Australia: The Land Where Time Began

A biography of the Australian continent 

Possible Hominin Footprints Dating to the Late Miocene of Crete about 5.7 Ma

In this paper Gierliński et al. describe footprints (tracks) of a tetrapod from the Trachilos locality of western Crete (Greece), which display characteristics that are human-like. The footprints were discovered in an emergent environment in what was otherwise a marine succession from the Messinian age (latest Miocene) that has been dated to about 5.7 Ma, just before the Messinian Salinity Crisis. It is indicated by the tracks that the maker of the tracks had no claws, and was bipedal, plantigrade, pentadactyl and strongly entaxonic (the inner digits, the big toe/thumb, more strongly developed than the outer). The impression of the large and non-divergent first digit (hallux) has a narrow neck and a bulbous asymmetrical distal pad. The impressions of the lateral digit become progressively smaller with the result that the digital region as a whole is strongly asymmetrical. Associated with the hallux is a large, rounded impression. It is shown by morphometric analysis that the footprints have outlines distinct from extant non-hominin primates, instead resembling those of hominins. Gierliński et al. suggest the interpretation of these footprints to be potentially controversial. It is suggested by the morphology of the print that the track maker was a basal member of the clade Hominini, though as Crete is well outside the known geographical range of pre-Pleistocene hominins the possibility must be considered that the trackmaker must be a primate from the Late Miocene that hitherto had been unknown that evolved a human-like foot convergently.

Information about the presence of a trackmaker at a moment in space and time is provided by fossil tracks. It is only possible to infer a trackmaker from a trackway where there is sufficient distinct morphological data to make the link between the trace and the culprit. Production of a track involves the interplay between the shape/anatomy of the foot and the pattern of the loading, which is mediated through a compliant substrate that is sufficiently elastic to deform, though rigid enough to retain the impression. There are complex variables involved here and a range of tracks may be made by a single trackmaker (e.g., Brand, 1996; Bennett et al., 2104; Milner & Lockley, 2016). Detailed pedal anatomy of a trackmaker may not be known in many cases. Therefore it is not surprising that ichnologists (the study of traces of animals and plants) practice parataxonomy (the sorting of samples into recognisable taxonomic units) when they classify traces; such linkages can have controversial implications, especial with body fossils present from comparable locations and stratigraphic intervals (e.g. Stӧssel, 1995; Niedźwiedzki et al., 2010; Voigt and Ganzelewski, 2010; Brusatte et al., 2011; Lichtig et al., 2017). In this paper Gierliński et al. report an example of making such inferences when the implications run counter to conventional views on the evolution of humans: footprints that are hominin-like in Crete from the Late Miocene, that have been dated to at least5.6 Ma, so about 2 Myr older than the hominin trackways from Laetoli in Tanzania (Leakey & Hay, 1979; Leakey & Harris, 1987; White & Suwa, 1987; Deino, 2011).

Geological setting and age

At Trachilos, to the west of Kissamos Harbour in western Crete, the coastal rocks are located within the Platanos Basin, where there is a succession of shallow marine carbonates and siliciclastics of the Roka Formation, which is a local development of the Vrysses Group (Freudenthal, 1969; van Hinsbergen & Meulenkamp, 2006) that date to the Late Miocene. This marine succession terminates abruptly at the top in the coarse-grained terrigenous sedimentary rocks of the Hellenikon Group, which formed when the Mediterranean Basin desiccated during the Messinian Salinity Crisis (van Hinsbergen & Meulenkamp, 2006), an event that has been dated to about 5.6 Ma (Govers, 2009). An emergent horizon is contained in the succession that has terrestrial trace fossils that are well preserved and sedimentary structures that are microbially induced lying immediately over shallow water ripplemark structures. 11 foraminiferan samples were taken at intervals through approximately 20 m of succession, which spanned the tracked horizon, terminating just below the Hellenikon Group conglomerate.  Hastigerina pelagica and Globigerina pseudobesa are both present in all the samples tested, which constrained the samples to a time interval between 8.5 Ma, the Late Miocene, the Tortonian, and 3.5 Ma, (Early Pliocene, Zanclean; Zacharisse, 1875; Kennett & Srinivasan, 1983; Boudagher-Fadel, 2013). Gierliński et al. concluded that the succession at Trachilos can be securely assigned to the Miocene and dated to the time interval 8.5 to 5.6 Ma based on:

1.     The end-Miocene Hellenikon Group is the only terrigenous incursion during this time interval into marine succession in western Crete;

2.     There are no index taxa that are dating to the post Miocene in the foraminiferan samples; and

3.     Deepwater marlstones, rather than shallow water carbonates, represents the Zanclean of Crete (Freudenthal, 1969; van Hinsbergen & Meulenkamp, 2006).

It is suggested by its position close to the Hellenikon contact that it represents the latest part of that interval, immediately before the desiccation event; Gierliński et al. approximate its age as 5.7 Ma to reflect this. A large number of representatives of the genus Elphidium in the benthic component of the foraminiferan assemblage indicates a shallow marine environment that was not more than 50 m deep with a relatively high salinity (35 ‰ - 70 ‰); Murray, 2006).

At the Trachilos tracksite the succession consists of an alternating series of conglomeratic beds and intraformational conglomerate/local breccia horizons (with skeletal debris lithoclasts); silty limestones, organodetric limestone; fine-grained, calcareous sandstone that was highly bioturbated, sometimes containing skeletal elements; and yellow marlstones that are poorly lithified, together representing a marginal marine environment. Horizons that contained many shallow water body fossils, such as algae; cf. Lithothamium, and trace fossils that are commonly seen in shallow water environments (e.g., Thalassinoides isp.) can be seen in the section. Also included in the fossil assemblage are bivalves, gastropods, echinoids, ostracods, foraminifera, fish bones and scales, and the bones of marine mammals.

The emergent horizon forms an exposed surface about 21 m x 6 m in maximum width, and can also be identified in section in neighbouring outcrops. The strata that lies immediately beneath contains ripples that are moderately large that have structures (wrinkles) that are related to microbial mats on their crests, which suggests deposition in water that is extremely shallow (Eriksson et al., 2010; Banerjee et al., 2014). The emergent is composed of a lower level (Surface A) which carries 2 distinct patches overlying sedimentary rock (surface B1 and B2), that are about 2-4 cm thick, that have edges that are sharply defined and upper surfaces appearing to represent a single bedding plane. According to Gierliński et al. B1 and B2 are clearly erosional elements of a sediment cover on top of A that had previously been continuous. Parting lineations  that had a SSW-NNE orientation are revealed by surface A, and truncation of fine laminae are shown locally at the contact with layer B, which suggests that the erosional event that had removed most of layer B may have occurred during the period of deposition, not when the strata became exposed in the recent past. Surfaces A and B dip gently to the east, and A is overlain by a soft marlstone that is yellow and poorly lithified. There is no sedimentary rock that has been preserved overlying B1 or B2, but it is inferred by Gierliński et al. that these patches were also covered by marlstone. The excellent preservation and clean surface of the B2 surface is probably explained by the contrasting hardness between the soft marlstone and the hard surface B2 that is well lithified.

There are 42 oval impressions between B1 and B2 that are filled with sediment on surface A, all of which are of approximately similar size and Shape that have long axes oriented SSW-NNE. On the NNE end of many of the impressions they are associated with a small field of ripple crests that are oriented perpendicular to the SSW-NNE long axis which suggests flow to the NNE. In the Messinian the general configuration of the reginal landscape was the same as it is in the present, with land to the south and sea to the north; a direction of the flow to the NNE, coupled with the complete lack of marine macrofossils in surface A, suggests, therefore that this flow of water could represent a temporary freshwater flooding event, possibly a small stream that burst its banks. Gierliński et al. interpret the impressions as poorly preserved tracks some of which are arranged in a linear series. It is not certain if they represent under-tracks or primary tracks on parts of surface B, or primary tracks that were made in shallow flowing water prior to the deposition of surface B. Surfaces B1 and B2 have different textures, in spite of representing the same level, and only B2 preserves ichnofossils: This, presumably, reflects differences of the substrate on a local scale.  B2 surface is very hard and its surface is dense, fine-grained sandstone that is well lithified calcareous sand, marked by parallel striations that are very subtle that have an alignment that is SSE-NNW alignment that appear to represent wind scour. Surface B contains no marine macrofossils, as is the case with surface A. Gierliński et al. interpret surface B2 as representing an area of sand that is aerially exposed which is close to the shoreline, and possibly represents part of a small river delta.


There are more than 50 ichnofossils on surface B2 in a total area of less than 4 m2. They range in size from less than 5 mm to 20 mm long, though this includes a small subset of small, irregular features, the origin of which is not known. The ichnites that are easiest to identify as footprints range from a size of 94 mm – 223 mm. Though it is difficult to discern individual trackways on the surface that has been densely trampled, the majority of footprints show a NE-SW long axis orientation. There are 2 identifiable trackways that conform to this pattern, and indicate that the trackmakers were travelling towards the southwest of the present. As there are no forelimb tracks it seems both narrow tracks were formed by a bipedal trackmaker. What appears to be a pair of left and right footprints may have been made by a stationary individual, the possibility cannot be excluded that this is a chance association.

The quality of preservation of the tracks is variable, though the best specimens have clear displacement rims (expulsion rims) and edges that are defined sharply. They are preserved as impressions that are shallow and partially excavated in concave epirelief that also contain pullup features that are associated with adhesion of the substrate and ejecta. They were impressed into a compact, slightly adhesive substrate. Tracks that were less well preserved lack detail and in some cases may have been modified by wave action. Included among the ichnofossils were some large complex structures, possibly representing multiple overprints.

The tracks are of similar size and may have consistent outlines across all the specimens. This is an oblique subtriangular shape that is formed by the combination of a heart-shaped, plantigrade sole with a narrow, tapering heel region, and a digital region that is asymmetrical with a large hallux and lateral digits that are progressively smaller that are all attached to the anterior margin of the sole. Therefore, the tracks are extremely entaxonic. There is no significant divide between the impressions of the first digit and lateral digits, though a gap of a few millimetres is sometimes visible in prints that have been well preserved, but in other examples the impressions are confluent. The entaxonic and the lack of a gap between the hallux and the other digits are evident even in the prints that are poorly preserved. Impressions of claws are not visible in any of the prints. The heel impression in a few of the prints appears to be bulbous rather than narrow, though this is an effect that is produced by an expulsion rim that is unusually large; the narrow and  pointed shape of the heel impression is clearly visible in prints in which the expulsion rim is small. There are 3 of the prints that are especially well-preserved, which provide morphological information about the trackmaker’s feet. The smallest print has a length of approximately 105 mm. This print has an expulsion rim that is strongly developed around the heel impression, across most of the sole there is a flat infill and 2 large blobs of adhering sediment in the heel region and behind the ball. The track shows a set of digit impressions that are well-preserved as well as part of a ball impression. A strongly asymmetrical curving array is formed by the digit impressions. The impression of the first digit is morphologically distinctive, is larger than the other impressions, and is slightly offset from them. There appear to be 4 lateral digits, though in the impressions the boundaries between them are to some degree indistinct. According to Gierliński et al. it is noteworthy that there is no trace of claw impressions, though the tips of the digits have dug into the sediment. The largest print shown in Fig. 9b of Gierliński et al. is approximately 135 mm long. With the exception of an L-shaped patch of sediment that is adhering that extends along the mid-lateral part of the sole, much of the plantar sole appears to be preserved, and a short limb that marks the crease between the sole and the digits.  A deep, rounded ball impression that has its own small expulsion rim is included in the plantar surface. The impression of the first digit is deep; it has a clear outline that shows a narrow neck that and an expanded, asymmetrically trapezoidal to oval pad. There are distal ends that are well-defined on the impressions of digits II-IV, though becoming less clear proximally. Digits II and III are slender and parallel-sided, and have squared-off ends; the impression of digit IV is shorter and oval with a tip that is slightly pointed. It is suggested by the depth of the impression of the ball and the apparent deflection towards the right of the digit impressions that the foot rotated clockwise on the ball during the step. The print in Fig. 9c of Gierliński et al. is one the largest prints, about 154 mm long. It is separated into separate anterior and posterior parts by a circular displacement features that surround the entire print, which shows that parts were generated by a single footfall, Impressions of a Hallux and 4 lateral digits are included in the anterior part.

Interpretations and implications

There are some characteristics of the Trachilos footprints that need to be explained: bipedality and plantigrade posture, pentadactyly with entaxony and the absence of claws; the bulbous first digit and short lateral toes; and in some of the tracks the presence a distinct ball. It is suggested by the morphometric analysis there is closer affinity to hominin track outlines than to those of extinct non-hominin primates. This leaves 2 possible interpretations:

1.     Gierliński et al. suggest the Trachilos tracks may have been made by a phylogenetically basal member of the Hominin clade. The combination of characteristics that are unique to the Hominin clade in the anterior part of the foot may be explained by this interpretation, such as pronounced entaxony, non-divergence and distal position of the hallux, the shape of the hallux and its size relationship to the ball, as well as the distal ends of digits 2-4, with a sole that is rather generic and relatively short, lacks an arch and has a heel that is narrow and tapered. Under this interpretation the tracks would represent a small, primitive hominin that was habitually bipedal that had human-like pedal digits and ball combination with an ape-like sole that lacked a bulbous heel. The non-divergent hallux and short lateral digits of the Trachilos tracks are not present in the skeleton of the foot of Ardipithecus ramidus that are known from Ethiopia (Lovejoy et al., 2009a, b), which are more than 1 million years younger (White et al., 2009). The age of the Trachilos footprints is not problematic for this hypothesis, though it is strikingly early: assuming the age of a bit more than 5.6 Ma, they are approximately coeval with Orrorin and Salhelanthropus (Sénut et al., 2001; Brunet et al., 2002; Almécija et al., 2013). There are no trackways that are known in Africa from hominins in the Miocene and almost nothing is known of the foot morphology of African hominins from the Miocene. The evidence for hominins of Miocene age in the European body fossil record was at best ambiguous, until recently (Spassov et al., 2912), though Graecopithecus, a primate from the Messinian, which was represented by 2 fragmentary specimens from savannah environment in Greece and Bulgaria (Bӧhme, 2017), was reinterpreted as a probable hominin based on dental characteristics (Fuss et al., 2017), while this paper was in review. According to Gierliński et al. with the Greek mandible dating to 7.175 Ma and isolated Bulgarian tooth dating to 7.24 Ma (Bӧhme, 2017), Graecopithecus is probably somewhat older than the Trachilos footprints. Obviously, it is highly relevant to the interpretation of the prints, though the fragmentary nature of the specimens and with no postcranial material limit the conclusions that can be drawn from it as yet.

2.     An alternative explanation, a hitherto unrecognised primate could be looked at, which is potentially not related to the Hominini, though having overall morphological similarities with this tribe. The characteristics which are hominin-like, the anterior placement of the first digit in particular, would reflect an example of convergent evolution, which is familiar in the fossil record (Emery & Clayton, 2004; Lockley et al., 2008; Parker et al., 2013). It is common to find fossils apes in Europe from the early Middle Miocene to the Late Miocene (middle Turolian; Harrison, 2010; Sénut, 2010; Spassov et al., 2012; Bӧhme et al., 2017). The cooling and drying of the climate has been suggested for the cause of the disappearance of apes from the region. A point that has been noted is that apes persisted somewhat longer in Eastern Europe than in Western and Central Europe. It has been argued that these late eastern apes had adapted to landscapes that were drier and more open (Casanovas-Vilar et al., 2011; Spassov et al., 2012). Ouranopithecus, which was a genus of large sized animals that was present in Greece and adjacent regions during the Late Miocene (9.6-8.7 Ma: Koufos & de Bonis, 2005), has been proposed to be a close relative of the Hominini or Hominidae (hominins, chimps and gorillas), though this is debated (de Bonis & Koufos, 1993; Begun et al., 2012; Koufos, 2015). The pedal morphology and locomotory behaviour are not known, because most of the available fossils are of craniodental material. This alternative hypothesis is therefore not implausible, though it should be noted that it is not positively supported by data from the pedal skeleton of any of the known primates from Europe.

It is clear that the first of these interpretations is more straightforward. Convergence with hominin morphology is not positively suggested by anything about the character complement or morphometrics of the Trachilos prints and, as noted above, there is no positive body fossil evidence for such a primate that is convergently hominin-like. In a formal sense the second interpretation therefore fails the Occam’s razor test of explanatory parsimony. According to Gierliński et al. they believe that explanation 2 should be entertained nerveless, as nature is not always parsimonious and, what is more important, there are major biogeographical implications that are associated with the first interpretation that also should be examined critically.

Crete was separating from the mainland during the Late Miocene by extension faulting and the formation of the Aegean Sea Basin (vans Hinsbergen & Meulenkamp, 2006). Faunas from the Late Miocene (Valesian-Turolian) of Crete include large mammals that were not endemic to Crete such as Hynaenids, proboscideans (gomphotheres and deinotheres), a hipparionine horse, pigs, a cervid, a bovid and tragulids (Benda et al., 1970; de Bruijn et al., 1971; Kuss; 1976; Leinders & Meulenkamp, 1978; van der Made, 1996; Athanassiou, 2004; Poulakakis et al., 2005a; Iliopoulos et al., 2012), which suggests that there was a land bridge that was still present. A final separation between western Crete and mainland Greece occurred by 5 My BP (Poulakakis et al., 2005b) is suggested by molecular biogeographic studies of extant lizards and gastropods. Gierliński et al. suggest the Trachilos trackmaker may have inhabited a peninsula of mainland Greece that was SE-trending instead of an island. It is clear that Crete never had a direct connection to the southern shore of the Mediterranean, whatever the exact timing of the process of separation. It is therefore implied by the identification of the prints as hominin that a minimum range extension for this group from Africa to encompass the Levant, Asia Minor, and southern Balkans.

Gierliński et al. suggest the question is whether such a range extension is credible. It seems doubtful, from a present-day perspective, because the Miocene hominin locations that are known in Chad, Ethiopia and Kenya are separated from the northeast coast of the Mediterranean by the expanse of the Sahara Desert, with a tenuous, and discontinuous, chain of mesic environments being provided only by The Nile Valley and Levant between the two. Conditions in the Messinian were, however, very different, with monsoonal rainfall over northeast Africa forming well-watered environments that drained northwards through the Eonile River and the Eosahabi Rivers (in Libya), and south into Lake Chad, which was much larger at that time than it is at present (Griffin, 2001). No evidence is known of inhospitable environments that would have formed a barrier to the dispersal of early hominins. Gierliński et al. concluded that a hominin interpretation of Trachilos footprints is not implausible biogeographically.


Gierliński et al. presented 2 alternative interpretations of the ichnites that were found at Trachilos. The hypothesis that the Trachilos trackmaker was a basal hominin has substantial implications for the biogeography of early hominins, as well as for the development of bipedality, and the entaxonic foot (Lockley et al., 2016). Gierliński et al. suggest it might be prudent to delay the taxonomic assignment given the challenging nature of this potential interpretation. These tracks are not poor trace fossils, in spite of their full 3D anatomy, was not preserved optimally. The outlines of these tracks are particularly clear, forming the basis of the morphometric analysis that was presented in this paper. Though it would be helpful if more trace fossils were available, but based on the currently available evidence the potential implications cannot be ignored, however challenging these implications may be. Gierliński et al. say the search for ichnofossils from the Late Miocene should be continued in the Mediterranean area in order to resolve the identity of the Trachilos trackmaker.

Sources & Further reading

  1. Gierliński, G. D., et al. "Possible hominin footprints from the late Miocene (c. 5.7 Ma) of Crete?" Proceedings of the Geologists' Association.





Author: M. H. Monroe
Last Updated 17/09/2017
Journey Back Through Time
Experience Australia
Aboriginal Australia
National Parks
Photo Galleries
Site Map
                                                                                           Author: M.H.Monroe  Email:     Sources & Further reading