This study quantitatively evaluates the bacterial leakage beneath crowns cemented with a glass ionomer cement compared with those cemented with a polycarboxylate cement or a varnish plus zinc phosphate cement. Bacterial samples were taken from beneath full crowns cemented with the test luting agents 3, 10, and 56 days after cementation. Two-way analysis of variance of the resultant data showed a significant increase in bacterial counts for zinc phosphate, a significant decrease for polycarboxylate, and no change for glass ionomer. These findings suggest that bacterial leakage may not be the reason for the reported poor clinical response associated with glass ionomer luting agents.