90 E. H. L. SCHWARZ 
occur on the Wynberg side of Table Mountain, but I have not the 
material before me; a large number of similar series of analyses 
that I have tried give the same result. 
Granite Dolerite Mean Mica-Schist ae 
SOF see eRe atone 76.1 50.2 63.1 66.2 SST 
AU Faas eae aay Corte 103} 54 I5.0 14.2 18.6 4.4 
INEOWNGHORS sco gob udioc it 68 16.9 g.1 Rok a 
IVS Oe eT eae es ee teae 2 5.8 3 112 ae 
CaO Gi Daids canon es 58 10.5 5-4 4 ae 
ONES Rest Wah eat realest 3 Boi DD 2.6 2.2 Las 
KO eee ane nes 4.9 T.4 Bott 3.9 8 
THE RESEMBLANCES BETWEEN THE DEVELOPMENT OF ANDALUSITE 
CRYSTALS AND IGNEOUS INJECTION 
In the case of the George andalusite we have a crystal substance 
injected into a rock. It first occupies cavities ready made for it, 
then with accession of further material, it dissolves, or as we should 
say In speaking of igneous rocks, melts up portions of the invaded 
rock, assimilates suitable substances and passes out the residue. _ 
Segregation bands develop where conditions are unfavorable, and ; 
included blocks or xenoliths entirely unaltered or more or less 
‘““metamorphosed”’ occur. All the phenomena of igneous injection 
are portrayed without the aid of any great heat, though hot water 
circulating under pressure was no doubt the agent by which the 
transference of material was accomplished. In the constancy 
of size also the masses of igneous rocks are simulated. The Zwart 
River crystals are about an inch in length—4 cm. perhaps would be 
the best average. None are much smaller or larger. A little 
farther on a similar but less perfect development of crystals occurs 
where the size is half that of those of the Zwart River material, and 
again in the chiastolite slates of the Transvaal a much smaller 
average is attained. So throughout the Karroo the dykes and sills 
are very constant in thickness; the laccolites of the Drakensberg 
Plateau and the granite bosses of the southwest are of approxi- 
mately equal bulk. The granite masses of Cornwall and Devon 
are another striking example. It would seem that as with crystals 
so with magmas, there is a limit of size. 
