MEM011I8 OF THE NATIONAL ACADEMY OF SCIENCES. 397 



cbi'onosco]i(') tlie reconleil times would scarcely ever be correct, and would bccoinc iiiiicli short- 

 ened in the course of an hour's work.' 



For each degree of tension of the spring there is one strength of current witii which flu^ time 

 of magnetizaticm, and demagnetization will be equal, C X" being in tins case the same as J) Y", 

 and the recorded time A" Y" is equal to the real time C J>. This strength of current may be 

 found empirically by letting the chronoscope measure a known interval and adjusting tlie current 

 until it gives tlie correct time. The times given by the chronoscope will tiien be constant so long 

 as the current remains constant. 



In order to secure such a standard time (and for otlier chrouometric purposes) we have 

 constructed an instrument which is, to a certain extent, the inversion of the princi])le of the 

 Atwood gravity machine. This is ai falling screen, which is shown in outline in the accompanying 

 figure (fig. 3). On a heavy triangular base, B B, a perpendicular iron column i' m. in hciglit is 

 fastened. This column njay be made exactly perpendicular by means of the heavy set screws 

 S S in the base. On this iron column two brass bars, 3 cm. scjuare, are bolted .j cm. apart and 

 exactly i)arallel (having been planed in position). On the inside of these brass bars or columns 

 are triangular grooves, in which a screen, S S, runs up and down. The screen is 30 cm. long. .5 

 cm. wide, and 2 cm. thick, and weighs 2 kg. On each side of the screen two wheels are inserted, 

 which barely touch the grooves and allow the screen to run uj) and down almost without friction. 

 This screen is held at the top of the columns by the electromagnet .1/ JY, or may be held at 

 any height by means of a second adjustible magnet not shown in the figure. When the current 

 suiiplying the electro magnet is broken the screen falls, and at the rate re(piired by the law s of 

 gravity, excepting in so far as it may be retarded by resistance of the air and friction. Owing to 

 the shape and weight of the screen the resistance of the air is slight, and as the column is exactly 

 perpendicular the wheels of the screen scarcely touch the grooves, and what sliglit friction there 

 would be is nearly obviated by the revolution of the wheels. We did not, however, depend on 

 the theoretical time of fall, but measured the time with a tuning fork, as described below. The 

 force with which the screen strikes the base is broken by rubber cushions on the bottom and sides. 

 The cushions on the sides are especially useful. The screen is slightly cut away, as shown in the 

 figure, and is gradually stopped by the projecting rubber bars (ui both sides; these obviate most 

 of the Jar and prevent the screen from rebounding. The .screen is lifted by ineans of the pulley 

 shown in the cut. The cord a a is pulled by the handle h and lifts a carriage behind the screen, 

 which in turn lifts the screen. The carriage moves in separate grooves back of the screen and 

 falls into its i)lace when the screen touches the electro-magnet. 



On the front of the brass columns are also grooves in which pins are inserted which can be 

 placed in any position. By means of these pins electrical contacts C G can be securely adjusted 

 to the columns at any height. The electrical contacts were made especially for us. They consist 

 of wheels ( W W, fig. 4) which turn on points. The circumference of each wheel is half platinum 

 (/' i'), the rest being rubber or being cut away. From the wheel a pin (not visible in the figure) 

 projects, and this is struck by the screen when it falls, causing the wheel to turn. A platinum 

 contact presses gently on the wheel, and as the wheel revolves the contact is brought against the 

 ])latinum circumference, thus closing one circuit and simultaneously breaking a second circuit. A 

 current is thus closed or broken (or two circuits may be simultaneously closed or broken, or one 

 circuit closed at the instant another is broken) at a given point in the fall of the screen. We 

 especially recommend this contact for closing a circuit; a circuit maybe readily broken, but in order 

 to close a circuit a mercury contact is nearly always used, which involves various inconveniences 

 and sources of error. By means of this wheel contact a current can be closed at a given instant 

 and remains closed, or the circuit could be closed for any given fraction of a hundredth of a 

 second, which might be useful for many experiments. The rubbing keeps the contacts clean. 



' In the Hipp chronoscope tho latent time of magnetization is mncU larger than necessary. The electro-magnet 



has a very high resistance and si-lf-induction, au<l the levers carried by the armature are needlessly complicated and 



heavy. We have, as stated, greatly reduced the latent time by rewinding the magnets with coarser wire. We .are 



-rure that a chronoscope could be constructed in every way better than that by Hipp, and it is only the expense 



which has prevented us from making one. 



