The Miguasha Fossil-Fish-Lagerstätte: a consequence of the Devonian land–sea interactions more

Palaeobio Palaeoenv (2011) 91:293–323 DOI 10.1007/s12549-011-0058-0 ORIGINAL PAPER The Miguasha Fossil-Fish-Lagerstätte: a consequence of the Devonian land–sea interactions Richard Cloutier & Jean-Noël Proust & Bernadette Tessier Received: 1 June 2011 / Revised: 18 August 2011 / Accepted: 24 August 2011 / Published online: 11 October 2011 # Senckenberg Gesellschaft für Naturforschung and Springer 2011 Abstract The evolution of vertebrate assemblages in terms of fluctuating environments has rarely been investigated for the Devonian period. Variation of biodiversity (richness, abundance and species composition) in the diverse Devonian fish assemblage of the Escuminac Formation (Quebec, Canada) is analysed in response to changes in lithofacies, depositional environment and taphonomy through time. Five sequences within an inner wave-dominated estuary show shifts in continentalisation. Although a ubiquitous fish assemblage is identified throughout the formation, species are more diversified and species composition is better structured during relative sea-level rise than during still-stand and relative sea-level fall. Konservat and Konzentrat FossilLagerstätte horizons occur in the transgressive phase of the sequences. Keywords Taphonomy . Sequence stratigraphy . Palaeoichthyology . Devonian fish . Lagerstätte R. Cloutier (*) Chaire de Recherche en Paléontologie et Biologie Evolutive, Université du Québec à Rimouski, 300 allée des Ursulines, Rimouski, Québec G5L 3A1, Canada e-mail: richard_cloutier@uqar.qc.ca J.-N. Proust UMR CNRS 6118, Géosciences, Université Rennes 1, Campus de Beaulieu, 35042 Rennes, France e-mail: Jean-Noel.Proust@univ-rennes1.fr B. Tessier UMR CNRS 6143 M2C, Université de Caen, 24, rue des Tilleuls, 14000 Caen, France e-mail: bernadette.tessier@unicaen.fr Introduction Major evolutionary events have shaped Devonian palaeodiversity in terms of richness, abundance and species composition. Such events include the diversification of land plants (Davies and Gibling 2009; DiMichele and Bateman 2005; Kenrick and Crane 1997), the establishment of primary forest (Algeo and Scheckler 1998; Bateman et al. 1998; Stein et al. 2007; Young 2007), the diversification of land invertebrates (DiMichele et al. 1992; Shear and Selden 2001) and aquatic lower vertebrates (Anderson et al. 2011; Cloutier 2009; Klug et al. 2010; Sallan and Coates 2010; Zhu et al. 2009), the origin of tetrapods (or limbed vertebrates) (Clack 2007), the largest development of reefal ecosystems in Earth history (Copper 2002), and one of the five major mass extinction events (McGhee 2001; Sandberg et al. 2002). Major environmental changes have also marked the Devonian, including the development of soil on land (Algeo and Scheckler 1998; Algeo et al. 2001; Davies and Gibling 2009), low atmospheric O2 levels (Algeo and Ingall 2007), global marine anoxic events (Joachimski and Buggisch 1993), enhanced organic carbon burial rates (Algeo and Scheckler 1998) and major sea-level changes (Haq and Schutter 2008; Johnson et al. 1985). Many of these biological and environmental changes are related to the continentalisation of ecosystems. The establishment of a continental plant ecosystem most likely altered profoundly the sedimentary regime of the coastal and marginal palaeoenvironments (Davies and Gibling 2009). One of the main critical interfaces is the estuarine environment, which has recorded continental and marine fluctuations. Furthermore, fluctuating environments, such as estuaries, are considered to be favourable to evolutionary changes for many reasons (Kussell and Leibler 2005; Levins 1968; Meyers and Bull 2002). Nevertheless, only a 294 Palaeobio Palaeoenv (2011) 91:293–323 few Devonian transitional, paralic or estuarine macrofaunal assemblages have been reported and documented (Blieck and Cloutier 2000; Cloutier et al. 1996; Selover et al. 2005; Wehrmann et al. 2005). Fluctuation of a Devonian vertebrate assemblage within a biostratigraphic unit has rarely been studied, in contrast to assemblages of Devonian marine sessile invertebrates, for which temporal and compositional variation have been well documented (Brett 1995; Brett et al. 2007). The mobility of vertebrate organisms during their life, the complex preservation of the vertebrate skeleton and the relative smaller sample size of vertebrate specimens (in comparison to invertebrate fossils) make the study of their variation more difficult. The Upper Devonian Escuminac Formation from eastern Canada (Fig. 1) offers a near perfect setting to evaluate empirically the relationships between palaeodiversity and palaeoenvironment in transitional environments. The Parc national de Miguasha, which protects the Escuminac Formation, has been included on the list of natural sites of the UNESCO World Heritage as being the palaeontological site most representative of the Devonian period. This inscription recognises the distinctiveness of the Escuminac assemblage in terms of: (1) its faunal representativeness of major groups of sarcopterygians (Table 1), (2) the representativeness of vertebrate evolutionary events (Cloutier and Lelièvre 1998), (3) the floristic and faunal representativeness of aquatic and continental assemblages (Cloutier et al. 1996), (4) palaeobiological representativeness [e.g., presence of ingested prey (Arsenault 1982; Arratia and Cloutier 1996; Cloutier 2009) and presence of fossilised ontogenies (Cloutier et al. 2009; Cloutier 2010)], (5) the quality of preservation in terms of anatomical completeness (Parent and Cloutier 1996), (6) the quality of preservation in terms of exceptional fossilisation (e.g., soft tissues; Janvier and Arsenault 2009; Janvier et al. 2006, 2007) and (7) the abundance of specimens (Table 1). The palaeobiological, taphonomic and palaeoenvironmental peculiarities of the Miguasha Lagerstätte allows us to characterise different levels and scales of variation in palaeodiversity. Variation in palaeodiversity is investigated here to identify potential sources of variation and/or bias that affected the richness (number of species), abundance (number of individuals of a species) and composition (associations among species) of the faunal assemblage. According to Brett et al. (2009), it is critical that taphonomic and palaeoecological studies of extraordinary fossil assemblages be placed in a sequence stratigraphic context, as this is the only way to gain a predictive understanding of where and when such biotas may occur. We will address the questions of whether the Escuminac palaeodiversity signature is lithofacies dependent, taphofacies dependent or sequence-dependent. The primary objectives of this study are: (1) to reinterpret the palaeoenvironments of the Escuminac Formation and their stacking patterns in terms of depositional sequences; (2) to describe the fluctuation of the palaeodiversity of the Escuminac Formation in terms of richness, abundance and composition; (3) to explore temporal and compositional variation through the Escuminac Formation taking into account palaeoenvironmental, lithological and taphonomic factors. Material and methods Stratigraphic sampling was carried out bed-by-bed throughout the Escuminac Formation on the east side of the syncline by one of the authors (RC) and completed partially by Miguasha national park personnel for the period between 1996 and 2008. The main strata on the east side of the syncline of the Escuminac Formation (Fig. 1) have been numbered from 1 to 394. In the section, each bed, recognised as a lithological change (Hesse and Sawh 1992), was assigned a number, where bed 1 is the first bed at the base of the Escuminac Formation, in paraconcordance with the Devonian conglomerate of the Fleurant Formation, and bed 394 corresponds to the last bed, in angular unconformity with the Carboniferous conglomerate of the Bonaventure Formation (Fig. 2); some beds have been subdivided (e.g. bed 21.1). A total of 273 different levels (some beds having up to 7 levels) out of 394 beds from the Escuminac Formation have been sampled. There have been approximately 60 days of fieldwork during the summer of 1993 and 1994. On average, each bed has been sampled for approximately 2–3 hours. Data for each bed were collected on an approximately 2-m-wide section throughout the thickness of the bed; specimens have been stored at the Musée d’Histoire Naturelle de Miguasha (MHNM). Sequences 4–6 have been sampled less due to the accessibility problem. Sequence stratigraphy was interpreted by one of us (JNP) based upon several features, including (1) patterns of upward change in lithology, granulometry (fining/coarsening upward trends), sedimentary structures and bedding (thinning/ thickening upward patterns); (2) evidence of inferred changes in water depth and lateral shifts in depositional environments; (3) tracing of sharp erosional surfaces and apparent flooding surfaces. Five different lithofacies are encountered in the Escuminac Formation (Fig. 2), namely, conglomerate (0), sandstone (1), siltstone (2), laminites (3) and shale (4). A decreasing granulometric gradient from 0–4, referred as the lithographic gradient, is used in the various analyses. The total animal richness is evaluated as the fish richness plus the invertebrate richness per bed. The presence of coprolites and their respective bony inclusions (if identifiable) were taken into account in the evaluation of the fish richness. The abundances of fish, conchostracans and plants illustrated in the stratigraphic column have been coded in Palaeobio Palaeoenv (2011) 91:293–323 66 30’ o 295 66 15’ o 48 10’ 48 10’ o o b Nouvelle Escuminac Pointe Fleurant Pointe-à-la-Garde Miguasha wharf Dalhousie Québec Pointe-à-la-Croix 48 00’ o Pointe Yacta 66 15’ o Campbellton New-Brunswick 66 30’ o 0 5 km Bonaventure Formation Miguasha Group Pirate Cove Formation La Garde Formation Dalhousie Group Chaleurs Group a Bonaventure Formation NW c SE Québec N-B USA Escuminac Formation (modified from Hesse and Sawh 1992). c General view of the east side of the syncline of the Escuminac Formation representing the main part of the Miguasha section (see localisation in b) Fig. 1 Geographical and geological maps of the Escuminac Formation, Miguasha, eastern Quebec, Canada. a Localisation of Miguasha, eastern Quebec. b Geological map of the Miguasha area; the Miguasha Group includes the Fleurant and Escuminac Formations terms of four categories (Fig. 2): absent (0), rare (1), common (2) and abundant (3). With regard to fish abundance, the four nominal categories correspond to an average number of specimens: 0, absent; 1, between 1 and 20 specimens; 2, between 20 and 50 specimens; 3, more than 50 specimens. The number of specimens per bed has been calculated based on the 334 specimens collected during the stratigraphic sampling plus an additional 1,121 specimens housed in the collection of the MHNM (as of October 2008) for which precise stratigraphic occurrence has been documented. The preservation of individual specimens of fishes has been classified into six taphonomic states, from poor (1) to excellent preservation (6). The taphonomic gradient is inferred from the descriptive and experimental work on the disintegration (or disarticulation) of fish carcasses (Elder and Smith 1984, 1988; Schäfer 1972; Smith and Elder 1985). The gradient emphasised the disintegration of floating carcasses because the scattering of bones by scavengers or the transport of elements by bottom current is minimal in the Escuminac Formation (Parent and Cloutier 1996). Fish specimens for which all bony elements are preserved in natural position and which do not show signs of decomposition are coded 6. Specimens coded 5 show signs of weak decomposition (mainly at the junction of bony elements suturing with cartilages). Specimens for 296 Table 1 Abundance and occurrence of the vertebrate species from the Escuminac Formation Taxon/species Abundancea Total Anaspida Euphanerops longaevus Endeiolepis aneri Osteostraci Escuminaspis laticeps Levesquaspis patteni Placodermi Bothriolepis canadensis Plourdosteus canadensis Acanthodi Diplacanthus horridus Diplacanthus ellsi Triazeugacanthus affinis Homalacanthus concinnus Actinopterygii Cheirolepis canadensis Actinistia Miguashaia bureaui Dipnoiformes Scaumenacia curta Fleurantia denticulata Porolepiformes Holoptychius jarviki Quebecius quebecensis Porolepiformes indet. “Osteolepiformes” Eusthenopteron foordi Callistiopterus clappi Elpistostegalia Elpistostege watsoni NA, Not applicable a Palaeobio Palaeoenv (2011) 91:293–323 Occurrenceb Total Relative Relative 24 66 57 9 6,726 145 18 6 4,278 941 352 31 2,265 24 15 32 2 3,064 1 3 0.13 0.37 0.32 0.05 37.24 0.80 0.10 0.03 23.69 5.21 1.95 0.17 12.54 0.13 0.08 0.18 0.01 16.97 0.01 0.02 6 5 5 2 63 4 4 NA 27 15 7 1 56 6 NA 3 1 47 NA 1 2.24 1.87 1.87 0.75 23.51 1.49 1.49 NA 10.07 5.60 2.61 maximum state of preservation per bed or multiple states for single beds were used. Relationships among parameters were evaluated with Spearman rank correlation (rs); most variables are not normally distributed and some variables are ordinal (e.g. abundance, lithofacies, preservation). Cluster analyses using the Dice (association index) similarity index with average linkage (Legendre and Legendre 1998) were calculated to recognise the composition of fish assemblages. The Dice index was preferred because it emphasises the double-presence of species between beds and minimises the effect of single-presence between beds because this difference might be completely or largely due to differences in sample size (Wolga 1981). To validate the robustness of clusters in a dendrogram, 10,000 bootstraps were performed on the original matrix. PAST version 1.56b (Hammer et al. 2001) was used to perform the cluster analyses, Kruskal–Wallis test, and Mann–Whitney test. SYSTAT ver. 9.01 (SYSTAT, Chicago, IL) was used to calculate descriptive statistics (i.e., mean, standard deviation, correlation coefficient). Lithofacies and depositional environments 0.37 20.90 2.24 NA 1.12 0.37 17.54 NA 0.37 The description of the lithofacies and depositional environments of the Miguasha Section is centered on the Escuminac Formation but includes the topmost Devonian Fleurant and the lowermost Carboniferous Bonaventure formations. It complements previous works from Hesse and Sawh (1992), Rust (1982) and Rust et al. (1989). Braided fluvial conglomerates Conglomerates are the least abundant lithofacies found in the Escuminac Formation (<0.1% of total thickness), but they are well exposed in the Fleurant Formation at the base of the Miguasha section (Figs. 1 and 2) in a 3-m-thick outcrop (Rust 1982; Rust et al. 1989) and at the top in the Bonaventure Formation. The 18-m-thick Fleurant Formation consists of a polymictic conglomerate of pebble size to large cobble size and occasional boulders, and lithic sandstone beds showing large-scale cross bedding and small-scale ripple lamination (Rust et al. 1989). Siltstone and mudstone are minor components. Pebbles and cobbles are moderately sorted and rounded and locally imbricated. They consist of limestone, with a lesser amount of sandstones, andesitic volcanics and minor plutonic rocks (Rust et al. 1989; Zaitlin 1981). The clasts were derived from older Devonian, Silurian and Ordovician (?) sedimentary successions located to the north and northwest of the Miguasha area. The Fleurant–Escuminac contact is sharp, with sandstone of the lower bed 1 of the Escuminac Formation moulding the upper surface of the underly- Species abundance (in number of specimens) is calculated based on previous museum survey (Parent and Cloutier 1996) plus new statistics from the Musée d’Histoire Naturelle de Miguasha (J. Kerr, personal communication). The relative abundance (in %) corresponds to the proportion on the total of 18,059 specimens b Occurrence corresponds to the actual number of beds from which the species has been found during the field sampling. The relative occurrence (in %) corresponds to the proportion on the total number of sampled beds (272) which only kinetic elements are missing (e.g. operculum, branchiostegal rays) are coded 4. Obvious signs of bacterial activity or fermentation, such as rupture of the abdominal cavity, are coded 3. Specimens represented only by part of the body are coded 2, whereas isolated bones (in most cases isolated scales) are coded 1. In the analyses, either the Palaeobio Palaeoenv (2011) 91:293–323 fl a ts 297 e ls Fr on Ch a 30 Siltst Sh Sdt Cgt To Ce n e m nn t tra lb a Fluvial bay head delta as & in 2 95 90 85 80 25 75 70 65 60 20 55 50 45 1 40 15 Abundance (nb of individuals per species) All species 0=absent; 1=rare; 2=common; 4=abundant Fish species 0=absent; 1=1 to 20; 2=20 to 50; 4= >50 35 Fish richness (nb of species) 1 to 7; exceptional value in circle Fish taphonomy 30 1 10 25 6 20 15 10 11 5 Sedimentary structures Ripple lamination Climbing ripple lamination Planar lamination Groove cast Concretion Paleo-current direction Fibrous calcite Slump Dish structures Lithologies Shale Rhythmite Siltstone Sandstone Conglomerate 5 Escuminac Fm. 0 Fleurant Fm. Fish taphon. 1 3 5 7 Fish richness Fish Asmusia Plant abundance Fig. 2 Detailed sedimentological exposed on the main section from Numbers on the left side of the Information (from left to right) is log of the Escuminac Formation René Bureau’s Cliff at Miguasha. log correspond to bed numbers. provided in five columns on fish taphonomy (six states from poor to excellent preservation; see text), fish richness, abundance of fish, conchostracans (Asmusia), and plants. The right column gives the depositional environment. The sedimentological log is modified from Hesse and Sawh (1992) 298 on t Palaeobio Palaeoenv (2011) 91:293–323 Ch an ne ls & m 60 Siltst Sh Sdt Cgt Fluvial bay head delta 240 235 230 225 220 55 215 Ce ntr al ba sin Fr To e b fla ts 210 3 50 205 200 195 190 185 45 180 175 170 165 160 155 150 40 145 140 135 130 35 125 120 115 110 105 30 100 2 Fish taphon. 1 3 5 7 Fish richness Fish Asmusia Plant abundance Fig. 2 (continued) Palaeobio Palaeoenv (2011) 91:293–323 fla ts t 299 Fr on an m Siltst Sh Sdt Cgt Ch Ce Fluvial bay head delta ntr al ne c ls & To ba sin e 5 90 355 350 85 345 340 80 335 4 330 325 75 320 315 310 305 300 295 290 285 70 280 275 270 65 265 260 255 3 250 245 Fish taphon. 1 3 5 7 Fish richness Fish Asmusia Plant abundance Fig. 2 (continued) 300 fla ts Palaeobio Palaeoenv (2011) 91:293–323 Ch an ne ls & d meters Bonaventure Fm. Siltst Sh Sdt Cgt Fluvial bay head delta Fr on t Escuminac Fm. To e Ce ntr al ba sin 6 115 390 110 385 380 375 105 5 100 370 8 365 95 360 Fish taphon. 1 3 5 7 Fish richness Fish Asmusia Plant abundance Fig. 2 (continued) ing pebble and cobble conglomerate. Lack of disruption of Fleurant pebbles by the basal Escuminac bed (Dineley and Williams 1968; Zaitlin 1981) implies prior lithification of the Fleurant Formation, suggesting a short hiatus in the sedimentation (Rust 1982). The Fleurant Formation is interpreted as a proximal gravel-dominated braided-plain deposit flowing toward the southeast (Rust et al. 1989; Zaitlin 1981), at the margin of the growing Acadian Appalachians. The approximately 375-m-thick Bonaventure Formation [as newly re-defined by Jutras et al. (2001)] is composed of a red, poorly sorted, polymictic conglomerate of grainsupported (gravel, pebble to large cobble size) and lithic sandstone beds showing large-scale planar tabular and trough Palaeobio Palaeoenv (2011) 91:293–323 301 cross bedding and mudstone (Zaitlin 1981). The clasts are composed predominantly of carbonates, with lesser amounts of basalt and sandstones that are most likely derived from the Fortin, Chaleurs and Matapédia Groups (Zaitlin 1981). Cross-bedding indicates a direction of sediment transport to the SE (Jutras et al. 2001). The contact with the underlying Escuminac Formation is sharp and erosional and characterised by a middle-Frasnian–Viséan (approx. 40 Ma) hiatus (Jutras et al. 2001). Most of the conglomeratic layers within the Miguasha section of the Escuminac Formation are poorly bedded and composed of a mixture of sub-rounded basement pebbles, clay chips and pebbles and isolated bones and bone fragments (predominantly of the placoderm Bothriolepis canadensis) in a fine-grained sandstone matrix. Some conglomerate beds are matrix supported, massive and structureless. Others are grain supported and exhibit some crude decimetre-scale planar tabular cross-bedding and pebble, and cobble imbrications to the SW. The coarsegrained nature and basement origin of the elements in the conglomerates imply that they were probably emplaced in a high-energy environment. The unsorted, matrix-supported and massive character of the beds are typical of debris flow deposits, and the well-sorted grain-supported beds with planar tabular cross-bedding are typical of losangic straight crest bars in a fluvial braided environment. The overall unsorted fabric as well as the high-energy environment are indicative of alluvial fan deposits close to elevated source area. Intertidal (slikke) siltstones and argillites (“laminites” facies) This facies is quite frequent but generally thin (7% of total thickness) and more common in the lower part of the Escuminac Formation. It occurs in beds varying in thickness from 1 to 89 cm, averaging approximately 12 cm. It is composed of a regular alternation of <1-mm-thick laminae of siltstone interbedded with <0.2-mm-thin shale laminae rich in amorphous organic matter (AOM); the two types of laminae form couplets. The detrital laminae are primarily composed of quartz and carbonate silt with angular to subangular grains and exhibit rough fining upward trends. Muscovite mica is abundant and plagioclase feldspar is present; Hesse and Sawh (1992) reported the presence of chlorite. The organic shale laminae are thin, continuous and slightly undulating. Locally, particularly towards the base of the section, the successive couplets display cyclic patterns of deposition with gradual thickening and thinning and some desiccation cracks. Each thickening–thinning package is comprised of 18–20 couplets. This facies has the highest percentage of CaCO3 (mean= 43.28%; N=12; El Albani et al. 2002) of the section. Detrital carbonate elements are abundant, and it is difficult to differentiate, in this percentage, detrital grains from diagenetic cement which also involves replacement. This facies is also characterised by a higher sulphur content than that found in shales and siltstones (El Albani et al. 2002). The laminites are not bioturbated but yield the greatest diversity and abundance of fossil fishes within the formation. Most fishes are well preserved (i.e. complete specimens). Some laminated beds are associated with mass mortalities of acanthodians (primarily Triazeugacanthus affinis). The regular alternation of detrital and organic-rich laminae is the result of a rhythmic sedimentation that could be due either to seasonal changes of sediment supply (varves in periglacial lakes), diurnal or semi-diurnal tidal ebb and flood oscillations or flood events possibly related to rain storms. The three hypotheses are supported by the high organic and sulphur content, but palaeogeography excludes the periglacial hypothesis. Couplets that are not organised in rhythmic bundles could correspond to the periodic influx of sediment in a standing body of water through low-density gravity flows; flood events producing the fining upward siltstone laminae when the silty organic rich laminae may correspond to the passive settling of clay in a standing body of water. The rarity of well-preserved algal structures in the organic-rich layer (Cloutier et al. 1996) may preclude the varve origin of the laminites (Stephenson et al. 2006) and favour a marine environment with periodic input of sediment by floods. The 18–20 couplet packages cut by rare desiccation cracks observed at the base of the section could correspond to the record of neap spring cycles (Ashton 1981; Choi et al. 2001; Mazumder and Arima 2005; Tessier 1993; Williams 1989) in the intertidal zone of the internal part of an estuary. This hypothesis is supported by the thickening–thinning pattern of couplet succession, the number of couplets fairly consistent with the number of days in the lunar month minus the weakest tide couplets in the upper intertidal zone owing to nondeposition during neap tides; the high carbonate content is consistent with sedimentation in an internal estuarine environment (Ashton 1981); the low bioturbation is due to a high sedimentation rate and a stressful, frequently desiccated environment supplied by freshwater sources. Bay head delta sandstones and siltstones (“sandstones” and “siltstones” lithofacies) Sandstones and siltstones can occur separately but are more frequently found in close association in the Miguasha Section, with the sandstone grading upward into the siltstones or, in places, the siltstones grading into the sandstones. Sandstones are poorly sorted, medium- to very finegrained, subarkosic arenites with detrital quartz, plagioclase 302 Palaeobio Palaeoenv (2011) 91:293–323 feldspar, biotite and muscovite. Sandstone beds (25% of total thickness of the formation) vary in thickness from 0.01 to 3 m and average 0.39 m. Most of the sandstone beds are fining upward and exhibit at their base erosional sedimentary structures (e.g. groove casts, flute casts, brush and prod casts, rill casts) and, in places, load casts. Internal structures include centimetric cross-bedding, climbing ripples and low-angle to parallel lamination. Most sandstone beds show a sequence of erosional sedimentary structures at the base, overlain by upper flow regime planar lamination, with current-ripple cross-bedding ending in lower flow regime parallel laminations. At the base of the section, stacked sandstone beds exhibit slightly concave-up, channel-like erosional sole imbricated in lateral accretion with desiccation cracks at the top. Some beds are structure-less. Others are slightly coarsening upward, ending up with a fining upward trend. In the upper part of the formation (Units VI and VIII; Fig. 2), where sandstone is more abundant, fishes are preserved frequently three-dimensionally, and large, well-preserved plant remains (primarily pennes of the progymnosperm Archaeopteris halliana) are abundant. Light-grey siltstones occur throughout the formation (22% of total thickness). They display bed thicknesses from 0.01 to 1.46 m, averaging approximately 0.15 m. The siltstones are made up of angular to subangular quartz grains, abundant mica, wood fragments and pyrite, as well as some millimetre-scale bony elements dispersed into the sediment. Fossils are preserved frequently as isolated bony elements (mainly sarcopterygian scales or isolated abraded bones of indeterminate fishes) concentrated at the base of the beds. Some beds are crudely normally graded. When the siltstones occur in separate beds, they show basal erosional structures similar to those of the sandstone beds. Internal sedimentary structures are scarce and mainly represented by current ripple-forms passing upward to faint horizontal planar to low-angle lamination. The sedimentary structures of the sandstone beds are indicative of a medium-energy subaquatic depositional environment. The fining-up trends together with the progressive decrease in the energy of the sequence of internal sedimentary structures is typical of episodic, waning flow energy. This character is amplified when the sandstones grade to the fining-upward siltstone bed where the upward change from current ripples to planar lamination is indicative of a progressive decrease in the energy of the depositional environment. The presence of some superimposed coarsening and of fining-up trends indicative of waxing and waning flow alternations together with the lack of oscillation wave ripples indicative of an open marine environment and the presence of stacked channel features with desiccation cracks are all arguments for a shallow-water, flood- influenced, depositional environment, probably close to a fluvial outlet, such as a fluvial bay-head delta. Some localised slump deposits (bed 242.1; Fig. 2) are indicative of a slight slope gradient along the depositional profile that may have favoured the downslope displacement of the sediment by gravity and the formation of some turbidity currents. However, it is still likely that, in this shallow marine environment, floods may constitute the main forcing mechanism at the origin of these deposits. Central semi-enclosed basin shales (“shales” lithofacies) Dark-grey, light-grey and greenish shales are the most abundant lithofacies (46% of total thickness). Shale forms beds varying in thickness from 0.02 to 4.5 m, averaging approximately 0.33 m. Among the five lithofacies, shales show the lowest percentage of CaCO3 (El Albani et al. 2002). Two types of lithofacies are encountered in the shale layers: (1) homogeneous shale and (2) shale with fine parallel layers of dark AOM. The laminated type shares a similarity with the laminites in terms of palynofacies characteristics, but differs in having lower CaCO3 and silty components. The layers of AOM are either present on the complete thickness of the shaly beds, restricted to the base, the middle or the top of the bed or they occur simultaneously at the base and the top of the bed. However, they occur most frequently throughout the bed. This laminated facies seems to be more frequent in beds thinner than 0.60 m. Most fishes found in the homogeneous shale are poorly preserved (with rare exceptions, such as bed 351) and are represented as isolated bony elements, such as scales. Most beds of laminated shale are devoid of fossil fish. A low-energy depositional environment may explain the massive character and the lack of sedimentary structure observed in the shales. The organic laminae are indicative of periodic starvation of the depositional environment with no detrital input. This hypothesis is supported by the low carbonate content that may correspond to periods of low oxygen, reducing sea-floor conditions. The poor conditions of life in the water column may explain the low abundance of fishes and fossil remains in the sediment. Low detrital input, low oxygen content and low-energy conditions are common features of semi-enclosed basins or ponds or very deep marine environments. The presence of the Devonian brackish water conchostracan Asmusia favours a nearshore probably lagunal, hypothesis. Sequence stratigraphy interpretation The detailed changes through time in depositional environments in the Escuminac Formation (see Fig. 2) show metre- Palaeobio Palaeoenv (2011) 91:293–323 303 thick variations from proximal to distal environments that can translate into deepening and shallowing upward variations or base-level changes. These metre-scale variations are stacked onto each other, drawing into four 10 m-scale trends of deepening and then shallowing up trends bounded by sharp erosional boundaries. These surfaces are interpreted as sequence boundaries overlain respectively by a transgressive system tract (deepening) and then a highstand system tract (shallowing) above a maximum flooding surface. The second depositional sequence in the section shows a different and more complete assemblage. It starts with low-amplitude variations in palaeowater depth that may correspond to a lowstand system tract capped by a transgressive surface overlain by the suite of a transgressive and a highstand system tract. The maximum palaeowater depth in the section is interpreted to be located in the second sequence at 90 m, whereas a minimum is reached in the basal and topmost fluvial conglomerates. Taxonomic and palaeoecological diversity (invertebrates and fishes) Invertebrates account for a minor part of the diversity of the Escuminac Formation with the presence of 12 species. The aquatic component of the invertebrate fauna includes the conchostracan Asmusia membranacea (Martens 1996), a parastylonurid, an eurypterid (Jeram 1996) and a scolecodont (Cloutier et al. 1996), whereas the continental component includes the millipede Zanclodesmus willetti (Wilson et al. 2005), the scorpion Petaloscorpio bureaui and a gigantoscorpionid (Jeram 1996). Fragments of arthropod cuticles have been found in palynological preparations (Cloutier et al. 1996), and some of these are likely referable to arachnids or trigonotarbids. Two aquatic ichnotaxa have been identified (Maples 1996; Schultze 1999) as well as an additional two types of ichnofossils. Ichnofossils have been found in six of the 272 beds sampled. Petaloscorpio was found in only three beds and Zanclodesmus in a single bed. Normal marine invertebrate fossils are lacking. The Escuminac Formation includes a total of 20 vertebrate species (Table 1; Fig. 3). Most major groups of Late Devonian vertebrates are represented in the Escuminac Formation: anaspids (Fig. 3a), osteostracans (Fig. 3b, c), placoderms (Fig. 3d), acanthodians (Fig. 3e–g), actinopterygians (Fig. 3h) and sarcopterygians (Fig. 3l–o). Specimens of 14 of the 20 species have been found during the bed-by-bed sampling of this study. Additional biostratigraphic occurrences were included for three species (Levesquaspis patteni, Miguashaia bureaui, Elpistostege watsoni) that have not been sampled for this study but for which stratigraphic information was available. The overall palaeoecology of the vertebrate assemblage is inferred from the gross morphology of the organisms, the evidence of predation and comparisons with closely related forms. Euphanerops longaevus and Endeiolepis aneri were most likely microphagous bottom feeders (Janvier and Arsenault 2007). Based on their relatively flattened morphology, three species (Escuminaspis laticeps, Levesquaspis patteni, Bothriolepis canadensis) are considered as bottom dwellers. Escuminaspis and Levesquaspis were detritivores most likely consuming both bottom detritus and particles and organisms suspended in the water column (Moloshnikov 2008). Bothriolepis canadensis, Plourdosteus canadensis and the dipnoans are considered benthivores (Moloshnikov 2008). The four species of acanthodians were most likely planktivores (Trewin 1986) and contributed as forage fish in the fish palaeocommunity. Cheirolepis canadensis was a small predator, whereas large predators include Miguashaia bureaui, Holoptychius jarviki, Quebecius quebecensis, the porolepiform indet., Eusthenopteron foordi and Elpistostege watsoni. Among the predators, C. canadensis (Arratia and Cloutier 1996) and E. foordi (Arsenault 1982; Cloutier 1996) have been found with ingested prey (including C. canadensis, Homalacanthus concinnus, Triazeugacanthus affinis and E. foordi). Conchostracans (most likely referable to Asmusia membranacea) were non-selective algal and detrital feeders (Orr and Briggs 1999) that have been found in the digestive track of Homalacanthus concinnus, Scaumenacia curta and Bothriolepis canadensis. Regurgitates and coprolites are fairly frequently found with bony inclusions (McAllister 1996). Coprolites have been found in 38 beds of 272 beds during this study; in 32 of these beds, coprolites were found in association with anatomical remains of fishes. Gregarious behaviour has the potential to influence abundance. Evidence of gregarious behaviour has been found for Bothriolepis canadensis, Triazeugacanthus affinis, and Scaumenacia curta. It is likely that the gregarious behaviour of these species is associated with particular periods of their respective life cycles because groups are generally composed of similar-sized individuals (age classes) (Parent and Cloutier 1996). Richness (invertebrates and fishes) The total invertebrate richness (including macro- and microremains as well as the ichnotaxa) is 12, although the invertebrate richness per bed varies between zero and three (one bed with three species, six beds with two species, and 101 beds with one species). With the exception of Asmusia, invertebrates are rare and therefore do not account for a large part of the total richness of beds (i.e., fish richness plus invertebrate richness per bed). 304 Palaeobio Palaeoenv (2011) 91:293–323 b a c d e f g h i k j o l n m Palaeobio Palaeoenv (2011) 91:293–323 305 ƒFig. 3 Representative lower vertebrates from the Escuminac Formation. a the anaspid Euphanerops longaevus, MHNM 01–02, b the osteostracan Levesquaspis patteni, MHNM 01-12-10b, c the osteostracan Escuminaspis laticeps, MHNM 01-09a, d the placoderm Bothriolepis canadensis, MHNM 02–2676, e the acanthodian Diplacanthus horridus, MHNM 03–734, f the acanthodian Homalacanthus concinnus, MHNM 03-860A, g the acanthodian Triazeugacanthus affinis, MHNM 03–729, h the actinopterygian Cheirolepis canadensis, MHNM 05–71, i the actinistian Miguashaia bureaui, MHNM 06–494, j the porolepiform Quebecius quebecensis (skull), MHNM 06–246, k the porolepiform Holoptychius jarviki, AMNH 11593, l the dipnoan Scaumenacia curta, MHNM 04-13A, m the dipnoan Fleurantia denticulata, MHNM 04–1392, n the osteolepiform Eusthenopteron foordi, MHNM 06–62, o the elpistostegalian Elpistostege watsoni (skull), MHNM 06–538 Although the complete assemblage includes 20 vertebrate species, the maximum number of vertebrate species encountered in a single bed is 11. Fish remains have been found in 134 beds of the 272 beds sampled (Fig. 4a). The fish richness per bed varies between zero and 11, where approximately 50% of the beds lack vertebrate remains (Fig. 4a). Most beds yielded one or two species, with an average of 1.15 [standard deviation (SD) = 1.72] species per bed; when considering solely the 134 beds that yielded fish remains, the average increases to 2.33 (SD = 1.80) species per bed. The mean fish richness differs among lithofacies (Fig. 4b; Kruskal–Wallis test: H = 60.5, P= 4.597E-10); richness associated with laminites is significantly higher than that in sandstones, siltstones and shales. Abundance (plants, conchostracans and fishes) Most plant debris are shorter than 5 cm, thus leaving most material as undetermined in terms of stems and raches. There is a larger proportion of beds with plant debris than beds devoid of debris in most lithofacies, with the exception of siltstone beds (Fig. 5). In general, the mean size of plant debris is not significantly different among the four most common lithofacies (Kruskal–Wallis test: H= 6.256, P=0.09981; with the exception of the shale, sample sizes are low). However, the mean size of plant debris tends 150 to be inversely correlated with the lithofacies gradient (conglomerate beds are underrepresented in the samples). Millimetre-scale plant remains are fairly frequent throughout the formation, whereas larger centimetre- and decimetre-scale remains are more abundant in the upper part of the formation. Although minute in size, plant debris are most abundant within shaly beds, but there are no significant differences among lithofacies (Kruskal–Wallis test: H=5.112, P=0.1638). The abundance of plants varies throughout the formation (Fig. 2). Conchostracans are the most abundant component of the invertebrate diversity. Although specimens of Asmusia are the most common invertebrates, most beds are devoid of them independently of the lithofacies (Fig. 6). However, the proportion of beds and the relative abundance of Asmusia are higher in laminites (Fig. 6). The abundance of Asmusia seems to be more important during the transgressive phases of the different sequences (Fig. 2). The average relative abundances of Asmusia and fish do not differ significantly (Mann–Whitney test: T=3.613E4, P=0.5852), although Asmusia is more abundant than fish in absolute numbers. More than 18,000 specimens of lower vertebrates have been found in the Escuminac Formation (Table 1). Some of the Escuminac vertebrate species are known from thousands of specimens encompassing various states of preservation. The most abundant species are Bothriolepis canadensis, Triazeugacanthus affinis, Scaumenacia curta and Eusthenopteron foordi, with these four species accounting for 90% of the total number of specimens discovered. These species also correspond to the most frequently encountered species because they have been encountered in the greatest number of beds (72%). The acanthodian Triazeugacanthus has a low percentage of occurrences, not proportional to its abundance, because a large number of these specimens have been found in only a few beds associated with mass mortalities. The rarer species are Callistiopterus clappi (one nearly complete juvenile specimen), a new undescribed porolepiform (one juvenile and one large specimen), Elpistostege watsoni (four partial specimens including one isolated scale), and Diplacanthus ellsi (two complete and four partial specimens) (Table 1). Number of beds a Bed 8 Mean fish richness 3 2 1 0 1 2 3 b 100 50 0 Bed 366 0 1 2 3 4 5 6 7 8 9 10 11 4 Fish richness Lithofacies Fig. 4 Fish richness in the Escuminac Formation. a Frequency distribution of the fish richness (number of fish species). b Comparison of the mean fish richness per bed in relation to the four lithofacies encountered in the Escuminac Formation: 1 sandstone, 2 siltstone, 3, laminites, 4 shale. Kruskal-Wallis test: H = 60.5, P= 4.597E-10 306 Abundance N Mean S.D. N Mean S.D. 1 3 N.A. 21 1.19 1.17 14 7.91 11.19 60 0.01 1.08 21 4.01 4.92 38 1.84 1.72 19 4.94 4.97 76 1.87 0.90 39 1.87 2.21 Palaeobio Palaeoenv (2011) 91:293–323 Fig. 7 Histograms of the occurrence of 12 fish species (a–l) found in„ the Escuminac Formation in relation to the lithofacies. Lithofacies: 1 sandstone, 2 siltstone, 3, laminites, 4 shale. A Absence of the species, P presence of the species. Black bars correspond to the number of beds for which the species has been found Size 1 N.A. N.A. 60 Number of beds 50 40 30 20 10 Congl. Sands. Silts. Lamin. Shale Fig. 5 Abundance, size and distribution of plant debris in the Escuminac Formation in relation to the lithofacies. The histogram corresponds to the number of beds without (white bars) and with (grey bars) plant remains per lithofacies. N Number of beds for the specific categories, N.A. not applicable, S.D. standard deviation, Congl conglomerates In total, 334 specimens have been found during the bed-bybed sampling. In general, fishes are rare throughout the formation. Most vertebrate species are not distributed uniformly among the lithofacies (Fig. 7), and only a small percentage of the sampled beds have yielded specimens. Bothriolepis, Scaumenacia and Eusthenopteron are the three most frequent occurrences in terms of the number of beds. Lithological distributions are either exclusively unimodal (Fig. 7c, e; Escuminaspis and Plourdosteus), primarily unimodal (Fig. 7d, f–g, k–l; Bothriolepis, Homalacanthus, Triazeugacanthus, Fleurantia and Eusthenopteron), with partial selection (Fig. 7a, h–i; Endeiolepis, Diplacanthus and Cheirolepis) or without lithological selection (Fig. 7c, j; Escuminaspis and Scaumenacia). Thus, the occurrence of some species seems to be lithofacies-dependent. Most species are more common in laminites, with the exception of some agnathans (Endeiolepis, Escuminaspis and Levesquaspis), which are more common in siltstone. The curve of fish abundance is punctuated by only a few beds with high abundance (Fig. 2). Among these beds, only beds 8 and 351 have yielded a few hundred specimens each. No fish has been found below bed 8 or in the Fleurant Formation. This absence of occurrence does not rely only on the sampling for this study but also on collecting for more than 20 years. Fishes are common to abundant in the lower part of the formation (between beds 8 and 32). Between beds 48 and 93, fish remains are present but rare. Between beds 104 and 174, the section is almost devoid of both, fish remains and conchostracans. Thus, most of sequence 3 is devoid of animal fossils, with the exception of rare occurrences in the upper part of the sequence. Between beds 164 and 347, fishes are rare, although the richness fluctuates. The abundance of fishes increases again in the upper part of the formation above bed 347. Fishes seem to be more abundant in the transgressive phase of the sequences. No significant correlation has been found between plant abundance and the abundance of Asmusia or fish. However, the abundance of Asmusia is significantly correlated with fish abundance (rs =0.483; N=239). Taphonomy (fishes only) A priori, the potential for fossilisation is not identical among the different species from the Escuminac Formation because of their different anatomy. The anaspids Endeiolepis and Euphanerops (Fig. 3a) have the lowest fossilisation potential because of their poorly ossified nature; they represent <0.5% of the total relative abundance of specimens (Table 1). Nevertheless, when present, representatives of these species are exceptionally well preserved with their soft tissues. The placoderms Bothriolepis (Fig. 3d) and Plourdosteus have the greatest potential for preservation because of the heavy dermal bones covering their head and trunk shields and Number of beds 60 40 20 Sandstone 60 N = 25 40 20 Siltstone N = 66 60 40 20 Laminites 60 N = 50 40 20 Shale N = 99 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 Abundance of Asmusia Fig. 6 Abundance of the conchostracan Asmusia membranacea from the Escuminac Formation in relation to the lithofacies. 0 Absent, 1 very rare or seen only in palynological preparation, 2 rare, 3 common, 4 abundant, 5 very abundant. N Number of beds for the specific categories Palaeobio Palaeoenv (2011) 91:293–323 307 a 270 240 Number of beds 210 180 150 120 90 60 30 0 6 5 4 3 2 1 b Endeiolepis 270 240 210 180 150 120 90 60 30 0 5 4 3 2 1 c Euphanerops 270 240 210 180 150 120 90 60 30 0 5 4 3 2 1 Escuminaspis A P 1 64 56 48 40 32 24 16 8 2 3 4 A P 1 2 3 4 A P 1 2 3 4 d 270 240 Number of beds 210 180 150 120 90 60 30 0 e Bothriolepis 270 240 210 180 150 120 90 60 30 0 5 4 3 2 1 f Plourdosteus 270 240 210 180 150 120 90 60 30 0 18 16 14 12 10 8 6 4 2 0 Homalacanthus A P 1 2 3 4 A P 1 2 3 4 A P 1 2 3 4 g 270 240 Number of beds 210 180 150 120 90 60 30 0 27 24 21 18 15 12 9 6 3 0 h Triazeugacanthus 270 240 210 180 150 120 90 60 30 0 5 4 3 2 1 0 i Diplacanthus 270 240 210 180 150 120 90 60 30 0 2 1 7 6 5 4 3 Cheirolepis A P 1 2 3 4 A P 1 2 3 4 A P 1 2 3 4 j 270 240 Number of beds 210 180 150 120 90 60 30 0 54 48 42 36 30 24 18 12 6 k Scaumenacia 270 240 210 180 150 120 90 60 30 0 6 5 4 3 2 1 l Fleurantia 270 240 210 180 150 120 90 60 30 0 45 40 35 30 25 20 15 10 Eusthenopteron A P 1 2 3 4 A P 1 2 3 4 A P 1 2 3 4 308 Palaeobio Palaeoenv (2011) 91:293–323 the sutures among plates, favouring fairly well-articulated shields. Bothriolepis is effectively the most abundant and most frequently encountered species. However, Plourdosteus represents <1% of the relative abundance and <2% of the relative occurrence of the Escuminac species (Table 1). The exoskeleton of the four species of acanthodians is composed of minute scales and head tesserae (Fig. 3e–g); most of these elements are not imbricated. The potential of disarticulation of these species is high. Nevertheless, Triazeugacanthus is the second most abundant species and the fourth most frequent species, whereas both species of Diplacanthus are rare (Table 1). In terms of fossilisation potential, Cheirolepis canadensis is comparable to acanthodians, except for its skull, which is composed of relatively large, thin dermal bones (Fig. 3h). The general anatomy of sarcopterygian fishes (Fig. 3i–o) is highly comparable and, therefore, their potential for fossilisation is also similar and fairly good. However, despite their similar fossilisation potential, their abundance and frequency vary among the nine species of sarcopterygians. Although the potential for fossilisation is unequal among the 20 species of Escuminac fish, their relative abundance and frequency do not corroborate that fossilisation potential is a major factor controlling their presence (and discovery). The state of preservation of the fish varies within single beds as well as through the Escuminac Formation (Fig. 2). Fish preservation ranges from completely articulated specimens with occasional traces of soft anatomy to completely disarticulated specimens; however, most fish are partly articulated. The taphonomic gradient (from isolated elements to perfectly preserved specimens) is represented almost in totality in each lithofacies (Fig. 8). In general, better preserved specimens occur in laminites (Fig. 8c). The average state of preservation decreases in the order laminites (highest), sandstone, shale and siltstone (lowest). Isolated elements (predominantly isolated scales), independently of the taxonomic identification, are the dominant category found in sandstone (Fig. 8a), siltstone (Fig. 8b) and shale (Fig. 8c). Each lithofacies has a specific taphonomic signature (Fig. 9) referring to the relationships among the six states of the taphonomic gradient. For example, with respect to the 15 sandstone beds having fish remains (Fig. 9a), parts of specimens are most frequently encountered in the same bed with specimens showing only minor signs of decay. Specimens showing loss of kinetic anatomical elements are not found in the same beds as the other types. In terms of the Dice similarity coefficients, the associations among the taphonomic states are fairly low, with the exception of laminites (Fig. 9). The signatures of sandstone and shale are the most similar. The taphonomic signature of the laminites shows that there are two types of preservation in laminites (Fig. 9c): (1) beds with well-preserved specimens (taphonomic states: complete, minor and kinetic) and (2) beds with poorly preserved specimens (taphonomic states: decayed, parts and isolated). Of the 26 laminite beds with well-preserved specimens, 24 are associated with transgressive phases of sequences, whereas 11 of the 31 laminite beds with poorly preserved specimens are associated with regressive phases of sequences. The differences in states of preservation and the slight differential differences in terms of sequence occurrence might indicate that the laminites represent two distinct depositional environments despite their sedimentological similarities. The maximum state of preservation observed within a bed is significantly correlated with the total richness (rs = 0.776, N=267) and the fish richness (rs =0.911, N=272); however, it is evident that fish richness is significantly correlated with the total richness (rs = 0.910, N =272) because fishes constitute the most important component of the assemblage. The average richness has a tendency to increase along the taphonomic gradient from state 1 to state 6 (Fig. 10a–f), with a special increase of richness with state 2; on average, 1.56 species of fish are found in the 32 beds for which only isolated elements were found (state 1, Fig. 10a), whereas 4.12 species are found in the 25 beds characterised by exceptional preservation (state 6, Fig. 10f). The average fish richness differs significantly between the better states of preservation (states 5 and 6; Table 2) and the poorest state of preservation (Kruskal–Wallis test: H= 25.01, P=0.0001385); mean fish richness for beds in which the maximum state of preservation corresponds to decayed specimens differs also from that for beds with the best preservation. However, there are no significant correlations between the richness and the lithofacies, nor between the lithofacies and the maximum state of preservation per bed (Fig. 11). Taking into account the maximum state of preservation, these correlations suggest that the richness of a bed is biased positively by a taphonomic filter. When looking at the complete spectrum of states of preservation within a bed (rather than solely the maximum state), it is clear that poor states of preservation are primarily associated with low richness (Figs. 10, 11). Furthermore, laminites possess the broader spectrum of richness and the greatest proportion of good preservation (Fig. 11). Narrow spectra of total richness per bed are associated with the conglomerate (a single conglomeratic bed was analysed) and siltstones, and are not dependent on taphonomic states (Fig. 11). Thus, Lagerstätte horizons within the Escuminac Formation have a greater potential to yield a diverse assemblage of fish. Beds with the most favourable taphonomic conditions and the greatest richness and abundance of fishes (Fig. 2) may be considered as Lagerstätte horizons, and they can be qualified either as Konservat Lagerstätten (conservation deposit), Konzentrat Lagerstätten (concentration deposit), Palaeobio Palaeoenv (2011) 91:293–323 Fig. 8 Frequency distribution along a taphonomic gradient (between poor and excellent preservation) per lithofacies (a-d) of the Escuminac Formation. The taphonomic coding corresponds to the maximum state of preservation encountered within a bed. Taphonomic states: 0 Absence of fish remains, 1 isolated bones, 2 parts of body, 3 obvious signs of bacterial activity, 4 specimen relatively complete but with kinetic elements missing, 5 specimen with weak decomposition, 6 specimens for which all bony elements are preserved in natural position and which do not show signs of decomposition. N Number of beds for the specific taphonomic states for each lithofacies. Mean (SD) Average taphonomic condition 309 a 20 Sandstone N = 33 Mean = 2.30 (1.64) 60 50 b Siltstone N = 69 Mean = 0.91 (1.77) Number of beds 15 40 30 20 10 5 10 0 0 0 1 2 3 4 5 6 0 1 2 3 4 5 6 c Number of beds 15 Laminites N = 57 Mean = 3.42 (2.06) 80 70 60 50 d Shale N = 108 Mean = 1.10 (1.83) 10 40 30 5 20 10 0 0 0 1 2 3 4 5 6 0 1 2 3 4 5 6 Maximum state of preservation per bed or both. Twenty-six beds have exceptional taphonomic conditions (state 6), 16 beds have a fish richness of five or more species, and five beds have abundant fish (>50 specimens). Among these horizons, beds 8, 35 and 366 qualify for the three categories and can be considered as both Konservat and Konzentrat Lagerstätte horizons. Eight beds (i.e. beds 16, 23, 24.1, 26, 30, 39, 242.1 and 369) rank highly for both richness and preservation of fish; these beds are considered as Konservat Lagerstätte horizons. Bed 351 ranks highly for both richness and abundance and is considered as a Konzentrat Lagerstätte horizon. Most of these beds occur in the transgressive phase of the sequences and primarily towards the maximum palaeowater depth. regressive phases of the different sequences within the formation. To contrast these three sub-analyses, we compare the results to a cluster analysis of the complete dataset, which we will refer to as the Escuminac comprehensive assemblage (Fig. 12). This analysis is assumed to be independent of the abundance of the different species, the taphonomic condition, the sedimentology and the depositional environment. Escuminac comprehensive assemblage Seventeen of the 20 species have been included in the comprehensive cluster analysis irrespective of the lithofacies in which fossils were found and irrespective of their state of preservation. Callistiopterus and Holoptychius have been excluded from the cluster analysis because their stratigraphic origin is unknown. The dendrogram is based on the vertebrate contents from 272 beds (Fig. 12); only the beds for which one species is present are included in the analysis. Most links between clusters vary between 0.2 and 0.55 (Dice similarity index); thus, none of the associations is strongly correlated. Four primary clusters divide the assemblage (Fig. 12): (1) Euphanerops and Porolepiformes indet., (2) Cheirolepis, Miguashaia, Diplacanthus (D. horridus and D. ellsi are considered simultaneously) and Plourdosteus, (3) Bothriolepis, Scaumenacia, Euthenopteron, Triazeugacanthus and Homalacanthus and (4) Levesquaspis, Endeiolepis, Species composition (fishes only) Species composition corresponds to the hierarchical structure of the fish assemblage independently of the abundance of the different species. This portrait of the association among species reflects their co-occurrence owing either to the living and/or the depositional environment and to their palaeoecological relationships. To investigate the structuring constraints on the fish assemblage, we examine how the species composition varies as a function of (1) the lithofacies, (2) the state of preservation and (3) the transgressive and 310 Palaeobio Palaeoenv (2011) 91:293–323 a Sandstone (N = 15; 45.5%) Complete Parts Minor Isolated Decayed Kinetic species with respect to lithofacies (Fig. 7d, j, l). These three species will be referred to as the ubiquitous component of the fish assemblage of the Escuminac Formation. Cluster 4 is primarily found, but not exclusively, in the upper part of the formation. Lithofacies control on species composition The clusters obtained for the four lithofacies differ from one another in terms of structure, richness and intensity of the association. From the comprehensive assemblage, cluster 1 is only recovered in the laminites, clusters 2 and 3 are partly recovered in the laminites and cluster 4 is partly recovered in the sandstone, siltstone and shale. However, the ubiquitous cluster is recovered in each dendrogram with a stronger link between Scaumenacia and Euthenopteron (Fig. 13). The comprehensive assemblage (Fig. 12) is most similar to the faunal assemblage in the laminites (Fig. 13c) but with minor differences. In the siltstone lithofacies (Fig. 13b), a strong cluster corresponding to the channel level (base of sequence 5) is distinct from the ubiquitous assemblage; a similar but weaker pattern is also present in the sandstone lithofacies. In the laminites, the ubiquitous assemblage clusters with the abundant acanthodians, whereas they never cluster together in the other lithofacies. In addition to this association in the laminites, there are two other clusters including species limited to this lithofacies and primarily found in the lower part of the formation. Two distinct clusters are found in the shale (Fig. 13d): one cluster including the ubiquitous assemblage with agnathans, and a second cluster including primarily acanthodians and rare species (Quebecius and Elpistostege). Taphonomic control of species composition The taphonomic condition of the specimens has a strong effect on species composition. When only the beds for which the preservation is very good are considered (i.e. kinetic, minor and complete), the hierarchy of the assemblage (Fig. 14a) is similar to that of the comprehensive assemblage (Fig. 12) and the assemblage found in the laminate lithofacies (Fig. 13c). However, when solely the beds with poor states of preservation are considered (i.e. isolated, parts and decayed; Fig. 14b), two weak clusters (the ubiquitous species and a group from the lower part of the formation) and a single species (Cheirolepis) are found. The assemblage is less diversified in beds characterised by the poorer states of preservation. This disparity suggests that the taphonomic condition not only affects the richness but also the species composition. Although the species composition is partly affected by the taphonomic condition of the specimens, general clustering is coherent with that recovered in the different analyses. b Siltstone (N = 20; 29.4%) Complete Decayed Isolated Parts Kinetic c Laminites (N = 50; 87.7%) Kinetic Minor Complete Decayed Isolated Parts d Shale (N = 38; 35.2%) Complete Decayed Parts Minor Isolated Kinetic 0 0.2 0.4 0.6 0.8 1 Dice similarity Fig. 9 Taphonomic signature per lithofacies (a–d) of the Escuminac Formation. The taphonomic signature corresponds to the relationships among the different states of the taphonomic gradient encountered within a single bed for a given lithofacies. States given in bold correspond to the most frequent state encountered for a given lithofacies. N Number of beds for a specific lithofacies. The percentage in parentheses corresponds to the percentage of sampled beds for which taphonomic states have been recorded Escuminaspis, Fleurantia, Elpistostege and Quebecius. Cluster 1 is poorly informative because a single specimen of the Porolepiformes indet. has been found in bed 10, whereas only five beds yielded specimens of Euphanerops. Cluster 2 is almost limited to the lower part of the Escuminac Formation. Cluster 3 regroups the most abundant species in the formation (Table 1). Furthermore, Bothriolepis, Scaumenacia and Eusthenopteron are the non-selective or poorly selective Palaeobio Palaeoenv (2011) 91:293–323 Fig. 10 Relationships between fish richness and the maximum taphonomic state of a bed of the Escuminac Formation. Taphonomic categories (states): 1 Isolated bones (a), 2 parts of body (b), 3 obvious signs of bacterial activity (c), 4 specimen relatively complete but with kinetic elements missing (d), 5 specimen with weak decomposition (e), 6 specimens for which all bony elements are preserved in their natural position and which do not show signs of decomposition (f). N Number of beds for the specific taphonomic state, Mean average fish richness for a given state of preservation 311 a 15 10 5 State 1 N = 32 Mean = 1.56 b 15 10 5 State 2 N=8 Mean = 2.36 c 15 10 5 State 3 N = 26 Mean = 1.70 Number of beds 0 2 4 6 8 10 0 0 2 4 6 8 10 2 4 6 8 10 d 15 10 5 State 4 N = 26 Mean = 2.18 e 15 10 5 State 5 N = 19 Mean = 2.89 f 15 10 5 State 6 N = 25 Mean = 4.12 0 2 4 6 8 10 0 0 2 4 6 8 10 2 4 6 8 10 Fish richness Sequential control on species composition Fish assemblages were recovered for the transgressive and regressive phases of the first three sequences, whereas only the assemblage of the transgressive phases of sequences 4 and 5 were calculated (Fig. 15). Regressive phases have been less sampled than the transgressive phases; the sequence division of the formation was done a posteriori, resulting in an unequal sampling effort among sequences and phases. In general, Dice similarity indices and bootstrap values are low, partly owing to the weak number of occurrences for most species. The fish composition of the transgressive phase of sequence 1 is well structured and similar to the total assemblage computed for the sequence. Three clusters comprise sequence 1: (1) a cluster of relatively frequent species, including the ubiquitous assemblage with forage-fish acanthodians (similar to cluster 3 from the comprehensive assemblage; Fig. 12), (2) a cluster of rare species mainly occurring in bed 8 (similar to cluster 2 from the comprehensive assemblage), and (3) a tandem of rare species (corresponding to cluster 1 from the comprehensive assemTable 2 Mann–Whitney pairwise comparisons among the mean fish richness in relation to the maximum state of preservation (1–6) encountered in a bed States Mean 2 (Parts) 3 (Decayed) 4 (Kinetic) 5 (Minor) 6 (Complete) 1 (Isolated) 1.56 2.408 10.47 5.042 0.02445* 0.00091* blage). The regressive phase of sequence 1 lacks the ubiquitous assemblage because of the absence of Bothriolepis. The transgressive phase of sequence 1 shows three species unique to this phase. Sequence 2 is distinctive from the remaining sequences because of the weak diversity and weak cohesion of the assemblage of the transgressive phase. Furthermore, Scaumenacia is the dominant species only in sequence 2. As in regressive sequence 1, the regressive phase of sequence 2 lacks Bothriolepis. There are only seven beds of 77 sampled beds belonging to the aggradational phase of sequence 3 (from bed 111 to bed 213) that yielded fish remains (Fig. 2). Only one species was found in each one of these seven beds: Homalacanthus, Diplacanthus and Cheirolepis were found solely in one bed each, whereas Bothriolepis and Eusthenopteron were both found in two beds each. A strong ubiquitous assemblage is only recovered in the total assemblage of sequence 3. The transgressive assemblage is fairly similar to the total assemblage of the sequence. Bothriolepis is the dominant species in the total sequence as well as the regressive phase of sequence 3, whereas Eusthenopteron is slightly more frequent during the transgressive phase. 2 (Parts) 2.38 3.943 10.07 7.628 1.632 3 (Decayed) 1.70 4 (Kinetic) 2.08 5 (Minor) 2.8 6 (Complete) 4.0 P values are corrected using the Bonferroni correction method. P values denoted with an asterisk are significant 7.695 0.1568 0.0049* 2.785 0.3282 2.491 312 14 Preservation 12 10 8 6 4 2 0 Complete Minor Kinetic Decayed Parts Isolated Palaeobio Palaeoenv (2011) 91:293–323 0 1 2 Lithofacies 3 4 Fig. 11 Relationships among total richness, taphonomy and the lithofacies of the Escuminac Formation. Lithofacies: 0 Conglomerates, 1 sandstones, 2 siltstones, 3 laminites, 4 shales The transgressive phases of sequences 4 and 5 are well structured but slightly distinct in terms of composition from each other. Data are unavailable for the regressive phases of sequences 4 and 5. In both sequences, the ubiquitous assemblage is recovered, and these three species are relatively frequent; in sequence 5, Endeiolepis is also frequent. Bothriolepis is also the dominant species in sequence 4. The transgressive phases of sequences 4 and 5 include a single unique species each. In summary, the composition of Escuminac fish assemblage fluctuates in relation to the progression of the depositional environment, although the ubiquitous assemblage has been recovered in all but sequence 2. An initial assemblage colonised rapidly the intertidal environment. Sequence 1 is the most diversified, whereas sequence 2 is the less diversified sequence. Sequences 1, 4 and 5 include species unique to their respective sequence. Sporadic invasions occurred during transgressive phases, but none of them was persistent. The regressive phases of the lower three sequences display a poorly diversified (four species) and/or structured assemblage; sample size and number of occurrences of individual species are low in regressive phases. Transgressive phases are more diversified Total richness Levesquaspis Endeiolepis Fleurantia Escuminaspis Quebecius Elpistostege Bothriolepis Scaumenacia Eusthenopteron Homalacanthus Triazeugacanthus Plourdosteus Diplacanthus Miguashaia Cheirolepis Euphanerops Porolepiformes indet. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 4 3 2 1 Dice similarity Fig. 12 Comprehensive fish assemblage of the Escuminac Formation. Cluster analysis of 17 of 20 species for 272 beds is based on the Dice similarity index with average linkage. Names given in bold correspond to the ubiquitous component of the assemblage Palaeobio Palaeoenv (2011) 91:293–323 313 Bothriolepis Scaumenacia Eusthenopteron Fleurantia Quebecius Escuminaspis Levesquaspis Endeiolepis Homalacanthus a Sandstone Discussion Palaeoenvironment of the Escuminac Formation The depositional environment of the Escuminac Formation has been variously considered to be either lacustrine, estuarine, coastal marine or open marine. An estuarine interpretation is the depositional setting that best accommodates the different lines of evidence provided by the fauna (Elliott et al. 2000; Schultze 1999; Schultze and Cloutier 1996), the palynofacies (Cloutier et al. 1996), the trace fossil assemblage (Maples 1996), the sedimentological and stratigraphic setting of the formation (Hesse and Sawh 1992; Prichonnet et al. 1996), the isotope geochemistry of the sediments (Chidiac 1996; El Albani et al. 2002), the 87Sr/86Sr, Na, F, S and La analyses of fish bones (Matton et al. 2006; Schmitz et al. 1991) and the Mg and Fe concentrations of the sediments (Vézina 1991). The present refinement of inferred depositional environments and sequence stratigraphic interpretation of the main section of the Escuminac Formation re-emphasises the nature of the estuarine system and provides the first dynamic interpretation of its history. The Escuminac Formation is interpreted as a wavedominated estuarine depositional environment. It shows the transgression of a very shallow nearshore marine environment over continental braided fluvial deposits of the Fleurant Formation. These transgressive shallow-marine deposits are comprised of shallow sandy channels overlain by carbonaceous, fine-grained micro-laminated silty-to-shaly sediments with tidal rhythmites and desiccation cracks, typical of the inner parts of modern estuaries (Dalrymple and Choi 2007; Tessier 1993). Up-section, fine-grained sandstones and siltstones show indications of waxing and waning flow conditions owing to the influence of flooding along a depositional profile with a low slope gradient, but significant enough to allow reworking by gravity through such events as slumps and turbidity currents. Such a shallow-marine, lowenergy, low-gradient, flood-dominated fan at the outlet of a fluvial feeder is interpreted as a bay-head delta. The sandstones and siltstones are interbedded with shales with low carbonate content, and there are indications of lowenergy, brackish water and deeper marine conditions. Organic matter is preserved from time to time in this standing body of water, indicative of a low sediment input and oxygen depletion in a periodically starved standing body of water. The overall lack of bioturbation is a consequence of a possible stressful environment with periods of either low oxygen content or freshwater input. The juxtaposition of a fluvial, intertidal, flood and gravity-influenced fan interbedding in a low-energy basin temporarily starved is typical of wave-dominated estuaries. The barrier encloses a broad central basin whose shape varies depending on the wave energy and the nature of the b Siltstone Bothriolepis Scaumenacia Eusthenopteron Levesquaspis Escuminaspis Endeiolepis Triazeugacanthus c Laminites Porolepiformes indet. Euphanerops Homalacanthus Triazeugacanthus Eusthenopteron Scaumenacia Bothriolepis Cheirolepis Fleurantia Plourdosteus Diplacanthus Escuminaspis Miguashaia d Shale Homalacanthus Triazeugacanthus Cheirolepis Diplacanthus Quebecius Elpistostege Bothriolepis Scaumenacia Eusthenopteron Escuminaspis Levesquaspis Endeiolepis 0.4 0.6 0.8 1 0 0.2 Dice similarity Fig. 13 Relationships between species composition and lithofacies (a–d) of the Escuminac Formation. Cluster analyses are based on the Dice similarity index with average linkage. Names given in bold correspond to the ubiquitous component of the comprehensive assemblage. Shaded area corresponds to species unique to the lithofacies and better structured than aggradational and regressive phases. The dominance or frequency of species varies among phases and sequences. With minor differences, only the transgressive assemblage of sequence 1 resembles the comprehensive assemblage (Fig. 12). Thus, a broad-scale assemblage, such as the comprehensive assemblage, could produce an artificial assemblage that never existed during depositional of the formation. 314 Fig. 14 Relationships between species composition and taphonomy. a Dendrogram regrouping the better states of preservation (states 4–6; i.e. kinetic, minor and complete, respectively). b Dendrogram regrouping the poorer states of preservation (states 1–3; i.e. isolated, parts and decayed, respectively). Cluster analyses are based on the Dice similarity index with average linkage. Names given in bold correspond to the ubiquitous component of the comprehensive assemblage Palaeobio Palaeoenv (2011) 91:293–323 a Excellent preservation Bothriolepis Scaumenacia Triazeugacanthus Eusthenopteron Homalacanthus Fleurantia Escuminaspis Plourdosteus Miguashaia Diplacanthus Cheirolepis Euphanerops Levesquaspis Endeiolepis Elpistostege Quebecius b Poor preservation Scaumenacia Eusthenopteron Bothriolepis Cheirolepis Plourdosteus Diplacanthus Homalacanthus 0 0.2 0.4 0.6 0.8 1 Dice similarity 0 0.2 0.4 0.6 0.8 1 coast (rock or soft sediment) (Chapman et al. 1982; Morrisey et al. 1995). The basin is filled by sediments provided by both terrestrial and marine sources. Seaward of the wave-dominated estuary, the tidal inlet, crossing the barrier island, allows the exchange of water and organisms between the central basin and the open sea, but this narrow entrance is periodically closed by the lateral drift of the barrier. This marine connectivity is suggested by the presence of acritarchs (Cloutier et al. 1996) as well as by the presence of a “transitional” fish fauna (Schultze and Cloutier 1996). Landward of the wave-dominated estuary, a fluvial bay-head delta built from fluvial sediment deposited at the mouth of the river provides fresh water, plants and coarse sediments to the central basin. The fluvial bay-head delta, which extends into the central basin, comprises a low-angle subaqueous fan overlain by organic-rich levees, channels and intertidal flats (Pasternack and Brush 2002; Webster et al. 2002). Sediments in a wave-dominated estuary consist of fine-to-coarse sands in the barrier and tidal inlet deposits, fine organic mud and sandy mud in the central basin and mud and coarse sands to gravels in the fluvial bay-head delta (Allen 1991; Nichol 1991; Pasternack and Brush 2002; Webster et al. 2002). A diverse range of marine and brackish, subtidal, intertidal and supratidal estuarine habitats is represented in wavedominated estuaries. Because of the narrow entrance, the exchange with the marine environment is typically low. Only a small portion of the central basin waters is exchanged at each tide; river flows are high and freshwater may expel marine water from the estuary. Turbidity is low except during flood periods. The basin is an efficient trap for organic material coming from the catchment, with a long residence time that allows organic matter processing (nitrogen load). Generally, wave-dominated estuaries contain euryhaline faunas as well as transitory faunas that come occasionally from the marine environment (Paterson and Whitfield 2000; Potter and Hyndes 1994; Rainer and Fitzhardinge 1981). They provide a diverse range of habitats: salt marshes around the edges of the central basin, sands and mud with episodic turbid waters during river floods at the fluvial outlet, sands with algae, clear waters and high-energy conditions (tides, waves) at the tidal inlet (Roy et al. 2001) and mud with micro- and macroalgae and benthic invertebrates in the low-energy conditions of the central basin. Storms frequently close the tidal inlet by the lateral drift of the littoral barrier, resulting in temporary disaerobic conditions in the central basin and good preservation of the organic matter. Schultze and Cloutier (1996) compared the Escuminac fish assemblage with 39 late Givetian and early-middle Frasnian fish assemblages from around the world to identify palaeoenvironmental (freshwater, “transitional” and marine) affinities. Based on their analysis at the generic level, the Escuminac assemblage clusters with fish assemblages from the Okse Bay Group (Ellesmere Island, Nunavut Territory, Canada), the Edenkillie beds (Whitemire, Scotland, UK) and the Pskov beds (Russia). The palaeoenvironment of the Escuminac Formation is fairly compatible with the Okse Bay Group and the Pskov beds; the Edenkillie beds are still interpreted as a lacustrine environment (Dineley 1999). The Okse Bay Group is reinterpreted as a nearly continuous terrestrial deposition of meandering stream or braided stream facies complexes (Algeo and Scheckler 1998; Daeschler et al. 2006). The Pskov beds represent a marine shallow-water environment (Zhuravlev et al. 2006). In addition, localities from the Middle Devonian Baltic Basin, which yielded numerous fish assemblages sharing faunal similarities with the Escuminac assemblage, have been interpreted as tide-dominated estuary environments (Ponten and PlinkBjorklund 2007, 2009). Sequence 1 Triazeugacanthus (1) 70 59 Sequence 2 Transgressive Euphanerops (2) 48 Sequence 3 Sequence 4 51 52 29 36 21 Scaumenacia (11) 40 28 71 100 41 48 100 6 Eusthenopteron (2) 53 17 Homalacanthus (1) 40 28 42 34 51 23 Total (beds) Bothriolepis (2) 39 9 33 N = 59 Palaeobio Palaeoenv (2011) 91:293–323 Porolepiformes indet. (1) Euphanerops (3) Homalacanthus (9) Bothriolepis (11) Triazeugacanthus (16) Eusthenopteron (17) Scaumenacia (14) Fleurantia (4) Plourdosteus (4) Diplacanthus (2) Escuminaspis (2) Miguashaia (1) Cheirolepis (3) N = 36 N = 110 Scaumenacia (8) Eusthenopteron (13) Bothriolepis (18) Diplacanthus (2) Homalacanthus (2) Triazeugacanthus (2) Cheirolepis (4) Fleurantia (1) 42 19 Bothriolepis (23) Scaumenacia (16) Eusthenopteron (10) Triazeugacanthus (6) Homalacanthus (2) Endeiolepis (1) Elpistostege (1) Quebecius (2) N = 41 50 Scaumenacia (9) 64 4 52 Fleurantia (1) Homalacanthus (1) 44 32 43 36 25 100 Euphanerops (1) 55 46 2 40 Eusthenopteron (2) 100 Diplacanthus (1) Triazeugacanthus (2) Cheirolepis (2) Bothriolepis (4) Scaumenacia (4) Sequence 5 Transgressive 42 31 52 23 Transgressive Bothriolepis (2) 55 34 17 N = 45 N = 21 Porolepiformes indet. (1) Euphanerops (3) Homalacanthus (9) Bothriolepis (11) Eusthenopteron (15) Triazeugacanthus (15) Scaumenacia (11) Fleurantia (3) Plourdosteus (4) Diplacanthus (2) Escuminaspis (2) Miguashaia (1) Cheirolepis (3) N = 28 Eusthenopteron (7) 34 Fleurantia (1) Triazeugacanthus (1) 65 47 76 80 56 Triazeugacanthus (1) Euphanerops (1) 65 Cheirolepis (1) 1 100 53 38 Scaumenacia (4) 40 35 Eusthenopteron (1) Scaumenacia (2) Homalacanthus (1) N=8 Eusthenopteron (4) 76 Fleurantia (1) Quebecius (1) Escuminaspis (2) Bothriolepis (6) Scaumenacia (7) Eusthenopteron (5) Levesquaspis (2) Endeiolepis (5) Homalacanthus (1) Bothriolepis (10) N = 16 N = 13 100 Regressive Scaumenacia (3) N = 14 Fig. 15 Relationships between species composition and sequence stratigraphy of the Escuminac Formation. Total dendrograms include all phases of a sequence. Cluster analyses are based on the Dice similarity index with average linkage. Names given in bold correspond to the ubiquitous component of the comprehensive assemblage. Shaded area corresponds to species unique to the phase or the sequence. Bootstrap values for 10,000 replicates are given at the dichotomies. Numbers in parentheses correspond to the number of beds in which the species was found 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Dice similarity Dice similarity Dice similarity Dice similarity 315 316 Palaeobio Palaeoenv (2011) 91:293–323 Duration of the Escuminac Formation The timespan encompassed by the complete 119-m thickness of the Escuminac Formation has never been evaluated precisely. Only two biostatistical studies (Cloutier 1997; Schultze 1984) have ever addressed this temporal issue because statistical samples have to originate from a short time interval and to belong to one population or species. Three criteria were used by Cloutier (1997) to estimate the timespan of the Escuminac Formation between 1 and 2 Ma: (1) a palynological criterion, (2) a sedimentological criterion and (3) a palaeoecological criterion. First, a single miosporal biozone (undifferentiated Zone BJ-BM) is found throughout the formation (Cloutier et al. 1996), thus most likely representing a timespan of less than 3–4 Ma (Richardson et al. 1984). Although conodonts are absent from the Escuminac Formation, it has been suggested that the miosporal biozone was located within the transitanshassi standard conodont zone (Cloutier et al. 1996; Streel and Loboziak 1996). Recent calibration of the Devonian time scale supports the idea that the complete Frasnian lasted 7.6 Ma (Kaufmann 2006) and that the transitans-hassi standard conodont zone could have lasted approximately 2.5 Ma. Thus, based on relative association between the two systems of biozonation and the new time scale, the Escuminac Formation would represent a timespan of less than 2.5 Ma. Second, the sedimentological criterion relies on the estimation of sedimentation rates. Based on the sedimentation rate calculated for the New York Devonian basin (approx. 45 m per Ma; Bayer and McGhee 1989), the 119-m thickness of the Escuminac Formation was assumed to have occurred in 2.6 Ma (Cloutier et al. 1996). Although the New York basin setting is not strictly comparable to that of the Escuminac Formation, similarities in the temporal and palaeogeographical contexts provide a crude estimate. On the other hand, Holocene sedimentation rates in central basins of wavedominated estuaries range from 0.2 to 2 mm/year (Brooke 2003), whereas it varies between 0.02 and 1 mm/year on tidal flats and delta plains (Einsele 1992); Brooke’s estimate takes out the few hundred years of human influence on sedimentation rates. Based on these recent sedimentation rates in comparable environments, the timespan estimate for the Escuminac Formation varies between 59,500 years and 5.95 Ma. The overall application of sequence stratigraphic principles, which explain how sediments are preserved in sedimentary basins due to the relative changes in the position of the baseline, is independent of temporal and spatial scales (Posamentier and Vail 1988; Posamentier et al. 1988; Vail et al. 1977). These principles were successfully applied in the interpretation of flume-scale experiments (Koss et al. 1994; Paola 2000; Paola et al. 2001; Wood et al. 1993) on continent-wide sedimentation (Sloss 1981, 1988) and to our study despite the lack of data on precise age dating and on the lateral extent of the depositional sequences. Recent reconstruction of the Palaeozoic global sea-level curve (Haq and Schutter 2008) shows that eustasy may explain the overall synchronicity of depositional sequences preserved on different cratonic margins worldwide. Most of these sea-level cycles are 0.5–2 Ma long (3rd order) with interferences of 0.4-Ma-long (4th order), long-term eccentricity-driven cycles at specific periods in the Palaeozoic, such as in the Middle to early Late Devonian (Haq and Schutter 2008). The latter Milankovitch-scale sequences are typically 1–10 m thick, whereas the third order ones are tens of metres to 100 m-thick. We recognised four complete and two incomplete 1- to 10-mscale sequences in the Escuminac Formation of the Miguasha Section. If none of those sequences is autogenic (e.g. lobe avulsion), a reasonable estimate of the duration of deposition for the Escuminac Formation would be approximately 1.6– 2.5 Ma. This estimate is close to the duration of the transitanshassi conodont zone extending from 380.3 to 382.8 (± approx. 3 Ma) (Kaufmann 2006). This time window comprises two major unconformities at the boundaries of three major sequences, dated at 380.8 and 382 Ma in the East Iowa, Alberta, Williston basin, Dinant Synclinorium, Czech Republic and Libya world-class Devonian sections (Haq and Schutter 2008). This comparison shows that some of the sequence boundaries observed in the Miguasha section of the Escuminac Formation might only be local and the expression of either simple autogenic processes or local tectonic deformation and not related to broad-scale tectonic or eustatic events. Finally, the stasis of a palaeoecological association composed of Escuminaspis laticeps, Bothriolepis canadensis, Scaumenacia curta and Eusthenopteron foordi without apparent morphological changes within each one of these species suggests a relatively short period of time (Cloutier 1997). Given the notion of palaeoecological stability (DiMichele et al. 2004) not having been addressed for a Palaeozoic fish assemblage (Brett et al. 1996; Ivany et al. 2009), the initial assumption was defensible. However, in the context of palaeoecological stability, a palaeoecological association is not necessary synonymous with a short timespan. DiMichele et al. (2004) presented four factors to explain the persistence of community composition over geological time: (1) ecological species interactions (e.g. mutualism, competition, predation), (2) significant overlap in species environmental tolerance, (3) geographic isolation and (4) the “law of large numbers”. The present study recognises an ubiquitous assemblage composed of three species rather than the original four species (Escuminaspis being excluded). Ecological interactions among the three species are unclear because of the lack of apparent interaction other than the interpretation that Scaumenacia and Bothriolepis are benthivores. Environmental tolerance might be comparable among the three species because they Palaeobio Palaeoenv (2011) 91:293–323 317 are all encountered in the different lithofacies with similar patterns of occurrence (Fig. 5). Although not at the species level, the three genera have a wide distribution (Schultze and Cloutier 1996), which suggests comparable broad environmental tolerance as well as broad palaeogeographic distribution. Thus, there is no clear indication of geographic isolation. Finally, the “law of large numbers”, even if it is not a robust criterion according to DiMichele et al. (2004), suggests the recognition of a persistent assemblage. The three species are among the most abundant species of the Escuminac Formation, and abundant larval–juvenile individuals have been found for these three species (Cloutier et al. 2009). The ubiquitous assemblage has strong potential to be considered a persistent palaeoecological assemblage. This assemblage might be considered stable (DiMichele et al. 2004) because it returned to its equilibrium state after each perturbation event (regressive phase of the sequence). Based on the different lines of evidence, the Escuminac Formation lasted between 59,500 years, which might correspond to the time effectively recorded by sediment (Couëffé et al. 2001), and 2.5 Ma, based on estimates with spore and conodont biozonation encompassing the time of nondeposition and erosion. Palaeodiversity of the Escuminac Formation Siluro-Devonian palaeocommunities of lower vertebrates are assumed to be correlated with their geographic areas, the sediment types and, more rarely, with biologic factors (Blieck and Janvier 1999). This correlation is based on the assumption that a palaeocommunity is a fixed entity. However, fish communities are dynamic assemblages that respond and/or adapt to their habitat, even more so if we are dealing with a fluctuating environment, such as an estuary. Our results show clearly that the Escuminac assemblage is a dynamic assemblage with a persistent component evolving in a fluctuating environment. Palaeodiversity of the Escuminac Formation was analysed quantitatively for the first time by Schultze and Cloutier (1996), who compared 40 Middle–Late Devonian fish lists to identify palaeogeographical and palaeoenvironmental similarities. Cloutier et al. (1996) provided the first cluster analysis of the Escuminac assemblage to recognise clusters of species; they observed that there was no clear biozonation within the formation. Differences between Cloutier et al.’s (1996) cluster analysis and our comprehensive cluster analysis stem from the usage of Euclidean distance rather than the Dice similarity index; the Euclidean distance could yield misleading results when applied to taxon-occurrence data because of the double-zero problem (Gagné and Proulx 2009). However, neither study considered intraformational variation in fish composition. The Escuminac assemblage was regarded as an entity (a faunal list), which is not the case, as demonstrated in our study. Fishes are autochthonous to the Escuminac environment. The Escuminac fish assemblage corresponds to the palaeocommunity and is not the result of a mixing owing to preburial transport or a selective or random taphonomic concentration of specimens. This interpretation is suggested because (1) articulated and complete specimens of all vertebrate species have been found (with the exception of Elpistostege watsoni), (2) exceptional preservation has been observed in a few species (Janvier and Arsenault 2009; Parent and Cloutier 1996), (3) coprolitic material co-occurs with fossil fishes, (4) the inclusions within the piscine coprolites and gastric cavities are representative of the macrofauna (Cloutier 2009; McAllister 1996) and (5) the most abundant species constitute an ubiquitous component of the assemblage. The Escuminac fish assemblage is fairly diversified compared to 180 Devonian fish localities (Fig. 16) (Cloutier and Lelièvre 1998; Schultze and Cloutier 1996). The fish richness in recent estuaries is fairly important since the environment supports its own resident fish community and is used by a variety of fish as nursery grounds, migration routes and refuge areas (Elliott et al. 2007). A large component of the assemblage, including the ubiquitous component, represents a resident fish community because of its presence in the various lithofacies as well as the transgressive and regressive phases of the sequences. Because of their extended distribution within the Escuminac Formation, it is likely that most fish species were euryhaline because they occupied different habitats within the estuarine system. The sporadic occurrence of other components of the complete assemblage might well represent species that 70 60 50 40 30 20 10 10 20 30 40 N = 180 Max = 61 Mean = 3.8 S.D. = 8.39 Escuminac Formation 50 60 70 Richness per Devonian fish localities Fig. 16 Histogram of fish richness for 180 Devonian fish localities around the world. Only localities with macro-remains have been included. Original data came from various sources (Blieck and Turner 2000; Cloutier and Lelièvre 1998; Dineley and Metcalf 1999; Janvier and Suarez-Riglos 1986; Schultze and Cloutier 1996) 318 Palaeobio Palaeoenv (2011) 91:293–323 used part of the Escuminac setting as a transitional habitat between truly marine and freshwater habitats. Larval–juvenile individuals of 14 out of 20 species (Cloutier et al. 2009) have been identified in various levels of the Escuminac Formation, with a high frequency in beds 8 and 366. Most species are using the palaeoestuary as nursery grounds. Some species might have used habitats outside the Escuminac palaeoestuary as spawning sites or nursery ground; however, it is also likely that larval–juveniles of some species have not been discovered yet within the Escuminac Formation. Various fluctuating regimes have structured the biotic components of the Escuminac Formation. Large-scale variation shows a continentalisation of the formation. This transition is illustrated by (1) an increase of abundance and size of continental plant debris, (2) an increase in the abundance and occurrence of continental arthropods, (3) a reduction of the abundance of conchostracans and (4) a reduction and depletion of acritarchs. Sea-level fluctuations seem to have played a role in the variation of richness and species composition of the fish assemblage. Fluctuations of shorter temporal scale fluctuations have influenced variation in the abundance of the various components (plants, Asmusia and fishes). Konservat and Konzentrat Fossil-Lagerstätten within the Escuminac Formation first type, specimens are found in laminites associated with tidal deposits, thus resulting from a rapid burial with a minimal benthic life. In the second type, specimens are found in thick siltstone units originating from rapid burial (obrution events) in higher energy environments. Other Devonian and Carboniferous fish Fossil-Lagerstätten have been associated with thinly laminated tidal lithofacies and anoxic events (Feldman et al. 1993, 1994; Kuecher et al. 1990; Schultze 1999; Trewin 1986; Upeniece 2001). Threedimensional preservation of articulated fish has also been reported in a few Devonian localities associated either with reefal (Long and Trinajstic 2010) or fluvial palaeoenvironments (Daeschler et al. 2006; Johanson 1997). Rare are the fossil-fish localities for which both types of preservation are encountered; the co-occurrence of these two types is most likely a result of the palaeoenvironmental shift in the Escuminac Formation. Considering benthic macro-invertebrates in a marine setting, Brett (1995) and Brett et al. (2009) suggested that Konservat/Konzentrat Lagerstätten are frequently associated with rapid burial and or anoxia and are, consequently, characteristic of the transgressive to early high-stand interval of a sequence. Both Konservat and Konzentrat Lagerstätte horizons of the Escuminac Formation fit with this model. Conclusions There is a basic assumption that only Konservat Lagerstätte horizons provide an accurate vision of a palaeocommunity (Kidwell and Flessa 1995). Although richness and abundance are greatly influenced by both Konservat and Konzentrat Lagerstätte horizons, all states of preservation help in defining the species composition of the assemblage as well as related biotic fluctuations. The Escuminac Formation is frequently considered to be a Fossil-Lagerstätte. Such a broad generalisation is similar to that for other fish sites, such as the Devonian Gogo Station (Long and Trinajstic 2010), the Carboniferous Bear Gulch (Feldman et al. 1994; Hagadorn 2002), the Jurassic Solnhofen Plattenkalk (Viohl 1994), the Triassic Grès de Vosges (Gall et al. 1974) and the Eocene Pescaria di Bolca site (Papazzoni and Trevisani 2006; Schwark et al. 2009). However, as shown in this study, most beds of the Escuminac Formation are devoid of fish remains, and isolated bony elements are the most frequent state of preservation encountered. Nevertheless, a few horizons (beds) can be recognised as Fossil-Lagerstätten. Two distinct types of Konservat-Lagerstätte horizons are recognised within the Escuminac Formation: (1) completely articulated, flattened specimens with soft-tissue preservation and (2) completely articulated, three-dimensional specimens with occasional soft-tissue preservation. In the Through the geological record, palaeodiversity has varied in terms of richness, abundance and species composition at different spatial and temporal scales. The record of this variation is influenced or biased by the palaeoenvironment in which the organisms were living and buried. During the Devonian, major evolutionary and environmental changes influenced palaeodiversity in relation to the continentalisation of ecosystems. In order to evaluate potential sources of variation in palaeodiversity at a local spatial scale and a relatively short temporal scale, we investigated the Late Devonian biota (primarily the fishes) of the Escuminac Formation, Miguasha, eastern Canada. The Parc national de Miguasha, which protects the Escuminac Formation, has been declared a UNESCO World Heritage because it is considered the palaeontological site most representative of the Devonian period. The depositional setting of the Escuminac Formation and its fauna is typical of transitional environments associated with land–sea interactions. The Escuminac Formation is interpreted as a wave-dominated estuarine depositional environment. A fluvial freshwater source fed the estuary, which was closed temporarily by a beach sand-bar. The environments correspond to the internal zones of an estuary, including a fluvial discharge system, a Palaeobio Palaeoenv (2011) 91:293–323 319 National de la Recherche Scientifique, France)–invited-chair position in Geosciences at the University of Rennes 1. We would like to thank Florentin Paris and Romain Vullo for stimulating discussions. Isabelle Béchard helped in preparing fig. 2. Johanne Kerr and Olivier Matton helped with the photography. Mark V. H. Wilson and Susan Turner constructively reviewed the manuscript. This research was funded by a research grant from the Natural Sciences and Engineering Research Council of Canada (238612) and the Research chair in Paleontology and Evolutionary Biology (UQAR-Parc national de Miguasha). Fieldwork (1993–1994) was supported by the CNRS. bay-head delta and the anoxic proximal part of a central basin. The living conditions were those of brackish water with weak energy subjected to flooding and undergoing temporary periods of anoxia when the basin was closed, particularly near the bottom of the central basin where reducing conditions can persist. The detailed changes through time in depositional environments of the Escuminac Formation show metre-thick variations from proximal to distal environments that can translate into deepening and shallowing upward variations or base-level changes. The timespan for the Escuminac Formation and its faunal assemblage is estimated to be between 59,500 years (minimal estimate based on sedimentation rate in similar depositional environments) and 2.5 Ma (maximal estimate based on spore and conodont biozonation tied to sea-level curves). The Escuminac assemblage is fairly diversified, including 20 vertebrate and 12 invertebrate species. A large component of the fish assemblage, including the ubiquitous component (the placoderm Bothriolepis canadensis, the dipnoan Scaumenacia curta and the osteolepiform Eusthenopteron foordi), represents a resident community of euryhaline fishes. However, sporadic species might have used part of the Escuminac setting as transitional habitat between truly marine and freshwater habitats. Most Escuminac species used the palaeoestuary as nursery grounds. The composition of the Escuminac fish assemblage fluctuates in relation to the progression of the depositional environment, although the ubiquitous assemblage is recovered in all environments. An initial assemblage rapidly colonised the intertidal environment. Sporadic invasion occurred during transgressive phases, but none of them was persistent. The regressive phases of the lower three sequences display poorly diversified and/or structured assemblages. Transgressive phases are more diversified and better structured than aggradational and regressive phases. Certain horizons of the Escuminac Formation are interpreted either as Konservat or Konzentrat Lagerstätten, with some rare horizons fulfilling both conditions. The laminites associated with tidal deposits are optimal in terms of both types of Lagerstätten. Fossil-Lagerstätten are associated with rapid burial and/or anoxia, and thus they are characteristic of the transgressive to early high-stand interval of a sequence. The new interpretation of the palaeodiversity and palaeoenvironment of the Late Devonian Escuminac assemblage provides a model for further studies to better understand one of the most important evolutionary events in the history of vertebrates, namely, the transition from fishes to tetrapods. Acknowledgments A draft of this paper was written during the 2008–2009 sabbatical leave of Richard Cloutier on a CNRS (Centre References Algeo TJ, Ingall E (2007) Sedimentary Corg:P ratios, paleocean ventilation, and Phanerozoic atmospheric pO2. Palaeogeogr Palaeoclimatol Palaeoecol 256:130–155 Algeo TJ, Scheckler SE (1998) Terrestrial-marine teleconnections in the Devonian: links between the evolution of land plants, weathering processes, and marine anoxic events. Philos Trans R Soc Lond B Biol Sci 353:113–128 Algeo TJ, Scheckler SE, Maynard JB (2001) Effects of the Middle to Late Devonian spread of vascular land plants on weathering regimes, marine biotas, and global climate. In: Gensel PG, Edwards D (eds) Plants invade the land. Evolutionary & environmental perspectives. Columbia University Press, New York, pp 213–236 Allen GP (1991) Sedimentary processes and facies in the Gironde estuary—a model for macrotidal estuarine systems. In: Smith DG, Reinson GE, Zaitlin BA, Rahmani RA (eds) Clastic tidal sedimentology. Can Soc Petrol Geol Mem 16:219–226 Anderson PSL, Friedman M, Brazeau MD, Rayfield EJ (2011) Initial radiation of jaws demonstrated stability despite faunal and environmental change. Nature 476:206–209 Arratia G, Cloutier R (1996) A reassessment of the morphology of Cheirolepis canadensis (Actinopterygii). In: Schultze H-P, Cloutier R (eds) Devonian fishes and plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 165–197 Arsenault M (1982) Eusthenopteron foordi, a predator on Homalacanthus concinnus from the Escuminac Formation, Miguasha, Quebec. Can J Earth Sci 19:2214–2217 Ashton M (1981) Carbonate tidal rhythmites from the Middle Jurassic of Britain. Sedimentology 28:689–698 Bateman RM, Crane PR, DiMichele WA, Kenrick PR, Rowe NP, Speck T, Stein WE (1998) Early evolution of land plants: phylogeny, physiology, and ecology of the primary terrestrial radiation. Annu Rev Ecol Syst 29:263–292 Bayer U, McGhee GR (1989) Periodicity of Devonian sedimentary and biological perturbation: implications for the Devonian timescale. N Jb Geol Paläont Mh 1989:1–16 Blieck A, Cloutier R (2000) Biostratigraphical correlations of Early Devonian vertebrate assemblages of the Old Red Sandstone Continent. Cour Forsch-Inst Senckenberg 223:223–269 Blieck A, Janvier P (1999) Silurian–Devonian vertebrate dominated communities, with particular reference to agnathans. In: Boucot AJ, Lawson JD (eds) Paleocommunities: a case study from the Silurian and Lower Devonian, vol 11. World and regional geology series. Cambridge University Press, Cambridge, pp 79–105 Blieck A, Turner S (eds) (2000) Palaeozoic vertebrate biochronology and global marine/non-marine correlation. Final Report of IGCP 328 (1991–1996), vol 223. Cour Forsch-Inst Senckenberg, Frankfurt Brett CE (1995) Sequence stratigraphy, biostratigraphy, and taphonomy in shallow marine environments. Palaios 10:597–616 320 Brett CE, Ivany LC, Schopf KM (1996) Coordinated stasis: an overview. Palaeogeogr Palaeoclimatol Palaeoecol 127:1–20 Brett CE, Bartholomew AJ, Baird GC (2007) Biofacies recurrence in the Middle Devonian of New York state: an example with implications for evolutionary paleoecology. Palaios 22:306–324. doi:10.2110/palo.2005.p05-027r Brett CE, Allison PA, DeSantis MK, Liddell WD, Kramer A (2009) Sequence stratigraphy, cyclic facies, and Lagerstätten in the Middle Cambrian Wheeler and Marjum Formations, Great Basin, Utah. Palaeogeogr Palaeoclimatol Palaeoecol 277:9–33 Brooke B (2003) The role of sedimentological information in estuary management. In: Thorn B, Coughlan P (eds) Proc Coast to Coast 2002–Source to sea. Tweed Heads, Coastal Council, Brisbane, pp 31–34 Chapman DM, Geary M, Roy PS, Thorn BC (1982) Coastal evolution and coastal erosion in New South Wales. Coastal Council of N.S. W., Sydney Chidiac Y (1996) Paleoenvironmental interpretation of the Escuminac Formation based on geochemical evidence. In: Schultze H-P, Cloutier R (eds) Devonian fishes and plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 47–53 Choi KS, Kim BO, Park YA (2001) Late Pleistocene tidal rhythmites in Kyunggi Bay, west coast of Korea: a comparison with simulated rhythmites based on modern tides and implications for intertidal positioning. J Sediment Res 71:680–691 Clack JA (2007) Devonian climate change, breathing, and the origin of the tetrapod stem group. Integr Comp Biol 47:510–523. doi:10.1093/icb/icm055 Cloutier R (1996) Taxonomic review of Eusthenopteron foordi. In: Schultze H-P, Cloutier R (eds) Devonian fishes and plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 271–284 Cloutier R (1997) Morphologie et variations du toit crânien du Dipneuste Scaumenacia curta (Whiteaves) (Sarcopterygii), du Dévonien supérieur du Québec. Geodiversitas 19:59–105 Cloutier R (2009) De l'eau à la terre. Parc national de Miguasha. 2nd edn. Muséologie In situ, Quebec Cloutier R (2010) The fossil record of fish ontogenies: insights into developmental patterns and processes. Semin Cell Dev Biol 21:400–413. doi:10.1016/j.semcdb.2009.11.004 Cloutier R, Lelièvre H (1998) Comparative study of the fossiliferous sites from the Devonian. Ministère de l'Environnement et de la Faune, Gouv Québec Cloutier R, Loboziak S, Candilier A-M, Blieck A (1996) Biostratigraphy of the Upper Devonian Escuminac Formation, eastern Quebec, Canada: a comparative study based on miospores and fishes. Rev Palaeobot Palynol 93:191–215 Cloutier R, Béchard I, Charest F, Matton O (2009) La contribution des poissons fossiles de Miguasha à la biologie évolutive du développement. Nat Can 133:84–95 Copper P (2002) Silurian and Devonian reefs: 80 million years of greenhouse between two ice ages. In: Kiessling W, Flügel E, Golonka J (eds) Phanerozoic reef patterns, vol 72. Special publication. Society for Sedimentary Geology (SEPM), Tulsa, pp 181–238 Couëffé R, Tessier B, Gigot P, Beaudoin B (2001) Le temps préservé sous forme de sédiments: résultats semi-quantitatifs obtenus dans la molasse marine miocène du bassin de Digne (Alpes-de-HauteProvence, Sud-Est de la France). C R Acad Sci Sér II 332:5–11 Daeschler EB, Shubin NH, Jenkins FA (2006) A Devonian tetrapod-like fish and the evolution of the tetrapod body plan. Nature 440:757– 763. doi:10.1038/nature04639 Dalrymple RW, Choi K (2007) Morphologic and facies trends through the fluvial-marine transition in tide-dominated depositional systems: a schematic framework for environmental and sequence-stratigraphic interpretation. Earth Sci Rev 81:135– 174. doi:10.1016/j.earscirev.2006.10.002 Palaeobio Palaeoenv (2011) 91:293–323 Davies NS, Gibling MR (2009) Cambrian to Devonian evolution of alluvial systems: the sedimentological impact of the earliest land plants. Earth Sci Rev 98:171–200 DiMichele WA, Bateman RM (2005) Evolution of land plant diversity: major innovations and lineages through time. In: Krupnick G, Kress WJ (eds) Plant conservation: a natural history approach. University of Chicago Press, Chicago, pp 3–14 DiMichele WA, Hook RW, Beerbower JR, Boy JA, Gastaldo RA, Hotton N, Phillips TL, Scheckler SE, Shear WA, Sues H-D (1992) Paleozoic terrestrial ecosystems. In: Behrensmeyer AK, Damuth JD, DiMichele WA, Potts R, Sues H-D, Wing SL (eds) Evolutionary paleoecology of terrestrial plants and animals. University of Chicago Press, Chicago, pp 205–235 DiMichele WA, Behrensmeyer AK, Olszewski TD, Labandeira CC, Pandolfi JM, Wing SL, Bobe R (2004) Long-term stasis in ecological assemblages: evidence from the fossil record. Annu Rev Ecol Evol Syst 35:285–322. doi:10.1146/annurev.ecolsys.35.120202.110110 Dineley DL (1999) Late Devonian fossil fishes sites of Scotland. In: Dineley DL, Metcalf SJ (eds) Fossil fishes of Great Britain. Geological conservation review series. Joint Nature Conservation Committee, Peterborough, pp 241–262 Dineley DL, Williams BPF (1968) Sedimentation and paleoecology of the Devonian Escuminac Formation and related strata, Escuminac Bay, Quebec. In: Klein GdV (ed) Late Paleozoic and Mesozoic continental sedimentation, northeastern North America. Special Paper. Geological Society of America, Boulder, pp 241–264 Dineley DL, Metcalf SJ (eds) (1999) Fossil fishes of Great Britain. The geological conservation review series. Joint Nature Conservation Committee, Peterborough Einsele G (1992) Sedimentary basins. Evolution, facies and sediment budget. Springer, Berlin Heidelberg New York El Albani A, Cloutier R, Candilier A-M (2002) Early diagenesis of the Upper Devonian Escuminac Formation in the Gaspé Peninsula, Québec: sedimentological and geochemical evidence. Sediment Geol 146:209–223 Elder RL, Smith GR (1984) Fish taphonomy and paleoecology. Geobios Mém Spéc 8:287–291 Elder RL, Smith GR (1988) Fish taphonomy and environmental inference in paleolimnology. Palaeogeogr Palaeoclimatol Palaeoecol 62:577–592 Elliott DK, Johnson HG, Cloutier R, Carr RK, Daeschler EB (2000) Middle and Late Devonian vertebrates of the western Old Red Sandstone Continent. Cour Forsch-Inst Senckenberg 223:291–308 Elliott M, Whitfield AK, Potter IC, Blaber SJM, Cyrus DP, Nordlie FG, Harrison TD (2007) The guild approach to categorizing estuarine fish assemblages: a global review. Fish Fisheries 8:241–268 Feldman HR, Archer AW, Kvale EP, Cunningham CR, Maples CG, West RR (1993) A tidal model of Carboniferous KonservatLagerstätten formation. Palaios 8:485–498 Feldman HR, Lund R, Maples CG, Archer AW (1994) Origin of the Bear Gulch beds (Namurian, Montana, USA). Geobios (spec issue) 16:283–291 Gagné SA, Proulx R (2009) Accurate delineation of biogeographical regions depends on the use of an appropriate distance measure. J Biogeogr 36:561–562. doi:10.1111/j.1365-2699.2008.01990.x Gall JC, Grauvogel L, Lehman J-P (1974) Faune du Buntsandstein. V. Les poissons fossiles de la collection Grauvogel-Gall. Ann Paléontol (Vertébr) 60:129–170 Hagadorn JW (2002) Bear Gulch: an exceptional Upper Carboniferous Plattenkalk. In: Bottjer DJ, Etter W, Hagadorn JW, Tang CM (eds) Exceptional fossil preservation. A unique view on the evolution of marine life. Columbia University Press, New York, pp 167–183 Hammer Ø, Harper DAT, Ryan PD (2001) PAST: Paleontological statistics software package for education and data analysis (version 1.56b). Palaeontol Electron 4. Available at: http:// palaeo-electronica.org/2001_2001/past/issue2001_2001.htm Palaeobio Palaeoenv (2011) 91:293–323 Haq BU, Schutter SR (2008) A chronology of Paleozoic sea-level changes. Science 322:64–68. doi:10.1126/science.1161648 Hesse R, Sawh H (1992) Geology and sedimentology of the Upper Devonian Escuminac Formation, Quebec, and evaluation of its paleoenvironment: lacustrine versus estuarine turbidite sequence. Atl Geol 28:257–275 Ivany LC, Brett CE, Wall HLB, Wall PD, Handley JC (2009) Relative taxonomic and ecologic stability in Devonian marine faunas of New York State: a test of coordinated stasis. Paleobiology 35:499–524 Janvier P, Arsenault M (2007) The anatomy of Euphanerops longaevus Woodward, 1900, an anaspid-like jawless vertebrate from the Upper Devonian of Miguasha, Quebec, Canada. Geodiversitas 29:143–216 Janvier P, Arsenault M (2009) La conservation exceptionnelle des fossiles du parc national de Miguasha. Nat Can 133:78–83 Janvier P, Suarez-Riglos M (1986) The Silurian and Devonian vertebrates of Bolivia. Bull Inst Fr Etudes Andines 15:73–14 Janvier P, Desbiens S, Willett JA, Arsenault M (2006) Lamprey-like gills in a gnathostome-related Devonian jawless vertebrate. Nature 440:1183–1185 Janvier P, Desbiens S, Willett JA (2007) New evidence for the controversial "lungs" of the Late Devonian antiarch Bothriolepis canadensis (Whiteaves, 1880) (Placodermi: Antiarcha). J Vertebr Paleontol 27:709–710 Jeram AJ (1996) Chelicerata from the Escuminac Formation. In: Schultze H-P, Cloutier R (eds) Devonian Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 103–111 Joachimski MM, Buggisch W (1993) Anoxic events in the Late Frasnian— Causes of the Frasnian-Famennian faunal crisis? Geology 21:675–678 Johanson Z (1997) New Remigolepis (Placodermi; Antiarchi) from Canowindra, New South Wales, Australia. Geol Mag 134:813– 846 Johnson JG, Klapper G, Sandberg CA (1985) Devonian eustatic fluctuations in Euramerica. Geol Soc Am Bull 96:567–587 Jutras P, Prichonnet G, Utting J (2001) Newly identified Carboniferous units (the Pointe Sawyer and Chemin-des-Pêcheurs formations) in the Gaspé Peninsula, Quebec; implications regarding the evolution of the northwestern sector of the Maritimes Basin. Can J Earth Sci 38:1–19 Kaufmann B (2006) Calibrating the Devonian time scale: a synthesis of U-Pb ID-TIMS ages and conodont stratigraphy. Earth Sci Rev 76:175–190. doi:10.1016/j.earscirev.2006.01.001 Kenrick P, Crane PR (1997) The origin and early evolution of plants on land. Nature 389:33–39 Kidwell SM, Flessa KW (1995) The quality of the fossil record— populations, species, and communities. Annu Rev Ecol Syst 26:269–299 Klug C, Kroger B, Kiessling W, Mullins GL, Servais T, Fryda J, Korn D, Turner S (2010) The Devonian nekton revolution. Lethaia 43:465–477. doi:10.1111/j.1502-3931.2009.00206.x Koss JE, Ethridge FG, Schumm SA (1994) An experimental study on the effects of base-level change on fluvial, coastal plain and shelf systems. J Sediment Res B64:90–98 Kuecher GJ, Woodland BG, Broadhurst FM (1990) Evidence of deposition from individual tides and of tidal cycles from the Francis Creek Shale (host rock to the Mazon Creek biota), Westphalian D (Pennsylvanian), Northeastern Illinois. Sediment Geol 68:211–221 Kussell E, Leibler S (2005) Phenotypic diversity, population growth, and information in fluctuating environments. Science 309:2075–2078 Legendre P, Legendre L (1998) Numerical ecology. Developments in environmental modelling, 2nd English edn. Elsevier Science, Amsterdam Levins R (1968) Evolution in changing environments: some theoretical explanations. Monographs in population biology. Princeton University Press, Princeton Long JA, Trinajstic K (2010) The Late Devonian Gogo Formation Lagerstatte of Western Australia: exceptional early vertebrate 321 preservation and diversity. Annu Rev Earth Planet Sci 38:255– 279. doi:10.1146/annurev-earth-040809-152416 Maples CG (1996) Paleoenvironmental significance of trace fossils in the Escuminac Formation. In: Schultze H-P, Cloutier R (eds) Devonian Fishes and Plants of Miguasha, Québec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 114–119 Martens T (1996) Conchostraca (Phyllopoda, Crustacea) from the Escuminac Formation. In: Schultze H-P, Cloutier R (eds) Devonian fishes and plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 112–113 Matton O, Stevenson R, Cloutier R (2006) Isotopic and geochemical analyses of fossil fish remains from the Upper Devonian Escuminac Formation (Miguasha, Quebec): implications for paleoenvironmental interpretation. J Vertebr Paleontol 26:97A Mazumder R, Arima M (2005) Tidal rhythmites and their implications. Earth Sci Rev 69:79–95 McAllister J (1996) Coprolitic remains from the Devonian Escuminac Formation. In: Schultze H-P, Cloutier R (eds) Devonian fishes and plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 328–347 McGhee GR (2001) The 'multiple impacts hypothesis' for mass extinction: a comparison of the Late Devonian and the late Eocene. Palaeogeogr Palaeoclimatol Palaeoecol 176:47–58 Meyers LA, Bull JJ (2002) Fighting change with change: adaptive variation in an uncertain world. Trends Ecol Evol 17:551–557 Moloshnikov SV (2008) The placoderm Plourdosteus livonicus (Eastman) in the Early Frasnian of the Central Devonian Field and the trophic structure of the Mikhailovskii fish assemblage. Paleontol J 42:607–614 Morrisey D, Underwood AJ, Chapman MG (1995) Estuaries. In: Underwood AJ, Chapman MG (eds) Coastal marine ecology of temperate Australia. University of New South Wales Press, Sydney, pp 152–170 Nichol SL (1991) Zonation and sedimentology of estuarine facies in an incised-valley, wave-dominated, microtidal setting, New South Wales, Australia. In: Smith DG, Reinson GE, Zaitlin BA, Rahmani RA (eds) Clastic tidal sedimentology. Can Soc Petrol Geol Mem 16: 41–58 Orr PJ, Briggs DEG (1999) Exceptionally preserved conchostracans and other crustaceans from the Upper Carboniferous of Ireland. In: Exceptionally Preserved Conchostracans and other Crustaceans from the Upper Carboniferous of Ireland. Spec Pap Palaeontology 62, The Palaeontological Association, London, pp 5–68 Paola C (2000) Quantitative models of sedimentary basin filling. Sedimentology 47:121–178 Paola C, Mullin J, Ellis C, Mohrig DC, Swenson JB, Parker G, Hickson T, Heller PL, Pratson L, Syvitski J, Sheets B, Strong N (2001) Experimental stratigraphy. GSA Today 11:4–9 Papazzoni CA, Trevisani E (2006) Facies analysis, palaeoenvironmental reconstruction, and biostratigraphy of the "Pesciara di Bolca" (Verona, northern Italy): an early Eocene Fossil-Lagerstätte. Palaeogeogr Palaeoclimatol Palaeoecol 242:21–35 Parent N, Cloutier R (1996) Distribution and preservation of fossils in the Escuminac Formation. In: Schultze H-P, Cloutier R (eds) Devonian fishes and plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 54–78 Pasternack GB, Brush GS (2002) Biogeomorphic controls on sedimentation and substrate on a vegetated tidal freshwater delta in upper Chesapeake Bay. Geomorphology 43:293–311 Paterson AW, Whitfield AK (2000) Do shallow-water habitats function as refugia for juvenile fishes? Estuar Coast Shelf Sci 51:359–364 Ponten A, Plink-Bjorklund P (2007) Depositional environments in an extensive tide-influenced delta plain, Middle Devonian Gauja Formation, Devonian Baltic Basin. Sedimentology 54:969–1006. doi:10.1111/j.1365-3091.2007.00869.x 322 Ponten A, Plink-Bjorklund P (2009) Regressive to transgressive transits reflected in tidal bars, Middle Devonian Baltic Basin. Sediment Geol 218:48–60. doi:10.1016/j.sedgeo.2009.04.003 Posamentier HW, Vail PR (1988) Eustatic controls on clastic deposition II—sequence and systems tract models. In: Wilgus CK, Hastings BS, Kendall CGSC, Posamentier HW, Ross CA, Van Wagoner JC (eds) Sea level changes—an integrated approach, vol 42. SEPM Spec Publ, San Francisco, pp 125–154 Posamentier HW, Jervey MT, Vail PR (1988) Eustatic controls on clastic deposition. I. Conceptual framework. In: Wilgus CK, Hastings BS, Kendall CGSC, Posamentier HW, Ross CA, Van Wagoner JC (eds) Sea level changes—an integrated approach, vol 42. SEPM Spec Publ, San Francisco, pp 110–124 Potter IC, Hyndes GA (1994) Composition of the fish fauna of a permanently open estuary on the southern coast of Australia, and comparisons with a nearby seasonally closed estuary. Mar Biol 121:199–209 Prichonnet G, DiVergilio M, Chidiac Y (1996) Stratigraphic, sedimentological and paleontological context of the Escuminac Formation: paleoenvironmental hypotheses. In: Schultze H-P, Cloutier R (eds) Devonian fishes and plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 23–36 Rainer SF, Fitzhardinge RC (1981) Benthic communities in an estuary with periodic deoxygenation. Aust J Mar Freshw Res 32:227–244 Richardson JB, Ford JH, Parker F (1984) Miospores, correlation and age of some Scottish Lower Old Red Sandstone sediments from the Strathmore region (Fife and Angus). J Micropalaeontol 3:109–124 Roy PS, Williams RJ, Jones AR, Yassini R, Gibbs PJ, Coates B, West RJ, Scanes PR, Hudson JP, Nichol S (2001) Structure and function of south-east Australian estuaries. Estuar Coast Shelf Sci 53:351–384 Rust BR (1982) Devonian and Carboniferous continental sedimentation of southern Gaspé peninsula. In: Hesse R, Middleton GV, Rust BR (eds) Paleozoic continental margin sedimentation in the Quebec Appalachians—11th Int Congress of Sedimentology: vol Excursion 7b. McGill University, Montreal, pp 69–96 Rust BR, Lawrence DA, Zaitlin BA (1989) The sedimentation and tectonic significance of Devonian and Carboniferous terrestrial successions in Gaspé, Québec. Atl Geol 25:1–13 Sallan LC, Coates MI (2010) End-Devonian extinction and a bottleneck in the early evolution of modern jawed vertebrates. Proc Natl Acad Sci USA 107:10131–10135. doi:10.1073/pnas.0914000107 Sandberg CA, Morrow JR, Ziegler W (2002) Late Devonian sea-level changes, catastrophic events and mass extinctions. In: Koeberl C, MacLeod KG (eds) Catastrophic events and mass extinctions: impacts and beyond, vol 356: special papers. Geological Society of America, Boulder, pp 473–487 Schäfer W (1972) Ecology and palaeoecology of marine environments (trans: Oertel I). Oliver and Boyd, Edinburgh Schmitz B, Aberg G, Werdelin L, Forey P, Bendix-Almgreen SE (1991) 87 Sr/ 86Sr, Na, F, Sr, and La in skeletal fish debris as a measure of the paleosalinity of fossil fish habitats. Bull Geol Soc Am 103:786–794 Schultze H-P (1984) Juvenile specimens of Eusthenopteron foordi Whiteaves, 1881 (osteolepiform rhipidistian, Pisces) from the Late Devonian of Miguasha, Quebec, Canada. J Vertebr Paleontol 4:1–16 Schultze H-P (1999) The fossil record of the intertidal zone. In: Horn MH, Martin KLM, Chotkowski MA (eds) Intertidal fishes: life in two worlds. Academic Press, San Diego, pp 373–392 Schultze H-P, Cloutier R (1996) Comparison of the Escuminac Formation ichthyofauna with other late Givetian/early Frasnian ichthyofaunas. In: Schultze H-P, Cloutier R (eds) Devonian fishes and plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Munich, pp 348–368 Palaeobio Palaeoenv (2011) 91:293–323 Schwark L, Ferretti A, Papazzoni CA, Trevisani E (2009) Organic geochemistry and paleoenvironment of the Early Eocene "Pesciara di Bolca" Konservat-Lagerstätte, Italy. Palaeogeogr Palaeoclimatol Palaeoecol 273:272–285 Selover RW, Gastaldo R, Nelson RE (2005) An estuarine assemblage from the Middle Devonian Trout Valley Formation of northern Maine. Palaios 20:192–197. doi:10.2210/ palo.2004.p04-16 Shear WA, Selden PA (2001) Rustling in the undergrowth: animals in early terrestrial ecosystems. In: Gensel PG, Edwards D (eds) Plants invade the land. Evolutionary & environmental perspectives. Columbia University Press, New York, pp 29–51 Sloss LL (1981) Comparative anatomy of cratonic unconformities. AAPG Bull 65:993–999 Sloss LL (1988) Forty years of sequence stratigraphy. Geol Soc Am Bull 100:1661–1665 Smith GR, Elder RL (1985) Environmental interpretation of burial and preservation of Clarkia fishes. In: Smiley CJ (ed) Late Cenozoic history of the Pacific Northwest. Interdisciplinary studies on the Clarkia Fossil Beds of Northern Idaho. AAAS, San Francisco, pp 85–93 Stein WE, Mannolini F, Hernick LV, Landing E, Berry CM (2007) Giant cladoxylopsid trees resolve the enigma of the Earth's earliest forest stumps at Gilboa. Nature 446:904–907. doi:10.1038/nature05705 Stephenson MH, Leng MJ, Michie U, Vane CH (2006) Palaeolimnology of Palaeozoic lakes, focussing on a single lake cycle in the Middle Devonian of the Orcadian Basin, Scotland. Earth Sci Rev 75:177–197. doi:10.1016/j.earscirev.2005.09.002 Streel M, Loboziak S (1996) Middle to Upper Devonian miospores. In: Jansonius JJ, McGregor DC (eds) Palynology: principles and applications, vol 2. AASP Foundation, Salt Lake City, pp 575–587 Tessier B (1993) Upper intertidal rhythmites in the Mont-St-Michel Bay (NW France)—perspectives for paleoreconstructions. Mar Geol 110:355–367 Trewin NH (1986) Palaeoecology and sedimentology of the Achanarras fish bed of the Middle Old Red Sandstone, Scotland. Trans R Soc Edinb Earth Sci 77:21–46 Upeniece I (2001) The unique fossil assemblage from the Lode Quarry (Upper Devonian, Latvia). Mitteilungen Mus Naturkunde Berlin–Geowissenschaftliche Reihe 4:101–119 Vail PR, Mitchum RM, Todd RG, Widmier JM, Thompson S, Sangree JB, Bubb JN, Hateilid WG (1977) Seismic stratigraphy and global changes in sea-level. In: Payton CE (ed) Seismic stratigraphy—application to hydrocarbon exploration. Mem Am Assoc Petrol Geol 26:49–212 Vézina D (1991) New observations on the environment of Escuminac Formation sedimentation (Upper Devonian, Frasnian) in Quebec. Can J Earth Sci 28:225–230 Viohl G (1994) Fish taphonomy of the Solnhofen Plattenkalk—an approach to the reconstruction of the palaeoenvironment. Geobios (spec issue) 16:81–90 Webster IT, Ford PW, Hodgson B (2002) Microphytobenthos contribution to nutrient–phytoplankton dynamics in a shallow coastal lagoon. Estuaries 25:540–551 Wehrmann A, Hertweck G, Brocke R, Jansen U, Konigshof P, Plodowski G, Schindler E, Wilde V, Blieck A, Schultka S (2005) Paleoenvironment of an Early Devonian land-sea transition: a case study from the southern margin of the Old Red Continent (Mosel Valley, Germany). Palaios 20:101–120. doi:10.2210/palo.2003.p03-33 Williams GE (1989) Late Precambrian tidal rhythmites in SouthAustralia and the history of the Earth’s rotation. J Geol Soc Lond 146:97–111 Wilson HM, Daeschler EB, Desbiens S (2005) New flat-backed archipolypodan millipedes from the Upper Devonian of North America. J Paleontol 79:738–744 Palaeobio Palaeoenv (2011) 91:293–323 Wolga H (1981) Similarity indices, sample size and diversity. Oecologia 50:296–302 Wood LJ, Ethridge FG, Schumm SA (1993) The effects of rate of base-level fluctuation on coastal-plain, shelf and slope depositional systems; an experimental approach. In: Posamentier HW, Summerhayes CP, Haq BU, Allen GP (eds) Sequence stratigraphy and facies associations, vol 18 (special publication). International Association of Sedimentologists, pp 43–53 Young GC (2007) Devonian formations, vertebrate faunas and age control on the far south coast of New South Wales and adjacent Victoria. Austr J Earth Sci 54:991–1008 323 Zaitlin BA (1981) Sedimentology of the Pirate Cove, Fleurant and Bonaventure Formations of the western Baie des Chaleurs area, Maritime Canada: a depositional and tectonic model. MSc thesis. University of Ottawa, Ottawa Zhu M, Zhao WJ, Jia LT, Lu J, Qiao T, Qu QM (2009) The oldest articulated osteichthyan reveals mosaic gnathostome characters. Nature 458:469–474. doi:10.1038/nature07855 Zhuravlev AV, Sokiran EV, Evdokimova IO, Dorofeeva LA, Rusetskaya GA, Malkowski K (2006) Faunal and facies changes at the EarlyMiddle Frasnian boundary in the north-western East European Platform. Acta Palaeontol Pol 51:747–758