That the great ice ages of the past several hundred thousand years march to an astronomical beat is among the more secure findings of paleoclimatology. The tilt of Earth's axis, the precession of its spin, and the shape of its orbit vary on periods of roughly forty-one thousand, twenty-three thousand, and one hundred thousand years, and these cycles redistribute sunlight across latitudes and seasons. Where deep-sea sediments record the waxing and waning of continental ice, the same periods reappear. The correspondence is too exact to be coincidence: orbital variation paces the glaciations. Yet pacing is not the same as forcing, and here the tidy picture frays. Of the three orbital rhythms, the one governing orbital shape - eccentricity, with its hundred-thousand-year period - produces by far the feeblest change in the total sunlight Earth receives. The variation in incoming energy attributable to eccentricity is a fraction of a percent, far too small, on any straightforward accounting, to drive the enormous ice sheets that advanced and retreated on precisely that hundred-thousand-year schedule over the past million years. The dominant beat of the ice ages is set by the weakest of the drivers. This is the hundred-thousand-year problem, and it has resisted easy solution for half a century. Two broad classes of response have emerged. The first holds that the climate system does not respond to orbital forcing in proportion to the forcing's size. A small but well-timed nudge can trigger a large reorganization if the system is already poised near a threshold; feedbacks internal to the ice - the reflectivity of its bright surface, the slow deformation of its base, the drawing down of atmospheric carbon dioxide as the oceans cool - can amplify a modest astronomical signal into a full glacial epoch. On this view eccentricity does not supply the energy; it supplies the timing, and the system's own nonlinearities furnish the magnitude. The second class of response is more radical. It proposes that the hundred-thousand-year rhythm is not a direct response to eccentricity at all but an internal oscillation of the coupled ice-ocean-atmosphere system, one that would occur even under a constant Sun and that the orbital cycles merely nudge into phase. If so, the striking match between eccentricity and glaciation is partly a coincidence of comparable periods - a resonance rather than a response. Adjudicating between these accounts is difficult precisely because both predict the observed correlation. Distinguishing them requires evidence the correlation alone cannot supply: independent constraints on how much amplification the ice and carbon systems can actually deliver, and records long enough to test whether the hundred-thousand-year beat persisted when eccentricity was weak. The available records hint that the strong hundred-thousand-year cycle is a comparatively recent feature, largely absent from the earlier part of the ice-age era, when a forty-one-thousand-year rhythm prevailed. That the dominant period should have shifted, while the orbit itself did not, tells against any account in which eccentricity simply and directly writes the climate record - and reminds us that a correlation robust enough to anchor a chronology may still be too weak to settle a mechanism.