THE LAST TEN YEAES. 359 



volume. Meanwhile he regarded the mechanical theory of 

 heat as 'very clever, and developed with great ingenuity/ 

 He sought information from Magnus upon the subject, as 

 appears from letters found among his papers. In writing to 

 Dove on November 6, 1857, he complains: 'I am still tor- 

 mented, though it is entirely my own fault, concerning this 

 mechanical theory of heat as upheld by Joule, Grove, Eankine, 

 &c. My physical and intellectual powers are on the wane, 

 therefore I am all the more industrious and anxious that I 

 may present you as soon as possible with the first part of the 

 last volume of " Cosmos " that work upon which I so inconsi- 

 derately entered. Pray avoid living to so unusual an age ! ' 

 This tormenting uncertainty is unmistakably evident in the 

 introductory pages to the last volume, published, after Hum- 

 boldt's death, by Buschmann, in 1862 ; as a general principle, 

 he no longer doubts in the possibility of the reciprocal con- 

 version of heat into force, but in individual instances too much 

 seemed to be built upon * somewhat capricious assumptions ; ' 

 the atomic theory still appeared to him a convenient form 

 of expression, and one universally current though full of c myths.' 

 The new form assumed by 'metaphysical science' he was 

 evidently incapable of receiving, and his inquiry is almost 

 touching as to ' whether " Cosmos " had remained true to the 

 plan originally devised he might almost say had kept within 

 the limits that had appeared to him advisable from his own 

 point of view, and from his acquaintance with science as it then 

 was.' Science had in the mean time undergone many important 

 changes, the epoch of the intellectual development of man- 

 kind, as evinced in the aspect of science in 1834, of which 

 ' Cosmos ' was the record, had already passed away. If to this 

 be added that only a year after c Cosmos ' had been brought to 

 a close by the death of the author, the discoveries of spectrum 

 analysis had annihilated the strong separation insisted on by 

 Humboldt between cosmical matter regarded only in the mass, 

 and telluric matter to be viewed in all the varieties of its 

 component materials, it is impossible not to be vividly reminded 

 of the passage in the introduction to the first volume, where 

 the possibility is suggested that ' Cosmos ' may become anti- 

 quated. 



