So, a while ago I was commenting on how glow worms imitate the night sky or not, as the case may be. Just did a quick check on things yesterday, and was finally able to resolve the issues, I think.
Gould's description of glow worms can now be found in the copy of his book at the internet archive. As all the citations suggested, Gould was only working with simulated data - he doesn't provide any real evidence to support his conjecture that the glow worms are too spread out. But I think I've finally found a picture that can be used as a primative test of Gould's hypothesis, based on this stock footage...
If you do the spatial statistics of this image using Ripley's K relative to a Poisson process, you find that the dots are a little more clustered toward the center of the picture, but also that the worms are not as close together as they should be - just as Gould conjectured. I'd show you the results, but I'm saving them for a homework problem ;)
Of course, this is a pretty crude analysis. The top of the cave, in particular, has its own geometry which is a confounding factor in identifying clustering. It seems a really convincing analysis remains to be done.