182 



Colonel H. Watkin 



[May 1, 



shows is very reliable. In this I depend on a very constant quantity, 

 viz. gravity. A weight is dropped from a given height, and in its fall 

 breaks two screens one after the other. Knowing the distance the 

 weight has to fall to the first screen, and the distance between the 

 screens, it is easy to calculate the time it has taken for the weight to 

 pass from the one screen to the other. 



The screen is made thus — see Fig. 3. A piece of hardened watch 

 spring A B, is pivoted in a brass frame B C, and capable of being 

 held up and pressed very lightly against the support D, so that two 

 pieces of platinum, one on the spring and one on the support, are 

 kept in contact. The fall of the weight E breaks the contact. An 

 exactly similar arrangement, A' B' C D' is placed about 3*77 

 inches below the top spring. Each screen is connected to the 

 primary wire of an induction coil, the secondary being led to the 

 recording points opposite the drum of the chronograph. It follows, 

 then, that the moment the weight touches the 

 first screen, a spark passes on the drum from 

 the steel points. The drum goes on re- 

 volving, and the weight continues to fall until 

 the second screen is reached, when again a 

 spark passes. The distance between the two 

 spots measured on the graduated circle, and the 

 known time taken by the weight to pass the 

 screens, gives the speed of the drum. The time 

 taken for the weight to fall below the screens 

 was '01894:8 second. As the result of a trial 

 before a committee, in which the record of two 

 weights was made on a rapidly revolving cy- 

 linder, the variation did not exceed 0*16 per 

 cent. To test whether the weights were appre- 

 ciably retarded by breaking the screens, a third 

 screen was inserted between the weight and the 

 first screen, and it was found that there was no 

 appreciable retardation. As a precaution I al- 

 ways employ two drop weights with entirely 

 independent circuits, so as to avoid the chance 

 of an experiment being wasted, from the pos- 

 sible failure of one ' of the screens not acting 

 through a bad contact; but I nearly always 

 obtain the double record. 

 The next difficulty I encountered in my experiments was the 

 means of reading the record of the sparks. Some days we might get 

 nice small records by carefully adjusting the strength of the current. 

 Another day the records would be much too large for any accuracy. 

 I tried every conceivable method of smoking, from the carbon de- 

 posited by gas flame, to that deposited by various kinds of oils, and 

 also that of burning camphor, but could not be certain of my records 

 being readable. I may here mention that for accurate experiments. 



Fig. 3. 



