DowLiNG — Sleachj and Turbulent Motion in Gases. 383 



thi'.ougli tlie tube, to another. In these experiments, it was noticed tliat the 

 conditions were much steadier, which was probably due to the fact that 

 the interior of the tube was completely protected from outside influences. 

 Those of tlie former curves with air, which did not appear satisfactory, 

 were accordingly repeated with the double gasometer arrangement. Such 

 curves are indicated below by the title " two gasometers." The method of 

 ■working in this case will be dealt witli more fully in a later part of the paper. 

 An examination of the curves (Nos. 1 to 1-3, Plates XXXTI. and XXXTII.) 

 shows us that in general for each fiibe there are two "discontinuities" on the 

 curve. Tliere avefoiir instances, however, where this is not tlie case. Curve 

 No. 1 refers to the experiment first carried out and already described. A 

 few numbers only were taken, and these were mostly in the vicinity of the 

 ordinary " turbulence " discontinuity. Curve No. 5, Plate XXXII. (for 3-cm. 

 tube), shows only one discontinuity; while Nos. 10 and 13, Plate XXXIII. 

 (for 7'4 cm. and 1'9 cm. copper tubes), also show only one. We will later 

 consider the significance of the second discontinuity on the other curves, 

 and for the present confine our attention to the turbulence effect. 



One remark may be made with regard to the form of the curve 

 immediately after a discontinuitj'. In many cases the curve simply 

 continues sloping upward, but at a rate different from before. In other 

 cases, however, tliere is a sudden drop, followed usually by an increase. 

 This occurs with the narrower tubes. Just before the discontinuity the 

 curve is in these cases becoming more and more horizontal, indicating 

 that, with the high velocity then used, the greater part of tlie ions formed 

 by the radium are being carried riglit down to B. When turbulence 

 sets in, fewer ions are left in the gas on its arrival at tlie electrode, and 

 consequently a falling off occurs in the electrometer readings. 



The following table (Table I.) collects the results for the turbulence effect. 

 The table includes only one number for each tube. In most cases, 

 however, several independent experiments were made on each separate 

 tube ; but, on the whole, tliese results agreed well amongst themselves. 

 It seems unnecessary to give more than a single result for each size of tube 

 (especially as the degree of accuracy attained would seldom appear to be 

 more than five per cent.). Tiie tube bores were measured with calipers across 

 two perpendicular diameters at each end. This method was accurate enough, 

 considering that the tubes were mostly of rather wide bore and fairly 

 uniform. The temperatures are those given by a thermometer in the air 

 near the apparatus. The pressures were taken from a barograph record. 

 The viscosities were calculated by Sutherland's formula (Kaye and Laby's 

 tables). The density was cnlculated from the pressure and temperature. 



SOlF.f^T, PEOC. R.D.S., VOL. XIII., NO. X.WI, 3 Jj 



