258 A. Schrauf and EF. S. Dana— Thermo-electrical 
(d.) On erystals of pyrite the + por- 1. 
tions are sometimes distributed very 
irregularly (fig. 1), and any explanation <= 7" — — 
of this by the supposition of a twin |-4 ~ 
structure is impossible. We should 
have to assume, in this case, a parallel- | -—~ i t 
wise interpenetration of the different 4. 2 
individuals or lamelle. ee. + 
(e.) The majority of the pyrite crys- 
tals are negative in relation to copper; “ 
the positive portions appear often to be only thin layers, of a dif- 
ferent nature from the mass of the crystal. Homogeneous + 
pyrite crystals are exceedingly rare. 
These results show clearly that an investigation of the 
thermo-electrical properties of minerals is of value only when 
their chemical composition is known. It is well known how 
- =e 
at Ge il = + 
oh ot . 
the same composi 
quently upon their ac trey characters. On this account our 
ollowing investigations, was directed espe- 
certain minerals are especially well ada ted; and, in fact, out 
of a long series of sulphids, arsenids and tellurids of cobalt, 
iron, nickel and bismuth, we were successful in finding some 
examples capable of showing the relation between the thermo- 
electricity and the chemical nature of the minerals. It was nec- 
3. O Its in regard to the thermo-electrical character of 
the minerals tested related throughout to their behaviour in 
contact with copper. € position of the wire employed was 
trical series as given by Seebeck (L. ¢.), but on the other hand 
It was to be expected, in conse- 
quence of this, that the larger number of substances would be 
negative in relation to copper wire ; and the observations were 
in r h this. A number of minerals which gave no 
rceptible str and are consequently marked 0 in the fol- 
owing list, might perhaps be + in contact with a different 
*The use made of pyrite in En (mo ight milli ; 
srry a Toye oh, in wt. 
