174 GREAT MEN OF SCIENCE 



evacuated space, taking precautions which excluded all 

 possibility of conduction of heat.^ 



Rumford himself carried out further very refined experi- 

 ments in order to determine a possible weight of heat-stujflF. 

 He had already found earlier (1785), that a heated body did 

 not sensibly weigh more than a cold one, and also not less - 

 as some people believed themselves to have proved - if only 

 care were taken to avoid disturbance by currents of air, which 

 a hot body when not enclosed always produces around itself. 

 'For a long time I feared to form a decisive opinion on the 

 matter,' he says, 'on account of the great difficulties of carry- 

 ing out weighings of this kind,' until (in 1799) he had inven- 

 ted a special device for eliminating all temperature differences 

 upon the balance, and other sources of error. The result, 

 again completely negative, he then communicated to the 

 Royal Society in London. ^ 



We may here remark by the way, as illustrating the in- 

 exhaustible surprises which nature prepares for us, that 

 since the fairly recent experiments of Hasenohrl (1904), there 

 can no longer be any doubt that heat (like every form of 

 energy) does possess weight, though its extraordinarily small 

 amount makes it quite immeasurable by present methods of 

 weighing. 



Rumford, whose original name was Benjamin Thompson, 

 was born in North America, of a family of English origin; he 

 was very poor in his youth, could not gain much schooling, 

 and was entered as an apprentice in a business house at the 

 age of thirteen. But he soon set up as a teacher with the 

 knowledge which he had acquired from time to time, and 

 then married a rich widow. When the American War of 



^ These experiments were carried out a year after Rumford 's announce- 

 ment; they are to be found in Davy's first publication. See the Works of 

 Sir Humphrey Davy, vol. 2, pages 1 1 flf. 



2 These communications are to be found in the Philosophical Trans- 

 actions, 1799, page 180. According to the figures he gives there, he was 

 able to show that a kilogram-calorie certainly weighs less than -0.013 

 milligrams. 



