A general causal cycle of events is represented in Figure 1(a). In this sequence A causes B which in turn causes C which causes D. Finally D causes event A completing the cycle. A biological application of this model is given in Figure 1(b). This represents four genes, A, B, C and D, of a single-celled organism with RNA or protein gene products a, b, c and d. Gene product a switches on gene B which produces gene product b; b switches on gene C and so on until gene A is switched on again to give product a. Negative feedback effects on synthesis, or dilution, or breakdown of a gene product could prevent its accumulation and thus ensure a true cycle. The period, or time between successive events of the cycle need not be constant if switching also depends on other external or internal forces. In Figure 1(b) the period of the cycle is kept in phase with the cell cycle by assuming that cell division is necessary for synthesis of a. Causal cycles analogous to that of Figure 1(b) are characteristic of living systems where for example the cell cycle consists of a number of distant sequential stages. Figure 1(b) illustrates a general principle, and fewer or more steps in the cycle could be envisaged with other forms of molecular interaction. The cycle of Figure 1(b) stores heritable information apart from that encoded in the genes A, B, C and D, because an external environmental factor or adaptive phenotypic response suppressing synthesis of say gene product b would terminate the sequence of switching events. The cycle would be broken and the phenotype associated with gene products a, b, c and d would be lost. There would be two heritable alternatives, 'cycle on' and 'cycle off' and switching between them achieved by external factors. The cycle might also be broken by a gene mutation and the 'cycle off' alternative could be lethal. Cell cycle events are obviously more complex than those of Figure 1(b) and perhaps form a network of interacting events. Possible implications of increased complexity are shown in Figure 1(c). Here a network of events can be interpreted as three separate cycles, I, II and III. The cycles show inter-dependence because gene products b and g are involved in switching events in more than one cycle. In this situation inhibition of an event in one cycle may subsequently break other cycles. Conversely a break in one cycle may be repaired by later switching on of the cycle by events in other cycles. This is summarised in Table 1 which gives the phenotypes just before cell division resulting from breakage of cycles II and III by inhibition of synthesis of c, f or k within one cell generation.