1. The aim of this article is to explore whether cost-benefit analysis of behaviour may help to understand the population dynamics of a species. The Great Tit is taken as an example. 2. The lifetime reproductive success in different populations of Great Tits amounts from 0.7 (Hoge Veluwe, Wytham) to 1.5 recruits per pair in an island population (Vlieland). As numbers stay roughly stable the difference points at dispersal as an important factor determining local recruitment. 3. The variation in lifetime reproductive success between individuals can be explained in the first place by variations in the recruitment rate, the chance that a fledgling returns as a breeding bird. Also variations in lifespan play a role in determining variations in lifetime reproductive success. Variations in beech crop index are not responsible for these effects. 4. Juvenile Great Tits start competing for territories in autumn. Both the density of resident territorial males and the density of candidate males affect the probability that a candidate will get a territory. A descriptive model shows that the effect of the density of resident males on the candidate settling chance is always negative. However, the effect of candidate density is dependent on the density of resident males. At low resident male densities the density of candidates has a negative effect, and at high resident male densities a positive effect on settling chance of a candidate. The consequences of this model for the settling pattern in other woods under study are discussed. 5. A speculative cost-benefit model is developed in order to explain the interaction between resident male density and candidate density in explaining settling chance as mentioned under 4. 6. Experimental manipulation of brood size is undertaken in order to measure costs and benefits associated to brood size. On basis of these experiments an optimalisation model is constructed that predicts the clutch size that maximizes first clutch recruitment. This is done on basis of a relation between manipulated brood size and nestling weight and a relation between nestling weight and recruitment probability. The prediction is far greater than the actual brood size. 7. The negative effect of an increase in brood size on the probability of producing a second clutch was included in the model, now predicting the maximal annual recruitment as a function of the size of the first brood. The predicted brood size shifts to a lower value close to the observed. 8. Measurements of energy content of female Great Tits in the incubation period in relation to temperature reveal a serious loss of fat reserves at low ambient temperatures. Whether these losses are also related to clutch size is not known. 9. The role that these models may play in understanding population dynamics is discussed.