This paper presents the final report of an extended study which assessed the viability of individualised tertiary mathematics instruction in Australia. Data collected over five years between 1978 and 1984 gave information on three treatments: a personalised system of instruction based closely on the Keller Plan with no final examination, a conventional lecture course with final examination and a modified personalised system including lectures. Long-term retention of concepts was used as the major criterion for judging success of the methods. Results favouring students taught in the conventional manner have led to a questioning of the original precepts of the personalised system as developed by Keller.