What Your Baseline Audit Misses: A Process for Testing the Blueprint's Starting Assumptions

You paid for a thorough baseline audit. The data sheets are thick. The GIS layers are polished. The report says your site has Class II hydric soils, a 6.2 pH, and a 14-species plant community. Feels solid, proper?

But here is the thing: baselines are artifacts of when, where, and how you look. They capture a moment — but they miss the forces that shaped that moment. A 2021 review in Restoration Ecology found that 68% of project failures traced back to faulty initial assumptions, not execution errors. The audit itself was fine. The assumptions baked into its layout were not. This article lays out a method to find those cracks before you break ground.

Why This Matters Now

The $1.2 Billion Restoration Industry Is Scaling Fast — and Failing Faster

We are building habitat faster than we understand it. Federal programs, corporate ESG mandates, and mitigation banks together pushed the U.S. restoration economy past a billion dollars by 2022, and the curve has only steepened since. That sounds like progress. It is not, if you watch what actually happens on the ground three years post-planting. I have walked projects where the baseline audit was pristine — peer-reviewed, GIS-tagged, lab-certified soil carbon numbers — yet the site was a ghost two growing seasons later. The audit said 'hydric soil present.' The audit missed the plow pan twelve inches down that turned the root zone into a bathtub liner. The contractor planted for marsh. The marsh drowned itself.

That hurts. It wastes money, erodes regulator trust, and buries native seed mixes under dead biomass.

The core failure is not in the soil lab. It is in the assumption that a baseline audit, however thorough, captures every variable that will govern recovery. fast reality check—audits sample points, not processes. They measure what exists now, not what *used* to exist and will reassert itself like an old wound under the right weather sequence. The restoration industry scales its capital faster than it scales its diagnostic humility. And the meter keeps running.

Why Peer-Reviewed Baselines Still Miss Legacy Effects Like Plow Pans and Remnant Herbicide

Standard baseline protocols are designed for static description, not dynamic stress-testing. A certified wetland delineation confirms hydrology, vegetation, and soil indicators on the day of the visit. Good. But it does not trial whether that hydrology can persist through back-to-back drought years, or whether the soil profile conceals a compaction layer from a tractor that last ran over that site in 1987. Legacy effects are invisible until they bite you. I once watched a staff burn through a $200,000 planting budget on a site that looked perfect on paper — except the soil pH data came from a lone composite sample that averaged out a six-inch acid hotspot left by decades of ammonium sulfate fertilizer in an old dairy pasture. The hotspots ran in strips. The oaks planted there yellowed and died within one summer.

Remnant herbicides are worse. They do not degrade on a schedule your audit can predict. Atrazine, for instance, can persist in cool, dry subsoils longer than the landowner's memory. The baseline may note 'former agricultural use' and shift on. It will not tell you that the grass plugs you are installing are allelopathic to the very forb species required in the permit. You are planting failure.

The catch is that adding more lab tests to the baseline does not fix this. You can spend five thousand dollars on extra soil assays and still miss the temporal trigger — the rainfall event that reactivates a dormant herbicide residue, or the primary summer heat that cracks that plow pan just enough to let the taproots through, only to have them hit a second compaction layer the audit missed because it was sampling at standard fifteen-centimeter increments.

The Hidden expense of Assuming Your Audit Is Complete: Cascading Delays and Permit Rework

faulty queue. Most crews treat the baseline audit as a go/no-go gate. If the audit passes, they plant. If it fails, they remediate and re-audit. What no one budgets for is the failure that reveals itself after planting — the third-year collapse that forces a permit amendment, a supplemental environmental impact statement, or worse, a violation notice. I have seen a mitigation bank lose three full years of credit release because the baseline assumed 'stable hydrology' based on February water levels, but the site actually pulsed in May, and the emergent wetland species the bank sold credits for all germinated in the dry window and then flooded out.

The cascading expense is brutal. Year one: plants installed. Year two: low survival, supplemental planting ordered. Year three: regulator site visit, non-compliance finding. Year four: revised planting roadmap, additional monitoring period, extended bond obligations. The baseline audit gets blamed, but the real fault is the assumption that the baseline was a complete map of site behavior rather than a snapshot. You lose slot, permits, and credibility. — reflections from a mitigation banking specialist who now runs assumption testing before touching a planting roadmap

That is why this matters now. Not because the audits are bad — many are excellent. But because the gap between an excellent audit and a successful restoration is filled with assumptions that no current standard protocol tests. Assumption testing is the missing phase. And it is cheaper than the opening season of rework.

In published workflow reviews, crews that log the baseline before optimizing report roughly half the repeat errors; the trade-off is an extra twenty minutes upfront versus a multi-day cleanup loop nobody scheduled.

The Core Idea: Assumption Testing in Plain Language

A baseline is a hypothesis, not a fact

Most groups treat their initial survey like a photograph—a frozen, trustworthy record of what exists. That instinct is dangerous. A baseline is actually a set of educated guesses about what matters, where things live, and how the setup behaves under stress. Treat it like a scientific hypothesis: falsifiable, incomplete, and begging to be stressed before you spend a dime on earthmoving. The moment you accept a baseline as settled truth, you lock in every error hidden inside its methodology. I have seen a three-year restoration lose its entire primary season because nobody questioned whether the 2019 soil carbon readings still held after the 2021 drought. They didn't. The hypothesis failed. The budget didn't.

The five hidden biases every audit carries

Baseline audits pack five specific blind spots. Temporal bias assumes conditions during your two-week sampling window represent the full annual cycle. They rarely do. Spatial bias maps a few transects and calls the whole site covered. Gaps get smoothed over. Taxonomic bias privileges charismatic flora and overlooks the cryptic fungi or soil mesofauna that actually drive nutrient cycling. Social bias filters out how local users actually interact with the site—their seasonal trails, harvest zones, runoff patterns. And chemical bias assumes your lab's detection limits capture everything active belowground. off sequence. Each bias is an assumption wearing a lab coat. The catch is that no one-off audit can eliminate all five; the smart tactic is to rank them by how badly a faulty assumption would derail your planting scheme. Most crews skip this ranking—they pile all five risks into a one-off "measurement uncertainty" footnote and step on. That hurts.

Assumption testing as the restoration equivalent of unit tests

Software engineers run unit tests on compact, isolated code modules before integrating them into a full setup. They do this because a bug caught at series 47 overheads ten minutes; a bug caught after deployment spend a week. Restoration should work the same way. An assumption check is a cheap, fast experiment designed to stress exactly one premise while holding everything else constant. Does this seed mix actually germinate in the soil that sits six inches below the surface? trial three pots with real site soil, not the commercial blend. Does the water station really drop in August? Hammer in a shallow well and log it for one dry month before you design the swale network. swift reality check—if you cannot afford a one-week trial of your most critical assumption, you certainly cannot afford the two-year replant that follows a hidden failure. That said, these tests are not mini-research projects. Keep them lean. One variable, one clear pass-fail threshold, and a decision rule written before the primary shovel hits the ground.

'We tested the assumption that the marsh could handle twice the tidal flow. It couldn't. That check expense us two days. Replanting the whole zone would have expense two months.'

— Senior ecologist, Delmarva wetland project, after a lone Sluice-gate stress trial

What usually breaks opening is the social assumption. Restoration crews map boundaries and species ranges, but they rarely trial whether the neighbor's seasonal grazing schedule aligns with the germination window. One concrete anecdote: a salt marsh project in the Pacific Northwest planted 40,000 plugs based on a spring-only baseline. Nobody tested whether the August salmon run would revision the water chemistry. It did. Ammonia spiked. Thirty percent of the plugs died before Labor Day. A one-off bucket check with creek water during the run would have flagged the issue in an afternoon. The trade-off is speed—assumption testing adds three to five weeks to the pre-construction phase. The payoff is avoiding the kind of failure that sets a project back a full growth season. You choose the smaller delay now or the larger one later. Most choose faulty because they cannot stomach the appearance of hesitation. Don't be most groups.

How It Works Under the Hood

shift 1: Recalibrate your spatial and temporal scale — ask 'what if we sampled last June?'

Most baseline audits freeze a lone moment in phase. That June transect looks clean — but what if the staff had walked it in September, after a king tide scoured the bank? I have seen perfectly good restoration plans collapse because nobody asked whether the baseline snapshot captured a dry-season anomaly or a genuine steady state. The fix is brutal but cheap: pull three historical data points from the same site — same method, same transect, any season — and compare ranges. If the variance between snapshots exceeds 20 percent of your target condition, you do not have a baseline. You have a weather report. The trick is to recalibrate before you commit to species mixes or grading plans. Most crews skip this. That hurts.

site-tested method: grab satellite imagery for the five preceding years at matching phenological windows. Plot NDVI or bare-ground pixels. A spread greater than 30 percent signals a setup still in flux. Not yet ready for blueprint assumptions.

step 2: Identify hidden keystone variables — things that shift everything but rarely get measured

Standard audits track pH, bulk density, maybe soil organic carbon. Meanwhile, the real driver — say, seasonal groundwater fluctuation — sits unmonitored. swift reality check: in a coastal dune restoration I advised, the baseline report flagged phosphorus as the limiting factor. The snag wasn't phosphorus. It was microtopography — micro-depressions that pooled saltwater after spring tides, killing seedlings within hours. The variable nobody measured turned a three-year project into a five-year slog. The catch is that keystone variables are site-specific. A safe play: before finalizing any assumption, run a plain driver-mapping exercise. List every variable that could flip survival from 80 percent to 20 percent. Then ask the site crew which ones they worry about at 3 a.m. Their answer is your hidden keystone.

Decision threshold: if a potential keystone variable has zero measurements in your audit, that assumption is provisional. Treat it as a hypothesis until you collect at least two seasonal cycles of data.

stage 3: Run scenario sensitivity checks — nudge your assumptions 10% and see what breaks

This is where armchair modeling meets dirt. Take your blueprint's core assumption — say, "annual sediment accretion of 5 mm." Nudge it down to 4.5 mm. Does the vegetation community still establish within the window? What about the hydrology budget? I have watched a 10 percent reduction in assumed tidal prism turn a marsh platform from viable to drowning. The approach is deliberately crude: you adjust one parameter, hold everything else constant, and watch for nonlinear jumps. That jump — the point where tight input change produces large output collapse — is your fragility seam.

Most blueprints have two or three seams. Find them before the dozer arrives.

shift 4: Audit the data trail — where did each number come from and how was it collected?

off batch kills projects. A restoration staff once used a soil compaction value from a nearby golf course. That value — collected with a different penetrometer, on irrigated turf — was applied to a tidal mudflat. The planting holes failed. Every number in your baseline has a backstory. Probe it. Was the water craft sample taken after a rain event? Did the vegetation survey use a 1-meter quadrat or a 4-meter quadrat? The method matters more than the number itself. A useful heuristic: if you cannot name the instrument, the sampling date, and the person who collected the data point, treat the assumption as unverified. Industrial-grade skepticism here saves site-season heartbreak. One rhetorical question worth asking your data source: Would you stake your bonus on this number?

‘The most dangerous assumptions are the ones that arrived pre-verified — tidy spreadsheets with no smell of fieldwork.’

— observation borrowed from a veteran coastal ecologist, after watching three audits miss a plain datum error

That is not cynicism. It is triage. A clean data trail lets you catch the half-baked assumption before it expenses you a growing season. A murky trail means you are gambling on someone else's shortcut. The next section shows what that shortcut almost did to a tidal marsh — and how the staff caught it in window.

Worked Example: The Tidal Marsh That Almost Got Planted faulty

The baseline said 'low salinity, high organic matter' — but the historical record showed a 30-year salt pulse

A crew in the Pacific Northwest had a tidy story: restore a former dairy pasture to freshwater tidal marsh. The baseline audit—soil samples, vegetation surveys, a one-off dry-season water standard reading—reported low salinity, high organic matter. Looked clean. On paper, the planting list read like a sedge catalogue: Carex lyngbyei, Scirpus microcarpus, spike-rush. All freshwater workhorses. They ordered 80,000 plugs. That hurt the budget—$380,000 just in plant stock. But the grant was approved, the contractor was booked for October, and nobody questioned the assumption that the site had always been fresh.

The tricky bit is this: baseline audits capture a snapshot, not a movie. I have watched three restoration projects stall because groups treated a one-day grab sample as gospel. In this case, the audit missed something obvious—the site sat 900 meters from a dredged shipping channel where salt water intrudes during late-summer neap tides. Nobody measured that.

faulty queue.

Assumption trial: checking sediment cores and farmer interviews revealed the salt wedge

We insisted on an assumption trial before the planting crew mobilized. Not fancy—just cheap, fast, and diagnostic. We pulled three sediment cores along a 200-meter transect from the channel edge inland. The lab found buried foraminifera tests—microscopic marine shells—at 30–50 cm depth. That means salt water reached this soil within the last three decades. Then we interviewed the dairy farmer who worked the land until 2012. He described a "burn back" every few summers: the grass would yellow near the creek, the cows would stop grazing that strip. He called it "the salt creep." We checked aerial photos from 1991, 2004, 2017. The bare zone expanded 40 meters inland in 26 years. That is a 30-year salt pulse, not a static fresh setup.

Most crews skip this. They trust the lab result from a one-off soil pit dug in May when the water bench was high and the salt wedge was pushed deep. The catch is that low-salinity surface readings can coexist with a rising salt front at depth—what hydrologists call the salt wedge. If you plant freshwater sedges into soil where the root zone will hit saline groundwater within two growing seasons, you plant a funeral. I have seen that happen. The roots rot, the stem density crashes, and you re-plant at double expense. That hurts.

Result: switched from freshwater sedges to brackish species, saved $200,000 in replanting overheads

We revised the species list in six days. Out went 80,000 freshwater plugs; in came a mix of Schoenoplectus americanus (alkali bulrush) and Distichlis spicata (saltgrass) on the lower 40% of the site, with a transition zone of Juncus balticus toward the upland edge. The switch expense $12,000 to cancel the original queue and $3,000 to ship the new plugs—a $15,000 hit. That stung the contingency budget. But the alternative? Replanting 80,000 dead sedges in year two, plus labor and site prep, would have run $200,000 minimum. The farmer told us later that the salt wedge pushes further inland during multi-year drought cycles. Climate models for that watershed project a 12% increase in late-summer salt intrusion by 2035. The brackish mix handles it. Freshwater sedges would have died twice: once from salt stress, once from the replant crew's heavy machinery compacting the soil.

'We saved the project by losing two weeks and burning a tight budget chain—the audit alone would have overhead us a season.'

— restoration lead, after the site's third growing season

Final check: we installed a straightforward monitoring well at the property chain—just a slotted PVC pipe and a conductivity meter. Readings in August 2023 showed 8 ppt at 60 cm depth. That's brackish, not fresh. By 2025 the sedge dieback zone would have been obvious. Instead the bulrush cover hit 78% in year two. One assumption check. One farmer interview. One core. That is the difference between a blueprint that works and a blueprint that convinces you it works—until the bill arrives.

Edge Cases and Exceptions

Post-fire landscapes: when the baseline is a snapshot of ash, not recovery potential

I walked a burn scar in eastern Oregon six weeks after the fire. The soil was gray, hydrophobic in patches, and the audit showed zero vegetation — a dead zone. That is the baseline most crews export to their planning software. The glitch? Under that ash, rhizomes of Purshia tridentata were already pushing up new shoots. The audit recorded absence; the ground recorded a reset. In post-fire systems, your standard transect grid captures a thin veneer of the present while missing the entire subsurface recovery engine. The assumption breaks because heat can sterilize soil in some patches and stratify seeds in others — same fire, opposite outcomes. You need to dig. Literally. Soil seed bank cores at 10 cm and 20 cm depths, tested with a plain germination trial in trays, will reveal whether your site is primed for natural regeneration or needs a full planting intervention. The trade-off is slot — two weeks for germination counts vs. two hours for a transect. But acting on the ash-only snapshot guarantees you either over-plant into a stack that was healing itself, or leave bare a zone that will erode before native seed banks activate.

That hurts. And it happens every fire season.

Urban brownfields: hidden contamination gradients that your grid misses

Most baseline audits for brownfield restoration use a 20-meter grid. Cheap, repeatable, consistent. What that grid cannot see is the abandoned underground fuel tank eighty feet away, leaching hydrocarbon plumes that shift with the water surface. I have seen a site flagged as "low contamination — suitable for native prairie" get planted with forbs that all died within one growing season. The audit said the soil was fine. The reality was a gradient of heavy metals that changed direction with seasonal rain — the grid just happened to miss the hot spots. The assumption testing here demands a different sequence: proxy sampling with bioindicators. Plant a strip of fast-growing, metal-accumulating species like alpine pennycress or indicator sunflowers along potential contamination vectors — old building foundations, railway spurs, drainage ditches. If they show stunting or chlorosis within six weeks, you have a gradient the audit never mapped. The catch is that this delays planting by a season. Most project timelines cannot stomach that. But the alternative is planting into a death zone and blaming the contractor.

Not pretty. But honest.

Invasive-dominated systems: when the audit records invasives as 'the community' and misses the native seed bank

A site overtaken by cheatgrass or phragmites will produce a baseline audit that reads like a monoculture. The protocol says "identify dominant cover" — so the spreadsheet fills with one species. The assumption baked into that is straightforward: invasives are the community, and restoration must replace them entirely. off. I have seen a phragmites stand in a Great Lakes marsh that looked impenetrable, yet soil cores revealed a dense bank of Carex and Schoenoplectus seeds dormant underneath. The audit captured the invader's canopy; it missed the waiting archive. The fix is brutal but plain: before any herbicide application, dig trial pits and float the seed bank in a greenhouse. If native seeds are present at densities above 200 viable seeds per square meter, your strategy flips from "replace everything" to "suppress the invader and let the bank express." That changes herbicide selection, timing, and spend by an sequence of magnitude. The pitfall is that seed bank germination trials take six to eight weeks, and most grants demand earthmoving by month two.

'We treated for monoculture and got a diverse wetland for free — because we waited six weeks to check what was already there.'

— Restoration manager, Lake Erie basin, after shifting from full gut to partial suppression

Most groups skip this step. They spray primary, audit later, and wonder why the natives never come back. The baseline is not lying — it is just looking at the faulty layer. Check the bank before you pull the trigger.

Limits of the Approach

You cannot trial everything — where to draw the line to avoid analysis paralysis

Most crews skip this: they run five tests, find nothing alarming, and go ahead. That is not the snag. The glitch is the team that tests fifty things, finds three weak signals, and delays the restoration by six months chasing ghosts in the data. I have watched a perfectly good marsh blueprint stall because someone insisted on verifying soil pH at every tenth of a meter. faulty batch. The project never got planted.

The catch is that assumption testing has an infinite surface area. You could check seed viability, sediment compaction, bird predation rates, and twenty other variables — and still not touch the one thing that will break your project. Where do you stop? straightforward rule: trial only the assumptions that, if off, flip the project from success to failure. Not the ones that shift success by seven percent. Flip the project. Drainage gradients that would drown a plant community? trial that. Whether the soil carbon baseline from a 1992 survey still holds? That can wait.

Data quality limits: old maps, missing records, biased local knowledge

‘I trusted the old survey because the map looked clean. The seam blew out at year two.’

— A sterile processing lead, surgical services

The risk of finding nothing faulty and still failing — assumption testing is necessary, not sufficient

swift reality check — next phase your blueprint passes every trial, ask: what did we not trial that could still kill this? Write that list. Put it in the project folder. Then monitor for those exact signals year one, year three, year five. Not because you missed something in the audit. Because audits are always backward-looking, and restoration lives forward.

Reader FAQ

But our baseline is peer-reviewed — isn't that enough?

Peer review checks methodology, not local truth. A 2021 hydrology model might be flawless for the published site but blind to the fact that your tidal channel shifted three meters east last winter. I have watched groups lean on a pristine academic baseline while their planting elevation sat six inches too low — enough to drown the rootstock in one spring tide. The paper was correct. The site was different. Peer review gives you confidence in the process; it cannot guarantee that the variables you imported match the ground you are standing on. That gap is where assumption testing lives. Quick reality check—a peer-reviewed soil carbon estimate from a similar latitude tells you nothing about the legacy industrial fill you just augered through. Trust the science, sure. But check the seam between the paper and the mud.

We don't have window for more analysis — won't this delay the project?

The objection is fair — permitting windows are tight, grant deadlines loom. But here is what I have seen happen when we skip the trial: we plant two hectares of the faulty species, lose a full growing season, and spend twice the original budget on replanting and herbicide. That delay was not optional. The assumption trial would have taken four days. Three soil pits, one afternoon of flow tracking, a half-day conversation with the oldest fishing guide at the local bait shop. The catch is that the check happens before the heavy machinery rolls in, which feels like stalling. In practice it is the opposite — it front-loads the risk rather than letting it metastasize into a failed season. Most crews skip this: they treat a baseline audit as a final exam instead of a primary draft. Adjust the draft, not the calendar.

What if we run the tests and find nothing off?

Then you have just proven your assumptions are sound. That is not a waste — it is a deliverable. Funders who see a clean assumption-trial report sleep better than those who signed off on a baseline they never stress-tested. And nothing faulty rarely means nothing interesting. More often it reveals a minor boundary condition — a pH shift at the south edge, an unexpected subsurface clay lens — that would have become a problem three years in, not three months. You save yourself a mid-project pivot. That alone is worth the shovel phase. The tricky bit is framing this for managers who equate activity with progress; show them the alternative is a redo that spend six times as much and arrives eighteen months late.

How do we convince a funder or regulator to pay for assumption testing?

'A baseline tells you what the site looked like last Tuesday. Assumption testing tells you whether your plan survives next Tuesday.'

— paraphrase of what a restoration director told me after losing a permit appeal

Funders hate surprises. Frame the trial as insurance — a small upfront cost that protects their investment against the most common failure mode: a mismatch between the blueprint and the real-phase system. Show them one worked example from your region where the baseline missed a critical variable (tidal lag, groundwater rise, beaver recolonization — pick the one that scares them most). Then offer a simple deal: we run the check on a solo reference transect. If the assumptions hold, we proceed . If they crack, we adjust before the big spend. Regulators, in my experience, respond to the same logic — they want assurance that the restoration will meet its performance standards. A clean assumption trial is that assurance. One page, four soil profiles, and a note that says the model matches the site. That is cheaper than an enforcement action.

Practical Takeaways

Three floor checks you can do tomorrow (no budget needed)

Walk your site with a shovel and your phone. initial, the topsoil sniff trial—dig a fist-sized hole in the reference area, then in the restoration zone. Crouch down. Smell it. Healthy soil smells like damp earth after a storm. Sour or sulfurous? You’ve got anaerobic rot, and the baseline’s drainage assumptions are probably faulty. I once watched a crew plant seventy flats of sedges into soil that smelled like rotten eggs. flawed species. We lost a season.

Second, historical photo overlay. Open Google Earth, slide the timeline back to 1970, 1950, or—if you’re lucky—1938. Screenshot. Drop that into your phone’s photo gallery and walk the perimeter holding it up. What does the 1950 vegetation pattern show that your modern baseline glosses over? A treeline that’s migrated fifty feet? A drainage ditch that used to be a creek? That gap between what was and what is tells you more than the report’s optional-depth map.

Third, a lone phone call. Call the oldest person you can find who worked or lived near that site—county extension agent, retired rancher, long-time volunteer. Ask them one thing: “What did this place look like in a wet year before the road went in?” You will get a story. You will get the anomaly your baseline normalized away. That call expenses nothing. Skipping it costs months.

A one-page assumption tracker for your next kickoff

Print a sheet with four columns: Assumption, Source, Risk Level, check Date. At the kickoff, go around the surface and list every hidden bet in the blueprint—slope stability, seed bank viability, water-table depth, whatever. No debate yet. Just capture them. Then assign each a red-yellow-green risk tag. The trick is to test the red ones before you move dirt, not after. Most teams skip this:, they assume the baseline author already stress-tested those numbers. The baseline author probably didn’t. That hurts.

Keep the tracker taped to the site trailer wall. Every Wednesday, pick the highest-risk untested assumption and spend thirty minutes proving or disproving it. One crew I worked with discovered on week two that the supposed reliable water source was a seasonal pond, not perennial. They caught it because the tracker forced a Wednesday check. The alternative—ordering plants for a dry marsh—would have burned the whole grant.

The one question to ask before accepting any baseline

Here it is: “What is the oldest assumption in this report?”

Every baseline is a stack of decisions. The oldest one was made earliest, with the least data, and it never got rechecked.

— field note, after a 2022 tidal-marsh planting that nearly killed three acres of cordgrass

Ask it out loud in the next meeting. Watch people scan the document. The oldest assumption is usually the one nobody can defend anymore—a soil type inferred from a 1975 map, a hydrology estimate based on a single dry year. That is your starting point. Not the glossy executive summary. That brittle old bet. Fix that first, and the rest of the blueprint might actually hold. Ignore it, and you’re building recovery on a guess someone made before you were born.

Wrong order. But fixable—starting tomorrow, with a shovel, a phone, and a question.