A mathematical model consistent with both visual and tactile reading data and accurately predicting at least one set of visual reading data is presented. This model is based upon the results of a free-rate tactile reading experiment in which word scanning times were measured. A thorough analysis of the results of the experiment showed that, although the average scanning time per letterspace is quite consistent over a wide range of word length and word familiarity (assumed to be inversely related to word rank), the distribution of the word scanning times contains a lot of information about the mental decision processes. From a study of the word scanning time histograms a four-degree-of-freedom model was derived and used to calculate a predicted visual constant-rate reading performance curve. In order to make the calculation, an equation describing the relationship between free-rate and constant-rate reading was also derived. The predicted performance curve is in excellent agreement with the measured data.