In the meantime Campbell had not been idle. The difficulty with the older measurements consisted in the fact that the absorption line of water vapour in the atmosphere of Mars occupies the identical place of the line due to vapour in the Earth’s atmosphere. There exists, however, a method, as already pointed out by Campbell in 1896, of separating the two, which method is available when Mars either approaches or departs from the Earth with sufficient velocity. The latter could be determined both from the known motions of the two planets and from the displacement of certain spectral lines of the Sun. These two determinations were in almost perfect agreement; for instance January 26–27, 1910, astronomical calculations gave a relative velocity of 19.1 km. (11.86 miles) per second and spectroscopical measurements 19.2 km. (11.93 miles) per second, while on February 3–4 the relative velocity was 18.1 km. (11.24 miles) a difference of 1 km. (.62 miles) per second. This trial shows the accuracy of the method. Among the absorption lines of water vapour and of oxygen there was, however, none due to the atmosphere of Mars. Campbell assumes that such lines would certainly have been visible if they had been only one-fifth as strong as the so-called tellurian lines. The advantage of this method is evidently that the “martian” and the “tellurian” lines lie close beside each other on the same plate so that differences in sensitiveness, exposure, and atmospheric conditions are entirely eliminated. From these and the following data we may calculate water content and temperature of the atmosphere on Mars anew: water vapour at the point of observation was 1.9 grammes (29.3 grains) per cubic meter (1.3 cu. yds.), zenith-distance of Mars 55° and incident as well as reflected sun-rays formed an angle of 70° with the surface of Mars; hence, the amount of moisture at the surface was only 0.12 gramme (1.85 grains) per cubic meter (1.3 cu. yds.), corresponding to -38° C. (-36.4° F.) for saturated air and to -27° C. (-16.6° F.) for air of 31 per cent. saturation. Oxygen content per cubic meter (1.3 cu. yds.) at the surface of Mars would be only a sixteenth part of corresponding numerical value on the Earth. This determination is more accurate than any of the previous ones and reduces the temperature another 10° C. (18° F.) below the lowest value derived earlier in this chapter. We should remember, however, that, during the trial of September, 1909, the sun stood practically in zenith on Mars, while in January and February, 1910, we are concerned with a point where sunrise had occurred about four and a half hours previously. The latter observation should give a value close to, but slightly above, the mean diurnal temperature on Mars. No determination comparable in precision with this one by Campbell appears to have been made. We must therefore recognize it as conclusive.