CISSUS-OPUNTIA. 29 



dry foresummer. It was now set in a sand-bed to allow it to take up a 

 renewed usual course of growth. 



The older parts of the roots showed crushed epidermal cells into which 

 some fungal hyphae had entered. Four or five layers of phelloderm were 

 distinguishable, arising from a phellogen in the outer cortex. The cortex 

 was everywhere loaded with starch. 



December 5, 1908. —A second one of the original preparations which was 

 successful was dissected. Only two roots had been formed, and the tips of 

 these were pressed into the pale-greenish tissue, the older parts being 

 brownish. The relatively small absorbent capacity developed by this plant 

 was correlated with an accentuated desiccation, which had killed the ter- 

 minal part of the insertion. 



The slips of Cissus placed in sand-beds came into proper condition during 

 November and were used in making preparations in the glass-house and in 

 the open during this month. But little noticeable activity was displayed, 

 however, until late in March. This and other tests gave definite evidence 

 that the most advantageous time for making successful preparations was 

 in the few weeks just previous to the beginning of the season of vegetative 

 activity of both host and parasite. Arrangements made in the earlier part 

 of the vegetative season were also successful. (Plate 5, A.) 



March 30, 1909. — One Cisszis taken from the dark-room and which had 

 a greened branch now displayed a broad leaf on one of the middle inter- 

 nodes. 



April 19, i9(99.— A third cutting of the trio which had survived from 

 the original operation was now dissected. The five internodes of this 

 cutting which had been exposed during the dry foresummer of 1908 had 

 flagged and been emptied of their contents, and the bud at the lower- 

 most internode had begun to develop. The submerged portion had under- 

 gone swelling into a tuberous formation, filling a cavity with a diameter 

 much greater than that originally made for it. Two large roots had been 

 formed, one of which had branched. One main root was 21 mm. in length 

 and penetrated the tissues directly by a median course to a point near the 

 center of the joint of the host. The second root was 32 mm. long and had 

 penetrated the tissues of the host in a direction more nearly horizontal, 

 giving off a branch which had attained a length of several millimeters. 



A very vigorous penetration of the tissues had been made, the anatom- 

 ical relations with the host being much the same as those described above. 

 The decorticating parts of the roots were coated with the brownish sub- 

 stance formed from the Optcntia, while the clear, actively-growing tissues 

 of the root were pressed against the live, green, turgid tissues of the host. 

 The portion of the roots capable of absorption was very short, perhaps not 

 much more than 5 mm. long, yet come material must have been acquired 

 by its aid. A similar portion of the stem of Cissus unattached would have 

 desiccated within a few days. This one had not merely maintained its 

 veg-etative activity, but had formed a swollen tuber at the base of the 

 stem. The slow fiagging' of the terminal internodes would have allowed 

 the withdrawal of some of the material to this tuber, but during the 14 

 months of the existence of the arrangement the presvimptive evidence 

 that much was obtained from the host is very strong. (Plate 5, B.^ 



