90 Dr. Tyndall on the Progress of the Pkijsical Sciences : 



tion of the current is the opposite, it passes from the warm metal 

 to the cold. This result is exceedingly remarkable, but it has 

 been proved by multiplied experiments. 



Another memoir on this subject by M. Franz of Berlin has 

 recently appeared in PoggendorfFs Annalen. He uses cubes of 

 bismuth. The cubes are placed between two small copper pillars 

 connected with a galvanometer; the pillars are moveable, and 

 thus permit of the cubes being pressed together. We will call 

 the direction from pillar to pillar the axial direction, and that 

 perpendicular thereto, the equatorial. In some cubes the plane 

 of most eminent cleavage formed two of the opposite sides, and 

 iii some the said plane was inclined at an angle of 30° or 60° to 

 two opposite sides. When two of the former were so placed that 

 the cleavage throughout both stood either axial or equatorial, no 

 current was observed on heating. When the cleavage of one 

 cube was axial and that of the other equatorial, there was a deflec- 

 tion of 45°. When a pair of the other cubes were placed so that 

 the cleavage of each made an angle of 30° with the plane of the 

 horizon, a current of 30° was observed ; when the angle with the 

 horizon was 60°, the deflection was 19 0, 7. Bismuth was also 

 found to change its thermo-electric power in contact with other 

 metals, when the position of the plane of most eminent cleavage 

 in relation to the plane of contact of both metals was altered. 

 These results appear to stand in intimate connexion with those 

 of M. Magnus. 



Application of the results ofM. Magnus to the solution of certain 

 difficulties encountered by M. Begnault. 



An exceedingly interesting memoir, " On the Measurement of 

 Temperatures by Thermo-electric Currents," by M. Regnault, 

 appears in the Philosophical Magazine for June 1850. In the 

 course of experiment some very perplexing and indeed unexplain- 

 able phfenomena presented themselves, the solution of which 

 appears to be furnished by the experiments of M. Magnus. This 

 does not appear to have been noticed by the latter philosopher, 

 as he is silent on the subject. I have carefully plotted the seven 

 series of results given by M. llcgnault ; taking the difference of 

 temperature of iron and platinum as abscissa?, and the difference 

 between bismuth and antimony as ordinates, and using a hori- 

 zontal scale of twenty, and a vertical scale of ten divisions to an 

 inch. In the curves formed by the plotting of the last three 

 series, where every pains was taken to remove all possible causes 

 of disturbance, the anomalies are most striking. Laying the 

 datum line of one upon that of another, and commencing at a 

 common point, the curves ought to superpose ; but they do not ; 



