86 Sahni. — On an Australian Specimen of Clepsydropsis . 
at which they stand to the vertical it maybe concluded that the leaves were 
probably almost erect in their basal region. These facts make it very 
probable that the plant was a fair-sized tree-fern resembling the fossil 
Osmundaceae. 
Anatomy . The preservation on the whole cannot be described as 
good, although some portions show well-preserved details of structure. 
Many of the roots have penetrated deep into the cortex of the leaf-stalks, 
and have even distorted the shapes of most of the petiolar strands. 
The roots probably all belong to the Clepsydropsis itself, although none 
of them is seen actually connected with the leaves— they may be arising 
from the more proximal regions of the latter and in part even directly from 
the stem, as appears to be the case in the Barra ba specimen, and as Stenzel 
( 1889 ) describes in C. kirgisica (p. 22, PI. IV, Fig. 38, w). They have 
a diarch xylem strand (PI. IV, Fig. 1) and the pitting of the metaxylem 
tracheides was finely scalariform as in the leaf-strand. The cortex of the 
petiole consists of a broad inner zone of large thin-walled cells with inter- 
spaces, succeeded on the outside by a narrower zone of cells gradually 
decreasing in diameter towards the epidermis (PI. IV, Fig. 2). In longitu- 
dinal section the cells of the inner cortex are nearly isodiametric, those of 
the outer cortex more elongated, with transverse or oblique end-walls 
(PI. IV, Fig. 3). The epidermis is not preserved, consequently the ‘ external 
glands or hairs’ mentioned by Mrs. Osborn are not seen. A few of the 
outermost layers of the cortex appear to be specialized as a thickened 
hypodermis , as described by Schenk in Rachiopteris L udwigii ( 1889 , p. 554 ) > 
but it is difficult to be clear on this point. The total radial thickness of the 
preserved cortex varies from 3 to 4 mm. ; between it and the petiolar strand 
the leaf tissues are nowhere preserved. In Schenk’s fossil just referred to 
this space was in part at least (loc. cit ., p. 553) occupied by a sclerotic sheath. 
The adjoining figures, from camera-lucida sketches of the petiolar 
strands, illustrate the extent to which the intruded roots (shown in rough 
outline only to indicate their positions) distort the normal clepsydroid 
shape of the leaf-traces. Only one of these happens to be completely 
undisturbed (Text-fig. 2, 1, 1). It was seen cut at two different levels (T 1, 
T 2) and its dimensions are given below : 
Total length, 6*4 mm. 
Thickness at the ‘ waist ’, 3 mm. 
Length of each peripheral loop, 1*7 mm. 
Distance between the inner ends of the two loops, 2*1 mm. 
The variations in size could be judged from the remaining figures, but 
the amount of distortion makes it unsafe to base any measurements upon 
them. The limits of the xylem could not in all cases be accurately 
observed, owing to imperfect preservation. In many places only rows of 
dirt-granules are seen, which, however, unmistakably marked the contours 
