STRUCTURE OF FLOWERS. 81 



of the Rutaceae, presents an example of it; here the five 

 sepals are reduced to scales, because the flowers are in a 

 compact head. We find more numerous and decided 

 examples in families where the ovary is adherent, and 

 the flowers capitate, such as the Dipsacess and Compo- 

 sitae. In these plants, the tube of the calyx is reduced 

 to a thin membrane adhering to the ovary, and the limb 

 appears under different forms. Sometimes it has five 

 foliaceous teeth, like those of ordinary calyces, as in 

 Catananche. These teeth are sometimes changed into 

 membranous scales, free, as in Centaurea Cupina, or 

 united together, as in Hymenopappus and Favonium ; or 

 into almost spiny points, as in a species of Cnicus ; or 

 into bundles of simple hairs, as in Sonchus; united 

 together, and thus appearing branched, as in Stcehelina ; 

 or feathery, as in Scorzonera. 



It is so certain that the pappus of the Compositae is 

 the true limb of the calyx, that it sometimes puts on 

 every appearance of it; thus M. Dufresne has shown 

 me a plant of Podospermum laciniatum, (PI. 18, fig. 1, 2,) 

 the pappus of which was replaced by five linear, slightly 

 foliaceous lobes. 



I shall revert to these different forms of the pappus 

 when speaking of the fruit, and I shall only remark here, 

 that every organ called Pappus is only the limb of the 

 adherent calyx of plants with capitate flowers, in which 

 this organ is half abortive, or deformed by the pressure 

 of neighbouring flowers ; sometimes even it is entirely 

 abortive, and we then say that the pappus is absent : it 

 is replaced by a small circular rim, as in most Umbel- 

 life rae. 



The Valerianeae, although having their flowers dis- 

 tinct and not capitate, present also a true pappus ; this 

 results from the limb of the calyx being rolled inwards 

 during flowering, and submitted consequently to a pres- 



VOL. II. <; 



