82 



THE COLOURS OF FLOWERS. 



or than the Cynaroidecu, which have no hgulate florets 

 at all. Hence we should naturally pcct them to be 

 blue or purple, whereas they are loi the most part 

 yellow of a very primitive golden type, while the 

 ray- florets of the Corymbifers are usually white or 

 pink, and all the florets of the Cynaroids are usually 

 purple. It is, of course, quite possible that a flower 

 might have progressed as far as the CoroUifloral stage 

 — might have joined all its petals into a united corolla 

 under the influence of insect selection — and yet might 



Fig. 27. — Vertical section of head of typical Hgulate ; yellow. 



never have got beyond its primitive golden-yellow 

 colouration. Something of the sort may be the 

 case with Chlora perfoliata diVnong the GentianacecE ; it 

 may even be the case with the LysimacJiia genus 

 among the Priviulacece ; and therefore it may also have 

 been the case with the primordial ancestor of the 

 Composites. But there are great difficulties in the way 

 of this explanation. In the first place, it would be 

 hard on that supposition to understand why the less- 

 developed Cynaroids should have outstripped the 

 more-developed Corymbifers and the most-developed 



