This study sought to determine the relative reliability of indirect cephalometric measurements and direct anthropometric ones taken in norma frontalis of 25 dry adult human skulls. Six of the 11 linear projective measurements were singular and located in the orbital, middle, and lower parts of the face, with two from each part. Two of the five paired measurements were taken in the orbital region on both sides, and the other three were taken in the middle to lower face between the midpoint of the facial axis and the landmarks lateral to it. Both singular and paired cephalometric distances were significantly shorter than the anthropometric distances. Mean numerical differences were much greater in paired measurements than in singular ones. The differences between these two sets of findings are a result of the uneven position of the landmarks used for measurement, located, as they are, on different planes of the face. These differences are undetectable by two-dimensional cephalograms.