Having many siblings, or none, may impair, or improve, a child's development compared to being part of a two-child family. Any effect may vary for different aspects of development. This note describes, cross-sectionally, the observed association between child development at ages 3 to 7 years and the number of co-resident siblings, at three sweeps of the UK Millennium Cohort Study. Indicators of cognitive development (verbal and non-verbal), are taken from surveys at ages 3, 5 and 7 years. Behavioural problems are reported at the three surveys on the Strengths and Difficulties Questionnaire. We analyze its five sub-scales separately, and also the Total Difficulties score. For each of 26 outcomes, we estimate the risk of falling into the most problematic 10% of the child population, depending on the number of siblings at each survey, controlling for the child's gender and the level of the mother's education. In this descriptive exercise, maternal education stands in for a host of possible social covariates; and allows for the least educated mothers having larger families. Children with 2 siblings generally showed increased odds of adverse outcomes, especially in cognition at age 3. This is in line with the hypothesis of resource dilution, but only-children tend, for some outcomes, to score worse than those in two-child families. The odds ratios (ORs) for boys were, except one, unfavourable, of similar magnitude to estimates for larger families. The disadvantages associated with low maternal education were generally greater. These explorations lay the ground for longitudinal modelling of causal pathways.