We studied timing of breeding in four species of hole-nesting passerines, t
he Siberian tit (Poecile cinctus), great tit (Parus major), pied flycatcher
(Ficedula hypoleuca), and redstart (Phoenicurus phoenicurus), and the abun
dance of the arthropods in their diet (ground-dwelling spiders, winged inse
cts, foliage-feeding lepidopteran and hymenopteran larvae) in Finnish Lapla
nd for 6-12 years. Densities of the invertebrate groups varied considerably
, both seasonally and annually. All the bird species started to breed in ea
rly summer, when weather conditions were often poor. At the beginning of th
e breeding season all species relied on ground-dwelling spiders, whose abun
dances peaked early and were less variable over the years than those of ins
ects. Abundances and timing of emergence of caterpillars in birch and pine
canopies (the major food sources for foliage-gleaners) were highly variable
, but each year their abundances peaked in July or August, after the nestli
ng period of the birds. This contrasts with reported cases in deciduous for
ests at midlatitudes and shows that at our site the birds did not time thei
r nestling period ultimately to coincide with peak food availability. Mean
clutch sizes declined seasonally in the tit species and the pied flycatcher
but not in the redstart. The number of fledglings did not depend on laying
date, except in the Siberian tit, for which dependence of fledgling number
on laying date varied among years. Although early broods were not more pro
ductive (i.e., greater number of fledglings) than later ones, early breedin
g is ultimately a necessity for recruitment because of the short subarctic
summer. Birds cannot delay breeding in the north because late breeding woul
d shorten the time available for moulting, food-hoarding, and preparation f
or migration or wintering. Our results show that at the time of egg laying,
passerines may receive proximate cues to allow them to predict food condit
ions during the nestling period. Still, exact prediction of food availabili
ty during later phases of nesting is constrained by an inherent discrepancy
between warm-blooded birds and poikilothermic invertebrates: once a bird h
as started egg laying, the subsequent phases follow nearly automatically on
certain calendar dates, whereas the development of the arthropods depends
on the rate of temperature increase.