In later trials, Campbell and Keeler have employed an improved method, using photographs of the spectra on sensitive plates, but neither has succeeded in discovering any water vapour in the atmosphere of Mars. Obviously photography offers a great advantage over direct ocular observation. The two pictures may be placed side by side and very accurate measurements may be made at leisure. We may also choose the moments for exposure when the two stellar bodies stand equally high over the horizon so that the sunlight reflected from them traverses equal distances in the humid atmosphere of the earth. It now devolved upon Lowell to test his theories by means of the magnificent resources at his disposal in Flagstaff observatory in the desert of Arizona 2200 m. (7200 ft.) above sea level. In the months of January and February the dew-point there is about -7° C. (+19.4° F.) _i. e._, each cubic meter (1.3 cu. yds.) of air contains 2.8 grammes (43.25 grains) of water vapour while saturated air at zero temperature (32° F.) holds nearly twice this amount or 4.8 grammes (74 grains) per cubic meter (1.3 cu. yds.). Slipher, working in this observatory, pushed the sensitiveness of his plates to the utmost then obtainable, and photographed the spectrum of Mars in January and February, 1908. He found that the most important rain-band always was more prominent in the spectrum of Mars than in the spectrum of the Moon photographed later during the same night. Peculiarly enough, it was only the rain-band designated “A,” and located in the red spectral field, that was of a marked difference in the two spectra. Other bands gave no indication of the presence of water vapour on Mars. This result did not directly contradict the conclusions reached by Campbell and Keeler, also by means of photography; they had investigated other bands than “A.” The “A”-line might therefore possibly be more sensitive to water vapour than the others.