364 
PROFESSOR A. M. WORTHINGTON ON THE 
Each of these two sources of error tends to make the bubble appear too long, %. e. 5 the 
measure of the strain too great; and it is, without doubt, to their action that the 
irregularity of the observations must be chiefly attributed. Fortunately, the degree 
of steadiness of the tonometer just before the liquid is made to let go enables the 
observer to discriminate somewhat between “good"’ and “less certain” observations. 
Diffic ulties and General Observations. 
It is in this connection, and before passing to the measurements themselves, that 
the difficulties attendant on the experiments may be alluded to. In the first place, 
considerable tensions (above 12 atmospheres) are not easy to attain, the liquid letting- 
go its hold unexpectedly and too soon; sometimes this will continue for an hour or 
two, the bubble appearing perpetually in the same place in one or other of the bulbs, 
or at one of the bends. This is, perhaps, attributable to the liberation of air from 
minute crevices in the glass, which air has to be dispersed before adhesion can be re¬ 
established. For this purpose I have often found the application of ice or of a freezing- 
mixture to the affected part to be efficacious, especially when accompanied by judicious 
sharp tapping of the apparatus, so as to secure an impulsive pressure of the liquid 
against the surface at the non-adherent part of the glass, whereby I imagine that the 
minute bubble or film of air is the better dispersed. It is also useful, by manipulation 
and variation of the temperature of different parts of the liquid, to produce currents 
which sweep away from the narrow tubes into the wider bulbs portions of the liquid 
that may have become somewhat charged with dissolved gas. Yet at all times the 
behaviour of the liquid, as regards its adhesion, is somewhat uncertain. Thus, there 
may be no difficulty in reaching, time after time, a tension, say, of 9 atmospheres, the 
liquid always separating from the wall at some particular place ; and when, after an 
hour or more of patient manipulation, the observer is beginning to think the disease 
incurable, and that the apparatus must be reopened and the wffiole process of filling 
and boiling-out gone through again, the difficulty will unexpectedly disappear, or 
change. 
Meanwhile, and especially when high tensions have been reached, there is constant 
danger that the suddenness of the release may break the whole apparatus; and this 
is the more to be feared when the release is an accidental one, taking- place in one of 
the bulbs, than when it is intentionally effected at the wire in the narrow tube, for in 
the latter case the friction of the long column of liquid against the walls of the tube 
appears to act as a brake, and the shock to the apparatus is less sudden. Four pieces 
of apparatus similar to that described broke in this way, under the sudden release, 
before I was able to obtain any determinate measures. I have endeavoured to 
diminish this danger by connecting the two stems, AG and DE (fig. 1), and the cross¬ 
connecting tube as rigidly as possible to a light board when experimenting, so as to 
prevent these parts acquiring any considerable momentum. 
