second and let the vibration be, say, 60,000 per second - then there would be 30 individual discharges in each break wheel period, understanding by this the time interval from one break to the next, or from one closure to the next. Since the break wheel effect is always seen, even when the streamer moves slowly, it would seem that the finer threads indicate the individual discharges. At any rate, with a stronger draught the latter can be evidently easily recorded and this might be a simple way of exactly determining the vibration of the system, certainly simpler than analysis by a revolving mirror.
This experiment I expect to perform on my return for it is indispensable to determine the vibration quite exactly. Up to the present this necessity was not imperative, the method used being satisfactory so long as the chief purpose was to perfect the apparatus and make general observations. But now quantitative estimates have become important. To return to the description, the strong sparks to the hood are particularly curious. The strong current of air is evident from their behaviour and also from the appearance of the upper streamers. Some very strong sparks pass to the coil in series with the extra coil. One of the streamers striking the floor ignites the wood. Many streamers are carried through the opening in the roof. One, exceptionally long, passes to the photographer in the corner of the building. The shock is but slight as might be expected. A spark might be fatal, but there was no possibility of a spark taking that course without being stopped by conducting objects nearer to the origin. Some of the upper streamers are chopped up curiously indicating the presence of small whirls or eddies in the air current passing through the roof opening. The switch was here also closed 100 times; the closures were very short and the other particulars remained as before.
