10 Transactions of the Royal Mieroscojoical Society. 
study of thin sections of rocks. For some time I feared that it 
never would be possible to obtain satisfactory results with sectious 
sufficiently thin to be useful for ordinary microscopical examination, 
since it appeared probable that the errors of observation would be as 
great as the differences between the indices of the various minerals 
usually met with in rocks. In order to identify these with confi- 
dence, there ought to be no considerable error in the second place 
of decimals of their measured indices. I, however, now find that it 
is possible to make sufficiently accurate measurements with sections 
less than rhu of an inch in thickness, which b^d been prepared 
many years ago, and ground down as thin as was thought desir- 
able for ordinary microscopical examination. Though my appa- 
ratus is still far from perfect, I have been able to obtain satis- 
factory results with sections only 5 ^^ of an inch in thickness, and 
even less. 
Aj)plicatio7i of the above Method to the study of thin Sectio7is of 
Rochs. 
When first I commenced to apply this method with thin 
sections, I very naturally dealt with them in the same manner as I 
bad previously adopted in the case of thicker specimens, and 
determined the values of t and d by means of the scale attached to 
the body of the microscope. The next step was to determine their 
values by means of the rotation of the roughly graduated circular 
head of a well-constructed, fine adjustment screw. I found that 
with a y%- object-glass of small aperture a difference in adjust- 
ment corresponding to yyy of a revolution was just certainly 
apparent, which corresponds to about tg^fo of inch. Conse- 
cutive measurements may vary by several such divisions, but by 
taking the mean of a number of observations it appears to me 
possible to measure to at all events tofof of an inch, and 
probably less. Since the true value of each revolution gradually 
diminishes from the top downwards, the mean value of d must be 
determined from observations made both at the upper and lower 
limits of the range of the adjustment required to determine the 
value of t. If, however, the range he small, no such precaution 
need be taken. Of course when thus employing a microscope for 
such extremely accurate quantitative observations it is absolutely 
necessary that all the movements should be thoroughly well made, 
and that no other part but the fine adjustment should by any 
accident move to such an extent as even to-ooo of an inch. A 
general construction, which would do well enough for mere mag- 
nifying purposes, might thus he totally unfit for such refined 
quantitative observations. 
The above results were obtained with my ordinary condenser 
