and in one test an alloy melting at 280 deg. C. began to soften. Working
with an experimental apparatus constructed on the "dripping" principle--
_i.e._, a generator in which water is allowed to fall in single
drops or as a fine stream upon a mass of carbide--with the deliberate
object of ascertaining the highest temperatures capable of production
when calcium carbide is decomposed in this particular fashion, and
employing for the measurement of the heat a Le Chatelier thermo-couple,
with its sensitive wires lying among the carbide lumps, Lewes has
observed a maximum temperature of 674 deg. C. to be reached in 19 minutes
when water was dripped upon 227 grammes of carbide at a speed of about 8
grammes per minute. In other experiments he used a laboratory apparatus
designed upon the "dipping" principle, and found maximum temperatures, in
four different trials, of 703 deg., 734 deg., 754 deg., and 807 deg. C.,
which were reached in periods of time ranging from 12 to 17 minutes. Even
allowing for the greater delicacy of the instrument adopted by Lewes for
measuring the temperature in comparison with the device employed by Caro,
there still remains an astonishing difference between Caro's maximum of
280 deg. and Lewes' maximum of 807 deg. C. The explanation of this
discrepancy is to be inferred from what has just been said. The generator
used by Caro was properly made of metal, was quite small in size, was
properly designed with some skill to prevent overheating as much as possible,
and was worked at the speed for which it was intended--in a word, it was as
good an apparatus as could be made of this particular type. Lewes' generator
was simply a piece of glass and metal, in which provisions to avoid
overheating were absent; and therefore the wide difference between the
temperatures noted does not suggest any inaccuracy of observation or
experiment, but shows what can be done to assist in the dissipation of
heat by careful arrangement of parts. The difference in temperature
between the acetylene and the carbide in Caro's test accentuates the
difficulty of gauging the heat in a carbide vessel by mere external