Continuous cultures of Skeletonema costatum, Chaetoceros muelleri, Rhinomonas reticulata and Pavlova lutheri were subjected to phosphorus or nitrogen limitation at two light intensities in order to manipulate their biochemical composition. They were then fed to Mytilus edulis larvae over a 2-week period, and the larval growth and mortality were assessed. All monospecific diets supported growth, sometimes equal to or better than a control diet containing a mixture of species ( R. reticulata and P. lutheri). Survival was not affected by the diets but significantly better larval growth was obtained with low-light, nitrogen-limited or non-nutrient-limited cells of S. costatum and high-light, nitrogen-limited or low-light cells of C. muelleri grown under no limitation or nitrogen limitation. High-light phosphorus-limited or low-light non-nutrient-limited R. reticulata cells and high- light phosphorus-limited cells of P. lutheri were also superior in relation to their counterparts. A novel computer-aided image analysis technique was used for measuring the length of the larvae. A multidimensional model was used in an effort to correlate algal biochemical components with larval growth. Some fatty acids were found to be significant in determining the algal nutritional value, with protein and carbohydrate playing a secondary “modifying” role. In P. lutheri, the 16:0 and saturated fatty acids (SAFA) were significantly positively correlated with larval growth while the contrary was found for dietary n−3 fatty acids, suggesting their strong negative effect on larval growth. Similar results were found in R. reticulata, although in this case, both protein and carbohydrate content were found to determine the algal nutritional value concurrently with some fatty acids. In the other two diatoms, S. costatum and C. muelleri, no consistent relationship could be established, thus suggesting that either there is a species-specific relationship or that other components are important in determining the algal nutritional value.