobster Bay began in 1982 and, up until 2018, the colony was known to nest annually on North Brother Island (43°38′10.99″N, 65°49′24.49″W; https://teddeon.com/ternrep.html), located 8 km southeast of Gull Island. North Brother Island is managed for tern breeding with measures to support recovery of the Canadian Roseate Tern population, including provisioning of artificial nest shelters and predator management (Environment Canada 2010, Pratte et al. 2021). In 2017, high rates of tern nest predation by American Crows (Corvus brachyrhynchos) and gulls were at least partially responsible for abandonment of the North Brother Island tern colony, and some of these terns immediately relocated to Gull Island where they resumed breeding (J. McKnight, personal observation). The tern colony returned to Gull Island in 2018, though it has been back on North Brother Island since 2019. Unlike most large Roseate Tern colonies in the Northeast and Northwest Atlantic Ocean, tern habitat enhancement, e.g., provision of nest shelters, and predator control, except for the removal of three Great Black-backed Gull nests, was not undertaken on Gull Island in 2018 (Seward et al. 2019, U.S. Fish and Wildlife Service 2020). The situation on Gull Island in 2018 thus provided a rare opportunity to assess habitat partitioning among these three tern species under conditions of minimal management.

Data collection

We visited Gull Island two to three times weekly during the 2018 breeding season that spanned mid-May to early August. Island visits did not occur during inclement weather, i.e., rain or heavy fog, and typically lasted no more than two to three hours. Tern nests were identified to species by observing incubating adults from a blind placed 5 to 25 m from focal nests. Peak nesting for all three tern species occurred during the second and third weeks of June. A total of 292 tern nests were identified to species: 81 Arctic, 176 Common, and 35 Roseate tern nests. We likely identified all Roseate Tern nests in the colony. We recorded nest-site coordinates with a high-precision GPS (error of 20 cm; EOS positioning systems, Terrebonne, Québec, Canada). A sub-sample of the 292 nests that included 48 Common, 43 Arctic, and the 35 Roseate tern nests were marked by placing a 15 cm long stick ~25 cm to the west of the nest bowl. Nest marking facilitated observations of incubating adults. Marked nests were followed until their fates were determined, and during each nest visit we noted clutch size and nest status as active or failed. Successful nests hatched ≥ 1 egg. We did not find cases in which terns abandoned their clutches, and only Great Black-backed Gulls were observed predating tern eggs. We therefore suspect that most, if not all, unsuccessful tern nesting attempts were attributed to avian predation.

At the conclusion of the breeding season, we measured habitat characteristics at all marked tern nests. We centered a 1-m diameter plastic ring on the nest bowl and obtained an overhead photograph of habitat within the ring. A 1 m long stick with 10-cm unit markers was placed on the ground and included in photos, with the stick serving as a size reference for rocks at nest sites. From photos, the number of rocks inside the ring was recorded for each of two categories representing a rock’s maximum length in cm parallel to the substrate: i) 20 to 29 cm; and ii) ≥ 30 cm. The two rock-length categories represented intermediate, i.e., 20 to 29 cm, and large, i.e., ≥ 30 cm, rocks in the tern colony. Rocks of these size classes were included in analyses because they were among the most distinct features of tern nesting substrate. Although we were unable to measure rock height from photos, we still had a general sense of the height of rocks at nest sites because lateral cover scores were a function of rock height. We recorded overhead cover at nest sites by placing a 10-cm diameter black disk with five 4.5-cm² white squares, positioned along two lines perpendicular to each other, in the middle of the nest bowl. The disk’s diameter represented the approximate size of a tern nest bowl (Arnold et al. 2020, Gochfeld and Burger 2020, Hatch et al. 2020). We then estimated the percentage of each of the five white squares that was not visible from 1 m above the ground and averaged the five estimates to determine an overhead cover score at the nest. We recorded lateral cover at nests by placing two artificial tern eggs in the center of the nest bowl and determining whether eggs were visible from each of the four cardinal directions. Observations were undertaken from 40 to 60 cm above the ground and laterally 30 cm from the nest. We chose a 40 to 60 cm height to simulate the height of an adult Great Black-backed Gull. Lateral cover was scored as 0 if any portion of either egg was visible or 1 if the eggs were not visible. The average lateral cover from the four directions served as a lateral cover score at the nest site.

For each tern nest at which habitat characteristics were measured, we selected a random site where we measured habitat characteristics as described above. A series of random numbers was used to choose directions and distances between nests and random sites. To ensure random sites were in available tern nesting habitat and within the colony boundaries, we limited the maximum distance between a nest site and its associated random site to 3.3 m, a value that represented the approximate maximum distance between nests along the periphery of the colony and other nests. We chose a new random site if the previous site was in unsuitable tern nesting habitat, such as an area flooded daily during high tides.

We determined tern nesting densities around marked nests and their random sites by calculating the number of interspecific and intraspecific nests within 1, 2, and 5 m using the Geosphere package (Hijmans et al. 2017). Geographic coordinates were not obtained for about 50 tern nests in the colony because of logistical constraints; however, the majority of these nests occurred in an area of the colony that did not include marked nests, so their exclusion from analyses was not expected to bias conclusions about the role of tern nesting densities on nest-site selection and nest success.

Data analyses

We tested for correlations between rock-size classes and lateral and overhead cover using Spearman rank correlations. Number of rocks ≥ 30 cm was correlated with lateral cover (ρ = 0.14, p = 0.03) and overhead cover (ρ = 0.24, p < 0.01). We included lateral cover in subsequent analyses because many tern nests were adjacent to large, flat rocks that offered minimal nest cover (N. Knutson, personal observation). Overhead cover was not observed at Arctic and Common tern nests or at their associated random sites, so we did not include overhead cover in analyses for these two species. In contrast, 14 of 35 (40%) Roseate Tern nests had at least some overhead cover, i.e., overhangs provided primarily by rocks ≥ 30 cm. For each of the Roseate Tern nest-site selection and nest-success datasets, we conducted binomial logistic regressions to compare the fit of a model with overhead cover as the only explanatory variable to that of a model with number of rocks ≥ 30 cm as the only explanatory variable. Both models included an intercept term. Response variables were site type, i.e., nest or random site, and nest fate, i.e., successful or unsuccessful nest. Model fit was assessed using the multi-model inference package (MuMIn; Bartoń 2015) and we compared models with values of Akaike’s Information Criterion corrected for small sample sizes (AICc; Akaike 1973). Lower AICc values indicated better model fit. For Roseate Tern nest-site selection, a model with overhead cover had a lower AICc value (81.21) than that with number of rocks ≥ 30 cm (90.20), whereas a model with number of rocks ≥ 30 cm had better fit (AICc = 46.82) than a model with overhead cover (AICc = 50.50) for discriminating between successful and unsuccessful nests. Accordingly, we included overhead cover, but not number of rocks ≥ 30 cm, in analyses for Roseate Tern nest-site selection, and rocks ≥ 30 cm, but not overhead cover, for Roseate Tern nest success analyses.

Binomial logistic regressions were used to assess differences in habitat between (i) nest and random sites and (ii) successful and unsuccessful nests, and a multinomial logistic regression model was used to evaluate differences in nest-site characteristics among the three tern species, i.e., habitat partitioning, with package nnet (Ripley and Venables 2025). Response variables were site type, i.e., nest or random site, for nest-site selection analyses; nest fate, i.e., successful or unsuccessful nest, for nest success; and tern species, i.e., Arctic, Common, or Roseate tern, for habitat partitioning. The logit link was used for all regression analyses. Partial predation was detected at some successful tern nests; however, we opted to compare habitat characteristics between successful and unsuccessful nests because complete nest predation exerts strong selection on nest-site selection behavior (Clark and Shutler 1999). The role of habitat characteristics in tern nest-site selection, nest success, and habitat partitioning was initially assessed using univariate logistic regressions. We included in final candidate model sets only those habitat characteristics that were deemed potentially informative for tern nesting behavior (p > 0.25; Hosmer et al. 2013; Table 1).

For each of the nest-site selection, nest success, and habitat partitioning analyses, we evaluated candidate models consisting of all single and additive effects of habitat characteristics deemed potentially informative using the multi-model inference package (MuMIn; Bartoń 2015). AICc was used to assess model fit and to identify the best approximating model(s), i.e., ΔAICc < 2 (Burnham and Anderson 2002). For each best approximating model, we calculated model weight (w), a measure of relative support for the model from the data. We model-averaged slopes of each habitat characteristic included in the best approximating model(s) and generated unconditional standard errors and 95% confidence intervals to accommodate model uncertainty (Burnham and Anderson 2002, Arnold 2010). Habitat characteristics with 95% confidence intervals that did not include 0 were considered informative for tern nesting behavior (Anderson et al. 2001).

We compared nest-success probabilities among Arctic, Common, and Roseate terns using a chi-squared test for equality of proportions. All statistical analyses were performed using R 4.5.1 (R Core Team 2025). For each of the three analyses, results of models with weights ≥ 0.01 are presented in Appendices 1 through 7.

RESULTS

Sub-colony placement

Each of the three tern species clustered their nests into sub-colonies (Figure 1). Arctic Terns formed two sub-colonies along the periphery of the colony near the high-water line, whereas Roseate and Common tern sub-colonies were at more interior regions of the colony. Lateral and overhead cover at tern nest sites was provided by rocks, most often rocks ≥ 30 cm (N. Knutson, personal observation), except for herbaceous plants that provided sparse lateral cover at some Common Tern nests.

Nest-site selection and nest success

For Common and Roseate terns, lateral cover was included in all models best approximating nest-site selection (ΔAICc < 2; Table 2), and summed weights for these models were 0.85 for Common Terns and 0.81 for Roseate Terns. Lateral cover was eight and ten times greater at nest sites compared to random sites for Common Terns (model-averaged β = 9.44; 95% CI: 0.75–18.13) and Roseate Terns (model-averaged β = 6.17; 95% CI: 1.27–11.07; Tables 3 and 4), respectively. Habitat characteristics measured at Arctic Tern nests did not vary from those at random sites (Table 5).

Nest success was similar among Roseate (63% of 35 nests), Common (58% of 48 nests), and Arctic terns (72% of 43 nests; χ² = 1.91; df = 2; p = 0.38). We found no support for an effect of lateral or overhead cover on tern-nest success (Tables 3, 4, 5). However, there was evidence for a positive association between the number of interspecific nests ≤ 2 m and nest success for Common Terns (model-averaged β = 1.34; 95% CI: 0.05–2.64). The number of interspecific nests ≤ 2 m was included in all four models best approximating nest success for Common Terns, and these models had a summed weight of 0.67 (Table 6). In contrast, there was no model support for an effect of nearby interspecific nests on nest success for Roseate Terns (model-averaged β = 0.16; 95% CI: -0.16–0.48) and Arctic Terns (model-averaged β = 0.17; 95% CI: -0.12–0.45).

Nest-site habitat partitioning

Lateral cover, number of rocks ≥ 30 cm, and number of inter- and intraspecific nests ≤ 5 m were included in the two best-approximating models of habitat partitioning (Table 7). On average, 30% of a Roseate Tern nest was covered laterally, which was considerably higher than lateral cover at nests of Common Terns (model-averaged β = 6.06; 95% CI: 0.66–11.46) and Arctic Terns (model-averaged β = 11.79; 95% CI: 4.85–18.74; Figure 2). Number of rocks ≥ 30 cm at Roseate Tern nests was also greater than that for Common Terns (model-averaged β = 0.95; 95% CI: 0.36–1.55), but not for Arctic Terns (model-averaged β = 0.30; 95% CI: -0.21–0.82; Figure 2). Although we did not measure rock height, the height of rocks ≥ 30 cm at Arctic Tern nest sites was generally lower than that at Roseate Tern nests, given that lateral cover was sparse at Arctic Tern nests and common at those of Roseate Terns (Tables 4 and 5).

Roseate Tern nests averaged 11 interspecific nests ≤ 5 m, which was greater than that for Common Terns (model-averaged β = 0.42; 95% CI: 0.22–0.62) and Arctic Terns (model-averaged β = 0.31; 95% CI: 0.13–0.49; Figure 2). The number of intraspecific nests ≤ 5 m from Common Tern nests (average 11 nests) was higher than that for Arctic Terns (model-averaged β = 0.31; 95% CI: 0.13–0.49), but not Roseate Terns (model-averaged β = -0.11; 95% CI: -0.44–0.01). These species-specific differences in the number of inter- and intraspecific nests ≤ 5 m reflected tendencies for (i) Roseate Terns to nest at high densities within Common Tern sub-colonies and (ii) Common Terns to nest at high densities throughout much of the colony (Figure 1).

DISCUSSION

When closely related species that share aspects of their habitat niches nest in sympatry, natural selection is expected to favor species-specific differences in nest-site selection that allow birds to partition habitat and minimize competition for nest sites (Schoener 1974, Martin et al. 2004, Samraoui et al. 2025). Our observations support habitat partitioning as an explanation for the existence of a dense colony of closely related Arctic, Common, and Roseate terns on Gull Island, Nova Scotia, because the three species demonstrated marked differences in nest-site habitat preferences. Terns clustered their nests into species-specific sub-colonies and, notably, Arctic Tern sub-colonies contrasted with those of Roseate and Common terns by occurring along the periphery of the colony where substrate consisted of relatively large, flat rocks providing little nest cover. Compared to Common and Roseate terns, Arctic Terns have short tarsi and long wings, and both traits may help make Arctic Terns well adapted to nesting in open, nearshore terrain where adults can fly directly to nest sites (Hawksley 1957). In the absence of nesting by other tern species, Arctic Terns may nonetheless place their nests adjacent to low cover (Kress and Hall 2004).

There were spaces for additional tern nests along the periphery of the colony, i.e., adjacent to Arctic Tern nests, yet Common and Roseate terns chose to nest at more interior sites and where Roseate Tern sub-colonies occurred within sub-colonies of Common Terns. Still, we found that the two species partitioned nest habitat according to the distribution of rocks ≥ 30 cm, and importantly, rocks of this size class were the primary source of lateral cover at tern nests, and they provided overhead cover to some Roseate Tern nests. At least one rock ≥ 30 cm was adjacent to most Roseate Tern nests, whereas the majority of Common Tern nests were void of rocks ≥ 30 cm, resulting in lateral cover at Roseate Tern nests being six times greater than that for Common Terns. Nest initiation among Common Terns generally began a few days prior to that for Roseate Terns (N. Knutson, personal observation), meaning that Common Terns had opportunities to select nest sites with relatively high levels of cover, but they avoided them. Thus, the patterns of nest-site selection observed among Common, Roseate, and Arctic terns are more consistent with habitat partitioning than competition for limited habitat. Nonetheless, availability of nesting habitat on Gull Island was not quantified, so we cannot fully dismiss competition as a factor potentially contributing to the limited niche overlap observed across species.

Across their range, Roseate Terns are more likely to nest closer to ledges, vegetation, logs, rocks, or other structures affording high levels of nest cover than sympatric Common Terns (Spendelow 1982, Ramos and del Nevo 1995, Donehower et al. 2007). Several hypotheses have been proposed to explain selection for nest cover by Roseate Terns, including the role of cover in minimizing effects of aggression from nearby terns (Burger and Gochfeld 1988b, Ramos 1998). Common Terns at our study site often exhibited aggression toward conspecifics and Roseate Terns nesting on adjacent territories. Rocks providing lateral and overhead cover, and particularly those rocks ≥ 30 cm, may help screen Roseate Tern eggs and chicks from neighboring terns. In the absence of nesting Common Terns, Roseate Terns will nest at sites with sparse cover, though it has been suggested that this tactic has evolved to reduce predation pressure, e.g., avoiding land crabs at sites in the Caribbean (Burger and Gochfeld 1988c), rather than reflect a niche shift linked to the absence of nesting Common Terns.

Large gull species have adverse effects on tern breeding colonies by predating adults, chicks, and eggs (Becker 1995, Whittam and Leonard 1999). Predation by Great Black-backed Gulls was the most important source of tern egg mortality during our study, and we occasionally observed gulls predating tern chicks. Nest success for all three tern species ranged from 58% for Common Terns to 72% in Arctic Terns. Rates of tern nest depredation by gulls on Eastern Egg Rock, Maine, were lower for Roseate Terns (6%) compared to Common (23%) and Arctic terns (32%; Donehower et al. 2007). The study’s authors attributed these species-specific differences in nest predation rates to nest-site selection because, unlike Common and Arctic terns that nested in relatively open areas, Roseate Terns nested at sites with cover sufficient for reducing gull predation risk (Donehower et al. 2007). On Gull Island, higher levels of nest cover did not translate into improved rates of tern nest success, suggesting that rocks adjacent to tern nests were generally inadequate in providing eggs, and possibly chicks, with a refuge from gull predation, and regardless of rock size.

Colonial nesting can enhance fitness of terns and other seabirds through factors that include improved information sharing, e.g., location of food sources (Goyert 2015), and reduced predation pressure (Anderson and Hodum 1993). Where nest predation risk is high, tern reproductive success can increase with colony size and nest densities (Hernández-Matìas et al. 2003, Byerly et al. 2021) because large groups of terns breeding synchronously may maximize effectiveness of antipredator tactics such as vigilance, mobbing, and prey dilution (Burger and Gochfeld 1988b, Burger et al. 1993). In tern sub-colonies on Gull Island, the relationship between nest success and intraspecific nest densities may have reached an upper plateau, because we failed to detect links between tern nest success and numbers of nearby intraspecific nests. In contrast, our results suggest that tern fitness can increase with greater densities of nests of other tern species, because we found support for a positive association between nest success and numbers of nearby interspecific nests for Common Terns. Model support for similar associations was not obtained for Arctic and Roseate terns, though we noted that successful nests generally had greater nearby densities of interspecific nests relative to unsuccessful nests for both species (Tables 4 and 5). Arctic, Common, and Roseate terns are aggressive defenders of their nests (Shealer and Burger 1992, Whittam and Leonard 2000). Accordingly, elevated nest predation risk, such as that observed on Gull Island, may favor selection for nest sites, at least for Common Terns, where locally high densities of intra- and interspecific nests facilitate efficient group predator defense or prey dilution (Burger and Gochfeld 1988a, Palestis 2005). In turn, such fitness advantages linked to antipredator strategies would be expected to help promote the evolution of nest-site partitioning in mixed colonies of terns.

Habitat partitioning observed among Arctic, Common, and Roseate terns reflected the response of a single colony, and thus these observations limit generalization of results. Still, similar to what unfolded on Gull Island during our study, historical observations from North Brother Island, i.e., primary location of the Lobster Bay mixed-tern colony, have showed that these three tern species select nest sites that vary spatially within the colony and along a range of concealment options provided by rocky substrate. Notably, Arctic Tern sub-colonies on North Brother Island are discriminated from those of Common and Roseate terns by occurring in open, more peripheral areas of the colony than the other two species, and Roseate Tern nests have greater lateral cover than nests of neighboring Common Terns. There thus may be a high level of repeatability in patterns of nest-site habitat partitioning within this mixed colony across islands in the Lobster Bay area. Accordingly, we encourage managers tasked with identifying islands potentially suitable for Roseate Tern nesting, i.e., islands with large colonies of Common and Arctic terns, to consider the potential for rocky substrate to provide cover for Roseate Tern nest sites, and particularly within sub-colonies of Common Terns.

LITERATURE CITED

Akaike, H. 1973. Information theory and an extension of the maximum likelihood principle. Pages 267-281 in B. N. Petrov and F. Csaki, editors. Second international symposium on information theory, Akadémia Kiadó, Budapest, Hungary.

Anderson, D. J., and P. J. Hodum. 1993. Predator behavior favors clumped nesting in an oceanic seabird. Ecology 74:2462-2464. https://doi.org/10.2307/1939597

Anderson, D. R., W. A. Link, D. H. Johnson, and K. P. Burnham. 2001. Suggestions for presenting the results of data analyses. Journal of Wildlife Management 65:373-378. https://doi.org/10.2307/3803088

Arnold, J. M., S. A. Oswald, I. C. T. Nisbet, P. Pyle, and M. A. Patten. 2020. Common Tern (Sterna hirundo), version 1.0. In S. M. Billerman, editor. Birds of the world. Cornell Lab of Ornithology, Ithaca, New York, USA. https://doi.org/10.2173/bow.comter.01

Arnold, T. W. 2010. Uninformative parameters and model selection using Akaike’s Information Criterion. Journal of Wildlife Management 74:1175-1178. https://doi.org/10.2193/2009-367

Bartoń, K. 2015. Package MuMIn: multi-model inference. R package version 1.48.11. R Foundation for Statistical Computing, Vienna, Austria. https://doi.org/10.32614/CRAN.package.MuMIn

Becker, P. H. 1995. Effects of coloniality on gull predation on Common Tern (Sterna hirundo) chicks. Waterbirds 18:11-22. https://doi.org/10.2307/1521394

Bridge, E., A. W. Jones, and A. J. Baker. 2005. A phylogenetic framework for the terns (Sternini) inferred from MtDNA sequences: implications for taxonomy and plumage evolution. Molecular Phylogenetics and Evolution 35:459-469. https://doi.org/10.1016/j.ympev.2004.12.010

Buckley, P. A., and F. G. Buckley. 2000. Patterns of colony-site use and disuse in saltmarsh-nesting Common and Roseate Terns. Journal of Field Ornithology 71:356-369. https://doi.org/10.1648/0273-8570-71.2.356

Burger, J. 1979. Resource partitioning: nest-site selection in mixed-species colonies of herons, egrets, and ibises. American Midland Naturalist 101:191-210. https://doi.org/10.2307/2424913

Burger, J., and M. Gochfeld. 1988a. Nest-site selection and temporal patterns in habitat use of Roseate and Common Terns. Auk 105:433-438. https://doi.org/10.1093/auk/105.3.433

Burger, J., and M. Gochfeld. 1988b. Defensive aggression in terns: effect of species, density, and isolation. Aggressive Behavior 14:169-178.

Burger, J., and M. Gochfeld. 1988c. Nest-site selection by Roseate Terns in two tropical colonies on Culebra, Puerto Rico. Condor 90:843-851. https://doi.org/10.2307/1368841

Burger, J., D. A. Shealer, and M. Gochfeld. 1993. Defensive aggression in terns: discrimination and response to individual researchers. Aggressive Behavior 19:303-311.

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Second edition. Springer-Verlag, New York, New York, USA. https://doi.org/10.1007/b97636

Byerly, P. A., S. Zaluski, D. Nellis, and P. L. Leberg. 2021. Colony characteristics influence nest survival of Caribbean Roseate Terns. Ornithological Applications 123:1-15. https://doi.org/10.1093/ornithapp/duaa069

Clark, R. G., and D. Shutler. 1999. Avian habitat selection: pattern from process in nest-site use by ducks? Ecology 80:272-287. https://doi.org/10.1890/0012-9658(1999)080[0272:AHSPFP]2.0.CO;2

Clode, D., J. D. S. Birks, and D. W. Macdonald. 2000. The influence of risk and vulnerability on predator mobbing by terns (Sterna spp.) and gulls (Larus spp.). Journal of Zoology 252:53-59. https://doi.org/10.1111/j.1469-7998.2000.tb00819.x

Committee on the Status of Endangered Wildlife in Canada (COSEWIC). 2009. COSEWIC assessment and update status report on the Roseate Tern (Sterna dougallii) in Canada. COSEWIC, Canadian Wildlife Service, Environment Canada, Ottawa, Ontario, Canada. https://www.canada.ca/en/environment-climate-change/services/species-risk-public-registry/cosewic-assessments-status-reports/roseate-tern-2009.html

Coulson, J. C. 2002. Colonial breeding in seabirds. Pages 87-114 in E. A. Schreiber and J. Burger, editors. Biology of marine birds. CRC Press, Boca Raton, Florida, USA.

Davies, N. B., J. Krebs, and S. A. West. 2012. An introduction to behavioral ecology. Fourth edition. Wiley-Blackwell, Hoboken, New Jersey, USA.

Descamps, S. 2019. Breeding synchrony and predator specialization: a test of the predator swamping hypothesis in seabirds. Ecology and Evolution 9:1431-1436. https://doi.org/10.1002/ece3.4863

Donehower, C. E., D. M. Bird, C. S. Hall, and S. W. Kress. 2007. Effects of gull predation and predator control on tern nesting success at Eastern Egg Rock, Maine. Waterbirds 30:29-39. https://doi.org/10.1675/1524-4695(2007)030[0029:EOGPAP]2.0.CO;2

Environment Canada. 2010. Amended recovery strategy for the Roseate Tern (Sterna dougallii) in Canada. Species at Risk Act Recovery Strategy Series, Environment Canada, Ottawa, Ontario, Canada. https://www.canada.ca/en/environment-climate-change/services/species-risk-public-registry/recovery-strategies/roseate-tern-amended-2010.html

Environment Canada. 2014. Action plan for the Roseate Tern (Sterna dougallii) in Canada. Species at Risk Act Action Plan Series, Environment Canada, Ottawa, Ontario, Canada. https://www.sararegistry.gc.ca/virtual_sara/files/plans/ap_roseate_tern_e_proposed.pdf

Gause, G. F. 1934. The struggle for existence. Williams and Williams, Baltimore, Maryland, USA. https://doi.org/10.1097/00010694-193602000-00018

Gochfeld, M., and J. Burger. 1987. Nest site selection: comparison of Roseate and Common Terns (Sterna dougallii and S. hirundo) in a Long Island, New York colony. Bird Behavior 7:58-66. https://doi.org/10.3727/015613887791918060

Gochfeld, M., and J. Burger. 2020. Roseate Tern (Sterna dougallii), version 1.0. In S. M. Billerman, editor. Birds of the world. Cornell Lab of Ornithology, Ithaca, New York, USA. https://doi.org/10.2173/bow.roster.01

Goyert, H. F. 2015. Foraging specificity and prey utilization: evaluating social and memory-based strategies in seabirds. Behavior 152:861-895. https://doi.org/10.1163/1568539X-00003260

Hatch, J. J., M. Gochfeld, J. Burger, and E. Garcia. 2020. Arctic Tern (Sterna paradisaea), version 1.0. In S. M. Billerman, editor. Birds of the world. Cornell Lab of Ornithology, Ithaca, New York, USA. https://doi.org/10.2173/bow.arcter.01

Hatchwell, B. J. 1991. An experimental study of the effects of timing of breeding on the reproductive success of common guillemots (Uria aalge). Journal of Animal Ecology 60:721-736. https://doi.org/10.2307/5410

Hawksley, O. 1957. Ecology of a breeding population of Arctic Terns. Bird-Banding 28:57-92. https://doi.org/10.2307/4510623

Hernández-Matìas, A., L. Jover, and X. Ruiz. 2003. Predation on Common Tern eggs in relation to sub-colony size, nest aggregation, and breeding synchrony. Waterbirds 26:280-289. https://doi.org/10.1675/1524-4695(2003)026[0280:POCTEI]2.0.CO;2

Hijmans, R. J., E. Williams, C. Vennes, and M. R. J. Hijmans. 2017. Package ‘geosphere.’ Spherical Trigonometry 1:1-45.

Hildén, O. 1965. Habitat selection in birds: a review. Annales Zoologici Fennici 2: 53-75. https://www.jstor.org/stable/23730835

Hosmer, D. W., S. Lemeshow, and R. X. Sturdivant. 2013. Applied logistic regression. Third edition. John Wiley & Sons, Hoboken, New Jersey, USA. https://doi.org/10.1002/9781118548387

Kokko, H., M. P. Harris, and S. Wanless. 2004. Competition for breeding sites and site-dependent population regulation in a highly colonial seabird, the common guillemot Uria aalge. Journal of Animal Ecology 73:367-376. https://doi.org/10.1111/j.0021-8790.2004.00813.x

Kress, S. W., and C. S. Hall. 2004. Tern management handbook - coastal northeastern United States and Atlantic Canada. U.S. Fish and Wildlife Service, Hadley, Maine, USA.

Laidlaw, R. A., T. G. Gunnarsson, V. Méndez, C. Carneiro, B. Þórisson, A. Wentworth, J. A. Gill, and J. A. Alves. 2020. Vegetation structure influences predation rates of early nests in subarctic breeding waders. Ibis 162:1225-1236. https://doi.org/10.1111/ibi.12827

Liljesthröm, M., S. Emslie, D. Frierson, and A. Schiavini. 2008. Avian predation at a southern rockhopper penguin colony on Staten Island, Argentina. Polar Biology 31:465-474. https://doi.org/10.1007/s00300-007-0372-1

MacArthur, R. H. 1958. Population ecology of some warblers of northeastern coniferous forests. Ecology 39:599-619. https://doi.org/10.2307/1931600

MacArthur, R., and R. Levins. 1967. The limiting similarity, convergence, and divergence of coexisting species. American Naturalist 101:377-385. https://doi.org/10.1086/282505

Martin, K., K. E. Aitken, and K. L. Wiebe. 2004. Nest sites and nest webs for cavity-nesting communities in interior British Columbia, Canada: nest characteristics and niche partitioning. Condor 106:5-19. https://doi.org/10.1093/condor/106.1.5

Martin, T. E. 1993. Nest predation and nest sites: new perspectives on old patterns. BioScience 43:523-532. https://doi.org/10.2307/1311947

Northfield, T. D., G. B. Snyder, A. R. Ives, and W. E. Snyder. 2010. Niche saturation reveals resource partitioning among consumers. Ecology Letters 13:338-348. https://doi.org/10.1111/j.1461-0248.2009.01428.x

Palestis, B. G. 2005. Nesting stage and nest defense by Common Terns. Waterbirds 28:87-94. https://doi.org/10.1675/1524-4695(2005)028[0087:NSANDB]2.0.CO;2

Pratte, I., R. A. Ronconi, S. R. Craik, and J. McKnight. 2021. Spatial ecology of endangered roseate terns and foraging habitat suitability around a colony in the western North Atlantic. Endangered Species Research 44:339-350. https://doi.org/10.3354/esr01108

Pulliam, H. R., and B. J. Danielson. 1991. Sources, sinks, and habitat selection: a landscape perspective on population dynamics. American Naturalist 137:50-66. https://doi.org/10.1086/285139

R Core Team. 2025. R: a language and environment for statistical computing. Version 4.5.1. R Foundation for Statistical Computing, Vienna, Austria. https://doi.org/10.32614/R.manuals

Ramos, J. A. 1998. Nest-site selection by roseate terns breeding on Aride Island, Seychelles. Waterbirds 21:438-443. https://doi.org/10.2307/1521658

Ramos, J. A., and A. J. del Nevo. 1995. Nest-site selection by Roseate Terns and Common Terns in the Azores. Auk 112:580-589.

Ripley, B., and W. Venables. 2025. Package nnet. R package version 7.3-20. R Foundation for Statistical Computing, Vienna, Austria. https://doi.org/10.32614/CRAN.package.nnet

Robertson, G. S., M. Bolton, W. J. Grecian, L. J. Wilson, W. Davies, and P. Monaghan. 2014. Resource partitioning in three congeneric sympatrically breeding seabirds: foraging areas and prey utilization. Auk: Ornithological Advances 131:434-446. https://doi.org/10.1642/AUK-13-243.1

Rock, J. C., M. L. Leonard, and A. W. Boyne. 2007. Foraging habitat and chick diets of Roseate Tern, Sterna dougallii, breeding on Country Island, Nova Scotia. Avian Conservation and Ecology 2:4. https://doi.org/10.5751/ACE-00126-020104

Samraoui, F., R. Nedjah, A. Boucheker, H. A. El-Serehy, and B. Samraoui. 2025. Sky high or grounded: nest-site selection of herons and egrets in a mixed-species colony. Ecology and Evolution 15:e70761. https://doi.org/10.1002/ece3.70761

Schoener, T. W. 1974. Resource partitioning in ecological communities. Science 185:27-39. https://doi.org/10.1126/science.185.4145.27

Schreiber, E. A., and J. Burger. 2002. Biology of marine birds. CRC Press, Boca Raton, Florida, USA. https://doi.org/10.1201/9781420036305

Seward, A., N. Ratcliffe, S. Newton, R. Caldow, D. Piec, P. Morrison, T. Cadwallender, W. Davies, and M. Bolton. 2019. Metapopulation dynamics of roseate terns: sources, sinks, and implications for conservation management decisions. Journal of Animal Ecology 88:138-153. https://doi.org/10.1111/1365-2656.12904

Shealer, D. A., and J. Burger. 1992. Differential responses of tropical Roseate Terns to aerial intruders throughout the nesting cycle. Condor 94:712-719. https://doi.org/10.2307/1369256

Spendelow, J. A. 1982. An analysis of temporal variation in, and the effects of habitat modification on, the reproductive success of Roseate Terns. Waterbirds 5:19-31. https://doi.org/10.2307/1521029

Stempniewicz, L. 1995. Predator-prey interactions between glaucous gull Larus hyperboreus and little auk Alle alle in Spitsbergen. Acta Ornithologica 29:155-170.

U.S. Fish and Wildlife Service. 2010. Caribbean Roseate Tern and North Atlantic Roseate Tern (Sterna Dougallii Dougallii). Five-year review: summary and evaluation. U.S. Fish and Wildlife Service, Boquerón, Puerto Rico, USA. https://semspub.epa.gov/work/01/518514.pdf

U.S. Fish and Wildlife Service. 2020. Roseate Tern Northeastern North American population (Sterna dougallii dougallii). Five-year review: summary and evaluation. U.S. Fish and Wildlife Service, Concord, New Hampshire, USA. https://ecos.fws.gov/docs/tess/species_nonpublish/3063.pdf

Weidinger, K. 2002. Interactive effects of cover, parental behavior, and predators on the survival of open passerine nests. Journal of Animal Ecology 71:424-437. https://doi.org/10.1046/j.1365-2656.2002.00611.x

Whittam, R. M., and M. L. Leonard. 1999. Predation and breeding success in roseate terns (Sterna dougallii). Canadian Journal of Zoology 77:851-856. https://doi.org/10.1139/z99-047

Whittam, R. M., and M. L. Leonard. 2000. Characteristics of predators and offspring influence nest defense by Arctic and Common Terns. Condor 102:301-306. https://doi.org/10.1093/condor/102.2.301