The timing of clutch initiation in migratory birds is determined by many factors and affects fitness. In polygynous species the date of egg laying may also be determined by mating status. In order to identify the habitat factors enabling Eurasian bittern females to breed early, a study was carried out on nine fishpond complexes in eastern Poland during 2003–2009. The habitat characteristics of the nesting sites, including fish stock type and the breeding success were recorded for 48 monogamous, 31 primary, 30 second and 12 third or fourth females. The analysis showed that early nesting females preferred a higher density of old reed stems and chose less attractive ponds with poorer food resources. Monogamous and primary females laid first eggs eight days earlier on average and had higher breeding success than the second, third and fourth females. The relationship between the fate of broods and the breeding success of neighbouring nests was found for second females. Our results suggest that Eurasian bittern females benefit from early nesting, provided they can find appropriate vegetation at the nesting site.
The timing of breeding can strongly affect fitness of many bird species (Lack 1968). Generally, early breeders enjoy a higher reproductive success than later ones (Price et al. 1988). The benefits and potential costs of early breeding may be especially pronounced in single-brooded birds (Svensson 1995). If birds start breeding too early, they run the risk of the eggs hatching before the food supply is optimal. In addition, laying eggs under harsh conditions early in the season increases energy costs (Perrins 1970). On the other hand, nest predation is another strong selection pressure forcing the early timing of reproduction, since re-nesting is then possible if the first clutch has failed (Svensson 1995). Moreover, many external factors, such as climate, territory quality and mating system, as well as internal ones like age, body condition and genetic differences among individuals, determine the date of egg laying (e.g. Sheldon et al. 2003, Gonzalez-Solis et al. 2004, Elmberg et al. 2005, Casagrande et al. 2006, Rubolini et al. 2007, Dobson et al. 2008).
The nesting of migratory birds in high latitudes should be relatively synchronous and peak when food conditions are optimal for chick rearing (Lack 1968). However, females generally turn up in the nesting areas in temporal waves after the males. Timing differences in the arrivals of females influence the start of nesting and may determine their status within a male territory (Orians 1969). Females sharing one male's territory often compete with each other for male parental care or resources, and such intrasexual competition may adversely affect reproductive output (Verner & Willson 1966, Lubjuhn et al. 2000, Moreno et al. 2002). Selective pressure on lower mating status females is particularly noticeable in species with biparental care of the young (Bensch 1996, Searcy & Yasukawa 1996, Huk & Winkel 2006).
In the Eurasian bittern Botaurus stellaris, a polygynous species, parental care on the part of the male is occasional, if it occurs at all (Voisin 1991, Klejdus 2007). The long-range vocalisation of the male is thought to serve as an advertising call for territorial defence and mate attraction (White et al. 2006). Booming season length is variable among sites, years and individuals (Gilbert et al. 1994, Puglisi et al. 2003) and could be an important indicator for females in mate choice (White et al. 2006). Consequently, the males best able to signal their presence to females may attract more mates (see Dale & Slagsvold 1990). However, no correlation has been found between booming parameters, and components of breeding success and harem size (Polak 2006).
In different European populations of the Eurasian bittern, from one to five female nests have been found within male territories (Puglisi & Bretagnolle 2005). On the fishponds of eastern Poland, in contrast to other habitats, clumped nesting is a common strategy and more than half the females nest in harems (Kasprzykowski & Polak 2013). In single nests (monogamous females), the first eggs were laid earlier than in the clustered nests (polygynous females), and the first-egg laying period was longer in nest clusters than in single nests (Kasprzykowski & Polak 2013). However, there were no differences in other reproductive parameters between these two categories of female. Owing to the severe winter in eastern Poland, first-egg laying is somewhat delayed and better synchronised than in west European populations; it is also correlated with the start of the booming period (Polak & Kasprzykowski 2010). Other factors favouring early nesting have not yet been analysed, however.
The aims of this study are to identify the factors allowing a female to breed early and to examine the effects of early nesting. One of our hypotheses is that the first-egg laying date depends on habitat quality. We expect that some habitat parameters could accelerate the onset of breeding. Because nesting time determines mating status of females in the polygynous system, we also include this aspect in the analysis. The second hypothesis is that the success of nests is linked with female category and the fate of neighbour females. We expect that late secondary females have lower success corresponding to the success of neighbouring nests.
Material and Methods
Study area and field procedures
The study was carried out in eastern Poland on nine fishpond complexes stocked with common carp Cyprinus carpio. The ponds varied in area from 15 to 203 ha; their combined area was 815 ha. Most ponds were extensively managed and partially covered by discrete stands of vegetation dominated by common reed Phragmites australis, lesser bulrush Typha angustifolia and sedges Carex sp. The mean density of the bittern breeding populations on these ponds ranged between 4.6 ind./100 ha for males and 5.2 ind./100 ha for females (Polak & Kasprzykowski 2009). The study areas were surveyed from the beginning of March to early July from 2003 to 2009. Nests were located by systematic walking through the emergent vegetation, not only in the booming males' territories. Special efforts were made to find nest clusters, and the surroundings of every nest found were searched. In a large reedbed, this was an area within a radius of 100 m from the nest. The positions of all the nests found were recorded on a 1:5000 scale sketch map of the area, and the coordinates of all the nests were obtained by GPS. During the chick-rearing period five inspections of the nest contents were usually made within the first 20 days of hatching. In the majority of nests the first-egg date was established by backdating, assuming 25–26 days for incubation (Mallord et al. 2000, Demongin et al. 2007, Longoni et al. 2007). To eliminate the year effect, the data were standardised against the median of the first egg date in the population in each year. Because two-week-old bitterns can disperse from the nest into the dense vegetation, nests were defined after Puglisi & Bretagnolle (2005) as being successful when at least one young bird had survived up to 15 days old. Females were assigned to four categories: monogamous (single female without cluster effect), primary (the first females hatching and feeding nestlings in cluster), second (females hatched later than primary) and third/fourth (females hatched later than the second ones) (see Sejberg et al. 2000, Huk & Winkel 2006, Kasprzykowski & Polak 2013). We collected data on the breeding parameters of 48 monogamous, 31 primary, 30 second and 12 third/ fourth females. The distances between the nearest active nests were measured using a tape or GPS and the nests closest each other were treated as neighbouring. Five habitat parameters which may govern the nesting of females on a specific semi-natural habitat like fish ponds were recorded (Polak 2007, Polak et al. 2008, Table 1). We used the methodology applied in studies of the Eurasian bittern in the UK (Tyler et al. 1998, Gilbert et al. 2005) with some modifications (Kasprzykowski & Polak 2013). The microhabitat variables, i.e. water depth and the number of dry and green reed stems, were measured in 126 quadrats (50 × 50 cm) placed around the nests. The macrohabitat parameters, i.e. distance to open water and distance to the fishpond dyke were measured from aerial photographs (GEOPORTAL, www.geoportal.gov.pl). All the measurements were made from the end of April to the end of May. Due to seasonal variation in water level, water depth under nests was recorded during the first visit at the beginning of breeding period. In a specific habitat type like fishpond complexes, an important factor related to food resources is the common carp year-class (Kasprzykowski & Polak 2013). Therefore, the fish stock in the separate ponds with active nests was used as a proxy of food abundance. After Kloskowski (2009), we divided common carp into three year-classes: young-of-year carp with individual weights of 1.5–3.0 mg, to reach 8–15 g in early July; medium carp after first wintering at 30–50 g, 2-year-old-carp at 150– 240 g and wild fish ponds, with different species and different age categories.
Mean, standard deviation (SD) and the range of habitat variables measured at Eurasian bittern nests (n = 126).
General Linear Model test of the influence of environmental parameters on the standardised first-egg laying date of Eurasian bittern females.
General Linear Model (GLM) with identity link function and normal error distribution was used to test first hypothesis stated in the introduction (Crawley 1993). The standardised first-egg laying date was treated as a dependent variable. Five continuous habitat parameters and four categories of fish stock were used as independent variables. Multi-co-linearity of variables was checked and correlations between them were < 0.6. To verify the second hypothesis, Generalized Linear Model (GLZ) with logit link function and binomial distribution was used (McCullagh & Neider 1989). The fate of brood (1 — success, 0 — loss) was dependent variable in the model and female mating status and success of neighbouring nest were the predictors. In addition, ANOVA was used to compare first-egg laying date in four categories of female mating status. Statistical processing was performed using the STATISTICA 10.0 package (StatSoft Inc. 2011).
Generalized Linear Model test of the influence of female mating status, the fate of neighbour nest and their interaction on success of Eurasian bittern females.
The first analysis showed that the first-egg laying date was negatively related with the number of old reed stems (Table 2). The other four habitat parameters: water depth, the number of green reed stems, the distance to open water and the distance to fishpond dyke did not have a significant effect. However, ponds with young-of-year carp negatively affected the onset of reproduction in female bitterns. The other three categories of fish stock: medium carp, 2-year-old-carp and wild fish ponds were not related with the first-egg laying date. In nests situated on ponds with the youngest fish, the first eggs were laid five days earlier on average than in the other ones (Fig. 1).
Accounting for female mating status showed that monogamous and primary females laid first eggs eight days earlier on average than the second, third and fourth females (ANOVA F3,118= 8.60, P < 0.001, Fig. 2). Moreover, the breeding success in monogamous and primary nests was higher (70.8 %, 71.0 % respectively) than in those of second (53.3 %) and third/fourth females (58.3 %). These differences between female categories were supported by GLZ analysis (Table 3). The success of nest was found to be associated with the breeding success of neighbouring nests. However, the interaction between both predictors was significant and positive only for second nest category but no such relationship was recorded for monogamous and primary females.
Although the Eurasian bittern inhabits a wide spectrum of habitats, reedbeds are still the most important nesting sites (White et al. 2006). Evolutionary adaptation to breeding in this kind of habitat is manifested by the perfect camouflage of the female plumage. Therefore, old (dry) reed stems play an important role in choosing nesting sites (Polak et al. 2008). The present study shows that this habitat parameter is also associated with early breeding. In the second part of April, when Eurasian bittern females begin nesting, water bodies are devoid of green emergent vegetation and, crucially, have just the right density of dry stems to mask the female on the nest. Water depth and other macrohabitat parameters appear to have a lesser effect on the timing of breeding. In turn, the richness of food resources, analysed as a qualitative factor of fish stock, affected the females' timing of breeding. We found a negative relation between the first-egg laying date and young-of-year common carp. Ponds with fish of this age class are unattractive compared to the one-year class (medium) and were sparsely populated by clustering females, because of the very small weight of the fish (Kasprzykowski & Polak 2013). On the other hand, one quarter of the monogamous females nested on this category of pond. This negative preference could be explained by the fact that solitary females belong to the group of early breeders, together with primary females. The late-nesting group included females of lower mating status in the polygynous system. Second, third and fourth females laid their first eggs relatively simultaneously and significantly later after the monogamous and primary females. These results suggest that Eurasian bittern females start breeding in two waves. When choosing nesting sites, early females are guided mostly by the densities of old stems and settle in less attractive habitats. In contrast, late females build nests on ponds offering a more suitable category of fish stock. Some late females may re-lay after a failed first clutch. Replacement clutches seem to be a fairly common behaviour in the Eurasian bittern, especially when nest failure occurs before the end of May (Dmitrenok et al. 2005).
Avoidance of aggression could explain the significant differences in the first-egg laying date between primary females and second, third and fourth ones that settled at the same time in a male territory (Slagsvold & Lifjeld 1994). Secondary females should benefit from the reduction in breeding overlap with primary females because resident females are usually the most aggressive during the early stages of breeding. Generally, aggression towards a female intruder is a reaction to the theft of materials, the displacement of the resident female from the nesting site and mate, the destruction of eggs or young, nest parasitism, increased nest predation as well as reduced investment in nest guarding by the male parent (Slagsvold & Lifjeld 1994). However, during our research we did not observe any aggression between females.
Our study shows that early nesting monogamous and primary females have higher breeding success. The greater total losses in polygynous bird species are most plausibly explained by the males' withholding paternal care from secondary broods and the frequently suboptimal food conditions available to late-breeding secondary females (Lubjuhn et al. 2000). These two factors may be of much less importance to a bittern population nesting on fishponds. Extensive fish management and abundant food resources were most likely responsible for the good breeding parameters obtained for this population and favoured the polygynous system of reproduction (Polak & Kasprzykowski 2010, Kasprzykowski & Polak 2013). Moreover, because bittern females do not normally receive help from males in the care of the young (Puglisi & Bretagnolle 2005), intrasexual competition appears to be reduced.
We found that the fate of secondary broods was linked with the breeding success of neighbouring nests, but no such relationship was recorded for primary nests. These results suggest increasing predation pressure during the lengthening breeding period when secondary females are nesting in a cluster. Another study showed that nearest-neighbour nests of waterfowl tended to share the same fate, suggesting that cluster nests either survived or were depredated as a group (Ringelman et al. 2012). Therefore, early and solitary (before cluster formation) bittern females seem to be less exposed to failure resulting from predation. For cryptically coloured birds such as Eurasian bittern females, nest dispersion has been proposed as an important mechanism for counteracting density-dependent predation (Tinbergen et al. 1967, Andren 1991). However, studies on waterfowl report different levels of relationship between nest density and predation, which are associated with different factors like nest type (artificial vs. natural), predator community or spatial scale (Ackerman et al. 2004, Gunnarsson & Elmberg 2008, Ringelman et al. 2012). To conclude, our study shows that early nesting females are more closely associated with appropriate vegetation at the nesting site than late nesters. On the other hand, early females choose less attractive habitats with poorer food resources. Later second, third and fourth females have lower breeding success, which shows the benefits of early nesting.
We would like to thank Janusz Kloskowski, Jarosław Krogulec, Marek Nieoczym, Ewa Zębek, Rafał Kuropieska and Marta Szaniawska for their help with the fieldwork and Peter Senn for correcting the English. We also thank the two anonymous referees for their helpful comments on the manuscript.