20 



to the petals of the outer row. Ovary inferior with its base sunk in the 

 rachis, ellipsoidal ; calyculus short and notched. Style very stout, shorter 

 than the petals. Stigna subcapitate, 3-lobed and oblique. Fruit \ inch 

 long, obovate or ellipsoidal, glabrous, yellowish when ripe, pulpy ; endocarp 

 membranaceous with 4-6 longitudinal ridges and covered with viscous sub- 

 stance. Seed ellipsoidal, smooth ; embryo straight, cylindrical in the axis 

 of a fleshy albumen. 



This is the plant first noticed by Franchet and Savatier in their 

 Enumeratio Plantarum Japonicarum Vol. II. under the name of Loranthus 

 (?) Tanakae. Since that time no details have been published about 

 this plant from the lack ol good specimens. 



In the summer (June 9th) of 1896, I was fortunate enough to collect 

 specimens of this rare parasite in flower at Kawamatamura, a village 

 situated at the foot of the northern slope of Kanayatoge near Yumoto in 

 Nikko, and obtained its fruit in the winter of the same year from the 

 same district. 



This species is usually found parasitic on the branches of Quercus 

 grosseserrata and Castanea vulgaris often in company with the common 

 Viscum album ; but the mode of growth of its root is very characteristic, 

 and is entirely different from that of the latter species. The primary root 

 of Loranthus Tanafcce sends out many lateral roots which push their way 

 through the cambium and the young wood of the branch of the host 

 plant, and behave exactly in the same manner as those of Loranthus 

 europaens. 



From the outer surface of the lateral roots there often arise many 

 arms which grow towards the cortex along the medullary rays and serve as 

 organs for taking nourishment from the surrounding tissues, and on the 

 surface of the roots we find a zigzag outline which is also the case with 

 Loranthus europeans. 



This outline is produced by the successive bending of the growing apex 

 of the root towards outside (caused by the lignification and hardening of 

 the wood fibres through which it grows) and its further growth in the 

 softer tissues of the newly formed wood. 



The number of the angles of the zigzag formed in the wood layer of 

 one year's growth varies from 4-6. 



Besides this outward bending the apex of the root presents a lateral 

 bending probably caused by the presence of a large medullary ray which 

 seems to supply nourishment from the cortex of the host to the root. 

 This shows a kind of chemotropism within the wood of the host. 



