NATURAL PHILOSOPHY. 157 



those metals so effectually as to render all other preservatives unne- 

 cessary. I afterwards replaced this portion of the instrument by 

 a very large Nicol's prism constructed by M. Ruhmkorff. It is 0.086 

 of a metre in length; the greater diagonal of the base is 0.036 of a 

 metre, and the lesser 0.028 of a metre. 



The body of the pile is sheltered against variations of temperature 

 by filling the entire chest with carded cotton. In the side opposed to 

 that which contains the analyser is a rectangular glazed window, 

 through which, by the means of a good thermometer, the temperature 

 of the envelope can be read off. Finally, a little hole pierced in the 

 bottom permits of the passage of two wires from the poles of the pile 

 to the rheometer. The whole is preserved in a place less warm than 

 the surrounding atmosphere, so that during the experiments the pile 

 must necessarily be affected by any accession of heat. The sense of 

 deviation of the rheometer serves to prove that this condition is ful- 

 filled. 



The chest furnished with axes of hard wood turns in a rectangular 

 frame, which permits of the pile being retained at any angle whatever 

 with the horizon, in the vertical plane which it describes. This angle 

 of declination is estimated on an appropriate dial by means of a plum- 

 met and an index which follows the chest in its motion. The frame, in 

 its turn, moves round a vertical foot, in which it is steadied by friction. 

 The azimuths are read on a fixed horizontal dial, which permits of the 

 adjustment of the apparatus. No magnetic metal ought to be used in 

 the construction of the latter. 



I have said that the temperature of the air is subjected to almost 

 perpetual fluctuations, which cause the corresponding variations in 

 the thermo-electric current. To lessen this grave inconvenience, the 

 pile was caused to act near the window of a closed room. In the 

 reading of the rheometric deviations, it is better to determine the 

 arcs described by the index at each change of the plane of analysis, 

 than the positions at which it tends to come to rest after a number of 

 excursions, which become less rapid the more nearly astatic is the 

 system of needles. The results agree exactly with those deduced 

 from fixed deflections, in those rare cases when the atmosphere is calm 

 and permits of the operation being carried on in the open air, as also 

 within doors. 



The success of these researches depends also upon the goodness of 

 the rheometer. I have obtained an excellent multiplier from M. 

 Ruhmkorff. It is composed of two short and thick wires rolled on a 

 frame of bone. The dial is of pure copper, with its graduated cir- 

 cumference silvered. The needles, suspended from a fiber of silk 

 extremely fine, and 0.15 of a metre in length, make only a single 

 oscillation in twenty-four seconds. When the calorific radiations are 

 weak, I found the compensator of M. Melloni to be of service, more 

 especially as the object was not to obtain absolute measures, but the 

 ratios of the deviations. Operating in the manner just described, it 

 is found there are two positions of the analyser 180 degrees apart, at 

 which the deviations are equal and maximum ; and two other positions 



