386 ANNUAL KEPORT SMITHSONIAN INSTITUTION, 1910. 



only be drawn on here with reservation, but they seem at least to 

 promise important help in solving the problem of a regulative system. 



A close analysis of the movements of the molecules of the outer 

 atmosphere leads to the belief that the prevailing states there are 

 distinctly different from the collisional states that prevail near the 

 earth. The movements in the outer atmospheres seem to be in an 

 essential part orbital in nature, and this orbital atmosphere may log- 

 ically be supposed to occupy in an extremely attenuated way some 

 large part of the whole sphere of the earth's control. There should, 

 under the same reasoning, be a similar orbital extension of the atmos- 

 phere of the sun, and this extremely attenuated extension of the sun's 

 atmosphere should embrace the earth and its atmosphere. Under 

 the laws of molecular activity, these two atmospheres should be inter- 

 changing molecules at rates controlled by the equilibrium that ex- 

 ists between them. It is logical to infer that any excess above this 

 state of equilibrium that may at any time come to affect the earth's 

 atmosphere would cause it to feed out into the sun's sphere of control 

 faster than the reverse feeding took place, and that any deficiency 

 relative to the equilibrium state that might at any other time come 

 to affect the earth's atmosphere would lead to a deficient feeding out 

 while it would facilitate a greater feeding in from the sun's orbital 

 atmosphere. If this logical inference is valid, and if it has the requi- 

 site efficiency — which is a vital question yet to be settled — the main- 

 tenance of the delicate atmospheric conditions requisite for the con- 

 tinuity of life is automatically secured by a cosmic process of a 

 fundamental nature. Under this view the future competencj^ of our 

 atmosphere is not left wholly dependent on losses and gains at the 

 earth's surface, but is abetted bj^^ a sj^stem of solar and interplanetary 

 exchanges of a broadly cosmic order. The endurance of the earth's 

 atmospheres is thus in a measure wrapped up in the continued effi- 

 ciency of the sun's activities. 



If the question of our future be thus wrapped up in the problem 

 of solar endurance, weight must be given to the fact that the sun is 

 sending forth daily prodigious measures of energy. But yet these 

 are not wholly without some gains by way of partial offset. So far 

 as present knowledge goes, however, the gains are greatly inferior to 

 the losses. So long as the radiance of the sun was supposed to be 

 dependent on ordinary chemical action, or on the fall of meteorites, 

 or on self-contraction, it did not seem possible to forecast an en- 

 durance of activity sufficient for the direct and indirect needs of ter- 

 restrial life beyond a few million years. These few million years of 

 probable endurance were of course a great advance on the estimates 

 of the endurance of terrestrial conditions suitable for life worked 

 out on the old method of estimate. But recent physical investiga- 



