NOV. 22, I9IS 
Improved Respiration Calorimeter 
333 
the copper wall. A spindle is pushed through the sleeve and into the 
thimble until the junctions projecting from its inner end are very close to 
the bottom of the thimble, actual contact being prevented by the adjust¬ 
ment of a small screw in the end of the spindle. A change in the tem¬ 
perature of the copper wall immediately affects the temperature of the 
thimble attached to it, and consequently that of the junctions within 
the thimble. The junctions in the recess at the other end of the spindle 
are within the sleeve attached to the zinc wall, and any change in the 
temperature of the zinc wall affects the sleeve and, hence, the tempera¬ 
ture of the junctions within it. Since both the sleeve and the thimble 
are short, neither affects the temperature of the wire in the elements any 
considerable distance from the junctions. The sleeve, however, projects 
slightly either side of the zinc wall, so that it will surround the junctions, 
even when they might come inside or outside the plane of the zinc, be¬ 
cause of inequalities of distance between the two metal walls. 
A short section of the spindle, between the recess and one end, projects 
from the outer end of the sleeve in the zinc wall and provides a firm stay 
for the terminals of the elements. 
There are 95 such thermoelements distributed in the walls of the 
chamber. If they were equally spaced there would be one for each 4.5 
dm. square of surface; but since the temperature of the chamber would 
tend to vary more at the top than at the bottom, more elements were 
installed in the upper than in the lower parts of the chamber to increase 
the sensitivity and integrate a larger number of sections of the walls. 
There are accordingly 16 elements in the ceiling and 10 in the floor. In 
the sides are five rows, with 14 elements in each row except the first one 
from the top, from which one is missing because the space in which it 
would be located is occupied by the window. The five rows are not quite 
equally separated, the two upper rows being slightly nearer together 
than the three lower ones, in accordance with the idea that the tempera¬ 
ture of the upper section would tend to vary more than that of the lower 
one. These thermoelements are joined in groups in such manner that a 
difference between the temperature of the copper wall and that of the 
zinc wall may be detected in certain portions of the walls without regard 
to conditions in other parts. One group includes the 16 elements in the 
top; another the 28 elements in the two upper rows of the sides, called 
the upper zone; a third, the 42 elements in the three lower rows of the 
sides, called the lower zone; and a fourth group, the 10 elements in the 
bottom. The thermoelements in each group are connected in series 
by heavily insulated No. 18 copper wire, and the same sort of wire leads 
from the terminals of each group to a multiple point switch on the 
observer's table by which the groups may be connected successively 
with the galvanometer. It is also possible to connect all 95 thermo¬ 
elements in series as a whole with the galvanometer and thus observe 
the average difference between the temperature of the copper wall as a 
whole and that of the zinc wall as a whole. 
