743 
Hemigymnia . j clvii. gramine^ (Stapf). 
reduced to clusters of 6-3-spikelets ; rhachis finely filiform, wavy, 
scabrid ; pedicels very fine with, discoid tips, scaberulous, J-l lin., 
or sometimes up to 2 lin. long and then flexuous. Spikelets ovate- 
lanceolate to linear- oblong, acute or acuminate, 14-lf lin. long, quite 
glabrous (in the African specimens). Glumes similar, ovate to ovate- 
lanceolate, acute, lower one-third to two-fifths the length of the 
spikelet, 3-nerved ; upper slightly longer, 5- (rarely 7 -) nerved. 
Lower floret : valve 7-nerved, nerves distinct. Upper floret pale, 
smooth ; anthers 1 mm. long. Grain oblong-elliptic, obtuse, sides 
slightly thickened. — Panicum arnottianum, Nees ex Steud. Syn. PI. 
Glum. i. 59. P. nodosum , Pranch. Contr. FI. Congo Fran 9 . in Bull. 
Soc. Hist. Nat. Autun, viii. 346 (reprint, 38) ; Hook. f. FI. Brit. Ind. 
vii. 43, and in Trim. Handb. FI. Ceyl. 145 ; not of Kunth. P. mala- 
baricum , Merrill in Philipp. Journ. Sc. iv. 248. 
Lower Guinea. French Congo : Bangui, Chevalier, 10956 ! Brazzaville, in 
woods along the Diowe river, Thollon, 4087 ! Mpila plain, Thollon, 386. Belgian 
Congo : Stanley Pool Distr, ; Wombali, Vanderyst, 4518 ! 4519 ! Kibambeli, 
Vanderyst, 4153 ! 
Widely distributed in the Indo-Malayan region, as far as the Philippines. 
Apparently very variable as to size and lignilication of the lower parts of the 
culms. All the African specimens examined have glabrous spikelets and leaves 
{except along the margins of the sheaths), whilst in the Indian material the 
glumes are frequently ciliolate and the leaves hairy. 
In the Philippine Journal of Science, iv. 248, Merrill identified this plant with 
Poa malabarica, Linn. , and he named it consequently Panicum malabaricum. This 
was done on Munro’s authority who had referred the original of Poa malabarica 
in the Linnean Herbarium to Panicum nodosum, and on the assumed identity of 
certain Philippine specimens with the plant figured in Rheede, Hort. Malab. 
xii. t. 45, and quoted by Linnaeus with his original description of Poa malabarica. 
The comparison of Panicum nodosum (sensu lat. auct.) with the plate quoted — 
Merrill saw only a rough tracing of it — and the accompanying text shows, 
however, at a glance that the two plants are not identical, even generically. 
Linnaeus must have come to the same conclusion ; for in the second edition 
of his Species Plantarum the reference to Rheede is wanting. What Rheede’s 
plate represents may be doubtful, but I would suggest that it was drawn from 
a specimen of Diplachhe fusca. However, Linnaeus’ quotation is of little im- 
portance in this instance. His description was not made from it but from a 
specimen in his collection. This is still in excellent preservation and is evidently 
a member of the “ Panicum nodosum ” group but characterised by its small 
spikelets. No locality or collector is quoted, but from the two facts 
(1) that Osbeck in his Voyage to China and the East Indies, ii. 29 and 343, 
quotes Poa malabarica and (2) that the Linnean specimen is a perfect match of 
specimens collected since in Hongkong (Happy Valley Woods, Wilford, no. 254) 
we may safely assume that Linnaeus had the grass from Osbeck who collected it 
near Canton on Dec. 17, 1751. The observation in Species Plantarum ed. i. 
“ Habitat in Indiae arenosis ” may be explained either by the loose way in which 
the term India was. used in those days or as a consequence of Linnaeus’ mistaken 
identification of the plant of Rheede who says of it (p. 83) “arenoso gaudens solo.” 
71. ICHNANTHUS, P. Beaiiv. Agrost. 56, t. xii. fig. i. 
Spikelets lanceolate to oblong or elliptic, acute or acuminate, more 
rarely obtuse, slightly laterally compressed, rounded on the back and 
