530 



SCIENCE 



[N. S. Vol. XXVI. No. 669 



bution we have recourse to an indirect 

 means of investigation, and the chief in- 

 strument of research is the pendulum. The 

 time of vibration of a pendulum varies 

 with the place where it is swung, and from 

 the observed times we deduce the values of 

 gravity at the various places, and it was 

 shown many years ago by Stokes that the 

 shape of the geoid can be inferred from the 

 variation of gravity over the surface. 



The question to which I wish to invite 

 your attention is a different one. If the 

 ocean could be dried up, the earth would 

 still have a shape. What shape would it 

 be? Why should the earth have that shape 

 rather than some other ^ In order to de- 

 scribe the shape we may imagine that we 

 try to make a model of it. If we could 

 begin with a model of the geoid we should 

 have to attach additional material over the 

 parts representing land and to remove 

 some material over the parts representing 

 sea. Our model would have to be as big 

 as a battleship if the elevations and depres- 

 sions were to be as much as three or four 

 inches. In thinking out the construction 

 of such a model we could not fail to be im- 

 pressed by certain general features of the 

 distribution of continent and ocean, and we 

 may examine a map to discover such fea- 



tures. Fig. 1 is a rough map of the world 

 drawn in such a way that to every degree 

 of latitude or of longitude there cor- 

 responds the same distance on the map. 

 Certain very prominent features have often 



been remarked: the tapering of America 

 and Africa towards the south, the dispro- 

 portion between the land areas of the 

 northern and southern hemispheres, the 

 excess of the oceanic area above the conti- 

 nental area, which occupies but litttle more 

 than one quarter of the surface; the wide 

 extent of the Pacific Ocean, which with the 

 adjoining parts of the Southern Ocean 

 covers nearly two fifths of the surface. 

 Another prominent feature is the antipodal 

 position of continent and ocean. South 

 America south of an irregular line which 

 runs from a point near Lake Titicaca to 

 Buenos Ayres is antipodal to a portion of 

 Asia which lies in an irregular triangle 

 with corners near Bangkok, Kiaochau and 

 Lake Baikal; but no other considerable 

 parts of the continental system have con- 

 tinental antipodes. The Antarctic conti- 

 nent is antipodal to the Arctic Ocean, 

 Australia is antipodal to the central At- 

 lantic, and so on. Another notable feature 

 is the skew position of South America to 

 the east of North America ; South America 

 lies to the east of the meridian 85° west of 

 Greenwich; most of North America lies to 

 the west of it. But, although we may ob- 

 serve prominent general features of the 

 distribution, we should find it far from 

 easy to attribute to the form of our imagi- 

 nary model anything that could be called 

 a regular geometrical figure. When we 

 begin to think about the removal of ma- 

 terial from the parts of the model which 

 are to represent oceans and seas, we require 

 a map which gives information about the 

 depth of the sea in diflierent places. 

 Around all the coasts there is a margin of 

 not very deep water. If some part of the 

 sea could be dried up, so that more land 

 was exposed around all the coasts, the area 

 of the surface of the sea would be dimin- 

 ished; and it is known that the depth of 

 water that would have to be removed m 



