ON A NEW SELF-RECORDING THERMOMETER. 337 



as there are in the half wheel on the cylinder end. As the 

 cylinder turns ; its half- wheel gears into the wheel on the screw and 

 turns the screw one turn, at the same time winding up a weight 

 which will turn the screw back directly the half wheel gets out of 

 gear. Now the nut of the screw is so arranged that when the 

 screw is turning, this nut is pushed forward •£$■ of an inch = one 

 turn of the screw, and it comes back suddenly as the weight falls. 

 We have then a regular motion forward and sudden retreat every 

 minute, i.e., every turn of the cylinder, so much for the recording 

 parts. The thermometer is a piece of thin zinc tube f- inch 

 in diameter and 20 inches long, it is held firmly in a brass 

 clamp at the end most distant from the recording parts, and from 

 its free end projects a pin which is held forward by a spring, and 

 is pushed back when the motion of the screw forces the nut against 

 it ; the needle is tipped with gold, and the part of the nut which 

 comes against it is also tipped with gold, to ensure good electrical 

 contacts. A battery connected with these parts, and the electrical 

 pen, complete our requirements, and then the clock being set in 

 motion events follow in this order. The cylinder turns slowly 

 under the pen, the wheel on its end gears into the wheel on the 

 screw, which pushes the nut forward, and the instant it touches 

 the pin in the end of the zinc, the electric current brings the pen 

 down on the paper and a record of the exact spot is made ; at 

 each revolution this is repeated and results in plotting out the 

 temperature curve with surprising accuracy. I have watched the 

 working a great deal, and find that a change of -fV of a degree in 

 a delicate standard thermometer is shewn also by a change in the 

 length of this piece of zinc tube, and appears distinctly on the 

 cylinder, being represented by j^q of an inch of paper. Close 

 observation has shewn that the zinc tube is more sensitive to 

 changes of temperature than the delicate standard thermometer 

 which I keep along side of it. 



You will observe that I have mounted the zinc tube on glass, 

 because the really effective change is only the difference between the 

 expansion of zinc and glass, the expansion of the glass tends to 

 push the end of the zinc tube away from the screw, while that of 

 the zinc brings them nearer together. 



If we take the expansion of zinc as -0000173, and that of glass 

 as -0000046 we find the effective expansion is -0000127 for one 

 degree, and therefore -00000127 for ^ of one degree, this on 20 

 inches amounts to 0-000025 or tctooo- part of an inch, and if as I 

 have shewn, this extreme delicacy can be got from a rough working 

 model, I have no doubt whatever that when properly made the 

 apparatus will record changes of only T Wo o o" part of an inch. 



This is not the time or place to point out the numerous possible 

 applications of the principle involved in this instrument to the 



V- November 7, 1888. 



