A Middle Pleistocene hippo tracksite at Gombore II-2 (Melka Kunture, Upper Awash, Ethiopia)

Highlights

• At Gombore II-2 (Melka Kunture) a 700 ka volcanic ash caps a fossil hippo trail.

• Hippopotamus cfr. amphibius printed deeply into unconsolidated sediments.

• A fossil hippo trail, up to 0.7 m deep and 2 m wide, was excavated over 6 m.

• Natural casts provide information on the outer appearance of hippos' feet and legs.

• Hippos were resilient to Pleistocene volcanic events.

Abstract

In this paper we describe exceptionally well-preserved evidence of hippo activity at Gombore II-2, a 700,000 year-old site at Melka Kunture, at 2000 m asl in the Ethiopian highlands, in the Upper Awash Valley. We excavated and made casts of footprints that had been left after a volcanic eruption. The 3D casts provide a detailed record of the outer anatomy and soft tissues of Pleistocene hippos. They reveal for the first time the shape of the legs and feet of the animals. Gombore II-2 also provides information on the behavior of past hippos. This prominent species of the African wildlife affects the vegetation and the landscape, and acts as a geomorphological agent. The animals wallow beside bodies of water and erode deep trails linking them to pastures. At Gombore II-2 the trampling marks converge toward an erosional feature that we interpret as a fossil trail, documenting an early Middle Pleistocene behavior similar to that of today. Signs of trampling and erosional features interpreted as hippo trails have also been described in Pliocene and early Pleistocene levels at Olduvai and Koobi Fora. We further underline that the hippo population of this tract of the valley was resilient and re-established itself not long time after the volcanic eruption.

Introduction

Hippopotamus amphibius, the common hippo, is a prominent and locally dominant sub-Saharan species (Field, 1968, Eltringham, 1999, Chansa et al., 2011). Living in groups, or “schools” (Eltringham, 1999), of up to several dozens of adults, these animals have a strong impact on the environment, both direct and indirect. At night, they leave the wallows, ponds and rivers where they spend most of the daytime, and set off to forage to places that may be a few km away. They tend to follow established trails to the same grazing spots, where they pluck the grass with their lips, producing characteristic closely-cropped patches known as “hippo lawns” (Lock, 1972). They also act as geomorphological agents, wallowing beside bodies of water where they enlarge and deepen their favorite resting places. Being huge animals—an adult weighs 1.5 ton or more—with relatively small feet (Eltringham, 1999), their trampling back and forth along the same trails erodes the ground into long shallow trenches that are easily recognizable in the landscape and sometimes become real gullies. As time goes on, the trail floors get harder and harder, and retain their profile even when flooded by seasonally rising waters. The geomorphological impact of hippos and the channeling of water along their trails has been described both in the Ngorongoro crater (Tanzania) and in the Okavango Delta (Botswana), where hippos are viewed as “ecosystem engineers” (McCarthy et al., 1998, Deocampo, 2002, Mosepele et al., 2009).

As in the present, hippos were a feature of past sub-Saharan wildlife. In East Africa their fossil remains are frequently unearthed at Pleistocene archaeological sites. At Koobi Fora, early evidence of hominin-hippo interaction was also discovered: cut-marks were identified on 1.5 Ma-old hippopotamid bones (Bunn, 1994, Pobiner et al., 2008). A “hippo turnover” occurred at ca.1 Ma, when Hippopotamus became the dominant hippopotamid group in Africa (Boisserie and Gilbert, 2008).

Here we will focus on evidence subsequent to this event. We will describe in detail new Middle Pleistocene evidence discovered at Melka Kunture, when Hippopotamus was a permanent part of the landscape.

Melka Kunture, a cluster of archaeological sites extending along the Upper Awash Valley on the Ethiopian highlands (Fig. 1), is well suited to provide evidence of past hippo behavior. The Pleistocene environment was characterized not only by the slow-flowing Awash river, which meandered in a half-graben depression, but also, at least seasonally, by a variety of abandoned meanders and still bodies of water (Raynal et al., 2004). Furthermore, the landscape was modified repeatedly by the thick volcanic ash fallout, which clogged the streams and led to the formation of shallow ponds and pools. Starting at ca. 1.7 Ma, the vast archaeological record includes a number of Oldowan, Acheulean, Middle Stone Age and Late Stone Age sites (Chavaillon and Berthelet, 2004, Piperno et al., 2008, Mussi et al., 2014). The fauna discovered at sites close to the main river is invariably dominated by hippos determined as Hippopotamus cf. amphibius (Geraads et al., 2004, Gallotti et al., 2010, Geraads et al., 2017).

The Melka Kunture sites are named after the gullies and valleys where they are located, i.e. the tributaries of the Awash that drain the area, many of them seasonally. The name of the gully or valley is followed by the consecutive Roman numeral identifying the specific site. Exceptionally well-preserved evidence of hippo activity, including footprints, was recently discovered at Gombore II-2, a 700,000 year-old site of Melka Kunture which is part of a geological sequence documenting environmental change at the Mid-Pleistocene Transition (Mussi et al., 2016). The record described here in detail was produced on layers of volcanic ashes soon after volcanic eruptions. It also makes it possible to address questions related to hippo population resilience.

Archaeological surveys and excavations were first undertaken in the area of Melka Kunture by Gérard Bailloud (1965), and then by Jean Chavaillon, who directed the French Archaeological Mission from 1965 to 1998. Since 1999, research at Melka Kunture has been under the responsibility of the Italian Archaeological Mission, first directed by Marcello Piperno and, since 2011, by one of the authors of this paper (M.M.).

Gombore II-2 is a site at the top of the early and Middle Pleistocene stratigraphic sequence of Gombore gully, on the right bank of the Awash River (Egels, 1971, Chavaillon and Berthelet, 2004, Gallotti et al., 2010, Kieffer et al., 2004, Morgan et al., 2012, Mussi et al., 2016, Raynal et al., 2004). The archaeology was first investigated in 1974, when members of the French team noticed some fossils eroding from the top of the gully (Chavaillon and Berthelet, 2004).

A test stratigraphic excavation grid just 20 cm deep was dug over 7 m². Outside the grid, a systematic collection of exposed materials was also made over another 7 m² area. A total of 126 objects were collected: 49 lithic implements and 77 faunal remains, most of which were hippo bones. After preliminary determinations by Denis Geraads, at least two individual hippos were identified, together with some giraffid and equid remains (Chavaillon, 1975, Chavaillon, 1976). The site, currently called Gombore II-2, was named at the time “Gombore IIA”.

More excavations were planned for the 1980s, but were delayed because of the difficult political situation in Ethiopia at that time (Chavaillon, 1981–1982). New research eventually started again in 1993. In 1993 and 1995, Jean-Luc Boisaubert excavated 26 m². More lithic industry was discovered, together with fossil hippo bones (Chavaillon and Berthelet, 2004). Since 2001, the area excavated by the French team has been open to tourist visits, together with casts of the bone and lithic finds, and a panel suggesting that this had been a prehistoric hippo butchering site (Chavaillon and Piperno, 2004).

In 2012, the Italian team started a new project aiming to gather updated information. This led to the discovery that the stratigraphic sequence was more complex than expected. For one thing, on top of the already researched archaeological site was a previously unrecorded layer displaying the features discussed below.

Raynal et al. (2004) provided a detailed description of the stratigraphic sequence of Gombore gully, from the early and Middle Pleistocene, while absolute dates are available in Morgan et al. (2012) for episodic volcanic deposits alternating with fluvio-lacustrine deposits. Close to the Awash and almost at the river's present-day level, the earliest site is Gombore I, containing Oldowan lithic industry. Just 100 m upstream lies Gombore II, which includes several sub-sites (Gombore II-1, Gombore II-3, Gombore II-4, Gombore II-5, Gombore OAM), all of them just above a tuff unit dated by ⁴⁰Ar/³⁹Ar to ca. 875 ka (Morgan et al., 2012). Thousands of Acheulean lithic implements were found with faunal remains. The record was dominated by Hippopotamus cf. amphibius, as was the case everywhere else at Melka Kunture (Gallotti et al., 2010).

Gombore II-2 was discovered in the upper part of the sequence, 5 m higher up in the stratigraphy than the other Gombore II sub-sites. In 2013, 2014 and 2015, we re-excavated the uppermost 1.5 m of the sequence described below. A volcanic layer at the top of the succession is dated by ⁴⁰Ar/³⁹Ar to 0.709 ± 0.013 Ma (MK27-09), i.e. at the beginning of the Middle Pleistocene (Morgan et al., 2012). It lies atop the layer where we discovered the features described in this paper.