Following rediscovery of the sclerosponges (Hartman 1969, 1979; Hartman and Goreau, 1970, 1975) and recognition of their similarities to the stromatoporoids (Hartman and Goreau, 1970, p. 221–228), classification and interpretation of fossil Porifera have undergone major revisions during the past ten years. Our interpretation of the “edges” of the phylum have been considerably expanded.

Demospongia have persisted as a major class of sponges since inception of the record of the phylum in the Cambrian. The class includes about 95% of all sponges in modern seas and all of the known freshwater forms, as well. These sponges occupy environments that range from warm, shallow subtidal, high energy to quiet, cold oceanic deeps. Demosponges range from exposed tide pool forms to cryptic dwellers and include the only boring sponges.

Calcareous sponges have a skeleton formed of spicules of calcium carbonate laid down as calcite. These spicules may be free in the sponge tissue, or may be in part fused together by secondary deposits of calcite and in part free, or may occur as free elements together with a reticulate or massive skeleton of crystalline calcite. The spicules are monactinal, diactinal, triactinal and tetractinal and do not in general occur in categories marked by distinct size differences as is true of the megascleres and microscleres of demosponges and hexactinellids.

When sponges were well-behaved, a calcareous skeleton was sufficient for assignment to the Class Calcarea. Since the discovery that some sponges secrete both siliceous spicules and a massive calcareous skeleton (Class Sclerospongia Hartman and Goreau, 1970) and more recently, the discovery of an apparent living Sphinctozoan with an aspicular calcareous skeleton and soft parts like those of a demosponge (Vacelet, 1977b) the picture has become more clouded. The groups to be discussed here originally included all the known sponges with massive calcareous material in the skeleton. They were named Pharetrones by Zittel (1878) who emphasized the skeleton of anastomosing fibers, containing, so he thought, spicules packed side-by-side like arrows in a quiver (pharetra is Greek for “quiver”). Steinmann (1882) noted that in some of the genera included by Zittel as well as in new forms which he described, the skeleton was sheet-like and outlined a series of hollow chambers. Steinmann set apart the chambered types as a subgroup Sphinctozoa(=“constricted animals”) and the fibrous types as the subgroup Inozoa(=“fiber animals”). Steinmann also pointed out that the skeleton of the Sphinctozoa was aspicular and spherulitic. Zittel had illustrated the spherulites but considered them diagenetic. Steinmann, however, observed that the spherulites were bored by endolithic algae or fungi, and were therefore probably original. He also noted that the spicules in some Inozoa were sinuous and unlike typical sponge spicules, and indeed considered the pharetronids not to be sponges at all. Hinde (1884) and others soon demonstrated that the fibers of many pharetronids, as well as the external cortical layer (including that of the sphinctozoan Tremacystia) contained typical triradiates of the Class Calcarea. The discovery of the first living pharetronids (Doederlein, 1892, Kirkpatrick, 1900, 1908) with soft-parts and flesh-spicules typical of the Calcarea, seemed to settle the matter. Rauff (1913) introduced the useful term sclerosome for the non-spicular material of the skeleton, which might range from simple cement between spicules to the entire skeleton. It is this material that recent studies by Wendt (1979, 1980), Cuif and others (1979), and Jones (1979) have shown to be quite heterogeneous and variable. We will now look at the morphology and fossil history of the Inozoa and Sphinctozoa without any commitment as to the biological unity of each group.

The Heteractinida are a relatively minor group of calcareous sponges, although Astraeospongium, one of the currently 14 known genera in the order, is one of the most commonly studied fossil sponges in the world. Principle record of the Heteractinida is from North America and Europe, although they have now been reported from Australia, Africa, Asia and South America. Occurrences in the latter areas generally are only isolated spicules or single specimens.

Hexactinellid sponges have a skeleton made up basically of six-rayed spicules occurring individually or fused together by supplementary secretions of silicon dioxide to form a rigid latticelike or reticulate skeleton. Microscleres of one of two basic types are always present. The axial canal of hexactinellid spicules is square in cross section.

The structure and evolution of the hexactinellid skeleton has been conditioned by: (1) the rectangular form of the spicule (ultimately due to the square cross-section of the axial canal), and (2) the sheet-like form of the choanocyte-membrane, properly the choano-syncytium (Reiswig, 1979) as individual choanocytes with proper nuclei do not exist, which is a continuous syncytial sheet bearing outpocketings (diverticula) that correspond to the choanocyte chambers of other sponge groups. In most species the diverticula are separate, and attached to one another by syncytial filaments, but their arrangement as part of a sheet-like structure is still apparent. In a few species the choanocyte membrane becomes a labyrinth of anastomosing tubes. Accompanying these are the essential absence of mesogloea and the paucity of collagen (spongin), which results in the soft parts being essentially confined to the space occupied by the skeleton. The place of mesogloea is taken by a syncytial network of filaments which also form the dermal and gastral membranes.

Sclerosponges secrete a basal crystalline and aspicular skeleton of calcium carbonate, either aragonite or calcite, above which lies a thin layer of living tissue which also secretes siliceous spicules and collagenous fibers. The tripartite skeleton of sclerosponges distinguishes them from all other sponges and also from all other multicellular animals, no one of which has an abundant quantity of two disparate minerals helping to make up its skeleton. The cell types and their organization as well as what little is known about their development indicate that the sclerosponges are related to the demosponges. Sclerosponges are inhabitants of shaded crevices, caves and tunnels on coral reefs in both the Caribbean and Indo-Pacific regions. The range of only one species, Merlia normani Kirkpatrick, extends from the tropics into the warm temperate waters of the Mediterranean Sea.

Chaetetids have recently emerged from relative obscurity on two fronts. On one, they have achieved a certain notoriety in their reclassification from tabulate corals to a recently erected class of sponges; on the other, they are becoming appreciated as locally significant faunal elements and bioherm framebuilders.

This summer in Washington one of the sessions of the Fossil Coral Symposium was devoted to the subject of stromatoporoids and their relatives; this fall a major part of this workshop on fossil sponges is being devoted to the same group. The closeness of these two meetings emphasizes the problematic nature of these fossils and focuses attention on a controversy over their affinity that is over 100 years old. Although these enigmatic fossils have many cnidarian features, I think you will agree that they are “in the right place” in this sponge workshop.

There are two levels at which the internal structure of a stromatoporoid is classified. The first level includes structures which can be discerned in a thin section using low magnifications (10X); these are called macrostructures. The second level is composed of structures that are not evident unless a thin section is magnified to a higher degree (50X-100X). These smaller scale aspects are referred to as microstructures.

The most influential early classification of the group appeared in H.A. Nicholson's great monograph on the stromatoporoids for the Palaeontographical Society (1886–1892). He believed that the stromatoporoids were similar to the two major groups of living hydrozoans and based his major divisions, the Hydractinoidea and the Milleporoidea, on this similarity. He divided the first of the two groups into the families Actinostromatidae (distinct laminae and pillars) and Idiostromatidae (dendritic forms), and the second group into the Stromatoporidae (amalgamate structures of coenosteles and coenostroms) and the Labechiidae (horizontal elements largely dissepiments). These basic divisions have continued to be recognized in later classifications but additional groups have been split off from Nicholson's original four and his bases of classification have been changed. Lecompte (1956) in the Treatise volume recognized, in addition, the Clathrodictyidae for genera with laminar subvesicular structure and compact tissue; the Stromatoporellidae for genera with dominantly laminar structure and porous tissue, and the Syringostromatidae for genera with well differentiated laminae and pillars and “vertically tabulate chambers”. He also continued a practice started by Dehorne (1920) of including Mesozoic stromatoporoids in Paleozoic families. He recognized the families Milleporellidae and Milleporidiidae for exclusively Mesozoic stromatoporoids. Most later classifications have ignored the Mesozoic stromatoporoids and grouped only Paleozoic genera. Recent works on the classification of Paleozoic stromatoporoids include those of Galloway (1957), Bogoyavlenskaya (1969), Khalfina and Yavorsky (1973), and Stearn (1980).

Controversy concerning the affinity of the stromatoporoids is not new; it began over 100 years ago with the first systematic studies of the group. Rosen's (1867) opinion was that these fossils were allied to the sponges. Nicholson and Murie (1878) agreed. Solomko (1885) introduced some of the arguments still used by the advocates of the poriferan affinity of the stromatoporoids such as the recognition of the astrorhizae as excurrent canal systems. H.A. Nicholson who had at first supported the hypothesis of sponge affinity became convinced that the closest living representatives were to be found in the hydrozoans. He compared the skeletons of the stromatoporoids like Actinostroma to the modern hydrozoan Hydractinia and the skeleton of amalgamate genera like Stromatopora to that of Millepora. Nicholson's monograph of British stromatoporoids (1886–92) and the papers he published in the Annals and Magazine of Natural History to accompany his revision of the group form the basis of all later work and hence his views on the affinity of stromatoporoids had immense influence.

The great majority of the published work on the paleontology of the stromatoporoids has appeared since 1950. This influx of information has made possible a fairly clear picture of their evolution. As mentioned earlier by Stearn, stromatoporoids were quite abundant in shallow marine environments from the Middle Ordovician through the Late Devonian. These are referred to as the “Paleozoic stromatoporoids.” The record of Carboniferous stromatoporoids is sparse, but they began to come back in the Permian, and were common again through much of the Mesozoic, until disappearing at the the end of the Cretaceous. The Permian through Cretaceous forms shall be referred to as the “Mesozoic stromatoporoids.”

For the non-specialist in stromatoporoid paleobiology, perhaps the greatest interest in these enigmatic organisms is in their potential value in paleoecology; Stearn (1982a) has noted the general preference of stromatoporoids for reef environments as a unifying aspect of the group. Stromatoporoid coenostea are usually obvious in the field but care must be taken not to confuse them with non-skeletal and skeletal algae, flatpebble conglomerate clasts or finely laminated carbonate sediment.

The Archaeocyatha were marine organisms developing mineral skeletons and using calcium carbonate for this purpose. Remains of their cups are found in carbonate shelfs and reef environments of the Early Cambrian seas. Few representatives of the family Archaeocyathidae are found through the Middle Cambrian to the Upper Cambrian (Debrenne, Rozanov and Webers, in press).

Archaeocyathans were shallow marine benthic filter-feeding organisms. They preferred carbonate-rich waters and typically lived in aggregates or communities which were dominated by calcareous algae or algae-like problematica. In addition, a wide range of metazoans either occupied these communities or lived in great abundance nearby. It is now clear that these communities or aggregates were capable of constructing carbonate buildups which were topographically and sedimentologically significant. The Archaeocyatha evolved for some 45 to 50 million years and so did the content and structural complexity of their buildups. This complexity reached its acme during the later part of the Lower Cambrian (late Atdabanian and Lenian of Soviet and European workers). Let us now shift from generalities and broad interpretations to a review of the supportive data.