In Saccharomyces cerevisiae, START has been shown to comprise a series of tightly regulated reactions by which the cellular environment is assessed and under appropriate conditions, cells are commited to a further round of mitotic division. The key effector of START is the product of the CDC28 gene and the mechanisms by which the protein kinase activity of this gene product is regulated at START are well characterized. This is in contrast to the events which follow p34CDC28 activation and the way in which progress to S phase is achieved, which are less clear. We suggest two possible models to describe the regulation of these events. Firstly, it is conceivable that the only post-START targets of the p34CDC28/G1 cyclin kinase complex are components of the SBF and DSC1 transcription factors. This would require that either SBF or DSC1 regulates CDC4 function either directly by activating the transcription of CDC4 itself or else indirectly by activating the transcription of a mediator of CDC4 function in a manner analogous to the way in which the control of CDC7 function may be mediated by transcriptional regulation of DBF4 (Jackson et al., 1993). Potential regulatory effectors of CDC4 function include SCM4, which suppresses cdc4 mutations in an allele-specific manner (Smith et al., 1992) or its homologue HFS1 (J. Hartley & J. Rosamond, unpublished). This possibility is supported by the finding that CDC4 has no upstream SCB or MCB elements, whereas SCM4 and HFS1 have either an exact or close match to the SCB. This model would further require that genes needed for bud emergence and spindle pole body duplication are also subject to transcriptional regulation by DSC1 or SBF. An alternative model is that the p34CDC28/G1 cyclin complexes have several targets post-START, one being DSC1 and the others being as yet unidentified components of the pathways leading to CDC4 function, spindle pole body duplication and bud emergence. This model could account for the functional redundancy observed amongst the G1 cyclins with the various cyclins providing substrate specificity for the kinase complex. We suggest that a complex containing Cln3 protein is primarily responsible for, and acts most efficiently on, the targets containing Swi6 protein (SBF and DSC1), with complexes containing other G1 cyclins (Cln1 and/or Cln2 proteins) principally involved in activating the other pathways. However, there must be overlap in the function of these complexes with each cyclin able to substitute for some or all of the functions when necessary, albeit with differing efficiencies. This hypothesis is supported by several observations.(ABSTRACT TRUNCATED AT 400 WORDS)