of Edinhurgli, Session 1882-83. 
33 
mimication with the boiler or the water supply. It was made of 
copper, that its conductivity might be as great as possible. 
The curved sides and bottom of the absorber were made as before 
of the stoutest commercial tinned iron. We thought it unnecessary 
to have it made more rigid, as the pressure of the water in the jacket 
was constant during each experiment. The copper plate bent under 
the pressure of the water and the steam, however, more than we anti- 
cipated, and the differences in the amount of the bending caused 
perceptible changes of form in the sides. The pressures given below 
are subject therefore to sudden slight variations, which depend upon 
sudden changes of volume of the absorber, not upon absorption. 
Hence the numbers given below cannot be taken as exact results. 
It will be noticed in the experiments made with this apparatus, 
that after the steam is turned on in the radiator, previously filled 
with running water, the pressure at first diminishes, and only after 
a certain time increases. This is not due to absorption, but to the 
fact that the water pressure in the laboratory was slightly greater 
than the boiler pressure which we used. When, therefore, the 
water pressure was taken off the sides of the radiator K, and the 
steam pressure was put on, the copper plate became less bent, and 
the reservoir therefore became suddenly larger, the pressure conse- 
quently falling. As might be expected, in the case of an apparatus 
not sufficiently rigid, the first experiments made with it were much 
more unsatisfactory than the last. The apparatus took gradually 
the set determined by the pressures employed. If the measurements, 
given in the tables below are plotted, the resulting curves will be 
found far smoother in the case of the later than in that of the earlier 
experiments. Tor more exact experiments a more rigid apparatus 
should be employed. 
Tor the following experiments a new manometer was used. It 
was made from a carefully selected tube of practically uniform bore. 
The mode of procedure in the experiments with the new reservoir 
was practically the same as that described above. Tests were 
applied more frequently to find if the reservoir was air-tight, and to 
find if the' drying tubes were efficient. Their efficiency was tested 
by weighing at intervals the last of the series. 
XIII. Air saturated with Water Vapour at 13° ‘2 C. — The 
absorber being full of air, a small quantity of water was introduced 
VOL. XII. 
c 
