MR. J. B. HANNAY ON THE MICRORHEOMETER, 
277 
arrived at a different result. Graham also published a table from 0° to 70°, but 
Guerout does not appear to be cognisant of Graham’s work. According to Guerout, 
the curve representing the effect of temperature would be a straight line, whereas both 
Poiseiulle and Graham’s numbers correspond with a curve. This shall be considered 
further on. 
About four years ago it occurred to me that as all the work hitherto done with the 
capillary tube had not fairly tried its usefulness (either by reason of not considering 
temperature, or of using percentage instead of normal solutions), a fuller investigation 
of the subject might yield important results. The term “normal solutions” is used in 
the sense which chemists attach to it, namely, solutions in which the molecular weight 
of the compound in grammes is dissolved in a litre of water—or, shortly, grammes 
equivalent per litre. Where this could not be obtained owing to the insolubility 
of the salt, the strongest possible solution was taken, and its proportional effect 
calculated. In using the tube of Poiseiulle, which was also used by Graham and 
Guerout, I found a good deal of inconvenience. The apparatus consists of a tube 
bent at right angles, the horizontal lower portion being the capillary, and the vertical 
part consisting of the measuring bulb and tube for applying pressure. The whole was 
immersed in the liquid whose rate was required, and alternate suction and pressure 
applied to fill and empty the bulb. This requires so much of the liquid to be experi¬ 
mented upon that I modified the apparatus, as shown in Plate 35, fig. 1, by placing a 
receiving bulb at the other end of the capillary. But even this had its disadvantages, 
as the pressure-tube had to be placed alternately at each end, and this disturbance 
sometimes caused dust to fall into the bulb and choke the tube. It had, however, this 
advantage over Poiseiulle’s tube—that liquids like bromine could be experimented 
upon. After working some time with that form, I finally decided to adopt the form of 
apparatus shown in Plate 35, fig. 2. It consists of a capillary tube at the bottom, 
attached to which are two bulbs, one for the purpose of measurement and the other as 
a receiver. Two tubes pass up from these bulbs and end in Y-pieces. One of the 
Y-tubes from each tube is connected with an inverted Y-tube, and each is also provided 
with a clip or stop-cock. One of the inverted Y-pieces is connected with the pressure 
chamber, while the other is left open to the air. By this arrangement the pressure 
can be turned on to either side, and moreover the apparatus can be filled with any gas 
other than air. To introduce the liquid to be experimented upon, it is only necessary 
to undo the connexion of one of the Y-tubes on the measuring bulb side, and run in 
sufficient liquid. A thermometer is placed between the two bulbs, and its bulb is 
situated at the mean point of the whole system, thus registering the average temperature 
of the experiment. In the cases of some mixtures it was found that after some time 
they deposited a slight sediment which choked the tube. To obviate this another form 
was devised, as at Plate 35, fig. 3, where the receiving bulb is at the lower end, so that 
the liquid may be allowed to settle here for some time, and the upper portion forced 
up into the bulb free from any solid matter. 
