Domain of validity of the Lee–Levitov mechanical universality theorem for phyllotaxis: explicit characterization of the admissible potential family

The mechanical theory of phyllotaxis of H.-W. Lee and L. Levitov proves that, for a family of repulsive pair potentials $U(r)$ on a growing cylindrical lattice, a continuous trajectory of energy minima traverses parastichy domains whose spiral counts form a generalized Fibonacci sequence — but the family is defined only implicitly, by two conditions (continuity/stability of the deformation and 'minimal topography' of the modular-invariant energy $E(z)$, i.e. no critical points beyond the triangular and square symmetry points), and the authors state that an explicit characterization of this family is lacking. The open problem posed by the source is exactly this characterization: explicit, verifiable conditions on $U(r)$ — necessary and sufficient if possible, or a counterexample repulsive potential — under which the universality theorem holds, settling how generally energy minimization under growth stress forces Fibonacci order. Two further blind spots of the theory — the dynamics of the core/creation region in disk and spiral geometries, whose study the source declares beyond the scope of the paper, and a quantitative probabilistic theory of the observed low frequencies of non-Fibonacci (mainly Lucas-type) exceptions, which the source treats only through a qualitative noise-triggered explanation — are documented in the Background section but are not part of the posed problem.

Background

Phyllotaxis is the regular arrangement of lateral plant elements — leaves around a stem, florets in a sunflower head, scales on a pine cone. Visible spiral rows of such elements are called parastichies, and in the vast majority of plants the numbers of clockwise and counter-clockwise parastichies are consecutive Fibonacci numbers $1,1,2,3,5,8,13,21,\dots$ (each term the sum of the previous two). The most common exceptions follow the Lucas sequence $1,3,4,7,11,18,\dots$, which obeys the same addition rule with different starting numbers; both are instances of a generalized Fibonacci sequence, any integer sequence satisfying $\Phi_{n+1}=\Phi_n+\Phi_{n-1}$. A generalized Fibonacci sequence starting from a pair $(m,n)$ is denoted $\langle m,n\rangle$; the standard Fibonacci sequence is $\langle 1,1\rangle$ and the Lucas sequence is $\langle 1,3\rangle$.

A cylindrical lattice is an infinite set of points on a cylinder generated by two basis vectors, one wrapping around the circumference and one advancing along the axis. Lee and Levitov model phyllotactic growth as such a lattice of repulsively interacting particles (the structural units of the plant) under a gradually increasing uniaxial stress, mimicking growth in a confined volume. Each lattice is parametrized by a complex number $z=x+iy$ in a fundamental domain of the upper half-plane, where $x$ is a normalized twist per growth step and $y$ is an aspect ratio of the lattice cell; equivalently the relevant observables are the twist $x$, a relative radial increment $y$, and a cell area $A$. The total pair energy $E(z)$, summed over the lattice for a given potential $U(r)$ and fixed cell area, is invariant under the modular group $SL(2,\mathbb{Z})$ acting on $z$, so the energy topography on the entire upper half-plane is obtained from a single Farey triangle — a triangle of the Farey tessellation of the hyperbolic plane whose vertices are rational points $m/n$ related by Farey addition $m/n\oplus m'/n'=(m+m')/(n+n')$ (the operation that generates the Stern–Brocot tree of rationals and underlies continued fractions). Each Farey triangle corresponds to a parastichy domain: a region of the $z$ plane in which a given pair $(m,n)$ of parastichy numbers, determined by the two shortest lattice vectors, is visible. The triangular ($\triangle$) and square ($\square$) points are the symmetry points of the Farey triangulation at which three or four parastichy domains of equal energy meet; they are the generic locations of energy minima and bifurcations.

Levitov's energetic approach (1991), refined in the Lee–Levitov review, established the main theorem: for interactions $U(r)$ in a certain family, a continuous trajectory of local energy minima in the $z$ plane — the path traced as growth-induced stress deforms the lattice — passes through parastichy domains whose pairs belong to a single generalized Fibonacci sequence, fixed by the beginning of the trajectory. Since growth of a disk-shaped plant head starts near the $(1,1)$ or $(1,2)$ domain, the observed sequence is the standard Fibonacci one. The theorem's potential family is specified only by two implicit conditions: (a) the deformation is continuous and never destabilizes the lattice, and (b) the topography of $E(z)$ is 'minimal', i.e. the simplest allowed by modular symmetry, with no critical points other than the $\triangle$ and $\square$ symmetry points. The authors note that this family 'apparently' includes all repulsive potentials, that they 'failed to find exceptions among repulsive interactions', but that 'an explicit characterization of this family is lacking'. Condition (b) in particular has never been verified in general: it is a property of the critical-point structure of a modular-invariant lattice sum whose verification for a given $U(r)$ requires controlling the energy landscape over the whole hyperbolic plane.

The theorem also has two structural blind spots. First, in disk and spiral geometries (sunflower heads), the analysis applies only outside a core region near the growth center where new units are created and no local lattice can be defined; the core sets the initial conditions from which the deformation starts, and hosts the early-stage 'jumps' between trajectories, but its dynamics is explicitly declared beyond the scope of the theory. Second, the theory is deterministic: it explains qualitatively why Lucas-type exceptions are the most common deviations (a single noise-triggered mistake at an early growth stage moves the trajectory onto the $\langle 1,3\rangle$ sequence), but real non-Fibonacci patterns occur rarely — at the few-percent level and below, as quantified by the large 2016 citizen-science sunflower survey — and no quantitative probabilistic theory predicting these frequencies from specified noise and growth-rate statistics exists within the energy framework.

Adjacent but distinct research traditions frame, without answering, these gaps. Douady and Couder demonstrated phyllotactic self-organization in ferrofluid-droplet experiments and in dynamical inhibition models; Atela, Golé and Hotton gave a rigorous dynamical-systems analysis of pattern transitions in the geometric tradition begun by van Iterson's 1907 construction of the bifurcation tree of successive disk-packings on a cone; Pennybacker and Newell and Golé, Dumais and Douady analyzed pushed pattern-forming fronts (fronts whose speed is set by the nonlinear state behind them, not the linear instability) and disk-stacking (Schwendener-type) placement rules; Mirabet et al. studied noise robustness of models based on inhibitory fields of auxin (the plant growth hormone) transport; Swinton and coworkers validated disk-stacking models against the sunflower statistics; Nisoli et al. verified phyllotactic ground states for magnetic-dipole repulsion in a 'magnetic cactus' experiment while noting that Levitov's energy assumption remained unproved; and Flack and Nechaev analyzed the renormalization-group flow of Levitov's model for the special conformal $1/r^2$ potential, conjecturing BKT-type (Berezinskii–Kosterlitz–Thouless-type, i.e. infinite-order, topological) critical behavior at strong compression. None of these characterizes the admissible potential family, models the energy-framework core dynamics, or derives the exception statistics.

Problem Statement

Determine the domain of validity of the Lee–Levitov mechanical universality theorem for phyllotaxis. Concretely, for the cylindrical-lattice energy model with modular-invariant lattice energy $E(z)$, $z=x+iy$, at fixed cell area, characterize explicitly the family of repulsive pair potentials $U(r)$ for which the theorem's two defining conditions hold, taking the class to consist of nondegenerate interactions ($U$ smooth, strictly positive, and strictly decreasing on $(0,\infty)$; degenerate potentials such as compactly supported or zero-range interactions are outside the family): (a) the growth-driven deformation is continuous and stability-preserving, and (b) the energy topography is minimal, i.e. $E(z)$ has no critical points other than the triangle ($\triangle$) and square ($\square$) symmetry points of the Farey triangulation. A complete answer gives verifiable necessary and sufficient conditions on $U(r)$ (the Lee–Levitov text conjectures the family 'apparently includes all repulsive potentials'); a valid alternative answer exhibits an explicit repulsive potential violating condition (b). A counterexample to condition (b) settles that the family of potentials satisfying conditions (a) and (b) does not include all repulsive potentials; the complete characterization must settle for which interactions every continuous energy-minima trajectory visits only parastichy domains belonging to a single generalized Fibonacci sequence $\langle m,n\rangle$ — which, by the theorem, is the family defined by conditions (a) and (b).

Related aspects documented in the source but not part of the posed completion: the core/creation region of disk and cone geometries — the region near the growth center where new units are created and no local lattice is defined — is explicitly declared 'beyond the scope' of the source paper, which assumes only that the core 'has essentially no effect on the lattice structure far away' and that its role is 'merely in setting right initial conditions' for the deformation; and the statistics of Lucas-type exceptions is treated in the source as qualitatively explained by a single early noise-triggered jump onto the $\langle 1,3\rangle$ sequence. The source work's fourth stated open issue — identifying which physiological mechanisms in actual plants supply the effective repulsion and anisotropic stress, across species — is an open-ended experimental program rather than a question with a determinate completion boundary; it is documented in the source but is not part of the posed completion.

References

1. H.-W. Lee and L. Levitov, 'Universality in Phyllotaxis: a Mechanical Theory', in Symmetry in Plants, Series in Mathematical Biology and Medicine (World Scientific, 1998), DOI 10.1142/9789814261074_0024, https://doi.org/10.1142/9789814261074_0024; posted as arXiv:2101.02652, DOI 10.48550/arXiv.2101.02652, https://arxiv.org/abs/2101.02652
2. L. S. Levitov, 'Energetic Approach to Phyllotaxis', Europhysics Letters 14, 533–539 (1991), DOI 10.1209/0295-5075/14/6/006, https://doi.org/10.1209/0295-5075/14/6/006
3. L. S. Levitov, 'Phyllotaxis of flux lattices in layered superconductors', Physical Review Letters 66, 224–227 (1991), DOI 10.1103/PhysRevLett.66.224, https://doi.org/10.1103/PhysRevLett.66.224
4. S. Douady and Y. Couder, 'Phyllotaxis as a physical self-organized growth process', Physical Review Letters 68, 2098–2101 (1992), DOI 10.1103/PhysRevLett.68.2098, https://doi.org/10.1103/PhysRevLett.68.2098
5. P. Atela, C. Golé and S. Hotton, 'A Dynamical System for Plant Pattern Formation: A Rigorous Analysis', Journal of Nonlinear Science 12, 641–676 (2003), DOI 10.1007/s00332-002-0513-1, https://doi.org/10.1007/s00332-002-0513-1
6. C. Nisoli, N. M. Gabor, P. E. Lammert, J. D. Maynard and V. H. Crespi, 'Static and Dynamical Phyllotaxis in a Magnetic Cactus', Physical Review Letters 102, 186103 (2009), DOI 10.1103/PhysRevLett.102.186103, https://doi.org/10.1103/PhysRevLett.102.186103
7. C. Nisoli, N. M. Gabor, P. E. Lammert, J. D. Maynard and V. H. Crespi, 'Annealing a Magnetic Cactus into Phyllotaxis', Physical Review E 81, 046107 (2010), DOI 10.1103/PhysRevE.81.046107, https://doi.org/10.1103/PhysRevE.81.046107
8. V. Mirabet, F. Besnard, T. Vernoux and A. Boudaoud, 'Noise and Robustness in Phyllotaxis', PLoS Computational Biology 8, e1002389 (2012), DOI 10.1371/journal.pcbi.1002389, https://doi.org/10.1371/journal.pcbi.1002389
9. M. Pennybacker and A. C. Newell, 'Phyllotaxis, Pushed Pattern-Forming Fronts, and Optimal Packing', Physical Review Letters 110, 248104 (2013), DOI 10.1103/PhysRevLett.110.248104, https://doi.org/10.1103/PhysRevLett.110.248104
10. C. Golé, J. Dumais and S. Douady, 'Fibonacci or quasi-symmetric phyllotaxis. Part I: why?', Acta Societatis Botanicorum Poloniae 85(4), 3533 (2016), DOI 10.5586/asbp.3533, https://doi.org/10.5586/asbp.3533
11. J. Swinton, E. Ochu and the MSI Turing Sunflower Advisory Group, 'Novel Fibonacci and non-Fibonacci structure in the sunflower: results of a citizen science experiment', Royal Society Open Science 3, 160091 (2016), DOI 10.1098/rsos.160091, https://doi.org/10.1098/rsos.160091
12. T. Yonekura, A. Iwamoto, H. Fujita and M. Sugiyama, 'Mathematical model studies of the comprehensive generation of major and minor phyllotactic patterns in plants with a predominant focus on orixate phyllotaxis', PLoS Computational Biology 15(6), e1007044 (2019), DOI 10.1371/journal.pcbi.1007044, https://doi.org/10.1371/journal.pcbi.1007044
13. A. Mughal and D. Weaire, 'Phyllotaxis, disk packing, and Fibonacci numbers', Physical Review E 95, 022401 (2017), DOI 10.1103/PhysRevE.95.022401, https://doi.org/10.1103/PhysRevE.95.022401
14. A. Flack and S. Nechaev, 'BKT in Phyllotaxis', arXiv:2310.05580 (2023), https://arxiv.org/abs/2310.05580; published in Russian as 'BKT transition in phyllotaxis' by S. K. Nechaev, Zhurnal Eksperimental'noi i Teoreticheskoi Fiziki 166, 277–289 (2024), DOI 10.31857/s0044451024080133, https://doi.org/10.31857/s0044451024080133
15. A. Flack, A. Gorsky and S. Nechaev, 'Generalized Devil's staircase and RG flows', arXiv:2304.07640 (2023), https://arxiv.org/abs/2304.07640
16. J. Swinton, 'Disk-stacking models are consistent with Fibonacci and non-Fibonacci structure in sunflowers', arXiv:2407.05857 (2024), https://arxiv.org/abs/2407.05857
17. J. Swinton, 'Mathematical Phyllotaxis', Surveys and Tutorials in the Applied Mathematical Sciences vol. 17 (Springer, Cham, 2025), DOI 10.1007/978-3-031-94013-2, https://doi.org/10.1007/978-3-031-94013-2
18. G. Facchini, M. A. Budroni, G. Schuszter, F. Brau and A. De Wit, 'Phyllotactic Structures in Radially Growing Spatial Symmetry Breaking Systems', Physical Review Letters 135, 018001 (2025), DOI 10.1103/7tnn-pq49, https://doi.org/10.1103/7tnn-pq49
19. M. F. Staddon, 'Phyllotaxis in a Keller-Segel model', Physical Review E 113, 034407 (2026), DOI 10.1103/9yk5-b4q2, arXiv:2509.06498, https://arxiv.org/abs/2509.06498