2004MNRAS.351..967S

From the 389 OGLE-III 2002 observations of Galactic bulge microlensing events, we select 321 that are well described by a point-source point-lens light-curve model. From this sample we identify one event, 2002-BLG-055, that we regard as a strong planetary lensing candidate, and another, 2002-BLG-140, that is a possible candidate. If each of the 321 lens stars has one planet with a mass ratio q = m/M = 10^–3 and orbit radius a = R_E, the Einstein ring radius, analysis of detection efficiencies indicates that 14 planets should have been detectable with Δχ² > 25. Assuming our candidate is due to planetary lensing, then the abundance of planets with q = 10^–3 and a = R_Eis n_p~ n/14 = 7 per cent. Conversion to physical units (Jupiter masses, M_Jup, and astronomical units, au) gives the abundance of `cool Jupiters' (m ~ M_Jup, a ~ 4 au) per lens star as n_p~ n/5.5 = 18 per cent. The detection probability scales roughly with q and (Δχ²)^–1/2, and drops off from a peak at a ~ 4 au like a Gaussian with a dispersion of 0.4 dex.