We study the thermodynamics of a dilute Bose gas. Not too close to the Bose
-Einstein transition we find results differing from the ones in the literat
ure. As a consequence the low momentum part of the distribution of quasipar
ticules is not a Bose factor, and the effective spectrum is not of the phon
on type. We also discuss the possibility of a Bose-Einstein transition in 2
D.