The Chemical Basis of Morphogenesis

How Two Diffusing Molecules Carve Leopard Spots, Zebra Stripes, and Fibonacci Spirals

Drop a bead of blue ink into a still glass of water and Fick's law of diffusion does only one thing. Random thermal motion drives ink molecules from regions of high concentration to regions of low concentration until the water turns a uniform, featureless pale blue. For centuries, that physical law posed a deep paradox for embryology. A fertilized animal egg begins its development as a nearly spherical, genetically uniform ball of cells called a blastula. Every cell carries the exact same genes, and chemical diffusion normally smooths out concentration differences. How, then, does a homogeneous embryo spontaneously break its own rotational symmetry to sprout six tentacles on a freshwater Hydra, stripes on a zebra, or five fingers on a limb bud?

After moving to the University of Manchester in 1948 and winning election as a Fellow of the Royal Society in March 1951 (nominated by Max Newman and Bertrand Russell), Alan Turing turned his attention from logic gates to mathematical biology. Even during the harrowing months of early 1952, he pressed forward with the calculations. Manchester police arrested him in early 1952 for a consensual relationship with another man under Victorian gross indecency statutes. The resulting March 1952 conviction stripped his security clearance and forced him to endure a year of synthetic estrogen injections at his home in Wilmslow, Cheshire. Throughout that ordeal, Turing worked night shifts on the newly installed Manchester Ferranti Mark 1 computer to finish his landmark study in biological physics.

Published in the August 1952 issue of the Philosophical Transactions of the Royal Society of London, Series B under the title The Chemical Basis of Morphogenesis, the paper introduced the concept of morphogens. These are diffusing chemical signaling molecules, such as hormones or skin pigments, that react with one another and instruct embryonic cells how to differentiate. In a continuous sheet or ring of tissue, Turing modeled two interacting morphogens with concentrations $u(x, t)$ and $v(x, t)$ governed by coupled reaction-diffusion partial differential equations:

Here $f(u, v)$ and $g(u, v)$ represent the local chemical production and degradation rates inside each cell. The operator $\nabla^2$ is the spatial Laplacian measuring curvature in concentration, and $D_u$ and $D_v$ are the physical diffusion coefficients of the two molecules across cell membranes.

Turing discovered a counterintuitive mathematical instability—now called a Turing or diffusion-driven instability. Adding diffusion to a locally stable chemical mixture can destabilize uniformity and carve stationary spatial waves. Suppose morphogen $u$ is a short-range autocatalytic activator that stimulates both its own synthesis ($\frac{\partial f}{\partial u} > 0$) and the synthesis of a second molecule $v$ ($\frac{\partial g}{\partial u} > 0$). Morphogen $v$, in turn, acts as an inhibitor that suppresses $u$ ($\frac{\partial f}{\partial v} < 0$) and degrades itself ($\frac{\partial g}{\partial v} < 0$).

In a well-stirred beaker without spatial diffusion ($D_u = D_v = 0$), the homogeneous equilibrium $(u_0, v_0)$ is completely stable, and any perturbation dies back to flatness. Suppose, however, that the inhibitor molecule $v$ diffuses through embryonic tissue significantly faster than the activator $u$ ($D_v > D_u$). When random thermal noise creates a tiny local peak of activator $u$, the slow activator stays trapped at that spot and amplifies itself. Simultaneously, the newly created inhibitor $v$ races outward into the surrounding tissue and prevents any neighbor peaks from forming within a characteristic inhibition radius.

By expanding spatial perturbations into Fourier modes $\cos(kx)$, Turing proved that a narrow band of spatial wavenumbers $k$ acquires a positive real eigenvalue. That dominant mode grows exponentially until chemical saturation locks in a periodic stationary wave of wavelength $\lambda = 2\pi / k_{\max}$. On a narrow cylinder, those waves form striped rings; on a two-dimensional skin surface—computed numerically by Turing on the Ferranti Mark 1—they break into mottled blotches and leopard spots. On a growing plant apex, which Turing analyzed in Wilmslow until his death on June 7, 1954, those coupled waves generate the Fibonacci parastichy spirals ($F_n \in \{2, 3, 5, 8, 13, 21, \dots\}$) of sunflower heads and fir cones. Four decades after Turing solved those coupled reaction-diffusion equations on the Manchester Mark 1, experimental chemists in 1990 and developmental biologists studying vertebrate skin and digit spacing confirmed stationary Turing waves operating inside living embryos.