208 W J MCGEE OUTLINES OF HYDROLOGY 



tends to extrude the extraneous sediment by internal movement, whicJi 

 again is in accord with the cryptocrystalline character and high density 

 of anchor ice formed on stream bottoms in rapids. The concept of the 

 liquid HjO as itself tending toward hexagonal forms similarly helps to 

 explain various optical phenomena, including refraction in bodies of 

 water, the brilliantly sparkling quality of pure sj^ring water, and perhaps 

 the form of refraction exhibited in rainbows and parhelia and also the be- 

 havior of bubbles and foams. The several analogies and inferences lead to 

 the inference that the water particle involved in stream-flow is a discrete 

 aggregation of molecules varying in size from considerably below that of 

 the air-formed drop indefinitely downward; with the suggestion that the 

 particles tend to assume hexagonal groupings at the surface and perhaps 

 on discontinuous planes ordinated therewith, and the further suggestion 

 that when in moderate motion the particles tend to increase in dimension;? 

 and concurrently to extrude extraneous matter by flocculation, freezing, 

 etcetera. Manifestly the size of the particle is measurably affected by 

 temperature (as it must be also by electrical tension, etcetera), and it is 

 convenient to think of it in a general way as varjdng inversely in size 

 ■ndth the temperature from a larger aggregation ready to form a crystal 

 of ice to a smaller aggregation ready to sever gravitative bonds and float 

 into the air as a vapor particle. 



In connection with this view of the water particle, it is noteworthy that 

 while the hexagonal tendency in water is so strong that crystals formed 

 by freezing will, as shown by Tyndall, normally assume no other form, 

 and that groups of bubbles are, as shown by Boys, stable only when every 

 point of complete contact is made up of three edges and six planes, each 

 pair 120° apart, the hexagonal structure is essentially bi-dimensional and 

 the structural units will not fall into stable or equipotential forms in the 

 tri-dimensional relation. Conformably, a special series of experiments 

 kindly conducted by Cameron and Free show that when similar and equal 

 spheres, such as shot (which if compressed in a single layer tend to form 

 similar hexagonal prisms), are brought together en masse they tend first 

 to gather into groups hexagonal on surfaces and occasional discontinuous 

 planes within, and that when further compressed they tend to lose the 

 hexagonal type and form indefinitely variable polyhedra, somewhat ap- 

 proaching yet never attaining either the rhombic or the pentagonal 

 dodecahedron or the icosahedron, and indeed display what may be con- 

 sidered an inherent tendency toward differentiation. All observations 

 indicate that if the water particles are uniform and regularly spheroidal 

 or polyhedral (though not cul)ical, which is inconceivable among cosmic 

 t3'pes) and form stable combinations in the bi-dimensional relation or in 

 a single layer, they can not form siable combinations in the tri-dimen- 



