152 Royal Society. 



IV. Flocculent striae, such as are usually seen in carbonic- acid 

 tubes, are a compound phenomenon. They are due to a succes- 

 sion of short-lived elementary striae, which are regularly renewed. 

 The positions at which they are renewed determine the apparent 

 proper motion of the elementary striae. If they are constantly 

 renewed at the same positions in the tube, the flocculent striae will 

 appear to have no proper motion and to remain steady. If they 

 are renewed at positions nearer and nearer to the positive ter- 

 minal, the proper motion will be the same as that of the ele- 

 mentary striae ; if they are renewed at positions further and far- 

 ther from the positive terminal, the proper motion will be reversed. 



V. The velocity of proper motion varies, other circumstances 

 being the same, with the diameter of the tube. This was notably 

 exemplified in the conical tube. In tubes constructed for spectrum- 

 analysis the capillary part shows very slight, while the more open 

 parts often show considerable proper motion. 



VI. Speaking generally, the discharge lasts longer in narrow 

 than in wide tubes. In spectrum-tubes the capillary part gives 

 in the mirror an image extending far beyond that due to the wider 

 parts. 



VII. The coil-discharge appears, in the earlier part of its de- 

 velopment at least, to be subject to great fluctuations in extent. 

 In all cases there is a strong outburst at first. This, although 

 sometimes appearing as a bright line, is always, I believe, really 

 stratified. Immediately after this there follows a very rapid 

 shortening of the column. The extent of this shortening varies 

 with circumstances ; but when, as is often the case, it reaches far 

 down towards the positive terminal, a corresponding diminution 

 of intensity is perceptible in the negative glow. The column of 

 striae, after rising again, is often subject to similar fluctuations, 

 These, which are sometimes four or five in number, are successively 

 of less and less extent, and reach only a short distance down the 

 column of striae. The rifts due to these fluctuations then disappear, 

 and the striae either continue without interruption, or follow 

 broken at irregular intervals, until the close of the discharge. 



VIII. The effect of the proper motion, taken by itself, is to 

 shorten the column of striae. But, as we have seen, the striae are 

 in many cases renewed from time to time. In regard to this 

 point, the head of the column presents the most instructive 

 features. After the cessation of these rifts, the general ap- 

 pearance of the field is that of a series of diagonal lines com- 

 mencing at successive points which form the bounding limit of 

 the column at successive instants of time. If the points are 

 situated in a horizontal line, the striae are renewed at regular inter- 

 vals at the same place ; and the length of the column is maintained 

 by a periodic renewal of striae, a new one appearing at the head 

 of the column as soon as its predecessor has passed over one dark 

 interval. If the boundary of the illuminated field rises, the length 

 of the column increases ; if it descends, the column shortens. In 

 every case, however, the growth of the column takes place by 

 regular and successive steps, and not irregularly. The intervals of 



