Well, it does seem that if JSTOR serves 7000 institutions around the world, having that number effectively jump to 7100 for a couple of months shouldn't be a big drama. Also, it's pretty surprising to me that the average number of downloads at MIT was about 1.4 per person per month. (2 million / 100 / 12000 people / 2 months.) I would have expected at least one order of magnitude higher.
if JSTOR serves 7000 institutions around the world, having that number effectively jump to 7100
I understand what you're thinking, but that's not what happened. He downloaded a hundred times MIT, not a hundred times the average institution. You're assuming that MIT is close to average, but there is no reason to think that. I strongly suspect that MIT is a bigger user than (for example) Rye Neck High School or the Scottish Agricultural College.
