The present study investigated the effects of small changes in child-staff ratio on observed child and staff behavior in 27 preschools. Numbers of children were manipulated in each preschool to produce a “low” ratio (7.7:1), an “average” ratio (9.2:1) and a “high” ratio (11.2:1). A survey of staff attitudes had revealed a belief that changes in ratio as small as this would have a measurable effect on both staff and child behavior. Observations were carried out using a “target person” procedure, so that only one person (child or staff) was observed at any given time. The results showed a greater effect of ratio on individual child behavior than on individual staff behavior. Children spent more time in large groups, annoyed and teased others more and were less absorbed in what they were doing under higher ratios. The ratio effects were mainly apparent under a ratio difference of about 4:1 with few and inconsistent effects for a ratio change of about 2:1. If the behavior of the whole group (i.e., all children or all staff, rather than the individual) was examined, it was evident that (a) staff were confronted with a substantial increase in problematic child behaviors as the ratio went up (even by 2:1) and (b) the child’s access to staff on an individual or tiny group basis was considerably reduced with the high ratios.