There's a half acre in the southernmost corner of the Floodplain Agricultural Basin where the Tifton just stops. In its place is a solid stand of one plant — waist-high, packed in edge to edge, ending right where the ground rises and the grass comes back. Twenty years of hay never planted it. The water did.

We keyed it out to common cocklebur (Xanthium strumarium), an annual that specializes in colonizing bare mud. The identification got confirmed from an unexpected direction: those yellow-ringed spots on the leaves are cocklebur rust, a fungus that infects almost nothing else. If the rust thinks it's cocklebur, it's cocklebur.

A stand that dense tells us the low spot ponded after the spring rains and held water long enough to kill the bermudagrass. Cocklebur seed waits in the soil for exactly that kind of opening. When the mud came up bare, it took all of it.

What the patch is telling us

The stewardship plan already marked the low ground in the south of the property for wetland work. What we didn't have was a boundary. Now we do. The cocklebur traces the ponding zone edge to edge — where it grows, water stood this spring; where the bermuda held on, it didn't. We couldn't have mapped that line more precisely with instruments, and we didn't have to.

About the cocklebur: We're not adding it to the R&D Library, because we won't be growing it. It's an annual, generally treated as native, though botanists are still arguing about where it originally came from. Each bur holds two seeds with staggered dormancy that stay viable in the soil for years, and the seedlings and seeds are poisonous to livestock. So the patch gets mowed this month before the burs harden, and clipped high next spring when the seed bank flushes.

What goes in the pocket

The plan is to plant the pocket the way the water dictates, zone by zone: the spot that ponds longest, the ring that floods and dries, the shaded edge along the pecans, and the rim. Plugs and dormant woody stock, not the seed drill. Here's the palette.

The wetland pocket palette

• Sand Spikerush (Eleocharis montevidensis) · Core · Spikerush
• Three-Square Bulrush (Schoenoplectus pungens) · Core · Bulrush · conditional
• Littletooth Sedge (Carex microdonta) · Ring · Sedge
• Grass-Leaved Rush (Juncus marginatus) · Ring · Rush
• Frogfruit (Phyla nodiflora) · Ring · Groundcover
• Common Sneezeweed (Helenium autumnale) · Ring · Forb
• Bushy Bluestem (Andropogon glomeratus) · Ring · Grass
• Texas Meadow Sedge (Carex perdentata) · Edge · Sedge
• Cherokee Sedge (Carex cherokeensis) · Edge · Sedge · trial
• Buttonbush (Cephalanthus occidentalis) · Rim · Shrub
• Bald Cypress (Taxodium distichum) · Rim · Tree

The full photo plate is on the site.

Picking this list took more range-checking than we expected. Our rule for the R&D Library is documented records within about forty miles of this ground, on the same limestone — the standard that already retired muscadine grape from our species list. Most of the palette clears it. Two species we originally wanted didn't, and the plan changed. We dropped Torrey's rush after the Wildflower Center's notes rated it intolerant of high-carbonate soils, which is the only kind we have, and put sand spikerush in its slot — same job, and it grows on limestone a county away.

Cherokee sedge is the exception we kept. NPSOT maps it to the country east of us, and the nearest record we can find is about forty miles away at Fort Cavazos. But it's the wet-meadow sedge nurseries actually sell, and the sedges with local records are hard to buy. So it goes in labeled as a trial, planted next to the local sedges so we can compare what establishes. We may be wrong about it. We'll find out.

Two placement details matter for the woody rim. Bald cypress seedlings drown if their crowns go underwater — mature trees stand in it indefinitely, but young ones establish on the drawdown line, so ours go on the rim rather than down in the bottom. And they stay out from under the pecans, because cypress wants full sun. Long term, that grove is the interesting part. Grown out along its wet flank, the pocket and the pecans could add up to about an acre and a half of pecan-and-cypress seasonal wetland. That's a couple of phases away. It's on the map now.

The calendar

• August, before the burs harden — mow the patch. Every bur that matures now is weeding we'll be doing for years in ground we're about to plant.
• Mid-September — the second glyphosate pass under the PUB program takes the regrowth. Cocklebur is highly susceptible.
• October–November — sedge, rush and spikerush plugs go in on the drawdown, with frogfruit sprigged between them. Not into standing water — plugs set in saturated soil rot before they root.
• December–February — dormant buttonbush and bald cypress on the rim, buttonbush cuttings to the propagation bench, and bushy bluestem seed in with the winter dormant sowing.
• March — the seed bank flushes and the cocklebur comes back. A high clip at eight to twelve inches knocks it down without touching the natives coming up under it.

There's a real unknown under all of this: we don't actually know how long the pond holds. This is the first year we've watched it, which is why the bulrush decision waits and why some of this plan will probably turn out wrong. When it does, we'll write down which parts.

Everything about this restoration eventually comes back to the soil. So before we drill the native seed and before we strike the first trays on the propagation bench, we spent a day with buckets, bags and a sharpshooter figuring out what the ground actually is. Not what the county soil survey says it should be — what it is, section by section, the year we start.

Three reasons to dig

• The record — soil tests are part of our NRCS EQIP obligations. The conservation practices we signed up for start from a documented baseline, and the lab results go into the same public file as everything else here.
• The seeding — the native grass-and-forb mix is going into ground that's carried fertilized hay for twenty years. Knowing the texture, pH and nutrient load across the sections tells us what the seed is walking into, and where the drill plan needs adjusting.
• The bench — we gathered soil as well as sampled it. Screened native soil from the tract goes into the propagation unit's mix, so the seedlings and cuttings we raise start in the same dirt they'll get planted back into.

How the day went

We pulled cores across the working sections and composited them by area, then bagged and labeled everything against the section atlas — samples for the lab in one set of bags, gathered soil for the bench in buckets. It's unglamorous work in July heat, and it's about the cheapest insurance the whole project buys: a seed mix or an amendment chosen against real numbers instead of a guess.

Local dirt, local plants: Raising restoration stock in store-bought potting media and then planting it into Lampasas County clay and sand asks every seedling to change worlds twice. Start them in screened native soil — cut with just enough drainage material to behave in a tray — and the roots meet their real home, biology and all, from day one.

What happens next

The samples go to the Texas A&M AgriLife Extension Soil, Water and Forage Testing Laboratory in College Station — the standard bench for Texas land work — and the numbers come back here. They'll shape the amendment plan for the converted acres, tune the drill mix where a section runs thin or alkaline, and set the recipe for the bench's first potting-mix trials. When the reports land, we'll post them in this journal with everything else.

Any honest restoration starts by looking hard at what you actually have. What we've got is a tired piece of Central Texas: a hay meadow that's been run as improved pasture for decades, river frontage thinned out by grazing, and uplands where a few tough introduced grasses have crowded out almost everything else. It's not ruined. But it's a long way from what this ground could be, and closing that gap is the whole project.

The land's been in the family four generations — cleared off cedar and farmed for seed wheat in the '40s, grazed for decades, then held in Tifton 85 hay for the last twenty years. Each of those was the job the land got asked to do at the time. That's what we inherited. Here's what it needs now.

The core problem: a monoculture that works too well

Most of the working acres are Tifton 85 bermudagrass, with King Ranch bluestem and Johnsongrass filling in the edges. They're excellent forage and stubbornly good at their job, and that's exactly the problem. They form dense, single-species mats that shade out native seedlings, give almost nothing to grassland birds or pollinators, and leave the soil biology a lot simpler than a native prairie's. A field of Tifton 85 is green, productive, and nearly silent.

The fixes, in order

Restoration here is a sequence, not one move. First, convert the exotic pasture back to native prairie under the Texas Parks & Wildlife Pastures for Upland Birds program — knock down the introduced grasses, then no-till drill a native grass-and-forb mix built for this ecoregion. Second, pull the river frontage out of production as a conservation strip and let a riparian buffer come back. Third, build the woody structure back — a shelterbelt on the high ground and scattered mottes — grown out on our own propagation bench so the stock is local and cheap. Each fix has its own section on the tract and its own care class.

Why native, why now: Native prairie isn't nostalgia. Deep-rooted natives hold the soil, take rain into the ground instead of running it off, and feed the insects that feed the birds. And once they're established they cost far less to keep than a pasture you have to mow, fertilize and spray to hold in one species.

The problems we already see coming

We'd rather name the risks than pretend they aren't there. Reinvasion is the first — bermudagrass and Johnsongrass don't quit after a single pass, and the first two or three years are going to be a spot-treatment fight. Establishment is the second: native seed is slow, and a dry fall after drilling can cost us a whole planting. Water's the third — new plantings need it before their roots go deep, which is why a watering trailer is on the build list. And the last one is just time and continuity. This isn't a project that pays off in a growing season.

The honest horizon

The seed we drill won't look like prairie for years. Realistically the mix needs five to ten years to knit into something that carries itself, and the opening arc of the whole thing — conversion, riparian buffer, shelterbelt, first woody plantings — is a fifteen-year job. After that it's not "done," just handed forward: grassland is a disturbance ecosystem, and it wants fire and grazing and attention for as long as anyone keeps it. We're planting things we won't see finished. We're fine with that.

Last week, in the last hot days of June, the final full cutting of Tifton 85 bermudagrass came off the meadow. It was a strange thing to stand next to — good hay, cut clean, and the last of its kind we mean to make here. With the bales off, the conversion can start.

Over the next window we're going to treat roughly 45% of the property: knock down the introduced grasses that run it — Tifton 85 bermudagrass, King Ranch bluestem, Johnsongrass — and get a clean seedbed ready for a native mix. This is the single biggest thing the restoration will do, so we want to be straight about exactly what it involves and why.

We're not converting everything at once. Roughly 45% stays in Tifton 85 for now — kept in hay until the native replanting proves itself — and we'll do our best to hold clean breaks between the standing Tifton and the new native stands. That retained hay brings in a small but steady income that helps pay for the restoration while the natives get going.

The treatment

The knockdown is a glyphosate application, mixed for the job. Away from water we use a standard terrestrial formulation; anywhere near the river and its drainages we switch to an aquatic-approved one, labeled for use over and around water. Here's exactly what goes in the tank.

Knockdown mix
Carrier — Water
Glyphosate · upland — RoundUp PowerMax 3
Glyphosate · near water — Aquaneat (aquatic-approved, in place of PowerMax 3)
Surfactant — Brimstone
Deposition / drift-reduction aid — Novita DownForce
Application — Drone-applied by a licensed applicator (Agrotech of Lampasas)

We're putting it down by drone instead of a tractor-pulled boom, and that buys us three things: much more targeted, accurate coverage; the ability to work around individual trees and mottes we want to keep; and reach into the hard places — the brushy, uneven ground along the river where a tractor and sprayer just can't go.

Applied by: The burndown and native reseeding are done by Agrotech of Lampasas (agrotechtexas.com), a local licensed applicator — our partner on the grassland conversion.

Two passes, then never again

We're not casual about spraying, and we've got no intention of farming with herbicide. The plan is a limited, deliberate knockdown — at most two passes to burn through the bermudagrass and Johnsongrass rhizomes — and then no broadcast glyphosate on these acres again. Once the natives are drilled, management shifts for good to the tools a prairie actually wants: prescribed fire, timed grazing, and spot-treating the individual reinvaders by hand.

Why glyphosate, and why it’s the right call here

Tifton 85 is a sterile hybrid bermudagrass that spreads by dense rhizomes and stolons, not seed — and it's a bastard to kill. At field scale there's no mechanical way to get rid of it that doesn't also wreck the soil structure we're trying to protect; tilling that fine would just invite erosion on our slopes and along the river. A targeted chemical knockdown followed by no-till drilling is the standard restoration practice, and it's exactly the method Texas Parks & Wildlife specifies under Pastures for Upland Birds: the right herbicide to kill the exotic grasses, then no-till seeding of natives.

Glyphosate is the right chemistry for a one-time conversion. It's non-selective, so it takes the bermuda, KR bluestem and Johnsongrass all together. It binds tightly to soil particles and gets broken down by soil microbes instead of hanging around or moving through most soils. And it leaves no residual soil activity, so native seed can germinate weeks later without a problem. Near the river we don't touch the standard product — the aquatic-labeled formulation (Aquaneat) is made and approved for exactly this situation, without the surfactant that makes terrestrial products unsafe over water.

Our standard: Two knockdowns to reset the field, an aquatic-safe product anywhere near the Lampasas, a licensed applicator, then a permanent switch to fire and grazing. If we can't explain a treatment in plain language, we don't do it.

What goes back down

The seedbed gets drilled with a native grass-and-forb mix Texas Parks & Wildlife designed for this ecoregion under PUB, sourced from Bamert Seed Company in Muleshoe, Texas. Deep-rooted bunchgrasses for structure and soil, plus a broad set of wildflowers and legumes for the pollinators and birds. The full mix is below.

The drill mix

• Little Bluestem “OK Select” (Schizachyrium scoparium) · 15%
• Big Bluestem “Earl” (Andropogon gerardii) · 10%
• Switchgrass “Blackwell” (Panicum virgatum) · 10%
• Indiangrass “Lometa” (Sorghastrum nutans) · 10%
• Illinois Bundleflower “Sabine” (Desmanthus illinoensis) · 6%
• Sideoats Grama “El Reno” (Bouteloua curtipendula) · 5%
• American Basketflower (Centaurea americana) · 5%
• Engelmann Daisy “Eldorado” (Engelmannia peristenia) · 5%
• Maximilian Sunflower (Helianthus maximiliani) · 5%
• Dotted Gayfeather (Liatris punctata) · 5%
• Blackeyed Susan (Rudbeckia hirta) · 5%
• Eastern Gamagrass “Pete” (Tripsacum dactyloides) · 5%
• Buffalograss “Texoka” (Bouteloua dactyloides) · 2%
• Canada Wildrye “Lavaca” (Elymus canadensis) · 2%
• Virginia Wildrye (Elymus virginicus) · 2%
• Vine Mesquite (Panicum obtusum) · 2%
• White Tridens “Guadalupe” (Tridens albescens) · 2%
• Mexican Hat “VNS” (Ratibida columnifera) · 2%
• Purple Prairie Clover “Cuero” (Dalea purpurea) · 1%
• Prairie Tea (Croton monanthogynus) · 1%

The full photo plate is on the site.

The long view

None of this looks like much for a while. A drilled field looks like bare dirt, then like weeds, and only years later like prairie — realistically five to ten years before the natives take over and the stand carries itself. The conversion is one move in a fifteen-year opening arc, and the grassland it makes will need fire and care long after that. The last harvest was an ending. This is the slow start of the next thing.

The first Propagation Unit is a three-shelf bench that raises native restoration stock from seed and cutting with as little daily babysitting as the work allows. It's deliberately small — one rack, off-the-shelf parts, and a controller anyone could rebuild in a weekend. The point was never to build something clever. It was to see how much of a nursery routine you can hand off to a cheap sensor and a valve, so the crew can spend its hours out in the sections instead of standing at a hose.

What it must do

Hold each shelf inside a target moisture band. Run the lights on a per-shelf photoperiod. Move air with a fan. Water a shelf only when it drops below its floor, and never more often than its cooldown allows. And log every reading and every cycle, so a season can be reviewed instead of just remembered.

• Rack — three growing shelves, 01 through 03, on a steel wire frame
• Controller — a single-board Piconium reading moisture, temperature and reservoir level
• Water — 120 L reservoir, diaphragm pump, one solenoid valve per shelf
• Light — full-spectrum bars on a per-shelf schedule
• Air — USB circulation fans on the rack
• Interface — a 5″ Waveshare touchscreen at the rack and a web dashboard

On the bench: The unit is instrumented, not autonomous. A person still decides what grows where and signs off on any change. The software just keeps the shelves honest between visits.

Bill of materials

The full parts list, with the engineering notes, lives in the repo at docs/HARDWARE.md. The short version:

• Compute — a Piconium single-board computer (x86, 4 GB RAM) running Debian
• Display — Waveshare 5″ 800×480 touchscreen (HDMI + USB touch)
• Relays — Numato 8-channel USB relay module
• Serial — an FTDI RS485 adapter (FT232RQ + SP485EN)
• Grow lights — Barina full-spectrum bars, 120 VAC
• Fans — 12 V USB circulation fans
• Pump — a SEAFLO 12 VDC diaphragm pump
• Valves — four U.S. Solid 12 VDC solenoid valves, normally closed (they fail shut on power loss)
• Soil probes — RS485 multi-parameter soil probes (moisture, temperature)
• Reservoir sensor — a submersible hydrostatic level transmitter (RS485)
• Reservoir — a 120 L tank
• Rack — a three-shelf steel wire frame

The interface is public. You can open the same rack screen and dashboard the crew uses over on Instruments & Infrastructure — in observation mode, so you can click through every screen without touching the live bench.

The code is open source too — the full build, controller and dashboard live at github.com/Short-s-Resort-Field-Station/propagation-controller.

We get help paying for this, and we want to be straight about how. Two agreements cover a good chunk of the work right now. The first is cost-share through the USDA Natural Resources Conservation Service under the Environmental Quality Incentives Program — NRCS EQIP — run through the Lampasas County field office. The second is a Texas Parks & Wildlife grant under Pastures for Upland Birds (PUB), which helps landowners turn exotic-grass pasture back into native prairie for grassland birds.

What the support covers

The NRCS EQIP cost-share is tied to specific, named conservation practices — riparian buffer establishment, brush management, native seeding, and a windbreak along the eastern boundary. Under PUB, Texas Parks & Wildlife supplies the herbicide for the knockdown, the native grass-and-forb seed mix, and technical guidance for converting our Tifton 85 hayfield to native prairie. Neither one covers the cabin, and neither goes to anyone as a salary.

Why say so: Public money should leave a public trail. Practices, acreages and outcomes get reported here as they're finished, in the same plain format as everything else on this station.

If you're thinking about similar work on your own ground, both programs are worth a conversation with your local NRCS office and your TPWD biologist. The paperwork is real, but the cost-share is what makes a project this size possible for a family holding.

It starts with a cabin above the Lampasas River. Our great-grandfather — everybody called him Short — got this ground in the 1940s, after the federal government took his wheat farm in Killeen to build what became Fort Hood. He grew seed wheat here for years, and in the early '60s he and his kids built the cabin down by the river. The family actually lived across the road on separate land, running goats and sheep and cutting hay. The cabin was the other thing — a place to cool off in the river after a day in the fields. They called it "Short's Resort," half joking, and our great-aunts painted the sign by hand.

No running water, no electricity, and four generations later it's still standing. We're the ones keeping it now, and we still camp there most times we come down to work — which is where a lot of this restoration actually gets planned, over a fire with the river going by. The name's a joke that stuck. There's no resort. There's a cabin, a river, and seven sections of ground that need work. But it says something we wanted to keep: this is a place people are welcome, and the work of fixing it up ought to be welcoming too — easy to follow, written down, and free for anyone to copy.

A field journal, not a blog

So that's what this is. A field journal. What we try, what we learn, how we keep track of it, and the problems we run into on the way. The sections are numbered, the species are recorded, the instruments are logged, the plans are open. There's an institution behind it — small, mostly one family — but we run it like one on purpose, because that's what keeps the work going past any single season or any one set of hands.

A note on names: You won't find anyone's name here. The station speaks for the work, not for people — partly for privacy, and partly because the land doesn't care who planted the tree, only that it got planted right.