Substrate Specificity of Lower Devonian Corals

---

Substrate specificity relates to the idea that coral polyps were selectively seeking certain objects on the sea floor as anchor points for colony formation. The premise is that rather than random settling on either the soft mud of the sea floor or the first hard object encountered (brachiopod, pebble, crinoid stem etc), the polyp was selecting a specific host upon which the colony could grow with maximum chance of survival. In many instances, determination of the original host can be made as it is often preserved within the base of the the coral. Most specimens show that the host was completely smothered by the coral and entirely incorporated into the basal theca of the colony. In rare instances the coral attached to a much larger host it was unable to incorporate.

A previous study on the species Pleurodictyum americanum (Brett and Cottrell 1982) from the Middle Devonian of New York proved that the coral was  on average selecting the shells of the gastropod Palaeozygopleura as a host about 35% of the time (depending on the bed studied this varied from 20% to 60%). This might not seem a high percentage, but considering the gastropod makes up only 2% of the fauna in any of the studied beds, even the lowest figure of 20% it is far above what would be expected if attachment was purely random. The authors determined that the Palaeozygopleura shells did not contain the living snail at the time of colonisation, which further reduced the number of shells that met the criteria required by the polyp. A number of other factors were identified as possible reasons for why Palaeozygopleura was chosen, but the authors concluded the Pleurodictyum polyps were actively selecting the snail as attachment sites for future colonies.

The aim of this study is to determine whether the corals Favosites conicus and P. trifoliatum  found in the Birdsong Shale of Tennessee and Bois d'Arc Formation of Oklahoma also exhibit selectivity in their settlement sites.  All specimens referenced and figured were collected during 1995/6 and 2006/7 from outcrops in Tennessee and Oklahoma.  The Birdsong specimens came principally from 3 different sites in the Birdsong - the Parsons roadcut and two Vulcan quarries - sites for which collections were available for comparative analysis. The Oklahoma specimens came from the White Mound locality  and the trilobite quarry near Clarita.

Reference

Brett, C E. & Cottrell, J F. 1982. Substrate Specificity in the Devonian Tabulate Coral Pleurodictyum. Lethaia Vol 15. p.247-262 

Authors Note:   Pleurodictyum lenticulare does occur on the Cravatt of Oklahoma but it is quite rare and a meaningful number of specimens suitable for analysis has not yet been accumulated. As a result, the primary analysis for this species is based on specimens from the Birdsong Shale of Tennessee. A second species from the Birdsong (P. trifoliatum - right) is not included in this study as it is considerably more uncommon and again an insufficient number of specimens were available.

For an introduction to the Birdsong Shale and the palaeovenvironment of the Ross Shelf please click here.

Pleurodictyum lenticulare is an uncommon member of the Ross Formation coral fauna, accounting for typically only 2 to 3% of recovered anthozoans. However, at a locality just north of Parsons, Tennessee the coral is considerably more common, making up 10% of the coral fauna. From this locality a considerable number of specimens were acquired resulting in almost 200 specimens being available for analysis.

The summary below lists all the identifiable and unidentifiable hosts that have been observed in specimens from the Birdsong Shale. The number of specimens without any visible host is also included.

Pleurodictyum lenticulare host and specimen count

Brachiopods

Atrypa oklahomensis - 7

Atrypina hami - 2

Dicoelosia varica - 1

Discomyorthis oblata  - 102

Kozlowskiellina perlamellosa - 1

Leptaena acuticuspidata - 3

Leptostophia beckii tennesseensis - 4

Levenea subcarinata pumilis - 1

Rhynchospirina formosa? - 1

Unidentified brachiopod - 13

Corals

Favosites foerstei - 1

Streptelasma strictum - 4

Crinoid Stem - 3

Bryozoa Undetermined - 2

Mollusc

Actinopterinea textilis - 1

Arthropods

Conchostracan - 1

Dalamanites retusus - 1

Paciphacops logani - 1

Undetermined host - 4

No host - 40

Total number of  identified hosts  - 18       Total number of studied specimens - 193

Summary: Despite 18 identifiable hosts being observed, the data shows that the brachiopod Discomyorthis oblata was overwhelmingly favoured by P. lenticulare with 102 (representing 52.8%) of the 193 studied specimens belonging to this species. Far behind this, the second most abundant host was another brachiopod, Atrypa oklahomensis, but with just 3.6% of identified hosts representing this species. 

Discussion: At all the Birdsong sites studied, A. oklahomensis accounted for more than 30% of the brachiopod macrofauna, with Discomyorthis representing 18 to 20%. When viewed as a percentage of the total identified macrofauna* these numbers were approximately halved, with Discomyorthis dropping to a level of around 10%. 

[*Microfossils were not included in the analyses, and bryozoans were almost entirely excluded due to the need to make thin sections for accurate identification (something that had not been done at the time specimens were counted) but if they had been included this 10% figure would be reduced even further. Inclusion of the microbrachiopods would also have reduced this number, but assessing the relative abundance of micro species vs macro species requires controlled sampling that was not possible at the time specimens were collected]. 

Comparing the relative abundance of Discomyorthis in the Birdsong fauna to the percentage of P. lenticulare specimens that used the brachiopod as a host produces a result that cannot be reconciled unless Pleurodictyum lenticulare was selecting the brachiopod Discomyorthis as a host. If settling was entirely random, one would expect the number of Discomyorthis specimens that acted as hosts to Pleurodictyum to more or less match the relative abundance of the species in the macrofauna as a whole (~10%), but this is clearly not the case. 

Further analysis of the data does not produce many meaningful results because the majority of the remaining identified species only occur once. However, the other most frequently used host species are all closely correlated to their relative abundance in the Birdsong Shale – Atrypa oklahomensis being the dominant brachiopod and Streptelasma strictum being the most common coral and third most abundant species after Atrypa and Discomyorthis respectively. This trend, as far as it can be tracked with the relatively low number of specimens available, does continue and is what would be expected if polyps were settling at random. Only with Discomyorthis does there appear to be any selectivity at all. 

But why?

With the exception of the rare Leptostrophia brachiopods, Discomyorthis was among the larger species on the shelf but more interestingly had a fairly smooth, large surface area. There are other brachiopods of similar size (such as Leptaena and Atrypa) but they differ to Discomyorthis in that in Leptaena the valves are strongly rugate, while in Atrypa the more strongly curved and often with margins that are strongly flexed. Only in Discomyorthis was the entire surface of the brachiopod relatively smooth across the entire width of the valve, including the commissure. Was Discomyorthis chosen because the valves offered an unobstructed area over which the coral colony could grow? What is perhaps more likely is that Pleurodictyum favoured the smoother surface of Discomyorthis. This idea is increasingly possible when one considers that after Discomyorthis, Pleurodictyum colonies without a host are the most commonly occurring. There would be little to choose between the ocean floor and a smooth shell if the seabed was sufficiently compact, and it may have been this was all that the polyp was searching for.

When the orientation of Pleurodictyum on the host was analysed, it was found that with the exception of Dicomyorthis the part of the host on which the polyp settled was entirely random – it could be on the outer surface of an articulated brachiopod, the inner or outer surface of a loose valve, at the edge or in the centre of the host etc. Only in Discomyorthis was there any evidence that the orientation of the valve was important. Based on the Birdsong specimens, the polyp was overwhelmingly selecting the outer surface of juvenile living Discomyorthis valves. This is determined by the fact that:

1.) Only 10% of specimens are attached to adult valves.

2.) Brachiopod valves, due to their shape, when disarticulated settle on the ocean floor with the curved outer surface facing into the mud – this being the most stable position for the saucer shape in a water current. If Pleurodictyum was just selecting Discomyorthis valves with no preference for living or dead animals then many more specimens would be found with the coral attached to the inner surface of the valves, but this is not the case as more than 90% of specimens show attachment to the outer surface – a percentage that could only occur if the brachiopod was alive when attachment took place. 

Attaching to a juvenile living host must have benefited the polyp in some way. Symbiosis is possible as the feeding habits of the brachiopods would have brought food particles in reach of the polyps while they grew. In return the coral could have provided some protection to the brachiopod if the polyps like other anthozoans contained stinger cells. Regardless, the symbiotic relationship only last until the coral was able to fend for itself. The 90% of specimens attached to juvenile brachiopods all show the same thing – they are either in the process of or have completely smothered the brachiopod - the entire commissure being covered by the theca of the coral, which has also continued to grow laterally far beyond the margins of the host’s shell so that there was no way the brachiopod could survive. What may have started as a symbiotic relationship soon became parasitic. The brachiopod valves show no evidence of damage or fracturing to suggest they were killed by storm activity, and in all instances look to have been killed by the growth of the coral. Discomyorthis specimens of all sizes are represented in the Birdsong Shale, with no evidence juveniles dominated, indicating Pleurodictyum was deliberately targeting the younger animals. Why it did so is difficult to answer as settling on an adult brachiopod would have provided a ready-made base across which the coral could grow to adulthood without smothering the host. Instead, 90% of Pleurodictyum colonies that grew on Discomyorthis killed the host. Killing the host may not have been deliberate but simply a result of the faster growing coral overwhelming the brachiopod, but again why weren’t adult specimens chosen more frequently. Was Pleurodcityum just mean?

What is perhaps more likely is that Pleurodictyum was attracted to the brachiopod by something that is not preserved in the fossil record such as the colour of the valves, or chemical stimuli, but this cannot be tested. It may be that Atrypa oklahomensis was producing a chemical that repelled Pleurodictyum and the coral’s polyps actually settled for the second best choice of large and abundant substrate on the shelf, which was Discomyorthis. Again though, this cannot be tested at the present time.  

The evidence as it currently stands is that Pleurodictyum lenticulare was for a statistically significant part of the time selectively choosing Discomyorthis oblata as a host. 

This fact raises the question of whether other coral species on the Ross Shelf were doing the same thing. Favosites conicus, the most abundant tabulate coral on the Ross, has a growth pattern and colony size similar to that of Pleurodictyum and it too has been studied to see if it also showed a preference for a particular host as it may confirm some of the findings or suggest alternative explanations for host vs symbiote/parasite relationships. 

Favosites conicus is the second most abundant coral on the Ross Shelf (after the rugosan Streptelasma strictum). It is the dominant tabulate coral found in the Birdsong and exhibits a growth pattern similar to that of Pleurodictyum in that it settled on the substrate and grew rapidly both laterally and to some extent vertically in order to keep the colony above the ocean floor. Like Pleurodictyum, F. conicus has a basal theca, which often allows the initial attachment point of the coral to be determined, be it a short stalk, host, or simply resting on the sea bed. 

The bases of over 100 specimens of F. conicus collected in the Birdsong Shale were examined, and in complete contrast to the Pleurodictyum, not a single specimen appears to ever have been attached to a host. In fact, in most instances all that is visible is a small cluster of circular voids in the centre of the theca that obviously were the point of initial colony formation. Even under magnification there is no evidence to suggest the coral was ever attached to a host. This result was somewhat surprising as it was thought that there would be at least some similarities between Pleurodictyum and F. conicus. The fact that there is no evidence of any hosts may suggest that perhaps the coral was deliberately choosing the sea bed and relied entirely on lateral/vertical growth to keep it from sinking further into the mud (see article on the presumed monocarpous lifestyle of F. conicus for more information). If this is the case, the one feature the two corals do share is a preference for smoother substrates – be it harder muds of the sea floor or brachiopods like Discomyorthis.

To try and determine the exact life habits of F. conicus a population of the coral from the corelated Bois d’Arc Formation in Oklahoma was studied. [See Bois d’Arc article for how these two deposits are correlated]. 330 specimens of F. conicus from the Cravat Member of the Bois d’Arc Formation were examined, and produced the following results:

No visible host – 163 specimens

Huntoniatonia oklahomae – 29 specimens (11 cephalic, 9 pygidial spines, 6 thoracic, 3 pygidia)

Strophonella bransoni – 29 specimens

Stropheodonta sp. – 20 specimens

Discomyorthis oblata – 9 specimens

Meristella atoka – 9 specimens

Atrypa oklahomensis – 6 specimens

Bryozoan branches (thin types) – 6 specimens

Orthostrophia strophomenoides parva – 6 specimens

Favosites conicus – 5 specimens

Actinopterinia textilis - 4 specimens

Leptaena acuticuspidata – 4 specimens

17 additional hosts ranging from 1 to 3 specimens.

The results appear to contrast greatly with those seen in the Birdsong Shale as even though many specimens again showed no evidence of being attached to a host it only amounted to 49.4% of the total, with more than half of all specimens showing attachment. However, of these, only 7 specimens preserved a complete host in the theca of the coral (4 of which were the microbrachiopods Coelospira and Dicoelosia, and the other 3 small specimens of the coral Streptelasma strictum), with the remaining 96% of specimens preserving only a fragment of the original host. The fact that in almost all instances the coral has fully grown beyond the edges of the host the specimens indicates the host was fragmentary when they were originally settled upon and have not been broken before or after fossilisation. 

From the data, colonies of F. conicus in the Cravatt were either settling on shell hash or the sea floor the majority of the time, and seemingly ignoring any living host. Of the hosts it was choosing, the preference was for larger, flatter pieces of shell hash – as evidenced by the use of the brachiopods Strophonella, Stropheondonta, Discomyorthis, Meristella and Atrypa, and the larger trilobite Huntoniatonia. These 6 species account for 62.5% of the host population and in all of them the shells have large, flat or gently curved surfaces lacking strong ridges or uneven surfaces. Few of the pieces of identified brachiopod observed as hosts represent the edges of the shells where the curvature could be greatest, and in Huntonitonia, the preference was clearly for the glabella and free cheeks. [The larger than expected percentage of coralla that used the long, thin pygidial spine of Huntoniatonia (31% of all trilobite specimens) seems anomalous as in contradicts the evidence from the brachiopods and can’t be explained at the present time]. This preference for smooth shells or the sea floor was also a feature seen in Pleurodictyum lenticulare.

Conclusion: Although it may initially seem that there was a great difference in the life habits of the Birdsong versus Cravatt specimens of F. conicus, this is really not the case. The differing palaeoenvironmental conditions between the two localities certainly had some effect but in reality in both the Birdsong and Cravatt F. conicus was using a non-living host as an anchor point. As was seen in Pleurodictyum, no visible host and shell hash are both examples of the coral polyps simply settling on what was the ocean floor – be it a muddy substrate or one with a higher amount debris. The Birdsong environment was shallower, and more frequently effected by wave and current activity, resulting in a physically smaller variety of F. conicus compared to that found in the calmer conditions of the Cravatt (that reduced the need for a monocarpous and produced the larger average colony size that is recorded in Oklahoma) but the life habits of both were practically the same.

F. conicus does show some selectivity in that both populations were choosing smoother surfaces, but the coral certainly was not as selective as Pleurodictyum. There is no evidence that F. conicus choose a living host, something that has been shown for P. lenticulare.