500 



SCIENCE. 



[N. S. Vol. XX. No. 511. 



atoms are ionized when they separate from 

 each other and that the same atom may be- 

 come either positive or negative. 



In following farther the thought of the 

 attraction between electrons as the cause 

 of chemical combination, we must suppose 

 that in addition to the effect of this attrac- 

 tion in holding together the atoms which 

 are immediately attached there is a residual 

 effect upon other atoms within the molecule. 

 This gives a rational explanation of the 

 very great difference in the stability of 

 the union between carbon atoms in differ- 

 ent compounds as, for instance, the insta- 

 bility of acetaeetic acid in comparison with 

 butyric acid, occasioned by the substitu- 

 tion of an oxygen atom for two hydrogen 

 atoms of the latter. The study of organic 

 compounds has given us a knowledge of a 

 large number of cases of this sort and our 

 text-books contain many empirical rules 

 about them, but there have been few, if 

 any, attempts to give for such facts any 

 rational explanation. 



In considering double unions three ex- 

 planations suggest themselves : (1) We may 

 suppose with Pfeiffer that such unions are 

 in reality single unions and free valences. 

 In this case the presence in adjacent car- 

 bon atoms of positive and negative elec- 

 trons which are uncombined would reduce 

 the attraction of each for the electrons of 

 another molecule, thus explaining why two 

 free valences are so much less active than 

 a single one. (2) We may suppose that 

 the carbon atoms are in reality doubly 

 united, but that, owing to the localization of 

 the electrons in definite parts of the carbon 

 atoms, the four electrons involved can not 

 approach as near to each other as is the 

 case in a single union. This is Baeyer's 

 theory of strain and is much better in ac- 

 cord than is the theory of free valences 

 with the fact that cyclopropane and propy- 

 lene appear to be about equally unsaturated 

 as evidenced by their heats of combustion 



and by their conduct toward bromine. On 

 the other hand, it seems to lead logically 

 to conclusions with regard to the addition 

 of bromine to triple unions which Pro- 

 fessor Michael has shown are contrary to 

 the facts. (3) Without a condition of 

 strain, we may suppose that the presence of 

 both a positive and a negative electron in 

 each of the atoms united by the double 

 union causes a lessening of the attraction 

 of the electrons. This would result in such 

 a union being less stable than a single 

 union. The second and third views appear, 

 at present, most in accord with the facts — 

 possibly the truth lies in some combination 

 of the two. 



Wliatever view we may take, it is note- 

 worthy that double unions are usually 

 formed by the loss of a positive and nega- 

 tive atom or group from adjacent carbon 

 atoms, as hydrogen and hydroxyl or hydro- 

 gen and bromine. It is also true that in 

 many double unions one of the carbon 

 atoms is more positive than the other, caus- 

 ing the addition of halogen acids in a 

 definite manner which may be predicted in 

 accordance with Michael's 'positive nega- 

 tive law. ' Applying this thought to conju- 

 gated double unions, we see that of the four 

 atoms involved the two central ones are 

 likely to be positive and negative respect- 

 ively and neutralize each other's attraction 

 for outside atoms, while an intensified at- 

 traction for outside atoms would be found 

 in the exterior atoms. The effect may be 

 analogous to that of the attractive forces of 

 a magnet which exhibit themselves chiefly 

 at the ends. 



But I have permitted myself to wander 

 much farther in the field of speculation 

 than was my first intention — farther than 

 is at all profitable, I fear, for these ques- 

 tions furnish, at present, few points for 

 experimental study, and speculations di- 

 vorced from experiment have usually been 

 profitless. I should be very sorry if what 



