It is widely assumed that degradation of mitotic cyclins causes a decrease in mitotic cdc2/CDC28 kinase activity and thereby triggers the metaphase to anaphase transition. Two observations made on the budding yeast Saccharomyces cerevisiae are inconsistent with this scenario: (i) anaphase occurs in the presence of high levels of kinase in cdc15 mutants and (ii) overproduction of a B-type mitotic cyclin causes arrest not in metaphase as previously reported but in telophase. Kinase destruction is therefore implicated in the exit from mitosis rather than the entry into anaphase. The behaviour of esp1 mutants shows in addition that kinase destruction can occur in the absence of anaphase completion. The execution of anaphase and the destruction of CDC28 kinase activity therefore appear to take place independently of one another.