1922] BENSON—HETEROTHECA GRIEVII 131 
It is not an organ sui generis, but partakes of the same potentialities 
as other branches. 
The widely accepted view that the Sphenopterid leaf is a meri- 
phyte, and that the pinna corresponds with the cladodified pri- 
mary branch axis, makes it easy to conceive of the microsporange 
showing in some cases the same anatomical elements as the pinna. 
The Sphenopterid type of frond as recorded by ARBER (2) had 
already made its appearance in the Devonian, and Heterangium 
Grievit may well have retained some reminiscence of the earlier 
unspecialized condition from which we conceive the sporange to 
be evolved. The curious specimens under discussion, which prob- 
ably owe their preservation in_a fossil form to their condition (as 
only two or three specimens of class I are so far recorded), are thus 
of considerable interest for the pinna-like character they exhibit. 
Nor is their more or less sterile condition without significance in 
any discussion as to the origin of the seed habit in Heterangium. 
Any comparable instability in the homosporous sporangia of the 
ancestor could well have been the beginning of a differentiation 
into megasporangia and microsporangia. A plant bearing vascular 
sporangia which retained but a few tetrads and relatively much 
surrounding vegetative tissue might well be regarded as on the 
high road to seed formation. 
Evidence of synangial origin of seed 
It must have occurred to many morphologists that a seed was 
a transformed synangium, the nucellus representing the one persist- 
ing fertile loculus, and the inner integument the vestigial peripheral 
part. In 1904, the evidence then available for this theory of the 
synangial origin of the seed was discussed by the writer (4) in con- 
nection with the description of Telangium Scottii, a synangium 
very common in the Coal Measures of the Upper Carboniferous 
rocks of the north of England, and the first species of that genus 
to be described showing structure. There were at that time certain 
difficulties in the general acceptance of the theory, chiefly because 
the structure of Telangium was such that, while explaining the 
integument, the absence of any central loculus rendered the explana- 
tion of the nucellar or central fertile loculus of the seed difficult. 
