There was a linear, somewhat variable, relationship
between cumulative rat damage, as measured by
the percentage of cut tillers, and yield (Fig. 2). The
slope of this regression was the same in the wet and
dry season, and for every 1% increase in rat damage,
there was a decrease of 58 kg/ha in rice yield.
Damage measured only at tillering could not define
this relationship because of smaller samples and
scattered data, whereas damage at booting or at
ripening produced similar relationships as that
shown in Fig. 2.
Damage to rice tillers assessed only during the 2
weeks prior to harvest needed to be multiplied by 4.2
(95% confidence limits 2.9–7.8) to obtain an estimate
of percentage cumulative yield loss