Over the past few decades, archaeologists across North America have unearthed pits with concentrations of rocks altered by intense heat. These archaeological features tend to be circular or oval, from 20 inches to several feet across, and lined with a couple of dozen to several hundred rocks. Some are remnants of heating a shelter, some of ceremonial activities, and many of underground ovens. The earliest earth ovens in North America are more than 8,000 years old. Also known as pit ovens, they were used to slow roast or steam roots, nuts, fish, mussels, and meats.

Recent archaeological excavations in Glendalough and Lake Shetek state parks uncovered fire-altered rocks that show evidence of earth-oven cooking. At Glendalough, archaeologists found over 500 cobbles arranged in a shallow pit. Almost all were granite, ranging in diameter from 4 to 8 inches. All had a reddened color, crumbling surfaces, and deep, angular cracks. To alter granite cobbles to this degree requires heat in excess of 900 degrees. A radiocarbon date, obtained from charcoal associated with the rocks, indicates the oven was used around A.D. 1350. Such a large earth oven would have taken considerable labor to construct. It might have been used for celebration, ceremony, or preparation of a communal harvest.

A common method of preparing an earth oven was to dig a shallow pit and line it with cobbles such as granite or basalt. A fire was tended on the rocks until they began to glow red. The embers were then either spread aside or removed, and a layer of vegetation such as green grasses or leaves was placed atop the hot rocks. Uncooked food was placed directly on the vegetation or first wrapped in leaves. More vegetation was piled on the food, and the pit was covered with earth. The food was left to cook for a couple of hours or, in some cases, up to 48 hours.

After encountering fire-altered rocks in archaeological excavations, I decided I could gain a better understanding of these features by constructing and using experimental earth ovens. I have done so several times. Once I covered an earth oven after using it, left it for five years, and then excavated it to compare the experimental oven with our archaeological finds. I excavated it carefully with a trowel, recording with notes and photographs the "reveal" at each 5-centimeter vertical depth, just as we do with archaeological investigations. The configuration of rocks was similar to some fire-cracked rock features I have encountered in excavations, including one at Father Hennepin State Park. I presented the results of my earth oven experiments at a conference in 2013 to share what I learned with other archaeologists and to solicit their input.

As an interpretive naturalist at Mille Lacs Kathio State Park, I have also built and used earth ovens as public education demonstrations. To avoid disturbing any archaeological site, my colleagues from the Minnesota Historical Society and I dig an oven pit in beach sand in the park's human-made swimming area. At one annual Archaeology Day event in September, we roasted two ducks and a couple of squash in an earth oven. We also cooked wild rice in a reproduction ceramic pot and boiled sweet corn using fire-heated rocks. After 90 minutes of cooking, we served moist, tender duck with squash, corn, and wild rice to event volunteers.

Experimental archaeology stimulates discussion among archaeologists and state park visitors. For example, while observing a demonstration, a military veteran told us about a cook who used earth ovens to prepare meals for soldiers serving in Afghanistan. The cook had learned how to use earth ovens in an advanced field skills course. In the high-tech world of military engagement, this ancient cooking method allowed for continuous, efficient feeding of soldiers 24 hours a day.

Each earth-oven demonstration raises new questions. Keeping food free of soil and ash is a considerable challenge. Historical records and oral histories tell of using maple, basswood, or wild grape leaves to wrap food cooked in earth ovens. Science helps fill out the story: Soil can be microscopically examined for starch grains and small silica bodies known as phytoliths. This analysis can reveal what plants were cooked and what vegetation covered the food.

In a recent experiment, I did not have access to fresh basswood leaves. Instead, I wrapped a duck with store-bought turnip greens and bound the package with the inner bark of basswood, which I had soaked in water. With earth ovens, I get to eat a delicious meal with each experiment.