Two studies are reported in which the behaviours of four infant class teachers and their respective classes were observed and recorded. The aim of the studies was to examine the effects of contingent teacher touch behaviour upon children's classroom behaviour. Baseline data were collected using an existing classroom observation schedule adapted to include systematic observation and recording of teacher touch behaviour as well as other relevant teacher behaviours and class behaviour. Following the collection of baseline data, teachers were instructed to attempt to restrict their touch behaviour to occasions when they were praising children's classroom behaviour (i.e. to use touch positively and contingently). Observations then continued as before. This intervention resulted in greatly increased contingent touch behaviour by all teachers. The effects on children's behaviour in all cases were to increase on-task behaviour and (where measured) to decrease disruptive behaviour. Reversals to pre-intervention baseline conditions were not attempted, but in the second of the two studies a multiple baseline design across classes was employed which effectively demonstrated experimental control.