We have investigated dissociative recombination (DR) of NH + with electrons using a merged beams configuration at the TSR heavy-ion storage ring located at the Max Planck Institute for Nuclear Physics in Heidelberg, Germany. We present our measured absolute merged beams recombination rate coefficient for collision energies from 0 to 12 eV. From these data we have extracted a cross section which we have transformed to a plasma rate coefficient for the collisional plasma temperature range from T pl = 10 to 18000 K. We show that the NH + DR rate coefficient data in current astrochemical models are underestimated by up to a factor of ∼ 9. Our new data will result in predicted NH + abundances lower than calculated by present models. This is in agreement with the sensitivity limits of all observations attempting to detect NH + in interstellar clouds.In the cold ISM, such as in molecular clouds, the most abundant nitrogen-bearing species are expected to be N and N 2 (Langer & Graedel 1989). However, direct observation of these is difficult. Atomic N does not have any low-lying, fine-structure levels that can be populated at molecular cloud temperatures and N 2 lacks a dipole moment, which is needed to provide reasonably strong ro-vibrational transitions. Thus, observations of tracers such as NH, NH 2 , NH 3 , or N 2 H + must be used to infer the N and N 2 abundances through chemical models.Neutral nitrogen hydrides are widely seen in the ISM. Ammonia (NH 3 ) was first detected by Cheung et al. (1968). Later NH 2 and NH were identified by van Dishoeck et al. (1993) and Meyer & Roth (1991), respectively. The observed abundances, however, do not match predictions from astrochemical models. In diffuse interstellar clouds, observed abundance ratio for NH/NH 3 are ∼ 1.7 and for NH 3 /H ∼ 3.2 × 10 −9 . These cannot be simultaneously explained by existing chemical models. The models can fit either one of the observed values but then the other predicted ratio is a factor of 10 off from the observation (Persson et al. 2010). Similarly, in dark clouds the abundance ratio for NH/NH 3 is underpredicted by more than an order of magnitude (Hily-Blant et al. 2010). These discrepancies possibly originate from using incorrect reaction rate coefficients in the models.