A/A' WILLIAM SIEMENS, F.R.S. 85 



well-ascertained fact that the conductivity of a copper wire in- 

 creases in a simple ratio inversely with its temperature. The in- 

 strument consists of a rod or tube of met.il about 18 inches long, 

 upon which silk-covered copper wire is wound in several layers so 

 as to produce a total resistance of, say 1,000 (Siemens) units at 

 the freezing temperature of water. The wire is covered for pro- 

 tection with sheet india-rubber, inserted into a tube and hermeti- 

 cally sealed. The two ends of the coil of wire are brought, by 

 means of insulated conducting wires, into the observatory, where 

 they are connected to measuring apparatus, consisting of a battery, 

 galvanometer, and variable resistance coil. The galvanometer 

 employed has two sets of coils, traversed in opposite directions by 

 the current of the battery. One circuit is completed by the 

 insulated thermometer coil, and the other by a variable 

 resistance coil of German silver wire. Instead of the differential 

 galvanometer, a regular Wheatstone's bridge arrangement may be 

 employed. 



You will readily perceive that if the thermometer coil before 

 described were placed in snow and water, and the variable resist- 

 ance coil were stoppered so as to present 1,000 units of resistance, 

 the currents passing through both coils of the differential galvano- 

 meter would equal one another, and produce, therefore, no deflec- 

 tion of the needle. If, however, the temperature of the water 

 should rise, say 1 Fahr., its resistance would undergo an increase 

 of 1,000 x -0021 = 2-1 units of resistance, necessitating an 

 addition of 2'1 units to the variable resistance coil in order to re- 

 establish the equilibrium of the needle. 



The ratio of increase of resistance of copper wire with increase 

 of temperature may be regarded as perfectly constant within the 

 ordinary limits of temperature ; and being able to appreciate the 

 tenth part of a unit in the variable resistance coil employed, I 

 have the means of determining with great accuracy the tempera- 

 ture of the locality where the thermometer resistance coil is 

 placed. Such thermometer resistance coils I caused to be placed 

 between the layers of the cable at regular intervals, connecting all 

 of them with the same measuring apparatus in the cabin. 



After the cable had been about ten days on board (having left 

 a wet tank on the contractors' works), very marked effects of heat 

 resulted from the indications of the thermometer coils inserted into 



