Red Shift - A Practical Test

[ma...@geoffgrayer.force9.co.uk]

If you have stopped laughing at my last email, I can suggest a simple test, which, if successful, would also give some information about the number of interactions, but which, if unsuccessful, would probably invalidate the idea.

 

The red shift is defined as (fe - fo)/fe, where fe is the emitted line frequency, and fo is the observed frequency. My claim is that:

 

    (fe - fo)/fo ~ N, the mean number of interactions (scatters).

 

But there will be an uncertainty in the number of scatters  ~ SQRT(N).

 

So the apparent increase in the width of the line will also be  ~ SQRT(N).

 

hence if one plots dfo (the observed width of a particular line) vs distance, one should see

 

    (dfo - dfe)/fo ~ SQRT(N)/N

 

For large red shifts, dfe is probably negligible, so one has

 

    dfo/fo ~ 1/SQRT(N)

 

Now, I have noticed that illustrations of spectra of objects with large red shifts normally exhibit very large line widths also. How is this explained in conventional

cosmology?

 

Geoff

 

Indeed (assuming the original width is negligible)