, Volume 101, Issue 2, pp 143–148

Earliest and first Northern Hemispheric hoatzin fossils substantiate Old World origin of a “Neotropic endemic”


    • Senckenberg Research Institute and Natural History Museum Frankfurt, Ornithological Section
  • Vanesa L. De Pietri
    • Senckenberg Research Institute and Natural History Museum Frankfurt, Ornithological Section
Original Paper

DOI: 10.1007/s00114-014-1144-8

Cite this article as:
Mayr, G. & De Pietri, V.L. Naturwissenschaften (2014) 101: 143. doi:10.1007/s00114-014-1144-8


The recent identification of hoatzins (Opisthocomiformes) in the Miocene of Africa showed part of the evolution of these birds, which are now only found in South America, to have taken place outside the Neotropic region. Here, we describe a new fossil species from the late Eocene of France, which constitutes the earliest fossil record of hoatzins and the first one from the Northern Hemisphere. Protoazin parisiensis gen. et sp. nov. is more closely related to South American Opisthocomiformes than the African taxon Namibiavis and substantiates an Old World origin of hoatzins, as well as a relictual distribution of the single extant species. Although recognition of hoatzins in Europe may challenge their presumed transatlantic dispersal, there are still no North American fossils in support of an alternative, Northern Hemispheric, dispersal route. In addition to Opisthocomiformes, other avian taxa are known from the Cenozoic of Europe, the extant representatives of which are only found in South America. Recognition of hoatzins in the early Cenozoic of Europe is of particular significance because Opisthocomiformes have a fossil record in sub-Saharan Africa, which supports the hypothesis that extinction of at least some of these “South American” groups outside the Neotropic region was not primarily due to climatic factors.




The South American Hoatzin, Opisthocomus hoazin, is an unusual folivorous bird, which is characterised by a unique foregut fermentation and nestlings with well-developed wing claws (Thomas 1996). The species is the single extant representative of the taxon Opisthocomiformes, and its phylogenetic affinities are not well resolved (Ericson et al. 2006; Hackett et al. 2008; McCormack et al. 2013). For a long time, the only fossil record was a fragmentary cranium from the middle Miocene of Colombia (Miller 1953), with earlier, Eocene, New World fossils linked to hoatzin origins being of uncertain affinities and either based on fragmentary undiagnostic material (Cracraft 1971) or morphologically very different from extant Opisthocomiformes (Olson 1992).

In recent years, however, more light has been shed on the evolutionary history of hoatzins. From the late Oligocene/early Miocene (22–24 million years ago [mya]) of Brazil, the up to now earliest unambiguous record was described as Hoazinavis lacustris (Mayr et al. 2011). This species is represented by wing and pectoral girdle bones, which closely resemble the corresponding skeletal elements of the modern Hoatzin. It is now also known that part of hoatzin evolution took place outside the Neotropic region, with hoatzin remains having been identified in the Miocene of Africa. These African fossils are Namibiavis senutae from the late early Miocene (17–17.5 mya) of Namibia (Mayr et al. 2011), which is also based on forelimb elements, as well as a tarsometatarsus from the middle Miocene (15 mya) of Kenya, which was tentatively assigned to Namibiavis (Mayr 2014).

A phylogenetic analysis showed Namibiavis to be outside a clade including the South American Hoazinavis and Opisthocomus. Based on the presumed absence of Northern Hemispheric hoatzin fossils and the predominant palaeowinds and -currents, a transatlantic dispersal of hoatzins from Africa to South America, possibly by oceanic rafting, was considered most likely (Mayr et al. 2011). With only a single stem group representative known from the Old World and this being younger than the earliest South American fossils, the centre of origin of hoatzins nevertheless remained elusive.

Here we describe hoatzin fossils from the late Eocene (about 34 mya) of Romainville near Paris, which remained unstudied since their discovery more than a century ago. The only previously described avian remains from this locality belong to an extinct anseriform taxon and a bird of indeterminate affinities, originally described as a heron (Brunet 1970; Mourer-Chauviré 1996; Mayr 2008). The new specimens come from the white marl (“marnes blanches supragypseuses”), which was deposited in a lacustrine palaeoenvironment (Thiry 1989). They are significantly older than all unambiguously identified previous hoatzin remains and constitute the first Northern Hemispheric record of Opisthocomiformes, as such providing critical new data for an understanding of early hoatzin evolution and the centre of origin of the group.

Material and methods

The fossils are deposited in Naturhistorisches Museum Basel, Switzerland (NMB), Museu de História Natural de Taubaté, Brazil (MHNT), and Geological Survey, Ministry of Mines and Energy, Windhoek, Namibia (GSMME).

To test our phylogenetic placement, we added scorings for the new species to the character matrix used for the assessment of the affinities of Namibiavis (Mayr et al. 2011). Of the 151 characters, the following four could be coded for the new fossil: 65:1 (foramen nervi supracoracoidei absent), 66:1 (pneumatic foramen below scapular cotyla present), 67:2 (dorsal surface of sternal extremity with large pneumatic opening), and 68:0 (coracoid and scapula not fused). The analysis was performed with the heuristic search modus of NONA 2.0 (Goloboff 1993) with all settings as in Mayr et al. (2011).

Systematic palaeontology

Aves Linnaeus, 1758

Opisthocomiformes Sclater, 1880

Opisthocomidae Swainson, 1837

Protoazin parisiensis gen. et sp. nov.

Holotype. Partial left coracoid, extremitas cranialis of left scapula, and distal portion of pedal phalanx (NMB PG.70); all from the same small block of matrix, found in June 1912 (Fig. 1).
Fig. 1

aiP. parisiensis gen. et sp. nov. from the late Eocene of Romainville in France (NMB PG.70), specimen before preparation to show the long procoracoid process (a), left coracoid in dorsal (b), medial (c), and ventral (d) view, left scapula in lateral (e) and medial (f) view, distal end of pedal phalanx in plantar (g), dorsal (h), and distal (i) view (latter not to scale). j, kN. senutae from the late early Miocene of Namibia (GSMME AD 210’98), right coracoid (reversed, from Mayr et al. 2011) in dorsal (j) and ventral (k) view. ln Extant O. hoazin, left coracoid in dorsal (l) and ventral (m) view, left scapula in lateral view (n). oqHoazinoides lacustris from the late Oligocene/early Miocene of Brazil (MHNT-VT 5332), right coracoid (reversed, from Mayr et al. 2011) in dorsal (o) and ventral (p) view, right scapula (reversed, from Mayr et al. 2011) in lateral view (q). Abbreviations: acr acromion, csc scapular cotyla, for pneumatic opening in sternal extremity of coracoid, fur furcula, iml intermuscular line, pnf pneumatic foramen in coracoid shaft, prp procoracoid process, rdg ridge along medial margin of shaft, sca scapula, ste sternum

Differential diagnosis. Slightly larger than the extant O. hoazin, characterised by a mediolaterally wide coracoid shaft, which is here considered autapomorphic; further differs from Namibiavis in the presence of a pneumatic opening below the scapular cotyla of the coracoid and from Hoazinavis and Opisthocomus in the much shorter acromion of the scapula.

Type locality and horizon. Romainville near Paris, France, white marl (“marnes blanches supragypseuses”); late Eocene, Priabonian (European biostratigraphic reference level MP 20; Mourer-Chauviré 1996).

Etymology. From proto (Gr.): first and hoazin; the species epithet refers to the type locality near Paris.

Measurements (in mm). Coracoid, maximum length as preserved, 45.6; estimated total length, ∼48–50; mediolateral width of scapular cotyla, 5.4. Scapula, maximum length as preserved, 26.7.

Description and comparison. The coracoid lacks the tip of the acrocoracoid process and part of the sternal end; the originally preserved procoracoid process (Fig. 1a) was destroyed during preparation of the fossil and is now missing. As in Namibiavis, Hoazinavis, and Opisthocomus but unlike most other birds, the scapular cotyla is medially elongated and extends onto the procoracoid process. This latter process was very long and strap-like and oriented perpendicular to the shaft as in other hoatzins. As in Opisthocomus and Hoazinavis but unlike in Namibiavis, there is a pneumatic foramen below the scapular cotyla, which opens into the shaft of the bone and is smaller than in Opisthocomus. The shaft itself is mediolaterally wider than in Namibiavis and Opisthocomus, but as in Opisthocomus its dorsal surface is flat. On the ventral surface of the bone, there is a low ridge along the medial margin of the shaft (Fig. 1b), and the medial portion of the sternal extremity has an irregular dorsal surface. As in other Opisthocomiformes, the dorsal surface of the sternal extremity exhibits a very large pneumatic opening, and although the sternal end of the bone is not completely preserved, the omal (upper) margin of this opening is well delimited. In its development, this opening corresponds to that of Namibiavis, whereas it is partially covered by an osseous sheet in Opisthocomus. Unlike in Namibiavis and Opisthocomus, there is no intermuscular line separating the supracoracoideus and coracobrachialis caudalis muscles in the lateral part of the ventral surface of the sternal extremity. In the extant Opisthocomus, the coracoid is solidly fused with the furcula and sternum (Fig. 1l, m). The ends of the Protoazin coracoid are not preserved, but fusion with the furcula is absent in Namibiavis and Hoazinavis, and by phylogenetic bracketing it can also be assumed to have been absent in Protoazin (the sternal end of the coracoid is unknown for Hoazinavis, but in Namibiavis this bone was not fused with the sternum).

The acromion of the scapula is angled laterally and much less elongated than in Hoazinavis and Opisthocomus (Fig. 1e, n, q); unlike in the latter taxon there is no pneumatic opening at its base (the greatly elongated acromion of Opisthocomus, which adjoins the procoracoid process of the coracoid and thus strengthens the triosseal canal, is likely to be correlated with the general rigidity of the pectoral girdle elements). The preserved portions of the Protoazin scapula indicate that it was bowed as in Opisthocomus.

The fragmentary pedal phalanx fits in size to the coracoid and scapula. Therefore and because avian remains are rare in the Romainville fossil locality, it is likely that it belongs to the same individual as the pectoral girdle bones. In distal view, the bone is, however, dorsoventrally deeper than the pedal phalanges of Opisthocomus, from which it further differs in the markedly asymmetric size of the two halves of the trochlea (Fig. 1i).


Assignment of Protoazin to Opisthocomiformes is supported by our phylogenetic analysis, which places Namibiavis outside a clade including Protoazin, Hoazinavis, and Opisthocomus, but does not resolve the interrelationships among the latter three taxa (Fig. 2a). Protoazin can be unambiguously assigned to Opisthocomiformes because its coracoid shares characteristic derived features with that of Namibiavis, Hoazinavis, and Opisthocomus, including a large opening in the dorsal surface of the sternal extremity, a long procoracoid process oriented perpendicular to the shaft of the bone (now broken, but see Fig. 1a), a medially elongated scapular cotyla that extends onto the base of the procoracoid process, and a pneumatic foramen below that latter process. Placement of the new species outside a clade including the South American taxa Hoazinavis and Opisthocomus is supported by the plesiomorphic morphology of the scapula, which lacks an elongated acromion (Fig. 2b; this character was not included in the character matrix, because acromion length varies continuously within birds, and any subdivisions for character scoring would be arbitrary).
Fig. 2

Phylogenetic relationships of fossil hoatzins. a Strict consensus tree of seven most parsimonious trees resulting from the phylogenetic analysis (length = 707, CI = 0.23, RI = 0.47) with the opisthocomiform clade highlighted in grey; bootstrap support values are indicated below the nodes, unsupported nodes were collapsed. b Preferred hypothesis based on the features discussed in the text, with the temporal occurrence of fossil taxa indicated by black squares. The question mark for Namibiavis denotes a tentative taxonomic referral to that particular taxon, whereas the question mark for Hoazinavis indicates an uncertain stratigraphic age. The dotted blueline marks the Grande Coupure in Europe, the dashed internodes indicate that divergence dates are hypothetical

P. parisiensis constitutes the earliest fossil record of Opisthocomiformes, predating the Brazilian H. lacustris by 10–12 Ma and the Namibian N. senutae by 16.5–17 Ma (Fig. 2). From a temporal perspective, the late Eocene Protoazin may thus be assumed to occupy a more basal phylogenetic position than the much younger, early Miocene, Namibiavis. However, Protoazin differs from Namibiavis and agrees with Hoazinavis and Opisthocomus in the presence of a large pneumatic opening below the procoracoid process (Fig. 1b). This feature is otherwise only found in the palaeognathous Tinamidae, as well as in few Otididae (Otis) and Cuculidae (Coua), and its occurrence in Opisthocomiformes is clearly derived. It is more parsimonious to consider the absence of the foramen in Namibiavis to be plesiomorphic rather than a secondary loss, which would require an additional step to explain character distribution; accordingly, a pneumatic foramen was optimised as a synapomorphy of Protoazin, Hoazinavis, and Opisthocomus in the analysis.

The occurrence of two early diverging lineages of hoatzins in Europe and Africa substantiates an Old World origin of Opisthocomiformes and a relictual extant distribution for the group in South America (Mayr et al. 2011). Extant Hoatzins live in riparian lowland forests, and all fossil Opisthocomiformes likewise stem from lacustrine or fluviatile deposits (Mayr et al. 2011; Mayr 2014), indicating similar habitat preferences for the stem group representatives. Although recognition of hoatzins in the late Eocene of France is unexpected, the European avian record from this period is sparse and there are few fossil deposits that stem from suitable palaeohabitats. This situation sharply contrasts with the well-studied and comprehensive early and middle Miocene avian fossil record of Europe, which is dominated by fossil sites that originated in a freshwater environment, but notably did not yield any hoatzin remains (Mlíkovský 2002). Most likely, thus, hoatzins disappeared from Europe in the Oligocene, and the evidence at hand suggests that their extinction in Europe predates that in Africa, where Opisthocomiformes persisted until the middle Miocene (15 mya; Mayr 2014).

Recognition of stem group Opisthocomiformes in the Miocene of sub-Saharan Africa renders it unlikely that their extinction outside the New World was primarily driven by climatic changes during the mid-Cenozoic. Not only are the Miocene African hoatzin localities on comparable latitudes to the extant distribution area of these birds (Fig. 3), but the extinction of hoatzins in Europe also appears to have been before the onset of climatic cooling after the Miocene Climatic Optimum (Böhme 2003). Middle Miocene climatic differences between sub-Saharan Africa and Central Europe were further less pronounced than they are today (Böhme 2003; Krapp and Jungclaus 2011), and European avifaunas of that epoch still included frugivorous avian groups with poor migration capabilities, which are today only found in tropical or subtropical regions, such as Musophagidae (turacos), Trogonidae (trogons), and Psittaciformes (parrots) (Mlíkovský 2002; Mayr 2011).
Fig. 3

Distribution of localities, which yielded hoatzin fossil on a palaeomap of the continents in the earliest Oligocene (35 mya; copyright Ron Blakey, Colorado Plateau Geosystems, Inc.): 1 Romainville, France (late Eocene); 2 Arrisdrift, Namibia (late early Miocene); 3 Maboko, Kenya (middle Miocene); 4 Taubaté Basin, Brazil (late Oligocene/early Miocene); 5 Villavieja, Colombia (middle Miocene); 6 distribution area of extant Opisthocomus. Two possible dispersal routes, via North America and across the Atlantic, are indicated

In addition to Opisthocomiformes, there were several other avian taxa in the Palaeogene of Europe whose extant representatives are confined to the Neotropics, such as Trochilidae (hummingbirds), Nyctibiidae (potoos), and Cariamiformes (seriemas) (Mayr 2009). It has been hypothesised that their extinction in Europe was due to immigration of new faunal elements from Asia after closure of the Turgai Strait between the two continents (Mayr 2011), which marked a turnover in the European mammalian faunas at the Eocene/Oligocene boundary known as the Grande Coupure (Hooker 2010). Occurrence of hoatzins just before that event conforms to this hypothesis.

Hoatzins are particularly poor fliers and suffer from high predation rates at their open nests, in response to which the nestlings evolved sophisticated escape strategies (Thomas 1996). Despite some morphological differences, the known bones of stem group Opisthocomiformes exhibit an essentially modern-type morphology. If these birds also had limited flight capabilities and open nests like their extant relatives, they may have been prone to new predators that immigrated into Europe during the Grande Coupure. Especially in insular ecosystems, introduced predators, mainly cats and rats, are known to have a great impact on native bird populations (Kawakami and Higuchi 2010; Hilton and Cuthbert 2010). Arboreal carnivorans, i.e. Felidae (cats) and Viverridae (civets, genets, and allies), first appear in Europe after the Grande Coupure (Rose 2006), whereas the earliest African records of both groups are from the early/middle Miocene (Morales et al. 2001; Werdelin 2011; Werdelin and Peigné 2010). Whether a correlation between the arrival of these or other mammalian taxa and hoatzin extinction indeed existed can nevertheless only be revealed by a better stratigraphic refinement based on future finds. In present-day South America, where placental carnivorans did not arrive before the Great American Interchange in the Pliocene (Marshall et al. 1982), nest predation is mainly by Cebus monkeys but has also been noted for arboreal carnivorans; predation on adult hoatzins appears to be rare and predominantly by accipitrid birds (Strahl 1988).

Recognition in the Palaeogene of Europe of Opisthocomiformes more closely related to the South American species may indicate that hoatzins dispersed into South America through a Northern Hemispheric route (Fig. 3). In the absence of North American opisthocomiform fossils, however, a transatlantic dispersal, which is also considered most likely for caviomorph rodents and platyrrhine primates (Poux et al. 2006; Antoine et al. 2012), remains the most plausible hypothesis to explain colonization of the New World by hoatzins.


We thank L. Costeur (NMB) for the loan of the fossils, M. Weick (NMB) for their preparation, and S. Tränkner (Senckenberg Research Institute) for taking the photographs of the Protoazin and Opisthocomus bones. We further thank three reviewers for comments, which improved the manuscript.

Copyright information

© Springer-Verlag Berlin Heidelberg 2014