in both is light (see paper by the writer in Trans. Philosophic Society).

19. These several facts all go to show that Mars possesses an atmosphere, which, since it contains water vapor, would, from the kinetic theory of gases also contain the heavier ones; nitrogen, oxygen and carbonic acid.

20. That the amount of the water vapor corresponds more to that over our deserts than anything else on earth.

21. That the only water on the planet exists in the atmosphere and in the polar snows. Mars, therefore, is a planet which is very badly off for water and can only get a surface distribution of it by the melting of the polar caps.

From the behaviour of the polar caps due to the fact that they can be nothing but hoar frost or snow, Ave have our first evidence that the temperature on Mars is by no means very low.

22. The fact that the snow caps disappear up to 87° of latitudine, and sometimes further, shows that the temperature at times must be decidedly high.

23. Now this has been confirmed by a mathematical investigation of the writer’s, published in the Philosophic Magazine, on what the theoretic temperature should be, taking into account all the factors in the case, many of which had, in previous determinations, been omitted. The out-come of this investigation was to show that the mean temperature of Mars was probably about 8° centigrade. This is somewhat but not very much colder than the mean temperature of the Earth, which is usually taken as 15° centigrade.

24. The whole aspect of the disk bears out this conclusion. The seasons at which the melting of the polar caps begins, and the time at which it is brought to an end, at what corresponds to about the 20 th of August in our calendar, — this being the time at which the first new snow of the on-coming winter makes its appearance — all tend to show that the mean temperature must be just about what the theoretic puts it at.

25. The variations in temperature between summer and winter must, however, be very great, because of the thinness of the Martian air, and this also seems to be the case from the latitude, below 60°, to which the first snowfall reaches and the time of year at which it does so.