This latter condition is one of essential importance. Bacteria cannot leave a moist surface either under evaporation or by means of air-currents.[22] Only when there is considerable molecular disturbance, such as splashing, can there possibly be microbes transmitted to the surrounding air. This fact, coupled with the influence of gravitation, is the reason why sewer gas and all air contained within moist perimeters is almost germ-free; whereas from dry surfaces the least air-current is able to raise countless numbers of organisms. Quite recently this principle has been admirably illustrated in two series of investigations made upon expired and inspired air. In a report to the Smithsonian Institution of Washington (1895) upon the composition of expired air, it is concluded that "in ordinary quiet respiration no bacteria, epithelial scabs, or particles of dead tissue are contained in the expired air. In the act of coughing or sneezing such organisms or particles may probably be thrown out." The interior of the cavity of the mouth and external respiratory tract is a moist perimeter, from the walls of which no organisms can rise except under molecular disturbance. The position is precisely analogous to the germ-free sewer air as established by Messrs. Laws and Andrewes for the London County Council. The popular idea that infection can be "given off by the breath" is contrary to the laws of organismal pollution of air. The required conditions are not fulfilled, and such breath infection must be of extremely rare occurrence. The air can only be infective when filled with organisms arising from dried surfaces. The other series of investigations were conducted by Drs. Hewlett and St. Clair Thompson, and dealt with the fate of micro-organisms in inspired air and micro-organisms in the healthy nose. They estimated that from 1500 to 14,000 bacteria were inspired every hour. Yet, as we have pointed out, expired air contains practically none at all. It is clear, then, that the inspired bacteria are detained somewhere. Lister has pointed out, from observation on a pneumo-thorax caused by a wound of the lung by a fractured rib, that bacteria are arrested before they reach the air-cells of the lung; hence it is at some intermediate stage that they are detained. Hewlett and Thomson examined the mucus from the wall of the trachea, and found it germ-free. It was only when they reached the mucous membrane and moist vestibules and vibrissæ of the nose that they found bacteria. Here they were present in abundance. The ciliated epithelium, the moist mucus, and the bactericidal influence of the wandering or "phagocyte" cells probably all contribute to their final removal.[23]