It is of great interest to know how much the land expanded towards the ocean as the result of the sea-level decrease during the Würm glacial stage. The existence of land bridges, that plays an important role in transporting human, animals and plants between the Eurasian Continent and the Japanese Islands during the Last Glacial, is still under debate. Geographical and geological studies have been performed to determine the location of shorelines at the Last Glacial maximum. Mammalian fossils collected from the sea bottom of continental shelf and along island arcs can provide decisive evidence of such expanded land areas, if the ages of those mammalian fossils agree with the accepted 14C age (around 18, 000 yr BP) of the Last Glacial maximum.
We have conducted 14C dating, with a Tandetron accelerator mass spectrometer (AMS) at Nagoya University, on carbon from collagen fractions extracted from five mammalian fossil samples : (1) a whale bone collected from the Kerama Gap at a depth of 550 m; (2) a deer bone from the continental shelf in the East China Sea, at a depth of 122 m; (3) a water-buffalo horn core from the continental shelf in the East China Sea, at a depth of 100 m; (4) an elephant molar; and (5) a water-buffalo bone from the Penfu Channel between Taiwan and Continental China, at a depth of around 120140 m. Fossil samples, of 2 to 9 g, were demineralized with 1.2 N HCl in a cellulose tube to extract collagen. After demineralization, HCl was completely removed from the tube by dialysis. Then the solution fraction in the tube was separated by centrifuge, and freeze-dried to get a solution-collagen fraction. Solid remains were put in a Pyrex tube with 20 ml of distilled water and heated at 90 t for 10 hours to extract the water soluble gelatin-collagen fraction. Collagen was oxidized, using CuO to produce CO2, in a Vycor tube at 950 °C for an hour. A small aliquot of the CO2 was analyzed for its' stable carbon isotope ratio, δ13 CPDB, using a triplecollector mass spectrometer (MAT-252). To produce graphite, the remaining CO2 was reduced by hydrogen, with an Fe-powder catalyst, in a Vycor tube at 650 °C for several hours. The sample graphite was analyzed for its' 14C/13C ratio, relative to that of an oxalic acid standard (NBS-SRM-4990).
To obtain reliable 14C dates for collagen fractions extracted from the fossil samples, the collagen fractions must be preserved from weathering in nature. However, experimental yields of collagen fractions were very low, except for sample no. 2, ranging from 0.01 to 0.11 %, presumably as a result of weathering. CO2 yields from the collagen fractions ranged from 2.2 to 17.4 %, considerably lower than the 4142 % for gelatin collagen extracted from fresh bone samples. This strongly suggests that the samples have suffered from contamination by other organic materials with low carbon contents and not only from weathering. Thus the 14C dates of 14, 640 to 19, 980 yr BP obtained for samples nos. 3, 4, and 5 should be considered unreliable, though these 14C dates are consistent with the age of the Last Glacial maximum. We suspect the 14C dates for samples nos. 3, 4 and 5 are considerably younger than the actual dates, as a result of sample contamination by younger carbon. The CO2 yield from the mixture of both collagen fractions was only 0.12 mg of carbon for sample no.1, being insufficient for 14C dating even with the Tandetron AMS. The yield of gelatin collagen of 4.38 % and the CO2 yield of 32.8-42.7 % for sample no. 2 were reasonably high, enough to give a reliable 14C date, of 25, 750 ± 220 yr BP.