144 Schaeberle—Collimation Constant of a Transit Circle. 
been questioned. This case simply illustrates the fact that 
most observers have not considered the possibility of two 
pyroxenes occurring together in the andesites, and the results 
obtained by Fouqué, who isolated both augite and hypersthene 
from the Santorin lavas, have been passed over as interesting 
but unique. 
In conclusion, the writer must not be misunderstood as 
h 
emical investigations are now. in progress upon several 
well-known rocks of Hungary which are cited by Zirkel and 
Rosenbusch as types of “ Augit-andesit,” and the results will 
be communicated at an early day. 
Denver, Colorado, December, 1882. 
Art. XVIL—A New Method for determining the Collimation 
Constant of a Transit Circle; by J. M. SCHAEBERLE. 
THE collimation constant of a transit instrument is usually 
determined by one of the four following methods: 
y observations of stars in reversed positions of the 
instrument. ; 
z. By reversal on a collimator or very distant terrestrial 
object. 
3. By combining observations made with the spirit level 
with nadir observations. 
B ssel’s method, in which two horizontal collimators 
(placed on opposite sides of the instrument and having their 
optical axes parallel) are used. 
he last two methods are usually employed in cases where, 
on account of the construction or size of the instrument, it 1s 
not advisable to reverse it very often; or, as in case of the 
Greenwich circle, where the instrument cannot be reversed at all. 
As the greatest possible accuracy is sought with instruments 
of this class, it is especially desirable to vary the processes by 
means of which the instrumental constants are obtained. The 
various results derived, besides serving as a check on data 
otherwise obtained, furnish means for investigating irregular 
variations in the so-called constants, due to changes in the 
material of which the instrument is made. 
thod which I am about to propose for finding the 
collimation constant, without reversing or disturbing the 1n- 
