Tropospheric photochemistry is central to terrestrial climate change and pollution effects and so will be modelled on global 3-D grids. The chemistry is also complex, numerically stiff and kinetically nonlinear. A packaged family based integrator has been developed specifically to combat the difficulties associated with computational modelling of atmospheric chemistry on the global scale. The present work describes its ability to reproduce the major nonlinear features of tropospheric kinetics--those relating the nitrogen oxides (NO x ) and oxidizing organics to ozone. It is shown that the family modules can duplicate typical changes in ozone production as a function of NO x level while consuming a minimum number of mathematical operations. The tests are first performed in the box model mode for a variety of pristine and pollutant scenarios. Zero-dimensional runs are patterned largely after the nonlinearity investigations of Liu and coworkers. The testing is then extended to column representations for vertical mixing of ozone precursors in convective storms. Here the calculations follow the climatology of ozone production enhancements assembled by Pickering and colleagues. Benchmarking is reported for a mechanism containing full inorganic kinetics as well as decomposition sequences for six nonmethane hydrocarbons. Chemical species in the simulation number 92. The operations count is roughly 10,000 per cell step for time increments of 1 h or more. The coding should thus enable decadal scale runs on massively parallel processors. Scaling experiments indicate full vectorization has been achieved. The chemistry packages are optimized not only for speed but also for convenience. Modularity and routines automating setup of solutions to the kinetic continuity equations are outlined as incidentals.