180 = Dr. Harrison on Variation with Temperature of 
the changes in both variables occur at practically the same 
temperature in nickel, while for iron the agreement is a 
little less definite. A paper on the Temperature Variation 
of the Coefficient of Expansion of Nickel, forming a fourth 
part of the research, appeared in the Philosophical Magazine 
of June last. 
Throughout this research the same specimens of metal are 
employed. 
Previous Observations. 
The fact that a magnet loses its magnetism at a high 
temperature was known to Gilbert in the year 1600; but 
the phenomenon appears first to have been investigated by 
Barlow * in 1822, and again by Faraday Tf twenty years after. 
Only qualitative results were obtained, owing to the difficulties 
of temperature measurement. Subsequent experimenters 
have been many, and include in the chronological order of 
their work: Rowland{, Baur§, Berson ||, Hopkinson 4, 
Curie**, Willst{, Wildett, D. K. Morris§§, and Pitcher |||. 
The ballistic method, with ring-specimens, has been used 
most frequently for measuring the induction. Berson, 
however, used a magnetometric method, and Wilde deduced 
the induction of pure cobalt at various temperatures from the 
weight required to detach a heated cube of the metal from an 
electromagnet. 
In the earliest experiments the metal was heated to 
whiteness, and was then allowed to cool in a magnetic field, 
but since Hopkinson’s work ring-specimens have usually 
been heated ina gas furnace, and their temperature calculated 
from the resistance of the secondary winding. Morris 
heated his ring-specimens by means of a coil of platinum 
* Barlow, Phil. Trans. 1822. 
Tt Faraday, Phil. Mag. vol. viii. p. 177 (1836) and vol. xxvii. p. 1 
(1845). , 
i Pee Phil. Mag. 1874. 
§ Baur, Annalen der Physik, vol. ii. (1880). 
|| Berson, Annales de Chimie et de Physique, tom. viii. (1886). 
] Hopkinson, Proc. Roy. Soe. vol. xliv. (1888) and Phil. Trans. 1889. 
** P. Curie, Comptes Rendus, tom. cxvili. pp. 796, 859, & 1134 (1894). 
tt Wills, Phil. Mag. July 1900. 
tt Wilde, Proc. Roy. Soc. 1891. 
§§ D. K. Morris, Phil. Mag. Sept. 1897. 
||| F. H. Pitcher, Phil. Mag.’ May 1889. Pitcher’s actual investi- 
gations are concerned with the effect of circular magnetization on a 
longitudinally magnetized iron wire. After some preliminary experi- 
ments he appears to have abandoned a platinum tube method of heating 
for direct heating by current in the wire. (Vide infra, p. 181.) 
