Temperature on the Viscosity of Gases, 203 



bottle L, of about 10 litres capacity, sometimes two bottles 

 each of that volume, connected between G and JV, served to 

 reduce the rate of pressure-fluctuations arising from irregu- 

 larities in the aspiration. At iVwas connected the pressure- 

 regulator, which I will describe in detail, as it is, so far as I 

 know, wholly new, and is available for many purposes. 



The aspirator under a given head of water can exhaust at a 

 definite rate, maintaining a certain exhaustion ; if the head be 

 increased or lessened, the rate of exhaustion will be increased 

 or diminished, and thus if the leakage into the exhausted space, 

 as through the capillaries in this case, be sensibly constant, 

 the pressure in the exhausted space must diminish or increase. 

 It is often impracticable, as it was for me, to maintain a con- 

 stant head of water, and some means is therefore necessary to 

 overcome this irregular action. The arrangement used was 

 that shown in the sketch. A large glass tube, M, open at 

 both ends, stands upright in an open mercury-trough of con- 

 siderably larger dimensions. Into its top is inserted a rubber 

 stopper with two borings, through one of which passes one 

 arm of a T-joint, whose other two arms are connected respec- 

 tively with the aspirator and the vessel to be exhausted. 

 Through the other boring passes a tube of 1 to 2 millim. dia- 

 meter bent twice at right angles, and dipping into the mercury 

 in the trough. Suppose that the vessel to be exhausted is close 

 (the apparatus covers cases ranging from this to a leakage of 

 about half the rate of exhaustion of which the pump is capable), 

 and the aspirator set in operation. As the exhaustion pro- 

 ceeds, the mercury rises equally in both large and small tubes 

 until the mercury in the trough drops below the point of the 

 latter ; whereupon, as soon as the excess of external pressure 

 is sufficiently great to overcome the friction of the mercury, 

 the column in the fine tube rises rapidly and flows over into 

 the upper part of the large tube and upon the mercury surface 

 in it. This overflow is of course immediately followed by an 

 inrush of air through the open point, and a consequent sudden 

 lowering of the mercury in the large tube with a correspond- 

 ing rise of level in the trough, which closes, or partly closes, D. 

 This first sudden action is followed by several of lessening 

 violence, and a steady condition is soon reached, in which 

 there is a continuous inflow of air and small drops of mercury 

 at the point, the proportion between the two depending upon 

 the relation between the sizes of the tubes, the capacity of the 

 aspirator, and the supply of gas from the vessel to be exhausted. 

 If a fluctuation occurs in either the rate of the aspiration or 

 the supply of gas to be exhausted, a greater or less proportion 

 of air will be taken in, with a corresponding but very small 



