56 
information about the distribution of the species. I am fully 
conscious of the imperfectionof my account in the ‘ Madras Flora,’ 
and I may very likely have made too many species, but | have 
tried to study and do my best with the material available. 
LytHracEar.—In this family I have adhered as nearly as pos- 
sible to the arrangement adopted in Koehne’s Monograph 
(Engler’s Pflanzenreich), more especially as regards the sub- 
division of Ammannia (Genera Plantarum), into the three genera 
—Rotala, Ammannia and Nesaea. The only difficulty I have found 
is in Rotala, in which genus it is not at all easy to separate the 
species /?. leptopetala, Koehne, and R. densiflora, Koehne. This 
has been fully pointed out by Blatter and Haltberg in a paper in 
the ‘ Journal of the Bombay Natural History Society, xxv 701,’ 
in which they finally settle to combine the two species in one, Ff. 
pentandra, Blatt. and Hallb. I should, however, think that if 
their view is adopted, the oldest name would be /?. densiflora. So 
far as the Madras material is concerned, I have succeeded, with 
Lagerstroemia. For the reasons given by Sir G. King in his 
‘Malay Materials’ and Dr. T. Cooke in his ‘ Flora of Bombay,’ 
I have adhered to the well-known name L. Flos-Reginae for the 
splendid flowering tree so much grown in India, instead of» L. 
speciosa, Pers. e two species, L. Thomsonii, Koehne, and L. 
kottleri, C. B.-Clarke, are apparently very little known and 
deserve to be carefully looked for by local botanists and forest 
officers in Madras. 
spermum being quite distinct, in a family of its own, Cochlo- 
spermaceae 
Casearia. I cannot agree with Sir D. Brandis (‘ Indian Trees,’ 
p. 343) in combining C. glomerata, Roxb., and C, graveolens, 
Dalz. The latter is a very widespread small tree of low levels in 
Northern India, deciduous and with its leav 
they fall, and is, I think, very distinct from the t 
evergreen, C. glomerata of the Sikkim forests, about 
. 
ense moist forests of the W. Ghats at 2000-3000 ft. 
‘ gree with C. esculenta, Roxb., 
as Clarke and Bourdillon have identified it, nor entirely, though 
ss with C. rubescens, Dalz. There are two specimens in the 
w Herbarium, collected by Bourdillon—(1) No. 181 from ever- 
