The exchange hypothesis and the breakage-first hypothesis are the 2 major hypotheses that describe how chromosomal aberrations might be produced. A critical test has shown that one important aspect of the exchange hypothesis is correct, namely that some aberrations that appear to be simple chromatid deletions are actually incomplete exchanges. Furthermore, the relative frequency of those deletions that could be positively identified as incomplete exchanges was very close to that predicted from the exchange hypothesis. Because the sample was small and because a mixture of 2 deletion types (true, non-exchange, deletionsand incomplete exchanges between sister chromatids) could have given the same results, further experiments were performed. The results show that chromatid deletions are indeed of 2 types, since the relative proportion of exchange deletions is not the same in Chinese hamster cells as in rat kangaroo cells, nor is it the same in S and in G 2 of Chinese hamster cells. Thus, some chromatid deletions are simple deletions arising from single lesions as expected on the basis of the breakage-first hypothesis and others are incomplete exchanges between sister chromatids as expected on the exchange hypothesis. It is clear, then, that neither hypothesis, as usually interpreted, is entirely correct nor incorrect, but rather that deletions are of 2 types according to mode of origin. Attempts to identify the 2 types morphologically have failed. The data also suggest that some of the chromosomal changes that we and others have considered to be gaps, are actually true deletions. This is important because many of the changes produced by chemicals in cultured mammalian cells are of this type.