of his voyage to the polar regions. He says, "I found no difficulty, in that cold and quiet air, in conversing with a man a mile away." It will be noticed that in both of these cases the air was mechanically homogeneous; that is, there were no alternating currents of hot and cold air.
In striking contrast with these may be mentioned the condition of the air as a vehicle for sound in the burned district of Boston, just after the fire had swept over it. There were many places, where there was a mixture of hot air, smoke, steam, and currents of cold air, in which the shouts of two people hardly a hundred feet apart, although audible, were so confused and indistinct as to make communication entirely impossible, and this too in quiet parts quite remote from the scene of conflagration. This effect was noticed when there were mingled currents of hot and cold air. That is, the air was mechanically heterogeneous. Humboldt speaks of the great difference in transmissive power for sound of the tropical air during the day and at night, and attributes this difference to the homogeneous condition of the air at night as compared to its heterogeneous condition in the day-time, due to convection-currents rising from the heated sands.
A large number of instances might be cited, and we should find that a clear homogeneous air transmits sounds readily, while an atmosphere broken by alternating convection-currents of warm and cold air is a very poor vehicle for sound. The explanation of this is not difficult. The original ray of sound, striking upon the first current of air, is partially reflected and partially transmitted.
The loss of the reflected portion causes a decrease in the intensity of sound. The transmitted portion, striking upon a second current, is likewise divided, and its transmitted portion continues to be so divided as many times as there are variations in the density of the air. Its reflected portion, as well as that of all the succeeding reflections, instead of being wholly lost, is interrupted in its backward course by the first current of air, and reflected along the path of the primary wave, but following it at an interval of time, depending upon the thickness of the current of air. Each reflection being thus again and again divided and reflected, we have, following close upon the primary wave, a multitude of secondary waves, which, falling later and later upon the ear, greatly mask the distinctness of the original sound, and give rise to indistinctness and confusion.
It is evident, then, that in order to procure the proper propagation of sound, one must do away with these air-currents. It must be remembered, however, that, when large numbers of people are crowded into halls, the air within is usually subjected to very considerable disturbances in order to obtain even indifferent ventilation. Registers, sprinkled here and there over the floor, send up their currents of hot and cold air. Opened windows or other cold-air ventilators send in their currents of cold air. If these currents could be made visible to the eye, as they can be detected by proper apparatus, we should see