Sponge detective: when faunal lists go bad

I set out to do something simple, really I did. All I wanted to write was an introduction to sponges and a quick description of the forms known from the Twin Cities region. I already had a list of appropriate species, and I knew that most of the original forms weren't actually sponges, which I thought would make things easier. "There's only a couple left, that's not too bad." Then I made the mistake of checking into those leftovers. It turns out that you can never assume a classification for early Paleozoic sponge-like things. There's always room for an argument. In paleontology, the answer to any question always includes "start digging," whether it be rocks or research, and, frankly, isn't some mystery more interesting than a list?

Background

Sponges have never received much respect from humans. This does not seem to bother them, of course, sponges not being noted for the complexity of their nervous systems. Sponges belong to the phylum Porifera. They are multicellular but do not have organs, tissues, or more than a vague idea of symmetry. The basic sponge lifestyle is as a stationary aquatic filter feeder, drawing water through pores and canals and extracting organic particles and tiny organisms. Lest you assume that a sponge is just a bag of cells that hang around together because they have nothing better to do, sponges actually contain a number of different types of cells, which can move throughout the sponge and take on different roles. Their bodies often include microscopic hard structural elements called spicules, which have a variety of simple shapes from rods to jacks or caltrops. These spicules may be composed of calcium carbonate (i.e. calcite, or lime) or silica. Many forms with silica spicules also have structural fibers composed of the protein spongin. These different structures have been used to sort sponges into several classes, traditionally Calcarea (sponges with calcitic spicules), Hexactinellida (sponges with silica spicules but not spongin; the glass sponges), and Demospongiae (sponges often with both silica spicules and spongin, but sometimes just spongin). Sponges have proved to be not quite this convenient and apparently the "demosponges" include two distinct lineages. Demosponges are the most abundant and diverse of the three, but have the poorest fossil record because of their relative lack of hard parts; when they die, their non-mineralized bodies disintegrate and whatever spicules were included disperse. This does mean that sponge spicules can be found as microfossils, though. Although sponges have a reputation for being, well, "spongy", the Calcarea and Hexactinellida are made of stern material (a glass sponge like our old friend Nunavutospongia could give you a drastic exfoliation if you tried to use it in the bath). The bath sponges are demosponges without extensive spicule networks (so yes, natural sponges are in fact dead sea critters).

There are also some extinct probable sponges, like the archaeocyathids of the Cambrian. The reef-building stromatoporoids of the Paleozoic (not to be confused with stromatolites) are also now regarded as sponges of some stripe. The receptaculitids, known informally as "sunflower corals", have generally alternated between being classified as sponges and calcareous algae over the years, with it being the algae's turn in recent history, but may well be neither. They are perhaps the characteristic fossil of the lower Cummingsville Formation, the next unit above the Decorah Shale. It's not really a surprise, but when dealing with organisms that don't necessarily have a set structure or an easily fossilized body, it can be difficult to tell how to classify a given fossil. Many mistakes have been made and will be made.

The suspects

Stauffer and Thiel (1941), in their work on the Paleozoic of Minnesota, included a section (starting p. 227) listing all of the fossil species known for the various formations. We don't call all of the formations by the same names, and many of the species are under different genera and sometimes are classified with much different groups, but their work makes a useful starting point to see what's out there if you're willing to put in the time to check it before you do anything important with it. If you need to refresh your memory concerning the rocks, I have a text overview, diagrams, a link to the most recent geologic map, and a Platteville-specific discussion. This post gives a thumbnail of the fossil groups. Some chewy stratigraphic exposition follows, and if you don't plan to check out the original document you may want to skip the rest of the paragraph. For the purposes of this discussion, we're looking at essentially the St. Peter, Glenwood, Platteville, Decorah, and lowermost Cummingsville formations. S&T included equivalents to all of these units in their list, some rougher than others. The St. Peter Sandstone is unchanged. The Glenwood Formation as we know it was described as the Glenwood Member of the Platteville Formation. The four beloved members of the modern Platteville Formation were lumped as the McGregor Member, although you can usually guess at the modern divisions in their stratigraphic sections based on the descriptions. The modern Carimona Member of the Decorah Shale was described as the Spechts Ferry Member of the Platteville Limestone, and the shaly part of the Decorah Shale was the Decorah Shale Member of the Galena Formation. The only really tricky unit is the modern Cummingsville, which apparently gave Stauffer problems, and was included in the Prosser Member of the Galena Formation along with what we would now called the Prosser Formation and some bits of the Decorah Shale and overlying Stewartville Formation. See Sloan (1987a, 1987b) and Mossler (2008) if you have an urge to trace these usages.

Stauffer and Thiel listed a handful of fossil sponges from the relevant formations, but most of them can be recognized as other things once you've acquired some experience with the nomenclature. The not-sponges include receptaculitids (Ischadites iowensis, Receptaculites oweni) and burrow casts (Rauffella filosa, R. fucoida, and R. palmipes; there was also a fourth species by that time, R. ulrichana, but we can forgive Stauffer and Thiel for missing it because apparently the only people who ever heard of it are Frederick Sardeson, who named it in 1925 without providing a type specimen or locality, Malcolm Weiss, who wrote his biography, and me). Pulling out these imitators leaves four species in three genera across four geologic units: Camarocladia rugosa, Cylindrocoelia minnesotensis, Hindia inaequalis, and Hindia parva. The record doesn't expand greatly if we invite in the other formations known from Minnesota; all that's added is the demosponge Heterospongia subramosa from the "Wykoff Member of the Maquoketa Formation" (modern Maquoketa Formation) and a couple of stromatoporoid sponges (Idiostroma and Stromatoporella erratica) from the Devonian, then thought to be "hydrozoans". (By the way, it's not worth your time to try to parse out the modern Devonian formations from Stauffer and Thiel; they had one, we have six.) So, what kind of sponges (or not) were these?

Camarocladia rugosa

Camarocladia rugosa was named by Edward Oscar Ulrich in a footnote on page xcv of The Geology of Minnesota 3(2) (1897), as another species of the genus Camarocladia, described from Illinois. Several species have been attached to this genus over the years, but they're all basically variants on these flattened branching things; C. rugosa happens to be a large variety. Stauffer and Thiel (1941) report C. rugosa only in the Decorah Shale. In case you're wondering, no, they don't let you name things in footnotes any more, no matter how authoritative the surrounding monograph is. Its humble origin is why it did not receive coverage in one of the lavish plates, so we are left to other sources to figure out what it looks like. Sardeson (1925) included a line drawing, reproduced a couple of paragraphs below (forgive the curves; it was scanned out of a thick volume), and photographs appeared in his folio of the Twin Cities (1916; as "fucoids") and a Maryland Geological Survey monograph (1919). The photos are quite telling; I realized I'd seen things like them before, but I always thought they were burrows.

Camarocladia rugosa bed from the Decorah Shale of St. Paul (Maryland Geological Survey 1919).

Ulrich's footnote described it as a branching "fucoid"-like natural cast of a sponge, usually with flattened furrowed branches showing no structure, and an internal framework of sub-cylindrical segmented rods and "oblique septa-like partitions". It's not entirely obvious what he's trying to describe, and I don't see anything like that in the photographs, but I digress. Both he and Sardeson (1925) identify it as the namesake "fucoid" of Sardeson's "Fucoid Bed" in the Decorah Shale (aka Sardeson Bed 5) as pictured in Sardeson's folio. Writing from his exile to the Pan-American Geologist* in 1925, Sardeson could find no evidence that Camarocladia rugosa was a sponge, and instead compared it to his Rauffella? fucoida from 1896. Sardeson was of the old school that saw trace fossils as evidence of seaweed and algae, and had the curious idea that the structures were algal roots and holdfasts that had been infiltrated by worms, stuffed like sausages (he'd been pitching this idea since at least 1913). Although this makes an amusing mental picture, it is difficult with hindsight to see why he didn't just skip the stuffing and flat-out call them burrows. I can't judge him too harshly, though. (You can if you want, but it wouldn't reflect well.) To him, both C. rugosa and Rauffella? fucoida were examples of these stuffed algal structures, and could even be synonyms. [2019/02/20: finessed the discussion of R.? fucoida after re-reading the article.]

[*In the 1920s, there were only a few options for publishing geological papers, including the USGS, state surveys, the Journal of Geology, the Geological Society of America Bulletin, and the American Journal of Science. If for some reason you couldn't go through these channels, for example because of personality conflicts, you published in the Pan-American Geologist.]

Sardeson (1925); 1 and 2 are Camarocladia rugosa (I assume from the text, given there is no formal explanation of the plate), 3 is Hindia parva, 4 is burrow Rauffella filosa, 5 and 6 are Rauffella ulrichana, 7 is Rauffella palmipes, and 8 is the true algae Chaetocladus sardesoni.

Rauffella? fucoida from Sardeson (1896); this is looking familiar.

Rigby and Nitecki (1968) in passing (p. 113) wrote that Camarocladia should probably be considered a trace fossil and not a sponge, which Rigby asserted again in 1976. If J. Keith Rigby, who saw more fossil sponges than probably anyone, thought Camarocladia is a trace fossil, and Frederick Sardeson, who saw more Minnesota fossils than probably anyone, thought roughly the same thing (if we cut out the algal middleman from his burrow sausages), and all of the photos and illustrations look like trace fossils, I can't help but figure that C. rugosa *is* a trace fossil, and obscurity and inertia are responsible for any lingering sponge identifications. For confirmation, I checked the most recent version of the Treatise on Invertebrate Paleontology concerning sponges (Fink et al. 2004), and Camarocladia was listed among the taxa formerly thought to be sponges, as a trace fossil. It occurs to me that I could have saved a lot of time if I'd thought of the Treatise sooner, but then I wouldn't have seen everything else. Also, it wouldn't have helped in the next case.

Camarocladia bed in Sardeson (1916). Depending on who you read, this would be a mass-death assemblage of small branching sponges, a veritable forest of seaweed holdfasts later filled in by worms, or a bed that was heavily burrowed by worms.

So much for Camarocladia rugosa. How about Cylindocoelia minnesotensis?

Cylindrocoelia minnesotensis

Ulrich named this genus and species back in 1889 along with a number of other sponges and spongy things. In fact, he came up with four species of the genus, the other three being from the Cincinnati beds of Kentucky. They essentially all differed by size and how cylindrical they were. Ulrich neglected to illustrate them, but Winchell and Schuchert (1895) did include figures for C. minnesotensis a few years later. These are the only figures I've been able to dig up for any Cylindrocoelia species. His C. minnesotensis specimens came from the Decorah Shale (then known as the Trenton Shale and thought to be Silurian in age) and were a few cm long and about 1 to 1.5 cm across, with very little tapering and pores 1 to 2.5 mm across. Apparently there were fewer and more irregularly distributed pores on this species than the others, but in the absence of illustrations it's hard to get a feel for what that means. Stauffer and Thiel (1941) also reported it from the "Spechts Ferry", aka the Carimona Member.

From Winchell and Schuchert (1895), Plate G: Cylindrocoelia is on the left, figures 1 to 3. 1 is a specimen partially encrusted by the inevitable bryozoan, at natural size (in the original plate). 2 is a larger fragment, and 3 is its end view, again at natural size (for the original size of the plate), from the upper part of the "middle third of the Trenton shales" (=Decorah) of St. Paul. Figures 4 through 6 are a bonus, Heterospongia from Spring Valley and Cincinnati.

How has history treated Cylindrocoelia? Not very well. Go ahead, think of a famous fossil sponge. You can't, because there aren't any, but that's not the point. You certainly didn't think of Cylindrocoelia. I thought there was little out there on Camarocladia, but it was positively attracting paparazzi compared to this genus; at least Camarocladia gets invoked sometimes when there is one of those Medusa-like beds. There is also a hidden handicap to using search engines to find information on Cylindrocoelia: note the "oe". Revisions to the taxonomic code have proscribed the various characters that have accents and so forth, but in the 19th century there were plenty of names with hyphens and diacritics and œ running around. This genus was originally spelled Cylindrocœlia, which gives search engines fits.

The Cincinnati folks identify their Cylindrocoelia covingtonensis as an algae, which doesn't help all that much. There are a million things that are called, have been called, or will be called algae, and many of them are only guilty by lifestyle. This isn't that big of an issue, though. The frustrating bit is that it's not apparent how they came to that conclusion. The source mentioned at the above link, Dalvé (1948), doesn't list it as an algae, just as of "uncertain position" with a few other malcontents, not all of which are algae. Instead, that citation is probably for the stratigraphy. The Treatise is also no help here; the most recent version doesn't include it at all, even as a dubious sponge or non-sponge.

Sardeson (1925) turned his baleful gaze on this genus, like he did with Camarocladia, and was again not impressed, interpreting C. minnesotensis as the bored and eroded siphuncles of straight-shelled nautiloids like Endoceras. Hilariously, Ulrich in his original description of type species C. endoceroidea wrote "A careless observer might mistake this sponge for the shell of an Endoceras or perhaps Orthoceras, to which the form and obscure annulations give it some resemblance. However, unusually acute powers of discrimination are not required to see that it is a very different organism." Take that, I guess. Anyway, looking at the few figures, I can squint as well as the next man. The irregularity in the size and distribution of the "pores" leads me to suspect that they are indeed borings or other biological modifications representing the attentions of more than one type of critter. In fact, I have seen coprolites covered with a similar array of irregularly positioned burrows of varying size; check out Figure 6b here. Whether or not the base structure for Cylindrocoelia minnesotensis is a chunk of nautiloid, I can't say without seeing the original fossils. I suspect that Sardeson was on the right track, though. The real mystery to me is the algal identification of C. covingtonensis, which is something I'd love to know the reasoning behind, and whether it would also be applicable to C. minnesotensis. In either case, I think it's safe to regard Cylindrocoelia minnesotensis as outside of the sponges.

Hindia spp.

This leaves us with one lonely genus of true sponge from the Cummingsville to the Platteville: Hindia (trust me, S&T do much better with things that aren't sponges, with most of the changes being due to differences in classification and not fossils of mistaken identity). This genus is a demosponge with spicules of silica, and has even got its own family of sponges, Hindiidae, which persisted into the Permian and achieved worldwide distribution. Hindia fossils look something like a breakfast cereal, round things a centimeter or less in diameter. Well-preserved specimens look more obviously like something biological, with a porous-looking (dare I say "spongy"?) surface, but at this size you'll need to be looking closely. Stauffer and Thiel recognized two species, with Hindia parva present in their Prosser and Decorah Shale members, and H. inaequalis present in their Spechts Ferry and McGregor (our Platteville) members. The two species differ in that one is spelled parva and one is spelled inaequalis. It's right on the label; how did you think we knew what to call them? To be more serious, H. inaequalis is apparently somewhat larger and has "radiating canals of unequal size", while H. parva has canals smaller than 0.27 millimeters in diameter (Ulrich 1889). If you've got one where the canals are not preserved well enough to tell, or you left your ruler with the sub-millimeter ticks marks in the car, just call it Hindia. Something turning out to be an actual sponge seems a little disappointing after all of the above, especially when its fossils are just tiny balls and their idea of fun is to have differently-sized canals, but not everything can be equally exciting. I imaging in life they were little bulbs attached to the sea floor.

Behold, the one and lonely Twin Cities fossil sponge, Hindia. These specimens are in the collection of the University of Minnesota. They came from the Brickyard at Lilydale, and were part of the Adamson thesis (1993) collection.

More of Plate G, Winchell and Schuchert (1895). 7 is three differently sized Hindia specimens, although you'd be forgiven for mistaking it for an artistic exercise in shadows on spheres. 8 is a vertical section and 9 is a tangential section, both at 10x. These are from the "upper third of the Trenton shales and Galena shales, Goodhue and Fillmore counties".

And now you are far more educated concerning sponges that you were before. Go forth and use this knowledge wisely!

References

Adamson, K. F. 1993. A numerical analysis of the fauna of the upper quarter of the middle Ordovician Decorah Shale, St. Paul, Minnesota. Senior thesis. University of Minnesota, Minneapolis, MN.

Dalvé, E. 1948. The fossil fauna of the Ordovician in the Cincinnatian region. University Museum, Department of Geology and Geography, University of Cincinnati, Cincinnati, OH.

Finks, R. M., R. E. H. Reid, and J. K. Rigby (coordinating author). 2004. Porifera (Demospongea, Hexactinellida, Heteractinida, Calcarea). Treatise on Invertebrate Paleontology. Part E (Porifera) Revised, volume 3. Geological Society of America, Boulder, CO, and University of Kansas, Lawrence, KS.

Maryland Geological Survey. 1919. Cambrian and Ordovician. Maryland Geological Survey systematic report 7. Johns Hopkins Press, Baltimore, MD.

Mossler, J. H. 2008. Paleozoic stratigraphic nomenclature for Minnesota. Minnesota Geological Survey, St. Paul, MN. Report of Investigations 65.

Rigby, J. K. 1976. Some observations on occurrences of Cambrian Porifera in western North America and their evolution. Geology Studies (Brigham Young University) 23(2):51–60.

Rigby, J. K., and M. H. Nitecki. 1968. Annotated bibliography of Lower Paleozoic sponges of North America. Fieldiana: Geology 18(1).

Sardeson, F. W. 1896. The Saint Peter Sandstone. Bulletin of the Minnesota Academy of Natural Sciences 4(1):64–88.

Sardeson, F. W. 1913. Characteristics of a corrosion conglomerate. Bulletin of the Geological Society of America 25:265–276.

Sardeson, F. W. 1916. Minneapolis–St. Paul folio. U.S. Geological Survey, Washington, D.C. Folio 201.

Sardeson, F. W. 1925. Ordovicic kelp, sponges and sea worms in Minnesota. Pan-American Geologist 43(4):271–286. [I have a scan]

Sloan, R. E. 1987a. History of study of the Middle and Late Ordovician rocks of the Upper Mississippi Valley. Pages 3–6 in R. E. Sloan, editor. Middle and Late Ordovician lithostratigraphy and biostratigraphy of the Upper Mississippi Valley. Minnesota Geological Survey, St. Paul, MN. Report of Investigations 35.

Sloan, R. E. 1987b. Tectonics, biostratigraphy and lithostratigraphy of the Middle and Late Ordovician of the Upper Mississippi Valley. Pages 7–20 in R. E. Sloan, editor. Middle and Late Ordovician lithostratigraphy and biostratigraphy of the Upper Mississippi Valley. Minnesota Geological Survey, St. Paul, MN. Report of Investigations 35.

Stauffer, C. R., and G. A. Thiel. 1941. The Paleozoic and related rocks of southeastern Minnesota. Minnesota Geological Survey, St. Paul, MN. Bulletin 29.

Ulrich, E. O. 1889. Preliminary description of new Lower Silurian sponges. The American Geologist 3:233–248.

Ulrich, E., W. Scofield, J. Clarke, and N. H. Winchell. The geology of Minnesota. Minnesota Geological and Natural History Survey, Final Report 3(2). Johnson, Smith & Harrison, state printers, Minneapolis, MN.

Winchell, N. H., and C. Schuchert. 1895. Sponges, graptolites, and corals from the Lower Silurian in Minnesota. Pages 55–95 in Lesquereux, L., C. Schuchert, A. Woodward, E. Ulrich, B. Thomas, and N. H. Winchell. The geology of Minnesota. Minnesota Geological and Natural History Survey, Final Report 3(1). Johnson, Smith & Harrison, state printers, Minneapolis, MN.