On a chromatogram, the height of a peak tells you how much of a compound there is. It does not tell you how strongly you smell it.
The two diverge so far that they can reverse a conclusion entirely.
Detection threshold
Each compound has a minimum concentration below which the human nose cannot detect it. That figure differs between compounds by millions of times.
Geosmin — the smell of first rain — is detected at a few parts per trillion. Limonene needs concentrations many orders of magnitude higher.
The consequence: a sample containing 40% limonene and 0.0001% geosmin may smell more clearly of earth than of peel.
Three examples from this dictionary
1-Octen-3-one — the compound behind the metallic smell — has one of the lowest thresholds ever measured. It forms in tiny amounts on your skin and is still enough for you to notice at once.
2,4,6-Trichloroanisole ruins a whole bottle of wine at a few parts per trillion. No ordinary measurement catches it, but the drinker knows at once.
2-Acetyl-1-pyrroline is present in sticky rice in tiny amounts relative to the starch, and it still defines the whole smell of the pot.
The reading rule
The horizontal axis of a chromatogram can be trusted. The vertical one needs care.
Reading horizontally means reading the order of evaporation, and that order reflects your experience quite accurately.
Reading vertically means reading quantity, and quantity is not proportional to perception.
Why this matters when buying
It explains why two bottles with almost identical analyses can still smell clearly different: the difference lies in trace components that the analysis records at 0.1 per cent or not at all.
It also explains why blended goods are poorer than the real thing even when the data table looks equivalent. The blender adds the big peaks; the dozens of trace compounds nobody checks are not added.
The next lesson looks more closely at those trace compounds.