Chemistry and Physics. 617 



tered light. The visual lines of the lamp do not appear in the 

 upper photograph at all, and the far ultra-violet lines are absent 

 in the lower one. This effect is of the same type as the blue color 

 of the sky, and it is fully accounted for by the accepted theory. 



Similar results were obtained with other gases. Hydrogen 

 gives much less scattering than air, oxygen about the same, and 

 carbon dioxide decidedly more. The scattered light was blue for 

 all of these gases. Finally it was found both visually and pho- 

 tographically that the light scattered by the gas molecules was 

 polarized, the vibrations being transverse to the direction of 

 propagation, just as theory predicts for scattering by particles 

 small compared with the wave-length. — Proc. Roy. Soc, 94 A, 

 453, 1918. h. s. u. 



8. The Occurrence in the Solar Spectrum of the Ultra-violet 

 Bands of Ammonia and of Water-vapor. — A question of interest 

 and importance in connection with the solar spectrum is that of 

 the origin of the thousands of unidentified faint lines which were 

 photographed and catalogued by Rowland and Jewell. In a 

 recent paper by A. Fowler and C. C. L. Gregory are given the 

 results of an investigation which was undertaken primarily in 

 order to determine whether group P in the ultra-violet region 

 of the solar spectrum might not be mainly due to the presence 

 of ammonia in the absorbing atmosphere of the sun. It was 

 already known that ammonia exhibits a remarkable band in this 

 region, having its maximum intensity at about A3360, but the 

 earlier records of the component lines were found inadequate for 

 comparison with the solar tables. Accordingly spectrograms were 

 taken with instruments of various dispersions, ranging up to that 

 of the third order of a grating of 10 feet radius of curvature, 

 an arc between copper electrodes in an atmosphere of ammonia 

 being employed as source in the latter case. 



With regard to the spectrum of ammonia the investigators 

 record the following facts. The chief ammonia band consists of 

 a bright central maximum about A3360, a secondary maximum 

 about A3371, and a number of lines, which occur in groups of 

 three, extending to a considerable distance in both directions. 

 The lines composing the two maxima are very closely crowded 

 together and have been found to be arranged in series of the 

 ordinary type. The components of the triplets are widely sep- 

 arated near the central maxima, but the intervals diminish 

 rapidly until there is final coalescence at A3450 toward the less 

 refrangible side, and at A3287 toward the more refrangible side, 

 where the lines fade out. The triplets, however, are not sym- 

 metrical with respect to the central maxima, and they show 

 marked peculiarities, so that they are very poorly represented by 

 the formulae usually employed for band spectra. 



A comparison of the lines produced by the laboratory source 

 with the corresponding region of the solar spectrum shows that 



