Skull Island sits in Massacre Bay, in Washington's San Juan archipelago. Here, in 1858, Haida raiders killed a band of Coast Salish and left the bones behind. I can think of other, perhaps more cheery spots to look for flowers, but Madrona Murphy's enthusiasm is unstanched.

"Look!" she calls as our boat nudges against shore. "There's some camas right over there." She hikes over the rocks along the beach toward some large blue flowers. Her bright orange safety vest must be visible from Canada, but she wants to appear official: The federal Bureau of Land Management manages Skull Island, and few regulations govern its use. "I don't want someone to see a lot of people crawling around here and get ideas," she says.

In a square-meter patch of the camas, she measures the tallest plant, which stands over two feet, then counts other plant species, and records the distance to a band of orange lichen along the shore. The lichen marks the reach of sea spray, which will likely climb if oceans rise as projected; camas, she explains, is not terribly salt tolerant.

Murphy, 31, is a botanist with the Kwiáht Center for Historical Ecology of the Salish Sea, based on nearby Lopez Island, her home. Since 2007, she has helped survey the 1,000 or so BLM acres patchworked throughout the San Juans for biological and cultural resources. With Russel Barsh, the director of the Kwiáht Center, and Nick Teague, the local BLM ranger, she notes the spread of invasive species like blackberry and counts insects, all with an eye to an eventual management plan.

But she keeps a special watch out for camas. For the Coast Salish of the San Juans, it was a dietary staple. Tribes cultivated it in large gardens, subdivided into family-owned plots passed down through generations. These were fertilized with seaweed, cleared of weeds and stones, and burned to control brush and grass. With European settlement, though, the potato became the tribe's tuber of choice. The gardens were swallowed by farms and developments, and camas faded into memory, just one among many beautiful wildflowers.

To Murphy, it is the botanical equivalent of deposed royalty, and it is time it was restored to its former station.

On Skull Island, Murphy points out camas, chocolate lily and wild onion, all of which the Coast Salish grew for food, though she doubts they were actively cultivated here. The patches are too random, and food plants have a way of hitchhiking hither and yon. We spot a small pear tree. "Probably someone had a picnic and tossed the core away," she says.

As she walks, her face brightens at the sight of some uncommon inch-tall plant with a tiny flower within a field of what to me seem its identical, far more common relatives. She became interested in plants as a child, she says, and stopped paying attention to anything else. She seems to know the Latin binomial of every plant she sees, as well as its associated lore. "Ethnobotanists use people to ask questions about plants," she says. "I use plants to figure out what people are doing."

Until recently, camas was thought rare in the San Juans. That may have been because people were looking in the wrong places. Ethnographic evidence suggests that the Coast Salish used small islands for camas gardening. It would have been easier to control pests there -- mainly deer -- and to weed out death camas, the bulbs of which look like camas but are decidedly inedible. Also, camas bulbs would have been easier to dig from the islands' thinner soils. (Camas has a contractile root. "It can literally crawl under rocks," Murphy says. "Wild harvest would be incredibly frustrating.")

Murphy has surveyed more than 40 uninhabited islands and found camas on over 50 percent of the vegetated ones. Most are hidden in small, wild patches in relatively inaccessible spots, but three sites she thinks were once gardens. Those especially interest Murphy, who wants to return camas to its old haunts, and perhaps cultivate it. The Coast Salish cooked the bulbs for a day and a half over hot coals, in pits covered with grass and dirt. They could be eaten immediately, used as a sweetener, or mashed into cakes and dried. "As food plants," Murphy says, "they would be really well-adapted: They don't need irrigation, they don't have pests." She herself has never eaten the bulbs -- the ones she keeps in a seed bank are too valuable -- but Barsh has. He says they taste like carmelized onion, but without the onion flavor –– just the sweet.

At day's end, Murphy takes me to a known relic garden she found five years ago, on Blind Island. It looks like most of the other patches of camas we've seen, but she points out a spot under an aggressive sprawl of blackberry where the turf has been pulled back, perhaps by otters. "Look at the soil," she says. "See how black it is." That means that charcoal was mixed in. The soil is also unstratified -- another sign of caretaking. She picks some up and rubs it in her hands. The earth stains her fingers.

Although the physical evidence goes back only 200 years, Murphy suspects that Native Americans cultivated camas for over 2,000. "The Coast Salish are frustrating to archaeologists," she says. "A lot of their culture was very compostable, so there aren't many beautiful artifacts." Instead, there is a rich biological record of their diet: middens stippled with shell and bone that sit under picnic tables at state parks, or meadows of wild onion growing beneath young pear trees, or relic camas gardens, tucked in shadow.

Murphy sets about her survey. The camas here is tall, and buzzing with bees. "We're here at the peak of blooming," she says. "Look at how the bees are almost pushing each other off the stamens, they're so excited." We watch for a few minutes as they smear themselves with pollen. Then we get back in the boat, leaving the garden to its newest tenders.