Seeds that predict weather
But I don't mean the persimmon way, charming as it is. Ah, the Old Farmer's Almanac - I do have a soft spot for it. Well I remember Christmas 1968, when I got it as a gift, and also listened to the Apollo 8 crew read from the Book of Genesis - that was a not merely a great Christmas but a Christmas with greatness. May every child have his or her One. Later, of course, I and my brothers would crack all manner of jokes about the OFA, especially about its numerous truss ads. I am still not sure what a truss even is. (My brothers are long-and-well-married and so I expect they don't know either.) But by way of a guess, I'll toss out a bedroom-comedy idea, and it goes: "What the hell's THAT!" "It's my truss."
But back, with difficulty, to the topic at hand. I think with some elementary statistics, it might be shown that certain seeds restrain themselves from sprouting unless they are convinced there will be enough rain to sustain their sprouts through growth, flowering, and new seed formation. Exactly how any seed could do that is not the point here. The point here is to establish the two things that would justify a search for the how. Those things are
I think one can reasonably restrict his weather-data searches to places near the Equator, and to places with soil that neither gets nor holds much water. (Only because I've been there, I wonder if Cape Verde qualifies. I mean the archipelago countrylet in the Atlantic, not the pointy west end of Senegal. Another possibility: the desert shores of northern Peru.) The idea is to ensure one's seeds aren't responding to distant, persistent stimuli from the past, or to stimuli which issue like solar clockwork; rather, they have nothing to go on but what's happening NOW...and what might happen later. Anyway, the information to check first is rain dates, rain amounts, soil saturation, and soil drying: if the periods that start with now-too-dry soil and end with now-wet-enough soil are themselves random, and also large fractions of the plant's lifespan, then we've at least started.
With such information in hand, one may then look at germination rates. If the plant of interest is growing in an area where rainfall really is spotty, then I imagine it's a plant that grows very opportunistically and quickly, yielding a great deal of seed which is prepared to lie dormant for long periods. There are of course germination strategies more plausible than clairvoyance. Seed could have highly randomized sensitivity to wettings, with only a small fraction responding to a single exciting shower, another fraction insisting on two, or three, etc. Or, none may germinate unless there is not merely a shower but a downpour. I think one must gather seed from a cohort of plants, then sow each one in separate open-air containers and note, every day over a period of several lifespans, percent of seeds germinated AND (obviously in retrospect) amount of rain received over the ensuing lifespan AFTER the first thorough dry spell AFTER germination. Plot these germination percentages versus rain amounts. The ideal is a signum function: 0% germination below a critical future rainfall amount, 100% above it.
I wonder what the statistical test for that is. Probably two tests, one for "before" and the other for "after."
UNDER CONSTRUCTION!
[Upon mature consideration]...I'm not constructing anything like it. Now that I've apprehended the poor definability of "random." I think what I was looking for was "patternlessness," or as it may be more formally termed, maximum Kolmogorov complexity. But even if a series of drought intervals was so "random" that the briefest way to describe it would be to recite it, there would still be a boundary to its wanderings that would be useful to a plant: the historical maximum drought length. If you can survive - or wait - that long, you're going to succeed. As, indeed, you evidently have.