We have studied the long-term variability of LMC X-3 in optical light curve
s spanning 6 yr, in order to search for optical signatures which could conf
irm or refute the suggestion that the 'modulation' is the result of accreti
on-rate variability rather than accretion-disc precession. We find that the
re is no stable period in the optical light curves, that the optical and X-
ray light curves are correlated, with an X-ray lag, and that there is no mo
dulation in the optical colours. We argue that these effects agree with the
variable mass-accretion rate model proposed by Wilms et at.