360 Takeda. — A Theory of ‘ Transfusion-tissue \ 
or is distributed more or less scattered round the bundle, and often even 
completely surrounds the latter, and occasionally comes into contact with 
the centripetal xylem, if any elements of this latter tissue are present at 
all. Worsdell ( 14 , p. 316) maintains that the £ transfusion-tissue’ is centri- 
petally developed in the cotyledons of Ginkgo and Cycas. But in these 
cases centripetal xylem is considerably more developed than the centrifugal 
part, so that it obscures tracing the course of development. Recent re- 
searches on the Conifers by Miss Carter ( 3 ), and on Welwitschia by me ( 13 ), 
clearly show that there is no room for doubt that the tissue is first formed 
not centripetally but laterally. Worsdell ( 1 . c., p. 318) also supports his 
hypothesis ‘ by the presence of the transfusion-tissue on the ventral side of 
the xylem ’, and ‘ by the very frequent extension of the lateral transfusion- 
tissue tozvards the ventral side of the bundle ’. This seems to me to have 
no value as a support of his view of an intimate relation between the ‘ trans- 
fusion-tissue ’ and centripetal xylem, since this presence and extension is 
not a starting-point but a terminal portion of the tissue in question. It 
is true that in Cycas the ‘transfusion-tissue’ comes into contact with the 
centripetal xylem, which is rather to be expected, for in this case the centri- 
fugal xylem is exceedingly feebly developed, and the function of the xylem 
is mostly carried out by the centripetal xylem. Connexion between £ trans- 
fusion-tissue ’ and xylem of the bundle is usually to be seen, even in Wel- 
witschia :, where the sheath of ‘ transfusion-tracheides ’ is generally separated 
from the vascular bundle, for one occasionally sees connexion on the lateral 
side. This does not favour the hypothesis that the tissue under con- 
sideration is a modification of centripetal xylem, for this connexion 
simply indicates the mutual function of these two tissues, which will be 
discussed later. 
Before we pass on to a consideration of the function of ‘ transfusion- 
tissue’, I will just remark here that most of the later-formed elements of this 
tissue owe their origin to the mesophyll-parenchyma, or, in case the vascular 
bundles are surrounded by an endodermis, to pericyclic cells, which we can 
clearly see by comparing the size, shape, and position of * transfusion- 
tracheides ’ with the adjacent cells. The elements also do not arise in 
a definite order, but in an irregular way, as one often sees immature tracheides 
mingled with full-grown ones. 
The function of the ‘ transfusion-tissue ’ has been discussed by those 
botanists who have dealt with this tissue. The view which is most 
universally accepted is, however, that the tissue in question is to conduct 
water, serving as an auxiliary conducting system. The functional 
character is also sometimes made use of in supporting the supposition that 
this tissue represents a modification of the centripetal xylem (cf. 14 , p. 316). 
It is difficult to accept this assumption in the case of Welwitschia , for the 
leaf of this plant is both longitudinally and transversely traversed by larger 
