94 C. J. HUMPHRIES 



by contrast, the time during which conditions are suitable for flowering may often be considerably 

 shortened, hence curtailing the time-span for producing successive generations of capitula. In 

 these situations there can be a considerable reduction in the number of developing branches, and 

 the condition in Fig. 3 B is reached. In extreme situations further branch reduction and loss of 

 leaves on the peduncles can occur to produce eventually habit types indicated in Figs 3 C, 3 D, 26. 

 These are commonplace in the east Mediterranean endemic A. nigellifolius and on rare occasions 

 in A. clavatus {A. capillifolius Maire). 



A seemingly independent trend is in contraction of the stem and the peduncles. Many plants of 

 A. radiatus and A. clavatus appear to have what looks like a protracted basal rosette, with branches 

 of the stem and peduncles emerging nearer the base of the plant. The leaf internodes are extremely 

 short, and the leaves emerge alternately. This condition is illustrated in Fig. 7 E. Often stem 

 reduction is so complete that branching occurs directly from the taproot and there is an almost 

 total loss of the larger leaves making up the rosette (e.g. in A. monanthos subsp. cyrtolepidioides). 

 In A. monanthos subsp. monanthos stem loss is combined with peduncle reduction to produce a 

 central acauline, sessile capitulum of the type indicated in Figs 3 F, 14. Lateral peduncles, when 

 present, emerge from the axils of the rosette leaves or branch directly from the vestigial stem below 

 the capitulum. A related trend is the fasciation and tumescence of peduncles as they become 

 further reduced so that a synflorescence with 2-6 capitula is formed (Fig. 3 G). The most derived 

 condition is produced when several or all of the capitula of the synflorescence fuse to form a 

 syncephalum. Here, the lateral peduncles can either emerge from rosette leaf axils directly below 

 the syncephalum or on extreme occasions even appear to emerge from the axils of involucral 

 bracts (Fig. 3 H). In well developed specimens of A. monanthos subsp. monanthos secondary 

 syncephala and synflorescences have been observed. 



In A. pyrethrum (Figs 3 I, 12), the only perennial species of the genus, a rather different situation 

 from the synflorescence and syncephalum of A. monanthos has developed; the stem is greatly 

 reduced and fused directly with a long woody taproot to form a submerged basal caudex. In 

 early spring large leaves emerge from the centre of the caudex to form a regular dense rosette. 

 Eventually these are succeeded by a whorl of peduncles, emerging from the axils of the innermost 

 leaves. The peduncles are commonly single-headed or only once or twice branched, providing a 

 distinct, although somewhat contracted, flowering period. During the winter the inflorescence 

 and leaves die right down to the ground, to be succeeded in the following season by new flowers. 

 The development in A. pyrethrum differs from that in A. monanthos for various reasons. It shows 

 no indication of fasciation or tumescence in the peduncles, does not have a synflorescence or 

 syncephalum and has only one generation of flowers, never having a secondary branching point 

 or any means of vegetative spread. The stem reduction in A. pyrethrum may well be an adaptation 

 to the alpine environment in which it grows. A. pyrethrum is an obligate outbreeder, with a sporo- 

 phytic, self-incompatible breeding system (Humphries, in press, a), and its flowers all exhibit the 

 common Anthemidean, radiate gynomonoecious condition. In A. monanthos, stem, peduncle 

 and capitulum modifications may be an adaptation to the hot, dry desert environment in which 

 it is found. It is self-compatible, and its relatively inconspicuous flowers are discoid-herma- 

 phrodite, which indicates that selfing must frequently occur. As we have seen, the evolution of 

 the A. pyrethrum and A. monanthos inflorescences appears superficially to be very similar. How- 

 ever, the reduced inflorescence in each species appears to be the derivative culmination of an 

 independent parallel trend, since they are unique conditions for the genus as a whole and indeed 

 for the sister group of the genus. This means that the evolution of the A. pyrethrum rosette-type 

 inflorescence must have passed through a number of morphological changes, if it did indeed 

 originate from a branched cyme with long peduncles, involvingthe loss of a determinate capitulum, 

 stem reduction and fusion with a woody taproot. It is interesting in this context that the extinct 

 'Magdeburg' officinal plant, A. officinarum (p. 114), which may simply be an annual derivative 

 of A. pyrethrum, produced a long central stem with alternate leaves. 



The prospect of secondary condensation cycles occurring in Anacyclus, culminating in a 

 rosette, synflorescence or solitary capitulum, leaves the ancestral condition in some doubt, since 

 similar morphological trends can occur in related genera. Achillea, for example, normally has a 

 tight corymbose cyme but can have species with solitary capitula, as can Leucocyclus. More 



