1. The abundance and fecundity-related body size variation of the cyclic autumnal moth Epirrita autumnata were monitored from the early increase phase and throughout the outbreak to the end of the density decline in northernmost Norway during 1999-2006. Another geometrid, the winter moth Operophtera brumata, did not increase in density until the autumnal moth had its post-peak in 2004, and was at its own peak concurrent with the steeply declining autumnal moth abundance in 2005-06. 2. The body size variables measured (forewing lengths of males and females and hind femur lengths of males) of the autumnal moth showed a similar density-dependent response, i.e. increasing density was associated with decreasing body size and fecundity. Nevertheless, regression analyses clearly ranked the pooled geometrid abundance without a time lag as the best predictor for the body size variation, ahead of the abundance of the autumnal moth or past abundance of all geometrids. 3. Nondelayed effects of lowered food quality and absolute shortage of foliage under congested conditions are the most plausible reasons for reduced body size. 4. Two most commonly proposed causal factors of the autumnal moth population cycle, i.e. delayed inducible resistance of the host plant (mountain birch Betula pubescens czerepanovii) and delayed density-dependent parasitism by specialized hymenopteran parasitoids, cannot easily explain the diverging population trends between the autumnal and winter moths. 5. We suggest that either the inducible resistance of the host tree or the host utilization of the most important parasitoids and/or pathogens have to be strictly species-specific between these closely related moth species to produce the population dynamics observed. That fecundity of the autumnal moth was best related to the pooled geometrid abundance weakens support for the former hypothesis, while our lack of host-specific information limits conclusions about the role of natural enemies.