The roughly three million year long, end-Ordovician (late Katian-Hirnantian) mass extinction was marked by at least three major glacial episodes on the North Gondwana Platform, extending from Morocco to Oman, and including Spain and Portugal [1]. Adjacent cold climate areas in the southern hemisphere were also glaciated [2,3], or had extensive marine ice shelf coverage [4]. Latest Ordovician (late Katian and Hirnantian) glaciations were accompanied by sealevel lowstands that exposed much of the paleocontinent of Laurentia, resulting in extensive hiatuses in the fossil record of reefs. However, in areas on the continental margins of Laurentia, with inner to mid-shelf facies, reefs developed within the mid to late Hirnantian [5,6] leaving a record of the coral and stromatoporoid faunas of the end Ordovician [7]. This Hirnantian reef fauna essentially vanished with only a few hold-overs, such as the phaceloid reef-builder Palaeophyllum. Anticosti Island (Figure 1), with a virtually uninterrupted 800 m thick section across the Ordovician-Silurian boundary, is an ideal location to study extinction and recovery impacts in the reef setting.

The earliest Silurian (Rhuddanian) reef record is very sparse worldwide [8,9,10]. One of the few places where earliest Rhuddanian coral reefs are present is at the eastern end of Manitoulin Island, Canada, in the Michigan Basin [11]. Bergström et al. [12] have suggested that the Manitoulin reefs were late Hirnantian in age. Nevertheless, the Manitoulin reefs lack the distinctive aulaceratid-clathrodictyid sponge and tetradiid-rugose coral faunas typical of the Hirnantian reefs of both Anticosti and Estonia. The Manitoulin reefs also overlie the occurrence of the brachiopod genus Zygospiraella, diagnostic of the early Rhuddanian of Anticosti Island and the Baltic region. Middle- to late Rhuddanian reefs are known from the Baltic region, where they are assigned to the Tammiku Member (Tamsalu Formation) and the Hilliste Formation [13,14]. These Rhuddanian reefs, of which there were three types in Estonia, are not presently described in great detail. The tabulate corals (initially auloporids with algae in the middle-late Rhuddanian Tammiku Member, then primarily favositids and halysitids in the Hilliste Formation of late Rhuddanian to early Aeronian age) formed a major role [15]. Colonial rugose corals were also reef dwellers (e.g., Paliphyllum, Palaeophyllum, Cyathophylloides, Petrozium, Entelophyllum), and have been described by Kaljo [16]. The coral fauna characteristic of the Rhuddanian Becscie Formation of Anticosti is similar to that of the Manitoulin reefs and the Rhuddanian reefs of Estonia. The whole Rhuddanian sequence of Estonia is no more than 64 m thick: on Anticosti the Rhuddanian is about 200 m thick.

At present, there is a lack of detailed knowledge about most of the earliest Silurian (Rhuddanian) reefs in terms of their faunal composition (such as what phyla and genera dominated), and structure (how these constructed or modified the reefs). On Anticosti, the Rhuddanian upper Becscie Formation (Chabot Member) and succeeding Merrimack Formation developed local clusters or concentrations of the phaceloid rugosans Nanophyllum, Palaeophyllum and Donacophyllum, but these did not build reefs. Even the early through mid- Aeronian Gun River and lower Menier formations show no reef growth. The late Aeronian East Point reefs are younger than those of Estonia, have a different fauna, and mark the first Silurian reef development on the Anticosti Platform. We provide information on the ‘first arrivals’ of taxa, in terms of reef builders for the Rhuddanian and Aeronian. For example, when or where precisely did the Silurian reef-builders (e.g., alveolitids, coenitids, multisolenids and pachyporids) originate, that subsequently dominated reef settings in the Middle Silurian (Wenlock) through Devonian? Did some of these arise in the transitional latest Ordovician (Hirnantian) stage, or were they a newly evolved component of the later recovery fauna? What is the fossil record of bio-eroders?

This paper is a contribution to the theme of recovery of reef taxa and reefs after the end Ordovician mass extinction events, as seen in the Anticosti Basin of eastern Laurentia. It records the transitional coral and stromatoporoid fauna, and associated reef inhabitants that marked reefs of Aeronian (middle Llandovery) age on the margins of the Laurentia paleoplate. These earliest Silurian reefs were thus built at a time of global re-warming and rising sealevels, following a Late Ordovician ice age that lasted about the same length of time as the Pleistocene glacial episodes, i.e., 3 million years. Coral and stromatoporoid biodiversity increases on Anticosti Island began already in middle Rhuddanian time, with rising sealevel and warming, as seen in the faunas of the Becscie and Merrimack formations [17,18]. Nevertheless, no reefs developed in the Anticosti section until deposition of the carbonates the East Point Member, in the Menier Formation, some 300 m above the base of the Early Silurian section. Compared with Estonia in the Baltic Basin, on the opposite side of the ancient Iapetus ocean [14,16,19], the Anticosti Silurian carbonate section is three to four times thicker, and fills in, for example, the faunal gaps of the more condensed, and hiatus-prone, Estonian sequence [7].

All samples mentioned or illustrated in this study are in the repository of the Geological Survey of Canada, Ottawa, with catalogued samples denoted by the prefix GSC, and locality numbers preceded by the letter “A”.

In 1856 James Richardson departed for an epic boat journey around Anticosti Island in the Gulf of St Lawrence, charged with mapping and measuring stratigraphic sections, and collecting fossils to establish their age [20]. In the process, he was the first to discover, and briefly describe, the fossil reefs at East Point (Figure 1), which he assigned to the top of his Division D. He described these strata as “25 ft, … charged with a multitude of corals…the bed assumed a hummocky character… some patches of the corals rising from one to five feet high”. Schuchert and Twenhofel [21], following field work from a boat in 1906, re-examined the East Point outcrops along the north coast, and were the first to note that this unit was “a coral reef… essentially of corals”, but they mistakenly correlated the strata to near the top of their freshly dedicated Gun River Formation, named from the south coast some 120 km to the west. Later, Twenhofel [22] placed the East Point ‘extensive coral reef’ at the base of the Jupiter Formation and confirmed that stratigraphic position for his “coral reef limestone of variable thickness” [23]. Schuchert and Twenhofel were never able to visit the 3–4 km long overhanging Goéland cliffs west of East Point, as these are located along a dangerous storm-swept part of the coast that is poorly accessible except by zodiac or helicopter, even today. Nor did they land at what is today called the “Baie aux Goélands”: if so, they would probably have understood that the base of their Jupiter Formation lay well below the East Point reefs.

In 1972 Bolton went to the east end of the island by float plane and examined the East Point reefs and the recessive shales above them, allocating the reefs to the base of the Jupiter Formation (as had Twenhofel in 1921 and 1928 before him). Neither did Bolton visit the Goéland cliffs section to the west. On the other hand, using helicopter support, Petryk [24,25] placed the East Point reefs at the top of the Gun River Formation, and used that to draw the boundary between the Gun River and Jupiter formations at the north side of the island. Thus the East Point reefs have had a mixed correlation history in the past. Copper first visited the East Point section by boat in 1978 (and subsequently), collecting samples from the reefs. Copper and Long [26], after spending several field seasons by boat circumnavigating the island, established the East Point reef unit as an independent member of the Jupiter Formation, which was reconfirmed in Copper and Brunton [8]. The East Point Member, with its base some 55 m above the base of the original Jupiter Formation, overlies the easily identifiable recessive shales with rich in situ pentamerid brachiopod faunas of the Goéland Member below it [27,28]. Since 1984, we have been able to map the reefs far inland by road, and their equivalent crinoidal grainstones some 160 km to the west near Cape Macgilvray, including outcrops along the coast and rivers [5] (Figure 1). Faunal correlations of the reefal stromatoporoids, e.g., the possibly endemic genus Tarphystroma [7] and rugose corals, with the presence of a Siberian genus, Palaearaea [29], confirm a late Aeronian age for the East Point Member. Apart from some preliminary data on the distribution of the East Point reefs as part of field guides [30,31], this is the first time the East Point reef-builders and their associated facies are described in some detail.

We herein include the Aeronian East Point reefal unit (Figure 2), and the Goéland Member below, in a new Menier Formation. In accordance with North American Stratigraphic Code, the definition of the Menier Formation is outlined as follows:

The Aeronian reef belt stretches from East Point itself, the most easterly protrusion of Anticosti Island, where it outcrops in the bluffs and along the tidal flats, to as far west as the Box River, some 70 km away (see Figure 1). The crinoidal meadow facies of the East Point Member thins to <1 m in coastal outcrops west of the Jupiter River. Reefs are thickest and most prominently and three-dimensionally exposed along the cliffs and tidal flats at, and to ca. 6 km west of East Point (Figure 3, Figure 4 and Figure 5). They taper out below overlying shales of the Richardson Member up section to the south around East Point (Figure 3e). The reefs at East Point are 5–10 m thick overall, but may be thicker locally. Diameters range from about 10 to 60 m in this area, but seem to be greater inland along the main road, ca. 40 km to the west and north. Some of the reefs are stacked multiple-levels in the bluffs and cliffs (Figure 4c and Figure 5), and others are displaced laterally above underlying reef levels. This suggests multiple reef-building events, and discontinuous growth, or amalgamation of adjacent and overlying reefs, over a wide area. In the tidal flat sections at East Point, the reef bases are not exposed, but they lie over a ca. 1 m thick grainstone bed (Figure 4c). They are only rarely and partly accessible in the lower bluff and cliff faces, and via erosional collapse of the cliff face in the coastal cliffs to the west (Figure 3f and Figure 5b).

The foundations of the reefs are better exposed in two road-cuts along the main road, showing also the reef bases, cores, tops and flanks beds, ca. 2 km to the east of the Box road junction (e.g., outcrops A863, A1198, A1199, with about 200 m separating the two reefs). In the Chaloupe River to Ferrée River areas, reefs give way down the carbonate ramp to both crinoidal or biostromal facies of less than 3 m thickness. The biostromal facies is well displayed along the east bank of the Dauphiné River (A1204, 12E/1, 42350:48280) and on the Petite Rivière de la Chaloupe (A1223, 12E/2, 34700:53580). West of the Ferrée River, towards the mouth of the Jupiter River and Triplesia Creek, and inland along the Firetower road, only the crinoid meadow facies is evident. The East Point Member there shrinks to <1 m in thickness; locally it may even be so thin and poorly exposed as to become difficult to find in road outcrop.

The three-dimensional outcrops of patch reefs along the East Point tidal flats point to two aspects: (1) most of the reefs have an irregular to circular outline, with diameters ranging from about 3 m to 50–60 m, as seen from above by helicopter and from the bluff tops (Figure 3 and Figure 4), and (2) they are relatively closely spaced, 10–50 m apart (Figure 3, Figure 4 and Figure 5). Some reefs that started separately appear to have amalgamated into a larger cluster (e.g., Figure 4c). Individual patch reefs appear to lack a definable windward or leeward side, in terms of a fore-reef rubble zone, or less developed leeward side. The patch reef complex as a whole, and individual reefs, show no orientation or elongation into a possible current or windward wave direction and dispersal appears to be random, and not in any way aligned. This suggests that for the East Point coastal area itself, there was no tidal influence, such as tidal channels or windward wave direction, which affected reef shape, orientation or density (as seen in the much younger Telychian reefs of the Chicotte Formation). This suggests that at the east end of the island, waters were probably slightly deeper and less exposed to storm effects, possibly growing at depths of 10 m or more. Comparable patch reef patterns today might be those of the inner Swain Reefs of the Great Barrier Reef shelf [32]. Alternatively, the East Point reef pattern resembles that seen in the lagoons of the Maldives in the Indian Ocean today [32], though on Anticosti there is no evidence for a protective back-reef barrier in any direction. However, reef tops and flanks show breakage of large coralla and stromatoporoids and reef rubble in the Box River-Dauphiné area, suggesting that these were in shallower waters of the inner shelf, more affected by storms and tidal currents. None of the reefs grew into the shallowest subtidal zone: no storm channels, spur and groove structures, or erosional discontinuities were evident, as such are seen in the much younger Chicotte reefs of Anticosti [33].

Reef bases appear to have been generally flat-bottomed, as seen in road outcrops and coastal bluffs. Reefs appear to have been too small to produce depressions in the underlying carbonate basal beds, such as the ‘philip structures’ seen in the late Llandovery of Gotland, Sweden.

The East Point reefs show a mixed fabric produced by frame-building constructors, encrusting and binding skeletons, bafflers, and interstitial sediment, primarily of carbonate silt to sand grade. Coarse reef rubble fabric (rudstone) was developed along the margins by storm-reworked carbonate-secreting builders, and fine to coarse-grained carbonate sediment (siliciclastics are absent). Some of the clasts were deposited between the builders in a mix of fine “polymud” to sand-sized fragments: cavities were infilled by layers of fine sediment, including peloidal grains, and a few show encrusting cryptic biota (bryozoans or enigmatic biota) that were suspended from cavity roofs. Cryptic “lithistid” demosponges, displaying internal spicules, were attached to the sides and in the cavity space between branching rugosans. Endosymbiont embedment structures (bioclaustrations, such as Chaetosalpinx and cornulitids) and borings (bio-eroders) were a common feature in the reef biota, but usually in the denser domal skeletons. Indeed, borers were more intense in off-reef stromatoporoids, producing a “Swiss cheese” fabric [7]. Intimate intergrowths of stromatoporoids with rugose corals such as Palaeophyllum were also a feature first seen on Anticosti in the East Point reefs at Box River [7].

Quantitative estimates of the various components are in general difficult to obtain (a) since almost every reef demonstrates different dominant components and fabrics, (b) such components vary through the reef from base to top or from core to flanks, and (c) for the smaller components, the percentage volume varies from thin section to thin section, and even within thin section, or from slab to slab. The East Point reefs cannot be classified as “mudmounds”, as muddy fractions formed only a very small part (<3–5%) of the reef fabric as a whole. Cavities and reef interstices of the reef core and flanks usually remained unfilled by contemporaneous or post-depositional cement; thus the East Point reefs are generally cement-poor. The East Point reefs also differed substantially from the younger Chicotte diagenetic, mud-dominated, bryozoan-sponge ‘buildups’ [34]. Interstitial fabrics of the remaining lower and upper Chicotte reefs are also different, being largely dominated by crinoidal grainstones and show common radial-fibrous cement (instead of abundant cavity space as in the East Point reefs), and more abundant East Point microbial and sponge encrusters (see Figures 6a,b, in the dark micritic crusts around Entelophyllum). East Point reefs have little in common with the older, Hirnantian Laframboise reefs of Anticosti in their biota, and they featured less calcimicrobial fabric such as Girvanella, seen in the Ellis Bay Formation, and lacked the common Ellis Bay red alga Solenopora. The East Point reefs also had no basal oncolite platform bed.

The East Point reefs were primarily constructed (1) via the upward individual growth of large (40 cm in diameter) coral skeletons, such as the overall domal forms of large cerioid Favosites, cateniform Cystihalysites, the aphroid rugosan Palaearaea, or phaceloid rugosan Entelophyllum (proportions differing in almost every reef, but primarily in reef tops and flanks), and (2) by vertical and lateral stacking of layers of smaller overgrown corals and stromatoporoids. These commonly roofed over cavities between the stacked skeletons or between adjacent intergrown skeletons. Mutlisolenia is relatively common, but formed small-sized colonies (<10 cm in diameter) in the East Point reefs. The bases of the reefs were established, where this could be determined, on meadows of crinoids, or a hash of broken small Catenipora and especially small syringoporids. Thus the term ‘coral patch reef’ would be the most appropriate description.

The East Point reef bases occur as outcrops at relatively few localities, and reef foundations were variable from reef to reef (as were the major reef builders). In some instances, the basal components did not lead to reef growth. Where the reef bases were exposed, they consisted of variable clasts ranging from crinoid ossicles (maybe the most common reef base), brachiopod shells (Stegerhynchus being most common), paleoporellids (Figure 10h), or tabulate coral hash of Catenipora (Figure 6f), and Syringopora. (Figure 9c). In the shallower water areas of the East Point facies at the northern edge of the outcrop belt these resulted in reef growth in deeper water facies to the south biostromes resulted, and the member thinned to <1–2 m.

In terms of environmental controlling factors, the reef-building intervals in the Upper Ordovician-Lower Silurian succession of Anticosti Island show a broad correlation with episodes of sealevel fall and marked positive excursions of δ¹³C and δ¹⁸O excursions [40]. On Gotland, a similar correlation between reef-building and positive C and O isotope excursions has been noted for the Wenlock and other horizons, which led to the observation that, in the Silurian low paleolatitude, periods of such positive isotope excursions are characterized by the growth of reefs and the formation of extended carbonate platforms [41,42]. It remains debatable if, during a general supergreenhouse climate of the Late Ordovician-mid Silurian, relative cooling (e.g., Hirnantian and Sheinwoodian glaciations) could create optimal conditions for reef growth in tropical latitudes by: (1) lowered super-heated ocean water temperature; (2) a drier tropical climate; (3) improved ocean circulation due to strengthened downwelling and psychrospheric flow generated around ice-capped polar regions; and (4) lowered sealevel encouraging shallow reef growth. On the other hand, increased phytoplankton production with nutrient upwelling during cooler episodes (triggering black organic-rich shale formation, as in the Calvert-Pedersen model) might have constrained reef size and distribution in the Hirnantian. Organic rich graptolitic shales of the deeper water Richardson Member cap the East Point reefs, and mark their demise. It is notable that in many areas worldwide, the Hirnantian, such as in the stratotype section of the O/S boundary at Dob’s Lin, is marked by organic rich black shales. These are absent at the O/S boundary section of Anticosti.