1902.] Antarctic Origin of the Tribe Schamece. 497 



All the material on which the map is founded has been seen and 

 determined by me; no fact in the distribution is copied out of a list 

 or depends on a name or a distribution number. The material 

 recorded by number or name of collector in my MSS. is about 1850 

 collections. The great mass of the commoner species, or commoner 

 habitats, is not recorded in my MSS., but is virtually included in the 

 map. 



The genera included in this sub-order are very close together, and 

 may be arranged on a different system from mine ■ this would not 

 affect at all the map, which deals only with species. I have doubtless 

 made some errors in the specific determinations. Moreover, any com- 

 petent man revising the material would have a different opinion from 

 myself as regards some species and varieties. The utmost alterations 

 that could thus be necessitated in the map would be two or three black 

 dots more (or less) in Australia and the Cape. 



The rings represent Schoenus nigricans, Linn., a cosmopolitan species, 

 and Schoenus ferrugineus, Linn., a species closely allied to it but confined 

 to the sub-area 1 — " Cooler Europe." [It is so closely allied as to be 

 sometimes confounded with it by learned cyperologists.] 



In an exactly parallel manner, the crosses represent Cladium 

 Jamaicense, Crantz, a cosmopolitan species, and Cladium triglomeratum, 

 Nees, a United States plant so closely allied that some competent 

 cyperologists call it a var. of C. Jamaicense. 



Five of the genera occur in Oceania and South America ; two occur in 

 South Africa and South America ; five occur in Oceania and the Cape ; 

 but the genera are so closely allied that little can be inferred from this. 

 Only two or three species are common to Australia and South 

 America. 



Numerical tabulations on a large scale are viewed with suspicion 

 by botanists, as they are often drawn in great part from books or 

 from lists ; the percentage of errors then introduced from wrong 

 identifications, diversities in nomenclature, variable limits assigned 

 to species and areas, and doubtfully wild species, is so large that it 

 invalidates the conclusions. No one of these sources of error taints 

 the present tabulation. 



The conclusions suggested are — 



(1.) The sub-order Schoenese originated in prehistoric time at some 

 centre on which the three streams of species (Patagonian, 

 Cape, Australian) converge, and has spread from that centre 

 northward. 



(2.) Two of the genera, viz., Schoenus and Cladium, have developed 

 largely, and produced each a great number of species. 



(3.) One species in each of these two genera has become cosmo- 

 politan, a common case in large genera from one end of the 

 Genera Plantarum to the other. 



