The development of mass spectrometric techniques for determination of 230Th abundance has made it possible to reduce analytical errors in 238U-234U-230Th dating of corals even with very small samples. Samples of 6 × 108 atoms of 230Th can be measured to an accuracy of ±3 percent (2σ) and 3 × 1010 atoms of 230Th can be measured to an accuracy of ±0.2 percent. The time range over which useful age data on corals can be obtained now ranges from about 50 to about 500,000 years. For young corals, this approach may be preferable to 14C dating. The precision with which the age of a coral can now be determined should make it possible to critically test the Milankovitch hypothesis concerning Pleistocene climate fluctuations. Analyses of a number of corals that grew during the last interglacial period yield ages of 122,000 to 130,000 years. The ages coincide with, or slightly postdate, the summer solar insolation high at 65°N latitude which occurred 128,000 years ago. This supports the idea that changes in Pleistocene climate can be the result of variations in the distribution of solar insolation caused by changes in the geometry of the earth's orbit and rotation axis.