RESEARCH METHODS Samples from the Dakota Formation were obtained from cores at Sergeant Bluff, Iowa (Witzke and Ludvigson, 1994). Samples from the Swan River Formation of Manitoba were obtained from unit 1 of outcrop section 57 of McNeil and Caldwell (1981, p. 349-350). Samples were impregnated with epoxy, and thin slabs and thin sections were cut perpendicular to bedding. Powdered

Palaeozoic lithic palaeoclimatic data were plotted at the series or stage level for Laurentia (North America) andEuramerica. Their distribution was used to infer the extent of humid and arid climatic belts across the continent for specified time intervals. Climatic data are potentially useful for constraining palaeolatitudes for times when zonal atmospheric circulation remains the dominant pattern. Arid lithic indicators (evaporites, carbonate oolite) and humid indicators (coal, bauxite) were constrained to provide the best fit of these data into a simple zonal climatic pattern delineating subtropical arid belts and equatorial and/or temperate humid belts, respectively. An attempt was made to minimize occurrences of arid indicators in interpreted equatorial regions to maximize a zonal fit. Climatic belts can be arranged in a temporally consistent zonal climatic scheme for much of the Palaeozoic, but monsoonal and orographic effects apparently disrupted the general zonal pattern during portions of the Late Palaeozoic. The interpreted climatic patterns suggest progressive southward movement of Laurentia during the Cambrian through Early Devonian followed by general northward drift of Euramerica during the Middle Devonian through Permian. Lithic palaeoclimatic data provide an independent means to test palaeolatitudes interpreted from palaeomagnetic data. There is general agreement between palaeoclimatic and palaeomagnetic interpretations, but some discrepancies, particularly for the Devonian and Early Carboniferous, warrant further attention.