'264 Dr. Thomas Thomsons Observations 



The first of these products is only half of 0-375. The 

 second is little more than half that number, doubtless, be- 

 cause our estimate of the specific heat is too high. The 

 true atomic weights of these bodies are double the numbers 

 of Berzelius or 4-5 for chlorine and 10 for bromine. Dr. 

 Turner's number for chlorine is double that of Berzelius's. 

 His atom of bromine is 9-8. Both of these are pretty near 

 the truth, but not exactly so. 



We have now taken a view of all the substances in the 

 first table except the four last ; namely, cobalt, phosphorus, 

 silver and iodine. 



The specific heat of cobalt was determined by Dulong 

 and Petit, and they give us no information either respect- 

 ing the purity of the metal, the mode of operating, or the 

 quantity of metal used in their experiments. The same 

 observations apply to the specific heat of nickel, which was 

 determined by the same gentlemen, and given by them 

 merely in a table without any details whatever. Now, as 

 the atomic weights of nickel and cobalt very nearly agree, 

 while their specific heats, according to Dulong and Petit 

 are as follow, 



Nickel, . . 0-1035 

 Cobalt, . . 0-1498 



there is reason to suspect some inaccuracy in one or 

 other of these determinations, or in both of them. Under 

 these circumstances it is not surprizing that the number 

 resulting from the multiplication of the atomic weight and 

 specific heat of cobalt instead of 0*375, is 0*4875 about -|d 

 greater than the truth. Berzelius's atom of cobalt is 

 3*68991 ; which multiplied by 015 gives 0*553, which de- 

 viates still farther from 0-375 than the quantity resulting 

 from my atomic number does. 



To clear up the anomaly regarding cobalt, new experi- 

 ments on its specific heat must be made. Were the specific 

 heat 0*1 15 instead of 0*1498, cobalt would coincide with the 

 other simple substances. 



Phosphorus and silver, when their atoms and specific 

 heats are multiplied together give 0*770 or 0*375 x 2. To 

 bring them to agree with the other simple substances we 

 must reduce the atom of each to one-half; we then have 



