Origin and Development of the Composites. 269 
is expelled at the first touch. This explosive movement has as yet 
been observed only in the Mutisieae, and is sometimes so energetic 
that the pollen is thrown over the edge of the capitulum. 
29. The movement in Beilis perennis is most active, as in 
many other species, between noon and 3 p.m., and is sometimes so 
strong that the anther tube is split against the comparatively thick 
and rigid style. 
30. The anther tube is frequently split in Celmisia verbascifolia, 
cp. notes 2, 6 and 17. This character is also recorded for Cratistylis 
conocephala (49, p. 267, note 8). 
31. When repeatedly tested with a cloud over the sun 
Erigeron Coulteri showed no movement, but an hour later with a 
few minutes of sunshine the “ A” type of movement was distinct. 
32. The movement is greatest in Erigeron multiradiatus when 
touched so that the reaction is towards the centre of the capitulum. 
33. Juel (28) records no irritability in Gaillnrdia aristata and 
the var. grandijlora showed none at 6.30 p.m., but was distinctly 
sensitive at 10.30 a.m. (Kew). 
34. The movement is greatest in Heliopsis senbra when it 
is towards the centre, cp. note 32. 
35. In var. gratissimci of the above species the anthers are 
forced apart when the filament is stimulated and the pollen is 
presented along the slit. In the evening the stamens are more or 
less free from each other, cp. notes 2, 6, 17 and 30. 
36. The lateral movement in Helianthus annuus is slight and 
slow, but Gray (13) previously recorded a distinct lateral movement 
(see Section A, Chap. III). 
The physiology of the movement is discussed in Section D 
and the phylogenetic significance in Section E, so that it is only 
necessary to point out some of the chief facts of the distribution of 
irritability in the family as shown by Table VIII. 
Of the three main types, A, B and C, of movement C seems to 
be the most primitive because it involves only one step—the con¬ 
traction of one of the five filaments ; B involves the transference of 
the stimulus to the other four filaments, and this occurs in an 
indefinite order, giving an indefinite lateral movement; A involves 
the transference of the stimulus to the other four filaments in such 
an organised fashion that all five contract simultaneously. 
In the cases where the A type of movement is recorded as 
slight it is possible that this is due to an incipient C motion, which 
would show as such in better conditions. The A type in the 
Anthemideae, Lindheimera, Coreopsis and Erigeron, however, seems 
