July 31, 1890] 



NATURE 



325 



numerous, and far from the simple rise and fall noted by 

 Rosenthal, Planer, Toaldo, and other early writers. A 

 very good illustration was afforded by my Redier baro- 

 graph on August 2-3, 1879, of which the curve is repro- 

 duced below. 



It might for a moment be supposed that the zig-zag about 

 2 a.m. was due to what is known as "pumping" in the 

 barometer, but that is not the case. "Pumping" 

 rarely takes a minute from its lowest to its highest 

 point, i.e. two minutes from one maximum to another/ 





Camden Square, London, Angus 



whereas these maxima are at intervals of about half 

 an hour. 



Several explanations of these oscillations have been 

 offered, and in my paper in the Proc. Roy. Soc. I have 

 endeavoured to give a summary of them, but I have long 

 felt that as a preliminary to a theory we ought to be sure 

 of the facts. 



Years ago, when Sir Francis Ronalds, F.R.S.,had his 

 collection of electrometers in the dome of Kew Observa- 

 tory, he devised what he called a storm clock, which was 

 really a paper going at a regular rate, so that the observer 

 needed only to record the phenomena, and the position of 

 his writing showed the time. 



In 1890 we ought to do better than half a century ago, 

 and, thanks to the great skill of MM. Richard Freres, of 

 Paris, the new machine, if it does not absolutely justify 

 its name " brontometer " (Bpom-r] fifrpoi; thunderstorm 

 measurer), is a very near approach to it. 



And first as to the object, we want to find out (i) the 

 nature of the oscillations already mentioned ; and (2) to 

 what they are due. 



The only way to do this is to get them on so large a 

 scale that they can be critically examined, and to find out 

 with what phenomena they are synchronous, or indefinite 

 relation as to sequence and time. Irrespective altogether 

 of these barometric oscillations there are several features 

 in thunderstorms not at all understood, such, for instance, 

 as whether the rush of rain which sometimes accompanies 

 an exceptionally fine flash of lightning is the cause or the 

 result of that flash. For this and other points abso- 

 lutely accurate time is of the highest importance, and 

 evidently all phenomena must be recorded on one sheet 

 of paper. 



A method adopted for some of his instruments by Mr. 

 H. C. Russell, F.R.S., of Sydney, might with advantage 

 be copied in some of the European Observatories. As 

 a general rule, and for ordinary phenomena, half an inch 

 of paper for an hour of time, i.e. 12 inches per diem, gives 

 a sufficiently open scale ; but when special phenomena 

 occur it is very handy to be able to accelerate the speed 

 five or ten times, and this Mr. Russell does with ease. 

 But even ten times times the ordinary speed, or 5 inches 

 an hour, would not enable one to read closer than to 

 quarter minutes, which would be useless for ascertaining 

 the details of a thunderstorm and the correlation of the 

 various phenomena with the peculiar oscillations already 

 mentioned. 



These, then, are some of the reasons which led me, 

 NO, 1083, VOL. 42] 



more than three years since, to consult MM. Richard as 

 to the construction of the brontometer, now at length 

 completed. 



It is provided with endless paper, 12 inches wide, tra- 

 velling under the various recording pens at the rate of 

 I '2 inch per minute, or 6 feet per hour. This is about 

 150 times faster than is usual in meteorological instru- 

 ments, and enables the time of any phenomenon to be 

 read off with certainty to a single second of time. 



The traces are made in aniline ink by a series of seven 

 Richard pens. 



The first pen is driven by the clock which feeds the 

 paper, so that the time scale and the paper must go 

 together. The pen usually produces a straight line, 

 which serves as the base line for all measurements, but 

 at 55 seconds after each minute the pen begins to go, at 

 an angle of about 45°, one-tenth of an inch to the left, 

 and at the sixtieth second it flies back to its original 

 position. 



The second pen is driven by one of Richard's anemo- 

 cinemog^raphs— a name which they have given to a pat- 

 tern of anemometer not yet known in England. The 

 external portion has some resemblance to the ordinary 

 windmill governor, but it differs from it in that the plates 

 are curved, not flat ; they are made of aluminium, and 

 are so light that they have little momentum, and have 

 thus a great advantage over cups, which run on for many 

 seconds after the wind-force has decreased or ceased. 

 The fans make one revolution for each metre of wind 

 that passes, and send an electric current to the bronto- 

 meter, where it acts on an electro-magnet, and tends to 

 draw this (second) pen towards the left ; but a train of 

 clock-work is constantly tending to draw the pen to the 

 right, the joint result being that the pen continuously 

 shows, not the total motion (as is the case with most 

 anemometers), but the actual velocity almost second by 

 second. It does this certainly with an error of less than 

 five seconds, for the fans will stop dead in less than that 

 time, and the clock-work train will bring the pen from 

 indicating a velocity of 70 miles an hour to 20 miles an 

 hour in three seconds, and down to a dead calm in seven 

 seconds. The trace will thus resemble that of a pressure 

 anemometer, but with a much more open scale than was 

 ever before available. 



The third pen is actuated by a handle, and can be set 

 at zero or at i, 2, 3, or 4 spaces from it. The author's 



' I believe that half a minute would be nearer, but until th brontometer 

 has been worked during a heavy gale no one knows. 



