RELAPSE AND RETROGRESSION. 89 



succeed, in certain positions, better than larger ones. 

 Moreover, they will recollect that in numerous in- 

 stances the larger blossoms of each family are so 

 exclusively adapted to insect fertilisation that they 

 cannot fertilise themselves ; while among the smaller 

 blossoms alternative devices for self-fertilisation com- 

 monly come into play after the flower has been open 

 for some time, if it has not first been cross-fertilised. 

 Structural considerations show us that in most such 

 instances the larger and purely entomophilous flowers 

 are the more primitive, while the smaller and occa- 

 sionally self-fertilising flowers are derivative and 

 degraded, having usually lost some of their parts. 

 Hence, in tracing the progressive law of colouration 

 in the families generally, it is necessary, for the most 

 part, to consider only the larger and more typical 

 species, setting aside most of the smaller as products 

 of degeneration. 



Moreover, as Mr. Henslow has shown, a large 

 number of minute species have fallen back almost 

 entirely upon self-fertilisation ; and these, we must 

 presume, when they retain their petals at all, retain 

 them either by mere hereditary habit, as functionless 

 relics, or else use them to effect an occasional cross 

 at long intervals. Such degenerate and dwarfed 

 species survive exactly as mites and other degraded 

 forms survive in the animal world — because they fill 

 certain holes and corners of the organic system better 

 than more highly developed forms could do. The 

 advantages of cross-fertilisation are seen in the large 

 stature and vigorous constitution of the truly en- 

 tomophilous or anemophilous plants, which have 



