2 
8  HwutTCcHINSON—THE MADDENI SERIES OF RHODODENDRON. 
there is no red or purple spotting, such as one gets in R, 
yunnanense, for example. R. burmanicum has a greenish-yellow 
and R. pachypodum a yellow corolla. 
Stamens.—These are numerous (15-25) in the Eumaddenia 
group, and usually 10 in the other two groups; they are, 
almost without exception, considerably longer than the corolla 
tube and rather unequal in length, corresponding to the slight 
zygomorphy of the corolla ; the filaments, except in R. Maddent 
and R. calophyllum, are pubescent in their lower part; the 
anthers are very large (8-13 mm.) in group Megacalyx, whilst 
in the other two groups they are much smaller, averaging 
about 5-6 mm. in Ciliicalyx. 
Ovary.—The number of ovary cells is usually consistent in 
each species. In the Eumaddenia group there are from Io to 12 
cells, and the ovary passes gradually into the style; in group 
Megacalyx the number is constantly 5, whilst in C7licalyx it 
is more often 6, and the transition to the style is usually very 
abrupt ; in all the species the ovary is densely lepidote, and girt 
at the base by a more or less shortly tomentose annular disk. 
Style.-—Except in R. ciliatum, where it is glabrous, the style 
is more or less lepidote; in R. ciliicalyx it is only so for a few 
millimeters above the base, and in this species there are a few 
hairs as well; I find the degree of scaliness of the style to be 
very constant in each species. 
Stigma.—The exact form of the stigma is not readily made 
out from dried specimens; in most species it appears to be 
depressed and lobulate, and more or less viscid. 
Capsule.—-This is very oblique in R. inaequale, less so in a 
few other species, but straight in the majority; the number 
of valves always corresponds with the number of ovary cells ; 
the capsule is constantly more or less lepidote, and the central 
axis, from the top of which the valves dehisce, is often tipped 
by the persistent style. 
Seeds.—Always small, brown or straw-coloured, mostly 
acute at both ends, and narrowly winged. 
Professor Balfour had already studied the indumentum of 
the Maddeni series when I took up this revision, and he very 
kindly allows me to insert the following notes of his own :— 
“These Rhododendrons have all of them lepidote leaves. 
Some of them have in addition in the young state a varying 
number of setose hairs which may disappear before the leaf 
reaches full size. The peltate scales occur upon both surfaces 
of the young leaf. They are sunk in pits of the leaf surface, 
these varying in depth, each scale showing a stout pluricellular 
stalk anda disk with typical umbo and fringe equalling the 
radius of the umbo. The outermost cells of the umbo at an 
