[jballanc]

I wouldn't give too much weight to that graph. To date, the biggest challenge with H5N1 is reporting. Primary outbreaks have occurred in parts of the world where the monitoring and reporting networks are sadly inadequate. It is likely that if you had full reporting on all infections, and not just those that resulted in mass fatalities, that the actual mortality would be less. Also (see my other comment), it is typical for a virus's mortality to be roughly inversely proportional to its infectivity.

Edit: I should add, however, that the H5N1 threat is very real and very troubling. Not end-of-the-world troubling, but set-humanity-back-a-handful-of-decades troubling to be sure...