Snow accumulates on the Antarctic ice sheet and is compressed first into firn, then ice. Dust and chemical traces settle with it. Air remains connected through pores until bubbles finally close, so the gas is younger than the surrounding ice — at glacial Vostok, sometimes by thousands of years. Oxygen and hydrogen isotope ratios provide temperature proxies whose interpretation depends on site and calibration, not a perfect thermometer stamped into every layer.

Together, these measurements reveal recurring glacial and interglacial states. In the Vostok record, atmospheric CO₂ generally ranged from roughly 180 ppm in glacial periods to about 280–300 ppm in interglacials; the often-cited temperature swing is an Antarctic estimate, not a single global thermometer. The curves move in striking relationship, while their leads, lags, and causes require analysis beyond what the eye alone can settle.

The wonder survives the caveat: air once thought irretrievable persists in bubbles small enough to disappear beneath a fingertip.

Deep ice can crack as a core is brought into different pressure and temperature conditions. It is tempting to hear ancient air speaking, but the sound alone cannot date a bubble. That work belongs to the core's measured depth, age model, and the physics of how firn closes.

In the laboratory, researchers extract gas from the bubbles and directly measure its composition. The concentration measurement is direct; assigning it a date is reconstructed, because air is sealed only after snow has compacted below the surface and is therefore younger than the surrounding ice. Memory is the metaphor. Measurement is the mechanism — and both begin with something astonishingly small surviving under pressure.

The Vostok core reaches 420,000 years. The EPICA Dome C core extends to 800,000. During the 2024–25 field season, the separate Beyond EPICA project reached bedrock at Little Dome C; by October 2025, the recovered core was reported to contain at least 1.2 million years of climate history, reaching into the Mid-Pleistocene Transition.

Why the rhythm of glacial cycles changed remains an active research question. The new ice may help answer it.

Each deeper core is a longer memory, but each has its own site, drilling history, and publication date.

Ice cores preserve more than gases. Some eruptions leave sulfate peaks, acidity changes, or rare particles of volcanic tephra. A sulfate spike can mark an eruption without identifying which volcano produced it; geochemically matched tephra offers firmer attribution, but is much less common. Atmospheric circulation, deposition, chronology, and later alteration all shape what reaches a given core.

Well-supported volcanic horizons can help align records across sites, much as a widely witnessed sky event can anchor more than one human chronology. The comparison is about cross-referencing, not identical methods: ice chemistry is physically deposited, while a Winter Count is a community's chosen mnemonic.

The ice does not remember every eruption. It preserves an uneven signal. The achievement lies in learning what that signal can — and cannot — say.

The long CO₂ curve joins measurements made in different ways: air extracted from ice for the ancient record, then direct atmospheric monitoring for the present. In May 2026, NOAA's preliminary global monthly mean was 428.73 ppm. The live number will move; the historical contrast does not. It stands far above the range measured in the 800,000-year EPICA record.

You are not looking at the graph from outside. You are living inside its newest line.

The planet kept no diary for us, but ice retained evidence. Reading it gives us something rarer than certainty about every detail: a long perspective on the atmosphere we are changing, and responsibility for what we do with that knowledge.