Top: Collisional part (for models 1 and 2) of the O VI λ1032 intensity along the equator. Bottom: Abundance inferred from these profiles (as well as those for Lyβ) applying eq. (7). For the results of both model 1 and model 2, we assumed that the streamer closed region oxygen abundance is depleted by a factor of 1/100. The departure of the inferred values from that number is due to projection effects, which become especially important at 1.75 ﹩R_{\odot }﹩ and higher.