The cross-cultural generalizability of two Western intelligence tests, the Stanford-Binet and the WAIS, in a non-Western culture was examined. Samples of 976 Australian and 1003 Chinese university and high school students participated in a rating task. Items from the two tests were given to students, and were rated on the two aspects of relevance and difficulty. On relevance ratings, the underlying three-factor structure of the Chinese sample accords very well with that of the Australian, suggesting a basic common structure for the concept of intelligence across cultures. The three factors were spatial-mechanical, verbal and memory. The results also showed that both cultures consider spatial-mechanical items most important and memory items least important. The two samples differed, however, in mean difficulty scores on the three dimensions. These differences were attributed to possible cultural differences in nurturing and providing opportunities to practice the different skills. It is concluded that the concept of intelligence is comparable between Australian and Chinese cultures, and that, for both cultures, the two tests are legitimate measures of the intelligence construct.