1. Top of page
  2. Abstract
  7. Acknowledgments

Diet, feeding behaviour and habitat selection of breeding Eurasian Woodcock Scolopax rusticola were studied by radiotracking birds from March to July in two contrasting situations: a 171-ha lowland plantation of Sycamore Acer pseudoplatanus, Beech Fagus sylvaticus and pine Pinus sylvestris/P. nigra in Derbyshire, central England, and an area of c. 900 ha of fragmented, naturally regenerating birch Betula pendula/B. pubescens woodland and hill margin in an upland glen in Angus, northeast Scotland. Earthworms were the most important diet component of adults and chicks in terms of biomass at both sites (50–80%), the rest comprising mainly spiders, harvestmen and beetles. In spring both sexes flew c. 1 km after dusk to feed on fields at night, with up to 94% of nocturnal radiolocations on fields in March, dropping to 18% in July. This behaviour probably reflected seasonal changes in the relative availability of earthworms in fields and woodland. Diurnal home range sizes were similar at both sites and the mean size of 30-day ranges was 62 ± 20 ha (± se), although Woodcock changed locations regularly and areas used for feeding on a daily basis were typically smaller than 1 ha. In the lowland plantation, Sycamore with Dog's Mercury Mercurialis perennis ground cover was highly used relative to availability. In the upland margin study area, the same held for dense sapling-stage birch. These habitats appear to represent a compromise between food availability and protection from avian predators. Recently documented changes in the structure of British woods during the last 30 years, suggestive of reduced management and increased grazing/browsing, are likely to have been detrimental to breeding Woodcock.

The Eurasian Woodcock Scolopax rusticola breeds widely throughout Britain, with notable absences only on the highest ground in parts of Scotland, and in southwest England and south Wales (Gibbons et al. 1993). Nevertheless, the species is currently ‘amber-listed’ as a bird of conservation concern because of an apparent long-term decline in breeding numbers (–76% during 1974–99) and range (–31% during 1968/72–1988/91, Gregory et al. 2002). The Woodcock is difficult to survey owing to its cryptic plumage, secretive behaviour and crepuscular displays, and hence the reliability of current population and trend estimates is questionable, but the magnitude of the estimated decline is sufficient to justify investigation of factors likely to affect Woodcock numbers.

Insufficient birds have been ringed and nests monitored each year to permit a demographic analysis for the Woodcock. The pragmatic approach is therefore to examine what constitutes optimal breeding habitat, to assess which elements of this might have changed in the last 30 years, and the factors responsible, and to determine a mechanism for restoring habitat at an appropriate scale. For example, annual ‘singing-ground’ monitoring of the closely related American Woodcock Scolopax minor has highlighted a long-term population decline which appears, at least in part, to be related to habitat loss and alteration on the breeding grounds (Dwyer et al. 1983, Sauer & Bortner 1991, Sepik et al. 1993).

The aim of this study was to compare home ranges, habitat selection and diet in relation to food availability in two contrasting landscapes representative of those used by breeding Woodcock in central and southern England and in parts of northern England and Scotland. All previous studies of breeding Woodcock have been undertaken in the lowlands (e.g. Hirons 1983, Hirons & Johnson 1987) and yet breeding birds are currently known to be more abundant in Scotland and the upland parts of northern England (Gibbons et al. 1993, Hoodless et al. 2006). Birch Betula pendula and B. pubescens woodland comprises a characteristic component of upland margins and one that appears to be widely used by breeding Woodcock, as judged by the presence of roding birds (A.N.H. pers. obs.). As such, its value to Woodcock as a breeding habitat merits study.

For habitat management to be successful, knowledge is needed about which woodland stand types are preferred and what structural features are important within stands, in addition to information on any landscape-scale factors influencing Woodcock distribution. Breeding Woodcock are known to favour large woods (> 80 ha) with sites of less than 10 ha rarely used (Fuller 1982). Deciduous woodland is thought to be preferred to conifers (Clausager 1972), although conifer plantations are used for nesting up to the thicket stage (Shorten 1974). In winter, when their diet consists predominantly of earthworms and fly larvae (Hirons & Bickford-Smith 1983, Granval 1987), Woodcock frequently fly to fields to feed at night, where they are able to benefit from a much higher biomass of earthworms than in woodland (Granval & Bouché 1993). Woodcock appear to employ this behaviour to enable them to satisfy their energy requirement in cold weather and when food availability in woodland is low (Duriez et al. 2005). Similarly, during the breeding season, habitat use is likely to be governed by the availability of suitable feeding areas which combine plentiful food and safety from predators. Hirons and Johnson (1987) found that young stands of trees with high earthworm availability and dense ground vegetation were highly utilized for feeding and brood rearing and areas with more open ground vegetation were used for nesting. We examine the scale at which habitat selection takes place, determine the extent of nocturnal foraging outside woodland and assess relative food availability in different habitats.


  1. Top of page
  2. Abstract
  7. Acknowledgments

Whitwell Wood, Derbyshire

Whitwell Wood (53°18′N, 1°13′W) is located 24 km southeast of Sheffield, at 140 m asl and is surrounded by predominantly arable agriculture. It consists of 171 ha of mainly deciduous plantations planted during the period 1932–47, and different stands (averaging approximately 15 ha) were thinned each winter between 1983/84 and 1991/92. The dominant tree species are Sycamore Acer pseudoplatanus (comprising 55% of the total area) and Beech Fagus sylvaticus (comprising 37%), sometimes planted in mixed stands, with smaller areas of Scots and Corsican Pine Pinus sylvestris and P. nigra (6%), Ash Fraxinus excelsior and birch Betula spp. Oak Quercus robur and Rowan Sorbus aucuparia are distributed irregularly throughout the wood.

The wood has a reduced shrub layer consisting of a few patches of Hazel Corylus avellana and some Holly Ilex aquifolium. Stands are typically dominated by one species of ground vegetation, generally Dog's Mercury Mercurialis perennis, bramble Rubus spp. or grasses in the Sycamore stands and Ramsons Allium ursinum, Bluebells Hyacinthoides nonscriptus or bramble in the Beech stands. Whitwell Wood has an extensive network of rides and tracks which facilitate access and radiotracking.

Millden, Angus

Millden is an upland estate of c. 50 km2 consisting predominantly of Heather Calluna vulgaris moorland bisected by the River North Esk running west to east through the centre. The valley bottom varies from 400 to 1000 m wide and consists of pasture fields supporting sheep and fragmented birch woodland. Fieldwork was focused on c. 9 km2 of woodland, marginal agriculture and dry heath below 300 m asl between Millden Lodge and Tarfside village (56°53′N, 2°47′W). The birch occurs in patches of irregular size and age, ranging from 5-year-old thickets to mature trees of at least 60 years old. There is no understorey and the predominant ground cover is Bracken Pteridium aquilinum or grasses. There was a 40-year Scots Pine plantation (11.6 ha) within the study area and small mixed stands (1–4 ha) of 60–70-year Scots Pine and Beech or Ash.


  1. Top of page
  2. Abstract
  7. Acknowledgments


Woodcock were caught at dawn and dusk in mist-nets set across rides and clearings, and in drop-traps with ‘lead-in fences’ in likely feeding areas. They were aged as first-year or adult birds according to Clausager (1973). The sexes of 85% of birds were determined from their subsequent behaviour (roding or nesting) and the remainder were sexed using a discriminant function based on morphological measures with a classification success of 94% (Hoodless 1994). In 1978 and 1980, birds were fitted with radio-transmitters weighing 5–8 g, and in 1986, 1992 and 1993 4-g transmitters (Biotrack Ltd, Wareham, Dorset, UK) were used. They were attached with figure-of-eight harnesses made from 5-mm elastic (Amlaner et al. 1978, Hirons & Owen 1982). Birds that yielded information for less than 10 days have been excluded and data are presented for 30 birds (25 males and five females) captured at Whitwell Wood during mid-March to May 1978, 1980, 1986 and 1992 and for ten birds (three males and seven females) captured at Millden in mid-March to mid-April 1992 and 1993. Woodcock were tracked from late March to July, for a mean of 24 ± 2 days (± se) following capture at Whitwell Wood and 41 ± 10  days (± se) at Millden.

Radiotagged birds were located with an AVM LA-12 receiver (AVM Instrument Company, Champaign, IL, USA) in 1978 and 1980 and subsequently with a Mariner Radar M57 receiver, both used in conjunction with a hand-held three-element Yagi antenna. Typical maximum detection ranges were 200–500 m at Whitwell Wood and 200–1000 m at Millden. Position fixes were determined by triangulation at distances of 30–40 m at Whitwell Wood and 50–150 m at Millden. Radiotagged birds could not be approached as closely at Millden owing to more open habitat and an increased risk of flushing the birds. The mean size of error triangles calculated from bearings taken on spare radiotags was < 0.01 ha at Whitwell Wood and 0.14 ± 0.04 (± se, n = 10) at Millden. Birds were located twice per day during daylight hours and position fixes were recorded using grid sizes of 10 × 10 m at Whitwell Wood and 50 × 50 m at Millden. On a few days at each study area, daily home ranges were estimated by obtaining 4–8 radiolocations per day to an accuracy of 10 × 10 m. At Whitwell Wood an attempt was made to locate each bird once per night within 30–120 min after the end of the evening roding period.

Feeding activity of radiotagged birds was detected by monitoring changes in received signal strength and attempts were made to pinpoint feeding locations at Whitwell Wood in May–June 1985 and 1992 and at Millden in May–June 1992. Birds were approached cautiously until visible through binoculars and prominent vegetation was marked, from which compass bearings, distances and descriptions were subsequently used to find the exact locations for habitat recording.

Habitat mapping and measurement

Habitats were recorded onto 1 : 5 280 (Whitwell Wood) or 1 : 10 000 (Millden) scale maps using ground surveys. Within woodland, the dominant species of trees and ground vegetation were recorded by walking to within 50 m of every point. Other habitats were recorded as Phase 1 categories (NCC 1990) using transects that took advantage of viewpoints. Data were entered into a GIS (MapInfo 7, MapInfo Corporation 2002).

At Whitwell Wood habitats were classified into five categories according to the dominant tree species and whether the ground vegetation community was more typical of either basic or neutral soil. The categories were (1) Sycamore or Sycamore–Ash stands with Dog's Mercury, (2) Sycamore or Sycamore–Ash with bramble and/or grasses, (3) Beech or Beech–Sycamore with Ramsons, (4) Beech or Beech–Sycamore with bramble and/or Bluebells, (5) Scots and Corsican Pine stands (planted 1940 and 1946) and an area of young Corsican Pine (planted 1974) with deciduous shrubs.

At Millden, five major habitats were defined as (1) woodland, including birch and Scots Pine, (2) marginal habitats, including Bracken and acid flush, (3) dry heath, (4) unimproved acid grassland and (5) improved grassland. Six fine-scale habitats comprising those mainly used by Woodcock were classed as (1) mature birch (> 12 m tall) woodland with Bracken ground cover, (2) mature birch woodland with grass ground cover, (3) sapling-stage birch and young birch thicket, (4) Scots Pine plantation, (5) Bracken and (6) acid flush, characterized by Juncus spp. and Sphagnum spp.

Four habitat variables, ground vegetation cover (%), leaf litter cover (%), basal area per tree (cm2) and tree density (no./ha), were measured at Woodcock feeding locations and an equal number of randomly selected points at Whitwell Wood in May to early June 1985 and 1992 (n = 58 feeding locations) and at Millden in late May to June 1992 (n = 17 feeding locations). Vegetation cover and leaf litter cover were estimated using 0.25-m2 quadrats. The basal area of trees and tree density were calculated from point-quarter measurements (Cottam & Curtis 1956).

Invertebrate sampling

Earthworm availability in Whitwell Wood relative to the surrounding fields was assessed at monthly intervals during late March to late June 1980. Sampling took place during the day using the formalin method (now illegal in the UK for reasons of human health), which involved clearing vegetation and leaf litter from a 0.25-m2 quadrat and applying a solution of 25 mL formalin in 4.5 L of water. All earthworms emerging in 10 min were counted. Adult Lumbricus terrestris were excluded from counts within the woodland because their burrows extend up to 2.5 m below ground and they would not have been readily available to Woodcock feeding during the day.

A comparison of earthworm availability between woodland habitats was undertaken during May 1992 at Whitwell Wood and May 1993 at Millden. For this comparison, earthworms were sampled by digging 25-cm square soil samples to a depth of 7 cm (the length of the Woodcock's bill) and handsorting. All samples were taken between 10:00 and 16:00 h. Handsorting was a more efficient method of sampling earthworms: a calibration of the two methods in 1992 revealed that, on average, handsorting produced 2.09× as many earthworms as the formalin method. For consistency in presentation, density estimates from formalin sampling have been multiplied by two, but analyses have been restricted to estimates derived by one sampling method.

Surface invertebrates were sampled by pitfall trapping. Traps were placed at least 10 m apart in representative areas of each habitat. The number of traps per habitat was approximately proportional to the area occupied by each habitat. Traps were one-third filled with a 50% solution of antifreeze with 5 mL of 70% ethanol added to reduce the surface tension, and were set for 10 days during late May to June.

Diet analysis

Fresh faecal samples were obtained when adult Woodcock were flushed or broods were captured for ringing during late April to early June 1991–92. Only three of 32 adult samples (9%) were from birds accompanying chicks that also yielded samples. Samples were stored in 70% ethanol and were washed through two sieves, with mesh sizes of 0.254 and 0.075 mm, prior to examination to remove fine granular particles that otherwise obscured food items. The first sieve retained most of the arthropod fragments and large earthworm chaetae, but the second was required to retain small earthworm chaetae.

Samples were examined in an 89-mm-diameter Petri dish with a 55-mm-diameter dish glued in the centre, forming a 17-mm annular channel. The underside of the dish was inscribed to divide the channel into 16 equal sectors. Samples were examined at 30× magnification under a binocular microscope. Counts of prey remains were restricted to items that were readily identifiable and carried in known numbers (Appendix). Fragments were identified using Moreby (1988), Royal Entomological Society keys to the main insect taxa, and by comparison with reference collections of invertebrates from both study areas.

No discrete structures were identified for millipedes and fragments of the characteristically striated posterior margin of the segmental ring exoskeleton were counted (Green & Tyler 1989). Most of these fragments were of fairly uniform size (about 1 mm along the ring circumference) and larger pieces were mentally subdivided when accumulating the fragment count. Large numbers of earthworm chaetae were present and the total number was estimated from counts of every fourth sector of the dish (one-quarter of the sample). Oribatid mites and plant fragments were not recorded as prey items as they were probably ingested accidentally as a consequence of probing through leaf litter. Data from different chicks of the same brood were pooled because samples from the same brood were not independent.

Correction factors from studies of captive waders of other species were applied in diet calculations to address the issue of differential digestion rates of different prey taxa (Appendix). Because the number of millipede fragments recovered was dependent on the size of animals ingested, the width of the striated posterior margin (w) was measured to 0.05 mm with a micrometer eyepiece and the number of fragments comprising a single millipede (n) was estimated using the function n = 403.8w2.994 from Green and Tyler (1989). The same correction factors were applied to diet calculations for adult Woodcock and chicks because there have been insufficient studies of differential digestion rates to enable age classes to be treated separately. However, it should be borne in mind when interpreting our results that there are likely to be differences in digestive efficiency between adults and chicks, and between Woodcock and the species used to determine the correction factors. Chaetae probably only provide a crude approximation of the number of earthworms in the diet because the number of chaetae per worm differs between species and varies in relation to worm size.

Prey biomass was calculated by multiplying the estimated number of animals ingested by mean dry weights for each taxon (Appendix). Dry weights were obtained by oven-drying ten individuals of each group (five representative specimens from each study area) for 3 days at 60 °C to constant weight and weighing them to the nearest 0.1 mg.

Statistical analyses

Home ranges were of interest in their own right and were described by minimum convex polygons (MCPs). MCP sizes were calculated in MapInfo 7.0 (MapInfo Corporation 2002). Analysis of covariance (ancova) was used to test for effects of age, sex, median fix date, study area and year (nested within study area) on log10-transformed diurnal home range areas. The number of days that each bird was tracked was included as a covariate to account for the relationship between home range area and the number of fixes. The frequency with which birds left Whitwell Wood at night was analysed using a generalized linear mixed model (GLMM) with a binomial dependent variable (left wood or stayed). Bird age, sex and month were fixed effects in the model and bird identifier and year were random effects.

To assess nocturnal habitat selection, the probability of visiting grass fields as opposed to arable fields was analysed with a GLMM with a binomial dependent variable, bird age, sex and month as fixed model effects and bird identifier and year as random effects. Relative earthworm availability between field types and between Whitwell Wood and pasture fields in 1980 was examined using earthworm numbers as the dependent variable in generalized linear models (GLMs) with negative binomial errors, a logarithmic link function and field type or habitat and month as explanatory factors, respectively. Habitat use by Woodcock roosting in Whitwell Wood from May onwards was analysed by comparing the proportions of radiolocations of 12 birds in Sycamore, Beech or pine stands with the availability of these habitats using compositional analysis (Aebischer et al. 1993).

Woodcock diurnal habitat selection was determined using compositional analysis of the radiotracking data at two levels: comparison of habitat proportions within home ranges relative to those in the study area and comparison of radiolocations within home ranges. At Whitwell Wood the woodland boundary was taken to delimit the study area because the surrounding farmland habitats were not used during the day. The study area contained a mean of 11 ± 3 (± se, range 5–21) spatially independent patches of each habitat category. At Millden the study area was defined as the composite minimum convex polygon of the home ranges of all the radiotagged birds because the woodland was fragmented and non-woodland habitats were also used. Means of 34 ± 13 (± se, range 6–81) spatially independent patches of major habitat types and 29 ± 8 (± se, range 8–61) patches of fine-scale habitats were present at Millden. Computer simulations have indicated that compositional analysis can result in inflated Type I errors when analysis includes habitats with small proportional availability and 0% use (Bingham & Brennan 2004). To avoid this problem, we ensured that the availability of all habitat categories within the study areas and within most home ranges was greater than 6% and we replaced the few null proportions with 0.006 as recommended by Bingham and Brennan (2004). Randomization tests were used to assess whether habitat use was significantly non-random and to determine whether use differed between pairs of habitats. At Whitwell Wood, data were collected over 15 years so time period (1978 + 1980 vs. 1986 + 1992) was included as a factor in the analysis to check for any differences in habitat selection related to changes in habitat structure over time.

For comparison of diets between study areas and between adults and chicks, estimated numbers of individuals ingested were summed into six invertebrate groups and prey frequencies were recalculated. These groups were earthworms, carabid + scarabaeid beetles, staphylinids + weevils + other beetles, beetle larvae + butterfly and moth larvae, woodlice + millipedes, and spiders + harvestmen. Ants and flies were excluded. Compositional analysis was then performed using the proportions of prey frequency.

Comparisons of earthworm abundance and invertebrate abundance/activity in woodland habitats in 1992–93 were made by analysing earthworm and invertebrate numbers as the dependent variables in GLMs with negative binomial errors and a logarithmic link function and study area or habitat as an explanatory factor. Pitfall trap samples provide a very poor measure of the relative abundance of different taxa in ground invertebrate assemblages owing to major differences in activity and trapability (Luff 1975, Adis 1979), but are adequate for assessing relative numbers of a species in different vegetation types (Greenslade 1964).

For analysis of structural features at Woodcock feeding sites, the four habitat variables were analysed using analysis of variance (anova) with location type (feeding or random) and habitat category as factors. For the Whitwell Wood data, year (1986 or 1992) was included as a factor and the location type × year interaction was tested. Vegetation cover and leaf litter cover were arcsin-transformed prior to analysis and the basal area per tree and tree density were log10-transformed. GLMs, GLMMs and randomizations were performed in GENSTAT 8.2 (Lawes Agricultural Trust 2005). All other statistics were calculated in SYSTAT 9 (SPSS Inc. 1999).


  1. Top of page
  2. Abstract
  7. Acknowledgments

Home range sizes and daily movements

There was no difference in home range size in relation to age (ancovaF1,30 = 2.55, P = 0.121), sex (F1,30 = 2.09, P = 0.159), date (F1,30 = 1.15, P = 0.293), study area (F1,30 = 2.22, P = 0.147) or year (F4,30 = 1.30, P = 0.294). Home range size was related to the number of days radiotracking (F1,30 = 10.87, P = 0.003) and mean range area at 30 days was 62.3 ± 19.8 ha (± se). Daily ranges were much smaller at both study areas: 0.13–0.67 ha at Whitwell Wood (n = 5) and 0.06–1.25 ha at Millden (n = 8).

In early spring, a high proportion of radiotagged males, and females that were not incubating clutches of eggs, left Whitwell Wood to feed and roost on fields at night. The proportion of nocturnal locations on fields declined steadily from 94% in March to 18% in July (GLMM inline image = 34.34, P < 0.001, Fig. 1), but there were no differences in the proportions of locations on fields between age classes or sexes (GLMM inline image = 0.05, P = 0.820 and inline image = 0.32, P = 0.574, respectively). The flights to fields were made immediately after the dusk roding period and the mean distance travelled was 1005 ± 51 m (± se, range 50–3915 m, n = 144).


Figure 1. Seasonal change in the frequency that radiotagged Woodcock left Whitwell Wood at night. Means are given +1 se and numbers above bars denote the number of birds monitored each month.

Download figure to PowerPoint

Nocturnal habitat selection

Of 176 nocturnal radiolocations on fields during March–May, permanent pastures were visited on 63.1% of nights, grass leys on 7.4%, grass margins of winter cereal fields on 5.1% and winter cereal (mainly wheat) fields on 24.4%. The mean proportion of nocturnal locations on grass fields or margins was 76 ± 12% (± se) and there was no difference between months (GLMM inline image = 2.25, P = 0.325), ages (inline image = 1.74, P = 0.187) or sexes (inline image = 0.68, P = 0.410). Only c. 5% of the fields within 1 km of Whitwell Wood were grass, but the Woodcock flew no further to feed on grass than to arable fields (mean ± se distances 941 ± 188 and 1088 ± 242 m, respectively, anovaF1,28 = 0.52, P = 0.478).

Permanent pastures had significantly higher earthworm availability in late March to early April than grass leys or winter wheat fields (means ± se 443 ± 167, 250 ± 85 and 88 ± 28 earthworms/m2, respectively, GLM F2,27 = 5.41, P = 0.004). The availability of earthworms in the topsoil declined rapidly in pasture fields from late March to late June, whereas their availability in woodland remained relatively constant until after late May (GLM habitat × month F3,75 = 3.67, P = 0.012, Fig. 2). Later in the season, when roosting at night in the woodland, the radiotagged Woodcock exhibited no differential use of Sycamore, Beech or pine stands relative to availability (Wilks’Λ = 0.781, n = 12, P = 0.289).


Figure 2. Seasonal change in the relative availability of earthworms in pasture fields and woodland at Whitwell Wood in 1980. Earthworms were sampled by the formalin method but density estimates have been corrected for the efficiency of this method relative to handsorting. Means are given +1 se and numbers above bars denote the number of samples.

Download figure to PowerPoint

Diurnal habitat selection

Compositional analysis revealed significant differences in the use of habitats by radiotagged Woodcock at the level of the home range and the radiolocation at both study areas. At Whitwell Wood, home ranges contained higher proportions of Sycamore–Dog's Mercury, Sycamore–bramble and pine relative to availability than Beech–bramble stands (Wilks’Λ = 0.673, n = 28, P = 0.029; no time period effect Λ = 0.833, P = 0.275, Table 1, Fig. 3). Within home ranges, Sycamore–Dog's Mercury was the most intensively used habitat, with Beech–Ramsons least used (Λ = 0.434, n = 28, P = 0.001; time period Λ = 0.741, P = 0.078).

Table 1.  Relative habitat use rankings (most used 5 to least used 0) based upon mean log-ratio differences between proportional use and availability from compositional analysis. Within each row, there was no difference in relative use between habitats with the same superscript letter (randomization test, P > 0.05).
Study areaComparisonHabitat ranking
  1. Habitat abbreviations: SycDm = Sycamore–Dog's Mercury, SycBram = Sycamore–bramble, BeRms = Beech–Ramsons, BeBram = Beech–bramble, YgBirch = young and intermediate age birch, BirchBk = mature birch–Bracken, BirchGr = mature birch–grass.

Whitwell WoodHome range vs. Study area SycDmaSycBramaPineaBeRmsa,bBeBramb
Whitwell WoodFixes vs. Home range SycDmaPinea,bBeBrama,b,cSycBramb,cBeRmsd
MilldenHome range vs. Study areaYgBirchaBrackenaBirchBka,bBirchGrbFlushb,cPinec
MilldenFixes vs. Home rangeYgBirchaBirchBka,bBirchGra,bBrackenbPinea,b,cFlushc

Figure 3. Mean (+1 se) habitat compositions of Woodcock diurnal home ranges and radiolocations relative to availability at (a) Whitwell Wood (n = 28) and (b) Millden (n = 10). Habitat abbreviations are: SycDm = Sycamore–Dog's Mercury, SycBram = Sycamore–bramble, BeRms = Beech–Ramsons, BeBram = Beech–bramble, YgBirch = young and intermediate age birch, BirchBk = mature birch–Bracken, BirchGr = mature birch–grass, Brack = Bracken.

Download figure to PowerPoint

At Millden, 89% of radiolocations (n = 699) were in woodland habitats, but Bracken, acid flush and heath were visited by 4–7 birds and two birds visited grassland. Compositional analysis of major habitat categories revealed greater relative use of woodland and marginal habitats than acid grassland, improved grassland and heath at the home range level (Λ = 0.184, n = 10, P = 0.014) and greater relative use of woodland than all other habitats at the level of the radiolocation (Λ = 0.032, n = 10, P = 0.003). Of the woodland and marginal habitats, home ranges contained high proportions of young birch and Bracken and low proportions of Scots Pine relative to availability within the study area (Λ = 0.135, n = 10, P = 0.041, Table 1, Fig. 3). Within home ranges, young birch was used intensively and acid flush was little used (Λ = 0.085, n = 10, P = 0.007).

Woodcock diet

Earthworms, spiders and harvestmen were the most numerous dietary components of Woodcock adults and chicks at both study sites, with earthworms by far the most important prey items in terms of biomass (Tables 2 & 3). A broad range of other taxa were eaten less frequently, including beetles of at least six families, beetle larvae, millipedes, woodlice, butterfly or moth and sawfly larvae, ants and flies. Only adults ate caterpillars and sawfly larvae and only chicks at Millden ate ants.

Table 2.  Estimated diet composition (mean ± se) of Woodcock at Whitwell Wood, based on analysis of faeces collected during late April to early June 1991–92 and mean (± se) pitfall catch of invertebrates per 10-day trapping session in late May 1991 (n = 158 traps). Mean ± se estimated numbers of animals ingested per sample were 8.8 ± 1.7 for adults and 5.5 ± 3.6 for chicks.
Prey orderAdult (n = 23)Chick (n = 7)Numbers inpitfall traps
% frequency% dry weight% frequency% dry weight
  • *

    Mainly Elateridae

  • Includes Curculionidae.

Lumbricidae30.2 ± 10.050.4 ± 14.434.9 ± 14.751.1 ± 16.0 
Carabidae (adult)3.3 ± 1.86.6 ± 4.29.2 ± 5.517.9 ± 10.78.8 ± 0.7
Staphylinidae (adult)8.4 ± 4.12.5 ± 1.213.8 ± 13.710.8 ± 10.827.8 ± 1.2
Curculionidae (adult)0.6 ± 0.60.1 ± 0.11.0 ± 1.00.1 ± 0.1 
Other adult Coleoptera*0.8 ± 0.90.1 ± ± 2.4
Coleoptera (larval)10.8 ± 4.66.3 ± 2.90.7 ± 0.70.3 ± 0.32.0 ± 0.2
Oniscoidea7.5 ± 3.04.1 ± 1.73.6 ± 3.61.6 ± 1.65.4 ± 0.4
Diplopoda12.9 ± 6.423.4 ± 11.27.4 ± 7.411.0 ± 11.08.7 ± 0.6
Araneae + Opiliones22.2 ± 4.64.8 ± 1.229.3 ± 11.07.1 ± 3.329.6 ± 1.7
Formicidae    0.0   0.0 ± < 0.1
Lepidoptera (larval)3.0 ± 3.01.8 ± 1.6    0.0 0.01.8 ± 0.2
Diptera0.4 ± 0.4< 0.1 ± < 0.10.2 ± 0.2< 0.1 ± < 0.130.1 ± 2.3
Table 3.  Estimated diet composition (mean ± se) of Woodcock at Millden, based on analysis of faeces collected during May 1992, and mean (± se) pitfall catch of invertebrates per 10-day trapping session in June 1992 (woodland habitats and bracken, n = 126 traps). Mean ± se estimated numbers of animals ingested per sample were 8.6 ± 1.1 for adults and 2.6 ± 1.2 for chicks.
Prey orderAdult (n = 9)Chick (n = 9) Numbers inpitfall traps
% frequency% dry weight% frequency% dry weight
  • *

    Mainly Elateridae.

  • Includes Curculionidae and Scarabaeidae.

  • Includes Symphyta larvae.

Lumbricidae32.8 ± 11.968.9 ± 11.849.8 ± 9.379.5 ± 9.3 
Carabidae (adult)5.7 ± 3.211.0 ± 7.60.9 ± 0.61.6 ± 1.11.8 ± 0.2
Scarabaeidae (adult)2.5 ± 0.92.0 ± 1.0     0.0        0.0 
Staphylinidae (adult)8.1 ± 2.72.4 ± 1.05.3 ± 3.11.5 ± 0.94.4 ± 0.5
Curculionidae (adult)5.9 ± 5.01.1 ± 1.11.3 ± 0.90.3 ± 0.2 
Other adult Coleoptera*5.3 ± 3.70.4 ± 0.32.5 ± 1.40.1 ± 0.15.0 ± 1.0
Coleoptera (larval)4.4 ± 5.43.1 ± 3.25.6 ± 3.74.2 ± 2.82.2 ± 0.3
Oniscoidea     0.0     0.02.4 ± 1.61.7 ± 1.30.3 ± 0.1
Diplopoda0.4 ± 0.50.9 ± 1.21.5 ± 1.54.3 ± 4.31.9 ± 0.2
Araneae + Opiliones19.7 ± 2.94.7 ± 0.428.4 ± 7.36.3 ± 2.037.0 ± 2.9
Formicidae     0.0     0.02.4 ± 2.40.4 ± 0.410.1 ± 0.8
Lepidoptera (larval)6.7 ± 5.44.5 ± 3.6     0.0     0.00.4 ± 0.1
Diptera8.6 ± 5.90.9 ± 0.7     0.0     0.00.8 ± 0.3

Compositional analysis of prey groups revealed significant differences in diet composition between study areas (Λ = 0.740, n = 48, P = 0.026) and between adults and chicks (Λ = 0.768, n = 48, P = 0.048). Millipedes and woodlice were eaten in greater numbers at Whitwell Wood compared with Millden, probably reflecting their greater abundance as measured by pitfall catches (GLM millipedes F1,282 = 89.52, P < 0.001 and woodlice F1,282 = 141.99, P < 0.001). Beetle larvae and caterpillars were taken more frequently by adult birds.

Compared with numbers caught in pitfall traps, Woodcock appeared to consume most surface invertebrates broadly in proportion to their abundance/activity (Tables 2 & 3). Possible exceptions to this were that beetle larvae comprised a relatively high proportion of adult diet at Whitwell Wood and small beetles comprised a low proportion of adult and chick diets there. Woodcock adults and chicks at Millden consumed small numbers of ants relative to their abundance/activity.

Food availability in woodland habitats

Earthworm availability differed significantly between habitats at Millden but not at Whitwell Wood (Table 4). Total surface invertebrate abundance/activity and that of spiders and harvestmen differed between habitats at both study areas. The most used feeding habitats of radiotagged Woodcock (Sycamore–Dog's Mercury at Whitwell Wood, young birch at Millden) did not have the highest earthworm densities, but supported densities which were statistically similar to the habitats with the highest values. The habitats with ground flora characteristic of basic soils (Sycamore–Dog's Mercury and Beech–Ramsons) had the highest surface invertebrate and arachnid abundance/activity at Whitwell Wood. At Millden, Bracken had significantly higher abundance/activity of surface invertebrates and arachnids than all other habitats and young birch had similar abundance/activity to acid flush and the other woodland habitats. There was no difference between study areas in earthworm density (GLM F1,205 = 1.25, P = 0.264), but surface invertebrate catches were 1.8 times higher at Whitwell Wood (GLM F1,282 = 73.48, P < 0.001).

Table 4.  Comparison of mean (± se) earthworm density (no./m2) and invertebrate abundance/activity (no. per 10 trap-days) between habitats at each study area. There was no difference in abundance between habitats with the same superscript letter within each comparison. Scots Pine was not sampled for surface invertebrates at Millden. Flies have been excluded from the total invertebrate count because these were not considered to be sampled efficiently by pitfall trapping.
HabitatSoil coresPitfall traps
nEarthwormsnSurface invertebratesSpiders + harvestmen
Whitwell Wood
 Sycamore–Dog's Mercury36127 ± 2180136 ± 7a36 ± 2a
 Sycamore–bramble27171 ± 323982 ± 6b20 ± 2b,c
 Beech–Ramsons 6176 ± 7020133 ± 14a34 ± 4a,b
 Beech–bramble1588 ± 23 962 ± 10b 9 ± 2c
 Pine 369 ± 421074 ± 11b25 ± 5a,b,c
 F4,82 = 1.46 F4,153 = 12.51F4,153 = 14.29
 P = 0.211 P < 0.001P < 0.001
 Mature birch–Bracken20116 ± 39a,b6051 ± 4c21 ± 2c
 Mature birch–grass20200 ± 62a2047 ± 6c25 ± 5b,c
 Young birch40111 ± 27a,b1062 ± 11b,c37 ± 9b,c
 Scots Pine2047 ± 20b 
 Bracken2037 ± 17b3688 ± 8a70 ± 9a
 Acid flush20111 ± 38a,b3070 ± 7b38 ± 6b
 F5,134 = 2.84 F4,151 = 7.46F4,151 = 15.04
 P = 0.014 P < 0.001P < 0.001

Habitat structure at feeding sites

Comparison of habitat measures at Woodcock feeding sites and random locations showed that birds at Whitwell Wood fed at sites with more ground vegetation and less leaf litter in areas with smaller trees (Table 5). The basal area per tree also differed between habitats, being lowest in pine stands (256 cm2, 95% CL 174–377 cm2), probably as a result of the relatively young planting, but averaging 456 cm2 (95% CL 413–503 cm2) in Sycamore–Dog's Mercury. At Millden there were no differences in habitat measures between feeding and random locations, although sample sizes were smaller than at Whitwell and leaf litter cover at both types of site was comparable with that at feeding sites in Whitwell Wood. The size of trees and tree density differed between habitats at Millden, with the young birch patches having an appreciably higher density (2180 trees/ha) of small trees (basal area per tree 53 cm2, 95% CL 26–110 cm2) than all the other habitats.

Table 5.  Comparison of habitat measures at Woodcock feeding sites and random locations during May to early June 1985 and 1992 at Whitwell Wood and during late May to June 1992 at Millden. Data are presented as back-transformed means and 95% confidence intervals. Analyses for Whitwell Wood included year as a factor and tested for a location type–year interaction. This interaction was significant for none of the variables. Sample sizes differed slightly at Whitwell Wood owing to missing values for some variables.
VariableFeedingRandomLocation typeHabitat
Mean (95% CI)Mean (95% CI)StatisticPStatisticP
Whitwell Wood
 Vegetation cover (%) 62 (47–75) 44 (32–56)F1,100 = 5.430.022F4,100 = 0.290.885
 Leaf litter cover (%) 35 (22–50) 69 (57–80)F1,99 = 19.58< 0.001F4,99 = 0.600.665
 Mean basal area per tree (cm2)343 (300–393)423 (376–475)F1,98 = 10.590.002F4,98 = 5.89< 0.001
 Tree density (no./ha)451 (335–608)392 (305–505)F1,98 = 0.010.938F4,98 = 2.380.057
 Vegetation cover (%) 82 (74–90) 84 (76–91)F1,28 = 0.140.712F4,28 = 1.470.237
 Leaf litter cover (%) 34 (24–46) 27 (17–38)F1,28 = 1.210.280F4,28 = 2.220.093
 Mean basal area per tree (cm2)508 (384–672)569 (430–753)F1,28 = 0.380.543F4,28 = 18.60< 0.001
 Tree density (no./ha)304 (173–535)187 (107–329)F1,28 = 1.800.191F4,28 = 13.46< 0.001


  1. Top of page
  2. Abstract
  7. Acknowledgments


Previous studies of Woodcock diet have used stomach contents and focused predominantly on winter because of the relative ease of obtaining material (e.g. Hirons & Bickford-Smith 1983, Granval 1987). Relatively little is known about spring diet composition and, to our knowledge, there is no published information on chick diet. It is clear from our analysis that in May, when Woodcock spend more time feeding in woodland (Hirons & Owen 1982, this study), earthworms remain the most important dietary component in terms of biomass ingested, but spiders, harvestmen and beetles assume greater importance than in winter. The high proportion of earthworms in the diet of chicks is not surprising because they receive food from the female for about the first week and occasionally longer (Workman 1954, Marcström & Sundgren 1977, Watson 2004). The difference in the proportions of prey biomass contributed by earthworms between Whitwell Wood and Millden might be explained by the greater abundance/activity of surface invertebrates at Whitwell Wood.

The range of prey items found in the diets of the Woodcock at Whitwell Wood and Millden was similar to that found by Bettmann (1961) and Koubek (1986) based on the examination of stomachs in spring, with the exception that Bettmann also found Earwigs Forficula auricularia and snails, and fly larvae comprised 13% of prey items in Koubek's analysis. The absence of fly larvae in our samples may be because larval densities were low relative to earthworm densities (20 per m2 at Whitwell Wood and 45 per m2 at Millden) or because our sample size was too small to detect them if they were only eaten in small quantities. The absence of molluscs from the diet of adult Woodcock at our study areas was unexpected. This may be a consequence of the time of season that the samples were collected, because Gruar et al. (2003) found a seasonal decline in earthworms in Song Thrush Turdus philomelos diet and a concurrent increase in the relative abundance of snails, which was most pronounced after mid-June. Other Song Thrush studies have concluded that snails are only taken in large quantities when more preferred prey, such as earthworms or Lepidoptera larvae, are unavailable (Davies & Snow 1965, Török 1985).

Plant material was considered unimportant in the diet of Woodcock in this study because the items found consisted solely of small amounts of leaf fragments, presumed to have been ingested incidentally during probing. However, Koubek (1986) reported the presence of seeds of six species and pine needles, with plant fragments comprising 21% of all the items identified in stomachs. The American Woodcock is also considered to take some plant material as food (Gregg 1984).

Without larger numbers of faecal samples and statistical comparison of prey ingested with their availability, it is difficult to assess whether Woodcock feed selectively on invertebrates other than earthworms. This comparison will always be limited by the poor measure of relative abundance of different taxa in ground invertebrate assemblages provided by pitfall samples (Luff 1975, Adis 1979) and the fact that certain species are active mainly at night, although pitfall traps may provide a reasonable measure of likely encounter rates with some diurnally active taxa for birds. Spiders and harvestmen were the most numerous taxa in our pitfall catches, suggesting that the Woodcock were probably not selecting them over other taxa, but were responding to their seasonal abundance/activity. The difference in the proportion of millipedes in the diets of birds at Whitwell Wood and Millden is also related to the numbers caught in pitfall traps and is suggestive that the birds were simply taking items that were easily caught in relation to their abundance/activity.

Use of fields

Woodcock regularly feed on fields at night in winter (Hirons & Bickford-Smith 1983, Wilson 1983). Recently, nocturnal activity has been shown to be inversely correlated with air temperature and with daylight foraging activity (Duriez et al. 2005). Thus, birds in woodland patches with low food availability or in cold weather, or both, flew to fields at night, where there was a relative superabundance of earthworms, to satisfy their energy requirements. Because earthworms are such an important food for Woodcock, it is not surprising that this behaviour continued into spring. March and April are likely to be months of high energy demand for both males and females because most males will be roding and females will be attempting to maintain or improve body condition prior to egg-laying. The preference for pasture fields at night is consistent with birds maximizing their energy intake, especially because the flights to pasture fields at Whitwell were no further than those to the few winter wheat fields that were used. The relative stability of earthworm densities in the woodland provides a likely reason for the seasonal change in foraging behaviour by the Woodcock, although our data were not particularly convincing. It is possible that the effectiveness of the formalin sampling method differed between woodland and pasture fields, perhaps as a result of soil compaction in pastures, and soil cores would have provided better estimates of absolute earthworm densities in the two habitats. Our analysis of seasonal changes in earthworm availability would also be more convincing with larger sample sizes, particularly for pasture in late May and late June. However, a decline in earthworm availability in pasture fields in summer due to progressive desiccation of the topsoil is well documented (Edwards & Lofty 1977, Peach et al. 2004). Hirons and Owen (1982) present activity data for radiotagged Woodcock demonstrating that actual feeding time in woodland and pastures changes from early spring to mid-summer.

Habitat selection

Compared with other soil invertebrate specialists breeding in woodland, the Woodcock's diurnal range is large (e.g. Song Thrush 3.3 ha, Peach et al. 2004). This is not surprising given its body mass (270–330 g in the breeding season), but it may also be due, at least in part, to the Woodcock's polygynous breeding system (Hirons 1983), which means that males are not required to defend territories. Although daily ranges were small, birds presumably had the freedom to change feeding areas more frequently and to explore larger areas for food-rich patches.

Woodcock home ranges at Whitwell Wood were established predominantly in Sycamore stands. It was apparent, however, that only small parts of home ranges were used intensively for feeding and that this took place predominantly in Sycamore stands where Dog's Mercury was the dominant ground cover. There are quite likely to have been subtle changes in habitat use over the 15-year period of our study at Whitwell Wood, related to changes in habitat structure and/or predator abundance. However, our analysis of diurnal habitat use was based on habitat categories which were probably too crude to detect this. Nevertheless, the fact that time period was not significant in the compositional analysis suggests that the combinations of tree species and dominant vegetation used to define categories remained important despite any changes over time in the structure of the understorey.

Our upland study area at Millden consisted of relatively simple but more fragmented habitats compared with Whitwell Wood and provided the first information on breeding Woodcock use of upland margins. Birch woodlands cover less than 2% of the Scottish uplands (Milne et al. 1998), but comprise the most abundant class of semi-natural woodland (MacKenzie 1987) and are important for their landscape and biodiversity value (Fenton 1984, Forestry Commission 1994). The Woodcock at Millden made greatest use of young regenerating birch, including high proportions within their home ranges and spending a large proportion of the daytime within it. Young birch was typically found encroaching onto unimproved grassland or acid flush at the edges of mature birch, and sapling-stage patches consisting of stems about 5–10 cm in diameter and 5–8 m tall seemed to be used most.

Habitat use represents a trade-off between the availability of food in a particular habitat and the predation risk its structure presents. Although Woodcock may comprise only a small proportion of the prey items in the diet of Sparrowhawks Accipiter nisus, Goshawks Accipiter gentilis and Tawny Owls Strix aluco (Zomerdijk 1983, Newton 1986, Toyne 1998), predation by these avian predators may be an important cause of mortality for Woodcock. Of ten ringed or radiotagged Woodcock killed by predators at Whitwell Wood, three were taken by Sparrowhawks and two by Tawny Owls, with the others killed by mammals. Sparrowhawks were observed chasing roding Woodcock at Millden and Woodcock feathers were found at plucking posts.

At Whitwell, Sycamore–bramble and Beech–Ramsons areas supported earthworm densities at least as high as those in Sycamore–Dog's Mercury stands, suggesting that habitat structure also influenced the choice of feeding locations. Dog's Mercury provides a good combination of overhead cover and ease of passage for Woodcock compared with bramble and Ramsons. Similarly, young birch areas did not support the highest earthworm densities at Millden, which were found under mature birch with grass ground cover. However, the latter areas were grazed by sheep and deer and, certainly early in the breeding season, would have provided little cover from avian predators. Areas of sapling-stage birch typically had sparse ground vegetation but the density of stems made them impenetrable to Sparrowhawks and Tawny Owls. Provided that overhead cover is sufficiently dense to reduce predation risk significantly, sparse ground cover and leaf litter presumably facilitate probing for food.

Early studies of the American Woodcock found similar habitat use. Areas highly used by American Woodcock were characterized by a high density of shrubs and reduced herb cover (Rabe 1977), but in woodland with a poorly developed shrub layer, birds selected for herb stem density and ground cover (Morgenweck 1977). Radiotracking studies of American Woodcock have also shown that individuals utilize intensively only small areas of even preferred habitat types (Dunford & Owen 1973, Morgenweck 1977). Our results at Whitwell Wood are consistent with this, with feeding locations having more ground cover, less leaf litter and smaller trees than random locations; this effect was not simply the product of the habitat chosen. Hirons and Johnson (1987) demonstrated that these patches contain higher earthworm densities than random sites at Whitwell Wood. In the American Woodcock, an experimental trial showed that the number of probes per capture decreased when earthworm density increased (22 probes at 26 earthworms/m2, i.e. a probing success of 0.05, decreasing to nine probes at 105 worms/m2, i.e. probing success of 0.11, Rabe et al. 1983). The density of earthworms above which Woodcock intake rate is not limited is not known, but this density is unlikely to be reached in woodland soils.

We were unable to detect structural differences between feeding sites and random locations after controlling for habitat at Millden, suggesting that structure was more uniform within habitats than at Whitwell Wood or that further selection was of no advantage to the Woodcock. In the United States, Alder Alnus glutinosa and Aspen Populus tremula between 10 and 20 years old constitutes the preferred feeding habitat of American Woodcock (Gregg 1984, Sepik et al. 1992). Throughout the eastern United States Woodcock populations have been declining as a result of the loss of early-successional forest habitats (Dessecker & McAuley 2001).

Could habitat change be responsible for a decline in breeding Woodcock?

This study has concentrated largely on habitats used by Woodcock for feeding. Clearly habitat suitable for displaying and nesting is also important if Woodcock are to breed successfully. Hirons and Johnson (1987) have shown that nesting areas at Whitwell are characterized by less dense ground vegetation than areas used for feeding and brood rearing. Based on just a small sample of nests (n = 10), the same seemed to be true at Millden, with half of nests found in bracken and heather outside the woodland (Hoodless 1994). Again, studies of American Woodcock in naturally regenerating Alder and Aspen woodland have shown that birds nest in relatively open areas whereas broods and solitary birds use heavier cover (Rabe 1977). Nevertheless, Hirons (1988) found that early season nests in Beech were near to the edges of stands, and hence close to Sycamore areas that provided better feeding.

Woodcock are likely to be sensitive to habitat change because they are specialist feeders and have specific habitat requirements. They need large woods containing a mosaic of clearings and rides, stands with relatively sparse, open ground cover, and areas with continuous ground cover or patches of dense shrubs and saplings. They are also dependent on fields near woodland for food for a large part of the year. Changes in woodland management, forestry policy and agricultural practice are therefore all likely to affect breeding Woodcock.

There has been a 76% increase in the total area of woodland in Britain since the 1940s (Warren & Key 1991). However, there has been a major change in the abundance of different types of woodland. There was a 199% increase in coniferous woodland during 1940–90, but managed coppice declined by 82% during the same period (Warren & Key 1991). Although some of the conifer areas at Whitwell Wood were highly utilized, these consisted of young Scots Pine (10 years old), which was still relatively open and supported dense ground vegetation. The mature Scots Pine areas at Millden ranked lowest for relative use. A recent national survey of breeding Woodcock showed that birds often rode over conifer forests, but that numbers of registrations of roding males were lower than in deciduous woods (Hoodless et al. 2006). Parslow (1967) supposed that Woodcock benefited from the large-scale planting of conifers in southern Scotland and north Wales in the 1960s, but the value of this habitat to Woodcock for breeding is probably now much reduced owing to closure of the forest canopy. Further work on Woodcock breeding ecology should focus on how birds use conifer forests, particularly in Scotland and Wales. In these countries Woodcock trends may be closely linked to the species, age and configuration of plantations. In many lowland Oak woods, Hazel coppice, particularly the younger growth stages, may provide habitat of broadly similar structure to young birch thickets and hence comprise important foraging habitat. However, active coppice management is likely to be important to ensure continued use by Woodcock.

There is now good evidence for changes in the structure of British broadleaved and mixed woods during the last 20–30 years. Kirby et al. (2005) found a reduction in small woody stems (< 10 cm dbh), an increase in the basal area of woody species and a decrease in open habitats, such as rides and glades, between 1971 and 2001, as well as a marked decline in ground flora richness. They concluded that without deliberate management intervention woods are, on average, likely to become older and darker in the next 20 years, leading to a decline in species associated with open space and young growth. Amar et al. (2006) recorded increases in Bracken, vegetation cover at 4–10 m and dead trees, along with a reduction in shrub diversity, between the 1980s and 2003/04, which they suggest are indicative of a reduction in woodland management and/or an increase in grazing/browsing. Both of these processes, and the resultant changes in woodland structure, are likely to lead to reductions in suitable habitat for breeding Woodcock, which require open habitats and young growth.

Declining management, such as the abandonment of coppicing, reduced stand thinning and general neglect resulting in the closure of rides and open spaces, may be widespread in lowland woods (Fuller et al. 2005) and could be a major factor in local Woodcock declines in southern and eastern England. There is currently widespread concern about increased deer numbers in lowland England and intensified grazing and browsing pressure on woodlands (Fuller & Gill 2001, Fuller et al. 2005). Circumstantial evidence indicates that grazing pressure can affect the abundance of many bird species (Donald et al. 1998, Perrins & Overall 2001), but a better understanding of the processes is required. A reduction in the density and height of shrubs and the removal of bramble is probably of most relevance to Woodcock.

Outside woodland, Woodcock may be susceptible to a number of the large-scale agricultural changes which have affected food availability for other declining birds such as the Starling Sturnus vulgaris, Song Thrush, Mistle Thrush Turdus viscivorus and Lapwing Vanellus vanellus. The conversion of permanent pasture to cereal production, the loss of livestock and organic fertilizers from arable farms and the installation of under-field drainage systems all result in lower densities of soil invertebrates like earthworms (Edwards 1984, O’Connor & Shrubb 1986, Peach et al. 2004).


  1. Top of page
  2. Abstract
  7. Acknowledgments

This work was funded by The Game Conservancy Trust, the Migratory Birds Commission of CIC, David Caldow and the Bernard Sunley Charitable Foundation. We are grateful to the Forestry Commission (Hugh Insley, Andrew Powers) and to Mrs Duffield for permission to work at Whitwell Wood and Millden, respectively. John Ellis, Fred Ellis, Gordon Evers and Tim Johnson assisted with fieldwork at Whitwell. We thank Jon Easton, Dennis Caithness and Richard Cook for help at Millden. Steve Moreby and Stewart Lowther helped with identification of arthropod fragments. The compositional analyses were performed using an Excel macro written by Dr P.G. Smith. Nicholas Aebischer made helpful comments on an earlier draft and supplied Genstat code for compositional analysis randomizations in analyses involving factors.


  1. Top of page
  2. Abstract
  7. Acknowledgments
  • Adis, J. 1979. Problems of interpreting arthropod sampling with pitfall traps. Zool. Anzeiger Jena 202: 177184.
  • Aebischer, N.J., Robertson, P.A. & Kenward, R.E. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74: 13131325.
  • Amar, A., Hewson, C.M., Thewlis, R.M., Smith, K.W., Fuller, R.J., Lindsell, J.A., Conway, G., Butler, S. & MacDonald, M.A. 2006. What's Happening to Our Woodland Birds? Long-Term Changes in the Populations of Woodland Birds. RSPB Research Report no. 19. BTO Research Report no. 169. Sandy: RSPB.
  • Amlaner, C.J. Jr, Sibley, R. & McCleery, R. 1978. Effects of telemetry transmitter weight on breeding success in Herring Gulls. Biotelemetry 5: 154163.
  • Bettmann, H. 1961. Die Waldschnepfe. Munchen-Solln: F.C. Mayer Verlag.
  • Bingham, R.L. & Brennan, L.A. 2004. Comparison of Type I error rates for statistical analyses of resource selection. J. Wildl. Manage. 68: 206212.
  • Clausager, I. 1972. Skovsneppen (Scolopax rusticola) som ynglefugl I Danmark. Danske Vildtundersogelser, hefte 19.
  • Clausager, I. 1973. Age and sex determination of the Woodcock (Scolopax rusticola). Dan. Rev. Game Biol. 8: 118.
  • Cottam, G. & Curtis, J.T. 1956. The use of distance measures in phytosociological sampling. Ecology 37: 451460.
  • Custer, T.W. & Pitelka, F.A. 1975. Correction factors for digestion rates for prey taken by Snow Buntings (Plectrophenax nivalis). Condor 77: 210212.
  • Davies, P.W. & Snow, D.W. 1965. Territory and food of the Song Thrush. Br. Birds 58: 161175.
  • Dessecker, D.R. & McAuley, D.G. 2001. Importance of early successional habitat to Ruffed Grouse and American Woodcock. Wildlife Soc. Bull. 29: 456465.
  • Donald, P.F., Fuller, R.J., Evans, A.D. & Gough, S.J. 1998. Effects of forest management and grazing on breeding bird communities in plantations of broadleaved and coniferous trees in western England. Biol. Conserv. 85: 183197.
  • Dunford, R.D. & Owen, R.B. Jr 1973. Summer behavior of immature radio-tagged woodcock in central Maine. J. Wildl. Manage. 37: 462469.
  • Duriez, O., Fritz, H., Binet, F., Tremblay, Y. & Ferrand, Y. 2005. Individual activity rates in wintering Eurasian Woodcocks: starvation versus predation risk trade-off? Anim. Behav. 69: 3949.
  • Dwyer, T.J., McAuley, D.G. & Derleth, E.L. 1983. Woodcock singing-ground counts and habitat changes in the Northeastern United States. J. Wildl. Manage. 47: 772779.
  • Edwards, C.A. 1984. Changes in agricultural practice and their impact on soil organisms. In Jenkins, D. (ed.) Agriculture and the Environment: 5665. Proceedings of ITE Symposium No. 13. Cambridge: NERC.
  • Edwards, C.A. & Lofty, J.R. 1977. The Biology of Earthworms. London: Chapman & Hall.
  • Fenton, J. 1984. The State of Highland Birchwoods. Edinburgh: Scottish Wildlife Trust.
  • Forestry Commission. 1994. The Management of Semi-Natural Woodlands: 6. Upland Birchwoods. Edinburgh: Forestry Commission.
  • Fuller, R.J. 1982. Bird Habitats in Britain. Calton: Poyser.
  • Fuller, R.J. & Gill, R.M.A. 2001. Ecological impacts of increasing numbers of deer in British woodland. Forestry 74: 193199.
  • Fuller, R.J., Noble, D.G., Smith, K.W. & Vanhinsbergh, D. 2005. Recent declines in populations of woodland birds in Britain: a review of possible causes. Br. Birds 98: 116143.
  • Galbraith, H. 1989. The diet of Lapwing Vanellus vanellus chicks on Scottish farmland. Ibis 131: 8084.
  • Gibbons, D.W., Reid, J.B. & Chapman, R.A. 1993. The New Atlas of Breeding Birds in Britain and Ireland: 1988–1991. London: Poyser.
  • Granval, P. 1987. Régime alimentaire diurne de la Bécasse des Bois (Scolopax rusticola) en hivernage: approche quantitative. Gibier Faune Sauvage 4: 125147.
  • Granval, P. & Bouché, M.B. 1993. Importance of meadows for wintering Eurasian Woodcock in the west of France. In Longcore, J.R. & Sepik, G.F. (eds) 8th American Woodcock Symposium: 135. Washington, DC: U.S. Fish and Wildlife Service.
  • Green, R.E. 1984. The feeding ecology and survival of partridge chicks (Alectoris rufa and Perdix perdix) on arable farmland in East Anglia. J. Appl. Ecol. 21: 817830.
  • Green, R.E. & Tyler, G.A. 1989. Determination of the diet of the Stone Curlew (Burhinus oedicnemus) by faecal analysis. J. Zool., Lond. 217: 311320.
  • Greenslade, P.J.M. 1964. Pitfall trapping as a method for studying populations of Carabidae (Coleoptera). J. Anim. Ecol. 33: 301310.
  • Gregg, L. 1984. Population Ecology of Woodcock in Wisconsin. Technical Bulletin no. 144. Madison, WI: Department of Natural Resources.
  • Gregory, R.D., Wilkinson, N.I., Noble, D.G., Robinson, J.A., Brown, A.F., Hughes, J., Procter, D., Gibbons, D.W. & Galbraith, C.A. 2002. The population status of birds in the United Kingdom, Channel Islands and Isle of Man: an analysis of conservation concern 2002–2007. Br. Birds 95: 410448.
  • Gruar, D., Peach, W. & Taylor, R. 2003. Summer diet and body condition of Song Thrushes Turdus philomelos in stable and declining farmland populations. Ibis 145: 637649.
  • Hirons, G. 1983. A five-year study of the breeding behaviour and biology of the Woodcock in England – a first report. In Kalchreuter, H. (ed.) Proceedings of the 2nd European Woodcock and Snipe Workshop: 5167. Slimbridge: IWRB.
  • Hirons, G. 1988. Habitat use by Woodcock (Scolopax rusticola) during the breeding season. In Havet, P. & Hirons, G. (eds) Proceedings of the 3rd European Woodcock and Snipe Workshop: 4247. Slimbridge: IWRB.
  • Hirons, G. & Bickford-Smith, P. 1983. The diet and behaviour of Eurasian Woodcock wintering in Cornwall. In Kalchreuter, H. (ed.) Proceedings of the 2nd European Woodcock and Snipe Workshop: 1117. Slimbridge: IWRB.
  • Hirons, G. & Johnson, T.H. 1987. A quantitative analysis of habitat preferences of Woodcock Scolopax rusticola in the breeding season. Ibis 129: 371381.
  • Hirons, G.J.M. & Owen, R.B. Jr 1982. Radio-tagging as an aid to the study of woodcock. Symp. Zool. Soc. Lond. 49: 139152.
  • Hoodless, A.N. 1994. Aspects of the Ecology of the European Woodcock Scolopax rusticola L. PhD thesis, University of Durham.
  • Hoodless, A., Lang, D., Fuller, R.J., Aebischer, N. & Ewald, J. 2006. Development of a survey method for breeding Woodcock and its application to assessing the status of the British population. In Ferrand, Y. (ed.) Proceedings of the 6th European Woodcock and Snipe Workshop: 4854. Wageningen: Wetlands International.
  • Kirby, K.J., Smart, S.M., Black, H.I.J., Bunce, R.G.H., Corney, P.M. & Smithers, R.J. 2005. Long-Term Ecological Change in British Woodland (1971–2001). English Nature Research Report no. 653. Peterborough: English Nature.
  • Koubek, P. 1986. The spring diet of the Woodcock (Scolopax rusticola). Folia Zool. 35: 289297.
  • Lawes Agricultural Trust. 2005. Genstat 8 Release 2. Reference Manual. Oxford: Oxford University Press.
  • Luff, M.L. 1975. Some features influencing the efficiency of pitfall traps. Oecologia 19: 345357.
  • MacKenzie, N. 1987. The Native Woodlands of Scotland. Edinburgh: Friends of the Earth (Scotland).
  • MapInfo Corporation. 2002. MapInfo Professional. New York: MapInfo Corporation.
  • Marcström, V. & Sundgren, F. 1977. On the reproduction of the European Woodcock. Viltrevy 10: 2740.
  • Milne, J.A., Birch, C.P.D., Hester, A.J., Armstrong, H.M. & Robertson, A. 1998. The Impact of Vertebrate Herbivores on the Natural Heritage of the Scottish Uplands – a Review. Scottish Natural Heritage Review no. 95. Edinburgh: Scottish Natural Heritage.
  • Moreby, S.J. 1988. An aid to the identification of arthropod fragments in the faeces of gamebird chicks (Galliformes). Ibis 130: 519526.
  • Morgenweck, R.O. 1977. Diurnal high use areas of hatching-year female Woodcock. In Keppie, D.M. & Owen, R.B. Jr (eds) Proceedings of the 6th Woodcock Symposium: 155160. Fredericton: New Brunswick Dept of Natural Resources.
  • NCC. 1990. Handbook for Phase 1 Habitat Survey – a Technique for Environmental Audit. Peterborough: NCC.
  • Newton, I. 1986. The Sparrowhawk. Calton: Poyser.
  • O’Connor, R.J. & Shrubb, M. 1986. Farming and Birds. Cambridge: Cambridge University Press.
  • Parslow, J.L.F. 1967. Changes in status among breeding birds in Britain and Ireland. Br. Birds 60: 97123.
  • Peach, W.J., Denny, M., Cotton, P.A., Hill, I.F., Gruar, D., Barritt, D., Impey, A. & Mallord, J. 2004. Habitat selection by Song Thrushes in stable and declining farmland populations. J. Appl. Ecol. 41: 275293.
  • Perrins, C.M. & Overall, R. 2001. Effect of increasing numbers of deer on bird populations in Wytham Woods, central England. Forestry 74: 299309.
  • Rabe, D. 1977. Structural analysis of Woodcock diurnal habitat in northern Michigan. In Keppie, D.M. & Owen, R.B. Jr (eds) Proceedings of the 6th Woodcock Symposium: 125134. Fredericton: New Brunswick Dept of Natural Resources.
  • Rabe, D.L., Prince, H.H. & Beaver, D.L. 1983. Feeding-site selection and foraging strategies of American Woodcock. Auk 100: 711716.
  • Sauer, J.R. & Bortner, J.B. 1991. Population trends from the American Woodcock singing-ground survey, 1970–88. J. Wildl. Manage. 55: 300312.
  • Sepik, G.F., Owen, R.B. Jr & Coulter, M.W. 1992. A Landowner's Guide to Woodcock Management in the Northeast. Miscellaneous Report 253. Maine: LSA Experiment Station, University of Maine.
  • Sepik, G.F., McAuley, D.G. & Longcore, J.R. 1993. Critical review of the current knowledge of the biology of the American Woodcock and its management on the breeding grounds. US Fish Wildl. Serv. Biol. Report 16: 98104.
  • Shorten, M. 1974. The European Woodcock (Scolopax rusticola). A Search of the Literature Since 1940. Game Conservancy Report no. 21. Fordingbridge: The Game Conservancy.
  • SPSS Inc. 1999. SYSTAT 9. Chicago: SPSS Inc.
  • Török, J. 1985. Comparative ecological studies on Blackbird and Song Thrush populations. I. Nutritional ecology. Opusc. Zool. (Budapest) 21: 105135.
  • Toyne, E.P. 1998. Breeding season diet of the Goshawk Accipiter gentilis in Wales. Ibis 140: 569579.
  • Warren, M.S. & Key, R.S. 1991. Woodlands: past, present and potential for insects. In Collins, N.M. & Thomas, J.A. (eds) The Conservation of Insects and Their Habitats: 155211. London: Academic Press.
  • Watson, A. 2004. Woodcock feeding young. Br. Birds 97: 98.
  • Wilson, J. 1983. Wintering site fidelity of Woodcock in Ireland. In Kalchreuter, H. (ed.) Proceedings of the 2nd European Woodcock and Snipe Workshop: 1827. Slimbridge: IWRB.
  • Workman, W.H. 1954. Notes on woodcock and snipe from Tynan Abbey. Irish Naturalists’ Journal 11: 232.
  • Zomerdijk, P. 1983. De houtsnip (Scolopax rusticola), deel I. De Graspieper, October 1983, 107121.
Table Appendix..  Prey remains, correction factors and dry weights used to estimate the number of individuals and dry weight of each prey taxon ingested. Correction factors were obtained from Custer and Pitelka (1975), Green (1984), Galbraith (1989) and Green and Tyler (1989), with the mean being used when more than one was available for a particular taxon. Dry weights are means calculated from ten individuals collected at Whitwell Wood and Millden. The mean carabid dry weight is based on Carabus spp. because carabid remains in Woodcock diet were mainly from relatively large species.
Prey orderCharacteristic fragments and correction factors to convert to one individual ingestedDry weight (mg)
LumbricidaeChaetae × 1/35331.4
Carabidae (adult)Maximum number from:26.9
mandibles × 0.65 or tibia × 0.20 
Scarabaeidae (adult)Maximum number from:11.6
mandibles × 0.65 or tibia × 0.20 
Staphylinidae (adult)Maximum number from: 4.1
mandibles × 0.73 or tibia × 0.21 
Curculionidae (adult)Maximum number from: 2.5
mandibles × 0.65 or tibia × 0.20 
Other adult ColeopteraMaximum number from: 1.0
mandibles × 0.73 or tibia × 0.21 
Coleoptera (larval)Maximum number from:  9.0
mandibles × 0.70 or tibia × 0.21 
OniscoideaMaximum number from: 8.9
tergites × 1.042 or legs × 0.52 
DiplopodaNumber of 1-mm ring fragments/29.4
(403.8 × posterior margin width2.994) 
Araneae + OpilionesMaximum number from: 3.4
mandibles × 0.835 or metatarsi × 0.209 
FormicidaeLegs × 0.20 1.7
Lepidoptera (larval)Mandibles × 0.5 7.7
DipteraWings × 0.75 1.2