DE. J. P. JOULE ON THE SUEFACE-CONDENSATION OF STEAM. 
135 
though very small, quantity of heat is thus conducted along the tube from the stopcock 
T as far as the india-rubber junction t. Any water condensed in P falls back again into 
the boiler ; that between the stopcock and t falls into the receiver ; so that the small 
quantity of conducted heat just mentioned is probably compensated by the trifling 
cooling effect of the atmosphere between T and the refrigerating water. 
The short continuation pipe q exposes to the water an eflective length of 3 inches, 
which, on account of the wideness of the channel there, could not generally have had 
an effect greater than that due to 2 inches in the narrower part. As, however, a length 
amounting to an inch and a half of the ends of the condensing tube is overlapped by the 
vulcanized tubing, the entire amomit of condensation may, without appreciable error, 
be laid to the account of the condensing pipe. 
The receiver R and the pipes C, P, and ^ are enveloped by a thick coating of cotton- 
wool and flannel, so as to prevent as far as possible the refrigerating effect of the atmo- 
sphere. 
Great pains were taken to make every part of the apparatus in which the pressure 
is below that of the atmosphere, perfectly air-tight. It will be seen that the form of 
the stopcock T effectually prevents any leakage except from the high pressure side into 
the atmosphere, which is of no consequence. The india-rubber junctions were at first 
made by simply binding on the ends of the tubes short lengths of vulcanized caoutchouc ; 
but it was soon found that enough air passed to vitiate the experiments, which were 
consequently rejected. The method afterwards adopted was to smear the ends of the 
tubes with melted vulcanized caoutchouc before the short india-rubber tubes were bound 
on. This plan was found to be so efficacious that air appeared to be perfectly excluded, 
and the vacuum wholly unimpaired, however long an experiment was carried on. 
The vacuum-gauge glass tube is 0‘45 of an inch in internal diameter. It is plunged 
into a wide dish of mercuiy, Aom the surface of which the height of the column is 
measm’ed. The temperature of the mercury in the gauge was always nearly that of 
the barometer w’hich registered the atmospheric pressure. During each experiment a 
small quantity of condensed water settled by degrees on the top of the mercury, the 
length of which, divided by 13‘56, gave the correction to be supplied to the height of the 
column. 
It will be observed that the pipe leading to the vacuum-gauge is inserted near the 
stopcock which admits the steam. It was important to ascertain whether the gauge 
would stand at the same level if it were connected with other parts of the vacuous 
space. To determine this, a pipe was attached to the receiver at r, and connected with 
a gauge placed side by side with the first gauge, and dipping into the same dish of 
mercury. The gauges were observed during rapid and slow condensation, at different 
and at varying pressures ; but the height of the columns appeared to be in general 
exactly the same : if any difference could be observed at any time, I would say that the 
receiver gauge indicated the less perfect vacuum of the two ; the difference, however, 
amounted in no case to more than - 3 -(jth of an inch. 
