115 



The perianth of the female flower in H. hnmliensis is of a 

 dull mnstarcl yellow, and if normal consists of five ovate acuminate 

 sepals coherent in tlieir lower third into a cup, which is of a 

 greener tint inside than the lobes. This cup is almost filled by the 

 ovary, around the base of which is a slight circular swelling being 

 the disc which may be just damp with honey; five slight thicken- 

 ings extend up the cup as the midribs of each part of it. In H. 

 eonfusa, the perianth lobes are ovate and blunt, and the cup ex- 

 tends to half their length; they and the cup are straw-coloured 

 with a magenta line down the middle from the tip or near it to 

 the very base inside. Outside they are covered with short hair. 

 The top of the ovary is conspicuously blunt with sessile stigmas. 



Often in H. brasiliensis the first flower of a panicle to open 

 is a male flower ; but after all the female flowers are over, there are 

 males that follow in considerable numbers. The male of H. 

 brasiliensis is like tlie female in perianth, but smaller and hardly 

 pale green within the cup. The staminal column carries two rings 

 of five anthers. The pose of the flower depends upon the axis 

 which bears it, and it may face in any direction. The male flowers 

 of H. confusa, like its female flowers, are smaller than those of 

 H. brasiliensis, blunter, different in colour, being straw-coloured: 

 they have fewer anthers, and by the bending of their pedicels they 

 face more or less earthwards. Outside they are hairy. This bend- 

 ing of the pedicels gives a very good distinguishing mark which the 

 herbarium student cannot note so well as the field student. 



The seeds are as figured by Hemsley in Hooker's I cones 

 Plantarnm, plate 2575 : but tree No. 1844 gave flowers with blunter 

 perianth-lobes than the figures on plate 2574. 



Hevea panciflora is known to produce h3'brids with //. brasi- 

 liensis, and so far it seems that these hybrids have no value. 



A sample of rubber from the tree was submitted to Dr. 

 Frankland Dent. Government Analyst, Straits Settlements, and 

 another to the Director of Agriculture, F. M. S., for kind submis- 

 sion to Mr. B. J. Eaton, Agricultural Chemist in the Department 

 of Agriculture. These two samples on analysis scarcel}' differed : 

 they contained about 95 per cent of a substance chemically rubber 

 but lacking the physical properties cequired in commercial rubber, 

 probably as Mr. Eaton suggested a polymer of caoutchouc; and 

 they contained also rather under 2 per cent of resins. The samples 

 were too small for a vulcanization test. They were small l)eeause 

 the tree yielded so grudgingly. 



I. H. BURKILL. 



MANGO PESTS IN SINGAPORE. 



In no other part of the tropics are mangoes more badly i^est- 

 ridden than in Singapore. Locally produced fruit is therefore 

 neither abundant nor of good quality. There is nothing in Singa- 

 pore to compare with the great quantities of fine " carabao " man- 



