84 Proceedings of Royal Society of Edinburgh, [jan. 17 , 
There would appear to be an advantage in favour of the C screen, 
which may be referred to here. We have seen that when the louvres 
of the Stevenson screen get dirty, they absorb more radiant heat, and 
so increase the error of the readings. Many screens in daily use 
must, from this cause, give too high readings. The screen C is, 
however, not much affected from this cause; indeed, I am not quite 
certain but that the screen will act quite as well if certain parts 
are black. Tor instance, the sides of the plates C and D which 
are exposed to the bulb of the thermometer might perhaps with 
advantage be blackened. I have not yet been able experimentally 
to determine this point, but many of my observations have been 
taken with a large black patch in the centre of each plate. My 
reason for testing this was, that if these surfaces are white, they 
will reflect to the bulb some of the heat which falls on them ; but 
if they are black, they will absorb this heat; and it seemed possible 
that the increased amount of heat radiated by the blackened sur- 
faces, together with the greater amount to which the air in contact 
with them is heated, might affect the thermometer less than the heat 
reflected by the white. It was not found possible to settle this point 
by readings taken with the screen under the two conditions and com- 
paring them with the silvered bulb, as the inertia of the two arrange- 
ments is so different, and the effect sought for very small. This, with 
many other matters connected with this screen, will be best settled 
by means of two screens similar in all points save the one we wish 
to test. For these trials, however, we must wait another warm 
season. Although this new screen has acted very satisfactorily 
up to the present time, giving readings almost exactly the same 
as the fine silvered bulb standard, and much below those of the 
Stevenson screen, yet it would be rash to conclude that it will be 
superior under all conditions of climate. The various influences 
to which thermometer screens are exposed are so numerous that 
the unexpected has many opportunities of happening and upsetting 
our hopes and expectations. 
[A Postscript to this Paper will be found immediately after the 
Proceedings of July 1887.] 
