Steve Kershaw

Institute for the Environment, Halsbury Building, Brunel University, Uxbridge, Middlesex, UB8 3PH, UK

1 Introduction and aims

This discussion paper is intended to stimulate ideas and debate amongst palaeontologists interested in stromatoporoids and their applications.The following points summarise key aspects of the current status of knowledge of stromatoporoids, followed by the aims of this paper.

1)Since the original description of stromatoporoids by Goldfuss (1826)a lot of work has been done on their taxonomy; the culmination of 180 years of study lies in the recent chapters of the revisedTreatise on Paleontology, available in Treatise Online, in 2010 to 2012 (see reference list for selected references).In recent years it has become clear that taxa based on the calcareous skeleton have been over-split and a more conservative view of stromatoporoid taxonomy is appropriate.Furthermore,there is recognition that Palaeozoic stromatoporoid taxonomy must use the calcareous skeleton because spicules used in modern, and many younger calcified sponges,are almost completely absent from Palaeozoic stromatoporoids (Stearn, 2010a).Nevertheless, there is general agreement that Palaeozoic stromatoporoids are calcified sponges, based on a range of features of the calcareous skeleton (Stearn, 2010b).

2)In contrast to extensive taxonomic work, only a few detailed studies of stromatoporoid ecology exist, although much general knowledge of stromatoporoid growth forms and the relationship with sedimentary environments is published (Kershaw, 2012).Thus there is a limited amount of high-resolution data that involve the three components required in stromatoporoid ecology research: taxonomy,growth form and sedimentary facies.

3)Little is known of the original mineralogical composition of stromatoporoids; diagenetic features do not allow their mineralogy to be interpreted easily.Therefore it is unclear how stromatoporoid mineralogy fits into the wellestablished concept of aragonite/calcite seas, of importance because of the widespread and abundant occurrence of stromatoporoids during middle Palaeozoic time.

The aim of this paper is therefore to assess the facets of stromatoporoid studies that could be brought into focus as the 21stCentury progresses, to develop stromatoporoids as environmental indicators both in terms of local facies and of the wider field of palaeogeography.The paper discusses stromatoporoid mineralogy, taxonomy, and aspects of applications of sedimentary environments where they occur.This study tries to show the potential value of a focused integrated approach to maximize on the information available from stromatoporoids.

2 Stromatoporoid palaeobiology: major features

Stromatoporoids are readily recognizable in field settings, with a range of growth forms from laminar to domical, bulbous, branching and irregular shapes (Fig.1); some taxa have only certain growth forms, while other taxa show a range of forms.Stromatoporoids thrived in shallow marine reef and reef-related environments, and were especially able to deal with fine-grained sediment, in contrast to modern calcified sponges (Kershaw, 1998).Stromatoporoid growth histories (Fig.2)demonstrate use of both sediment and dead skeletons as substrates, and individual stromatoporoids recorded events on the sea bed while they were alive.A synthesis of these features is given in the revised Treatise Online by Webby and Kershaw (2011)and Kershaw (2012); nevertheless, published information provides only a relatively small dataset of available material, emphasizing that there is considerable scope to expand this knowledge.Many taxonomic studies provide valuable basic data on stromatoporoid growth forms,e.g.Stock (1982), Stearn (1983), Webby and Zhen (1993).Detailed studies focussed on stromatoporoid palaeoecology have been made on certain Silurian and Devonian stromatoporoid assemblages and show the need for many more similar studies to fully characterise the relationship between stromatoporoids and their environments (e.g.Kershaw, 1998; Da Silvaet al., 2011a, 2011b).This work requires a combination of growth form, low-level taxa and sedimentary data.

3 Stromatoporoid mineralogy

Constructing calcareous elements of stromatoporoids are used in taxonomy, their architecture has been extensively studied by numerous authors, summarized by Stearn (2011).However, even the (apparently)best-preserved specimens are substantially recrystallized and their structure needs to be understood, considered briefly here.

In all cases observed by the author, of brachiopods in the same samples and same facies as stromatoporoids, brachiopod shells are laminated and well-preserved (Fig.3)in significant contrast to the stromatoporoids, even those stromatoporoids considered to be well-preserved.This difference was also reported and illustrated by Rush and Chafetz (1991).Therefore, stromatoporoids were clearly NOT originally low-Mg calcite (LMC).However, simple demonstrations of the physical appearance of stromatoporoids in thin section draw attention to the differencesbetween stromatoporoids and other fossils, described in the next sections.

Fig.1 Examples of stromatoporoids in various settings in the Palaeozoic.A-Laminar form of Labechia from reef core, Coates Quarry, Wenlock of England; B-Domical and bulbous forms in a biostrome, Ludlow, Gotland; C-Back-reef facies with large bulbous stromatoporoid lower right, and numerous cross-sections through branching corals and stromatoporoids; Polished slab of facing stone on a building in London, UK, of the Upper Devonian Ashburton Limeston, Devon, England.Scale bar: black and white squares are 1 cm; D-Irregular stromatoporoid from Upper Silurian Rondout Formation of New York.

3.1 Aragonite recrystallization and dissolution

Apparently well-preserved stromatoporoids, showing architectural elements of vertical and horizontal structures,commonly occur in the same beds, even the same samples,as completely recrystallized mollusc shells.Furthermore,repeated observations by the author, of mollusc shells used by stromatoporoids as substrates throughout the Silurian of Gotland and England, demonstrate dissolution of the mollusc shell and collapse of the external mould onto the internal mould, yet the stromatoporoid has not suffered any dissolution (Figs.4 and 5).Such differences are also reported by Rush and Chafetz (1991)and Smosna (1984)in Devonian stromatoporoids, from New York and Virginia respectively, and in personal communication from Carl Stock (2013)on unpublished observations from the Pridoli (latest Silurian)of New York.This consistent difference between stromatoporoids and aragonitic mollusc shells creates a significant problem for interpretations that stromatoporoids were originally aragonitic.For reference note that the two modern calcified sponges most similar to stromatoporoids,CalcifibrispongiaandAstrosclera, both have aragonite calcareous skeletons (Stearn, 2010c).

3.2 Cathodoluminescence (CL)appearance

Stromatoporoids occur very commonly with crinoids(presumed originally high-Mg calcite, HMC)and both show a prominent speckled appearance in CL, which, in stromatoporoids, is sharp-bounded against the galleryfilling calcite cement.The CL view may show the original fabric of the stromatoporoid, but this depends on the interpretation of CL features, discussed by Kershaw (1994).However, crinoids have large overgrowths of non-luminescent cement, reflecting their single-crystal composition,in contrast to the small non-luminescent first generation cement on stromatoporoid fragments (Fig.6).The similar speckled appearance is circumstantial evidence of similarity between the two fossils’ mineralogy, but CL is not a reliable guide to mineralogy of carbonates and the similarity may be coincidental.Nevertheless, Rush and Chafetz(1991)demonstrated dolomitic microcrystals formed by diagenesis within stromatoporoid skeletons from the Devonian of New York, inferring that the stromatoporoids were HMC.The CL images in Figure 6 are interpretedhere as evidence of the original relationship between the stromaporoid skeleton and the gallery cement, such that the galleries were most probably infilled with cement after soft tissue decayed and was replaced by water in the galleries.The cements show a sequence of evolution from non-luminescent (probably oxygenated water in shallow burial), through bright luminescence (probably shallow anoxic position just below the redox boundary), to dull luminescent (probably deeper burial); see Scoffin (1987)for discussion of environments of different CL phases.However, the next section illustrates a later process of alteration.

Fig.2 Details of stromatoporoids, their life histories and substrates.A-Stromatoporoid grew on a dead halysitid colony, and the stromatoporoid formed a framework infilled with micrite.It is not clear whether the frame was formed as a primary cavity or was due to sedimentation that partly killed the stromatoporoid as it grew (see Kershaw et al., 2006 for discussion); Lea Quarry, Wenlock of England; B-Complex domical stromatoporoid growth involving at least two taxa and interrelation with the substrate.Successive layers of stromatoporoid growth used earlier dead stromatoporoids as substrate Visby Formation, Wenlock of Gotland; C-A domical stromatoporoid growing on the overturned skeleton of a solitary rugose coral Schlotheimophyllum, Visby Formation, Wenlock of Gotland; D-Stromatoporoid completely encasing a gastropod, with geopetal infill.The sample must have been turned over at least once on the sea floor to encase the gastropod.Note the white area shows partial silicification of the stromatoporoid in diagenesis; Lea Quarry,Wenlock of England.

Fig.3 Vertical thin section of Densastroma pexisum which was interrupted during growth, so that an atrypid brachiopod was enveloped by the recovered stromatoporoid growth (inset photograph).The main picture is an enlargement showing the well-preserved laminated brachiopod shell contrasting the altered stromatoporoid skeleton, discussed in the text.Note that the stromatoporoid shows characteristic irregular elongated crystals arranged normal to the growth surface, overprinting the skeletal structure, discussed in the text and further illustrated in Figs.7-10.SEM photographs of the contrast between stromatoporoids and brachiopods are given by Rush and Chafetz (1991).

Fig.4 Cross-section through a stromatoporoid (S)that grew on the dead shell of an orthoconic nautiloid (O)(left); the nautiloid has been susbstantially dissolved by diagenesis, whereas the stromatoporoid is unaffected.w= outer wall of nautiloid shell, si= siphuncle;septa are also visible as sharp changes in the micrite fill in the nautiloid.A heliolitid coral (C)lies on its side, right, and a recrystallized gastropod (G)is lower right; Visby Formation, Wenlock of Gotland.

Fig.5 Thin section (main picture)and polished slab (inset)of Densastroma pexisum, Visby Formation, Wenlock of Gotland.The stromatoporoid grew on a dead mollusc shell that was subsequently dissolved in diagenesis and the sediment collapsed, leaving a thin line to show the location of the shell.The stromatoporoid is unaffected.The sediment was burrowed (lower centre of photo), revealing unconsolidated sediment.The stromatoporoid used the shell as substrate thereby avoiding the loose sediment; however, other samples,not illustrated in this paper, indicate that stromatoporoids were capable of growing directly on the muddy substrate (see http://earthsurfaceprocesses.com).

3.3 Crossed polarized light (XPL)appearance

XPL provides a different view from CL and there is no primary relationship between them; the cements viewed in XPL crosscut the fabrics seen in CL and are interpreted as later diagenetic change in the stromatoporoid.However,this alteration of stromatoporoids is more than just recrystallization of the skeleton; stromatoporoid skeletons exhibit a feature that is probably unique in fossils.In XPL, vertical sections of the skeleton of almost all species shows an arrangement of irregular elongated calcite crystals orientated normal to the growth surface, crossing the lamination(coenostromes), thereby cutting both the stromatoporoid gallery cements and skeleton alike.Figure 7 shows the edge of a fragment of stromatoporoid in grainstone of several fossil groups, but the crystal structure stops abruptly at the stromatoporoid margin; these bioclasts were from shelly organisms that grew in environments and formed part of a sequence of stromatoporoid biostromes, so were deposited all together, yet the diagenesis within the stromatoporoid affected only the stromatoporoid.This observation, that stromatoporoids have irregular elongated crystals in contrast to other fossils, is repeated by the author in different facies of Silurian and Devonian stromatoporoids.Figure 7 particularly shows the contrast with crinoids, since (a)crinoids do not show such irregular crystals and (b)stromatoporoids lack syntaxial cements.Figure 8 shows apparently well-preservedHabrostromafrom Silurian of New York(sample provided by Carl Stock), demonstrating that the irregular crystals, which in transverse section are approximately equant.Figures 9 and 10 show two further taxa(Eostromatopora impexaandDensastroma pexisum)fromWenlock of Gotland (Sweden), showing the same irregular calcite cement cross-cutting the stromatoporoid structure,irrespective of stromatoporoid taxa (see also Fig.12).This characteristic irregular cement is so pervasive that even badly recrystallized stromatoporoids can be recognized as stromatoporoids in cross-polarized light, including cases where any skeleton-based taxonomic features are further altered beyond recognition.

The diagenetic character of stromatoporoids illustrated in Figures 7-10 is poorly reported in the literature, the only XPL illustrations that I am aware of are in Smosna(1984)and Rush and Chafetz (1991).The probable reason for its uncommon description is that stromatoporoid taxonomy normally uses thin sections of 50-80 microns thickness, in which the irregular crystal overprint is not visible because of the large refractive index of calcite; nevertheless even in PPL the fabric is visible if the sections are thin enough (see also Figs.3, 5 and some plates inDong, 2001).Thus thin sections need to be the normal 30 micron thickness, or less, for easy observation of this fabric.Smosna (1984, p.1004)provided a concise description of the recrystallization into irregular crystals that exactly matches the observations made in this paper of such alteration in stromatoporoids from various ages and facies.Smosna (1984)also noted that the crystals cross stromatoporoid lamination in vertical section.Furthermore, Smosna (1984)observed that the crystals do not pass through areas of a stromatoporoid where sediment interrupted growth, further emphasising that the diagenetic change is restricted to the stromatoporoid.Finally, Smosna (1984)recorded undulose extinction in the crystals, which can also be appreciated in Figures 8-10.

Fig.7 Fragment of stromatoporoid in vertical section (lower half of photo)and crinoidal-shelly grainstone above, from Ludlow of Gotland.A-Plane-polarised light; B-Cross-polarised light.B shows the characteristic irregular elongated calcite crystals typical of stromatoporoids, with the crystals cross-cutting the stromatoporoid skeleton, but terminating sharply at the stromatoporoid margin,demonstrating that the internal diagenetic alteration is limited to the stromatoporoid.

The process of stromatoporoid diagenesis took place not just in the stromatoporoid skeleton, but also in the galleryfilling cement.Figure 6 shows CL reveals the sequence of cements in galleries, which is overprinted by the elongated irregular crystals.The process of alteration remains unexplained and is an avenue for future investigation.Smosna(1984)interpreted the irregular calcite as having formed by inversion to calcite from the original mineralogy in freshwater environments; this may or may not apply in all cases, particularly in view of the CL evidence of later burial cement in the particular samples shown in Figure 5,but is certainly possible in other cases because of the relatively shallow water environment of stromatoporoids.The recrystallized structure survives further alteration of theskeletons (because it is found in specimens so altered that constructing elements are largely lost).

Fig.8 Stromatoporoid (Habrostroma)from the Upper Silurian of New York.A and B-PPL and XPL views in vertical section showing the irregular elongate calcite crystals in B, typical of stromatoporoids; C and D-Enlargements of A and B, respectively, showing the detail of relationship between irregular crystals in XPL and the stromatoporoid structure; E and F-Transverse section of the same specimen, showing the irregular crystals are approximately equant in transverse view.Thin sections provided by Carl Stock.

Fig.9 Vertical sections of Eostromatopora impexa from the Visby Formation, Wenlock of Gotland in both PPL and XPL.The photographs emphasize that the irregular calcite crystals in XPL cross-cut the stromatoporoid skeleton and continue into the gallery space,demonstrating the alteration of this skeleton, regardless of its apparently well-preserved structure.In C, the gallery space is shown by small equant areas of clear calcite cement left and right.Compare this figure with Figure 10.

In summary, because of the differences between stromatoporoids and both molluscs and crinoids with which they occur, the issue of the original mineralogy of stromatoporoids remains a problem unlikely to be resolved by light microscopy but more data are required to provide a comprehensive view of these features.

3.4 Scanning electron microscope (SEM)study

SEM images of stromatoporoids have been used by several authors to investigate the structure of the stromatoporoid skeleton (Stearn, 1977, 1989a; Smosna, 1984; Stearn and Mah, 1987; Rush and Chafetz, 1991).In each case,polished surfaces were etched and examined with secondary electrons, illustrations being of the microtopography of etched surfaces.In each case the stromatoporoid skeleton is revealed as having a sharp contact with the surrounding gallery cement, as can be seen in thin sections in PPL.There is no description in the above references of crystal boundaries passing from the skeleton into the cement, identified in XPL in Figures 7-10.However, careful examination of published SEM photographs in those references above shows curving and irregular lines in the structure subject to greater etching (e.g.Stearn, 1977; Stearn and Mah, 1987, Figs.1C, 1E, 1F; Stearn, 1989a, Fig.1B;Rush and Chafetz, 1991, Fig.3).Such lines, also mentioned by Stearn (1977), may be interpreted as boundaries of the large irregular crystals that cross-cut the stromatoporoid skeleton and overprint the smaller crystals making up the skeleton itself.Thus the visual evidence from published SEM photographs may be considered as being compatible with the diagenetic feature recognizable in XPL that overprints the stromatoporoid skeleton.

Fig.10 Low, medium and high power views of vertical sections of Densastroma pexisum from the Visby Formation, Wenlock of Gotland in both PPL (left)and XPL (right).Compare this figure with Figure 9, they are very different stromatoporoid skeletal structures(taxa)in the same environment but with the same style of alteration.Examples in Figures 7 and 8 are from two more different taxa and settings, so that all four stromatoporoid taxa show the same characteristic irregular elongated structure, visible also in Figures 3 and 5 in PPL.See text for discussion.

3.5 Implications for palaeoecology and palaeogeography

Cherns and Wright (2000)demonstrated loss of aragonite fossils by dissolution in molluscs in contrast to a key Silurian example of exceptional preservation of originallyaragonitic shells in silica.This work illustrates the underrepresentation of these organisms in the fossil record due to diagenetic dissolution.However, stromatoporoids occurring together with dissolved molluscs, as illustrated in Figures 4 and 5, do not exhibit dissolution loss, so this shows that stromatoporoid taxonomic assemblages are not under-represented, and that all taxa of stromatoporoids are likely to be equally represented in the fossil record.This key point provides confidence for palaeoecological and palaeogeographic studies on stromatoporoids.Thus the reconstructions of biogeographic distributions of stromatoporoids, such as those provided by Stock (1990)for the Devonian, are robust within the limitations of the accuracy of continental positioning and completeness of stromatoporoid collections.

4 Stromatoporoid taxonomy

Stromatoporoid study requires two thin sections per specimen, carefully orientated so that constructing elements can be viewed in as precisely horizontal and vertical attitudes as possible.Oblique sections, and oblique parts of sections, are misleading, and prevent the accurate identification of constructing elements, thereby diminishing the validity of identification.Studies on taxonomy must take care of this issue; the need for careful preparation prior to identification is one of the barriers to stromatoporoid study.

The validity of stromatoporoid taxonomy came under scrutiny from 1985 onwards as a result of the discovery in modern calcified sponges that the spicule-based classification used by living-sponge researchers is not consistent with the calcareous-skeleton-based classification of ancient calcified sponges (Vacelet, 1985).Furthermore,Reitner and Engeser (1987)described three species of calcified sponge, based on spicules, in different specimens that had an identical calcareous skeleton.The calcareous skeleton is a secondary skeletal structure, that forms after the spicules in a modern calcified sponge, and is considered a grade of organization rather than a higher-level taxonomic feature.In the case of stromatoporoids, this is the stromatoporoid grade.Spicules are almost absent in Palaeozoic stromatoporoids, only 2 cases are reported so far: Upper Devonian (Frasnian, Da Silvaet al., 2011c)and Upper Carboniferous (Woodet al., 1989).Thus the question of how to reconcile these contradictory classifications continues.Indeed the issue is complicated by the fact that some modern calcified sponges lack spicules or lose their spicules on death.Nevertheless, Kershaw (1997)argued that, irrespective of the presence of spicules, the calcareous skeleton shows such large and consistent differences between low-level taxa at least at genus level, it is only logical to accept such taxa as having validity.

Whether such distinction remains valid at the level of traditional stromatoporoid species is less certain in the light of the work by Reitner and Engeser (1987).Furthermore,certain low-level taxa are limited to certain growth forms(Kershaw 1984, 1990, 1998; Da Silvaet al., 2011a, 2011b)providing confidence of the validity of at least genus-level taxa based on the calcareous skeleton (discussed further below).Nevertheless, the spicule vs calcareous-skeleton argument reported above informs workers on Palaeozoic stromatoporoids, because it indicates that caution should be used in this classification.Thus the spicule-based classification suggests that higher-level groupings, based on the calcareous skeleton, into orders may not have validity.

Consequently, a pragmatic approach to stromatoporoid taxonomy for palaeobiological investigations is to accept that generic-level distinctions are reliable, but investigators should perhaps not expect too much from further divisions of taxa; this is an argument against over-splitting of taxa, and was applied by Da Silvaet al.(2011a, 2011b)in comprehensive palaeobiological studies of stromatoporoids from the Frasnian of southern Belgium.The discovery of three different spicule-based species within an identical calcified skeleton (Reitner and Engeser, 1987)diminishes the value of finer divisions based on the calcified skeleton alone.Nevertheless, numerous examples exist of calcareous-skeleton-based taxa as different species within one genus demonstrating distinct and consistent differences which lead to the conclusion that they are different low-level taxa.A good example isPetridiostroma simplexwhich occurs in the same outcrops asP.linnarssoniin the Lower Wenlock Visby Formation on Gotland (Kershaw,1984), leaving no doubt that they are different taxa.The issue that then arises is whether or not these are different species within one genus, or whether they are different genera biologically yet share the same morpho-taxon genus name.This issue is common in palaeontology and should not be a cause for concern; it underlies the need to remember that the names given to many fossils are morpho-taxa names, and that it is the distinctness of taxa thatmatters most.Nevertheless, in stromatoporoids there is potential danger of misinterpretation in biogeographic work.For example, it would be difficult to prove whether a taxon in one area is the same taxon in another area thousands of kilometers distant, in cases where the skeletal structure of the two occurrences are very similar, yet consistently different.Again, this is a problem common in palaeontology but it empowers investigators to remain cautious of interpretations and emphases the need for large sample collections with careful assembly of associated growth form and sedimentological data to achieve the most robust information.The possibility of invalidity of higher groupings of stromatoporoids promotes caution in interpreting their ecology at order level, but does not affect genus-level ecological work; thus focus on the interrelationships between genera, their growth forms and environments is likely to provide reliable conclusions.

Despite the previous comments, although stromatoporoid taxonomy is considered here to be valid at genus level, there are two potential problems with this approach,about which investigators should take care: (1)it is still possible for more than one stromatoporoid biological species to exist in any one setting, within identical calcareous skeletons; (2)consequently, in comparisons between sites, variations in the growth form of a single skeletonbased taxon, may not necessarily be ecophenotypes, as interpreted by Kershaw (1997), but may be due to different biological species within the same skeleton-defined taxon, which are impossible to discriminate since spicules are absent.The latter point has potential implications for palaeobiological interpretations of growth form variations within taxa between different environments.However, in one detailed case study there is good reason to accept the existence of different growth forms within one taxon based on the calcareous skeleton (Kershaw, 1997).In this case, at the Grogarnshuvud locality in the Ludlow of Gotland, Sweden, two biostromal deposits, one overlying the other, with a different sedimentary setting, contain the same assemblages of stromatoporoid taxa.One taxon has a different growth form in the lower biostrome compared to the upper biostrome, whereas other taxa in the assemblage have similar growth forms in both biostromes.Furthermore, these middle Ludlow-age stromatoporoidrich biostromes are widespread across eastern Gotland,with the same stromatoporoid assemblage across the area(Sandstr?m and Kershaw, 2008); although it is possible for different species (with identical calcareous skeletons,but biologically distinct)to exist in different biostromes,it is much more likely that the taxa are the same, across this region of normal marine waters in a wide carbonate shelf where shallow ocean water is expected to have circulated without restriction, thereby widely distributing zooplankton.

Finally, Stearn (1989b)drew attention to variations of the skeletal structure from older to younger parts within single thin sections of one stromatoporoid taxon, a feature which is widespread across taxa, environments and locations.Stromatoporoids clearly interacted with their environments throughout their lives in complex and responsive ways, so that growth characteristics of a single individual stromatoporoid varied through life.Work on stromatoporoids that have large variation of skeletal structure within a single thin section, and between different individuals in an assemblage, benefit most from a genus-level approach.

There is a tendency in publication of stromatoporoid taxonomy to publish photographs of only small areas of thin sections; but if stromatoporoids show intra-sample variation, then illustrations benefit from publication of large areas of a single thin section, with separate photographs showing enlargement of several areas, to illustrate the variations.In small stromatoporoids, if the entire fossil can be displayed in a single thin section, this aids appreciation of not only the taxonomy but also the growth history of the sample, for palaeoecological interpretations (see Figs.3-5, 11 and 13-14).

5 Stromatoporoid growth banding

Although growth banding in stromatoporoids is commonly considered to be annual (e.g.Gao and Copper,1997), there is no empirical evidence to support this(Young and Kershaw, 2005).A principal issue is the relationship between stromatoporoid banding and the deposition rate of sediment; this may be investigated at the margins of the stromatoporoid, and it is essential to have the margins available for study, in order to investigate whether or not sedimentation events may have caused the banding(Young and Kershaw, 2005).Furthermore, Kershawet al.(2006)noted that stromatoporoids may have grown primary cavities at their margins in times when sedimentation was slow or stopped, and this may be reflected in growth bands.Thus the lack of margin information prevents the fullest analysis of the banding, so that study should be made only where margins are preserved.Figure 11 shows polished slabs demonstrating banding in stromatoporoids;more detailed images in Young and Kershaw (2005)demonstrate the details of these bands.The extent to whichgrowth banding may be used to interpret annual, and possibly subannual, climatic changes affecting the places where they grew is a study area within its infancy, and would benefit from expansion of effort.

Fig.11 Polished vertical sections of stromatoporoids showing growth banding.A-Bands seem to relate to the sedimentation at the margins, as demonstrated in detail by Young and Kershaw (2005); B-Banding is more regular and not apparently related to sedimentation at the margins (A and B are both Densastroma pexisum from the Visby Formation, Wenlock of Gotland); C-Bands end sharply at the margin of a large domical stromatoporoid; there is no relationship with any marginal features (Unidentified stromatoporoid, Ludlow of Gotland); D-Partly silicified domical form of Plectostroma scaniense, Ludlow of Gotland, showing the effects of diagenesis on growth banding; whether the diagenetic alteration to silica (white areas)is a reflection of an original growth character is open to interpretation.

6 Intergrown organisms

Stromatoporoids include a range of shelly organisms which grew along with the stromatoporoid, resulting in the shells becoming encased within the stromatoporoid skeleton (Stearn, 2011).Intergrown organisms are common in reef stromatoporoids, although occur also in examples in bedded limestone.Figures 12-1A-12-1F showsPetridiostroma convictumfrom biostromes in the Ludlow of Gotland, with both syringoporids and branching rugose corals in the same specimen.Figures 12-1C-12-1F also shows the skeletons of all three organisms in thin sections prepared thinner than normal (15 microns in contrast toca.50 microns normally used to study stromatoporoid taxonomy, as used in Figs.12-1A and 12-1B).15 micron sections are highly instructive because they demonstrate that the stromatoporoid skeleton is, in this case, represented by a fine “dusty” appearance, referred to as specks and likely rich in fluid inclusions, by Stearn (2010c), completely superimposed by calcite cement crystals in diagenesis in contrast to the better preservation of the corals.

Intergrown organisms have considerable value in palaeoecological studies of stromatoporoids because they allow exploration of the relationship between stromatoporoids and other taxa while both are alive together.Corals(both tabulate and rugosa)are the most abundant, but numerous tube fossils, probably worms (e.g.Tourneuret al.,1994; Zhen and West, 1997), also occur.There is very little detailed and comprehensive work on intergrowths, and there is great potential for expansion of such study in cases where intergrowths occur.Some show a specific relationship between certain stromatoporoid taxa and intergrown organisms, the most common of which are syringoporid tabulates (see Da Silvaet al., 2001a, 2001b, and Stearn,2011, for photographs and descriptions).In the majority of cases, the stromatoporoids seem unaffected by the presence of the intergrown organisms (Figs.12-1A-12-1B,

also see Kershaw, 1987), although an example from the Middle Devonian of France has been interpreted to represent parasitism on the stromatoporoid by spiral tubes(Zapalski and Hubert, 2010)because of downturning of the successive stromatoporoid laminations around the spiral tubes.A comparable example from the Wenlock of England in Figure 12-2G shows a mixture of response, where some intergrown tubes are accompanied by down-flexing of the stromatoporoid laminae,while others appear to be unaffected.Even individual tubes show a varying relationship with stromatoporoid laminae in different parts of the same thin section (Fig.12-2G).Figure 12-2G thus indicates that a variation of response by the stromatoporoid host to the guest organism, even within one specimen.In another example, spiral tubes initiated on growth interruptions in stromatoporoids,while others seem to have initiated between growth interruptions (Fig.12-2H, where the origin points of the spiral tubes are out of the plane of section).Spiral tubes in Figures 12-2G-12-2H show how the calcareous tube expanded in diameter over a few millimeters of vertical growth, suggesting they grew quickly, using the stromatoporoid as a base; the lower parts of the tubes then became encased in the stromatoporoid skeleton as its growth progressed and possibly overwhelmed the spiral tube.If so, then the spiral tubes may represent an opportunistic shelly organism using the stromatoporoid as a firm base to grow, but then potentially killed by their host.The intriguing possibility that the spiral tubes represent an organism which shed larvae before becoming overwhelmed by its stromatoporoid host needs to be tested and could reveal the dynamic relationship between guest and host organisms.In some cases

intergrown organisms may have been soft-bodied, leaving cavities inside stromatoporoids that are otherwise difficult to explain (see Stearn, 2011).All these examples require careful and systematic analysis to produce a detailed dataset of all the variability before comprehensive interpretations are possible.Figure 13 shows a rare example of a different type of intergrowth, as possible competitive growth of two adjacent organisms (stromatoporoid and bryozoan in this case).

Fig.12-1 Intergrown corals inside a stromatoporoid, Ludlow of Gotland.Photos A-F show Petridiostroma convictum stromatoporoid.A and B-Vertical and tangential thin sections of different samples showing the close intergrowth of corals and stromatoproid; A has syringoporid coral tubes that developed as the stromatoporoid grew, and their growth rates were presumably well-matched; B has both syringoporids (small circles)and branching rugosan; A and B are negative photographs; C and D-XPL views of very thin section (15 microns)of detail of structure of the Petridiostroma convictum stromatoporoid; the stromatoporoid skeleton is visible as a dusty appearance on the calcite cement, demonstrating pervasive alteration of the stromatoporoid, even though its taxon is clearly identifiable at more normal thickness in A and B; E and F-XPL views of very thin sections of Petridiostroma convictum stromatoporoid with rugose coral (E)and syringoporid (F).In E, only the rugosan is visible, but shows its very well-preserved wall structure; in F, the syringoporid wall structure is partly altered, but is better preserved than the stromatoporoid “dusty” fabric in C and D.A-F therefore demonstrate not only the intimate relationship between corals and stromatoporoids in intergrowth, but also the differences in preservation.

Fig.12-2 In G and H (both from the Wenlock of England), spiral tubes grew inside a stromatoporoid, some apparently initiating at growth interruption surfaces (red arrows)others apparently initiating between interruptions (green arrows).G also shows vertical tubes,some associated with downflexing of the stromatoporoid growth layers (blue arrows), others showing the stromatoporoid was not affected (yellow arrows).Even some tubes with blue arrows show that only some parts of individual tubes are related to downturned laminae; other parts of the same tubes show no effect, emphasizing the complex relationship between tubes and stromatoporoid host.

Fig.13 A-Vertical section of unidentified stromatoporoid, Lea Quarry, Wenlock of England, showing intergrowth with another organism, in this case a bryozoan; B-Thin section detail of an adjacent part of the sample, demonstrating the tight intergrowth between two different organisms in close contact with each other, and may represent competitive growth.Such cases are rare in stromatoporoids,but have potential to help explain the mechanisms of stromatoporoid growth process.In this case a reasonable interpretation is that the growth rate of either or both organisms varied, allowing each to expand laterally in turn, to generate the interlayered growth.

7 Geotropic growth

Some researchers have suggested that stromatoporoids responded to light (see Kershaw, 1998 for discussion)and Kazmierczak (1976)went further to interpret stromatoporoids as cyanobacteria, contrasting the sponge interpretation discussed earlier.However, a poorly reported aspect of stromatoporoids provides evidence of geotropic growth,as illustrated by columnar features that maintain a vertical attitude regardless of the position of the stromatoporoid.Figure 14 illustrates variation of angle of repose of the basal surface of three examples yet in each case the columnar features all have a vertical attitude.Whether or not this is light-controlled is open to interpretation.Figure 14A,Parallelostroma typicum, and Figures 14C,14D,Clathrodictyon mohicanum, illustrate taxa that are found in outcrop close to specimens of the same taxon (not figured here)lacking vertical columns, suggesting that light was not acontrol unless the specimens with columns were shaded when alive on the biostromal reef surface where they occur (see Kershaw, 1990).However, there is no physical evidence within the skeletal structure of the stromatoporoids of the control on the formation of the columns and is an area for future investigation.The possibility of a phototropic response in stromatoporoids may relate to the undisputable conclusion that Palaeozoic stromatoporoids grew quickly, possibly as fast as modern corals, in contrast to the extremely slow growth rates of modern calcified sponges.Palaeozoic stromatoporoids can be many metres in diameter, which can be explained only by rapid growth(see Kershaw, 1998 for discussion).

Fig.14 Geotropic growth in stromatoporoids; small columns of stromatoporoid skeleton grow vertically irrespective of the attitude of the fossil.A- Laminar stromatoporoid (Parallelostroma typicum)on a very gently sloped substrate developed vertical columns;B- Unidentified stromatoporoid comprising columnar structure; black arrows: vertically-orientated columns grew up from a horizontal base; yellow arrows: vertically-orientated columns grew up from a gently sloping base.C and D- Different views of a fragment of a large low profile stromatoporoid (Clathrodictyon mohicanum)that grew on a ca.20-degree slope; short broad columns grew vertically.Scale in C shows size of features in D, and way-up arrow of sample in outcrop A, C and D from Kuppen, Hemse Group (Ludlow),Gotland.B from Halls Huk, H?gklint Formation (Wenlock), Gotland, sample donated by Nigel Watts.

8 Conclusions

The following points support the views presented in this discussion paper that stromatoporoids are important and valuable fossils for palaeoenvironmental reconstruction in Palaeozoic shallow marine carbonate environments where they occur:

1)Stromatoporoid taxonomy at generic level is a valid and powerful tool for investigation of palaeobiology, in conjunction with growth forms and sedimentary information.

2)Stromatoporoid mineralogy was not originally low-Mg calcite, and was not likely to have been aragonite;the mineralogy may have been high-Mg calcite because of preservation of skeletal features; but the overprinting of skeletal fabrics by diagenetic cements differs from the appearance of abundant crinoidal fossils that occur in stromatoporoid-bearing facies.Nevertheless, comparison with molluscs shows that stromatoporoids did not suffer diagenetic loss by dissolution and therefore assemblages of their taxa are representative of the sponge assemblages alive in the Palaeozoic.Thus, palaeoecological and palaeogeographical interpretations of stromatoporoids will not suffer from data loss.

3)Stromatoporoid growth banding is closely related to features at the margins of their skeletons, and it is essential to study the margins as well as the banding itself.Currently there is no certainty that stromatoporoid banding represents annual growth, thus estimates of growth rates of stromatoporoids, and therefore further applications, for example, sediment deposition rates, should be viewed with great caution.

4)Stromatoporoids are commonly closely associated with other organisms as intergrowths and these provide a rich potential source of information on the controls on stromatoporoid growth.

5)Whether stromatoporoids were influenced by light remains equivocal, but columnar features in some specimens reveal a geotropic response.

6)For stromatoporoid studies to have maximum palaeoecological value, sample numbers need to be large in high-resolution studies, and careful thin section preparation is required.

Acknowledgements

I am grateful to Carl Stock, Gong Yiming and Wang Yuan (surnames in upper case)for valuable comments on an earlier version of this manuscript.Carl Stock kindly provided the sample ofHabrostromafor Figure 8 and Nigel Watts the sample for Figure 14B.This paper is a contribution to both IGCP591 and IGCP596.

Cherns, L., Wright, V.P., 2000.Missing molluscs as evidence of large scale, early skeletal aragonite dissolution in a Silurian sea.Geology, 28: 791-794.

Da Silva, A.C., Kershaw, S., Boulvain, F., 2011a.Stromatoporoid palaeoecology in the Frasnian (Upper Devonian)Belgian platform, and its applications in interpretation of carbonate platform of carbonate platform environments.Palaeontology, 54: 883-905.

Da Silva, A.C., Kershaw, S., Boulvain, F., 2011b.Sedimentology and stromatoporoid palaeoecology of Frasnian (Upper Devonian)carbonate mounds in southern Belgium.Lethaia, 44: 255-274.

Da Silva, A.C., Kershaw, S., Boulvain, F., Reitner, J., 2011c.Longexpected! - First record of demosponge-type spicules in a Devonian stromatoporoid (Frasnian, Belgium).In: Aretz, M., Delculée, S., Denayer, J., Poty, E.(eds).11thSymposium on Fossil Cnidaria and Sponges, Liège, August 19-29, 2011, Abstracts.K?lner Forum Geol.Pal?ont., 19: 32-33.

Dong Deyuan, 2001.Stromatoporoids of China.Beijing: Science Press, 423 (in Chinese with English abstract).

Gao Jianguo, Copper, P., 1997.Growth rates of Middle Paleozoic corals and sponges: Early Silurian of eastern Canada.Proceedings of the Eighth International Coral Reef Symposium, 2:1651-1656.

Goldfuss, A., 1826.Petrefacta Germaniae (1sted).Verlag von List and Francke, Dusseldorf, 761.

Kazmierczak, J., 1976.Cyanophycaean nature of stromatoporoids.Nature, 264: 49-51.

Kershaw, S., 1981.Stromatoporoid growth form and taxonomy in a Silurian biostrome, Gotland.Journal of Paleontology, 55: 1284-1295.

Kershaw, S., 1984.Patterns of stromatoporoid growth in level-bott-om environments.Palaeontology, 27: 113-130.

Kershaw, S., 1987.Stromatoporoid - coral intergrowths in a Silurian biostrome.Lethaia, 20: 371-382.

Kershaw, S., 1990.Stromatoporoid palaeobiology and taphonomy in a Siluran biostrome, Gotland, Sweden.Palaeontology, 33(3):681-705.

Kershaw, S., 1994.Cathodoluminescence of Silurian stromatoporoids from Gotland, Sweden.Courier Forschungsinstitut Senckenberg, 172: 307-318.

Kershaw, S., 1997.Palaeoenvironmental change in Silurian stromatoporoid reefs, Gotland, Sweden.Boletin Real Sociedad Espa?ola de Historia Natural (Seccion Geologicas), 91(1-4): 331-344.

Kershaw, S., 1998.The Applications of stromatoporoid palaeobiology in palaeoenvironment analysis.Palaeontology, 41: 509-544.

Kershaw, S., 2012.Paleoecology.Part E, Volume 4, Chapter 13,Hypercalcified Porifera.Lawrence Press, University of Kansas,Treatise Online, 31: 1-24.

Kershaw, S., Brunton, F., 1999.Palaeozoic stromatoporoid taphonomy: ecologic and environmental significance.Palaeogeography,Palaeoclimatology, Palaeoecology, 147: 1-16.

Kershaw, S., Wood, R., Guo, L., 2006.Stromatoporoid response to muddy substrates in Silurian limestones.GFF, 128: 131-138.

Reitner, J., Engeser, T.S., 1987.Skeletal structures and habitats of Recent and fossil Acanthochaetetes (subclass Tetractinomorpha,Demospongiae, Porifera).Coral Reefs, 6: 13-18.

Rush, P.F., Chafetz, H.S., 1991.Skeletal mineralogy of Devonian stromatoporoids.Journal of Sedimentary Petrology, 61: 364-369.

Sandstr?m, S., Kershaw, S., 2008.Palaeobiology, ecology, and distribution of stromatoporoid faunas in biostromes of the mid-Ludlow of Gotland, Sweden.Acta Palaeontologica Polonica, 53:293-302.

Scoffin, T.P., 1987.An Introduction to Carbonate Sediments and Rocks.Blackie, Glasgow & London, 274.

Smosna, R., 1984.Diagenesis of a stromatoporoid patch reef.Journal of Sedimentary Petrology, 54: 1000-1011.

Stearn, C.W., 1977.Studies of stromatoporoids by scanning electron microscopy.Burreau de Recherches Géologiques et Minières,Mémoir, 89: 33-40.

Stearn, C.W., 1983.Stromatoporoids from the Blue Fiord Formation(Lower Devonian)of Ellesmere Island, Arctic Canada.Journal of Paleontology, 57: 539-559.

Stearn, C.W., 1989a.Specks in the microstructure of Paleozoic stromatoporoids.Proceedings of 5th International Symposium of Fossil Cnidaria, Brisbane.Association of Australasian Palaeontologists Memoir, 8: 143-148.

Stearn, C.W., 1989b.Intraspecific variability and species concepts in Palaeozoic stromatoporoids.Association of Australasian Palaeontologists Memoir, 8: 45-50.

Stearn, C.W., 2010a.Morphological affinities of Paleozoic Stromatoporoidea to other fossil and Recent groups.Part E, volume 4, Chapter 9E, Hypercalcified Porifera.Lawrence Press, University of Kansas, Treatise Online, 7: 1-9.

Stearn, C.W., 2010b.Paleozoic Stromatoporoidea: general introduction.Part E, Volume 4, Chapter 9A, Hypercalcified Porifera.Lawrence Press, University of Kansas, Treatise Online, 5: 1-3.

Stearn, C.W., 2010c.Microstructure and mineralogy of Paleozoic stromatoporoids.Part E, Volume 4, Chapter 9D, Hypercalcified Porifera.Lawrence Press, University of Kansas, Treatise Online,6: 1-25.

Stearn, C.W., 2011.Internal morphology of the Paleozoic Stromatoporoidea.Part E, Volume 4, Chapter 9C, Hypercalcified Porifera.Lawrence Press, University of Kansas, Treatise Online, 18:1-37.

Stearn, C.W., Mah, A.J., 1987.Skeletal microstructure of Paleozoic stromatoporoids and its mineralogical implications.Palaios, 2:76-84.

Stock, C.W., 1982.Upper Devonian (Frasnian)Stromatoporoidea of north-central Iowa: Mason City Member of the Shell Rock Formation.Journal of Paleontology, 56: 654-679.

Stock, C.W., 1990.Biogeography of the Devonian stromatoporoids.In: McKerrow, W.S., Scotese, C.R.(eds).Palaeozoic Palaeogeography and Biogeography.Geological Society Memoir, 12:257-265.

Tourneur, F., Lachkhem, H., Mistiaen, B., 1994.Trypanopora conilinov.sp.(Annelida?)from the Couvin Limestone, Eifelian of the southern margin of the Dinant Synclinorium (Belgium).Biological affinities and relationships with its hosts.Mémoires Institut Géologique de l’Université Catholique de Louvain, 35: 83-122.

Vacelet, J., 1985.Coralline sponges and the evolution of Porifera.Special Publication of the Systematics Association, 28: 1-13.

Webby, B.D., Kershaw, S., 2011.External morphology: shapes and growth habits.Part E, Volume 4, Chapter 9, Hypercalcified Porifera.Lawrence Press, University of Kansas, Treatise Online, 25:1-73.

Webby, B.D., Zhen Yong-Yi, 1993.Lower Devonian stromatoporoids from the Jesse Limestone of the Limekilns area, New South Wales.Alcheringa, 17: 327-352.

Wood, R., Reitner, J., West, R.R., 1989.Systematics and phylogenetic implications of the haploscerid stromatoporoidNewellia mira nov.gen.Lethaia, 22: 85-93.

Young, G., Kershaw, S., 2005.Classification and controls of internal banding in Palaeozoic stromatoporoids and colonial corals.Palaeontology, 48: 623-651.

Zapalski, M.K., Hubert, B.L.M., 2010.First fossil record of parasitism in Devonian calcareous sponges (stromatoporoids).Parasitology, 138: 132-138.

Zhen Yong-Yi, West, R.R., 1997.Symbionts in a stromatoporoidchaetetid association from the Middle Devonian Burdekin Basin,north Queensland.Alcheringa, 21: 271-280.