Abstract.Fast solar wind is known to emanate from polar coronal holes. However, only recently attention has been given to the problem of where, within coronal holes, fast wind originates. One way to gain information on whether the fast solar wind originates from plumes or interplume regions could consist in comparing the elemental abundances in these regions with the ones characterizing the fast wind. Here we present a first attempt to determine the oxygen abundance in the interplume regions by using spectra taken at times of minimum in the solar cycle (when it is easier to identify these structures) by the SUMER spectrograph aboard SoHO. To this end, we analyze spectra taken in 1996 in polar regions, at altitudes ranging between 1.05 and 1.3 R¢ , finding a value £ 8¤ 5 for the oxygen abundance in the interplume regions. From the analysis of the O VI 1032 to 1037 line intensity ratio we also find no evidence of outflow velocities below 1.2 solar radii in interplume regions, while there are indications that outflow motions start to be significant above 1.5 solar radii. The method used and the assumptions made are discussed in light of the derived values. Our values are compared with previous determinations in the corona and solar wind.