REVISION OF ANACYCLUS 



129 



Fig. 21 Anacyclus homogamos'. A - habit, B - involucral bract, C - receptacular scale, D - disc 



floret, E - cypsela, F - anthers. 



The evidence will not be presented here since the details form part of a critical experimental 

 study (Humphries, in press, a). However, the principal reasons for accepting the second hybrid 

 hypothesis are that: (a) In almost all herbarium specimens, the ligules are extremely short and are 

 frequently masked entirely by the involucral bracts. They fall into one particular size class and 

 do not form a gradual series between rayed and rayless forms. There are fewer ligules in short- 

 rayed morphs than can ever be found in a radiate plant, and these are irregularly arranged at 

 the periphery of each capitulum. (b) This condition can readily be synthesized in artificial Fj 

 hybrids, especially in crosses involving A. homogamos, A. clavatus and A. radiatus (Humphries, 

 in press, a), (c) In most herbarium specimens it is extremely difficult to identify ligule colour, 

 but it is possible to show that in a number of cases both white and yellow short-rayed morphs 

 exist, (d) The short-rayed plants are often found in mixed populations with the putative parents 

 and invariably have sympatric distributions with one or both of them, (e) There are numerous 

 citations in European literature (see synonyms) of collections believed to be rayless forms of 

 A. clavatus and A. radiatus found in Europe. However, truly rayless forms are extremely rare in 

 Europe and occur in abundance only in north Africa (Fig. 22). 



