The Anti‐Atlas of Morocco is well known for its extensive Lower Palaeozoic outcrops. Silurian strata, including some beds of high latitude, cold‐water carbonates, are exposed in scattered outcrops forming a long, narrow belt in the NE Anti‐Atlas (Lubeseder 2008; Fig. 1). The Tafilalt region, located in the NE‐most part of the outcrop‐belt, is one of the main areas within the Anti‐Atlas where Silurian strata crop out. Here, the total thickness of the Silurian is 100–150 m. The Silurian succession is dominated by graptolitic shales, into which rare, thin beds of nautiloid limestone are intercalated. These beds are generally only a few to a few tens of centimetres thick, but the Ludlow Orthoceras‐limestone may reach 1 m or more in thickness. The shale is often poorly exposed and forms wide flat plains from which beds of limestone protrude as small ridges (Fig. 2B). Such ridges are very distinctive and can be traced for kilometres.

The upper Silurian carbonates, mainly represented by nautiloid limestones of outer shelf origin, were recently characterized in detail by Lubeseder (2008). Although Lubeseder (2008, figs 3, 4) also illustrated lower Silurian strata, he did not discuss them because of their less extensive outcrops, generally poor exposure and lack of stratigraphical precision.

In the studied region, the base of the Silurian is highly diachronous and records the step‐wise transgression of the Silurian sea following the Late Ordovician glaciation. Earliest Silurian deposits (the Rhuddanian Tizi Ambed Formation) are limited to a N–S trending palaeovalley some 20 km to the east of the study area (Destombes et al. 1985). In the surrounding palaeovalley regions late Aeronian to early Telychian graptolite shales of the Tamaghrout Formation directly overlie Ordovician sandstones (Fig. 2C). The hiatus between the Ordovician and Silurian strata is of variable duration (Lüning et al. 2000). The predominantly shaley Tamaghrout Formation ranges in thickness from 30 m in the north (at its type locality near Hamar Laghdad; Jaeger & Massa 1965) to 50 m near Taouz in the south of the Tafilalt. Locally, the remnant end‐Ordovician palaeo‐relief caused the Silurian sedimentation to start even later, in the Wenlock. In some places in these areas of positive palaeo‐landforms siltstone and fine sandstone intercalations occur in the basal part of the Tamaghrout Formation (Destombes et al. 1985). This type of succession is characterictic of the Bou‐Maiz syncline including also the studied section. Here, the thickness of the formation is commonly about 10 m or even less (Fig. 2A). This area is also known for its pronounced Middle to Late Devonian palaeo‐relief. In that time, the Maider and Tafilalt basins were separated by a swell (e.g. Wendt 1988; Lubeseder et al. 2010; Fig. 1B).

The data discussed in this paper come from a sample collected by SL from the Bou‐Maiz (El Kachla) section (31°08.281′ N; 04°16.731′ W) in the Tafilalt region of the NE Anti‐Atlas, Morocco (Fig. 1). In Lubeseder (2008, figs 1, 4) this section is referred to as section s6. The measured thickness of Silurian strata in this section is about 100 m. The sample represents the 2.5–3.0 cm thick, in section 11, up to 35 cm thick basal limestone bed of the Tamaghrout Formation (Fig. 2). This limestone was also mentioned (but not discussed and not included in their figure 56) by Destombes et al. (1985, p. 260). According to the geological map sheet of the Tafilalt region (Fetah et al. 1986), the limestone bed lies immediately above fossiliferous (brachiopods, trilobites, bryozoans) shoreface sandstones and flat‐pebble conglomerates of the Lower Ktaoua Formation (middle Caradoc) in section s6, and just above sandstones of the Upper Tiouririne Formation (upper Caradoc to early Ashgill) in section s11 (in modern nomenclature, middle Katian and upper Katian, respectively). Although these sandstones are generally assigned to the Ordovician (Destombes et al. 1985; Fetah et al. 1986), the uppermost few metres of them containing the flat‐pebble conglomerates and ravinement deposits are genetically linked to the Silurian transgression and thus most likely are also Silurian (Wenlock) in age.

The sample processed for conodonts comes from a strongly weathered, finely crystalline ferruginous limestone. As a result of weathering, the sample is yellowish‐brown to brown in colour. The sample is rich in silt‐size terrigenuous material. Sub‐rounded to rounded quartz grains of medium to coarse sand size are also relatively common. The sample yields abundant poorly preserved graptolites and fragments of large (diameter up to 20 mm) thin‐walled shells (bivalves?) are common (Fig. 3). Some small iron‐rich concretions up to 10 mm in diameter occur in the lower part of the sample (Fig. 3C). The matrix of the limestone consists of silty, Fe‐stained micrite. Graptolites appear to be aligned in one direction and the mud‐ to floatstone is interpreted to have been deposited in calm water, possibly below storm wave base, but under the influence of bottom currents.

The weight of the sample processed for conodonts was 1068.0 g. To extract the conodonts, the sample was processed applying traditional methods: buffered acetic and formic acids were used to dissolve the sample, and bromoform to separate light and heavy fractions. Conodonts were picked from the heavy fraction of the residue. In total, 113 identifiable conodont specimens were found. The conodonts are reasonably well preserved, but identification of many specimens is complicated due to adhered quartz grains, cemented to the surface of the specimens by iron oxides (Fig. 4). The conodont elements are black, some of them even closer to grey in colour (Colour Alteration Index at least five, probably between five and six) indicating diagenetic temperatures over 300°C.

The graptolites are preserved in three dimensions in calcite; no periderm is preserved. The weathering renders identification of the graptolites very difficult. On the polished surface of the sample and in thin section cross‐sections of graptolites can be recognised as circular to sub‐circular bodies. They are most abundant in the central part of the sample processed for conodonts (‘graptolite layer’ in Fig. 3A, C). In the thin section coming from a different sample from the same bed the graptolites are most abundant in its lowermost part (Fig. 3B).

The studied material (conodonts, graptolites and rock sample) is housed in the Institute of Geology at Tallinn University of Technology (collection GIT 598).

The conodont assemblage is dominated by Dapsilodus obliquicostatus (Branson et Mehl) (Figs 4O–R). In total 47 specimens were found. Also very common are Wurmiella excavata (Branson et Mehl) (Figs 4I–N; 34 specimens) and Pterospathodus pennatus procerus (Walliser) (Figs 4A–H; 27 specimens). Additionally, two specimens of Pseudooneotodus beckmanni Bischoff et Sannemann, one poorly preserved Pb element of Kockelella? sp. (Fig. 4S) and two probable elements of Oulodus? sp. occur.

Graptolites in the studied sample are poorly preserved as a result of weathering. However, Monograptus priodon (Bronn), Monoclimacis vomerina (Nicholson) and Monoclimacis geinitzi (Bouček) have been confidently identified. It is possible that other species with narrower and/or curved rhabdosomes are also present, but identification even to generic level is not possible.

The assemblage of conodonts recognized is characteristic of the later part of the Ireviken Event, particularly of distal graptolitic environments (in the Aizpute‐41 core section in the Baltic region; Loydell et al. 2003). The composition of the fauna, particularly the co‐occurrence of P. p. procerus and D. obliquicostatus, indicates that this sample comes from strata corresponding to the P. p. procerus Superzone sensuJeppsson (1997). Although in distal environments P. p. procerus occurs also in strata older than the Ireviken Event (Männik 1998), D. obliquicostatus sensu stricto is not known (or is extremely rare) below datum 3 of the event (below the lower boundary of the P. p. procerus Superzone; Jeppsson 1998; Loydell et al. 2003). Based on co‐occurrences of conodonts and graptolites in the Baltic region ( Loydell et al. 1998, 2003) the P. p. procerus conodont Superzone correlates with the topmost Cyrtograptus murchisoni graptolite Biozone, Monograptus firmus graptolite Biozone, and with the lowermost Monograptus riccartonensis graptolite Biozone (Männik 2007b). Accordingly, it is evident that the lowermost Tamaghrout Formation in the studied section is early Wenlock in age. This dating agrees with the graptolite data from the same sample.

All three of the graptolites identified in the studied sample are long ranging, with first appearances in the Telychian and last appearances at various levels within the Wenlock. Monoclimacis geinitzi is the most useful, as it has not been recorded above the M. firmus Biozone (the highest record is from Latvia: Loydell et al. 2003). Graptolite assemblages from the M. riccartonensis Biozone (and particularly the lower part of this biozone) are almost always heavily dominated by the eponymous species. On this basis, and considering also the graptolite‐conodont biozone correlations of Loydell et al. (2003), the sample is from either the upper C. murchisoni Biozone or M. firmus Biozone. This is consistent also with Destombes et al.’s (1985) records of upper Sheinwoodian and lower Homerian graptolites from higher levels in the Tamaghrout Formation of the Tafilalt region.

As noted above, the conodont assemblage in the studied sample yielded taxa characteristic of the later part of the Ireviken Event. All of the conodonts are known also from several other regions and suggest that there was no major difference in conodont faunas in northern Gondwana and, for example, in Baltica and Laurentia at this time. Although the number of species recovered from our sample is lower than that in the Aizpute‐41 core section in the Baltic region, for example (Loydell et al. 2003), in both regions the fauna is dominated by D. obliquicostatus, and W. excavata and P. p. procerus are common. More data from Morocco are needed to assess the possible palaeoecological significance of the lower conodont diversity in the studied sample. It may be that this can be explained simply by differences in sample size.

The sample studied comes from the P. p. procerus conodont Superzone, either from the upper C. murchisoni or from the M. firmus graptolite Biozone.

In the studied section, the base of the Tamaghrout Formation can now be confidently dated as of early Wenlock age and at least the upper part of the Ireviken Event is represented in the section.

At least during the late stage of the Ireviken Event there was no major difference in conodont faunas in northern Gondwana and, for example, in Baltica and Laurentia.

The authors are grateful to the anonymous referees for very valuable comments and suggestions. Mr. G. Baranov (IG TTU) prepared the photos of the lithological sample. The study by PM was supported financially by the target project SF 0140020s08 and by the Estonian Science Foundation (grant No 7138). Fieldwork by SL was sponsored through the North Africa Research Group (Manchester).