258 Groom.—Remarks on the O ecology of Coni ferae. 
Sargent (’ 05 ) as growing on dry gravelly slopes and as having leaves ‘thick 
and firm or subcoriaceous 
2. The nine species with the narrowest tracheae are all evergreen. 
But of these, the five with the widest tracheae are what I may term ‘ sub¬ 
evergreen ' or ‘ sub-deciduous as all their old leaves are cast off at the same 
time, or very shortly after, the new ones unfold in spring, while the four species 
having the narrowest tracheae are the only four that are typical evergreen. 
Miss Beatrice Rhodes also measured the mean comparative diameters 
of the largest vessels of two to four-year-old twigs of three species of Fagus. 
They are less in the two evergreen South American species, Fagus 
Cunninghami (0-0155 mm.) and F. betuloides (0-0225), than in the deciduous 
F. sylvatica (0-03) and F. obliqua (0-043), latter °f which is also South 
American. But the wood of an older piece of F. Cunninghami from Nord- 
linger’s series gave a diameter of 0-057, and that of an older stem of 
F. sylvatica a diameter of 0-052. Undoubtedly the twig of F. Cunninghami 
was correctly named, as it came from a plant growing at Kew: there is 
no equal guarantee that Nordlinger’s specimen was genuine. 
But just as in desert types there seem to be two alternative methods 
of reducing the conducting channels, namely, by decreased number or 
decreased calibre of the tracheae, so in temperate evergreens the same 
appears to be the case. For I find that in comparing the wood of Prunus 
domestica with that of P. lusitanica the tracheae are of approximately the 
same size in the two species, but they are far more scanty in the latter 
species. 
That desert types and these evergreen species should show two different 
methods of bringing about the same result, namely, a limitation in the 
number of wide tracheae, suggests strongly that the phenomenon is not 
merely a case of inevitable correlation, but represents a structural change 
directly beneficial to the species. 
The remarkably exact agreement between deciduous and evergreen 
habit and width of tracheae in Quercus , the evidence supplied by Buxus 
and IleX) and the divergent case of Prunus , all suggest that even if the 
ancestors of the Coniferae had possessed deciduous leaves and tracheae, the 
assumption of the evergreen habit, coupled with the acquisition of small 
xerophytic leaves, would have been associated with a reduction in the 
number or calibre of the tracheae, until the wood more or less perfectly 
agreed in structure with that of existing Coniferae. 
Thus the question arises : Is there any evidence of evergreen dicotylous 
woody plants having undergone such a process of reduction in their wood ? 
The Magnoliaceae and Trochodendraceae supply examples whose true 
interpretation may perhaps be the one suggested. Drimys, the magnolia- 
ceous genus characterized by having its water-conducting tubes consisting 
of tracheides only, is evergreen and in austral regions marks the limit of its 
