142 G. H. KNIBBS. 
1V.—InpicaTIONS FOR A FURTHER DETERMINATION, 
28. Apparatus.—Some observations in regard to the gener 
disposition and dimensions of apparatus for a further determin 
tion by the efflux method of the viscosity constant, may notlt : 
inappropriate in concluding this examination of the subject — 
Fig. 4 indicates the general arrangement of the essential pa | 
best adapted to the purposes of accurate measurement'—it i ; 
essentially Poiseuille’s. The pipe P is connected with the sour 
of pressure—compressed air contained in a large reservoir it : 
Poiseuille’s' experiments,—continuous with it is the pipe Mf cm 
municating with the manometer. When the cock-valves X and 
are turned the pressure compels the contained liquid in the 
reservoir bulb V, to flow through the tube JA, the capillary 44, | 
and the efflux tube B HZ, whence it emerges into the water in 8 | 
heating bath W, carefully maintained at constant temperature 
and height, the temperature—that at which the viscosity is to be : 
determined—being measured by the thermometers T, 7. boul 
tubes JA and B E ought to be of the same diameter, and ee 
terminals 4 and B of the capillaries may if possible be inset — 
well within them. The small bulb U7 is intended to hold suffice 
liquid to permit the flow to be thoroughly established under al 
general conditions of an experiment, before the measure 
actually commences, and, with the rest of the apparatus men ion 
is wholly submerged in the heating bath. 2 is a micros¢ 
short-focus telescope—for noting the instant at which the ! 
of the concave surface of the fluid crosses the line J on the we 
This microscope after the observation has been made should 
lowered exactly the distance IJ, so as to observe the instant 
passage across the line at J, the interval between the two ™ 
the time of efflux (7). A chronographic record of these inst 
is the more satisfactory. 
The valves X and LZ should be closed immediately on the term 
ation of an experiment, so that impurities may not entet © 
- The figure is not even approximately to-scale; it is essentially oO . 
