The research undertaken in this thesis was based upon the desire to establish as early as possible in school life those pupils who have the potential to do well in science. Evidence is presented from the literature to suggest that spatial ability is an important indication regarding this and one that is not normally tested within schools. The fact that males show an average superior performance on such tests is discussed and the current available theories for this are explored. A possible view which integrates these presently dissociated theories is given. As formal thinking and the appearance of the sex differ­ence in performance on spatial tests are both associated with adolescence, the former is discussed and an argument is made to support the view that formal thinking possesses a spatial element. Based on the evidence linking spatial and scientific abilities a test of the former was developed and after pretesting was used in schools to compare pupil perform­ance in school with their attainment in the test. The results of these testings are given and their implica­tions discussed both regarding indication of scientific ability and also development of spatial ability in adolescence. The latter raises the problem associated with this and similar studies where the length of study prevents a great deal of longitudinal investigation. The thesis concludes that the developed test is of benefit to the classroom teacher and argues that the 'scientific nature' of such studies is not lost by the inclusion of subjective views.