178 
Chakles Austin Gardner. 
Frutex 30 eni. altus; folia usque 1,3 cm longa; calyx 0.5 mm. longiis, lobi 
3.2 mm. longi; ^■exillum 10 mm. iongum ac latuni^ alae 8.5 mm. lougne, carina 
7 mm.; OTariiim cum stylo 6.5 mm. Iongum. 
Hab. in distr. Coolgavdie ad fines meridionali Young Kiver, 
Gardner Jany. 1936. 
The three species described above may be differentiated by the foilo-\ving 
key : 
A. Calyx very unequal, the hvo upper segments united 
almost to the summit into a truncate or emarginate 
lip. 
a. Kacemo'^ slender, loose, the rhachis angular and 
shortly pubescent; leaves 4-4.5 cm. long; 
shrub 1-2 metres high. . . . . . . . . M. longifolia 
b. Kacenu^s dense; rhachis not angular, densely 
villous; h*aves 1-2 cm. long; shrnb less than 
a metre high . . . . . . . . . . M. taxifolia 
E. Calyx regular or almost so, the two ].>osterior lobes 
united to the middle; leaves 1-1.5 cm. long . . M. densiflora 
With the exception of M. ojcyclada F. MuelL, a species referred with 
some doubt to this genus, the three species described above differ from all 
otlier \Yes1eni Aiisfrnlian species in the number of the ovules which is con- 
stantly two. This brings the group close to 1/. reticuliita and 1/. aotoules, 
species restricted to Clueensland and Ncav South Wales. The habit of the 
plants is much like that of O.njlohiuni § Podolobium, but the ovary in each 
case is completely 2-eelled when young. The character upon which Mirhelia 
is established, namely the longitudinal division of the pod into two cells is 
not always readily observable in ilowering specimens. In 3/. dilaiata and 
3f. florihimda it is at once npY>ar(mt, but in most of the other AVesteru Aus- 
tralian species the partial intrusion of the false dissepiment from either one 
or both sutures is (d’ten very imp(*rfect and easily overlooked. The develop- 
ment of this fnlso diss(q>iment reijuires further study. In 3/. racemosa Turez.,. 
for example, it is entirely absent, and was described as such hy 
Tin’czaninow and Eentham is in reality a development of cellular tissue 
characteristic of CalHstachys and some species ()f OiJCfjlohmmj e.g. 
O. graniticfini. 37. racewom is in fact a true Oaylohium, and ivas described 
by the author under the name Oxylohium BennettsH in this Journal. The 
type species has been matclied with the type number of Mirhelia racemosa 
at Kew, aiid the tAvo are identical. 
Oxylohium racemosum (Turez,) C. A. Gardn. comb. nov. 
{Mirhelia racemosa Turez.; Oxylohium BennettsH C. A. Gardn.) 
Mirbelia seorsifolia {P\ MuelL) C. A. Gardn. comb. nov. 
The ]Aod of this ])Iant is typical of Mirhelia. Alueller described the 
ovules as being tAVO in number, but in reality 4 and 6 ovules are commonly 
present. The sj)ecies had already been regarded as an anomalous 
Gastrolobium. — Gastrolohium seorsifoliim F. Aluell. 
