468 THE ICE AGE IN NORTH AMERICA. 



10,500 years from the present time, the winter in the north- 

 ern hemisphere will occur in aphelion instead of perihelion ; 

 and, if the supposition be correct concerning the influence of 

 this increased length of the winter and distance of the earth 

 from the sun, the conditions would favor the return of a 

 glacial period 10,000 or 11,000 years hence, and would imply 

 that similar favorable conditions existed 10,000 or 11,000 

 years ago. According to this theory, also, there should have 

 been a succession of glacial periods every 21,000 years during 

 long ages past. 



But there is still another periodicity in the movements of 

 the earth about the sun with which to combine the preced- 

 ing. The shape of the earth's orbit is not permanent, but 

 through the influence of the attraction of the planets upon it 

 is subject to periodic changes. In astronomical terms, the 

 " eccentricity " of the earth's orbit is subject to variations ; 

 that is, there are sometimes very much greater differences 

 than at present between the longer and the shorter diameters 

 of the earth's orbit. When this difference is greatest it 

 amounts to no less than 7,000,000 miles ; so that at certain 

 times the earth is 14,000,000 miles farther from the sun in 

 winter than in summer, and vice versa. At the time of 

 greatest eccentricity, also, the difference in length between 

 the summers and winters would amount to thirty-six days, 

 instead of seven or eight as now. 



These periods of greatest eccentricity in the earth's orbit 

 during which, on Mr. Croll's theory, the conditions were ex- 

 tremely favorable for the production of glacial epochs, are 

 somewhat unevenly distributed. One of them culminated 

 200,000 years ago; another, 750,000; another, 850,000; 

 another, 2,500,000 ; and another, 2,600,000. In the future 

 they will occur 500,000, 800,000, 900,000 hence. In the 

 present condition of the earth's orbit this supposed cause is 

 at its minimum. 



Of the astronomical changes affecting the eccentricity of 

 the earth's orbit, we are certain. But the value of Croll's 

 theory depends upon the correctness of the original assump- 



