acta ethologica

, Volume 10, Issue 2, pp 73–79

The function of habitat change during brood-rearing in the precocial Kentish plover Charadrius alexandrinus


    • Department of Biology and BiochemistryUniversity of Bath
  • Tamás Székely
    • Department of Biology and BiochemistryUniversity of Bath
  • Innes C. Cuthill
    • Centre for Behavioural Biology, School of Biological SciencesUniversity of Bristol
Original Paper

DOI: 10.1007/s10211-007-0032-z

Cite this article as:
Kosztolányi, A., Székely, T. & Cuthill, I.C. acta ethol (2007) 10: 73. doi:10.1007/s10211-007-0032-z


Food availability is often variable during the breeding season. Parents with nonmobile, altricial young have no choice but to accept changes in local food availability, whereas in precocial animals, the parents may lead their young away from poor sites to areas that have rich resources and/or are safe from predators. We investigated the latter hypotheses in the Kentish plover Charadrius alexandrinus, a precocial shorebird that raises its young in two habitats: on lakeshore and in saltmarsh. Parents move with their broods from saltmarsh to lakeshore, especially late in the breeding season, and we hypothesized that lakeshores provide more food than the saltmarsh. Consistent with our hypotheses, plover chicks grew faster on the shore, and the difference in growth rates between the two habitats was amplified later in the breeding season. In addition, brood survival was higher on lakeshore than in saltmarsh and decreased with hatching date. Taken together, our results suggest that Kentish plover parents increase their reproductive success by switching brood-rearing habitats strategically.


Brood movementParental careHabitat choicePrecocial bird


The environment rarely stays constant during a breeding season. Thus, ecological conditions may change radically between the time the young are born and when they become independent. Unless they can be carried, in altricial species, the parents can only buffer their young from any fluctuations in the environment by changing their foraging site and/or the intensity of foraging. However, in precocial species (i.e., species in which the young are mobile shortly after birth), the parents have the option of leading the young to different habitats until they attain full independence. For example, in many precocial shorebirds and waterfowl, the parent(s) leads away their downy chicks from the nest to sites where the chicks feed, grow, and fledge (Walters 1984; Sedinger 1992).

The young may gain several sorts of benefits by moving between sites. First, they can be better protected from predators. For instance, ungulates and macropods that live in dense habitats hide their young in the vegetation, whereas in species that live in open habitats, the young follow their mother, presumably because an offspring attempting to hide in the open would have a poor chance of survival (Fisher et al. 2002). Second, parents may change habitat to avoid competition with, or infanticide by, conspecifics (Laing and Raveling 1993). For instance, mammals with mobile offspring appear less vulnerable to infanticide by conspecific females and therefore do not defend a territory (e.g., hare Lepus europaeus, sea otter Enhydra lutra, Hystricognathi rodents), whereas species with altricial offspring (e.g., rabbit Oryctolagus cuniculus, river otter Lutra lutra, Sciurognathi rodents) defend a territory (Wolff and Peterson 1998). Third, parents may take their young to sites with better resources (Diesel et al. 1995; Hagen et al. 2005; Johansson and Blomqvist 1996; Pearce-Higgins and Yalden 2004). For instance, Sandwich terns Sterna sandvicensis lead their chicks away from the nest to reduce food robbery by black-headed gulls Larus ridibundus, and this strategy accelerates the growth of their chicks (Stienen and Brenninkmeijer 1999).

The benefits of switching sites, however, should be traded off against costs. First, in ungulates and macropods with following young, weaning is longer than in species with hiding young, presumably because young expend the energy gained from milk on movement rather than on growth (Fisher et al. 2002). Second, chick mortality may increase during extensive movements because of higher predation or risks of starvation (Blomqvist and Johansson 1995). Precocial chicks use considerable amount of energy for thermoregulation and locomotion, and they need to keep their energy balance within narrow margins to grow and survive (Schekkerman and Visser 2001). Thus, a reduced food intake during brood movements may push the chicks into a negative energy balance. Third, a brood-rearing habitat may be beneficial from one aspect whilst costly from another; for instance, habitats with abundant food may be associated with high predation risk (Houston et al. 1993).

We investigated a small shorebird, the Kentish plover Charadrius alexandrinus (body mass approximately 42 g), that nests on the ground, and both parents incubate the clutch (Fraga and Amat 1996; Kosztolányi and Székely 2002). However, after hatching of the precocial chicks, either the male or, more often, the female may desert the brood (Warriner et al. 1986; Székely and Lessells 1993; Amat et al. 1999). Thus, some broods are raised by both parents, whereas in others, either only the male or only the female cares for the chicks. Kentish plovers raise their young in two habitats in Southern Turkey: on lakeshore and in saltmarsh, and broods move between these two types of habitat. Most broods hatch on the saltmarsh, however, they spend more time on lakeshore as the breeding season proceeds. This movement of families is consistent with the changes in the food intake rates in the two habitats (Kosztolányi et al. 2006).

We tested two main hypotheses on the function of habitat change raised by the results in Kosztolányi et al. (2006). First, as more food is available on the shore than in saltmarsh, as indicated by higher food intake rate of parents and their chicks, we predicted that the chicks should grow faster on the lakeshore and should have higher survival because of low risk of starvation. However, the effect of better food supply on brood survival may be counterbalanced by predation and/or infanticide. The chicks feed on the shore in the open, thus they may be more exposed to predators than in the saltmarsh where they can hide in the vegetation. Furthermore, the density of plovers is higher on the shore, especially late in the breeding season (Kosztolányi et al. 2006), and adults in neighboring families do kill chicks (Székely and Cuthill 1999). Thus, the intense competition between neighboring families on the shore may lead to lower brood survival. Second, more food is available on the shore; however, there are families that spend most of their time in the saltmarsh. This may indicate that some parents are unable to lead their chick to the habitat providing more resources; thus, we predicted that there are differences in quality of parents regarding where they raise their chicks.

Materials and methods

Study area and field methods

Fieldwork was carried out at Lake Tuzla (36° 43′ N, 35° 35′ E), Southern Turkey, between 1996 and 1999 (see details in Kosztolányi et al. 2006). Kentish plovers reared their chicks on the north shore of the lake (‘shore’ henceforward) and on a 50- to 800-m-wide strip of alkaline grassland on the north side of the lake (‘saltmarsh’ henceforward). The study area was divided into two sites: Site 1 (152 ha) and Site 2 (51 ha). Both sites included both types of habitats. These sites overlap with but are not entirely the same as the sites mentioned in previous studies of the same population (see Székely and Cuthill 1999; Székely and Cuthill 2000).

We investigated 144 families in total (n = 38, 24, 47, 35 broods in the 4 years, respectively). Both parents were caught and ringed with a metal ring and a unique combination of color rings. Parents were caught either during late incubation or shortly after hatching of the clutch. We measured their body mass (nearest 0.1 g) with a spring balance and their right tarsus length (nearest 0.1 mm) with calipers. Most chicks were captured and ringed within 0–2 days after hatching when we also measured their body mass (nearest 0.1 g) and length of their right tarsus (nearest 0.1 mm). We attempted to recapture and to remeasure broods every fourth day. We visited the broods and recorded the number of chicks and the number and sex of attending parent(s) at least every other day until the chicks died or reached the age of 25 days. Kentish plover chicks fledge at about 29 days of age (Cramp and Simmons 1983).

Statistical analyses

We considered each brood as the unit of analysis. We had several observations per broods; therefore, to avoid pseudoreplication, we took the mean of records for each brood. Samples sizes may be different between analyses because of missing values. Brood-rearing habitat was defined as the proportion of time the brood spent on the shore during brood visits (see above). For the purpose of illustration (e.g., Fig. 1), brood-rearing habitat was sometimes recoded as a binary variable: broods spending <0.5 of the time on shore were allocated to saltmarsh, whereas broods spending ≥0.5 of the time on shore were allocated to shore. Hatching date was calculated as number of days since 1st March. If the exact hatching date was unknown, we used tarsus length of chicks to estimate their hatching date (see Székely and Cuthill 1999). If a parent was measured several times during a breeding season, we took the mean of these measurements. The age of the parents was calculated using their minimum age (in years): for instance, the minimum age of a parent in 1997 that was caught as a breeding adult in 1996 was 2 years.
Fig. 1

Growth of Kentish plover chicks in relation to hatching date in two brood-rearing habitats (a, c; open circles and broken line, saltmarsh; solid circles and solid line, shore) and in relation to hatching date in two brood-rearing sites (b, d; open triangles and broken line, Site 1; solid triangles and solid line, Site 2). Note that the time spent on the shore was a continuous variable in the analyses; however, it has been converted to a binary variable for the purpose of illustration (see “Materials and methods”)

Chick growth was estimated by the slope of growth curves (Székely and Cuthill 1999): the change in body mass (g × day−1) was calculated as the slope of the linear regression of body mass (in g) on chick age (in days), whereas tarsus growth rate (mm × day−1) was calculated as the slope of the linear regression of the right tarsus length (in mm) on chick age (in days). For each brood, a separate regression line was fitted for body mass and tarsus. Tarsus growth was approximately linear before 25 days of age, whilst for body mass, the small number of recaptures precluded more sophisticated growth rate models (e.g., Ricklefs 1967).

The influences of habitat selection and breeding season on body mass change and tarsus growth of chicks were investigated by general linear models (GLMs). Chick growth has a quadratic relationship with hatching date (Székely and Cuthill 1999); therefore, we included hatching date as a second order polynomial in models of chick growth. Growth of chicks may be different between sites (see Székely and Cuthill 1999); therefore, site was included as a factor in GLMs. We included all two-way interactions in the initial models and deleted the nonsignificant interaction terms stepwise. Growth of chicks may differ between years; however, after controlling for hatching date and site, the year did not influence growth rate (results not shown); thus, including year in GLMs would not alter our conclusions.

Brood survival was estimated by a maximum likelihood method (Noszály et al. 1995; Székely and Cuthill 1999). The survival estimate varies between 0 (all chicks died before the age of 3 days) and 1 (all chicks survived until the age of 25 days). The distribution of chick survival was skewed, therefore, we used Spearman correlations and partial rank correlations for the analyses of this variable (Daniel 1990). The correlation between survival and brood-rearing habitat was similar in the 4 years (Spearman correlations 1996–1999: rs = 0.308–0.560, p = 0.006–0.072); therefore, year effects were not considered in the analyses of brood survival.

Several experimental manipulations were carried out during the study (e.g., Székely and Cuthill 1999; Székely and Cuthill 2000), and these manipulations may potentially influence our results presented here. However, none of the experimental manipulations influenced the growth rates of chicks (analyses of variance [ANOVAs], 1996–1999, body mass: p = 0.211–0.547; tarsus length: p = 0.140–0.875), brood survival (Kruskal–Wallis tests, 1996–1999: p = 0.122–0.948), or habitat use of the broods (ANOVAs, 1996–1999: p = 0.079–0.792) in the current dataset; therefore, manipulations were not considered in the analyses.

The influences of parental quality on habitat selection were investigated by entering both the male and the female parents in GLMs. In broods where one parent deserted (n = 49), the brood visitation records after the desertion event (at brood age of 12.1 ± 1.00 days, mean ± SE) were excluded from these analyses. Habitat selection may depend on time in the breeding season (see Kosztolányi et al. 2006); thus, we also included the observation date in models of parental quality. None of the two-way interactions were significant; therefore, they were removed from the GLMs. Habitat selection was not different between years (F3,106 = 0.952, p = 0.418); therefore, year was not included in these analyses. Statistical analyses were carried out using R 2.1.0 (R Development Core Team 2005). In the GLMs, Type III sums of squares were used, and two-tailed probabilities are given.


Chick development

Both body mass change and tarsus growth of chicks were influenced by time spent on the shore and hatching date (Table 1). In families that spent more time on the shore, body mass increased faster than in families that stayed in the saltmarsh, and the difference between habitats was larger late in the breeding season (Fig. 1a). Tarsus growth tended to increase over the breeding season on the shore, whereas it decreased late in the breeding season in the saltmarsh (Fig. 1c). Growth rates were similar between the two sites early in the breeding season and then tended to decrease more rapidly in Site 1 than in Site 2 (Fig. 1b,d).
Table 1

Growth rates of Kentish plover chicks in relation to time they spent on the shore, hatching date, and site (error df: body mass change, 120; tarsus growth, 119)



Body mass change (g × day−1)

Tarsus growth (mm × day−1)





Time spent on the shore






Hatching date (quadratic term)












Time spent on the shore × hatching date






Hatching date × site






Brood survival

Brood survival increased with the time spent on the shore (Fig. 2, Spearman correlation, rs = 0.362, n = 143, p < 0.001). The strength of this relationship increased after controlling for hatching date and site by partial rank correlation (habitat: rs = 0.426, n = 143, p < 0.001; hatching date: rs = −0.269, p = 0.001; site: rs = 0.331, p < 0.001).
Fig. 2

Brood survival in the Kentish plover chicks in relation to brood-rearing habitat. See text for statistics

Parental quality

Broods attended by heavy females spent more time on the shore than light ones (Fig. 3, GLM, F1,106 = 6.483, p = 0.012), whereas the body mass of males did not influence the time their brood spent on the shore (F1,106 = 0.079, p = 0.779; observation date: F1,106 = 14.213, p < 0.001).
Fig. 3

The brood-rearing habitat in relation to body mass and tarsus length of male (a, c) and female (b, d) parents (least-squares regression, male body mass: B = 0.005 ± 0.012, n = 110, p = 0.678; female body mass: B = 0.027 ± 0.010, n = 110, p = 0.010; male tarsus: B = −0.017 ± 0.028, n = 110, p = 0.528; female tarsus: B = 0.088 ± 0.033, n = 110, p = 0.009). See text for GLMs

Similarly, broods attended by large females (as indicated by their longer tarsus) spent more time on the shore than small ones (Fig. 3, GLM, F1,106 = 7.521, p = 0.007), whilst there was no relationship between brood-rearing habitat and body size of males (F1,106 = 0.001, p = 0.982; observation date: F1,106 = 14.086, p < 0.001).

Finally, the age of parents was unrelated to habitat use (GLM, male age: F1,106 = 0.323, p = 0.571; female age: F1,106 = 0.192, p = 0.663; observation date: F1,106 = 13.254, p < 0.001).


Both chick development and survival were related to the brood-rearing habitat in the Kentish plover. As predicted, chicks grew faster on the shore, and brood survival was also higher. Thus, by taking their chicks to the lake shore, parents may increase their reproductive success. This behavior, however, is risky, because by leading their chicks away from the nest toward the lake, the chicks may be attacked and killed by parents of other families (Kosztolányi and Székely, personal observation). Some parents take arduous journeys with their freshly hatched young toward the lakeshore that negatively affect the survival prospects of the chicks (Erdogan, Székely and Kosztolányi, unpublished data). Mostly, inexperienced parents choose these long journeys; whereas the old ones breed close to lakeshore, and so, their family needs to only cover a short distance (Erdogan, Székely, and Kosztolányi, unpublished data).

Plover chicks may survive better on the shore because the probability of starvation is lower. Thermoregulatory ability of precocial chicks increases with body mass (Visser and Ricklefs 1993); therefore, well-fed chicks are more likely to cope with extreme ambient temperatures and evade attacking predators. In line with our results, piping plover Charadrius melodus chicks grow faster and survive better in habitats where food availability is higher (Loegering and Fraser 1995). Furthermore, the predator-detecting ability of the parents may be higher on the lakeshore because the vegetation is sparse, so the parents can warn their chicks earlier of any danger. Consistent with the latter argument, incubating Kentish plovers in open sites respond earlier to approaching predators than in concealed nest sites (Amat and Masero 2004).

Our results are consistent with the notion that size and mass of female parents influenced the movements of their brood. Thus, large, and presumably more competitive, females were able to take their chicks to the shore and defend them from the neighboring parents (Kosztolányi et al. 2006). We need further research to establish how body size and mass relate to fighting behavior and, ultimately, competitive ability. In this paper, we only note that parents spend substantial time on fighting, especially on the shore (Kosztolányi et al. 2006). Many of these fights are vicious and involve grappling with opponents; thus, size and body reserves appear to influence their outcomes.

High food abundance is suspected to facilitate female desertion and the evolution of polyandry (Andersson 2005). However, in the Kentish plover, the opposite relationship is apparent: high food availability facilitated biparental care (Kosztolányi et al. 2006). In this study, we advanced this work by showing that females staying with their chicks on the shore are larger and presumably more competitive. Thus, it seems that food availability selects for biparental care, and those females that stay with their broods are more competitive. These results suggest that circumstances that select for female desertion are different from the ones that evoke female–female competition and sex role reversal (Erckmann 1983; Oring 1986; Andersson 2005).

In conclusion, Kentish plovers can increase their current reproductive output by leading their chicks to lakeshore. However, all parents are not equally likely to move to the shore because large females spend more time on the shore with their brood than small ones. Thus, the competitive ability of females appears to influence the choice of brood-rearing habitat. Experimental manipulations, for example by restricting the movement of families, are needed to separate the effect of food abundance and female quality on reproductive success.


The project was funded by grants from the NERC (GR3/10957, to A. I. Houston, ICC, and J. M. McNamara), BBSRC (BBS/B/05788 to TS, ICC, A. I. Houston, and J. M. McNamara), Hungarian Scientific Research Fund (T043390, to TS), by the Eötvös Scholarship of the Hungarian State to AK and by the Royal Society/NATO Postdoctoral Fellowship to AK. Permission for the field work was provided by the Turkish Ministry of National Parks, Tuzla Municipality, and Mr. E. Karakaya, the Governor of Karatas District.

Copyright information

© Springer-Verlag and ISPA 2007