RELAPSE AND RETROGRESSION. 87 



iimbellates themselves. Be this as it may, the family 

 is one which clearly lays itself out to attract a large 

 number of miscellaneous insects, as Miiller has shown ; 

 and its prevailing colours are white and pale yellowish 

 green. The flowers are all adapted to such small 

 visitors as prefer these hues. 



" The position of the honey on a flat disk," says 

 Sir John Lubbock, " which renders it accessible to 

 most insects, has the opposite result as regards the 

 Lepidoptera, which therefore, as might naturally be 

 expected, are but rare visitors of the Umbellifevce. 

 I have sometimes wondered whether the neutral tints 

 of these flowers have any connection with the number 

 of species by which they are frequented." This 

 pregnant hint is full of meaning for the student of 

 floral colouration. 



After so many instances of more or less probable 

 Retrogression, it will not surprise the reader to learn 

 that in an immense number of other cases there is 

 good reason to suspect some small amount of dwarf- 

 ing or even Degeneration. These cases might perhaps 

 be properly treated in the next chapter ; but their con- 

 nection with our present subject is so close that they 

 fall into place more naturally here. It may have struck 

 the reader, for example, when we were dealing with the 

 Crucifers, that many of the smaller white forms were 

 apparently lower in type than large and brilliant 

 yellow flowers like the charlocks. That is quite 

 true ; but then, many of these small types are de- 

 monstrably dwarfed and slightly degraded, as, for 

 example, Cardamine /lirsuta, which has usually only 

 four stamens instead of six, thus losing the most 

 characteristic mark of its family. In Soicbieva didyma, 



