We analyze a simple dynamic model of the interaction between terrorists and authorities. Our primary aim is to study optimal counterterrorism and its consequences when large terrorist attacks lead to a temporary increase in terrorist recruitment. First, we show that an increase in counterterrorism makes it more likely that terrorist cells plan small rather than large attacks and therefore may increase the probability of a successful attack. Analyzing optimal counterterrorism we see that the recruitment effect makes authorities increase the level of counterterrorism after large attacks. Therefore, in periods following large attacks a new attack is more likely to be small compared to other periods. Finally, we analyze the long-run consequences of the recruitment effect. We show that it leads to more counterterrorism, more small attacks, and a higher sum of terrorism damage and counterterrorism costs. On the other hand, it leads to fewer large attacks and less damage.