404 REGINALD A. DALY 



same temperature crystals of olivine (3 . 40) , augite (3 . 30) , magnetite 

 (5.00J, labradorite (2.70); and anorthite (2.75) would have specific 

 gravities of, respectively, about 3.30, 3.20, 4.85, 2.61 and 2.66. 

 These values are only approximate, but they show the order of 

 the density differences to be expected between the phenocrysts and 

 their mother-liquor. It appears probable that all the crystals except 

 labradorite would slowly sink in the mother-liquor. 



Some idea can be obtained of the rate at which the heavier pheno- 

 crysts would sink. For a spherical body the velocity of sub- 

 sidence {x) at a time when steady motion is reached in a highly 

 viscous fluid, may be found by solving the equation : 



9 V ' 



where g= the acceleration of gravity, r, the radius of the sphere; 

 d, the density of the body; d', the density of the fluid; v= viscosity.^ 

 This equation has been experimentally verified by Ladenburg who 

 found that steel spheres, ranging from about 0.075 ^^ about 0.2 cm, 

 in radius took, respectively, from 570 seconds to 3,858 seconds to 

 fall through a 20-centimeter column of Venetian turpentine — a 

 substance a hundred thousand times more viscous than water.^ 



In an experiment by Jamin, pieces of stone sank through a layer 

 of pitch in the course of several days (quelques joursj, and corks 

 simultaneously rose through the pitch, which, at 6° C, is much over 

 1,000,000,000,000 times as viscous as water.^ 



It is thus clear that even if the viscosity of the lava, within the 

 temperature interval of early phenocrystic development, be many 

 thousand times that of water, the phenocrysts must tend to sink. 

 So long as such a crystallizing lava-column remains in its conduit and 

 there undergoes cooling through the temperature interval, 1200°- 

 1050° C. — a process necessarily involving a long period of time — the 

 settling of the magnetite, olivine, arid augite crystals will continue, 

 though at a continuously slower rate. Theoretically the settlement 



1 Cf. Poynting and Thomson, Textbook of Physics, Properties of Matter. (Lon- 

 don, 1902), p. 222. 



2 R. Ladenburg, Annalen der Physik, Vol. XXII, 1907, p. 287. 



3 Jamin et Bouty, Cours de Physique (Paris, 1888), Tome i. 2^ fascicule, 1888, 

 P- 135- 



