[yungchin]

It's still surprising that arxiv is that high up. You'd think most people would cite the definitive publications, not the preprints at arxiv?

Anyway, given that in some fields _everything_ that's written goes to arxiv, arxiv will [by definition][1] have a very high h-index. The thing is, the h-index was conceived to compare individuals, not journals.

[1]: http://arxiv.org/abs/physics/0508025

[streptomycin]

Google Scholar also indexes papers that are only published on arxiv, which likely often include citations to other things on arxiv.