This thesis describes a very rapid survey of the nutritional state of children in a small island, Montserrat, West Indies. The anthropometric measurements were analysed in a standard manner to show the degree to which these children were "malnourished". Comparisons with data collected by other workers in the Caribbean were used to demonstrate that the findings in Montserrat were very similar and that any further detailed analysis of the data might therefore be applicable to far larger populations. During the course of the analysis it became apparent that the usual methods of classifying malnutrition were crude and that insufficient attention had been paid to growth, retardation and a deficit in height with its attendant deficit in weight. Attempts were made to develop a new system of analysing the data by expressing the data as percentiles. This analysis proved to be too complicated for routine use and the description of standard data was too limited to permit satisfactory and internally consistent results to be obtained. However, the analysis highlighted the importance of the accurate measurements of length or height. Two studies wore therefore conducted in Northampton and in London nurseries to assess the accuracy of height measurements. The choice o£ cut-off points for each anthropometric measurement, was critically examined end it was concluded that most workers had chosen values arbitrarily. Methods were therefore developed for establishing the appropriate cut-off points for height, weight for height, muscle and arm circumference as well as triceps skinfold thickness. It was concluded that the major problem in Montserrat, traditionally ascribed to malnutrition in fact related to a marked deficit in height with very few children being truly wasted. Arm measurements alone were considered to be somewhat unreliable for assessing nutritional state on their own because they were relatively insensitive measures which were also particularly liable to errors. Finally it wan concluded that the timing and the organisation of the initial survey allowed measurements to be made which are sufficiently detailed for most nutritional assessment surveys.