and Retractions of Rods of Iron and Steel. ] 79 



be accurately centred in the helix h. This helix is 60*25 inches 

 long, and has an outside diameter of 1*75 inch. It is wrapped 

 on a tube of brass of '8 inch internal diameter, slit longitudinally 

 throughout its whole length. At mis the micrometer abutting- 

 screw, against which the end of the rod is firmly pressed by two 

 heliacal springs, which are stretched between hooks on the brass 

 mounting of the screw and a rod which passed through the ter- 

 minal brass cap of the rod. These brass caps at both ends of 

 the rod are terminated by pieces of agate. The other end of the 

 rod is in contact with a slide c, with triangular section, which 

 accurately moves, between guides, in a direction which is the 

 axis of the rod prolonged. To this slide is attached a delicate 

 fusee-chain, which is coiled once round the vertical axis of the 

 mirror d. This chain is prevented from slipping by a steel pin, 

 which securely attaches it to the axis. The slide carrying the 

 chain is firmly pressed against the terminal agate of the rod by 

 means of an heliacal brass spring. Thus the rod is at one end 

 firmly pressed against the micrometer-screw, while against the 

 other end presses the slide, which is connected with the mirror 

 by the intervention of the fusee-chain, which latter is also tightly 

 stretched. The well-joined framework e supports the springs 

 s,s, and also a divided circle,/, from whose centre depended either 

 a fibre of silk or a filament of glass, which supported a magnet 

 of very hard steel, g. From the oscillations of this magnet, or 

 from its deflections, by means of the divided circle and glass fila- 

 ment were determined the intensities of the residual magnetism 

 of the rods. The deflections of the mirror, caused by the elon- 

 gations or retractions of the rods, were determined by means of 

 the telescope and scale represented at / (fig. 3). 

 The telescope and scale were placed between 5 ^ ' 



and 6 metres from the mirror ; and the scale was 

 divided into millimetres. Each unit of division 

 given in the experiments corresponds to one 

 centimetre of the scale. 



The above apparatus was placed in a room 

 without windows and entirely underground; 

 and the room was always entered by descending from a door 

 whose bottom was on a level with the ceiling of the room. Du- 

 ring the course of the experiments the range of variation of tem- 

 perature of this room did not exceed 0°*1 C. in twenty-four hours. 



Examination of the stability and degree of precision of the 

 apparatus. 



A rod was accurately centred in the helix, and the micrometer- 

 screw and the slide of the mirror were brought into contact with 

 its ends. After the heat imparted to the rod and apparatus bj 



O 2 



