OUR prospect of finding differences in the chemical composition of the air is, of course, better the higher the sample is obtained in the stratosphere. An important part of the whole research is therefore the collection of air samples from great altitudes. For this purpose, sending up automatic devices in unmanned balloons is the most efficient method. Aeroplanes cannot attain sufficiently high altitudes; even Squadron Leader F. R. D. Swain in his record flight last September reached only 15 km., and the inconveniences of an airtight. Continued from p. 182. suit preclude complicated scientific operations. Balloon ascents in closed gondolas, as introduced by A. Piccard in 1931, give more freedom for observations, and can attain greater heights. A year ago, Capt. Stevens and Capt. Anderson in a stratosphere flight arranged by the National Geographic Society and the U.S. Army Air Corps reached 22 km.; but it cannot be said that the scientific results of the expedition justified the immense costs. The varying conditions of the atmosphere make numerous observations necessary, and only the cheap flights of sounding balloons can be repeated frequently enough; moreover, they can ascend far higher than manned balloons. (Heights of more than 30 km. have been recorded.) Fig. 2 shows the altitudes so far reached by the most successful flights of aeroplanes, manned and unmanned balloons. The nights from which air samples have been brought back are underlined. A number of other air samples collected by sounding balloons are not recorded in the figure; the heights attained can be seen in Table 3.