138 BOTANICAL GAZETTE [OCTOBER 
mum, Trigonocarpus Parkinsonii, and T. Shorensis show an obvious 
series in the reduction of the compartmental character of their 
integument (text fig. 8, at, bt, c'). In the two former the six bundles 
obviously alternate with compartments as in Sphaerostoma and 
Heterotheca, but in T. Shorensis all trace of the buttresses in the 
plane of the bundles has disappeared, as in Lagenostoma, and the 
bundles appear to occupy a central position. 
The numerical relations of the bundles in Heterotheca and the 
Trigonocarpeae correspond, if we may assume in the latter a simple 
fusion of the two central bundles to supply the megasporange 
(nucellus), which in this series of seeds is free from the integument, 
and thus necessarily requires a central water supply. The tracheal 
mantle which becomes such a prominent organ in some of the 
Neuropterid seeds, for example, Stephanospermum (OLIVER 12); 
needs no explanation if the nucellus were derived from an ancestral 
sporange such as Heterotheca, with its large water storage equip- 
ment. Further, the succulent sarcotesta of the Trigonocarpeae is 
extraordinarily similar in structure to the succulent ground tissue 
of Heterotheca, so far as respects the macaroni-like tubules (text 
fig. 3). There is thus a considerable sum of evidence in support 
of the homology of the Pteridosperm seeds with a microsynangium 
such as Heterotheca. 
Let us now turn to a consideration of the other microsporangia 
attributed with general assent to Pteridosperms. I shall only refer 
to Crossotheca Hoeninghausi (Kipston 9g), Telangium Scotti 
(BENSON 4), and Codonotheca caduca (SELLARDS 15), Of these, 
Telangium Scotti is the only one in which the details of the anatomy 
have been published, but the preservation of the other sporangia 
as incrustation fossils is exceptionally good, and it has been possible 
to learn much of their organization. Crossotheca Hoeninghaust 
shows eight peripheral pairs of loculi, each pair showing a single 
vascular bundle running up in its dividing wall. In the young 
condition the eight pairs are seen to be in close lateral approxima- 
tion, so that they form a body resembling a stout undehisced 
Telangium with a wide base. In specimens recently provided by 
Mr. HEMINGWAY, there seems no reason to doubt that this younger 
condition (occurring nearer the tips of the fronds) represents an 
