July 3, 1919] 



NATURE 



347 



owe to him in great measure the enormous im- 

 provement in technique which distinguishes 

 modern determinative work of this kind. Such 

 work will not pass muster to-day unless it is per- 

 formed with the scrupulous regard to detail and 

 conscientious search for causes of error and for 

 means to avoid them which characterise the 

 determinations he has directed. 



America, moreover, is to be congratulated in 

 possessing a publishing agency hke the Carnegie 

 Institution of Washington, which undertakes the 

 printing and distribution of important scientific 

 memoirs which might seriously tax the means of 

 most scientific societies, and which, on account 

 of their specialised character, no ordinary pub- 

 lisher would be likely to accept as a business 

 profKJsition. 



A recent publication by the Carnegie Institution 

 is concerned with the results of a determination 

 of the atomic weights of boron and fluorine by 

 Messrs. Edgar F. Smith and Walter K. van 

 Haagen.' As it presents some features of general 

 interest, an account of the w^ork may not be 

 unacceptable. 



The redetermination of the atomic weight of 

 boron has revealed the unexpected fact that the 

 value for this constant hitherto accepted is at 

 least I per cent, too high — a remarkable circum- 

 stance, all things considered. Boron, of course, 

 is a common and widely distributed element, and 

 the estimation of its atomic weight has been 

 made by at least half-a-dozen experimenters since 

 the time of Berzelius with such concordant 

 results that it might be assumed that it was fairly 

 well known. But there are certain considerations 

 connected with these determinations which might 

 occasion doubt. To begin with, there is no great 

 choice of methods in this particular case of a 

 sufficiently valid character upon which to base 

 determinations. Practically all the numbers 

 depend upon the analysis of borax, either hydrated 

 or anhydrous. We have here an instance of what 

 has been frequently deprecated in atomic weight 

 work. A determination based upon the amount 

 of water in a hydrated salt rests upon a faulty 

 principle. It presupposes that the amount of 

 water in a hydrated salt is absolutely definite and 

 constant, and that adventitious water can be 

 separated from that which is supposed to be 

 normal to the constitution of the salt, of which 

 there is no absolute proof. It further assumes 

 that the salt can be completely dehydrated under 

 the particelar conditions of the experiment, which 

 may or may not be the case. Now, as all the 

 previous determinations of the atomic weight of 

 boron rest upon practically the same basis, they 

 may involve the same fortuitous errors, and 

 Messrs. Smith and van Haagen's investigation 

 shows that, as a matter of fact, they do. The 

 substantial uniformity of the previous results is 

 therefore misleading. It is a recognised canon in 

 atomic weight work that a value can be accepted 



1 "The Atomic Weights of Boron aiul Fluorine." By Edgar F. Smith 

 and Walter K. van Haagen. (Washington : The Carnegie Institution of 

 Washington, 191 8.) 



NO. 2592, VOL. 103] 



with confidence only if it is based upon 

 methods involving different principles and modes 

 of manipulation free from known sources of error. 

 In these analyses of borax the manipulative pro- 

 cesses were of the simplest possible character, 

 and of themselves not liable to introduce error if 

 properly conducted. The main error is traceable 

 to the water and to an imperfect knowledge of 

 the conditions under which the borax could be 

 completely dehydrated. 



The persistent retention of water by substances, 

 even when exposed to high temperatures, is, of 

 course, no new fact, and many instances might 

 be given of it. No rational explanation of the 

 phenomenon is known. In the case of borax 

 Messrs. Smith and van Haagen offer an explana- 

 tion which has at least the merit of ingenuity, if 

 not of generality. In effect it is as follows : When 

 the hydrated salt is heated the water of crystallisa- 

 tion is evolved, and at first passes through the 

 Hquid state before escaping as steam, forming 

 droplets of an aqueous solution of borax, which 

 is then hydrolysed as follows : — 



Na,B40- + H.,Oi::r2NaB02 + 2HB02 

 Na2B4O- + 3H20c:=2NaOH+4HB0j. 

 This process is known to occur in weak aqueous 

 solutions of borax. The sodium metaborate and 

 hydroxide on concentration slowly recombine with 

 the boric acid, reforming borax. It may be that 

 on heating the borax the expulsion of water takes 

 place more rapidly than the recombination of base 

 and acid, and therefore heated borax may contain 

 more or less sodium metaborate or hydroxide and 

 free boric acid, and that the recombination is only 

 complete after prolonged fusion. 



"According to this view," say the authors, 

 " the last traces of w^ater expelled from fused 

 borax are not merely the last portions of the water 

 of crystallisation proper, but are to be looked 

 upon as water of neutralisation, resulting from 

 the recombination of sodium metaborate (or 

 hydroxide) with boric acid, both of which were 

 produced by a transient hydrolysis during the 

 earlier stages in the dehydration ; and this view 

 explains why the last traces of water should be 

 removed with greater difficulty than the bulk. 

 Hence the final loss of water in the dehydration 

 of borax may in all probabilitv be due to the 

 completion of such reactions as the following : — 

 2NaBO, + 2HBO, = Na.,B407+H,0 

 2NaOH'+4HB02'=Na2B407 + 3H20." 

 In support of this hypothesis the authors point 

 to other instances in which salts which are ex- 

 tensively hydrolysed in solution retain the last 

 traces of water with great tenacity. There are, 

 however, cases to which this reasoning scarcely 

 applies. Indeed, even in the particular instance 

 of borax the authors point out that it is not neces- 

 sary to assume this hydrolytic action. Borax in a 

 state of fusion may dissociate into sodium meta- 

 borate and boric anhydride : — 



Na2B40;:;=r2NaB02 + B2O3. 

 This dissociation may begin before the water is 

 completely expelled, and the hygroscopic boric 



