314 APPENDIX. 



may be more correctly appreciated by what is known to be present rather than what 

 is problematically absent. The composition of the Panama flora, so far as investigated 

 by Seemann and Hayes, is so decidedly South American that there is no necessity for 

 insisting upon it, and northward extensions are prominent. Taking some of the 

 mainly tropical orders, such as the Dilleniacese and Anonaceee, we note a gradual 

 thinning out northward, and an extension into the southern province of many South- 

 American species. The essentially eastern South-American Lecythidese, a suborder of 

 the Myrtaceae, is represented by four genera and seven species, one of which is common 

 in Nicaragua, the northern limit of these trees. Podocarpus replaces Pinus in the 

 mountains of Costa Rica, and the Cyclanthacese are relatively numerous in the lower 

 regions. Noteworthy examples of southern limits of northern types are offered by 

 Liquidambar *, Sabiaceae, and Juglandeae in Costa Rica, and Pinus in Northern Nica- 

 ragua. The oak vegetation of the Volcan de Chiriqui comprises at least three species ; 

 and Arbutus and Arctostaphylos give way to South-American genera of the Vacciniacese 

 in the mountains generally. Chamwdorea, the characteristic genus of palms in the 

 oak-forests of South Mexico, is represented in the southern province by at least half 

 a dozen species, but the majority of the palms belong to genera having their greatest 

 development south of Panama. Many other examples of a change in the vegetation 

 nearly coincident with the northern boundary of Nicaragua might be given, but it 

 seems needless. 



An examination of the specimen of the mountain flora (pp. 282-299) brings to 

 light the fact that nearly, if not quite, all the genera there recorded from 8000 feet 

 and upwards in our southern province are such as range from Mexico to the Andes of 

 South America, and some of them wider. As bearing on this question it may be 

 mentioned that the alpine forms of the Andes of South America belong for the greater 

 part to the same genera which inhabit the higher regions of the Andes of Central 

 America and Mexico, though the species are very rarely identical. Engler very fully 

 discusses the relationships f of the vegetation of the upper regions of the mountains 

 of South America and Mexico and the northward and southward migrations that may 

 have taken place. We feel convinced that there has been a northward extension of 

 temperate and alpine forms as well as of tropical, and should regard such genera as 

 Drimys, Fuchsia, Colobanthus, Calceolaria, Boupala, &c. % as of southern origin. There 

 are other genera peculiar to these mountains so equally developed north and south of the 

 Isthmus of Panama that they may have migrated in either direction, always assuming 

 that each type originated in only one place. 



Finally, it may be briefly stated that the foregoing attempt to analyze the flora of 



* Inadvertently left out of our distribution tables, 

 t Versucb, &c. ii. pp. 215-256. 



$ For further particulars on this subject see Botany of the 'Challenger' Expedition, i., Introduction, 

 pp. 52-65. 



