The establishment of the Carnegie Institution of Washington's Geophysical Laboratory in 1905 and the pathbreaking work conducted there by Norman L. Bowen in the 1910s and 1920s are commonly considered to have rendered igneous petrology an experimental science. A closer examination of the work of American igneous petrologists outside the Laboratory reveals, however, that consideration of experimental data did not become an integral part of petrological practice until after World War II. To be sure, igneous petrologists celebrated the Geophysical Laboratory and its experimental approach in speeches and historical reviews throughout the interwar period. In their actual research, though, most igneous petrologists ignored the large body of experimental results gathered by the Geophysical Laboratory and treated Bowen's theory of differentiation as merely another speculative petrogenetic hypothesis to be tested against field data. These petrologists, many of whom were engaged in mapping for state and federal surveys, regarded laboratory data on simple anhydrous systems as simply inapplicable to real rocks. Not until after World War II, when the Geophysical Laboratory began large-scale experimentation on hydrous systems and natural rocks, did field petrologists generally accept the relevance of experimental data. At the same time, the institutional framework of igneous petrology changed in the late 1940s and 1950s as a number of experimental petrologists took advantage both of support from new government agencies, and of the rapidly increasing prestige of geochemistry within the discipline, to establish numerous new experimental laboratories. This change in the structure of the discipline, coupled with the greater reliance on hydrous systems and natural rocks in experimental work, finally led to the general incorporation of experimentation into petrological practice.
