October 25, 1907] 



SCIENCE 



531 



order to make the area of the sea just half 

 the total area, is about 1,400 fathoms. The 

 contour-line at this depth would divide the 

 surface into two regions of approximately 

 equal area— the continental region and the 

 oceanic region. Pig. 2 represents the eon- 



Fig. 2. 



tour-line at 1,400 fathoms, or the line of 

 separation of the continental and oceanic 

 regions. The continental region is shaded. 

 In drawing this map I have omitted a num- 

 ber of small islands, and I have also 

 omitted a few enclosed patches of deep 

 water. Two of these are in the Mediter- 

 ranean, one in the Arctic Ocean, and others 

 are in the Gulf of Mexico and the Carib- 

 bean Sea. The Red Sea, the Mediter- 

 ranean, and the Arctic Ocean belong to the 

 continental region, and so do the Gulf of 

 Mexico and the Caribbean Sea. At this 

 depth Asia and North America are joined 

 across Behring's Strait, and Europe is 

 joined to North America across the British 

 Isles, Iceland, and Greenland; Australia is 

 joined to Asia through Borneo and New 

 Guinea, and the Australasian continental 

 region nearly reaches the Antarctic region 

 by way of New Zealand. At this depth 

 also South America does not taper to the 

 south, but spreads out, and is separated 

 from the Antarctic region by a very nar- 

 row channel. By going down to great 

 depths our problem is very much simpli- 

 fied. We find that the surface of the earth 

 san be divided into continental and oceanic 

 i-egions of approximately equal area by a 



curve which approaches a regular geo- 

 metrical shape. By smoothing away the 

 irregularities we obtain the curve shown in 

 Pig. 3, which exhibits the surface as di- 

 vided up into a continuous continental 

 region and two oceanic regions — the basin 



Fig.3. 



of the Pacific Ocean and the basin of the 

 Atlantic and Indian Oceans. We may take 

 our problem to be this: to account on 

 dynamical grounds for the separation of 

 the surface into a continental region and 

 two oceanic regions which are approxi- 

 mately of this shape. 



The key of the problem was put into our 

 hands four years ago by Jeans in his theory 

 of gravitational instability. If there are 

 any differences of density in diiferent parts 

 of a gravitating body, the denser parts 

 attract with a greater force than the rarer 

 parts, and thus more and more of the mass 

 tends to be drawn towards the parts where 

 the density is in excess, and away from the 

 parts where it is in defect. In every 

 gravitating system there is a tendency to 

 instability. In a body of planetary dimen- 

 sions this tendency, if it were not checked, 

 would result in a concentration of the mass 

 either towards the center or towards some 

 other part. But concentration of the mass 

 means compression of the material, and it 

 can not proceed very far without being 

 cheeked by the resistance which the ma- 

 terial offers to compression. There ensues 

 a sort of competition between two agencies : 

 gravitation, making for instability, and the 



