Norton—Breaking of Watch Mainsprings. 655 
sium cyanide solution; otherwise the springs would break as 
soon as they were used, if they did not do so while still in the 
solution. 
Some twenty years ago I had two clock movements, each of 
which had two springs. The springs were under stress and 
had been for several weeks. I dipped the movements, with 
the springs wound up, into a weak solution of potassium cyanide 
for a few minutes; upon removing them from the solution, one 
spring broke while in my hand, another in a few minutes, the 
third within fifteen minutes, and the fourth the next day. The 
springs were of good quality, polished and blued. The bluing 
was not effaced. 
We know that leather and hoofs of animals, as also prussiate 
of potassium, bound around iron which is then subjected to a 
given heat and while hot plunged into water, will case-harden 
the iron and convert its surface into a kind of steel. We also 
know that upon filing the surface of a block of tool steel and 
then rubbing the surface with the hand, the surface becomes 
hardened so that upon filing the steel again the file does not 
‘‘bite” readily. 
The facts above mentioned incline me to think that during 
excessively warm and sultry weather there emanate from the hu¬ 
man body certain substances which have an effect upon the al¬ 
ready hardened steel similar to that cited above—possibly some 
subtle emanation not yet discovered—and of a kind that is less 
active in the cold months. 
Assuming this cause to be eliminated, the springs, in the 
nature of things, would break, but not in greater number in any 
one month than in another. 
The two decades mentioned would seem to be a fairly good 
basis upon to reason. 
The diagram accompanying this paper shows the number 
of springs broken during each month for the ten years from 
June 10, 1893, to June 10, 1903. 
