A researcher uses the Benjamini-Hochberg (BH) procedure to control the False Discovery Rate (FDR) at level $q = 0.05$ across 100 independent tests. Given sorted $p$-values $P_{(1)} \le P_{(2)} \le \dots \le P_{(100)}$, what is the condition for rejecting the hypothesis $H_{(i)}$?