University of Texas at Austin wordmarkUniversity of Texas at AustinCollege of Liberal Arts wordmarkCollege of Liberal Arts
Texas Beyond History
TBH Home

What the Middens Tell Us

photo of an overgrown midden next to contour map
Views of a Midden. In spite of centuries of overgrowth by grass and trees, this Camp Bowie midden (left) still maintains a characteristic donut-shape with outer ring of fire-cracked rock and central depression where the baking pit was located. Archeologists mapped the surface topography of the midden so the structure could be seen more clearly (contour map at right). Photo courtesy Texas Army National Guard; map by the Center for Archaeological Research, University of Texas at San Antonio.
photo of archeologists setting up an excavation
Archeologists from UTSA set up screens and shelter to begin the difficult process of excavating through layers and layers of burned rock. Their efforts will provide information on the size, structure and contents of the Camp Bowie middens, data which can be compared with other middens across the state.
profile drawing of midden
This profile (top) and plan view (bottom) of a Camp Bowie midden show a central pit lined at the bottom with large slabs of rock and surrounded by discarded pieces of smaller, fractured rock. Graphic by CAR-UTSA. Click to see full image.
geologic map of Camp Bowie area
Geologic map of the Camp Bowie area showing formations containing sandstone, limestone, and chert, resources important to prehistoric peoples. Areas marked in gold-brown are Holocene or Pleistocene deposits of primarily limestone. Map adapted from V. E. Barnes, Geologic Atlas of Texas, Brownwood Sheet, UT Bureau of Economic Geology 1976. Click for more detail.
photo of A camas
An eastern camas (C. scilloides) plant in full flower. Prehistoric peoples dug up bulbs of this plant and other species and baked them in hot-rock ovens to make them edible. Photo by Clarence A. Rechenthin, USDA-NRCS PLANTS Database.

Studying the burned rock middens at Camp Bowie was a painstaking but rewarding process. After the last shovelfull of dirt was sifted and the last rock sorted and counted, researchers at the Center for Archaeological Research (CAR) at UTSA had compiled a wealth of information that helps us better understand how prehistoric peoples lived in the region. Animal bones and plant remains—particularly the surprising prevalence of flower bulbs—provide a fascinating look at prehistoric diet as well as an indication of what the environment was like during the time of the middens. Chipped stone artifacts and oven construction techniques brought prehistoric technologies into focus. Radiocarbon dates of charcoal samples helped establish a timeline for midden use.

When all of the different types of evidence was combined, a picture of late prehistoric life in what is now Brown County, Texas, emerged. On a larger scale, the Camp Bowie data can be compared with that from other investigated middens across the state.

Midden Structure: What It Takes to Cook with Rock

The burned rock middens at Camp Bowie are "central feature" middens. The mounds, which are generally 10-15 meters (33-49 feet) in diameter, have a central depression within which most of the cooking took place. In order to understand how the middens were constructed, the archeologists excavated test pits (units) within the centers of the middens and on the "rings" of the middens. These 1-x-1 meter units (3.3-x-3.3 feet) were excavated in 10-centimeter levels (uniform artificial layers about 4 inches thick), and the artifacts found within each level were collected. The burned rocks from each level were sorted by size, counted, and then discarded.

Most of the Camp Bowie middens had a clearly defined central feature—a pit—usually containing high concentrations of charcoal and lined with rocks. The archeologists knew that any charred plant remains they found in the pits were likely either the remains of food cooked in the pit or of the wood used to fuel the fires.

Texas archeologists have learned more about how middens formed by creating experimental earth ovens using various techniques. The experiments have shown that it takes a considerable quantity of wood and stone to heat an earth oven to a temperature hot enough to cook the food in it. For a pit approximately 1.5 meters (5 feet) in diameter and .5 meters (1.6 feet) deep, roughly 200 kilograms (441 pounds) of rock and 250 kilograms (551 pounds) of wood are needed to create an earth oven that would stay hot for at least 36 hours.

These experimental data also provide an estimate on the amount of refuse (fractured rock) produced after several firings. Based on these experimental findings, one of the larger burned rock middens at Camp Bowie may represent more than 300 firings. Collectively, this would have required 86,500 kilograms (190,698 pounds) of wood and 48,250 kilograms (106,150 pounds) of rock over the life of the midden! It is also estimated that up to 5,000 plant bulbs could have been cooked during a single firing. (Keep in mind that these bulbs were small—about the size of a scallion or green onion.)

With so much wood and rock needed, it seems likely that earth ovens would be built (and middens would form) close to concentrations of these resources to reduce the amount of labor needed. Camp Bowie is situated in an area with several geological formations that would provide abundant stone for rock-lined earth ovens. The large floodplain of Pecan Bayou may have had large stands of oak trees, and eastern camas, wild onions, and other bulbs. The availability of rock and wood close to where the bulbs grew made the Camp Bowie area the perfect setting for the earth ovens.

Food and Fuel: Identifying the Plant Remains

Finding prehistoric plant remains in open-air sites (i.e., unlike protected sites, such as rockshelters or caves) is uncommon. Unprotected from the elements, most plant remains would rot in the ground long before any archeologist found them. The plant fragments archeologists recovered from the Camp Bowie midden sites had been burned and carbonized. If left intact, carbonized material, such as wood charcoal, will preserve for hundreds or even thousands of years. While the burning process served to preserve certain plants fragments, it also tended to make species identification more difficult. In the case of bulbs, heat and pressure can change the bulb shape. In many cases, however, the structure of the plant was sufficiently intact for the analyst to identify it using comparative collections of unburned plants and identification keys.

The charred plant samples recovered during excavations consisted of wood charcoal and what the archeologists thought were burned seeds. This material was sent to paleobotanist and archeologist Dr. Phil Dering for identification. It turned out that most of the charred remains identified as seeds were in fact plant bulbs or geophytes. Dering identified nearly 400 bulbs and pieces of bulbs and only one seed—a charred mesquite seed—as well as several species of charred wood, including oak, mesquite, juniper and willow. Oak was by far the most common wood species identified.

Bulbs are not roots but instead are modified leaves arranged in a rosette fashion around a compressed, central stem similar to an artichoke. The thickened, modified leaves of a bulb are termed bulb scales. Because they are leaves, the epidermal cells (outer skin) of bulb scales have distinctive shapes that are duplicated within a species, and are often unique to that species. The epidermal pattern provides a key diagnostic feature for identifying the genus or species of the plant because the overall shape of any bulb is modified during charring.

In order to identify prehistoric geophytes from this and other sites, Dering has established a reference collection of modern geophytes that would have been present in the area and used in prehistoric times. Photographs of the inner and outer epidermal layers of the modern examples (taken using a scanning electron microscope) were used for comparison to the charred archeological specimens. Dering identified three species of bulbs from the Camp Bowie burned rock middens: eastern camas, wild onion, and dog's-tooth violet. These are all species of bulbs that have been used by Native Americans in historic times.

photo of bags of rock
Bags of rock were a common sight during midden excavations. Archeologists mapped, sorted, and counted burned rock as they excavated downward into the middens. These types of data helped investigators plot the parameters and shape of midden features.

Click images to enlarge

photo of a plant bulb
A modern example of a bulb from plants that grow in the Camp Bowie area. These were used as comparative samples to aid in identification of the charred, prehistoric bulbs. Photo by Phil Dering. Click to see more examples.
photo showing several charred plant bulbs on a red background
These charred camas bulbs are over 8,000 years old. They were found amid the remains of an ancient earth oven at the Wilson-Leonard site near Cedar Park, Texas. This find shows that the burned rock middens of the Camp Bowie area were part of an economic pattern established thousands of years earlier. Photo by Phil Dering.
photos of electron microscope views
The outer "skin," or epidermis, of bulbs of three plant species, as seen through a scanning electron microscope. Dering found the cell patterns in these and inner layers were the most useful characteristics when identifying the prehistoric charred bulbs from the middens. Each species has its own cell pattern so that, even if charred, the bulbs usually can be identified. Left to right, eastern camas, Drummond onion, and dog's-tooth violet.
photo of arrow points
Arrow points, a hallmark of the Late Prehistoric period (about A.D. 800 to 1600), were among the chipped stone artifacts found within the middens. These artifacts were made during the same period that radiocarbon evidence suggests was the principal period of midden use. Shown from left are: a triangular arrow point blank and Scallorn arrow points from the early part of the Late Prehistoric. Click to enlarge.
photo of a stone mano
A stone mano from Camp Bowie. These hand-held tools were used to grind seeds and nuts into meal. Photo courtesy of CAR-UTSA.
chart showing probability distribution of calibrated radiocarbon dates of Bowie middens
This chart shows the probability distribution of calibrated radiocarbon dates derived from 31 samples from Camp Bowie middens. The blue shading marks the primary time period during which the middens were used. Graphic by CAR-UTSA. Click to enlarge and see complete chart.
map of burned rock midden density across Texas
Burned rock midden density across the state. The relative density by percentage of recorded middens in counties is shown by color, grading from peach tone (low) to red (high concentration). The areas with no middens present are in counties that either lack burned rock middens or have extremely low percentages of middens. Black star marks location of Camp Bowie. Graphic by CAR-UTSA. Click to enlarge.
photo of Phil Dering
Paleobotanist and archeologist Dr. Phil Dering, shown sampling the base of a lechuguilla plant after baking it in an experimental hot rock oven. Dering, who identified prehistoric plant remains from the Camp Bowie middens, suggests that sotol and lechuguilla are the most likely plant foods processed in middens in the southern section of the Plateau, as well as to the west.
photo of an acorn
Acorns may have been an important food resource on the Edwards Plateau. While some researchers have suggested that burned rock middens represent acorn-processing stations, the evidence for this interpretation is largely circumstantial. No ethnographic accounts document the intensive use of earth ovens for acorn processing.
photo of dart point
This Ensor-style dart point was found outside one of middens at Camp Bowie. This style was made at end of the Late Archaic period just before the intensive period of midden use in the area began.

The majority of the identified bulbs were of eastern camas (Camassia scilloides, also called wild hyacinth). Central Texas is at the southwesternmost extent of the historic distribution of eastern camas. Eastern camas tends to grow in moderately wet areas. Its range may have extended farther south and west during wetter periods in the past. The bulbs have the highest nutritional value in the spring, about the time the flower stalks begin to rise, and this is likely when they would have been collected. It is also the time of year when other plant resources are scarce and most mammals are in poor condition (having burned up stored fat to survive the winter).

The use of earth ovens to "bulk process" camas is best known in the Pacific Northwest and northern Rocky Mountains. Archeologist Alston V. Thoms (now at Texas A&M) combined early historic accounts and archeological evidence to show how the native use of western camas (Camassia quamash) intensified over time. Initial use began at least 7,000 years ago and the plant became a staple in the region around 4,000 years ago even though it was a "high-cost" resource (i.e., one that required lots of labor relative to caloric return). Thoms linked the intensification of camas exploitation in the Pacfic Northwest to increasing population pressure. Texas researchers have applied many of Thoms' ideas to understanding the burned rock middens of central Texas.

The plant food remains recovered from Camp Bowie middens are of significance to researchers in Texas and beyond. According to Dering, the collection is the largest preserved assemblage of plant food from earth ovens recovered from sites on the periphery of the Southern Plains. Further, the identification of the dog's-tooth violet marks the first identification of this plant from an archeological site. The best record of its use by historic tribes comes from observations of the Thompson River Indians in southwestern British Columbia. The Thompson Indians understood that the raw bulbs of dog's-tooth violet were poisonous, and they pit-baked the bulbs in order to remove the poison.

Chipped Stone Tools and Bones

Archeologists found a variety of chipped stone tools and animal bones at the Camp Bowie sites. In all, however, only 16 diagnostic artifacts—those representing distinctive styles known to date to certain time periods—were found within the middens. These include, from oldest to youngest, a Middle Archaic dart point, seven Late Archaic dart points, and eight Late Prehistoric arrow points. With the one Middle Archaic exception, these diagnostics fall into the same time periods as the radiocarbon evidence discussed below).

A variety of different types of chipped stone tools used for butchering animals, scraping hides, chopping wood, and other daily tasks also were found, along with quantities of chipping debris left behind by prehistoric tool makers.

Of interest were ground stone artifacts and features related to food processing—the grinding or pounding of plants, seeds, and nuts. Hand-held manos, or grinding stones, such as the one shown at left, would have been used in combination with grinding slabs, broad stones with shallow depressions that served as receptacles for nuts and seeds as they were pulverized or ground into meal.

Near one of the midden sites, archeologists discovered 12 bedrock mortar holes. These deep, circular holes in flat limestone bedrock have been worn from the pressure of prehistoric people pounding food with stone or wooden pestles. The mortars were grouped in clusters of three and nine and were flanked by two middens. They are roughly 8 to10 cm (3 to 4 inches) in diameter and 6 to 10 cm (2.5 to 4 inches) deep, although several were as large as 20 to 25 cm (8 to 10 inches) in diameter with depths up to 20 cm. A map showing distribution of the mortars and middens can be seen by clicking on the mortar image.

Archeologists also recovered animal bones in the middens, though they were not common. Bone tends to weather and deteriorate over time, whether on the surface or buried in the ground. Among the identifiable bones were those of white-tailed deer, bison, jackrabbits, cottontail rabbits, birds, and turtles.

Dating the Middens at Camp Bowie

In addition to determining what the burned rock middens at Camp Bowie were used for, the archeologists also wanted to know when the middens were used. The most reliable dating evidence at most prehistoric sites is radiocarbon dating. Radiocarbon dates or assays are statistical estimates of the age of the dated material. Dates were obtained on 31 samples of charcoal from the middens—most of the dated charcoal represents fragments of firewood. The results indicate that most of the Camp Bowie middens were used primarily during a relatively brief period spanning some 650 years.

The radiocarbon date chart on the left shows that 28 of the dates fall during the Late Prehistoric period between A.D. 750 and 1400. A single date falls slightly earlier at A.D. 600 and two dates fall even earlier, within the Late Archaic period, at 600 B.C. and 1200 B.C .

The charcoal samples chosen for radiocarbon dating came from different levels within excavation units in 17 of the middens. The sampling strategy was to try to get multiple dates from different levels within the same 1-x-1-m excavation unit. Whenever possible, the samples were obtained from units positioned on the midden ring rather than within the center of the midden. The thought was, if the middens form from repeated raking out of a central feature, with the removed material being deposited in the ring, then deposits located in the ring would likely be less disturbed than those in the central midden. It was assumed that the central areas of the Camp Bowie middens had been repeatedly churned up as earth ovens were built and rebuilt there.

Multiple samples were not available from some middens and in these cases the archeologists attempted to obtain samples from the lowest possible level within the midden. The radiocarbon date chart presents the probability distribution for each of these 31 (calibrated) midden dates from Camp Bowie. The dates are grouped by site, and within each site, by increasing depth. Dates from the same unit within a site are identified by different color groupings. The samples from sites with only a single date, or dates from different units within the same midden, are shaded in black.

There are 10 sites that have multiple dates from the same unit. These 10 color groupings in the chart are represented by 23 different dates. In all cases with multiple dates from different levels the radiocarbon dates form a pattern as expected, with older dates occurring at increasing depth. That is, in no case is there a statistically significant reversal of the dates (such as older above younger). This striking consistency is somewhat surprising in view of the fact that many of the charcoal pieces were quite small and could have moved up or down through the midden by gravity and disturbances such as tree roots and animal burrows. Given these results, it would appear that the midden rings gradually accumulated and were seldom disturbed.

The Distribution of Burned Rock Middens Across Texas

CAR researchers used their understanding of middens developed from the Camp Bowie data to look at broader patterns in midden use. Their first task was to look at the distribution of burned rock middens across Texas. They reviewed over 9,000 site forms and compiled all site records of burned rock middens for over 900 sites from more than 50 counties. These data were combined with an earlier study by archeologist Darrell Creel (TARL) to produce a data base of over 1,400 burned rock midden sites from 70 counties across the state. When the density of burned rock midden sites is plotted by county, the greatest concentration of sites is on the Edwards Plateau and in southwest Texas.

Because using pit ovens required a great deal of wood, it also is useful to compare the distribution of burned rock middens to the distribution of woody vegetation, especially oak, which is usually the dominate fuel wood found in burned rock middens (also, charred acorns are sometimes found in middens). However, because maps only exist for modern vegetation patterns, the distribution findings are only a broad indicator of differences in wood types.

The map on the left indicates that the distribution of burned rock middens within the state is tightly confined, concentrated primarily on the Edwards Plateau and in west Texas. Comparisons of midden distribution with plant distribution patterns (map on right) shows that, while the distributions of oak and middens partially overlap, there are high densities of middens found well southwest of the Edwards Plateau, in areas where oaks are uncommon. It could be argued that a slightly expanded prehistoric distribution of oak-dominated settings, to the southwest and to the west, could accommodate as much as half of the overall distribution of burned rock middens.

Nevertheless, even if we equate oak presence with acorn processing, it is unlikely that oak accounts for the entire distribution of burned rock middens. The distribution to the west and southwest of the Edwards Plateau is clearly not associated with oak, and by extension, clearly not associated with acorn processing. Archeologists such as Glenn Goode and Phil Dering have suggested that sotol and lechuguilla are the most likely candidates for processing in burned rock middens in the southern section of the Plateau, as well as to the west, where sotol and lechuguilla are more common. A number of ethnographic accounts document the use of these plants, processed in earth ovens, throughout New Mexico, Northern Mexico, Arizona, and into California. In addition, charred agave has been recovered from earth ovens and middens in southwest and far west Texas.

While the northern and eastern distribution of burned rock middens coincides with oak-dominated vegetation and could be correlated with areas of prehistoric acorn processing, the area also lies within the southwestern end of the distribution of geophytes (second map on right). Unlike oak wood, geophytes could not have been used as a fuel resource and clearly were being used for food, as indicated by their presence in the Camp Bowie middens. While the idea that acorns were being processed at burned rock midden sites cannot be entirely ruled out, the fact that oak is the major wood identified in Central Texas burned rock middens suggests that the occasional charred acorns found in middens may simply represent incidental burning associated with the use of oak fuel wood.

If geophytes—bulbs—rather than acorns, were the focus of most middens in north-central Texas, why are there no middens to the north? Eastern camas distribution, for example, continues to the north, and other geophytes are certainly in this northern section of the state also. The distribution of oak continues to the north as well. We have food and fuel, why don't we have middens? At present, archeologists have yet to seriously address that question. We can suggest, however, that if we are correct about geophytes, then the answer probably lies in considering alternative resources.

Geophytes are "low return" resources, they require a lot of work for the amount of nutrition they provide, and potentially were used only for a limited time during the early spring. We suspect that to the north other plant resources may have been available that reduced the importance of geophytes. At present, however, we have no suggestions as to what these alternative resources might have been. The absence of burned rock middens in many areas can be explained by the lack of suitable rock. Earth ovens can be built without rocks, but they may be harder for archeologists to recognize.

The southwestern distribution of burned rock middens probably reflects the use of lechuguilla and sotol, since it is unlikely that high densities of geophytes ever were present in this relatively dry region. Counties with the highest percentages of burned rock middens seem to be on the border between the areas to the west, probably associated with sotol and agave processing, and the areas to the north. While these higher percentages might reflect both processing of sotol and acorns, we would suggest that sotol, lechuguilla, and geophytes are a more likely mix.

Summing It Up

Between A.D. 750 to 1400, Late Prehistoric peoples repeatedly came to the Camp Bowie area to process large quantities of bulbs in earth ovens heated by hot rocks. As favored oven pits were reused over successive visits, burned rock middens formed at many places on the landscape.

The radiocarbon evidence suggests several periods of intensive use, separated by periods of lower use frequency. These patterns may be related to fluctuations in the availability of what was being processed in the middens, fluctuations in availability of wood resources, more general changes in settlement and subsistence, or a combination of these and other factors.

We believe that patterns in fuel wood resources are a critical component in understanding patterns of midden reuse. Because wood collecting was so labor intensive and because wood resources would have been locally depleted over time, the pit ovens may have been abandoned periodically until the area's trees had time to rejuvenate.

In addition, the differential distribution of these fuel wood resources may help to account for the strong association of burned rock middens with oak, an association that has previously been seen as related to the use of acorns.

Considering the spatial distribution of burned rock midden sites across the state, it is likely that two different sets of plant food resources (sotol and geophytes), each with different spatial distributions, might be the major food resources processed in middens within the state.

There is a strong likelihood that the Camp Bowie middens were ovens where geophytes were processed. Over hundreds of years, prehistoric people traveled to the area to harvest and process bulbs, particularly camas— the principal resource—during a short seasonal window of time, in the early spring. The evidence of their culinary handiwork still remains in the numerous burned rock middens and in an array of other data.

Managing Camp Bowie's Cultural Resources

Fourteen of the Camp Bowie burned rock midden sites were determined eligible for the National Register of Historic Places. For National Guard training purposes, these sites are considered "sensitive environmental areas" and are officially "off limits." These areas are monitored regularly by TXARNG training site managers and TXARNG cultural resource staff to ensure that they remain undisturbed. In the event that training activities or military construction threatens to impact any of these sites, the state historic preservation office (Texas Historical Commission) as well as the Native American tribes that historically lived on what is now Camp Bowie would be consulted and a reasonable compromise would be sought among the various interests.

Through the cultural resource management process, the Texas Army National Guard is able to continue its training programs at Camp Bowie to help ensure national security, while minimizing the loss of important historic information. This balanced consideration is precisely the outcome envisioned by Congress when the National Historic Preservation Act was passed in 1966.


photo of white dog-tooth violet in bloom
White dog's-tooth violet in bloom, a similar variety to the type recovered from a Camp Bowie midden. The finding of the prehistoric specimen of this bulb in an archeological context was a first in North America. Photo by Thomas G. Barnes, USDA-NRCS PLANTS Database. Click to enlarge.
photo of Late Archaic dart points
Late Archaic dart points were the other midden artifact types useful in dating the middens. Two radiocarbon samples also fell within the Late Archaic period, at 600 B.C. and 1200 B.C. The dart on left has the tell-tale "pot-lid" fractures and pinkish coloration derived from burning; it may have been inadvertently included within the earth used to cap the ovens.
photo of mortar hole
This bedrock mortar, one of 12 discovered in a cluster at Camp Bowie near two burned rock middens, likely was used with a pestle for pounding foods such as seeds or nuts. Photo courtesy of TXARNG. Click to see distribution map of middens and mortar holes.

FAQ: What are the Archaic and Late Prehistoric periods?

The Archaic period or era is a very long span of human history in North America that began about 9,500 years ago (7,500 B.C.) and lasted, in the Camp Bowie area, until about 1200 years ago (A.D. 800). The Archaic concept was originally conceived ... >>read more<<

map of wood vegetation across Texas
Wood vegetation across central and west Texas. The area in purple is predominately oak; the green area is mesquite-juniper. This map, a simplified compilation of various modern distribution maps, is not an accurate reflection of prehistoric patterns which, over time, were dramatically altered by both climate change and radical changes in land-use practices during the last 125 years. Nevertheless, the map provides some indication of large-scale differences in distribution of wood types . White star marks Camp Bowie location. Click to view full map.
graphic of camas distribution across the US
Camas distribution across the United States. Eastern camas still grows in central Texas, but it was probably much more common during certain times in prehistory when the climate was more moist. Most of the bulbs found in Camp Bowie middens were of this species. Black star marks Camp Bowie location.
photo of common camas in flower
Common camas (Camassia quamash) in flower. These grow primarily in the northwestern United States and have been found in archeological sites there as well. Photo by William and Wilma Follette.

Fourteen of the Camp Bowie burned rock midden sites were determined eligible for the National Register of Historic Places. For National Guard training purposes, these sites are considered "sensitive environmental areas" and are officially "off limits."

Through the cultural resource management process, the Texas Army National Guard is able to continue all of its training programs at Camp Bowie to help ensure national security, while minimizing the loss of important historic information.