SNOW-FLAKES 13 



and asked why. Cur autem sexangula? His most 

 hopeful suggestion was that chemists should find out 

 whether snow contained salts, and if so, what they 

 were. Descartes, to whom Meteorology and Optics 

 owe the first good explanation of the rainbow, figures 

 rather conventionally, several kinds of snow-stars, 

 which he observed at Amsterdam in February 1635. 

 The six-rayed star, branched and unbranched, the 

 rosette, the six-rayed star with intermediate rays, and 

 the pair of lamellar crystals joined by a prismatic rod 

 are all shown, and if we make some allowance for the 

 small size of the figures and the rudeness of the en- 

 graving, we may say that nothing is shown which does 

 not occur in nature. Robert Hooke in his Micro- 

 graphia (1665) gave many fairly good figures, and first 

 noticed that in branched snow-stars " the branchings 

 from each side of the stems were parallel to the next 

 stem on that side," an immediate consequence from 

 the fact that the secondary branches, in this case of 

 60, make the same angle with the primary axes, as 

 these make with each other. Rosetti of Leghorn in 

 1 68 1 first observed the extremely minute cavities 

 (" capillary cavities " or " air-spaces " of modern writers) 

 to be found in snow-flakes. Scoresby in 1820 pub- 

 lished an elaborate account of the form of the snow- 

 flakes of the Arctic regions, figuring 96 forms, and 

 classifying them under five principal heads. The figures 

 were completed symmetrically, and the dimensions 

 given. Glaisherin 1855 published 150 figures of snow- 

 crystals, all completed symmetrically. Scoresby's 

 and Glaisher's figures have been copied in many 

 common text-books. The next step was to photo- 



