A review of amber and copal occurrences in Africa and their paleontological significance

Fossilized resins represent important sources of paleontological data for reconstructing ancient terrestrial ecosystems. A single fossiliferous deposit can provide abundant, various, and often exquisitely preserved organic inclusions (mostly arthropods, plants, protists, and fungi; more rarely microalgae, molluscs, and vertebrate remains). Fresh liquid resins solidify and progressively transform into fossil resins during burial through the chemical process of polymerization, during which their volatile compounds are lost (Anderson and Crelling, 1995; Clifford and Hatcher, 1995; Seyfullah et al., 2018). Amber is a fully polymerized, cross-linked resin that can at most softened using organic solvents. A resin that is incompletely cross-linked and not fully polymerized can dissolve in various solvents and is sometime called copal, although the use of this term is ambiguous (Anderson and Crelling, 1995; Langenheim, 2003; Vávra, 2009; Lambert et al., 2012; Penney, 2016; Delclòs et al., 2020). The time needed for full polymerization greatly varies depending on the resin composition and the burial conditions, so that there is no temporal feature to characterize copal and amber. Commonly, however, copals are considered as young, semifossilized resins ranging from Recent up to the Pliocene age (5.3 Ma) (Poinar, 1992; Grimaldi, 1996; Delclòs et al., 2020). But there are older resins that are only partially polymerized and can readily dissolve, or that remain sticky and aromatic when cut or polished, and should therefore be copals. Nonetheless, these are generally classified as ambers because of their antiquity (e.g., Eocene Cambay “amber” from India, or Miocene “amber” from Indonesia; see Rust et al., 2010; Mazur et al., 2014; Narudeesombat et al., 2014). In the present review of all known fossil resins from Africa, “copal” and “amber” are used in this temporal sense rather than chemical sense, i.e. regardless of their level of polymerization.

The vast majority of amber deposits known today occurs in the Northern Hemisphere, particularly from the former Laurasian continent. The paucity of Gondwanan amber deposits may results from a long lack of prospection for this geological material in the Southern Hemisphere, combined with typically limited outcrops due to the dense plant covering in intertropical zones. Amber deposits of Gondwanan origin comprise Indomalayan occurrences in India (early Eocene; Rust et al., 2010) and possibly Myanmar (early Cenomanian; at the time probably an island in the Tethys Ocean but originating from Austral Gondwana, see Rasnitsyn and Öhm-Kühnle, 2018, Poinar, 2019 or Westerweel et al., 2019 for arguments); Australasian occurrences in Australia (Triassic to Neogene; Hand et al., 2010; Sonibare et al., 2014; Stilwell et al., 2020) and New Zealand (Cretaceous, Oligocene and Miocene; Schmidt et al., 2018; Mays et al., 2019; Stilwell et al., 2020); South American occurrences in Brazil (early Cretaceous; Pereira et al., 2011), Ecuador (Aptian–Albian; Cadena et al., 2018), Venezuela (Miocene, Pérez et al., 2016), and Peru (Miocene; Antoine et al., 2006); and rare African occurrences that are detailed and reviewed hereafter along with the known copal deposits.

Copal had once a major export value for its industrial use in varnishes and lacquers. A rich literature exists on African copals, their botany and their trade that developed during colonialism in the nineteenth and twentieth centuries. Abundant copal resources have been exploited mostly from two regions of the continent: Zanzibar and the German East Africa, that became a British territory after WWI, and are now corresponding to the modern Tanzania (Kirk, 1871; Schlüter and von Gnielinski, 1987; Sunseri, 2007); and the Democratic Republic of the Congo (DRC, formerly Congo Free State, then Belgian Congo) (Pynaert, 1924; De Wildeman, 1927; Vervloet, 1932; Hellinckx, 1935; Anonymous, 1942; Billing, 1944). Copal is more generally known from several countries in East Africa and West Africa (Fig. 1). All African copals derive from legume trees in the family Fabaceae, mostly Hymenaea or Guibourtia in the Caesalpinioideae group, but also Colophospermum, Copaifera, Daniellia or Tessmannia in the Detarioideae group (Howes, 1949; Langenheim, 2003; McCoy et al., 2017; Delclòs et al., 2020).

Pleistocene or even younger copal has been reported from various countries along the African East coast: Kenya, Tanzania, Mozambique, Madagascar (Fig. 1). Copal was traded into Indian Ocean networks, including India, by Swahili communities as soon as in the eighth century AD, but a rise in the trade came with the arrival of European and American ships in Zanzibar between the sixteenth and nineteenth centuries (Hedley Barry, 1932; Howes, 1949; Sunseri, 2007; Delclòs et al., 2020). Most of the copal traded on the Zanzibar market, however, originated from the coastal forests of mainland Tanzania. This brings confusion as to whether the bioinclusions described from the so-called Zanzibar copal are actually from there or from Tanzania (Schlüter and von Gnielinski, 1987; Stroiński, 2007). Here we use Tanzanian copal in its broad sense including deposits from coastal hinterlands of Tanzania as well as the Zanzibar archipelago. The trade of East African copal, particularly that of Madagascar, has been recently reviewed by Delclòs et al. (2020).

During the eighteenth and nineteenth centuries when the trade of Tanzanian copal was most intense, interest was almost exclusively industrial or botanical (Gilg, 1898; Zimmerman, 1908). The (sub)fossil content was briefly mentioned or studied only in rare instances (Dalman, 1825; Kirk, 1871; Quedenfeldt, 1885; De Saussure, 1890). The true paleontological interest for East African copal bioinclusions started in the twentieth century, with the pioneering studies of Meunier (> 20 papers between 1900 and 1917; e.g., Meunier, 1900a, 1900b, 1905, 1917) who mostly focused on Diptera and Hymenoptera from Tanzanian and Madagascar copals. Tanzanian copal has also yielded arthropods in the orders Araneae, Scorpiones, Embiodea, Psocoptera, Blattodea, Isoptera, Lepidoptera, Auchenorrhyncha, Heteroptera, and Coleoptera (Bervoets, 1909a; Rosen, 1913; Schlüter and von Gnielinski, 1987; Stroiński and Szwedo, 2002, 2011; Stroiński 2007; Grichanov, 2008). Copals from Madagascar, and in a lesser extent from Kenya and Mozambique, have yielded various Araneae, Pseudoscorpiones, Scorpiones, Heteroptera, Neuroptera, Hymenoptera, and Coleoptera (Quedenfeldt, 1885; De Saussure, 1890; Bervoets, 1909b; Meunier, 1910a; Krinsky, 1985; Schlüter and von Gnielinski, 1987; Lourenço, 1996; Stroiński and Szwedo, 2002; Bosselaers, 2004; Wunderlich, 2004, 2008; Selden et al., 2009; Bosselaers et al., 2010; Azevedo et al., 2010; Guilbert and Heiss, 2016; Bukejs and Legalov, 2017). Finally, some insects have been reported or described without further details on their origin than “East African copal” (e.g., bees; Zeuner and Manning, 1976). In fact, most of these studies lack data on the geographical provenance and age of the copal pieces, so that the scientific value of these bioinclusions is depreciated. In Madagascar copal, for instance, many bioinclusions have been described as probably “recently extinct, subfossil”, but several are species still living in African forests today, and recent studies have shown the Madagascar copal to be a Recent resin, up to a few hundred years old only (Delclòs et al., 2020).

All East African copals are assumed to derive from Hymenaea trees based on chemical characterization using gas-chromatography (GC-MS), nuclear magnetic resonance (¹³C NMR), or Fourier transform infrared spectroscopy (FTIR) (Lambert et al., 1995, 2012; McCoy et al., 2017; Delclòs et al., 2020); and more specifically Hymenaea verrucosa, the sole species found in East Africa and Madagascar, based on botanical considerations (Kirk, 1871; Gilg, 1898; Zimmerman, 1908; Howes, 1949; Langenheim, 1990, 2003; Delclòs et al., 2020).

Copals have been reported from the following countries in Western or Central Africa (Fig. 1): Angola (“Benguela copal”), Congo / DRC / Cameroon / Gabon (generally grouped under the term “Congo copal”), Liberia / Benin / Togo / Ghana / Nigeria (sometimes grouped under “Accra copal”), Sierra Leone, and Guinea (Howes, 1949; Mantell, 1950). All are estimated Pleistocene or more recent in age, although the precise stratigraphic provenance is unknown. From these, only the Congo copal (Fig. 2A) has been exploited in great amount, particularly during Belgian colonialism in the Interwar period when it surpassed the trade of Tanzanian copal (Howes, 1949). Reports from the nineteenth century were thus essentially focused on the plant sources, collecting methods, chemical properties or commercial value of Congo copal (Rackwitz, 1907; Pynaert, 1924; De Wildeman, 1927, 1933; Vervloet, 1932; Aubréville, 1933; Louis, 1933; Mertens, 1933; Hellinckx, 1935; Anonymous, 1942; Billing, 1944; Howes, 1949; Léonard, 1950). Other West African copals have been comparatively less studied.

According to botanical observations in the above references, the legume tree Guibourtia (with species formerly assigned to the genus Copaifera) is considered to have produced most of the Congo and other West and Central African copals. But some Congo copal is also derived from Tessmannia, and copal from Angola is thought to originate from Colophospermum. Copaifera or Daniellia are also considered as important producers of copals from Benin, Ghana and Togo (Léonard, 1950; Langenheim, 2003). Resins of all these genera and Hymenaea are relatively similar in their molecular composition and are thus difficult to distinguish chemically (Lambert et al., 2002, 2009).

Only few arthropod inclusions have been reported in West African copals, most of them described in the early twentieth century: Quedenfeldt (1885) provided the earliest account from Benguela copal of Angola, with the description of two beetles in the families Elateridae and Chrysomelidae; diverse Diptera (Cecidomyiidae, Chloropidae, Muscidae, Phoridae, Psychodidae, Sciaridae) and one Neuroptera Coniopterygidae were also described from Togo and Accra copals by Meunier (1906, 1909, 1910a, 1910b, 1910c, 1910d); and a few Coleoptera and Isoptera were described from copals of Benin, Ghana, Guinea, or Sierra Leone, by Hagedorn (1905–07) and Rosen (1913), respectively. Additionally, Handlirsch (1906–08) mentioned some undescribed insects in the orders Isoptera, Blattodea, Orthoptera, Thysanoptera, Coleoptera, Hymenoptera, Lepidoptera, and Diptera, in various copal pieces from West Africa (labelled “Benin” and “Guinea”; see Schlüter and von Gnielinski, 1987). It also seems that a platypodid beetle, originally described from Baltic amber by Nunberg (1959), was in fact from a piece of West African copal (Schlüter and von Gnielinski, 1987). Finally, several Mymaridae were described from copal of unspecified localities by Meunier (1900a, 1900b), who otherwise apparently accessed East and West African copals only, so these mymarids likely originated from African copals as well.

Amber has been comparatively less frequently encountered than copal in Africa (Fig. 1). The rare occurrences until now have mostly consisted in small deposits yielding only few amber pieces apparently devoid of organic inclusions. Recently, however, the first two fossiliferous African ambers have been discovered, in Congo and Ethiopia (Schmidt et al., 2010a, 2010b; Perrichot et al., 2016, 2018). The study of these deposits and their fossil content is still in a nascent stage, so that only a preliminary review is given herein.

The oldest fossilized resin known from Africa dates back to the Upper Triassic, in the Molteno Formation, and consists of few millimeter-sized amber drops found fossilized within gymnosperm cones (Ansorge, 2007). The seed-producing gymnosperm was evidently the plant source of the resin, but its precise affinity remains unknown and a chemical characterization of such a tiny amber sample remains challenging. The tiny size and preservation within seeds has also prevented the finding of any organismal inclusions within this material.

Numerous small (max. 7 mm) pieces of yellowish to red amber have been found in the Middle–Upper Valanginian (Lower Cretaceous) Kirkwood Formation (Gomez et al., 2002a, 2002b). The amber-bearing rock also yielded abundant remains of Brachyphyllum, a conifer genus of ill-defined position within the Araucariaceae or Cheirolepidiaceae, and possibly the resin-producer. This amber is devoid of inclusion. Paleobotanical and taphonomical studies suggest that the resin was produced under a warm to hot, semi-arid climate, and was washed from its primary environment into a fluvial deposit during a flooding event (Gomez et al., 2002a).

The only fossiliferous Cretaceous amber from the continent has been discovered in 2011 near Doumanga, in the Mayombe belt, 70 km NE of Pointe-Noire. Amber and plant debris, including lignite and fusinite, have accumulated in lags within the Vembo Shales, a member of the Chéla Formation that was dated as middle Aptian based on the palynomorphs and ostracods from the amber-bearing shales (Delhaye-Prat et al., 2016; Perrichot et al., 2016). An excavation by one of us (VP) in 2013 yielded about 5000 amber pieces, all dark, red to orange in color (Figs. 2B and 2C), commonly pebble-shaped, and up to 12 cm in diameter. Analysis of small, translucent pieces in light microscopy revealed the first arthropods, while synchrotron imaging has been used to radiograph a first portion of dark, larger pieces. The studied sample appears moderately fossiliferous, with a dozen families identified in 9 arthropod orders (Tab. 1; Figs. 3A and 3B). Plant debris (stellate hairs, wood fibers, spores) or fungal/bacterial remains also occur in many pieces.

The molecular composition of Congolese amber indicates a Group A in ¹³C-NMR (sensu Lambert and Poinar, 2002) and Class Ib in THM-GC-MS (sensu Anderson, 1994), indicative of resins from the conifer families Araucariaceae, Cupressaceae, Podocarpaceae, and Cheirolepidiaceae. The wood remains associated in the amber-bearing rock belong to Brachyoxylon and Agathoxylon, and the palynomorphs comprise Araucariacites and dominant Classopollis, suggesting an Araucariaceae or Cheirolepidiaceae as the plant source. Amber and plant debris were washed into a saline, lacustrine environment in a small, inland, rift basin connected to the ocean, and accumulated in flood related gravity driven deposits, as indicated by sedimentological and taphonomical studies. In addition with paleobotanical data, it also suggests a hot paleoclimate with contrasted, humid and dry seasons (Delhaye-Prat et al., 2016).

Small, isolated pieces of reddish amber have been collected in 2016 at Takalmaoua, Department of Dakoro, in Maradi Region. The amber was found in a well, at a depth of about 50 m, within rocks of Cenomanian/Senonian age (A. Moumouni, personal communication). A chemical characterization of this amber by thermochemolysis gas-chromatography-mass spectrometry (THM-GC-MS) has indicated a Class Ib amber (sensuAnderson, 1994), thus originating from a conifer (Perrichot, unpublished data). The few available pieces did not provide any inclusion.

A Lutetian or Lower Bartonian (Eocene) amber, named Amekit, has been found in the Ameki Formation near Umuahia, in southeastern Nigeria (Wilson, 1925; Arua, 1979; Savkevitch and Arua, 1990). Amber chunks occur within three successive sedimentary facies of this formation, and are generally associated with lignite and various marine organisms, suggesting they were washed into a marginal lagoon environment from a nearby mangrove-swamp area (Arua, 1979, 1986). An angiosperm source has been suggested for Amekit based on the absence of gymnosperms among the fossil plants from the amber-bearing strata (Rao and Kumaran, 1988). But the molecular characterization of the amber by infrared spectroscopy and gas chromatography-mass spectrometry (Py- and THM-GC-MS) indicates a Class Ib amber typical of conifers, possibly the family Cupressaceae (Sonibare et al., 2012). No organismal inclusions has ever been mentioned from this amber despite large pieces of yellow or dark red, translucent amber (Arua, 1979). However, there has been virtually no material available for screening by paleoentomologists since the formal description of this amber four decades ago.

It is possible that an older amber also exists in southern Nigeria, as mentioned by Le Gall et al. (2010). Indeed, some pieces were found associated with mosasaur teeth and ammonites in the Upper Cretaceous (Maastrichtian) strata that outcrop between Enugu and Port Harcourt, thus in the surroundings of the Ameki Formation (Ph. Courville, personal communication). Whether this corresponds to a true Cretaceous, primary deposition, or a redeposition after the erosion of above layers would require further investigation. But fieldwork in Nigeria is currently hampered by the geopolitical situation.

The richest sources of amber in Africa are outcrops located in the gorges of the Wenchit, Jemma, and Mugher rivers incising the northwestern Ethiopian plateau in the North Shewa Zone of the Amhara Region. At least 20 kilograms of amber are extracted annually by villagers, apparently with all excavations in a same siltstone that was erroneously attributed to a Cretaceous unit, the Debre Libanos Sandstone (Schmidt et al., 2010a; Kiefert, 2015). The Cretaceous age of the rock was suggested by local geologists based on field observations and correspondence with the geological map of Belay et al. (2009); it was also estimated based on the analysis of the chemical features of the amber and the sporomorphs of the associated sediment (Schmidt et al., 2010a). However, two of the sporomorphs that had been identified as Cretaceous taxa were mistaken, and other sporomorphs that had been left undetermined have eventually been identified as long ranging Cenozoic taxa (Perrichot et al., 2016). Recent investigation of additional amber material has provided arthropod and plant inclusions that belong almost exclusively to extant genera (Tab. 1), therefore modern lineages which definitely rule out a Cretaceous age. Additionally, the rocks associated with this new material have yielded palynomorph assemblages that are indicative of an Early Miocene age (Perrichot et al., 2018). According to geological studies, the only sedimentary rocks known in this area are either Upper Cretaceous, referred to as Debre Libanos Sandstone or Upper Sandstone; or Upper Miocene, referred to as “Tertiary sediments” (Belay et al., 2009). Data are yet insufficient to explain this incongruence with the Early Miocene age suggested by our paleontological results.

Ethiopian amber is hard, translucent, with colors ranging from orange to yellow greenish, light green, or dark green (Figs. 2D–2F), and pieces are commonly 5 cm or more in size. Analysis of nine amber samples in THM-GC-MS has consistently recovered a Class Ic amber, regardless of the color, and comparison with the spectra obtained from Dominican and Mexican ambers suggests the angiosperm family Fabaceae as the plant-source (unpublished pers. data). The finding of fossil leaves and stamens of Hymenaea in several amber pieces similarly suggests this legume tree as the resin producer (J. Szwedo, personal communication).

The first study of Ethiopian amber had already recovered a diversity of arthropod and fungal inclusions (Schmidt et al., 2010a, 2010b). Newly investigated material confirms the richness of Ethiopian amber inclusions and undoubtedly shows taxonomic affinities with Cenozoic biota (Tab. 1; Figs. 3C and 3D). Remarkably, this amber yields a fairly high proportion of plant inclusions, among which frequent remains of liverworts (Marchantiophyta) and angiosperms (Fabales), more rarely leafy mosses (Hypnales) and lichens (Lecanorales). Most plant and insect taxa found as inclusions are reported for the first time from the African fossil record (e.g., Frullaniaceae, Lejeuneaceae, Pylaisiadelphaceae, Formicidae: Dorylinae, Pseudomyrmecinae, Mycetophylidae). And many taxa are the first fossil occurrence of extant genera, thus bridging a gap in the evolutionary history of the corresponding lineages. A particular focus has been made on ants, revealing at least 13 genera, 10 of which are still extant. Several of these contemporary genera have relatively recent crown-group origins, for instance with the following age estimates obtained from molecular phylogenetic analyses (“Mean” values, in Ma): Hypoponera (30-37) for Ponerinae (Schmidt, 2013); Ravavy (27) and Technomyrmex (23-33) for Dolichoderinae (Ward et al., 2010); Cataulacus (18.3-20.4) and Trichomyrmex (20-22.7) for Myrmicinae (Blaimer et al., 2018). These occurrences support a Miocene maximal age dating for Ethiopian amber.

Finally, the palynomorph assemblages are suggestive of a lowland tropical rainforest, without any marine input. The presence of abundant liverworts assignable to Lejeuneaceae similarly indicates a humid tropical forest. Ethiopian amber was thus derived from Hymenaea trees likely growing in evergreen rainforests, rather than in dry environments like some recent Hymenaea species (Langenheim et al., 1973).

Several evolutionary trends can be seen from the few African deposits of fossil resin (Fig. 4). Ancient, Cretaceous to Eocene resins were all derived from conifers (class Ib type resins) while younger, Miocene to Recent resins derive from angiosperms in the Fabaceae family (class Ic type resins). Also, Araucariaceae or Cheirolepidiaceae were apparently the dominant amber-producers during the Early and mid-Cretaceous (South-Africa, Valanginian; Congo, Aptian), possibly replaced by Cupressaceae in the Eocene (Nigeria). This is similar to the global pattern, observed from amber deposits worldwide, of a shift to dominance of resin-production by Cupressaceae or Pinaceae in the Late Cretaceous, and by (sub)tropical angiosperms in the mid-Cenozoic, with the exception of the kauri copal produced by Agathis (Langenheim, 2003; Nohra et al., 2015; Seyfullah et al., 2018).

This change in the resin-producing trees was accompanied by a change in climate and environment. For the Cretaceous, sedimentary and fossil data in South Africa and Congo indicate a warm, subtropical climate with at least temporary arid conditions. From the Eocene onward, ambers from Nigeria and Ethiopia, and copals from eastern and western-central Africa have derived from lowland, tropical evergreen rainforests with (Nigeria) or without (Ethiopia) marine influence. Further study of fossil inclusions of fungi, plants, and arthropods from Congolese and Ethiopian ambers will thus likely further substantiate such contrasted habitats.

We are grateful to all colleagues and persons who contributed to the discovery or study of Congolese amber: V. Delhaye-Prat and colleagues from Total Exploration–Production Congo (Pau, France), Anaïs Boura (UPMC, France), Johanne Esnault and Laurent Jeanneau (Univ. Rennes, France), Carmen Soriano (Argonne Photon Source, Chicago). We are also indebted to all persons who contributed the access or the study of Ethiopian amber: Brendon Boudinot (Univ. California, Davis), Cédric Chény (Univ. Rennes, France), Jim Cole (Tie-Point Geoscience Ltd., UK), Yale Goldman (Collinsville, Connecticut), Alexander R. Schmidt (Univ. Göttingen, Germany), Matthias Svojtka (Univ. Vienna, Austria), Jacek Szwedo and Blazej Bojarski (Univ. Gdańsk, Poland), Bo Wang (NIGPAS, China), Benyam Teferi (Addis Ababa, Ethiopia), and Amde Zewdalem (Jacksonville, Florida). We warmly thank Ali Moumouni (Univ. Maradi, Niger) and Philippe Courville (Univ. Rennes, France) for information or samples of amber from Niger and Nigeria, and SP Demonvielle (Brazzaville) for samples of Congolese copal. And we thank the guest editors Jean-Paul Saint Martin and Simona Saint Martin for inviting us to contribute this thematic volume, and Alexander Schmidt and an anonymous reviewer for their helpful comments. This work was partly supported by the Tellus-INTERRVIE program of CNRS INSU (project AMBRAFRICA to V.P.).