In the helium method of age determination the age of the material to be tested is given by the ratio of the content of helium to the content of radioactive elements in the material. In an unaltered igneous rock, where presumably the minerals are of the same age, it would be expected that the distribution of helium in the various minerals should correspond with the distribution of the radioactive elements. In each mineral the ratio of helium to radioactivity should be the same. This was found to be not the case. Pyroxene and feldspar were separated from six samples of Triassic diabase. Age measurements on each yielded mean age ratios of 103 million years for the pyroxene samples and 36 million years for the feldspar samples. Six samples of late Triassic magnetite from ore occurrences in West Virginia, Pennsylvania, and Nova Scotia all gave age ratios closely grouped about a mean value of 134 million years. The constant age ratio, despite widely varying contents of radium and thorium, suggests that the magnetite has retained most of its helium. Additional evidence on the high helium retentivity of magnetite is shown by further work on magnetite specimens of various geological ages. These age ratios appear to show a sequence and spacing compatible with geological knowledge, and the results are in fair agreement with corresponding age ratios given by the lead method.