318 THE FERTILISATION OF FLOWERS. — [parr mt, — 
female only, and in this case all insect-visitors must lead to the ; 
crossing of separate capitula (Carduus). Lastly, in the few 
Composite in which the florets are all unisexual, either the male 
and female florets are confined to separate capitula or the outer 
florets in all the capitula are female and the inner male (Calendula, 
Silphiwm, etc.). In the former case the crossing of separate 
capitula is obviously inevitable, and in the latter (since the outer 
florets develop first) it is probable, and more so than in the 
case described where centripetal development proceeds slawhy in 
a capitulum of hermaphrodite florets. 
In absence of insects self-fertilisation is often rendered possible 4 
in the hermaphrodite flowers by the stigmas curving backwards till ‘ 
their papillee come in contact with the pallang : 
The irritability of the filaments in Centaurea, Onopordum, 
Cichorium, Hieracium, ete., was observed more than a hundred : 
years ago by Koelreuter (396, Pt, 11). Severin Axell gives a — 
résumé of more recent observations on this point by Morren, Cohn, : 
Unger, and others (17). fi 
Sprengel recognised the significance of the enlarged margaial 
florets, the closing of the capitulum during rain, the brushes on the 
style or stigmas, and the distinct proterandry (702, pp. 365-384). 
Hildebrand made the floral mechanism of Composite, especially 
the distribution of the brushes and the stigmatic papillz, the sub-_ 
ject of a very elaborate paper containing many magnified figures , 
of styles (357). - In the same year (1869) Delpino gave a full ~ 
account of the adaptive modifications of Composite (178), and 
appended some notes on the bees seen by him visiting Composite. 
In a more recent work (180) he tries to prove the gencticg 
connection of Composite with Campanulacez and Lobeliacez, and 
to trace the gradual passage from entomophilous Senecionide to i 
distinctly anemophilous Artemisiacew. I am unacquainted with 
most of the forms which Delpino treats of, and am therefore ~ 
unable to criticise his conclusions, so I confine myself to a general — 
reference to his interesting research. 
Tribe Hupatoriacee. 
My brother Fritz Miiller tells me by letter that at Itajahy, 
Adenostemma, growing at the borders of woods, is the haunt of 
Zygenide and Glaucopide. 
-* 
My 
214. EUPATORIUM. CANNABINUM, L.—The capitulum contains 
five, sometimes only four florets, in each of which the tube is” 
