145 
extension and growth of the cellulose membrane. The conidium is now 
attached to the basidium by a mass of callose in the form of a little cup 
embracing the slightly pyriform base of the conidium. The base of 
this cup is convex or plane, but the center often shows a little pit which 
is the last vestige of the previous funnel-form orifice. 
No division of the connecting cupule into three layers, as describedby 
Zalew'ski, was observed. At this stage it is pure callose. Soon the 
cupule contracts, its superior edges being reduced progressively, and 
it shortly takes on the form of a cylindric fragment uniting the conidia, 
but the cellulose membrane of the conidium or of the basidium is not 
yet continuous across the callose. Subsquently the cellulose wall of 
the upper part of the basidium is continued along the base of the cal¬ 
lose plug or through it when the latter projects. A similar process takes 
place a little later at the lower end of the conidium when the cup form 
has almost disappeared. Sometimes this cross wall is outside of the 
callose band, sometimes it grows through it, imprisoning a portion 
within the conidium. It is generally only when the conidium is second 
in rank from the basidium that the cellulose membrane is completed. 
Up to this point the changes in the callose band have been due to ab¬ 
sorption, but not so subsequently. The band of callose now changes 
chemically so as to become strongly hygroscopic and completely solu¬ 
ble in water or even in the vapor of water. This primary septum or 
connective band contains no pectin compounds and does not swell and 
become gelatinous previous to solution, as stated by de Bary and Za- 
lewski. It is simply a very neat case of liquefaction. 
New conidia are developed under the old ones in the following man¬ 
ner : The end of the basidium elongates by intercalary growth and a 
new ring of callose appears. This is not always in the same plane, but most 
often for each conidium or group of conidia it appears in a region nearer 
the summit, so that the lateral wall of the basidium presents a series of 
thickenings, which when stained appear as stride. Each one of these 
striae corresponds to a group of conidia, for they are always less numer¬ 
ous than the conidia successively developed from a basidium. 
The formation of conidia finally ceases, and in old sori, long ruptured, 
it is easy to find such exhausted basidia drawn out to a naked point or 
terminated by a single conidium which appears to be incapable of 
completing its development. The membrane of the basidium is then 
notably thickened in the terminal region and more or less deformed. 
The strim above mentioned are often visible and, finally, several rows of 
internal button-shaped thickenings are almost always present. These 
thickenings are composed either of pure callose or of a mixture of callose' 
and cellulose. 
The statements here given were also found to hold good for Cystopus 
cubicus and the closely related C. spinulosus. A somewhat similar 
method of growth and delimitation was studied in a form of Plasmo - 
para found on Epilobium montanum. Here, however, the cellulose wall 
