452 K D. Preston— Study of the Earth's 



of the theory. Then again, it seems quite well ' established 

 that our present continental forms are very old, which would 

 indicate that whatever form the contracting earth may be tak- 

 ing, it has been gradually settling into this shape for many 

 millions of years. Xo reference is made at present to changes 

 on the earth's surface, consequent upon outside conditions. 

 The changes in the eccentricity of the orbit which has its 

 greatest value at intervals of about 2,500,000 years certainly 

 produce vast changes in the distribution of matter, from the 

 accumulation of ice and snow, and from the shifting of the 

 ocean currents ; but these are purely surface phenomena, and 

 do not probably affect the permanent shape of the contracting 

 nucleus. Besides, any change produced at a time when the 

 eccentricity was at a maximum, would be counterbalanced 

 when the next minimum occurred somewhat more than a 

 million years later. But the earth's contraction goes on indefi- 

 nitely throughout all time. 



This brings us to a consideration of 



The EartlCs mean density. 



The pendulum has recently been employed in such deter- 

 mination. 



It is well known that during the early part of this century 

 Dr. Hutton conceived the idea of determining this constant by 

 comparing the attraction of the earth with that of a mountain 

 of known dimensions. The method of course rests on the 

 assumption that the volume of the mountain is a determinable 

 quantity and also that the mountain is solid. Latitude obser- 

 vations were made on the north and south flanks and the 

 results were compared with the actual differences of latitude 

 obtained by connecting the points by triangulation. This way 

 of getting the earth's mean density has been employed in a 

 number of cases since Dr. Hutton's time and always with 

 approximately the same result. It is evident now that if we 

 have a means of getting a value for the density of the mountain 

 that is independent of the latitude observations we get a rig- 

 orous check on the final result. This modification of the 

 problem was applied with entire success in the Sandwich 

 Islands in 1887; and not only was the mountain much larger 

 and higher than in Dr. Hutton's work, but its form was much 

 more accurately known. The entire island of Maui rises to 

 an elevation of 10,000 feet and has on its summit the crater of 

 Haleakala, which is twenty miles in circumference and half a 

 mile deep The whole mountain has been contoured from the 

 sea to the summit giving differences of elevation for every 500 

 feet. This gave a means of calculating with a high degree of 



