Wisconsin tills in Minnesota may be assigned to different ice lobes depending primarily upon the color, texture, stone content, and other petrographic characteristics. These petrographic features have their origin chiefly in distinctive types of bedrock in Minnesota and adjacent areas. The Precambrian rocks that underlie the drift over most of the state include red sandstone and slate in a belt extending SSW. from the head of Lake Superior; basalt, felsite, gabbro, anorthosite, and related rocks in the highland N. of Lake Superior; Fe formation; and various metamorphic and granitic rocks. Paleozoic carbonate rocks are found principally in southern Manitoba but also in southeastern Minnesota. Cretaceous shale occurs in the area of the Red River Valley and in southwestern Minnesota. The configuration of the several ice lobes was determined partly by the presence of lowlands in the bedrock, notably one following the Red River-Minnesota River valleys (Des Moines lobe) and one following a lowland extension of the Lake Superior basin (Superior lobe). The central part of the state was occupied by the Wadena lobe and the Rainy lobe, and their directions of movement were probably controlled by mutual interference as well as by preexisting topography. The Wisconsin glacial history in Minnesota was marked also by repeated advances of the several ice lobes. In the case of the Superior lobe the tills assigned to the Cary and Valders advances may be identified texturally because the Valders till was formed by the incorporation of lake clays deposited in the Lake Superior basin during the Two Creeks interstadial just preceding, where the Cary Superior till is sandy. A Mankato Superior till, identified on morphologic grounds, appears to have an intermediate (silty) texture. Of the 2 tills of western source, the Wadena is the more sandy, probably because the Des Moines till passed over Cretaceous shale. The tills of the Rainy lobe and its 2 sublobes (Pierz and Brainerd) are all sandy and are stonier than the tills of the other lobes. Stone counts are perhaps the most diagnostic petrographic characteristic of Minnesota tills. They are easily made and generally point to the correct assignment. Difficulties occur where tills contaminate one another in contact zones. Color distinctions, as measured by the Munsell color chart or in the laboratory by the Clark-Maxwell method, are also generally distinct. Heavy-mineral analyses and clay-mineral analyses are of limited use in differentiating tills. They are time-consuming and not particularly diagnostic.