Chemistry and Physios. 1 11 



tors, and in recent times it lias been generally admitted that 

 Prout's hypothesis does not hold true, but it is recognized that 

 there are a number of atomic weights that are curiously near to 

 whole numbers on the basis of hydrogen as unity, and especially 

 on the basis of oxygen as exactly 16. Recently William I). 

 Harkins of the University of Chicago has brought forward 

 again this frequently revived' idea and holds the view that the 

 elements are very probably inter-atomic compounds of hydro- 

 gen. He believes that hydrogen first forms helium (He = 4) 

 and that this becomes a secondary unit of fundamental impor- 

 tance in the formation of the higher atomic weights. He finds 

 that the elements taken in their serial order have atomic weights 

 that are multiples of helium in the even numbers of the series, 

 while they are multiples of helium plus three hydrogens in the 

 odd numbers. Thus, beginning with helium, the second ele- 

 ment: He = 4, Li = 4 + 3, Be (exception), B = 8 + 3, C = 

 12, N (exception), = 16, F = 16 + 3, Ne — 20, Na = 20 + 3, 

 Mg = 24, Al = 24 + 3. Si = 28, P = 28 + 3, S = 32, CI = 

 32 + 3, A = 40, K = 36 4- 3, Ca = 40, Sc (exception), Ti = 48, 

 V = 48 + 3, Cr = 52, Mn = 52 + 3, Fe = 56, Co = 56 + 3. 

 The three exceptions noticed are that Be is accepted as 9-1 

 instead of 8, N as 1401 instead of 15, while Sc is given in our 

 tables as 44 instead of 43. Although other variations from the 

 exact whole numbers occur, Li = 6-94, Mg = 24-32, Al = 27-1, 

 Si = 28-3, P = 31-02, S = 32-07, CI = 35-46, K = 39-10, Cu = 

 4007, Ti = 481, Mn = 5493, Fe = 55-84, and Co = 58-97, there 

 is a striking nearness to whole numbers and to the proposed rule 

 in many cases. However, it must be emphasized that these dif- 

 ferences, especially in the cases of chlorine, magnesium, silicon, 

 iron and sulphur, where the atomic weights have been deter- 

 mined with great accuracy, and have thus overthrown Prout's 

 hypothesis, without considering the other exceptions and without 

 considering the fact that the rule gives many more exceptions 

 and wider variations when it is applied to the higher atomic 

 weights, are sufficient to overthrow the practically identical 

 hypothesis under consideration. 



In his last article on this subject the author has shown that 

 the elements of even numbers in the series, those that he supposes 

 to have atoms made up of multiples of helium, appear to be 

 much more abundant than the "nHe -f- 3H" elements. There 

 is evidence of this in a general way from a consideration of the 

 composition of the earth's crust, but it is much better shown 

 from the average composition of meteorites, and the author 

 believes that the latter probably represents matter in general 

 more accurately than the segregated material on the surface of 

 the earth. This generalization appears to have too many excep- 

 tions to be worthy of very serious consideration. For example, 

 the very rare element beryllium between lithium and boron, and 

 the very rare element neon between fluorine and sodium should 



