Mariana stepped out of the small conference room on the second floor of RMBL365 and stood for a moment in the hallway, letting her ears ring down. The Niwot lead had not exactly conceded, but she had stopped saying "the Hesperophylax clade" as though it were one thing, which was what Mariana had come for. Two hours of taxonomic argument before lunch, and her Spanish-language slides had loaded without complaint on the shared platform, the genus names underlined where the synthesis engine had flagged a treatment conflict with Andrews, with Niwot, with the H. J. Andrews lead's 2031 revision. The system had pulled her old INECOL records into the same view without her asking. She would never get used to that, even though by now she should.
In the hallway the heating ticked. Outside, Crested Butte was in one of its bright cold February stretches: nine degrees Fahrenheit at the thermometer by the door, sun strong enough that the south-facing snowbanks were already cupping. Someone had propped open a window at the far end of the corridor and she could smell woodsmoke and the dry mineral smell of plowed snow. She thought of Cofre de Perote, where this morning the cloud forest would be wet to the knees and her field assistant Tonalli would be netting at the upper pool with cold hands. She had texted him before the session: Encontré aliados. La latitud sí importa. He had not yet answered. He would be driving back down to Xalapa, probably, with no signal on the switchbacks.
She had fifteen minutes before the Andrews lead disappeared into the next working group. She had been thinking about how to use them since Sunday.
Down the stairs, past the wall where someone had hung a print of the Gothic meadows in late June, glacier lilies thick as foam — except they wouldn't be that thick now, not in late June, the lily window had crept into the second week of May at lower plots and the print was from 2009, she'd checked. Print as artifact. Past the kitchen where the lunch trays were stacked: chile verde, a tray of those small bowls of stewed greens someone in the kitchen made on Tuesdays, a thermos of coffee with a hand-lettered sign saying fresh at 12:40. She poured a cup. The Andrews lead was at one of the small tables by the south windows, looking at a tablet, eating a tamale with one hand.
"Diego."
He looked up and smiled. "I was hoping you'd find me. Sit."
He had been at RMBL since Sunday too, and they had known each other since her postdoc year at Andrews in 2009 — she had been a confused thirty-year-old in his lab meetings, he had been a fourth-year grad student who took pity on her English. Now he was a full professor and the lead on the Andrews caddisfly time series and she ran a station and they were both, by some shared bureaucratic miracle, here.
"I want to talk about the synthesis paper," she said. "Specifically figure three."
"I knew you did. Sit, please."
She sat. She set the tablet down between them and woke the synthesis platform. The platform pulled up the draft figure — a multipanel showing community composition trajectories from 1998 to 2038 across the partner stations, currently with five panels, none of them below 36 degrees north. The empty space at the bottom of the latitudinal axis was, to her eye, very loud.
"You don't have a low-latitude anchor," she said.
"We don't. We were going to caveat it."
"Don't caveat it. Use mine."
He chewed for a moment. The platform, sensing her gesture, populated a sixth tentative panel from her INECOL Cofre series — eighteen years, not forty, but at 19 degrees north latitude and 4,000 meters elevation, with a hydroperiod regime where snowmelt was not the dominant cue and the question of community reassembly looked, frankly, different. The platform had been ingesting her data since November, since she had loaded it under the new shared schema. She had spent three weeks of evenings doing the loading. It had been worth it for this moment alone.
"The Trichoptera composition shift," Diego said, looking at her panel.
"Starts in 2026. Accelerates after the 2031 drought year. We lose two limnephilids by 2034 — gone, not reduced, gone — and we gain a hydropsychid I had to send to Costa Rica to confirm because nobody at INECOL had seen it above 3,500 meters before. The pattern is the same as yours in shape but the driver isn't snowmelt. It's dry-season pool persistence."
"Which means your panel does work mine can't do."
"Which means my panel does work yours can't do."
He was quiet for a moment. She watched him think. He was, she remembered, slow in this particular way — not slow at understanding, slow at committing. She had learned in 2009 to let him be slow.
"The MRS lead is going to want it framed as confirmation," he said finally. "That the pattern is general. I don't want it framed that way."
"Neither do I. I want it framed as: where snowmelt isn't the cue, the reassembly takes a different shape, and the difference tells us about the mechanism." She tapped her panel. "Look at the recovery years. 2032, 2036. The community doesn't reassemble back toward the old composition. It reassembles toward a new one, and the functional rates" — she swiped to a second tab, where the platform had pulled detritus-processing data from her colleague's parallel project — "don't track composition the way the basin's do. Not yet, anyway. That's a paper. That's not a caveat."
Diego set down his tamale. "Send me the panel as a working draft tonight. I'll talk to the MRS lead before the four o'clock. The framing argument we can have over dinner."
"Tonight I'm eating with Rosa and the Sierra Nevada people. Tomorrow."
"Tomorrow then."
She let herself smile. The coffee was bitter and very hot. Outside the window two ravens were doing something idiotic on the snowbank, one of them trying to carry off what looked like a glove.
"Mariana," he said. "I'm sorry the travel reimbursement was so late."
"It came through Friday. Maite already used it to pay the dentist."
"I heard your institution cut everything in March."
"Everything except salaries. So far." She shrugged. The shrug was, she felt, the most honest gesture she had made all week. "The network fund covered the flight and three nights. I covered the fourth. I would have come anyway."
"I know you would have."
She thought about telling him about Tonalli, who was twenty-six and brilliant and had been quietly applying to Canadian postdocs because he was tired of running a field program on someone else's emergency funding. She thought about telling him that she had decided, on the flight up, that if the synthesis paper went well — if her panel went in as a panel, not a caveat — she would write Tonalli a letter for the Andrews postdoc call in the fall and not feel guilty about losing him. The platform had a tab for postdoc exchanges across the partner stations now. She had looked at it twice on the plane.
She didn't tell him. There would be time tomorrow.
"Three o'clock session," he said, glancing at the tablet. "You're presenting first."
"I'm presenting the question about cue substitution. You should heckle me."
"I will heckle you."
"Good."
She stood, picked up her coffee, and walked back toward the conference room. Halfway down the hall her phone buzzed: Tonalli, finally. Llegué. La carretera estaba cerrada por un derrumbe pequeño. Mañana voy al sitio 4 a checar las trampas. Cuídese, doctora. She read it twice. Sitio 4 was the upper pool, the one where the limnephilids had been gone since 2034 and where she suspected, this dry season, the hydropsychid was establishing. She thumbed back: Mándame fotos del agua. Y de las trampas. Estoy en una pelea aquí y voy ganando.
The reply came as she reached the door of the conference room: a thumbs-up, and then, a second later, Va.
She slipped her phone into her pocket and went in to set up her slides.