Jl. Telleria et T. Santos, DISTRIBUTIONAL PATTERNS OF INSECTIVOROUS PASSERINES IN THE IBERIAN FORESTS - DOES ABUNDANCE DECREASE NEAR THE BORDER, Journal of biogeography, 20(2), 1993, pp. 235-240
In this paper we analyse whether the abundance of breeding insectivoro
us birds decreases towards the South along the Iberian forests (as pre
dicted by the model of Brown (1984)), and try also to relate these cha
nges of abundance to several environmental features. Abundances of six
foliage insectivorous passerines that are not tree-hole nesters, Trog
lodytes troglodytes (L., 1758), Erithacus rubecula (L., 1758), Sylvia
atricapilla (L., 1758), Phylloscopus collybita (Vieillot, 1817), Regul
us ignicapillus (Temminck, 1820), and Aegithalos caudatus (L., 1758) w
ere recorded in fifty-eight large woodlands along a 850 km belt crossi
ng the Iberian Peninsula. By means of simple, partial and stepwise mul
tiple regression analyses bird abundances were related with distances
to the north of each woodland and with the mean scores of climatic (me
an temperature and precipitation), physiognomic (tree trunk densities
under 20 cm diameter and over 30 cm) and floristic (conifer v. broadle
aved tree species) variables. Five of the six species showed significa
nt, negative correlations between their abundances and the distances o
f forests to the north, thus corroborating Brown's model. A. caudatus
did not show, however, any clear pattern of abundance distribution. Si
mple correlation analyses showed also the importance of climatic and p
hysiognomic variables in predicting bird abundance. When the problem o
f interaction between variables was solved, all the five species showe
d significant positive partial correlation with precipitation, althoug
h distance continued to be important for three species (T. troglodytes
, R. ignicapillus and P. collybita), suggesting some additional role o
f this variable in the processes involved in the observed distribution
. The importance of precipitation and distance to the north as predict
ors of abundance was also showed by the stepwise regression analysis.
The mean variance explained by these models was around 30%, higher tha
n or similar to that obtained in other studies performed at the local
or intra-habitat level. This suggests the potential usefulness of mult
ivariate approaches in studies of bird distribution undertaken on larg
e scales.