The teaching methods employed by six tutors from three teachers' colleges in Jamaica were classified as investigative or authoritative after observing each twice. The attitudes towards mathematics of 90 of their second-year students, all prospective primary school teachers, were then measured. It was found that students of low and middle achievement indicated more enjoyment of mathematics under an investigative method, whereas high-achieving students enjoyed mathematics more under authoritative methods. Students taught by an investigative method, when compared to those taught by an authoritative method, had a greater tendency to perceive mathematics as a problem-solving process rather than a set -of rules. Belief in mathematics as suitable for the average child was unrelated to teaching method.