Chapter IV: The Skeleton (1)

(§§ 101-140).

101. _The Significance of the Skeleton._--The skeleton of the Radiolaria is developed in such exceedingly manifold and various shapes, and exhibits at the same time such wonderful regularity and delicacy in its adjustments, that in both these respects the present group of Protista excels all other classes of the organic world. For, in spite of the fact that the Radiolarian organism always remains merely a single cell, it shows the potentiality of the highest complexity to which the process of skeleton formation can be brought by a single cell. All that has been brought to pass in this direction by single tissue-cells of animals and plants does not attain the extremely high stage of development of the Radiolaria. Only very few Rhizopoda of this very rich and varied class fail to exhibit the power of forming this firm supporting and protecting organ--indeed, only ten of the seven hundred and thirty-nine genera which are enrolled in the list of the Challenger collection, namely, six genera of SPUMELLARIA (five Thalassicollida, _Actissa_, _Thalassolampe_, _Thalassopila_, _Thalassicolla_, _Thalassophysa_, Pl. 1, and one genus of Collozoida, _Collozoum_, Pl. 3), and in addition two genera of NASSELLARIA (the Nassellida, _Cystidium_ and _Nassella_, Pl. 91, fig. 1), and two genera of PHÆODARIA (the Phæodinida, _Phæocolla_ and _Phæodina_, Pl. 101, figs. 1, 2). These skeletonless forms of Radiolaria are, however, of extreme interest, since they include the original stem-forms of the whole class as well as of its four legions. All Radiolaria which form skeletons have originated from soft and skeletonless stem-forms by adaptation, and that polyphyletically, for the skeletal types of the four legions have been developed independently of each other (§ 108).

102. _The Chemical Peculiarities of the Skeleton._--The chemical composition of the skeleton shows very marked variations in the different legions of the Radiolaria. The two legions SPUMELLARIA and NASSELLARIA (united formerly as "Polycystina") form their skeleton of pure silica (see note A, below); the legion PHÆODARIA of a silicate of carbon (see note B), and the ACANTHARIA of a peculiar organic substance--acanthin (see note C). This explains the well-known fact that the deposits of fossil Radiolaria (or Polycystine marls) are composed exclusively of the skeletons of SPUMELLARIA and NASSELLARIA, those of the ACANTHARIA and PHÆODARIA being entirely absent (in the case of the last group, however, exception must be made in favour of the Dictyochida, or those PHÆODARIA {lxix}whose skeleton is made up of isolated scattered tangential siliceous fragments). The enormous deposits of Radiolarian skeletons in the deep sea of today, which constitute the Radiolarian ooze, consist, like the fossil Polycystine marls, almost exclusively of the shells of SPUMELLARIA and NASSELLARIA, though here the acanthin skeletons of the ACANTHARIA may be present in very small numbers, and the silicate skeletons of the PHÆODARIA, which offer more resistance to the solvent action of sea-water, somewhat more abundantly. Calcareous skeletons do not occur in the Radiolaria (see note D).

A. The pure siliceous skeletons of the Polycystina were first recognised
in 1833 by Ehrenberg in chalky marls (L. N. 2, p. 117). Since the two
legions ACANTHARIA and PHÆODARIA were entirely unknown to Ehrenberg, his
name Polycystina has reference only to the SPUMELLARIA and NASSELLARIA.

B. The silicate skeleton of the PHÆODARIA was formerly taken by me for a
purely siliceous one. When I described the first PHÆODARIA in my
Monograph in 1862, I was only acquainted with five genera and seven
species, whilst the number of PHÆODARIA here described from the
Challenger amounts to eighty-four genera and four hundred and sixty-five
species. In the vast majority of these (though not in all) the skeleton
becomes more or less intensely stained by carmine, and is also more or
less charred at a red heat, in some even becoming of a blackish-brown. In
many PHÆODARIA, furthermore, the hollow skeletal tubes are destroyed by
the continued action of heat. They are also, for the most part, strongly
acted upon, or even destroyed by boiling caustic alkalis, whilst boiling
mineral acids have no effect upon them. The best method of cleaning the
skeletons of PHÆODARIA from their soft parts is to heat them in
concentrated sulphuric acid, and then add a drop of fuming nitric acid;
in this they are not dissolved even on prolonged heating. From these
facts it would appear that the skeletons of the PHÆODARIA consist of a
compound of organic substance and silica, or a "carbonic silicate." The
more intimate composition yet remains to be discovered, as also the
manifold differences which the various families of PHÆODARIA seem to show
in respect of its composition. The small skeletal fragments of the
Dictyochida (the only remains of PHÆODARIA which occur as fossils) appear
to consist of pure silica.

C. The acanthin skeleton of the ACANTHARIA was first described as such in
my Monograph (1862, pp. 30-32). Johannes Müller, the discoverer of this
legion, took them for siliceous skeletons and defined the #Acanthometra#
as "Radiolaria without lattice-shell, but with siliceous radial spines"
(L. N. 12, p. 46). I formerly supposed that the acanthin skeletons in
some of the ACANTHARIA were partially or wholly metamorphosed into
siliceous skeletons, but, according to the investigations of R. Hertwig,
this does not appear to be the case; he showed that the skeletons of the
most varied #Acanthometra# and #Acanthophracta# are completely dissolved
under the longer or shorter action of acids, and supposes that in all
ACANTHARIA, without exception, the skeleton is composed of acanthin
(1879, L. N. 33, p. 120). Quite recently Brandt has found that the
acanthin spines dissolve not only in acids, alkalis, and "liquor
conservativus" (as I had shown), but also in solutions of carbonate of
soda (1 per cent.), and even of common salt (10 to 20 per cent.); he
concludes from this that they consist of an albuminoid substance
(vitellin) (L. N. 38, p. 400). I am unable to share this view, for I have
never been able to see some of the most important reactions of albumen in
any of the skeletons which I have examined, such for example as the
xanthoproteic reaction, the red coloration with Millon's test, &c. They
do not become {lxx}yellow either with nitric acid or with iodine. In
dilute mineral acids they dissolve more rapidly than in concentrated. My
usual method of cleansing the skeleton of ACANTHARIA (which has been
practised with the same result on thousands of specimens) consists in
heating the preparation in a small volume of concentrated sulphuric acid
and then adding a drop of fuming nitric acid; all other constituents (the
whole central capsule and the calymma) are thus very rapidly destroyed;
the skeleton remains quite uninjured and withstands the combined action
of the mineral acids for a longer or shorter time, though on prolonged
heating it also is dissolved. I do not therefore regard acanthin as an
albuminous substance, but as one related to chitin.

D. Calcareous skeletons have not been certainly demonstrated in the
Radiolaria, and probably do not occur. Sir Wyville Thomson in his
Atlantic (1877, L. N. 31, vol. i. p. 233, fig. 51) described under the
name _Calcaromma calcarea_, a Radiolarian which contained scattered in
its calymma numerous calcareous corpuscles "resembling the rowels of
spurs." These are identical with the "toothed bodies, recalling crystal
balls," which Johannes Müller figured in the Mediterranean _Thalassicolla
morum_ so early as 1858, and compared with the "siliceous asterisks of
_Tethya_" (L. N. 12, p. 28, Taf. vii. figs. 1, 2). I formerly regarded
these peculiar calcareous corpuscles, whose solubility in mineral acids I
had observed, as spicules of a Thalassicollid, and hence described the
species in my Monograph as _Thalassosphæra morum_ (L. N. 16, p. 260). I
have, however, seen reason to change my view, and am now led to suppose
that those peculiar calcareous corpuscles, which may be named
"_Calcastrella_," are not formed by the Radiolarian itself, but are
foreign bodies which have been accidentally incorporated into the calymma
of a Thalassicollid (_Actissa_). These corpuscles occur, often in large
numbers, in many preparations in the Challenger collection, and in the
calymma of other Radiolaria, chiefly #Discoidea#, hence it would appear
that they are foreign bodies taken up by the pseudopodia and carried into
the calymma by the circulation of the sarcode. The Radiolaria which Sir
Wyville Thomson figured as _Calcaromma calcarea_, and Müller as
_Thalassicolla morum_, I regard as species of _Actissa_ (see p. 13),
perhaps _Actissa radiata_ of the Pacific, and _Actissa primordialis_ of
the Mediterranean (compare the description of the Thalassosphærida of the
Challenger collection, pp. 30, 31).

103. _The Physical Properties of the Skeleton._--The skeletons of all Radiolaria are characterised pre-eminently by a high degree of _firmness_, which fits them to serve as protective and supporting apparatus. This is obvious in the case of the pure siliceous shells of the Polycystina; but the acanthin framework of the ACANTHARIA also possesses a degree of stiffness but little inferior, whilst the silicate skeletons of the PHÆODARIA seem on the whole to be not so firm. The hollow skeletal tubes of the last-named, which are filled with gelatinous material, are very brittle on account of the delicacy of their walls. Their _elasticity_ also is very small, whilst that of the acanthin spines is considerable. The thin long needles of many ACANTHARIA are very elastic, as are also the bristle-like siliceous spicules of many SPUMELLARIA. The _refractive power_ of the skeleton in the various legions is very different, depending upon the chemical constitution. The siliceous skeleton of the Polycystina (SPUMELLARIA and NASSELLARIA) and the silicate skeleton of the PHÆODARIA have the same refractive index as glycerine, and hence become invisible when mounted in that fluid; they then become visible only on addition of {lxxi}water, and are clearer in proportion to the quantity of water which is added. The refractive index of acanthin is, however, very different from that of glycerine, so that the skeletons of ACANTHARIA are readily visible when mounted in this fluid. In water, the skeletons of all Radiolaria appear about equally refractive, as also in Canada balsam. The substance of the skeleton appears almost entirely hyaline, colourless, and transparent. Very rarely it is faintly coloured (in some ACANTHARIA). A cloudy opaque constitution is seen in some PHÆODARIA (especially in the "porcellanous shells" of Tuscarorida and Circoporida, Pls. 100, 114-117); when dried, these appear by reflected light milky-white or yellowish-white; the cause of this opacity lies partly in the peculiar "cement-like structure" of these porcellanous shells, partly in their fine porosity, and the minute air-bubbles contained in their thick walls.

104. _The Elementary Structure of the Skeleton._--The general constitution of the skeleton--or more accurately expressed, of the morphological elements of which the skeleton consists--is of such a nature that it may be termed structureless. Both the organic acanthin skeletons of the ACANTHARIA and the silicate skeletons of the PHÆODARIA, as well as the inorganic siliceous skeletons of the SPUMELLARIA and NASSELLARIA, appear under the microscope perfectly homogeneous, transparent, colourless, and crystalline. Only very rarely do they show traces of a concentric striation, which arises from the deposition of the skeletal substance in layers; as, for example, the thick spines of some PHÆODARIA (Pls. 105-107, &c.). Some of the PHÆODARIA, however, form an exception to this rule, inasmuch as their partially tubular skeletal elements possess a remarkable porcellanous structure. In the tubular or Cannoid skeleton, which occurs in most CANNOPYLEA, the lumen of the thin-walled flinty tube is filled with jelly, and frequently a thin siliceous thread runs in its axis, and is connected with the wall by transverse threads (§§ 127, 139). The elementary structure of the opaque porcellanous shells, which distinguish the two families Circoporida (Pls. 114-117) and Tuscarorida (Pl. 100), is quite peculiar. Numerous fine siliceous spicules lie scattered irregularly in a finely granular or porous matrix.

105. _Complete and Incomplete Lattice-Shells._--In the great majority of Radiolaria (in all four legions) the skeleton has the form of a delicate lattice-shell or a receptacle in which the central capsule is enclosed. In a small minority, however, this is not the case. The skeleton then consists only of isolated rigid pieces (radial or tangential spicules), or of a simple ring (sagittal ring of the #Stephoidea#), or of a basal tripod with or without a loose tissue of trabeculæ, &c. (#Plectoidea#); the central capsule is then not surrounded by a special latticed receptacle, but only rests upon the skeletal trabeculæ. According to these different arrangements, two principal groups or sublegions may be distinguished in each legion, of which one set (Cataphracta) are characterised by a complete {lxxii}lattice-shell, whilst the others (Aphracta) are without it. The RADIOLARIA APHRACTA, then, or Radiolaria without a complete skeleton, are the #Collodaria# (p. 9), the #Acanthometra# (p. 725), the #Plectellaria# (p. 895), and the #Phæocystina# (p. 1543). On the other hand, the RADIOLARIA CATAPHRACTA, or Radiolaria with a complete skeleton, are the #Sphærellaria# (p. 49), the #Acanthophracta# (p. 791), the #Cyrtellaria# (p. 1015), and the #Phæocoscina# (p. 1590).

Upon this basis the first subdivision of the Radiolaria was made by
Johannes Müller, who recognised three groups:--"I. _Thalassicolla_,
without receptacle, naked or with spicules; II. _Polycystina_, with a
siliceous receptacle; III. _Acanthometra_, without receptacle, but with
siliceous radial spines" (L. N. 12, p. 16).

106. _The Ectolithia and Entolithia (Extracapsular and Intracapsular Skeletons)._--The relation of the skeleton to the central capsule in the Radiolaria is very various in many respects; in the first instance two great groups, _Ectolithia_ and _Entolithia_ (see note A), may be distinguished topographically by mere external observation; in the former the skeleton lies entirely outside the central capsule; in the latter, partially at all events, within it. The _Ectolithia_, with a completely extracapsular skeleton, include all NASSELLARIA and PHÆODARIA, as well as a great part of the SPUMELLARIA (all #Collodaria# and the most archaic forms of #Sphærellaria#); the _Entolithia_, on the other hand, in which the skeleton lies partly within, partly without the central capsule, include all ACANTHARIA and the majority of the SPUMELLARIA (most #Sphærellaria#, see note B).

A. The difference between Ectolithia and Entolithia was applied in my
Monograph in 1862 (p. 222) to separate the Monocyttaria into two main
groups. The arrangement was, however, quite artificial, being contrary to
the natural relations of the larger groups, as was shown seventeen years
later by the discovery of the different structural relations of the
central capsule.

B. Among the ACANTHARIA, which all possess primitively an intracapsular
and centrogenous skeleton, the remarkable _Cenocapsa_ (Pl. 133, fig. 11),
seems to furnish the single exception; in it the skeleton consists of a
simple spherical shell which encloses the concentric central capsule. The
exception is, however, only apparent; the twenty perspinal pores of the
shell show that they were originally in connection with twenty
centrogenous acanthin spines, and that those have disappeared by
retrograde metamorphosis.

107. _Perigenous and Centrogenous Skeletons._--Much more important than the topographical relation of the skeleton to the central capsule, according to which the Ectolithia and Entolithia are separated from each other (§ 106), is the original development of the skeleton within or without the central capsule, which gives rise to the distinction between perigenous and centrogenous skeletons. _Centrogenous skeletons_ are found only in the ACANTHARIA, which are further distinguished from all other Radiolaria by their skeleton being formed of acanthin; in all ACANTHARIA the formation of the skeleton begins in the middle of the central capsule, from which twenty (the number is inconstant only in the {lxxiii}small group #Actinelida#) radial spines are centrifugally developed. The three other legions, on the contrary, possess on the whole a _perigenous_ skeleton, which _originally_ develops outside the central capsule and never in its middle. In the NASSELLARIA and PHÆODARIA the skeleton retains this extracapsular position, as also in the #Beloidea# and part of the #Sphærellaria# among the SPUMELLARIA; in the great majority of the latter, however, the primary perigenous skeleton is subsequently enveloped by the growing central capsule, so that it lies partially within it (§ 109).

108. _Polyphyletic Origin of the Skeleton._--The skeleton of the Radiolaria has undoubtedly originated polyphyletically, for it is impossible to derive its manifold varieties from a single ground-form, or to regard them as modifications of one type. It is much more probable that the different skeletonless Radiolaria have entered upon different ways of skeleton formation quite independently of each other. At the outset it is quite clear that the skeletons of the _four legions have originated independently of each other_. Further, it is certain that within the legion of the SPUMELLARIA the Beloid skeletons of the #Collodaria# are not connected with the Sphæroid skeletons of the #Sphærellaria# and the forms derived from them (see § 109). In the same way the skeletons of the PHÆODARIA are polyphyletic; probably in this legion the Beloid, Sphæroid, Cyrtoid, and Conchoid skeletons have been developed quite independently (see § 112). In the NASSELLARIA, on the other hand, it is possible that all the skeletal forms are to be derived monophyletically from a single simple primitive form (either the sagittal ring or basal tripod?) (see § 111). Still more probable is it that the ACANTHARIA have arisen monophyletically, for all the forms of their acanthin skeleton may be derived without violence from _Actinelius_ (see § 110).

109. _The Skeleton of the Spumellaria._--The skeletons of the SPUMELLARIA or PERIPYLEA consist of silica, and are very different and of independent origin in the two orders of this legion. The first order, #Collodaria#, have either no skeleton whatever (#Colloidea#, p. 10, Pls. 1, 3), or their skeleton is _Beloid_, a loose extracapsular envelope of spicules, consisting of numerous unconnected portions; the separate parts are usually disposed tangentially, either as simple or compound siliceous spicules (#Beloidea#, p. 28, Pls. 2, 4). The second order of SPUMELLARIA, on the other hand (#Sphærellaria#, p. 49), develops a siliceous lattice-shell which consists of a single piece, and is remarkable for the extraordinary variety of its forms (pp. 50-715, Pls. 5-50). To this order belong not less than three hundred genera and seventeen hundred species of the Challenger Radiolaria (that is, about two-fifths of all the genera and species). In spite of this extreme richness in different forms this large group must be regarded as _monophyletic_, since all its forms may be quite naturally derived from a common stem-form, a _simple lattice-sphere_ (_Cenosphæra_, p. 61, Pl. 2). The twenty-eight families of #Sphærellaria# may be distributed in four suborders, among which the #Sphæroidea# constitute the {lxxiv}stem-forms, since they retain the original spherical shape (Pls. 5-8, 11-30). In the other three suborders a vertical main axis is developed, which in #Prunoidea# is longer, in #Discoidea# shorter than the other axes of the shell. Hence the shell of the #Prunoidea# (p. 284, Pls. 13, _bis_, 17, 39, 40) is ellipsoidal or cylindrical, that of the #Discoidea#, on the other hand, lenticular or discoidal (p. 402, Pls. 31-38, 41-48). Finally, the shell of the fourth suborder, #Larcoidea#, is lentelliptical; it has the ground-form of a triaxial ellipsoid, and is characterised by the possession of three unequal dimensive axes, or three isopolar axes of different lengths perpendicular to each other (p. 599, Pls. 9, 10, 49, 50).

110. _The Skeleton of the Acantharia._--The skeletons of the ACANTHARIA or ACTIPYLEA are distinguished from those of all other Radiolaria by two very important peculiarities; in the first place, they consist not of silica but of a peculiar organic substance, _Acanthin_, and secondly, their development is centrogenous, numerous radial spines or acanthin spicules being formed which are united in the middle of the central capsule. Hence the ACANTHARIA are the only Radiolaria in which the skeleton originates from the first in the middle of the central capsule. The number of radial spines is primitively indefinite, variable, and often considerable (more than a hundred), but in the great majority it is limited to twenty. In accordance with this the legion may be divided into two orders, the more archaic small group Adelacantha, with an indefinite number of spines, and the more recent group, Icosacantha, which has been developed from them and possesses twenty regularly disposed spines; of the three hundred and seventy-two species of ACANTHARIA which have been hitherto described, about five per cent. belong to the former, about ninety-five per cent. to the latter division (see note A, below). The numerous genera of Icosacantha may then be again divided into two suborders, of which the #Acanthonida# (p. 740, Pls. 130-132) produce no complete lattice-shell, and thus agree with the #Actinelida#, with which they may be united as #Acanthometra# in the broader sense (or ACANTHARIA without a lattice-shell). The #Acanthophracta#, on the other hand (p. 791, Pls. 133-140), produce a complete lattice-shell, usually by means of two opposite or four crossed transverse processes, which arise from each radial spine and unite with each other (see note B, below). In most #Acanthophracta# the lattice-shell remains single; only in the Phractopeltida does it consist of two concentric lattice-spheres (p. 847, Pl. 133, figs. 1-6). Furthermore, the whole order #Acanthophracta# may be subdivided into two suborders according to the different ground-form of the lattice-shell; this remains spherical in the #Sphærophracta# (the three families Sphærocapsida, Dorataspida, Phractopeltida, Pls. 133-138). On the other hand, it assumes another form in the #Prunophracta#; it becomes ellipsoidal in the Belonaspida (Pl. 136, figs. 6-9), discoidal or lentiform in the Hexalaspida (Pl. 139); and finally takes the shape of a double cone in the Diploconida (Pl. 140).

{lxxv}A. The group Adelacantha consists only of the suborder
#Actinelida#, with the three families Astrolophida, Litholophida, and
Chiastolida (p. 728, Pl. 129, figs. 1-3); the number of the radial spines
is very different and variable, sometimes only from ten to sixteen, but
usually from thirty to fifty, and often more than one hundred; they are
generally irregularly distributed, and not as in the second main
division. This latter, the Icosacantha, always possesses _twenty_ radial
spines, which are regularly disposed according to a constant law, the
so-called "Müllerian" or "Icosacanthan" law; the twenty spines are always
so placed between the poles of a spineless axis that they form five zones
each of four spines; the four spines of each zone are equidistant from
each other, and also from the same pole, and alternate with those of the
neighbouring zones, so that the whole twenty lie in four meridian planes,
which cut out an angle of 45° (compare pp. 717-722, Pls. 130-140). In
spite of the manifold variations in form which are developed in the
Icosacantha, they may all be derived from a common stem-form,
_Acanthometron_ (p. 742), since the law of distribution of the twenty
spines is constantly inherited.

B. An exception is found in the peculiar family Sphærocapsida (p. 797,
Pl. 133, figs. 7-11; Pl. 135, figs. 6-10). Here the shell is composed of
innumerable small, perforated plates, which arise on the surface of the
calymma independently of the spines.

111. _The Skeleton of the Nassellaria._--The skeletons of the NASSELLARIA or MONOPYLEA consist of silica, and are never composed of separate portions, but constitute always a single continuous piece. The ground-form is originally monaxon, corresponding to that of the central capsule, with a constant difference between the two poles of the vertical main axis. The ground-form is never spherical or polyaxon as in the lattice-shells of the SPUMELLARIA, and the skeleton never consists of hollow tubes, as in the PHÆODARIA. The legion NASSELLARIA may be divided into two orders; in the #Plectellaria# (three suborders #Nassoidea#, #Plectoidea#, #Stephoidea#) the skeleton does not form a complete lattice-shell; in the #Cyrtellaria#, on the other hand, which are derived from these, the siliceous skeleton forms a complete lattice-shell enclosing the central capsule. The number of forms thus developed is astonishingly great, so that among the NASSELLARIA no less than two hundred and seventy-four genera and sixteen hundred and eighty-seven species may be distinguished, almost as many as in the #Sphærellaria#. In spite of this great variety of forms the legion MONOPYLEA is probably monophyletic; at least all the different skeletal forms may be derived from three elements which are combined in the most manifold fashion; (1) the _sagittal ring_, a simple siliceous ring, which lies vertically in the sagittal plane of the body, encircles the central capsule and comes into contact with it at the basal pole of the main axis (§ 124); (2) the _basal or oral tripod_, composed of three diverging radial spines, which meet in the middle of the basal pole of the central capsule (or in the centre of the porochora) (§ 125); (3) the _cephalis_, or lattice-head, a simple ovoid or subspherical lattice-shell, which encloses the central capsule and stands in connection with it at the basal pole of its main axis. Any one of these three important structural elements of the NASSELLARIAN skeleton may possibly be the starting-point {lxxvi}for all the remaining forms of the MONOPYLEA; the great difficulty in their phylogenetic derivation lies in the facts that, on the one hand, any one of the three elements may alone constitute the skeleton, and on the other hand, in the great majority of the legion, two or three are united together (compare §§ 182-185).

112. _The Skeleton of the Phæodaria._--The skeleton of the PHÆODARIA or CANNOPYLEA is always extracapsular, usually consists of a silicate of carbon (more rarely of pure silica), and in the majority of the legion is composed of hollow cylindrical tubes, whose siliceous wall is very thin, and whose lumen is filled with gelatinous material (§ 127). The manifold and remarkable skeletal forms occurring in this legion are not monophyletic, since they cannot be derived from a common stem-form; they are, on the contrary, polyphyletic, various skeletonless PHÆODARIA (Phæodinida) have independently acquired skeletons of different form and composition. The legion PHÆODARIA can be subdivided into four orders, the skeletons of which present the following important distinctions:--(1) The #Phæocystina# possess only incomplete Beloid skeletons (§ 115), composed of many separate pieces, sometimes tangentially (Cannorrhaphida, Pl. 101), sometimes radially arranged (Aulacanthida, Pls. 102-105). (2) The #Phæosphæria# form Sphæroid skeletons (§ 116), usually only a simple lattice-shell without special aperture (Pls. 106-111); two concentric shells united by radial bars occur only in the Cannosphærida (Pl. 112). (3) The #Phæogromia# are distinguished by the formation of a simple Cyrtoid skeleton (§ 123) resembling that of the Monocyrtida; the monothalamus lattice-shell is usually ovoid or helmet-shaped, more rarely polyhedral or almost spherical; a vertical main axis can always be distinguished, at the basal pole of which is an aperture usually armed with teeth or spines (Pls. 99, 100, 113-120). (4) The #Phæoconchia# are distinguished from all other Radiolaria by the possession of a bivalved shell like that of the Conchifera; the two valves of this Conchoid skeleton must be distinguished as dorsal and ventral, as in the Brachiopoda (Pls. 121-128). The fifteen families of PHÆODARIA which are arranged in the four orders just mentioned, present such great differences among themselves, that the skeleton must be regarded as probably polyphyletic even within the limits of each order.

113. _Types of Skeletal Formation._--No less than twelve different principal forms may be distinguished as morphological types of the formation of the skeleton in the Radiolaria; some of these are peculiar to a single legion or even to a smaller group; but sometimes the same form occurs in several legions. Some types occur only in an isolated manner, independently of the others, but most exist in various combinations with other types. Of the twelve described below the Conchoid and Cannoid occur only in the PHÆODARIA; the Plectoid and Circoid only in the NASSELLARIA; the Astroid only in the ACANTHARIA; the remaining seven types are found in several legions in the same form and hence are polyphyletic.

{lxxvii}114. _The Astroid Skeleton._--Under the name "Astroid" we place the peculiar star-shaped skeletons of the ACANTHARIA in opposition to those of all other Radiolaria, for they are separated from them not only fundamentally by reason of the chemical nature of their substance (Acanthin, § 102), but also by their centrogenous origin, and the resulting stellate form (Pls. 129-140). The ACANTHARIA are the only Radiolaria in which the skeleton arises within the central capsule by the formation of numerous rays or radial spines of acanthin which project on all sides from the centre. Originally these are united at this point, their conical or pyramidal points meeting and being supported one upon another. In the great majority of ACANTHARIA this loose apposition is constant, so that when the soft parts are destroyed the skeleton falls to pieces. Only in a few forms in this legion are the central ends of the spines fused so that the whole skeleton forms a connected star (_Astrolithium_). The small group Chiastolida (or Acanthochiasmida) is characterised by the fact that the two rays which are opposite to one another in each axis unite and form a diametral bar. The skeleton is almost always composed of twenty radial spines, which are regularly disposed (Icosacantha), only in the small primitive group #Actinelida# is the number variable (Adelacantha, § 110).

115. _The Beloid Skeleton._--As Beloid or spicular skeletons are grouped together all those which consist of several disconnected portions; these always lie outside the central capsule, either within the calymma or on its surface. Such extracapsular Beloid skeletons are entirely wanting in the ACANTHARIA and NASSELLARIA; they occur only in the #Beloidea# among the SPUMELLARIA, and in the #Phæocystina# among the PHÆODARIA; the individual Beloid portions of the former are solid, those of the latter hollow. In both groups the simplest forms of the separate portions are simple unbranched needles (_Thalassosphæra_, _Thalassoplancta_, _Physematium_, _Belonozoum_, among the SPUMELLARIA; _Cannobelos_ and _Cannorrhaphis_ among the PHÆODARIA); usually these spicules are disposed tangentially over the surface of the calymma. Among the #Beloidea# branched spicules occur more commonly than these simple ones; they are either stellate (with many rays united in a centre) or twin-like, with a tangential bar, from each pole of which two or three (seldom more) radial branches project (Pls. 2, 4). Among the PHÆODARIA the subfamily Dictyochida is characterised by the annular shape of its Beloid portions, either simple rings, or hat-shaped or pyramidal bodies with a latticed cap over the ring (Pl. 101, figs. 3-14; Pl. 114, figs. 7-13). The family Aulacanthida among the PHÆODARIA, alone possesses hollow _radial tubes_, which penetrate the whole calymma, and project distally over its surface, whilst their proximal ends rest upon the surface of the central capsule. Although in these cases the enclosed proximal end is always simple, the free distal end develops the most various processes in adaptation to its prehensile functions (Pls. 102-105).

{lxxviii}116. _The Sphæroid Skeletons or Lattice-Spheres._--The "lattice-spheres" or sphæroid skeletons are the simplest and most primitive forms of lattice shells, and are widely distributed in the three legions SPUMELLARIA, ACANTHARIA, and PHÆODARIA, whilst they are entirely wanting in the NASSELLARIA. The round lattice-shell is either a true sphere in the geometrical sense, or an endospherical polyhedron, _i.e._, a polyhedron, all whose angles lie in the surface of a sphere (§ 25). In general, _primary_ and _secondary_ lattice-spheres may be distinguished, of which the former are secreted on the outer surface of the primary, the latter on that of the secondary calymma (§ 85). Furthermore, _simple_ and _compound_ lattice-spheres may be distinguished, the latter of which consist of two or more concentric lattice-spheres firmly united by radial bars; in such cases the innermost lattice-sphere is always to be regarded as the oldest or primary, all the succeeding ones as secondary, and the outermost as the youngest (§ 129). The simple lattice-spheres are usually to be regarded as primary; they may, however, occasionally be secondary, in which case the primary shell, originally enclosed, has been lost by degeneration (as, for example, in the case of the Aulosphærida and some #Sphærellaria#).

117. _The Lattice-Spheres of the Spumellaria._--The lattice-spheres or Sphæroid skeletons of the SPUMELLARIA exhibit in spite of their simple type of structure, an extraordinary variety in the formation of the lattice-work and radial apophyses, so that in the systematic portion of this work no less than one hundred and seven genera and six hundred and fifty species are distinguished; these are united in one suborder, the #Sphæroidea# (pp. 50-284, Pls. 5-8, 11-30). It may be divided into two main divisions, the _Monosphærida_ with a single primary lattice-sphere (Pls. 12-14, 21, 26, 27), and _Pliosphærida_ (or Sphæroidea concentrica) whose skeleton consists of two or more concentric lattice-spheres united by radial bars. The latter are subdivided into Dyosphærida with two concentric lattice-spheres (Pls. 16, 19, 20, 22, 28); Triosphærida, with three spheres (Pls. 17, 24, 29); Tetrasphærida, with four (Pls. 23, 30); Polysphærida, with five or more (Pls. 15, 23); and Spongosphærida, with spongy lattice-spheres (Pls. 18, 25). A special group is made up of the simple lattice-spheres of the social Collosphærida (or Sphæroidea polyzoa) (Pls. 5-8); these are usually more or less irregular, and characterised by the development of peculiar tubular processes; the latter are generally wanting in the Sphæroidea monozoa, whose lattice-shell is very regularly formed. This distinction is interesting and important, inasmuch as the regular lattice-spheres are explained by the independent development of the free-swimming Monozoa, whilst the irregular spheres are due to the mutual dependence of the social Polyzoa.

118. _The Lattice-Spheres of the Acantharia._--The lattice-shells or Sphæroid skeletons of the ACANTHARIA are immediately distinguishable from those of all other Radiolaria by their centrogenous development and the central union of the radial spines by which they are supported; the only exception is furnished by the remarkable genus _Cenocapsa_ {lxxix}(Pl. 133, fig. 11), in which the radial spines are absent, not primitively, however, but in consequence of degeneration; for the twenty cross-shaped perspinal pores, originally due to the twenty radial spines, are still present. In the most nearly allied genera, _Porocapsa_ (Pl. 133, fig. 7) and _Cannocapsa_ (Pl. 133, fig. 8), the proximal part of the twenty radial spines is still present, while their distal portion has degenerated; hence in this case they do not stand in direct communication with the spherical shell. On the other hand, this primitive connection persists in the genera _Astrocapsa_ (Pl. 133, figs. 9, 10), and _Sphærocapsa_ (Pl. 135, figs. 6-10). The five genera just mentioned form the peculiar family Sphærocapsida (pp. 795-802); the spherical shell is in these cases composed of very numerous small plates disposed like a pavement, each plate or aglet being perforated by a pore canal; in addition to which there are twenty larger (perspinal) pores (or twenty cross-shaped groups each of four aspinal pores) at those important points where primitively the twenty radial spines penetrate the calymma. This peculiar porous "pavement shell" has probably been developed (independently of the twenty radial spines) upon the calymma of the #Acanthonida# (_Acanthonia_, p. 749) by the action of the sarcodictyum; it has, therefore, quite a different morphological significance from the spherical lattice-shell of the Dorataspida, which is composed of tangential apophyses of the twenty Acanthonid spines (pp. 802-847, Pls. 134-138). Each radial spine here forms either two opposite or four crossed transverse processes, and since their branches spread over the surface of the spherical calymma and are united suturally at their extremities, the peculiar lattice-sphere of the Dorataspida arises. This extensive family is again divided into two subfamilies:--the Diporaspida (Pls. 137, 138) possess always only two opposite apophyses, and form by the union of their branches two opposite primary apertures or aspinal meshes. The Tessaraspida, on the other hand (Pls. 135, 138), have always four crossed transverse processes, and form by their union four primary aspinal meshes. From the Diporaspida are probably to be derived the Phractopeltida (p. 847, Pl. 133, figs. 1-6), the only ACANTHARIA which possess a double lattice-sphere; their double concentric spherical shell may be compared with that of the Dyosphærida.

119. _The Lattice-Spheres of the Phæodaria._--The lattice-spheres or Sphæroid skeletons of the PHÆODARIA, which are generally developed quite regularly, though occasionally in a modified form, fall in the order #Phæosphæria# into two groups of very different structure, each of which includes two families. The first group (_Phæosphæria inarticulata_) contains the families Orosphærida (Pls. 106, 107) and Sagosphærida (Pl. 108); the lattice-work of the former consists of irregular polygonal meshes and very coarse, partially hollow trabeculæ; in the latter, on the other hand, it consists of triangular meshes and very slender filiform trabeculæ; in both families the whole sphæroid skeleton forms a single unsegmented piece as in most #Sphæroidea#. In the second group of {lxxx}#Phæosphæria# (_Phæosphæria articulata_), on the other hand, the lattice-sphere is segmented in quite a peculiar manner, and composed of hollow cylindrical tangential tubes, which are separated by astral septa at the nodal points of the network; this remarkable structure characterises the two families, Aulosphærida (Pls. 109-111) and Cannosphærida (Pl. 112); the segmented lattice-sphere of the former is simple and hollow; while that of the latter is connected by centripetal radial tubes with a simple concentric inner shell, which is sometimes solid, sometimes latticed, and provided with a main-opening corresponding to the astropyle of the enclosed central capsule. Since in the Aulosphærida also, hollow centripetal radial tubes project from the segmented lattice-sphere, it is possible that they have been derived from the Cannosphærida by the loss of the primitive internal shell. A special peculiarity of many #Phæosphæria# (_Oroscena_, _Sagoscena_, _Auloscena_, &c.) consists in the fact that the whole surface of the lattice-sphere is regularly covered with pyramidal or tent-shaped prominences (Pl. 106, fig. 4; Pl. 108, fig. 1; Pl. 110, fig. 1). A simple lattice-sphere quite similar to that of most Monosphærida also constitutes the skeleton of the Castanellida (Pl. 113), but since it possesses a special main-opening, it must be referred promorphologically to the Cyrtoid shells of the #Phæogromia#.

120. _The Prunoid Skeleton or Lattice-Ellipsoid._--The "lattice-ellipsoids" or Prunoid skeletons have arisen from the lattice-spheres or Sphæroid skeletons by more energetic growth and elongation of one axis; this is the main axis of the body and is probably always vertical; its two poles are commonly equal. The Prunoid skeleton is either a true ellipsoid in the geometrical sense or an "endellipsoidal polyhedron" (_i.e._, a polyhedron, all the angles of which lie in an ellipsoidal surface). By further elongation of the main axis, the ellipsoidal form passes over into the cylindrical, the polar surfaces of the cylinder being usually rounded, rarely truncated. The rich order #Prunoidea# (pp. 284-402) contains numerous modifications of this form of shell which arise on the one hand by the formation of transverse constrictions, on the other by the apposition of concentric secondary shells. In respect of the latter, simple and compound Prunoid shells can be distinguished as in the case of the Sphæroid shells. In the compound Prunoid shells either all the concentric lattice-shells may be ellipsoidal or the inner may be spherical. More important differences are found in the transverse annular constrictions, which give the Prunoid skeleton a segmented appearance; in this respect, three principal forms may be distinguished (p. 288):--(A) _Monoprunida_, with unsegmented shell, having no transverse constriction (Pls. 15-17); (B) _Dyoprunida_, having a shell with two segments and one (equatorial) transverse constriction (Pl. 39); (C) _Polyprunida_, with three or more parallel transverse constrictions, by means of which the shell is divided into four or more segments (Pl. 40). In the same manner as the #Prunoidea# have arisen from the #Sphæroidea# among the SPUMELLARIA by greater {lxxxi}development of the vertical main axis, the ellipsoidal Belonaspida have arisen from the spherical Dorataspida among the ACANTHARIA (p. 859; Pl. 136, figs. 6-9; Pl. 139, figs. 8, 9). The main axis of the ellipsoid in this case is always occupied by the opposite equatorial spines of the hydrotomical axis (pp. 719, 860). In the legion PHÆODARIA a similar prolongation of the main axis rarely occurs; it is found, however, in _Aulatractus_ (Pl. 111, figs. 6, 7), the lattice-shell of this Aulosphærid being sometimes truly fusiform, sometimes rather ellipsoidal or even double-conical.

121. _The Discoid Skeletons or Lattice-Discs._--The "lattice-discs" or Discoid skeletons are characteristic of the SPUMELLARIAN group #Discoidea#, and have arisen from the lattice-spheres of the #Sphæroidea# by a less development of one axis, which is the main axis of the body, and is probably usually vertical; its two poles are always equal. The Discoid lattice-shell is either a biconvex lens (with a thin margin), or a plane disc (a shortened cylinder with thick margin), or some form intermediate between the two. All Discoid shells show a horizontal median plane or equatorial plane, by which they are divided into two equal halves, an upper and lower; the margin of the lens itself is originally the equator. The main axis, the shortest of all the axes of the shell, stands vertically in the centre of the equatorial plane. Among the PHÆODARIA Discoid shells rarely occur (_Aulophacus_), as also among the ACANTHARIA (Hexalaspida).

122. _The Larcoid Skeleton or Lentelliptical Lattice-Shell._--The lentelliptical lattice-shells, which may be shortly designated "Larcoid," are especially characteristic of the #Larcoidea#, a large order of SPUMELLARIA (pp. 599-715; Pls. 9, 10, 49, 50). In addition they recur among the ACANTHARIA, in the small family Hexalaspida (p. 872, Pl. 139), and the family Diploconida (p. 881, Pl. 140), which is derived from it. These lentelliptical lattice-shells are all characterised by the clear differentiation of three unequal, but isopolar dimensive axes, _i.e._, the three geometrical axes, perpendicular to one another, which determine the form of the shell, are of unequal length; the two poles of each are, however, equal. The geometrical ground-form is, therefore, a triaxial ellipsoid (§ 34). In the rich order #Larcoidea# the lentelliptical lattice-shell shows many variations in its development.

123. _The Cyrtoid Skeleton._--Cyrtoid skeletons are those lattice-shells which possess a vertical main axis with two different poles (Monaxonia allopola); the upper pole is usually termed the apical, the lower the basal. Such Cyrtoid shells are characteristic of the great majority of the NASSELLARIA or MONOPYLEA (and especially of the #Cyrtellaria#); they are also found in a large division of the PHÆODARIA (the #Phæogromia#), and in some SPUMELLARIA. In general the manifold Cyrtoid shells may be divided into two large groups, those with one and those with several chambers. The _monothalamous_ Cyrtoid shells are usually ovoid, conical, cap- or helmet-shaped; their {lxxxii}internal cavity is simple, without constrictions or septa. Among the NASSELLARIA they occur in the Monocyrtida (Pls. 51-54, 98), where they have received the name "Cephalis." A form of shell, essentially the same, is found amongst the PHÆODARIA in the order #Phæogromia#, more especially in the Challengerida (Pl. 99), Medusettida (Pls. 118-120), and Tuscarorida (Pl. 100), many of these latter closely resembling many Monocyrtida. Such monothalamous Cyrtoid shells occur much more rarely among the SPUMELLARIA (_e.g._, among the #Prunoidea# in _Lithapium_, _Lithomespilus_, _Druppatractus_, Pls. 13, 14, &c.). Polythalamous Cyrtoid shells (Pls. 55-80) occur exclusively in the NASSELLARIA, and exhibit in this legion an astonishing variety of structure; they are distinguished from the monothalamous forms by the development of internal septa, or of annular incomplete diaphragms, which usually correspond to the external constrictions; their interior is thus divided into two or more communicating compartments. Among the polythalamous Cyrtoid shells may be distinguished three principal groups, the Stichocyrtid, Zygocyrtid, and Polycyrtid. Zygocyrtid shells are characteristic of the #Spyroidea# (Pls. 84-90), and are distinguished by a bilobate cephalis (cephalis bilocularis); the median sagittal ring, or a corresponding constriction, divides the shell into right and left compartments. Polycyrtid shells (Pl. 96) are peculiar to the #Botryodea#, and characterised by a multilobate cephalis (cephalis multilocularis). Stichocyrtid shells are those in which the primary cephalis remains simple, and new joints are successively added to its basal pole; such shells occur in the majority of the #Cyrtoidea#. Secondary chambers are sometimes added in the other two groups (#Botryodea# and #Spyroidea#). When, as often happens in these polythalamous Cyrtoid shells, two or three distinct joints follow each other, the first is called the "cephalis," the second the "thorax," and the third the "abdomen" (Tricyrtida Pls. 64-75).

124. _The Circoid Skeleton._--This is a very important and remarkable type of skeletal formation, which occurs exclusively in the legion NASSELLARIA, where it plays a very prominent part; its characteristic element is the "sagittal ring," a simple, vertical, siliceous ring, which surrounds the central capsule in its sagittal plane, and is specially differentiated in its basal portion. This "primary sagittal ring" whose vertical allopolar main axis coincides with that of the Monopylean central capsule embraced by it, is characteristic of all members of the order #Stephoidea# (p. 931, Pls. 81-83, 92-94); here it forms by itself the skeleton of the Stephanida (Pl. 81); in the Semantida (Pl. 92) it is combined with a horizontal basal ring, in the Coronida (Pls. 82, 93) with a vertical frontal ring and in the Tympanida (Pls. 83, 94) with two horizontal rings, an upper mitral and a lower basal. In the great majority of these #Stephoidea# there often develop in definite places characteristic processes or apophyses, whose branches combine to form a loose tissue or an incomplete lattice-shell. This becomes complete in the #Cyrtellaria#, the majority of which retain more or less {lxxxiii}distinct traces of the sagittal ring. Hence the skeletons of all NASSELLARIA may be derived monophyletically (Hypothesis A, p. 893) from a simple sagittal ring (_Archicircus_ and _Lithocircus_, Pl. 81). This theory, however, encounters the great difficulty that in many #Stephoidea# (_Cortina_, _Cortiniscus_, &c.) it is combined in a remarkable manner with the basal tripod of the #Plectoidea#, whilst in these latter it is entirely wanting (compare p. 894).

125. _The Plectoid Skeleton._--Those forms are distinguished as Plectoid in which three, four, or more radial siliceous spines proceed from a common point, which lies excentrically outside the central capsule and at the basal pole of its vertical allopolar main axis. This peculiar type of skeletal formation only occurs in the legion NASSELLARIA, and is specially characteristic of the order #Plectoidea# (p. 898, Pl. 91). But since the essential elements of this remarkable skeleton also occur in many other NASSELLARIA, sometimes combined with the Circoid, sometimes with the Cyrtoid skeleton, it perhaps has a fundamental significance in this legion; at all events it is possible to derive monophyletically all the other forms of this legion from it (Hypothesis B, p. 893). The simplest form of the Plectoid skeleton is a tripod, the three feet of which either lie in a horizontal plane (_Triplagia_, Pl. 91, fig. 2), or correspond to the three edges of a low pyramid (_Plagiacantha_). A fourth ray is sometimes added, which stands vertically upon the summit of the pyramid (_Plagoniscus_, _Plagiocarpa_, Pl. 91, figs. 4, 5). In other #Plectoidea# three secondary rays are intercalated between the three primary (Hexaplagida, &c.); seldom the number is greatly increased (Polyplagida, &c.). The rays are rarely simple, but usually branched; in the Plagonida (Pl. 91, figs. 2-6) the branches remain free; in the Plectanida (Pl. 91, figs. 7-13) they are united to form a loose wicker-work. From such a web a perfect Cyrtoid shell may arise. Several forms of Plagonida may also be readily confounded with the isolated triradiate or quadriradiate spicula of many Beloid skeletons (_Sphærozoum_, _Lampoxanthium_, &c.).

126. _The Spongoid Skeleton._--From the simple lattice-skeleton which the majority of Radiolaria possess, some of them develop a spongy shell; the trabeculæ of the lattice-work, situated in one plane in the former, are developed in the latter in different planes and cross irregularly in all directions; thus arises a kind of wicker-work of more or less spongy structure, usually with very thin trabeculæ and irregular meshes. Such Spongoid shells are most common among the SPUMELLARIA, especially in the #Sphæroidea# (Spongosphærida, Pl. 18) and #Discoidea# (Spongodiscida, Pls. 41-47), more rarely in the #Prunoidea# and #Larcoidea#. Lattice-work of similar spongy structure occurs very seldom among the NASSELLARIA, _e.g._, in some #Plectoidea# (Pl. 91) and #Cyrtoidea# (_Spongocyrtis_, _Spongopyramis_, _Spongomelissa_, &c., Pl. 56, fig. 10; Pl. 64, figs. 5-10, &c.). Among the PHÆODARIA spongy skeletons are very rare; they {lxxxiv}are to be seen in some #Phæosphæria# (_Oroplegma_, Pl. 107, fig. 1; _Sagoplegma_, Pl. 108, fig. 2; _Auloplegma_, Pl. 111, fig. 8). No Spongoid skeletons are known among the ACANTHARIA.

127. _The Cannoid Skeleton._--Cannoid or tubular skeletons are those which are composed of hollow tubes; they occur exclusively in the PHÆODARIA or CANNOPYLEA. Tubular processes, nevertheless, occur in some other Radiolaria, as, for example, among the SPUMELLARIA in a portion of the Collosphærida (_Siphonosphæra_, _Caminosphæra_, Pls. 6, 7), and of the #Prunoidea# (_Pipetta_, _Cannartus_, &c., Pl. 39, figs. 6-10, &c.), also among the NASSELLARIA in _Theosyringium_ (Pl. 68, figs. 4-6), _Cannobotrys_ (Pl. 96, figs. 3, 4, 8-11, 20-22), &c. In all these cases, however, the tubes are direct processes of the cavity of the shell, the trabeculæ of the lattice-work being solid. Only in the CANNOPYLEA are the lattice-bars themselves, the radial spines and appendicular organs, generally tubular (hence the designation "Pansolenia"). The lumen of the thin-walled siliceous tubes is filled with jelly, and hence the specific gravity of the relatively large skeleton is considerably diminished. This peculiarity is not found in all CANNOPYLEA; it is wanting in all Sagosphærida and Concharida, as well as in a part of the Orosphærida and Castanellida; in the latter there are found intermediate stages between hollow and solid skeletal rods. Very often a fine siliceous thread runs in the axis of the tubes, which is connected with its wall by lateral branches (Pl. 110, figs. 4, 6; Pl. 115, figs. 6, 7). More seldom the tubes are divided by horizontal septa into a series of chambers (Medusettida, Pls. 118-120). The two families Aulosphærida (Pls. 109-111) and Cannosphærida (Pl. 112) are distinguished from all other PHÆODARIA by the fact that their tubes are separated by astral septa in the nodal points of the lattice-shell (§§ 112, 134).

128. _The Conchoid Skeleton._--By the name "Conchoid skeletons" are distinguished the bivalved lattice-shells which occur exclusively in the legion PHÆODARIA; they are quite characteristic of the #Phæoconchia# or _Phæodaria bivalvia_, which embrace three families:--Concharida (Pls. 123-125), Coelodendrida (Pls. 121, 122), and Coelographida (Pls. 126-128). The two valves of the lattice-shell of the Concharida are simple, hemispherical, or boat-shaped, whilst in the Coelodendrida and Coelographida tubes grow out from them, which branch and usually give rise by anastomosis to a second external bivalved shell. In all #Phæoconchia# the two valves are so disposed about the central capsule that an open slit remains between them, into which open the apertures of the central capsule; and since all these _Phæodaria conchoidea_ are TRIPYLEA, with three typical openings in the central capsule, and since the two lateral accessory openings lie at either side of the aboral pole, and the unpaired main-opening at the oral pole of the main axis, it follows that the two valves are to be regarded as dorsal and ventral as in the Brachiopoda (not right and left as in the Lamellibranchiata). The dorsal and ventral {lxxxv}valves are usually equal, but in a portion of the Concharida they present constant differences. In this family the two valves are attached to each other by their free edges, just as in the bivalved Mollusca and Diatoms; and these edges may either be smooth (Conchasmida, Pl. 123, figs. 1-6), or dentate (Conchopsida, Pls. 124, 125); the valvular connection of the latter is sometimes strengthened by a special ligament which unites the two valves at the aboral pole (Pl. 123, figs. 8, 9). The form of the valve is sometimes hemispherical, sometimes boat-shaped, with a sagittal keel.

129. _Medullary and Conical Shells._--In all Radiolaria whose skeleton consists of a double shell or of two concentric lattice-shells united by radial bars, an inner medullary shell (testa medullaris) and an outer cortical shell (testa corticalis) may be distinguished (see note A, below). The medullary shell is usually to be regarded as a primary, the cortical as a secondary structure. Such double shells occur among the SPUMELLARIA in the Dyosphærida (Pls. 19, 20), as well as in many #Prunoidea# (Pls. 39, 40), #Discoidea# (Pls. 33, 34), and #Larcoidea# (Pls. 9, 10); among the ACANTHARIA only in the family Phractopeltida (Pl. 133); among the NASSELLARIA only in very few #Cyrtoidea# (_e.g._, _Periarachnium_, Pl. 55, fig. 11), and finally among the PHÆODARIA in the Cannosphærida (Pl. 112) as well as in part of the Coelodendrida (Pl. 121) and Coelographida (Pls. 127, 128). In most cases (if not always?) the cortical shell arises by the growth of radial spines from the surface of the medullary shell; these become united at equal distances from the centre by transverse apophyses, the surface of the secondary calymma furnishing the basis for their secretion (§ 85). Nevertheless, it seems that in many #Sphærellaria# the formation of the whole cortical shell proceeds simultaneously (at a definite dictyotic period) like that of the primary medullary shell (see note B). Whilst in the PHÆODARIA, ACANTHARIA, and NASSELLARIA, at most two concentric shells are formed, in many SPUMELLARIA their number increases continuously with additional growth; in many #Sphærellaria# it rises to four, eight, or even more, as well as in many #Discoidea# (if the concentric, peripherally disposed rings of chambers be regarded as incomplete flattened shells). In these cases either only the innermost primary lattice-shell is to be styled "medullary shell," or at most the two innermost (inner and outer medullary shells), all the others being cortical.