Abstract Lack of terrain data contributed significantly to the high costs of lives and operations during the Pacific campaign of World War II. After the war the U.S. Army Corps of Engineers contracted with the Military Geology Branch of the U.S. Geological Survey to gather detailed terrain information about the occupied islands under direct U.S. jurisdiction in the event they or comparable oceanic islands became sites of future military operations. The U.S. Geological Survey established a headquarters in Tokyo and initiated field studies of Okinawa during 1946. Subsequent detailed studies were launched at the Palau Islands (1947), Yap Islands (1947), Saipan (1948), Tinian (1949), Guam (1951), Pagan, Marianas Islands (1954), Truk (1954), Ishigaki and Miyako (1955), and the Marshall Islands (reconnaissance, 1951). Initial plans for detailed studies of all mandated islands were abandoned for lack of time, but members of the field parties briefly visited nearly all. Field teams included geologists, hydrologists, soils scientists, a plant ecologist, and a climatologist. The Tokyo office gathered and translated existing Japanese literature about the islands; more than 600 articles were translated. A by-product was the establishment of a joint U.S.-Japanese project to compile and publish a series of 1:250,000 geologic maps of formerly held Japanese territories, including Korea, Manchuria, northeast China, southern Sakhalin Island, and the Kuriles. Results of the field studies were published in a series of military geology folios composed of both basic and interpretive chapters. U.S. Geological Survey professional papers presented many of the scientific results.

Petrographic examination of temper sands in prehistoric Oceanian pottery collected by archaeologists from island groups spread across the tropical Pacific Ocean shows that the sands vary compositionally in geographic patterns that are governed by geotectonic setting and vagaries of local bedrock exposure on individual islands. The small islands serve as virtual point sources of sediment derived exclusively from the restricted array of rocks that form each island. Both natural and manually added tempers can be traced to bedrock sources by the same petrographic methodology, but independently sourcing clay bodies requires geochemical comparison of clay pastes with potential clay sources. Oceanian tempers include calcareous as well as terrigenous sands, but only the latter can be associated unequivocally with specific islands or island groups because the nature of reef tracts is similar throughout the tropical Pacific. Exotic tempers can be distinguished from indigenous tempers because their compositions are incompatible with the geology of the islands where the exotic sherds are found. Human migration into islands of the Pacific Ocean was the last main stage in human dispersal over the planet, with no human occupation of the small islands lying beyond island Southeast Asia and Australasia until 1500 B.C. The earliest inhabitants possessed a ceramic culture, and ceramic traditions evolved over subsequent centuries to produce a varied succession of ceramic phases. Lapita pottery, which is the oldest ware in southwest Pacific island groups, is especially notable because its production was limited to a time frame short enough to allow Lapita sherds to serve a role akin to index fossils. Temper sands in Lapita and post-Lapita sherds from the same locales are indistinguishable and show that salient temper contrasts are controlled by island geology rather than habits of ancient potters. Prehistoric collecting sites for temper sand were not necessarily identical to places where modern sand accumulates because of severe environmental changes on many islands. The compositions of terrigenous temper sands in Pacific Oceania reflect the complex pattern of circum-Pacific plate boundaries and intra-Pacific hotspot chains, and define oceanic basalt, andesitic arc, postarc-backarc, dissected orogen, and tectonic highland temper classes composed of different associations of grain types. The geographic distribution of different temper classes reflects not only the current geotectonic setting of each island group but also their paleotectonic settings when exposed rock assemblages were formed. Temper aggregates include beach, stream, and rarely dune sands, as well as grog (brokensherd) and crushed-rock particles in some island groups. Terrigenous grain types in Oceanian temper sands are subdivided by petrographic analysis into three main groups: light mineral grains including quartz and feldspars, heavyferromagnesian mineral grains including opaque iron oxides and ferromagnesian silicates, and a variety of polycrystalline lithic fragments that are dominantly of volcanic derivation in most temper suites. Useful triangular compositional diagrams plot relative proportions of grain types for populations of total terrigenous grains, mineral grains exclusive of lithic fragments, ferromagnesian silicate mineral grains, all non-ferromagnesian grains, only transparent mineral grains, and exclusively quartz and feldspar mineral grains. Supplemental grain parameters or indices express ratios of grain types among quartz and feldspar mineral grains, ferromagnesian grains, and lithic fragments. Oceanic basalt tempers are mineralogically simple volcanic sands derived from basaltic to basanitic volcanic assemblages of intraoceanic hotspot chains erupted in the interior of the Pacific plate in the eastern Caroline Islands, along the northern Melanesian borderland, in Samoa and American Samoa, and in the Marquesas Islands. Andesitic arc tempers are volcanic sands displaying more compositional variability and are the most abundant tempers within the region of Oceanian ceramic cultures, occurring along island arcs flanking the Philippine Sea plate, bounding the Banda Sea in eastern Indonesia, within the Bismarck Archipelago east of New Guinea, along the reversed-polarity Solomon and Vanuatu arcs, on the Fiji platform and the Lau remnant arc, and in Tonga. Postarc and backarc volcanic sand tempers, variously displaying affinities with both oceanic basalt and andesitic arc tempers, are known from the Bismarck Archipelago, the Vanuatu backarc region, the Horne Islands of the northern Melanesian borderland, and both the Fiji platform and the Lau remnant arc. All volcanic sand tempers of Pacific Oceania are composed of phenocrystic mineral grains and volcanic lithic fragments. Most are quartz-free or quartz-poor, but quartzose variants are present locally along island arcs where silicic eruptions accompanied more typical andesitic to basaltic activity, and within backarc settings where bimodal igneous assemblages are exposed. Most quartzose Oceanian temper sands are either dissected orogen tempers containing dominantly igneous but not exclusively volcanic detritus, or tectonic highland tempers containing recycled sedimentary detritus. Dissected orogen tempers with quartz-ose plutonic detritus occur in selected sherd suites from the Torres Strait Islands, the Bismarck Archipelago, and the Solomon Islands, but are especially characteristic from the south coast of Viti Levu in Fiji. Quartzose tectonic highland tempers occur in sherds from the outer Banda arc, the Aru Islands in the Arafura Sea, the D'Entrecasteaux Islands of the Solomon Sea, and New Caledonia. Nonquartzose tectonic highland tempers derived from ophiolitic rocks of uplifted oceanic crust are present in sherds from Yap and New Caledonia. Comparisons of temper compositions among temper classes indicate that oceanic basalt and basaltic backarc tempers contain significantly higher proportions of olivine mineral grains than arc and postarc tempers, which include a varied array of temper types containing different proportions of pyroxenes and hornblendes. Dissected orogen and quartzose andesitic arc tempers display varying proportions of quartz, plagioclase, and K-feldspar within the compositional field typical for circum-Pacific orogenic sands. Tectonic highland tempers contain distinctly higher proportions of nonigneous lithic fragments than other temper classes. The presence of exotic sherds containing temper sands incompatible with the geology of the islands from which they were recovered documents 106 instances of ceramic transfer between different islands, mostly lying within the same island groups, but also between island groups lying far apart. Two-thirds of the instances of ceramic transfer involved interisland distances of less than 200 km, and most of the remainder involved distances in the range of 200–600 km, but a few cases of ceramic transfer for 1000 km or more are known from temper analysis.