273 The Rev. T. R. Robinson on the Stratification 



l'"203 and the other 2'*9!25 ; but, as before, I could get no dis- 

 charge without strata. Then I examined the sparks, which were 

 actually showing them splendidly, in the revolving mirror at the 

 highest speed which I could manage, 120 in a second. There 

 may have been a few multiple, but only as exceptional ; for in 

 almost every instance they were certainly single with that speed, 

 which would easily show -^q^qq of the second. I substituted 

 for the egg one of Mr. Gassiot's magnificent Torricellian tubes 

 (for which, as well as much precious information and personal 

 Idndness, I have to thank that gentleman). It is 1' diameter, 

 and 31''3 between the platinum points : whenever a spark passed 

 the micrometer, its peculiar strata appeared in their noi'mal 

 beauty; but Avith the brush, they had the peculiar character 

 which Mr. Gassiot has shown to belong to discharges made by 

 induction thro.ugh the glass ; and placing the tube axial on a 

 powerful electro-magnet, it showed, as in his trials, the brush 

 discharges to be double in opposite directions, or, as he calls 

 them, " reciprocating." 



These are not the only trials which I have made; but I inva- 

 riably found that no length of spark prevents the formation of 

 strata, and I am obliged to conclude that Mr. Grovels rule is 

 not absolute. The difference between our results depends 

 doubtless on some unnoticed difference in the conditions under 

 which we operated. Among possible causes may be named 

 (1) the nature of the vacuum. He used air with phosphorus 

 vapour diffused in it ; I hydrogen : I prefer that, as definite in 

 nature, as with equal air-pumps giving a rarefaction fourteen 

 times greater, and as exhibiting these strata better than any 

 medium with which I am acquainted, except mercurial vapour. 

 (3) The sort of electrodes. Believing the strata to depend on a 

 peculiar mode of disruptive discharge, I think they will be 

 formed most certainly when the electric power is concentrated 

 on a narrow area, as in the guarded points which I used. 



(3) The direction of the current. "When the upper electrode is 

 positive, they are better developed than when it is negative ; in 

 the latter case there are often only a few large ones near the 

 lower point, and the rest are lost in luminous haze, the ascend- 

 ing currents of the heated medium probably confusing them. 



(4) Still more important is the intensity of the electricity : with 

 very weak power (which would give an air-spark of 0''1 or 0'*2) 

 I have failed in obtaining them, even in the Gassiot tube. On 

 the other hand, excessive power fails also, but in a different way, 

 producing but concealing them. If, when my six coils are 

 highly charged, the discharge of one, two, &c. be passed in suc- 

 cession, it is seen that the bright strata throw out cloudy ap- 

 pendages into the dark intervals as the intensity increases, so 



