A series of seven insight problems was given to 28 male and 28 female college students. Each problem required that an object be used in a novel manner to reach solution, but both the critical object and its required use varied across problems. Subjects showed moderate improvement in performance on the second half of the problem series. Males solved more quickly than females; relatively stable and predictable individual differences in performance were observed. Results support the conclusions that insight problems form a class involving common abilities and that general transfer effects can occur if sufficient practice is given.