A few years after Arrhenius published his hypothesis, another scientist in Sweden, Knut Ångström, asked an assistant to measure the passage of infrared radiation through a tube filled with carbon dioxide. The assistant ("Herr J. Koch," otherwise unrecorded in history) put in rather less of the gas in total than would be found in a column of air reaching to the top of the atmosphere. The assistant reported that the amount of radiation that got through the tube scarcely changed when he cut the quantity of gas back by a third. Apparently it took only a trace of the gas to "saturate" the absorption — that is, in the bands of the spectrum where CO2 blocked radiation, it did it so thoroughly that more gas could make little difference.
...........
These measurements and arguments had fatal flaws. Herr Koch had reported to Ångström that the absorption had not been reduced by more than 0.4% when he lowered the pressure, but a modern calculation shows that the absorption would have decreased about 1% — like many a researcher, the assistant was over confident about his degree of precision.(9*) But even if he had seen the1% shift, Ångström would have thought this an insignificant perturbation. He failed to understand that the logic of the experiment was altogether false.
The greenhouse effect will in fact operate even if the absorption of radiation were totally saturated in the lower atmosphere. The planet's temperature is regulated by the thin upper layers where radiation does escape easily into space. Adding more greenhouse gas there will change the balance. Moreover, even a 1% change in that delicate balance would make a serious difference in the planet’s surface temperature. The logic is rather simple once it is grasped, but it takes a new way of looking at the atmosphere — not as a single slab, like the gas in Koch's tube (or the glass over a greenhouse), but as a set of interacting layers.