Why is "number of unique species" an interesting metric? If island A had three species of deer and island B had one species of deer and one species of wolf, island A would win under this measure, though B might seem more diverse.<p>Additionally, defining a species is a really messy problem (see <a href="https://en.wikipedia.org/wiki/Species_problem" rel="nofollow">https://en.wikipedia.org/wiki/Species_problem</a>). A more objective measure, like fraction of shared genome, might make more sense in this context.
It would be nice to see the absolute numbers of species for the other contenders. I don't feel that percentage of endemic species is a useful metric.