Fifty four five-year-old children in three London Primary Schools were tested using the British Ability Scales (NFER). Every 'product' of the children's construction work for the term was photographed. Three intervention groups were withdrawn from their classroom for one hour per week for focused instruction. 'Access only' groups were also withdrawn but were not provided with instruction. The children in the second control group received no additional experience of making beyond that normally provided in the classroom. A total of 450 products were constructed by the children during the intervention phase and each has been categorised from an analysis of the photographs and according to principles of structures and mechanisms. The evidence suggests that while the intervention group constructions became increasingly complex, the performance of the 'access only' group actually deteriorated in terms of both quantity and quality. No significant improvement in BAS scores between the pre-test and post-test was found yet the evidence did suggest that learning had taken place, and that children in the intervention groups produced more elaborate constructions than any of those in the control groups.