Oct. 13, 1 881] 



NATURE 



571 



these considerations is that the discharges at the two terminals 

 of the tube are so far independent as to be primarily deter- 

 mined each by the ccmditions at its own terminal, and only 

 in a secondary degree, if at all, by the conditions that exist at 

 the opposite terminal. And since the discharges are not neces- 

 sarily the same at both terminals, the tube must contain free 

 charges at different time-. A tube is therefore not like a con- 

 ductor, but is an independent electrical system, holding much 

 the same po--ition as the air-vessel in a forcing-pump. All the 

 electricity that gnes into it goes out again, but this is true only 

 when we consider the whole discharge from the beginning to 

 the end, and it may not he true even approximately during a 

 small finite time. Tliis independence of the discharges from 

 the two terminals in the passage of electricity through rarefied 

 gases dissipates the error of seeking anal igies in metallic con- 

 duction ; and shows that any obedience to regular laws as to 

 change of potential as we proceed along the tube, resistance, 

 &c., must arise from the fact that the effects measured are really 

 average effects over an interval of time very long compared 

 with the duration of the individual di--"charges. 



The importance which attaches to the negative end of the 

 discharge has led experimenters to examine whether the features 

 appertaining to it could not be still further enlarged. And the 

 only thing requisite to carry the experiments to a limit was an 

 instrumental method of improving the vacuum to the degree 

 required. This was furni-hed by Mr. Crookes' refinements on 

 the Sprengel pump. In a series of most remarkable experi- 

 ments he has shown, as mentioned above, not only that the 

 striated column may be reduced to zero, but that the anchored 

 stria itself may be so driven back that the blank space in 

 question may be made to occupy the entire length of ihe tube. 



When an exhaustion mch as that described, or an exhaustion 

 nearly equal to it, has been reached, a phenomenon, previously 

 noticed, but not before made the subject of serious inquiry, 

 presents it elf. Certain parts of the interior surface of the tube 

 become luminous with phosphorescent light. The colour of 

 this light depends on the nature of the glass, and not in any 

 way on the nature of the residual gas within the tube, nor on the 

 substance of w hich the terminal is made. With German glass 

 the phosphorescent light is green, with English glass it is blue. 

 But the portion of the glass thus rendered luminous depends 

 upon the form and po-ition of the negative terminal. This 

 phenomenon is supposed to be due to the .■ treams of gaseous 

 particles shot off from the neighbourhood of the negative ter- 

 minal during the discharge. Although there is reason to think 

 that the e streams are an accompaniment rather than an integral 

 part of tlie discharge, yet the particles would seem to be them- 

 selves charged with electricity, inasmuch as their paths are 

 affected by a magnet, just as is a movable conductor carrying a 

 current, or a charged body in rapid motion. 



The whole subject of these streams, their power of heating 

 metals and other substances, the shadows cast by bodies inter- 

 posed in their path, and other properties of them, have been so 

 well and so fully illustrated by Mr. Crookes both in published 

 memoirs and in a lecture before this Association, that it is un- 

 necessary for me now to dwell upon the subject in detail. 



Their nature and properties, however, having been thus in 

 the main determined, these streams have proved a valuable 

 auxiliary in an investigation of what have been called the 

 "small time-quantities" involved in the discharge. The dis- 

 charge is, as has been already shown, a complex phenomenon, 

 the various parts of which, although not entirely separable, may 

 be shown to occupy different periods of time ; and the length of 

 these periods may be compared w ith one another, and with other 

 known electrical phenomena. We cannot, it is true, make any 

 abso ute determinations of the time occupied in any of them, 

 but we may still form a table of relative miguitudes of these 

 small time-quantities. And I will now endeavour in a fewr 

 words to give some idea of the nature of these quantities, and 

 of the method whereby they have been measured. 



If we take a tube of such high exhaustion as to cause the 

 discharge to become intermittent, or if we use a positive air- 

 spark of sufficient length with a tube of fair exhaustion, phos- 

 phorescent light, due to molecular streams, will be seen on the 

 inner surface of the glass near the negative terminal. If then 

 a patch of tinfoil, connected with earth, be placed on any other 

 part of the tube, it will cause negative relief discharges to take 

 place from the glass immediately within it, producing phosphor- 

 escence on the opposite side of the tube. 



If any solid object, such as a piece of wiie, should be present 



in the tube below the point of contact, it will cast a shadow on 

 the phosphorescence, precisely as in Crookes' experiments with 

 the streams from the negative terminal. If there be two points 

 of relief contact, the same object will throw two shadows, in 

 directions conformable with radiations from each. To these, 

 other experiments might be added. 



A determination of the precise directions in which these mole- 

 cular streams issue from a relieving surface is not a very simple 

 problem ; and we must here content ourselves with showing that, 

 in the case of intermittent discharges at least, the streams do 

 not is>ue normally. If a strip of tinfoil placed along the tube 

 be used as a relieving surface, the phosphorescence takes the 

 form of a sheet wrapped round the tube ; if the strip be 

 wrapped round the tub', the phosphorescence takes the form of 

 a sheet laid along the tube. If contact be made with the finger 

 over a finite surface, or by a ring of wire laid close upon the 

 tube, the phosphorescence takes the form, approximately, of 

 the evolute of an ellip-e. In all these cases the illuuiination is 

 somewhat irregular ; but the geometrical elements of which the 

 phosphorescent figure is composed, and the stripes or striations 

 of more intense light, are always formed at right angles to the 

 longer dimension of the contact piece. This being so, suppose 

 that we place on the tube a strip in such a curve that the normal 

 planes to the curve will pass through the tangent at the corre- 

 sponding point of the image of the curve, i.e. the curve on the 

 opposite side of the tube, each point of which is exactly oppo- 

 site to a point on the tinfoil. In such a case all the striations 

 will lie along the curve formed by the locus of the central 

 patches of phosphorescence, and the re-ult will be a single 

 bright curved line of phosphorescence without any spreading 

 out or striated margin. The cun'e fulfilling these conditions 

 will be a heli.x, whose pitch is half a right angle. Experiment 

 confirms the anticipation. 



One more step in the study of these molecular streams is 

 necessary for our present purpose, namely, an application to 

 them of the >ame method which we have used with the electrical 

 discharges themselves ; viz. we must examine the effect of an 

 inductive stream produced ab extra upon a du'ect stream due to 

 the discharge inside the tube. These effects may be described 

 generally a^ the interference of molecular sti'eams. 



If the finger be placed upon a highly exhausted tube through 

 which a discharge with a positive air-spark is passing, the phos- 

 phorescence due to the molecular streams from the negative 

 terminal is seen to fade away from the place where the finger 

 rests, and from a region lying thence in the direction of the 

 positive terminal. The effect is that of a shadow over that part 

 of the tube ; and as this is produced not by any real intervening 

 object, but by an action from outside, we have termed it a 

 " virtual shadow." The phenomenon is due to a beating down 

 of the streams of molecules coming from the negative terminal, 

 by the transverse streams from the side of the tube immediately 

 within the part touched. 



The interference of two molecular streams may be further 

 illustrated by a variety of experiments, and in particalar by 

 arranging within the tulie a conductor of some recognisable form 

 — say skeleton tetrahedron. If the tube be touched at a place 

 opposite to this object a shadow of the latter will be formed in 

 the relief phosphorescence; but if the tube be touched also at a 

 point on which the conductor rests, the shadow will be splayed 

 out in a striking manner. This splaying or bulging of the 

 shadow is due to the interference of the molecular streams issuing 

 from the surface of the conductor, which then acts as a quasi 

 negative terminal, with the original relief streams issuing from 

 the first point of contact. 



With the help of these properties we are able, by connecting 

 a patcli of tinfoil on the tube with earth, or with the negative 

 terminal itself, or with a second patch elsewhere on the tube, to 

 detect the presence or absence of a demand for negative elec- 

 tricity ; to localise the main seat of such demand ; and even to 

 compare the electrical condition of different parts of the tube at 

 the same time or of the same parts at different times during the 

 very passage of the discharge. In this way we approach the 

 que-tion of the small time-quantities involved in the discharge. 



And, in the first place, it must be understood that the whole 

 duration of one of these intermittent discharges is comprised 

 within a period of which the most rapidly revolving mirror has 

 been incompetent to give any account. It may be in the recol- 

 lection of a few of my audience that when the discharge from 

 my great mduction-coil was exhibited at the Royal Institution 

 with tubes on a revolving disk the discharge showed a durational 



