The headache arrived after lunch.

  Not sharp. Not sudden.

  A pressure behind the eyes, like something gently pressing inward from all sides. As if the inside of her skull were slowly filling with cotton wool, muting thought and pushing everything just out of focus.

  Lena sat in the shade of the equipment tent, laptop balanced on her knees, fingers moving quickly as she typed up preliminary observations. Data first. Feelings later. That was the rule.

  Always start with what you could measure.

  She pulled up reference images from comparable Indus sites stored in her research folder. Mohenjo-daro. Harappa. Dholavira. Kalibangan. Lothal. Smaller satellite settlements scattered across Gujarat and Rajasthan, many barely excavated, known only from surface surveys and test pits.

  Collapse layers everywhere.

  Ash deposits indicating fire. Flood silt from rivers changing course. Earthquake damage. Collapsed walls, tilted floors, pottery broken in place. Evidence of reoccupation, sometimes in multiple phases, as people returned to salvage materials or squat among the ruins. Gradual abandonment as trade networks thinned and water sources failed.

  Mess. Chaos. The expected signatures of decline.

  Nothing like what she had just seen.

  She overlaid timelines, pulling data from three different databases. Climate proxies from lake cores and speleothems showing the 4.2 kiloyear event. Monsoon weakening around 1900 BCE, documented in oxygen isotope ratios. Trade disruptions traced through ceramic distributions and metallurgical analysis.

  The model refused to cooperate.

  Every variable she added made the picture worse, not better. Like adding pieces to a puzzle only to realize they came from an entirely different box.

  Lena frowned and stripped out assumptions one by one. Maybe the monsoon failure was slower than currently modeled. Maybe regional variation mattered more. Maybe trade networks were more resilient than scholars assumed.

  Still nothing.

  The clean boundary should not exist.

  Not in nature.

  Not in history.

  But it did.

  She leaned back in the camp chair, eyes closed, letting the headache throb. One-two. One-two. Her pulse flashed faintly behind her eyelids.

  This was how bad data felt. Resistant. Uncooperative. As if it resented being questioned.

  As if it were hiding something and daring her to keep looking.

  Her phone buzzed on the folding table beside the laptop.

  She ignored it.

  It buzzed again.

  With a sigh, she opened her eyes. The brightness made them sting. She checked the screen.

  Supervisor: Quick check-in. How’s the dig?

  Lena stared at the message.

  Dr. Sharma never checked in mid-season. He was hands-off by preference, trusting his researchers to manage their own fieldwork. “You’re adults,” he always said. “You’ll tell me if there’s a problem.”

  That was the whole point of being a PhD candidate.

  So why now.

  She typed: Normal. Making good progress on the occupation sequence. Some interesting pottery in the latest trench.

  The word normal looked wrong even as she sent it.

  Three dots appeared immediately.

  Lena’s pulse quickened.

  The dots vanished.

  Returned.

  Unauthorized duplication: this tale has been taken without consent. Report sightings.

  Vanished again.

  Like he was writing something, deleting it, trying again.

  Finally: Good. Let me know if you need anything. Funding committee is asking for mid-season updates from everyone.

  She read it twice.

  That was wrong too.

  Sharma was precise. He asked specific questions. Depth measurements. Stratigraphic relationships. Preliminary dating contexts. “Details, Lena. The devil and the divine both live in details.”

  This was generic.

  Like a form message.

  Like he was ticking a box.

  Or like someone had asked him to check on her specifically.

  Lena set the phone face-up on the table and stared at it.

  Paranoia, she told herself. Heat. Isolation. The kind of thinking that crept in when you’d been in the field too long.

  Except noticing patterns that shouldn’t exist was her job.

  It was why she was here.

  She opened a new file and ran a simple distribution analysis on the depth variance of collapse layers across known Indus sites. A crude check. Something solid. Something numerical.

  The dataset was enormous. Decades of excavation reports and surveys, digitized and uploaded by patient archivists. Over three hundred recorded terminal contexts.

  She filtered for sites within a hundred-kilometer radius.

  Seventeen hits.

  Not ideal. Enough.

  She pulled the depth measurements marking the transition from occupation to abandonment and ran a standard deviation calculation.

  The variance should have been high. Collapse happened differently everywhere. Timing varied. Causes overlapped. Water, trade, soil, politics. No two sites failed the same way.

  The results populated on screen.

  Her breath caught.

  The variance curve flattened.

  Not naturally.

  Not randomly.

  Artificially.

  She zoomed in, heart hammering now, fingers unsteady on the trackpad.

  The curve wasn’t smooth because of error or missing data.

  It was smooth because the outliers were gone.

  Removed.

  Sites that collapsed too early were absent. Sites that lingered longer, showed recovery, or failed in odd ways were gone as well.

  Only the clean failures remained.

  Lena sat very still.

  This wasn’t just her site.

  The pattern ran through the entire dataset she had trusted for years. The dataset she had cited in exams. The dataset underpinning her dissertation proposal.

  Someone had been curating the collapse.

  Not interpreting it.

  Managing it.

  A sharp knock struck the tent pole.

  Lena flinched, snapping the laptop shut. The screen went dark.

  “Lena?” Rajiv’s voice again, hesitant. “There’s someone here. From the regional office.”

  The headache spiked, pain flaring white behind her eyes.

  She blinked hard and forced her voice steady. “Did you call them?”

  “No,” Rajiv said. “They said they were already nearby. Routine inspection. Documentation review.”

  Of course.

  Lena saved her files twice, once locally and once to the cloud, then ejected the USB drive and slipped it into her pocket. The small weight grounded her.

  Footsteps approached. Measured. Unhurried.

  A man stepped into the tent.

  Mid-forties. Polite smile that didn’t reach his eyes. Crisp white shirt already wilting in the heat. Government messenger bag. Wedding ring. Expensive watch. Shoes too clean for a dig site.

  “Dr. Vairavan,” he said, extending a hand. “Apologies for the interruption. Rajesh Kumar. Regional archaeological survey.”

  She shook his hand briefly. His grip was dry.

  “We weren’t expecting oversight today,” she said.

  “Of course not,” he replied, glancing around the tent with casual interest. “Just a courtesy visit. We’ve been reviewing regional data trends and wanted to see how things were progressing.”

  Trends.

  The word rang in her skull.

  He glanced toward the covered trench. The canvas glowed in the sunlight.

  “Anything interesting?” he asked.

  His tone was light.

  His eyes were not.

  “No,” Lena said.

  The lie lodged in her chest like something sharp.

  He smiled. “Excellent. Consistent with the regional pattern. That’s reassuring.”

  Reassuring.

  Archaeology was not supposed to be reassuring.

  He tapped on a tablet, scrolling. “We just want to ensure findings are properly contextualized. Isolated interpretations can cause confusion. Especially outside academic circles.”

  He paused. Looked up.

  “Your funding review is coming up, isn’t it?”

  “Yes.”

  “Well then. We’ll make sure your file reflects clean methodology. No complications.”

  He met her eyes.

  “No complications.”

  She nodded once.

  He smiled and packed the tablet away.

  After he left, Lena listened to his car fade into the distance. Listened to the dig resume around her.

  Then she sat down hard and bent forward until the dizziness passed.

  When she opened the laptop again, the variance graph was still there.

  Too smooth.

  Too calm.

  Beneath it, in her notes, the line she had written earlier waited.

  Failure is clean.

  She added one more sentence.

  Because someone makes it clean.

  Lena just discovered that the collapse of the Indus Valley Civilization isn’t just suspiciously tidy at her site - it’s suspiciously tidy everywhere. Someone has been quietly tidying up the data for decades.

  Outliers? Gone. Inconvenient lingering occupations? Poof. Messy human failure? Not on our watch, thank you very much.

  And then Mr. Crisp-White-Shirt shows up for a “routine” visit right after she starts noticing. Coincidence? In this story? Please.

  Questions I’m flinging at you like suspiciously clean pottery sherds:

  Is Rajesh Kumar just a mildly threatening bureaucrat, or is he the latest in a very long line of people whose job title is basically “Keeper of the Clean Failure”?

  Why does the word “reassuring” feel more terrifying here than “cursed tablet” ever could?

  And the big one: if you were Lena, would you email that variance graph to a trusted colleague… or would you start encrypting everything and sleeping with the USB drive under your pillow?

  Next chapter coming soon. Bring water, sunscreen, and a healthy sense of distrust.

  Still digging (nervously),

  The author who just triple-checked their own references