2>7^ 



NA TURE 



[^Augusi 1 8, 1887 



without decomposition, recourse must be had to indirect 

 means in the determination of the molecular weight, and 

 there is consequently more or less uncertainty in such 

 determinations. As regards the third operation in the 

 construction of a constitutional formula— the arrangement 

 of the atoms so as to satisfy their valencies, and at the 

 same time to account for the reactions of the compound 

 in question — both parts of this process, but particularly 

 the latter, involve more or less uncertainty. The valency 

 of an element is frequently a variable quantity, and the 

 validity of a constitutional formula for a compound will 

 depend upon our attributing to each element the valency 

 which it really possesses in that compound. In the case 

 of organic compounds, however, this source of uncertainty 

 is reduced to a minimum. Carbon is, with one certain 

 and one or two doubtful exceptions, always a tetrad ; 

 hydrogen and oxygen are constant in their valency ; and 

 the character of nitrogen as a triad or as a pentad is 

 generally easy to determine. The chief source of uncer- 

 tainty lies in the difficulty of expressing the reactions of a 

 compound by its constitutional formula, great scope being 

 left here for arbitrary interpretation. To this is due the 

 fact, on which the opponents of constitutional formulae 

 lay so much stress, that in the case of numerous com- 

 pounds the accepted formulce have varied from time to 

 time. This could, however, scarcely be otherwise. A 

 formula constructed on the basis of an insufficient number 

 of reactions would have to be altered as soon as new re- 

 actions were discovered with which it was not in harmony. 

 And it must be admitted that, in the case of most well- 

 studied compounds, very few changes have been made in 

 the constitutional formulas since these were constructed 

 on the principles of valency. In the case even of the 

 more complex compounds, the constitutional formulae 

 show a tendency to become finally settled as soon as 

 sufficient experimental material has been accumulated. 



In this light, then, a constitutional formula is to be 

 regarded as a symbolic expression, constructed according 

 to the laws of valency, and embodying in a very condensed 

 form the reactions of a compound. By a knowledge of 

 the rules according to which such a formula is constructed 

 — by a knowledge of chemical precedent, as it were — we 

 ought, from an inspection of the formula, to be able to 

 predict the reactions of the compound ; to say before- 

 hand, for example, how many substitution-products of a 

 particular class a given compound ought to yield, and so 

 on. The value of a good working hypothesis hes in the 

 fact that it can predict : if it can predict nothing, it is 

 worth nothing. Now, with regard to the question before 

 us, we find that if we correctly embody in a constitutional 

 formula a certain number of reactions — a number sufficient 

 to warrant its construction — it will correctly predict an 

 enormous number of reactions which were not in the 

 least contemplated during its construction. Let us take 

 the example of acetic acid ; — 



Starting with methyl iodide, which admits of only one 

 constitutional formula, we convert it into methyl cyanide 

 by heating it with potassium cyanide, CH3I -\- KCN = 

 CH3(CN) -j- KI, thus substituting a monad group, CN, 

 for the monad iodine atom. The constitution of this 

 methyl cyanide is, however, not rendered clear by this 

 reaction : it might be either CH3 . CN or CH3 . NC accord- 

 ing as the cyanogen group is united to the carbon of the 

 methyl by means of carbon or by means of the nitrogen. 

 Both these compounds are in fact known. The one 

 formed in the foregoing reaction has the first of these two 

 constitutions, inasmuch as, when heated with acids or 

 alkalies, it parts with its nitrogen in the form of ammonia, 

 yielding acetic acid,CH3.CN -f aH.O = CH3.COOH 

 + N H3. In this hydrolysis, or decomposition of the com- 

 pound with assumption of the elements of water, the 

 nitrogen atom of the cyanogen group is removed, whilst 

 the carbon atom remains in combination ; and we there- 

 fore conclude that it was by means of this carbon atom, 



and not by means of the nitrogen atom, that the CN- 

 group was united to the carbon of the methyl group — a 

 conclusion confirmed by the behaviour of the isomeric 

 methyl cyanide, in which the carbon atom of the NC- 

 group can be eliminated, leaving the nitrogen attached to 

 methyl. The only question remaining to be solved is the 

 constitution of the monad group, COO H, which might be 



O O 



ii i\ 



either — C — O — H, or — C — O. From what is known 



I 

 H 



from other sources concerning the mechanism of the pro- 

 cess of hydrolysis, the first formula is the more probable : 

 that it is correct is shown by the behaviour of acetic acid 

 towards the trichloride of phosphorus, by which reagent 

 the hydroxyl group (OH) can be removed and replaced 

 by a monad chlorine atom, whilst the resulting acetyl 

 chloride (CH3.COCI) may be reconverted by the action 

 of water into acetic acid (CH3.CO.OH). Acetic acid 

 therefore contains a monad group, OH, exchangeable for 

 chlorine ; and the first formula is correct. Uniting this 

 COOH-group (carboxyl) to the methyl group, and expand- 

 ing these radicle?, we have the fully-dissected formula of 

 H O 



I II 

 acetic acid, H — C — C— O— H. 



H 

 What does this formula tell us ? What does it predict 1 

 Of the four hydrogen atoms, three are directly united to 

 carbon, and one is distinguished from the others by being 

 indirectly united to carbon by means of oxygen. We 

 know that hydrogen, when directly united to oxygen (for 

 example, in water) may be displaced by electro-positive 

 elements such as metals ; and we find that in acetic acid 

 one hydrogen atom is distinguished from the others by 

 this property. We know that hydrogen atoms in direct 

 union with carbon (as in the hydrocarbons) may be dis- 

 placed by electro-negative elements such as the halogens. 

 This we find to occur in the case of acetic acid : there are 

 three atoms of hydrogen which may be successively dis- 

 placed by chlorine and other electro-negative atoms or 

 groups. That these three chlorine atoms are attached to 

 the same atom of carbon, and that therefore the three 

 atoms of hydrogen which they have displaced are also 

 so attached, is shown by the fact that potassium tri- 

 chloracetate, when warmed with a solution of potassium 

 hydroxide, yields chloroform (CHCI3) — 



CCls-i-CO.OK. 

 H-FOK 



The same reaction with ordinary potassium acetate (a 

 higher temperature being, however, required in this case) 

 yields marsh gas (CH4) — 



CHa-^CO.OK. 

 H-tOK 



In both these reactions the molecule is divided at the 

 point of union of the carbon atoms. Apart from this- 

 disruption, each carbon atom retains the same atoms in 

 combination with it after the separation which were 

 attached to it before. In the reaction with phosphorus 

 trichloride already referred to, the molecule of acetic acid 

 is divided at the point of union of the hydroxyl group 

 (OH) with the acetyl group (CH3.CO). That such 

 separations are possible without disturbing the atomic 

 arrangement of the separated groups renders the con- 

 struction of constitutional formulae possible. But the 

 point to be noted is that all the foregoing reactions and 

 many others — in fact all the reactions of acetic acid— arc 



I 



