the Mycorrhiza of the Marat tiaceae. 85 
either in a terminal (Text-fig. 7), or more often in an intercalary (Text- 
fig. 8) position on the hyphae, and were at first mistaken for ripe oospores, 
to which they bear a remarkably close resemblance, especially during the 
earlier stages of their development. However, the complete absence both 
of an oogonial wall and of antheridia is sufficient proof of their non- 
sexual nature. 
a. Development of the resting spores. 
A terminal resting spore originates as a more or less spherical enlarge- 
ment of the terminal portion of an ordinary young hypha, from which 
it is usually delimited by a septum (Text-fig. 8, i, ii, iii). The dense granu- 
lar contents of this enlargement are multinucleate. At this stage the 
immature resting spore closely resembles a young spherical vesicle, from 
which, however, it can readily be distinguished by its denser contents and by 
the uniform thickness of its wall (cf. Text-fig. 8, i, ii, with Text-fig. 6, a). 
The intercalary resting spores (= chlamydospores) are formed in the 
following manner : Swelling takes place in the course of a young hypha, 
either equally all round, in which case globular or ovoid enlargements 
are produced, or in one direction only, in which case the swollen portion has 
a sub-globose form. In either case the enlargements may or may not 
be delimited by septa from the supporting hyphae (Text-fig. 7, i). The 
wall of these swellings, whether terminal or intercalary, is at first similar in 
thickness to that of the hyphae from which they arise, but differs slightly in 
its staining capacity. For example, the wall of the young spore stains 
readily with Kleinenberg’s haematoxylin, whereas the hyphal walls take up 
this stain with difficulty. Very early in the development of the resting 
spore the middle wall originates as a thin highly refractive layer, in which 
irregular deep yellow areas can easily be distinguished (Text-figs. 7, ii, and 
8, iv, v). At a slightly later stage the first indications of canals or pores can 
be identified in this wall (Text-figs. 7, ii, and 8, iv, v). The thickening 
of the middle wall proceeds apace : meanwhile the contents of the spore 
become more densely granular, and oil globules, 1 which are so conspicuous in 
the contents of the ripe spore, appear in the cytoplasm and completely mask 
the subsequent behaviour of the nuclei (Text-figs. 7, iii, iv, and 8, vi, vii). 
So far the only indication of the innermost wall is the closer aggrega- 
tion of the granular contents in the peripheral zone of the cytoplasm (Text- 
figs. 7, iii, iv, vi, and 8, vi, vii). A number of spores which were artificially 
crushed at this stage showed no signs of an inner wall ; hence it must be con- 
cluded that the innermost wall is laid down very late in the development of 
the spore. Thickening of the middle wall continues until an average thick- 
ness of 6 /x is attained, but shortly before this result is arrived at, the inner 
wall begins to develop. 
1 It may be noted in passing that the presence of oil as a reserve material is characteristic of 
‘ resting ’ bodies. 
