Methodological errors in the precise measurement of gamma-ray energies with Ge detectors are examined basing on a model of photopeak shape function. The model revealed that the centroid of the normal distribution embedded in the phenomenal photopeak is to be compared with the energy scale. The peak difference rate is introduced to evaluate the peak difference, the essential quantity of experimentally giving secondary energy standard, relative to the energy difference. Varied value of the gap difference-rate index with several physical parameters has indicated that the secondary gamma-ray energies have a definite amount of methodological errors. A typical methodological error in the gamma-ray energy at around 300 keV is in the range of -263 +/- 125 approximately -54 +/- 123 eV (or -877 +/- 418 approximately -180 +/- 411 ppm). Commonly used "conventional" energy-calibration methods by the use of the most probable peak-height position have been reviewed. Origins of methodological errors are discussed. A new energy-calibration method using the centroid is proposed.