We examine a variety of observations that shed light on the orientation of the semi-major axis of the \eta Carinae massive binary system .
Under several assumptions we study the following observations : The Doppler shifts of some He I P-Cygni lines that is attributed to the secondary ’ s wind , of one Fe II line that is attributed to the primary ’ s wind , and of the Paschen emission lines which are attributed to the shocked primary ’ s wind , are computed in our model and compared with observations .
We compute the hydrogen column density toward the binary system in our model , and find a good agreement with that deduced from X-ray observations .
We calculate the ionization of surrounding gas blobs by the radiation of the hotter secondary star , and compare with observations of a highly excited [ Ar III ] narrow line .
We find that all of these support an orientation where for most of the time the secondary - the hotter less massive star - is behind the primary star .
The secondary comes closer to the observer only for a short time near periastron passage , in its highly eccentric ( e \simeq 0.9 ) orbit .
Further supporting arguments are also listed , followed by discussion of some open and complicated issues .