The results of a national survey of wintering Skylarks Alauda arvensis unde
rtaken by the British Trust for Ornithology (BTO) between November 1997 ari
d February 1998 are reported here. Over three visits, volunteers counted Sk
ylarks and mapped habitats in 541 1-km squares selected from the Skylarks w
inter range based oil BTO Winter Atlas data and a stratified random samplin
g approach. Four landscape strata were defined from the Institute of Terres
trial Ecology landscape classification: arable, pastoral, marginal upland a
nd saltmarsh. The survey counts underestimated Skylark abundance, but were
good measures of relative abundance across habitat types. The two best pred
ictors of Skylark presence-absence at the landscape scale were the availabi
lity of coastal and farmland habitats. Squares with saltmarsh had the highe
st densities and occupancy (80% of squares). At the patch scale crop stubbl
es, especially weedy cereal stubbles, were used significantly more than exp
ected by chance. Oilseed rape was positively selected whereas cereal crops
were used in proportion to availability and grazed grass was avoided. Skyla
rks avoided fields smaller than 2.5 ha and selected fields larger than 7.5
ha. We estimate that in midwinter there may be less than 1-2 ha of weedy ce
real stubble per 1-km square. We recommend the retention of over-winter stu
bbles for the conservation of Skylarks arid other farmland birds, and resea
rch oil stubble management and effects on grain availability and arable wee
d regeneration on Skylark use.