During 1997, high mortalities in the first 8 days of the experiment resulted in the
experiment being restarted, with four individuals carried over; three individuals from the
high rations and one individual on low rations. The ration provided to the S. lessoniana
juveniles was doubled from the start of the experiment because high mortality rates in
1996 were thought to be a result of underfeeding. Details of the feeding regime are
outlined in Jackson and Moltschaniwskyj, in press. Ration levels were lowest in the
1996 experiment, and highest in 1997 experiment (almost four times higher). The net
result was that highest ration in 1996 was similar to the lowest ration level in 1997
(Table 2a). Therefore, to allow sensible comparisons we identified three ration levels
across the two experiments. These ration levels were designated low, intermediate and
high (Table 2b), where low and intermediate levels are 45 and 75%, respectively of the
highest ration level.We used the average total food intake per individual over the whole
experiment, however the experiments ran for different times and the animals differed in
age at the termination of the experiments (Table 1). Therefore, to account for these
differences, the age of the animals at the end of the experiment was used as a covariate
in the analyses. If the age of the individuals was significant the adjusted means of the
variable of interest are shown. In the case of S. elliptica, two ration levels were used and
compared, and all individuals were the same age (Table 2c).