Oliver. — Vegetation and Flora of Lord Howe Island. 113 



with New Caledonia. A stream of migration is described as branching ofT 

 in New Guinea and traversing the Solomons and the New Hebrides. It 

 then turns to New Caledonia, sends an offshoot to Lord Howe Island, and 

 ultimately reaches New Zealand. 



Hemsley (Ann. Bot., vol. x, p. 282, 1897) favours a land connection to 

 explain the character of the flora of Lord Howe Island when considered in 

 connection with that of New Zealand, Norfolk Island, and east Australia. He 

 quotes Engler as having the same view, but Drude as thinking a land con- 

 nection between New Zealand and Lord Howe Island improbable on account 

 of the endemic character of the flora of the latter. 



My own conclusions, derived from a study of the birds (Trans. N.Z. 

 I list., vol. 44, p. 216, 1912) and of the flora, are that a land connection is 

 necessary to account for the remarkable set of animals and plants endemic 

 in the island ; but a considerable element due to a trans-oceanic immigration 

 of Australian forms is present. The evidence on which these conclusions 

 are based will now be considered. 



Genera. 



Of the 169 genera of vascular plants represented in Lord Howe Island 

 four are endemic. Cohneiroa and Hedyscepe are allied to New Zealand forms, 

 Negria to both New Zealand and New Caledonian genera, while Hoivea is 

 related to Malayan and tropical Australian genera. If the five species 

 belonging to these four genera be taken as modified descendants of species 

 which arrived by a land bridge, then they would indicate a New Caledonia - 

 New Zealand migration, with the land connection severed first at the southern 

 end, thus accounting for a greater degree of peculiarity for the New Zealand 

 related species. 



Of the non-endemic genera ninety-five are widely distributed, occurring 

 in New Zealand, Australia, and Polynesia ; and forty-seven more range 

 widely through tropical countries, but do not reach New Zealand. Their 

 presence may largely be due to the accident of latitude, and stamps 

 the flora as subtropical. The remaining genera have the following range : 

 New Zealand, 1 (Carmichaelia) ; Australia, 5 ; Polynesia, 3 ; New Zealand 

 and Australia, 11 ; New Zealand and Polynesia, 2 ; Africa, 1 (Moraea). 

 Numerically, therefore, Australian genera (158) preponderate ; Polynesia 

 comes second with 147, and New Zealand has 109. The presence of a large 

 proportion of widely distributed genera and species might have been expected 

 in the flora of an isolated island, because species possessing facilities for 

 wide dispersal would naturally form the bulk of immigrants after land 

 connection had been severed. 



Species. 



There are seventy endemic species of vascular plants in Lord Howe 

 Island, or 33 per cent, of the total number of species occurring there. If 

 these be divided according to their affinities, which can only be done 

 approximately, the following result is obtained : Related to species found 

 in New Zealand only, 17 ; Australia only, 11 ; Polynesia only, 10 ; New 

 Zealand and Australia, 3 ; New Zealand and Polynesia, 2 ; Australia and 

 Polynesia, 11 ; widely distributed, 16. Stated in this way the relationships 

 of the endemic species are about evenly balanced between the three regions 

 named. As supporting evidence of a former land connection it may be 

 pointed out that, compared with the Australian forms, the Polynesian and 

 New Zealand elements in the endemic plants are far higher than they are 

 in the non-endemic species. 



