rAR, typica from Kliaaia,
2 6 4 EUGOl lPHOSTEMMA.
PLATE 94. Gomphostemma lueidum Wall. Specimen of
Callatli/ N. 243.
This species is nearly related to Q. oHongum, nnJ may usually bo distinguished by its serrate (not
crenate) leaves which ai'e narrowa' and smaller, ns well as Ly its distLaotly smnllei' flowers, the tomentum
on which is generally rufous, find is always much donsor than on those of G. oblongum. But the leaf
charactoi- ftiils, since Aiadamans gathetings of Q. oblongum skew ci'enate and smato leaves on the same
broneh. Mr. Bt?ntliaui has used ns diagnostic the chai'acter " calycibus . . . dentibus . . . . abbrevilitis"
for G. luoidum, and " dentibus lancoolato-lineoiibus aoutis" for G. oblongum—o.ix expression which makes
his description of G. lueidum refoi' esclusivcly to WtilHoli's Silhet plant, and which excludes ^^'allioh'8
Tenasserim one. There is no specimen of Wall. Cat. u. 2156/1 at Calcutta, but Mr. Hemsley has
very kindly compared for me at tlio Linnean Society's rooms a specimen from Amlierst {Falconer n.
1033) which agrees with it exactly, and Falconer n. 1033 ia tmn is precisely tlio same as a plant
(from the herbtuium of the late Ih. Elurz) cultivated in the Buitcnzorg garden. This plant is more
noai-ly related to G. javanicwn, with which it agrees (at the samo time differing fi'om 6. lueidum) in
liaving a glabrous style aad in having a calyx with long teeth nnd distinct pedicels. But tlioiigh
f r om these specimens it is evident that JFall. Cat. n. 2156/1 is specifically distinct from Cat. n. 215G/2,
I have refrained from restiicting tlie name G. liicidnm to "Wallich's n. 2156/1. For Falconer n. 1033
and Kurz n. 1104 difíei' from tme O. ja»anicum mainly in being densely tomentoso, and might
well be treated as a variety of that plant. But were this reduction made it would be necessary
t o submerge G. javanicum (1833) in G. icciditm (1828)—a com-se tkat would bo Hglily inexpedient;
for while it is troe that G. lueidum (1S28) did include Wall. Oat. u. 2156/1, it is equally ti-ue
that G. lueidum (1831) as described in Walt. Tl. As. Har. and i'l all subsequent icorl.s refers precisely to
n. 21-56/2 and explicitly eicltides n. 2156/1 by the character "dentes calyeini -iix lineam longi." In
the present paper therefore a figure has been given of, and the name G. lueidum is hereby expressly
limited to, the plant which foi-ms Wall. Cat. n. 21Ó6/2.
As in the case of G. Ecaneanum, the two varieties are very distinct, nnd yet there is no chai'aetar
of calyx or corolla wherein they difíer.
18. Gomphostemma phlomoides Benth. (1835).
E l a t a erecta caule paa-um sulcato, /••liis longlus petiolatis late ovatis apice acuminatis
b a s i abruptius attenuatis marinine crenato-dentatis, supra molliter breve Iiirsutis snbtus
p e t i o l i s caulibusque dense floccoso-tomentosis, verticillastris pauciñoris, hracteis majusculis
o v a t o - l a n c e o l a t i s acutis, bracteolis jninimis, càlyeis breve pedicollati dense tomentosi
d e u t i b u s triangularibus tubo duplo brcvioribus, corolla lutea antice incurva calyce duplo
l o n g i o r c , extus (lobo labii antici medio excepto) parce ffiquabiliter pubescente, styli
s u m m o tríente parce pilis diverg-entibus hirsuto, miculis immaturis nigrescentibus apice
p i l o s i s sublffivibu.s.—G. phlomoides Bcntk, Lah. Gen. et Sp. 649 (1835) et DO. Prodr.
x i i , 551 (1848); Walp., Rep. iii, 898 (184o) ; 3IorUzi, Syst. Vcrzeiclm. 53 (1846); Miq.,
Fior. Ind. Bat. ii, 985 (1856).—Prasium phlomoides Reintv. ex Blume, Bijdr. 840 (1826).
I n archipel. JfAL.A.TAìiO ; in ins. Java in sylvis montosis altioribus, Blume, Rcimoardt,
Eorsfield n. 136! Sasskarl n. 1597! in ins. Madura, apnd Petjceden, Teìjsmann a. 1814!
C a u l i b u s 60—100 cm. altis, petiolis 3'5—6 cm. longis, laminis 13—16 cm. lungis bis
8 — 1 1 cm. latis, verticillastris nunc densis iiunc i n ramos foliolatulos excrescent ibus, bracteis
\b—-¿5 nim. longia bis 3—5 mm. Litis, bracteolis 5—7 min. longis linearibus, pedicellis
2 — 3 mro. longis, calyce 12 mm. longo hoc 6 mra. lato—tubo calycia 8 nnn. longo,
deiitiLius 4 mm. longis bis basi 'Ò'O mm. latis, corolla 35 mm. longa—parte tubi acquali
LUCIDA. 26;>
1 3 mm, longa, limbo 0 mm. lato, labio postico integro 6 mm. longo hoc 8 nim. Iato,
l a b i o antico 9 mm. longo hoc 19 mm. lato, lobis lateralibus majusculis, nucuHs maturis
n o n d u m communicatis.
PLATE 104. Gomphostemma phlomoides Benth. Specimens from Java, JIonfiM.
This species is vary distinct from the others of the group in often having the whorls disposed ou
short axillary branchlets with smaller leaves than those of the main stem, In this respect therefoi-e
i t approaches the group Pedunculala. But its whorls ture axillai-y and not on the old wood, and it ia
further distinguished by its hii-sute ovary and ite calyx prominently ribbed. Even when the whorls oi-h
condensed it is at once distinguished from the remaining Lucida by its smaller corolla veiy abruptly
and considerably widened at the top; and from O. oblontjum, with which Mr. Bentham and Dr. Miquel
particulai'ly compare it, it is fm'ther distinguished by its sparingly hirsute style. Tho slander part of
the tube, though shorter than in any of the other Lucida, is not, as Miquel says, shorter' than the calyx.
The various foi-ms that ore hei'e included in the gi'oup Lucida have been in this paper kept apart in
order to maintoin as unifoi-mly as possible the ti'eatment adopted by the late Mr. Bentham. But
though the intei-mediate forms which would waiTaut a reduction of these forms have not as yet boon
reported, there is evidence enough to show that reductioiis \vill ultimately bo necessary, and it
does not seem out of place to indicate here those that may have to be efiected. Abandoning all
refcroace to diSerenees in the leaf edges and in the stj'le, which even in some of the forms now
differentiated aro found to fail, and relying solely on differences in flower and fr'uit, wo find that only
tlu'ee plants are at all clearly separable. These aa'e:—
A. G-. oniosGUM Wall., am^l.; corolla tubo calyce multo longioro, nuoulis matmis hii-sutis.
VAK. a. tifpica; foliis majusculis hand I-ugosis breve petiolatis = &. oblongum Wall. Cat. 2154.—
G. Philippinarum Benth.—G-. Scortechinii Prei"«,
VAR. parvi/olia ; foliis minoribua subi-ugosis vel I-ugosis distincte petiolatis = G. javanicum
Benth.—Q. rugosum Frain (G-. lueidum Wall, quoad Cat. n. 2156/1).
AREA GEOGR. ; Andamans, Tenasserim, Malaya, Java, Philippinaa.
B. G. LrciDUM Wall., etnend.; corolla tubo calyce multo longioro, nuculis maturis glabrii.
TAR. a. hjpica; foliis oblanceolatis vel lanceolatis = G. Incidum Wall. Cat. 215G/2 taulim.
Y a r . /3, canescens; foliis eliiptico-ovatis snbtus albo-tomentosis = G. lueidum v/
Frain uti supra.
AUEA GEOGE.; Himalaya oricntalis, Assam, Bui'ma superior, Timnan.
G.
Benth.; corolla tubo calycem parum excedente.
AREA
Java, Madui'a.
Even after the wholesale reduction indicated above has been effected, it will be noted—1, that
O. lueidum is manifestly only the Indo-Chinese representative of the Malayan G. oblongum ; and 2, that
the only tangible chavactei-which sepmates ' these two is thot the ovarian tomentum is altogether deciduous
in Q. lueidum, whereas it pei-sists to some extent in G. oblongum. The two species therefore do not
difier from each other so much as do G. pedunculaum and G. Curtidi (which two Mi'. Hemsley in liti.
has suggested the advisability of uniting), and, as in the case of these lattei', it may be suggested that
ultimately G. lueidum would bo quite sufficiently diScrontiated if regai-ded as a local sab-speoies of
G. oblongum, as has been indicated abeady in TABLE I (p. 230).
Y I I . Peduticulata ; verticillasfris rmotis azillaribus vel secus cauloin aphyll
calyce hand costato intus tuho dentibusquc glabro, ovario glabro, nuculis ovatis.
üiií