This thesis questions whether the positive manifold effect, first observed by Spearman (1925) is equally influential across samples differing in intelligence level and age. In particular, the studies reported here ask whether positive manifold is less evident in older, compared to younger, children and in samples composed of subjects of higher, as opposed to lower IQ. They also attempt to test Anderson's (1986) theory regarding the role of mental speed in accounting for the strength of Spearman's g. In the first phase of the research (encompassing studies 1 and 2) approximately five hundred subjects, evenly dispersed between the ages of eight and twelve years, were examined using a battery of ten psychometric tests. The amount of common variance amongst tests was assessed in the five age groups and across IQ ranges. There was no consistent evidence to indicate a reduction in the positive manifold effect with increasing age, after correcting for statistical bias. Similarly, there was no tendency for Spearman's g to be more or less pervasive in groups differing in IQ level. In the second phase (encompassing studies 3, 4 and 5) subsamples of nine and twelve years old (c. 60 at each age) were tested with a battery of cognitive tasks thought to access mental speed. These included Inspection Time and two response time measures based on Shepard's Shape Rotation Task and Posner's Letter Discrimination Task (mean RT was the key variable in each): There was no consistent evidence of an age difference in Inspection Time. Similarly, IT-IQ correlations were no smaller in twelve year olds than in nine year olds. Response times were significantly slower in nine than in twelve year olds, for both the Shape and Letter discrimination tasks, but there was no evidence to indicate that RT and IQ correlated to a greater extent in the younger children. Correlations between IT and IQ showed no tendency to be stronger in samples of lower, compared to higher, general intelligence and there was no consistent evidence of stronger RT-IQ correlations in samples of lower IQ. A general tendency was observed for correlations between Inspection Time and IQ to be non-significantly stronger in samples with slower-than-average IT, compared to those with faster-than-average IT. Similarly, correlations between response times for the Shape and Letter discrimination tasks were non-significantly stronger in the group having slower mean Inspection Time. In contrast, no tendency was observed for either measure of response time to correlate more strongly with IQ in the sample having faster IT, neither was there any evidence to indicate that RT-IQ correlations are stronger in samples with longer mean response times. The results, with respect to age comparisons, are interpreted as offering some support for Anderson's theory that age is unrelated to positive manifold, since basic mental speed (here indicated by IT) does not change during development but knowledge (which contaminates RT performance) does. Although there is no evidence that IQ level influences the strength of g, the hypothesis that individual differences in mental speed predict extent of positive manifold has some support, although the relevant results are non-significant. Further research using larger samples and more robust methods for estimating Inspection Time would help to substantiate this theory.