Sponges

The genus Euspongia, to which belong all the finer bath sponges, is a typical example of the true " horny " sponges or Euceratosa, characterized especially by the fact that the skeleton is not composed of spicules but of so called horny fibres. A living bath sponge appears as a dark-coloured, irregular or sometimes cup-shaped mass attached by the under surface to the sea-bottom. The outer surface is covered by a skin or dermal membrane, elevated in innumerable minute conuli by the growing apices of the primary ? " ?J ?G o 'JO`)?o=. F,: ^ J?

???.;: li.' i.?? 4G`j????0 '::-' n` "Y,?v? ?F?'? V ??

n. - 94,j? .?

J ??n?FW s ? 1 `?m? ????is G> op'"o, 3, ? f),...9..„,_,?? pr ?!?06Q ??

?© C -.:!5 "'?` 1 ? 3.. r?' e. 5 ?J ".)??Jpu b. '`s?(30V.- S"-pp'-; fii?? oe: ?306 ??'?; ?? o ?'? © ? oo 0 O. :1 ? ? O. 3 ? G'?. <'- 1G J??c),;)€),..,-1,-ii. ;,a , O q- ?oc,('c',??Doo;?.;?.

e) ? 4f,c? oe ., 0,2c,0,% ? 3 ?°' `' to; v U C ? C, o` p c^:?c;,o??-,c? ???.,???`?I. i ? :'1n ! `4 (J GJ .Ll`'?O'?O?? !;'l.el (? (??. V V. .. ? V. ?;.,1) ?r. ,C7e0 O ? S s J a. 13 '`c ? V ? o?n??. c3?? ???6.O???c:aoac>c'<: i O Cn:.jEl'+J --cr. o?F!,0?,.

o.?.??:'>. ? Fi:: ? ? Z c 1 ? p?;` speak of under the name " sponge." It consists of a very close network of spongin fibres (closely resembling silk in chemical composition), some of which, known as primaries, run towards the surface at fairly regular intervals, while others, known as secondary (After F. E. Schulze. From a coloured plate in Zeits. fur Wissen. Zoologie. by permission of Wilhelm Engelmann.) FIG. 6. - Euspongia officinalis (bath sponge). Part of vertical section showing general arrangement of skeleton and canal-system.

p.f, Primary fibre of skeleton. i.c, Inhalant canals.

s.f, Secondary fibres. e.c, Exhalant canals.

d.p, Dermal pores (inhalant). f.c, Flagellated chambers. fibres, connect the primaries in all directions and themselves branch and anastomose freely. The primary fibres contain particles of sand or foreign spicules which are taken in by their growing (After F. E. Schulze.) FIG. 4. - Plakina monolopha. a, Ciliated embryo (the central e, Rhagon stage, viewed as a part should be shaded). transparent object, show b, Part of section of ciliated ing the inhalant pores on embryo. the surface and the flagel col, Inner cell-mass. lated chambers in the ec, External, columnar cells. interior; the osculum is not fl, Flagella. shown.

c, Attached embryo, viewed f, Part of vertical section from above, with the gasthrough adult sponge, tral cavity appearing in showing the folded choano the interior. somal lamella or spongo d, Vertical section of attached phare.

embryo. ov, Ova. bl, Embryo.

skeleton fibres. This skin is pierced by a vast number of inhalant dermal pores of microscopic size, and by a much smaller number of comparatively large vents or oscula. When the sponge is removed from the water the soft tissues rapidly decay and leave behind only the elastic ' ` horny " skeleton, which is what we usually (After F. E. Schulze. From a plate in Zeitschriftfiir Wissen. Zoologie, by permission of Wilhelm Engelmann.) FIG. 5. - Plakina monolopha. Spicules, a - e, tetracts or calthrops; f - k, triacts or triradiates; l - t, diacts, showing how the monaxon form (I) may be derived from the primitive tetract (a) by suppression of actines.

dZ (After F. E. Schulze. From Lankester's Treatise on Zoology.) FIG. 7. - Euspongia officinalis (bath sponge). Skeleton. Fibre surrounded by spongoblasts. sp.f, Spongin fibre; sp.bl, Spongoblasts. Coll, Collencytes. apices at the surface of the sponge, and the presence of which may greatly injure the quality of the sponge. The connecting fibres are only about 0.035 mm. in diameter, or even less, and the primaries are a little thicker, while the meshes between the fibres are so narrow as to permit of the soaking up of water by capillary attraction, (After F. E. Schulze.) FIG. 8. - Euspongia officinalis (bath sponge). Diagram of the arrangement of the canal-system as seen in vertical sections of two young individuals.

d.p, Dermal pores; o, Oscula; r, Rock to which the sponges are attached.

the property upon which the economic value of the bath sponge depends. In the living sponge the fibres are embedded in the mesogloea, where they are secreted by special cells known as spongoblasts, which are often found thickly clustering around them (fig. 7). The canal-system (figs. 6, 8) is very complex and shows but little indication of its origin from a folded rhagon. The inhalant pores lead each into a short, narrow, inhalant canal; these unite in roomy subdermal cavities lying in the ectosome, and from these in turn the main inhalant canals come off. The latter divide and subdivide, and thus ramify through the deeper parts of the sponge amongst the flagellated chambers, to each of which a small number of slender canaliculi are ultimately given off (fig. 9). The chambers themselves, lined by the usual collared cells, are small and approximately spherical, and each one discharges its water through a short and narrow exhalant canaliculus (fig. 9). The openings of the inhalant canaliculi into the chambers, of which there are several, correspond to the prosopyles of an Olynthus, while the single exhalant opening, or apopyle, may possibly correspond to an Olynthus osculum. The exhalant canaliculi unite together to form larger and larger canals which finally lead the stream of water to the vents on the surface of the sponge (fig. 8). The various parts of the canal-system, other than the chambers themselves, are lined by a flat pavement-epithelium, and the mesogloea, occupying all the spaces between the different parts of the canal-system, contains cells of various kinds, embedded in a very granular matrix.

Comparative Anatomy. External Characters. - Amongst the simpler calcareous sponges, which are all of comparatively small size, the external form is usually symmetrical and is evidently a kind of outward expression of the arrangement of the canal-system. This is well seen in the simplest form of all, the sac-shaped Olynthus, and also in its simpler Syconoid and Leuconoid derivatives (described later on), which may be regarded either as individuals of a higher order or as colonies of Olynthus persons grouped around a central individual whose large gastral cavity opens to the exterior through the single osculum. In the more complex Leuconoids, however, the process of colony formation becomes very irregular and may give rise to great compound masses, with many vents. In these masses we may perhaps recognize the presence of individuals of three orders: (I) the primitive Olynthus persons, represented by the individual flagellated chambers; (2) the Leuconoid persons, indicated each by its osculum; and (3) the entire colony formed by the union of many such Leuconoid persons in an irregular manner. It is, however, very doubtful how far the flagellated chambers in such forms as this can be regarded as morphologically equivalent to Olynthus persons.

In the non-calcareous sponges we are always dealing with individuals of a high order, which usually form complex aggregates (colonies) of large size and very various shape. As a general rule the form of those non-calcareous sponges which grow in shallow water is extremely irregular and variable while at great ocean depths the shape is usually definite, constant and often exquisitely symmetrical, a fact which may perhaps be accounted for in part by the absence of disturbing influences such as are met with in shallow water. Perhaps the most extraordinary external form yet discovered is that of Esperiopsis challenchallengeri: a deep-water geri, discovered by the " Challenger " Monaxonellid Sponge. expedition in deep water off the Philip pine Islands (fig. to), a form which reminds one strikingly of a number of flowers arranged in a raceme, except that the largest and oldest member of the compound colony is at the top of the stalk and the smallest at the bottom. In other deepwater species the external form may frequently be explained (Alter Ridley and Dendy. From "Challenger" Reports, xx., by permission of the Controller of H. M. Stationery Office.) FIG. II. - Cladorhiza longipinna: a deep-water Monaxonellid Sponge, showing the "Crinorhiza " form, adapted for support on soft ooze.

as an adaptation to the special exigencies of the environment. Thus, for example, many species are provided with long stalks which lift up the body of the sponge out of the soft in which it would otherwise be smothered, while the bottom of the stalk is frequently extended in root-like processes which serve to attach it to some solid object (e.g. Stylocordyla). In other cases the sponge supports itself on the surface of the ooze by long stiff processes, formed of bundles of spicules which radiate from the central, cap-shaped body; this is known as the " Crinorhiza form," and is met with in several distinct genera (fig. I 1). Amongst the Hexactinellida, which are essentially a deep-water group, many very beautiful external forms are met with, the best known, perhaps, being the so-called Venus's flower basket (Euplectella, fig. 12).

Flabellate (orfan-shaped) and cupshaped forms are frequently met with even amongst shallow-water sponges, and in widely separated genera, such as Poterion (the great Neptune's cup sponge) and Reniera testudinaria. In Phyllospongia the flabellate and cup-shaped forms pass insensibly into one another, the cup being apparently merely a folded lamella. Slender branching forms are also not uncommon in shallow water, as seen in the common Chalina oculata of the British coast. Spherical forms, such as Tethya, likewise occur. By far the greater number of shallow-water sponges, however, are quite irregular in shape and either form crusts of varying thickness on the surface of rocks and sea-weed, or large and massive aggregates which may rise to a considerable height above the substratum. In the boring sponges (Family Clionidae) the sponge occupies an elaborate system of chambers and passages which it excavates for itself in the shells of Mollusca and other calcareous organisms. The common British Cliona celata begins (After F. E. Schulze. From a plate in Zeits. fiir Wissen. Zoologie, by permission of Wilhelm Engelmann.) FIG. 9. - Euspongia officinalis (bath sponge). Part of a section such as is shown in fig. 6, more highly magnified, showing three flagellated chambers, with inhalant canaliculi on the left and exhalant canaliculi on the right.

(After Ridley and Dendy. From a plate in " Challenger " Reports, xx., by permission of the Controller of H.M. Stationery Office.) FIG. Io. - Esperiopsis ooze (After F. E. Schulze. From a plate in " Challenger " Reports, xxi., by permission of the Controller of H.M. Stationery Office.) FIG. 12. - Euplectella aspergillum, " Venus's Flower Basket ": a Hexactinellid Sponge.

life in this way, but soon outgrows the housing capacity of its host, whose shell then serves merely as a base of attachment for the large independent sponge-colony.

One of the most striking features of living sponges is their colour, which is often very brilliant. Yellow, red, orange, purple, brown, black, green and blue are all met with, in varying degrees of purity and intensity, amongst the commoner Non-calcarea; whilst the calcareous sponges are usually white. It appears probable that the colour is more or less constant for each species, and may therefore afford a useful guide to specific identification. As a rule the colour is lost in spirit-preserved or dry specimens, but a noteworthy exception is found in the brilliant purple Suberites wilsoni of Port Phillip, in which the colour, though soluble in water, is permanent in dry specimens and in alcohol. The colouring matter is sometimes lodged in special pigment cells belonging to the sponge itself, and sometimes in symbiotic algae, with which the mesogloea is frequently filled.

Canal-system. - Whether we start with the primitive Olynthus form of the Calcarea or with the more advanced Rhagon of many Non-calcarea, it is evident that further advance in the complication of the canal-system is arrived at either by budding or folding, or by a combination of these processes. As, however, the canalsystems of the calcareous and of the main types of non-calcareous sponges have been evolved along perfectly independent lines it will be well to consider them separately.

In the genus Leucosolenia (Calcarea Homocoela) the primitive Olynthus form may, as we have already seen, give rise, by branching and anastomosing, to complex reticulate colonies of the Clathrina type, in which a pseudoderm, pierced by inhalant pores, may cover over a system of inhalant canals which are simply the interspaces between the branching tubes of which the colony is made up, while at the same time a centrally placed pseudogaster, which is simply a space enclosed by upgrowth of the colony around it, may form the main exhalant canal and open to the exterior through a well-defined vent or pseudosculum. In this direction perhaps the most remarkable modification arrived at is that of Leucosolenia cavata, in which the Clathrina tubes, lined, z by collared cells, widen out into large irregular spaces, while the inhalant interspaces t to spaces become constricted into narrow canals lined by collared cells on the outside. We have here a kind of inversion of the ordinary Clathrina canalsystem, but a perfectly gradual transition from the ordinary to the ! inverted condition is seen as we pass from the older to the younger parts of the colony.

In Leucosolenia (Dendya) tri- r? podifera (fig. 13) we find a totally different type of colony formation, which is of great importance as - p indicating in its canal-system the n possible starting-point of a line, of evolution which culminates in the highest Calcarea. Here a large central individual, whose i spacious gastral cavity is lined }" by collared cells, gives off radial;' t ': ? y ? ') 0 6 0 o ('c? buds from all sides, which branc 000 slightly and terminate in blind 3° ends in contact with one another, -4 > " `, Yrj Y '?,'? so that the entire colony has an o?'oc, approximately even surface. The ooc

?;' inhalant canals are represented by X .

°: .? ? ?Ithe interspaces between the radial

, tubes, between the blind extremi o ?c i ties of which the water finds its a 6:

way in from the outside. There 'gie' I is only a single vent, situate at (,() the extremity of the central cavity.

o This cavity must be regarded as the original gastral cavity of a parent Olynthus, from which the radial tubes have been produced by budding.

(After Dendy. Simplified from a coloured We have next, amongst the plate in Trans. Roy. Soc. of Victoria, Calcarea Heterocoela, the Sycon Melbourne, vol. iii. pt. r.) Y e FIG. 13. - Leucosolenia tripodi- from o the foregoing in hl that the fern, with part of the spongecollared cells of the central gastral wall cut away to show the cavity are replaced by pavementarrangement of the radial outepithelium. The radial tubes now growths. form definite flagellated chambers, pierced as before by numerous prosopyles through which the water enters from the spaces between the chambers, while the original gastral cavity forms a central exhalant canal terminating in the single vent, a true osculum, corresponding to the osculum of an Olynthus. In the simplest Syconoid forms (Sycetta) the radial chambers remain perfectly straight and unbranched. They do not touch one another at all and there is no trace of an ectosome or dermal cortex, and hence there are no true inhalant canals, and the water circulates without interruption between the chambers. In the genus Sycon (fig. 14) the walls of adjacent chambers come into contact with one another and fuse together and thus give rise to more or less well-defined inhalant " inter-canals." The chambers themselves may branch, and in some species of Sycon a thin, pore-bearing dermal membrane connects to gether their distal extremities (From Dendy, in Quart. Journ. Micro. Sci., and covers over the entrances new series, xxxv., by permission of J. and to the inhalant canals. The A. Churchill.) canal-system now exhibits all FIG. 14. - Sycon carteri, part of the different parts found in a transverse (horizontal) section, the most highly-organized showing three radial chambers, the sponges: viz. dermal pores, middle one cut open.

inhalant canals, flagellated fl.ch, Flagellated chamber. chambers, exhalant canal and ex.op, Its exhalant opening osculum. In the genus Grantia apopyle.

and its allies (e.g. Ute, fig. 1 5) pros, Prosopyle.

the thin dermal membrane c.g.c, Central gastral cavity. of Sycon is converted into a i.c, Inhalant canal. well-developed cortex, cover- g,cor, Gastral cortex.

ing the extremities of both g.q, Gastral quadriradiate the inhalant canals and the cule.

radial chambers, and some- s.g.s, Subgastral sagittal trira times containing a system diate spicules, forming the of special cortical inhalant first joint of the articulate canals. We may now distubar skeleton.

tinguish between an ectosome t.ox, Tufts of monaxon spicules at (the dermal cortex), which the ends of the chambers. contains no flagellate cham bers, and a choanosome in which chambers are present. The next stage has probably been arrived at by a kind of fold (After Polejaeff.) FIG. 15. - Ute argentea, part of transverse section, showing the Syconoid canal-system, and thick dermal cortex containing huge longitudinally placed monaxon spicules whose cross-sections are represented by concentric circles.

radially, not around the central gastral cavity but around diverticula of the latter which form special exhalant canals. This condition, sometimes called the " sylleibid " type, is not characteristic of any particular genus or family, but occurs in a few isolated species, such as Leucilla connexiva (fig. 16). A somewhat similar condition may be arrived at by branching of the radial flagellated chambers, as in Heteropegma (fig. 17). The next stage is marked by great reduction in the size of the chambers, which may become almost spherical, and by further folding of the choanosome, so that in a section of the sponge-wall we see the small chambers scattered irregularly in the mesogloea between the numerous branches of complicated inhalant and exhalant canals. Each ‚pros. or spi ing of the choanosome, for find the chambers arranged we chamber still has several prosopyles, through which it receives water from the ultimate branches of the inhalant canals, while it opens into a relatively large exhalant canal by a wide apopyle. This is the highest type of canal-system met with amongst the Calcarea. It is sometimes known as the Leucon type and is seen in most species of the genus Leucandra, as well as in many others.

(After Polejaeff.) FIG. 16. - Leucilla connexiva, part of transverse section, showing " sylleibid " type of canal system with folded chamber layer and exhalant canals (E) into which the chambers open.

It is almost identical with one of the types commonly found in non-calcareous sponges (e.g. Plakina, fig. 4), but has of course been evolved independently. The various types of canal-system met with in the Calcarea are connected together by numerous intermediate forms, thus forming a very interesting evolutionary series, while both the Sylleibid and Leuconoid types appear to have been independently evolved several times, thus affording excellent examples of the phenomenon of convergence, a phenomenon which is very frequently met with amongst sponges.

(After Polejaeff.) FIG. 17. - Heteropegma nodus-gordii, part of transverse section, showing branching flagellated chambers and huge subdermal quadriradiate spicules, with greatly reduced tubar skeleton.

In describing the anatomy of Plakina as a type of non-calcareous sponge, we have traced the development of a fairly complex canalsystem from the so-called Rhagon form. We can, however, hardly regard the Rhagon as representing a fundamental type of canalsystem common to all the Non-calcarea, for in some of the Myxospongida, which are the most primitive of all, and again in the Hexactinellida, we find a type characterized by the presence of elongated sac-shaped flagellated chambers resembling those of the Sycon type amongst the Calcarea, and these chambers are arranged radially around the exhalant canals (Halisarca, Hexactinellida). The first recognizable stage in the evolution of the canal-system of the Non-calcarea would thus appear to be a condition not unlike that of Sycon, with a number of elongated chambers arranged radially around a central gastral cavity and having their blind outer extremities covered over by a dermal membrane. This stage is very nearly reproduced in the young form of a Hexactinellid sponge, Lanuginella pupa. From some such form the Rhagon type may perhaps be derived by flattening out of the lower end of the sponge into a broad base of attachment, and by reduction in the size of the flagellated chambers, accompanied by a more irregular arrangement.

Starting from the primitive Myxosponge ancestor, with large sac-shaped chambers, radially arranged, the Non-calcarea have apparently developed along four main lines, giving rise to the existing Myxospongida, the Hexactinellida (Triaxonida), the Tetraxonida osc. (After F. E. Schulze. From Lankester's Treatise on Zoology.) FIG. 18. - Lanuginella pupa. O.S., Vertical section of a young specimen (spicules omitted).

d.m, Dermal membrane. g.m, Gastral membrane. sd.tr, Subdermal trabecular layer. G.C, Gastral cavity.

fl.c, Flagellated chamber. osc, Region of future osculum. sg.tr, Subgastral trabecular layer.

and the Euceratosa. The Myxospongida have retained the large size of the chambers in certain forms (Halisarca, Bajalus), but have lost this primitive character in the more advanced members of the group (Oscarella). The Hexactinellida have retained the large size and radial arrangement of the flagellated chambers throughout their entire series. The chamber layer, however, tends to become more or less folded (fig. 19), and always lies between two layers of nt. p?((After Schulze. From Lankester's Treatise on Zoology.) FIG. 19. - Section of the Body-wall of Bathydorus fimbriatus, F.E.S. (spicules omitted).

ex.c, Exhalant canals. sg.tr, Subgastral trabecular layer.

d.on, Dermal membrane. g.m, Gastral membrane. sd.tr, Subdermal trabecular layer. G.C, Gastral cavity. fl.c, Flagellated chambers.

loose trabecular tissue in which the canals are represented by irregular spaces. The Tetraxonida appear to have suffered reduction in the size of the flagellated chambers at a very early date, and it is of this group especially that the Rhagon type is characteristic (e.g. Plakina, fig. 4). The Euceratosa exhibit a beautiful series, beginning with forms (Aplysillidae) having large sac-shaped chambers like those of Hexactinellids and ending with forms (Spongiidae, Euspongia, figs. 6, 8, 9) having small spherical chambers.

Along all four lines of descent it is probable that folding of the choanosome, or chamber-bearing layer of the sponge-wall, has played a very important part in the evolution of the canal-system. This folding is very clearly seen in the Hexactinellida and in such forms as Oscarella (Myxospongida) and Plakina (Tetraxonida). By this process inhalant and exhalant canal-systems have been formed, and then the ends of the inhalant canals have in most cases been closed in by development of an ectosome, as in Plakina trilopha and Stelletta phrissens (fig. 20). In the majority of cases (e.g.

(After Sollas.) FIG. 20. - Young specimen of Stelletta phrissens (Sollas). Vertical section through the osculum (o), showing the choanosome folded within the ectosome.

Euspongia) the folding has become so complex that it is no longer recognizable as such, and the origin of the now well-defined inhalant and exhalant canals is completely disguised. In many cases the principal exhalant canals may be surrounded by a layer of tissue of considerable thickness in which there are no flagellated chambers at all, known as the endosome, so that the folded choanosome may be sandwiched in between ectosome on the outside and endosome on the inside.

The manner in which the flagellated chambers communicate with their respective branches of the inhalant and exhalant canal (After Sollas.) FIG. 21. - Transverse section across an exhalant canal and surrounding choanosome of Cydonium eosaster (Sollas), showing the aphodal flagella,ed chambers.

system varies considerably in different forms, and the following types are recognizable, though by no means sharply distinguished from one another. In the more primitive forms (e: g. Hexactinellida, Aplysillidae, Spongeliidae) each chamber is provided with several prosopyles and receives its water supply direct from relatively large inhalant canals or even lacunae, discharging it again through a wide mouth (apopyle) into a relatively large exhalant canal or lacuna which also receives water directly from other chambers.

To this type (fig. 4, f) the name "eurypylous " has been given, and we may include in it cases where there is only a single prosopyle, and perhaps even a short, narrow inhalant canal. In more advanced forms the water is discharged from each chamber through a narrow exhalant canaliculus (aphodus) peculiar to itself, and thence into wider canals. This is known as the " aphodal " type (e.g. Cydonium, fig. 21). In the " diplodal " type there is a special inhalant canaliculus (prosodus) as well as a special aphodus to each chamber, with usually, at any rate, only a single prosopyle (e.g. Corticium, fig. 22). The progress from the eurypylous to the diplodal condition is accompanied by a corresponding increase in the development of the mesogloea, whereby the canals are greatly restricted in diameter, and at the same time the mesogloea tends to lose its transparent FIG. 22. - Part of a section of gelatinous character and to Corticium candelabrum, O.S., showbecome compact and granular. ing diplodal type of canal-system. With the growth of the ectoThe canal shown on the left is some we necessarily get a inhalant and that on the right (e) corresponding development of exhalant.

the proximal portion of the inhalant canal-system. At first the ectosome is merely a thin membrane, the dermal membrane, pierced by the inhalant pores, which are usually arranged in groups.

Beneath the groups of pores (pore-areas) lie spacious subdermal cavities which form the commencement of the inhalant canal-system in the choanosome. In more advanced types the ectosome becomes greatly thickened and may be specially strengthened in a variety of ways to form a cortex. The inhalant pores now no longer lead directly into the subdermal cavities, but first into a series of cavities lying in the cortex and known as chones, which may be separated from the underlying subdermal cavities (sub-cortical crypts) by definite sphincters (Cydonium, fig. 23).

The arrangement of the oscula and pores on the surface of the sponge varies greatly in different types, and sometimes gives rise to very striking modifications of the external form. The oscula or vents are usually relatively large openings situated on the more prominent parts of the sponge, often on special elevations. Occasionally they are replaced by sieve-like oscular areas (e.g. Geodia perarmata), a modification which doubtless serves to prevent foreign bodies from entering the wide exhalant canals. The inhalant pores may be irregularly scattered over the surface of the sponge or collected in more or less well-defined pore-areas. In cup-shaped sponges the pores are usually confined to the outer and the oscula to the inner surface. In flabellate sponges we find pores on one side and oscula on the other. In Tedania actiniiformis, a deepsea form, the pores are restricted to a narrow band surrounding the columnar body of the sponge just beneath the flattened top, which bears the vents; thus they are kept from being choked up by the soft ooze on which the sponge lies. In Xenospongia, a flattened discoid form, they are confined to narrow grooves on the upper surface, the chief of which run round the margin of the disk. In Esperella murrayi the pores are also confined to special grooves on the surface of the sponge, and in both these cases the grooves can apparently be opened and closed by special bands of musclefibres, and the supply of water thus regulated. In some species of Latrunculia we find the surface of the sponge covered with (After F. E. Schulze.) (After Sollas.) FIG. 23. - Section through the cortex and part of the choanosome of Cydonium eosaster (Sollas), showing a pore-sieve and underlying chone in the cortex. The chone communicates below with a subcortical crypt, from which the inhalant canals originate. The cortex contains numerous sterrasters, connected with one another by fibrous bands.

conspicuous projections of two kinds, some conical and bearing each a single vent, others truncated at the top and bearing the inhalant pores.

The original ancestral form (Protolynthus) from which all the Porifera are supposed to be descended, probably possessed no proper skeleton at all, and this condition has been retained in the existing Myxospongida, although these sponges have made considerable progress in the evolution of their canalsystem. There appears to be little doubt that the Myxospongida are primitively devoid of skeleton, and in this respect they must becaref ully distinguished from thegenus Chondrosia, in which the skeleton has been secondarily suppressed, as well as from numerous and divers species in which the proper skeleton has been more or less completely replaced by grains of sand or other foreign bodies. The Calcarea, Triaxonida, Tetraxonida and Euceratosa, except in cases of extreme degeneration, all possess a well-developed proper skeleton. As this skeleton has been independently evolved in each of these great groups it is necessary to deal with it separately in each case.

The skeleton in this group is composed of spicules of crystalline carbonate of lime (usually calcite), developed within special mother-cells or scleroblasts. Each spicule is enclosed in a delicate membranous spicule-sheath and contains an axial thread of organic matter. Three main types of calcareous spicule are met with, triradiate, quadriradiate and monaxon (fig. 24). The triradiates and quadriradiates, however, are not simple spicules, but spiculesystems formed of three or four rays each originating independently from its own scleroblast (actinoblast) and all uniting together secondarily. There is reason to believe that this may also sometimes be the case with the monaxon or oxeate spicules. In the most primitive triradiate spicules all three rays lie in the (After E. A. Minchin. From Lankester's Treatise on Zoology.) FIG. 24. - Spicules of Calcareous Sponges.

same plane. Three chief varieties may be distinguished: (I) Regular (fig. 24, li), with all the rays and all the angles equal; (2) Sagittal (fig. 24, c, d, 1, &c.), with two of the rays or two of the angles forming a pair, differentiated in some respect from the remaining ray or angle, the paired rays being termed " oral " and the odd ray " basal "; (3) Irregular (fig. 24, p), when conforming to neither of the above types. It has been proposed to draw a very sharp distinction between " equi-angular " triradiates and " alate " forms (in which the angle between the oral rays differs from the paired angles), but it may be doubted whether such a distinction has any great value. The quadriradiate (fig. 24, e, f, k, m) is formed by the addition of an " apical " or "gastral " ray to the three " facial " rays of the triradiate; this ray lies in a plane at right angles to that of the facial rays. The monaxon spicules (fig. 24, h, i, q, r, s) are straight or curved and the two ends are usually more or less sharply differentiated from one another. In all these spicules the form and arrangement of the rays is usually clearly correlated with their position in the sponge in such a manner that they are specially adapted for the work which they have to do.

The arrangement of the spicules in the case of the genus Leucosolenia has been dealt with above, and we must pass on at once to the Calcarea Heterocoela. In this group the skeleton exhibits an evolutionary series no less remarkable than that of the canal-system. We may take as a convenient starting-point the genus Sycetta, a typical Syconoid form, with the flagellated chambers radiating independently from the central gastral cavity. The wall of the gastral cavity is supported by a gastral skeleton of triradiate or quadriradiate spicules. These may be sagittal, in which case the oral rays are turned towards the osculum while the basal ray is directed downwards. If there is an apical ray it projects into the gastral cavity. The walls of the radial chambers are supported by a special " tubar " skeleton (cf. fig. 14), consisting exclusively of triradiates with their basal rays directed towards the distal end of each chamber. The oral rays are spread out at right angles to the length of the chamber, and as several spicules generally lie at the same level the tubar skeleton forms a series of more or less definite joints and is said to be " articulate." This type of skeleton is almost invariably associated with the Syconoid type of canalsystem. In the genus Sycon itself we find the distal ends of the chambers specially protected by tufts of monaxon spicules (fig. 14), but the next great advance in the evolution of the skeleton is brought about by the development of a dermal cortex, in which a special dermal skeleton is developed. This is well seen in the genus Ute (fig. 15). After this the skeleton of the chamber layer in the spongewall begins to undergo modifications, some of which are obviously correlated with the gradual change of the canal-system from the Syconoid to the Leuconoid condition (cf. figs. 16 and 17). Finally all trace of the articulate tubar skeleton is lost, and we get a " parenchymal " skeleton of scattered radiate spicules in the chamber layer. The skeleton of the chamber layer, no matter what the type of canal-system, may be supplemented by large subdermal sagittal triradiates or subdermal quadriradiates (fig. 17), whose basal or apical rays project inwards from the dermal cortex (Heteropidae and Amphoriscidae). Very generally a special " oscular " skeleton is developed in the form of a fringe of long monaxon spicules around the vent.

Various aberrant types of skeleton are met with in the group. In the genus Lelapia we find a partly fibrous skeleton, in which the fibres are composed of bundles of triradiates shaped like tuningforks (fig. 24, o), and in Petrostoma the main skeleton is formed of calcareous spicules actually fused together. In Astrosclera (fig. 25) a very anomalous type of calcareous skeletdn is found, consisting of spherical masses of arragonite, each originating in a special scleroblast and having a radiate structure, recalling that of a siliceous (After W. J. Sollas.) FIG. 26. - Typical Siliceous Megascleres.

(oxeate). g, Sterraster (often regarded as a microsclere).

h, Part of section of sterraster, showing two rays united by intervening silica.

A (After J. J. Lister.) Diactinal monaxon Style.

Triact.

Primitive tetraxon (calthrops). Hexact.

Polyaxon desma.

sterraster. These bodies become closely packed together over large areas, and give the sponge a stony hardness. Hexactinellida.-In this group the skeleton is composed of spicules of colloidal silica deposited in concentric lamellae around slender axes of an organic substance which in life occupies the " axial canal " of the spicule. Although varying greatly in detail and often exhibiting great complication or, it may be, reduction in structure, these spicules are all referable to the same fundamental triaxonid and hexactinellid type, characterized by the possession of three axes intersecting each other at right angles and each thereby divided into two rays or actines (fig. 26, e). According as one, two, three, four or five of these actines are suppressed we distinguish between pentact, tetract, triact, diact and monact spicules, and these may be further subdivided according to special modifications of the rays due to secondary branching, ornamentation by spines, (After F. E. Schulze.) FIG. 27.-Derivatives of the Hexact type of Spicule, found in Hexactinellida.

a, Dagger. d, Amphidisc. f, Tetract (staurus).

b., c, Pinuli. e, Pentact. g, Diact (rhabdus).

knobs, &c., or curvature, or to excessive development of certain rays as compared with the remainder. Some of the most characteristic of these special types are represented in figs. 27 and 28. Two of them require special notice on account of their importance in the classification of the group. These are the hexaster and the ti (After F. E. Schulze.) FIG. 28.-Derivatives of the Hexact type of Spicule, found in Hexactinellida.

a, Uncinaria; b, Clavula; c, Scopula.

amphidisc. A hexaster (=rosette) is a perfectly symmetrical hexact whose actines branch out into secondary or terminal rays, in a star-like manner (fig. 30, t). Various sub-types are distinguished according to the character of the rays (floricome, plumicome, &c.). An amphidisc (fig. 27, d) is a diact spicule consisting of two opposite rays each of which terminates in a disk-like or spherical expansion surrounded by marginal teeth.

In some cases the spicules all remain disconnected from one another (Lyssacine condition), in others some of them may be united by siliceous cement into a continuous framework (Dictyonine condition), and the distinction between these two types of arrangement was for a long time regarded as indicating a primary subdivision of the Hexactinellida into Lyssacina and Dictyonina, but this subdivision has now been abandoned. The term prostalia is applied to spicules which project freely from the surface of the sponge, and these are further distinguished as basalia, pleuralia and marginalia, according to their position at the base of the sponge, on the sides, or round the margin of the osculum. The basalia frequently form a root-tuft for attaching the sponge to the substratum (Hyalonema, Euplectella) and commonly have anchorlike distal extremities. They may be extremely long, as in the wellknown " glass-rope " of Hyalonema. In the remarkable genus Monorhaphis we find a single gigantic diact spicule, which may attain a length of two or three feet and the thickness of a lead pencil, transfixing the body of the sponge like a skewer from above downwards. A special dermal skeleton is usually formed by a number of spicules distinguished as dermalia, and a gastral skeleton may be similarly formed by special gastralia surrounding the central gastral cavity. Between the dermal and gastral skeletons another set of spicules, known as parenchymalia, form the most important part of the skeleton, supporting the chamber-layer and adjacent tissues. The distinction into large megascleres and small microscleres is perhaps less well marked in this group than in the Tetraxonida.

Tetraxonida.-Here, again, the spicules are composed of colloidal silica deposited around organic axial threads. The starting-point in the evolution of the very complex series of tetraxonid spicules is the primitive tetract or calthrops, characteristic of the most primitive members of the group (e.g. Plakina). This fundamental ground-form (fig. 26, d) consists of four rays or actines of equal length, which all meet one another at equal angles in the centre of the spicule, while their apices would occupy the four angles of a regular pyramid whose sides are four equilateral triangles. It is thus both tetraxonid (with four axes) and tetractinellid (with four rays). In Plakina the spicules are all of about the same size, neither very large nor very small, but in higher forms we usually find some of the spicules enlarged to form megascleres and others reduced to form microscleres. The megascleres play the principal part in building up the skeleton while the microscleres are usually scattered through the mesogloea.

Triaene Series of Megascleres.-When three rays (cladi) of the tetract resemble one another, while the fourth (shaft) differs in some respect the spicule is termed a traene. The simplest form is the plagiotriaene (fig. 29, 2), with three short simple cladi and an elon- s. 13, Tylostyle. 23, Sterraster. form).

14, Acanthotylostyle. 24, Spiraster.

gated shaft, the angles all remaining approximately equal. If the angles between the cladi and shaft become approximately right angles we have an orthotriaene. If the cladi point forward, we have a protriaene (fig. 29, 6). If the cladi are turned backwards towards the shaft we have an anatriaene (fig. 29, 5). If .the cladi branch each into two we have a dichotriaene (fig. 29, 3). If the cladi are expanded laterally and fused together to form a plate, while the shaft is reduced, we have a discotriaene (fig. 2 9, 4). The cladi may be reduced in size or even suppressed (fig. 29, 7, 8), leaving only the shaft, which may be either sharp at each end (oxeate) or sharp at the apex and rounded at the base (stylote). The spicule has now become monaxonid or monaxonellid (i.e. with a single axis) and monactinellid (with only a single ray); but this condition may also be arrived at in a different way, as we shall see directly.

The tetracrepid desma (fig. 29, 9), characteristic of many Lithistids, has been derived from the primitive tetract by ramification of the ends of all the rays.

Monaxonid Series of Megascleres.-We have already seen, in Plakina, how a diactinellid spicule may arise by suppression of two rays of the tetract (fig. 5). At first the two remaining axes FIG. 29.-The Tetraxon type of Spicule and its derivatives, found in Tetraxonida.

14a, 14b, Pseudaster 15, Cladotylote.

16, Acanthoxeate. 16a, Pseudaster (am phidisc).

17, Strongyle.

18, Tylote.

19, Cladostrongyle.

20, Rhabdocrepid (monocrepid) desma.

21, Aster.

22, Spheraster.

1, Primitive tetract.

2, Plagiotriaene.

3, Dchotriaene.

4, Discotriaene.

5, Anatriaene.

6, Protriaene.

7, 8, Reduced triaenes, becoming monaxon.

9, Tetracrepid desma. Io, Primitive diact. II, Oxeate.

12, Style.

25, Chiaster.

26, Oxyaster. 27,Aster with branching rays.

28, Rhaphis or trichite.

29, Trichodragma. 30, Sigmata.

31, Isochela.

32, Anisochela.

33, Diancistron.

34, Toxon.

35, Labis (forcipi are distinctly indicated by the presence of an angle in the middle of the spicule (fig. 29, to); by straightening out of this angle we reach a monaxonid but diactinellid condition-the diactinellid oxeate, with the organic centre of the spicule in the middle (fig. 29, II). By rounding off of both ends this form passes into the strongylote (fig. 29, 17), then if both ends become enlarged into knobs it is said to be tylote (fig. 29, 18). If one end only is rounded off, which apparently usually takes place by suppression of one ray, while the other remains sharp, the spicule is termed stylote (fig. 29, 12). It is now monactinellid as well as monaxonid. If the blunt end of the style enlarges to form a knob we have the tylostyle (fig. 29, 13). Acanthoxeates (fig. 29, 16), acanthostyles and acanthotylostyles (fig. 29, 14) are formed by the development of spines on the surface of the spicule. The development of large recurved spines at the apex of a tylostyle gives us the cladotylote or grapnel spicule (fig. 29, 15), which simulates an anatriaene. By enlargement of the spiny base of an acanthotylostyle and suppression of the shaft we get forms which simulate astrose microscleres and may be called pseudasters (fig. 29, 14a, 14b). Pseudasters may also be developed by shortening up of acanthoxeates, accompanied by enlargement of the spines (e.g. Spongillinae, fig. 29, 16a). The exotyle appears to have been formed by enlargement of the outer end of a radially placed oxeate at the surface of the sponge. By ramification of both ends of a diactinal megasclere we get the monocrepid desma (fig. 29, 20), characteristic of certain Lithistids and closely simulating the tetracrepid desma. By ramification of one end of a strongylote spicule we may get a cladostrongyle (fig. 29, 19).

Daactinal Series of Microscleres.-The starting-point of this series is the primitive angulate, diactinal oxeate (fig. 29, to). This has given rise to long hair-like forms or rhaphides (fig. 29, 28), short hair-like forms associated in bundles and called trichodragmata (fig. 29, 29), bow-shaped forms or toxa (fig. 2 9, 34), and Cand S-shaped forms or sigmata (fig. 29, 30). From the sigmata may be (After Sollas.) FIG. 30.-Typical Microscleres. a, b, Sigmata (sigmaspires).

c, Toxon.

d, Spiraster.

e, Sanidaster.

f, Amphiaster.

g, Sigma.

h, k, Isochelae. j, End of a chela, showing the teeth.

derived the diancistra (fig. 2 9, 33), shaped like pocket-knives with a blade half open at each end, and the wonderful series of chelae (fig. 29, 31, 32), in which each end branches into a number of sharply recurved teeth. These chelae are characteristic of the family Desmacidonidae, and exhibit great variations in detail, while each particular form is remarkably constant in the species in which it occurs. The most curious and aberrant are those of Melonanchora (fig. 30, 1) and Guitarra. In isochelae the two ends of the spicule are equal, in anisochelae they are unequal.

Astrose or Polyactinal Series of Microscleres.-For the beginning of this series we must go back to the primitive tetract. Reduction in size, sometimes accompanied by increase in the number of rays, has given rise to the oxyaster (fig. 29, 26), with sharp rays and no conspicuous centrum. The development of a distinct centrum from which numerous rays come off gives us the spheraster (fig. 29, 22). In the sterraster (fig. 26, g, h), characteristic of the family Geodiidae, numerous slender rays become fused together side by side to form a solid ball. In the spiraster (fig. 29, 24) the centrum appears to have become elongated and twisted into a spiral. The rays of the aster may terminate in knobs as in the chiaster (fig. 29, 25), or they may become branched (fig. 29, 27).

Arrangement of the Skeleton in the Tetraxonida.-The most primitive type of skeleton arrangement in this group was probably very similar to that which we still find in Plakina or Dercitopsis, but without any special dermal spicules, the skeleton consisting exclusively of small isolated tetracts irregularly scattered through the mesogloea between the chambers. We may call this the scattered or diffuse type of skeleton. With the development of an ectosome -whether thin dermal membrane or thick cortex-a special dermal skeleton arose. Sometimes this consists of small specially differentiated dermal spicules-candelabra in Plakina, oxeates in Dercitopsis-but a much more important series of modifications was (After W. S. Sollas.) FIG, 31.-Section of a young Stellettid Sponge, showing radial arrangement of skeleton.

initiated by the development of the triaenes. The cladi of these spicules are commonly extended in or beneath the ectosome and form a very efficient dermal skeleton, while the shafts are directed centripetally through the choanosome. In the genus Discodermia the discotriaenes form a continuous dermal armour of siliceous plates. When anatriaenes and protriaenes are developed their cladi commonly project beyond the surface of the sponge and render it more or less strongly hispid, thus forming a protection from the attacks of enemies. The shafts of the triaenes, though greatly reduced in Discodermia, usually become very much hypertrophied and may be grouped together in bundles, often associated with oxeate spicules. These spicules, or bundles of spicules, now form the principal part of the skeleton, and inasmuch as they radiate from the interior towards the surface of the sponge we distinguish this as the radiate type of skeleton. The skeleton of the vast majority of Tetraxonida is either actually radiate in structure or derived from the radiate type by further modification. In many Stellettidae, for example (fig. 31), we have a typical radiate skeleton in which a large number of the spicules retain the primitive tetractinellid form, though associated with oxeates, while in Tethya the skeleton is arranged in a similar manner but only monaxonid spicules are present. From the radiate we pass to the reticulate type of (After Minchin and Dendy. A, 13, C from Lankester's Treatise on Zoology, D from Trans. of Zool. Soc. of London, vol. xii.) FIG. 32.-Evolution of the Pseudoceratose Reticulate type of Skeleton, as seen in A, Reniera; B, Pachychalina; C, Chalina; D, Spinosella plicifera. sp, Spicules; spg., Spongin; m.f., Primary fibres; c.f., Secondary (connecting) fibres.

A 1, Modified isochela of Melonanchora. m, Spheraster.

n, o, p, Oxyasters.

q, r, Reduced asters.

s, Microxeate.

t, Hexaster (rosette).

skeleton which characterizes the majority of the so-called Monaxonellida. This is derived from the former by the establishment of secondary spicule-bundles connecting the primary or radial bundles together, and the transition is usually accompanied by loss of the cladi of the triaenes and by the development of a massive irregular form on the part of the entire sponge. An intermediate condition is found in some of the massive species of Tetilla (e.g. T. limicola), in which the spicule-bundles are very well defined and form distinct primary " fibres " in the i....erior of the sponge, but no distinct secondary or connecting fibres are yet developed.

(After Lendenfeld. Modified from Lendenfeld's Horny Sponges, by permission of the Royal Society of London.) FIG. 33. - Dendritic, Euceratose Skeleton of Dendrilla rosea. In the Sigmatomonaxonellida, derived from the .Tetillidae, the reticulate type of skeleton is almost universal, and in this group an entirely new element is introduced into the skeleton with the development of a " horny " cementing material (spongin) which unites the spicules together in the fibres. At first small in quantity (Reniera, fig. 32, A), the spongin cement gradually increases in proportion to the spicules until in many Chalininae (fig. 32, B, C) and Desmacidonidae the spicules become completely embedded in it, and the fibres may be formed chiefly of spongin, with only a core of spicules. The complete enclosure of the spicules by spongin at a very early stage cuts off their food supply and causes arrest of development. Finally, in some Chalininae (fig. 32, D) and Desmacidonidae the spicules entirely disappear from the interior of the fibre, and if at the same time they happen to be absent from the intervening mesogloea we get a skeleton composed exclusively of horny matter or spongin, to which the term pseudoceratose may be applied. In the sub-family Ectyoninae the skeleton becomes modified in an interesting manner by the development of "echinating " spicules, usually acanthostyles or acanthotylostyles, whose bases are cemented on to the fibre by spongin while their apices project into the surrounding soft tissues. These doubtless serve as a defence against internal parasites. In Agelas these echinating spicules may persist after the spicules have entirely disappeared from the interior of the strongly developed horny fibre. In the Axinellidae all the spicules in the fibres are typically more or less echinating in character and the fibres become plumelike.

Very frequently a special dermal skeleton is developed in the ectosome altogether distinct from that formed by the cladi of the triaenes (when these are present). Thus in the Geodiidae (fig. 23) the thick cortex is almost filled with densely packed sterrasters. In many forms there is a dense layer of small radially arranged monaxons at the surface of the sponge, whose projecting apices form an efficient protection. In the reticulate forms the ectosome is usually a thin dermal membrane supported by a reticulate dermal skeleton of slightly different structure from the " main " skeleton. In cases where a special stalk or a root-tuft is developed we also find a special and appropriate skeleton in connexion therewith.

In the so-called Lithistida alone amongst the Tetraxonida do we find the spicules (desmas) united together by silica to form a coherent skeleton, sometimes of stony hardness, very different from the elastic, flexible skeleton resulting from the development of spongin, and analogous to the condition met with in the Dictyonine Hexactinellids.

The microscleres usually play quite a subordinate part in the formation of the skeleton, being scattered irregularly throughout the mesogloea, though sometimes (Geodia, Tethya) the asters may form a definite cortical layer.

In the true horny sponges, if we neglect for the moment the presence of foreign bodies, we may say that the skeleton consists from the first exclusively of spongin, secreted (by special spongoblasts) in concentric layers to form very well defined fibres. In the most primitive forms (Aplysillidae) this horny skeleton is dendritic in arrangement (fig. 33), composed of fibres which rise vertically upwards from the base of the sponge (where they may be expanded to form a horny basal cuticle which serves for attachment) and ramify towards the surface, where their apices push against the dermal membrane and cause it to project in the form of " conuli." No reticulation is formed in the simplest cases (Aplysilla, Dendrilla), but in Megalopastas secondary connecting fibres are established (in relation, doubtless, to the increase in size and massive form of the sponge), and the skeleton thus simulates the pseudoceratose reticulate type of the Sigmatomonaxonellida. In Darwinella we have, in addition to the dendritic skeleton, isolated " spicules " of spongin scattered irregularly through the mesogloea. The presence of these spicules, which are sometimes, though by no means always, hexactinellid in form, has given rise to much speculation as to the possible relationship of the Aplysillidae to the siliceous Hexactinellida. Until we know more about their origin, however, we may perhaps best regard them simply as detached portions of the general skeleton secreted by isolated groups of spongoblasts. The genus Megalopastas forms a natural transition to the Spongeliidae, in which the reticulation of the horny skeleton is an almost constant feature, and in which the tendency to supplement or replace the spongin by foreign bodies (sand, broken spicules) is very strongly marked. In extreme cases the skeleton is composed almost exclusively of sand (e.g. Psammopemma), and the whole sponge looks like a mass of sand stuck together by a minimum of soft tissues and spongin cement. Such " arenaceous " sponges also occur in other groups (e.g. Desmacidonidae). The culminating point in the development of the true horny skeleton is found in the Spongiidae (e.g. Euspongia), but even in the bath sponge (fig. 6) we commonly find sand grains or other foreign matter in the interior of the primary fibres. The value of the sponge for domestic purposes depends upon the softness and elasticity of the fibre, the closeness of the meshes, and the relative absence of sand.

Histology. There are two primary tissue-forms in sponges, the flat pavement epithelium and the epithelium composed of choanocytes or collared cells. The former covers the whole of the external surface of the sponge and, except in the simpler Calcarea Homocoela, it also lines a considerable portion of the canal-system. The latter lines practically the whole of the primitive gastral cavity in the Calcarea Homocoela, but in all higher types becomes restricted to welldefined " flagellated chambers." A gelatinous " mesogloea," which must be regarded primarily as an intercellular substance, appears between the primitive outer and inner layers of the sponge-wall. This contains primitive amoeboid wandering cells (archaeocytes), (After Dendy. From Quart. Journ. Micro. Science, new series, vol. xxxv., by permission of J. and A. Churchill.) FIG. 34. Histology.

I, Pavement epithelium from the upper surface of an oscular diaphragm of Vosmaeropsis wilsoni. 2, Chamber diaphragm of Vermaeropsis macera; mus.c, Myocytes; ex.op, Exhalant aperture of flagellated chamber.

3, 4, 5, Amoebocytes of Leucandra phillipensis (the one shown in 5 appears to be feeding by means of pseudopodia upon the collared cells (c.c.) of a flagellated chamber).

6, Section across an inhalant canal (i.e.) of Ute syconoides, showing an ovum (ov.), suspended from the wall, apparently awaiting fertilization; sp, spicules.

which give rise to the ova and spermatozoa, and also various other cells which are now generally believed to migrate into it from the primitive pavement epithelium (dermal epithelium) of the outer surface, such as scleroblasts, various connective tissue elements and contractile fibres.

Pavement Epithelium (fig. 34,I). - This always consists of a single layer of polygonal cells, which are usually flat and very rarely (Oscarella) provided with cilia or flagella. They may be glandular and may secrete a definite cuticle (as in many Euceratosa). They may also be highly contractile.

In certain Calcareous sponges (Leucosolenia) it has been shown (by E. A. Minchin) that the primitive inhalant pores (prosopyles) are formed as perforations in certain of the pavement epithelium cells, which acquire a tubular form and extend through the mesogloea from the dermal to the gastral surface. The outer portion of each porocyte forms a contractile diaphragm which doubtless regulates the admission of water to the gastral cavity. The porocytes are sometimes conspicuous on account of their highly granular character.

We may distinguish three kinds of scleroblasts, according to the chemical character of the skeletal material which they secrete; these are calcoblasts, silicoblasts and spongoblasts. The calcoblasts and silicoblasts (fig. 35, h - n) form their respective spicules, at any rate in the first instance, as intra-cellular (perhaps sometimes intra-syncytial) secretions, though we must suppose (After Schulze and Soll :;.) FIG. 35. - Histology.

a, Collencytes from Thenea muricata. b, Chondrenchyme (with spicules) from Corticium' candelabrum. c, Cystenchyme, from Pachymatisma johnstoni. d, Desmacyte, from Dragmastra normani. e, Myocytes and collencytes, from Cinachyra barbata. f, Thesocyte, from Thenea muricata. g, Collared cell (choanocyte), from Sycon raphanus. h - n, Silicoblasts or mother-cells, in which different forms of siliceous spicules are being secreted.

that in the case of large spicules the later stages in growth are accomplished by the activity of several or many scleroblasts in co-operation. The spongoblasts (fig. 7) appear to co-operate with one another in the formation of the spongin fibre from the beginning. They are found only around the young, growing fibres, where they occur in large numbers, forming a kind of sheath of somewhat flask-shaped cells, each placed at right angles to the surface of the fibre and with the nucleus in its broad distal end. The spongin is secreted in concentric lamellae and is obviously intercellular in origin, and probably of the same nature as the cuticle which often occurs on the surface of the sponge.

The following are the chief forms assumed by the mesogloea according to the nature of its connectivetissue cells and intercellular substance. (a) Collenchyme, consisting of a clear gelatinous matrix with branching stellate collencytes (fig. 35, a) embedded in it; (b) Sarcenchyme, in which the quantity of intercellular matrix is greatly reduced and the connective-tissue cells are closely packed together; (c) Cystenchyme (fig. 7, Coll., fig. 35, c), consisting of close-packed, oval, vesicular cells with fluid contents and strands of protoplasm radiating from the nucleus to the periphery; (d) Chondrenchyme (fig. 35, b), somewhat resembling cartilage in texture and with a very large amount of intercellular matrix.

The name desmacytes has been given to certain slender connective tissue fibres (fig. 35, d) often united in dense bundles or layers, which occur especially in the ectosome of many Tetraxonida, giving rise to a fibrous cortex of leathery consistence.

Contractile Fibres. - Muscular fibres or myocytes (fig. 35, e) are of common occurrence, especially in relation to various parts of the canal-system, the diameter of which appears to be regulated by their agency. They may form definite sphincters around the vents or in other places (fig. 34, 2), or they may form transverse bands. lying in the floor of pore-bearing grooves, by the contraction of which the lips of the groove are doubtless approximated and the incurrent stream of water shut off (Esperella murrayi, Xenospongia patelliformis). Endothelial Cells. - In many sponges the developing embryos are enclosed in definite capsules composed of flattened polygonal cells, the whole being embedded in the mesogloea. The origin of the endothelial cells forming the capsules is doubtful. They sometimes aid in the nutrition of the developing embryo (e.g. in Stelospongus flabelliformis). No nervous elements, nor sensory cells of any kind, have as yet been recognized with any degree of certainty in sponges, in spite of various heroic attempts to demonstrate their existence.

Collared Cells or Choanocytes (fig. 35, g). - These are quite the most characteristic histological elements met with in sponges. Although exhibiting various minor differences in structure, and still more as, regards size, they always show the same essential features. Each consists usually of an oval or rounded body (frequently appearing polygonal from the pressure of its fellows) surmounted by a more or less cylindrical or funnel-shaped collar, which surrounds a single long, whip-like flagellum projecting from the apex of the cell. The collar is a filmy, transparent extension of the cytoplasm (cell - protoplasm), which can be completely withdrawn. The flagellum may also be withdrawn, and in preserved specimens neither collar nor flagellum is usually visible. The cell is usually broadest at the base and narrowed to form a neck or " collum," beneath the collar. The nucleus may be situated either at the base or at K. the apex of the cell-body or between the two. The collar itself is often a more complicated structure than (After F. E. Schulze.) appears at first sight. It FIG. 36. - Collared Cells of Schaudinnia may be provided with one arctica. or two transverse hoops, n, Nucleus; , fl, Flagellum; c, Collar. presumably serving to stiffen it (Ascandra falcata). In many cases the collars of adjacent choanocytes have been observed to be connected by a definite membrane which stretches from one to the other at the level of their. margins. This is known as Sollas's membrane, but it is apparently not a permanent structure, and the circumstances under which it appears require elucidation. In the Hexactinellida the form of the collared cells appears to be somewhat unusual (fig. 36).

The term " archaeocytes " has been applied to certain undifferentiated amtieboid cells which make their appearance: at an extremely early stage in the ontogeny, and some of which persist throughout life, with little, if any, modification, as the amoebocytes of the adult sponge, while others become germ-cells, differentiated into ova and spermatozoa.

} .0 Amoebocytes. - These are amoeboid cells closely resembling the leucocytes or white blood corpuscles of higher animals. They commonly have blunt, lobose pseudopodia and the cytoplasm is generally more or less densely charged with refractive granules. They have the power of wandering from place to place through the mesogloea (fig. 34, 3-5).

(After Polejaeff and Schulze.) FIG. 37. - Spermatozoa.

a - h, Development of Spermatozoa in Sycon raphanus; h, Mature 'Spermatozoa; j, Sperm-ball in Mesogloea of Oscarella lobularis; k, Mature Spermatozoon.

The ova (fig. 34, 6) are formed from amoebo-, cytes, which grow to a large size and finally withdraw their pseudopodia and acquire a rounded form. They have large nuclei with a very distinct nuclear membrane and commonly a conspicuous nucleolus. The spermatozoa (fig. 37) closely resemble those of higher animals, consisting each of a small " head," composed chiefly of chromatin material, and a slender vibratile " tail " composed of cytoplasm. In this case the amoebocyte gives rise to a single sperm mother-cell (spermato- .cyte) sometimes enclosed in one or two covering cells. The nucleus of the spermatocyte undergoes repeated mitosis and a " sperm-ball " is produced which is either enclosed in the .covering cell or in a special endothelium similar to that which surrounds the segmenting ovum. The germ-cells occur scattered through the mesogloea and are not aggregated in gonads, so that we cannot speak of " ovaries " and " testes " as in higher types.

Reproduction. Reproduction in sponges may be effected in one of three ways: (I) The first is by vegetative budding, followed by separation of the buds and thus differing from the ordinary budding which leads merely to increase in the size of the sponge-colony. This process has been observed in many cases (e.g. Leucosolenia, Oscarella, Lophocalyx, Aplysilla). (2) The second way is by the formation of specialized reproductive bodies known as gemmules. This process is best known in the fresh-water sponges (Spongillinae), where it has been developed as a special means of tiding over unfavourable periods during which the parent sponge is liable to be destroyed by cold or drought. Each gemmule consists of an aggregation of amoeboid cells (statocytes) densely charged with nutrient granules and enclosed in a protective horny envelope which may be strengthened by a layer of special spicules. The ripe gemmule is very resistant to adverse conditions and is capable of remaining dormant for a lengthened period, and of developing into a new sponge on the return of favourable conditions. In temperate climates the gemmules remain dormant throughout the winter and develop in the spring, the development being very similar to that of an ordinary fertilized ovum except that it begins at the " morula " stage, with the numerous statocytes representing the blastomeres. (3) The third way is by the union of ova and spermatozoa to form zygotes, which undergo segmentation and develop into the adult through a more or less complex series of ontogenetic stages. Previous to fertilization the ovum undergoes a process of maturation accompanied by the extrusion of two polar bodies, as in higher animals. Very little is known about the actual process of fertilization, but it appears probable that this is effected in the inhalant canals of the parent sponge, where the ova have been observed suspended from the epithelial lining of the canal (e.g. in Ute, fig. 34, 6). After fertilization they appear, usually at any rate, to migrate back into the mesogloea, where they become surrounded by endothelial capsules and undergo segmentation. In Stelospongus flabelliformis the cells of the capsule are of gigantic size and are attached to the superficial blastomeres of the developing embryo by protoplasmic processes, through which, no doubt, nutriment is passed from the parent to the embryo.

Embryology. The segmentation of the ovum appears to be in all cases complete or holoblastic, and the young sponge usually leaves the parent in the form of a free-swimming ciliated larva, which, after fixing itself to some object, undergoes a metamorphosis and then grows into the adult form. The details of development appear to differ widely in different species and various interpretations have been placed upon somewhat limited and discrepant observations.

One of the best-known cases is that of the calcareous genus Sycon (fig. 38). The fertilized ova develop into ciliated larvae within the parent sponge, embedded in the walls of the radial chambers, in their endothelial capsules. Each divides first into two, then into four, and then into eight equal and similar blastomeres by successive vertical clefts. The eight-celled stage (fig. 38, b, c,) has the form of a somewhat flattened cushion, with an axial cavity which is the beginning of the blastocoel or segmentation cavity. A horizontal cleft now divides each blastomere into a somewhat smaller upper and a somewhat larger lower portion, and the sixteen blastomeres arrange themselves in the form of a hollow sphere surrounding the blastocoel. The smaller cells multiply rapidly and become columnar, while still remaining as a single layer. Each one presently acquires a flagellum (" cilium ") at its outer end. The larger cells multiply more slowly and are characterized by their coarsely granular appearance. They are destined to give rise to the dermal layer and its derivatives (including archaeocytes ?) and never become flagellated.' The blastosphere or blastula (fig. 38, d, e,) is now complete, the blastocoel being completely surrounded by a single layer of cells differentiated, however, into two groups, gastral and dermal. The large granula (dermal) cells now become invaginated, but this Invagination of flagellated cells.

Gastrula attached by oral face. Young sponge (Olynthus stage).

Top view of young sponge.

is only a temporary condition, probably to be explained as the mechanical result of the pressure of the spicules of the parent sponge. The so-called " pseudogastrula " thus formed escapes by rupture ' According to E. A. Minchin, the first-formed granular cells are " archaeocytes," which migrate into the interior of the larva while their place is taken by granular cells formed by modification of the neighbouring flagellated cells. The later-formed granular cells are destined to give rise to the dermal layer of the adult, while the remaining flagellated cells form the gastral layer.

(After F. E. Schulze.) FIG. 38. - Development of Sycon raphanus. Ovum.

c, Embryo with 8 blastomeres (b, top view, c, side view). Blastosphere (blastula).

Larva at time of escape from parent.

a, b, :1, of the parent tissues into a radial flagellated chamber and passes to the exterior with the outgoing stream of water. The invaginated dermal cells are pushed out again and the "amphiblastula " swims away (fig. 38, e). (Possibly the granular dermal cells, by proliferation, may form a solid mass blocking up the blastocoel completely, so that we have a solid embryo.) The larva now fixes itself by the anterior flagellated pole (which, according to Schulze, becomes permanently invaginated, thus giving rise to a true gastrula, fig. 38, f, g), and the dermal cells spread themselves out over the gastral cells, which they completely cover. The fixed larva (" pupa ") consists of a solid mass of gastral cells enclosed in a single layer of now flattened dermal cells. Presently the gastral cavity appears (or reappears) in the middle, around which the gastral cells arrange themselves in a single layer. The young sponge elongates upwards, some of the dermal cells form porocytes which become perforated by prosopyles, others migrate into the gelatinous mesogloea and form scleroblasts, from which spicules are developed. The cells of the gastral layer acquire collars in addition to their flagella, an osculum is formed by perforation at the apex, and the young sponge begins to feed. It is now in the Olynthus condition (fig. 38, h) and is exactly comparable to a simple Leucosolenia individual. As it grows older radial flagellated chambers are budded out around the central gastral cavity and the collared cells lining the latter are replaced by pavement-epithelium derived from the dermal layer.

An interesting account of the development of Leucosolenia (Clathrina) blanca has been given by E. A. Minchin. Segmentation is regular and complete, resulting in the formation of a hollow, ciliated, oval blastula (fig. 39, A), with a large blastocoel and a wall composed of a single layer of columnar flagellated cells and a pair of very large granular cells at the posterior pole. The latter are primitive archaeocytes and are destined to give rise to the amoebocytes and germ-cells of the adult. The flagellated cells will give rise to all the other cells of the adult, both dermal and gastral. The larva becomes free-swimming in this condition. Here and there individual flagellated cells (destined to form the cells of the dermal layer) lose their flagella and, becoming amoeboid, migrate into the blastocoel, which presently becomes completely filled with such cells. The larva is thus converted into a solid " parenchymula," in which the archaeocytes remain unchanged in their original position at the posterior extremity. It now fixes itself and flattens out upon the substratum in the pupal condition. During the metamorphosis which now ensues the majority of the cells of the inner mass (dermal cells) pass out to the exterior again between the flagellated cells (gastral cells), over which they spread themselves in the form of a dermal layer of flattened epithelium. Some of the dermal cells, however, remain in the inner mass as porocytes; the primitive archaeocytes have divided up into amoebocytes; and porocytes, amoebocytes and the cells of the gastral layer are all crowded together in the interior of the pupa. The pupa now elongates vertically. A gastral cavity appears in the interior. The cells of the gastral layer arrange themselves around this cavity and develop their collars and flagella. At first, however, the gastral cavity is lined by the porocytes, which presently separate and migrate outwards.' Scleroblasts migrate inwards from the dermal layer and secrete spicules. An osculum and prosopyles are formed as in Sycon and the Olynthus stage is reached.

The development of sponges in general appears to be characterized by a remarkable want of uniformity in the arrangement of the different kinds of cells of which the larva is composed. Two, or possibly three, primary groups of cells are universally present; the flagellated cells, which will give rise to the collared cells of the adult, the non-flagellated (granular) cells, which will give rise to the dermal layer and its derivatives, and possibly the primitive archaeocytes (perhaps to be regarded as undifferentiated blastomeres). It may be considered as doubtful, however, whether the primitive a.rchaeocytes can in all cases be distinguished from the primitive dermal cells. The latter are in some cases (amphiblastula type) grouped at the posterior pole of the larva (Sycon), while in other cases (parenchymula type) they may pass inwards and completely fill the interior, blocking up the blastocoel and perhaps also freely projecting at the hinder end (fig. 39, F). At the time of the metamorphosis the dermal cells pass to the outside and come to completely enclose the gastral cells, so that the two layers acquire their proper relative positions. The sponge larva in many respects closely resembles the Coelenterate " planula," with its ectoderm and endoderm, but it is very doubtful how far this comparison is valid, and in the present state of our knowledge it is perhaps better to avoid the use of the terms ectoderm and endoderm in dealing with the sponges altogether. The idea naturally suggests itself that the two primary layers of the Sponge correspond to those of the Coelenterate, but in a reversed position, the inner layer of the one being the outer layer of the other, and vice versa, and this idea has found expression in the name Enantiozoa which has been proposed for the group by Yves Delage, but which has not met with general acceptance.

Physiology. Comparatively little is known of the physiology of sponges. The most obvious expression of the vital activity of the organism is the stream of water which flows in through the dermal pores or ostia and out through the vents or oscula. That this stream is maintained by the undulatory movements of the flagella of the collared cells there can be no doubt, but the fact that the movements of the flagella of different cells are not co-ordinated, so that they do not act in unison, indicates that the mechanical problem involved is not so simple as is usually supposed. There can be no doubt that the incoming stream brings with it minute food-particles, consisting of fragments of organic matter, alive or dead, and also the oxygen required for purposes of respiration; while the outgoing stream removes faecal products and waste matter (excreta). The rate of flow appears to be regulated by the opening and closing C ,? of the pores and vents, or of intermediate apertures such as the apopyles or exhalent openings of the flagellate chambers. This opening and closing may be effected by the activity of definite muscular sphincters (fig. 34, 2) or, in the case of some prosopyles, by the contractility of the porocytes themselves.

The ingestion of the food particles is no doubt effected in large measure by the collared cells, which seem to feed much in the same manner as independent collared monads (Choanoflagellata). It seems not improbable that Sollas's membrane may be a temporary structure which assists in arresting food particles as they pass through the flagellate chambers. There is reason to believe also that amoebocytes (in this case therefore phagocytes) may capture minute organisms on their way through the canal system, and even porocytes are sometimes credited with this power.. Digestion, no doubt, is, at any rate chiefly, intracellular. The amoebocytes probably serve not only to ingest food themselves but also to receive surplus food from the collared cells and distribute it through the sponge (fig. 34, 5).

Nothing definite is known as to the function of excretion,. but here, as in the case of nutrition, it seems likely that collared cells and amoebocytes are both concerned.

1 The position of the porocytes inside the collared cells appears at first sight very anomalous, but Minchin has. shown that this condition is actually repeated in the adult sponge every time the gastral cavity is obliterated by contraction.

(After E. A. Minchin.) FIG. 39. - Types of Sponge Larvae (semi-diagrammatic). The ciliated (gastral) cells are left blank; the dermal cells are shaded, and the archaeocytes are granulated.

A, Larva of Leucosolenia (Clathrina) blanca. B, Of Leucosolenia (Clathrina) reticulum. C, Young larva of Leucosolenia (or pseudogastrula stage of Sycon). D, Late larva of Leucosolenia (or newly hatched larva of Sycon). E, Larva of Oscarella. F, Parenchymula larva of a siliceous Monaxonellid (Myxilla). Sponges, as we have already seen possess no special nervous system and no special sense organs, and the power of response to stimuli appears to be very limited. Many sponges probably have the power of contracting as a whole, which may in some cases be due, in part at any rate, to the presence of bands of muscular fibres, and Sollas observes that in Pachymatisma irritation of the oscular margin is invariably followed after a short interval by a slow closure of the sphincter. The power of movement in adult sponges is, however, chiefly confined to individual cells acting independently. The young larvae, on the other hand, swim vigorously about by means of their cilia or flagella, whose movements must obviously be co-ordinated in order to ensure the progress of the entire organism in definite directions.

The rate of growth of sponges appears to be very rapid. A British species of Hymeniacidon is said to form a crust measuring a foot in diameter in so short a period as five months. With this rapidity of growth must be associated the fact that many sponges, marine as well as fresh-water, appear to be annual.

Distribution. The vast majority of sponges are marine, only a single sub-family, the Spongillinae, having acquired the habit of living in fresh water. The Spongillinae are, however, very widely distributed, being found in lakes and rivers in all parts of the world. Marine sponges occur everywhere, from low-water mark to the greatest depths, but certain localities, such as the Gulf of Manaar, Port Phillip and Port Jackson, appear to be much richer than others both in individuals and species. The Hexactinellida are essentially a deep-water group and are therefore much more rarely met with than other forms. The Tetraxonida and Euceratosa abound in shallow and in moderately deep water, and a comparatively small number of species of Tetraxonida occur at great depths. Both are dominant groups at the present day, represented by very large numbers of species and individuals. The Myxospongida are comparatively rare and represented by very few species. The Calcarea are common in the littoral region, especially in sheltered situations amongst rocks and seaweed.

Most families and even genera of sponges enjoy a very wide geographical range, very many being cosmopolitan. Species are usually much more restricted in distribution, but even here there are some noteworthy exceptions, and future researches will probably show that many species from different localities which are at present regarded as distinct are connected by intermediate forms living in intermediate situations.

There appears to be a well-marked relation between temperature and the power of spongin-secretion, and as a result we find that sponges with a really well-developed horny skeleton (whether Euceratosa or Pseudoceratosa) are usually only met with in comparatively warm waters. This fact brings about a striking contrast between the sponge-faunas of different latitudes.

Classification. The classification of the Phylum Porifera, the characters of which have already been given, is as follows Sub-phylum and Class Calcarea. - Sponges with a skeleton composed of carbonate of lime, commonly in the form of isolated spicules whose most usual shape is triradiate.

Calcarea in which the gastral cavity and its outgrowths are lined throughout by collared cells. This order is sometimes divided into two families, Clathrinidae and Leucosoleniidae, but it is doubtful if this distinction can be maintained, and by some writers only a single genus (Leucosolenia) is recognized.

Calcarea in which the original lining of the gastral cavity is partly replaced by pavement epithelium, so that the collared cells are confined to separate flagellated chambers. This order includes the living families Leucascidae, Sycettidae, Grantidae, Heteropidae, Amphoriscidae and Pharetronidae (with only two living representatives but numerous fossil forms). The relationships of the anomalous Astrosclera (fig. 25), for which the family Astroscleridae has been proposed by J. J. Lister, must still be regarded as problematical.

Sub-phylum Non-Calcarea. - Sponges without any calcareous skeleton.

Class and Order MYxosPONGIDA. - Sponges with no skeleton; with simple canal system and usually large flagellate chambers. (The absence of skeleton is primitive and not due to degeneration.) This class is sometimes divided into two families - Halisarcidae, with elongated, sac-shaped chambers, and Oscarellidae, with more or less spherical chambers.

Class Triaxonida (= Hexactinellida). - Sponges With a skeleton composed of siliceous spicules, either isolated or cemented together by silica, and either triaxonid and hexactinellid in form or derivable from the triaxonid and hexactinellid type. The canal system is simple and the flagellated chambers are large and sacshaped, and more or less radially arranged in a network of trabecular tissue. Spongin is never formed.

Triaxonida with characteristic amphidisc spicules, but no hexasters, and with a root-tuft of anchoring spicules. The family Hyalonematidae, including the well-known glass-rope sponges of the genus Hyalonema, is the only family recognized in this order.

Triaxonida whose most characteristic spicules are hexasters. To this order belong the living families Euplectellidae, Asconematidae, Rossellidae, Euretidae, Melittionidae, Coscinoporidae, Tretodictyidae and Maeandrospongidae, and a number of extinct families such as the Ventriculitidae so commonly met with in the Jurassic and Cretaceous rocks.

Class TETRAxoNIDA. - Sponges with a skeleton composed of siliceous spicules, either isolated or cemented together (by silica or by spongin), and either tetraxonid and tetractinellid in form or derivable from the tetraxonid and tetractinellid type. The canal system is usually complex, with small, more or less spherical flagellated chambers.

Grade TETRAcTINELLIDA. - Tetraxonida in which some, at any rate, of the megascleres retain the primitive tetractinellid form. No desmas are developed.

Order 1. Homosclerophora. - Tetractinellida in which microscleres and megascleres are not yet sharply differentiated from one another and no triaenes are developed. The canal system is comparatively simple. This order includes the family Plakinidae (see Plakina, ante) which forms the starting-point of the evolution of the class.

Order 2. A strophora. - Tetractinellida with triaenes and with astrose microscleres, without sigmata. This order includes the families Pachastrellidae, Theneidae, Stellettidae, Geodiidae.

Order 3. Sigmatophora. - Tetractinellida with triaenes, with sigmata for microscleres (when present), without asters. This order includes the families Tetillidae and Samidae.

Grade (? order) Lithistida. - Tetraxonida in which the megascleres form desmas, typically united with each other by siliceous cement to form a continuous skeleton, often of stony hardness. This group includes both tetractinellid and monaxonellid forms and may possibly be of polyphyletic origin. The Lithistida bear the same relation to the other Tetraxonida that the dictyonine Hexactinellids bear to the lyssacine forms, but in the present state of our knowledge it is hardly possible to trace the natural affinities of the numerous members of the group, many of which are only known in the fossil state. The following are the principal families: Tetraladidae, Desmanthidae, Corallistidae, Pleromidae, Neopeltidae, Scleritodermidae, Cladopeltidae, Azoricidae, Anomocladidae.

Grade MoNAXONELLIDA. - Tetraxonida in which the primitive tetraxonid and tetractinellid condition of the megascleres has been entirely lost through suppression of some of the spicule rays, so that none but monaxonellid megascleres remain. No desmas are developed. Owing to the extreme reduction or modification of the skeleton, leading in many cases to convergence, the classification of this group is extraordinarily difficult and the group is obviously not monophyletic.

Monaxonellida in which the microsclere, when present, is some form of aster. The members of this order are to be regarded as descended from aster-bearing tetractinellid ancestors.

Epipolasidae, Tethyidae, Spirastrellidae (including Placospongiidae), Clionidae (the boring sponges), Suberitidae, Chondrosiidae. (In Chondrosia the skeleton is entirely suppressed, so that it simulates the Myxospongida.) Order 2. Sigmatomonaxonellida. - Monaxonellida in which the typical microscleres are sigmata, or other diactinal forms. Normal astrose rnicroscleres are absent (though secondary pseudasters are occasionally present). The members of this order are to be regarded as descended from sigma-bearing tetractinellid ancestors.

Haploscleridae (chief sub-families: Gelliinae, Renierina Chalininae, Spongillinae), Desmacidonidae (chief sub-families: Esperellinae, Ectyoninae), Axinellidae.

Class and Order EucERATOsA. - Non-calcareous sponges without siliceous spicules, but with a skeleton composed of horny fibres developed independently, i.e. not in relation to any pre-existing spicular skeleton. The skeleton is often supplemented, or even largely replaced, by foreign bodies. This group includes the bathsponges and their very numerous relations.

Aplysillidae, Spongeliidae, Spongiidae.

There are two groups of palaeozoic fossil siliceous sponges which apparently do not fit into the above system, viz. the Octactinellida and Heteractinellida of G. J. Hinde. The former, represented by the genus Astraeospongia, have octactinal megascleres. The latter, represented by the genera Tholiasterella and Asteractinella, have polyaxon megascleres with an indefinite number of rays. These may indicate the former existence of two distinct classes of siliceous sponges A B (After G. J. Hinde.) FIG. 40. - A, octactine and B, hexactine spicules of Astraeospongia. which are, so present day.

(After G. J. Hinde.) FIG. 41.---Spicules of Heteractinellida.

A, Typical polyactine. B, Rosette-like form. C, D, E, Nail-like forms.

Phylogeny. The most recent views as to the evolution and inter-relationships of the principal groups of sponges above enumerated may be conveni ently expressed by the accompanying phylogenetic tree (fig. 42). Starting entire with the hypothetical Protolynthus as the ancestral form of the enti y am ' group, we see how two divergent o lines of descent are very early established according to whether or not a calcareous skeleton is developed. The Calcarea are at first simple Olynthus forms, Homocoela, differing only from the Protolynthus in the presence of the calcareous spicules. From these are derived, by the process of budding, on the one hand reticulate forms (Clathrina) and on the other radiate forms (e.g. Leuco- solenia tripodifera), and some of the latter (now probably extinct) form the starting point for the evolution of the Calcarea Heterocoela, ning with simple Syconoid forms and ending with complex Leuconoids, in which the original process of simple budding has been followed up Protolynthus by elaborate modifications of both FIG. 42. - Phylogenetic Tree, skeleton and canal system.

showing the supposed relationTurning to the other main line of ships of the principal groups descent we find at once a conspicuous of sponges to one another. gap between the Protolynthus and the simplest known non-calcareous sponge; though the analogy of the Calcarea makes it easy to understand how the almost Syconoid canal system of the simplest Hexactinellids, or the primitive Rhagon type of other groups, may have been derived from the Protolynthus ancestor in the first instance by simple budding. This line of descent may be regarded as continued straight on into the existing Myxospongida, with increase in the complexity of the canal system, due to folding of the chamberbearing layer and the accompanying development of inhalent and exhalent canal systems, but without the development of any skeleton. The Triaxonida and Euceratosa would seem to have branched off independently at a very early stage from the Myxosponge line, before the flagellated chambers had suffered that reduction in size which occurs in some existing Myxospongida and in all Tetraxonida. In the Triaxonid line of descent the evolution of the siliceous skeleton of primitively hexactinellid spicules is the leading feature, the canal system preserving remarkable uniformity throughout the group. In the Tetraxonida also the skeleton has played the principal part in the evolution of existing species, but the canal system too has undergone great modifications. The primitive tetraxonid, tetractinellid siliceous spicules must have arisen quite independently, their fundamental form being totally different from that of the triaxonid hexactinellid type. The appearance of differentiated microscleres in this group introduced new possibilities of variation, ' of which full advantage has been taken, and we are confronted with most interesting evolutionary series, terminating in many very remarkable and at present inexplicable spicule forms (fig. 29). In many of the more advanced Tetraxonida, especially in the Chalininae, the development of spongin cement also appears as a new factor in the process of evolution. At first serving merely to glue the megascleres together into a continuous framework, it ultimately, in some extreme cases, completely replaces the siliceous skeleton and gives, rise to a purely " horny " skeleton in which all traces of spicules. have been lost by degeneration. Thus we arrive at a " Pseudoceratose " condition (fig. 32, D) which must be carefully distinguished from the condition of the Euceratosa, which have apparently branched off quite independently from Myxosponge ancestors. Here we have another typical example of that phenomenon of " convergence " which has rendered the classification of sponges so very difficult. In the Euceratose line of descent we start with forms (Aplysilla) with large sac-shaped chambers and altogether primitive canal system, accompanied by an arborescent horny skeleton (fig. 33) of an entirely different type from that of the pseudoceratose Tetraxonida. From this we can trace the evolution gradually through the Spongeliidae to the Spongiidae, the skeleton becoming reticulate and the canal system gradually more complex with accompanying reduction in size of the chambers. The bath sponge perhaps represents the culminating point in this direction. Thus it appears that both the horny type of skeleton and the siliceous. spicular type have been twice independently produced in the evolution of the Non-calcarea. An analogous case of convergence is. seen in the union of originally separate spicules into a coherent skeleton by means of cement of the same chemical composition as themselves. This has taken place independently in the Calcarea (Petrostoma), the Dictyonine Hexactinellida and the Lithistid. Tetraxonida.

Affinities of the Porifera. Three main views have been put forward with regard to the position of the Sponges in the animal kingdom: (I) that they are colonies of Protozoa; (2) that they form a subdivision of the Coelenterata; (3) that they are not Protozoa but have originated from !Protozoon ancestors quite independently from other Metazoa (Enterozoa). The first of these views, associated especially with the names of James Clark and Saville Kent, is. supported by the relative independence of the constituent cells in the sponge-body and by the extraordinary resemblance of the collared cells to the choanoflagellate or collared Monads. It is also supported by the existence of a remarkable colonial form of Choanoflagellata (Proterospongia) in which the collared Monads are partially embedded in the surface of a gelatinous matrix, in the interior of which amoeboid cells are found. E. A. Minchin has shown that even in the adult Leucosolenia (Clathrina) the collared cells and porocytes have the power of changing their relative positions, while migration of dermal and gastral cells. and consequent inversion of the layers appears to be a common feature of the sponge larva at the time of metamorphosis. These facts are certainly suggestive of Protozoon colonies rather than of Metazoa. On the other hand it must not be forgotten that migratory amoebocytes (leucocytes) occur in probably all groups of Metazoa, while the degree of integration and the amount of histological differentiation in Sponges are far greater than in any other Protozoon colonies known to us. It has been argued that the process of sexual reproduction by means of ova and spermatozoa is fatal to the Protozoon-colony theory, but this argument is completely disposed of by the discovery of spermatozoa and ova in the unicellular Sporozoa. On the other hand the occurrence of collared cells has been held to distinguish the Sponges from all other Metazoa, and this argument has also been answered by the discovery of collared cells in the larva of Echinocyamus (an Echinoderm) by H. Theel. It would, in short,. be difficult to frame a definition of the Protozoa which should absolutely exclude the Sponges, while at the same time our conception of the nature of Protozoa will have to be profoundly modified if we are to admit the Sponges within the limits of that group.

The second view, that the Sponges constitute a subdivision of the Coelenterata, is maintained by some very eminent continental authors such as Ernst Haeckel and F. E. Schulze. This view is supported by the structure of the Olynthus type, which, as we have seen, forms the starting-point of Sponge evolution. The dermal layer of the Olynthus is regarded as ectoderm, the gastral layer as endoderm and the mesogloea with its contained cells as mesoderm, more highly developed far as we know, totally unrepresented 'at the A 'C'' _Z) than in most Coelenterates. It is also supported by a considerable amount of agreement in the early stages of development, up to the formation of the ciliated larva. According to this view the Olynthus, or at any rate the imaginary " Protolynthus " is only a slightly modified gastrula, and the Sponges are therefore Enterozoa without any coelom, or in other words Coelenterata. The extraordinary histological differences between the Sponges and other Coelenterates (Cnidaria), combined with the highly characteristic canal system and the absence of tentacles, are, however, alone sufficient to throw grave doubts upon the probability of a close relationship between the two groups, and these doubts are greatly strengthened by recent embryological researches, which tend to show that the so-called ectoderm and endoderm are not homologous in the two cases.

There remains the third view, in accordance with which the Sponges are multicellular animals which have originated quite independently from Choanoflagellate Protozoon ancestors, and this is the view which at present seems to have most in its favour. It is especially associated with the name of. W. J. Sollas, who invented the term " Parazoa " for the group. In support of this view it may be pointed out that the tendency to form hollow, spherical colonies, resembling the blastosphere stage in the development of Enterozoa, is met with in very distinct groups of Protozoa (e.g. Volvox, Sphaerozoum). This form of colony is obviously polyphyletic in origin. The fact that the segmentation of the ovum leads to such a form in both Sponges and Enterozoa is therefore by no means conclusive evidence that Sponges and Enterozoa have originated from the same Protozoon group. While, as has been repeatedly pointed out, the universal and characteristic collared cells of sponges point emphatically to a Choanoflagellate ancestry, it is impossible, in the present state of our knowledge, to indicate the particular Protozoon group which has given origin to the Enterozoa. We may then consider the Metazoa, or many-celled animals, as a polyphyletic, or at any rate diphyletic group, including two perfectly distinct lines of descent from the ancestral Protozoa, the Sponge-line on the one hand, which leads to nothing higher than Sponges, which retain in many respects the characters of Protozoa, and the Enterozoon line on the other, which leads through the Coelenterata to the Coelomata and so on to the highest divisions of the animal kingdom.

Economics. All the bath sponges belong to the two genera Euspongia, Bronn, and Hippospongia, Schulze, subdivisions of the old genus Spongia, auctorum, distinguished from one another by the fact that in Hippospongia the body of the sponge is traversed by wide ramifying canals or vestibules, in addition to the proper canal system of the sponge. Species of these two genera occur in many parts of the world, probably wherever the temperature of the sea-water is sufficiently high and the depth and bottom suitable. It is only in a few localities, however, that they occur in sufficient numbers and of sufficiently good quality to render a sponge fishery practicable. The sponges of commerce are obtained chiefly from the Mediterranean, the coast of Florida and the Bahama Islands. From the Mediterranean three distinct species are obtained - (1) Euspongia officinalis, which includes the " fine sponges," with two chief varieties, mollissima (the Levantine sponges, very soft and often cup-shaped), and adriatica; (2) Euspongia zimocca, including the " hard" or Zimocca sponges; (3) Hippospongia equina, the " common " or horse " sponge.

Of the Florida sponges five principal kinds are recognized by the dealers - (1) the sheep's wool sponge (Hippospongia gossypina) - this appears to be by far the most abundant in the market and also the most valuable; (2) the yellow sponge (Euspongia agaricina), resembling the Zimocca sponge of the Mediterranean; (3) the grass sponges (including both Hippospongia graminea and H. cerebriformis); (4) the velvet sponge Hippospongia maeandriniformis), which is not so common as the others; (5) the glove sponge (Euspongia tubulifera), which is the least valuable. In the year 1900 the Florida sponge fisheries yielded 418,125 lb of sponges, valued at $567,685. The Bahama sponges appear to be very similar to those of Florida.

Bath sponges occur in comparatively shallow water and are obtained by diving, by dredging, or by means of a trident or long-handled fork. The preparation of the sponges for the market is extremely simple. The slimy soft tissues very soon begin to decay and run off when they are removed from the water; after this has gone on for some time the sponges are washed and beaten until the skeleton is clean, they are then threaded on string and dried. They are frequently " loaded " with foreign matter by the dealers in order to increase their weight; rock-salt, glucose, molasses, lead, gravel, sand and stones being used for the purpose. They are also often bleached by means of chemicals to give them a better colour, but though their appearance is thereby greatly improved, their durability is said to be impaired.

In spite of the undoubted rapidity with which sponges grow, as shown by the fact that on the coast of Florida marketable sponges are found commonly in places that had been stripped of saleable specimens in the preceding year, there appears to be considerable danger of injury to the sponge industry by overfishing and by the reckless destruction of young specimens, and it has been found necessary to introduce special legislation in America to counteract these evil tendencies. The question of the artificial propagation and cultivation of sponges has also been much discussed, but although some very interesting experiments have been made, they have not as yet led to any great practical results. As far back as 1862 Oscar Schmidt showed that " cuttings " of sponges will attach themselves and grow. This idea was followed out in the experiments of G. Buccich on the Island of Lesina, from 1863-1872, but these experiments were brought to a close by the hostility of the native fishermen. Similar experiments have since been made on the Florida sponge-grounds. The possibility of rearing sponges in this way from cuttings has thus been fully demonstrated, but whether it can be done profitably is another question. According to the experience of G. Buccich it appeared that it would take seven years for the cuttings to attain marketable size in the Mediterranean. The Florida experiments, on the other hand, indicate a much more rapid rate of growth, and it has been stated that under favourable conditions the cuttings will attain marketable size in as short a time as one year. It has been doubted, however, whether the total weight of sponges produced by cuttings would be greater than the weight of the sponges from which the cuttings were taken if these sponges were allowed to continue their growth undisturbed. H. V. Wilson has suggested that sponges may be artificially reared from the eggs, in the same way that fishes or oysters are reared. The eggs of the bath sponge, like those of other sponges, develop into free-swimming ciliated larvae, and these might be made to attach themselves, like oyster-spat, to suitable objects, on which the young sponges could be cultivated under appropriate conditions. Detailed experiments are required to demonstrate the feasibility or otherwise of this interesting suggestion.

For further information on the economic aspect of the subject the student should consult the annual Bulletin and special papers of the United States Bureau of Fisheries and also the work of Seurat referred to in the bibliography.

A very full list of the literature of the group up to 1889 is given in Lendenfeld's work on the Horny Sponges, published by the Royal Society. We have only space here to refer to a very limited number of memoirs. Other references will be found in the works cited.

(1) J. S. Bowerbank, Monograph of the British Spongiadae (Ray Society); (2) H. J. Carter, a long series of memoirs, chiefly systematic, in the Annals and Magazine of Natural History (1847 to 1887); (3) Y. Delage, Embryogenie des eponges (Arch. Zool. Exp. (2), x. 1892); (4) A. Dendy, " Monograph of Victorian Sponges," pt. i., Trans. Royal Soc., Victoria III. (1891); (5) idem, " Studies on the Comparative Anatomy of Sponges," pts. i. - vi., Quart. Journ. Mic. Sci. (1888-1894); (6) idem, " Report on the Sponges collected by Professor Herdman at Ceylon in 1902 " (Royal Society, 1905); (7) E. Haeckel, Die Kalkschwdmme (Berlin, 1872); (8) G. J. Hinde, Monograph of British Fossil Sponges (Palaeontological Society, London); (9) A. Hyatt, " Revision of the North American Poriferae," Mem. Boston Soc. Nat. Hist. (1875-1877), vol. ii.; (io) R. Kirkpatrick, Descriptions of South African Sponges (Marine Investigations in South Africa; Cape of Good Hope Department of Agriculture, 1902-1903); (I I) W. Lundbeck, Porifera (Danish Ingolf-Expedition, vol. vi., 1902, &c.); (12) E. A. Minchin, " Materials for a Monograph of the Ascons," I., Quart. Journ. Mic. Sci. (1898), vol. xl.; (13) idem, " Sponges," in Lankester's Zoology, pt. ii. (1900); (14) N. Polejaeff, " Calcarea," " Challenger " Reports, " Zoology " (1883), vol. viii.; (15) S. O. Ridley and A. Dendy, " Monaxonida," " Challenger " Reports, " Zoology " (1888) vol. xx.; (16) F. E. Schulze, " Untersuchungen fiber den Bau and die Entwicklung der Spongien," Zeitschrift fisr wiss. Zoologie (1875-1881); (17) idem, " Hexactinellida " (" Challenger " Reports, " Zoology," vol. xxi.); (18) idem, Amerikanische Hexactinelliden (Gustav Fischer, Jena, 1899); (19) idem, Hexactinellida of the " Valdivia" Expedition (Jena, 1904); (20) L. G. Seurat, L'Eponge histoire naturelle; Peche; " Acclimatation "; Spongiculture, Bull. soc. nat. d'acclimatation de France, 48th year (1901); (21) W. J. Sollas, " Tetractinellida" (" Challenger " Reports, " Zoology," vol. xxv.); (22) I. B. J. Sollas, " Sponges," Cambridge Natural History (1906), vol. i.; (23) E. Topsent, Etudes monographiques des spongiaires de France (Arch. Zool. Exp. (3), 1894, vol. ii. &c.); (24) idem, " Contribution a l'etude des spongiaires de l'Atlantique Nord " (Campagnes scientifiques du prince de Monaco, 1892, vol. ii.); (25) idem, " Spongiaires des Acores " (Campagnes scientifiques du prince de Monaco, 1904, vol. xxv.); (26) G. C. J. Vosmaer, " Spongien," Bronn's Klassen and Ordnungen des Thierreichs (1887), vol. ii.; (27) H. V. Wilson, " On the Feasibility of Raising Sponges from the Egg," Bulletin of the United States Fish Commission (1897), vol. xvii. (A. DE.)