In both Tung's textbook and course notes, the following approximation was used: "If a new node sends out an attachment to an existing node having k degree with probability Mk, then from the total m attachments sent out by the new node, there will be a fraction mulk of nodes with degree k having a new attachements." This statement turns out to be only approximately true. The reason is that if sending out the m attachements are truely independent, then there will be some probability an existing node receiving two new attachments. This is clearly not allowed; this implies that the m attachements send out by the new node are not really independent. If Nx is the total number of nodes with degree k, what is the exact formula for the fraction of the Nk nodes that have a new attachment? What happens if Nk > m?