Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges with a Secondary Calcareous Skeleton (Tetractinomorpha, Demospongiae) J. REITNER Introduction In contrast to the nonrigid modern demosponges and Calcarea, the sponges with an additional calcareous skeleton are extremely rare and restricted to dark and cryptic habitats within tropical reefs and submarine caves ofthe Mediterranean Sea.Only 14 recent genera ofthe so-called sclerosponges or coralline sponges and pharetronids are known (nine demosponges and five calcisponges) (Kirkpatrick 1908,1910, 1911; Lister 1900; Däde.rlein 1897; Vacelet 1964, 1981 ; Hartman and Goreau 1970, 1975; Hartman 1979, etc.). Coralline sponges are quite common in the fossil record. During the Paleozoic and Cretaceous ihey were important reefbuilders or members of reefal communities. Traditionally, most of these sponges are classified as Stromatoporoidea, Chaetetida, Sphinctozoa, or Pha- retronida. These terms describe only the type of soft tissue arrangement within these particular sponge skeletons and have no taxonomic value (Reitner 1987a,b,c; Wood 1987). Hartmanand Goreau (1972) haveestablished a newclass of sponges wh ich incorporates all sponges with siliceous spicules and an addi- tional calcareous basal skeleton. However, based on the similarities ofthe spicules and soft tissue with nonrigid demosponges and Calcarea, van Soest (1984), Vacelet (1985), Reitner (1987a,b,c) and Wood (1987) have proposed a poly- phyletic origin ofthe class "Sclerospongiae". It is possible to recognize different phylogenetic lineages of fossil and modern calcified demosponges which demonstrate in some cases extremely long-ranging clades (Reitner and Engeser 1985, 1987; Wood et al. 1989). The purpose of this chapter is to demonstrate the polyphyletic nature of calcified hadromerida and to establish a phylogenetic model for this group. Material and Methods This study is based on the type material of the species Boswellia mortoni Gray (1980) (British Museum of Natural History (BMNH) no. R4429), Cassianoth- a/amia zardini Reitner (1987b) (Institut für Paläontologie Freie Universität Berlin (IPFUB) no. 87/112) and other newly collected fossil material deposited within the IPFUB. The Recent counterparts were loaned from the BMNH, Institute of Taxonomic Zoology Amsterdam (ITA), and the marine lab Endoume, University J . Reitner and H. Keupp (Eds.) Fossil and Recent Sponges © Springer-Verlag Berlin Heidelberg 1991 180 J. Reitner of Marseille. Some were collected in Bermuda and the western part of the .Mediterranean Sea near Banyuls sur Mer. From all sponges thin- or histological sections were made and studied under a light microscope.The histological sections were stained with acid fuchsin and toluidine blue to study the soft tissue arrangement. In order to separate the spicules from the soft tissue, the sponges were treated with H20 2. To expose the spicules from the calcareous basal skeletons, polished surfaces were etched with 5% Tritriplex-III solution (ClOH14N2ol8x2H20). After 1 day of treatment the sampies were c1eaned with distilled water and then dried. Subsequently the etched surfaces were examined with a scanning electron microscope (SEM). Micro- and ultrastructures were studied on freshly broken surfaces. The chemical analyses ofthe calcitic skeletons were performed with atomic absorption spectroscopy (AAS), energy dispersive X-ray analysis (EDAX), and electron microprobe analysis. Systematic Position of the Hadromerida The Hadromerida are sponges which possess monactine megasc1eres as tylostyles or subtylostyles. The microsc1eres are always asters or modified asters. The megasc1eres are generally organized in a radial pattern but, in most cases, this radial pattern is ill developed. The microsc1eres, if present, are concentrated within the cortical zone of the sponge. Within the Hadromerida, spongin is present but never as fibers. The hadromerid sponges are oviparous, wherever they have been observed. Based on spicule morphology, oviparity, and simple spongin, these particu- lar sponges are c1assified within the subc1ass Tetractinomorpha (Levi 1973; Bergquist 1978). The Hadromerida are one of the best-characterized sponge groups. Skeletal Types of Extant Hadromerid Sponges Within the Hadromerida a reduction of the spicular skeleton is observed in various families. This reducted part of the primary skeleton is substituted by strong collagenous skeletal elements. It is possible to distinguish different stages of sc1ere reduction: a) reduction of microsc1eres (Suberitidae, Polymastidae), b) red uction of megasc1eres (Chondrilla) or a partial red uction, as seen in Spirastrella, c) reduction of micro- and megasc1eres (Chondrosia) A special type of indirect skeletal formation is observed within some species of Spirastrella and the genus Cliona. Cliona excavates burrows in any calcareous material and the entire sponge, except the marginal inhalant and exhalant systems, lives inside the burrows (Rützler 1974; Keupp and Reitner this Vol.). Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 181 Calcareous basal skeletons, the main topic ofthis chapter, are observed only within the extant genus Acanthochaetetes, which possesses a spicular skeleton similar to Spirastrella (Reitner and Engeser 1987). Within the fossil record, calcareous basal skeletons are observed in several taxa. Fossil and Modern Hadromerida with a Secondary Calcareous Skeleton Remarks: The author has decided to erect new genera for new fossil taxa. The distinct types of calcareous skeletons of these particular fossils should be recognizable as morphotypes for all fieldworking paleontologists, since spicules are not seen or preserved in all cases. Using the principles of phylogenetic systematics new genera must not be erected in all cases of calcareous skeletons. Family Suberitidae Schmidt The Suberitidae are characterized by tylostyle and subtylostyle megascleres and rare microstrongyle microscleres. The typical hadromerid radial arrange- ment of the scleres is present only at the surface of the sponge. In the center part of the sponge the orientation of the spicules is confused. The spicules wh ich cover the exhalant canals often show a loose plumose orientation (Fig.3a,b). Chaetetes (Boswellia) mortoni Gray (Figs. 1,2) 1980 Chaetetes (Boswellia) mortoni Gray, p. 808, PI. 103 This species from the Lower Carboniferous (Lower Asbian) from Northern Wales is the oldest known coralline sponge with hadromerid character. Primary skeleton: Chaetetes mortoni possesses tylostyle megascleres (length 170-210 /-Lm) which are arranged in bundles as observed in Suberites (Fig. 3a,b). These bundles are relicts of a radial pattern, characteristic of the Hadromerida. Microscleres are not observed. The spicules are preserved as diagenetic silica or calcite (Fig. 2c). Secondary skeleton: The spicules are covered by a clinogonal ("water jet" structure, or fasicular fibrous) calcite (probably Mg-calcite) (Fig. lc), which is common in coralline sponges. In this case the secondary skeleton is analogous to the primary collageous/spicular skeleton, wh ich is substituted by Mg-calcite (Fig. lb). The entire basal skeleton exhibits a chaetetid structure: that means the calicles are separated by tabulae which mark the boundary between the living sponge and the dead part of the basal skeleton (for measurements see Gray 1980" p.809). 182 J. Reitner l~ a c Fig.1. Chaeleles (Boswellia) morloni Gray. a Tylostyle megasclere. b Reconstruction ofthe spicular arrangement which is similar to the extant genus Suberiles. c Primary and secondary skeleton of a single wall Fig. 2. a Horizontal section of Ch.(B.) morloni Gray demonstrating the spicule bundles. Gray's para type (British Museum Na!. His!. no R 4429). Scale = 200 J.'m . b Detail of a Spicules partly preserved in pyrite. Scale = 100 J.'m . c Tylostyle megasclere of Ch.(B.) morloni. Scale = 50 J.'m Phylogenetic Aspects a nd New Descriptions o f Spicu le-Bea ring Hadro merid Sponges 183 Fig.3. a Suberiles carnosus; view on the surface ofth e dermallayer. Cora llingene, Banyuls sur Mer, marine station Arago. b Deta il of derma llayer of the specimen of a . The spicules are a rranged in a subradia te way Remarks: The spicular skele ton of Chaetetes mortoni is comparable with the dermallayers of the families Suberitidae and Polymastiidae. Both groups have only rare microscleres a nd tylostyle megascleres. Microscleres are not known from C. mortoni. Significa nt are the bundles of megascleres, a rranged in a loose per- pendicular orientation , wh ich is cha racteristic for th e Suberitidae (F ig. 3a,b). This species is the oldest known had romerid sponge. Chaetetopsis favrei (Deninger 1906) (Figs. 4,5) 1906 Monotrypa favre i Deninge r, p. 64, PI. VI , Fig. 5a ,b 1930 Chaetetopsis fav rei (Deninger); Peterh ans, p. 35 , PI. I ,ll 1970 Chaetetopsis favrei (Deninger) ; F ischer, p. 190 1979 Chaetetopsis fav rei (Den inger) ; Kazmierczak, F igs. 2- 4 This early Cretaceous coralline sponge speci es is very common in Barremian and Aptian reefs. Kazmierczak (1979) repo rted the first monaxonic spicule remains fro m a specimen from the Crimea. The spicules a re preserved in rows offramboid pyrite. The exact sha pe of the spicules is not shown in this particular speci men . Newly collected material from the Albian ofGreece a llows one to distinguish th e spicule type and a rrangement, as well as the micros tructure of the calcareous , skeleton. 184 Primary skeleton: J. Reitner Fig. 4. Chaetetopsis favrei (Deninger). a Horizontal sec- tion exhibiting the spicule arrangement (black dots). Re- drawn from Kazmierczak 1979. b Vertical cut of a specimen from Arachova (Aptian? from Greece). c Tylostyle megasclere of the specimen from Arachova The tylostyle spicules (120-450 JLm length, 5-10 JLm thick) (Figs. 4c, 5c-e) are arranged in loose bundles (Figs. 4,5c). The specimen from the Barremian of Crimea covers partially the inner wall ofthe calic1es (Fig. 4a). Horizontal spicule elements and microsc1eres are not observed. The loose arrangement of the spicules indicates a prominent collagenous skeleton wh ich has a matrix function for the secondary calcareous skeleton. Secondary skeleton: The entire basal skeleton demonstrates a chaetetid structure (Figs. 4c, 5a). The calic1es are subdivided by tabulae. The tabulae are regular and thin (10-20 JLm) and composed of minimicrite. The vertical wall exhibits a fasicular fibrous microstructure, as seen in Paleozoic Chaetetes (Fig. 5b). The thickness ofthe wall va ries only a litde (300- 700 JLm). The spicules are intramural. Some ofthe spicules protrude from the calcitic wall. In this case the spicules were primarily not covered Fig. 5. Chaetetopsis favrei (Deninger). a Vertical section of the specimen of Arachova (near Delfi, Greece; Aptian? component in an Eocene "Qphiolithe Conglomerate"). Scale = 200 J.Lm. b Fascicular/ hemispherulitic microstructure of the calicle wall of the Arachova specimen. Scale = 75 J.Lm. c Tylostyle megasclere of the Arachova specimen. The spicule is partly pre- served by pyrite spherules (framboid pyrite, probable formed by bacteria). Scale = 50 J.Lm . d Tylostyle megasclere of the Arachova specimen. The spicule is preserved by a granular calcite. Scale = 50 J.Lm. e Intramural tylostyle megascleres of the Arachova specimen. Scale = 75 J.Lffi Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 185 186 J. Reitner by collagenous soft tissue. The basal skeleton reflects the primary sponge skeleton. The primary microstructure is, in most ca ses, not recognizable. The skeleton exhibits a diagenetic granular or blocky calcite. Only in a few examples are relicts of the original structure observed. Based on this observation, an aragonitic primary mineralogy is very probable. Remarks: Spicule type and the relict radially oriented spicules point toward a classification within the hadromerids. The axially oriented intramural tylostyle scleres in- dicate that Chaetetopsis as weIl as Chaetetes mortoni Gray is also a member ofthe family Suberitidae. Family Chondrosiidae Schulze The Chondrosiidae are sponges with a strong collagenous skeleton lacking, in most cases, the spicular skeleton. Only aster microscleres, mostly euasters, are present in the genus Chondrilla . Megascleres are unknown. All members ofthis group are characterized by a strong cortex in which the microscleres are enriched if present. The cortex is formed by fibrillar collagen. The sponges are encrusting or massive. Genus Chondrochaetetes n.gen. Type species: Chondrochaetetes longitubus n.gen. n.sp. Derivatio nominis: Combined from the hadromerid family Chondrosiidae "Chondro-" and the term "Chaetetes" for basal skeletons with subdivided by tabulae into calicles. Diagnosis: Hemispherical sponges with a chaetetid secondary calcareous basal skeleton. The microstructure of the vertical walls is fascicular fibrous. The tabulae are micri- tized. Intramural aster microscleres are common within the vertical walls. Chondrochaetetes longitubus n.gen. , n.sp. (Fig. 6a-e) Derivatio nominis: After the Latin word "longitubus" since this species pos- sesses very long calicles. Fig. 6. a Chondrochaetetes longitubus n.gen. n .sp., holotype from the Carboniferous of the USSR. Horizontal cut exhibiting asterose microscleres (arrows) (BMNH no. R 27318). Scale = 500 pm. b Vertical section ofthe holotype of Chondrochaetetes. The very thin tabulae and the calicle wall are covered by a granularcalcitic cement. Scale = 500 pm. c Detail ofthe calicle wall of Chondrochaetetes demonstrating the fascicular fibrous wallstructure and intramural asterose microscleres. Scale = 200 /Lm. d Intramural asterose microsclere of Chondrochaetetes. Preservation is in granular calcite. Scale = 50 /Lm . e Intramural asterose microscleres (euasters?) of Chondrochaetetes. Black cores of the spicules are pyrite. Some of them exhibiting relicts of the star rays (arrow). Scale = 50/Lm Phylogenetic Aspects a nd New Descriptions of Spicule-Bearing Hadromerid Sponges 187 188 J. Reitner Holotype: British Museum Natural History no. R 27318 (labeled as " Chaetetes radians" Fischer). Diagnosis: See genus diagnosis Locus typicus : Kaluga Borovitch, Yaldai, Russia Stratum typicum: Carboniferous Limestone Description of the holotype : Primary skeleton: The observed spicules are asters, probably euaster, and they are irregularly distributed within the calcitic wall. The spicules are preserved in calcite and have a diameter of25- 35 /Lm (Fig. 6c-e). Secondary skeleton: The uncut specimen has a diameter ofnearly 5 cm. The calicles have a diameter of200-300 /Lm and exhibit a radial pattern. They are separated only by few (3 or 4 tabulae/cm) thin (ca. 30 /Lm) tabulae (Fig. 6b). In some areas a dense packing oftabulae is observed with more than lO/cm. The microstructure ofthe vertical walls are fascicular fibrous (Fig. 6c). The tabulae have a minimicritic structure. Walls and tabulae are covered by a late diagenetic granular cement (Fig. 6b). Differential diagnosis: The new species differs from Chaetetes mortoni in possessing aster microscleres only. Other differences ofthe new species Calcistella tabulata include a missing canal system within the calicle wall. The microstructure of Chondrochaetetes is fascicular fibrous in contrast to the dense micritic structure of Calcistella. The microscleres are significantly smaller with a mean diameter of 30 /Lm in comparison with microscleres of Calcistella, which exhibit a mean diameter of55/Lm. Remarks : This Carboniferous spicule-bearing calcified demosponge has, based on the aster (euaster?) scleres, close affinities to the family Chondrosiidae. However, an enrichment of microscleres in any zone of the chaetetid skeleton is not observed and there are no traces of a dermallayer. The spicules are irregularly distributed, probably caused by early diagenetic dissolution of the scleres and rapid epitac- tical cementation ofthe molds{cf. Reitner 1987b,c). The new form is probably related to the genus Chondrilla , wh ich contains euaster microscleres only. In contrast to the tylostyle megasclere-bearing species, the chaetetid basal skel- eton of this particular form is not analogous to the primary spicular/organic skeleton. Genus Calcistella n.gen. (Fig. 7a-d) Type species: Calcistella tabulata n.gen., n.sp. Derivatio nominis : After the Latin words "calcium" and "stella" = star. Diagnosis: The aster microscleres-bearing sponge possesses a modified chaetetid secondary micritic skeleton. Diameter ofthe calicles is exceptionally large. The calicle walls Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 189 are structured by numerous pores and canals. The asters are restricted to the ' mineralized micritic part of the vertical walls. ' Differential diagnosis: See species description Calcistella tabulata n.gen., n.sp. (Fig. 7a-d) Derivatio nominis: After the Latin word "tabula" = plate Holotype: Institut für Paläontologie der Freien Universität Berlin (IPFUB, . JRl/89). Locus typicus: Arachova near Delfi, Greece; sampled by H. Keupp Stratum typicum: Albian component within the "Ophiolithe-Conglomerate" of Eocene age. Diagnosis: See generic dia gnosis Description ofthe holotype: Primary skeleton: The spicular skeleton is composed ofaster microscleres only (Fig. 7c,d). The asters are preserved as spherules and exhibit only, in a few cases, relicts of spines. Based on this observation, the asters were probably euasters. The diameter of the spherules va ries little (50-65 /Lm with a mean value of 55 /Lm) . The asters are restricted to dark micritic areas within the calicle wall (Fig. 7c). The dark areas form a more or less irregular network within the vertical skeletal elements, interrupted by pores wh ich mark complicated canal systems inside the calicle wall. The canal pores are separated from the dark micritic areas by thin granular sparitic layers. The diameter of the canal systems are 140-400 /Lm. The enrich- me nt of spicules in the micritic parts only indicates a thick fibrillar primary collagenous skeleton. The canal systems are probably excurrent canals. Spicules are never observed within the tabulae. Secondary skeleton: The entire sponge exhibits a hemispherical chaetetid structured basal skeleton (6 cm diameter measured in thin section) with extraordinarily big tubes (Fig. 7a,b). The diameter of the tubes va ries from l.2-3 mm (mean l.8 mm). The micro- structure oftabulae and the vertical wall is micritic. Sparitic lamellae are common above the tabulae, at the base ofthe tubes, and within the excurrent canals (Fig. 7b). These ca. 20 p.m thick lamellae are interrupted by thin micritic layers and mark the upward migration ofthe living sponge tissue. The sparitic layers mark the gap between the secondary skeleton and pinacoderm ofthe living sponge. The primary mineralogy of the basal skeleton was a high Mg-calcite. All associated aragonitic organisms are recrystallized. EDAX analysis show Mg anomalies within the wall and support this assumption. Remarks: This type of chaetetid basal skeleton is analogous to the primary collagenous skeleton ofthe living sponge. The collagenous skeleton was probably very strong ; and became calcified synvivo, as indicated by the sparitic growth oflamellae. The , microscleres are concentrated within the thick fibers only. The calicle wall was !p 190 J. Reitner Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 191 also penetrated by choanosomal soft tissue with prominent excurrent canals. This is in contrast to all other observed chaetetid sponge skeletons. The very big microscleres, the strong collagenous skeleton, and the canal systems are comparable with the family Chondrosiidae, particularly with the ge- nus Chondrilla (Figs. 7e,lOb). The euasters of Chondrilla grandistellata show dimensions as measured in Calcistella (Figs. 7e,lOb). The mineralized internal canal systems were probably formed by the endopinacoderm layer. The calical walls became mineralised by calcitication ofthe collagenous tibers, as observed in the extant coralline sponge Vaeeletia (Reitner 1987b). The tabulae were probably formed by the exopinacoderm layer, as observed in the extant ha- dromerid sponge Spirastrella (Aeanthoehaetetes) wellsi. This may explain the different observed structures ofthe calcitic secondary rigid body. Differential diagnosis: This new form differs from the genus Chondroehaetetes n.gen. by having a complicated excurrent canal system within the calicle walls, micritic micro- structure, and aster microscleres 55 /Lm in diameter in contrast to 30 /Lm measured in Chondroehaetetes. The chaetetid species from the early Cretaceous of China, Pseudomillestroma retieulata Deng (Deng 1982), differs from the new species in having smaller asters and missing a canal system inside the calicle wall. Genus Calciehondrilla n.gen. (Fig. 8a-d) Type species: Caleiehondrilla crustans n.gen., n.sp. Derivatio nominis: After the Latin word "Calcium" and the genus name "Chondrilla" for hadromerid sponges which have only euaster microscleres. Diagnosis: Crustose sponges which have large as'ter microscleres only. The scleres are enriched in the primary strong collagenous parts of the skeleton. Prominent excurrent canals do exist. The entire primary collagenous skeleton of the sponge is calcitied. Differential diagnosis: See species diagnosis Caleiehondrilla erustans n.gen., n.sp. (Fig. 8a-d) Derivatio nominis: After the Latin word "crustae" for crusts. Holotype: Deposited in the IPFUB, JR2/89 (Fig. 8a-d) ... Fig.7. a Tangential cut of the basal skeleton of Calcistella tabulata n.gen. n.sp. (holotype) from the Aptian? of Arachova (Greece). Within the thick calicle walls the characteristic canal systems are visible. Scale = I mm. b Calicles of Calcistella exhibiting the tabulae and the growing lamellae ofthe upward moving soft tissue preserved in a granular calcite. The calicle wall exhibiting numerous aste rose microscleres. Scale = I mm. C Detail of ashowing the internal canal systems and the spicules. Scale = 800 ,..m. d Calcite filled scleres molds (euasters?) of Calcistella. Scale = 50 ,..m. e Euaster microscleres of Chondri/la grandistellata (SEM microgaph). Compare d 192 J. Reitner Fig.8. a Total view (oblique cut) of the holotype of Calcichondrilla crUSlans n.gen. n .sp. from the Middle Albian ofSaturarran (northern Spain). Scale = I mm. b Detail ofa exhibiting the excurrent canals and the spicular skeleton . Scale = 400 /Lm. c Probably euasters of Calcichondrilla . Molds preserved in a polycrystalline calcite. Scale = 100 /Lm. d Lamellar structure ofthe calcified primary strong collagenous skeleton of Calcichondrilla. Scale = 200/Lm Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 193 Paratypes: Deposited in the IPFUB, JR3/89 (Fig. 9a-c) Locus typicus: Beach of Saturarran, Basque Lands in Northern Spain Stratum typicum: Middle Albian elast from a debris fiow of Late Albian age. Description ofthe holotype: Primary skeleton: The incomplete specimen is found in a thin section and is 5 mm in diameter. The specimen is cut obliquely and shows different niveaus ofthe excurrentcanals (Fig. 8a,b). The diameter of the excurrent canals varies between 100 and 400 pm. The excurrent canal systems are radially arranged. The spicules are enriched in areas between the canals. The inner surface ofthe canals are free ofspicules (Fig. 8b,d). The spicules are big euasters (75-100 /Lm diameter, mean 80 /Lm). The asters are preserved as calcitic spherules which exhibit, in some cases, relicts of short rays (Fig. 8c). The entire skeleton demonstrates irregular lamellae wh ich refiects the ' fibrillar character of the strong primary collagenous structures (Fig. 8d). No separate dermallayer enriched in microseleres is present. Secondary skeleton: The whole primary skeleton of the sponge is calcified. The entire skeleton is becoming calcified and is analogous to the primary collagenous skeleton. The primary mineralogy was a high Mg calcite. The microstructure is irregular. In most cases, the calcite is brownish in color. Description ofparatype I (Fig. 9a,b) The specimen was found in a thin seetion of a Lower Albian rock of the. Euzkadiella-facies of the red lagoonal rudist limestone of Ereno, 30 km west of Saturraran. Calcicondrilla forms a thin crust (mean 200 /Lm) on the surface of an already dead specimen of the coralline haploselerid sponge Euzkadiella eren- oensis. In some areas the sponge exhibits small buds (200- 250 /Lm). The diameter of the seleres varies from 75-115 /Lm. The secondary skeleton has the same irregular brownish calcite. Because of the thin crust, no excurrent canals are observed. Description of paratype 11 (Fig. 9c) The specimen was cut horizontally and is located in a thin section of the Lower Albian gray reeflimestone ofErefto. Calcichondrilla crusts on microsolenid corals ofthe Acanthochaetetes community. The thick part ofthe crust exhibits horizontal cuts of the excurrent canals. The very thin crusts overgrow biogenes, similar to bryozoans and algae. On the sponge a thecidellinid brachiopod is cemented with its fixed valve. This indicates the rigid character ofthe living sponge. The spicules are very densely arranged. Relict structures of the collagenous skeleton, as seen in the holotype, are not observed. The entire sponge represents first-order framebuilders of the Acanthochaetetes community. Remarks: The new genus Calcichondrilla exhibits many similarities with the genus Chon- drilla (Figs. 7e, lOa,b) especially the type of microseleres and the internal structures. The species Chondrilla grandistellata bears the same large spicules , (euaster diameter 75-150 /Lm , mean 100 /Lm) , the same type and dimensions of . 194 J. Reitner Fig. 9. a Paratype I of Calcichondrilla crustans n.gen. n.sp. from the Late Aptian "Red Rudist Limestone" of Ereflo (northern Spain). The specimen is growing on the haplosclerid stromatoporoid Euzkadiella erenoensis Reitner. Scale = I mm. b Detail of the spicular skeleton of paratype I. Scale = 300 ILm. c Paratype 11 from the la te Aptian gray reeflimestone ofEreflo (northern Spain). Scale = 500 ILm excurrent canals, and the same lamellae in the collagenous skeleton. All species of Chondrilla exh ibit a characteristic dermallayer with an enrichment ofspicules. Within the holotype no dermal layer is observed, but the thin crusts seen in paratype land Il may be interpreted as dermallayers. The systematic position within the family Chondrosiidae and a close relationship with the genus Chon- drilla is very probable. This species is part ofthe Acanlhochaeteles comm unity in the A ptianl Albian reefs (Reitner 1989a). Calcichondrilla is observed in northern Spain and within the Mural Limestone in Arizona (Reitner et al. , in press). Differential diagnosis : The new species differs from Calcistella tabu/ata in having larger microscleres, the crustose growing pattern, and the lack of chaetetid basal skeleton. The closely related genus Chondrilla is missing a calcareous secondary skeleton. Family Spirastrellidae Rid/ey & Dendy 1886, Sensu Hentschel1909 Emend. diagnosis The Spirastrellidae are sponges with tylostyle megascleres and spiraster microscleres with great variability. The sponges encrust or form massive bodies. Fig.l0. a Vertical section through the dermallayer ofRecent Chondril/a grandislel/ala exhibiting the enrichment of euasters. Scale = 500 JLm. b Euasters ofa. Scale = 100 JLm. c Choanosomal area of Ch. grandislel/ala exhibiting the excurrent canals (compare Fig. 8b). Scale = 500 JLm In some cases prominent astrorhizal excurrent canals are visible on the surfaces ofencrusting specimens (Figs. 11e-g, 13e) . In very few cases, excavating species are observed. High Mg calcite chae- tetid-structured secondary skeletons are present in the subgen us A canthochae- tetes since the early Cretaceous (J urassic?). Genus Spirastrella Schmidt 1868 Synom. list: See Wiedenmayer 1977 Type species: Spirastrella eunetatrix Schmidt 1868 Subgenus A canthochaetetes Fischer 1979 n.stat. 1970 - A canthoehaetetes n.gen. - Fischer, p. 199 1976 - Tabulospongia n.gen. - Mori , p. 5 Type species: A eanthochaetetes seunesi Fischer 1970 Remarks : The validity of the taxon "A canthochaetetes" is questionable. Rios and Almela (1944) have described a chaetetid sponge, Septachaetetes eocenus, from the Eocene of the southern Pyrenees, which exhibits a basal ske leton similar to A canthochaetetes. New collections ofthis taxon do not show any spicu le remains. 196 J. Reitner Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 197 The calicles ofthis species exhibits only few spines, or in most cases, the spines are missing. Septachaetetes is the older synonym of Acanthochaetetes. Ifboth taxons are similar, Septachaetetes is the valid name. But this question remains unans- wered, because spicules are missing. Diagnosis: Sponges with tylostyle megascleres and spiraster microscleres are similar to Spirastrella with an additional high Mg-calcite chaetetid skeleton. Within the chaetetid secondary skeleton modified archaeocytes could be enclosed (modified gemmules). Spirastrella (Acanthochaetetes) wellsi (Hartman and Goreau) n.comb. (Fig lla-d) Selected synonym list: 1975 - Acanthochaetetes wellsi n.sp. - Hartman and Goreau, pp. 2, Figs. 1-14. 1976 - Tabulospongia horiguchii n.gen., n.sp. - Mori, pp. 2, PI. 1-5 1977 - Tabulospongia japonica n.sp. - Mori, pp. 2, PI. 1-2 1987 - Acanthochaetetes wellsi - Reitner and Engeser, Figs. 1,3,5,7,8 Diagnosis: See subgenus diagnosis Description of the primary skeleton and soft tissue arrangement: The spicules are straight or slightly curved tylostyle megascleres (200-360 /Lm length) (Fig. 11 b). In contrast to the extremely abundant microscleres the tylostyles are rare and demonstrate no preferred orientation. The microscleres ar~ spirasters with extreme variation in shape (Fig. 11 b,c). Simple star-shaped scleres and all variations up to elongated amphiaster types are observed. The microscleres are enriched within the dermallayer (Fig. 11 b). Collagen fibrils are present to anchor the soft tissue with the calcitic basal skeleton. These fibrils are seen only inside of the calicle wall and are never observed within the upper part ofthe spines and tabulae. Vacelet and Garrone (1985) describe additional intracellularly collagen fibrils and large bundles of fibrils between cells at the bottom ofthe calicles forming a matrix for the tabulae. Spongin is not observed. Fig. 11. a SEM micrograph of Spirastrella (Acanthochaetetes) wellsi (Hartman & Goreau) from Guam (Pacific) demonstrating the chaetetid character of the basal skeleton. Only the youngest part is free of early diagenetic cements. b Histological section of the dermallayer S.(A.) wellsi exhibiting the microscJeres coat and the tylostyle megascJeres (Vacelet coll.) Scale = 100 /Lm . c Spiraster microscJere of S .(A .) wellsi (SEM micrograph). d Lower part ofthe youngest part ofa calicJe in S.(A.) wellsi. The base ofthe tube is covered by choanosomal tissue exhibiting the choanocyte chambers (a). The boUom of the calicJe is cJosed by an organic tabu la (b). The black part (c) are stained totipotent archaeocytes encJosed by tabulae. The lower tabula (d) is becoming mineralized. The archaeocytes below the first archaeocyte chamber are totally encJosed by mineralized tabulae (j) . (Vacelet coll.) Scale = 300 /Lm. e Spirastrella cunctatrix Schmidt from the Mediterranean Sea near Banyuls sur Mer. Tbe specimen exhibiting strong vertical canal systems comparable with the calicJe systems in S. (Acanthochaetetes) wellsi. (SEM micrograph). f Spirasters of Spirastrella cunctatrix (SEM micro- • graph). g Tylostyles of Spirastrella cunctatrix (SEM micrograph) , 198 J . Reitner The soft tissue covers only the youngest calicle (Fig. 1Ia). The choanosome covers the calicle wall. The biggest space ofthe calicle is needed for the aquiferous excurrent system. The choanosome is anchored with the endopinacoderm at the basal skeleton. The choanocyte chambers are small and have a mean diameter of 20 p.m (Hartman and Goreau 1975; Vacelet 1981) (Fig. lId). The surface ofthe sponge is covered by a dermallayer which is penetrated by incurrent pores and canals as weIl as superticial excurrent systems demonstrating an astrorhizal pattern. For further detailed descriptions, see Hartman and Goreau (1975). Very important is the occurrence of totipotent archaeocytes which are captured by tabulae in older parts of the calicles (Fig. lId). Secondary skeleton: The sponge has an unique high Mg-calcite skeleton. The entire skeleton exhibits a classical chaetetid structure (Fig. lla). The microstructure of the skeleton is characterized by irregular lamellae of elongated Mg-calcite crystals (for detailed descriptions and chemical compositions see Reitner and Engeser 1987). The calcite lamellae are present within a thin mucus layer between the exopinacoderm and the rigid calcitic skeleton. The thickness of the mucus layer is ca. 200 nm, measured by one discrete, newly formed crystallayer. The newly formed crystals are needle-like and show a mean length of2-4 p.m and diameter ofless than 250 nm. These primary crystals increase size by rapid epitactical crystal growth. The separate crystals grew together and formed a rigid calcareous body. Within the tabulae, the crystals are oriented more or less parallel. In some cases the crystals exhibit a random orientation. Within the empty calicles a randomly oriented high Mg-calcite cement ("back-till") is observed (Fig. lla). Very common are asexual buds. The buds have their roots in one single calicle of the parent sponge. This budding feature sterns from escaped, mobile, to- tipotent archaeocytes, forming a genetically identical sponge. The size of the calicles are mostly smaller than those seen in the parent sponge. These simple procedure for skeletal formation is supported by stable isotope geochemical data. An equilibrium with seawater (high positive delta C13 (+ 4) and delta 0 18 (-0.9 data) is indicated (Reitner 1989b; Reitner and Grothe 1988). Spirastrella (Acanthochaetetes)·seunesi (Fischer) n.comb. This species is very common in Late Aptian to Late Albian Urgonian reefs with a worldwide distribution (Reitner 1989a). Selected synonymy list: 1970 - Acanthochaetetes seunesi n.gen., n.sp. - Fischer, p. 201 , Fig. 32, PI. F, Figs. 3- 5 1972 - Acanthochaetetes seunesi - Fischer and Lafuste, Figs. 1-7, PI. 7 non 1973 - Acanthochaetetes seunesi - Cuif et aI. PI. 1, Fig. 6 1983 - Acanthochaetetes seunesi - Reitner and Engeser, PI. 1, Figs. 4-8 1987 - Acanthochaetetes seunesi - Reitner and Engeser, Figs. 2,4,6,8,10. Description of the skeleton: This sponge has straight or slightly curved tylostype megascleres and star-shaped spiraster microscleres. The microscleres dci not show the extreme variation as seen Phylogenetic Aspects and New Descriptions ofSpicule-Bearing Hadromerid Sponges 199 in S.(A.) wellsi. The spicules are embedded within the calcitic skeleton. The megascleres have no preferred orientation within the secondary skeleton. The microscleres are quite often enriched within the tabulae. Megascleres are ne ver entrapped within the tabulae. Within the calicle walls, microscleres are rare (further details see Reitner 1982, 1987d; Reitner and Engeser 1983, 1987). The structure and chemistry of the chaetetid basal skeleton is similar to S.(A .) wellsi. Some specimens show the same budding features as seen in S.(A.) wellsi. Spirastrella (Acanthochaetetes) dendroformis n.sp. (Figs. 12b, c; 13d) 1983 - Acanthochaetetes ramulosus. - Reitner and Engeser, pp. 774- 775, PI. 1, Figs. 1- 3. 1987d - Acanthochaetetes ramulosus. - Reitner, PI. 18, Fig. 2 (not A. ramulosus Michelin) Derivatio nominis : After the Creek word "dendr-" = tree shaped branches Holotype: Depository IPFUB, JR4/89 (Figs. 12a-c, 13a-d). Material: Over 50 specimens Locus typicus: Pefta Albeniz near the villa ge Araya, northern Spain. Stratum typicum: LaJe Albian fore reef sediments of the Albeniz reef mound. Diagnosis : This sponge exhibits a ramified chaetetid basal skeleton. Structure and chemistry is similar to other S. (Acanthochaetetes). Only the axial part of the calcareous skeleton is irregularly arranged. In this part only few calicles separated by tabulae are visible. The intramural tylostyle megascleres are arranged radially within the Fig. 12. Spirastrella (A canthochaetetes) dendro{ormis n.sub. gen., n.sp. a Draw- ing of the ontogenetic older parts of the central zone of S.(A.) dendroformis ex- hibiting an irregular stromatoporoid structure of the basal skeleton and the spicule arrangement (aquiferous sys- tems dotted). b Tylostyle megascJere. c Spiraster ("amphiaster") $pirastrella (Acanthochaetetesl dendroformls ---.... 10um Fig.13. a Micrograph ofthe holotype of Spiraslrella (A canihochaeleleS) dendro(ormis demonstrating the central irregular zone and the chaetetid outer zone. Specimen from Late Albian reefs near Araya (northern Spain). Scale = I mm. b Detail ofthe central zoneofthe holotype exhibiting the megascleres arrangement. Scale = 500 p.m.c Tylostyle megasclere of S.(A.) dendro(ormisetched with titriplex acid. SEM micrograph. d Spiraster ("amphiaster") microsclere of S.(A.) dendro(ormis etched with titriplex acid. (SEM micrograph). e Spiraster microsclere of Spiraslrella cunClalrix Schmidt (SEM micrograph) Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 201 axial part. Microscleres are modified ("amphiaster") spirasters, which are en- riched within the axial part and tabulae of the outer part of the sponge. l Description of the holotype: Spicular skeleton: The spicular skeleton is composed of tylostyle megascleres with a length of 325-450 /Lm (mean 365 /Lm) (Figs. l2a-c, l3b,c) and spiraster microscleres mostly of amphiaster shape (Figs. l2a-c, 13d). The diameter varies from 25-35 /Lm (mean 28 /Lm). The tylostyles are common in the central part and arranged in the same fan shaped pattern as observed by the tube systems (Fig. l3a,b). In some cases the spicules are arranged in small bundles which reflects the original radial pattern (Fig. l2a-c). Some of the megascleres protrude into the open space of the calicle. Microscleres are common and mostly irregularly distributed (Fig. ·12a-c). An enrichment is observed within the tube walls of the central part and the tabulae of the outer part. Secondary skeleton: From the holotype, three thin sections were made: the horizontally, vertically, and oblique section. The vertical median section has a length of7.5 cm. T4e stern has a thickness of l. 7 cm. The irregular center part of the sponge varies from 5 mm- 1O mm. The center part is characterized by irregularly ascending tubes. The whole skeletal structure is fan shaped (Fig. l3a,b). Only a few tubes are separated by thin tabulae. The boundary between the center and the outer part is marked bya prominent tabulae layer. The outer part is characterized by a true chaetetid structure as seen in other acanthochaetetids. Microstructure and chemistry ofthe calcitic skeleton is similar to other Spirastrella (A canthochaetetes) species. The inner calicle diameter varies from 320-520 /Lm (mean 433 /Lm) . The outer calicle diameter varies from 480-1000 /Lm (mean 703 /Lm). Buds as seen in hemispherical species of S.(A.) are not present. Differential diagnosis: The new species exhibits close similarities with the secondary skeleton of Spi- rastreIla (Acanthochaetetes) ramulosus (Michelin). But, in S.(A.) ramulosus, only few megascleres are observed and never spiraster microscleres. In S.(A.) seunesi (Fischer), no "amphiaster" -type spirasters are observed. Tylostyles are rare, never arranged in bundles, and never exhibit any radial pattern. Remarks: The spicular skeletons of all basal skeleton bearing spirastellids are similar to the species Spirastrella cunctatrix Schmidt (Figs. lle-g, l3e). The extant S.(A.) wellsi (Hartman and Goreau) exhibits a wide variability ofspiraster microscleres which is never observed within fossil forms. In S. cunctatrix microscleres are very abundant. The tylostyles are arranged in small bundles with a maximum of 20 scleres. Most bundles possess five to six spicules (Fig. llg). Within the extant S.(A .) wellsi this feature is also observed. Within all fossil species, onl)( S.(A.) dendroformis exhibits this feature . 202 J. Reitner An unique feature is the presence ofmodified archaeocytes within the older portions ofthe calic1es. These archaeocytes are comparable with gemmule bodies of freshwater sponges (Fig. lId). In so me investigated a prominent vertical exhaltant tube system is observed, perhaps comparable with the calic1es in Spirastrella (A canthochaetetes) (Fig. lle). Family Cassianothalamiidae Reitner (l987b) The affinities ofthis extinct coralline sponge group from the Late Triassie (Lower Carnian) with the order Hadromerida is based on the presence of spiraster microsc1eres. Besides spirasters, sterrasters and rare monaxonic megasc1eres are also observed. Atypical are the fusiform spicules. Cassianothalamia zardinii Reitner (1987b) 1985 -? Stylothalamia n.sp. - Reitner and Engeser, p. 170; PI. 4, Figs. 8-12 1987 - Cassianothalamia zardinii n.gen., n.sp. - Reitner, p. 573, PI. 1-3; PI. 4, Fig.l Primary and secondary skeleton: The entire sponge exhibits a thalamid structure, convergent with the modern Vaceletidae. The thalamid collagenous skeleton was subsequently mineralised by an irregular micritic high Mg-calcite. A prominent spongocoel is present. In- tramural spicules are rare. Mostly present are small spirasters, some bigger sterrasters, and very few monaxonic megasc1eres which rarely protrude into the open space of the chambers. The primary organic skeleton is marked by men- iscus-shaped structures which cover the megasc1eres in the outer part of the vertical pillars (for further details see Reitner 1987b). Remarks: Cassianothalamia is probably a hadromerid sponge, based on the spirasters. The relationship to any extant Hadromerida is not known. The problem lies with the sterrasters, wh ich are observed only in the Geodiidae, order Choristida, while the characteristic triaene megasc1eres are missing in Cassianothalamia. The ha- dromerid sponge Placospongia exhibits sterraster-like sc1eres (selenasters) beside small spirasters. Any further relationships are not known. The thalamid or sphinctozoid rigid calcitic skeleton exhibits the same shape as the extant ceractinomorph sphinctozan sponge Vaceletia crypta (Vacelet). Functional Interpretation of the Observed Basal Skeleton Types Seven different secondary skeletal types are present within the Hadromerida. Five of them exhibit a chaetetid structure, one a thalamid, and one a modified crust, wh ich is the simplest basal skeleton type. Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 203 Chaetetid skeletons are known from two extant demosponges, Spirastrella (A.) wellsi Hartman and Goreau, linked to the Hadromerida, and Merlia normani Kirkpatrick, linked to the Poecilosclerida. One species from the Calcarea, Pe- trobiona massiliana Vacelet, linked to the Calcaronea, is known, which contains a modified chaetetid skeleton (Reitner 1987a). In all three examples, the soft tissue ofthe active living sponge covers only the top ofthe skeleton and occupies only the youngest calicle. These sponges also produce totipotent archaeocytes enclosed in the older parts of the calicles. In S.(A.) wellsi the archaeocytes are separated from the active living sponge by organic or mineralized tabulae. In Merlia normani separation is also made by mineralized tabulae with a central hole and in Petrobiona the archaeocytes are located at the end of narrow canals within the basal skeleton (Vacelet 1988; Reitner 1987a; Reitner and Keupp 1989). Besides the rigid Merlia normani a "soft" form is known which does not hav!! any calcareous skeleton (Merlia deficiens sensu Vacelet 1980). Van Soest (1984) pointed out that both Merlia species are the same. This is important because the soft form does not construct any gemmule bodies. Both forms are probably ecotypes but the problem is, they occur together in same ecological niches. This feature is also observed in Spirastrella. The spicular skeleton of the subgenus Acanthochaetetes is closely related to Spirastrella. Acanthochaetetes has been restricted to deep fore reef areas and cryptic niches of coral reefs since the Lower Cretaceous and Spirastrella is not part of the A canthochaetetes community (Reitner 1989a). An ecological controlling factor in the formation of the basal skeleton is proposed for Acanthochaetetes. This may be true for the calcisponge Petrobiona, which is located only in submarine caves and under huge rocks (Vacelet 1964; Reitner 1989a). The formation of a chaetetid or modified chaetetid skeleton is linked, in these cases, with ecological parameters. All modern demosponge chaetetid skeletons are secreted by the exopinacoderm and are not analogous to the primary skeleton of the sponge. Within Chaetetes (Boswellia) mortoni, Chaetetopsis favrei , and Chondro- chaetetes longitubus, the secondary skeleton is analogous to the primary skeleton. The living sponge occupied in these cases the ontogenetic youngest parts of the secondary skeleton. The thickness ofthe soft tissue is marked by the last tabulum. Archaeocytes may have been present within the extinct taxa. Besides the classic chaetetid skeletons, the cera toporellid type of chaetetid skeleton is known, which is characterized by a rapid, early diagenetic, epitactical cementation ofthe older calicles. In this case, no totipotent archaeocytes can be protected . Probably all open chaetetid skeletons had the function ofh uge modified gemmule bodies. This idea is supported by the observation ofsmall buds, which are connected to certain calicles of the parent sponge, as seen in S.(A.) wellsi. The primary collagenous skeleton of the thalamid hadromerid sponge Cassianothalamia became calcified immediately after new chamber formation (Reitner 1987b). Common small asexual buds linked with the parent skeleton are observed. A formation from modified gemmules is probable. Two further noncalcified hadromerids form gemmules, Cliona lobata and several types of the Suberites taxon. Gemmulue formation is therefore not restricted to haplosclerids. 204 J. Reitner In S.(A.) dendroformis two ontogenetic stages of secondary skeletal forma- tion are observed. In the early stage calcitic skeleton is analogous to the primary skeleton. The irregular tubes probably reflect the primary aquiferous system. Megascleres are common, regularly oriented, and exhibit, in some cases, the primary aggregates as seen in Spirastrella. The ontogenetic late stage is charac- terized by anormal acanthochaetetid skeleton which is not analogous to the primary sponge skeleton. In Calcichondrilla the entire primary sponge skeleton became calcified. No chaetetid structure or special cavities for totipotent archaeocytes are observed. This basal crust is occasionally an important reef-frame stabilizer. Phylogenetic Theories (Figs. 14,15) To demonstrate possible phylogenetic relationships, the method ofphylogenetic systematics was used. The Hadromerida and Tetractinellida are a monophylum, based on a number ofwell-defined synapomorphies (Hartman 1982; van Soest 1987). Van Soest (1987) has discussed three different cladograms based on the classification of Hartman (1982), his own ideas, and a cladogram based on a computer program developed by W.N. Ellis. In all three cladograms the Ha- dromerida and Tetractinellida are classified as a monophylum. Within the computercladogram, the Chondrosiidae are grouped with the Vaceletidae, based on the lack ofany spicules. Within the genus Chondrilla, asters are present. A close relationship between Chondrilla and Chondrosia is based on the similarity ofthe soft tissue and collagenous skeleton. Chondrosia is a chondrillid wh ich lost its spicules. The Chondrosiidae must be classified within the monophylum Tetractinellida/ Hadromerida. There are three views on the significance ofthe secondary skeletons. Hartman and Goreau (1972) proposed that all calcareous basal skeleton bearing sponges are a monophylum and have created aseparate class ofsponges (Sclerospongiae). Van Soest (1984) and Vacelet (1985) pointed out the similarities ofcertain spicule skeletons with noncalcified demosponges. Van Soest (1984) suggested that calcified basal skeletons were developed only once and therefore an important plesiomorphy of the demosponges. Vacelet (1985) showed a polyphyletic origin ofthe calcified demosponges based on spicules similarities with noncalcified taxa . This idea was supported by Reitner (1987a,b), Reitner and Engeser (1985, 1987), Reitner and Keupp (1989), and Wood (1987), Wood and Reitner (1988), Wood et al. (1989) based on findings of many different spicule-bearing groups of calcified demosponges. These authors also pointed out the convergent nature of chaetetid, stromatoporoid, crust, and thalamid type of basal skeletons. The two cladograms discussed here are based mainly on skeletal features to derive the phylogenetic significance of the secondary calcareous skeletons. The analysis is based on the following features: 1. Radial pattern of megascleres 2. Asterose microscleres Phylogenetic Aspects and New Descriptions ofSpicule-Bearing Hadromerid Sponges 20~ 3. Tetractinellid spicules 4. Loss of tetractinellid spicules 5. Tylostyle or modified tylostyle spicules 6. Reduction ofradiate pattern (subradiate) 7. Spiraster microscleres 8. Calcareous secondary skeleton (css) 8a. Chaetetid high Mg-calcite basal skeleton combined with internal stroma- topororid basal skeleton and a remaining radial pattern of tylostyles. 8b. Chaetetid high Mg-calcite microlamellar-structured secondary skeleton without any similarities to the primary sponge skeleton. 8c. Chaetetid high Mg-calcite irregularly structured secondary skeleton a- nalogous to the primary collagenous/spicular skeleton. 8d. Chaetetid high Mg-calcite fascicular fibrous basal skeleton analogous ~o the primary collagenous/spicular skeleton. 8e. Chaetetid aragonitic fascicular fibrous basal skeleton analogous the primary skeleton. 8f. Thalamid high Mg-calcite secondary skeleton analogous to the primary skeleton. 8g. Crust type or modified stromatoporoid Mg-calcite secondary skeletol). analogous to the primary skeleton. 9. Modified totipotent archaeocytes 10. Loss of asterose scleres 11. Reduction of megascleres Ila. Strong collagenous primary skeleton 12. No intramural spicules 13. Loss of megascleres 14. Loss of the entire spicular skeleton 15. Loss of the calcareous secondary skeleton 16. Burrowing ability in calcareous material 17. Clionid -type spirasters Theory I (Fig. 14) This theory is based on a close relationship between the Tetractinellida (Choristida) and Hadromerida. Two strong synapomorphies ofthese groups are the radiate or subradiate architecture ofthe megascleres (I) and the presence of asterose microscleres (2). The Choristida are characterized by four-rayed spicules (3) which are not present or lost within the Hadromerida (4). An important apomorphy of the Hadromerida are tylostyle or modified tylostyle scleres. Van Soest (1987) has placed the Tethyidae as a sister group of the Astrophorida (Choristida) because both groups exhibit a strong radiate pattern never observed in other Hadromerida and do not possess tylostyle spicules. The Cassiano- thalamidae (8f) may show some affinities with the Tethyidae and must separate from the Hadromerida s.st. A typical hadromerid group are the Spirastrellidae, wh ich are characterized by spirasters (7) and so me excavating species (16). The Clionidae based on spicu!~ morphology and burrowing ability are a possible sister group of the Spirast'rd- . 206 J. Reitner TheoryI Fig. 14. Theory I. Cladogram demonstrating possible phylogenetic relationships based on a poly- phyletic origin of the calcareous secondary skeletons. Character numbers see text lidae. Important is the occurrence of a particularly high Mg-calcite chaetetid basal skeleton (8b) (Acanthochaetetes); a strong apomorphy ofthis sublineage. Related to this feature are modified totipotent archaeocytes. The oldest known repre- sentatives ofthis extant lineage were found in fore-reefand cryptic environments of Lower Cretaceous reef carbonates (occurrences from the Late Jurassic are doubtful) (Fischer 1970; Reitner and Engeser 1983,1987). The specimens ofthe Lower Cretaceous exhibit a relatively high diversity (ca. six species) and some of them have intramural spicules. The younger forms have lost the intramural spicules (12). Besides the characteristic "adapted" chaetetid skeleton, one form is present which exhibits a combined chaetetid/stromatoporoid skeleton (8b). Probably this form is the primary one. The Suberitidae/Polymastiidae are a sponge group in which the asterose microscleres are lost (10). The affinities to the Hadromerida is based on the relict radial pattern of the scleres. Boury-Esnault (1987) discussed this problem and concluded that the systematic position of this particular is open at the moment. For this present study the Suberitidae/Polymastiidae are placed within the Hadromerida. The oldest known (Lower Carboniferous) hadromerid sponge, is Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 201 Chaetetes (Boswellia) mortoni. Boswellia possesses a chaetetid skeleton analogous to the primary skeleton, which is similar to the genus Suberites (8d). The Lowell Cretaceous form Chaetetopsis [avrei exhibits nearly the same type of chaetetid skeleton, and is probably evolved from Boswellia. In contrast to Boswellia, Chaetetopsis exhibits an aragonitic chaetetid skeleton (8e) and a reduced density of intramural spicules (11). The primary skeleton is probably replaced by a strong collagenous skeleton (1la). The third group of Hadromerida wh ich exhibits ca1careous secondary skeletons are the Chondrillidae within the Chondrosida. The genus Chondrilla has lost the tylostyle megascleres (13) and is characterized by asterose microscleres enriched in an ectosomal crust. An important apomorphy of the Chondrosida is a strong collagenous skeleton (11 a) . Three different basal skeletons are observed. A classic chaetetid skeleton with a fascicular fibrous microstructure and ,rare intramural aster microscleres from the Carboniferous (Chondrochaetetes longi- tubus) (8d), a chaetetid skeleton with an irregular micritic microstructure with preserved remains of the aquiferous systems within the calicle walls and numerous astercise spicules (Calcistella tabulata) (8c), and the crust type or modified stromatoporoid ca1citic basal skeleton ofthe genus Calcichondrilla (8g) is observed. This form demonstrates a very close relationship with the extant nonca1cified species Chondrilla grandistellata. All three basal skeletons are anal- ogous to the primary skeleton. Within the genus Chondrosia, the whole spicular skeleton is lost (14). Based on this theory, the secondary ca1careous basal skeletons have developed independently seven times from nonca1cified ancestors and are therefore convergent. The entire structure and function of the secondary skeleton are different. The chaetetid type skeleton is probably linked with the occurrence of totipotent archaeocytes and therefore these skeletal types are modified gemmule bodies. Basal skeleton formation is probably related to ecological parameters. Theory II (Fig. 15) The main problem with theory I is the independent development of secondary skeletons. According to van Soest (1984), ca1careous secondary skeletons are a plesiomorphic feature of the Demospongiae. This eliminates the problem of a repeated origin for ca1careous basal skeletons. Within Ca1carea, both subclasses, the Ca1cinea and Ca1caronea, evolved from rigid basal skeleton-bearing forms. The Murrayonida are probable ancestors of the modern Ca1cinea and the Minchinellida ofthe nonrigid modern Ca1caronea (Reitner 1987a). Interpreting the basal skeletons as a plesiomorphic feature of the Hadromerida clade, it is much easier to explain the common occurrence of ca1careous secondary skele- tons. Based on this theory, the modern nonca1cified forms are evolved from basal skeleton-bearing forms. This view is supported by the fossil record . Most fossil Hadromerida have only formed a basal skeleton, excepting Cliona. In this clade Cliona is related to Spirastrella as a sister group. Both have lost the ca1careous skeleton (15) and have a burrowing ability (16). All clionids burrow and are characterized by elongated spirasters (17) , a good autapomorphy of this taxon« Isolated asterose and tylostyle spicules wh ich are common in Mesozoic sedimentsr . 208 J. Reitner TheoryJ[ Fig. 15. Theory 11. Cladogram based on the monophyletic origin of the calcareous basal skeleton. Character numbers see text may be of hadromerid origin. But both spicule types occur also in other sponge groups. The main criticism about this theory is the different micro- and ultrastruc- tures as weIl as different mineralogies of the basal skeletons. To decide which theory is correct, we must know more about the ontogeny ofyoung sponges after settlement of the larva. No data are available at this tim.e from these particular sponges. Basal skeletons are definitely not a plesiomorphic character ofthe entire Demospongiae, hut single c1ades can evolve from ca1careous rigid skeleton bearing ancestors. A monophyletic origin of the ca1careous skeletons of the Hadromerida cannot be fully exc1uded, but this opinion must be proven by studies ofthe early ontogeny of these sponges (e.g., Acanthochaetetes). A further problem are the preservation potentials ofnonrigid sponges. We do not know much about "soft" demosponges in the past. Only isolated sc1eres are common and in most cases the entire spicula sponges skeleton is not known. It is very probable that "soft" Hadromerida have existed since early Mesozoie time. Based on this current knowledge a polyphyletic/convergent nature of the basal skeletons of the Hadromerida is most probable. Phylogenetic Aspects and New Descriptions of Spicule-Bearing Hadromerid Sponges 209 Conclusions a) The Hadromerida S.str. are a monophyletic sponge group in which three types of skeletal formations are observed: (1) primary skeleton composed of collagenous skeletal material and siliceous micro- and megasc1eres which can be reduced or lost; (2) secondary ca1careous basal skeletons; and (3) indirect basal skeleton formation by excavating burrows. b) Five different chaetetid skeletons are observed. Four ofthem are analogous to the primary sponge skeleton. Only one does not exhibit any similarities with the primary skeleton (Acanthochaetetes). In all cases the chaetetid skeletons can be explained as modified gemmule bodies in which totipotent archaeocytes are present. Besides the chaetetid skeletons, one crust type or modified stromatoporoid skeleton (Calcichondrilla), and one thalamid skeleton are observed (Cassianothalamia). c) Only three new asterose spicules-bearing taxa are described: Calästella tabulata n.gen. n.sp. from the Lower Cretaceous ofGreece, Chondrochaetetes longitubus n.gen. n.sp. from the Carboniferous ofRussia, and Calcichondrilla crustans n.gen. n.sp. from the Lower Cretaceous ofnorthern Spain. All three new genera are c1assified within the chondrosiidae. The chaetetid genus Acanthochaetetes is inc1uded as a subgenus within the genus Spirastrella based on the similarity of the spicular skeleton. One new species is described: Spirastrella (Acanthochaetetes) dendroformis n.sp. It is characterized by an internal stromatoporoid skeleton with radially oriented spicules and anormal chaetetid adult structure. d) The Cassianothalamiidae are probably linked with the Tethyidae. e) The fossil species Chaetetes (Boswellia) mortoni Gray and Chaetetopsis favrei (Deninger) are c1assified within the tylostyle-bearing Hadromeria (Suberitidae/Polymastiidae). f) Two phylogenetic theories in which the origin and significance of the different basal skeletons are discussed. A polyphyletic origin is, based on the current knowledge, probable. A monophyletic origin cannot be fully ex- c1uded, if the young adult sponges do not possess any ca1careous skeletons. Acknowledgmenls. I am indebted to Prof. J. Vacelet (Station Marine Endoume Marseille), Prof. H . Keupp, Dipl. Biol. F. Grothe (Freie Universität Berlin), Dr. R. van Soest (Zoological Institute Amsterdam), and Miss S. Stone, cu ra tor ofthe sponge section ofthe British Museum N.H. in London, for the permission to study specimens in their collections and some donations. I thank Dr. B. Gierlowski-Kordesch (Freie Universität Berlin) far translational assistance. The Deutsche Forsc- hungsgemeinschaft is acknowledged for financing this investigation (Re 6651 - 1). References Bergquist PR (1978) Sponges. Hutehinson, University ofCalifornia Press, pp 268 Boury-Esnault N (1987) The Polymastia species (Demosponges, Hadromerida) ofthe Atlantic Area. In: Vacelet J, Boury-Esnault N (eds) Taxonomy of Porifera. Nato Asi Ser G 13 , Springer, Berlin Heidelberg New York Tokyo, pp 29- 66 210 J . Reitner Cuif J-P, Feuillee P, Fischer J-C, Pascal A (1973) Presence d'Astrorhizes chez les Chaetetida Mesozoiques. CR Acad Sci Paris 277:2473- 2476 Deng Z (1982) Mesozoic Milleporina and tabulatomorphic corals from Xizang. Paleontol Xizang 4:184-188 Deninger K (1906) Einige neue Tabulaten und Hydrozoen aus Mesozoischen Ablagerungen . N Jahrb Min Geol Paläontoll :61-70 Döderlein L (1897) Über die Lithonina, eine neue Gruppe von Kalkschwämmen. Zoo I Jahrb 10: 15- 32 Fischer J-C (1970) Revision et essai de classification des Chaetetida (Cnidaria) post-paleozoiques. Ann Paleontol (lnvert) 56: 151-220 Fischer J-C, Lafuste J (1972) Nouvelles observations sur la paleohistologie du genre A canthochaetetes (Hydrozoa, Chaetetida). Bull Soc Geol Fr 14(7):320-324 Gray D (1980) Spicule pseudomorphs in a new Paleozoic chaetetid, and its sclerosponge affinities. Palaeontology 23 :803- 820 Hartman WD (1979) A new sclerosponge from the Bahamas and its relationship to Mesozoic stromatoporoids. CollInt CNRS 291 :467- 474 Hartman WD (1982) Porifera . In: Parker SP (ed) Synopsis and classification of living organisms. McGraw-Hill, New York, pp 640- 666 Hartman WD, Goreau TF (1970) Jamaican coralline sponges: their morphology, ecology, and fossil relatives. In: Fry WG (ed) The biology of Porifera. Symp Zool Soc Lond 25:205- 243 Hartman WD, Goreau TF (1972) Ceratoporella (Porifera: Sclerospongiae) and the chaetetid "corals". Trans Conn Acad Arts Sci 44:133-148 Hartman WD, Goreau TF (1975) A Pacific tabulate sponge, living representative ofa new order of sclerosponges. Postilla 167: 1-14 Kazmierczak J (1979) Sclerosponge nature of chaetetids evidenced by spiculated Chaetetopsis {avrei (Deninger 1906) from the Barremian ofCrimea. N Jahrb Geol Paläontol Mh 1979:98-108 Kirkpatrick R (1908) On two new genera ofrecent pharetronid sponges. Ann Mag Nat Hist 8:503- 514 Kirkpatrick R (1910) On a remarkable pharetronid sponge from Christmas Island. Proc R Soc 83: 124- 133 Kirkpatrick R (1911) On Merlia normani, a sponge with a siliceous and calcareous skeleton. QJ Microsc Sci 56:657- 702 Levi C (1973) Systematique de la classe Demospongiaria (demosponges). In: Grasse PP (ed) Traite de Zoologie lll(l) Masson, Paris, pp 577-631 Lister JJ (1900) Astrosclera willeyana, the type of a new family of sponges. In : Willey A (ed) Zool Results Part 4:461 - 482 Mori K (1976) A new sclerosponge from Ngargol, Palau Island and its fossil relatives. Tohoku Univ Sci Rep 2 Ser Geo146: 1- 9 Mori K (1977) A calcitic sclerosponge from the Ishigaki-shima coast, R yukyu Islands, Japan. Tohoku Univ Sci Rep 2 Ser Geo147: 1- 5 Peterhans E (1930) Etude du genre Chaetetopsis Neumayr et classification nouvelle des Chaetetides. Eclogae Geol Helv 22:81 - 85 Reitner J (1982) Die Entwicklung von Inselplattformen und Diapir-Atollen im Alb des Basko- Kantabrikums (Nordspanien). N Jahrb Geol Paläontol Abh 165:87- 101 Reitner J (1987a) Phylogenie und Konvergenzen bei rezenten und fossilen Calcarea (Porifera) mit einem kalkigen Basalskelett ("Inozoa, Pharetronida"). Berl Geowiss Abh 86:87- 125 Reitner J (1987b) A new calcitic sphinctozoan sponge belonging to the Demospongiae from the Cassian Formation (LowerCamian; Dolomites, Northern Italy) and its phylogenetic relationship. Geobios 20:571 - 589 Reitner J (I 987c) Euzkadiel/aerenoensisn.gen. n.sp. ein Stromatopore mitspikulärem Skelett aus dem Oberapt von Ere.o (Prov Guipuzcoa, Nordspanien) und die systematische Stellung der Stromatoporen. Paläontol Z 61 :203-222 Reitner J (l987d) Mikrofazielle, palökologische und paläogeographische Analyse ausgewählter Vorkommen ftachmariner Karbonate im Basko-Kantabrischen Strike Slip Fault-Becken-System (Nordspanien) an der Wende von der Unterkreide zur Oberkreide. Doc Nat 40: 1- 248 Reitner J (1989) Lower and Mid-Cretaceous Coralline Sponge Communities of the Boreal and Tethyan Realms in Comparison with the Modern Ones - Palaeoecological and Palaeogeographic Phylogenetic Aspects and New Descriptions ofSpicule-Bearing Hadromerid Sponges 21! Implications. - In: Wiedmann J (ed) Cretaceous of the Western Tethys. Proceedings 3rd International Cretaceous Symposium, Tübingen 1987, pp 851-878. E Schweizerbart'sche Ver lagsbuchhandlung, Stuttgart • Reitner J (l989b) Struktur, Bildung und Diagenese der Basalskelette bei rezenten Pharetronidert unter besonderer Berücksichtigung von petrobiona massiliana Vacelet & Levi 1958 (Minchinellida, Porifera). Berl Geowiss Abh (A) 106 :343-384 Reitner J, Engeser TS (1983) Contributions to the systematics and the paleoecology of the family Acanthochaetetidae (Fischer 1970) Order Tabulospongida, Class Sclerospongiae. Geobios 16 :773-779 Reitner 1, Engeser TS (1985) Revision der Demospongier mit einem thalamiden, aragonitischen Basalskelett und trabekulärer Internstruktur("Sphinctozoa" pars). Berl Geowiss Abh 60 : 151-193 Reitner J, Engeser TS (1987) Skeletal structures and habitats of Recent and fossil Acanthochaetetes (subclass Tetractinomropha, Demospongiae, Porifera). Coral Reefs 6: 151 - 157 Reitner 1, Grothe F (1988) Origin of the basal skeleton of Petrobiona massiliana Vacelet (Calcaronea, "Pharetronida"). Berl Geowiss Abh 100:29 Reitner 1, Keupp H (1989) Basalskelette bei Schwämmen - Beispiel einer polyphyletfschen Entwicklung. Geowissenschaften 1989:71 -78 Reitner 1, Cuffey RJ, Scott R W Albian coralline sponge, Caribbean Province, Arizona. N Jahrb Geol Paläont Mh (in press) Rios 1M, Almela YA (1944) Un Chaetetido dei Eoceno espaftol. Notas Comun Inst Geol Min Esp 12: 19-37 Rützler K (1974) The burrowing sponges of Bermuda. Smithsonian Contrib Zool 165 :203-2 16 Vacelet 1 (1964) Etude monographique de I'eponge calcaire Pharetronide de Mediterranee, Pe. trobiona massiliana Vacelet & Levi. Les Pharetronides actuelles et fossiles. Recl Trav St Mar Endoume 50:1-125 Vacelet J (1980) Squelette calcaire facultatif et corps de regeneration dans le genre Merlia , eponges apparentees aux Chaetetides fossiles. CR Hebd Seances Acad Sci Paris 290:227- 230 Vacelet 1 (1981) Eponges hypercalcifiees ("Pharetronides", "Sclerosponges") des cavites recifs coralliens de Nouvelle-Caledonie. Bull Mus Natl Hist Paris 4 Ser 31981 sec a 2:313- 351 Vacelet 1 (1985) Coralline sponges and the evol~tion of Porifera. In : Conway Morris S, George JD, Gibson R, Platt HM (eds) The origins and relationships oflower invertebrates. Syst Assoc Spec VoI28 :1-13 Vacelet 1 (1988) The storage cells of calcified relict sponges. In : Proceedings Third Internat conf Sponge Biol. Smithsonian Inst Press, Washington, DC Vacelet J, Garrone R (1985) Two distinct populations of collagen fibrils in a "sclerosponge" (Porifera). In : Bairati A, Garrone R (eds) Biology ofinvertebrate and lower vertebrate collagens. NATO ASI Ser ser A: Life Sciences 93 . Springer, Berlin Heidelberg New York, pp 183- 189 Van Soest RWM (1984) Deficient Merlia normani Kirkpatrick, 1908, from the Curacao reefs, with a discussion on the phylogenetic interpretation of sclerosponges. Bijdr Dierkd 54:211- 219 Van Soest RWM (1987) Phylogenetic exercises with monophyletic groups of sponges. In : Vacelet 1, Boury-Esnault N (eds) Taxonomy ofPorifera. NATO ASI Ser, serG: Ecol Sci 13. Springer, Berlin Heidelberg New York, pp 227-241 Wiedenmayer F (1977) Shallow-water sponges of the western Bahamas. Exper Suppl 28:287 Wood R (1987) Biology and revised systematics of some Late Mesozoic stromatoporoids. Spec Pap Paleontol37: 1- 89 Wood R, Reitner 1 (1988) The "chaetetid" demosponge Stromatoaxinella irregularis (Michelin) and its systematic implications. N Jahrb Geol Paläontol Mh 1988:213- 224 Wood R, Reitner 1, West R (1989) Systematics and phylogenetic implications of the haplosclerid stromatoporoid Newellia mira nov. gen. Lethaia 22