388 Penfield and Pratt—Relation of Manganese and 
cleavage the following plan was adopted: A fragment show- 
ing the desired cleavage was fastened to a glass plate and while 
the cement was still soft the mineral was shifted in position by 
means of two small wedges running at right angles to one 
another until the reflection of a distant object, as a window 
bar, from the glass plate and the cleavage coincided, when the 
preparation was turned in different positions, thus showing 
that the cleavage was parallel to the glass. Then at the sides 
of the fragment glass plates of equal thickness were cemented, 
so as to form a large wearing surface, and the mineral ground 
away until even with the glass. The surface was next pol- 
ished and a plate or prism prepared as occasion demanded. 
The chemical composition of this group of minerals having 
been fully established by the complete analyses already 
referred to, the ferrous iron determinations were the only ones 
required for the present investigation, and these were made by 
dissolving the minerals in sulphuric acid and titration with 
potassium permanganate. The composition of the pure end 
products is as follows: 
PO, =45'3 | ( P.O = 458 
Lithiophilite § MnO=45: Triphylite < FeO =45°5 
ii O: = 9)6 ) Ly Oe=a85 
Therefore if the amount of FeO is known, the percentage of 
MnO may be told within 1 per cent by deducting the percent- 
age of FeO from the mean percentage of the bivalent oxides, 
45°3. | 
The results of the investigation are given in the following 
table : 
As seen from the table, the optical characters undergo a very 
considerable change due to the interchange of manganese and 
iron, and one that is perhaps greater than might be expected ~ 
considering how nearly alike the two elements are in atomic 
weights. With an increase in iron there is an increase in the 
indices of refraction, and also the divergence of the optical 
axes changes rapidly. The material from Grafton happened 
to be of such a composition that it illustrates this latter varia- 
tion in a very striking manner. Jn this special case the min- 
eral is practically uniaxial for yellow light, although there was a 
slight variation in the character of the material, for when the 
section that was prepared was moved about in the polariscope 
there were some parts where the interference cross opened up 
to a very slight extent. At that particular part, however, 
where the measurements were made the material was uniaxial. 
Owing to the strong dispersion characteristic of the group, the 
plane of the optical axes for green was 001, as in the previous 
eases, while for red it was at right angles to the latter, or 100. 
The interference figure, therefore, as seen in the polariscope 
— 
