The achievement scores of 1532 Year 10 students in the Northern Territory on a common external assessment task in mathematics were analysed to ascertain whether or not there were significant differences, both statistical and educational, between the performances of males and females. The students were streamed by schools into one of three levels; Level One (advanced), Level Two (standard) and Level Three (basic). The analyses were conducted on overall aggregated scores, aggregated curriculum strand scores, individual question scores, individual question part scores, students' responses, and on the distributions of males' and females' scores in particular mark ranges. The study found significant differences in the performance of males and females in both Level One and Level Three, but no significant differences at all in Level Two. In Level One, the differences were found in the mean achievement scores in the Number and Measurement strands and on individual questions from these strands. In Level Three, gender differences were found in overall aggregate scores, in scores for the Number, Measurement, Geometry, and Functions and Graphs strands, in questions from these strands and in the proportional representation of males and females in the top decile of scores. Results of content analyses found educationally significant patterns in the responses of males and females which suggest that females sometimes perform at levels below those of males on some multiple choice tasks, that males in Level One had a better understanding of exponential growth than their female counterparts and that males in Level Three were more able to use correctly the mathematical conventions in rounding amounts of money. Whilst these three findings were isolated, they support the findings of much previous research which indicate that males and females do perform differently on some types of mathematical tasks and on different strands of mathematical content.