The American Naturalist. 
[August, 
■i' 
^iL 
r for the sake of clearness not shown. The filter, c, is easily made a 
follows : take a piece of glass tubing twice the 
length of the required filter, heat it red hot, 
draw it out to arm's length and break it in 
two ; remove all the capillary part except about 
three inches on each half, heat, and bend into 
the required form (c) ; next carefully heat the 
capillary portion in a small alcohol flame, draw 
out exceedingly fine, and break off" so as to 
To use the differentiator, proceed as follows : 
Suppose objects fixed by corrosive sublimate 
are to be studied in balsam after staining with 
borax-carmine. Fill the filter with sublimate 
solution and insert a plug of cotton (previously 
boiled in water to remove the air) at the 
u-bend. Join the object-box to the filter, fill 
up with sublimate solution, and push a plug of 
cotton into the lower end of the box, avoiding 
bubbles. Wrap the cotton in fine linen if the subjects are minute. 
Put the objects into the box, plug the upper end in the same manner 
as the lower end, and finally join the box and filter thus filled to the 
empty reservoir a, and hang the instrument up in the position here il- 
lustrated. The objects are now transferred to 33 per cent, alcohol in 
the following manner : mix equal parts of sublimate solution and 33 
cent, alcohol (call this mixture 2). Mix equal parts of 2 and subli- 
mate solution (call this i). Mix equal parts of 2 and 33 per cent, 
alcohol (call this 3). Add mixture i to the reservoir until it is one- 
fourth full, mixture 2 until it is half full, mixture 3 until it is three- 
fourths full, and then fill up with 33 per cent, alcohol. If the suc- 
cessive mixtures are added with sufficient care, they will, owing to 
difference in specific gravity, remain distinct. If forced rapidly in, a 
nearly uniform mixture of about equal parts sublimate solution and 2>?t 
per cent, alcohol will result. The desirable procedure lies between 
these two suppositions, and gives rise to a uniform gradation of differ- 
entiation from sublimate solution to jj per cent, alcohol in passing upwards 
through the reservoir. The flow, which at once commences drop-wise 
from the point of the filter, should be so regulated, either by tipping 
the instrument, or by breaking off" more or less of the capillary part 
of the filter, as to cause the reservoir to be emptied in from two to five 
hours, when the objects will of course be in 33 per cent, alcohol, hav- 
