By E. van Beneden[54] they were observed in a Teleostean ovum to

appear nearly simultaneously in considerable numbers in the granular

matter beneath the blastoderm. Van Beneden concludes from the

simultaneous appearance of these bodies that they develop

autogenously. Kupffer at an earlier period arrived at a similar

conclusion. My own observations on these nuclei in Elasmobranchii on

the whole support the conclusions to be derived from Lankester's,

Kupffer's and Van Beneden's observations. As mentioned above, the

nuclei in Elasmobranchii do not appear simultaneously, but

increase in number as development proceeds; and it is possible that

Van Beneden may be mistaken on this point. No evidence came before me

of derivation from pre-existing nuclei in the blastoderm. My

observations prove however that the nuclei increase by division. This

is shewn by the fact that I have found them with the spindle

modification (fig. 46, _nx´_), and that in most cases they

usually exhibit the form of a number of aggregated vesicles[55], which

is a character of nuclei which have just undergone division. It should

be mentioned however that I failed to find a spindle modification of

the nuclei in the later stages. Against these observations must be set

those of Bobretzky, according to which the nuclei in Oniscus are

really the nuclei of cells which have migrated from the blastoderm.

Bobretzky's observations do not however appear to be very conclusive.

[54] _Quart. Journ. of Micr. Science,_ Vol. XVIII. p. 41.

[55] At the time when my observations on Elasmobranchii were

carried out, this peculiar condition of the nucleus had not been

elucidated.

It must be admitted that the general evidence at our command appears

to indicate that the nuclei of the yolk in meroblastic ova originate

_spontaneously_. There is however a difficulty in accepting this

conclusion in the fact that all the other nuclei of the embryo are

descendants of the first segmentation nucleus; and for this reason it

still appears to me possible that the nuclei of the yolk will be found

to originate from the continued division of one primitive nucleus,

itself derived from the segmentation nucleus.

The existence of these nuclei in the yolk and the formation of a

distinct cell body around them is a strong piece of evidence in favour

of the view above maintained, (which is not universally accepted,)

that the part of the ovum of meroblastic ova which does not segment is

of the same nature as that which does segment, and differs only in

being relatively deficient in active protoplasm.

The following forms have meroblastic ova of the first type: the

Cephalopoda, _Pyrosoma_, Elasmobranchii, Teleostei, Reptilia,

Aves, Ornithodelphia (?). The second type of meroblastic segmentation

occurs in many Crustacea, (parasitic Copepoda, Isopoda _Mysis_,

etc.). It is also stated to be found in _Scorpio_.

The ova of the majority of groups in the animal kingdom segment

according to one of the types which have just been described. These

types are not sharply separated, but form an unbroken series,

commencing with the ovum which segments uniformly, and ending with the

meroblastic ovum.

It is convenient to distinguish the ova which segment uniformly by

some term; and I should propose for this the term alecithal[56], as

implying that they are without food-yolk, or that what little

food-yolk there is, is distributed uniformly.

[56] For this term as well as for the terms telolecithal and

centrolecithal I am indebted to Mr Lankester.

The ova in which the yolk is especially concentrated at one pole I

should propose to call telolecithal. They constitute together a

group with an unequal or partial segmentation.

The telolecithal ova may be defined in the following way: ova in which

the food-yolk is not distributed uniformly, but is concentrated at one

pole of the ovum. When only a moderate quantity of food-yolk is

present the pole at which it is concentrated merely segments more

slowly than the opposite pole; but when food-yolk is present in very

large quantity the part of the ovum in which it is located is

incapable of segmentation, and forms a special appendage known as the

yolk-sack.

There is a third group of ova including a series of types of

segmentation nearly parallel to the telolecithal group. This group

takes its start from the alecithal ovum as do the telolecithal ova,

and equally with these includes a series of varieties of segmentation

running parallel to the regular and unequal types of segmentation

which directly result from the presence of a greater or smaller

quantity of food-yolk. The food-yolk is however placed, not at one

pole, but _at the centre of the ovum_. This group of ova I

propose to name centrolecithal. It is especially characteristic of the

Arthropoda, if not entirely confined to that group.

Centrolecithal ova. As might be anticipated on the analogy of

the types of segmentation already described, the concentration of the

food-yolk at the centre of the ovum does not always take place before

segmentation, but is sometimes deferred till even the later stages of

this process.

Examples of a regular segmentation in centrolecithal ova are afforded

by Palæmon (Bobretzky) and Penæus (Hæckel). A type of unequal

segmentation like that of the Frog occurs in _Gammarus locusta_

(Beneden and Bessels), where however the formation of a central yolk

mass does not appear to take place till rather late in the

segmentation. More irregular examples of unequal segmentation are also

afforded by other Crustaceans, _e.g._ various members of the

genus _Chondracanthus_ (Beneden and Bessels) and by Myriapods. In

all these cases segmentation ends in the formation of a layer of cells

enclosing a central mass of food-yolk.

[FIG. 48. SEGMENTATION OF A CRUSTACEAN OVUM (PENÆUS). (After

Hæckel.)

The sections illustrate the type of segmentation in which the yolk

is aggregated at the centre of the ovum.

_yk._ central yolk mass.

1 and 2. Surface view and section of the stage with four segments.

In 2 it is seen that the furrows visible on the surface do not

penetrate to the centre of the ovum.

3 and 4. Surface view and section of ovum near the end of

segmentation. The central yolk mass is very clearly seen in 4.]

The peculiarity of the centrolecithal ova with regular or unequal

segmentation is that (owing to the presence of the yolk in the

interior) the furrows which appear on the surface are not continued to

the centre of the egg. The spheres which are thus distinct on the

surface are really united internally. Fig. 48, copied from Hæckel,

shews this in a diagrammatic way.

Many ova, which in the later stages of segmentation exhibit the

characteristics of true centrolecithal ova, in the early stages

actually pass through nearly the same phases as holoblastic ova.

Thus in _Eupagurus prideauxii_[57] (fig. 49), and probably in the

majority of Decapods, the egg is divided successively into two, four

and eight distinct segments, and it is not till after the fourth phase

of the segmentation that the spheres fuse in the centre of the egg.

Such ova belong to a type which is really intermediate between the

ordinary type of segmentation and that with a central yolk mass.

Eupagurus presents one striking peculiarity, viz. that the nucleus

divides into two, four and eight nuclei, each surrounded by a delicate

layer of protoplasm prolonged into a reticulum, before the ovum itself

commences to become segmented. The ovum before segmentation is

therefore in the condition of a syncytium.

[57] Mayer, _Jenaische Zeitschrift_, Vol. XI.

[FIG. 49. TRANSVERSE SECTION THROUGH FOUR STAGES IN THE SEGMENTATION

OF EUPAGURUS PRIDEAUXII. (After P. Mayer.)]

The segmentation of Asellus aquaticus[58] is very similar to that of

Eupagurus, etc. but the ovum at the very first divides into as many

segments (viz. eight) as there are nuclei.

[58] Ed. van Beneden, _Bull. d. l'Acad. roy. Belgique_,

2me série, Tom. XXVIII. No. 7, 1869, p. 54.

In Gammarus locusta the resemblance to ordinary unequal segmentation

is very striking, and it is not till a considerable number of segments

have been formed that a central yolk mass appears.

In all the above types, as segmentation proceeds, the protoplasm

becomes more and more concentrated at the surface, and finally a

superficial layer of flat blastoderm cells is completely segmented off

from the yolk below (fig. 49 D).

In cases like those of Penæus, Eupagurus, etc., the yolk in the

interior is at first nearly homogeneous, but at a later period it

generally becomes divided up partially or completely into a number of

distinct spheres, which may have nuclei and therefore have the value

of cells. In many cases nuclei have however not been demonstrated in

these yolk spheres, though probably present; yet, till they have been

demonstrated, some doubt must remain on the nature of these yolk

spheres. It is probable that _not_ all the nuclei which result

from the division of the first segmentation nucleus become concerned

in the formation of the superficial blastoderm, but that some remain

in the interior of the ovum to become the nuclei of the yolk spheres.

[FIG. 50. SEGMENTATION AND FORMATION OF THE BLASTODERM IN CHELIFER.

(After Metschnikoff.)

In A the ovum is divided into a number of separate segments. In B a

number of small cells have appeared (_bl_) which form a blastoderm

enveloping the large yolk spheres. In C the blastoderm has become

divided into two layers.]

In _Myriapods_ (_Chilognatha_) a peculiar form of segmentation has

been observed by Metschnikoff[59]. The ovum commences by undergoing a

perfectly normal, though rather irregular total segmentation. But

after the process of division has reached a certain point, scattered

masses of very small cells make their appearance on the surface of the

large spheres. These small cells have probably arisen in a manner

analogous to that which characterizes the formation of the superficial

cells of the blastoderm in the types of centrolecithal ova already

described. They rapidly increase in number and eventually form a

continuous blastoderm; while the original large segments remain in the

centre as the yolk mass. In the interesting Arachnid _Chelifer_

segmentation takes place in nearly the same manner as in Myriapods

(fig. 50).

[59] _Zeitschrift für wiss. Zool._, Vol. XXIV. 1874.

[FIG. 51. FOUR SUCCESSIVE STAGES IN THE SEGMENTATION OF THE EGG OF

TETRANYCHUS TELARIUS. (After Claparède.)]

It is clear that it is not possible in centrolecithal ova to have any

type of segmentation exactly comparable with that of meroblastic ova.

There are however some types which fill the place of the meroblastic

ova in the present group, _in as much as they are characterised by

the presence of a large bulk of food-yolk which either does not

segment, or does not do so till a very late stage in the development_.

The essential character of this type of segmentation consists in the

division of the germinal vesicle in the interior, or at the surface of

the ovum into two, four, etc. nuclei (fig. 51). These nuclei are each

of them surrounded by a specially concentrated layer of protoplasm

(fig. 51) which is continuous with a general protoplasmic reticulum

passing through the ovum [not shewn in fig. 51]. The yolk is contained

in the meshes of this reticulum in the manner already described for

other ova.

The ovum, like that of Eupagurus before segmentation, is now a

syncytium. Eventually the nuclei, having increased by division and

become very numerous, travel, unless previously situated there, to the

surface of the ovum. They then either simultaneously or in succession

become, together with protoplasm around them, segmented off from the

yolk, and give rise to a peripheral blastoderm enclosing a central

yolk mass. In the latter however many of the nuclei usually remain,

and it also very often undergoes a secondary segmentation into a

number of yolk spheres.

The eggs of Insects afford numerous examples of this mode of

segmentation, of which the egg of Porthesia[60] may be taken as type.

After impregnation it consists of a central mass of yolk which passes

without a sharp line of demarcation into a peripheral layer of more

transparent (protoplasmic) material. In the earliest stage observed by

Bobretzky there were two bodies in the interior of the egg, each

consisting of a nucleus enclosed in a thin protoplasmic layer with

stellate prolongations. This stage corresponds with the division into

two, but though the nucleus divides, the preponderating amount of yolk

prevents the egg from segmenting at the same time. By a continuous

division of the nuclei there becomes scattered through the interior of

the ovum a series of bodies, each formed of nucleus and a thin layer

of protoplasm with reticulate processes. After a certain stage some of

these bodies pass to the surface, simultaneously (in Porthesia) or in

some cases successively. At the surface the protoplasm round each

nucleus contracts itself into a rounded cell body, distinctly cut off

from the adjacent yolk.

[60] Bobretzky, _Zeit. f. wiss. Zool._, Bd. XXXI. 1878.

The cells so formed give rise to a superficial blastoderm of a single

layer of cells. Many of the nucleated bodies remain in the yolk, and

after a certain time, which varies in different forms, the yolk

becomes segmented up into a number of rounded or polygonal bodies, in

the interior of each of which one of the above nuclei with its

protoplasm is present. This process, known as the secondary

segmentation of the yolk, is really part of the true segmentation, and

the bodies to which it gives rise are true cells.

Other examples of this type may be cited. In Aphis[61] Metschnikoff

shewed that the first segmentation nucleus divides into two, each of

which takes up a position in the clearer peripheral protoplasmic layer

of the egg (fig. 52, 1 and 2). Following upon further division the

nuclei enveloped in a continuous layer of protoplasm arrange

themselves in a regular manner, and form a syncytium, which becomes

segmented into definite cells (fig. 52, 3 and 4). The existence of a

special clear superficial layer of protoplasm has been questioned by

Brandt.

[61] Metschnikoff, "Embry. Stud. Insecten," _Zeit. für wiss.

Zool._, Bd. XVI. 1866. My own observations on this

form accord in the main with those of Metschnikoff.

[FIG. 52. SEGMENTATION OF APHIS ROSAE. (Copied from Metschnikoff.)

In all the stages there is seen to be a central yolk mass surrounded

by a layer of protoplasm.

In this protoplasm two nuclei have appeared in 1, four nuclei in 2.

In 3 the nuclei have arranged themselves regularly, and in 4 the

protoplasm has become divided into a number of columnar cells

corresponding to the nuclei.

_w._ pole of the blastoderm which has no share in forming the

embryo.]

In _Tetranychus telarius_, one of the mites, Claparède found on

the surface of the ovum a nucleus surrounded by granular protoplasm

(fig. 51); which is no doubt the first segmentation nucleus. By a

series of divisions, all on the surface, a layer of cells becomes

formed round a central yolk mass. The result here is the same as in

Insects, but the nucleus with its granular protoplasm is from the

first superficial. In other cases, such as that of the common fly[62],

a layer of protoplasm is stated to appear investing the yolk; and in

this there arise simultaneously (?) a number of nuclei at regular

intervals, around each of which the protoplasm separates itself to

form a distinct cell. Closely allied is the type observed by

Kowalevsky in Apis. Development here commences by the appearance of a

number of protoplasmic prominences, each forming a cell provided with

a nucleus, the nuclei having no doubt been formed by previous division

in the interior of the ovum. They appear at the edge of the yolk, and

are separated from one another by short intervals. Shortly after their

appearance a second batch of similar bodies appears, filling up the

interspaces between the first-formed prominences. In the fresh-water

Gammarus fluviatilis the protoplasm is stated first of all to collect

at the centre of the ovum, where no doubt the segmentation nucleus

divides. Subsequently cells appear at numerous points on the surface,

and by repeated division constitute an uniform blastoderm investing

the central yolk mass. This mode of formation of the blastoderm is

closely allied to that observed by Kowalevsky in Apis.

[62] _Vide_ Weismann, _Entwicklung d. Dipteren_; and

Auerbach, _Organologische Studien_.

Between ova with a segmentation like that of Insects, and those with a

segmentation like that of Penæus, there is more than one intermediate

form. The Eupagurus type, with the division of the first nucleus into

eight nuclei before the division of the ovum, must be regarded in this

light; but the most instructive example of such a transitional type of

segmentation is that afforded by Spiders[63].

[63] _Vide_ Ludwig, _Zeit. f. wiss. Zool._, 1876.

[FIG. 53. THREE STAGES IN THE SEGMENTATION OF PHILODROMUS LIMBATUS.

(After Hub. Ludwig.)]

The first phenomenon which can be observed after impregnation is the

conglomeration of the yolk spheres into cylindrical columns, which

finally assume a radiating form diverging from the centre of the egg.

In the centre of the radiate figure is a protoplasmic mass, probably

containing a nucleus, which sends out protoplasmic filaments

through the columns (fig. 53 A). After a certain period of repose the

figure becomes divided into two rosette-like masses, which remain

united for some time by a protoplasmic thread: this thread is finally

ruptured (fig. 53 B). The whole egg does not in this process divide

into two segments, but merely the radiate figure, which is enclosed in

a finely granular material. The two rosettes next become

simultaneously divided, giving rise to four rosettes (fig. 53 C): and

the whole process is repeated with the same rhythm as in a regular

segmentation till there are formed thirty-two rosettes in all (fig. 54

A). The rosettes by this time have become simple columns, which by

mutual pressure arrange themselves radiately around the centre of the

egg, which however they do not quite reach.

[FIG. 54. SURFACE VIEW AND OPTICAL SECTION OF A LATE STAGE IN THE

SEGMENTATION OF PHILODROMUS LIMBATUS (Koch). (After Hub. Ludwig.)

_bl._ blastoderm; _yk._ yolk spheres.]

When only two rosettes are present the protoplasm with its nucleus

occupies a central position in each rosette, but gradually, in the

course of the subsequent subdivisions, it travels towards the

periphery, and finally occupies, when the stage with thirty-two

rosettes is reached, a peripheral position. The peripheral protoplasm

next becomes separated off as a nucleated layer (fig. 54 B). It forms

the proper blastoderm, and in it the nuclei rapidly multiply and

finally around each an hexagonal or polygonal area of protoplasm is

marked off; and a blastoderm, formed of a single layer of flattened

cells, is thus constituted. The columns within the blastoderm now form

(fig. 54 B) more or less distinct masses, which are stated by Ludwig

to be without protoplasm.

From observations of my own I am inclined to differ from Ludwig as to

the nature of the parts within the blastoderm. My observations have

been made on _Agelena labyrinthica_ and commence at the close of

the segmentation. At this time I find a superficial layer of flattened

cells, and within these a number of large polyhedral yolk cells. In

many, and I believe all, of the yolk cells there is a nucleus

surrounded by protoplasm. It is generally placed at one side and not

in the centre of a yolk cell, and the nuclei are so often double that

I have no doubt they are rapidly undergoing division. It appears to me

probable that, at the time when the superficial layer of protoplasm is

segmented off from the yolk below, the nuclei undergo division, and

that a nucleus with surrounding protoplasm is left with each yolk

column. For further details _vide_ Chapter on Arachnida.

Although by the close of the segmentation the protoplasm has travelled

to a superficial position, it may be noted that at first it forms a

small mass in the centre of the egg, and only eventually assumes its

peripheral situation. It is moreover clear that in the Spider's ovum

there is, so to speak, an attempt at a complete segmentation, which

however only results in an arrangement of the constituents of the ovum

in masses round each nucleus, and not in a true division of the ovum

into distinct segments.

It seems very probable that Ludwig's observations on the segmentation

of Spiders only hold good for species with comparatively small ova.

In connection with the segmentation of the Insects' ovum and allied

types it should be mentioned that Bobretzky, to whose observations we

are largely indebted for our knowledge of this subject, holds somewhat

different views from those adopted in the text. He regards the nuclei

surrounded by protoplasm, which are produced by the division of the

primitive segmentation nucleus, as so many distinct cells. These cells

are supposed to move about freely in the yolk, which acts as a kind of

intercellular medium. This view does not commend itself to me. It is

opposed to my own observations on similar nuclei in the Spiders. It

does not fit in with our knowledge of the nature of the ovum, and it

cannot be reconciled with the segmentation of such types as Spiders or

even Eupagurus, with which the segmentation in Insects is undoubtedly

closely related.

The majority if not all the cases in which a central yolk mass is

formed occur in the Arthropoda, in which group centrolecithal ova are

undoubtedly in a majority. In Alcyonium palmatum the segmentation

appears however to resemble that of many insects.

One or two peculiar varieties in the segmentation of ova of this type

may be spoken of here. The first one I shall mention is detailed in

the important paper of E. Van Beneden and Bessels which I have already

so often had occasion to quote: it is characteristic of the eggs of

most of the species of Chondracanthus, a genus of parasitic

Crustaceans. The ovum divides in the usual way but somewhat

irregularly into 2, 4, 8 segments which meet in a central yolk mass;

but after the third division instead of each segment dividing into two

equal parts it divides _at once_ into four, and the division into

four having started, reappears at every successive division. Thus the

number of the segments at successive periods is 2, 4, 8, 32, 128, etc.

In another peculiar case, an instance of which[64] is afforded by

_Asellus aquaticus_, after each of the earlier segmentations all

the segments fuse and become indistinguishable, but at the succeeding

segmentation double the number of segments appears.

Although, as has been already stated, it does not seem possible to

have a true meroblastic segmentation in centrolecithal ova, it does

nevertheless appear probable that the apparent cases of a meroblastic

segmentation in the Arthropoda are derivatives of this type of

segmentation. The manner in which the one type might pass into the

other may perhaps be explained by the segmentation in _Asellus

aquaticus_[65]. In this ovum large segments are at first formed

around a central yolk mass, in the peculiar manner mentioned in the

previous paragraph, but at the close of the first period of

segmentation minute cells, which eventually form a superficial

blastoderm, are produced from the yolk cells. They do not however

appear at once round the whole periphery of the egg, but at first only

on the ventral surface and later on the dorsal surface. If the amount

of food-yolk in the egg were to increase so as to render the formation

of the yolk cells impossible, and at the same time the formation of

the blastodermic cells were to take place at the commencement, instead

of towards the close of the segmentation, a mass of protoplasm with a

nucleus might first appear at the surface on the future ventral side

of the egg, then divide in the usual way for meroblastic ova, and give

rise to a layer of cells gradually extending round to the dorsal

surface. A meroblastic segmentation might perhaps be even more easily

derived from the type found in Insects. It is probable that the cases

of Scorpio, Mysis, Oniscus, the parasitic Isopoda, and some parasitic

Copepoda belong to this category; and it may be noticed that in these

cases the blastopore would be situated on the dorsal and not on the

ventral side of the ovum. The morphological importance of this latter

fact will appear in the sequel.

[64] [65] Ed. van Beneden, _Bull. Acad. Belgique_, Vol. XXVIII.

1869.

The results arrived at in the present section may be shortly restated

in the following way.

(1) A comparatively small number of ova contain very little or no

food-yolk embedded in their protoplasm; and have what food-yolk may be

present distributed uniformly. In such ova the segmentation is

regular. They may be described as alecithal ova.

(2) The distribution of food-yolk in the protoplasm of the ovum

exercises an important influence on the segmentation.

The rapidity with which any part of an ovum segments varies _ceteris

paribus_ with the relative amount of protoplasm it contains; and

the size of the segments formed varies inversely to the relative

amount of protoplasm. When the proportion of protoplasm in any part of

an ovum becomes extremely small, segmentation does not occur in that

part.

Ova with food-yolk may be divided into two great groups according to

the eventual arrangement of the food-yolk in the protoplasm. In one of

these, the food-yolk when present is concentrated at the vegetative

pole of the ovum. In the other group it is concentrated at the centre

of the ovum. Ova belonging to the former group are known as

telolecithal ova, those to the latter as centrolecithal.

In each group more than one type may be distinguished. In the first

group these types are (1) unequal segmentation, (2) partial

segmentation. The features of these three types have been already so

fully explained that I need not repeat them here.

In the second group there are three distinct types, (1) equal

segmentation, (2) unequal segmentation. These two being externally

similar to the similarly named types in the first group. (3)

Superficial segmentation. This is unlike anything which is present in

the first group, and is characterized by the appearance of a

superficial layer of cells round a central yolk mass. These cells may

either appear simultaneously or successively, and their nuclei are

derived from the segmentation within the ovum of the first

segmentation nucleus.

The types of ova in relation to the characters of the segmentation may

be tabulated in the following way:

_Segmentation._

(1) alecithal } regular

ova }

(2) telolecithal } (_a_) unequal

ova } (_b_) partial

(3) centrolecithal } (_a_) regular (with segments united in

ova } central yolk mass)

} (_b_) unequal " " " "

} (_c_) superficial.

Although the various types of segmentation which have been described

present very different aspects, they must nevertheless be looked on as

manifestations of the same inherited tendency to division, which

differ only according to the conditions under which the tendency

displays itself.

This tendency is probably to be regarded as the embryological

repetition of that phase in the evolution of the Metazoa, which

constituted the transition from the protozoon to the metazoon

condition.

From the facts narrated in this chapter the reader will have gathered

that similarity or dissimilarity of segmentation is no safe guide to

affinities. In many cases, it is true, a special type of segmentation

may characterize a whole group; but in other cases very closely allied

animals present the greatest differences with respect to their

segmentation; as for instance the different species of the genus

Gammarus. The character of the segmentation has great influence on the

early phenomena of development, though naturally none on the adult

form.

EXTERNAL FEATURES OF SEGMENTATION.

(105) E. Haeckel. "Die Gastrula u. Eifurchung." _Jenaische

Zeitschrift_, Vol. IX. 1877.

(106) Fr. Leydig. "Die Dotterfurchung nach ihrem Vorkommen in d.

Thierwelt u. n. ihrer Bedeutung." _Oken Isis._ 1848.

PART I.

_SYSTEMATIC EMBRYOLOGY._

PART I.

SYSTEMATIC EMBRYOLOGY.

INTRODUCTION.

In all the Metazoa the segmentation is followed by a series of changes

which result in the grouping of the embryonic cells into definite

layers, or membranes, known as the germinal layers. There are always

two of these layers, known as the epiblast and hypoblast; and in the

majority of instances a third layer, known as the mesoblast, becomes

interposed between them. It is by the further differentiation of the

germinal layers that the organs of the adult become built up. Owing to

this it is usual, in the language of Embryology, to speak of the

organs as derived from such or such a germinal layer.

At the close of the section of this work devoted to systematic

embryology, there is a discussion of the difficult questions which

arise as to the complete or partial homology of these layers

throughout the Metazoa, and as to the meaning to be attached to the

various processes by which they take their origin; but a few words as

to the general fate of the layers, and the general nature of the

processes by which they are formed, will not be out of place here.

Of the three layers the epiblast and hypoblast are to be regarded as

the primary. The epiblast is essentially the primitive integument, and

constitutes the protective and sensory layer. It gives rise to the

skin, cuticle, nervous system, and organs of special sense. The

hypoblast is essentially the digestive and secretory layer, and gives

rise to the epithelium lining the alimentary tract and the glands

connected with it.

The mesoblast is only found in a fully developed condition in the

forms more highly organized than the Coelenterata. It gives origin

to the general connective tissue, internal skeleton, the muscular

system, the lining of the body cavity, the vascular, and excretory

systems. It probably in the first instance originated from

differentiations of the two primary layers, and in all groups with a

well-developed body cavity it is divided into two strata. One of them

forms part of the body wall and is known as the somatic mesoblast, the

other forms part of the wall of the viscera and is known as the

splanchnic mesoblast.

A very large number not to say the majority of organs are derived from

parts of two of the germinal layers. Many glands for instance have a

lining of hypoblast which is coated by a mesoblastic layer.

[FIG. 55. DIAGRAM OF A GASTRULA. (From Gegenbaur.)

_a._ blastopore; _b._ archenteron; _c._ hypoblast; _d._ epiblast.]

The processes by which the germinal layers take their origin are

largely influenced by the character of the segmentation, which, as was

shewn in the last chapter, is mainly dependent on the distribution of

the food-yolk. When the segmentation is regular, and results in the

formation of a blastosphere, the epiblast and hypoblast are usually

differentiated from the uniform cells forming the wall of the

blastosphere in one of the two following ways.

(1) One-half of the blastosphere may be pushed in towards the other

half. A two-layered hemisphere is thus established which soon

elongates, while its opening narrows to a small pore (fig. 55). The

embryonic form produced by this process is known as a gastrula. The

process by which it originates is known as embolic invagination, or

shortly invagination. Of the two layers of which it is formed the

inner one (_c_) is known as the hypoblast and the outer (_d_) as the

epiblast, while the pore leading into its cavity lined by the

hypoblast is the blastopore (_a_). The cavity itself is the

archenteron (_b_).

(2) The cells of the blastosphere may divide themselves by a process

of concentric splitting into two layers (fig. 56, 3). The two layers

are as before the epiblast and hypoblast, and the process by which

they originate is known as delamination. The central cavity or

archenteron (_F_) is in the case of delamination the original

segmentation cavity; and not an entirely new cavity as in the case of

invagination. By the perforation of the closed two-walled vesicle

resulting from delamination an embryonic form is produced which cannot

be distinguished in structure from the gastrula produced by

invagination (fig. 56, 4). The opening (_M_) in this case is not

however known as the blastopore but as the mouth.

[FIG. 56. DIAGRAM SHEWING THE FORMATION OF A GASTRULA BY

DELAMINATION. (From Lankester.)

Fig. 1. Ovum.

Fig. 2. Stage in segmentation.

Fig. 3. Commencement of delamination after the appearance of a

central cavity.

Fig. 4. Delamination completed, mouth forming at _M_.

In fig. 1, 2 and 3 _Ec._ is ectoplasm, and _En._ is entoplasm.

In fig. 4 _Ec._ is epiblast and _En._ hypoblast.]

When segmentation does not take place on the regular type the

processes above described are as a rule somewhat modified. The yolk is

usually concentrated in the cells which would, in the case of a simple

gastrula, be invaginated. As a consequence of this, these cells become

(1) distinctly marked off from the epiblast cells during the

segmentation; and (2) very much more bulky than the epiblast cells.

The bulkiness of the hypoblast cells necessitates a modification of

the normal process of embolic invagination, and causes another process

to be substituted for it, viz. the growth of the epiblast cells as a

thin layer over the hypoblast. This process (fig. 57) is known as

epibolic invagination. The point where the complete enclosure of the

hypoblast cells is effected is known as the blastopore. All

intermediate conditions between epibolic and embolic invagination have

been found.

[FIG. 57. TRANSVERSE SECTION THROUGH THE OVUM OF EUAXES DURING AN

EARLY STAGE OF DEVELOPMENT. (After Kowalevsky.)

_ep._ epiblast; _ms._ mesoblastic band; _hy._ hypoblast.]

[FIG. 58. TWO STAGES IN THE DEVELOPMENT OF STEPHANOMIA PICTUM.

(After Metschnikoff.)

A. Stage after the delamination. _ep._ epiblastic invagination to

form pneumatocyst.

B. Later stage after the formation of the gastric cavity in the

solid hypoblast. _po._ polypite; _t._ tentacle; _pp._ pneumatophore;

_ep._ epiblastic invagination to form pneumatocyst; _hy._ hypoblast

surrounding pneumatocyst.]

In delamination, when the segmentation is not uniform, or when a solid

morula is formed, the differentiation of the epiblast and hypoblast is

effected by the separation of the central solid mass of cells from the

peripheral cells (fig. 58 A).

In the case of epibolic invagination as well as in that of the type of

delamination just spoken of, the archenteric cavity is in most cases

secondarily formed in the solid mass of hypoblast (fig. 58 B).

In ova with a partial segmentation there is usually some modification

of the epibolic gastrula.

Many varieties are found in the animal kingdom of the types of

invagination and delamination just characterized, and in not a few

forms the layers originate in a manner which cannot be brought into

connection with either of these processes.

[FIG. 59. EPIBOLIC GASTRULA OF BONELLIA. (After Spengel.)

A. Stage when the four hypoblast cells are nearly enclosed.

B. Stage after the formation of the mesoblast has commenced by an

infolding of the lips of the blastopore.

_ep._ epiblast; _me._ mesoblast; _bl._ blastopore.]

The mesoblast usually originates subsequently to the two primary

layers. It then springs from one or both of the other layers, but its

modes of origin are so various that it would be useless to attempt to

classify them here. In cases of invagination it often arises at the

lips of the blastopore (fig. 57 and 59), and in other cases part of it

springs as paired hollow outgrowths of the walls of the archenteron.

Such outgrowths are shewn in fig. 60, B and C at _pv_. The cavity

of the outgrowths forms the body cavity, and the walls of the

outgrowths the somatic and splanchnic layers of mesoblast (fig. C.

_sp._ and _so._). The archenteron is in part always converted into a

section of the permanent alimentary tract and the section of the

alimentary tract so derived is known as the mesenteron. There are

however usually two additional parts of the alimentary tract, known as

the stomodaeum and proctodaeum, derived from epiblastic invaginations.

They give rise respectively to the oral and anal extremities of the

alimentary tract.

[FIG. 60. THREE STAGES IN THE DEVELOPMENT OF

SAGITTA. (A and C after Bütschli and B after Kowalevsky.) The

three embryos are represented in the same positions.

A. Represents the gastrula stage.

B. Represents a succeeding stage in which the primitive archenteron

is commencing to be divided into three parts, the two lateral of

which are destined to form the mesoblast.

C. Represents a later stage in which the mouth involution (_m_) has

become continuous with alimentary tract, and the blastopore has

become closed.

_m._ mouth; _al._ alimentary canal; _ae._ archenteron; _bl._ _p._

blastopore; _pv._ perivisceral cavity; _sp._ splanchnic mesoblast;

_so._ somatic mesoblast; _ge._ generative organs.]

BIBLIOGRAPHY.

(107) K. E. von Baer. "Ueb. Entwicklungsgeschichte d. Thiere."

Königsberg, 1828-1837.

(108) C. Claus. _Grundzüge d. Zoologie._ Marburg und Leipzig, 1879.

(109) C. Gegenbaur. _Grundriss d. vergleichenden Anatomie._ Leipzig,

1878. _Vide_ also Translation. _Elements of Comparative Anatomy._

Macmillan and Co., 1878.

(110) E. Haeckel. _Studien z. Gastræa-Theorie_. Jena, 1877, and also

_Jenaische Zeitschrift_, Vols. VIII. and IX.

(111) E. Haeckel. _Schöpfungsgeschichte._ Leipzig. _Vide_ also

Translation. _The History of Creation._ King and Co., London, 1876.

(112) E. Haeckel. _Anthropogenie._ Leipzig. _Vide_ also Translation.

_Anthropogeny_ (Translation). Kegan Paul and Co., London, 1878.

(113) Th. H. Huxley. _The Anatomy of Invertebrated Animals._

Churchill, 1877.

(114) E. R. Lankester. "Notes on Embryology and Classification."

_Quart. J. of. Micr. Science_, Vol. XVII. 1877.

(115) A. S. P. Packard. _Life Histories of Animals, including Man, or

Outlines of Comparative Embryology._ Holt and Co., New York, 1876.

(116) H. Rathke. _Abhandlungen z. Bildung und Entwicklungsgesch. d.

Menschen u. d. Thiere._ Leipzig, 1833.

CHAPTER IV.

DICYEMIDÆ AND ORTHONECTIDÆ.

DICYEMIDÆ.

The structure and development of these remarkable parasites in the

renal organs of the Cephalopoda have recently been greatly elucidated

by the researches of E. van Beneden; and although a male element has

not been discovered, yet the embryos originate from bodies which have

a close similarity to ordinary ova.

Van Beneden has shewn that Dicyema consists in the adult state of (1)

a single layer of ciliated epiblast cells, somewhat modified

anteriorly to form a cephalic enlargement; and of (2) one large

nucleated hypoblast cell enclosed within the epiblast. There are two

kinds of embryo, both developed from germs which originate in the

hypoblast cell. The two kinds of embryo arise in individuals of

somewhat different forms. The one kind, called by Van Beneden the

vermiform embryo, arises in the more elongated and thinner examples of

Dicyema which have been named Nematogens. These embryos pass directly

into the parent form without metamorphosis.

The second kind of embryo, called infusoriform, is very different from

the parent, and has a free existence. Its eventual history is not

known. It originates in the shorter and thicker individuals of

Dicyema; which have been called Rhombogens.

The Vermiform Embryos. The germs or cells which give rise to

the vermiform embryos originate endogenously in the protoplasmic

reticulum of the axial hypoblast cell. They appear as small but

well-defined spheres, with a minute body in the centre. In

these spheres a cortical layer becomes differentiated, which gradually

increases in thickness and gives rise to the body of a cell, the

nucleus and nucleolus of which are respectively formed from the inner

part of the original sphere and the minute central body. These germs

can originate in all parts of the hypoblast cell and are frequently

very numerous.

[FIG. 61. A. GASTRULA STAGE OF DICYEMA TYPUS. B. VERIFORM EMBRYO OF

DICYEMA TYPUS. (From Gegenbaur, after E. van Beneden.)]

The germ when completely formed undergoes a segmentation very similar

to that of an ordinary ovum. It divides first into two and then into

four approximately equal segments. Of the four segments one, however,

remains passive for the remainder of the development. The other three

divide and arrange themselves so as partially to enclose in a cup-like

fashion the passive cell (fig. 61 A). The six cells resulting from

their division again divide, giving rise to twelve cells, which nearly

enclose the passive cell, leaving only a small aperture at one point.

The whole process by which the central cell becomes enclosed is, as E.

van Beneden points out, identical with a gastrula formation by

epibole, and the space where the central cell is left uncovered is the

blastopore. The central cell itself gives origin to the hypoblast cell

of the adult, and the peripheral cells to the epiblast.

By this time the embryo has assumed an oval form, and the blastopore

is situated at the pole of the long axis of the oval where the

cephalic enlargement is eventually formed.

The subsequent development consists mainly in the closure of the

blastopore, and an increase in the number of the epiblast cells.

Before the development is completed, and while the embryo is still in

the body of the parent, two germs, destined themselves to give rise to

fresh embryos, appear in the hypoblast cell, one on each side of the

nucleus (fig. 61 B). The embryo continues to elongate, while the

anterior cells become converted into the polar cells. Cilia appear

simultaneously over the general surface, and the embryo makes its way

out of the body of the parent, usually at the cephalic pole, and

becomes itself parasitic in the renal organ of the host in which it

finds itself. At the time of birth the embryo may contain a

number of germs and sometimes even developing embryos.

Infusoriform Embryos. The infusoriform embryos are capable of

living in sea-water and almost certainly lead a free existence. In

their most fully developed condition so far known they have the

following rather complicated structure (fig. 62 D, E, F, G).

The body is somewhat pyriform, with a blunt extremity which is

directed forwards in swimming, and a more pointed extremity directed

backwards. The former may be spoken of as the anterior, and the latter

as the posterior extremity or tail. At the anterior extremity are

situated a pair of refractive bodies (_r_) which lie above an

unpaired organ which may be called the urn.

The structure of the urn, the refractive bodies, and the tail may be

dealt with in succession.

The urn consists of three parts: (1) a wall (_u_), (2) a lid

(_l_), and (3) contents (_gr_). The wall of the urn is hemispherical

in form, and composed of two halves in apposition (fig. F). Its

concavity is directed forwards, and in its edge are imbedded a number

of rod-like corpuscles which appear as a ring near the surface in a

full-face view (fig. D). The lid has the form of a low pyramid with

its apex directed outwards. It is made up of four segments (fig. D).

The contents of the urn, which completely fill up its cavity, are four

polynuclear cells arranged in the form of a cross which appear with

low powers as granular bodies (fig. F). They are frequently ejected,

apparently at the will of the embryo.

The refractive bodies (_r_), two in number, one on each side of

the middle line, are composed of a material which is not of a fatty

nature, and which is passive to the majority of reagents. Each is

enveloped in a special capsule, and at times more than one refractive

body is present in each capsule. The tail is a conical structure

formed of ciliated granular cells.

No plausible guess has been made as to the function either of the urn

or of the refractive bodies.

The infusoriform embryos originate from germs, which have however a

different origin to the germs of the vermiform embryos. One to five

cells appear in the axial hypoblast cell, in a way not clearly

made out, and each of them gives rise by an endogenous process to

several generations of cells, all of which develop into infusoriform

embryos.

[FIG. 62. INFUSORIFORM EMBRYO OF DICYEMA.

A. B. C. Three of the later stages in the development.

D. E. F. Three different views of the full-grown larva. D. from the

front, E. from the side, and F. from above.

G. side view of urn.

_u._ wall of urn; _l._ lid of urn; _r._ refractive bodies; _gr._

granular bodies filling the interior of the urn.]

The primitive cell is called by Van Beneden a Germogen. In its

protoplasm a number of germs first appear endogenously, but the

nucleus of the germogen does not assist in their formation. They

eventually become detached from the parent cell, around which they are

concentrically arranged. A second and then a third generation of germs

are formed in the same way, till the whole of the protoplasm of the

primitive cell is absorbed in the formation of these germs, and

nothing of it remains but the nucleus. The germs so formed are

arranged in about three concentric layers, of which the innermost is

the youngest. One to five masses of germs may be present in a single

Rhombogen. The germs undergo a division, in the course of which their

nuclei exhibit very beautifully a spindle modification. In the course

of the segmentation the embryo gradually assumes its permanent form,

and four of the cells composing it can be distinguished from the

remainder by their greater size (fig. 62 A, _u_). The two largest

of these give rise to the wall of the urn, and also give origin to

four smaller cells (fig. 62 B, _gr_) which eventually become

polynuclear and constitute the four granular cells in the urn. The two

other cells become the lid of the urn. The parts of the urn lie at

first side by side, but in the course of development the cells which

form the wall of the urn travel inwards, and the four granular cells

are carried into their concavity. At the same time the cells which

form the lid of the urn alter their position so as to overlie the wall

of the urn. The two cells immediately above the urn give rise to the

refractive bodies (fig. 62 A, B, C, _r_) and the remainder of the

cells of the embryo become the tail (fig. 62 C). The embryo becomes

ciliated, and attains its nearly full development before leaving the

parental tissues. It usually passes out at the cephalic extremity.

As has already been stated, it is probable that the infusoriform

embryos leave the renal organs of their host and lead a free

existence. What becomes of them afterwards is not however known,

though there can be little doubt that they serve to carry the species

to new hosts.

Till the further development of the infusoriform embryo is known it is

not possible to arrive at a definite conclusion as to the affinities

of this strange parasite. Van Beneden is anxious to form it, on

account of its simple organization, into a group between the Protozoa

and the Metazoa. It appears however very possible that the simplicity

of its organization is the result of a parasitic existence; a view

which receives confirmation from the common occurrence of the process

of endogenous cell formation in the axial hypoblast cell. It has been

clearly shewn by Strasburger that endogenous cell formation is

secondarily derived from cell division; so that the occurrence of this

process in Dicyema probably indicates that the hypoblast was

primitively multicellular. It is not improbable that the enigmatical

infusoriform embryo may develop into a sexual form, the progeny of

which are destined to complete the cycle of development by becoming

again parasitic in the renal organ of a Cephalopod.

BIBLIOGRAPHY.

(117) E. van Beneden. "Recherches sur les Dicyemides." _Bull. d.

l'Académie roy. de Belgique_, 2e sér. T. XLI. No. 6 and T. XLII.

No. 7, 1876. _Vide_ this paper for a full account of the literature.

(118) A. Kölliker. _Ueber Dicyema paradoxum den Schmarotzer der

Venenanhänge der Cephalopoden._

(119) Aug. Krohn. "Ueb. d. Vorkommen von Entozoen, etc." _Froriep

Notizen_, VII. 1839.

ORTHONECTIDÆ.

A number of minute parasites infesting various Nemertines,

Turbellarians, and Ophiuroids have recently been studied by Giard and

Metschnikoff, the former of whom has placed them in a special group

which he calls the Orthonectidæ. They were first discovered by W. C.

McIntosh.

In the adult state they are[66] (Metschnikoff) somewhat pear-shaped

bodies formed of a kind of plasmodium of cells with irregular lobate

processes. In the interior of this body are eggs in all stages of

development. In the type observed by Metschnikoff (Intoshia gigas) the

ova undergo a regular segmentation, resulting in the formation of a

blastosphere in which an inner layer is subsequently formed by

delamination. A smaller and a larger kind of embryo are formed; but

all the embryos in each female belong to one type. The larger become

females and the smaller males.

[66] This at any rate holds true for the type investigated by

Metschnikoff. The full history of other forms is not yet known.

The female embryos are ovoid. The outer layer of cells or epiblast

becomes ciliated, and divided into nine segments, of which the second

is marked off from the remainder by the absence of cilia, and by being

provided with refractive corpuscles. The inner layer which surrounds a

central cavity, and might be supposed to be the hypoblast, becomes

according to Metschnikoff converted into ova.

The male embryos are more elongated than the female, from which they

further differ in only having six segments. The cells of the inner

layer eventually divide up into spermatozoa.

The larvæ probably become free, and while in the free state

impregnation would appear to be effected. When the female larvæ become

parasitic they undergo a metamorphosis, the stages of which have not

been observed; but in the course of which the epiblast cells probably

unite into a plasmodium.

The observations of Giard are in several points irreconcilable with

those of Metschnikoff, but from the statements of the latter it

appears possible that Giard has made two genera from the males and

females of one species; and that Giard's account of an unequal

segmentation followed by an epibolic gastrula, in one of his species,

has arisen from two segmenting ova temporarily fusing together. Giard

has given a description of internal gemmiparous reproduction, upon the

accuracy of which doubts have been thrown by Metschnikoff. The

affinities of the Orthonectidæ are as obscure as those of the

Dicyemidæ; though there can be but little doubt that their

organization has been much simplified in correlation with their

parasitic habits. The origin of the genital products in the axial

tissue is a feature they have in common with the Dicyemidæ.

BIBLIOGRAPHY.

(120) Alf. Giard. "Les Orthonectida classe nouv. d. Phylum des Vers."

_Journal de l'Anat. et de la Physiol._, Vol. XV. 1879.

(121) El. Metschnikoff. "Zur Naturgeschichte d. Orthonectidæ."

_Zoologischer Anzeiger_, No. 40-43, 1879.

[Ch. Julin. "Rech. sur l'organization et le devel. d'Orthonectides."

_Arch. Biol._ Vol. III. 1882.

E. Metschnikoff. "Untersuchungen üb. Orthonectidæ." _Zeit. f. Wiss.

Zoologie_, Vol. XXXV. 1881.

For general account of Orthonectidæ, _vide_ Spengel. _Biolog.

Centralblatt_, No. 6.]

CHAPTER V.

PORIFERA.

Although within the last few years greater advances have probably been

made in our knowledge of the development of the Porifera than of any

other group, yet there is much that is still very obscure, and it is

not possible to make general statements applying to the whole group.

Calcispongiæ. The form which has so far been most completely

worked out is _Sycandra raphanus_, one of the Calcispongiæ

(Metschnikoff, Nos. 132 and 134, F. E. Schulze, Nos. 139 and 142), and

I shall commence my account with the life history of this species.

The ovum in Sycandra as in other Spongida has the form of a naked

amoeboid nucleated mass of protoplasm. From the analogy of the other

members of the group, there is no doubt that it is fertilized by a

male spermatic element, though this has not as yet been shewn to be

the case--and the changes which accompany fertilization are quite

unknown.

[FIG. 63. SUCCESSIVE STAGES IN THE SEGMENTATION OF SYCANDRA

RAPHANUS. (Copied from F. E. Schulze.)

A. stage with eight segments still arranged in pairs, from above.

B. side view of stage with eight segments.

C. side view of stage with sixteen segments.

D. side view of stage with forty-eight segments.

E. view from above of stage with forty-eight segments.

F. side view of embryo in the blastosphere stage, eight of the

granular cells which give rise to the epiblast of the adult are

present at the lower pole.

_cs._ segmentation cavity; _ec._ granular cells which form the

epiblast; _en._ clear cells which form the hypoblast.]

The segmentation and early stages of development take place in the

tissues of the parent. The segmentation is somewhat peculiar, though a

modification of a regular segmentation. The ovum divides along a

vertical plane, first into two, and then into four equal segments. But

even when two segments are formed, each of them has one end pointed

and the other broader. The pointed ends give rise to the ciliated

cells of the future larva, and the broad ends to the granular cells.

Instead of the next division taking place, as is usually the case, in

a horizontal (equatorial) plane, it is actually effected along two

vertical planes intermediate in position between the two first

planes of segmentation. Eight equal segments are thus formed, each of

which has the form of a pyramid. All the segments are situated in a

single tier, and are so arranged as to give to the whole ovum the form

of a flat cone, the apex of which is formed by the pointed extremities

of the constituent segments (fig. 63 B). The apices of the segments do

not however quite meet, but they leave a central space, which is an

actual perforation (fig. 63 A) through the axis of the ovum, open at

both ends. The first indications of this perforation appear when only

four segments are present, and it is to be regarded as the homologue

of the segmentation cavity of other ova. The next plane of division is

horizontal (equatorial), and the apices of the eight cells are

segmented off as a tier of small cells. At the completion of this

division (fig. 63 C), the ovum is formed of sixteen cells arranged in

two superimposed tiers. The ovum now assumes somewhat the form of a

biconvex lens, in the axis of which the central perforation is still

present. At the close of the next stage, forty-eight cells are

present arranged in four tiers (fig. 63 D and E), the two outer tiers

containing eight cells each, and the two inner sixteen. The two inner

tiers probably arise by the simultaneous appearance of two equatorial

furrows dividing the original tiers into two, and by the subsequent

simple division of the cells of the two inner of the tiers so formed.

At the close of the stage the eight basal cells become granular (fig.

63 F). At the same time the central part of the segmentation cavity

becomes enlarged, while its terminal apertures become narrowed and

finally, shortly after the end of this stage, closed. The axial

perforation thus acquires the character of a closed segmentation

cavity. While the ovum itself becomes at the same time a blastosphere.

[FIG. 64. LARVA OF SYCANDRA RAPHANUS AT PSEUDOGASTRULA STAGE, IN

SITU IN THE MATERNAL TISSUES. (Copied from F. E. Schulze.)

_me._ mesoblast of adult; _hy._ collared cells forming hypoblast of

the adult; _en._ clear cells of larva which eventually become

involuted to form the hypoblast; _ec._ granular cells of larva which

give rise to the epiblast, which at this stage are partially

involuted.]

This stage nearly completes the segmentation: in the next one, the

cells of the poles of the blastosphere increase in number, and

the cells of the greater part of the blastosphere become columnar and

ciliated, (fig. 64 _en._) while the granular cells (_ec._) increase to

about thirty-two in number and appear to be (partially at least)

involuted into the segmentation cavity, reducing this latter to a mere

slit. This stage forms the last passed by the embryo in the tissues of

the parent. The general position of the embryo while still in this

situation may be gathered from fig. 64, representing the embryo _in

situ_. The embryo is always placed close to one of the radial canals.

From this situation it makes its way through the lining cells into a

canal and is thence transported to the surrounding water. By the time

the larva has become free, the semi-invaginated granular cells have

increased in bulk and become everted so as to project very much more

prominently than in the encapsuled state. To the gastrula stage, if it

deserves the name, passed through by the embryo in the tissues of the

parent, no importance can be attached.

[FIG. 65. TWO FREE STAGES IN THE DEVELOPMENT OF SYCANDRA RAPHANUS.

(Copied from Schulze.)

A. Amphiblastula stage.

B. A later stage after the ciliated cells have commenced to become

invaginated.

_cs._ segmentation cavity; _ec._ granular cells which will form the

epiblast; _en._ ciliated cells which become invaginated to form the

hypoblast.]

The larva, after it has left the parental tissues, has an oval form

and is transversely divided into two areas (fig. 65 A). One of these

areas is formed of the elongated, clear, ciliated cells, with a small

amount of pigment near their inner ends (_en._), and the other and

larger area of the thirty-two granular cells already mentioned

(_ec._). Fifteen or sixteen of these are arranged as a special ring on

the border of the clear cells. In the centre of the embryo is a

segmentation cavity (_c.s._) which lies between the granular and the

clear cells, but is mainly bounded by the vaulted inner surface of the

latter. This stage is known as the amphiblastula stage. During the

later periods of the amphiblastula stage a cavity appears in the

granular cells dividing them into two layers. After the larva has for

some time enjoyed a free existence, a remarkable series of changes

take place, which result in the invagination of the half of it formed

of the clear cells, and form a prelude to the permanent attachment of

the larva. The entire process of invagination is completed in about

half an hour. The whole embryo first becomes flattened, but especially

the ciliated half, which gradually becomes less prominent (fig. 65 B);

and still later the cells composing it undergo a true process of

invagination. As a result of this invagination the segmentation cavity

is obliterated, and the larva assumes a compressed plano-convex form,

with a central gastrula cavity, and a blastopore in the middle of the

flattened surface. The two layers of the gastrula may now be spoken of

as epiblast and hypoblast. The blastopore becomes gradually narrowed

by the growth over it of the outer row of granular cells. When it has

become very small the attachment of the larva takes place by the flat

surface where the blastopore is situated. It is effected by

protoplasmic processes of the outer ring of epiblast cells, which,

together with the other epiblast cells, now become amoeboid. They

become at the same time clearer and permit a view of the interior of

the gastrula. Between the epiblast cells and the hypoblast cells which

line the gastrula cavity there arises a hyaline structureless layer,

which is more closely attached to the epiblast than to the hypoblast,

and is probably derived from the former. A view of the gastrula stage

after the larva has become fixed is given in fig. 66.

[FIG. 66. FIXED GASTRULA STAGE OF SYCANDRA RAPHANUS. (Copied from

Schulze.)

The figure shews the amoeboid epiblast cells (_ec._) derived from

the granular cells of the earlier stage, and the columnar hypoblast

cells, lining the gastrula cavity, derived from the ciliated cells

of the earlier stage. The larva is fixed by the amoeboid cells on

the side on which the blastopore is situated.]

There would seem according to Metschnikoff's observations (No. 134) to

be a number of mesoblast cells interposed between the two primary

layers, which he derives from the inner part of the mass of granular

cells.

After invagination the cilia of the hypoblast cells can no longer be

seen, and are probably absorbed; and their disappearance is nearly

coincident with the complete obliteration of the blastopore, an event

which takes place shortly after the attachment of the larva.

Not long after the closure of the blastopore, calcareous spicules make

their appearance in the larva as delicate unbranched rods pointed at

both extremities. They appear to be formed on the mesoblast cells

situated between the epiblast and hypoblast[67]. The larva when once

fixed rapidly grows in length and assumes a cylindrical form (fig. 67

A). The sides of the cylinder are beset with calcareous spicules which

project beyond the surface, and, in addition to the unbranched forms,

spicules are developed with three and four rays as well as some with a

blunt extremity and serrated edge. The extremity of the cylinder

opposite the attached surface is flattened, and, though surrounded by

a ring of four-rayed spicules, is itself free from them. At this

extremity a small perforation is formed leading into the gastric

cavity, which rapidly increases in size and forms an exhalent osculum

(_os._). A series of inhalent apertures is also formed at the

sides of the cylinder. The relative times of appearance of the single

osculum and the smaller apertures are not constant for the different

larvæ. On the central gastrula cavity of the sponge becoming placed in

communication with the external water, the hypoblast cells lining it

become ciliated afresh (fig. 67 B, _en._) and develop the peculiar

collar characteristic of the hypoblast cells of the Spongida (_vide_

fig. 64, _hy._). When this stage of development is reached we have a

fully formed sponge of the type made known by Haeckel as Olynthus.

[67] Metschnikoff was the first to give this account of the

development of the spicules in Sycandra, but Prof. Schulze has

informed me by letter that he has arrived at the same result.

[FIG. 67. THE YOUNG OF SYCANDRA RAPHANUS SHORTLY AFTER THE

DEVELOPMENT OF THE SPICULA. (Copied from Schulze.)

A. View from the side.

B. View from the free extremity.

_os._ osculum; _ec._ epiblast; _en._ hypoblast composed of ciliated

cells. The terminal osculum and lateral pores are represented as

oval white spaces.]

When young examples of Sycandra come in contact shortly after their

attachment they appear to fuse together temporarily or else

permanently. In the latter case colonies are produced by their fusion.

Amongst other calcareous sponges the larva of _Ascandra contorta_

(Haeckel No. 126, Barrois No. 122) presents the typical amphiblastula

stage, and so probably does that of _Ascandra Lieberkühnii_ (Keller

No. 128). In _Leucandra aspera_ (Keller No. 128, Metschnikoff No. 134)

the larva passes through an amphiblastula stage, but the characters of

the cells of the two halves of the larva do not differ to nearly the

same extent as in Sycandra.

Although the majority of calcareous sponges appear to agree in their

mode of development with Sycandra, nevertheless the concordant

researches of O. Schmidt (No. 138) and Metschnikoff (No. 134) have

shewn that this is not true for the genus Ascetta (_As. primordialis_,

_clathrus_ and _blanca_).

The larvæ of these forms are very differently constituted to those of

Sycandra. They have an oval form and are composed of a single row of

ciliated columnar cells: their two extremities only differ in the

cells at one extremity being longer than those at the other.

Especially at the pole where the shorter cells are situated (Schmidt)

a metamorphosis of the cells takes place. One after the other they

lose their cilia, become granular, and pass into the interior of the

vesicle. Here they become differentiated into two classes

(Metschnikoff); one of larger and more granular cells, and the other

of smaller cells with clearer protoplasm. Cells of the former class

are mainly found at one of the poles. When the larva becomes free the

cells in the interior of the vesicle increase in number and nearly

fill up its central cavity. After a short free existence the larva

becomes fixed, and the epiblast cells lose their cilia and become

flattened. At a later period the large granular cells assume a radiate

arrangement round a central cavity and become clearly marked out as

the hypoblast cells. The smaller cells become placed between the

epiblast and hypoblast and constitute the mesoblast.

Myxospongiæ. In this group Halisarca has been investigated by Carter

(No. 123), Barrois (No. 122), Schulze (No. 141) and Metschnikoff (No.

134). The ova develop in the mesoblast, and when ripe occupy special

chambers lined by a layer of epithelial cells. Schulze has found the

spermatozoa of this genus of sponge and has been able to shew that the

sexes may be distinct, though many species of Halisarca are

hermaphrodite.

The segmentation is, roughly speaking, regular, and a segmentation

cavity is early formed, which is never, as in Calcispongiæ, open at

the poles. When the larva leaves the parent it is an oval vesicle

formed of a single layer of columnar ciliated cells. Slight

differences may be observed between the two extremities of the larvæ

of most species. One of these--the hinder extremity--is directed

backwards in swimming.

The further history of the larva has been investigated by

Metschnikoff. He has found that the interior of the vesicle becomes

gradually filled with mesoblast cells of a peculiar type, called by

him rosette-cells, which are probably derived from the walls of the

vesicle.

When the metamorphosis commences, the larva assumes a flattened form,

and cells of a new type, viz. normal amoeboid cells, grow in amongst

the rosette cells. The new cells are also derived from the epiblast.

The larvæ appear to fix themselves by the hinder extremity. The cilia

gradually disappear, and the epiblast cells flatten out and form a

kind of cuticle. For some time the larva remains in the two-layered

condition, but gradually canals (? ciliated chambers) lined by

hypoblast cells become formed. They appear as closed spaces with walls

of ciliated cells derived from the amoeboid cells, and the different

parts of the system of chambers are established independently. In _H.

pontica_ the ciliated chambers are formed before the attachment of the

larva. The development was not followed up to the formation of the

pores placing the canal system in communication with the exterior.

The young sponges at a somewhat later stage have been studied by

Schulze and Barrois. They are formed of an external layer of flattened

cells, not clearly ciliated as in the adult, within which are a normal

mesoblastic tissue, and several spherical chambers lined by ciliated

cells exactly like the ciliated chambers of the full-grown sponge.

Irregular invaginations of the epiblast give to the young sponge a

honeycombed structure. The ciliated chambers in the youngest condition

of the sponge are closed; but in slightly older examples they come

into communication with the passages lined by hypoblast, and so

indirectly with the external medium.

Ceratospongiæ. Amongst the true Ceratospongiæ the embryos of two of

the Aplysinidæ, and of Spongelia and Euspongia have been to some

extent worked out by Barrois and Schulze. The form worked out by

Barrois is called by him _Verongia rosea_. The segmentation is nearly

regular, but from the first the segments may be divided according to

their constitution into two categories. At the close of segmentation

the embryo is oval and covered by a single layer of columnar ciliated

cells; these cells may however be divided into two categories,

corresponding with those observable during the segmentation. A certain

number are coloured red and form a definite circular mass at one pole,

while the remainder, which constitute the major part of the embryo,

have a pale yellowish colour. Those at the red pole lose their cilia

in the free larva, but around the area formed by them is a special

ring of long cilia. The chief peculiarity of the embryo (made known by

Schulze) consists in the fact that the layer of cells which covers the

embryo does not, as in other sponge embryos, simply enclose a space,

but the interior of the embryo is formed of a mass of stellate cells

like the normal mesoblast of full-grown sponges.

This feature is also characteristic of the embryos of Spongelia and

Euspongia.

The embryo of the Gummineæ (_Gummina mimosa_) has been investigated by

Barrois (No. 122), and has been shewn closely to resemble the typical

larvæ of calcareous sponges; one-half being formed of _elongated

ciliated cells_ and the other of rounded granular ones.

Silicispongiæ. The development of marine silicious sponges is but very

imperfectly understood. The larvæ of various forms--Reniera

(Isodyctia), Esperia (Desmacidon), Raspailia, Halichondria,

Tethya--have been described. Barrois has shewn that the egg segments

regularly and that in the earlier stages a segmentation cavity is

present. In the later stages the embryo appears to become solid.

Externally there is a layer of ciliated cells, and within a mass of

granular matter in which the separate cells cannot be made out. The

granular matter projects at one pole, and forms a prominence possibly

equivalent to the granular cells of Sycandra. In some forms, _e.g._

Reniera, the edge of the unciliated granular prominence may be

surrounded by a row of long cilia. In later stages the granular

material may project at both poles or even at other points. One

remarkable feature in the development of the Silicispongiæ is the

appearance of spicula between the ciliated cells and the central mass,

while the larva is still free.

Professor Schulze has informed me that these spicula are developed in

mesoblast cells; while the horny fibres of the sponge are developed as

cuticular products of special mesoblast cells (spongioblasts).

The attachment and accompanying metamorphosis are so diversely

described that no satisfactory account can be given of them. The

general statements are in favour of the attachment taking place by the

posterior extremity where the granular matter projects.

Carter especially gives a very precise account, with figures, of the

attachment of the larva in this way. He also figures the appearance of

an osculum at the opposite pole[68].

[68] Keller (No. 129) has recently given an account of the

development of Halichondria (Chalinula) fertilis. He finds that

there is an irregular segmentation, followed by a partial

epibolic invagination, the inner mass of cells remaining exposed

at one pole and forming there a prominence, equivalent to the

granular prominence in the larvæ of other Silicispongiæ. The

free-swimming larva resembles the larva of other Silicispongiæ in

the possession of spicula, etc., and after becoming laterally

compressed attaches itself by one of the flattened sides. A

central cavity is formed in the interior with ciliated chambers

opening into it, and is subsequently placed in communication with

the exterior by the formation of an aperture which constitutes

the osculum.

A very elaborate account of the development of Spongilla has been

published in Russian by Ganin, of which a German abstract has also

appeared (No. 124).

The ovum undergoes a regular segmentation and becomes a solid ova

morula. An epiblast of smaller cells is early differentiated, and in

the interior of the inner cells an archenteron becomes subsequently

formed. The inner cells next become divided into an hypoblastic layer

lining the archenteron, and a mesoblastic layer between this and the

now ciliated epiblast. At the narrow hinder end of the embryo the

mesoblast becomes thickened, and largely obliterates the archenteron.

In this part of the mesoblast silicious spicula are formed. The larva

becomes attached by its hinder extremity, and in the course of this

process flattens itself out to a disc-like form. From the nearly

obliterated archenteric cavity outgrowths take place which give rise

to the ciliated chambers. These are not placed directly in

communication with the exterior, but open, if I understand Ganin

rightly, into a space in the mesoblast, which subsequently acquires an

exterior communication--the primitive osculum. The subsequent pores

and oscula are also formed as openings leading into the mesoblastic

cavity, which communicates in its turn with the ciliated chambers.

It appears that in the present unsatisfactory state of our knowledge

the larvæ of the Porifera may be divided into two groups: viz. (1)

those which have the form of a blastosphere or else of a solid morula;

(2) those which have the amphiblastula form.

In the former type the mesoblast and hypoblast are formed either from

cells budded off from the outer cells of the blastosphere or from the

solid inner mass of cells; while the outer ciliated cells become the

epiblast. This type of larva, which is found in the majority of

sponges, is very similar in its general characters and development to

many Coelenterate planulæ.

The second type of larva is very peculiar, and though in its fully

developed form it is confined to the Calcispongiæ, where it is the

usual form, a larval type with the same characters is perhaps to be

found in other sponges, _e.g._ amongst the Gumminæ, and amongst

the Silicispongiæ where one-half of the embryo is without cilia,

though in the case of the Silicispongiæ the cells of the ciliated part

of the embryo correspond to the granular cells of the larva of

Sycandra.

The later stages in the development of the larvæ of the Porifera are

not similar to anything we know of in other groups.

It might perhaps be possible to regard sponges as degraded descendants

of some Actinozoon type such as Alcyonium, with branched prolongations

of the gastric cavity, but there does not appear to me to be

sufficient evidence for doing so at present. I should rather prefer to

regard them as an independent stock of the Metazoa.

In this connection the amphiblastula larva presents some points of

interest. Does this larva retain the characters of an ancestral type

of the Spongida, and if so, what does its form mean? It is, of course,

possible that it has no ancestral meaning but has been secondarily

acquired; but, assuming that this is not the case, it appears to me

that the characters of the larva may be plausibly explained by

regarding it as a transitional form between the Protozoa and Metazoa.

According to this view the larva is to be considered as a colony of

Protozoa, one-half of the individuals of which have become

differentiated into nutritive forms, and the other half into locomotor

and respiratory forms. The granular amoeboid cells represent the

nutritive forms, and the ciliated cells represent the locomotor and

respiratory forms. That the passage from the Protozoa to the Metazoa

may have been effected by such a differentiation is not improbable on

_a priori_ grounds.

While the above view seems fairly satisfactory for the free-swimming

stage of the larval sponge, there arises in the subsequent development

a difficulty which appears at first sight fatal to it. This difficulty

is the invagination of the ciliated cells instead of the granular

ones. If the granular cells represent the nutritive individuals of the

colony, they, and not the ciliated cells, ought most certainly to give

rise to the lining of the gastrula cavity, according to the generally

accepted views of the morphology of the Spongida. The suggestion which

I would venture to put forward in explanation of this paradox involves

a completely new view of the nature and functions of the germinal

layers of adult Spongida.

It is as follows:--When the free-swimming ancestor of the Spongida

became fixed, the ciliated cells by which its movements used to be

effected must have to a great extent become functionless. At the same

time the amoeboid nutritive cells would need to expose as large a

surface as possible. In these two considerations there may, perhaps,

be found a sufficient explanation of the invagination of the ciliated

cells, and the growth of the amoeboid cells over them. Though

respiration was, no doubt, mainly effected by the ciliated cells, it

is improbable that it was completely localized in them, but they were

enabled to continue performing this function through the formation of

an osculum and pores. The collared cells which line the ciliated

chambers, or in some cases the radial tubes, are undoubtedly derived

from the invaginated cells, and, if there is any truth in the above

suggestion, the collared cells in the adult sponge must be mainly

respiratory and not digestive in function, while the epiblastic cells,

which in most cases line the inhalent passages through its

substance[69], ought to be employed to absorb nutriment. The recent

researches of Metschnikoff (No. 134) on this head shew that the

nutriment is largely carried into the mesoblast cells, which in

Sycandra appear to be derived from the granular cells, and also that

it is taken up by the cells which line the passages, though not by the

superficial epiblast cells. Whether the collared cells generally

absorb nutriment is not clear from his statements: but _he finds that

they do not do so in Silicispongiæ_.

[69] That the greater part of the flat cells which line the

passages of most Sponges are really derived from epiblastic

invaginations appears to me to be proved by Schulze's and

Barrois' observations on the young fixed stages of Halisarca.

Schulze's (No. 140) observations have however proved that the

flat cells lining the axial gastric chamber of Sycandra are

hypoblastic in origin, and the observations of Keller (No. 129)

and Ganin (No. 124) have led to the same result for the flat

epithelium lining part of the passages of the Silicispongiæ.

Professor Schulze has informed me by letter that he finds the collared

cells to be respiratory in function, while the cells derived from the

granular cells in Sycandra are nutritive. Carter[70], on the contrary,

from his observations on Spongilla, has fully satisfied himself that

the food is absorbed by the cells lining the ciliated chambers.

[70] "On the Nutritive and Reproductive Processes of Sponges."

_Ann. and Mag. of Nat. Hist._, Vol. IV. Ser. V. 1879.

If it is eventually proved by further experiments on the nutrition of

sponges, that digestion is mainly carried on by the general cells

lining the passages and the mesoblast cells, and not for the most part

by the ciliated cells, it is clear that the epiblast, mesoblast and

hypoblast of sponges will not correspond with the similarly named

layers in the Coelenterata and other Metazoa. The invaginated

hypoblast will be the respiratory layer and the epiblast and mesoblast

the digestive and sensory layers; the sensory function being probably

mainly localized in the epithelium on the surface, and the digestive

one in the epithelium lining the passages and in the mesoblast. Such a

fundamental difference in the primary function of the germinal layers

between the Spongida and the other Metazoa, would necessarily involve

the creation of a special division of the Metazoa for the reception of

the former group.

BIBLIOGRAPHY.

(122) C. Barrois. "Embryologie de quelques éponges de la Manche."

_Annales des Sc. Nat. Zool._, VI. ser., Vol. III. 1876.

(123) Carter. "Development of the Marine Sponges." _Annals and Mag. of

Nat. Hist._, 4th series, Vol. XIV. 1874.

(124) Ganin[71]. "Zur Entwicklung d. Spongilla fluviatilis."

_Zoologischer Anzeiger._ Vol. I. No. 9, 1878.

(125) Robert Grant. "Observations and Experiments on the Structure and

Functions of the Sponge." _Edinburgh Phil. J._, Vol. XIII. and XIV.,

1825, 1826.

(126) E. Haeckel. _Die Kalkschwämme_, 1872.

(127) E. Haeckel. _Studien zur Gastræa-Theorie._ Jena, 1877.

(128) C. Keller. _Untersuchungen über Anatomie und Entwicklungsgeschichte

einiger Spongien._ Basel, 1876.

(129) C. Keller. "Studien üb. Organisation u. Entwick. d. Chalineen."

_Zeit. f. wiss. Zool._, Bd. XXVIII. 1879.

(130) Lieberkühn. "Beitr. z. Entwick. d. Spongillen." Müller's

_Archiv_, 1856.

(131) Lieberkühn. "Neue Beiträge zur Anatomie der Spongien." Müller's

_Archiv_, 1859. (132) El. Metschnikoff. "Zur Entwicklungsgeschichte

der Kalkschwämme." _Zeit. f. wiss. Zool._, Bd. XXIV. 1874.

(133) El. Metschnikoff. "Beiträge zur Morphologie der Spongien."

_Zeit. f. wiss. Zool._, Bd. XXVII. 1876.

(134) El. Metschnikoff. "Spongeologische Studien." _Zeit. f. wiss.

Zool._, Bd. XXXII. 1879.

(135) Miklucho-Maklay. "Beiträge zur Kenntniss der Spongien."

_Jenaische Zeitschrift_, Bd. IV. 1868.

(136) O. Schmidt. "Zur Orientirung über die Entwicklung der Schwämme."

_Zeit. f. wiss. Zool._, Bd. XXV. 1875.

(137) O. Schmidt. "Nochmals die Gastrula der Kalkschwämme." _Archiv

für mikrosk. Anat._, Bd. XII. 1876.

(138) O. Schmidt. "Das Larvenstadium von Ascetta primordialis und Asc.

clathrus." _Archiv für mikrosk. Anatomie_, Bd. XIV. 1877.

(139) F. E. Schulze. "Ueber den Bau und die Entwicklung von Sycandra

raphanus." _Zeit. f. wiss. Zool._, Bd. XXV. 1875.

(140) F. E. Schulze. "Zur Entwicklungsgeschichte von Sycandra." _Zeit.

f. wiss. Zool._, Bd. XXVII. 1876.

(141) F. E. Schulze. "Untersuchung üb. d. Bau, etc. Die Gattung

Halisarca." _Zeit. f. wiss. Zool._, Bd. XXVIII. 1877.

(142) F. E. Schulze. "Untersuchungen üb. d. Bau, etc. Die Metamorphose

von Sycandra raphanus." _Zeit. f. wiss. Zool._, Bd. XXXI. 1878.

(143) F. E. Schulze. "Untersuchungen ü. d. Bau, etc. Die Familie

Aplysinidæ." _Zeit. f. wiss. Zool._, Bd. XXX. 1878.

(144) F. E. Schulze. "Untersuchungen ü. d. Bau, etc. Die Gattung

Spongelia." _Zeit. f. wiss. Zool._, Bd. XXXII. 1878.

[71] There is a Russian paper by the same author, containing a

full account, with clear illustrations, of his observations.

CHAPTER VI.

COELENTERATA[72].

[72] I. HYDROZOA.

1. Hydromedusæ. {_Hydroidea._

{_Trachymedusæ._

2. Siphonophora. {_Calycophoridæ._

{_Physophoridæ._

3. Acraspeda.

II. ACTINOZOA.

1. Alcyonaria. (Octocoralla.)

2. Zoantharia. (Hexacoralla.)

III. CTENOPHORA.

Hydroidea. The most typical mode of development of the Hydroidea is

that in which the segmentation leads directly to the formation of a

free ciliated two-layered larva, known since Dalyell's observations as

a planula. The planula is characteristic of almost all the Hydromedusæ

with fixed hydrosomes including the Hydrocoralla (Stylasteridæ and

Millepora), the most important exceptions being the genus Tubularia

and one or two other genera, and the fresh-water Hydra.

In a typical Sertularian the segmentation is approximately regular[73]

and ends according to the usual accounts in the formation of a solid

spherical mass of cells. A process of delamination now takes place,

which leads to the formation of a superficial layer of cubical or

pyramidal cells, enclosing a central solid mass of more or less

irregularly arranged cells.

[73] For a detailed description of the development of a single

species the reader referred to Allman's description of Laomedia

flexuosa, No. 149, p. 85 _et seq._

The embryo, in the cases in which it is still contained within the

sporosack, now begins to exhibit slight changes of form, and one

extremity of it begins to elongate. It soon becomes free, and rapidly

assumes an elongated cylindrical form, while a coating of cilia, by

means of which it moves sluggishly about, appears on its outer

surface. A central cavity appears in the interior, and the inner cells

form themselves into a definite hypoblast. The larva has now become a

planula, and consists of a closed sack with double walls. It continues

for some few days to move about, but eventually drops its cilia, and

becomes dilated at one extremity, by which it then becomes attached.

The base of attachment becomes gradually enlarged so as to form a

disc, which spreads out and is frequently divided by fissures into

radiating lobes. The free extremity becomes enlarged to form the

eventual calyx.

Over the whole exterior a delicate pellicle--the future perisarc--now

becomes secreted. Round the edge of the anterior enlargement a row of

tentacles makes its appearance. These, in the embryos of the

Tubularian genera, lie some little way behind the apex of the body.

After a certain time the perisarc, which has hitherto been continuous,

becomes ruptured in the region of the calyx, and the tentacles become

quite free. At about the same period a mouth is formed at the oral

apex.

[FIG. 68. THREE LARVA STAGES OF EUCOPE POLYSTYLA. (After

Kowalevsky.)

A. Blastosphere stage with hypoblast spheres becoming budded off

into the central cavity.

B. Planula stage with solid hypoblast.

C. Planula stage with a gastric cavity.

_ep._ epiblast; _hy._ hypoblast; _al._ gastric cavity.]

The development of Eucope polystyla (fig. 68), one of the

Campanularidæ, deviates according to Kowalevsky (No. 147) in somewhat

important points from the usual type. The whole development takes

place after the deposition of the ovum. The segmentation results in

the formation of a single-walled blastosphere with a large central

cavity (fig. 68 A). This cavity, somewhat as in Ascetta, becomes

filled up with a not clearly (?) cellular material derived from the

walls of the blastosphere, which must be regarded as the hypoblast

(fig. 68 B). The larva elongates and becomes ciliated, and the

epiblast at its two extremities becomes thickened, and is stated by

Kowalevsky also to become divided into two layers. The alimentary

cavity appears as a slit in the middle of the hypoblast (fig. 68 C).

The cilia after a time disappear, and the larva then becomes fixed by

one extremity. It flattens itself out into a disc-like form, becomes

divided into four lobes, and covered by a cuticle (perisarc). From the

disc the stalk grows out which dilates at its free extremity into the

calyx.

[FIG. 69. LONGITUDINAL SECTION THROUGH A LARVA OF TUBULARIA

MESEMBRYANTHEMUM WHILE STILL IN THE GONOPHORE. The lower end is the

oral one.

_ep._ epiblast; _hy._ hypoblast of tentacle; _en._ enteric cavity.]

In both the groups (Tubularia and Hydra) which are exceptional in not

having a ciliated planula stage, its absence may be put down to an

abbreviation of the development, and in fact a two-layered quiescent

stage, through which the embryo passes, may be regarded as

representing the planula stage.

The development of Tubularia, which has been described in detail by

Ciamician, takes place in the gonophore[74]. The segmentation is

irregular and leads to the formation of an epibolic gastrula, four

large central cells constituting the hypoblast[75]. The larva now

elongates, and grows out laterally into two processes which constitute

the first pair of tentacles. At this stage it closely resembles the

larvæ of some Medusæ. Additional tentacles are soon formed; and a

central cavity appears in the hypoblast, the cells of which have in

the meantime become more numerous (fig. 69). The tentacles are

directed towards the aboral side, which is considerably more prominent

than the oral one. They contain a hypoblastic axis. The aboral end

continues to grow and the tentacles gradually assume a horizontal

position. A constriction now appears, dividing the larva into an

aboral portion which will eventually form the stalk, and an oral

portion. At the apex of the latter a row of short tentacles--the

future oral tentacles--now appears. The larva has at this stage the

form known as Actinula. In this condition it becomes hatched, and

shortly afterwards it becomes fixed by the aboral end and grows into a

colony.

[74] _Vide_ Ciamician, _Zeit. f. wiss. Zool._, Bd. XXXII. 1879.

[75] In examining the segmentation by means of sections I have

failed to detect an epibolic gastrula or such irregularity as is

described by Ciamician. Prof. Kleinenberg informs me that he has

been equally unsuccessful.

The development of Myriothela (Allman, No. 150) takes place on the

Tubularian type. The ovum invested by a delicate capsule becomes freed

by the rupture of the gonophore, and is then taken up by the

remarkable claspers characteristic of the genus. In the claspers it

becomes fecundated and undergoes its further development. After

segmentation a gastric cavity is formed, and provisional tentacles

arise as a series of conical involutions which subsequently become

evoluted. Permanent tentacles are formed as conical papillæ on a

truncated oral process. After hatching it has a few days' free

existence, and then becomes attached, and loses its provisional

tentacles.

Although Hydra itself constitutes the simplest type of Hydrozoon, its

development, which has been fully investigated by Kleinenberg (No.

161), is in some respects a little exceptional. The segmentation is

regular, but a segmentation cavity is not formed. The peripheral layer

of cells gradually becomes converted into a chitinous membrane, which

is perhaps homologous with the perisarc of marine forms. Between the

membrane and the germ a second pellicle makes its appearance. The

above changes require about four days for their completion, but there

next sets in a period of relative quiescence which lasts for some 6-8

weeks. During this period the remaining development is completed. The

cells of the germ first fuse together. In the interior of the

protoplasm a clear excentric space arises, which gradually extends

itself and forms the rudiment of the gastric cavity. The outer shell

in the meantime becomes less firm, and is finally burst and thrown

off, owing to the expansion of the embryo within.

The outermost layer of the protoplasm becomes, relatively to the inner

layer, clear and transparent, and there thus arises an indication of a

division of the walls of the archenteric cavity into two zones, or

layers. These layers, which form the epiblast and hypoblast, are

definitely established on the appearance of cells with contractile

tails[76] in the clear outer zone, between which the interstitial

epiblast cells subsequently arise.

[76] These cells are the so-called nerve-muscle cells. Their

nature is discussed in the second part of this work.

The embryo, still forming a closed double-walled sack, elongates

itself, and at one pole its wall becomes very thin. And at this point

a rupture takes place which gives rise to the mouth. Simultaneously

with the mouth the tentacles become formed as hollow processes,

according to Mereschkowsky two being formed first and subsequently the

others in pairs. Very shortly afterwards the hitherto uniform

hypoblast becomes divided up into distinct cells. The thin inner

pellicle which persists after the rupture of the outer membrane

becomes in the meantime absorbed. With these changes the embryo

practically acquires the characters of the adult.

Trachymedusæ. Amongst the Trachymedusæ, which as has now been

satisfactorily established develop directly without alternations of

generations, the embryology of species both of the Geryonidæ and the

Æginidæ has been studied.

In all the types so far investigated the hypoblast is formed by

delamination, and there is a more or less well-marked planula stage.

[FIG. 70. DIAGRAMMATIC FIGURE SHEWING THE DELAMINATION OF THE OVUM

OF GERYONIA. (Copied from Fol.)

_cs._ segmentation cavity; _a._ endoplasm; _b._ ectoplasm. The

dotted lines shew the course of the next planes of division.]

The development of Geryonia (Carmarina) hastata has been studied by

Fol (No. 155) and Metschnikoff (No. 163)[77]. The ovum, when laid, is

invested by a delicate vitelline membrane and mucous covering. Its

protoplasm is formed of an outer granular and dense layer, and a

central mass of a more spongy character. The segmentation is complete

and regular, and up to the time when thirty-two segments have appeared

each segment is composed of both constituents of the protoplasm of the

ovum. A segmentation cavity appears when sixteen segments are formed,

and becomes somewhat larger at the stage with thirty-two. At this

stage the process of delamination commences. Each of the thirty-two

segments, as shewn in the accompanying diagram (fig. 70), becomes

divided into two unequal parts. The smaller of these is formed almost

entirely of granular material; the larger contains portions of both

kinds of protoplasm. In the next segmentation the thirty-two large

cells only are concerned, and in each of these the line of division

passes between the granular and the transparent protoplasm. The

sixty-four lenticular masses of granular protoplasm thus formed

constitute an outer closed epiblastic vesicle, within which the

thirty-two masses of transparent protoplasm form an hypoblastic

vesicle. The embryo at this stage is shewn in optical section in fig.

71.

[77] In the succeeding account I have followed Fol, who differs

in some minor points from Metschnikoff.

The epiblastic vesicle now grows rapidly, while the hypoblastic

vesicle remains nearly passive and becomes somewhat lens-shaped. At

one point its wall comes in close contact with the epiblast. Elsewhere

a wide cavity is developed between the two vesicles which becomes

filled with gelatinous tissue. At this period cilia appear on the

surface, and the larva becomes a planula.

[FIG. 71. EMBRYO OF GERYONIA AFTER DELAMINATION. (After Fol.)

_ep._ epiblast; _hy._ hypoblast.]

The succeeding changes lead rapidly to the formation of a typical

Medusa. Where the epiblast and hypoblast are in contact the former

layer becomes thickened and forms a disc-shaped structure. The centre

of this becomes somewhat protuberant, fuses with the hypoblast and

then becomes perforated to form the mouth (fig. 72 _o_). The edge

of the disc forms a thickened ridge, the rudiment of the velum

(_v_), which is entirely formed of epiblast. At its edge six tentacles

(_t_) arise, into which are continued solid prolongations of the wall

of the now somewhat hexagonal gastric chamber. The hypoblastic axes of

the tentacles soon lose their connection with the gastric wall.

[FIG. 72. OPTICAL SECTION THROUGH THE ORAL POLE OF GERYONIA AFTER

THE APPEARANCE OF THE GELATINOUS TISSUE OF THE DISC. (After Fol.)

_o._ mouth; _v._ velum; _t._ tentacle.

The shaded part represents the gelatinous tissue.]

Up to this time the larva has retained a more or less spherical form,

and the cavity on the under side of the umbrella has not yet become

developed. The latter now becomes established by the whole disc

assuming a vaulted form with the concavity directed downwards. The

lining of the cavity so formed is derived from the epiblast of the

disc already spoken of.

The exact mode of formation of the gastrovascular canals has not been

worked out. It has however been established by the researches of the

Hertwigs (No. 146) and Claus (No. 153) that the radial and circular

vessels of this system are connected together in adult Medusæ by an

hypoblastic lamella; so that these canals would seem to be the

remnants of an once-continuous gastric cavity. This mode of formation

is established in the case of the medusiform buds; and it would

therefore seem, as pointed out by the Hertwigs, a fair deduction that

it occurs in the larva--a conclusion which is confirmed by the

primitive extension of the gastric cavity to the edge of the disc at

the time when its walls give rise to the solid axes of the tentacles.

In the course of the subsequent retirement of the gastric cavity from

the edge of the disc the gastrovascular canals probably take their

origin, though Fol was unable to follow the changes which result in

their formation.

On the completion of the above changes the larva has become a fully

formed Medusa, but it undergoes a not inconsiderable metamorphosis

before the attainment of the adult state.

[FIG. 73. A THREE-DAYS' LARVA OF ÆGINOPSIS WITH TWO TENTACLES.

(After Metschnikoff.)

_m._ mouth; _t._ tentacle.]

Two species of Æginidæ have been studied by Metschnikoff (163), viz.

_Polyxenia leucostyla_ (_Ægineta flavescens_), and _Æginopsis

mediterranea_. In both of these forms the segmentation results in the

formation of an elongated two-layered ciliated planula, without a

central cavity. The two ends of this grow out into two long

processes--the rudiments of a pair of at first aborally directed

arms--which contain a solid hypoblastic axis (fig. 73). At this stage

the larva closely resembles the larva of Tubularia. An alimentary

cavity is hollowed out in the centre of the hypoblast which soon opens

by a wide oral aperture (_m_). A second pair of arms becomes formed,

which are at first much shorter than the original pair; with their

formation a radial symmetry is acquired. Sense organs become at the

same time developed, and the whole embryo assumes a medusiform

character. Fresh tentacles arise, the velum and cavity of the umbrella

become established, but these changes do not involve any points of

very special interest.

Siphonophora. The development of the Siphonophora has been the

subject of careful investigation by Haeckel (158) and Metschnikoff

(163). The ova are large and usually (except Hippopodius) without a

membrane.

They are formed of a peripheral denser layer of protoplasm and a

central spongy mass. They usually undergo their entire development in

the water. In some instances they have been successfully reared by

artificial impregnation.

As an example of the Calycophoridæ I shall take Epibulia aurantiaca, a

form allied to Diphyes, the development of which has been studied by

Metschnikoff[78].

[78] In my description of the development of the Siphonophora I

employ Huxley's terminology.

[FIG. 74. THREE LARVAL STAGES OF EPIBULIA AURANTIACA. (After

Metschnikoff.)

A. Planula stage.

B. Six-days' larva with nectocalyx (_nc_) and tentacle (_t_).

C. Somewhat older larva with gastric cavity.

_ep._ epiblast; _hy._ hypoblast; _so._ somatocyst; _nc._ nectocalyx;

_t._ tentacle; _c._ large yolk cells; _po._ polypite.]

There is a regular segmentation, unaccompanied by the formation of a

segmentation cavity. At its close the ovum becomes a spherical

ciliated embryo. This embryo soon becomes elongated, and its cells

differentiate themselves into a central and a peripheral layer--the

epiblast and the hypoblast (fig. 74 A). At this stage the larva has

the typical planula form. The epiblast is especially thickened at a

pole, which may be called the oral pole, and towards the side of this,

which will be spoken of as the ventral side. Adjoining this thickened

layer of epiblast a special thin layer of hypoblast becomes

differentiated, which in opposition to the main mass of large

nutritive cells forms the true hypoblastic epithelium (fig. 74 B,

_hy_). On this thickening two prominences make their appearance (fig.

74 B). The oral of these is the rudiment of a tentacle (_t_), and the

aboral of a nectocalyx (_nc_).

[FIG. 75. AN ADVANCED LARVA OF EPIBULIA AURANTIACA WITH ONE LARGE

NECTOCALYX. (After Metschnikoff.)

_so._ somatocyst; _nc._ second imperfectly developed nectocalyx;

_hph._ hydrophyllium; _po._ polypite; _t._ tentacle.]

The former of these elongates itself in succeeding stages into a

process of both epiblast and hypoblast. The central part of the

nectocalyx on the other hand appears to originate from a thickening of

the epiblast in which the cavity of the bell becomes subsequently

hollowed out. Between this part and the external epiblast which gives

origin to the outermost layer of the nectocalyx a layer of hypoblast

is interposed. When the nectocalyx has become to a certain extent

established a cavity--the commencement of the primitive gastrovascular

cavity of the adult--appears in the general hypoblast between the

epithelial and nutritive layers in the immediate neighbourhood of its

attachment. This cavity becomes prolonged into the nectocalyx to form

the four gastrovascular canals; while the hypoblast at the upper end

of the nectocalyx forms the somatocyst (fig. 74 C, _so_). The

primitive enteric cavity once formed rapidly extends, especially in an

oral direction (fig. 74 C), and forms a widish cavity in the oral part

of the embryo. At the pole of this part (fig. 74, _po_) is eventually

formed the opening of the mouth, and the contained cavity becomes in a

special sense the gastric cavity. This region of the embryo may be

spoken of as the polypite. The nectocalyx grows with great rapidity

and soon forms by far the most prominent part of the larva (fig. 75).

The true gastric region or polypite (fig. 75, _po_) continues also to

grow, and a mouth becomes formed at its extremity. The aboral end of

the original body of the embryo gradually atrophies.

At the junction of the nectocalyx and polypite the coenosarc becomes

formed, and rudiments of a second nectocalyx (_nc_) and second

polypite early become visible; while a hydrophyllium is formed as a

bud which covers over the first polypite and tentacle (_hph_). With

the development of the hydrophyllium the first segment, if the term

may so be used, is complete. The second segment of which a rudiment is

already present as a second polypite is intercalated between the first

segment and the nectocalyces.

[FIG. 76. TWO STAGES IN THE DEVELOPMENT OF STEPHANOMIA PICTUM.

(After Metschnikoff.)

A. Stage after the delamination. _ep._ epiblastic invagination to

form pneumatocyst.

B. Later stage after the formation of the gastric cavity in the

solid hypoblast, _po._ polypite; _t._ tentacle; _pp._ pneumatophore;

_ep._ epiblastic invagination to form pneumatocyst; _hy._ hypoblast

surrounding pneumatocyst.]

Amongst the Physophoridæ there is a considerable range of variation in

development; though the variations concern for the most part not very

important points. The simplest type hitherto observed is that of

_Stephanomia_ (Halistemma) _pictum_. The segmentation and formation of

a two-layered planula (fig. 76) take place in the usual way. Between

the solid central mass of nutritive hypoblast cells and the epiblast

an epithelial hypoblastic layer becomes interposed which undergoes a

special thickening at the aboral pole. At this pole a solid involution

of epiblast next becomes formed, to which a layer of hypoblast becomes

applied. The structure so formed is the rudiment of the pneumatocyst

(_ep_). In the next stage the air-cavity of the pneumatocyst becomes

established within the epiblast.

The gastrovascular cavity is formed in the midst of the nutritive

hypoblast cells, which then become rapidly absorbed leaving the

gastrovascular cavity entirely enclosed by the epithelial layer of

hypoblast (fig. 76 B).

By the above changes the more important organs of the larva have

become established. The one end forms the pneumatophore, and the

other, the oral part, the polypite. Between the two there is already

present the rudiment of a tentacle, and a second tentacle soon becomes

formed. The mouth arises as a perforation at the oral end of the

larva.

The pneumatophore contains a prolongation of the gastrovascular

cavity, the fluid in which bathes the outer hypoblastic wall of the

pneumatocyst. It has however no communication with the enclosed cavity

of the pneumatocyst. In the later developmental stages the size of the

pneumatophore becomes immensely reduced in comparison with the

remainder of the larva.

The development of Physophora agrees closely with that of Stephanomia

except in one somewhat important point, viz. in the development of a

provisional hydrophyllium. This arises as a prominence at the

aboral pole, containing a prolongation of the gastrovascular cavity.

Between the epiblast and hypoblast of the prominence gelatinous tissue

becomes deposited, and the hydrophyllium is thus converted into a

large umbrella-like organ enclosing the polypite. The two together

have a close resemblance to an ordinary Medusa, the polypite forming

the manubrium, and the hydrophyllium the umbrella. The hydrophyllium

is eventually thrown off.

An important type of Physophorid development is exemplified in

Crystalloides, a genus closely allied to Agalma. In this type the

greater part of the original ovum, instead of directly giving rise to

the polypite, becomes a kind of yolk-sack, from which the polypite is

secondarily budded (fig. 77, _yk_). _Agalma sarsii_ is in this respect

intermediate between Crystalloides and Physophora. Both these types

are remarkable for developing a series of provisional hydrophyllia

(fig. 77, _h.ph._). In both genera the first of these develops as in

Physophora, and for a long time is the only one functional.

The conclusions to be drawn from the above description may be summed

up as follows. In all the Siphonophora, so far observed, the starting

point for further development is a typical ciliated two-layered

planula. The inner layer or hypoblast is mainly formed of large

nutritive cells. From these cells an epithelial hypoblastic layer

becomes secondarily differentiated, the exact relations of which

differ somewhat in the various types. The nutritive cells themselves

do not appear to become directly converted into the permanent

hypoblastic tissues. The development of the adult from the planula

commences by the thickening of the epiblastic layer, usually at one

pole (the future proximal or aboral pole), and the formation at this

pole of a series of bud-like structures (in the growth of which both

embryonic layers have a share), which become converted into the

hydrophyllia, nectocalyces etc. The main oral part of the planula

becomes generally converted into the polypite, though in some

instances (Crystalloides) it remains as a yolk-sack, and only

secondarily gives rise to a polypite.

Two very different views have been taken as to the nature of the

various component parts of the Siphonophora, and the embryological

evidence has been appealed to by both sides in confirmation of their

views. By Huxley and Metschnikoff the various parts--nectocalyces,

hydrophyllia, hydrocysts, polypites, generative gonophores etc. are

regarded as simple organs, while by Leuckart, Haeckel, Claus etc. they

are regarded as so many different individuals forming a compound

stock. The difference between these two views is not merely as

to the definition of an individual[79]. The question really is, are

these parts originally derived by the modification of complete zooids

like the gonophores and trophosomes of the fixed Hydrozoa stocks, or

are they structures derived from the modification of the tentacles or

some other parts of a single zooid?

[79] From the expressions used by Huxley, _Anatomy of

Invertebrated Animals_, p. 149, it appears to me possible that

his opposition to Leuckart's view is mainly as to the nature of

the individual.

[FIG. 77. LARVA OF CRYSTALLOIDES, (After Haeckel.)

_h.ph._ hydrophyllium; _h._ hydrocyst; _t._ tentacle; _pp._

pneumatophore; _po._ polypite; _yk._ yolk-sack.]

The difficulty of deciding this point on embryological evidence

depends on the fact that ontologically a tentacle and a true bud arise

in the same way, viz. as papilliform outgrowths containing

prolongations of both the primitive germinal layers. The balance of

evidence is nevertheless in my opinion in favour of regarding the

Siphonophora as compound stocks, and the views of Claus on this

subject (_Zoologie_, p. 271) appear to me the most satisfactory.

The most primitive condition is probably that like Physophora in an

early stage with an hydrophyllium enclosing a polypite (cf. Haeckel

and Metschnikoff). In this condition the whole larva may be compared

to a single Medusa in which the primitive hydrophyllium represents the

umbrella of the Medusa, and the polypite the manubrium. The tentacle

which appears so early is probably not to be regarded as a modified

zooid, but as a true tentacle. The absence of a ring of tentacles is

correlated with the bilateral symmetry of the Siphonophora.

The primitive zooid of a Siphonophora stock is thus a Medusa. Like

Sarsia and Wilsia this Medusa must be supposed to have been capable of

budding. The ordinary nectocalyces by their resemblance to the

umbrellas of typical Medusæ are clearly such buds of the medusiform

type. The same may be said of the pneumatophore, which, as pointed out

by Metschnikoff, is identical in its development with a nectocalyx.

Both are formed by a solid process of epiblast in which a cavity--the

cavity of the nectocalyx or pneumatocyst--is eventually hollowed out.

Around this there appears a double layer of hypoblast containing a

prolongation of the gastrovascular cavity; and this is in its turn

enclosed by a layer of epiblast which forms the covering of the convex

surface of the nectocalyx and the external epiblast of the

pneumatophore.

The generative gonophores are clearly also zooids, and the

hydrophyllia are probably a rudimentary form of umbrella. In many

cases (Epibulia, Stephanomia, Halistemma etc.) the hydrophyllium of

the primitive polypite (manubrium) is absent. In such instances it is

necessary to suppose that the umbrella of the primitive zooid of the

whole colony has become aborted. Leuckart originally took a somewhat

different view from the above in that he regarded the starting point

of the Siphonophora to be a compound fixed Hydrozoon stock, which

became detached and free-swimming.

Acraspeda[80]. The embryonic development of several of the

forms of the Acraspeda has been investigated by Kowalevsky (No. 147)

and Claus (No. 153). Their observations seem to point to an invaginate

gastrula being characteristic of this group.

[80] I use this term for the group, often known as the

Discophora, which includes the Pelagidæ, Rhizostomidæ, and

Lucernaridæ.

Amongst the forms with alternations of generations and a fixed larval

form Chrysaora and Cassiopea have been most fully investigated. The

ovum of the former undergoes the first embryonic phases while still in

the ovary. In the latter it is enclosed amongst the oral processes. A

complete and more or less regular segmentation leads to the formation

of a single-walled blastosphere with a small segmentation cavity. The

wall of the blastosphere next becomes invaginated, giving rise to an

archenteron (fig. 78 A). The blastopore soon closes up, and the

archenteron is converted into a closed sack completely isolated from

the epiblast (fig. 78 B). The surface of the larva becomes in the

meantime covered with cilia. The free larval stage thus reached is

similar to the ordinary Hydrozoon planula. After the closure of the

blastopore the larva becomes elongated, and one end becomes narrowed.

By this narrowed extremity the larva soon attaches itself, and at the

opposite and broader end a fresh involution of the epiblast appears

(fig. 78 C); this gives rise to the stomodæum, which is placed in

communication with the archenteron on the absorption of the septum

dividing them. The relation of the stomodæum to the original

blastopore has not been determined.

At the point of attachment there is developed a peculiar pedal disc,

and around the mouth there appears a fold of epiblast which gives rise

to an oral disc (fig. 78 D). Two tentacles first make their

appearance, but one of these is primarily much the largest, though

eventually the second overtakes it in its growth. A second pair of

tentacles next becomes formed, giving to the larva a 4-radial

symmetry. Between these four new tentacles subsequently sprout out,

and in the intermediate planes four ridge-like thickenings of the

hypoblast, projecting into the cavity of the stomach, make their

appearance. They imperfectly divide the stomach into four chambers, to

each of which one of the primary tentacles corresponds; they may be

regarded as homologous with the mesenteries of the Actinozoa. The

number of tentacles goes on increasing somewhat irregularly up to

sixteen. All the tentacles contain a solid hypoblastic axis. Muscular

elements are developed from the epiblast.

[FIG. 78. FOUR STAGES IN THE DEVELOPMENT OF CHRYSAORA. (After

Claus.)

A. Gastrula stage.

B. Stage after closure of blastopore.

C. Fixed larva with commencing stomodæum.

D. Fixed larva with mouth, short tentacles, etc.

_ep._ epiblast; _hy._ hypoblast; _st._ stomodæum; _m._ mouth; _bl._

blastopore.]

With the above changes the so-called Hydra tuba or Scyphistoma form is

reached (vide fig. 85). The peculiar strobilization of this form is

dealt with in the section devoted to the metamorphosis.

Aurelia is stated by Kowalevsky to develop in the same way as

Cassiopea; and the one stage of Rhizostoma observed is that in which

it has a (probably invaginate) gastrula form.

In Pelagia the ovum directly gives rise to a form like the parent. The

segmentation and the invagination take place nearly as in Cassiopea,

but the archenteric cavity is relatively much smaller, and the large

space between it and the epiblast becomes filled with the gelatinous

tissue which forms the umbrella. The blastopore does not appear to

close but to become directly converted into the mouth. As in Cassiopea

the larva takes a somewhat four-sided pyramidal form. The mouth is

placed at the base. The pyramid becomes subsequently flatter, and at

the four corners four tentacles grow out which increase to eight by

division. The flattening continues till the larva reaches a form

hardly to be distinguished from the Ephyra resulting from the

strobilization of the fixed Scyphistoma form of other Acraspeda.

Alcyonidæ. In the Alcyonidæ the segmentation appears always to

lead to the formation of a solid morula, which becomes a planula by

delamination. The true enteric cavity is formed by an absorption of

the central cells, but the axial portion of the gastric cavity and

mouth are formed by an epiblastic invagination.

The development of these types has been mainly studied by Kowalevsky

(147), and my knowledge of his results is derived from German

abstracts of the original Russian memoirs.

In _Alcyonium palmatum_ the impregnation is external. The segmentation

is very exceptional in character. It commences with the formation of a

series of irregular prominences on the surface of the ovum, which

become segmented off to form a superficial layer of epiblast cells.

The inner mass of protoplasm then divides up into polygonal cells to

form the hypoblast, which would thus seem to be formed by a kind of

delamination. In _Clavularia crassa_ (No. 168) there is a complete

segmentation followed by a delamination. The larva of _Al. palmatum_

elongates and becomes ciliated, and so assumes the characters of a

typical planula. The central hypoblast is formed of an outer granular

stratum with imperfectly differentiated cells--the true hypoblast--and

an inner homogeneous mass with vacuoles.

Some of the larvæ become fixed, while others coalesce together and

form a large mass, the fate of which has not been further studied. An

invagination of epiblast takes place at the free end of the fixed

larva, which gives rise to the so-called gastric cavity, _i.e._

the axial portion of the general enteric cavity, which would appear to

be in reality a kind of stomodæum. Around the gastric cavity the

hypoblast forms eight mesenteries, the chambers between which are

filled with the homogeneous material which occupied the centre of the

ovum in the previous stage. It is to be presumed, though not stated,

that by an absorption of the blind end of the stomodæal invagination

the gastric chamber is placed in free communication with the

spaces between the mesenteries[81]. During the next stage the young

Alcyonium also acquires eight tentacles, which arise as hollow papillæ

opening into the eight mesenteric chambers. By this stage also the

matter filling up the mesenteric chambers is nearly absorbed.

[81] The German abstract is very obscure as to the formation of

the mouth.

Between the epiblast and hypoblast there is formed an homogeneous

membrane, which penetrates in between the two layers of hypoblast

which form the mesenteries. On the outer side of this membrane, and

therefore presumably derived from the epiblast, is a layer of

connective-tissue cells, which eventually gives rise to the abundant

gelatinous tissue (coenenchyma) in which the skeletal elements are

deposited. In _Sympodium coralloides_ Kowalevsky (No. 168) has

shewn still more completely the derivation of the stellate mesoblast

cells from the epiblast. He finds that the calcareous spicula develop

in these cells as in the mesoblast cells of sponges. The branched

gastrovascular canals in this tissue are outgrowths of the primitive

enteric cavity. A layer of circular muscles is formed at a late period

from the epiblast, but the longitudinal muscles of the mesenteries on

the inner side of the homogeneous membrane are regarded by Kowalevsky

as hypoblastic.

A ciliated planula with delaminated hypoblast is also found in

Gorgonia and _Corallium rubrum_. In the former genus at the time

when the larva becomes fixed, the hypoblast is formed of two strata,

an outer one of columnar cells, and an inner one of round ciliated

cells lining a central enteric cavity. The inner layer is believed by

Kowalevsky to become eventually absorbed and to be homologous with the

inner granular mass of Alcyonium.

Zoantharia. Amongst the Zoantharia several forms have been

investigated by Kowalevsky (147) and Lacaze Duthiers (170), of which

some are stated by the former author to pass through an invaginate

gastrula stage, while in other instances the hypoblast is probably

formed by delamination.

To the first group belongs an edible form of Sea Anemone found near

Messina, Cerianthus, and perhaps also Caryophyllium. In the first of

these segmentation results in the formation of a blastosphere. A

normal invagination obliterating the segmentation cavity then ensues,

and the blastopore narrows to form the mouth. The borders of the mouth

bend inwards and so give rise to the gastric cavity (stomodæum) which

as in the Alcyonidæ is lined by epiblast. Simultaneously with the

formation of the mouth there appear the two first mesenteries.

In Cerianthus the segmentation is unequal, the early stages are the

same as in the Actinia just described, but the hypoblast cells give

rise to a mass of fatty material filling up the enteric

cavity, which becomes eventually absorbed.

In the majority of the Zoantharia so far investigated, including

species of Actinia, Sagartia, Bunodes, Astroides, Astræa, the

segmentation, which is often unequal[82] and not accompanied by the

formation of a segmentation cavity, results in a solid two-layered

ciliated planula. In these forms the impregnation takes place in the

ovary, and the early stages of development are passed through in the

maternal tissues.

[82] I have this on the authority of Kleinenberg. The existence

of an unequal segmentation probably indicates an epibolic

gastrula.

One end of the planula becomes somewhat oval and develops a special

bunch of cilia. At the other end a shallow depression appears, which

becomes deeper and forms an involution lined by epiblast. This

involution is the stomodæum, and becomes the so-called gastric cavity.

The true enteric cavity lined by hypoblast is for some time filled

with yolk material. The larva always swims with the aboral end

directed forwards.

Between the two embryonic layers a homogeneous membrane is formed,

similar to that already described in the Alcyonidæ.

The further development of the larvæ especially concerns the formation

of mesenteries, tentacles and calcareous skeleton. With reference to

this subject the observations of Lacaze Duthiers are especially

valuable and striking.

In the adult it is usually possible to recognise in the tentacles a

symmetry of six. There are six primary tentacles, six secondary,

twelve tertiary, twenty-four quaternary, etc. In the hard septa of the

skeleton the same law is followed up to the third cycle, but beyond

that, in the cases where the point can be verified, there appear to be

only twelve septa in each additional cycle. The observations of Lacaze

Duthiers have shewn that this symmetry is only secondarily acquired

and does not in the least correspond with the succession of the parts

in development.

His observations were conducted on three species of Zoantharia without

a skeleton, viz. Actinia mesembryanthemum, Sagartia, and Bunodes

gemmacea; while Astroides calycularis served as the type for his

investigations on the corallum. It will be convenient to commence with

his results on Actinia mesembryanthemum which served as his type.