OF BUCKLEYA QUADEIALA, B. ET H. 31 



whatever in its structure from other parts behind it, as I have 

 already pointed out {cf. Sec. IV). During the secondary growth, 

 the cambium of the sucker, after having joined with that of the 

 host, produces new elements in the same direction and nearly 

 with the same activity as the cambium of the host produces its 

 own elements, and so the sucker is able to expand more and 

 more laterally. Hence, after the obliteration of the primarily 

 formed cortical tissue of the host-root, which at first had enveloped 

 the lateral sides of the sucker, the parallel growth in thickness 

 of both haustorium and host no longer allows the cortex of the 

 host to grow over the sucker, or the cortex of the haustorium 

 to overlap the host, as it was the case in the primary growth 

 (Figs. 4, 9, 10). 



In this way the contact surface, which is at first wavy and 

 irregular, especially when the folds are numerous, becomes simple, 

 the greater part of the surface being now occupied by the front 

 of the sucker (Figs. 9, 10). At the same time the sucker, which 

 we have conveniently distinguished as such in the primary growth 

 by its being imbedded in the host, can no longer be distinctly 

 separated from the other part in such an old haustorium. 



Together with all these changes w T e see that on the whole 

 surface of the haustorium, excepting however a small portion of 

 its apex, the corky layer is formed successively under the super- 

 ficial layer of the cortical parenchyma, which is torn off sooner 

 or later from the haustorium, while new cortical parenchyma is 

 constantly produced by the cambium. During this process the 

 striated band, which is originally situated along the median portion 

 of the cortex, changes its position gradually towards the periphery, 

 until finally it is cast off together with the corky layer and the 

 cortical parenchyma. 



