A model relating chromatographic signal characteristics and calibration curve nonlinearity has been developed. The relationship between signal and concentration is represented as a powers series In concentration. The sample peak in the detector is assumed to be perfectly Gaussian, and the detector response Is presumed to be instantaneous. Calculations indicate that the peak center, peak symmetry, and odd higher moments are Insensitive to detector nonlinearity. In contrast, peak height is most sensitive to nonlinearity, area is less so, and peak variance (width) is least sensitive. A 10% quadratic departure from linearity causes a deviation of 7 and 3.5% in area and width, respectively. Extreme nonlinearity, to the extent that the calibration curve has a maximum, leads to the appearance of dual (split) peaks.