Predation is a major cause of nest failure for many birds and hence it is an important source of natural selection that shapes avian behaviour and life-history traits (Martin 1995). Nest predation has been suggested to be the primary cause of losses across a wide range of bird species, habitats, and geographic regions, accounting on average for 80% of nest losses (Ricklefs 1969, Martin 1993). In particular, breeding populations of some ground-nesting birds of open habitats, such as waders and galliformes, are highly vulnerable to a wide range of mammalian and avian predators, sometimes to the point of affecting population size (Marcström et al. 1988, Tapper et al. 1996, Grant et al. 1999). By contrast, the evidence that numbers of breeding passerines are limited directly by nest predation is weak (Gibbons et al. 2007). Population declines of many ground-nesting passerines of open habitats in Western and Central Europe are primarily caused by loss and degradation of breeding habitat (Donald 2004, Grzybek et al. 2008, Menz et al. 2009). However, interactions between habitat changes and nest predation rates frequently occur and can obscure the primary causes of bird populations declines (Evans 2004).

Nest predators differ among habitats and species, but some general patterns exist. Red Foxes Vulpes vulpes, martens and corvids are among the important nest predators in European woodland (Jędrzejewska & Jędrzejewski 1998, Weidinger 2009, Mallord et al. 2012) and in farmland (MacDonald & Bolton 2008, Teunissen et al. 2008, Praus & Weidinger 2010). Predation studies in intensive farmland have focused on nest success and nest predator identity of ground-breeding farmland birds (Teunissen et al. 2008, Morris & Gilroy 2008, Praus & Weidinger 2010). Some studies have also investigated the nesting success of these species in more natural areas like coastal, steppe and heathland habitats (Aunins et al. 2001, Pearce-Higgins & Grant 2002, Wright et al. 2009). However, direct evidence of the identity of nest predators in natural habitats is anecdotal and detailed studies are lacking.

The Skylark Alauda arvensis and the Woodlark Lullula arborea are ground-nesting species of open landscapes that face high predation rates (Donald 2004, Dolman 2010). The Skylark was originally widespread in a variety of open habitats including steppes, natural grasslands, heathlands and saltmarshes, but nowadays the majority of European birds breed in agricultural landscapes, where Skylarks are rapidly declining (EBCC 2012). The Woodlark inhabits semi-natural habitats, especially restocked conifer plantations and lowland heathlands across Europe (Sitters et al. 1996, Wotton & Gillings 2000, Mallord et al. 2007). Various predators and variable nest predation rates have been identified for Skylark nests in West and Central European farmland (Morris & Gilroy 2008, Praus & Weidinger 2010) and Woodlark nests in lowland heathlands of Great Britain (Dolman 2010). However, information about predation pressure and relative importance of nest predators in different habitats and across different geographic regions is scarce (Delius 1965, Yanes & Suarez 1995).

In this study, we quantified nest success and identified nest predators in a rather high-density population of Skylarks and Woodlarks breeding in a semi-natural heath- and grassland area in The Netherlands. Populations of both species co-occur in this area, but the two species differ in the selection of their nest sites; while Skylarks avoid forest edges and even the proximity of single trees, Woodlarks usually nest close to trees and/ or forest edges. Thus our study tentatively allowed us to explore whether different nest site selection within an area relates to species-specific vulnerability to nest predation.

METHODS

We monitored fates and identified predators of lark nests in the Aekingerzand, part of the National Park Drents-Friese Wold in the northern Netherlands (52°55′N, 6°18′E) in 2012. The study area (c. 400 ha) is characterised by nutrient-poor soil. Dominating vegetation types are heather, Calluna vulgaris and Erica tetralix, and different succession states of grass, moss and Juncus spec. Furthermore, patches of open sand and groups of trees are spread through the area which is surrounded by coniferous forest. Suitable nesting habitat covers approximately 240 ha for Skylarks and 220 ha for Woodlarks. Population densities are rather high compared to most modern agricultural areas of Western Europe and the local Skylark (80–100 pairs) and Woodlark (60–80 pairs) populations have been intensively studied since 2006 (Hegemann et al. 2010, Hegemann & Voesten 2011). About 500 sheep graze the study area year-round, keeping the vegetation short and succession limited. The National Park is surrounded by intensive farmland with maize, potatoes and cereals as the main crop types, and to a lesser extent by intensive grasslands (Geiger et al. 2014).

Fieldwork was conducted from early May to late July; roughly 80 days were devoted to nest searching in the whole study area. This period covers the entire breeding season of the local Skylark population (Hegemann et al. 2012, 2013) and the period of second and third broods in Woodlarks (Tieleman et al. unpublished data). Although we missed the first Woodlark broods, this does not pose a major problem, because one aim of our study is to compare predation rates and predators between the two species breeding in the same area and at the same time, rather than to obtain season-long estimates of breeding performance. Nests were found either through direct observation of adults returning to the nest or by systematic searches of spots where intense mating behaviour had been observed. Age of nestlings was estimated by visual clues from their development (e.g., opening of eyes, feather development); first-egg laying date was back-calculated from brood size and hatch date, or from clutch size and the stage of incubation estimated by egg floatation.

We found a total of 58 Skylark and 40 Woodlark nests. A randomly selected subset of these nests (Skylark: n = 37 nests, 247 active nest-days; Woodlark: n = 16, 92) were monitored by continuous video recording. The remaining nests served as control nests without video monitoring. Video systems consisted of a small camouflaged video camera with IR diodes, a portable security digital video recorder (DVR), and a 12V/40Ah deep cycle battery (for details see Praus & Weidinger 2010). We used local natural material (dry moss, stones) to mask the camera; all other parts (DVR, battery, cables) were buried under the ground (Figure 1). We set the DVR to record continually with a frequency of 10 fps at 640 × 480 pixel resolution. These settings allowed for 4.5 days of recording on a 16 GB memory card. Cameras were deployed about 2 days (range 0–4) after nest discovery. We visited the nests usually (80% cases) every fourth day (mean 3.4, range 1–5) to check nest content and to change the battery and memory card; we visited nests with cameras as well as control nests according to the same time schedule (mean 3.0, range 1–6). When a nest was found empty, we searched in the immediate vicinity for signs of nest failure (eggshells, feathers, dead nestlings) or nest success (alive fledglings, droppings).

Figure 1.

The typical layout of a video camera (arrow) at one of the monitored nests, here with incubating Woodlark female, 18 May 2012 (Photo Libor Praus).

For video monitored nests we determined the exact survival time, nest fate and the species of nest predator by inspection of video recordings. Survival time of failed nests without a video camera was estimated as a midpoint between the last visit to an active nest (eggs or nestlings) and the first negative visit (empty nest, dead chicks or deserted clutch). Survival time of successful nests without a video camera was terminated by the 8^th day of chick age for both Lark species (Skylark mean fledging age 8.6 days, range 7–12 days, n = 19 video-monitored nests; Woodlark 9.3 days, range 9–10, n = 10); disappearance of younger chicks was considered as predation. The observed mean age of fledging might have been influenced by the disturbance associated with ringing; chicks left the nest within one day after ringing. However, nestlings also regularly fledge at an age of 7 or 8 days, without ringing or other research activities at the nest, especially if ambient temperature is high and nestlings need to escape the direct sunlight (Hegemann et al., unpublished data). Exposure period for all nests was measured in nest-days since the day of discovery of an active nest or since camera deployment (see below). We treated each nest-day as an independent binary observation (survived or failed) and estimated the daily survival rate (DSR) as a simple ratio of survived to exposed nest-days. We calculated DSR separately for the egg (laying and incubation) and nestling stage. To calculate DSR we considered either nest losses owing to predation, or total mortality. Nest survival was estimated as DSR^t where t = 14 (egg stage including laying period), 8 (nestling stage) or 22 (total) days for Skylarks and where t = 15 (egg stage including laying period), 9 (nestling stage) or 24 (total) days for Woodlarks. Egg stage is based on a mean clutch size of 4 eggs for both species and an incubation period of 11 days for Skylarks (Glutz von Blotzheim & Bauer 1985, own unpublished data) and 12 days for Woodlarks (Nick Horrocks & Stef Waasdorp, pers. comm.) starting with the last egg laid. Duration of nestling stage is based on the video-recordings (see above). The limited sample size precluded formal statistical analyses, therefore we present only descriptive data. To check for an effect of research disturbance on the risk of nest predation, we compared daily predation rates between samples of nest-days with and without video camera, and between the first day after an observer's visit and subsequent days. Results of these comparisons are presented as odds ratios with 95% confidence limits.

Table 1.

Daily survival rates (DSR) of Skylark and Woodlark nests in a semi-natural heath- and grassland in the northern Netherlands during the breeding season of 2012.

Table 2.

Summary of video-recorded Skylark and Woodlark nest depredation events during the 2012 breeding season in a semi-natural heath- and grassland in The Netherlands.

RESULTS

DISCUSSION

The overall nest success of Skylarks in a semi-natural breeding habitat in the northern Netherlands was 33% and thus within the range of values reported (also using the ‘Mayfield method’) from arable fields in the UK (23–40%; Wilson et al. 1997, Chamberlain & Crick 1999, Donald et al. 2002), The Netherlands (27%; Kragten et al. 2008), Germany (22%; Jeromin 2002), Switzerland (22%; Weibel 1999) and Czech Republic (17%; Praus & Weidinger 2010). In our study area, Woodlarks had a slightly lower nest success (22%) than Skylarks. The nest success rates we found are also lower than those reported for Woodlarks breeding on heathlands in southern England (47%; Mallord et al. 2007). Predation accounted for 76% (13/17) and 92% (12/13) of nest losses in Skylarks and Woodlarks, respectively. These values are in accordance with data from Skylark populations breeding on farmland (generally >70%; Weibel 1999, Donald et al. 2002, Praus & Weidinger 2010) and in coastal dunes (>85%; Delius 1965). The second most important cause of nest failure in this study was nestling death. All cases of nestling death in Skylarks (3/53) and Woodlarks (1/34; Table 1) were likely a consequence of predation on parents. Video recordings of one Skylark nest showed that the male kept on bringing food to the 3 days old nestlings after the female had disappeared, but without brooding by the female the nestlings died of hypothermia. We also video-recorded an incubating Woodlark female being taken at night by a fox. The fox did not take the clutch but these remained unattended in the nest for four hours until taken by a Carrion Crow.

No nest with nestlings was deserted after we installed the camera equipment and only one female Skylark did not resume incubation after camera deployment. We are reasonably confident that neither the presence of a camera at the nest, nor the repeated visits associated with the video recording, increased nest predation rate. In fact, there was a non-significant trend in the opposite direction. In both species predation was marginally lower on nests with a camera. Furthermore, in Skylarks it was also marginally lower on the first day after an observer visit to the nest compared to subsequent days. In both cases confidence intervals are wide and do not allow any firm conclusions. Yet, our data are in line with the current opinion that nest cameras do not increase nest predation (Ribic et al. 2012).

Combining both lark species, nest predation occurred during darkness in 55% cases, despite the relatively longer daylight periods during the breeding season (68% of recording time). This pattern was mainly attributed to a predominance of predation by Red Foxes which mainly forage at night (Doncaster & MacDonald 1997). Red Foxes accounted for 55% (6/11) of all video recorded predation events. Foxes are known to be the most important predators of wader nests across The Netherlands (Teunissen et al. 2008). Nevertheless, bird nests might represent only an alternative prey for foxes, whose foraging behaviour and/or abundance is rather affected by their primary preypopulations of small rodents (Marcström et al. 1988, Tomkovich & Zharikov 1998, Kjellander & Nordström 2003). If so, foxes may actively switch to bird nests when the primary prey is scarce. Or alternatively, incidental predation on bird nests may increase as a side effect of increased abundance of the primary prey (Yanes & Suarez 1996). In 2012 Common Vole Microtus arvalis and Wood Mouse Apodemus sylvaticus, two of the main prey species of Red Foxes (Lloyd 1980), showed relatively high population abundances in Drents-Friese Wold, where our study area is located (Bijlsma 2013).

The second most important nest predators were corvids, accounting for more predation losses in Woodlarks (3/4) than in Skylarks (1/7). This is not unexpected because corvids probably use trees as observation posts, and consequently prey more on Woodlark nests, which are located generally closer to trees than Skylark nests (Table 2). In spite of limited sample size, we documented one case of a European Adder Vipera berus preying on Skylark nestlings. This, together with evidence from a broadly similar habitat in southern England (16% of 7 nests; Dolman 2010) suggests that snakes might be locally important predators of lark nests in heathlands. Although the available data suggest that foxes were the main predators during the single study season, a higher sampling intensity over multiple seasons would likely reveal a wider range of nest predator species and may change the importance of different predator species in our study area (e.g., Dolman 2010, Praus & Weidinger 2010).

In summary, nest success of both lark species, intensity of predation and species composition of nest predators in Aekingerzand were similar to what is known from other regions and habitats. We realize that our sample size is limited and our data are mainly based on nests found in the nestling stage, which may influence the robustness of our conclusions. Additionally, we studied only one breeding season and might have missed yearly variation in predation rates. We therefore suggest sampling across years to help address these issues. Monitoring the abundance of predator species identified in this study and the abundance of their primary prey may further help to understand predation patterns. Finally, combining a study on nest predation with tracking of predators foraging behaviour would help to clarify whether foxes actively search for nests or find them incidentally while searching for their primary prey.