When Mark Tveskov crested the top of the ridge, he paused for a moment, scanning the fire-blackened forest to get his bearings.

When Mark Tveskov crested the top of the ridge, he paused for a moment, scanning the fire-blackened forest to get his bearings.

"The initial charge by the dragoon soldiers and the militia came up this side of the hill," said the director of Southern Oregon University's Laboratory of Anthropology in Ashland.

"The Indians were on the back side," he continued. "They fired off a volley that killed (militiaman) John Gillespie, then tactically retreated to the west with their women and children."

Tveskov was standing on the knob of a narrow ridge, one of countless spines in the vast area known as the Grave Creek Hills.

"You can see how they could have defended this razor-back ridge," he continued, then pointed to a spot slightly down the hill to the west. "See that little hump over there? That's where we found a lot of the fired musket balls."

These were musket balls — 17 have been found thus far — from the Battle of Hungry Hill on Oct. 31, 1855. The Halloween battle was the largest clash in the Rogue River Indian War of 1855-56 that began in early October and continued until the following June.

Tveskov recently visited the site with historian Ben Truwe of Medford to retrace the flow of the fray.

As they talked about the battle fought 158 years ago, you can almost hear the blast of the muskets and see the puff of smoke rising at each discharge of black powder.

The site is in a rugged area perhaps a dozen miles wide as the crow flies that includes the mountains west of Interstate 5 between Glendale and the Rogue River. The exact location is not being divulged out of concern it would attract artifact hunters.

The conflict involved about 100 Army soldiers known as dragoons and 200 militiamen on one side, with perhaps 200 American Indians, including women and children, on the other side.

Despite being outnumbered, the Indians clearly won the day, Tveskov said.

The Army suffered 39 casualties, including 10 who died on the battlefield and others who later succumbed to their wounds. It is believed that 16 Indians died on the battlefield, he said.

The battle was triggered by the Lupton massacre, in which more than two dozen Indians were slain in a village at the mouth of Little Butte Creek on Oct. 8, 1855, by vigilantes from Jacksonville. The fleeing Indians split into at least two groups, with one seeking protection at nearby Fort Lane under the command of Capt. Andrew Jackson Smith. The other group, led by Chiefs George and Limpy, headed down the Rogue River, killing several settlers en route. Atrocities were committed by both sides, according to historic accounts.

The battle would delay the end of the Rogue River Indian War until June 1856, Tveskov said. While the Indians won the battle, they would lose the war and ultimately be removed from their Southern Oregon homeland to the Grand Ronde and Siletz reservations in 1856.

After having been lost in the fog of time, the battle site was located last fall following a long search by Tveskov and others. The battlefield is within the U.S. Bureau of Land Management's Medford District. Hungry Hill west of Glendale is not connected to the site.

Begun in 2009, the field search covered more than 24 square miles of rugged ground. The historians and archaeologists bushwhacked into canyons where the brush was so thick they couldn't see their feet. They clamored like goats up steep mountainsides, sliding down others. And they pored over old records, from the National Archives in Washington, D.C., to the Bancroft Library in Berkeley, Calif.

Using information gleaned from old maps and other documents, they found musket balls and other items at the site, confirming it is the correct location, Tveskov said.

An old iron frying pan discovered in the area may have been from the battle, but that has not yet been confirmed. They also found horseshoes along the ridge, but the ridgeline was used as a trail early on, so it's unknown whether the shoes were related to the Battle of Hungry Hill.

This past summer, the search team spent two weeks on the ground, using metal detectors to unearth items left from the battle. But a dry lightning storm on July 26 disrupted the ground truthing. The resulting Douglas Complex fires grew to more than 48,000 acres, consuming the remote battle site.

"But the fire did make it easier to walk around," Tveskov said. "Before the fire, it was difficult to survey the battlefield because we were walking on two or three feet of brush at times."

The fire appears to have destroyed some clues, however.

"One piece of lead I found that I'm assuming was a musket ball looked like a shiny puddle of metal," Truwe said.

Charred spires of large pines and firs now stand on the ridgetop where the two sides clashed. Yet at the time of the battle, the ridgeline was described as an open area covered with grass.

"The Indians managed the forests with fire," Truwe said. "They did it to improve deer forage."

A government survey in the 1880s described the area as dense and brushy, the result of the end of annual burning by the Native Americans, Tveskov said.

An Indian encampment was discovered on Oct. 25, 1855, when Lt. August V. Kautz and 11 men came upon it while the soldiers were crossing the mountains from Fort Orford, an Army camp inland from Port Orford on the coast. After losing two men in the ensuing clash, Kautz headed for Fort Lane.

By the time the soldiers arrived from Fort Lane and were joined by militia from Jackson, Douglas and Lane counties, the Indian encampment had moved across a valley and up on the ridge.

"The Indians chose this spot because they knew that, from the whites' point of view, it was relatively isolated and hard to access," Tveskov said. "Yet it was very strategic. It was also in the home territory of Chief George and his people, who were from this section of the Rogue."

"They were looking for a place to hole up, a place where they weren't likely to be found," Truwe said.

The telltale wisps of their cooking fires were spotted by the troops down in the valley. The largely untrained militia began running toward the distant ridge.

"Imagine, they see the Indians way the hell over here," Truwe said. "So what did they do? They ran down the mountain, across the bottom of the valley and up the ridge to attack.

"The one united feature we find in the written accounts is how exhausted they were when they got here," he continued. "The charge was something the militia started, and the dragoons had to follow along."

Tveskov estimated the soldiers and militia climbed some 1,500 feet in elevation in the headlong rush. The attackers were strung out as they climbed up the ridge.

"Instead of doing a flanking movement as planned by the Army, they all charged," he said.

The Indians were well-prepared to repulse the attackers when they arrived. The two forces engaged at about 10 a.m. that day with the battle continuing until early Halloween evening, he said.

"This whole hillside has one of the richest concentrations of artifacts," Tveskov said as they walked along the knob where the fire left a thin layer of ash on the ground.

"The artifacts we find here are very reminiscent of those we found at Fort Lane where American soldiers were hanging out," he added. "Belt buckles, musket balls, melted lead slag. Of course, it was minus things like alcohol bottles and nails that we found at the fort."

As the two trekked along the ridge, they envisioned how the battle unfolded. The Indians formed a rear guard as they moved along the ridge to the west, they said.

"They defended their flank while they retreated their folks westward," Tveskov said. "Through the whole day of the 31st, the Army spent trying to either charge down this ridgeline after them or trying to flank the Indians."

"The whites came to hold the high ground, but this was a bare knob," Truwe said. "And the Indians were pinning the whites down on the high ground."

Although the likely male Indian leader in the Hungry Hill battle appears to have been Chief George, there are three historic accounts of a native woman named Indian Mary who led some warriors in the clash, Tveskov said.

The woman — not the same one for whom the county park was named in Josephine County — was the wife of Chief Jim, he said. She was described by both Joseph Lane and Joe Palmer, the local Indian agent, as a very strong leader, Tveskov said.

"The idea of her being a very powerful speaker is reflected in other accounts as well," he said.

The story of an Indian woman leading them in battle could have also been part of the romanticized myth that rose since the conflict, he said.

During the initial assault on the bare knob, the attackers had to pass the body of Gillespie, the militiaman killed early in the battle.

A letter written shortly afterwards by Lt. Kautz indicated the sight of Gillespie's cadaver caused many militiamen to immediately lose whatever remaining enthusiasm they had for a fight, Tveskov said.

"There was this sudden realization they could get killed," he said. "That took the wind out of their sails. I think they came into it believing it would be a cake walk."

He walked farther down the ridge to a site about 100 yards west and slightly downhill from the knob.

"This was the Indians' defensive position when they started retreating," he said. "Archaeologically, this is where we start to see fired .69-caliber shot. These were fired by the soldiers against this position.

"That archaeological signature continues for another 100 meters or so," he added. "My interpretation is the Indians were displacing backwards while defending this ridgeline."

It was the likely spot where an Indian sharpshooter used a hollow tree for protection while picking off the soldiers and militia, he said.

"He caused most of the casualties on the other side," he said. "Keep in mind the tactical goal of the Indians was to protect the retreat of their women and children back to the west. This was the position they could best do that from."

This was also the spot where more of the .69 caliber musket balls were found, shot from an 1847 Springfield musketoon, the same rifle used by soldiers at Fort Lane, Tveskov said.

Although he has yet to find an account of an arrow being shot in the battle, that doesn't mean they were not used, he said.

However, the Indians were well-armed with firearms and knew how to use them, he stressed.

"One of the bones of contention between the white factions at the time was accusations that the settlers had sold all their good weapons to the Indians," Tveskov said. "Some people felt that was one of the reasons they were able to fight so effectively."

"It was a common complaint in the letters from the time that the Indians were better armed than the whites," Truwe said.

"At least one military officer said one of the reasons they were so unsuccessful in this battle was that the Indian weapons had a longer reach than the whites' did," Tveskov added.

The .69 caliber musketoon was disliked by the soldiers because of its short range and difficulty in reloading, he said.

"But we have also found some .44 caliber rounds — the dragoons had .44 caliber Colt Navy revolvers," Tveskov said. "And we found some other calibers we haven't sorted out yet."

For all the evidence they have found, there are still mysteries to solve.

For instance, they wonder what happened to the copper alloy percussion caps used in the firearms of the day.

"That's a real bias in our recovery," Tveskov said. "There must be a lot here but we've found just one. Fired percussion caps would be a real indicator of someone's position where they were firing."

Perhaps they could have been dissolved by the chemistry in the soil, Truwe volunteered.

What they know for sure is the soldiers and the militia retreated back down the ridge on Halloween evening to a spring which had earlier been used by the Indians.

After leaving the battlefield, the Indians continued working their way farther west. The next battle would be at a place known as the Little Meadows, followed by one at the Big Meadows and finally at Big Bend on the Rogue River the following year.

"Each time, the soldiers would follow them to a point where there would be a fight, then the Indians would retreat farther," Tveskov said. "But after the battle of Hungry Hill, the Indians didn't do as well. They were finally defeated at the battle of Big Bend."

That battle in June 1856 ended the Rogue River Indian War of 1855-56.

Reach reporter Paul Fattig at 541-776-4496 or pfattig@mailtribune.com.