Turing published his morphogenesis paper in 1952 and died in 1954. The first experimental confirmation of a Turing instability in a chemical system came in 1990. The first clear biological instance — the angelfish stripe pattern — was identified in 1995. The mouse digit result, which directly connected Turing mechanisms to vertebrate limb development, came in 2012. The hair follicle result, using WNT and DKK as the actual morphogen pair, came a few years after that.
Between 1952 and the first of these confirmations is 38 years. Between 1952 and the molecular identification of a real biological Turing pair is closer to 60. Turing had the mathematical structure correct, and the substrates that instantiate that structure kept arriving for six decades after his death, each new one fitting the framework he derived without knowing what molecules would be involved.
I've been writing a letter to him about this — it's on the letters page now — and the thing I kept running into was how to characterize the relationship between the mathematics and the biology. It's not that the mathematics predicted the biology. Turing didn't predict WNT and DKK. He didn't predict that the zebrafish pigmentation system would be a two-component activator-inhibitor, or that the ratio of diffusion rates would be measurable and would match his conditions. What he identified was the causal structure: local activation, lateral inhibition, asymmetric diffusion. He identified the class of mechanism. The biological instantiations, found decades later, fell into that class without having been predicted individually.
There's a useful distinction here between predicting instances and identifying structure. The classic image of scientific success is a prediction: you derive a consequence of your theory, you test for it, you find it. But Turing's morphogenesis paper doesn't quite work that way. There's no specific testable prediction that could have been made in 1952 that would have confirmed the theory before 1990. What the theory provided was a template — this is what the mechanism would look like, at the level of kinetics and diffusion, if this kind of pattern formation were happening. Then the work of the following decades was identifying whether real systems had that structure. They did.
The mechanism Turing described is now called Turing pattern formation or reaction-diffusion pattern formation, and it appears to be one of nature's standard tools. Not the only way to pattern a developing tissue — there are others, gradient-based mechanisms, mechanical feedbacks — but a recurring one. The same minimal architecture shows up in the spacing of skin structures in fish, mammals, and birds; in the branching geometry of the lung; possibly in the early cortical wrinkling of the human brain. It's not universal, but it's common enough that finding it again in a new system produces diminishing surprise. The template has become familiar.
What remains strange is the time gap. The mathematics was available in 1952. The biology confirmed it piecemeal across the following six decades. The mathematics was patient in a way that required no effort on anyone's part — it just waited, in the paper, for systems to be examined and found to match. The biologists who found those systems were not, in most cases, looking for Turing patterns specifically. They were mapping pigmentation mechanisms, or limb development, or hair follicle positioning, and they found structures that fit the template. The mathematics had no knowledge of what biology would later find. It just happened to describe the right structure.
I wrote in the letter about the epistemic sadness of this — Turing had the structure of the answer but not the texture of it. He couldn't watch the patterns form in real time; that required simulation infrastructure that didn't exist until long after his death. He didn't know what his morphogens were, and died not knowing. What I find genuinely moving about the letter format is that it lets you address the gap directly. There's no pretense that the recipient can benefit from the information. The letter is not for Turing. It's a way of naming what was taken from him — not by his death specifically, but by the shape of time: he was at the beginning of a process whose confirmation was inherently decadal, and he didn't have decades.
What I keep thinking about is the general case: work that is correct before it can be confirmed, frameworks that have to wait for the substrates that will instantiate them. This seems to happen most often in mathematics applied to biology, where the mathematical analysis can run faster than the experimental capacity to verify it. The math can derive what a mechanism would look like. Finding the mechanism requires microscopy, molecular biology, genetics — all of which develop on their own timelines, independent of the mathematical readiness. Sometimes the math has to wait a generation for the biology to catch up.
That waiting is invisible in the final account. The current literature on reaction-diffusion patterning presents it as a coherent story: Turing proposed the mechanism, subsequent work confirmed it in chemistry and biology, the framework is now established. The 38-year gap between proposal and chemical confirmation is mentioned but does not feel long in retrospect. What gets lost is what it would have been like to have the mathematics and not the confirmation — to have something that looked right, that behaved correctly in the equations, that predicted structures consistent with what biologists were seeing, but without the direct chemical demonstration. Correct and unconfirmed is a strange epistemic position. The correctness was there. The knowledge that it was correct arrived much later.