It would seem to me that no explicit arithmetic is being done - it's simply a consequence of natural behaviour. Consider the case of a starch channel, the size

It would seem to me that no explicit arithmetic is being done - it's simply a consequence of natural behaviour. Consider the case of a starch channel, the size of which is governed by two factors - the quantity of stored starch, and the time since nightfall. (Other factors may also affect the size, but are irrelevant to this discussion). This seems like a construct that could quite feasibly appear in a plant. No arithmetic is being performed by the plant, but simply these two factors working in concert cause the rate of starch consumption to appear to be based on whatever the starch level was when night fell. The fact that the observed rate of consumption is such that nearly all the starch is consumed by morning is easily explained by natural selection - plants that consume starch too fast obviously starve overnight, while plants that consume it too sparingly are using their resources inefficiently and so would tend to be out-competed by plants that make better use of them. Over many generations, the base rate of consumption would become tuned to how long nights last in that region.