an archaic type of Seed from the Palaeozoic Rocks. 87 
trains of satellites — are cut obliquely more or less in the tangential plane, 
whilst three are cut more nearly at right angles. It is evident that 
geometric symmetry is lost. 
Again, in PL V, Fig. 4, the positions of the right and left longitudinally 
cut arms are quite asymmetrical, whilst the sections of four other tentacles 
above the pollen-chamber (t 1 . . . . / 4 ) show these also to have been in 
disorder. In contrast to these, others are met with in which the arms 
preserved a conical posture above the pollen-chamber, as in Dr. Scott’s 
section S. 1753, W.C. 1440, and a beautiful preparation in Mr. Watson’s 
Collection, prepared with his own hands (Text-fig. 7, p. 92). 
The matter need not be pursued further : it is evident the arms 
afforded adequate protection for the pollen-chamber after pollination had 
taken place, and it may be conjectured that at an earlier stage they had 
diverged, permitting access to its orifice. Of this, however, no direct 
evidence has been found, as almost all our seeds appear to be, approximately, 
at the same stage of development. 
1. The Epidermis. 
The epidermal system of Physostoma merits detailed consideration on 
account of the remarkable tubular prolongations into which its cells are 
drawn out along the convex, outwardly directed faces of the tentacles and 
ribs — a feature which gives to almost every section of the seed a unique 
and beautiful appearance. 
If a transverse section of a tentacle be examined at the level of the 
pollen-chamber, such, for instance, as /. in PI. V, Fig. 7, the flat, centrally- 
directed face is found to be clothed with a close-fitting layer of small, 
almost cubical, cells having a radial diameter of about 30 /x. 
Passing round to the flanks, the height of the cells increases. Many 
of the cells of intermediate height interlock with corresponding cells on the 
adjacent tentacles (cf. PI. V, Fig. 7), thus giving the ring of tentacles in this 
region much of the character of a united tube. That real fusion is lacking 
is evident from the readiness with which the tentacles separated without 
injury to the cells. The interlocking, though less in degree, may be com- 
pared to the well-known case of the syngenesious anthers of a Composite 
flower. In not a few cases among the numerous transverse sections at the 
level of the pollen-chamber that have come under observation, close con- 
tact between the adjacent tentacles did not persist till fossilization, though 
the close correspondence of the indentations of the epidermal layers show 
this displacement to be the result of some post-mortem change. Directly 
the region of contact is passed, the epidermal cells, now much broader 
tangentially, are found expanded in their full width as large, cylindrical, 
unicellular hairs or processes with substantial walls which stand out at 
right angles to the surface of the tentacle. It is along the middle line 
