While zinc oxide is a promising material for blue and UV solid-state lighting devices, the lack of p-type doping has prevented ZnO from becoming a dominant material for optoelectronic applications. Over the past decade, numerous reports have claimed that nitrogen is a viable p-type dopant in ZnO. However, recent calculations by Lyons, Janotti, and Van de Walle [Appl. Phys. Lett. 95, 252105 (2009)] suggest that nitrogen is a deep acceptor. In our work, we performed photoluminescence (PL) measurements on bulk, single crystal ZnO grown by chemical vapor transport. Nitrogen doping was achieved by growing in ammonia. In prior work at room temperature, we observed a broad PL band at ∼1.7 eV, with an excitation threshold of ∼2.2 eV, consistent with the calculated configuration-coordinate diagram. In the present work, at liquid-helium temperatures, the PL emission increases in intensity and red-shifts by ∼0.2 eV. A peak is observed at 3.267 eV, which we tentatively attribute to an exciton bound to a nitrogen acceptor. Our experimental results indicate that nitrogen is indeed a deep acceptor and cannot be used to produce p-type ZnO.