Some models for flare statistics predict or assume that there is a relation
ship between the times between flares and the energy of flares. This questi
on is examined observationally using the WATCH solar X-ray burst catalogue.
A rank correlation test applied to the data finds strong evidence for a co
rrelation between the time since the last event, t(b), and the size (peak c
ount rate) of an event, and for a correlation between the time to the next
event, t(a), and the size of an event. A more sophisticated statistical tes
t, taking into account a probable bias in event selection, does not support
the hypothesis that event size depends on t(b) or t(a).