MY NEW IDEAS 407 



and adduced a completely new argument for this mode of 

 origin of the valley lakes of glaciated countries. This is 

 founded on their surface and bottom contours, both of which 

 are shown to be such as would necessarily arise from ice- 

 action, while they would not arise from the other alleged 

 mode of origin — unequal elevation or subsidence. 



12. In a new edition of " Stanford's Compendium, Austral- 

 asia," vol. i., when describing the physical and mental char- 

 acteristics of the Australian aborigines, I stated my belief 

 that they were really a low and perhaps primitive type of the 

 Caucasian race. I further developed the subject in my 

 " Studies," and illustrated it by photographs of Australians 

 and Ainos, of the Veddahs of Ceylon, and of the Khmers of 

 Cambodia — all outlying members of the same great human 

 race. This, I think, is an important simplification in the clas- 

 sification of the races of man. 



Bees' cells. — But besides these more important scientific 

 principles or ideas, there are a few minor ones which are of 

 sufficient interest to be briefly mentioned. In the article on 

 the " Bees' cells," (referred to in chapter xxviii.) I called 

 attention to a circumstance that had been, I think, unnoticed 

 by all previous writers. An immense deal of ingenuity and 

 of mathematical skill had been expended in showing that the 

 two layers of hexagonal cells, with basal dividing-plates in- 

 clined at a particular angle, gave the greatest economy of 

 space and of material possible ; and the instinct of the bees in 

 building such a comb to contain their store of honey was held 

 to show that it was a divinely bestowed special faculty. But 

 all these writers omitted to take into account one fact, which 

 shows their whole argument to be fallacious. This is, that 

 the combs are suspended vertically, and that when full of honey 

 the upper rows of cells have to support at least ten times as 

 much weight as the lowest rows. But there is no corres- 

 ponding difference in the thickness of the walls of the cells ; 

 so that, as the upper rows are strong enough, the lower 

 must be quite unnecessarily strong, and there is thus a great 



