A self-contained, Fourier-matched derivation is presented that extends the classical Jackson–Hunt lamellar spacing-selection law from binary eutectic solidification to multicomponent solid-state eutectoid transformations in steels. The resulting selection law, \(V\lambda _*^2 = B/A\) , retains the elegant structure of the classical result while embedding full multicomponent thermodynamic content. The selection constant B/A depends explicitly on the substitutional composition \(\varvec{x}_\infty \) through the conditional solvus-slope difference \(g_\Delta (\varvec{x}_\infty )\) , the carbon chemical diffusivity \(D_{\text{C}}^\gamma (\varvec{x}_\infty )\) , and the capillarity parameter \({\mathcal {G}}(\varvec{x}_\infty )\) , thereby providing a direct, closed-form link between alloy chemistry and pearlite microstructural refinement. The derivation exploits a dual Péclet-number ordering—carbon fast, substitutionals frozen at the lamellar scale—to reduce the multicomponent diffusion problem to a single-species Fourier solution, while the parametric influence of every substitutional alloying element is preserved through conditional phase-equilibrium quantities. Capillarity (Gibbs–Thomson) undercooling is treated through trijunction force balance and circular-arc curvature, interface mobility (kinetic) undercooling is incorporated as an additive contribution, and a framework for boundary-diffusion corrections that interpolates smoothly between the volume-diffusion ( \(V\lambda ^2\) ) and boundary-diffusion ( \(V\lambda ^3\) ) limits is presented. Comparison with four independent experimental datasets spanning 0 to 3.5 wt pct Mn in Fe–C and Fe–C–Mn steels demonstrates that the theory captures approximately Mn-induced growth rate retardation without adjustable parameters, correctly ranks the alloy systems, and identifies the transition to partitioning pearlite as the natural model boundary. A Robin boundary condition extension is outlined that provides a clear path toward quantitative treatment of the partitioning regime.