MECHANICAL STRETCHING OF LIQUIDS. 
363 
of a bubble of definite length, which he reads off while observer 0 A notes the height 
to which the mercury in the tonometer has again risen. 
It is not always possible in practice to prevent the liquid from letting go its hold 
too soon, and it often happens that the bubble forms in one of the bulbs, and not in 
the tube, where only its volume can be read off, but this fact need not prevent an 
observation of the volume from being made, for, if the observer 0 3 only heats the 
platinum wire for an instant, the liquid below it, which has been clinging to it, 
separates, owing to its weight, and at once a bubble of hot vapour is formed, which, 
by its greater pressure, causes the bubble in the bulb, which also is at a lower level, 
to close up and vanish. Thus the bubble can be instantly transferred to the place 
where it can be measured. 
The method of liberating the bubble where it was wanted by means of the heated 
platinum wire was only hit upon after many other methods had been tried unsuccess¬ 
fully, but I had not foreseen that it would carry with it the advantage of enabling 
the observer to transfer the bubble in the manner just mentioned. This was a piece 
of good luck. 
Corrections to be applied to the Measure of the Bubble. 
(1.) Correction for Capillary Curvature. —This was applied in the usual way, the 
ends being considered as hemispheres of the diameter of the tube, which was 
2’20 millims. 
(2.) Correction for Inequality of the Bore near the Wire. —This was determined by 
floating the bubble when very small to a point in the tube 3 or 4 centims. below the 
platinum wire, and then allowing the whole liquid to cool to a steady temperature, 
thus forming a long bubble in a uniform part of the tube. The length of this bubble 
was then carefully measured, and then this bubble was “ transferred ” in the manner 
mentioned, so that its upper end was at the wire. The difference of the lengths 
gives the correction to be applied. The manner in which the upper end of the 
bubble was entangled with the platinum wire was not, however, always the same, 
but appeared to be always one or other of two configurations, and the correction to be 
applied for the variation of bore was -f- 1*55 millims. , or + ’015 millims., according to 
the configuration which was observed and recorded on each occasion. 
Sources of Error in Determining the Length of the Bubble. 
(1.) It was often impossible so to regulate the influx or efflux of heat as to secure 
that a stationary stage had been reached when the bubble was formed. Thus the 
liquid generally continued to shrink by cooling after having let go the walls, and on 
this account the bubble required to be read immediately. 
(2.) On the other hand, if the reading was made too soon there was not time for 
the liquid still clinging to the side of the tube to drain down. 
3 a 2 
