170 APPENDIX. 
the same plant in Sir Wm. Hooker’s herbarium, I noticed one 
very important character that has been quite overlooked by all 
preceding observers: the anthers are here decidedly extrorse, 
instead of the usual introrse direction before assigned to them. 
This circumstance brings Dudoisia in close connexion with the 
two following genera, and at once removes them from the tribe of 
the Salpiglossidee. 
8. Anthocercis.—I was glad to avail myself of the opportu- 
nity of investigating the structure of the flowers in this genus 
from a plant in the living state of 4. viscosa. It agrees with 
the figure given by Endlicher in his ‘ Iconographia,’ tab. 68, of 
A, littorea, with the exception of the very important feature of 
the structure of the anthers, which, as in the preceding genus, 
offer the very distinct peculiarity of being affixed extrorsely just 
above the sinus upon the filament, so that the lines of dehiscence 
are towards the tube of the corolla, not introrsely towards the 
centre of the flower, as appears represented in the plate above 
referred to. The estivation of the corolla in 
Anthocercis viscosa is also very peculiar: at 
first sight it would be said to be induplicato- 
valvate, but upon more careful examination it {¥/ 
will be observed that each lobe of the border is (¥¥' 
distinctly supervolute, one of its edges being | § 
rolled inwards and overlapped by its opposite 9 ¥/ 
edge ; these are not all turned in one direction, (ug 
two being dextrorsely, and the other three coiled ng 
up alternately in a smistrorse order. This mode \¥J f 
of wstivation is certainly extremely unusual and {| 
peculiar, approaching that observed in the Goode. © 
noviacee, on which on a former occasion (Lond. 
Journ. Bot. vii. p.59) I have made some observa. 
tions. There exists between them this difference, 
that here each lobe is longitudinally and super- 
volutely coiled round upon itself, in a somewhat 
spiral form, while in Goodenia the winged margins are respec- 
tively folded back over one another, upon the plane of the cen- 
tral portion of each segment. I have also examined in the dried 
state the flowers of A. littorea, A. albicans, A. Tasmanica and 
A. scabrella, and they all appear to offer the same kind of zsti- 
vation and similarly extrorse anthers, so that these appear to be 
constant characters. It is worthy of remark, that the peculiar 
smell of the leaves and flowers of Anthocercis viscosa resembles 
that of the Myoporacee, and that its pedicels are bibracteated, 
which is also a feature in that family ; but its extra-axillary pe- 
duncles, the estivation of its corolla, the position of its stamens, 
its bilocular ovarium with numerous ovules attached to a thick- 
