DEGENERATION. 97 



debased bisexual form. Its stamens are numerous, 

 and they hang out to the wind, as do also the 

 feathery stigmas in the female flowers, to catch the 

 pollen from neighbouring heads. But the closely- 

 allied Sanguisorba officinalis (Fig. 30) is evidently 

 an entomophilous variation on the same ancestral 

 form ; for it resembles Poterium in every respect 

 except in its flowers, which have very few stamens, 

 enclosed in the purple calyx-tube. This interesting 

 case shows us that when a flower has once lost its 



4'V4 "^* 



F"jc. 29. Flower of salad burnet FIG. 30. Flower of great burnet (Satigui- 

 (Poterium sanguisorba) ; green sorba officinalis) ; purple and entomoplii- 



and anemophilous. lous. 



petals and become anemophilous, it cannot re-develop 

 them if it reverts to insect fertilisation, but must 

 acquire a coloured calyx instead. The same lesson 

 is perhaps elsewhere enforced by Glaux maritima 

 among the Primulacea, and by Clematis among the 

 Ranunculacece. 



Mr. Darwin remarks that anemophilous flowers 

 never possess a gaily-coloured corolla. The reaspn 

 is clear. Such an adjunct could only result in the 

 attraction of stray insects, which would uselessly ea-t 

 up the pollen, and so do harm to the plant. Hence 



H 



