A description is given of an alpha-ray spectrograph, consisting of the Princeton cyclotron magnet and a Plexiglas deflection chamber, in which the alpha-particles can be bent into a semi-circle of about 80-cm maximum diameter. Three different methods of detection have been employed according to the different strengths of the radioactive sources. They are the ordinary photographic method, the counting method, and the method of photographic tracks. The behavior of the spectrograph has been investigated with Po alpha-particles by the three methods of detection. The forms of the energy distribution have been determined, respectively, by these three methods and agree fairly well with one another. The half-width of the main line under favorable conditions is less than \textonehalf{} mm, which is equivalent to about 0.01 Mev. The intensities a few millimeters away from the maximum are less than 0.1 percent of the main line intensity. The track method of detection has been used to study both the low and high energy regions of the main line of the Po alpha-particles. Microscopic examination reveals distinctly a series of weak groups in the low energy region, while in the high energy region no indication of any discrete group has been found. These experiments have been carefully repeated with three different sources and under different experimental conditions. In each case similar results have been obtained. The emission of these different alpha-particle groups from Po may, as in the case of ordinary fine structure, leave the product nucleus, Pb, in different excited states. Then, when the Pb nucleus falls from one of these quantum states to the normal state, a gamma-ray quantum is given off. As far as energy and intensity are concerned, these weak groups of alpha-particles are generally compatible with the gamma-ray lines from Po as measured by Bothe. However, their intensities are not in agreement with the current theory of alpha-decay, as can be seen from the abnormally large spin changes deduced from the theory for the different alpha-transformations, and also from the large deviation of the Geiger-Nuttall curve for these groups from the curve for the members of the Ra family. Possible explanations of these short-range alpha-particles from Po have been suggested and discussed.