Abstract

Kleptoparasitism (food theft) is a tactic used opportunistically by many foraging birds, but little is known about its fitness benefits. Here we show that habitual kleptoparasitism by individual parent roseate terns ( Sterna dougallii ) is associated with consistently superior reproductive performance relative to nonkleptoparasitic (“honest”) parents, as measured by growth and survival to fledging among their offspring. In broods of two, both chicks of kleptoparasitic parents exhibited superior growth performance during the middle and later stages of the rearing period, relative to chicks of honest parents. This difference was especially pronounced in second-hatched chicks, whose survival is highly variable among years and dependent on food availability. Over a 10-year period, average productivity (number of chicks fledged per pair) was significantly higher among kleptoparasites than among honest parents, with a larger relative difference during years of food shortage. Our study indicates that kleptoparasitism in roseate terns is an important component of parental quality and provides the first evidence that food stealing is associated with enhanced fitness in a facultatively kleptoparasitic seabird.

The term “parental quality” is used to characterize individual animals whose reproductive performance is consistently high (or low) in successive breeding attempts, relative to other individuals ( Catry et al., 1999 ; Cobley et al., 1998 ). Previous studies of long-lived birds, especially seabirds, have shown such consistency in many aspects of reproductive performance, including probability of breeding ( Cam et al., 1998 ), laying date ( Catry et al., 1998 ), clutch size, egg size, hatching success, chick growth rates, and chick survival ( Coulson and Porter, 1985 ; Nisbet et al., 1998 ; Wendeln, 1997 ), as well as annual productivity (chicks raised per year) and lifetime reproductive success ( Coulson and Thomas, 1985 ; Mills, 1989 ; Thomas and Coulson, 1985 ). These aspects of reproductive performance are often correlated with each other and have been used as indices of parental quality ( Coulson and Porter, 1985 ; Mills, 1989 ; Nisbet et al., 1998 ; Wendeln, 1997 ; Wendeln and Becker, 1999 ).

Parental quality can be regarded as a phenotypic character that arises during development and is manifested during all or part of a lifetime, although it may have a genetic component also ( Cam and Monnat, 2000 ). However, little is known about its origins, ontogeny, or morphological, physiological, or behavioral characteristics that are correlated with it, other than physical size in some species ( Coulson, 1968 ; Mills, 1989 ; Wendeln, 1997 ; Wendeln and Becker, 1999 ). Although behavioral characteristics such as foraging specializations may predict breeding success of individuals within years ( Wendeln, 1997 ), none of these characteristics has been shown to predict consistently higher-than-average breeding performance in successive years.

Kleptoparasitism (food theft) is a well-known foraging tactic in birds ( Brockmann and Barnard, 1979 ) and is particularly common in seabirds ( Furness, 1987 ). Some species such as skuas ( Catharacta spp.), jaegers ( Stercorarius spp.), and sheathbills ( Chionis spp.) are considered obligate kleptoparasites, particularly during the breeding season or on migration. Other species such as gulls and terns (Laridae) may practice kleptoparasitism opportunistically at breeding colonies or in areas where birds congregate at abundant food resources, such as at landfills or behind fishing vessels ( Hudson and Furness, 1988 ; Steele and Hockey, 1995 ). For opportunists, kleptoparasitic behavior appears to be context dependent and is probably related to variation in food availability ( Broom and Ruxton, 1998 ; Furness, 1987 ; Oro, 1996 ) or to constraints on the ability of individuals to obtain food ( Duffy, 1980 , 1982 ; Hamilton, 2002 ).

In roseate terns ( Sterna dougallii ), kleptoparasitism is widespread, having been reported in England ( Dunn, 1973 ), Australia ( Hulsman, 1976 ), Puerto Rico ( Shealer et al., 1997 ), and the United States ( Shealer and Spendelow, 2002 ). Until recently, it was not known whether kleptoparasitism is practiced regularly by only a small percentage of the population or occasionally by most or all individuals, and the fitness consequences of engaging in this behavior had not been examined. At a colony site in Connecticut between 1995 and 1998, we identified and observed 10 individual adult roseate terns that regularly stole fish from other terns to feed their own chicks ( Shealer and Spendelow, 2002 ), indicating that, for this species at least, habitual kleptoparasitism is practiced by only a small number of individuals. Here we provide the first evidence that kleptoparasitism by parent birds is associated with superior growth performance in their chicks and consistently high breeding success in individuals, relative to nonkleptoparasitic (“honest”) foragers.

METHODS

Behavioral observations of roseate terns have been conducted annually since 1990 at Falkner Island (41° 13′ N, 72° 39′ W), Connecticut, USA. During the study period 1990–1999, Falkner Island supported 100–180 nesting pairs of roseate terns and 3300–4400 pairs of common terns ( Sterna hirundo ) each year. Observations were made from semipermanent hides located above each of the breeding areas of roseate terns on the island. Observations began when adults were establishing pair bonds and prospecting for nest sites. Individuals were identified by unique combinations of color rings and/or by reading numbers on field-readable (FR) or U.S. Bird Banding Laboratory (BBL) rings. Since 1992, we have marked all roseate tern chicks with a BBL ring on one leg and a four-character FR ring on the other leg. Since 1994, we have marked each adult with a unique six-ring combination of two metal and four color rings. More than 95% of the birds breeding at this site have been ringed, about 70% are of known age and about 95% are of known sex, which was determined by behavior (e.g., copulatory position and mate feeding) or sex-specific molecular markers ( Spendelow et al., 2002 ).

Kleptoparasitic roseate terns were identified during the breeding seasons 1995–1999 by observers who spotted and then followed, with binoculars or telescope, individuals that appeared to be patrolling the skies above the island. The behavior of birds engaged in aerial piracy is obvious and has been described in detail elsewhere ( Hulsman, 1976 ; Shealer and Spendelow, 2002 ). The birds were then followed visually to their nests and identified by reading color-ring combinations or ring numbers. Other criteria used for identifying kleptoparasitism included deliveries of parts of fish (rather than whole fish) to chicks at nests or feeds to chicks in rapid succession by a single adult, indicating that food was stolen in the vicinity of the colony. Most (8/10) kleptoparasites were female, and none had a kleptoparasitic mate (see Shealer and Spendelow, 2002 , for more detail).

Once egg laying was initiated, subcolony areas were checked every afternoon until the last known chick had fledged. Nests were marked with numbered yellow tongue depressors. Eggs were numbered in laying order with a nontoxic marker, measured (length and width) with calipers, and weighed with an electronic balance. Roseate tern eggs normally hatch at 23–24 days, so eggs were checked starting around day 21 for signs of hatching.

Chicks that hatched were given a BBL ring when found and an FR ring after they had reached about 25 g. Chicks soon start to move from the nest but were weighed daily (when found) to the nearest 0.1 g on an electronic balance and were placed either back in their original nest location or where they were found on that day. At about 21–22 days after hatching, a unique three-color pattern of waterproof marker was applied to the backs and wings of chicks. Observations were made from the hides and other locations on the island to confirm fledging of individuals. Methods for determining growth and estimating productivity and the descriptions of the kleptoparasitic behavior of roseate terns, including the 10 kleptoparasites in the present study, have been described elsewhere ( Dunn, 1973 ; Hulsman, 1976 ; Nisbet et al., 1995 , 1998 ; Shealer and Spendelow, 2002 ).

We compared measures of growth (1995–1999) and survival (1990–1999) of the offspring of these 10 kleptoparasites ( n = 40 brood-years for growth, n = 77 brood-years for survival) to those of nonkleptoparasitic parents and to a larger group of birds of unknown behavior that represented the average performance of the entire colony. The “nonkleptoparasitic” group contained a date-matched sample of nests ( n = 10 year −1 ) that were watched intensively in at least two different years to verify that neither parent behaved kleptoparasitically. The third category represented pairs that were not watched intensively but probably did not contain a kleptoparasitic parent ( Shealer and Spendelow, 2002 ). This group was not date matched to kleptoparasite nests but consisted of all remaining nests ( n = 50–60 year −1 ) for which detailed growth and survival records of chicks were obtained ( n = 320–350 total chicks over the 5-year period).

Analysis of variance (ANOVA) was used to analyze differences in growth performance of chicks according to hatching order (first or second), type of parent (kleptoparasite, nonkleptoparasite, or unknown), and their interaction. Growth performance was assessed by examining three parameters of the growth curve, corresponding to different stages of growth. Mass at day 3 (M3) is the body mass of each chick on day 3 (hatch = day 0) and provides a measure of early growth, which is a good predictor of subsequent survival of chicks to fledging ( Nisbet et al., 1999 ). Linear growth rate (LGR) is defined as the slope of a regression line fitted to mass data during the quasilinear period of growth where chicks are growing most rapidly (3–12 days for A-chicks and 4–13 days for B-chicks, where A and B refer to the first and second chicks that hatched in each brood, respectively). Asymptotic mass (AM) is defined as the mean of all masses (minimum of two) measured during the period of near-constant mass (17–28 days for A-chicks and 18–29 days for B-chicks; Nisbet et al., 1998 ) and most closely reflects the average masses of chicks at fledging age. At Falkner Island, broods consist of either one (approximately 30%) or two (approximately 70%) chicks. Growth rates of A-chicks from broods of one and two are similar ( Nisbet et al., 1995 ), so these two categories were combined for all analyses.

Analysis of covariance (ANCOVA) was used to analyze the simultaneous effects of categorical variables (year, hatch order, and parental type), continuous variables (egg mass and hatch date), and their interactions on the three growth parameters (M3, LGR, and AM) defined above. Although inclusion of other factors and covariates in the models was possible, previous research on this species ( Nisbet et al., 1995 , 1998 ) indicated that the variables listed above were likely to be most important in explaining differences in growth performance among chicks. We could not use parental age as a candidate predictor variable in the model because the exact ages were known for only 4 of the 10 kleptoparasites. In a previous study, however, parental age was shown to be relatively unimportant in explaining differences in early growth among roseate tern chicks ( Nisbet et al., 1998 ). Model selection followed Nisbet et al. (1995 , 1998 ). Type III sums of squares were used to evaluate significance for each analysis. F tests for inequality of slopes were significant only for M3 ( F14,471 = 2.14, p = .009), so an unequal-slopes model was used for testing this variable.

RESULTS

ANOVA models ( Table 1 ) revealed significant effects ( p < .001) of parental type and hatching order on early (M3), middle (LGR), and later (AM) stages of chick growth. However, a significant interaction between parental type and hatching order also was present for each growth parameter. For parental type, A- and B-chicks from nests with a kleptoparasitic parent exhibited superior growth performance relative to A- and B-chicks from nests with nonkleptoparasitic parents and to the overall average ( Figure 1 ). For hatching order, A-chicks exhibited superior growth performance relative to B-chicks. The nature of the interaction effect was such that for A-chicks, both LGR and AM were significantly higher in chicks with a kleptoparasitic parent than in either of the other two categories ( p < .05 in all cases), and for B-chicks, M3, LGR, and AM were all significantly higher in chicks with a kleptoparasitic parent than in either of the other two categories ( p < .01 in all cases). In fact, B-chicks with a kleptoparasitic parent grew almost as fast as A-chicks in the other two categories ( p > .05 in both cases, Figure 1 ).

Figure 1

Growth performance of first-hatched (A-chicks) and second-hatched (B-chicks) roseate terns with one kleptoparasitic parent (white bars), two nonkleptoparasitic parents (black bars), and parents of unknown foraging behavior (gray bars). (a) Mass at day 3 is a measure of early growth, (b) linear growth rate is the average daily mass gain during the middle growth period, and (c) asymptotic mass provides an estimate of fledging mass. Given are means (±SD), with sample size (number of chicks) above bars.

Figure 1

Growth performance of first-hatched (A-chicks) and second-hatched (B-chicks) roseate terns with one kleptoparasitic parent (white bars), two nonkleptoparasitic parents (black bars), and parents of unknown foraging behavior (gray bars). (a) Mass at day 3 is a measure of early growth, (b) linear growth rate is the average daily mass gain during the middle growth period, and (c) asymptotic mass provides an estimate of fledging mass. Given are means (±SD), with sample size (number of chicks) above bars.

Table 1

Results of ANOVA models for growth parameters of roseate tern chicks at Falkner Island, Connecticut, USA, according to parental type (one kleptoparasitic parent, two nonkleptoparasitic parents, and parents of unknown foraging behavior) and hatch order (first or second) in a brood

  Dependent variables
 
   
Source of variation
 
df
 
M3
 
LGR
 
AM
 
Parental type 26.4*** 18.4*** 21.5*** 
Hatch order  173.8 ***  78.1 ***  108.3 *** 
Parental type × hatch order  11.7 ***  4.3 *  5.7 ** 
Error df
 

 
655
 
619
 
533
 
  Dependent variables
 
   
Source of variation
 
df
 
M3
 
LGR
 
AM
 
Parental type 26.4*** 18.4*** 21.5*** 
Hatch order  173.8 ***  78.1 ***  108.3 *** 
Parental type × hatch order  11.7 ***  4.3 *  5.7 ** 
Error df
 

 
655
 
619
 
533
 

M3 = mass at day 3, LGR = linear growth rate, AM = asymptotic mass (see text for details). Given are F ratios with associated probabilities (* p < .05, ** p < .01, *** p < .001).

ANCOVA models ( Table 2 ) indicated that M3 was significantly related to egg mass and to parental type ( p < .001 in each case), but the latter relationship was significant only in B-chicks. There also were significant interactions ( p < .05 in each case) with hatch date × parental type (M3 was not related to hatch date in chicks with a kleptoparasitic parent but was negatively related in others), egg mass × hatch order (M3 was more strongly positively related to egg mass in B-chicks than in A-chicks), and hatch date × hatch order (M3 was more strongly negatively related to hatch date in B-chicks than in A-chicks). LGR and AM were significantly related to year, hatch order, and parental type ( p < .001 in each case), with no significant interactions. Egg mass and hatch date were weakly significant predictors of middle-stage growth (LGR), but neither was associated with AM ( Table 2 ).

Table 2

Results of ANCOVA models for growth parameters of roseate tern chicks at Falkner Island, Connecticut, USA

  Dependent variables
 
  
Independent variables
 
M3
 
LGR
 
AM
 
Categorical variables ( F ratios)     
    Year ns 29.16*** 13.42*** 
    Parental type ns 14.04*** 8.38*** 
    Hatch order ns 56.00*** 51.03*** 
    Parental type × hatch order 7.67*** ns ns 
Continuous variables (regression coefficients)    
    Egg mass  +0.732 day −1 ***   −0.069 day −1 *  ns 
    Hatch date ns  −0.019 g day −2 *  ns 
Interactions ( F ratios)     
    Egg mass × year ns — — 
    Egg mass × parental type ns — — 
    Hatch date × parental type 3.90* — — 
    Egg mass × hatch order 4.99* — — 
    Hatch date × hatch order 4.03* — — 
Model    
     R2 0.444 0.306 0.287 
     F ( p )  15.03*** 19.31*** 16.32*** 
    df
 
25, 471
 
11, 485
 
11, 445
 
  Dependent variables
 
  
Independent variables
 
M3
 
LGR
 
AM
 
Categorical variables ( F ratios)     
    Year ns 29.16*** 13.42*** 
    Parental type ns 14.04*** 8.38*** 
    Hatch order ns 56.00*** 51.03*** 
    Parental type × hatch order 7.67*** ns ns 
Continuous variables (regression coefficients)    
    Egg mass  +0.732 day −1 ***   −0.069 day −1 *  ns 
    Hatch date ns  −0.019 g day −2 *  ns 
Interactions ( F ratios)     
    Egg mass × year ns — — 
    Egg mass × parental type ns — — 
    Hatch date × parental type 3.90* — — 
    Egg mass × hatch order 4.99* — — 
    Hatch date × hatch order 4.03* — — 
Model    
     R2 0.444 0.306 0.287 
     F ( p )  15.03*** 19.31*** 16.32*** 
    df
 
25, 471
 
11, 485
 
11, 445
 

M3 = mass at day 3, LGR = linear growth rate, AM = asymptotic mass (see text for details). F ratios are given for categorical predictor variables and interactions, and regression coefficients are given for continuous variables. ns = not significant, * p < .05, *** p < .001, — indicates not included in the model.

We compared breeding success over the 10-year period (1990–1999) for which we had reproductive data on some or all of these kleptoparasitic individuals ( n = 4–10 birds year −1 ). Average productivity (number of chicks fledged per pair) was significantly higher among pairs with a kleptoparasitic parent than among nonkleptoparasites (paired t9 = 6.18, p < .001, Figure 2 ). This difference was primarily due to enhanced survival of B-chicks with a kleptoparasitic parent (overall mean productivity, 1.20 fledglings pair −1 ± 0.20 SD) relative to those without (0.83 ± 0.25). The slope of the regression line (0.517 ± 0.200 SE) in Figure 2 is significantly lower than 1 (95% confidence interval of slope = 0.065–0.969), indicating that the relative difference between the breeding success of pairs with a kleptoparasitic parent and nonkleptoparasites was greater in years of low overall productivity.

Figure 2

Productivity (mean number of chicks fledged per pair) comparisons between kleptoparasites and nonkleptoparasites at Falkner Island, 1990–1999. Each point represents a single year. Diagonal line represents equal reproductive success between groups.

Figure 2

Productivity (mean number of chicks fledged per pair) comparisons between kleptoparasites and nonkleptoparasites at Falkner Island, 1990–1999. Each point represents a single year. Diagonal line represents equal reproductive success between groups.

DISCUSSION

Consistent with studies of other species ( Moreno, 1998 ; Williams, 1994 ), egg mass and early breeding are reliable predictors of reproductive success in roseate terns ( Nisbet, 1978 ; Nisbet et al., 1990 , 1995 ), and they have been proposed as indices of parental quality ( Nisbet et al., 1998 ). However, ANCOVA models indicated that these effects were negligible when kleptoparasitism was included as a predictor variable ( Table 2 ). Although egg mass was important in explaining differences in early growth (M3), it became less important as the chicks grew larger, and hatching date alone was not a significant predictor of early growth. These results suggest that the components of parental quality that enabled parents to lay large eggs or to lay early in the season gave way in importance to their ability to provide food to chicks. The significant hatching order × parental type interaction for M3 indicates that differences in early growth among B-chicks depended on whether or not these chicks had a kleptoparasitic parent ( Table 1 ; Figure 1 ).

Three factors were consistent predictors of middle (LGR) and late (AM) stages of growth: year, hatching order, and whether the chick had a kleptoparasitic parent ( Table 2 ). The year effect most likely reflects annual fluctuations in food availability near the colony; the hatching-order effect is due to physiological constraint whereby eggs are laid, and hence hatch, 2–3 days apart ( Gochfeld et al., 1998 ), giving the older A-chick a competitive advantage over the B-chick. However, having a kleptoparasitic parent was nearly as important as a chick's hatching order in explaining differences in growth parameters. For B-chicks with a kleptoparasitic parent, middle and later stages of growth were nearly equal to A-chicks from nonkleptoparasitic nests ( Figure 1 ).

Average productivity (number of chicks fledged per pair averaged among pairs within years) was significantly higher among pairs with a kleptoparasitic parent than among nonkleptoparasites, primarily due to enhanced survival of B-chicks in the kleptoparasite group. Thus, assuming equal reproductive lifetimes, kleptoparasitic individuals would be expected to produce about 45% more fledglings per capita than nonkleptoparasitic birds, representing a considerable fitness advantage.

The relative difference in reproductive success was more pronounced in years of low overall productivity ( Figure 2 ), suggesting that reproductive consequences of food shortage affected kleptoparasites less severely than nonkleptoparasitic foragers. In terns, male parents assume most of the feeding duties during the first few days after chicks hatch ( Fasola and Saino, 1995 ; Wiggins and Morris, 1987 ; authors' unpublished data for roseate terns), whereas females attend the brood. A 10-year study at the same site ( Nisbet et al., 1998 ) revealed that the survival of B-chicks to fledging was strongly dependent on their growth during the first few days after hatching. Thus, having a kleptoparasitic parent enhanced the growth and subsequent survival probability of the B-chicks relative to nonkleptoparasitic parents.

Because the survival of the B-chick depends on its early growth ( Nisbet et al., 1998 , 1999 ), the extent to which the female parent contributes to the feeding of her chicks during the early nestling period is probably a primary determinant of whether the B-chick survives or starves. It is therefore noteworthy that 8 of 10 kleptoparasites were females ( Shealer and Spendelow, 2002 ). By stealing fish near their nests, females do not compromise their role as protectors of the brood while males are away on foraging trips, which at Falkner Island average 45–60 min (Shealer, unpublished data). Average prey delivery rates by kleptoparasites to their chicks ranged from 2 to 13 times higher than honest parents ( Shealer and Spendelow, 2002 ), suggesting that the ability to obtain prey quickly and satiate the chicks is indeed the mechanism by which such high success is achieved.

We have identified here a behavioral attribute of individual marine birds that is associated with consistently superior breeding performance relative to the colony average. Consistently high breeding performance is part of the definition of phenotypic quality ( Catry et al., 1999 ; Cobley et al., 1998 ), and we have previously identified hatching date, egg mass, and early growth of chicks as indices of parental quality in roseate terns ( Nisbet et al., 1998 ). This study, however, suggests that hatching date and egg mass are only minor predictors of parental performance after controlling for kleptoparasitism, and we conclude that kleptoparasitism is a more important index of phenotypic quality in roseate terns. It remains to be determined whether kleptoparasitism is a primary component of phenotypic quality or whether it results independently from some other factor. Simple cross-fostering experiments, whereby eggs or nestlings are transplanted between nests of kleptoparasites and nonkleptoparasites, would be useful in helping to tease apart correlative from causative factors.

Recent theoretical papers have attempted to model facultative kleptoparasitism as a context-dependent foraging strategy, whose frequency in a population is influenced by food availability, contest duration ( Broom and Ruxton, 1998 ), or distribution and proficiency of competitors ( Hamilton, 2002 ). These models predict that individuals should switch their foraging tactics (searching for or contesting for food) depending on environmental conditions. The situation with roseate terns at Falkner Island appears inconsistent with this prediction because, despite apparent annual variation in food availability (as indexed by significant year effects in chick growth and survival), the number of kleptoparasites remained constant. The fact that the pool of kleptoparasites comprised the same individuals year after year further suggests that kleptoparasitic behavior becomes fixed at some point in life. The low frequency (3–5%) of habitual kleptoparasitism by roseate terns in this population suggests that some mechanism (e.g., high cost or phenotypic constraint) prevents most individuals from adopting this strategy. The availability of victims does not appear to be limiting because roseate terns steal fish both from conspecifics and from common terns ( Shealer and Spendelow, 2002 ), which outnumber them 30-fold on the island.

If kleptoparasitism is simply a component, rather than an independent expression, of overall quality in roseate terns and if individual quality is heritable to some degree, we might expect to see this behavior expressed among a certain fraction of the offspring when they reach breeding age. However, of the 29 chicks reared by kleptoparasitic parents and seen subsequently as breeders on Falkner Island (oldest F1 individual = 14 years in 2003), none has been identified as a kleptoparasite. We likewise have not identified any morphological characteristics (e.g., body mass and wing length) shared by kleptoparasites that might predispose them to this tactic. To date, it remains unclear under what circumstances individual roseate terns become kleptoparasitic or why only a small fraction are able to do so successfully. Investigation of the ontogeny of this behavioral strategy may reveal its constraints.

We thank the U.S. Fish and Wildlife Service for support and permission to work on the Falkner Island Unit of the Stewart B. McKinney NWR, the many research assistants for their help over the years with the fieldwork, and the following organizations (in alphabetical order) for their logistic and/or financial support: Colgate University, Connecticut Audubon Society, Connecticut Chapter of The Nature Conservancy, Connecticut Department of Environmental Protection, Fulton Foundation, Little Harbor Laboratory, Menunkatuck Audubon Society, U.S. Fish and Wildlife Service, and U.S. Geological Survey (Patuxent Wildlife Research Center). We thank A.F.G. Bourke and two anonymous reviewers for helpful comments on the manuscript.

References

Brockmann HJ, Barnard CJ,
1979
. Kleptoparasitism in birds.
Anim Behav
 
27
:
487
–514.
Broom M, Ruxton GD,
1998
. Evolutionary stable stealing: game theory applied to kleptoparasitism.
Behav Ecol
 
9
:
397
–403.
Cam E, Monnat JY,
2000
. Stratification based on reproductive state reveals contrasting patterns of age-related variation in demographic parameters in the kittiwake.
Oikos
 
90
:
560
–574.
Cam E, Hines JE, Monnat JY, Nichols JD, Danchin E,
1998
. Are adult nonbreeders prudent parents? The kittiwake model.
Ecology
 
79
:
2917
–2930.
Catry P, Ratcliffe N, Furness RW,
1998
. The influence of hatching date on different life-history stages of great skuas Catharacta skua .
J Avian Biol
 
29
:
299
–300.
Catry P, Ruxton GD, Ratcliffe N, Hamer KC, Furness RW,
1999
. Short-lived repeatabilities in long-lived great skuas: implications for the study of individual quality.
Oikos
 
84
:
473
–479.
Cobley ND, Croxall JP, Prince PA,
1998
. Individual quality and reproductive performance in the grey-headed albatross Diomedea chrysostoma .
Ibis
 
140
:
315
–322.
Coulson JC,
1968
. Differences in the quality of birds nesting in the center and on the edge of a colony.
Nature
 
217
:
478
–479.
Coulson JC, Porter JM,
1985
. Reproductive success of the kittiwake Rissa tridactyla : the roles of clutch size, chick growth rates and parental quality.
Ibis
 
127
:
450
–466.
Coulson JC, Thomas CS,
1985
. Differences in the breeding performance of individual kittiwake gulls, Rissa tridactyla (L.). In: Behavioural ecology: ecological consequences of adaptive behaviour (Sibly RM, Smith RH, eds). Oxford: Blackwell; 489–503.
Duffy DC,
1980
. Patterns of piracy by Peruvian seabirds: a depth hypothesis.
Ibis
 
122
:
521
–525.
Duffy DC,
1982
. Patterns of piracy in the seabird communities of the Galápagos Islands and southern Africa.
Cormorant
 
10
:
71
–80.
Dunn EK,
1973
. Robbing behavior of roseate terns.
Auk
 
90
:
641
–651.
Fasola M, Saino N,
1995
. Sex-biased parental-care allocation in three tern species (Laridae, Aves).
Can J Zool
 
73
:
1461
–1467.
Furness RW,
1987
. Kleptoparasitism in seabirds. In: Seabirds: feeding ecology and role in marine ecosystems (Croxall JP, ed). Cambridge: Cambridge University Press; 77–100.
Gochfeld M, Burger J, Nisbet ICT,
1998
. Roseate tern ( Sterna dougallii ). In: The birds of North America, no. 370 (Poole A, Gill F, eds). Philadelphia: The Birds of North America, Inc.
Hamilton IM,
2002
. Kleptoparasitism and the distribution of unequal competitors.
Behav Ecol
 
13
:
260
–267.
Hudson AV, Furness RW,
1988
. Utilisation of discarded fish by scavenging seabirds behind whitefish trawlers in Shetland.
J Zool Lond
 
215
:
151
–166.
Hulsman K,
1976
. The robbing behaviour of terns and gulls.
Emu
 
76
:
143
–149.
Mills JA,
1989
. Red-billed gull. In: Lifetime reproduction in birds (Newton I, ed). London: Academic Press; 387–404.
Moreno J,
1998
. The determination of seasonal declines in breeding success in seabirds.
Etología
 
6
:
7
–31.
Nisbet ICT,
1978
. Dependence of fledging success on egg-size, parental performance and egg-composition among common and roseate terns, Sterna hirundo and S. dougallii .
Ibis
 
120
:
207
–215.
Nisbet ICT, Burger J, Safina C, Gochfeld M,
1990
. Estimating fledging success and productivity in roseate terns ( Sterna dougallii ).
Colon Waterbirds
 
13
:
85
–91.
Nisbet ICT, Hatfield JS, Link WA, Spendelow JA,
1999
. Predicting chick survival and productivity of roseate terns from data on early growth.
Waterbirds
 
22
:
90
–97.
Nisbet ICT, Spendelow JA, Hatfield JS,
1995
. Variations in growth of roseate tern chicks.
Condor
 
97
:
335
–344.
Nisbet ICT, Spendelow JA, Hatfield JS, Zingo JM, Gough GA,
1998
. Variations in growth of roseate tern chicks. II. Early growth as an index of parental quality.
Condor
 
100
:
305
–315.
Oro D,
1996
. Interspecific kleptoparasitism in Audouin's gull Larus audouinii at the Ebro Delta, northeast Spain: a behavioural response to low food availability.
Ibis
 
138
:
218
–221.
Shealer DA, Floyd T, Burger J,
1997
. Host choice and success of gulls and terns kleptoparasitizing brown pelicans.
Anim Behav
 
53
:
655
–665.
Shealer DA, Spendelow JA,
2002
. Individual foraging strategies of kleptoparasitic roseate terns.
Waterbirds
 
25
:
436
–441.
Spendelow JA,
1982
. An analysis of temporal variation in, and the effects of habitat modification on, the reproductive success of roseate terns.
Colon Waterbirds
 
5
:
19
–31.
Spendelow JA, Nichols JD, Hines JE, Lebreton J-D, Pradel R,
2002
. Modelling postfledging survival and age-specific breeding probabilities in species with delayed maturity: a case study of roseate terns at Falkner Island, Connecticut.
J Appl Stat
 
29
:
385
–405.
Steele WK, Hockey PAR,
1995
. Factors influencing the rate and success of intraspecific kleptoparasitism among kelp gulls ( Larus dominicanus ).
Auk
 
112
:
847
–859.
Thomas CS, Coulson JC,
1985
. Reproductive success of kittiwake gulls, Rissa tridactyla . In: Reproductive success (Clutton-Brock TH, ed). Chicago: University of Chicago Press; 251–262.
Wendeln H,
1997
. Body mass of female common terns ( Sterna hirundo ) during courtship: relationships to male quality, egg mass, diet, laying date and age.
Colon Waterbirds
 
20
:
235
–243.
Wendeln H, Becker PH,
1999
. Effects of parental quality and effort on the reproduction of common terns.
J Anim Ecol
 
68
:
205
–214.
Wiggins DA, Morris RD,
1987
. Parental care of the common tern.
Ibis
 
129
:
533
–540.
Williams TD,
1994
. Intraspecific variation in egg size and egg composition in birds: effects on offspring fitness.
Biol Rev
 
69
:
35
–59.

Author notes

a Department of Biology, Loras College, 1450 Alta Vista, Dubuque, IA 52004-0178, USA, b U.S. Geological Survey, Patuxent Wildlife Research Center, Laurel, MD 20708, USA, and c ICT Nisbet & Company, North Falmouth, MA 02556, USA