Recall from (3.5) that one is assuming to be much larger than any given function of , or equivalently that is larger than any given function of that goes to zero. In particular, one can use (3.5) to make less than, say, , and then by choosing appropriately one can obtain the display after (3.8).

]]>makes it unclear what type of c function is intended. ]]>

Specifically, I’m stuck on page 36 on the equation that occurs after the phrase “whenever i >= 16/ (epsilon ^ 2).

I understand equation A 1. which involves a sum of products

of terms f_i(v) * g_i(w).

However, when you multiply one row the of adjacency matrix by g_i(w), and

take the sum, I’m very puzzled that so many of the f_i(v) terms seem to melt away.

Thanks for your help.

Paul Epstein

]]>