The dwarf planet Eris (2003 UB313, formerly known also as ``Xena'') is the largest KBO up to now discovered. Despite being larger than Pluto and bearing many similarities with it, it has not been possible insofar to detect any significant variability in its light curve, preventing the determination of its period and axial ratio. We attempt to assess the level of variability of the Eris light curve by determining its BVRI photometry with a target accuracy of 0.03 mag/frame in R and a comparable or better stability in the calibration. Eris has been observed between November $30^\mathrm{th}$ and December $5^\mathrm{th}$ 2005 with the Y4KCam on-board the 1.0m Yale telescope at Cerro Tololo Interamerican Observatory, Chile in photometric nights. We obtain 7 measures in B, 23 in V, 62 in R and 20 in I. Averaged B, V, and I magnitudes as colors are in agreement within $\approx 0.03$ mag with measures from Rabinowitz et al. (2006) taken in the same nights. Night-averaged magnitudes in R shows a statistically significant variability over a range of about $0.05\pm0.01$ mag. This can not be explained by known systematics, background objects or some periodical variation with periods less than two days in the light-curve. The same applies to B, V and to less extent to I due to larger errors. In analogy with Pluto and if confirmed by future observations, this ``long term'' variability might be ascribed to a slow rotation of Eris, with periods longer than 5 days, or to the effect of its unresolved satellite ``Dysnomea'' which may contribute for $\approx0.02$ mag to the total brightness.