17 



varying in length including petiole, from i-JO inches ; the largest ones occurring 

 only near base of female plants, and usually not more than one or two on each 

 plant. .Flowers, Male in axillary racemes, 2-6 in each axil, 1-J 2 inches long, the 

 flowering branches usually terminating in a leafless compound raceme ; rachis 

 pilose, many flowered, flowers greenish white, pedicels short, bract ovate, acuminate, 

 "pubescent, 2-5 lines long, twice to thrice longer than flowers. Perianth 6 lobedj 

 lobes oblong-acuminate, in two rows, outer broader, ^tomentulose, inner narrower, 

 sub-glabrous. Perfect stamens ,3, with 3 staminodes, inserted on base of perianth 

 lobes, >and opposite to them filaments very short, anthers small, ovate, 2 celled, 

 staminodes sub-spathulate, larger than fertile stamens, opposite to outer lobes of 

 perianth. Ovary abortive. Female flowers in axillary spikes, 1-3 together, rachis 

 densely pilose, laxly flowered, lengthening in fruit, bract lanceolate, acuminate, 

 very much longer than perianth. Perianth 6 parted, lobes ovate, acute, strongly 

 inflexed at apex, 3 outer ones pubescent, inner ores glabrous, each enclosing a 

 minute barren anther at base. Ovary inferior, triangular, densely pilose, stigmas 

 3, recurved. Capsule oblong, broadly 3 winged, 3 celled, 1 inch long, 7-8 lines 

 wide. Seeds discoid, winged at one side, glabrous. 



Habitat: NATAL: Coast district. 



Drawn and described from specimens gathered near Durban, February, 1898. 



It will be noticed that in several points the description given above differs 

 from the description given by the author of the species in the Flora Capensis. 

 Tho leaflets ;n c there said to be ovate, and in Wood's No. 1618, one of which is in 

 the Government Herbarium, and from which gathering the species was probably 

 described, they are so. but in the variety figured here, they are narrow oblong; 

 both varieties are to be found near Durban, but there does not appear to be any 

 other appreciable difference between the two varieties. Then, in the Flora 

 Capensis, no mention is made of the staminodes, which are undoubtedly pre-ent in 

 both forms. We find that in the dried specimens these staminodes are difficult to 

 detect, but in the fresh plant they are quite conspicuous ; we find also that the 

 fresh flowers are visited by a minute insect, which seems quickly to devour these 

 delicate organs. The genus Dioscorea contains about 150 species, of which 10 or 

 1 2 are natives of Natal ; many of the species are cultivated for their esculent 

 tubers, and ,'ire popularly known as Yams, and in most tropical countries they 

 form an important article of diet for all classes, and some of the varieties are 

 much esteemed. The Natal species, so far as we are aware, have never been 

 exploited in this direction, and as the juice of the tubers of this species produces 

 itching when applied to the skin, it is not likely that they will be found to be of 

 much value as a culinary vegetable. A rough experiment which we have made 

 with them shows that they contain little if any starch. 



Fig. 1 , Branch of male plant with flowers and leaves. 



2, Branch of female plant, with flowers, leaves, and young fruit. 



MALE PLANT. FEMALE PLANT. 



Fig. 3, Flower. Fig. 13, Flower with bract. 



4, Flower with bract removed. 14, Outer lobe of perianth with 



5, Outer perianth lobe with anther abortive stamen. 



6, Inner perianth lobe with stain- 1 5, do. side view. 



inode. 16, Inner lobe of perianth, front 



7, Bract. view. 



