PARAMAGNETIC & DIAMAGNETIC SUBSTANCES 295 



The susceptibilities of solutions of potassium ferrocyanide and 

 potassium ferri cyanide were the same as if water alone were 

 present so that when iron is present in the acid radicle it must be 

 at least 100 times less magnetic than where it takes the part of a 

 metal in a salt. 



The paper contains an interesting suggestion. Imagine a magnet 

 set spinning about its centre in the earth's field. The N pole will 

 go round the semicircle on the Northern side more rapidly than it 

 will go round that on the Southern side. Thus the end towards the 

 North will be a longer time S than N, and the average effect at a 

 distance will be that of a reversed magnet. If the molecules of a 

 body consisted of small magnets rotating about their centres under 

 the action of a field, the body would appear diamagnetic. 



We may note here that iron carbonyl, the liquid compound 

 Fe(CO) 5 discovered byMond and Quincke in 1891,* is diamagnetic. 



Experiments of Fleming and Dewar on oxygen at 

 low temperatures. Dewar had already found that liquid 



jen was strongly paramagnetic.f Fleming and Dewar made 

 investigations on its susceptibility. The first method J which they 

 used was an induction method in which the primary and secondary 

 (oils were immersed in liquid oxygen and then in gaseous oxygen 

 at nearly the same temperature. But this was not so satisfactory 

 >iid method j in which balls of silver, of bismuth, of copper, 

 and of gl.-i^s containing mercury, were weighed above the pole of 

 a strong electro-magnet. When the magnet was excited there 



was a change in weight of (/c 1 -Ac a )VH-^-, where V is the volume 



of the ball, /^ is its susceptibility, and /c, is that of the surrounding 

 medium and x is measured upwards. The field H was determined 

 for different values of x and for different exciting currents by means 

 of an exploring coil connected to a ballistic galvanometer. 



The susceptibility of each ball was first found by weighing it in 

 air with the current off and on, and assuming that 10 6 /c for air is 

 0'024. To test the method the silver ball and a glass ball were 

 each weighed also in water with the current off and on, and the 

 mean value of 10V for water was 0'74, agreeing fairly with the 

 values obtained by others. The balls were then weighed in liquid 

 oxygen at 182 C. with the current off and on. The susceptibilities 

 of the balls were estimated at 182 C.,and thus the susceptibility 

 of liquid oxygen was found. The mean result was 10 6 /c = + 324, 

 whence /x = 1 *004. 



Curie found that for gaseous oxygen 



1ft , ,v* 33700 

 0x= -^-, 



* /r.ins. Chem >V. , l-'.M, vol. 50. p. 604. 



t /'.//..V. v..I. I. p. 247. { 76trf., vol. lx. p. 283 (1896). 



. v.,1. Ixiii. p, 311 (1 *'.*), 



