the ' Crown-Gall ' Fungus of Lucerne 363 



and the branching cavity formed by the invaded cells may be of 

 any shape. Once the wall is thickened the fungus never grows 

 through into the adjoining cells, and cell division ceases in the 

 tissues surrounding the older parts of the cavity. It is sometimes 

 observed that a portion of the host tissue is completely isolated by 

 the fungus, and slowly dissolved without actual invasion of the 

 cells composing it. In sectioned material it is not at all easy to 

 make out the method of growth of the hyphae or of the resting 

 spores, but if young and actively growing galls are dissected out 

 and the tissues stained after fixing in bulk, it is possible to re- 

 construct the fungus thallus with great certainty. 



Development of the fungus. 



Figs. 2 to 7 are camera-lucida sketches from preparations 

 made in this way. The youngest hyphae {Hy, figs. 3 and 5) are seen 

 to possess a very narrow lumen and thin wall, their diameter being 

 about -5^. They are terminated by a swollen portion (C, figs. 2, 5 

 and 6) containing rather dense protoplasm, and at first one nucleus 

 only. This swollen portion will be referred to as the 'collecting 

 cell,' since similar terms have been employed by other writers for 

 analogous structures. 



The extreme end of each of these collecting cells develops a 

 short, very dehcate and much branched process, which is considered 

 by Jones and Drechsler to have an absorptive function {Ha, figs. 2 

 and 7). It could never be determined witli certainty whether this 

 process was a branching hypha or an outgrowth of the wall only: 

 a similar structure is described by Biisgen* for CladocJiytrium 

 Butomi, and by Schroeterf for Physoderma ( = Urophlyctis) pulposa. 



The collecting cell increases in size, developing from 10 to 

 15 nuclei, until it is about 10/x in thickness. Details of nuclear 

 ■division have not been made out, the resting nucleus shows one 

 ■deeply staining mass of chromatinic material, but very httle normal 

 reticulum. Fine cross walls are then laid down, obhque to the axis 

 of the collecting cell, cutting off 2 to 4 uninucleate masses of 

 protoplasm, peripheral in position, from a central, multinucleate 

 portion. The former give rise to branch hyphae, the latter to one 

 resting spore; in each case by a process of prohferation. From each 

 of the peripheral cells a papilla arises, the end of which enlarges; 

 into this the single nucleus and contents of the cell are passed; it 

 rapidly elongates to form another hypha of hmited growth exactly 

 like the one first considered. The resting spore arises from the 

 centre of the apical haustorial process, as a rounded cell on a short 

 stalk, simultaneously with the branch hyphae. Into this cell the 



* Biisgen, M. (1887). Beitrag zur Kenntniss der Cladochytrien. Cohn's Beitr. 

 Biol. Pflanzen, Bd 4, pi. 15. 



t Schroeter, J. (1882). Bot. Centbl. Bd 11, Nos. 31, 32, p. 219. 



