124 Sahni. — On Certain Archaic Features in the Seed of Taxus 
On a general survey of this assemblage it was noticed that if a certain 
well-marked tendency (which I shall presently explain) was kept in view, 
it was possible to arrange the seeds in a series, as follows : 
Cardiocarpus ; 1 Cycadinocarpus ; Rhabdospermum ; 2 Mitrospermum ; 
Taxospermum . 
Of course, it is not by any means suggested that this series represents 
the actual course of evolution, but it appears as if the tendency it expresses 
may have been a real factor in the history of these ancient seeds, which, as 
already remarked, are all approximately of the same age. And this view 
at least acquires some interest when it is noticed that, so far as this tendency 
is concerned, the series appears to be continued into the living genera 
Torreya and Cephalotaxus in one direction, and into Taxus in another. See * 
Fig. 7, Nos. 2— 1 1. 
A word of explanation is necessary in connexion with the figures 
referred to. All the figures are purely diagrammatic, but with two excep- 
tions they do not represent any essential facts not already published. 
These exceptions are : 
(a) The dotted lines in Mitrospermum (No. 6) represent the ‘ inner ’ 
(nucellar) vascular system, and have been inserted on the presumption that 
their absence in Mrs. Arber’s material was due to imperfect preservation. 3 
The resemblance of Mitrospermum to Rhabdospermum (in which the ‘ inner ’ 
strands are clearly seen in Brongniart’s figures) is far-reaching enough to 
justify this step. 
(b) In Taxospermum (Fig. 7), too, which is likewise fundamentally 
similar to Mitrospermum , traces of an £ inner ’ system have been observed 
by C. F- Bertrand (1907, p. 216) at a level higher than the inner openings 
of the obliquely ascending lateral canals in the shell. The continuation 
downwards of these strands is thus not entirely unjustified, although their 
exact points of insertion must remain doubtful till better-preserved material 
is available. 
It will be seen that whereas in the first member of the series ( Cardio- 
carpus, Fig. 2) the ‘ outer ’ system of strands is entirely outside the stone, 
as we pass along the series its points of origin from the main supply 
1 As defined by Professor C. E. Bertrand (1908). 
2 A generic name proposed by Professor Seward (1917, p. 341 ) for seeds of the Rhabdocarpus 
type showing anatomical structure, Berger’s original name Rhabdocarpus being reserved for 
impressions. 
3 Since writing the above I have, through the great kindness of Dr. D. H. Scott., F.R.S., been 
able to examine some well-preserved sections of Mitrospermum compressum ; although these sections 
did not show any sign of an internal vascular system, it may be stated that none of them happened to 
pass quite through the right plane. The question as to the existence of internal bundles in this seed 
is therefore open. It is quite possible that no internal bundles existed in Mitrospermum , which in 
this respect would hold the same relation to Rhabdospermum as, among living genera, Taxus 
does to the new Taxinean genus recently discovered in New Caledonia by Prof. Compton ( vide 
supra). 
