SEPTEMBER 10, 1897.] 
these indices, obtained from all the tissues, 
members and organs of a group of plants, 
are judiciously considered they mark out 
lines of descent with great distinctness. 
We have thus open to us three lines of 
investigation in the study of the phylogeny 
of plants, namely: (1) the historical, in 
which the materials are supplied by phyto- 
paleontology; (2) the ontogenetic, in which 
the development of the individual supplies 
us with the necessary data; and (3) the 
morphological, in which the different de- 
velopment of homologous parts is our index 
of relationship. In this paper I propose to 
bring these three lines of investigation to 
bear upon the problem of the phylogeny of 
the angiosperms. * ** * * 
From all the foregoing we may pretty 
safely proceed to construct the hypothetical 
phylogeny of the angiosperms, to serve as 
the basis of their taxonomy. And let it be 
fully understood that this is not presented 
as final, or as entirely satisfactory ; it is 
merely a working hypothesis which claims 
no other merit than that ofan attempt at con- 
formity to the suggestions, sometimes faint, 
sometimes doubtful, from paleontology, from 
embryology (ontogeny) and from morphol- 
ogy. That some of these suggestions have 
been misinterpreted or that others have 
been overlooked is altogther likely, but in 
this I beg the indulgence of systematists, 
who may well realize the difficulties sur- 
rounding the problem here undertaken. 
HYPOTHETICAL PHYLOGENY OF ANGIOSPERMS, 
The angiospermous phylum parted very 
early into two sub-classes, the Monocotyle- 
dons and Dicotyledons. This separation 
took place while the flower strobilus was 
still apocarpous, and before any of the 
strobilar leaves had undergone much if 
any modification. At this stage. the vege- 
tative characters of the sporophyte were so 
well established that no profound modifica- 
tions have been undergone since. 
SCIENCE. 
399 
The modifications which gave us the 
main lines of monocotyledons were, first, the 
fusion of the carpels with one another and 
the production of a syncarpium, and sec- 
ond, the progressive fusion of the syncar- 
pium with the other strobilar leaves. These 
resulted in the phylum which begins with 
Apocarpze and passes to Coronariex, Epig- 
ynee and Microspermee. In some Apocar- 
pe and many plants of the type of the 
Coronariez the perianth has been more or 
less reduced (by aphanisis), in some cases 
amounting to complete suppression, as in 
palms, aroids, and sedges and grasses. 
The primitive dicotyledons were apocar- 
pous plants which soon developed along 
two diverging lines, characterized in the 
one case by the tendency of the leaves of 
the strobilus to fuse with each other in a 
transverse direction (transverse symphy- 
sis), while in the other the tendency was 
to a fusion of the leaves in two directions 
(transverse and longitudinal symphysis). 
The phylum resulting from the predomi- 
nance of transverse symphysis began with 
the apocarpous Ranales, soon developing 
into the synearpous Caryophyllalesand Mal- 
vales. The type of the Caryophyllales be- 
came slightly modified in the Primulales 
by the transverse symphysis of the inner 
perianth-whorl, resulting in gamopetaly. 
In the Polemoniales the type of the 
Primulales began to undergo modification 
by aphanisis, resulting in a reduction of 
the microsporophylls to five, and the car- 
pels in the syncarpium to two or three. 
Increasing aphanisis produced the Person- 
ales and Lamiales with their four or two 
microsporophylls and irregular perianth, 
and in the latter group with each carpel re- 
stricted to the production of but one or 
two macrosporangia. 
The phylum in which both transverse 
and longitudinal fusion are well marked 
proceeds from the apocarpous roseworts to 
the syncarpous saxifrages of the Rosales, 
