5G 
PROFESSOR W. RAMSAY AXD DR. M. M'. TRAVERS 
a serie.s of ascending densities; they are represented by the numbers 2 to 6 in 
the annexed table. The method employed for their further fractionation was as 
follows :— 
Table I. 
The lieaviest portion G was placed in the gasholder, and liquefied in the bulb; a 
portion of it was allowed to l3oil oft' into the gasholder; and the remainder (the 
heavier portion 12) was transferred to a tube. The next heaviest portion 5 was 
then placed in the gasholder, li(|uefied together with the gas already in the bulb, and 
sidosequently divided by evaporation into two portions. This operation was repeated 
in the case of each fraction, the proportion of gas taken hack into the gasholder 
increasing at each operation so as to produce six a 2 )proximately equal fractions, 
represented in the table l)y numbers 7 to 12. 
This process, as is clearl}^ showm in the table, resulted in the sej^aration of the 
heavy and light constituents of the mixture with a certain degree of shaiqDness. 
Fraction 7, wdieu examined spectroscopically, was found to contain only a trace of 
argon, wdiile fraction 12 consisted almost entirely of that gas; beyond this, it was 
found that the gas of density 97 could not be liquefied at —200° under a pressure 
of two atmospheres. 
In order to determine wdiether fraction 7 consisted of homogeneous gas, mixed 
wfith only a trace of argon, it w’as necessary to fractionate it further, and to submit 
each fraction to a rigid sjDectroscopic examination. Since the gas would not liquefy, 
it WAS mixed with about eight times its volume of oxygen; the mixture was liquefied 
in the ajqDaratus, and the liquid was allowed to evaporate in three fractions. Nos. 13, 
14, and 15. After the oxvwn had been removed from the inactive residue, it was 
found that in the third fi'actlon (15) alone, could argon be detected by visual 
observation of the sjDectrum. When, however, tlie sjjectrum of the lighter fraction 
WAS 23hotographed, it was found that the lines rejiresented by the wave-lengths 587G 
and 501G were j^resent. The gas contained lielium, wdu’ch is accordingly present in 
the atmosphere ; Kayser’s and Feiedlaxder’s oh.servati(ms w'ere confirmed. After 
knowdng that they w’ere present, it w^as jJt^ssible to see the principal helium lines 
wfith a direct-vision S2)ectroscope. 
