The two main aims of the research presented in this thesis are firstly; to develop theoretically valid methods of distinguishing between dyslexic and non dyslexic poor readers (ND-PR) through the use of behavioural tests in a range of primitive skill areas and secondly; to develop and evaluate three intervention programmes for learning disabled children based on sound theoretical principles. The implications of these findings for the traditional discrepancy definition of dyslexia are considered. In the first set of studies, dyslexic and ND-PR were tested at 8 and 10 years of age on a battery of theoretically chosen tests of primitive skills. As expected from the literature, both the dyslexic and ND-PR showed difficulties in phonological skills (Bradley and Bryant, 1983; Snowling et al., 1986; Vellutino, 1979). However, by contrast the dyslexic children also showed difficulties in certain tests of cerebellar dysfunction, whereas the ND-PR did not. This dissociation presents evidence for the cerebellar impairment hypothesis (Nicolson, Fawcett and Dean 1995). Findings give early support for the supposition that the phonological deficit theory may be subsumed within a broader causal framework of cerebellar impairment. Findings also suggest that there is value in retaining the discrepancy definition of dyslexia. The second set of studies compared three types of training on skill acquisition for children of varied ability ages 5 to 6 years. The groups were given systematic training over a period of several weeks using a phonological, motor and arithmetic skill programme. Training was designed for administration by a relatively unskilled instructor. Parents successfully delivered the first training programme, the author the second. Promising results were reported for both small-scale studies. Persistent performance improvement in all training groups were shown in measures of reading and spelling age and IQ, together with cautious evidence of skill transfer. Interpretation of the results suggests that each of the training programmes had generic value for metacognition. The findings provide a demonstration of both the effectiveness and cost-effectiveness of this type of structured and regular training intervention with young children, particularly those with learning difficulties. Implications for both the phonological and cerebellar impairment hypotheses are discussed. It is proposed that a motor skill training programme in conjunction with a phonological training programme has potential in any home/school based intervention.