REVISION OF ANACYCLUS 



85 



Fig. 1 Capitulum, floret & cypsela measurements: A -involucre, B-ligule, C-disc floret, D- 

 receptacular scale, E - involucral bract, F - cypsela. 



Association Committee for Descriptive Biological Terminology (1962). The capitula are usually 

 single and terminal at the end of the peduncle. Sometimes the peduncles are structurally reduced 

 so that the capitula are aggregated into central sessile clusters (a synflorescence) or, on rare 

 occasions, into a single head (a syncephalum). The locations of floret and cypsela measurements 

 are given in Fig. 1. Descriptions of variation in corolla morphology generally follow the scheme 

 used by Jeffrey (1977). 



Taxonomic concepts 



The taxonomic concepts used in this revision are based on morphology, to some extent from the 

 results of hybridization studies (Humphries, in press, a, b), distribution data and observations 

 on ecology. 



In generic revisions, as Bremer (1976) quite rightly points out, the concept of the genus, and of 

 other categories for that matter, is seldom adequately discussed. Most flowering-plant taxonomists 

 still follow the methods of the Aristotelian essentialist philosophical tradition as practised by 

 Linnaeus and his generations of followers or, by contrast, by the nominalist ideals of Adanson 

 and his contemporary protagonists, the pheneticists, as exemplified by the considerable following 

 botanists have given to the empirical methods of Sokal &Sneath ( 1 963) and Sneath & Sokal ( 1973). 

 Consequently, in nearly all revisions, particularly in the distinctive families such as the Compositae 

 or Umbelliferae, genera are defined on one or a few a priori characters (monothetic groupings) or 

 many uncritically evaluated characters, from which totally artificial, polythetic groupings 'emerge'. 

 This approach is of course unacceptable in phylogenetic studies, since it in no way reflects 

 evolutionary relationships. Generic names, or those of any higher rank, can only really be 

 applied, in phylogenetic classifications, to monophyletic groups. To produce satisfactory mono- 

 phyletic groups which can be called genera (or tribes, families, etc.) the concept of 'resemblance' 

 must be resolved, so that those features which have undergone transformation (evolution) may 

 be recognized. Then, using only those characters which define monophyletic groups, it is possible 

 to identify 'sister groups', i.e. those groups of species which have demonstrable shared common 



