34 Dansk Botanisk Arkiv, Bd. 2. Nr. 8. 



VII (1878) 167) that they are connected by a series of forms 

 which makes it impossible to draw a separating line between them, 

 at least not until growing experiments have been tried, and these 

 would probably result in a fair number of distinct micro-species, 

 not in two. Therefore, I do not find it allowable from systema- 

 tical and phytogeographical points of view to maintain two forms 

 the extremes of which may be discernible, while the main bulk 

 arc indiscernible. That Hooker's T. centrocarpa is an extreme 

 form, seems probable from the fact that it has not been collec- 

 ted since Drummond's original specimens ; while specimens referred 

 to the less sharply defined T. nana have often been found. 



The supposed distinction marks are: T. centrocarpa has ap- 

 pressed, sessile fruits, 3,8—5 mm long, and the backs of the car- 

 pels are subcarinate, while T. nana has erect, pedunculate fruits, 

 2 — 3 mm long, and rounded backs of the carpels. 



Now we have in another species, T. calcitrapa, specimens 

 with sessile and larger fruits and others with pedunculate and 

 smaller fruits, without separating them into two species, and conse- 

 quently there is no reason for doing it here. I have myself col- 

 lected a number of specimens of T. centrocarpa at Armadale, near 

 Perth (No. 143, 20. Sept. 1914) which show in some individuals 

 sessile (see uppermost specimen in PL IV, Fig. 2), in others pe- 

 dunculate fruits (see the larger specimens in PL IV, Fig. 2), and 

 the back of the carpels varying from carinate to rounded. The 

 length of the fruits is also highly variable, and there seems, usu- 

 ally, to be a correlation in such a manner that the sessile fruit 

 is larger than the pedunculate one. 



Neither does the geographical range show any distinction 

 between the forms: I have seen specimens of T. nana collected 

 by F.V.Müller himself in Victoria (Station Peak, 1867 ; PL IV, Fig. 4) 

 and they do not differ in any essential point from West Austra- 

 lian ones ; neither do Tasmanian specimens collected by R. Gunn 

 and quoted by F. v. Müller (Fragm. VI (1867) 82) differ. Also 

 from South Australia I have seen specimens (see PL IV, Fig. 1). 



The only two instances in which my examination leaves a 

 little doubt are the following: 



(1) var. brevicarpa nov. var. (fructus oblongo-linearis, brevis, 

 2—2,5 mm longus, basi haud calcarata). Some specimens collected 

 by myself at Yallingup Cave (No. 145, 26.— 27. Sept. 1914; see 

 PL VI, Fig. 3) are rather large, and have shorter and more long- 

 stalked fruits than usual, and the basal spur of the carpels is 

 much less developed than usually in T. centrocarpa, where it is 



