Osterburg, Parthasarathy, Raghavan, and Sclove (1977) developed a model for assigning a probability to a fingerprint, including a partial print, based on the number of Galton characteristics (ridge endings, forks, etc.) present. A grid of 1-mm squares was placed over a fingerprint. A multinomial model for the cells of the grid was used, the categories being the individual characteristics and combinations of them. In the present article this model is extended to consider dependence among the cells of the grid. The occurrence of the characteristics is modeled as a two-dimensional Markov-type process.