1914] on X-Rays and Crystalline Structure 208 



photograph and on the spectrometer. The reflections from the 

 various series of planes of the latter crystal show spacings consonant 

 with an arrangement in the simplest cubical array, of which tlie 

 smallest element is a cube at each corner of which is placed the same 

 group, a single atom or molecule, or group of atoms or molecules. 

 In the case of rocksalt, the indications are that the crystal possesses 

 a structure intermediate between the very simple arrangement just 

 described and one in which the smallest element is a cube having a 

 similar group of atoms or molecules at every corner and at the middle 

 point of each face. The arrangement is called by crystallographers 

 the face-centred cube. The substitution of the sodium for the 

 potassium atom must transform one arrangement into the other. 

 This can be done in the following way, if we accept various indica- 

 tions that atoms of equal weight are to be treated as equivalent. 

 Imagine an elementary cube of the crystal pattern to have an atom 

 of chlorine at every corner and in the middle of each face, and an 

 atom of sodium or potassium as the case may be, at the middle point 

 of each edge and at the centre of the cube. We have now an arrange- 

 ment which fits the facts exactly. The weights of the potassium and 

 chlorine atoms are so nearly the same as to be practically equivalent, 

 and when they are considered to be so, the arrangement becomes the 

 simple cube of sylvine. But when the lighter sodium replaces the 

 potassium as in rocksalt the arrangement is on its way to be that of 

 the face-centred cube, and would actually become so were the weight 

 of the sodium atoms negligible in comparison with those of chlorine. 

 Of course the same result would follow were two or three, or any 

 number of atoms of each sort to take the place of the single atom, 

 provided the same increase were made in the number of the atoms of 

 both sorts. We might even imagine two sorts of groups of chlorine 

 and metal atoms, one containing a preponderance of the former, the 

 other of the latter, but so that two groups, one of each kind, contain 

 between them the same proportion of chlorine and metal as the 

 crystal does. We must merely have two groups which differ in 

 weight in the case of rocksalt, and are approximately equal in weight 

 in the case of sylvine. But it was best to take the simplest supposi- 

 tion at the outset : and now the evidence that the right arrangement 

 has been chosen is growing as fresh crystals are measured. For it 

 turns out that in all crystals so far investigated, the number of atoms 

 at each point must always be the same. Why, then, should it be 

 more than one ? Or in other words, if atoms are always found in 

 groups of a certain number, ought not that group to be called the 

 atom ? 



As soon as the structure of a crystal has been found we can at 

 once find by simple arithmetic the scale on which it is built. For 

 we know from other sources the weight of individual atoms, and we 

 know the total weight of the atoms in a cubic centimetre of the 

 crystal. In this way we find that the nearest distance between two 



