PHYLOGENETIC CLASSIFICATION

OF

ANIMALS

(FOE THE USE OF STUDENTS),

/

BY

W. A. HEEDMAN, D.Sc, . ^, '^\^^

F.L.S., F.R.S.E

PEOrEBSOR OF NATURAL HISTORY IN UNIVERSITY COIiLEGE, LIVERPOOL

N^'

WITH ILLUSTRATIONS.

LONDON :

MACMILLAN AND CO. ;

LIVEBPOOL: ADAM HOLDEN.

1885.

LIVERPOOL :

MABPLES AND CO. LIMITED,
LORD STREET.

PREFACE

The accompanying Genealogical Table was drawn up in
May, 1884, mainly from various partial schemes of classifica-
tion which I have been in the habit of using in my lectures
for several years ; and a brief description was read in the
following December before the Literary and Philosophical
Society of Liverpool. While preparing this paper for pub-
lication * it occurred to me that in an extended form it
might prove serviceable to students of Biology : hence its
issue in the present condition.

As it is intended to be used along with a good text-book,
or as a supplement to a course of lectures on Zoology, no
attempt has been made to give the characters of the various
groups, and the facts of Anatomy and Embryology are
only referred to when they indicate the probable course of
Phylogeny, Most attention has been devoted to hypothetical
ancestral forms which are rarely, if ever, mentioned in the
text-books.

An explanation of the meaning to be attached to the
various lines in the table will be found on page 76. The
tree-like arrangement is admittedly the best way of repre-
senting on a flat surface the affinities of organisms, but it
should be remembered that it is after all only a substitute for
the model or actual tree, which would much more correctly
represent the lines of evolution branching through all the
three dimensions of space.

* Proceedings of the Literary and Philosophical Society of Liverpool
for 1884-85.

Some of my statements in regard to the ancestry or
relations of the various groups may doubtless appear very
dogmatic. They are not intentionally so, and have been
put in a positive form simply to avoid circumlocution and
the constant use of "probably," "possibly," and similar
expressions.

It is obvious that a classification such as this can only be
in a limited sense original. It must of necessity agree in
many respects with older schemes, amongst which the well-
known diagrams of Professor Haeckel, published first in
1866, * are the most notable of those in a tree-like form.

In working out the details of the table many books have
been consulted, and I have tried to incorporate the views of
the latest authorities so far as they commended themselves
to my judgment. I may expressly mention the extensive use
that has been made of various books and papers by Huxley,
Ray Lankester, Moseley, Haeckel, Glaus, and others; and
particularly of that invaluable work, Balfour's Treatise on
Comparative Embryology,

No one can be more profoundly impressed than I am
with the temporary nature of such a table as this. The
rapid advances of biological investigation will probably very
soon necessitate additions and corrections to any such phylo-
genetic scheme. The utmost that can be desired is that it
should express diagrammatically the present state of know-
ledge as to the natural classification of animals.

W. A. HERDMAN.

University College,

Liverpool, Januanj, 1885.

* Generelle Morphologic. See also Anthwpogenie, etc., Leipzig, 1877,
and Naturliche Schopfungsgeschichte, Berlin, 1879.

3(,

V-

A PHYLOGENETIC CLASSIFICATION OF
ANIMALS.

In the following account of the probable Phylogeny of
Animals — exhibited diagrammatically in the Table opposite
page 76 — I have begun with the lowest Protozoa and
worked upwards. The Metazoa have been discussed in the
following order (the number refers to the page on which
the group is commenced) : — Porifera (15), Coelenterata (16),
Platyelmia (25), Mollusca (29), Enteropneusta (40), Echino-
dermata (40), Nematelmia (44), Gephyrea (44), Brachiopoda
(44), Polyzoa (45), Chsetognatha (45), Rotifera (45), Crusta-
cea (46), Tracheata (51), Discophora (54), Chsetopoda (55),
Tunicata (58), Cephalochorda (61), Cyclostomata (62),
Pisces (62), Amphibia (64), Reptilia {jq^), Aves do^), Mam-
malia {^^)^

At the base of the Table all Animals and Plants are
represented as having arisen from a single organism of
extreme simplicity, to which the name Protamoeha is attached.
The simplest form of life known to science is Haeckel's
Protamoeha primitiva/^^ (fig. 1) and it is probable that the first

^ ^

i^m ^^^v

^^^m^^^ jfe^^S^ A^f^^M^

Fig. 1. Protamoeha primitiva, Haeckel. Three stages in the life-history, showing
reproduction by fission.

formed organism was a particle of unmodified protoplasm,
* Haeckel, Studien iiber Moneren, Leipzig, 1870, p. 43.

2

which agreed closely with that form in the absence of all
visible structure and differentiation. From such an ancestor
the derivation of the various Monera now living, and of other
allied organisms which have probably existed at different
periods, can be readily imagined to have taken place by
slight changes of form, habit, and life-history, effected by
means of natural selection. It is quite possible that some
existing or extinct groups of lowest organisms may have
arisen independently of others, but there is no evidence in
favour of this " polyphyletic " arrangement, and therefore it is
simpler to understand, and on the whole more probable, that
all forms of life have been derived from a single common
ancestor. When once a particle of living protoplasm had,
under certain conditions of which we are absolutely ignorant,
become evolved from inorganic materials, there would be no
need of any further points of origin. The various Protista
(the lowest animals and plants) can all be satisfactorily
regarded as being derived from one another, or from hypo-
thetical intermediate forms.

The line leading straight * upwards from Protamoeha to
near Protomyxa is supposed to run through a series of ances-
tral Monera which were successively less and less absolutely
undifferentiated, and from which many side branches (long
or short, according to the amount of variation displayed)
have diverged. The few of these which are shown in the
table may be taken as representing some of the more im-
portant groups of Monera which are known. The inter-
mediate forms have become extinct.! The table would be a
more correct representation of nature if every line and branch
had been shown bristling with innumerable short twigs,
extending in all directions, of different lengths, and many of

* All the straight lines in the table must be regarded as very fine zig-
zags, since each ancestral form diverged a little from its predecessor.
t In regard to this, consult The Origin of Species, 6ih ed., p. 293.

8

them branched, and giving off twigs in their turn. These
would represent different slight modifications or varieties,
most of which have died out. They would have added
greatly to the complication of the table, and can readily be
imagined, therefore they have been omitted.

Fig. 2. Protomyxa aurantiaca, Haeckel, A. The Plasmodium stage. B. The
encysted condition. C. The protoplasm inside the cyst breaking up into
mastigopods. D. The mastigopods, set free by the rupture of the cyst, becoming
myxopods, and then uniting to form small Plasmodia.

Protomyxa (fig. 2) is decidedly above Protamoeba and
others of the Monera, and has a complicated and very
instructive life-history,* the main points of which are that a
myxopod stage (the plasmodium, fig. 2, A) passes into an
encysted condition (fig. 2, B), and then breaks up into a
number of mastigopods (fig. 2, C), each of which after
escaping from the cyst becomes a myxopod (fig. 2, D).

* See Haeckel, Studien iiher Moneren; or, Huxley's Invertebrata, p. 81.

These then fuse together, in small numbers, to form
Plasmodia, such as the first stage, and thus complete the
cycle. The life-histories of many of the Myxomycetes
exhibit series of stages very closely resembling those just
enumerated, and therefore it is probable that Protomyxa
and the Myxomycetes possessed a common ancestor, or that
the Myxomycetes diverged from the main stem at the top
of the Monera, and close to where Protomyxa is placed.
This is a very important spot, as it is the place where the
series of plants diverged from the series of animals.

In the table, the plants are seen extending outwards
to the right, opposite Protomyxa, while the animals
extend upwards and occupy the rest of the table. The
long line with short lateral branches which represents
the series of Myxomycetes, is shewn as horizontal, not
because there is no upward evolution in the organisms
it indicates, but simply in order to emphasise its diver-
gence from the main stem of the lower animals. The
Myxomycetes lead to the various groups of Protophyta, or
lower plants, from which the detailed classification of the
Metaphyta might be continued onwards to the right. In
making a phylogenetic table of plants alone, it would be
better to show the series of Myxomycetes and the main axis
of the Protophyta more in a straight line with the Monera ;
and in a scheme embracing all organisms the main stems of
the plants and the animals should be made to diverge at
nearly equal angles from the common ancestor, close to
Protomyxa.

The rest of the lower animals, from Protojnyxa up to the
first dotted line running across the table, are the Protozoa,
and may be considered as comprising four great groups, which
diverge from the main stem, two of them, the Infusoria and
the Gregarinida, near to Protomyxa, and the other two, the
Foraminifera and the Radiolaria, near to Amoeba, further up.

The large and rather heterogeneous group of organisms,
known as Infusoria,* may be traced back to a point of origin
at the top of the Monera, and the ancestral forms were
probably simple Monads, resembling closely the mastigopod
stage in the life-history of Protomyxa, If one of the doubt-
less numerous varieties of Protomyxa-like organisms, which
have existed, had its mastigopod stage emphasised, so as to
become the most important condition in its life, while the
other stages were partly suppressed or modified, it would
gradually become a Monad, or simple Flagellate Infusorian
(fig. 3), and such was probably the mode of origin of this
group of the Protozoa.

Fig. 3. A Monad, one of the Flagellate Infusorians.

The main line of the Infusoria leads upwards from
these simple ancestral forms to the very much higher
and more complex Ciliata; but there are several aberrant
groups, such as some of the Flagellata, the Catallacta
{Magosphcera), Noctiluca, and allied forms, which must have

Fig. 4. MagosphcBra planula, Haeckel. Optical section of the colonial stage in the

life-history.

diverged from the axis far back, and have become evolved
in different directions. Magosphcera (fig. 4) exhibits affini-

* For figures of this group, the student should consult Saville Kent's
Mamial of the Infusoria.

ties with Protomyxa, and with the Flagellata, but is more
highly differentiated ; while some of the Cilioflagellata (e.g.,
Peridinium) shew such striking resemblances to some of the
lower plants that it is a question with some biologists
whether or not they belong to the animal series. But the
resemblances are really not surprising when we consider the
close relationship between the lower Infusoria and the Proto-
phyta. The common Protomyxa-like ancestor is not far
removed from either of them.

In passing up to the higher Infusoria, we find that
the Holotricha {e.g., Paramoecium) — where besides internal
differentiation, the body has acquired a uniform covering of
cilia — are most nearly in the direct line of development {i.e.,
are most nearly allied to the probable ancestral forms),
and may be regarded as the central group of the Ciliata.
The Hypotricha, the Heterotricha, and the Peritricha,
are more or less divergent groups, in which the cilia do not
form a uniform coating, and they may be conveniently
represented as being at the ends of side branches from
ancestors of the Holotricha. The Peritricha {e.g., Vorti-
cella) are probably further removed from the main line
than any of the others. The Opalinida, a small group
of parasitic Infusoria, found in the intestine of some
Amphibia, may be regarded as having degenerated from the
Holotricha, which they resemble in the arrangement of their
cilia.

The Tentaculifera (or Suctoria, e.g., Acineta) are a rather
isolated group, the exact relations of which are difficult to
determine. Saville Kent traces them back to a point on the
main axis of the Protozoa, near to Amoeba, but it is more
probable that they have diverged from the stem of the
Infusoria, distinctly below the Ciliata, as shown in the table.
They are placed on a long side branch, which does not
rise much in its course. This indicates that they are

considerably divergent, but have not attained such a high
grade of organisation as is found in the Ciliata.

So many modifications of form and structure occur in the
higher Infusoria, and so many of these are intermediate or
transition forms between the different groups, that it is com-
paratively easy to imagine the process of evolution, and to
trace the course by which common ancestral forms became
gradually modified through successive generations into
Heterotricha on the one hand, or Peritricha on the other, or
were slowly degraded into the Opalinida. Before leaving the
Infusoria, it is well to notice the great range of organisation
in the group. The difference between such a simple form as
one of the Monads, and such a highly differentiated Pro-
tozoon as Paramoecium, or Euplotes, or Stentor, or Vorticella
is very great. This is indicated in the table by the length of
the line from the point of origin to the top of the Peritricha.
It is greater than that of any other group of the Protozoa.

The Gregarinida,* like all parasitic organisms, are diffi-
cult to place, as there is always a probability that they have
been considerably modified, or even degraded from the ances-
tral type, in consequence of their habits. They are placed in
the table at the end of a long branch springing from the
main stem of the Protozoa, close to the highest Monera, and
extending outward and upward so as to reach a point a
little above the level of Amoeba, but far from the axis. The
length of the line shows the considerable amount of differen-
tiation attained by the group f and its somewhat isolated
position, while its point of origin indicates the relationship
which probably exists with the Monera, There is a similarity
with the life-history of Myxastrum,l and the ancestors of the

* E. van Beneden, Bull, de VAcad. Roy. de Belgique, 2nd ser., T. xxxi.
t E. van Beneden, B^ill. de VAcad. Roy. de Belgique, 2nd ser., T. xxxiii ;
and Quart. Journ. Mic. Sc, new ser., vol. xii, p. 211.

\ Haeckel, Studien iiber Moneren, or Huxley's Iiivertebrata, p. 79.

8

Gregarinida may have diverged from the other Protozoa at a
point close to this form, or one of the other allied Monera.
On the other hand, it is possible that the Gregarinida may
have degenerated from one of the higher Protozoa — from
some form above Amoeba — or even from still higher animals.
The dotted line in the table, stretching downwards from the
base of the Metazoa, may serve to recall the possibility that
the Gregarinida are a much degraded offshoot from some
group of Gastrea-like organisms.

The two remaining large groups of the Protozoa — the
Foraminifera and the Radiolaria — may be satisfactorily traced
back to ancestors which must have been closely allied to
Amoeba. As it is improbable that Amoeba has remained
absolutely unchanged since the time when the Foraminifera
and Radiolaria diverged, it has been placed in the table not
on the main line but upon a short side branch considerably
above the Monera. The ancestral forms which occupy the
axis between Protomyxa and the point nearest to Amoeba
must have gradually acquired a well-marked endoplast, while
at the same time, the protoplasm became more and more
differentiated into two layers, the ectosarc and endosarc, in
the former of which a contractile vacuole * was developed.
It is possible that the AmoebidaB, the Foraminifera, and the
Radiolaria may all have had a common ancestor from which
the three lines started.

Taking the Foraminifera first, we can trace their origin
from the Amceba-like form on the main stem through the
ancestral Lobosa. The stages by which a shell of some kind

* PosBibly this structure was not present in the ancestral forms, and has
been acquired since by Amoeba, Actinosphcerium, &c. I think that the
contractile vacuole in the higher Infusoria must be regarded as having been
evolved independently in that group. The only alternative is to place the
point of origin of the Infusoria much higher up on the main axis, above an
ancestral form possessing a contractile vacuole, and to consider the lower
Infusoria as degraded forms.

9

was first acquired may readily be understood by passing from
the common Amoeba to forms where the pseudopodia are
restricted to one part of the surface ; and then to Difflugia,
where the rest of the body is enclosed by a case formed of
small sand grains, picked up and attached by the protoplasm ;
and Arcella, where a delicate shell of a chitinous nature is
secreted by the surface layer of protoplasm. In such forms
the pseudopodia are still short and thick, as in Amoeba, but
the change from these to the higher Foraminifera, where
the pseudopodia are long and delicate, is a slight one, which
can be readily understood. Finally, the Foraminifera have
branched out into a large number of modifications which
difier comparatively little from one another.

In tracing the history of the Kadiolaria, we start from
much the same point as that occupied by the ancestral
Foraminifera, but diverge in a difi'erent direction. If we
imagine an Amoeba-like form becoming more and more regu-
larly spherical in shape, while the pseudopodia get longer
and thinner and more regularly arranged, and the ectosarc
becomes more clearly distinguishable from the endosarc, we
shall have it gradually passing into an ancestral Radiolarian
or a Heliozoon (such as Actinos'phcermm) , for Haeckel has
shown* that the four existing groups of Kadiolaria, the
Acantharia, the Spumellaria, the Nassellaria, and the Phaeo-
daria, may be traced back to a common ancestor, which
agrees in all particulars with an ancestral Heliozoon in
which the endosarc has become separated from the ectosarc
by a membrane, thus forming a central capsule. This form,
to which Haeckel has given the name Actissa (fig. 5), is so
closely allied to Actinosphcerium that it is certain the two
forms must have had a common ancestor not far back, and
only differing from Actissa in having no capsule membrane.
The Heliozoa of the present day split off at this point, and
* See Nature, vol. xxix, pp. 274 and 296. 1884.

10

have remained at much the same level of organisation, while
the main line continued upwards to Actissa. A side branch
from this point leads to Actinelius, the ancestral form of the
Acantharia. This was derived from Actissa, according to

Fig. 5. Actissa (after Haeckel).

Haeckel, by the hardening of the firmer axial part of the
radiating pseudopodia into spicules of acanthin, the only
form of skeleton found in the Acantharia. The central cap-
sule in this group remains simple and spherical, and is
pierced on all sides by fine pores.

The Spumellaria are more nearly in the direct line
of development than the other three groups, and may be
traced back to Actissa as an ancestral form. The central
capsule remains in its simple form, but most of the
Spumellaria have acquired a siliceous skeleton, or shell,
which serves to distinguish them from the Acantharia. The
Nassellaria have diverged from the primitive Spumellaria,
their probable ancestor Cystidium being derived from
Actissa by the pores of the capsule membrane becoming

11

restricted to an area situated at one of the poles in place of
being scattered equally all over. The Nassellaria agree with
the Spumellaria, and differ from the Acautharia in possessing
a siliceous skeleton.

The Phaeodaria are the most divergent and most highly
differentiated group of the Kadiolaria. They may be con-
sidered as derived from Phceodina, an ancestral form which
has arisen from Actissa by a considerable amount of modifi-
cation. The capsule membrane became double, and the
pores probably first became restricted to one pole, and then
this porous area was modified into a single opening provided
with a radiated operculum. Two small accessory openings at
the opposite pole of the capsule may also be present. A
peculiar pigment body (the phseodium) became developed
outside the capsule membrane, close to the principal open-
ing. These characters are found throughout the Phaeodaria,
and the majority of the group possess in addition a skeleton
formed of hollow siliceous bars, a feature distinguishing them
from all other Radiolaria.

The Metazoa, to which we must now pass on, includes
the animals above the Protozoa, and may be distinguished by
two important characteristics : — 1. The body is always
multicellular, being formed of more than one cell, usually of
a very large number. The few Protozoa which are composed
of more than a single cell (such as some of the Infusoria —
see fig. 6), are clearly colonies formed of a number of inde-
pendent members, each of which is unicellular. 2. Repro-
duction, though it may also be effected by budding, or some
other asexual method, is always performed sexually by ova
and spermatozoa ; while in the Protozoa, these reproductive
elements are not found, and consequently true sexual repro-
duction cannot take place. These two characters might be
considered as one, since the second really depends upon the

12

first. As the sexual elements are equivalent to cells, they can
obviously only be produced in a multicellular body.

As to the method by which the unicellular Protozoa
became multicellular Metazoa, it is probable that the passage

Fig. 6. Magosphcera planula, Haeckel. Optical section of the colonial stage in the

life-history.

was eJBfected by some unknown colonial forms which may be
placed above Amceha, This is a more probable position for
the transition forms to occupy than the tops of any of the
Protozoa groups, such as Infusoria, or Foraminifera, or
Radiolaria would be, and the line stretching straight
upwards from near Amoeba may be supposed to pass through
the hypothetical compound Protozoa. In the development of
any one of the Metazoa, we see a unicellular organism

Fig. 7. Blastula" stage in the development of a Metazoon.

becoming multicellular (e.g., the blastula stage seen in
fig. 7), but the exact process difi'ers considerably in different
groups, and there is great difficulty in determining in which

13

case the ancestral evolution is most closely followed. Con-
sequently, although it is generally admitted that some
embryonic stages in the development of Metazoa probably
repeat the unknown transition forms, still there is great
difference of opinion amongst zoologists as to which em-
bryonic forms actually represent the ancestral Metazoa.
Haeckel * has founded upon the prevalence of the embryonic
stage known as the Gastrula (fig. 8) throughout many

Fig. 8. Gastrula stage in the development of a Metazoon ; a, epiblast ; c, hypoblast ;
bl, blastopore ; en, archenteron.

groups of the Metazoa, his " Gastrea " theory, which is,
that the Gastrula stage in embryology represents the Gastrea,
an ancestral organism formed of two layers of cells, the
outer epiblast and the inner hypoblast, enclosing a central
cavity which communicates with the exterior at one end
(see fig. 8). Lankester,! on the other hand, considers that
the Planula (fig. 9), an embryo formed like the Gastrula
of two layers of cells, but differing from it in having no
opening, is more probably the far back common ancestor
of the Metazoa. Eecently Butschli | has brought forward
arguments in favour of the " Placula " (fig. 10) — a simple

* SUidien zur Gastrcea-theorie, Jena, 1877.
+ Notes on Embryology and Classification. London, 1877.
X Aiinals and Magazine of Natural History^ for May, 1884.

14

flat plate or disc formed of two layers of cells, the upper
epiblastic and the lower hypoblastic — being regarded as the
earliest common ancestor of the Metazoa. The Placula is

Fig. 9.

Planula stage in the development of a Metazoon ; A, planulawith solid hypo-
blast; B, planulawith a cavity in the hypoblast.

certainly a form which might be very naturally assumed by a
small colony of unicellular organisms, and Butschli has
shown how it might readily be modified during its formation
so as to assume either a Planula or a Gastrula structure.

From the early Metazoon represented by one of these,
or some closely related embryonic form (the point marked
Gastrea in the table), several lines must have diverged.

Fig. 10.

Placula— a hypothetical ancestral form (after Butschli).
development are shewn.

Two stages in its

One of these, on the right, slopes considerably downwards to
end in the Dicyemida and the Orthonectida. These two
lowly organised groups are parasitic, and probably degene-
rate. They agree with the Metazoa in being multicellular,
but differ from them in having the endoderm represented by
a single cell only, and they have some peculiarities in regard

15

to their reproduction. On these grounds they have been
distinguished bj^ E. van Beneden,* and others, as Mesozoa.
Probably they are merely degraded and modified offshoots
from an early group of the Metazoa.

The two great series of the Sponges, or Porifera, and the
Coelenterata, probably diverged from the main stem, close to
Gastrea, at or near the same point, or possibly they may
have arisen together by a short side branch representing a
few common ancestors after they had left the main stem. It
is possible, on the other hand, that the Porifera may have
arisen from a group of the higher Protozoa, independently
of the other Metazoa. f In this case, they would have no
close relationship with the Coelenterata.

The Physemarial have been placed upon a separate
branch, close to the base of the Porifera, and arising from
Gastrea. This is the position assigned to them by Haeckel,
who described them§ as simple Gastrea-like organisms,
related to the lowest Sponges ; but more recent investigations
by Eay Lankester, || and others, have thrown very grave
doubts upon this interpretation of their structure, and it is
not improbable that they may all turn out to be merely large
and somewhat abnormal Foraminifera.

The ancestral Sponges probably divided at an early period
in the history of the group, into two series, now represented
by the forms with calcareous spicules and the rest. The
simplest calcareous Sponges (HaeckeFs Ascones) may be
distinguished as Homocoela,1T from the more complex forms
(the Sycones, the Leucones, and the Teichones) or Hetero-
coela.

* "Eecherches sur les Dicyemidea,^^ Bulletin Acad. Roy. de Belgique, 1876.

t See Balfour, Comp. Embiyol., v. ii, p. 285.

I Haeckel, Jen. Zeitsch., Bd. x, and Huxley's Invertebrata, p. 645.

§ Studien zur Gastrcea-theorie, iii, Die Physemarien.

II Quart. Journ. Micros. Sc. vol. xix, p. 476, 1879.

1] Pol6jaeff, " Challenger' Zoological Reports, vol, viii, Part xxiv (1883).

16

The remaining groups of the Porifera have become con-
siderably modified from the primitive ancestor, and their
phylogeny is not very clear. The Myxospongiae, in which
no spicules are present, are probably the least differentiated,
but may have degenerated somewhat, as is shewn by the line
in the table sloping slightly downwards. The great group of
the Fibrospongise, in which an extensive skeletal apparatus,
formed of kerotose fibres, and generally of siliceous spicules
also, is developed, have diverged with a considerable amount
of differentiation from a still more advanced point, leaving
the Hexactinellidse (the vitreous sponges) as the termination
of the Porifera branch, and farthest from the main stem
of the table. The Fibrospongiae have split up into a large
number of smaller groups.

The great series of the Coelenterata arose from Gastrea,
the sac-like ancestor, with its wall formed of two layers of
cells, and having a central cavity and a mouth opening (see
fig. 11). Some of the descendants of this form diverged

Fig. 11. Gastmla stage in the development of a Metazoon; a, epiblast; c, hypoblast;
6i, blastopore; en, archenteron.

from the primitive Sponges on the one hand, and from
the ancestors of the higher Metazoa on the other, and by
the development of tentacles or outgrowths from the body.

17

formed by both cell layers, became the more immediate pro-
genitors of the Coelenterata. The organism* which occupied
the point where the Coelenterate branch first divided was
probably a short, wide, sac-like form, fixed by its aboral end,
and having large tentacles in multiples of four placed equa-
torially, and possibly also smaller tentacles around the mouth
opening (fig. 12, A). This form, probably, on becoming

Fig. 12. Primitive Hydrozoa. Three diagrammatic vertical sections (after
Lankester), A, Ancestral Hydrozoon shewing a condition intermediate between
the Hydra-form and the Medusa-form. B, Simple Hydra-like ancestor produced by
modification of A. C, Simple Medusa-like ancestor produced by modification
of A. m, mouth; en, enteron; t, tentacle; g, reproductive organs. A, Indicates
the same point in each form, viz., the base of the tentacle. In A it is at the
equator of the globose body, in B at the top of the cylindrical body, and in C at
the margin of the beU-shaped body. The dark inner wall of the enteron is the
endoderm, the lighter outer layer of the body is the ectoderm.

sexually mature detached its base, and, by enlarging its
equatorial region into a disc, and finally into a bell concave
on the oral face and provided with circularly-placed bands
of muscle, acquired a swimming organ by the contractions
of which it could be propelled through the water (see
fig. 12, C). The ova and spermatozoa were then developed
upon the inner walls of the enteric cavity of this Medusa-
like form.

This common ancestor of all the Coelenterata (which is
most nearly represented at the present day by the Hydra-

* See Ray Lankester's article, "Hydrozoa," in Ency. Brit., 9th edition,
p. 552.

18

tuba stage in the development of the Discomedusse) must
have given rise to two divergent series of forms — the one
the ancestors of the Hydromedusse and the Ctenophora,
and the other the ancestors of the Scyphomedusse and the
Actinozoa.

In the first series the hydriform fixed stage became
emphasised, the body became longer, and the enteric cavity
remained a simple tube (see fig. 12, B) ; and, in place of
becoming detached and modified into a Medusa-like form
when mature, it acquired the power of budding to a remark-
able extent, the buds being in the form of processes from
the body, and formed of both cell layers. By means of
this property, fixed tree-like colonies were formed, on which,
after a time, a special set of buds were produced which, in
place of remaining like their predecessors in a hydriform
condition, became modified and were detached as free-
swimming medusiform persons in which the reproductive
elements were developed. This specialisation of the two
sets of buds in the common ancestor of the Hydromedusas
brought about the state of affairs generally known as
"alternation of generations." The ovum of the medusiform
person developes into a hydriform person which produces
by budding a tree-like colony upon which certain buds
become medusiform persons, are detached, and produce
ova and spermatozoa.

In the ancestors of the Trachylarida the life-history
became simplified, probably by the hydriform fixed stage
being more rapidly hurried over until it came to be of little
importance, and was finally suppressed altogether — the result
being that in that group (including the Trachomedusse and
the Narcomedusae) at the present day we find the ovum
developing directly into the Medusa.

On the other hand, in many of the Calyptoblastea and
Gymnoblastea, we find that the medusiform persons become

19

modified by developing their reproductive organs while still
attached to the hydriform colony ; this results in their
ceasing to become detached, and in some forms they become
more and more degenerate, until eventually the sporosacs of
Hydractinia, and the simple reproductive organs of Hydra,
are reached.

In the Hydrocorallina and some Gymnoblastea, "poly-"
morphism" has added greatly to the complication of the
colony. We find various sets of buds developing into
differently shaped persons which are specially fitted to per-
form certain functions in the colony. In Hydr actinia ^"^^
for example, there are nutritive persons, reproductive
persons, tentacular persons, and defensive persons in the
one colony.

On its outer surface, in the hydriform person in many of
the Hydromedusae, the ectoderm (the outer layer of cells)
forms a horny layer, the perisarc. This may be very slightly
developed and confined to the aboral end, or it may cover
the entire body and project .beyond the oral region in the
form of a calycle or hydrotheca, and cover groups of medusi-
form buds as a gonangium (as in the Calyptoblastea). The
Hydrocorallina probably diverged from the base of the
Gymnoblastea, their ancestors having acquired the property
of forming calcareous deposits in the ectoderm, so as to
produce a hard, stony corallum. Polymorphism is found
here in an advanced condition.

The Siphonophora diverged from the ancestral Hydro-
medusae, and acquired the characteristic of never becoming
fixed at any period of their life-history. Probably at an
early point in their independent history the hydriform person
developed from the ovum commenced to bud while very
young, and produced medusiform persons before becoming
fixed. The result of this would be that the pulsations of

* AUman's Gymnoblastic Hydroids, Part ii, p. 220, Ray Soc, 1872.

20

the Medusae would propel the entire colony, consisting of
both hydriform and medusiform persons, through the water,
and locomotion being thus effected, it would be unnecessary
for the medusiform persons to become detached. Then
polymorphism produced changes in both the hydriform and
the medusiform persons, resulting in the very complicated
free-swimming colonies of Siphonophora found at the
present day.

The Ctenophora were probably derived from an ancestral
Hydromedusa near to the common ancestor of the Gymno-
blastea and the Siphonophora. The primitive Ctenophora
must first have lost their hydriform stage in the same
manner as it was lost by the ancestral Trachylarida. This
would result in their becoming free-swimming Medusae, like
the medusiform persons of the Gymnoblastea, but difi'ering
from them in developing directly from the egg. They must
then have undergone a series of changes which may be seen
partly effected in the remarkable transition form described by
Haeckel as Ctenaria, which resulted in the narrowing of
the margin of the bell so as to produce a nearly spherical
form with a small mouth opening (as in Pleurobrachia) , and
in the formation of eight bands of modified ciliated ectoderm
running meridionally down the outside from pole to pole.
The enteric cavities also became modified, and the tentacles
were reduced to two, and became retractile into laterally
placed sacs. The evolution of the various groups of Cteno-
phora from a Pleurobrachia-like common ancestor is easy to
trace.

Keturning to the common ancestor of all the Coelenterata,
we find that the second series of forms diverging from this
point leads to the primitive Scyphomedusae and Actinozoa,
and is characterised by the hydriform stage remaining
simple and single in place of producing a colony. It acquired,
however, the power of giving off pieces of its body as buds.

21

each of which developed into a free- swimming Medusa-like
form, in which reproductive elements were produced. We
see this ancestral process repeated at the present day in the
life-history of the Discomedusse {e.g., Aurelia), where the
Scyphistoma stage (hydriform person) produces by trans-
verse fission a number of Ephyrae (medusiform persons). In
some Scyphomedusae {e.g., Pelagia), the hydriform stage has
become suppressed, and the ovum developes into a Medusa
directly, just as it does in the case of the Trachylarida.

The ancestral Scyphomedusa must have developed solid
tentacle-like filaments projecting from the body wall into the
enteric cavity, as we find such gastral filaments* present in
all members of the Scyphomedusae. The Lucernarida are
probably more in the direct line of development than any
of the other Scyphomedusae, as they have retained an inter-
mediate condition between a Hydra and a Medusa form.
They do not produce Medusae by transverse fission (strobila-
tion), but develop genital organs from the endoderm.

The Discomedusae are the typical large Medusae, with
often very complicated sense-organs (modified tentacles)
around the margin of the bell. The mouth may remain
simple or become greatly complicated (in the Khizostomag).

The Cubomedusae and Peromedusae are two small groups,
which differ from the Discomedusae and from each other
mainly in the arrangement of the sense-organs and the
enteric cavities. They probably diverged from the base of the
Discomedusae.

The Actinozoa were in all probability derived from
the ancestral Scyphomedusae, but they have attained a
more advanced condition both as regards their internal
cavities and also their general histology than is found in any
of the Hydrozoa (Hydromedusae and Scyphomedusae). Con-

* Haeckel's "Phacellae," '• Challenger Zoological Reports, vol. iv, part xii.
The Deep-Sea Medusae, lutroduction, p. Ixxiii.

22

sequently a considerable space must be allowed between their
point of origin from the Scyphomedusse and the immediate
common ancestor of the Zoantharia and Alcyonaria. This
form, which must have been polype-like in appearance (the
medusiform condition having been finally lost some way
back), was intermediate in its characters between a Sea-
Anemone and a simple Alcyonarian, but less difi'erentiated
than either. It must have possessed the typical Actinozoon
arrangement of internal cavities — a stomodseum or gastric
tube, and an enteron sub-divided by radiating mesenteries- —
as this structure is found in both the Zoantharia and the
Alcyonaria. It probably had the power of budding off
polypes like itself, which were not free-swimming, but
remained fixed to the parent form. From this *' Proto-
polype " two ancestral series must have arisen, the one lead-
ing onward to the Proto-Alcyonaria, and through them to
the various groups of Alcyonarians now known, and the
other diverging to become the ancestors of the Antipatharia,
the Actiniaria, and the Madreporaria.

The Proto-Alcyonaria must have acquired the character of
having only eight tentacles and the same number of mesen-
teries, while the tentacles became more or less pinnate or
fringed. This ancestral stock is most nearly represented at
the present day by such simple Alcyonaria as the genera
Monoxeniciy Haimeaf and Hartea; while the remaining
Alcyonarians have advanced from this point along two diver-
gent series — the one branch (to the right in the table)
leading through such forms as Sarcodictyon and Clavularia
to the Tubiporidae, and the second forming the common stem
from which the ancestral Helioporidae, Alcyonidae, Gorgon-
idse, and Pennatulidae, have arisen. In the first of these
series the property was acquired of forming calcareous
deposits in the mesoderm (or deeper layer of the ectoderm)
of the body wall. They are found in the form of detached but

23

numerous spicules in Sarcodictyon,"^ and united to form a
continuous tube in Tuhipora. t

The four families forming the second series have been
differentiated in different directions, and differ from one
another mainly in their methods of budding to form colonies,
and in the kind of skeleton the}^ produce. The Helioporidge
are remarkable on account of their having twelve tentacles
and a peculiar corallum which shows certain resemblances to
the skeleton in the true corals or Madreporaria, a group with
which the HelioporidsB have no direct relationship. | In the
Alcyonidae the skeleton is in the form of scattered spicules,
while in the two remaining families, the Pennatulidse and the
Gorgonidae, a solid axis is formed which traverses the centre
of the colony. In the Pennatulidse the colony is of elongated
form, and the axis is unbranched ; while in the Gorgonidae
both are frequently of complicated and tree-like forms.

The second diverging series of forms from the Proto-
polype, must have acquired the characteristic of having the
tentacles simple and in multiples of six, since those con-
ditions are found in the Antipatharia, and in nearly all
Actiniaria and Madreporaria. The ancestors of the Anti-
patharia probably were the first offshoots from the primitive
Zoantharia, and in their descendants at the present day we
find a form of skeleton developed which differs from that of
all other Zoantharia, and agrees with that of most Alcyon-
aria, in being an axial rod. The primitive Zoantharia must
then have split up into two series, of which one was the
stock of the Actiniaria, and the other of the Madreporaria.
The ancestors of the Sea-Anemones lost the habit of repro-
ducing by gemmation, and formed no hard skeleton of any

* Herdman, «« On the Structure of Sarcodictyon," Proc. Roy. Phijs. Soc,

Edin., vol. viii, p. 31 (1883).

t Hickson, " On the Structure and Relations of Tubipora," Quart. Journ.,

Mic. Sc, vol. xxiii, No. xcii, p. 556 (1883).

I Moseley, ** Challenger " Zoological Beports, vol. ii, part vii, p. 102.

24

kind. The tentacles and mesenteries have become very
numerous in most of the forms existing at the present day.

The ancestral Madreporaria, on the other hand, must have
acquired the property of producing a calcareous skeleton
which was not confined to the axis of the colony, as in the
Alcyonaria and the Antipatharia, but was developed in the
mesoderm, not only of the colony, but of the polypes also,
so as to form a skeleton for each member of the colony. In
most of the groups of Madreporaria now existing, repro-
duction by budding takes place to a great extent, large and
complicated colonies being the result.

Returning now to the main stem of the tree, we find that
above Gastrea it passes up into that intensely interesting
region which is the origin of the various groups of lower
Vermes. From this it may be traced upwards through the
starting points of all the great groups of higher Metazoa,
the Mollusca and the Chordata, the Crustacea and the
Tracheata, to its termination in the Polychaeta — the highest
Vermes. From Gastrea to the base of Polychaeta then, the
main axis of the tree may be considered as consisting of a
series of ancestral worm-like forms extending from the most
primitive, the first modification of a Gastrula, up to the
immediate progenitors of the higher Annelides.

What the changes were by means of which the Gastrea
passed into one of the ancestral lower Vermes is difficult to
determine. The body probably became elongated, and a
mesoderm was developed between the two primary cell
layers ; but whether the elongation took place along the
antero-posterior axis of the Gastrea, so that the aperture
remained as a terminal mouth-opening, or at right angles to
that axis, so as to convert the aperture into an elongated slit
placed upon one surface, is a disputed point. Balfour* and

* Comp. EmbryoL, vol. ii, p. 308.

25

Sedgwick* have advocated the derivation of all the higher
Metazoa from a form having the aperture pulled out at right
angles to the axis, and more or less dilated at its angles, as in
many of the Actinozoa, and characterised also by having its
enteron prolonged into a series of radially placed diverticula
resembling the intermesenteric spaces of the Actinozoa.
Such a structure might readily be produced by modifications
of Gastrea (see fig. 13), and there are strong arguments in

Fig. 13. Hypothetical forms shewing stages in the conversion of the Gastrula
into a primitive worm. A, Gastrula. B, Elongation has taken place along a line at
right angles to the long axis of the Gastrula, and the opening is becoming reduced
to a slit dilated at the ends. C, Further elongation has taken place, and the original
Gastrula opening is now converted into two apertures (the mouth and the anus),
united by a curved tube, formed by the archenteron of the Gastrula. D, The anus
has moved further back, and the alimentary canal has become longer and straighter.
E, The diffused nervous system has commenced to be concentrated in the anterior
part of the body in front of the mouth, and as two longitudinal bands running
along the sides of the body, a, anus; in, mouth; p, prse-oral lobe; g, supra-
cesophageal ganglion ; d, lateral nerve cord.

favour of the first ancestral Vermes having had such an
arrangement, but it is very improbable that it was related
in any w^ay to the similar structures in the Actinozoa, and
must have been acquired independently.

The groups of lower Vermes usually united under the term
Platyelmia,f are represented as springing from the main

* Quart. Journ, Micros. Sc, vol. xxiv, No. xciii., p. 43 (1884).

+ Lang {Mittheil. Zool. Stat. Neapel, Bd. iii. p. 187, 1882 ; and Fauna
und Flora des Golfes von Neapel, XI. Monographie : Die Polycladen, p. 645,
1884) tries to show that the lower Vermes have been evolved from the
Ctenophora, a course which seems very improbable and will require a great
deal of evidence in its favour before it can be accepted.

26

stem by a common root, from which the group Cestoda is
given off. The existing forms are placed at the end of a
lin-e which is long, to indicate considerable divergence from
the ancestral form, and slopes downwards to show that the
group is a degenerate one.

The Dendrocoelous and Rhabdocoelous Turbellaria are
probably the least modified members of the Platyelmian
branch, and are therefore placed most nearly in the direct
line of development. Lang* considers the Polycladidae as
the most ancestral forms known, and derives from them the
Tricladidse, the Rhabdocoela, the Acoela, and the Trematoda.
The three last named groups are probably all more or less
degenerate and have been placed in the table on side
branches sloping downwards from the ancestral Dendrocoela.
The Accela were given off first and have degenerated more
than the Ehabdoccela.

The Trematoda are more divergent than any of the
Turbellarian groups. They show affinities in structure both
with the Dendrocoela and with the probable ancestral form of
the Cestoda. Probably they diverged from the Dendrocoelous
Turbellarians, and have been considerably specialised and
somewhat degraded. As a result of their generally parasitic
habits they have lost their original coating of cilia and have
acquired various organs of attachment.

The Nemertida are the highest group of the Platyelmia.
They probably arose from the ancestral Turbellarians very far
back and seem to have undergone a considerable amount of
evolution. The body is elongated and more worm-like than
in the other Platyelmia, and the alimentary canal is more
highly differentiated. An anal opening is present, and a
protractile frontal proboscis, often of very large size and
armed with chitinous styles, is a characteristic feature.

* Fauna und Flora des Golfes von Neapel, XI Monographie : Die
Polycladen.

27

The nervous system, which in the ancestral Platyelminth
was probably diffused over the general surface, with per-
haps a tendency towards concentration around the mouth
opening, has become lost altogether in the very degraded
Cestodes, and is aggregated to form a pair of anteriorly
placed ganglionic masses from which nerves stretch back
along the sides of the body (see fig. 13, E), in the
Trematodes and the higher Turbellarians. In the Nemer-
teans the same general arrangement is found, but the
concentration is more complete and the anterior ganglionic
masses are more definite. The two longitudinal nerves are
long, and they are placed quite laterally. In some forms
there is in addition, a certain amount of diffused nervous
tissue in the body-wall, and in the walls of the proboscis.
The Palseonemertea are the most nearly related to the primi-
tive Nemerteans, while the Schizonemertea and the Hoplo-
nemertea form two divergent series at the top of the branch.

The points of origin of the Mollusca and of the far-back
ancestors of the Vertebrata from the Vermes were probably
close together, and a little way above the origin of the
Platyelmia. We can obtain a certain amount of information
in regard to the probable ancestors of these higher groups
of the Metazoa by comparing their larval forms. MoUuscan
larvsB may be referred to the same type as the Trochosphere
(fig. 14), the typical larval form of the Chaetopoda and some
other groups of Vermes which will be discussed further on.

This larva is a bilaterally symmetrical body with a more
or less rounded dorsal surface, which is prolonged anteriorly
in front of the mouth as a prae-oral lobe. The alimentary
canal is a slightly bent tube concave ventrally.* The mouth
is ventral and the anus posterior, and the tube may be
divided into three regions, the oesophagus, the stomach, and

* Compare with the primitive worm represented by fig. 13, E.

28

the rectum. A circle of long cilia is placed on the praeoral
lobe immediately in front of the mouth, and there may be
also a perianal and several other post-oral bands of short
cilia. There is often an epiblastic thickening of a nervous

Fig. 14. Trochosphere. p, prae-oral lobe ; q, post-oral or abdominal part of
body ; c, prae-oral circle of large cilia ; p.c, perianal circle of cUia ; m, mouth ;
oe, oesophagus (stomodaeum) ; s, stomach ; i, intestine ; a, anus ; g, supra-oesopha-
geal ganglion ; o, eye-spot.

nature in the prae-oral lobe, and simple sense organs may be
developed in connection with it. Paired excretory organs
may also be present in the form of small tubes opening to
the exterior and communicating with the body cavity which
surrounds the alimentary canal.

In the Echinodermata the typical larval form is charac-
terised by having no ganglion or sense organs on the prae-oral
lobe, and by the chief band of cilia being post-oral in posi-
tion. In these respects it differs from the Trochosphere, and
hence it has been argued* that the Echinodermata and the
Mollusca cannot have had a common point of origin, but
Tornaria, the larva of Balanoglossus (Enteropneusta), shows
in some respects characters which are intermediate between
those of the Echinoderm Bipinnaria larva and those of the
Trochosphere. Balfour, however, has suggested that the
resemblances of Tornaria to the Trochosphere are adaptive,
and do not indicate close genetic affinity. If this should

* Comp. Emb. vol ii, p. 316.

29

turn out to be the case, the common origin of the Chordata,
the Enteropneusta, and the Echinodermata (for these three
groups are undoubtedly related) must be moved back on the
main axis so as to be at a different point from the origin of
the Mollusca, and nearer to the Gastrea.*

Turning now to the branch representing the phylum
Molluscaf we find that no existing forms are very primitive.
Probably the most nearly related to *he ancestral Molluscs
are those composing the group Isopleura, but even they are
very considerably differentiated. To express this a long line
has been left between the origin of the Molluscan branch and
the first point of division into the ancestral forms of the
existing groups of the Mollusca.

Professor Kay Lankester has reconstructed the probable
common ancestor of the Mollusca, the form which occupied
the point of division of the branch into primitive Isopleura,
Lamellibranchiata, and Cephalophora. This Archi-Mollusc
(fig. 15) is bilaterally symmetrical and elongated antero-
posteriorly. It has a flat ventral and a rounded dorsal
surface. The head is well marked, and has a region in
front of the mouth (prostomium) on the upper surface of
which are placed a pair of cephalic tentacles. The mouth
is placed in the middle line anteriorly, but behind the
prostomium. The anus is placed in the middle line pos-
teriorly, and rather on the dorsal surface. A pair of renal
excretory organs (nephridia) open one at each side of the
anus, while close to the nephridial apertures, but further
forward, are the openings of the paired reproductive organs.
The ventral surface of the body is formed of a thick muscular

* Balfour advocates the view that the Echinodermata at least have
sprung from a radially symmetrical ancestor, and that the bilateral,
symmetry of the larval forms is secondary {Comp. Emb., vol. ii, p. 318.)

f The student is referred to Lankester's article ** Mollusca," Ency. Brit.
9th ed., p. 632.

30

mass extending from behind the mouth backwards to the
posterior end of the body. This is the characteristic
molluscan "foot." (Fig. 15, A).

Fig. 15, Archi-Mollusc. Three diagrammatic views taken from Lankester's
figures. A, Vertical antero-posterior section. B, Dorsal view, in which the body-
wall is supposed to be transparent, allowing the circulatory, renal, reproductive,
and other organs to be seen through. C. Ventral view shewing the nervous system,
a, cephahc tentacle; &, head; c, free edge of mantle skirt; cZ, anus; e, edge of the
foot; /, edge of the shell follicle ; £/, gonad, or reproductive organ; g\ opening of
duct from gonad ; h, osphradium ; i, ctenidium ; Tc, liver ; I, nephridium ; l^, opening
of nephridium ; m, mouth ; n, posterior end of foot ; o, cephalic eye ; r, auricle of
heart; s, pericardium; v, ventricle of heart, giving off an anterior and a posterior
vessel; y, otocyst; g.a, abdominal ganglion; g.c, cerebral ganglion ; g^.o, olfactory
ganglion; g.p, pedal ganglion; g.pl, pleural ganglion; g.v, visceral ganglion;
n.p, pedal nerve ; n.v, visceral nerve.

Covering the rounded dorsal surface (the " visceral
dome ") is a cap-shaped shell, which is exposed in the
centre but has its edge all round embedded in a flattened sac
in the body-wall — the primary shell follicle (fig. 15, A andB).
On the edge of the visceral dome the wall of the body

81

projects all round in the form of a flap (the mantle skirt)
overhanging the head and the sides of the foot. The
space between the mantle skirt and the sides of the body
is the subpallial chamber, and into this the anus and the
nephridia and generative apertures open. From each side
of the body, rather far back, there grows out into the
subpallial chamber a process of the body-wall, consisting
of an axis (containing two blood vessels) from which a
double row of filaments hang down. These organs are the
gills or ctenidia (fig. 15, B and C). Near the base of the
ctenidium on the body-wall there is a sense organ (the
osphradium) composed of modified epithelium and having a
special ganglion and nerve in connection with it. Its probable
use is to test the quality of the water flowing over the gills.
The mouth leads into the stomodseum, an anterior portion
of the alimentary canal which is epiblastic in origin, and
which forms a muscular bulb. At the posterior end of the
intestine there is a short proctodaeum, which is also derived
from the epiblast. The alimentary canal between these two
terminal parts is the remains of the archenteron and is
therefore hypoblastic. An outgrowth from this region on each
side forms the liver. Around the alimentary canal is a cavity
(the coelom), which has been formed from diverticula of the
archenteron, corresponding probably to the intermesenteric
spaces in the Actinozoa, and in which the blood flows.
The ccelom is more or less broken up into small spaces,
but a large open region is found dorsally to the alimentary
canal, and in this (the pericardium) lies a muscular tube, the
heart, which propels the blood through ill-defined vessels and
lacunae, which are merely parts of the coelom. The heart
consists of two lateral dilated thin-walled sacs, the auricles,
which receive the purified blood brought back from the gills
and pour it into the median thick-walled muscular tube, the
ventricle, which on contracting drives the blood out through

32

two terminal vessels, the anterior and posterior aortas. The
nephridia are coiled tubes which open by one end into the
pericardium (coelom), and by the other to the exterior, close
to the anus.

The nervous system of the Archi-Mollusc is rather com-
plicated and consists of eleven ganglia and various connecting
nerve-cords (see fig. 15. C). There are two cerebral ganglia
above and in front of the mouth, and connected by the
cerebral nerve commissure. At each side of the stomodaeum
is a pleural ganglion, and below it a pair of pedal ganglia.
The two pleurals and the two pedals are connected below
the stomodaeum by a commissure, and each ganglion is
connected also with the cerebral of its own side. From
each pleural ganglion a visceral nerve is continued backwards
below the alimentary canal to a visceral ganglion from which
a nerve is given off to the small ganglion connected with
the osphradium at the base of the gill. The two visceral
ganglia are joined posteriorly by a visceral loop in the middle
of which, below the intestine, a median abdominal ganglion
is placed. Finally each pedal ganglion is united to the
pleural of its own side and gives off also a posteriorly
directed nerve which runs backwards along the lateral part of
the foot. A pair of auditory organs (the otocysts) are placed
in the front of the foot, and an eye-spot is present at the base
of each cephalic tentacle.

From such an ancestral form as this* all the existing
groups of Mollusca have probably been derived. The des-
cendants of the Archi-Mollusc must have split into two
great divergent series which have resulted in the produc-
tion of the Lamellibranchiata on the one hand, and the
Cephalophora on the other.

In the first of these ancestral lines, the primitive Mol-

* For further details see Eay Lankester's article, "Mollusca," Ency.
Brit., 9tli ed., p. 635.

33

luscan characteristics have been modified as the result of a
more or less stationary mode of life. The head region,
including the prostomium, the cephalic tentacles, and the
eyes, has become more and more rudimentary so as to be
finally suppressed altogether. The primitive bilateral sym-
metry is retained, but the body is much compressed from
side to side, the result being that the foot is either nearly
aborted or is reduced to a narrow projection. In conse-
quence of the flattening, the mantle is also changed and
takes the form of two large laterally placed lobes, on the
outer surface of each of which a valve or plate of the shell is
formed. The ctenidia have become enormously enlarged to
form two great gill-lamellae on each side of the body.
Besides being respiratory in function, they also, by their
extended ciliated surface, cause currents of water which
bring food particles to the mouth — an important matter to
an animal with very limited powers of movement. The
cerebral, pleural, and visceral ganglia of each side of the
body have coalesced to form a single mass, the so-called
cerebral ganglion, placed alongside the mouth. The pedal
ganglia are normal, and the osphradial are very large
and are usually known as parieto-splanchnic. Two large
muscles run transversely across the body, one in front of and
above the mouth, and the other below the anus. They serve
to approximate the valves of the shell, and are the anterior
and posterior adductor muscles.

In the primitive Lamellibranchs, these muscles were
probably of much the same size, and we find the most direct
descendants of such forms in the Isomya of the present
day, and especially in the Integropalliata, such as Area,
Nucula, and Trigonia, which may be regarded as being
more in the direct line of development than any of the
other Lamellibranchiata. From this stock two divergent
lines have sprung. On the one hand, the Sinupalliata have

34

been evolved upwards ; and on the other, the Heteromya and
the Monomya have degenerated.

The Sinupalliata have remained Isomya, both adductors
being well developed, but the posterior parts of their mantle
lobes have become prolonged and united to form a pair
of siphons through which the water flows into and out
of the pallial cavity, and the line of attachment of the
mantle to the shell has been inflected so as to form a
sinus into which the siphons may be more or less completely
withdrawn.

In the other series we have, starting from one of the
primitive Integropalliate Isomya, the anterior adductor
muscle becoming more and more reduced so as to produce
first the Heteromya, including such forms as Mytilus, where
the muscles are very unequal in size, and then the Monomya
(Ostrea, Anomia, Pecten, &c.), where the anterior adductor
is entirely absent. That the Monomya have really degene-
rated from Isomya is very clearly proved by Huxley's
discovery that the oyster {Ostrea) when very young has both
adductors well developed, but loses the anterior one later on,
thus recapitulating in its development the ancestral history.
The various groups of Lamellibranchiata differ from one
another also in the structure of the gill-lamellae, which may
become very complex, and in the special development of the
pallial siphons and the foot.

In the second ancestral line, starting from the Archi-
Mollusc, the prostomial region has been retained and
developed into a well-marked head with cephalic sense
organs, and a special and very remarkable organ, the
odontophore, has been developed in connection with the
stomodgeum. This organ consists essentially of a band,
covered by transverse rows of chitinous teeth, stretched over
a cartilaginous mass placed on the floor of the buccal cavity,
and capable of being protracted and retracted by special

35

muscles.* From this ancestral Cephalopborous or Glosso-
phorous Mollusc, which is continued up into the primitive
Gastropoda, two lines have diverged. The first, with compa-
ratively little modification, to the Isopleura, and the second,
with a considerable amount of divergence, to the Scaphopoda.

In the Isopleura, the primitive bilateral symmetry has
been retained, and most of the systems remain very much in
the condition in which they were found in the Archi-
Mollusc, with, of course, the addition of the odontophore.
The Isopleura include three groups : — the Polyplacophora,
or Chitons, in which the simple dorsal shell has been
multiplied so as to form a series of eight valves ; and the
Neomenise and Chaetoderma in which the shell is repre-
sented by numerous minute calcareous plates or spines, and
the mantle and foot are much reduced, and the body is
worm-like in form. These are the most primitive Gastro-
podous Molluscs which are known.

The Scaphopoda {Dentalium) are much more modified,
and possibly somewhat degenerate. They have retained the
primitive bilateral symmetry, but the body has become
greatly elongated antero-posteriorly, and the foot is especi-
ally produced anteriorly, and adapted for burrowing in sand.
The mantle-skirts have fused ventrally below the foot, so as
to produce a cylindrical body-form, and around this the shell
is developed as a cylindrical tube, open at both ends. The
heart seems to have been lost during the stages which have
intervened since the Scaphopoda diverged from the primitive
Gastropods.

Eeturning to the line leading upwards to the higher
Cephalopborous Molluscs, we find that after the ancestral
Scaphopoda had diverged, the visceral dome must have
increased greatly in size, so as to form a great dorsal pro-

* For a description, with figures, of the odontophore, see Lankester's
article, '* Mollusca," Ency. Brit., 9th edition, p. 640.

36

jection, containing the greater part of the digestive and
reproductive viscera. Then two great lines of descent were
produced, one leading to the Pteropoda and Cephalopoda, the
other to the Gastropoda.

Following up the last of these we find that a large
shell became developed as a protection to the visceral mass,
and that (possibly, as Lankester suggests, as a result of
the weight of the shell falling on one side) the visceral dome
became twisted round spirally to the right, so that the anus
which was originally posterior came to be placed first on the
right side of the body, and then anteriorly above the head.
The whole visceral mass has also in most cases come to be
more or less coiled spirally, and the shell covering it has taken
the same form. By these changes the bilateral symmetry was
entirely destroyed. The ctenidia and the nephridia and other
paired organs shared in the torsion of the visceral mass, and,
in most cases, became unequally developed on the two sides
in consequence. In many cases the ctenidium and nephri-
dium, which were originally on the left side of the anus,
and which, after the changes, came to be upon the right
side, have become atrophied.

There is reason to believe that these ancestral Gastropods
split into two series, in the one of which the visceral nerve
loop, with the visceral ganglia, became implicated in the
rotation of the visceral mass, the result being that the loop
was twisted into a figure of eight, and the ganglia changed
sides ; while in the other series, possibly on account of the
deeper position of the nerves in the body, they were not
affected by the other changes. For the first of these series,
in which the nerves cross, Spengel has proposed the name
Streptoneura ; and for the second, in which the visceral
nerves remain straight and unaltered, Euthyneura.

The primitive Streptoneura divided into two series :
the Zygobranchiata {Patella, Haliotis, &c.) in which the

37

ctenidium and nephridium on the right side of the body did
not atrophy, but are as large, and sometimes larger, than
those on the left side; and the Azygobranchiata, in which
the ctenidium and nephridium on the right side of the body
have been lost. This last is a very large group, including
the majority of the Gastropoda, and is considerably more
modified than the Zygobranchiata.

The Heteropoda are a divergent group of the Strepto-
neura. They agree with the Azygobranchiata in their
essential characters, but have become adapted to a free-
swimming existence. The foot is especially modified in
some forms {e.g., Carinaria) into a flattened fin-like organ,
provided with a sucker. The visceral mass may become
greatly reduced, and the shell may be entirely lost.

The Euthyneura are characterised by their straight vis-
ceral nerves, and by the atrophy of the paired organs on the
right side of the body. They are all hermaphrodite, and in
many cases the shell is absent. The ancestral Euthyneura
are represented by the Opisthobranchiata of the present
time, including two series, the Nudibranchiata and the
Tectibranchiata ; while the Pulmonata are a divergent group,
derived from the primitive Opisthobranchiata. In the Opis-
thobranchiata the heart is placed anteriorly to the base of
the gill. In the Nudibranchiata the mantle-skirt and the
shell are both absent. Some extraordinarily modified forms
belong to this latter group. In many the ctenidium is
absent, and respiration is performed by other processes from
the body wall.

The Pulmonata have become adapted to a terrestrial life.
The ctenidium has been lost, and the paUial cavity has
become converted into a respiratory sac, communicating
with the exterior by a small aperture. In some forms
{LimaXy &c., the slugs) the visceral mass is greatly reduced,
and the shell is either very small or altogether absent.

38

Returning now to the primitive Cephalophora, after the
ancestors of the Scaphopoda had diverged, we must trace the
line leading up to the Cephalopoda and the Pteropoda. In
this series the primitive bilateral symmetry has been retained,
and the visceral mass has been greatly enlarged, so as to
produce a great dorsal projection, but no twisting of this
region takes place, the anus remaining in its primitive
posterior position. The foot has been greatly modified, in
accordance with the free-swimming habits of the animals.
Its front part has grown upwards and forward, so as to
surround the head, the result being that the mouth is
placed in the anterior portion of the foot, and opens ven-
trally. In most cases this region of the foot is drawn out
radially into paired processes, which may be provided with
suckers. The median part behind the head is developed
into a pair of flaps (the epipodia), which may be used as
swimming organs, or modified to form a posteriorly placed
funnel. The remainder of the foot is undeveloped. A shell,
if present, is light and fragile.

Such an ancestral form as this is most nearly represented
at the present day by the Pteropoda, a group which diverged
far back, and probably have undergone comparatively little
modification. Some of them are probably degenerate. The
epipodia have remained in their primitive condition as large
muscular flaps, and are used as swimming paddles. The
other parts of the foot may remain in a rudimentary con-
dition. The ctenidia have been lost, and in one section of
the group (the Thecosomata), where a mantle and shell are
present, the walls of the large subpallial cavity act as organs
of respiration. In the Gymnosomata the mantle skirt is
absent, and there is no shell.

The ancestral Cephalopoda, after the Pteropoda had
diverged, must have had their epipodia modified, so as to
become more or less completely united in the middle line

39

posteriorly, to form a tube (the siphon or funnel), open at
both ends, which would serve to conduct the water outwards
from the subpallial chamber. The front of the foot, sur-
rounding the mouth, must have become well developed, and
drawn out on each side into four or five projections. The
other systems remained in their primitive typical condition.

This ancestral line divided into two branches, the one
leading to the Tetrabranchiata {Nautilus and extinct forms,
e.g., Ammonites), and the other, after a considerable amount
of further evolution, to the Dibranchiata, including the
ordinary cuttlefishes. The Tetrabranchiata possess two pairs
of ctenidia and two pairs of nephridia, and the epipodia are
not completely united. The lobes of the front part of the
foot bear numerous tentacular processes, but no suckers. A
large external shell, divided into a series of chambers, is
present.

The ancestral Dibranchiata after separation from the
Tetrabranchiata must have undergone some further changes.
The fusion of the epipodia became complete. The lobes of
the fore-foot developed rows of suckers (acetabula), and the
shell became enclosed in folds of the mantle so as to be
internal. The nervous system and sense organs became
more highly evolved, and finally, an "ink" sac was developed.
The Dibranchiata are divided into two groups (Octopoda and
Decapoda) according to the number of processes developed
from the foot surrounding the mouth. The very long line
occupied by the series of ancestral Cephalopoda indicates
the extensive modifications the group has undergone during
its evolution. The Dibranchiata occupy the highest point in
the Mollusca and are very far above the place of origin of the
phylum from the Vermes. This shows the great range of
organisation which is found amongst the Mollusca.

The Echinodermata, it has already been seen, probably

40

arose along with the Enteropneusta and the early Chordata
by a common root from the Vermes, close to the point of
origin of the primitive Mollusca, or possibly a little further
back. This branch is the largest and most important in the
table. It probably very soon broke up into two series of
ancestral forms ; (1) those which lead upwards to the Verte-
brata, and to which we shall return later on ; and (2), the
common ancestors of the Enteropneusta and Echinodermata.
This latter series is now represented by Balanoglossus,
which may be regarded as the termination of the primary
branch having as lateral off-shoots the Echinodermata on
the one hand, and the Proto-Chordata on the other.

The relationship between the Echinodermata and the
Enteropneusta is shown by Tornaria, the larva of Balano-
glossus agreeing with the typical Echinoderm larva in all its
most important characteristics.* The ancestral forms of this
branch may be regarded as being derived from the Trocho-
sphere (or possibly from a rather more generalised form
which had not yet acquired all the special Trochosphere
characters), after separating from which, they must have
acquired such peculiar features as the longitudinal post-oral
band of cilia, and the derivation of a water-vascular system
from the archenteron. From such an ancestor Balanoglossus
has probably been derived very much as we see it developing
at the present day from the Tornaria stage, by the disappear-
ance of the longitudinal band of cilia, the conversion of the
praB-oral lobe into the proboscis, the elongation of the body
into a worm-like form, and other changes.

The line along which the Echinodermata were evolved is
not so easy to trace. Probably their first ancestors (fig. 16)
differed from the ancestors of the Enteropneusta in that
portion of the coelom (diverticula from the archenteron),
which formed the water-vascular system being developed
* See Balfour, Comp. EmbryoL, vol. i, p. 485.

41

radially, or as a series of pouches around the front part of
the alimentary canal, so as to produce the first traces of the
more or less complete radial symmetry which is such a
marked feature in all Echinoderms. If the nervous system
was still in its primitive difi'used condition, it may readily
be imagined that the formation of radially arranged regions
of the coelom, which were being evolved into vessels with
tentacle-like projections to the exterior, would result in the
concentration of the nervous tissue along these lines ; and
if, as is probable, there was previously a nervous concentra-
tion around the mouth, then the newly formed nerve bands
would naturally radiate outwards from the circum-oral ring
(see fig. 16). The other characters which the common

Fig. 16. Hypothetical ancestor of the Echinodermata. m, mouth; a, anus;
n, nerve cord; r, water- vascular diverticulum from archenteron.

ancestor of the Echinodermata must have acquired are : — a
tendency towards the formation of the water-vascular caeca
in fives, and as a result the pentagonal symmetry of most of
the systems of the body ; the development of a large ccelom
or body cavity; and the deposition of calcareous matter in
the deeper parts of the integument.

The ancestral line then probably split into two series : —
the one being continued into the progenitors of the
Asteroids, the Ophiuroids, and the Crinoids, and the other
being evolved into the primitive Echinoids and Holothurians.

In the latter series the primitive body-form was most nearly
retained, but all the systems became considerably differen-
tiated.

In the ancestral Holothuroidea the shape became elon-
gated antero-posteriorly, and, as a result, the pentagonal
arrangement of many of the parts was masked. The calcar-
eous deposits in the integument remained as scattered
spicules, while strong bands of muscle were developed to
strengthen the body wall. At the anterior end, around the
mouth, a part of the water-vascular system was prolonged
outwards to form a series of large tentacles, while at the
posterior end of the body a pair of large branched tubular
organs (the " respiratory trees ") in all probability place the
coelom in communication with the exterior through the
cloaca, and act as nephridia. From such an ancestral form
the various groups of Holothuroidea may be readily derived.
Some of them (e.c/., Synapta) are probably slightly degen-
erate, the respiratory trees and some parts of the water-
vascular system being absent.

In the ancestral Echinoidea the body became more
globular, and the calcareous deposits in the integument
were increased greatly, so as to form eventually a continuous
shell composed of regularly arranged plates, and bearing
spines and other calcareous projections on its outer surface.
Some of the existing groups of Echinoids have become
highly differentiated, but all the systems may be traced back
to those of the ancestral form. The water-vascular system
has become an important and complicated series of organs,
and a very elaborate calcareous apparatus, bearing teeth
(Aristotle's lantern), has been evolved in connection with the
mouth.

Returning now to the main branch of the Echinodermata,
and tracing it onwards towards the Asteroidea, Ophiuroidea
and Crinoidea, we lind that in the common ancestors of

48

these groups the body form must have become considerably
changed. It must have been flattened antero-posteriorly, so
as to reduce the oro-anal axis, and then pushed out radially
along the lines of the chief parts of the water-vascular
system, so as to produce, first a pentagonal, and then a
stellate shape. There being then only half the body in a
position for creeping, the water- vascular system was developed
only on the lower (oral) surface, in place of extending almost
to the aboral pole, as in Holothurians and Echinoids.

From such an ancestor the Asteroidea were probably
given off, and they have retained all the essential characters,
while some of their systems have become more highly
differentiated. In the Ophiuroidea and the Crinoidea the
radial processes of the body or arms have become more
marked, and the formation of calcareous matter has increased
greatly, the result being that the arms are almost entirely
solid, and in great part formed of calcareous ossicles. These
changes took place after the separation of the ancestral
Asteroids, so that probably the Ophiuroids and Crinoids
have had a rather longer common ancestry.

The ancestral Crinoids, after the Ophiuroids diverged,
then probably became fixed by the aboral end of the body,
and this necessitated a change in the position of the anus,
resulting in both openings of the ahmentary canal coming
to lie upon the same surface of the body. The extinct
Cystoidea and Blastoidea may be placed as neighbouring
side branches from the primitive Crinoids. Probably these
three groups are all somewhat degenerate.

The remarkable characteristics acquired by most Echino-
derms during their larval stages (e.g., Pluteus) are certainly
merely adaptive and have no phylogenetic significance, and
this explains the sudden, and in some cases, very profound
changes which may take place when the animal throws off its
temporary larval characteristics and acquires its adult structure.

44

Returning to the main stem of the table, we find a
number of comparatively small groups, lying between the
primitive Echinoderms and the higher worms, which are
probably all derived from side branches given off from what
may be called the middle third of the vermean axis — the
lower third being the region below the point of origin of the
Mollusca, and the upper third above the origin of the
Arthropoda. These small groups are all rather aberrant,
and some of them are certainly degraded.

The Nematelmian worms are represented by a branch
extending upwards to the right above the common origin of
the Chordata and the Echinodermata. The members of this
group are less unlike the typical Vermes than is the case with
most of the other branches from the middle third of the
vermean axis. The highest forms of the group are the
Nematoda, while the Acanthocephala (Echinorhynchus) must
be regarded as a degraded offshoot, sloping downwards from
near the base of the Nematelmian branch. The degenera-
tion of Echinorhynchus, and probably of some of the
Nematoda also, is a result of their parasitic mode of life.

The Gephyrea have probably arisen from the main stem,
not far from the point of origin of the Nematelmia. They
have developed a spacious coelom, and in most cases a long
convoluted alimentary canal. On the other hand, they retain
the characters of primitive Vermes in the absence of true
segmentation, and in the larval prae-oral lobe developing into
an important part of the front of the body. In most, a small
number of paired nephridia are present, and in some cases
(e.g,y Bonellia), two of these placed at the posterior end of
the body become greatly enlarged to form branched organs,
placing the coelom in communication with the proctodaeum,
as in the case of the respiratory trees in the Holothuroidea.

The affinities of the Brachiopoda are still unsettled. The
larvae of some forms {e.g., Argiope) appear to shew close

46

relationship with the Chaetopoda, but there is a want of
agreement between the groups in some points which makes
the matter doubtful, and therefore it is safer at present not to
regard the Brachiopods as degraded Chaetopoda, but as
derived from a lower point in the vermean series. The exact
position of that point of origin is very doubtful ; probably it
lay above the primitive Gephyrea, but below the origin of the
Kotifera. The present Brachiopods must be regarded as
having degenerated in accordance with their sessile mode
of life.

The Polyzoa are also degenerate forms belonging in all
probability to this division of the Vermes. The larvae are
free-swimming ciliated forms, which may be compared with
the Trochosphere stage found in the development of so many
Vermes. Balfour* regards them as Trochospheres, which
became fixed in the adult by the extremity of the prse-oral
lobe. He also shows that there is reason to consider the
Polyzoa as exhibiting alternation of generations. The ovum
develops into a free-swimming form (the so-called larva),
which never becomes sexual, but produces by budding the
attached form (the adult Polyzoon), which develops repro-
ductive organs. From the fact that Cyphonautes, the larva of
Membranipora, an ectoproctous Polyzoon, is itself entoproct-
ous, it is probable that the Entoprocta (Pedicellina) are more
primitive than the Ectoprocta. Both groups are, however,
degenerate and considerably modified.

The Chaetognatha (Sagitta) are a small group with
obscure affinities, which are best placed in this part of the
vermean series close to the origin of the Arthropoda.
Possibly they may be more closely related to the Nematelmia
than is shown by the table.

The Rotifera are certainly degraded, but what they have
been derived from is somewhat doubtful. Pedalion appears

* Comp. EmbryoL, v. i, p. 255.

46

to show Arthropod affinities in some of its characters, and
it is possible that the group may be connected with the base
of the Crustacean series. On the other hand, the Rotifera
retain in the trochal disc an organ which is apparently the
ciliated prae-oral lobe of the larvae of so many of the Vermes,
and this seems to shew that they have arisen directly from
the vermean axis. Some of the groups of Rotifera have
degenerated to a very great extent.

We now come to the important region where the two
great Arthropodan series, the Crustacea and the Tracheata,
diverged from the base of the higher worms at or about the
same point. It is probable, from a consideration both of the
anatomy and of the development of the two groups, that they
do not belong to one great series, but have been evolved
independently, and therefore, having no common ancestors
nearer than the Vermes, they must have acquired separately
such Arthropodan characteristics as they possess in common.
The Tracheata are descended from some primitive Annelidan
form allied to Peripatus, while the Crustacea must have had
as their common ancestor a primitive Phyllopod, and the
differences between these two types — Peripatus and the
Phyllopod — are so great that they cannot have had any
common ancestry nearer than the Vermes. The similarity of
some of the organs in the two series may be explained by
considering them as similar differentiations of parts inherited
from their Annelidan ancestors, while other common charac-
teristics may be regarded as being merely adaptive, and due,
in some cases, to similar modes of life and habits.

At the point where the Crustacea and Tracheata diverge
the Vermes must have already acquired some of the charac-
teristics of their higher groups, such as true segmenta-
tion, or the formation of metameres ; and rudimentary
appendages, as simple processes developed in pairs, one from

47

each metamere. Renal organs were present in the form of
paired nephridia. In the ancestral Crustacean, derived from
this primitive segmented worm, the lateral appendages must
have become developed into more efficient locomotory organs,
by their elongation and the formation of joints upon them.
The increased power of movement which this change con-
ferred probably resulted in an increase in the muscular
system connected with the appendages, and this, in order
that the new organs might be fully taken advantage of, would
require the formation of a hard skeletal tissue, to which the
muscles might be attached. To meet this want, the more or
less calcified cuticular *' shell " of the Crustacea was devel-
oped, and, in the higher members of the group, has been
brought to a high degree of perfection as an exoskeleton.

The further changes which have taken place during the
evolution of the Crustacea, have been mainly in the direction
of producing " heteronomy " in the segments of the body.
Division of labour amongst the appendages has resulted in
great modifications in structure, some being converted into
jaws, others into sense organs, walking legs, swimming legs,
and other organs ; while the segments themselves have also
undergone change, the exoskeleton becoming in many cases
fused over large tracts to form a continuous hard covering or
carapace. Some of the metameres at the front of the body
more or less completely unite with the prse-oral lobe to form
a distinct " head."

The free larval stage known as the Nauplius, is of such
constant occurrence throughout the Crustacea, though exhib-
iting slight peculiarities in each group, that it is almost
certain that it represents more or less closely a far-back
common ancestor. The Nauplius of the Phyllopoda (fig. 17),
the most nearly primitive group of Crustaceans, is an ovate
body, provided with three pairs of appendages, attached to
the anterior (cephalic) region. These appendages afterwards

48

become the two pairs of antennae and the mandibles. The
first pair of antennae are comparatively slight, consist of one

Fig. 17. Phyllopod Nanplius (Apus) when hatched (after Claus). 1, first pair of
antennEe ; 2, second pair of antennae ; 3, mandibles ; I, labrum ; o, median eye ;
s, segments of the post-cephalic region of the body.

branch only, and are probably sensory organs. Between
their bases is placed a large upper lip or labrum, behind
which lies the mouth. The second antennae are very much
larger, and are biramous, with a spine projecting from near
the base inwards towards the mouth. These appendages
are the main organs of locomotion of the Nauplius. The
mandibles are rather smaller. They represent the mandi-
bular palps of the adult, and are biramous. On the front
of the head is placed a median unpaired eye. The post-
cephalic region shews traces of five segments and their
appendages, and in this respect diff'ers from most Nauplii,
which are unsegmented.

Probably this Phyllopod Nauplius is the nearest form
known to the primitive Crustacean, but it must remain
doubtful whether or not additional post-cephalic segments
with simple biramous appendages were present. From
this form the existing group of Phyllopoda was probably
evolved very much as we find Apus developing from the
Nauplius stage at the present day. The modifications of
form and structure take place gradually, without any
sudden metamorphosis. This ancestral series of Proto-
Phyllopods has formed the main axis of the Crustacean
branch from which the various existing groups have been

49

giveD off at different points. It ends in the existing Phyllo-
poda, the nearly allied Cladocera being shown as a slightly
divergent side branch.

The Copepoda, the Cirripedia, and the Ostracoda have all
probably arisen independently from the ancestral stem, and
have become considerably modified in different directions.
The Copepoda are probably the oldest group. They retain
the median Nauplius eye and have simple biramous append-
ages. A number of the Copepoda are certainly degenerate
forms, and some of them are very much modified.

The Cirripedia probably arose further up the main
stem, and in the ancestral condition were characterised
by the possession of a large carapace, in the form of a
bivalve shell. We see the remains of this stage in the
" Cypris " larva of existing Cirripedes. The group has
undergone very great modifications since, due doubtless to
the change from a free-swimming to a sessile existence. A
considerable amount of degeneration has also taken place
amongst the Cirripedia.

The Ostracoda probably arose from a point still further
up in the series of the Proto-Phyllopods, and have under-
gone a good deal of modification. A bivalve shell has been
developed to such an extent as to enclose the entire body.
In some members of the group degeneration resulting in the
loss of the heart and the compound eyes has taken place.

The higher Crustacea have arisen from the main axis
of the group, below the point of separation of the Cladocera
from the Phyllopoda, but probably not much further back,
and the NebaliadaB are, according to Claus,* the nearest
forms we know to the primitive Malacostraca. Possibly after
this, as suggested by Balfour, t the thoracic appendage

* Ueber den Bau und die systematische Stellung von Nehalia. Zeitsch. f.

Wiss., Zool, Bd. xxii (1872).

+ See Balfour, Gomp. EmhryoL, vol. i, p. 420.

E

50

became reduced, and a form was assumed which is more or
less nearly represented by the Zoaea stage, which occurs in
the development of most of the higher Crustacea.

The Edriophthalmata must have sprung from the primi-
tive Malacostraca far back. They have no Zoaea stage, and
consequently may have diverged during that first Phyllopod
condition of the Malacostraca which Nehalia still represents,
or they may have arisen later on and have lost the Zoaea
stage in their development since. The Edriophthalmata
probably underwent a considerable amount of divergence
before breaking up into the groups now existing. The
Stomatopoda and Cumacea may be represented as short
lateral branches from the Malacostraca after the Zosea-like
ancestral stage. The Cumacea in some respects show
resemblances to the Edriophthalmata.

The ancestral Podophthalmata then split into the Schizo-
poda, a small group which did not diverge much, and which
may be taken as representing an ancestral stage (Mysis),
and the Decapoda, which after some further evolution,
including the loss of the Schizopod character— the presence
of exopodites on the posterior thoracic appendages — became
broken up into the ancestral forms of the existing groups.
The Macrura and the Brachyura form two somewhat divergent
series, while the Anomoura have undergone considerable
modification or degeneration. Penceiis, amongst the Macrura,
has been shown by Fritz Miiller* to leave the egg as a
Nauplius, and to go through a series of larval stages, which
probably represent more nearly than any other forms the more
important ancestral stages in the phylogeny of the Crustacea.

In the Decapoda the evolution of the sense organs, the
efficiency of the exoskeleton, the heteronomy of the segments,
and the specialization of the appendages reach their highest
degree of perfection.

* Facts for Darwin, London, 1869.

51

Recent researches* on the structure and embryology of
Peripatus (Onychophora in table) have shown that while it
is distinctly one of the Tracheata, in as much as it possesses
respiratory organs in the form of tracheae, it still exhibits
the Annelidan feature of paired nephridia corresponding to
the somites. Peripatus from its geographical distribution
must be regarded as a very ancient type, and in all prob-
ability it represents the ancestral Tracheata close to the point
of their divergence from the Vermes. It has probably
become slightly modified by degeneration, but still is
extremely useful in helping us to form an idea of the series
of changes by which the Tracheata were evolved. The
ancestral worm from which the branch started must be
regarded as having attained about the same level of high
organisation as the form from which the series of Crustacea
arose, consequently the two great Arthropodan branches may
have diverged from a common ancestor in the Vermes.

The primitive Tracheata had an elongated worm-like
body, which was divided into segments, probably with com-
paratively little heteronomy. Each segment was prolonged
laterally into a pair of processes, the appendages, which were
beginning to be transversely segmented or jointed. The
muscular system was becoming more highly differentiated,
and the cuticle had partially hardened to form an exoskeleton.
The characteristic excretory organs of the Vermes were still
retained in the form of laterally placed nephridia, a pair in
each segment ; but the typical respiratory system of the
Tracheata had commenced to develop, probably at first as a
series of slight ectodermal depressions which gradually
worked their way deeper and deeper into the tissues until
they formed a system of tubes, branching through the body

* Moseley on the Structure and Development of Peripatus capensis,
Phil. Trans., v. 164 (1874) ; and Balfour, Quart. Jour. Micros. Sc, vol. xix
(1879) and vol. xxiii, no. xc, p. 213 (1883).

52

and lined by a delicate chitinous cuticle. The appendages at
the front of the body were beginning to be specialised in
connection with the head and mouth opening, the first pair
becoming elongated as sensory antennae, and the three
following pairs being modified into jaws.

It is such an ancestral form as this (Proto-Tracheata)
that Peripatus represents, and that was continued upwards
to the primitive Myriapoda. Above the ancestral stage
perpetuated by Peripatus the nephridia were lost, and the
segmentation of the appendages and the formation of the
tracheae became more perfect. The Myriapoda have remained
in very much this condition, the body being still formed of
a large number of segments showing very little heteronomy.

From the ancestral Myriapoda the line leading to higher
Tracheata diverged and may be traced upwards to the base
of the series of Arachnida and Insecta. Near to the
ancestral Arachnida may be placed a few aberrant groups
with somewhat doubtful affinities. Of these the most
distinct is the Pantopoda (Pycnogonida), a class* which is
not closely allied to any of the neighbouring groups, and is
best regarded as forming a branch by itself which has
diverged from the line leading upwards from the Proto-
Tracheata and Myriapoda. In the Pantopoda the number of
segments in the body has become greatly reduced, the
posterior ones (those forming the abdomen in higher
Tracheata) being in a rudimentary condition. The append-
ages have become enormously elongated, and those around
the mouth considerably modified. The tracheae have been
altogether lost.

The Tardigrada and the Pentastomida are more closely
allied to the lower Arachnids than to the Insects, and
were probably divergent and degenerate off- shoots from the

• See Hoek, Challenger Zoological Reports, vol. iii, part x, " The
Pycnogonida," p. 145.

53

ancestral Arachnida. The extinct Trilobites were possibly
derived from this part of the Tracheate series, while the
Poecilopoda* {Limidiis) , and their ancestral allies the
Eurypterida form a branch probably from this region of the
primitive Arachnida.

From about the same point must have diverged the
ancestral series leading upwards to the higher Arachnida.
Of these the Arthrogastra {Scorpio) are probably the most
nearly in the direct line of development, while the Araneina
and the Acarida are two divergent and more modified series.
The Acarida are probably somewhat degenerate, and the
Araneina are the most highly differentiated group in this
branch of the Tracheata. In these higher Arachnids pul-
monary sacs to a large extent replace tracheae as respiratory
organs.

At the point where the ancestral Insecta (Hexapoda)
diverged from the primitive Arachnida and Pantopoda, the
number of segments in the body must have become restricted
to about seventeen or eighteen, and were arranged in three
groups — four constituting the head, three the thorax, and
ten or eleven the abdomen. At first these segments were
provided each with a pair of appendages, and the Thysanura
and Collembola (constituting together the Aptera of the
table) may be regarded as the degenerate representatives of
such a stage. They are decidedly the most nearly related of
all Insects to the primitive wingless ancestors,! and retain,
along with other ancestral characters, rudiments of abdominal
appendages (Campodea).

Above the point where the Aptera diverged, certain
important changes must have taken place. The abdominal

* See Bay Lankester, Limulus an Arachnid, Quart. Journ. Micros. Sc,
vol. xxi, Nos. Ixxxiii and Ixxxiv. (1881).

+ See Lubbook, Monograph on Collembola and Thysanura, Bay Society,
1873 ; and Balfour, Corny. Emhryol., vol. i, p. 353.

54

appendages were lost, and those of the thoracic region alone
remain as the three pairs of legs characteristic of the Insecta.
Wings then developed, as outgrowths from the dorsal region
of the second and third thoracic segments, one pair on each.
The Orthoptera most nearly represent this ancestral stage.
They are certainly more primitive than any of the higher
orders of Insects, and probably, along with their allies, the
Neuroptera, the Pseudo-Neuroptera, and the Dermatoptera
(forming altogether Packard's " Phyloptera "*), constitute a
branch which diverged about this point.

Higher up, the main stem probably split into two
branches, one leading through the ancestral Hemiptera,
after considerable modification, to the Coleoptera,t while
the other forms the line from which the ancestral Diptera,
Lepidoptera, and Hymenoptera successively diverged. Some
of the Diptera have undergone considerable degeneration.
The Hymenoptera are the most perfectly differentiated
forms. They attain to a higher level of organisation than
any of the other Tracheata, and probably occupy the highest
point among Invertebrates. The Insecta is a very extensive
class, but, notwithstanding the very large number of species
it includes, and the high position attained by some of its
members, the range of organisation in the group is com-
paratively slight. I

Keturning once more to the vermean axis, we must next
examine the upper third, the region above the origin of the
Arthropoda. Near this point, and distinctly below the Anne-
lides, is placed, on a side branch, the group Discophora or
Hirudinea. It is possible that this may not be the true

* Ann. and Mag. Nat. Hist., v. xii, p. 146 (1883).

f See Packard, Guide to Study of Insects, p. 105 (New York, 1876).

+ In regard to the modifications of the larval condition, and the meaning
of the metamorphosis of insects, see Lubbock's Origin <ind Metamorphosis
of Inure la, Nature Series (l874j.

55

position of this somewhat obscure group, and when fuller
knowledge as to its affinities has been acquired, it may turn
out that the Discophora are really allied, not to the higher
worms, but to the Platyelmia."*'

From the ancestral Chaetopoda, which compose this
upper part of the vermean axis, the Oligochseta probably
diverged in one direction, and the small group Achgeta
(Polygordius) in another. The various larval stages in the
development of Polygordius show how the metameric seg-
mentation, which is found in all the Chaetopoda, was acquired
in their ancestral history. Beginning with the Trochosphere
stage (fig. 18), successive segments are formed at the post-

Fig. 18. Trochosphere. p, prse-oral lobe ; q, post-oral or abdominal part of
body ; c, prse-oral circle of large cilia ; 2^ c, perianal circle of ciUa ; m, mouth ;
oe, oesophagus (stomodseimi) ; s, stomach ; i, intestine; a, amis; g, supra-oesophageal
ganglion ; o, eye-spot.

erior end of the body, gradually converting it into an elon-
gated worm -like form, while the greater part of the original
Trochosphere remains as the head region of the adult worm.
In the Oligochaeta and the Achaeta no appendages are present,
although in the former there are chitinous bristles (setae)
which project from the body wall. The primitive Oligochaeta
have diverged into several small groups, all of which are
hermaphrodite, and possess highly specialised reproductive
organs,. more or less resembling those of Lumbricus.

* See Bourne, Quart. Journ. Mlc. Sc, vol. xxiv, No. xcv, p. 493 (1884).

56

Above this point the ancestral Polychseta must have
acquired parapodia, the characteristic lateral appendages of
the Annelides, upon which the bundles of setae are borne.
Various processes more or less of a sensory nature were also
developed from the anterior end of the body (the prae-oral
lobe), making it more of a specialised head. From this
ancestor or some very similar form must be derived the
curious little group of Myzostomida* which are found as
ecto-parasites on the Crinoidea. They are certainly degene-
rate, and have been much modified since they left the main
stem.

The series of ancestral Vermes finally terminates by
breaking up into the progenitors of the numerous closely-
allied groups of Polychseta now existing. These are therefore
the highest members of the large and very heterogeneous
phylum Vermes. Unlike the Oligochaeta they are nearly all
dioecious and marine forms, and have no highly specialised
reproductive organs. In some of the groups (Tubicolous
forms) a considerable amount of heteronomy is exhibited.

There remains now only the great and important phylum
Chordata, which contains the whole series of vertebrate
animals. The ancestral Chordata as we have already seen
arose from the main axis of the Vermes rather low down.
Hubrecht t maintains that there is a closer relationship
between the Chordata and the Nemertida than my table
shows ; he regards the sheath of the protractile proboscis of the
Nemertean as being the origin of the notochord, and considers
that the Proto-Chordata arose from the Nemertida. Semper!

* Beard, Mitthl. Zool. Stat. Neapel., Bd. v., p. 544 (1884) ; and also
L. von Graff, Challenger Zoological Reports; " The Myzostomida," vol. x,
part xxvii (1884). From the adult anatomy Graff considers the Myzosto-
mida as being allied to the Tardigrada.

■\ Quart. Journ. Micros. Sc, vol. xxiii, no. xci, p. 349, July, 1883.

+ Arbeit, a.d., Zool.-Zoot. Instit., Wiirzburg, vols, ii and iii (1875-77).

57

and others on the contrary regard the Chordata as being
derived from some group of the Annelida, chiefly on account
of the similarity in their nephridia. I prefer to look upon
the Proto-Chordata as having arisen from a more generalised
form than either the Nemerteans or the Annelides. Their
point of origin was certainly very far back, and may have
been even lower down on the vermean axis than its position
in the table. From the striking resemblance between the
respiratory organs of Balanoglossus and those of some of the
lower Chordata it is extremely probable that the Proto-
Chordata started from ancestral Enteropneusta, and it has
been already seen (page 40) that the Enteropneusta and the
Echinodermata probably possessed a common ancestor out-
side the vermean series, consequently the Proto-Chordata,
the Enteropneusta, and the Echinodermata most probably
diverged from the main axis together as a branch which very
soon gave off the first ancestors of the Chordata in one
direction and the primitive Echinoderms in another.

The early Proto-Chordata had probably a worm-like seg-
mented body (fig. 19), with a ventrally placed mouth in front

Fig. 19. Hypothetical ancestor of the Proto-Chordata. a, anus ; b, anterior
respiratory part of the alimentary canal; c, notochord ; d, dorsally placed nervous
system ; g, supra-oesophageal ganglion forming part of brain ; i, intestine ; m, mouth ;
n, nephridium ; n^, modified anterior nephridia (hypophysis cerebri?) ; p, front end
of body formed of prae-oral lobe ; s, one of the segments or metameres.

of which projected a large prae-oral lobe. The anterior
portion of the alimentary canal had become respiratory in
function by giving off a series of lateral diverticula which
were placed in communication with the exterior by a series of
vertical slits on the side of the body. Renal organs were
present in the form of paired nephridia, placing the coelom in

58

communication with the exterior. Possibly, as Sedgwick* has
suggested, the gill slits may, like the nephridia, be merely
specialised parts of the original coelom, which was probably
formed as a series of diverticula from the archenteron. The
nervous system in place of being concentrated in a pair of
cords, running more or less ventrally, as in all the higher
Vermes and most of the groups derived from them, appears in
the primitive Chordata to have moved dorsally (probably in
consequence of some peculiarity in the habits or mode of life
of the animal), so that eventually the two bands came to
occupy a position above the alimentary canal and in the
median dorsal line (compare fig. 13, E, and fig. 19). The
anterior portion of the primitive nervous system (as seen in
the vermean ancestor) which occupied the prae-oral lobe
remained as the foremost part of the Chordate nervous
system, which no longer crossed the alimentary canal (as it
does in the higher Vermes and Arthropoda) but lay entirely
upon its dorsal side. The chief modification in the structure
of the Proto-Chordata was the formation of a stiff axis, the
notochord, running longitudinally between the alimentary
canal and the dorsal nervous system. It was probably
formed of endoderm cells from the dorsal wall of the alimen-
tary canal, but what its primary function was, or what organ
in any existing group of Vermes corresponds to it is
unknown.!

The lowest forms of the Chordata which we know are
the Urochorda or Tunicata, but they are certainly a
degenerate group which diverged from an ancestral Proto-
Chordate more or less resembling Ajipendicidaria, or better
still, the tailed larval Ascidian. In these forms the notochord

* Quart. Jourv. Micros. Sc, vol. xxiv, No. xciii, p. 43.

t Hubrecbt {Quart. Joiiru. Micros. Sc, vol. xxiii, July, 1883,) has
suggested that it was derived from the sheath of the proboscis of the
Nemerteaiis.

59

is limited to the posterior part of the body, and does not
extend forwards into the region occupied by the chief parts
of the nervous system and alimentary canal. Whether this
is the result of degeneration or indicates the primitive con-
dition of the notochord in the Proto-Chordata is uncertain.
Probably it is to some extent at least the result of degenera-
tion or adaptation of the primitive Tunicata to surrounding
conditions.

It may readily be imagined that these free-swimming
surface organisms might find it advantageous that the
alimentary canal, which performed both nutritive and respi-
ratory functions, and the main part of the nervous system
in connection with which sense organs were being devel-
oped, should be as much as possible concentrated in the
anterior part of the body so as to leave the posterior part free
to become modified into an efficient locomotory organ.
Under such circumstances it would be natural that the
notochord, the sole internal skeleton, should be restricted to
the posterior part of the body where it would form an axis
around which the muscles were placed.

From this ancestral Tunicate (fig. 20), represented at the
d

Fig. 20. Hypothetical ancestor of the Tunicata. yn, mouth; c, notochord;
b, anterior respiratory part of the aUmentary canal; d, dorsally placed nervous
system; a, anus; s, one of the segments or metameres; i, intestine; g, anterior
ganglionic part of the nervous system ; p, prse-oral lobe forming front of head ;
n^, anterior nephridia forming neural gland (hypophysis cerebri?).

present day by the Appendiculariidae two diverging lines
seem to have started, both showing degeneration. In the
first the organism remained free-swimming but acquired the
power of passing currents of water through its respiratory
system of cavities in such a way as to propel itself through

60

the water. This resulted in the great enlargement of these
cavities and the development in the body-walls of transversely
placed bands of muscle which served to drive out with force
the contained water. The posterior region of the body,
or tail, being now superseded as an organ of locomotion,
became suppressed, and thus the Thaliacea {Salpa, &c.) were
evolved.

In the second ancestral line the animal changed its
habits, and became sessile ; and, just as we find the tailed
larval Ascidian at the present day, after a short free-swim-
ming existence, becomes attached, loses its locomotory organ,
and undergoes other changes, so the ancestral Tunicate, when
it settled down on some object to lead a stationary existence,
went through a similar but gradual process of degeneration.
The tail, with its contained notochord, being no longer
necessary, became rudimentary and disappeared. The well-
developed sense organs, which were most important to a
locomotory organism, had become almost useless through
the change of life, and they also became suppressed, and, as
a result, the nervous system, which was in most intimate
connection with the sense organs, underwent considerable
degeneration. On the other hand, the alimentary canal, and
especially its respiratory portion, became greatly enlarged
and somewhat complicated. This resulted in the formation
of an ancestral form somewhat resembling the Simple Asci-
dians, and probably more like a solitary Clavelina than any
other existing form. It is probable that this creature then
acquired the power of reproducing by gemmation, so as to
form small colonies, like those of the " Social " Ascidians
(Clavelinidse). From this point two lines started. In one
series of ancestors the power of budding was retained, and
the members of the resulting colonies became more closely
united. This led up to the Compound Ascidians, which
have branched out into many smaller groups. The other set

61

of primitive Social Ascidians gradually lost the power of
reproducing by gemmation, and so became the first solitary
or Simple Ascidians.*

The long line extending from the point of separation
of the most primitive Chordate from the Enteropneusta
upwards to the place where the ancestral Tunicata diverged,
indicates the considerable amount of evolution which must
have taken place in the series of Proto-Chordata. It was
during this period that the notochord and other character-
istic structures of the Chordata, some of which are referred
to above, were gradually acquired.

The Cephalochorda (Amphioxus) are, like the Tunicata,
degenerate, and probably their ancestors diverged from the
main stem nearly at the same point. As the notochord
extends to the anterior end of the body in the Cephalo-
chorda, it is highly probable that its restriction to the
posterior part in the Tunicate larva and the Appendicu-
lariidae is, as suggested above, the result of adaptation. The
anterior part of the nervous system is in a degenerate
condition in Ainphioxus, and sense organs are very feebly
developed. The whole head is in a most rudimentary
condition, probably as the result of degeneration.

Above the Proto- Chordata! from which the Cephalo-
chorda arose, important changes took place resulting in the
formation of a series of ancestral forms which Balfour has
called the Proto-Vertebrata. In this group the notochord
was no longer the only skeleton, neural arches extending
upwards at the sides of the dorsal nervous system had made
their appearance in the mesoderm. The anterior part of the
nervous system became more highly developed so as to form

* For the phylogeny of the Simple Ascidians, see Herdman, Challenger
Zoological Reports, vol. vi, part xvi, p. 286 (1882).

t In this region of the primitive Chordata and the lower Vertebrata the
phylogeny given by Balfour has been followed (see Go7np. Embryol., vol. ii,
p. 271).

62

a brain in connection with which sense organs were evolved.
Probably as a result of this, the anterior end of the body
became greatly modified and a well-marked head was formed
by the differentiation and partial fusion of some of the
anterior segments, including those, possibly reduced in
number, in which the gill-slits were present.

From the Proto-Vertebrata the Cyclostomata were derived.
They are probably degenerate forms, and possibly represent
what was formerly a large group of primitive Vertebrata.
From these the primitive Fishes arose by a series of changes
resulting in the forms which Balfour has called the Proto-
Gnathostomata. In this group the internal skeleton became
considerably increased. Cartilaginous branchial bars were
formed in the intervals between the gill-slits. True jaws
appeared for the first time in the history of the Vertebrata.
The vertebral column and the skull, formed in the meso-
derm around the notochord and the dorsal nervous system,
became much more completely developed than they were in
the Proto-Vertebrata from which the Cyclostomata diverged.
Finally, paired fins, or limbs, made their appearance.
Probably, as suggested by Balfour and others, these were
first formed as continuous lateral fins, one extending along
each side of the body from close behind the gill-slits to near
the termination of the alimentary canal. Afterwards concen-
tration took place at two points on each side, resulting in the
formation of the primitive pectoral and pelvic tins, while the
part between became suppressed.

The Proto-Gnathostomata is represented by the existing
Elasmobranchii, a group which have probably undergone
comparatively little change and are, therefore, of great value
as showing many of the typical vertebrate systems in a more
or less primitive condition. The Holocephali are the modi-
fied descendants of an offshoot from the primitive Elasmo-
branchs.

63

The main line then passed upwards towards the higher
Fishes. This ancestral series above the common origin of
the Elasmobranchii and Holocephali, is the Proto-Ganoidei
of Balfour, * called Pneumatocoela by Bridge f because the
air-bladder probably first appeared as a distinct organ in this
group. From it the primitive Ganoids and Teleosteans
diverged together.

The Teleostei are the most highly differentiated group of
Fishes, and have been derived from the ancestral Ganoidei.
The endo skeleton which was cartilaginous in the Elasmo-
branchii and Proto-Ganoidei, becomes more or less com-
pletely ossified in the higher Fishes, and especially in the
Teleostei. In many respects the Ganoidei more closely
resemble the common ancestry. They have separated into
two groups, of which one retains in a less modified condition
the primitive characteristics, while the other resembles the
Teleosteans in the differentiations which have been effected.
The Teleostei have been evolved into a number of distinct
groups, some of which have diverged considerably from the
common ancestor.

The Dipnoi must be regarded as representing an offshoot
from the Proto-Ganoidei at the point where the common
ancestors of the Ganoids and Teleosteans diverged from the
line leading upwards to the higher Yertebrata. In some
respects the Dipnoi exhibit features seen in the less modified
Ganoidei and the Elasmobranchii, and in virtue of such
characters they are persistent Proto- Ganoids; but on the
other hand, the Dipnoi show in some points that they have
undergone evolution in the direction of the line leading
towards the primitive Amphibians. The most important
character in which they exhibit an advance upon other
groups of Fishes is the development of the air-bladder, which

* Comp. EmhryoL, v. ii, p. 272.
t Fhil. Trans., 1878, Part ii.

64

in Ganoids and Teleosteans is mainly hydrostatic in function,
into an accessory organ of respiration, which corresponds to
the lungs of higher Vertebrata.

One of the most important changes which took place in
the series of ancestral forms leading upwards from the top
of the Proto-Ganoidei was the conversion of the ichthyoptery-
gium, or fin-like limb, found in Fishes, into the cheiroptery-
gium, or true limb, characteristic of higher Vertebrates.
This change was probably effected by the elongation of the
basal cartilage (metapterygium) of the fin to form the
primitive humerus or femur, which became directed outwards
from the body instead of lying more or less parallel with it.
The other skeletal elements of the cheiropterygium are
probably derived from modified fin rays.''' The group in
which this higher form of limb was first developed is
Balfour's Proto-Pentadactyloidei. This must have split up
into two lines of descent; the one forming the common
ancestors of the Amphibia and the Labyrinthodontia, and the
other leading upwards to higher forms — the Proto-Amniota.

The Urodelous Amphibia probably represent with com-
paratively little modification the first of these two ancestral
lines. Lungs had now become completely developed, and
their establishment as the chief respiratory organs had
caused changes in the course of the circulation and in the
structure of the heart. The Anura are a somewhat more
modified group which have sprung from the Urodela, while
the Gymnophiona are clearly degenerate forms, which prob-
ably diverged from the ancestral Urodela. The extinct
Labyrinthodontia may be regarded as a lateral ofi'shoot from
the primitive Amphibians.

The Proto-Amniota divided into two great series of
ancestral forms, the primitive Sauropsida leading upwards to
Reptiles and Birds, and the early Mammalia. Just below
* Comp. EmbryoL, v. ii, p. 510.

65

this point of divergence may be placed the extinct Ichthyo-
sauria and Plesiosauria as independent branches from the
Proto-x\mniota, between the ancestral Amphibia and Eeptilia.
Their position is however a little doubtful. They may
possibly have arisen further back than they are placed in
the table, and may really be offshoots from the Proto-
Pentadactyloidei, close to, but independent of the ancestral
Amphibia.

In the primitive Sauropsida, the epidermis must have
come to take a prominent part in the formation of the exo-
skeleton. The internal skeleton also, and especially the
skull, became more completely formed and more thoroughly
ossified. This group then probably split into two ancestral
series, of which the one eventually produced the Crocodilia,
the Ornithoscelida, and the Aves ; and the other the
Dicynodontia, the Chelonia, the Pterosauria, the Lacertilia,
and the Ophidia.

The latter branch must be regarded as the more direct
continuation of the primitive Sauropsida, and is probably
most nearly represented by the Lacertilia of the present
day. They form the central group in the Reptiles just
as the Urodela do amongst Amphibians. The Ophidia are
a degenerate group of Reptiles allied to the Lizards ; they
may be regarded as having diverged from the ancestral
Lacertilia. The extinct Pterosauria branched off from the
Lacertilian stem further back, and became considerably
modified ; while the Chelonia and the extinct Dicynodontia
diverged still earlier, and probably not far above the point
where the primitive Sauropsida branched. The Chelonia
and the Dicynodontia probably left the ancestral Lizards
at much the same point, or they may possibly have had a
short common ancestry. The Chelonia have become greatly
modified, chiefly in respect to their exoskeleton and body-
form.

66

The second main branch of the Sauropsida leads upwards
to the extinct Ornithoscelida (Deinosauria), the Crocodilia
being given off as a side branch in the direction of the
other Reptilia^ and rising to a rather higher level of organisa-
tion.

The Aves are derived from the Ornithoscelida, and are
mainly distinguished from other Sauropsida by the modifica-
tion of the epidermal exoskeleton into feathers. The
extinct groups belonging to this region shew a most
interesting series of transition forms between Reptiles and
Birds, and most of the Ornithoscelida lead gradually up
through various extinct forms known as fossils to the Birds
of the present day.

The Saururae {Aixhceopteryx) were more primitive than
any existing forms, and possessed a long tail, to which how-
ever feathers were attached. From this group the remaining
Aves, in which the tail is short and formed of coalesced
vertebrae, have descended in two series : the Ratitae, in which
the sternum is retained in its primitive flat condition, and the
Carinatae, in which it is provided with a more or less well
developed keel. The Aves have become differentiated into a
large number of groups which, however, show very little
structural modification. They attain a higher level of
organisation than any of the other Sauropsida.

The first ancestral Mammalia probably arose from the
Proto-Amniota, at the base of the Sauropsida. Caldwell's*
recent investigations show that the most primitive Mammals
now existing differ from the Amphibia, and agree with Rep-
tiles and Aves in having a large amount of food-yolk in the
ovum. Very possibly, this character was acquired by the
Proto-Amniota, and inherited by the two lines of descent
(Sauropsida and Mammalia) into which that group divided.

* See Nature, vol. xxx, p. 577. 1884.

67

It was afterwards lost far back in the Mammalian series. It
is, I think, more probable that the first Mammalia and first
Sauropsida diverged from a common ancestral stock in
which the ovum had much food-yolk, than that the Mammals
were derived from some group of the Sauropsida. The
arrangement shewn in the table, besides taking account of
the condition of the ovum in the Monotremata, also explains
the very decided Amphibian characteristics of the Mammalia
pointed out by Huxley,* such as the condition of the pelvis,
the skull, and some other parts of the skeleton, if these
Amphibian characters can be regarded, as I think they
may, as being found in the Proto-Pentadactyloidei, from
which both Amphibia and Proto-Amniota are derived. If
this view is correct, then the features in which Mammalia
resemble Amphibia are really ancestral Proto-Pentadacty-
loidean characteristics, and any in which they agree with
Sauropsida only were acquired later by the Proto-Amniota.
The ancestral Mammals, — which must have developed
mammary glands for the nourishment of the young, while the
epidermic exoskeleton assumed the form of hair, and the left
fourth aortic arch became the chief artery leaving the heart,
— are most nearly represented at the present day by the
Monotremata, a group in which the brain and some parts of
the skeleton exhibit a lower stage of organisation than is
found in other Mammalia. The oviducts also remain in the
primitive condition of two tubes opening independently into
a cloaeal cavity, and the mammary glands are not well devel-
oped. The few existing forms belonging to the Monotremata
are probably the representatives of a group which diverged
from the earliest Mammals, and to which Huxley f has given
the name Prototheria. The present Monotremata are doubt-
less somewhat modified, but retain the characters of the

* Proceedings Roy. Soc, v. xxviii, p. 395 (1879).
t Proceedings Zool. Soc, for 1880, p. 649.

Prototlieria sufficiently to show that they are distinctly below
any other groups of the Mammalia.

The Marsupialia may also be regarded as the existing
representatives of an ancient group which diverged from the
ancestral stem far back, but above the origin of the
Prototheria. This group is Huxley's Metatheria ; while all
the remaining Mammalia, the descendants of the ancestral
series, after the separation of the Prototheria and Metatheria,
are the Eutheria.

The existing Marsupials show in their anatomical charac-
ters a condition intermediate between that of the Prototheria
and that found in the Eutheria. The mammary glands in
particular are much more highly developed than they are in
the Monotremata; but in the structure of the brain and some
parts of the skeleton they are still distinctly below the other
groups of the Mammalia. The Marsupialia, although
retaining the essential characters of the Metatheria, have pro-
bably undergone considerable evolution, which has affected
the less important points of structure, and given rise to
several distinct groups presenting differences mainly in
external features, and evidently related to their very diverse
habits.

In the evolution of the early Eutheria, the third group of
ancestral Mammalia,"^' one of the most important changes
was the acquisition of the allantoic placenta for the nourish-
ment of the embryo. The placenta is present in all of the
Eutheria, which are therefore called '* Placental Mammals,"
and is variously modified in the different groups.!

The Edentata were probably the first important divergent
series from the ancestral Eutheria. The existing members

* For an account of the Mammalia see Flower's article *' Mammalia," in
Encijcl. Brit., 9th edition, v. xv, p. 347.

fFor a comparative history of the Placenta, see Balfour, Com]).
KmhnjoL, v. ii, p. 198.

69

of this group have become greatly specialised in accordance
with their habits of life. Many extinct forms are known
from their fossil remains, and it is probable that the
Edentata were once a much larger groap than they are
now. They probably divided at an early period into several
distinct branches, the representatives of some of which {e.g.,
Manis and Orycteropus) are very different from the other
members of the group. The existing Edentata ought pro-
bably to be considered as having degenerated somewhat.

The Sirenia are probably the remains of a degenerate off-
shoot from the older Eutheria above the point of origin of
the Edentata. They are not closely related to any of the
higher groups of Mammals, and, therefore, probably diverged
before the Eutherian stem began to split up. In the Sirenia
the hind limbs are wanting, hair is almost entirely absent
from the skin, and the body is modified to suit the aquatic
habits of the animals.

The remaining groups of the Eutheria are all more or
less closely allied to one another, but it is very difficult to
determine the lines of their evolution. As they can be
nearly all traced back to a generalised form, with characters
which may have been common to the ancestral Ungulata,
Carnivora, and Prosimise, the probability is that the main
stem split up soon after the separation of the primitive
Sirenia, into various branches leading to the main groups of
higher Mammals. These may be arranged in two series, the
one consisting of the ancestral Ungulata, from which probably
the Cetacea, Hyracoidea, Proboscidea and Kodentia arose at
different points in the series; and the other comprising
the common ancestors of the Carnivora, lusectivora and
Cheiroptera, and of the Prosimi® and Primates.

The former of these two series is shown in the table
as diverging to the left and leading upwards through a long
series of ancestral Ungulata. The Cetacea probably arose

70

from near the base of this branch, but have diverged greatlj^
Like the Sirenia (with which they have no direct or close
relationship) they are modified to suit an aquatic existence,
and in some respects — especially in the condition of the
limbs and the almost total absence of hair— they exhibit
considerable degeneration.

The primitive Ungulata were continued upwards into the
ancestors of the Artiodactyla and Perissodactyla, the two
groups into which the existing Ungulata fall, and of
numerous extinct groups which unite the Artiodactyla and
Perissodactyla, or diverge from them to varying extents. The
ancestral Rodentia, Hyracoidea and Proboscidea, and many
extinct groups known by their fossil remains, arose from the
series of primitive Ungulates.*

The Rodentia are the most important of these divergent
branches. They have undergone very considerable evolution
since their origin, and are now a well-defined group of large
size, and containing a number of distinct families. But a
passage can be traced through the extinct forms Mesotherium
and Toxodon from undoubted Rodents to typical Ungulates.
Similarly the Proboscidea, although now a very isolated
group, may be joined to the Ungulata by the extinct
Proboscidea, the Dinotheria, and the Dinocerata. The
primitive Hyracoidea probably arose from the Ungulate stem,
close to the point of origin of the Proboscidea, but diverged
in a different direction, and the group as it is now known
(Hyrax) occupies a very isolated position. Thus from the
primitive Ungulates, which arose from the generalised
Eutherian, ancestral lines of descent diverged towards such
very different groups as the Cetacea, the Hyracoidea, the
various series of living and fossil Ungulata, the Proboscidea
and the Rodentia ; but the relations of these different

* See Flower, article " Mammalia" in Encijcl. Brit., 9th edition, v. xv,
p. 372.

71

branches to one another, and their probable lengths and
directions, are still very obscure.

The second main branch of the higher Mammalia pro-
bably split far back into the ancestral forms of two great
groups — the first of which became evolved into the Cheirop-
tera, the Insectivora, and the Carnivora ; while the second
was the stock from which the Prosimiae and the Primates
have descended.

In the first of these series, the extinct Creodontia
probably represent the early ancestral Carnivora, from
which the Insectivora diverged. The Cheiroptera agree
with the Insectivora, and with the primitive Carnivora in
many respects. Possibly they may have arisen as a side
branch from the Insectivora, and not from the early common
ancestors of Carnivora and Insectivora, as shown in the
table. The Cheiroptera have undergone great modification
since they diverged, the anterior extremities having become
converted into organs of flight. The Bats now existing fall
naturally into two main groups, which however have probably
had a long common ancestry.

The Insectivora retain many of the characters of the
primitive Carnivora, from which they are descended. They
have undergone less modification than either the Cheiroptera
or the Carnivora, and some of them probably represent
better than any other existing forms the primitive higher
Mammal from which all the groups above the Sirenia have
been derived. The aberrant genus, Galeopitliecus, is best
placed as a side branch from the Insectivora.

The primitive Carnivora must have divided into two
ancestral series, the one leading to the Carnivora proper
(Fissipedia), while the other diverged to the Seals and their
allies (Pinnipedia). The Fissipedia are probably the most
direct descendants known of the ancestral Carnivora. They
have undergone a considerable amount of evolution, and have

72

split up into three groups, the ^luroidea (Cats, etc.), the
Cynoidea (Dogs, etc.), and the Arctoidea (Bears, etc.).
Some of the extinct Caruivora shew transition forms between
these now distinct sections. The Pinnipedia have been
greatly modified in accordance with their aquatic habits.
They are most nearly allied to the Arctoid section of the
Fissipedia, and, notwithstanding the great modification, a
passage may be traced from the typical Seals (Phocidge)
through Trichechus (Walrus) and Otaria (Sea-Bear) to the
true Carnivora.

Turning now to the important line leading- upwards to
the Primates, the highest Mammals, we find that soon after
diverging from the primitive Ccrnivora a side branch is
given off which leads outwards to the ProsimisB (Lemurs).
This group has not undergone much evolution, and as it
probably arose very far back, the existing members may be
taken as representing the early ancestral forms of the
Primates. A considerable gap exists, however, between the
Lemurs and the Primates as now known. Some fossil
Prosimiae show resemblances to the primitive Ungulates, but
such features may be merely characteristics inherited by both
groups from the generalised higher Mammal which was their
common ancestor. Certain North American fossil forms
are in some respects transitional from the Prosimias to the
Primates. Probably they represent the earlier and less
specialised Lemurs, or possibly the ancestral Primates from
which the Lemurs arose. Tarsms and Chiromys must be
regarded as forming divergent branches, while the rest of the
existing Prosimiae do not exhibit much modification.

The Primates are shown in the table as forming five
more or less divergent branches from a main stem, viz.,
the Arctopitheci (Hapalidae or Marmosets), the Platyrrhini
(Cebidse or American Monkeys), the Cynomorpha (Cercopi-
thecidse or Old -World Monkeys), the Anthropomorpha (Simi-

73

idse or Man-like Monkeys), and the Anthropida (Hominidae
or Men). The Cynoraorpha and the Anthropomorpha
together constitute the group Catarrhini.

Of these the Arctopitheci probably diverged first from
the ancestral Primates. They form a small group which
has not undergone any great modification. The Platyrrhini
and the Catarrhini shew very marked differences, and con-
sequently must have diverged greatly from the common
ancestral series. Probably the Platyrrhine branch was given
.off first, and it has not risen to so high a level of organisa-
tion as the remaining Primates have.

The common ancestors of the Catarrhini and the
Anthropida, after the separation of the primitive Platyrrhini,
must have acquired the narrow partition between the nostrils
and the arrangement of teeth characteristic of man and
some of the higher monkeys. This ancestral series then
gave rise to two divergent lines of descent, the one leading
to the Catarrhini, and the other to the first men.

The Catarrhine branch probably split at an early period
into two parts — a lower, which has resulted in the Cyno-
morpha, a group containing all the ordinary Old-World Apes,
quadrupedal forms with ischial tuberosities and usually a
long tail ; and a higher, represented by the Anthropomorpha
or Anthropoid Apes, which assume a semi-erect position
and have no tail. This last group includes four genera
which are decidedly the highest of Apes, and those which
approach most nearly to the Anthropida, although, as may
be seen from the table, they are not in any sense ancestral
forms. They are: Troglodytes (the Chimpanzee), Gorilla,
Simla (the Orang), and Hylobates (the Gibbons).

The Anthropida, the highest product of evolution, pro-
bably arose from the ancestral Catarrhini. It is possible,
however, that the point of origin was on the Catarrhine
branch, near the place where the Cynomorpha and Anthropo-

G

74

morpha diverged. In the table, the more primitive and less
specialised ancestor has been adopted as the starting point.
From this early Anthropoid, or generalised Catarrhine Ape,
a series of forms arose, in which the erect position, accom-
panied by slight adaptations of the vertebral column, was
more and more perfectly assumed, while, at the same time,
the proportion of the upper to the lower limb, and of the
facial to the cranial region of the skull, was reduced, the
hallux or great toe became relatively longer, and the power
of bringing it into opposition with the other toes was
gradually lost. Finally, by the reduction in size of the male
canine teeth, and the modification of the hairy covering of
the body, these ancestral forms were converted by slight
changes into the primitive man.*

This remote progenitor, who lived at least as far back
as Pliocene times, was like all other organisms subject to
the laws of variation and heredity, and, under the influence
of Natural Selection became gradually evolved into the
ancestors of the various races or varieties of the genus
Homo. There is no reason to regard these races as con-
stituting more than one species.

* For an account of the Anthropoid Apes, and their relations to the
Anthropida, the student should consult Huxley's Anatomy of Vertebrated
Animals, p. 474, and Darwin's Descent of Man, chapters ii and xxi.

76

EXPLA.NATORY.

The lowest organisms are placed at the foot of the Table, the highest
at the top. The line, straight or zig-zag, traced from the very base
upwards to any name indicates the probable course of the evolution of
the group of animals to which the name belongs. If a line stretches
upwards it shows an advance in structure; if it is nearly horizontal it
means that Uttle or no upward evolution has taken place ; if it slopes
downwards, that indicates degeneration or degradation. The propor-
tional lengths and angles of the various lines are meant to represent
roughly the amount and the nature of the evolution which has taken
place.

In no case has the line representing the evolution of one group
been allowed to pass through another group. All existing animals
are represented as being at the ends of lines or branches.

Most of the larger groups or phyla have been enclosed in dotted
red lines in order that their limits might be easily seen at a glance.

It is scarcely necessary to point out that horizontal lines could not
be drawn across this table in such a way as to divide it into sections
representing the Fauna of the various geological periods. In order to
show that, a very different table would require to be constructed in
which distance along a line stretching upwards from the base* would
indicate merely the age of the group and not evolution or advance in
organisation as in the present table. Degeneration could not be
represented, and it would be impossible to distinguish except, perhaps,
by colour, between groups which had undergone rapid evolution and
those which had remained comparatively stationary.

* In such a table the Foraminifera and other groups of Protozoa would extend
from the base to the top of the table, although they have undergone comparatively
little evolution. Some of Haeckel's phylogenetic schemes (see Generelle Morpho-
logic der Organismen, Berlin, 1866) are of this palaeontological nature and show the
probable distribution in time of the various groups of animals and plants.

^h