Abstract The functional prediction of the Individual Optimization Hypothesis (Pettifor et al. 1988) is explicitly tested, namely that ‘the optimal clutch size is n for birds that choose to lay n eggs, regardless of the actual value of n’ (Högstedt 1980). Over a 10‐year period, we examine the recruitment from 1392 nests of great tits, 575 of which had their brood size experimentally manipulated. We defined final manipulation as the difference between the number of young ringed and the original clutch size. Mean nestling mass declined with final manipulation: those broods from which young were removed had a significantly higher mean fledging mass than control broods, which in turn fledged at a higher mass than those from nests to which young had been added. There was a significant difference in the strength of this effect between years. Addition nests (final manipulation ≥ 2 young) fledged significantly more offspring (x̄ = 10·96, SE = ± 0·16, n = 167) than control nests (final manipulation =−1, 0 or +1 young; x̄ = 8·68, SE = ± 0·07, n = 646), which fledged significantly more young than removal nests (final manipulation ≤ 2 young; x̄ = 5·68, SE = ± 0·09, n = 579). Recruitment per nest (the key measure upon which selection would act since parental survival was independent of manipulation in this study) was greatest for parents rearing a brood size equal to their own clutch size. Averaged over the 10 years, a Poisson model of the number of young recruited per nest showed maximal recruitment to occur at a final manipulation of 0·42 extra young added to the original clutch size. This final manipulation value was not significantly different from zero manipulation (i.e. when parents rear the same size of brood as their original clutch size). Using a restricted data set, where there was no loss of young through egg‐hatch failure and/or loss of very small young, and examining recruitment per nest with respect to experimental manipulation, indicated maximal recruitment to occur at a manipulation of an additional 0·44 young. The recruitment from this experimental manipulation was not significantly different from zero manipulation (i.e. control broods). The original size of clutch laid by a female was indicative of the ‘condition’ of the parents. Manipulation of brood size showed that birds naturally laying larger clutches were ‘fitter’ than birds laying smaller clutches in that, when comparing birds that were all made to rear the same brood size, those with young removed recruited more young than control parents, which in turn recruited more young than those parents to which young had been added to the nests. There were significant differences between years in the shape of recruitment per nest relative to the extent of both final and, separately, experimental manipulation, but in no year did addition broods recruit significantly more offspring than control broods. All these results are consistent with the Individual Optimization Hypothesis – individual optimization via phenotypic plasticity is the most parsimonious functional explanation for the observed variation in the range in size of clutch laid by individuals.