PRESIDENTIAL ADDRESS. 431) 
respective weights, and this no matter how the atoms are constructed and that 
we shall be satisfied with them so long as they disport themselves in chemical 
changes as indivisible entities. And further, I assume that we are satisfied 
with our molecules and their respective weights, as determined by the application 
of Avogadro’s hypothesis. Whether the molecular weight is obtained by direct 
determination of gaseous density or by taking advantage of the properties of 
lilute solutions, in either case the molecular weight which results is the weight 
of a supposed gaseous molecule, for the latter method depends for its justifica- 
tion on the former. All our molecular weights are weights of molecules in the 
gaseous state or are supposed to be; they are not necessarily applicable to 
liquids, and much less to solids: solids and liquids may well consist of far more 
complex particles. 
Gradually the central problem of chemistry has become more and more the 
study of the internal structure of molecules—of gaseous molecules. ‘The enormous 
number and variety of the compounds of carbon, with which so many workers 
have enriched the science during the last hundred years, and the special adapta- 
bility of these compounds to the experimental study of molecular structure, has 
led investigators to make use of them rather than of the so-called inorganic com- 
pounds: thus out of inquiries into the intramolecular structure of these com- 
pounds arose and were developed the theories of types of Gerhardt, Williamson 
and Kekulé. These are now, however, looked upon more as aspects of the 
general problem. More fruitful has been the study of the compound radicals or 
individualised groups of Wohler and Liebig. But gradually these molecular struc- 
tures have been regarded, in agreement with the views of Dumas, as complete 
wholes; like fairy temples, which from different points of view show different 
parts in relief, accentuating, it may be, this or that column or frieze or pediment. 
Kekulé’s brilliant and suggestive theory of chain compounds and ring compounds 
did more than any other theory to guide and stimulate research in chemistry in 
recent times. Like Gay-Lussac’s theory of gaseous combination, though built 
in the first place only upon a few facts, this theory has proved true of the 
thousands of others with which we have since become acquainted; there seems 
indeed to be a need of a new psychology to account for such truly marvellous 
foresight as is here exhibited. ‘I'he atoms forming these varied structures were, 
however, regarded as being arranged in a plane, until the great discoveries of 
Pasteur made it necessary for chemists to extend their conceptions and to frame 
hypotheses of three dimensions. Thus have arisen in the hands of Le Bel and 
van’t Hoff and others our modern theories of stereo-chemistry. When isomerism 
occurs in an element Berzelius names it allotropy. 1t seems to me that now, 
when molecules of the elements do not differ essentially from molecules of com- 
pounds, there is no longer any distinctive meaning in the term, and that it 
might well be abandoned. I would like also to make another suggestion respect- 
ing nomenclature : that when we distinguish ring compounds as cyclic we might 
appropriately adopt the word hormathic (from the Greek word for a chain or 
a row) for chain compounds. 
But in order to understand the linking of atoms in these molecular edifices 
some combining value had to be assigned to the different atoms. This idea of 
valency of the atoms was, no doubt, implied in Gerhardt’s theory of types; 
but it did not gain much attention until later, when Frankland and Kolbé 
formulated an empirical. theory of variable valency. Kekulé thought that atoms 
could not vary in their valency; but the alternative formule which he put 
forward to explain cases of difficulty would appear to be, rather, an attempted 
explanation of variable valency. It might be more correct to say that Kekulé’s 
formule constitute an anticipation of Werner’s theory of auxiliary valencies, 
the theory which seems to find most favour at the present day. Fixed valency 
can scarcely now be defended, in view of the existence of such compounds, for 
example, as the two fluorides and the two chlorides of phosphorus; the two 
oxides of carbon, ammonia and ammonium chloride; and, for example, the two 
series of compounds respectively of iron, mercury and copper. Variable valency 
of atoms is empirically, at least, an established fact. 
By the latest conceptions of variable atomic valency and its extension almost 
without limit—so that, for example, oxygen may be regarded as quadrivalent 
and even sexivalent—no doubt the existence of numerous compounds which 
