To date, in Flanders, the Common Woodpigeon breeding season is assumed to last from March to September. Recent preliminary surveys, however, indicate a later onset (Huysentruyt & Casaer 2009). To determine the start of the Woodpigeon breeding season in Flanders, we collected adult woodpigeons that were shot by hunters around the alleged start of the breeding season and examined their reproductive systems for indications of breeding activity.

Methods

Timing of breeding in Woodpigeons depends on light cycle, food availability, spring temperature and nest availability (Murton & Westwood 1977), on average starting much later in rural than in urban environments (Cramp 1985). We limited our scope to rural Flanders, where two distinct geographic regions were chosen for sampling (representing extremes on the axis from open farmland to a more closed region with woodland) to anticipate possible differences in start of the breeding season throughout Flanders (Figure 1) due to habitat differences.

To determine the onset of breeding, we examined the gonadal development of pigeons. Since Woodpigeons under the age of one year are not yet reproducing, only data from adult woodpigeons were analysed (Cramp 1985). We collected pigeons shot by 20 hunters in six weeks between 14 February and 1 May 2011 (14-20/2, 28/2-6/3, 14-20/3, 28/3-3/4, 11-17/4 and 25/4-1/5/2011). We aimed to collect 20 pigeons per region per week. Hunters were asked to label each pigeon with shooting date and location. We determined age class (2^nd calendar year or adult) based on primary moult pattern and sex.

Determination of the breeding season by gonadal development is well documented for pigeons (Murton et al. 1963, Lofts et al. 1966, Guttiérrez et al. 1975, Frith et al. 1976, Murton & Westwood 1977). In male pigeons, Lofts et al. (1966) used testis volume as a metric variable, calculated as v = 4/3π×(length/2)× (width/2)². In this study, testes length and width were measured to the nearest 0.01 mm using Mitutoyo CD-20DC digital callipers. A linear regression between testis volume and testis weight showed that weight was a good proxy for testis size (P < 0.001; R²= 0.88, n = 103). To reduce measurement errors we used weight as the metric variable and only weight of the largest testis was used. Testes were weighed to the nearest 0.01 g using a Sartorius TE612 digital scale. In females, the diameter of the largest follicle was measured to the nearest 0.01 mm (Guttiérrez et al. 1975). When eggs were present in the oviduct the follicle diameter was classified as 20 mm, i.e. approximately the maximal follicle size. All Woodpigeons with ovaries larger than 5.5 mm were classified as sexually active (Lofts et al. 1966, Guttiérez et al. 1975). For each of the sample weeks we described the ratio of sexually active versus non-active female Woodpigeons. For the analysis of the metrical variables, we used generalized linear models (GLM) with Poisson distribution in both sexes (McCullagh & Nelder 1989). In these models we tested the hypothesis that regional differences played no role in the start of the breeding season of Woodpigeon in Flanders. To do this, for both sexes a model which included region, sample week and their interaction term was compared to a model without the interaction term and a model which included only week. Models were compared using a likelihood ratio test.

Figure 1.

Location of both study regions in Flanders with woodland indicated. Region #1 is located mainly in farmland, region #2 is located in woodland. Black lines indicate province boundaries.

Results

A total of 272 woodpigeons were collected out of which 185 were adults (104 males, 81 females). Forty-nine were collected from farmland, 136 were shot in woodland. The gonads of one male Woodpigeon from wood-land could not be analysed. For region 1 in week 1 no pigeons were shot and in week 4 no female pigeons were collected. This way, the interaction term in the model for male pigeons only compared five weeks in both regions (df = 4) and in the model for female pigeons, the interaction term only compared four weeks (df = 3).

Table 1.

Comparison of generalized linear models to test effects of region, week and interactions on testis weight and follicle size.

For testis weight, the best model comprised both week and region as independent factors (the interaction term was not retained in the final model, Table 1). Testis size started to increase from the first week of March onwards (Figure 2). Woodpigeons in farmland had a slight advance compared to those in the more wooded parts of Flanders. In both regions, average testis sizes were maximal in late April.

The best model for follicle size in adult female Woodpigeons was the model containing only week as factor (Table 1). Before late March/early April only four female Woodpigeons in which follicle size surpassed 5.5 mm, indicating sexual activity, were found. In the first sample week 0 out of 2 Woodpigeons had reached this state, in week 2 this ratio was 2/15, in week 3 2/17. The boxplots presented in Figure 3 show that these follicle size were outliers within the follicles examined for those weeks. From late March onwards, a drastic increase in the proportion of reproductive female Woodpigeons was observed (7/11 in week 4, 8/15 in week 5 and 8/21 in week 6). From this period on, the boxplots in Figure 3 also show a difference in follicle size between the first and last three weeks of our study.

Figure 2.

Mean maximal testis weight (g) in male adult woodpigeons in farmland and woodland (+/- SE).

Figure 3.

Boxplots of maximal follicle size (mm) in adult female woodpigeons. Horizontal lines represent median, box limits represent the 25th and 75th percentile, the interquartile range (IQR). Whiskers are drawn to the nearest value not beyond 1.5×IQR. Any points beyond this value (outliers) are drawn individually. The reference line at 5.5 mm indicates sexual activity.

Discussion

The analysis of follicle sizes shows that the first female Woodpigeons in Flanders become sexually active by late February/early March, followed by the majority of the females a month later. Testis size started to increase earlier, namely in early March. This difference in onset of sexual activity between males and females is well known in Woodpigeons (Lofts et al. 1966). In the UK, Lofts et al. (1966) observed that active testes were found from March to September but breeding only took place when environmental conditions were favourable and couples spent sufficient time together in their breeding territories. Courtship behaviour is considered to trigger the physical development that allows ovulation (Murton et al. 1963, Lofts et al. 1966). Our own results indicate that male Woodpigeons in Flanders prepare for breeding from early March on, followed slightly later by females. The first phase of follicle growth from 1 to 5 mm, in Flanders in March, is influenced by gonadotropine, the luteinising hormone LH and the follicle stimulating hormone FSH (Murton & Westwood 1977). A second phase in follicle development starts during pair formation, when courtship and nesting behaviour have a positive feedback on the emission of FSH en LH (Murton & Westwood 1977). Only when the ovaries produce oestrogen and follicles develop further, egg laying is possible (Murton & Westwood 1977).

The start of the breeding season is often studied using nesting observations and the moment egg laying starts. The study of gonads provides a useful addition to these studies since it directly measures the physiological capacity of the species for breeding onset. Various studies from the UK describe that egg laying between 1955–1998 in Woodpigeons in rural and suburban conditions only starts from May/June on and peaks around July/August (Inglis et al. 1994, Slater 2001). Additional observations from Denmark and the Netherlands from the late 1970s and late 1980s also describe the first eggs laid early April, with laying numbers culminating for a first time by early May (Bijlsma 1980, Søndergaard 1996). A start of sexual activity by late March/early April, as found in this study, is similar to the timing in southern England in the 1960s (Lofts et al. 1967a,b), where woodpigeons started to develop their gonads from late March onwards. Behavioural studies indicate that by late March, pair formation has taken place and breeding territories are occupied, although egg laying has not yet started (Murton & Isaacson 1962, Murton et al. 1974, Schnock 1981). Observations from Belgium in the 1980s also describe the first manifestations of sexual activity to be often observed starting late March/early April (Schnock 1981). The lag of time between the onset of sexual activity and egg laying is probably a response to inadequate feeding conditions earlier in the season (Murton & Isaacson 1962, Murton et al. 1974). All observations confirm the importance of photo-periodicity in gonadal development in Woodpigeon (Lofts et al. 1967a). The gonads of Woodpigeons are stimulated into breeding condition at a daily photo-period-equivalent to that of late March in southern England (Lofts et al. 1967a,b, Murton et al. 1974). Although feeding conditions between Flanders and southern England may vary significantly, daylength does not. Therefore, our expectation is that, independent of feeding conditions, onset of sexual activity will not vary much between years and will start throughout Flanders on average in late March/early April. Subsequently, a delay in egg laying can be the result of the local inadequate feeding conditions.