Einstein derived the expression for stellar aberration by relating the ray direction cosine in the moving frame to that in the stationary frame. See Article Zur Elektrodynamik bewegter Körper
P 911-912. On Page 911, the direction cosines are related by the expression a' = (a-v/V)/(1- a v/V) where a' is the direction cosine of the ray in the moving system, a the direction cosine in the stationary system, v the velocity of the moving frame and V the velocity of light. For the stellar aberration formula, Einstein explicitly put in the angles, giving cos(ϕ′) = (cos(ϕ)-v/V)/(1- cos(ϕ) v/V). However, in presenting his formula, Einstein says "If we call the angle between the wave-normal (direction of the ray) in the moving system and the connecting line “source-observer” ϕ′, the equation for ϕ′ assumes the form: cos(ϕ′) = (cos(ϕ)-v/V)/(1- cos(ϕ) v/V)". As far as my understanding goes, here the angle ϕ′ is being being replaced by the difference of ϕ′ and ϕ; which is not allowed. It has been pointed out to me by somebody elsewhere that Einstein, later on, changed his original text by replacing the phrase "connecting line 'source-observer' with the expression "direction of motion". (See Note 29, https://einsteinpapers.press.princeton.edu/vol2-doc/345). But this put the stellar aberration angle corresponding to ϕ=Pi/2, (arccos(-v/c)), in the second quadrant, which is contrary to experimental observations.