The baryon fraction in clusters, combined with constraints from primordial
nucleosynthesis is currently used to provide a robust upper limit on the co
smological density parameter Omega (0). Current analyses lead to gas fracti
ons at virial radii which are typically of the order of 0.20h(50)(-3/2), fa
voring a low density universe. In this work, we examine critically this iss
ue through the analysis of the baryon distribution in clusters. We find tha
t the currently derived gas fraction profile increases regularly from the i
nner part to the outer part, up to the virial radius, and beyond. Such a sh
ape contrasts with what is expected from numerical hydro-dynamical simulati
ons, in which the gas fraction is more or less constant in the outer region
, reaching a plateau when the contrast density falls off below 10(4). We ar
gue that such a difference is hardly explained by reheating effects, while
taking into account various factors entering into the determination of clus
ters gas fraction may erase such a difference. Indeed, using recent estimat
es on gas content in the outer part of clusters (Vikhlinin et al. 1999) and
applying the correction factor due to the effect of gas clumping, we find
that the gas fraction shape over the range 200 < <delta> < 10(5) is roughly
consistent with hydro-dynamical simulations for a universal gas fraction i
n the range 8-11% (for h(50) = 1), the mass estimators being calibrated fro
m numerical simulations. In contrast, values of the order of 20% do not giv
e acceptable fit to the data on any scale. We conclude that high values of
<Omega>(0) cannot be ruled out on the basis of the baryon fraction argument
.