I drive into the Driftless in the dark. The straight, flat highway in southeastern Minnesota suddenly drops into a winding set of switchbacks, left and then right, down and then up. Here the Driftless is like an inverted mountain range, sunk into the earth instead of rising from it, and from the edge, you can see flat farm country right across the top. But before you drops a slashing valley, dark with thick hardwood and pine forest, bounded by bluffs and knife-like ridges. At the bottom of each slope, tumbling forever downhill toward the Mississippi River, is a stream.

And in that stream are brook trout, I think to myself, as I drive through Lanesboro in the early morning to the DNR fisheries office. Or at least there should be. Used to be.

The brook trout, Salvelinus fontinalis, is Minnesota's only native stream trout. In the small, cold streams where it lives, the brookie is a predator of smaller fish and invertebrates, and it in turn is preyed upon by birds, otters, and turtles. It's loved by anglers as a willing biter and fighter, tasty in the skillet but often released because of its small size and startling beauty. A black and olive maze drapes across its back, unique to each fish, and fades down its side in a silver gradient, speckled with red, yellow, and blue halos. Pure white and black leading edges set off flame-orange fins.

The angler holding a brook trout feels like he has stumbled upon a secret, artistry transcending natural selection, like a jeweler who cuts into a stone and finds otherworldly light and color inside.

A diamond, though, is hard, indestructible. The brookie is delicate. It needs cold, clean water and abundant insect life to eat. It needs in-stream debris to hide behind and upwelling gravelly springs to spawn over.

The Driftless Area precisely meets these needs. The bedrock here is close to the surface, and the region wasn't bulldozed by the Wisconsin Glaciation that buried the rest of the Midwest in gravel, sand, and clay. Rainwater soaks into the rock and percolates down to ancient aquifers. On the valley floors, this cold, mineral-supercharged groundwater burbles from springs, forming the headwaters of thousands of miles of just-right-for-brook-trout streams.

But in the 1800s, settlers cleared, plowed, and burned the land to grow crops and graze cattle. Rampant erosion left great gulley-wounds and smothered the valleys in sediment. Runoff happened so fast that streams either trickled or flooded, and were forced into muddy trenches with high, slumping banks. By 1900, there were few, if any, wild brook trout left.

"Every day I show up to work and I see the effects of mankind on these streams," Travis Viker, fisheries technician at the DNR Lanesboro office, says from the back seat of the truck. "It's impossible not to, even on some of the more preserved brookie streams."

Vaughn Snook, the assistant fisheries supervisor, is driving. We wind down into the Root River drainage, past old farmsteads, fields of standing corn, and pastures with winding streams and grazing dairy cattle.

Snook pulls the truck over. I follow Snook and Viker into the thick brush—me carefully threading my fly rod through the tangle—which opens into a small valley of dappled green, with the stream meandering downslope toward us. It's a trickle, only a couple of feet wide in most spots. Just ahead is a waist-deep pool, walled on one side by a mossy, seeping cliff face. The water is clear and gray-green.

We can see the effects of mankind that Viker spoke of. Old earthworks and culverts impede springwater inputs, and patches of invasive garlic mustard and burdock dot the banks. At outside bends, the stream is eroding fragile mounds of topsoil that used to be on top of the ridges, but now slump into the stream with every rainfall, smothering gravel and filling in pools. Still, Coolridge Creek is beautiful. It might not look like it did 200 years ago, or be as just-right for brook trout, but it looks a lot better than it did 100 years ago.

Starting in the 1930s, water- and land-use practices improved. The DNR and conservation groups began restoring streams, stabilizing banks, and adding trout-friendly structure. Erosion slowed, base flows increased, and the streams got cleaner, colder, and more stable. Unfortunately for the brookies, the European brown trout, introduced in the late 1800s, is more tolerant of degraded habitat and got a head start. They flourished as the streams healed. Most habitat work and DNR management since then has focused on the browns, which are loved by anglers—myself included—and they still dominate most Driftless streams.

If brook trout parents were to tell scary stories to their fingerlings, Salmo trutta would be the boogeyman. Brown trout get much bigger than brookies, and they eat them or push them out of prime resting and feeding areas. They can thrive in areas that are marginal but seasonally critical for brook trout.

Today, most Driftless brook trout are in the colder and smaller headwaters streams, like Coolridge Creek. These refuge areas have allowed the brook trout to survive, and it turns out they're doing quite well.

It wasn't until the 1970s that the DNR began brook trout restoration efforts. Back then, a brook trout was a brook trout, and fisheries workers sourced them from hatcheries as far away as Massachusetts, New Hampshire, and West Virginia, and stocked them here for decades.

But today, the trout in Coolridge Creek and others across the region don't match the genetics of those hatchery strains. In fact, geographically isolated populations around the region seem to match each other, and even resemble native fish from Iowa and Wisconsin Driftless streams.

This was uncovered in 2015 when the DNR conducted a genetic study to map out the brook trout of the Minnesota Driftless. The goal was to learn about recent ancestry of southeastern Minnesota brook trout and, maybe, their origins. They found that brook trout are now in almost 70 percent of southeastern Minnesota's streams—up from about 50 percent in the 1990s, and from just a handful of streams in the 1970s. They also found that some of these trout might just be carrying remnant native Minnesota brook trout genetics. The DNR is cautiously calling these fish heritage brook trout, and Coolridge Creek is one of their strongholds.

We can see them at the head of the pool, slipping side to side in the current, shadowy with an occasional flash of spots or white fin-edges. They're small, five inches long or so. Snook explains that they're 2-year-olds at that length, probably already parents.

"Lucky for us," he says, "they reproduce like fruit flies."

I make a few casts, but they don't even look at my little grasshopper imitation. We work upstream and try different flies as we go, but the trout are not biting.

Just like the brook trout I've fished for up north of Lake Superior, or in Michigan. They look like regular brook trout to me. I ask Snook, Why do we think these ones are special?

"It's like those cheek-swab DNA tests you can do," Snook says. They can't tell you who your parents are, he explains, unless your parents also happened to take the test. But they can tell you that you match other people whose parents' parents were from a certain corner of the world.

As it turns out, native is a very strong word. We like to think that DNA testing leads to quick, certain answers, but when it comes to tracing ancestry, it's more complicated than that. Especially since we don't have any genetic samples from for-sure native Minnesota fish, which were last seen—for sure—in the early 1900s.

I ask Loren Miller, fisheries research scientist for the DNR and the University of Minnesota and an author of the 2015 study, about the use of the word heritage instead of native. You can't trace a fish's lineage directly up the family tree, he says. Instead, by looking for specific genetic markers, you can place a fish within known regional groupings—or find that it doesn't fit anywhere, which can be just as informative.

"There's still uncertainty," Miller tells me about the heritage trout in Coolridge Creek. "But we found no association with the known hatchery strains."

And this is why Miller, Snook, and Viker aren't hung up on "heritage." The true nature of these fish—whether they are actually the descendants of Minnesota brook trout that evolved here—isn't really the point. The point is that these fish are genetically distinct, despite a century of mixing with known eastern fish.

"They're persisting clean in a genetic sense," Miller says, which suggests that the heritage trout do better here in the Minnesota Driftless streams than other strains of brook trout. If that's true, they could be an important fishery management tool for the DNR in the future, in the face of a warming climate, changing land-use patterns, and other challenges.

And the timing is good. After a 2015 disease outbreak at the Crystal Springs hatchery, the DNR has to restart its brook trout hatchery program from scratch. A native strain would be ideal, and a heritage strain that's likely native and proven successful would be the next best thing. These trout would be useful both for new reintroductions in streams that have only brown trout, and for improving or supporting existing brook trout populations.

"For a stream that already has eastern fish in it," Snook says, "why not throw heritage fish on top of it and see what happens?" Over time, the genetics of the brook trout in the stream could shift toward heritage.

"You might never get to 100 percent heritage fish," Viker says. "But just approaching it would be cool."

The Crystal Springs hatchery and the heritage trout will be tested for diseases for a few years before any new hatchery population can be created. In the meantime, Snook and Viker are keeping an eye on the trout and helping where they can.

On Coolridge, they removed the browns altogether and installed an "immigration barrier" downstream to keep them out. But the browns are tenacious: Even with the four-foot barrier and four electroshock removals per year for six years, a few browns still turn up every year. We saw one today, a 14-incher, looking outsized like a wolf in the flock at the bottom of a deep pool.

It's evening now, and photographer Mike Dvorak and I are fishing our way up another stream—this one I won't name—that holds heritage brook trout. It's also got plenty of browns, and so far, that's all we've caught. We take turns at each bend in the stream, which runs through a private dairy pasture with a DNR angler-access easement. It's a warm night, and caddisflies and mayflies flash all around us in the low sunset.

Dvorak is a good caster. He kneels at the edge of the bank and lays his fly where a one-foot-wide chute of fast water has scoured a neat cut beneath the grassy bank. A triangular snout appears and the fly is gone. Dvorak stays on his knees while playing the fish, which has nowhere to go, and nets it neatly. At last, it's a brook trout, nearly a foot long. Old, relatively—maybe three years.

Maybe older. Maybe 10,000 years old. It's a male, with a deep jaw and a scarred nose, that looks almost black in the fading light. Does the trout care if he's swimming in his ancestral water? That he sometimes chokes on sediment that should be 200 feet up the hill? That he belongs here, more than the cows, the brown trout, than us?

Well, we care. In our lives, Dvorak and I have caught a lot of brook trout, but few of them could be argued to be native, in their native water. Right now, it seems to matter. Dvorak cradles the trout in the cold springwater with one open hand, and it fins free, back to its undercut.