April 20, 1900.] 



SCIENCE. 



607 



have the same chemical, and in general the 

 same physical properties ; their action on 

 polarized light, however, would be the same, 

 but in opposite directions, and their crys- 

 talline form, if unsymmetrical, would be so 

 in opposite senses. In short, the isomer- 

 ism would not be chemical, but physical or 

 geometrical, like the pairs in question. It 

 was further observed that in every case of 

 this kind the molecule contains an ' asym- 

 metric carbon atom,' a carbon atom united 

 with four atoms or groups each differing 

 from the others : 



a 



b—C—d 



If we regard the groups a, h, c, d as inter- 

 changeable in position, or as rotating inde- 

 pendently about the central carbon atom, 

 we cannot explain the apparent right- aad 

 left-handedness of the molecule; there could 

 be no fixed difference between the two com- 

 pounds. If, however, as van't Hofif and 

 Le Bel pointed out, we suppose the four 

 valences to extend in the directions of the 

 apices of a tetrahedron and to be fixed in 

 these directions ; then when the combined 

 groups are all different, we obtain two forms 

 of molecule which are identical in every re- 

 spect except that .the one is like the re- 

 flected image of the other. 



The figures represent two tetrahedra the 

 centers of which are supposed to be occu- 

 pied by a carbon atom, the four groups a, 

 6, c, d being located in the direction of the 

 apices. It is easily seen that the one cor- 

 responds to the reflection of the other in a 



mirror. It is to be distinctly understood 

 that it is intended to represent only the di- 

 rection of the valences, not the shape of the 

 carbon atom. As, in general, these geo- 

 metrical isomers are not readily transformed 

 into each other, it follows that the com- 

 bined groups or atoms have a strong ten- 

 dency to retain their relative positions ; in 

 short, that the direction of the valences is 

 practically fixed. The same holds true in 

 the case of pairs of compounds in which 

 there are two doubly united carbon atoms 

 as 



1>G = G<1 and >C=C<^^ 



a simple rotation of one half the molecule 

 about its axis, or an interchange of position 

 on the part of a and 6 or c and d in one of 

 these forms would convert it into the other, 

 yet in reality this does not occur, and in 

 general, the two forms represent distinct 

 compounds. The theory has been applied 

 with great success in predicting new com- 

 pounds and in explaining the nature of sub- 

 stances containing several asymmetric car- 

 bon atoms in the same molecule, notably the 

 sugar group. It does not imply that the 

 molecule is rigid, but merely that there are 

 certain fixed directions of attraction, about 

 which, within limits, the combined atoms 

 may vibrate. The concordance of a very 

 great number of facts with this hypothesis, 

 and the absence of any noteworthy ex- 

 ceptions, lend to it a high degree of plausi- 

 bility. Whether the fixed direction of the 

 valences is true of other atoms than those 

 of carbon, is as yet uncertain. Carbon com- 

 pounds lend themselves with especial ease 

 to such studies, and the valency of carbon 

 is practically fixed at four. The effect of 

 varying valency, as in the case of iron, is 

 unknown, but in the case of nitrogen, and 

 to some extent in that of certain metals, 

 some evidence has been accumulated, tend- 

 ing to show that the rule is a general one. 

 It will be seen that the theory of the 



