Comparative model tests on capsizing of container ships navigating in quartering seas have been performed. Two types of the hull form are used. One has a large stability variation in waves and the other has a small one. The model tests are carried out for both irregular and regular waves in the 80 m square basin of the Ship Research Institute. The parameters of the experiment are the GM value, the model speed and the encounter angle with the wave for the tests in the irregular wave, and also the model speed, the encounter angle, the wave length and the wave height for the tests in the regular waves. The encounter angle is varied precisely every 10 degrees from 0° (following wave) to 90° (beam wave) by using the auto-pilot steering device in order to specify the dangerous range. The Froude number of the model speed is varied every 0.01 to clarify the critical speed for the capsizing. The length and height of the regular wave are also varied in a fine interval to examine the tendency of the capsizing for the various waves. Among the total 1643 runs, 490 capsizings have been observed. From these results the following three characteristic features are confirmed to be common features for any type of ships. Firstly, the most dangerous encounter angle is 20° to 40° or 50°, and the range spreads to 0° to 70° as the ship speed increases. Secondly, as the ship speed decreases, the capsizing also decreases, and there exists a critical ship speed below which the capsizing never occur. Thirdly, the direction of the capsizing is always to the lee side. Although these results are almost the same as the previous model tests for an unusual hull form, which has an extraordinary tendency of inward heel in turning, it has been considered necessary to reconfirm that the above features of capsizing are common for more general hull forms. A previously proposed hypothetical explanation on the mechanism of capsizing in quartering seas can also explain the present results. The capsizing due to the parametric rolling, which has been regarded as one of the typical mode of dynamical capsizing, has never been observed in both model tests. Instead, it has been also reconfirmed that there is a new mode of dynamical capsizing accompanied with the period bifurcation phenomenon, which is regarded as a precursor of the chaos.