CHEMICAL INVESTIGATIONS — RICHARDS. 217 



helium, cand why should it have an intense affinity for sodium and no 

 affinity whatever for argon or fluorine? No man can answer these 

 questions; he can discover the facts, but can not yet account for 

 them. The reasons are as obscure and elusive as the mechanism of 

 gravitation. But we shall not really understand the material basis 

 upon which our life is built until we have found answers to questions 

 of this sort. 



In order to correlate the properties of the elements, and to attain 

 any comprehension of their significance, one must first exactly ascer- 

 tain the facts. Therefore, my endeavor has been to institute sys- 

 tematic series of experiments to fill the gaps in our knowledge of the 

 actual phenomena. In much of this work I have had the invaluable 

 aid of efficient collaborators, for which I am grateful. 



The atomic weights were the first of the fundamental properties of 

 the elements to receive attention in carrying out this plan. These, 

 as everyone who has studied elementary chemistry knows, represent 

 the relative weights in which substances combine with one another. 

 They are called atomic weights rather than merely combining pro- 

 portions, because they can be explained satisfactorily only by the 

 assumption of definite particles which remain indivisible during 

 chemical change. Even if some of these particles or so-called 

 " atoms " suffer disintegration in the mysterious processes of radio- 

 active transformation, the atomic theory remains the best interpre- 

 tation of the weight-relations of all ordinary chemical reaction. In- 

 deed, it is entrenched to-day as never before in man's history. 



The determination of atomic weights is primarily a question of 

 analytical chemistry — a question of weighing the amount of one 

 substance combined with another in a definite compound — ^but its 

 successful prosecution involves a much wider field. First, the sub- 

 stances must be prepared and weighed in the pure state, and, next, 

 they must be subjected to suitable reactions and again weighed with 

 proof that in the process nothing has been lost and nothing acci- 

 dentally garnered into the material to be placed on the scale pan. 

 These requirements involve many of the principles of the new 

 physical chemistry, so that the accurate determination of atomic 

 weights really belongs as much in that field as in the field of ana- 

 lytical chemistry. 



At Harvard during the last thirty years the values of the atomic 

 weights of thirty of the most frequently occurring among the eighty 

 or more chemical elements have been redetermined. From data 

 secured here and elsewhere is compiled an international table of 

 atomic weights, revised from year to year by an authoritative com- 

 mittee composed of representatives of various nations. The values 

 thus recorded are in daily use in every chemical laboratory through- 

 out the world, serving as the basis for the computation of count- 



