ME. A. FOWLER ON THE SPECTEUM OF MAGNESIUM HYDRIDE. 451 



the poles became heated and entered into combination with the magnesium vapour 

 in the outer parts of the arc. Other experiments have shown that hydrogen is 

 occluded in considerable quantity in the metal, and, in fact, the C and F lines were 

 often present in the spectrum, although no hydrogen was otherwise introduced into 

 the globe. Attempts to completely expel the occluded hydrogen from the metal 

 employed have not been successful. 



In this experiment, if a high vacuum was first produced, the oxide bands were not 

 seen at all, but the hydride spectrum gradually appeared as before. Again, when the 

 residual air was first washed out by hydrogen, the hydride flutings were brilliantly 

 visible as soon as the arc was struck, and the oxide flutings were not observed. 

 When dry nitrogen at atmospheric pressure was introduced into the globe, neither 

 the hydride nor the oxide flutings were seen. In hydrogen at atmospheric pressure 

 the hydride spectrum was present, but was less bright than when the pressure was 

 reduced. 



In bis earlier paper, Mr. BROOKS suggested that water vapour might be an essential 

 factor in the production of the hydride spectrum,* and the experiment on the arc in 

 hydrogen was accordingly repeated, with a quantity of phosphoric anhydride intro- 

 duced into the globe. After standing for 24 hours, however, the flutings appeared 

 at once on striking the arc, as in previous experiments, and with undiminished 

 brightness. In fact, the absence of water vapour would seem to be sufficiently 

 indicated by the absence of the oxide group 5007. Subsequent experiments made by 

 Mr. BROOKS! have similarly led him to conclude that the presence of water vapour 

 cannot be regarded as an essential condition for the production of the hydride 

 spectrum. 



All the experiments accordingly tend to show that magnesium and hydrogen are 

 together concerned in the production of the flutings, and the simplest supposition is 

 that they originate in the combination magnesium hydride. Assuming such com- 

 bination of the two elements, it appears to take place directly under the influence of 

 the arc, especially at low pressures, or of some forms of the spark discharge, or by the 

 combustion of the metal in an atmosphere containing hydrogen either free or in 

 combination. 



It should be noted that the production of the hydride flutings is not the only effect 

 of hydrogen on the magnesium arc spectrum. The enhanced lines also appear very 

 prominently,^ but they are especially developed close to the poles, while the flutings 



* ' Roy. Soc. Proc.,' A, vol. 80, p. 223 (1908). 



t ' Astrophys. Jour.,' vol. 29, p. 187 (.1909). 



| The enhanced lines which appear in the visible spectrum are 4481-31, and the pairs of lines for 

 which the revised wave-lengths are 4434-20, 4428-20 and 4390-80, 4384-86. The polar localisation of 

 these lines as compared with the flutings is particularly marked in a photograph taken with one pole 

 of magnesium and one of iron. In another photograph, where the images of the poles appear to 

 have been successfully kept off the slit, 4481 was relatively weak, and the feebler enhanced lines were 

 wanting. 



3 M 2 



