[plt-scheme] Suspected Error in lngamma Function

This was sent as a PR about PLT Scheme, but it seems to be about the
science.plt code in PLaneT:

mrgordon at mit.edu writes:
> I wanted the equivalent of the gammaln function in Matlab, so I
> checked the online repositories. I included (require (planet
> "science-with-graphics.ss" ("williams" "science.plt" 2 0))). As far
> as I can tell, the lngamma function is supposed to achieve the
> performance of the Matlab gammaln function. Both functions produce
> the same values for arguments greater than .5 or so, but for values
> lower than this the scheme function is WAY off. Are these functions
> not supposed to be analagous or is this indeed a bug?

Robby