As I have already mentioned, the proportion of β-rays increases with increase of distance from the source of radiation. These rays never occur alone, and for great distances the presence of γ-rays is always discernible. The presence of very penetrating, undeflected rays in the radiation of radium was first observed by M. Villard. These rays constitute only a small portion of the radiation measured by the electrical method, and their presence escaped our notice in our first experiments, so that we believed falsely that the radiation at great distances contained only rays capable of deflection. The following are the numerical results obtained with experiments made by the electrical method with an apparatus similar to that of Fig. 5. The radium was only separated from the condenser by the surrounding air. I shall indicate by the letter _d_ the distance from the source of radiation to the condenser. The numbers of the second line represent the current subsisting when the magnetic field is acting, supposing the current obtained with no field equal to 100 for each distance. These numbers may be considered as giving the percentage of the total α- and γ-rays, the deflection of the α-rays having been scarcely observable with the conditions employed. At great distances there are no α-rays, and the undeflected radiation is therefore of the γ kind only. Experiments made at short distances:— _d_, in centimetres 3·4 5·1 6·0 6·5 Percentage of undeflected rays 74 56 33 11 Experiments made at long distances with a product considerably more active than that which was used for the preceding series:— _d_, in centimetres 14 30 53 80 98 Percentage of undeflected rays 12 14 17 14 16 ─────────────────────────────────────────────── _d_ 124 157 Percentage of undeflected rays 14 11 It is thus evident that after a certain distance the proportion of undeflected rays in the radiation is approximately constant. These rays probably all belong to the γ species. The following is another series of experiments in which the radium was enclosed in a very narrow glass tube, placed below the condenser and parallel to the plates. The rays emitted traversed a certain thickness of glass and air before entering the condenser:— _d_, in centimetres 2·5 3·3 4·1 5·9 7·5 9·6 11·3 Percentage rays not deflected 33 33 21 16 14 10 9 ──────────────────────────────────────────────────────────────── _d_ 13·9 17·2 Percentage rays not deflected 9 10 As in the preceding experiments, the number of the second line approximate to a constant value, when the distance _d_ increases, but the limit is reached for smaller distances than in the preceding series, because the α-rays have been more completely absorbed by the glass than the β- and γ-rays.