Photoperiodic induction of diapause in the spider miteTetranychus urticae is the net result of at least two processes: time measurement (the photoperiodic ‘clock’) and the accumulation of the photoperiodic information contained in a sequence of light-dark cycles (the photoperiodic ‘counter’). In this paper an analysis is presented of the operation of the photoperiodic counter in the spider mite. 1. Mites which experienced a sequence of longnight cycles during their entire sensitive period showed 100% diapause; no diapause was observed in continuous darkness. When an increasing number of long-night cycles was applied to the mites against a ‘background’ of continuous darkness, diapause incidence was found to rise steadily: only 3 cycles sufficed to induce diapause in about half the population, whereas a minimal number of 6 cycles was required for 90–100% diapause to be attained. At the test temperature of 18.5°C the sensitive period lasted 11–12 days, comprising the complete post-embryonic developmental period, up to the final moult. Photoperiodic sensitivity was found to vary slightly over the whole sensitive period of the mites, the highest sensitivity being observed around days 3–6. 2. Short-night cycles were also shown to be accumulated, but with an effect opposite to that of long-night cycles. If the mites received a number of short-night cycles before being transferred to a long-night regime, the effect of the short-night cycles had to be levelled first by a number of longnight cycles, before the accumulation of the diapause-inducing effect of the long-night cycles was started. 3. Special attention has been given to the effect of the aperiodic signals continuous light and continuous darkness. It could be shown that continuous light has a slightly reversing effect on diapause induction if applied after a series of long-night cycles; continuous light is more or less ‘neutral’ (i.e. neither reversing nor promoting diapause induction) if it precedes the long-night cycles. 4. In experiments in which the effects of continuous light and continuous darkness were compared it could be shown that continuous darkness is equivalent to one long night: the minimal number of long-night cycles required for diapause induction was found to be one more if the long-night cycles were given after the mites had received continuous light instead of continuous darkness during the first part of the sensitive period. 5. A critical test, based on the photoperiodic counter principle and devised to discriminate between single and repeated nightlength measurements (using nights of 12 h and 36 h long), showed that all nights are counted only once, irrespective of their lengths: all nights longer than the critical nightlength were found to be about equally inductive. This shows that the photoperiodic clock in the spider mite does not operate according to oscillator kinetics: a clock of the oscillator type resets itself in longer dark phases and would have performed two consecutive acts of time measurement in a night of 36 h long. Consequently, the photoperiodic clock inT. urticae is either an hourglass or an instantly damped circadian oscillator, the kinetics of either of which would produce the results observed in the experiments reported here.