April 23, 1S85J 



NATURE 



577 



give us the outlines of a self-consistent theory ; but the 

 chapters on descriptive chemistry employ the term " mole- 

 cule " in the vaguest and widest way, e.g. (p. 67) these 

 formulae are given as molecular (KO) ; , (0»Zn), (NH 9 ) S , 



f° 



and these as scmimolecular OK, - Zn, NH.. 



to 



Indeed, all through the book little or no distinction is 

 made between the formula? of gases and those of solids ; 

 all are treated as molecular. The disadvantage of doing 

 this becomes very apparent when we turn to our authors' 

 treatment of the much-vexed questions connoted by the 

 term " valency " or " atomicity." 



Here the reviewer would protest against the use 

 of the term " atomicity " as synonymous with " valency 

 of atoms." On p. 30 we are told that the molecules 

 of hydrogen, oxygen, chlorine, &c, are diatomic, and 

 the molecule of ozone is triatomic ; if, therefore, we 

 meet with the statement that oxygen is a diatomic 

 element, we should naturally interpret this to mean 

 that the molecule of oxygen is twice as heavy as 

 the atom ; but we find that it means something quite 

 different : it means, according to this book, that oxygen 

 has an atom-fixing power equal to twice that of one atom 

 of hydrogen. 



The treatment of valency, or equivalency, of atoms by 

 Profs. Frankland and Japp is, in our opinion, open to the 

 gravest objections. 



The statement on p. 57 that the atomic weight of an 

 element "is the smallest proportion by weight in which that 

 element enters into, or is expelled from, a chemical com- 

 pound" (italics are ours), we think, strikes the keynote of 

 the confusion which immediately becomes evident. If for 

 the words in italics are substituted the words, a molecule 

 of a chemical compound, and if the definition of molecule 

 as given by Clerk Maxwell or other physicists and as 

 practically adopted by our authors (pp. 26-7), is rigidly 

 adhered to, the confusion, we are convinced, would 

 vanish. 



It may be said that the word 'molecule' is understood 

 in the definition quoted, and also in the statements that 

 appear on p. 57 and elsewhere — e.g. in the mutual action 

 of zinc and steam, " one atom of zinc expels jrom the 

 steam two atoms of hydrogen " (italics are again ours) ; but 

 the frequent reiteration of the word would do something 

 to restrain the chemical student from giving the reins to 

 his fancy and plunging into dreams of graphic formula; 

 supposed to represent the structure of molecules, the 

 existence of which is unproved. 



Each element is said (p. 58) to have a certain atom- 

 fixing power, and we are told " each unit of atom-fixing 

 power will be named a bond." But when we come to 

 study the formula; which are constructed on this hasis, we 

 find that a bond is not a unit of atom-fixing power or of 

 any other " power " at all. We find that an element with 

 two bonds is simply an element one atom of which usually 

 combines with two atoms of hydrogen or chlorine, &c, 

 but the " power " cannot be measured by the number of 

 atoms fixed. It is in our opinion altogether erroneous to 

 speak of a " bond " as a unit of power, unless one is pre- 

 pared to employ the term " unit" in a sense in which no 

 known science has been bold enough to use it, and the 

 word "power" in no particular sense at all. 



The valency of many elementary atoms varies accord- 

 ing to the nature of the other atoms with which the\ .ire 

 combined in various compound molecules. The valency 

 of an atom is, as a rule, expressed by an odd or an even 

 number (there are more exceptions to this rule than the 

 authors seem willing to admit on p. 60). "These remark- 

 able facts can be explained by a very simple and obvious 

 assumption, viz. that one or more pairs of bonds belonging 

 to the atom of an element can unite and, liaving saturated 

 each other, become, as it were, latent'' 



One is obliged to ask here, Is this a scientific explana- 

 tion ? Does the explanation explain anything ? What 

 are these bonds which " become, as it were, latent " ? Are 

 not the facts much more " simple and obvious " than the 

 explanation ? What is the explanation ? 



Then we are told (p. 61) that "the apparent exception 

 to this hypothesis [one asks, What hypothesis?] nearly all 

 disappear on investigation. Thus, iron, which is a dyad 

 in ferrous compounds as (FeCl 2 ), a tetrad in iron pyrites 

 (FeS 2 ), and a hexad in ferric acid (Fe0 2 (OH) 2 ), is ap- 

 parently a triad in ferric chloride (FeCl 3 ) ; but the vapour- 

 density of ferric chloride shows that its formula must be 

 doubled — that, in fact, the two atoms of the hypothetical 

 molecule of iron (Fe 2 ) have not been completely sepa- 

 rated." Then follow structural formulae (so called) of the 

 iron compounds already named. If this is the kind of 

 explanation that the bond hypothesis has to give of facts, 

 we may well doubt whether any scientific advance is to 

 be hoped for by using this hypothesis. 



There is, it would seem, something metaphorical in the 

 statement that when the bonds have satisfied each other 

 they " become, as it were, latent " (italics ours) : and 

 " when a metaphor comes to be regarded as an argument, 

 what an irresistible argument it always seems " ! 



One is so apt in chemistry to prove a fact by a hypo- 

 thesis. We cannot but think that this method is too often 

 followed in the book before us. For instance, the fact 

 that water of crystallisation is generally easily removed 

 by heating the crystalline salt, is explained (?) by the 

 statement that " in the formation of such compounds no 

 change takes place in the active atomicity of any of the 

 molecules." 



Great advances have been lately made in the study of 

 chemical affinity. We turn with pleasure to Chapter XII., 

 hoping to have our views on this subject rendered clear 

 and definite. 



Chemical affinity " may be measured as regards its ex- 

 tent and as regards its intensity." Relative extent of 

 affinity is measured, we are told, by the number of atoms 

 of a standard element with which two or more given ele- 

 ments (? elementary atoms) can combine. " Extent of 

 affinity is thus directly connected with atomicity." " Rela- 

 tive intensity of affinity of two or more elements for any 

 given element refers to the resistance which their com- 

 pounds with this element offer to decomposition. The 

 measure of this intensity is the quantity of heat evolved 

 in combination or required for decomposition." 



" Extent of affinity " seems to be here closely connected 

 with the atom of the elements ; we are left in doubt 

 whether " intensity of affinity " is or is not similarly con- 

 nected with these atoms. 



The measure of the intensity of affinity seems to have 

 something of the nature of an atomic bond, it is so very 



