A method of representing multivariate data as cartoon faces was originated by Chernoff (1971). It is useful for a first look at or overview of a set of data. The ruga pattern is a complex structure which under purely visual observation is meaningless, and as a set of data it also says nothing to the casual observer. As a Chernoff face, however, the visual impact is immediate. Palatal rugae as Chernoff faces have now been used to try to establish family groupings and possibly to identify the parentage of a child. Three families out of a sample of 17 were randomly selected. Ruga pattern classification was carried out and 10 of the variables which are not affected by growth were used to draw faces. Twenty-four observers were asked to match each child with a set of parents. The number of correct matchings were then counted, as were the incidences of similar matching correct or incorrect. It was found that matching was easier in certain families (observers consistently achieved higher scores), that the incidence of 2, 3 or 4 correct placings was constant and that observers tended to be consistent in their matching (right or wrong). ; Article