Several of the numerical results contained in this table may be verified by calculation.
For, when two plates of different kinds are exposed together to the radiation of the source, their position relative to the entrance and the issue of the calorific rays does not affect the quantity of heat which passes through this system. This is easily proved by putting the first plate in the place of the second; for the thermomultiplier, notwithstanding this change of order, continues to mark the same degree of its scale. Let us now take two plates and place them alternately in each of the two positions, for instance, the plate of alum and the chromate of potash. These two substances, exposed separately to 100 rays of heat emanating directly from the source, transmit 9 and 34 respectively. The quantities of heat that should fall on each of the two plates in order that 100 may emerge in each case is easily determined by these simple proportions;
| 9 | : | 100 | : : | 100 | : | x, |
| 34 | : | 100 | : : | 100 | : | x, |
which give 1111 for the alum and 294 for the chromate of potash. Now we know by experiment that chromate of potash exposed to 100 rays issuing from alum transmits 57, and that alum exposed to 100 rays issuing from chromate of potash transmits 15.
But the order of succession has no influence on the transmission of the pair: let us therefore reverse the system only in one case or the other. We shall then have the same plates exposed in the same manner to the two radiations of 1111 and 294. The quantities transmitted under both circumstances should accordingly be proportional to the incident quantities, as is actually proved within the limits of approximation compatible with the nature of the experiments; for we have,
57 : 15 :: 1111 : 294.
The table contains ten pairs which are submitted in both ways to the radiations of the source; there are in it consequently twenty numbers which should be in proportions analogous to the preceding. It is evident too that these calculations require that the five plates emitting the 100 rays which fall successively on the whole series of diathermanous bodies should be those that are indicated by the same names in the first column. I have accordingly taken care that this condition should be satisfied.
The bodies submitted to the heat emerging from the screens present no longer the same order of transmission that they presented under the immediate action of the radiation of the lamp. The changes which take place have no apparent regularity whether we compare one series with another or consider only the different terms of the same series. Thus glass, Iceland spar, and rock crystal are more diathermanous to the heat emerging from the five screens than to that which comes di-
