280 Wisconsin Academy of Sciences, Arts and Letters. 
pressures and whether with long or short tubes or columns, the 
departures from the law of capillary flow have been systemati¬ 
cally in series either plus or minus instead of in single obser¬ 
vations, first one side and then the other, as should be expected 
if the observed departures were due to errors of observation or 
of manipulation. 
It has, from the first, been recognized that under certain 
relations of length and diameter of tube to velocity of flow 
the law of capillary flow did not hold and that the discharge 
fell below the amount computed. This falling away in the 
amount of visible work done has been explained by the absorp¬ 
tion of energy at the ends and within the body of the tube 
through the setting up of vortex or other movements of the 
fluid more or less transverse to the axis of the main stream, 
which necessarily absorb more or less energy, tending to pro¬ 
duce flow and thus to diminish the amount of discharge in unit 
time. But the fact that the flow may increase faster than the 
pressure does not appear to have been generally recognized or 
considered significant and we have nowhere seen it explicitly 
stated in any discussion of the law of capillary flow. 
It is true that a flow increasing more rapidly than the pres¬ 
sure could occur only under conditions where the resistance to 
flow became less as the pressure became greater, and as the 
earlier theoretical investigations of JIaxwell lead to the dedue- 
tion that the viscosity of gases must be independent of pressure 
there was no apparent means of explaining a flow which in¬ 
creased more rapidly than the pressure. 
The later experimental investigations of Bongten, 1 Warburg 
and Kundt, 2 Warburg and Sachs, 1 and Cohen, 3 appear to have 
established the fact that the viscosity of both water and air does 
change with pressure, that of water appearing to decrease in 
some amount or ratio with an increase of pressure. At any 
rate their observations indicate that for water the flow through 
capillary tubes increases faster than the pressure, even when 
1 Annalen der Physik und Chemie, Vol. XXII, 1884, pp. 510, 518. 
2 Annalen der Physik, Vol. CLV, 1875 p. 337. ( 
s Annalen der Physik und Chemie, Vol. XLV, 1892 p. 666. 
