388 PROF. F. W. OLIVER ON THE STRUCTURE 
entanglements would impede the movements of the autherozoids. The inferred degemera- 
tion of the tracheides, however, solves the difficulty. The view here stated, though 
somewhat hypothetical in character, rests in a measure on the observed facts. 
It will be appropriate to consider here very briefly the fertilisation-mechanisnh of 
recent Cycads. In Zamia the archegonia open upon a shallow depression termed the 
archegonial chamber. The pollen-tubes hang down, so that their ends approximately 
touch the necks of the archegonia. As yet the chambers (the pollen- and archegonial 
chambers) are dry or merely-moist. The pollen-tubes burst and the antherozoids are 
discharged together with the fluid that distends the pollen-tubes—a mechanism recalling 
that of the stinging-hair of a nettle. The necessary fluid for the antherozoids is there- 
fore supplied from the pollen-tubes just before fertilisation. This mechanism is a special 
one correlated with the reduction in number of antherozoids produced by each pollen- ` 
grain, and marks a high degree of precision and certainty in bringing the antherozoids 
to the archegonia, indeed it is the perfection of attainable accuracy possible with free- 
swimming antherozoids. 
In Stephanospermum caryoides the arrangement was more primitive. The antherozoids 
were liberated from the pollen-grains at a distance from the archegonia relatively remote. 
Correlated with the distance to be traversed by the antherozoids we note their production 
in larger numbers than in recent Cycads*. Whilst in the latter two are formed from 
each pollen-grain, in S. caryoides—assuming two internal cells to each pollen-grain and 
that each of them produces a packet of four or five antherozoid mother-cells—the 
minimum number of antherozoids per pollen-grain may safely be estimated at eight. In 
the pollen-chamber of our seed (not counting pollen-grains lost in the cutting and grinding 
of the sections) there are ten internal cells accounted for (see text-fig. 2, p. 380)—4. e., 
ten packets of, say, four antherozoids each. This gives a minimum of forty antherozoids T, 
a number that would necessitate pollination in the Cyead by at least twenty pollen-grains. 
The supposed antberozoids of S. caryoides are small, having average dimensions of 
17 X Dän, These dimensions are confirmed by those of the individual cells of the 
secondary reticula. Those of Cycads are large and in some cases immense. They range 
from 82" X 49 u in Ginkgo, 160 u X 70 n in Cycas, to 222-322 u x 222-3064 in Zamia. 
Whether S. caryoides belongs to a stock from which recent Cycads have originated, 
or whether, as seems probable, the series are distinct, they stand on quite different 
planes of remoteness from Pteridophytes. In the Cycads we have a highly specialised 
pollen-tube-antherozoid-mechanism with reduction in number and increase in size of 
the antherozoids. ln A caryoides, so far as the antherozoids are concerned, we have a type 
relatively near the Pteridophytes—a condition analogous to that which might obtain 
if, in such a type as Selaginella, a pollen-chamber and integument be superadded to the 
macrosporangium, and the microspores being introduced into this chamber followed their 
normal development. The antherozoids (if such they be) have dimensions comparable to 
those of Pteridophytes rather than those of Cycads, whilst all the facts observed are 
consistent with their liberation at a considerable distance from the archegonia, 
-© * A point anticipated on general grounds by D. H. Scott, ‘ Studies in Fossil Botany,’ p. 435. 
+ In S. akenioides, assuming one antherozoid to arise from each of the (say) twenty cells in a pollen-grain, we 
have in Pl. 41. fig. 3 seven non-abortive grains, i. e. one hundred and forty potential antherozoids. 
