Gabbroid Rocks of Minnesota. — Winchell. 207 
and crystallized between other minerals. The brass-yellow 
color with strong metallic lustre in reflected light is very char- 
acteristic. The refringence is higher than that of augite. 
Alteration to limonite has not been observed in the diabases, 
but pyrite seems frequently to furnish iron for the formation 
of biotite and clinochlore. 
Ilmcnite seems to occur in small quantity accompanying 
the much more prevalent magnetite. 
Finally vermicular quartz has been found in small quantity, 
and in one case primary quartz seems to occur in anhedral 
form. 
Inclusions. The labradorite (as well as the other 
minerals) is usually quite fresh and its inclusions can be 
studied to advantage. It nearly always contains the minute 
black dust-like inclusions which have already been described 
as occurring in the normal gabbro; they are usually less 
abundant or altogether wanting about masses and grains of 
magnetite. The black acicular inclusions so often found in 
gabbros are often present in their usual form, and are some- 
times very abundant. The ilmenitic colored microplakites and 
microphyllites have not been found in the diabases. The 
opaque black needles are nearly always markedly lacking 
about grains and masses of magnetite, as if the latter had in 
crystallizing drawn to and incorporated in themselves the 
elements which elsewhere form these inclusions. Therefore 
the needles are considered to be of the same composition as 
the masses. To determine the nature of the latter the heavy 
minerals were separated out of the rock powder with tetra- 
bromide of acetylene; from these heavy minerals the black 
metallic mineral was separated out by careful mechanical 
selection, grain by grain; finally these grains were carefully 
examined with the microscope, and every one containing any 
foreign element was separated out. The weight of absolutely 
pure mineral thus obtained was .002 gr. This amount was 
fused with potassium bisulphate, and the fused mass was tested 
by putting a bit in a few drops of sulphuric acid to which had 
been added about .0005 gr. of morphine. A similar test was 
made using hydroquinone instead of morphine. Both 
methods* gave remarkably good tests for titanic acid. Re- 
*The methods followed give the color reactions proposed by L. 
Levy: Sur quelques reactions colorees des acidcs titanique, niobique, 
tantalique, et stannique. C. Rendus, CIII, 1886. 
