One way in which air or any other gas is cooled is by expanding against pressure. Some of the energy in the gas, originally manifesting itself as heat, is applied to the work of expansion, and thus ceases to be heat. Hence the temperature of the gas falls. Conversely, if a mass of gas is compressed, the mere process of compression raises its temperature. The heat produced in pumping up a bicycle tire is the classic example of the latter fact. Heating by compression and cooling by expansion are called, respectively, “dynamic heating” and “dynamic cooling.” The processes thus described are of the utmost importance in meteorology. If air of average humidity is admitted to the receiver of an air pump in the usual way and suddenly expanded by partial exhaustion, a cloud of moisture is seen to form in the receiver. This moisture is condensed and made visible by the dynamic cooling of the air. If, however, after the receiver is exhausted air of the same humidity as before is admitted through a filter of cotton wool, and is then similarly cooled by expansion, no cloud will form. Evidently the filter has removed from the air something that is essential to the process of condensation. Perhaps it will occur to the reader that, in some obscure way, the filter has prevented water vapor itself from entering the receiver. There are several methods by which we can ascertain whether such is the case. One of the simplest is to admit a little smoke to the receiver before expanding the filtered air. In this case the cloud _does_ form, showing that moisture is present, and also showing that smoke, though a perfectly dry substance, aids the formation of the water drops. Such experiments have led to the conclusion, now universally admitted, that when water drops form in the atmosphere they always form around “nuclei” of something that is not water. These nuclei are often referred to as “dust particles,” but it is recognized that a vast proportion of them are very much more minute than the dust that worries housewives. They are largely beyond the power of the microscope, and some of them, indeed, appear to be of molecular size, consisting of molecules of hygroscopic gases, such as the oxides of sulphur and of nitrogen.