124 Floivcrs and ihcir Pcdii^^nrs, 



Ai^ain, \ou may have observed that I said just 

 now the primitive ancestor of the goose-grass had five 

 petals. lUit tlie present united corolla has only four 

 lobes instead of five, and it is this arrangement, ap- 

 parently, which has gained for the whole tribe the 

 name of stellate. Now the tropical Rubiace.e, which 

 we saw reason to believe represent an earlier stage of 

 development than the goose grass group, have usually 

 five lobes to the corolla ; and in this respect they 

 agree in the lump with the whole great class of dico- 

 tyledonous plants to V ich they belong. Therefore 

 we ma\- fairly conclude that to have four lobes instead 

 of five is a mark of further specialisation in the 

 stellates ; in other words, ir is they that have lost a 

 lobe, not the other madder-worts that have added 

 one. This, then, gives us a further test of relative 

 development— or perhaps we ought rather to say of 

 relative degeneration — among the stellate tribe. Wild 

 madder, whose flowers are comparatively large, has 

 usuallv five lobes. Yellow crosswort has most of its 

 blossoms four-lobed, interspersed with a few five-lobed 

 specimens. Goose-grass occasionally produces large 

 five-lobed flowers, but has normally only four lobes. 

 The still smaller skulking species have almost inva- 

 riably four only. In fact, the suppression of one 

 original petal seems to be due to the general dwarfing 



