Three fossil horizons redraw Waco's Dig Shelter

Columbian mammoth remains exposed inside Waco's Dig Shelter. Their close arrangement once made a single catastrophic flood feel intuitive; the Dig Shelter sequence now resolves at least three fossil horizons. NPS photograph.[2]

Under the Dig Shelter in Waco, Texas, tusks curl through pale sediment like enormous parentheses. Skull fragments, ribs, limb bones, and vertebrae crowd the excavation floor. The remains are close enough, and some are articulated enough, to invite a cinematic explanation: a mammoth group entered one bad flood and became one frozen scene.

That is no longer the National Park Service's account of the deposit inside the shelter. The NPS visitor page, last updated June 8, 2026, says recent work has ruled out catastrophic flood deposits there. It describes three distinct fossil layers formed on a low-energy floodplain and points toward repeated environmental stress as an active line of investigation.[2] The sharper technical account is still preliminary, but it changes the question. The sediment may record how carcasses were buried without recording what killed every animal.

This revision does not make Waco's famous nursery herd disappear. Adult females and young still occur together, and that association remains rare evidence about the social structure of the Columbian mammoth, Mammuthus columbi.[2][5] What has changed is the scale of the tableau. The in-situ shelter sequence can no longer be compressed into one flood, one layer, or one cause.

The shelter makes one moment look plausible

Waco's story began in 1978, when Paul Barron and Eddie Bufkin found a large bone eroding from a ravine near the Bosque River. Baylor's Strecker Museum organized excavations, and by 1990 the first phase had exposed about sixteen mammoths in varying degrees of articulation. Later work found more mammoths and other vertebrates. Some remains were removed in plaster jackets; others stayed in place beneath the shelter that opened in 2009.[1]

The first group had an age-and-sex composition consistent with a family unit: adult females with juveniles and subadults. That is why “nursery herd” is more than a promotional nickname. The 2016 foundation document for the national monument treated the arrangement as the nation's first and only recorded evidence of a Pleistocene Columbian mammoth nursery herd. It also called the group the product of a single catastrophic natural event.[5]

A peer-reviewed 2015 study supplied a detailed version of that event. Lee Nordt and colleagues reconstructed most of the mammoths dying on a gravel bar of the ancestral Bosque River beside a collapsing tributary wall. Optically stimulated luminescence dating placed the principal event at a weighted mean of 66.8 ± 5.0 thousand years ago. The paper framed the assemblage as North America's largest Columbian mammoth herd lost in one catastrophe, while acknowledging that the timing, cause, and number of death assemblages had remained controversial.[4]

That model had real evidence behind it: herd-like age structure, close association, an early reading of limited scavenging, and sediment deposited by water. The scavenging premise was challenged almost immediately. In 2016, Logan Wiest, Don Esker, and Steven Driese documented vertebrate and invertebrate modifications to bones that required post-mortem exposure, then proposed a diminished watering hole and drought as a more plausible explanation than rapid fatal burial.[4][7]

That peer-reviewed challenge did not supply today's three-horizon stratigraphy, and its title deliberately called the scavenging evidence preliminary. It did establish a crucial separation: “water moved the sediment” is not equivalent to “a violent flood killed the animals.” Nor does physical proximity guarantee simultaneity. Those are separate propositions, and the new work tests them at the scale of grains, soil horizons, and vertical position.[7]

Three horizons break the single-scene spell

Ashley Gonzalez, Daniel Peppe, and Lindsey Yann presented the new stratigraphic case at the Geological Society of America's 2025 South-Central meeting. Their composite section through the Dig Shelter is about 3.5 metres thick and contains at least three discrete mammoth-bearing intervals. It is dominated by clay, interrupted by sand lenses, and organized into a fining-upward succession with weakly developed ancient soils. The team interpreted the setting as a distal floodplain under seasonal, semi-arid to dry sub-humid conditions—not the product of one dramatic channel-incision event.[3]

The NPS research report translates that section into the animals visible at the site. Its first layer contains the original nursery-herd horizon and a western camel. A second contains a bull mammoth with a juvenile draped over its tusks. A third holds an articulated female and an isolated pelvis. Clay- and silt-sized sediment buried the carcasses with little reworking or transport, but the horizons accumulated separately.[1]

That combination matters. A low-energy overbank deposit can cover a body without having killed it. An articulated skeleton can show that a carcass was not rolled far down a channel without showing whether drought, illness, entrapment, or some other stress caused death. Waco still has a fluvial history; sand lenses and water-laid sediment make that clear. The revision rejects a single major-flood deposit, not the presence of floods in the landscape.[1][3]

The word attritional can sound as if the new interpretation reduces the site to a meaningless scatter. It does the opposite. It says the assemblage grew through more than one mortality and burial interval. Vertical separation becomes information: the same floodplain setting accumulated carcasses in at least three separate intervals, and enough anatomical integrity survived for those episodes to remain distinguishable.[1][3]

Bone growth adds an independent stress signal

At the 2024 Society of Vertebrate Paleontology meeting, Dava Butler, Yann, and Peppe reported transverse growth-arrest features known as Harris lines in Waco mammoth limb bones. Their visual survey covered 65 elements: twenty humeri, fifteen ulnae, twenty femora, and ten tibiae. Among the twenty humeri, 40 percent had strongly developed lines and another 25 percent were classed as partial or ambiguous. The team radiographed six Waco humeri and two non-Waco comparison specimens; imaging confirmed the visual calls and detected lines in one additional Waco humerus. In juveniles, the lines lay in the most recent growth—consistent with prolonged physiological stress near death.[6]

The distinction between stress and drought must stay visible. Harris lines can form after malnutrition, systemic illness, or trauma; they do not name the environmental trigger by themselves. The conference abstract joins them to taphonomic and isotopic evidence consistent with recurring drought, making drought a coherent hypothesis rather than a diagnosis written directly into bone.[6]

The NPS's 2026 wording is appropriately restrained. It says the mammoths experienced multiple periods of extreme stress, that juveniles show stress near the time of death, and that active projects are testing harmful environmental conditions.[2] Drought is now a stronger candidate, but no equally finished replacement movie is warranted. Stratigraphy, climate proxies, and pathology are beginning to converge while the work remains active.

The publication ladder matters too. The 2015 catastrophic-event model and the 2016 scavenging-and-drought challenge both appeared as full peer-reviewed papers.[4][7] The newer contribution is stratigraphic: the three-horizon reinterpretation is currently represented by a 2025 conference abstract and an NPS account of Gonzalez's completed master's research.[1][3] The Harris-line case is also reported in a conference abstract.[6] Those results deserve attention because they are specific and because the NPS has already revised its public interpretation around them. Full publication will expose the new methods and comparisons to more exacting scrutiny.

The nursery herd survives the revision

It would be easy to overcorrect and say that if the shelter assemblage accumulated over time, the nursery herd was an illusion. The sources do not support that. The adult-female-and-young association remains concentrated in the lower group, and the monument continues to identify it as the only known Columbian mammoth nursery herd.[2][5] Three Dig Shelter horizons show that every mammoth did not arrive together; they do not automatically split one horizon into unrelated individuals.

The limit is more precise. Gonzalez's work focused on the in-situ sequence inside the Dig Shelter, while many first-phase nursery-herd fossils were removed from the area outside the modern shelter. The NPS report explicitly says more work is needed to understand that first-phase group and its associated remains.[1] The responsible update is therefore not “there was no herd.” It is “the herd cannot stand in for the formation history of the entire site.”

Even the head count is being reopened. In 2025 the NPS described a separate student project devoted to determining how many mammoths the collection actually preserves.[1] That is possible because the collection is both unusually rich and unusually difficult. Fossils exposed for years suffered modern weather and flood damage; excavation maps and notes are incomplete in places; fragments from one limb can occupy many boxes. In October 2024, Baylor transferred thousands of fossils and their associated data and archives to the national monument, while the Mayborn Museum continued as repository.[1]

This curatorial history is not housekeeping at the edge of the science. Matching fragments to individuals can revise the minimum count. Reconstructing a limb makes pathological features available for X-ray study. Recovering old maps can clarify a bone's position in the section. At Waco, the new field report is partly being written from the field notes and jackets of the old one.[1]

A better scene has more than one clock

The cover photograph remains spectacular after the flood story loosens. The tusks still curve through the sediment; nearby bones still preserve the physical company of mammoths. What changes is how long the eye should linger. Instead of reading the floor as one instant, it now has to move vertically through at least three horizons and analytically among three clocks.

One clock dates the deposits. Another records the repeated formation of floodplain sediment and ancient soils. A third sits inside growing bone, where interruptions register periods of physiological stress. None supplies the cause of death alone. Together they make the old single-catastrophe account increasingly difficult to sustain for the Dig Shelter assemblage.[1][2][3][6]

That is the scientific gain. Waco has not become less exceptional because its cleanest story was revised. It has become a better fossil site: one where herd structure, mortality, burial, preservation, and later excavation can be separated instead of collapsed into the same event. One flood was a defensible model; three horizons now require a more complex one.

cronfeed.work