312 
0-75 mm. diametro, glabrum, olivaceum. Stylus 3 cm. longus; 
stigmata obovato-oblonga. Ovula pro loculo biseriata, circiter 
20 pro serie. Capsula (immatura tantum visa) circiter 3 dm. 
West Inpizs. Trinidad: St. Clair; Knapp Hill, Aug. 1911, 
Broadway 4164 (type); and without precise locality, Fendler 517. 
Petastoma Broadway? is the first species of Petastoma recorded 
from the West Indies. It agrees with P. leucopogon in the 
general shape of the calyx, but differs in most other respects, 
and does not appear to be nearly related to any of the species 
hitherto described. 
XXXIX. THE FLORA OF MADRAS: IV. 
The fourth part (pp. 57 rel Rubiaceae to Sapotaceae, and 
‘part of the Ebenaceae of the Flora of Madras, has now been 
published, and the following notes have been drafted by Mr. 
Gamble. We are indebted to him for the opportunity of placing 
these on record for the information of workers in herbaria, 
relative to conclusions with regard to particular species dealt 
with in the work that differ from those arrived at by earlier 
writers. 
hae Nores oN THE Fiona or Mapras.* 
% J. S. GAMBLE. 
' RuBiacEaE. Oldenlandia. I have thought it best to follow 
other botanists and especially K. Schumann in the “ Natiirliche 
Pflanzenfamilien ’’ in placing Hedyotis under Oldenlandia owing 
to the ies in finding characters to keep them separate. 
Mussaenda. The rearrangement of the species of this genus 
follows that which Mr. J. Hutchinson has made in the Kew 
Herbarium. Mr. Hutchinson kindly placed at my disposal his 
draft of a Key to the 8. Indian and Ceylon species and this has 
been followed in the main. 
Randia. R. dumetorum, Lamk. As a Forest Officer in 
India I was much puzzled over this species especially after the 
appearance of Part vii (Vol. iii) of the ‘‘ Flora of British India.” 
The first form of it whose acquaintance I made about 1873 
was a small tree very common in the Sal forests and mixed 
forests of the Darjeeling Terai. It had thin obovate leaves 
often 4 to 5 in. long and an apple-like fruit used by the Mechis 
and Garos and other forest residents for poisoning fish. In 
forest shade, the spines were usually small: and thin, but 
I remember that they were longer and thicker when the tree 
was found in open places. I have seen the same form in the 
* Previous notes were erie in the Bulletin for 1916, p. 57, for 
1918, p. 222, and for 1920, p. 
