Several imaging spectrometers, involving three distinct rotation modulation collimator designs, have been simulated on a computer. Data produced by the Monte-Carlo technique were used to reconstruct images using the method of rotational aperture synthesis'. Common to all the simulations were the use of nonposition-sensitive detectors, cylindrical collimator geometry, data accumulation over 360°, and grids producing consecutive harmonics of the lowest spatial frequency. The statistical behavior of 'simple and complex images were studied including the noise produced by the reconstruction method and by the detector background. Aberrations or artifacts were produced by the following: variations of detector efficiency, misalignment of the instrument and rotation axes, time varying backgrounds, parallax (as encountered in laboratory testing), and using only one detector of each frequency pair. Methods were devised to eliminate each of these aberrations or artifacts; most of these methods are exact and do not increase the background noise. To determine the optimum number of detectors to use for a given total area, the reconstructed image of a single point was studied as a function of resolution. If, for a constant flux, one expands the image field proportional to the number of detectors used, the peak height, peak width, sidelobe structure, and background characteristics remain unchanged within statistics. Thus, this method imposes no constraint on resolution; rather, geometry and instrument complexity determine the practical limit.