860 Mr. Gr. A. Hemsalech cuid the Comte de Gramont on 



liquid air (b) , in nitrogen (c), and in hydrogen (d). AH these 

 spectra were obtained with arcs at 80 volts. The photographs 

 show up clearly the remarkable behaviour o£ the spark line 

 A 4481 under these various conditions. The relatively great 

 strength and breadth of this line in liquid air and in hydrogen 

 are in marked contrast to its feeble development in air and 

 to its tenuity in nitrogen. The considerable expansion of 

 this line during the first phase of the hydrogen arc is another 

 feature to be noted. Compare also the " winged " lines of the 

 Rydberg series in a, b, and d with the better denned and more 

 perfectly developed lines of the same series in nitrogen (c). 

 Further note the relative feebleness of the low-temperature 

 flame line A, 4571 in liquid air and in hydrogen and the sym- 

 metrical widening of the blue hydrogen line A4861 as a result 

 of the Stark effect. 



Plate XVIII. gives examples of the effect of liquid films 

 upon various types of magnesium lines. Note the great 

 expansion of the spark line \ 4481 in the first phase of the 

 water arc (a) and the strong red wings of the two lines A 4^352 

 and A, 4703 of the Rydberg series. Remarkable is the great 

 feebleness of these lines during the first phase of the paraffin- 

 oil arc. On the other hand, note the enhancement of the 

 triplet at A 5184 in the first phase of the same spectrum. 

 Another feature of these three spectra, which were obtained 

 with 80 volt arcs, is the rapid falling off in intensity of 

 the spark line X 4481 after the end of the first phase. 



The first two photographs on Plate XIX. were obtained 

 with a 200 volt arc, a being the spectrum of the ordinary 

 arc between magnesium electrodes in air, and b that of the 

 same arc with a drop of water between the poles. The 

 most outstanding feature of these two specira is the extra- 

 ordinary behaviour of the spark line A, 4481. Hardly 

 perc. ptible in the ordinary arc, it occupies a position of 

 significant prominence in the water arc not merely by 

 reason of its relative strength, but more particularly on 

 account of its sudden cessation during the second phase 

 whilst all the lines of other types continue to be emitted 

 without any change in intensity. Note the relative weak- 

 ening of the flame line A, 4571 and the better definition of 

 A, 4352 and A, 4703 of the Rydberg series in the second 

 phase of the water arc. The third photograph {c) represents 

 a portion of the ultra-violet region of the magnesium spectrum 

 as given by a capacity spark, water arc and ordinary arc, 

 all taken between electrodes of metallic magnesium. Note 

 the reversals of the flame line of X2852. 



