The combination effect in baking was studied with two-component mixtures of wheat flours of different baking properties during three years. Loaves were prepared of the flour mixtures according to the standard method, and the kneading properties were studied by Valorigraph measurements. Evaluation of the data was carried out by principal component analysis. Assuming the properties of the pure components utilized as additive, the values of the properties studied to be expected were calculated, and these were subtracted from the data measured. On the basis of principal component analysis of the differences obtained, it was established that the differences in loaves properties did not show any dependence on the year of harvest. On the contrary, the dough properties were dependent on the year. The combination effect was expressed most markedly by the baking value according to TIBOR (1933) and, for the dough properties, by the dough formation time. The baking value improved to a greater extent than expected in the cases when mixing prolonged the dough formation time. On the other band, dough formation time as shortened upon the influence of mixing, deteriorated the product quality as compared to the calculated values.