Magmatic Differentiation In Effusive Rocks (593801f6-d89d-478d-99f3-52be1aad4963)

Discussion of the paper of SIDNEY POWERS and ALFRED C. LANE, presented at the New York meeting, February, 1916, and printed in Bulletin No. 110, February, 1916, pp. 535 to 548. N. L. BOWEN, Washington, D. C. (communication to the Secretary*). -This detailed study and. complete demonstration of differentiation as a result of the sinking of crystals has a great general significance in petrogenesis. If this action can be demonstrated in an effusive body, even though a very thick one, how can one doubt its possibility in large intrusive bodies? The authors speak of the sinking of augite and the rising of plagioclase but I should consider that they should depend entirely on the former action to obtain their results. It is true that in my experiments, to which the authors refer, a floating of crystals was in one case obtained, viz., in the case of tridymite crystals, but it must be remembered that tridymite has an exceptionally low density. Plagioclase crystals, on the other hand, are of nearly the same density as basaltic liquid. Under magmatic conditions they might be a little heavier than the liquid or a little lighter, we do not know which, but the margin would be exceedingly small and any motion resulting, whether upward or downward, would be correspondingly small in the usual case. For this reason I should think that the sinking of pyroxene should be considered probably the sole factor in producing the observed result, for which purpose it appears to be entirely adequate. The measurements of grain have also a certain significance in this connection. In the experiments referred to a difference of grain was obtained in different parts of the charge, which was entirely independent of the rate of cooling but connected with the fact that a sinking crystal does not depend entirely upon diffusion for the collection of its material from the surrounding liquid. The crystals which sank a considerable distance were therefore exceptionally large. Now in the Cape Spencer flow the augite crystals are normally (top and bottom) somewhat smaller than the plagioclase crystals, but in the layer somewhat below the middle,