150 Transactions. 



with species occurring in Lord Howe and Norfolk Islands (one or botL,), 

 Australia, and Polynesia, but not in New Zealand. The last row shows that 

 34 species, or 30 per cent, of the flora, are found in all four regions named 

 in the table ; and so on. The number of species extending to each of the 

 four regions is shown at the foot, whilst the total number occurring in 

 the Kermadecs — namely, 114 — may be obtained by adding together the 

 numbers for each row. 



The last column shows where the 12 endemic species are included in the 

 table, and the following list is in the order in which they appear : — 



Cyathea kermadecensis, Boehmeria dealhala, 



Imperata Cheesemani, Rapanea kermadecensis, 



Ascarina lanceolata, Sccevola gracilis, 



Homolanthus polyandrus, Coprosma acutifolia, 



Poa polyphylla, C. petiolata, 



Veronica hreviracemosa, Cyathea Milnei. 



The large Polynesian element represented in the above species seems 

 to indicate that arrivals are less frequent from the north than from New 

 Zealand. 



Before going into details regarding the geographical relationships of the 

 plants of the Kermadec Islands, a few remarks are necessary to clear the 

 way. The figures given above include 38 ferns and fern-allies, which, on 

 account of the lightness and abundance of the spores they produce, are of 

 little value in estimating the amount of affinity of the flora with those of 

 the surrounding lands. Again, I consider plants taking essential parts in 

 the different formations are of more importance than those occurring spora- 

 dically, and sometimes having the appearance of having been introduced, 

 but in the absence of direct evidence retained in the flora. As belonging 

 to the latter category I should name Rumex flexuosus, Eleusine indica, 

 Kyllinga brevifolia, Halorrhagis erecta, Calystegia sepium, Oxalis corniculata, 

 Solanum nigrum, Bidens pilosa, and several others. The Australian element 

 may be disposed of at once, and will be disregarded when considering the 

 relationships of the plants. Of the 78 species, all but 9 are found in New 

 Zealand, and not one Australian species does not occur in one or more of the 

 other three regions named in the table. It is probable that no plants have 

 reached the Kermadecs from Australia direct — a conclusion arrived at by 

 Mr. Cheeseman after his visit to the group in 1887 (1 ; p. 160). 



In the following remarks an endemic species is considered as equivalent 

 in its distribution to its most closely allied species. 



The 34 species of Kermadec Islands plants common to the four regions 

 named in the above table need not be further considered. They include 

 Pisonia Brunoniana, some coastal plants such as Mesembryanthemum aus- 

 Irale, Tetragonia expansa, Apium prostratum, and Sicyos australis, 19 ferns 

 and fern-allies, and 10 others. 



As pointed out in the introduction to this paper, the Kermadec Islands 

 form part of the New Zealand biological region. The figures given in the 

 table leave no room for doubt on this point. To the 16 species confined to 

 the Kermadecs and New Zealand must be added the 11 also found in Aus- 

 tralia, making in all 24 per cent, of the flora ; and assuming Lord Howe 

 and Norfolk Islands to be part of the New Zealand region the proportion 

 would be 47 per cent. The New Zealand species include six forest-trees 

 (Myoporum Ixptum, Coriaria sarmentosa, Corynocarpus la-vigata, Melicope 

 ternata, Nothopanax arboreum, Pittosporum crassifolium), Poa polyphylla 



