Mortality levels near 40% during the experiments using juvenile S. lessoniana were
higher than desirable. Although there was no evidence that mortality was due to ration
level there is insufficient replication required for a rigorous frequency analysis. The
increase in feeding levels in the second experiment was explicitly intended to minimise
mortalities. The isolation of the animals is an artificial condition, particularly since S.
lessoniana juveniles display schooling behaviour several weeks after hatching. However,
this isolation provided us with the ability to control and record exactly the food intake of
each juvenile. We were also able to remove the effects of feeding hierarchies and
cannibalism that occurs among individuals when schools of juveniles are maintained in
captivity.