262 C. SKOTTSBERG 



of the species inhabit extratropical S. America, but many of them are so closely 

 related that they are little more than microspecies and in not few cases based 

 on one or two specimens from a single locality, and the number of separable 

 taxa will perhaps be reduced when more material becomes available. I^e this as 

 it may, 7 sections are represented in America, 5 in the Australian-X. Zealand 

 area, 3 are common to both, one of them ranging to the African sector, where 

 another section is endemic. As regards A. i)iasafuera7ia see p. 206 above. I sup- 

 j)ose we can draw no other conclusion from this distribution than that Acaeiia is 

 an Antarcto-tertiary genus, 2 sections having developed numerous species in the 

 Andes and Patagonia. 



Mari:;yricaypus, a small Andean genus, is so closely allied to Acaciia that they 

 have j)roduced a bigeneric hybrid in Juan Fernandez. Two more genera are found 

 in the Andes, TetvaglocJii}! and Polylepis. In S. Africa we have the large genus 

 Cliffortia. The remaining genera Sangiijsorba, Poteriuiu and Beiicomia (Macarone- 

 sia) belong to the X. hemis{)here. 



Sophoni sect. Tetrapterae is austral-circumpolar: New Zealand (3 species), 

 Chatham I. (i). Lord Howe I. (i). Austral Is. (i), Rapa (i), Marquesas (i), Hawaii 

 (i), I'^aster I. (i), Juan Fernandez (2), Chile (2; S. Diacrocarpa, however, rather unlike 

 all the others), Diego Alvarez (i), and Reunion (i). With the exception o'i macrocarpa 

 and the Hawaiian cJirysopJiylla the remaining species used to be united under 

 tetraptera Ait., otherwise endemic in New Zealand. They are very closely related, 

 but distinct; it is of minor importance if we call them species or geographical 

 subs{)ecies. Unless we believe that kS. tetraptera was carried by water from island 

 to island and was transformed into a new species wherever it landed, we must 

 look upon Antarctica as a one-time centre of a polymorphous population, which 

 radiated in various directions; we shall not discuss here how this may have hap- 

 pened. We have not to do with litoral but with inland plants; the pods are 

 ada|:)tc(l to float, assisted by the four narrow wings, Joiiow says, but some of the 

 forms have no wings at all, and even if they have, the pods open on the tree and 

 discharge their seeds. 



I'agara inayu and exterjia form their own section. When HC'RGER said [4.1. 19) 

 that I'agara had migrated to Juan Fernandez from the primeval forests of Peru 

 and Colombia he oxerlooked that the affinity is with palaeotropical rather than 

 with neotropical si)ecies; there are numerous species scattered from Australia and 

 Xew Caledonia to Polynesia and Hawaii, where many are endemic. Rutaceae were 

 })erhaps represented in the Antarctic in Tertiary times, and we have too look for 

 a route across to the American sector. 



The family liuphorbiaeeae is pantroj)ical, let alone that Euphorbia has at- 

 tained a world-wide distribution and flourishes also in temjierate climates. Dysopsis 

 is Andean, Seidelia (2) and Leidesia (i) South African, the fourth genus of the 

 JSPercurialis grouj), Mercurialis (<S), ranges from Xorth l^urope to the Mediterranean 

 and is found in \\. Asia. The southern genera seem to be more closely con- 

 nected mutually than with Mereiirialis. The disjunctions are interesting and difficult 

 to explain, unless we can find good reason to look for a common source in the 

 Antarctic. 



