Re: Cat's eyes

[John Denker <jsd@MONMOUTH.COM>]

At 02:06 PM 9/3/00 -0400, Ludwik Kowalski wrote:
> If I understand you correctly, John, you are saying that my
> eyes reflected the kitchen light toward cat's eyes and then
> the light was reflected back to me.

Please read what I wrote.  I said "face" not "eyes".  I also mentioned
objects projectively near your face (as seen from the cat's
viewpoint).  Also note the light is diffusely scattered by your face, not
specularly "reflected".

> I am not convinced that this is a correct explanation. The orientation of
> my head would be critical, but it was not. Cat was also not in a frozen
> position.

I don't see why orientation should be critical to a diffuse scattering
situation.

The orientation of the retroreflector should be immaterial over a wide
range.  That's the defining property that distinguishes a cat's-eye
retroreflector from a simple mirror.

> While quickly browsing Feynman's Chapter 35 I could not
> find anything about this effect. Where does he discuss the
> retroreflection of the eye? That word is not in the index.

My apologies, the information is in chapter 36 not 35.  See the small
middle paragraph on page 36-4 and contrast it with the middle paragraph on
page 36-9.  The reflector part is not explicitly mentioned, but if you
think about it you'll realize that by putting a mirror right behind the
active layer you can make the light pass twice through the active
layer.  Most vertebrates that are active at night have such mirrors.

> I am familiar with the retroreflection in a corner cube. If cat's
> eyes were corner cubes then they would always reflect light
> back to a source regardless of cat's motion.

Yes.  That is consistent with the observation (reported above) that the cat
was not in a frozen position.

> Why do cat's eyes act as retroreflectors? Are there any
> corner cubes inside of them?

Please read what I wrote.  I said that the cat implemented the
retroreflector as a lens with a mirror in the focal plane, and I contrasted
this with the corner-cube implementation just to be extra-explicit.  I also
suggested that you diagram a lens with a mirror in the focal plane and
trace a few rays.  If you had done that you would have known that this
makes a fine retroreflector, with performance comparable to the other
implementation.