i6 DARWINIAN AND SPENCERIAN 



that what is apparently non-selectional Evolution may- 

 hereafter be shown to be strictly Darwinian. 



This last point is of such fundamental importance in 

 relation to those border subjects where physical and 

 biological science meet, that no excuse need be offered for 

 inviting its fuller discussion. In the domain of chemistry, 

 for example, I have often pondered the question whether 

 some principle of Selection or Survival may not be 

 applicable to the case of the synthesis of organic (i.e. car- 

 bon) compounds, many thousands of which are now known 

 to chemists.^ The special interest attaching to these 

 compounds from the present point of view is of course 

 their relationship to the lowest form of living matter. 

 The great majority of these compounds are purely artificial, 

 i.e. they are laboratory preparations which are unknown 

 as products of vital activity. But great as have been 

 the achievements of chemists in the way of producing 

 new compounds, it is becoming more and more evident 

 that from a scientific point of view negative results may 

 be as important as positive results. In other words, with 

 the progress of knowledge it is becoming apparent that 

 the possibilities of developing new atomic groupings are 

 subject to definite limitations. 



This is expressed in modern terms by saying that certain 

 configurations of atoms are possible and others impossible. 

 Although tentative hypotheses connecting the stability of 

 certain types of compounds with particular configurations 

 have been suggested, it must be confessed that our know- 

 ledge is still in an empirical stage. We cannot deduce 

 from known data with any degree of precision why the 

 possibilities of synthesis are restricted in this, that, or the 

 other direction. It may be said in a general way that the 

 stability of any atomic system must be ultimately ex- 

 plicable in mechanical or dynamical terms, and that the 

 greatest future development of our science may therefore 



' The new edition of Richter's Lexikon proposes cataloguing 150,000 

 formulae. 



