When using the Benjamini-Hochberg procedure for $m=200$ tests at a False Discovery Rate $\alpha = 0.05$, if the sorted $p$-values are $P_{(1)} \le P_{(2)} \le \dots \le P_{(200)}$, what is the specific condition that determines the largest index $k$ such that all null hypotheses $H_{(1)}, \dots, H_{(k)}$ are rejected?