The third mode of junction, by central concrescence of all twenty spines, was formerly regarded by me as an important peculiarity, sufficient for the separation of subfamilies and genera (Monogr. d. Radiol., 1862, pp. 399, 401; Prodromus, 1881, p. 466). But I found afterwards that in many species where the twenty spines commonly remain separated, accidentally they grow perfectly together and form one single piece of acanthin—a starrulet with twenty rays. Therefore I now think it is more natural to divide those species only into different subgenera.

A fourth and a very different mode of junction, quite sufficient for the distinction of different families, is the concrescence in pairs of every two opposite spines, lying in one diameter (in Acanthochiasma and Chiastolus). Here we obtain a number of "diametral spines" (each composed of two originally opposed radial spines) and all these diametral spines are crossed loosely near the central point of the body without any solid and permanent apposition (Chiastolida). However, in some species of this peculiar family the central part of the diametral spines is twisted like a screw or spirally convoluted (Pl. 129, figs. 2, 3).

The Form of the Radial Spines in the Acantharia is extremely varied, and constitutes the main characters for the distinction of nearly four hundred species. But all these different forms may be reduced phylogenetically to three different fundamental forms:—(a) the cylindrical (with circular transverse section), (b) the two-edged (with elliptical or lanceolate transverse section), and (c) the four-edged (with square transverse section). No doubt the first (a) is the original primitive form, from which the two others are secondarily derived. Triangular spines never occur in the Acantharia, whilst, however, they are common in the Sphærellaria. The first and original form, the cylindrical spine, is either a true cylinder of equal thickness in its whole length, or it is more or less conical. Rarely the spine is in the distal half spindle-shaped, and thicker than in the basal half. The second form, the two-edged spine, is more or less compressed from two opposite sides; its two edges are either more blunt, rounded, or more acute, sharp; its transverse section in the former case is elliptical, in the latter case lanceolate or rhomboidal. Sometimes the two edges are broader and in the form of two thin opposite wings. The two-edged spines may be occasionally shorter, triangular or lanceolate, at other times longer sword-shaped or linear. The third form, the four-edged spine, has constantly a square transverse section; the sides of this square are either even or concave; in the latter case the four edges are broadened and wing-like, but in the former case not. The quadrangular spines are either prismatic (of equal breadth throughout their whole length) or pyramidal (becoming gradually thinner towards the distal apex).

The Apex of the Radial Spines, or their free distal end, is in the majority of Acantharia simple, conical. In the minority it is either truncated or bifid, or four-sided pyramidal, often with two, rarely with four prominent parallel teeth. In some forms the bifid spines are so deeply cleft that they become forked. Much more interesting and more varied than these different forms of the distal end are those of the apophyses of the radial spines.

The Apophyses of the Radial Spines, or their "lateral transverse processes," are of the greatest importance for the morphological development of the whole subclass. Only in sixteen among the sixty-five genera of Acantharia are the apophyses perfectly wanting; in the other genera they determine in the first place their general character. In the Acanthometra the apophyses remain perfectly free, whilst in the Acanthophracta their meeting ends or branches compose the latticed shell. All differences in form and shape of the apophyses can be reduced to only two primary modes; either the spine bears two opposite or four crossed apophyses; correspondingly all Acantharia apophysaria may be divided into two different main groups, the Zygapophysica (with two opposite lateral processes) and the Staurapophysica (with four crossed lateral processes opposite in pairs). Both groups have probably no direct phylogenetic connection, but seem to be derived independently from different stocks, and produce different families. The Zygapophysica are probably derived from Astrolonchida with two-edged spines (Zygacantha), and from this group arise the Diporaspida, the ancestral group of the majority of Acanthophracta. On the other hand the Staurapophysica are probably derived from Astrolonchida with four-edged spines (Acanthonia), and from this group arise the Tessaraspida. The apophyses of the Acanthonida are partly simple, partly branched or even latticed; the apophyses of the Acanthophracta are never simple, constantly branched and commonly latticed.

The Malacoma (or the whole soft body of the Acantharia as opposed to the skeleton) exhibits some peculiarities which distinguish them from the other Radiolaria, as well in the structure of the central capsule and its nucleus as in that of the enveloping extracapsular body and the pseudopodia.

The Central Capsule is constantly spherical in the far greater number of the Acantharia, viz., in the following six families:—Astrolophida, Chiastolida, Astrolonchida, Dorataspida, Sphærocapsida, and Phractopeltida. Among these six families the Astrolonchida and Dorataspida are far greater and far richer in different forms than all the other families. The central capsule becomes ellipsoidal or cylindrical, prolonged in one axis, in the three families, Amphilonchida, Belonapsida, and Diploconida; it becomes discoidal or lenticular, by the shortening of one axis, in two families, viz., in the Quadrilonchida and Hexalaspida. Finally, the peculiar family Litholophida is distinguished by the conical form of its central capsule.

The Membrane of the central capsule in all Acantharia is simple, commonly thin, sometimes very delicate; in some species it seems to be developed late, just immediately before the formation of the spores; but in no species is it completely missing. The membrane is constantly pierced by innumerable fine pores, for the emission of the pseudopodia; but in many species (and probably more or less in all Acantharia) there is recognisable a certain regularity in the disposition of the numerous pseudopodia and of the pores by which they radiate from the capsule. Sometimes these pores are disposed in a regular network of ramified lines, whilst the meshes of this network are devoid of pores; in other cases they form regular tufts or bushes between the radial spines. Probably in no Acantharia are the pores of the capsule membrane so numerous and so equally distributed throughout as in the Spumellaria; we may therefore call the former Actipylea (in opposition to the latter, as Peripylea).

The Nucleus of the Acantharia is constantly excentric, whilst it is originally constantly central in the Spumellaria. This excentric position is a necessary consequence of the centrogenous development of the radial spines. Probably connected with this peculiarity is the other, that the nucleus assumes a peculiar, complicated structure, and that in the greater number of Acantharia it becomes cleft very early, and that this cleavage is effected by a peculiar kind of gemmation, first detected and very accurately described by R. Hertwig (compare his Organismus d. Radiol., 1879, pp. 10-24). However, in the young Acantharia the nucleus is constantly simple, and in a certain number of species its cleavage takes place late (as in the greater number of Spumellaria).

The Endoplasm, or the intracapsular sarcode, exhibits in the greater number of Acantharia a more or less distinct radial arrangement; but this is often concealed by the different enclosed products of the endoplasm—oil-globules, vacuoles, red or different coloured pigment-granules, crystals, &c. Often it encloses a variable number of "yellow cells" (becoming green by mineral acids) to be considered as symbiotic xanthellæ.

The Calymma or the jelly-veil, including the central capsule, in the Acantharia is more or less voluminous, and commonly envelops the skeleton perfectly. In its surface is sometimes developed a peculiar network of "supporting fibres." A very peculiar product are the remarkable "Myophrisca" of the Acanthometra, which are wanting in the Acanthophracta; they were first detected by Johannes Müller, and figured as "Cilien-Kränze," afterwards explained by Hertwig as "contractile Fäden," similar to muscular fibrillæ (compare below).

The Matrix, placed between the calymma and central capsule, in the majority of the Acantharia is a rather thin layer of granular exoplasm.

The Pseudopodia arising from it are not so numerous as in the Spumellaria, and not so equally disposed over the whole surface. Also their tendency to ramify, anastomose, and form networks seems to be much less developed. Commonly they are simple or little ramified. In many cases (and perhaps everywhere) there may be distinguished two different kinds of pseudopodia:—(1) Axopodia, or permanent pseudopodia (with axial filaments?), piercing the wall of the central capsule, and arising from the central mass of endoplasm; and (2) Collopodia, or variable pseudopodia (without axial filaments), arising outside the capsule from the matrix of extracapsular sarcode or from the exoplasm on the surface of the calymma. These and other differentiations seem to indicate that the pseudopodia in the Acantharia are more highly developed than in the Spumellaria, and justify the denomination of the former as "Actipylea."

Synopsis of the Orders and Suborders of Acantharia.

Order III. ACANTHOMETRA, Johannes Müller, 1855.

Definition.—Acantharia without complete latticed shell.

The order Acanthometra, the third order of Radiolaria, comprises all those Acantharia in which the acanthinic skeleton is only composed of radial spines arising from one common central point, but never forms a complete latticed shell. By the absence of such a latticed or fenestrated shell the Acanthometra differ principally from the nearly allied Acanthophracta, the second order of Acantharia, which constantly possess such a complete shell.

Johannes Müller, who first detected and described the Acanthometra (in 1855, loc. cit.), defined them as follows:—"Radiolaria without shell, with siliceous radial spines" (1858, Abhandl. d. k. Akad. d. Wiss. Berlin, p. 46). He described and figured eighteen species of them, disposed in four genera (Acanthometra with fifteen species, and Zygacantha, Lithophyllium, Lithoptera, each with a single species). Among those eighteen species, however, were two "Acanthometræ cataphractæ," appertaining to the following order, the Acanthophracta.

In my Monograph (1862, p. 371) all true Acanthometra were united into a single family, Acanthometrida, with the following definition:—"Skeleton composed of a number of radial spines, piercing the central capsule and united in its centre, without latticed shell." In the majority of them I observed that the skeleton did not consist of silex, but of a very peculiar organic substance, which I called "acanthin." At that time I divided the family Acanthometrida into four subfamilies:—(1) Acanthostaurida, (2) Astrolithida, (3) Litholophida, (4) Acanthochiasmida. The two former now represent the suborder Acanthonida, the two latter the suborder Actinelida. The number of genera which I distinguished in my Monograph amounted to nine, the number of species to fifty. By the rich collections of the Challenger this number is so much increased that we can here describe twenty-seven genera and one hundred and sixty species.

Richard Hertwig in his work on the Organismus der Radiolarien (1879, pp. 6-25) adopted my family Acanthometrida, and gave a very accurate description of its anatomical structure. He confirmed my observations that the radial spines of this family are never hollow, but solid, and that their chemical substance is not silex, but the organic matter "acanthin." He found that the simple nucleus of the Acanthometrida is commonly very early cleft, and that the peculiar brushes of filaments on the calymma, described by Johannes Müller and by me as "Gallert-cilien," are peculiar "contractile filaments," comparable to the "muscle-fibrillæ" of some Infusoria, or the "Myophan-filaments" (Myophrisca).

The order Acanthometra is here divided into two different suborders of very unequal extent and value, the Actinelida and Acanthonida. The first may be regarded as the common ancestral stock, not only of the second, but of all Acantharia. In the small group of Actinelida the number of radial spines is variable and commonly indefinite, often very large (more than a hundred); they are therefore Adelacantha. The second suborder, the Acanthonida, comprise by far the greatest part of the order, and possess constantly twenty radial spines, regularly disposed after the Müllerian law; they are therefore (like all Acanthophracta) Icosacantha (compare above, p. 717).

The Actinelida possess constantly simple radial spines, without any apophyses; their form is commonly very simple and primitive. This suborder comprises three small but very different families, the Astrolophida, Litholophida, and Chiastolida. The first family, the Astrolophida, is the original ancestral group. A large and variable, commonly indefinite number of radial spines is here united in the centre of the spherical central capsule and radiating within a spherical space. In the second family, the Litholophida, a small and variable number of radial spines (between ten and twenty) is united in the apex of a conical central capsule and radiating within the quadrant or octant of a spherical space. In the third family, the Chiastolida, a variable number of radial spines is grown together by pairs, in such a manner that every two opposite spines (placed originally in one axis of the spherical central capsule) forms a single "diametral spine"; all these diametral spines are not united in the centre of the central capsule but only crossed loosely near the centre.

The Acanthonida, the second suborder of Acanthometra, embraces by far the greatest number in this order, viz., all those forms in which twenty radial spines are regularly disposed after the Müllerian law—Icosacantha (compare above, p. 717). The radial spines of this suborder are either simple or provided with transverse processes (either two opposite or four crossed apophyses). They are commonly united in the middle of the central capsule by their opposed basal ends, forming small pyramids; the meeting triangular faces of the neighbouring pyramids being propped one upon another. Above these small basal pyramids often arises a basal leaf-cross formed by four broad triangular leaves or wings with straight edges; the meeting thin edges of the neighbouring spines serve for strengthening the basal junction and form hollow pyramidal spaces or compartments, filled with the contents of the central capsule (compare p. 721). The suborder Acanthonida comprises three different families, the Astrolonchida, Quadrilonchida, and Amphilonchida. The first family, the Astrolonchida, comprises by far the greater number of the Acanthonida; those genera in which all twenty spines are perfectly equal or nearly equal in size and form. In the second family, the Quadrilonchida, the four equatorial spines are much larger (and often also of another form) than the sixteen other spines (often also the eight tropical larger than the eight polar spines). The third family, the Amphilonchida, is distinguished by the preponderating development of only two opposite equatorial spines, which are much larger (and often also of another form) than the eighteen other spines.

Synopsis of the Suborders and Families of Acanthometra.

Suborder I. ACTINELIDA, Haeckel, 1882.

Definition.—Acanthometra with a variable number of radial spines, which are commonly irregularly disposed, not according to the Icosacantha.

Family XXXIII. Astrolophida, Haeckel.

Astrolophida, Haeckel, 1881, Prodromus, p. 469.

Definition.—Acantharia with a variable number of simple radial spines, radiating within a spherical space from one common central point which is the centre of the spherical central capsule. No lattice shell.

The family Astrolophida comprises the simplest and the most primitive forms among all Acantharia, and may therefore be regarded as the common ancestral stock of this whole legion or subclass of Radiolaria. The acanthinic skeleton is composed of a variable number of quite simple radial spines, which are united in the centre of the spherical central capsule and radiate, piercing its walls and the surrounding jelly-veil, within a spherical space.

The first observed form of this family is the ancestral genus Actinelius, two different species of which I detected in 1864 in the northern Mediterranean, at Villafranca, near Nice (compare Zeitschr. f. wiss. Zool., 1865, Bd. xv. p. 364, Taf. xxvi. fig. 4). Three other species of the same genus were afterwards found by me in the Challenger collections. Whilst in this Actinelius all radial spines are of the same size, a new nearly allied genus, Astrolophus (with two species), differs from it by the different size of the radial spines, a small number of very large spines being intermingled with a very large number of small spines. In these two genera, Actinelius and Astrolophus (the true "Astrolophida" sensu strictiori), the number of the radial spines is quite indeterminable and their arrangement quite irregular and variable.

A third remarkable genus, Actinastrum, differs from these two genera in the definite number and regular order of thirty-two radial spines, and may therefore perhaps better represent a peculiar family, Actinastrida. In this genus (of which two species were observed) the thirty-two radial spines are disposed in such a regular manner that they lie in four meridian planes, and that their distal ends fall into five parallel zones. These five zones and these four planes are the same as we find in all Icosacantha (compare above, p. 717). Also the constant twenty spines of these latter are present in Actinastrum; but their number is here enlarged by twelve other spines missing in the Icosacantha; four of these are secondary or interradial equatorial spines, lying opposite in pairs between the four primary or perradial equatorial spines; and eight are perradial tropical spines, lying between the eight interradial tropical spines. Therefore the distal ends of the thirty-two radial spines are disposed regularly in five parallel zones, and while two zones (the two polar) contain only the points of every four spines, three zones (the single equatorial and the two tropical) contain the points of every eight spines. The four meridian planes are in Actinastrum the same as in the Icosacantha, crossed in the spineless axis at angles of 45°. But in the Icosacantha each of the two perradial meridian planes contains six radial spines (two equatorial and four polar), each of the two interradial meridian planes only four tropical spines. Whereas in Actinastrum each of the two primary or perradial meridian planes contains ten spines (two equatorial, four tropical, and four polar), each of the two secondary or interradial meridian planes six spines (two equatorial and four tropical). We find therefore altogether thirty-two radial spines in three orders; eight equatorial, sixteen tropical, and eight polar spines.

Only one other genus of Radiolaria exhibits the same characteristic disposition of thirty-two radial spines as Actinastrum, and this is Chiastolus; but here the two opposite spines of each pair are grown together and form one diametral spine; and the sixteen diametral spines are crossed in the centre of the capsule. In Actinastrum, as in Astrolophus and Actinelius, the central ends or bases of all the spines are pyramidal, and the triangular faces of the neighbouring spines rest one upon another (as in the greater number of Acanthonida). The form of the radial spines in all Astrolophida is quite simple, without lateral processes or apophyses; chiefly cylindrical, more rarely compressed, two-edged or quadrangular.

The central capsule in all Astrolophida is spherical, and in the younger specimens contains a single large concentric and lobed nucleus, but in the older specimens a large number of small nuclei. The surrounding jelly-veil or calymma seems commonly to envelop the spines perfectly. The piercing pseudopodia radiate everywhere between the spines, and are very numerous and thin. The circulating granules in them are sometimes red (Actinelius purpureus).

Synopsis of the Genera of Astrolophida.

Genus 317. Actinelius,^[357] Haeckel, 1865, Zeitschr. f. wiss. Zool., Bd. xv. p. 364.

Definition.—Astrolophida with a variable and undetermined number of simple radial spines, all of equal size, united in the centre of the spherical central capsule.

The genus Actinelius comprises the most simple and primitive forms among all Acantharia, and may be regarded as the common ancestral stock of this whole legion. The spherical central capsule is pierced by numerous simple radial spines of equal size, the pyramidal bases of which are supported one upon another with their triangular faces in the centre of the capsule. The number and position of the spines are quite indefinite and variable. We may derive Actinelius either from Actissa (Colloidea) by development of acanthinic radial spines, or directly from Actinosphærium (Heliozoa) by formation of a central capsule.

Subgenus 1. Actinelarium, Haeckel.

Definition.—Radial spines cylindrical, conical, or spindle-shaped, their transverse section circular.

1. Actinelius primordialis, n. sp. (Pl. 129, fig. 1).

Spines sixty to eighty or more, cylindrical, at the distal end thickened, spindle-shaped. Apex simple. Base a small slender pyramid. Central capsule yellow. Granules of the sarcode colourless.

Dimensions.—Length of the spines 0.3 to 0.4, breadth in the distal part 0.02, in the basal part 0.008.

Habitat.—Central Pacific, Stations 265 to 274, surface.

2. Actinelius purpureus, Haeckel.

Spines thirty to forty or more, cylindrical, very thin, a little thinner towards both ends. Apex simple. Base a small sulcate pyramid. Central capsule opaque, purple. Granules of the sarcode also purple.

Dimensions.—Length of the spines 0.2 to 0.3, breadth 0.002.

Habitat.—Mediterranean (Nice), Haeckel.

Subgenus 2. Actinelidium, Haeckel.

Definition.—Radial spines compressed, two-edged; their transverse section elliptical or lanceolate.

3. Actinelius protogenes, n. sp.

Spines fifty to sixty, compressed, two-edged, gradually broadened towards the truncated distal end. Basal or proximal end thin, pyramidal. The spines of this species are similar to those of Actinastrum pentazonium (p. 733) and of Chiastolus amphicopium (Pl. 129, fig. 3), but much more numerous, smaller, and not regularly disposed. These latter two Actinelida must be separated on account of the regular disposition of the thirty-two spines.

Dimensions.—Length of the spines 0.2, basal breadth 0.008, distal breadth 0.02.

Habitat.—South Pacific, Station 165, surface.

Subgenus 3. Actinelonium, Haeckel.

Definition.—Radial spines quadrangular, prismatic, or pyramidal, their transverse section square.

4. Actinelius pallidus, Haeckel.

Spines eighty to one hundred and twenty or more, quadrangular, prismatic, of equal breadth throughout their whole length. Apex simple, truncate or pyramidal. Base a four-sided slender pyramid. Central capsule pale yellowish. Granules of the sarcode colourless.

Dimensions.—Length of the spines 0.2 to 0.3, breadth 0.005.

Habitat.—Cosmopolitan; Mediterranean, Atlantic, Pacific.

5. Actinelius polyacanthus, n. sp.

Spines two hundred to three hundred or more, quadrangular, pyramidal, gradually thinned towards the simple apex. Base a small three-sided pyramid. Central capsule opaque.

Dimensions.—Length of the spines 0.12 to 0.18, basal breadth 0.012.

Habitat.—South Pacific, Station 291, surface.

Genus 318. Astrolophus,^[358] Haeckel, 1881, Prodromus, p. 469.

Definition.—Astrolophida with a variable and undetermined number of simple radial spines of different sizes (large and small spines intermingled), which are united in the centre of the spherical central capsule.

The genus Astrolophus differs from the nearly allied ancestral genus Actinelius only in the unequal size of the numerous radial spines. In both observed species very numerous small spines are intermingled with a small number of large spines, and between them numerous spines of medium size. The small spines fill up the hollow spaces between the basal parts of the large spines.

1. Astrolophus stellaris, n. sp.

Radial spines from one hundred to two hundred, of very different sizes, but of similar form; about sixteen to twenty very large spines, forty to fifty of medium size, and one hundred to one hundred and twenty much smaller. All spines cylindrical in the greater part of their length, with simple apex, gradually thickened towards the central part, conical, without edges. The base itself is a slender pyramid with four to eight edges.

Dimensions.—Length of the largest spines 0.3 to 0.4, of the majority 0.1 to 0.2, of the smallest 0.05 to 0.1.

Habitat.—South Pacific, Station 288, surface.

2. Astrolophus solaris, n. sp. (Pl. 132, figs. 12a, 12b).

Radial spines from two hundred to three hundred, of very different sizes, but of similar form; about twenty to thirty very large spines, sixty to eighty of medium size, and one hundred and twenty to one hundred and fifty much smaller. All spines cylindrical in the greater part of their length, with simple apex, gradually thickened and four-edged towards the central base. The base itself is a slender pyramid with four to eight edges; partly the faces, partly the edges of these basal pyramids rest one upon another, the points of the larger spines meeting in the centre.

Dimensions.—Length of the largest spines 0.4 to 0.5, of the majority 0.2 to 0.3, of the smallest 0.1 to 0.16; basal thickness of the largest spines 0.015.

Habitat.—South-east Pacific (off Juan Fernandez), Station 296, surface.

Genus 319. Actinastrum,^[359] n. gen.

Definition.—Astrolophida with thirty-two simple radial spines, regularly disposed within four meridian planes in such an order that their distal ends fall into five parallel zones. Central ends of the thirty-two spines supported one upon another in the centre of the spherical central capsule.

The genus Actinastrum differs from the two preceding genera in the definite number and order of the thirty-two radial spines, which are disposed in a very remarkable manner. Twenty radial spines are disposed after the Müllerian law of Icosacantha (compare above, p. 717). The remaining twelve spines are four equatorial spines lying in the two secondary meridian planes, and eight tropical spines lying in the two primary meridian planes. We have therefore together eight equatorial, sixteen tropical, and eight polar spines (compare above, p. 729).

1. Actinastrum legitimum, n. sp.

All thirty-two radial spines of equal size and similar form, cylindrical, conical at the distal end, at the central base pyramidal. Central capsule pellucid, colourless.

Dimensions.—Length of the radial spines 0.3, breadth 0.004.

Habitat.—South Pacific, Station 288, surface.

2. Actinastrum pentazonium, n. sp.

All thirty-two radial spines of equal size and similar form, compressed, two-edged, gradually becoming broader and thinner from the pyramidal central base towards the truncated distal end. Central capsule dark, opaque. (Compare the similar Chiastolus amphicopium, Pl. 129, fig. 3.)

Dimensions.—Length of the radial spines 0.2, breadth at the base 0.005, at the distal end 0.02.

Habitat.—South Pacific (west coast of Patagonia), Station 302, surface.

Family XXXIV. Litholophida, Haeckel.

Litholophida, Haeckel, 1862, Monogr. d. Radiol., p. 401.

Definition.—Acantharia with a variable number of simple radial spines radiating within a conical space (or within the quadrant of a sphere) from one common central point, which is the apex of the conical central capsule. No lattice-shell.

The family Litholophida, represented only by a single genus, Litholophus, differs from all other Acantharia in the remarkable fact that the common point, from which the radial spines arise, is not the geometrical central point of the whole body, but is quite excentric in position, the apex of the conical or pyramidal central capsule. Therefore the spines form together a kind of brush or broom.

When I founded the family Litholophida in my Monograph (1862, p. 401) I knew only a single species, Litholophus rhipidium, observed very frequently in Messina. Another species, Litholophus ligurinus, was afterwards (1864) found by me at Nice. Six other species were detected in the preparations of the Challenger, some of them very frequent. All these eight species of Litholophus are very nearly allied, and exhibit only slight differences in the form and number of the radial spines; their mode of excentric connection and the structure of the peculiar soft body is everywhere the same.

The radial spines in all observed Litholophida possess the form of the genus Acanthonia, i.e., they are quite simple, four-sided prismatic or quadrangular, with square transverse section; their four edges are sometimes smooth, at other times elegantly denticulate, commonly more or less prominent or wing-shaped. In the greater number of species they are very long and of nearly equal breadth, prismatic; in some species they are more pyramidal, thinned towards the distal end; the latter is commonly truncated or broken off, sometimes pyramidal. The central end is everywhere thinned, more or less pyramidal, and the neighbouring spines are propped one upon another by the triangular faces of their small basal pyramids. A slight pressure is sufficient to destroy their connection.

The number and disposition of the radial spines seem to be variable and irregular, but require further researches. In four of the observed eight species I found constantly ten spines, in two other species from ten to twenty (commonly twelve or sixteen), and in two species twenty or more. A certain order or disposition of the spines within the conical space in which they radiate could nowhere be ascertained.

When I first observed Litholophus, I supposed that it might only be a mutilated or altered form of an Acanthonia. Afterwards, observing many specimens with ten spines, I was led to the suggestion that they were produced by self-division of an Acanthonia, and that the number of the spines in each half of the body might be afterwards doubled. But this suggestion seems to be refuted by the fact that in no other genus of the numerous Acantharia is self-division observed, and that many hundreds of Litholophus which I observed exhibit quite constantly only a single form of radial spines, that of Acanthonia—simple quadrangular spines without any apophyses.

Genus 320. Litholophus,^[360] Haeckel, 1862, Monogr. d. Radiol., p. 401.

Definition.—Litholophida with a variable number of quadrangular diverging radial spines, united with pyramidal bases in the apex of the conical central capsule.

The genus Litholophus, the only one of this family, exhibits the peculiarities just described, but might more nearly be defined as a typical "genus" by the quadrangular form of the radial spines, identical with those of Acanthonia.

The central capsule of Litholophus is constantly conical or pyramidal, commonly opaque, of a dark brownish or reddish colour; it contains many small nuclei. It envelops the basal half of all radial spines in such a manner that their basal parts are united in its apex, and their distal parts pierce the rounded base of the conical capsule (Pl. 129, fig. 2).

The calymma or the jelly envelope of the central capsule is only developed at its base, where the spines radiate; at the conical mantle of the capsule it is very thin. The spines seem to be perfectly enclosed in the calymma and connected with it by the same contractile retinacula or "myophrisca" which we observe in the Acanthonida. The pseudopodia arise only from the rounded base of the conical capsule, and radiate between the spines, piercing the calymma, diverging within the conical space occupied by the fascicle of spines.