Tropospheric N2O climatologies are stimulated with the GFDL general circulation/tracer model for three idealized source specifications: (1) a constant surface flux of 1.44¿109 molecules cm-2 s-1 distributed uniformly over the earth's surface with a global source strength of 17 tg N2O yr-1 and an atmospheric lifetime of 131 yr; (2) a 13 month integration of the stable N2O field from experiment (1) with its N2O source removed: (3) a constant surface flux of 2.24¿1010 molecules cm-2 s-1 over only those land areas with precipation in excess of an arbitrary limit of 127 cm yr-1 but with the same global strength (17 tg N2O yr-1) and atmospheric lifetime (131 yr) as in experiment (1). In the boundary layer the model produces an interhemispheric gradient with a minimum in the northern hemisphere (N.H). This is due to the greater downward transport in the N.H. which results in more dilution of N.H. boundary layer N2O mixing ratios by the N2O poor air from the lower stratosphere. The boundary latyer distribution of N2O is also influenced by the distribution of the surface source. The lack of an N2O maximum in the model's N.H. boundary layer suggests that, unlike the model's idealized source, the true source has a large excess in the northern hemisphere. Above the boundary layer the north-south gradient is controlled by the large-scale vertical transport which produces a N.H. minimum in N2O mixing ratio. The impact of the surface source distribution is small. Current measurements at 500 mbar have too low a precision to confirm or disprove the model prediction of an interhemispheric gradient with N.H. minimum in the middle troposphere. The sources of temporal variability in the model's N2O fields are transient motions of all scales acting on mixing ratio gradients, both vertical and horizontal. The model finds that a small surface source of 17 tg N2O yr-1, sufficient to balance stratospheric destruction, is more than able to maintain the observed variability in the boundary layer (?1.0%). The empirical Junge rule relating temporal relative standard deviation to lifetime clearly does not apply to measurements in the boundary layer. There, the variability is dependent on local meteorology and source strength rather than on atmospheric lifetime. It ranges from <0.1% in regions far from any source to 1.0--2.0% in regions with both a source and weak vertical motion. However, the measurement of local mixing ratio time series in the boundary layer, when combined with local meteorological data, should give considerable information about the surrounding sources.