146 Proceedings of Royal Society of Edinburgh. [sess. 
simply an arrangement for moving the piston accurately to short 
distances by means of a screw. The male screw is fixed to the end 
of the tube L, while the nut closes the end of the tube, as shown. 
This screw should have a very wide pitch. One turn of the screw 
should advance the piston at least 2 mm. — that is, according to size 
of cylinder, should push the piston forwards sufficient to displace 
i c.c. The object of the wide pitch is to enable the movement to 
be made quickly. The nut is provided with a projection which 
comes against a stop on the male screw at the point furthest in, 
as shown in fig. 5. The circumference of the nut has a scale engraved 
on it. The rod S fits easily inside the tube L, and presses against 
the closed end of the nut. This rod should be of such a length as 
to prevent the piston being drawn out more than \ or 1 cm. 
The reason for using the rod S, and not allowing the piston rod to 
be drawn out to the end of its stroke, and there come into contact 
with the screw, is, that the great amount of air in the long length 
of the cylinder might be so much influenced by change of tempera- 
ture as to make it impossible accurately to measure very small quan- 
tities. When using this part of the apparatus the rod S is put in 
its place, and this reduces the stroke of the pump to less than 1 cm. 
Before drawing out the pump, the screw R is turned back from its 
stop till the index points at the quantity we wish to measure, say 
1, on the scale. The piston is then drawn out till stopped by the 
rod S, and after the stopcock is turned the screw R is turned till its 
movement is arrested by its stop. By these movements the desired 
quantity is rapidly and accurately measured, and little time is given 
for the dust to settle in the apparatus. After the quantity of air is 
measured in this way, the stopcock F should be turned before the 
receiver stopcock is opened ; and before making the next test the 
cylinder B should be emptied and a fresh charge taken in. 
The tightness of the piston in the measuring cylinder should be 
tested from time to time, because it is evident that any leakage will 
give rise to errors in the number of dust particles counted. The 
manner of testing is as follows : — The stopcock F is put up against 
the stop Hj, and the air in the receiver is thoroughly purified. The 
stopcock is now turned against the stop H 2 , and two or three strokes 
of the measuring piston made. By these movements filtered air is 
taken out of C, and again returned. If the piston is not quite tight, 
