A study of 304 selected biotite analyses, with 17 chemical variables (AlIV, FeIV, AlVI FeVI, Mg, Mn, Ti, Li, Na, K, Rb, Ca, Ba, OH, F, Cl,Θ), was carried out using variation diagrams and statistical methods. It was our aim to verify the existence of characteristic variation patterns in the crystal chemistry of igneous biotites related to the geological setting and chemistry of the parent rocks. In the absence of a range of analyses covering the whole spectrum of igneous rocks, the biotites were grouped “a priori” either as volcanic (rhyolites, rhyodacites and trachyrhyolites, dacites and trachytes, “andesites”, trachybasalts and nephelinites) or as plutonic (granites, granodiorites, tonalites, diorites, “gabbros”). Variation diagrams failed to distinguish clearly between the different groups of biotites; the data overlapped considerably and were accordingly difficult to interpret. Statistical analysis, on the other hand, revealed clear chemical differences; moreover, simple statistical equations and careful coefficients were established which make it possible to evaluate the degree of discrimination between the different groups and to classify unknown biotites. The most important petrological factors affecting biotite chemistry are temperature of crystallization, rock “acidity” and, probably, rock “alkalinity” and the presence of other Fe-Mg minerals. Factors, such as/tf\(f_{H_2 O}\) and\(f_{O_2 }\), can cause more limited variations.