336 MR. G. BENTHAM ON COMPOSITA. 
Page | Page 
Table 10. Summary.of Ame- 7. Connexions be- 
rican distribution ...... 517 tween distant 
Notes: 1. Mexican Region... 517 Old-World R ... 551 
2. United-States R... 519 Insular Regions ............... 554 
3. West-Indian R.... 521 | 1. Sandwich Islands (tab. 
4. Andine Region... 522 IS 555 
5. Brazilian Region. 524 2. Galapagos (tab. 14) ... 556 
6. Chilian Region ... 526 3. Juan Fernandez, Masa- 
7. Connexions be- fueram, &e...... es 557 
tween distant American 4. South-Sea Islands...... 558 
Fegions feces 527 5. Atlantie Isl. (table 15) 559 
Table 11. Number of Com- | 6. St, Helena and Tristan 
posite in the several | d'Acunha o 563 
Old-World Regions ... 529 7. Mascarene Isl.(tab. 16) 564 
Table 12. Summary of Old- "8. New Caledonia......... 566 
World distribution...... 539 9. New Zealand (tab. 17) 567 
Notes: 1. Mediterranean R. 539 D. Colonizing Composite or 
2. Europxo-AsiaticR. 542 Introduced species...... 568 
3. Tropical African R. 544 Table 18. Escapes from cul- 
4. Tropical Asiatic R. 546 UTI E1816) ors a 569 
5. South-African R... 547 Table 19. Weeds of cultiv... 570 
6. Australian Region 549 Conclusion (csc. re cee 576 
I. IxTrRODUCTION. 
THE Composite are at once the largest, the most distinct, and 
the most uniform, and therefore the most natural, of all orders of 
Pheenogamous plants. Nearly ten thousand known species are 
separated from each other by characters most of which are usually 
considered as only of secondary importance; and I cannot recall 
a single ambiguous species as to which there can be any hesita- 
tion in pronouncing whether it does or does not belong to the 
order. The very few cases where species have been erroneously 
referred to or excluded from it have been the result of conjec- 
tural determinations of imperfect specimens, or of gross ignorance 
on the part of the observer. The andrecium, gynecium, and 
fruit, as to all the essential characters of number of parts and rela- 
tive position, the seed and its embryo in every particular, are 
absolutely uniform throughout the order; or in the very few 
cases of a slight variation (as, for instance, in the shape of the 
cotyledons) the differences are no more than specifie, varying in 
one and the same small genus. To distribute, therefore, these ten 
thousand species into thirteen tribes and above seven hundred 
