42 



SCIENCE 



[N. S. Vol. XLIV. No. 1124 



fair to give us in many ways an entirely 

 new insight into the innermost structure 

 of the atom. 



During the progress of the study of the 

 combining proportions of the elements, it 

 became more and more evident to me that 

 the atomic weights should be considered 

 not only in relation to one another, but also 

 in relation to many other essential distin- 

 guishing properties of the elements. This 

 wider problem involved a great extension of 

 the experimental field. 



Among other attributes of the various 

 forms of matter, compressibilities, surface 

 tensions, densities, dielectric constants, 

 heats of reaction and electromotive forces 

 have begun to receive attention, and already 

 many new data have been accumulated. 

 The explanation of the nature of these re- 

 searches would take us far beyond the scope 

 of this present address, but their object de- 

 serves attention. This object is the corre- 

 lation of the various properties into a con- 

 sistent whole, in the hope of tracing the 

 unknown physical influences which deter- 

 mine the nature of the elements. 



The rigorous science of thermodynamics 

 enables us to predict in logical and precise 

 fashion some of the relations between phys- 

 ical properties. My hope is not only to aid 

 in providing accurate experimental basis 

 for calculations of this kind, but also to 

 achieve the correlation of different prop- 

 erties, apparently independent of one an- 

 other from a thermodynamic point of view, 

 thus, perhaps, enabling one by inductive 

 reasoning to penetrate further into the 

 causes which lie back of all the attributes 

 of matter. 



In attempting to follow this inductive 

 path comparisons of the properties of the 

 elements have been made in two different 

 ways. 



On the one hand a given property of ©ne 

 element has been compared with the same 



property of another. For example, the 

 question, "Which of the two elements, 

 cobalt or nickel, has the heavier atom?" 

 was answered by parallel determinations, 

 using the same methods, conducted side by 

 side in the laboratory. Cobalt was found to 

 possess the higher atomic weight. 



On the other hand, the attempt has been 

 made to discover a relation between the 

 different, apparently quite distinct, prop- 

 erties of a single element. For example, 

 one may ask: "Have the low melting and 

 boiling points of phosphorus any connec- 

 tion with its small density and its large 

 compressibility ? ' ' Here one compares vari- 

 ous properties of the same element, and 

 one seeks to discover if all are based upon 

 some common, ultimate characteristic of 

 phosphorus, of which the properties are 

 merely symptoms. 



The inductive methods used in compari- 

 sons of this sort can not be explained here 

 to-day. They are partly statistical, partly 

 mathematical and partly graphical. From 

 the nature of the problem, which involves 

 many unknown variables, perfect mathe- 

 matical exactness is not to be expected. 

 Nevertheless, little by little, one may hope 

 to trace the conflicting tendencies, and as- 

 cribe them to a few common causes. 



"With the help of these methods the ten- 

 tative conclusion has been reached that the 

 space occupied by the atom and molecule in 

 solids and liquids is highly significant. The 

 actual atomic bulk or volume is diminished 

 but slightly by moderate mechanical pres- 

 sures, and by cooling even to the absolute 

 zero; but it is very greatly affected, appar- 

 ently, by the mutual attractions of the 

 atoms, called cohesion and chemical affinity. 

 Usually the less volatile a substance (that 

 is to say, the more firmly it is held together 

 by cohesion) the greater is its density and 

 the less is its compressibility, other things 

 being equal. Greater cohesion is associated 



