208 CARNEGIE INSTITUTION OF WASHINGTON. 



photographic plate. Heretofore only about one explosion out of three was 

 recorded. A new and accurately made slit has also been provided. 



The explosion of tungsten, aluminum, and lead wires in a very high vacuum 

 was undertaken by Smith, and after many trials was accomplished successfully. 

 In order to be sure that the wires exploded properly, each flash was recorded 

 on a photographic plate. The photographs show plainly what occurs, 

 especially in the case of tungsten wire. If the vacuum is insufficient, the 

 wire, when it becomes hot enough to melt, breaks into a number of short 

 segments, some of which may be melted into little globules. These are 

 projected against the sides of the glass bulb, where they are reflected, a small 

 globule or short segment of the wire sometimes traversing the space within 

 the bulb several times before it cools to such an extent that its light no longer 

 affects the photographic plate. Since the wire may break into dozens or even 

 hundreds of small pieces, the photograph occasionally is very striking. If 

 the vacuum is sufficiently good, however, the explosion is normal, and the 

 photograph shows only the uniform cloud of incandescent metallic vapor. 



In every case of a perfect explosion no trace of gas was found, the bulb 

 remaining non-conducting after the explosion, with 26,000 volts applied. 

 It is therefore certain that with the apparatus employed no decomposition of 

 the heavy metallic atoms into helium or other gases takes place. On the 

 average, about a week was required to prepare the bulb, bake it out, and 

 secure the necessary vacuum for these experiments. 



Interesting results have been obtained with the vacuum-spark spectra of 

 metals. The discharge of the large condenser was passed between metallic 

 electrodes separated about 2 mm. in a good vacuum. The maximum value 

 of the momentary current flowing through the gap was about 10,000 am- 

 peres, and in the case of most of the metals tried the spark was extremely 

 brilliant. The spectrum which was photographed with a concave-grating 

 spectrograph in both first and second orders is very rich in lines. The lines, 

 due to the residual gas, are weak or entirely absent, depending upon the metals 

 which are used as terminals. With aluminum a few oxygen lines show, while 

 with zinc or cadmium no certain trace either of oxygen or nitrogen has been 

 found. Calcium, the first metal investigated, gives a spectrum consisting of 

 about 900 lines in the ordinary photographic region. Of these, approximately 

 75 are known hues of calcium or of impurities, mainly magnesium and stron- 

 tium. The remaining 800 or more lines appear to be chiefly lines of calcium 

 in a higher state of ionization than that occurring in the arc or ordinary spark. 

 Many of the lines are very strong. 



A large condenser is now under construction for work on wire explosions and 

 the vacuum spark and, in July, the 218 glass plates were received from the 

 manufacturers. The proper design of this condenser, to insure permanence 

 and efficiency in operation, is a problem which demands a considerable amount 

 of experimental work. When completed, the condenser will have nearly the 

 same capacity as the one now in use, and if, as we hope, it will withstand 5 

 times the voltage used at present, it will furnish approximately 25 times the 

 energy now available. 



Anderson is continuing the study of the vacuum-spark spectra of metals. 

 Spectrograms have been secured for the following elements: aluminum, carbon, 

 calcium, cadmium, copper, iron, magnesium, lead, and zinc. Spectra of the 

 remaining available metals will be photographed as soon as possible. 



