95 



The lowest roots, J, c, d, e, rise out of the same opening, and first 

 the foot-knot, /, appears : it is a circle of a few transparent cells, 

 out of which other roots, g, h, grow. The portion of the stem 

 between this knot and the lower roots consists of one short cell, 

 forming the lowest part of the stem. 



The plant so far, being colourless and transparent, shows the cir- 

 culation in perfection, as indicated by the arrows. 



All the succeeding portions of the plant are elongations from the 

 head, i, but should that be injured, a second or a third stem may 

 be produced from the upper foot-knot, /, but none from the seed, 

 leading therefore to the conclusion that it is a single seed and not a 

 seed-vessel. Indeed, the whole contents of the seed appear to be ab- 

 sorbed by the production of a single plant, and the skin falls away 

 empty. 



Fig. 5 is a plant about a week old ; fig. 6 the top or head, i, of one 

 a few weeks older : its several circulations are shown by arrows. As 

 the growth proceeds one of the head-sprouts becomes larger than the 

 rest, and the plant thus far consists of single tubes. But from the 

 midst of these head-sprouts the succeeding portions of the plant arise, 

 having a covering of smaller tubes : b* is a continuation of the largest 

 root, b, to show five secondary roots, j, branching from it. 



Fig. 7. A portion of the plant of the natural size. Every knot may 

 produce roots, but it is remarkable that they always proceed from 

 the upper surface of the knot and then turn downwards, so that it is 

 not peculiar that the first roots also should rise upwards with the plant 

 and come out at the top of the seed-skin, and then turn downwards. 



Stems precisely like the original one may be produced from any of 

 the knots; the only projections being the similar number of arms 

 at each knot, all of which are liable to bear fruit. 



That the stem and aims are tubular, and entirely covered with 

 smaller tubes, and that the circulation can mostly be observed in 

 these tubes, is well known. But the structure of this covering of 

 tubes, and the peculiar mode and order of their development, are the 

 circumstances deserving more particular attention. 



The number of fruit-bearing arms in each cluster is variable, seven, 

 eight, but most frequently nine are found, and as I have occasionally 

 found ten, I cannot correctly state the number due to a complete 

 plant. 



Nine blossoms and seeds generally grow on each arm, but I have 

 sometimes had the plant with two globules and two seeds at each 



K 2 



