Polarisation by Reflection - oscillation direction I'm currently studying polarisation by reflection, and have come across two pieces of information from the same source, which I can't seem to understand on how they differ.


*

*The oscillation direction of the field polarised in the plane of incidence is parallel to the direction of the reflected ray.

*As EM waves are transverse waves, this polarisation cannot propagate along the reflected ray.
Could someone please explain, the direction of oscillation of the reflected ray and what each piece of information means.
 A: Finally found out, what I was getting so confused about, here's the answer (credit to Dr Sebastian Steinlechner) with a relevant diagram.
The incoming light is assumed to be unpolarised. We can, however, describe it as a combination of two orthogonal polarisations: one is polarised in the plane of incidence (the arrows in the picture), and the other is perpendicular to that plane (indicated by the dots). 

Looking along the ray hitting the glass, you thus have one E field component that oscillates left and right (dots) and one that oscillates up and down (arrows). Now, look at what happens to these two components at the interface between glass and air. The incoming ray is split into two new rays, the refracted and the reflected ray. Their electric fields need to come from the original polarisation components, i.e. their polarisation is given by some combination of the polarisation components of the incident ray. 
Now imagine standing at the surface of the glass, looking along the refracted ray. Then the dot-polarisation component is still oscillating left and right, while the arrow component is oscillating up and down. 
At Brewster's angle, the reflected ray is exactly at 90deg from the refracted ray, i.e. looking towards the reflected ray you'd still see the dotted component oscillating left and right, but the arrow component would now oscillate along the direction of the reflected ray. 
Remember that electro-magnetic waves are transverse waves, i.e. their direction of oscillation always has to be perpendicular to the propagation direction. Here, however, the arrow component would be in the direction of propagation. 
Thus, only the dotted component will get reflected, while the arrow component is completely refracted/transmitted -- you're left with a perfectly polarised reflected ray, and a slightly polarised refracted ray.
A: I have a site which also explains this phenomenon clearly here.
In short here you would find that when an unpolarised light hit at the interface of the two medium the reflected light will be generated such that only that component of electric field is oscillating, which is perpendicular to the direction from point of incidence to the point of observation which will also be the path for propagation.
And one would see that this component is parallel to the plane of incidence and and a component of the unpolarised light. 
