672 Studies in the Theory of Descent. 



and again others may be seen in which the 

 differences predominate. This last deviation 

 would then be designated by many systematists 

 as a new species, but not so by others. 



The " specific type " is thus indeed a kind of 

 mosaic-work, but it is a structure to which all the 

 single characters the stones of the mosaic 

 belong and build up one ha.rmonious whole, and 

 not a meaningless confusion. Some of the stones 

 or groups of stones can be taken away and re- 

 placed by others differently coloured without the 

 structure being thereby necessarily distorted, i. e. 

 destroyed as a structure ; but the larger the stones 

 which are exchanged the more necessary will 

 corrections in the other parts of the structure 

 become, in order that the harmony of the whole 

 may be preserved. 



Still more weighty than those insensible tran- 

 sitions which in various groups of animals so 

 frequently connect species with species, appear 

 to me, however, the facts made known in the 

 second essay of the second part of this volume, 

 which prove that the two forms in which one 

 species appears can change entirely independently 

 of one another. The caterpillar changes and 

 becomes a new variety or even species (according 

 to the form-value of the change), wHilst the 

 butterfly remains unaltered. How could this 

 occur if some other law than that of physiological 

 equilibrium linked together the parts or charac- 



