i looked at the code, and it looks like the top observers by species list probably used to work as expected, but a recent change to the page introduced the problem. it looks like the intent of the change was to make fewer API calls, but i wonder if instead of making 1 API call to return 200 records, it could have been just as or even more efficient to make 2 API calls to return 2 x 5 = 10 records?
@pleary will probably be the one who would have to say whether the new behavior that sacrifices accuracy for performance is the best solution.