HOW MOLECULES MAKE MASSES 47 



though it may not be true of quartz, should clarify the mystery 

 Graham refers to. 



Petrographers have long noticed the incipient crystal forms 

 (crystallites) which appear in media where high viscosity and quick 

 cooling oppose crystallization, as in obsidian, pitchstone, etc. 

 Such forms can be obtained artificially from a solution of sulfur 

 in carbon disulfide, made viscous by the addition of some Canada 

 balsam. If a drop of this solution is rubbed out on a microscope 

 slide and blown upon to speed the evaporation of the solvent, any 

 or all of the following stages may be seen on microscopic examina- 

 tion: (1) tiny globules (possibly containing ultramicroscopic crys- 

 tals); (2) aggregations of these globules into bead-like strings or 

 margarites (Greek, margaritis, a pearl), or into geometrical groups; 

 and (3) tiny crystals surrounded by empty spaces or lacunae, indi- 

 cating that the surrounding material had gone to make up the 

 crystal. 10 



The various intermediate steps in the formation of visible 

 crystals are very often observable in window-pane ice, when con- 

 ditions are right. In fact, during a half-hour ride in a Fifth 

 Avenue bus in New York City on a very cold day, the transition 

 from the "ground-glass" effect of globules, to dendrites (tree or 

 fern-like formations), and ice-crystals with lacunae, took place be- 

 fore my eyes. Mr. Wilson A. Bentley of Jericho, Vermont, had 

 for years made photographs of a great many varieties of the frozen 

 forms of water, including snow, hoar-frost, hail, and window-pane 

 ice. On seeing some of these printed in the National Geographic 

 Magazine of 1923, and later in a copy of the Monthly Weather 

 Review for Nov. 1907, which Mr. Bentley kindly sent me, I recog- 

 nized each of the phenomena above mentioned. Besides, there 

 were a number of cases where specific dendritic forms indicated 

 the presence on the glass of sodium chloride and some protective 

 colloid, both perhaps coming from the soap used to clean the 

 window-pane. 



The usual course of crystallization seems to proceed in the fol- 

 lowing steps: (1) as drop in temperature or loss of solvent lessens 

 the capacity of the magma or solvent mass to hold the atoms, ions, 

 or molecules of the dispersed substance in true solution, they form 

 groups which clump together into tiny crystals or amorphous 

 heaps (crystallogens); (2) when these clumps become large enough, 

 their kinetic activity drops, and they are aggregated by surface 

 forces into spherulites; (3) the spherulites then unite to form 



