Meteorological observations were also of great importance in connection with the new process of locating distant guns known as “sound-ranging.” This process consists, briefly, in determining the exact instant of arrival at several points of the sound waves propagated through the air from the gun that is being located. If sound traveled at a uniform speed, these observations would show the exact distance of each of the observing points from the gun, and a simple geometrical construction would indicate the position of the latter. The speed of sound waves in the air is, however, affected by both wind and temperature. Accordingly, allowance had to be made for these varying factors, and the necessary data were supplied by the meteorological units. The observation and prediction of winds favoring the use of poisonous gases by friend or foe was one of the most delicate tasks allotted to the army meteorologists. The flow of such gases is determined by the winds close to the surface of the earth, and these are greatly affected by topography. Local air currents controlled by the slope of the ground were especially utilized for gas attacks. Strong winds were unfavorable, because they quickly dissipated the gas cloud. The meteorologists not only advised their own troops when to use gas, but also gave warning when the atmospheric conditions were such that gas was likely to be used by the enemy. The use of gas shells was less dependent for its success upon the wind than the liberation of gas clouds, but even when shells were used the wind and weather at the objective point were factors of importance. The exigencies of warfare developed several new features of meteorological practice, the utility of which did not cease with the war. Thus it became customary to measure the degree of “visibility” of distant objects, for the benefit of aviators and gunners, and this element was included in the routine weather reports. Scales of visibility, ranging from “very bad” to “excellent,” etc., were adopted, and eventually certain forms of apparatus (“visibility meters”) were devised for getting fairly precise measurements of this weather factor. Another novel practice that deserves to be perpetuated was the plan adopted by the military forecasters of adding to their predictions a statement as to their probable accuracy; this was expressed on a numerical scale of “odds,” instead of by use of the vague terms “probably” and “possibly,” which have generally served the purpose of the dubious forecaster. The war brought about many improvements in the instruments and methods used in sounding the upper air; and the intensive campaign of pilot-balloon observations carried out at the military stations provided a body of data for study quite unparalleled in the history of meteorology.