And here we find ourselves face to face with a question of great delicacy and importance. Both as a radiator and as an absorber, carbonic acid is, in general, a feeble gas. It is beaten in this respect by chloride of methyl, ethylene, ammonia, sulphurous acid, nitrous oxide, and marsh gas. Compared with some of these gases, its behavior, in fact, approaches that of elementary bodies. May it not help to explain their neutrality? The doctrine is now very generally accepted that atoms of the same kind may, like atoms of different kinds, group themselves to molecules. Affinity exists between hydrogen and hydrogen and between chlorine and chlorine, as well as between hydrogen and chlorine. We have thus homogeneous molecules as well as heterogeneous molecules, and the neutrality so strikingly exhibited by the elements may be due to a quality of which carbonic acid furnishes a partial illustration. The paired atoms of the elementary molecules may be so out of accord with the periods of the ultra red waves--the vibrating periods of these atoms may, for example, be so rapid--as to disqualify them both from emitting those waves, and from accepting their energy. This would practically destroy their power, both as radiators and absorbers. I have reason to know that a distinguished authority has for some time entertained this hypothesis. We must, however, refresh ourselves by occasional contact with the solid ground of experiment, and an interesting problem now lies before us awaiting experimental solution. Suppose two hundred men to be scattered equably throughout the length of Pall Mall. By timely swerving now and then, a runner from St. James's Palace to the Athenaeum Club might be able to get through such a crowd without much hinderance. But supposing the men to close up so as to form a dense file crossing Pall Mall from north to south; such a barrier might seriously impede, or entirely stop, the runner. Instead of a crowd of men, let us imagine a column of molecules under small pressure, thus resembling the sparsely distributed crowd. Let us suppose the column to shorten, without change in the quantity of matter, until the molecules are so squeezed together as to resemble the closed file across Pall Mall. During these changes of density, would the action of the molecules upon a beam of heat passing among them at all resemble the action of the crowd upon the runner?