Since the preliminary announcement of the discovery of radium, Giesel[19] has devoted a great deal of attention to the separation of radium, polonium and other active bodies from pitchblende. He was indebted for his working material to the firm of P. de Haen, of Hanover, who presented him with a ton of pitchblende residues. Using the method of fractional crystallization of the bromide instead of the chloride, he has been able to prepare considerable quantities of pure radium. By this means the labour of final purification of radium has been much reduced. He states that six or eight crystallizations with the bromide are sufficient to free the radium almost completely from the barium. =13. Spectrum of radium.= It was of great importance to settle as soon as possible whether radium was in reality modified barium or a new element with a definite spectrum. For this purpose the Curies prepared some specimens of radium chloride, and submitted them for examination of their spectrum to Demarçay, an authority on that subject. The first specimen of radium chloride examined by Demarçay[20] was not very active, but showed, besides the lines due to barium, a very strong new line in the ultra-violet. In another sample of greater activity, the line was still stronger and others also appeared, while the intensity of the new lines was comparable with those present due to barium. With a still more active specimen which was probably nearly pure, only three strong lines of barium appeared, while the new spectrum was very bright. The following table shows the wave-length of the new lines observed for radium. The wave lengths are expressed in Ångström units and the intensity of each ray is denoted by a number, the ray of maximum intensity being 16. Wave length Intensity Wave length Intensity 4826·3 10 4600·3 3 4726·9 5 4533·5 9 4699·6 3 4436·1 6 4692·1 7 4340·6 12 4683·0 14 3814·7 16 4641·9 4 3649·6 12 The lines are all sharply defined, and three or four of them have an intensity comparable with any known lines of other substances. There are also present in the spectrum two strong nebulous bands. In the visible part of the spectrum, which has not been photographed, the only noticeable ray has a wave length 5665, which is, however, very feeble compared with that of wave length 4826·3. The general aspect of the spectrum is similar to that of the alkaline earths; it is known that these metals have strong lines accompanied by nebulous bands.