THE ORIGIN OF ANGIOSPERMS. 43 
floral parts as reduced. Here the gradation from hermaphrodite flowers with 
a complete perianth to unisexual naked types presents all stages of reduction. 
At the same time we can trace the evolution of a complicated inflorescence. 
The attractive function of the perianth in a less highly evolved genus, such as 
Acorus, is transferred in many of the higher members to the spathe, which may 
become petaloid and envelop the whole inflorescence. In fact the inflor- 
escence practically comes to function as a single flower. 
Here we find our opponents adopting the very views which we, in common 
with Hallier and others, urge as applicable to all such cases where naked flowers 
are aggregated in dense inflorescences. If Acorus and its near allies were 
non-existent, would this interpretation. of the family have met with equal 
acceptance? Because these stages cannot be so easily traced in other groups 
such as the Piperales and Amentifers, the absence of а perianth in these 
flowers has been too readily accepted as a primitive feature, 
Though Engler regards the hermaphrodite flowers of a genus like Acorus 
as the most primitive types in the order, Campbell *, оп the other hand, has 
decided that the unisexual flower, with a single carpel and a solitary basal ovule, 
e. g., Spathicarpa, Aglaonema, and Nephthytis, is really the least highly evolved. 
This conclusion, based on embryological considerations, appears to us to rest on 
far tooslender evidence, especially in view of the fact that no general agreement 
exists as to which features presented by a study of the embryo-sac may be 
regarded as primitive. 
THE PRIMITIVE Form or THE ORGANS OF THE EUC-ANTHOSTROBILUS OR FLOWER. 
We have seen that Engler and others regard certain orders, where the flowers 
are devoid of perianth and often unisexual, as the more primitive members 
of both existing Dicotyledons and Monocotyledons. But it must not be over- 
looked that Engler's Theory, like the Strobilus Theory diseussed here, is but 
a working hypothesis, the truth of which is to be sought for in its application. 
The, at present, prevailing view has the merit of simplicity. We start with 
something simple, and from it derive the more complicated types оЁ flowers, 
possessing a biseriate perianth, and at the same time the hermaphrodite or 
amphisporangiate condition. But its application as a working hypothesis does 
not assist us in our search for a clue to the phylogeny of the Angiosperms as 
a whole. Nor does it help to bring this group into line with any of those 
now known to us in the fossil state. On the other hand, the Strobilus Theory, 
which postulates that the monosporangiate Apetalee were derived by reduction 
from an amphisporangiate strobilus possessing a distinct perianth, leads us back 
‘naturally to a great group of Mesozoic plants, the Bennettiteze, which afford 
the key to the ancestry of the race in question. 
* Campbell (1905). 
