242 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1911. 
dier? has shown, by means of his chromo-actinometer, that every 
quartz mercury-vapor lamp gradually loses its power of giving out 
ultra-violet rays because of the increasing opaqueness of the quartz 
tube. At the end of a service of about 500 hours, he states, the 
lamp emits only one-seventh of its original quota of these rays. 
This defect is without remedy and therefore decisive. If his results 
are real, then immersion is necessary for cooling the lamp, so as to 
prevent this change. 
Immersion, therefore, has these advantages, economy and long 
service. 
It may be objected that since the immersion cools the lamp, with a 
given current of electricity there will be a smaller amount of ultra- 
violet rays emitted. This would be expected in accordance with the 
results of Retchinsky (1906). However, a lamp thus cooled can be 
made to emit just as much ultra-violet radiation as a warm lamp 
(new and unused, see the experiments of H. Bordier) if a greater cur- 
rent of electricity is used. ‘The increased cost of this greater current 
is practically of small importance compared with the increased 
efficiency of the immersed lamp where all the radiation is utilized. 
That is, though an immersed lamp uses a greater amount of elec- 
tricity; it will sterilize a far greater amount of water. 
To sum up: The immersion of the lamp, economically considered, 
is preferable. If the results of Bordier are confirmed, immersion will 
be the only method giving a long life to the sterilizing power of 
the lamps. 
NATURE OF THE PROCESS OF STERILIZATION OF WATER BY THE ULTRA- 
VIOLET RADIATIONS. 
The ultra-violet rays kill the microbes in the water by a direct 
bactericidal action, and not indirectly by a chemical modification of 
the water. 
t might at first have been surmised that the sterilization was due 
to the production of ozone. That is wholly untrue. With Th. 
Nogier and Rochaix,’ I have shown that during the time required 
for sterilization (only a few seconds or a minute) not even a trace of 
ozone was produced. If with a longer time it is produced, it would 
be after sterilization had occurred and have nothing to do with the 
iatter process. It would never occur in practice. The production 
of ozone is a separate laboratory process. Moreover, sterilization 
will take place in the absence of oxygen. (Cernovodeanu and V, 
Henri.) 
1. Bordier, Quantitométrie des rayons ultra-violets. Archives El. Méd., No. 285, p. 396. 
2J. Courmont, Th. Nogier et Rochaix. C.R. Ac. des Sciences, 12 juillet 1909. 
