vacceae alll 
AUSTRALIAN SPECIES OF BREUTELIA 17 
character, though less strongly marked than in other species of 
the Section. But other specimens determined by Mitten himself 
show an intermediate type of base, e.g. the Waitangi River plant 
referred to above, as also a plant in Herb, Hook., “ Allan Cunning- 
ham, No. 96, es Z., 1834,” first named B. pendula, then deter- 
mined by Mitten as B. comosa; while a specimen ‘“ B. comosa, 
N.Z., Knight,” i Herb. Mitten, 9 = leaf-base of Acoleos pure 
and simple. Moreover, specimens of B. pendula in Herb. Hook., 
notably the Campbell Island aeunes of Hooker’s Antarctic 
Expedition, show considerable variation in degree of robustness, 
the more slender forms agreeing ivigh the type of B. pendula, the 
more robust agreeing in habit and leaf-base emer ee sokeceas s 
B. comosa from Waitangi River, New Zealand, cite 
f 
type of B. pendula, showing the extreme Acoleos form, from which 
a complete set of gradations can be traced, culminating in the 
robust tomentose form with fairly wall. marked ‘ Hubreutelia” 
base of Mitten’s type of comosa. If therefore B. comosa is to be 
retained as a species, it must be on the ground of the type of base 
and robust habit being correlated with the only other character 
given by Mitten, vez. the capsule being suberect or horizontal in- 
stead of being pendulous as in B. pendula. The specimens in Herb. 
Kew., however, and others that I have received through the kind- 
that this character is absolutely unreliable. Thns robust 
specimens with a clearly marked Hubreutelia base, sent me by 
Dr. Cockayne, hawa distinctly pendulous capsules, and the same 
is the case with ‘“B. comosa, Kelly's Range, New Zealand, 
T. W. N. Beckett, No. 1030” at Kew; but the most convincing 
examples are several in which capsules i in both positions occur in 
the same tuft ; wees - markedly the case with “ B. comosa, Otira 
Ra”. 1030 B”’ at Kew, where among numerous 
capsules, ‘all ppaas and ripe, most are pendulous, several 
horizontal, and a few suberect. ‘“B. pendula, VY. D. Land, 
o. 2282” in Hisch. Hooker, as also ‘‘ V. D. Land, Laurence and. 
Soa No. a ” equally show both positions of capsule in the 
same specim 
I have Gurnlors no hesitation in concluding that B. pendula 
Hook. is a highly variable plant in size, branchi f 
» aegree 
tomentosity, form and size of leaves, and position and size of 
capsule; standing on the border line between 
Eubreutelia, it must be held to include B. comosa. Mitt., which 
in view of the ambiguous 9 Papa referred to above cannot 
even be clearly characterized as a variety. In view, indeed, of its 
better —— the comosa jo would properly be considered 
while Menzies’ eats (the type of B. pendula Hook.) 
Brotherus implies a Jonbk as ~~. yeceren B. pendula occurs 
on Kerguelen Island—‘nach Mitten auch auf den Insel Ker- 
JournaL or Borany.—Vot. 53. [January, 1915.] c 
