betaFeedback

WorksMathematics

Orbits of pp-adic flows: balls, time-one maps and minimal group actions

Abstract

Let ff be a vector field on Zpd\mathbb{Z}_p^d given by polynomials, or by power series converging on Zpd\mathbb{Z}_p^d, whose coefficients are divisible by pcp^c, where c≥1c \ge 1 (c≥2c \ge 2 when p=2p = 2). Its flow exists for every pp-adic time. I show that t↦φt(x)t \mapsto φ_t(x) multiplies distances by exactly ∥f(x)∥p\|f(x)\|_p, in every dimension. This gives explicit flow-box coordinates, closed formulas for hitting times, return times and orbit averages, and a Newton iteration that computes them with quadratic convergence; in dimension one it recovers Mukhamedov's theorem that interpolated orbits are balls. I then show that the analytic maps gg congruent to the identity modulo pcp^c, coefficientwise, are exactly the time-one maps of such fields, and that the field is congruent to g−idg - \mathrm{id} modulo pc+1p^{c+1}. So the interpolated iterates of Bell and Poonen are flows of this kind, and I obtain a quantitative minimality criterion: if the displacements gi(x0)−x0g_i(x_0) - x_0 of dd such maps span pkZpdp^k\mathbb{Z}_p^d, the group they generate acts minimally on the ball of radius p−kp^{-k} about x0x_0. Where the argument of Cantat and Jang gives an orbit closure containing some open set, this names the ball, at every level kk and in every dimension; for d=2d = 2 and k=1k = 1 it is Jang's criterion for Markoff surfaces, which I test numerically at 78307830 points. Finally, the formal logarithm of Gm\mathbb{G}_m or of an elliptic curve is the rectifying coordinate of an explicit field, so the local divisibility vp([t]z)=vp(z)+vp(t)v_p([t]z) = v_p(z) + v_p(t) for vp(z)≥1+cv_p(z) \ge 1 + c, deep in the kernel of reduction, follows from the metric identity alone.

Introduction

Over the real numbers the flow of dx/dt=f(x)dx/dt = f(x) is usually known only approximately, and only for a bounded time. Over the pp-adic integers the situation is different. If f∈pZp[x1,…,xd]f \in p\mathbb{Z}_p[x_1,\ldots,x_d] (with 4∣f4 \mid f if p=2p = 2), the flow φt\varphi_t exists for every t∈Zpt \in\mathbb{Z}_p and is given by a power series in tt that converges on all of Zp\mathbb{Z}_p, and its geometry can be read off from that series. This paper is about that geometry: what the orbits look like, which maps are time-one maps of such flows, and when a group generated by such maps acts minimally on a ball.

The idea. Write the flow as φt(x)=x+tf(x)+∑n≥2tnEn(x)\varphi_t(x) = x + tf(x) + \sum_{n\ge2} t^n E_n(x). Over R\mathbb{R} the higher terms are small only for small tt, and the flow-box theorem holds on a neighbourhood of unspecified size. Over Zp\mathbb{Z}_p there is no such trade-off. At every point xx, each higher term is smaller than f(x)f(x) by at least one pp-adic digit, for all t∈Zpt \in\mathbb{Z}_p at once (Lemma 3.1). By the ultrametric inequality the orbit of xx is then, up to that one digit, the straight line x+tf(x)x + tf(x), and distances along it are exactly the distances along the line. So local questions about the flow reduce to linear algebra over Zp\mathbb{Z}_p on the vectors f(x)f(x), and, through time-one maps, questions about groups generated by maps close to the identity reduce to linear algebra on the displacements g(x)−xg(x) - x. This is why the answers below come with explicit radii, and why their hypotheses can be checked on finitely many digits of finitely many values.

Context. In arithmetic dynamics such flows usually appear as interpolations. If a map gg is close enough to the identity, the iterates gn(x)g^n(x) are the values at integers of a function analytic in n∈Zpn \in\mathbb{Z}_p. Bell used this to prove the dynamical Mordell–Lang conjecture for automorphisms of affine varieties [2], and Poonen gave the sharp condition on gg [27]; the method is the backbone of the pp-adic approach surveyed in [3]. Cantat and Xie turned it into a pp-adic method for actions of discrete groups [8], and Cantat and Jang used it to study orbit closures of automorphism groups of affine surfaces over pp-adic fields [7]. Jang used it to give a new proof of a result announced by Bourgain, Gamburd and Sarnak, that for the Markoff surface transitivity modulo pp implies minimality on the Zp\mathbb{Z}_p-points, and to extend it, for p>3p > 3, to the surfaces x2+y2+z2=xyz+θx^2 + y^2 + z^2 = xyz + \theta with θ≡0(modp2)\theta\equiv0 \pmod{p^2} or (θ−4p2)=1\left(\frac{\theta-4}{p^2}\right) = 1 [16 Theorem 1]. Transitivity modulo pp is the conjecture of Bourgain, Gamburd and Sarnak [5], now proved for all but finitely many pp [9]. In one variable, Mukhamedov showed that the interpolated orbit of a non-fixed point fills a ball [26 Theorem 9.3].

The iteration of a single map is also well understood: there are complete descriptions of minimal subsystems of polynomial and 11-Lipschitz maps of Zp\mathbb{Z}_p [14, 1] and of their cycles modulo pnp^n [12, 21], a local theory of analytic germs [22, 28, 17, 4], and, in several variables, results on periods modulo pnp^n, box dimensions of orbit closures and discrepancy [24]. Almost all of this work starts from a map, and builds a flow only when it needs one. I start from the vector field and read the geometry off its Lie series.

Why time is pp-adic. A continuous map from R\mathbb{R} to Zpd\mathbb{Z}_p^d is constant, since R\mathbb{R} is connected and Zpd\mathbb{Z}_p^d is totally disconnected. Paths in Zpd\mathbb{Z}_p^d parametrised by real time are trivial, and the natural time domain of a pp-adic flow is Zp\mathbb{Z}_p itself. This changes familiar notions: time has no order, so “the first time an event happens” only makes sense after restricting to integer times, and the set of times at which something happens is a closed subset of Zp\mathbb{Z}_p, typically a finite union of balls.

Results

  1. Orbits (§ 4). For a point xx with f(x)≠0f(x) \ne0, the map t↦φt(x)t \mapsto\varphi_t(x) multiplies distances by exactly ∥f(x)∥\lVert f(x)\rVert, in every dimension (Theorem 3.2). In dimension one the orbit is therefore the whole ball B(x,∣f(x)∣)B(x,\lvert f(x)\rvert), which is Mukhamedov’s theorem for the time-one map. In higher dimension the interpolated orbit is an analytic arc in a polydisc, and Mukhamedov notes that there is then no ball-counting interpretation [26 Section 14]; Theorem 3.2 says that this arc is an isometric copy of Zp\mathbb{Z}_p, scaled by ∥f(x)∥\lVert f(x)\rVert. Cantat and Jang prove an analogue in normed Banach algebras, which they apply to composition operators on the Tate algebra [7 Lemma A.2]; the pointwise statement is finer, because ∥f(x)∥\lVert f(x)\rVert can be much smaller than the Gauss norm of ff. As consequences the flow has flow-box coordinates on B(x,∥f(x)∥)B(x,\lVert f(x)\rVert), first hitting times, first return times and orbit averages have closed formulas, and the integer orbit meets exactly pmax⁡(m−k,0)p^{\max(m-k,0)} classes modulo pmp^m, where k=v⁡(f(x))k = \operatorname{v}(f(x)) (Corollary 3.7). For time-one maps this is an exact form of the period growth ≍pm\asymp p^m proved for polynomial maps in [24 Theorem C].

  1. Time-one maps (§ 4.2). That an analytic map g≡id⁡(modpc)g \equiv\operatorname{id} \pmod{p^c}, coefficientwise, with cc as in (1), is the time-one map of a unique analytic field is the Bell–Poonen theorem [2, 27] (in the form of [7 Theorem 2.1]; see also [8]), and Cantat and Jang write the field as a logarithm [7 (A.14)]. Proposition 3.8 adds the converse, so that X↦φX1X \mapsto\varphi_X^1 is a bijection from pcZp{x}dp^c\mathbb{Z}_p\{x\}^d onto id⁡+pcZp{x}d\operatorname{id} + p^c\mathbb{Z}_p\{x\}^d, and the congruence X≡g−id⁡(modpc+1)X \equiv g - \operatorname{id} \pmod{p^{c+1}}. The congruence is what lets the criterion below be checked on the values gi(x0)g_i(x_0), without computing a field. As Cantat and Jang observe [7 Remark 2.2], the congruence g≡id⁡g \equiv\operatorname{id} must hold for the coefficients and not only pointwise (see also Remark 3.9).

  1. Minimal actions (Proposition 3.10). If g1,…,gdg_1,\ldots,g_d are such maps and the matrix of displacements (gi(x0)−x0)/pk(g_i(x_0)-x_0)/p^k is invertible over Zp\mathbb{Z}_p, the group they generate acts minimally on B(x0,p−k)B(x_0,p^{-k}), for every level kk and every dd (necessarily k≥ck\ge c). For d=2d=2 and k=c=1k=c=1 this is [16 Proposition 4]. In the proof of [7 Theorem B] the fields only span the tangent space over Qp\mathbb{Q}_p, and the orbit closure contains a neighbourhood of x0x_0 of unspecified size; with the Zp\mathbb{Z}_p-span I get the neighbourhood and an isometry onto it. Levels k>1k>1 cover points where no element of the group moves by exactly p−1p^{-1}, the situation of [16 Section 7.1].

  1. Markoff surfaces (§ 4.3). I test the first-order version of the criterion numerically at 7830 points. It holds at all but 28 of them, and those 28 are congruent modulo p2p^2 to the finite orbits described in [16 Section 7.2], in line with the conjecture stated there.

  1. Examples. Analytic fields behave like polynomial ones (§ 3.1), flow-box coordinates are computed by a Newton iteration that doubles the number of correct digits at each step (§ 5), and the formal logarithm of Gm\mathbb{G}_m or of an elliptic curve is a flow-box coordinate of an explicit field (§ 6).

Notation. v=vpv=v_p is the pp-adic valuation and ∣x∣=p−v(x)|x|=p^{-v(x)}. On Zpd\mathbb{Z}_p^d I use ∥x∥=max⁡i∣xi∣\lVert x\rVert=\max_i|x_i|, and v(x)=min⁡iv(xi)v(x)=\min_i v(x_i). For a polynomial gg with coefficients in Zp\mathbb{Z}_p (or in Zpd\mathbb{Z}_p^d), v(g)v(g) is the minimum valuation of its coefficients. B(x,ρ)={y:∥y−x∥≤ρ}B(x,\rho)=\{y:\lVert y-x\rVert\le\rho\}. For the vector field ff I write c=v(f)c=v(f) and always assume

c>1p−1,that is,c≥1 for p≥3,c≥2 for p=2.(1)c>\frac{1}{p-1},\qquad\text{that is,}\qquad c\ge1\text{ for }p\ge3,\qquad c\ge2\text{ for }p=2. \tag*{(1)}

A field with pp-integral coefficients satisfies (1) after multiplication by pp (by 4 if p=2p=2), which amounts to rescaling time. The bound cannot be lowered even for linear fields: the flow eλtxe^{\lambda t}x of f(x)=λxf(x)=\lambda x is defined for all t∈Zpt\in\mathbb{Z}_p exactly when v(λ)>1/(p−1)v(\lambda)>1/(p-1), because the exponential series converges exactly on ∣z∣<p−1/(p−1)|z|<p^{-1/(p-1)}. This is the rank-one case of Dwork’s transfer theorem, whose proof bounds the same Taylor series ∑ntnDn/n!\sum_n t^nD^n/n! through the spectral norm of DD [18]. Finally, v(n!)=(n−sp(n))/(p−1)≤(n−1)/(p−1)v(n!)=(n-s_p(n))/(p-1)\le(n-1)/(p-1) by Legendre’s formula, where sp(n)s_p(n) is the sum of the base-pp digits of nn.

The flow and its Lie series

Let D=∑jfj∂jD=\sum_j f_j\partial_j be the derivation of Zp[x]=Zp[x1,…,xd]\mathbb{Z}_p[x]=\mathbb{Z}_p[x_1,\ldots,x_d] attached to ff, and put

En=Dnxn!∈Qp[x]d,φt(x)=∑n≥0tnEn(x),E_n=\frac{D^n x}{n!}\in\mathbb{Q}_p[x]^d,\qquad\varphi_t(x)=\sum_{n\ge0}t^nE_n(x),

where DnxD^n x is applied componentwise to the coordinate functions.

Lemma 2.1. v(Dng)≥v(g)+ncv(D^n g)\ge v(g)+nc for every g∈Zp[x]g\in\mathbb{Z}_p[x], and v(En)≥nc−v(n!)≥n(c−1p−1)+1p−1v(E_n)\ge nc-v(n!)\ge n\left(c-\frac{1}{p-1}\right)+\frac{1}{p-1}. In particular En∈Zp[x]dE_n\in\mathbb{Z}_p[x]^d for all nn, v(En)≥1v(E_n)\ge1 for n≥1n\ge1, and v(En)→∞v(E_n)\to\infty.

Proof. The partial derivatives preserve Zp[x]\mathbb{Z}_p[x] and multiplication by fjf_j adds at least cc to the valuation, which gives the first claim by induction. The second follows from Legendre’s formula. By (1) the lower bound is positive for n≥1n\ge1 and grows linearly.

Consequently the series φt(x)\varphi_t(x) converges uniformly on Zp×Zpd\mathbb{Z}_p\times\mathbb{Z}_p^d and takes values in Zpd\mathbb{Z}_p^d. Write Φt=∑ntnDn/n!\Phi_t=\sum_n t^nD^n/n! for the corresponding operator on Zp[x]\mathbb{Z}_p[x]; by Lemma 2.1 it is well defined on Tate algebra of power series with coefficients tending to zero. Since DD is a derivation, Φt\Phi_t is a ring homomorphism, so Φt(g)=g∘φt\Phi_t(g)=g\circ\varphi_t for polynomials gg, and the usual identities for the exponential of a derivation hold:

Φs+t=ΦsΦt,∂∂tΦt(g)=Φt(Dg).\Phi_{s+t}=\Phi_s\Phi_t,\qquad\frac{\partial}{\partial t}\Phi_t(g)=\Phi_t(Dg).

Applied to g=xg=x they give φs+t=φt∘φs\varphi_{s+t}=\varphi_t\circ\varphi_s, φ0=id⁡\varphi_0=\operatorname{id} and ∂tφt(x)=f(φt(x))\partial_t\varphi_t(x)=f(\varphi_t(x)). Hence φ\varphi is a flow on Zpd\mathbb{Z}_p^d with time Zp\mathbb{Z}_p, and t↦φt(x)t\mapsto\varphi_t(x) is a solution given by a power series in tt that converges on all of Zp\mathbb{Z}_p; among such solutions it is the only one, since the coefficients of a power series solution are determined recursively by the equation.

Remark 2.2 (Nilpotent fields). For a field with pp-integral coefficients, only the decay of v⁡(Dnx/n!)\operatorname{v}(D^n x/n!) is needed for the Lie series to converge and define a flow, so the coefficientwise bound cc may be replaced by the spectral condition lim inf⁡nv⁡(Dnx)/n>1/(p−1)\liminf_n \operatorname{v}(D^n x)/n>1/(p-1), as in the transfer theorem. For example f=(p,x1)f=(p,x_1) has c=0c=0, but D2x1=0D^2x_1=0 and D3x2=0D^3x_2=0, and its flow is the polynomial map (x1+pt,x2+tx1+pt2/2)(x_1+pt,x_2+tx_1+pt^2/2). The metric results of § 4 use cc itself, and I keep (1) throughout.

Remark 2.3. Uniqueness has to be stated within analytic solutions. A pp-adic function can have zero derivative everywhere without being constant (any locally constant function does, and there are injective examples [30]), so the equation alone does not determine a merely differentiable solution.

Truncation. Each φt\varphi_t with ∣t∣≤1|t|\leq1 is an isometry of Zpd\mathbb{Z}_p^d: by Lemma 2.1, φt(x)−φt(y)=(x−y)+∑n≥1tn(En(x)−En(y))\varphi_t(x)-\varphi_t(y)=(x-y)+\sum_{n\geq1}t^n(E_n(x)-E_n(y)) and the sum has norm at most p−1∥x−y∥p^{-1}\lVert x-y\rVert. More importantly for computation, since v⁡(tnEn)≥nc−⌊(n−1)/(p−1)⌋\operatorname{v}(t^nE_n)\geq nc-\left\lfloor(n-1)/(p-1)\right\rfloor for t∈Zpt\in\mathbb{Z}_p, let

N(K;c)=min⁡{N≥1:nc−⌊n−1p−1⌋≥K for all n≥N}.(2)N(K;c)=\min\left\{N\geq1:nc-\left\lfloor\frac{n-1}{p-1}\right\rfloor\geq K\text{ for all }n\geq N\right\}. \tag*{(2)}

The expression in nn is non-decreasing (its increments are cc or c−1c-1), so N(K;c)N(K;c) is the first nn at which it reaches KK. Then

φt(x)≡∑n<N(K;c)tnEn(x)(modpK),t∈Zp, x∈Zpd.(3)\varphi_t(x)\equiv\sum_{n<N(K;c)}t^nE_n(x)\pmod{p^K},\qquad t\in\mathbb{Z}_p,\ x\in\mathbb{Z}_p^d. \tag*{(3)}

The right-hand side is a polynomial in (t,x)(t,x) with coefficients in Zp\mathbb{Z}_p, so φt(x) mod pK\varphi_t(x)\bmod p^K depends only on tt and xx modulo pKp^K. If t∈pmZpt\in p^m\mathbb{Z}_p, then v⁡(tnEn)≥n(c+m)−⌊(n−1)/(p−1)⌋\operatorname{v}(t^nE_n)\geq n(c+m)-\left\lfloor(n-1)/(p-1)\right\rfloor and N(K;c+m)N(K;c+m) terms suffice. For example N(19;1)=36N(19;1)=36 for p=3p=3, while N(19;4)=6N(19;4)=6.

Analytic fields

Nothing above, and nothing in § 4, uses that ff is a polynomial. Let ff be a vector of power series ∑efexe\sum_e f_e x^e with fe∈Zpf_e\in\mathbb{Z}_p, fe→0f_e\to0, and min⁡ev⁡(fe)=c\min_e\operatorname{v}(f_e)=c satisfying (1); these are the elements of the Tate algebra, the series converging on Zpd\mathbb{Z}_p^d. The derivation D=f⋅∇D=f\cdot\nabla maps the Tate algebra to itself and raises valuations by at least cc, so Lemmas 2.1 and 3.1 and Theorem 3.2 hold verbatim.

Lemma 2.4. Let γ=v⁡((N−1)!)\gamma=\operatorname{v}((N-1)!) with N=N(K;c)N=N(K;c). Then φt(x) mod pK\varphi_t(x)\bmod p^K depends only on f mod pK+γf\bmod p^{K+\gamma}.

Proof. DnxD^n x is a polynomial with integer coefficients in the coefficients of ff and their products with monomials; changing ff by a multiple of pK+γp^{K+\gamma} changes DnxD^n x by a multiple of pK+γp^{K+\gamma}. For n<Nn<N, dividing by n!n! loses at most γ\gamma digits, and the terms n≥Nn\geq N vanish modulo pKp^K by (2). □

Since the coefficients of ff tend to zero, f mod pK+γf\bmod p^{K+\gamma} is a polynomial. So φt(x) mod pK\varphi_t(x)\bmod p^K for an analytic field equals the same quantity for a polynomial field, and is computed exactly by finite arithmetic.

Orbits are balls

The following bound is the source of all geometric statements about the flow. It sharpens Lemma 2.1 at a given point.

Lemma 3.1. Let x∈Zpdx \in\mathbb{Z}_p^d and k=v⁡(f(x))k = \operatorname{v}(f(x)). For n≥2n \ge2,

v⁡(En(x))≥k+(n−1)c−v⁡(n!)≥k+1.\operatorname{v}(E_n(x)) \ge k + (n-1)c - \operatorname{v}(n!) \ge k + 1.

Proof. For n≥1n \ge1 and each coordinate ii, Dnxi=D(Dn−1xi)=∑jfj∂j(Dn−1xi)D^n x_i = D(D^{n-1}x_i) = \sum_j f_j \partial_j(D^{n-1}x_i). Evaluating at xx and using v⁡(∂jDn−1xi)≥(n−1)c\operatorname{v}(\partial_jD^{n-1}x_i) \ge(n-1)c from Lemma 2.1 gives v⁡((Dnxi)(x))≥k+(n−1)c\operatorname{v}((D^n x_i)(x)) \ge k + (n-1)c. Dividing by n!n! gives the first inequality. For the second, v⁡(n!)≤(n−1)/(p−1)<(n−1)c\operatorname{v}(n!) \le(n-1)/(p-1) < (n-1)c by (1), and v⁡(n!)\operatorname{v}(n!), (n−1)c(n-1)c are integers, so (n−1)c−v⁡(n!)≥1(n-1)c-\operatorname{v}(n!) \ge1.

For p=2p=2 the hypothesis c≥2c \ge2 is used here: with c=1c=1 and n=2n=2 the bound would read k+1−1=kk+1-1=k, and the strict inequality fails.

Theorem 3.2. Let x∈Zpdx \in\mathbb{Z}_p^d with f(x)≠0f(x) \ne0. Then for all s,t∈Zps,t \in\mathbb{Z}_p,

∥φt(x)−φs(x)∥=∥f(x)∥⋅∣t−s∣.\lVert\varphi_t(x)-\varphi_s(x)\rVert=\lVert f(x)\rVert\cdot\lvert t-s\rvert.

Hence t↦φt(x)t \mapsto\varphi_t(x) is a homeomorphism of Zp\mathbb{Z}_p onto the orbit of xx, which lies in B(x,∥f(x)∥)B(x,\lVert f(x)\rVert). If d=1d=1 the orbit is the whole ball B(x,∣f(x)∣)B(x,\lvert f(x)\rvert).

Proof. Put an=En(x)a_n=E_n(x), so a1=f(x)a_1=f(x) and φt(x)=∑nantn\varphi_t(x)=\sum_n a_nt^n. Then

φt(x)−φs(x)=(t−s)(a1+∑n≥2anhn(s,t)),hn(s,t)=∑i=0n−1tisn−1−i∈Zp.\varphi_t(x)-\varphi_s(x)=(t-s)\left(a_1+\sum_{n\ge2}a_nh_n(s,t)\right),\qquad h_n(s,t)=\sum_{i=0}^{n-1}t^is^{n-1-i}\in\mathbb{Z}_p.

By Lemma 3.1 each term of the sum has norm <∥a1∥<\lVert a_1\rVert, and the terms tend to zero, so the bracket has norm exactly ∥a1∥=∥f(x)∥\lVert a_1\rVert=\lVert f(x)\rVert. This proves the identity, and with it injectivity, continuity of the inverse and the inclusion in the ball (take s=0s=0).

Let d=1d=1 and k=v⁡(f(x))k=\operatorname{v}(f(x)), and write F(t)=φt(x)F(t)=\varphi_t(x). For m≥0m\ge0, the identity shows that FF induces a well-defined injective map from Zp/pmZp\mathbb{Z}_p/p^m\mathbb{Z}_p to the set of balls of radius p−k−mp^{-k-m} contained in B(x,p−k)B(x,p^{-k}). Both sets have pmp^m elements, so the map is a bijection. Given y∈B(x,p−k)y\in B(x,p^{-k}), let tmt_m be the unique class modulo pmp^m with F(tm)≡y(modpk+m)F(t_m)\equiv y\pmod{p^{k+m}}. By uniqueness the classes are compatible, so they define t∈Zpt\in\mathbb{Z}_p, and F(t)=yF(t)=y by continuity.

Remark 3.3. In dimension one, the fact that the orbit is a ball was proved by Mukhamedov [26 Theorem 9.3] for maps x+pqh(x)x+p^qh(x) with h∈Zp⟨x⟩h\in\mathbb{Z}_p\langle x\rangle and q(p−1)>1q(p-1)>1, which include φ1\varphi_1, through the analytic interpolation of their iterates; he writes the interpolated orbit as a+δWa(t)a+\delta W_a(t), with δ=φ1(a)−a\delta=\varphi_1(a)-a and WaW_a an analytic automorphism of the unit disc. The proof above starts from the field instead, and its metric identity holds in every dimension. The time-one map φ1\varphi_1 is therefore minimal on each ball B(x,∣f(x)∣)B(x,\lvert f(x)\rvert) with f(x)≠0f(x)\ne0, since Z\mathbb{Z} is dense in Zp\mathbb{Z}_p. This agrees with the minimal decomposition theory of Fan and Liao [14] for polynomial maps, although φ1\varphi_1 is in general not a polynomial. The setting is opposite to that of Bell and Poonen [2, 27], who interpolate the iterates of a given map close to the identity; § 4.2 shows that the two settings coincide. Here the field is given, and by Mahler’s theorem [23], φt(x)=∑n(tn)Δn(x)\varphi_t(x)=\sum_n \binom{t}{n}\Delta^n(x) with Δn(x)=∑j≤n(−1)n−j(nj)φj(x)\Delta^n(x)=\sum_{j\le n}(-1)^{n-j}\binom{n}{j}\varphi_j(x), so the integer iterates determine the flow at every pp-adic time.

Flow-box coordinates

Fix xx with f(x)≠0f(x) \ne0, put k=v⁡(f(x))k = \operatorname{v}(f(x)), choose an index ii with v⁡(fi(x))=k\operatorname{v}(f_i(x)) = k, and let

B=B(x,p−k),Σ={w∈B:wi=xi}.B = \mathbb{B}(x,p^{-k}), \qquad\Sigma= \{w \in B : w_i = x_i\}.

For w∈Bw \in B, f(w)−f(x)f(w)-f(x) has valuation at least c+k>kc+k>k, so v⁡(f(w))=v⁡(fi(w))=k\operatorname{v}(f(w))=\operatorname{v}(f_i(w))=k on all of BB, and by Theorem 3.2 the orbit of every point of BB stays in BB.

Proposition 3.4. The map Ψ(s,w)=φs(w)\Psi(s,w)=\varphi_s(w) is a bijection from Zp×Σ\mathbb{Z}_p \times\Sigma onto BB, and it is an isometry for the distance max⁡(p−k∣s−s′∣,∥w−w′∥)\max(p^{-k}|s-s'|,\lVert w-w'\rVert) on Zp×Σ\mathbb{Z}_p \times\Sigma. In these coordinates φt(Ψ(s,w))=Ψ(s+t,w)\varphi_t(\Psi(s,w))=\Psi(s+t,w).

Proof. Write φs(w)−φs′(w′)=(φs(w)−φs(w′))+(φs(w′)−φs′(w′))=A+C\varphi_s(w)-\varphi_{s'}(w')=(\varphi_s(w)-\varphi_s(w'))+(\varphi_s(w')-\varphi_{s'}(w'))=A+C. Since φs\varphi_s is an isometry, A=(w−w′)+RA=(w-w')+R with ∥R∥≤p−1∥w−w′∥\lVert R\rVert\le p^{-1}\lVert w-w'\rVert, and Ai=RiA_i=R_i because wi=wi′w_i=w'_i. By Theorem 3.2 and v⁡(f(w′))=k\operatorname{v}(f(w'))=k, ∥C∥=p−k∣s−s′∣\lVert C\rVert=p^{-k}|s-s'|, and the same argument applied to the ii-th coordinate alone shows ∣Ci∣=p−k∣s−s′∣|C_i|=p^{-k}|s-s'|. If ∥w−w′∥>p−k∣s−s′∣\lVert w-w'\rVert>p^{-k}|s-s'|, then ∥A+C∥=∥w−w′∥\lVert A+C\rVert=\lVert w-w'\rVert. Otherwise ∣(A+C)i∣=∣Ri+Ci∣=p−k∣s−s′∣|(A+C)_i|=|R_i+C_i|=p^{-k}|s-s'| (when s≠s′s\ne s'), while ∥A+C∥≤max⁡(∥A∥,∥C∥)=p−k∣s−s′∣\lVert A+C\rVert\le\max(\lVert A\rVert,\lVert C\rVert)=p^{-k}|s-s'|. In both cases ∥Ψ(s,w)−Ψ(s′,w′)∥\lVert\Psi(s,w)-\Psi(s',w')\rVert equals the stated distance, which gives isometry and injectivity.

For surjectivity, let z∈Bz\in B and G(s)=(φ−s(z)i−xi)/fi(z)G(s)=(\varphi_{-s}(z)_i-x_i)/f_i(z). Arguing as in the proof of Theorem 3.2 for the ii-th coordinate, ∣G(s)−G(s′)∣=∣s−s′∣|G(s)-G(s')|=|s-s'|, and G(0)=(zi−xi)/fi(z)∈ZpG(0)=(z_i-x_i)/f_i(z)\in\mathbb{Z}_p, so GG maps Zp\mathbb{Z}_p to Zp\mathbb{Z}_p isometrically. An isometry of Zp\mathbb{Z}_p into itself is onto (count classes modulo pmp^m), so G(s)=0G(s)=0 for some ss, and w=φ−s(z)∈Σw=\varphi_{-s}(z)\in\Sigma satisfies Ψ(s,w)=z\Psi(s,w)=z. The last claim is the group property of the flow.

For d=1d=1 the transversal is the single point xx, and s(z)s(z) is a rectifying coordinate: φt(z)=φs(z)+t(x)\varphi_t(z)=\varphi_{s(z)+t}(x).

Corollary 3.5. Let d=1d=1, xx with k=v⁡(f(x))<∞k=\operatorname{v}(f(x))<\infty, and m≥km\ge k.

(a) For y∈B(x,p−k)y\in\mathbb{B}(x,p^{-k}) the set {t∈Zp:φt(x)∈B(y,p−m)}\{t\in\mathbb{Z}_p:\varphi_t(x)\in\mathbb{B}(y,p^{-m})\} is the ball s(y)+pm−kZps(y)+p^{m-k}\mathbb{Z}_p. The first non-negative integer time at which the orbit of xx enters B(y,p−m)\mathbb{B}(y,p^{-m}) is the representative of s(y)s(y) in [0,pm−k)[0,p^{m-k}). If y∉B(x,p−k)y\notin\mathbb{B}(x,p^{-k}), the orbit never enters B(y,p−m)\mathbb{B}(y,p^{-m}).

(b) The first return time of xx to B(x,p−m)\mathbb{B}(x,p^{-m}) under φ1\varphi_1 is pm−kp^{m-k}.

(c) The sequence n↦φn(x) mod pmn\mapsto\varphi_n(x)\bmod p^m is periodic with period pm−kp^{m-k} and visits each of the pm−kp^{m-k} balls of radius p−mp^{-m} in B(x,p−k)\mathbb{B}(x,p^{-k}) once per period. Hence for every FF constant on balls of radius p−mp^{-m},

lim⁡M→∞1M∑n<MF(φn(x))=p−(m−k)∑n<pm−kF(φn(x)).\lim_{M\to\infty}\frac{1}{M}\sum_{n<M}F(\varphi_n(x))=p^{-(m-k)}\sum_{n<p^{m-k}}F(\varphi_n(x)).

In dimension dd, (b) and the periodicity in (c) hold with the same proof, and the orbit visits pm−kp^{m-k} classes modulo pmp^m once per period (Corollary 3.7). Statement (a) holds for yy on the orbit of xx, with s(y)s(y) defined by φs(y)(x)=y\varphi_{s(y)}(x)=y; for other yy one decides whether the orbit meets B(y,p−m)\mathbb{B}(y,p^{-m}) through the flow-box coordinates of Proposition 3.4.

Proof. All three follow from ∣φt(x)−φt′(x)∣=p−k∣t−t′∣|\varphi_t(x)-\varphi_{t'}(x)|=p^{-k}|t-t'| and Theorem 3.2: φt(x)∈B(y,p−m)\varphi_t(x)\in\mathbb{B}(y,p^{-m}) iff ∣t−s(y)∣≤p−(m−k)|t-s(y)|\le p^{-(m-k)}.

Remark 3.6 (Ergodic maps and orbit closures). Let d=1d=1, f(x0)≠0f(x_0)\ne0 and k=v⁡(f(x0))k=\operatorname{v}(f(x_0)), and rescale the time-one map to g(y)=(φ1(x0+pky)−x0)/pkg(y)=(\varphi_1(x_0+p^k y)-x_0)/p^k on Zp\mathbb{Z}_p. By Corollary 3.5, gg is an isometry of Zp\mathbb{Z}_p that is conjugate to y↦y+1y\mapsto y+1 through the rescaled rectifying coordinate, so it is ergodic, and each

of its orbits visits the pmp^m classes modulo pmp^m once per period: it is a low-discrepancy sequence in the sense of [24]. Its Mahler coefficients an=Δgng(0)a_n = \Delta_g^n g(0) can be computed exactly from g(0),…,g(n)g(0),\ldots,g(n). Every EjE_j has coefficients of valuation at least cc, so the coefficient of yey^e in gg has valuation at least k(e−1)+ck(e-1)+c for e≥2e \ge2, and expanding yey^e in the binomial basis gives

a0≡f(x0)/pk≢0(modp),a1≡1(modpc),v(an)≥k(n−1)+c+v(n!)(n≥2).a_0 \equiv f(x_0)/p^k \not\equiv0 \pmod p,\qquad a_1 \equiv1 \pmod{p^c},\qquad v(a_n) \ge k(n-1)+c+v(n!)\quad(n\ge2).

These are well within Anashin’s sufficient conditions for ergodicity, as stated in [24 Theorem 6.2]. In any dimension the orbit closures of time-one maps satisfy the dimension alternative of [24 Theorem B], stated there for polynomial maps, by Corollary 3.7. As a pseudorandom sequence it is predictable: by (3) its terms modulo pKp^K are a polynomial of degree less than N(K;c)N(K;c) in nn, so the nn-th term costs O(K)O(K) operations independently of nn, and its index is recovered from it by Newton’s method (§ 5) in O(log⁡K)O(\log K) steps. The argument uses analyticity, so it does not apply to T-functions built with OR, such as the map x↦x+(x2∨C)x\mapsto x+(x^2\mathbin{\vee} C) of [20].

Corollary 3.7 (Orbit counts). Let x∈Zpdx\in\mathbb{Z}_p^d with f(x)≠0f(x)\ne0 and k=v(f(x))k=v(f(x)). For every m≥0m\ge0 the orbit {φn(x):n≥0}\{\varphi_n(x):n\ge0\} meets exactly pmax⁡(m−k,0)p^{\max(m-k,0)} classes modulo pmp^m. Hence its closure {φt(x):t∈Zp}\{\varphi_t(x):t\in\mathbb{Z}_p\} has box dimension one, while an equilibrium has box dimension zero. This is the dimension alternative proved for polynomial maps in [24 Theorem B], here for time-one maps of analytic fields, in any dimension.

Proof. For m≤km\le k the orbit lies in B(x,p−k)B(x,p^{-k}), inside one class modulo pmp^m. For m>km>k, Theorem 3.2 gives φt(x)≡φs(x)(modpm)\varphi_t(x)\equiv\varphi_s(x)\pmod{p^m} exactly when t≡s(modpm−k)t\equiv s\pmod{p^{m-k}}, and the non-negative integers meet every class modulo pm−kp^{m-k}. So the closure is covered by exactly pmax⁡(m−k,0)p^{\max(m-k,0)} balls of radius p−mp^{-m}, and log⁡pmax⁡(m−k,0)/log⁡pm→1\log p^{\max(m-k,0)}/\log p^m\to1.

Interpolated iterates and minimal group actions

Every analytic map g≡id⁡(modpc)g\equiv\operatorname{id}\pmod{p^c}, coefficientwise, is the time-one map of a unique analytic field. This is the Bell–Poonen theorem [2, 27], in the form of [7 Theorem 2.1], and the field is the logarithm of TgT_g applied to id⁡\operatorname{id} [7 (A.14)]. I give a short proof that also yields the converse and the congruence X≡g−id⁡(modpc+1)X\equiv g-\operatorname{id}\pmod{p^{c+1}}, which is the part used below. For g∈Zp⟨x⟩dg\in\mathbb{Z}_p\langle x\rangle^d write Tgh=h∘gT_g h=h\circ g and Δg=Tg−Id⁡\Delta_g=T_g-\operatorname{Id}, acting on Zp⟨x⟩\mathbb{Z}_p\langle x\rangle and componentwise on vectors.

Proposition 3.8 (Time-one maps). Let cc satisfy (1). The map X↦φ1XX\mapsto\varphi_1^X is a bijection from pcZp⟨x⟩dp^c\mathbb{Z}_p\langle x\rangle^d onto id⁡+pcZp⟨x⟩d\operatorname{id}+p^c\mathbb{Z}_p\langle x\rangle^d. Its inverse is

X=∑j≥1(−1)j−1jΔgj(id⁡),X=g−id⁡(modpc+1),X=\sum_{j\ge1}\frac{(-1)^{j-1}}{j}\Delta_g^j(\operatorname{id}),\qquad X=g-\operatorname{id}\pmod{p^{c+1}},

and the flow of XX is the interpolation Φ(t,x)=∑j≥0(tj)Δgj(id⁡)(x)\Phi(t,x)=\sum_{j\ge0}\binom{t}{j}\Delta_g^j(\operatorname{id})(x) of [27].

Proof. If X∈pcZp⟨x⟩dX\in p^c\mathbb{Z}_p\langle x\rangle^d, then v(En)≥nc−v(n!)≥cv(E_n)\ge nc-v(n!)\ge c for n≥1n\ge1 by Lemma 2.1, so φ1X∈id⁡+pcZp⟨x⟩d\varphi_1^X\in\operatorname{id}+p^c\mathbb{Z}_p\langle x\rangle^d. Conversely let g∈id⁡+pcZp⟨x⟩dg\in\operatorname{id}+p^c\mathbb{Z}_p\langle x\rangle^d. Since h∘g−hh\circ g-h lies in the ideal generated by the gi−xig_i-x_i, Δg\Delta_g raises valuations by at least cc, so v(Δgj(id⁡))≥jcv(\Delta_g^j(\operatorname{id}))\ge jc. Expanding (tj)=∑ks(j,k)tk/j!\binom{t}{j}=\sum_k s(j,k)t^k/j! with integer Stirling numbers s(j,k)s(j,k) gives Φ(t,x)=∑ktkGk(x)\Phi(t,x)=\sum_k t^kG_k(x) with Gk=∑j≥ks(j,k)Δgj(id⁡)/j!G_k=\sum_{j\ge k}s(j,k)\Delta_g^j(\operatorname{id})/j!, and v(Gk)≥min⁡j≥k(jc−v(j!))≥k(c−1p−1)v(G_k)\ge\min_{j\ge k}(jc-v(j!))\ge k(c-\frac{1}{p-1}), so Φ\Phi converges on Zp×Zpd\mathbb{Z}_p\times\mathbb{Z}_p^d. For integers n≥0n\ge0 the binomial theorem gives Φ(n,⋅)=Tgn(id⁡)=gn\Phi(n,\cdot)=T_g^n(\operatorname{id})=g^n, so Φ(s+t,x)=Φ(s,Φ(t,x))\Phi(s+t,x)=\Phi(s,\Phi(t,x)) for integers s,t≥0s,t\ge0, and by continuity for all s,t∈Zps,t\in\mathbb{Z}_p. Both sides are power series in ss converging on Zp\mathbb{Z}_p, and a non-zero such series has finitely many zeros in Zp\mathbb{Z}_p [34], so their coefficients of ss agree: ∂tΦ(t,x)=G1(Φ(t,x))\partial_t\Phi(t,x)=G_1(\Phi(t,x)). Thus t↦Φ(t,x)t\mapsto\Phi(t,x) is a power series solution of x′=G1(x)x'=G_1(x) with Φ(0,x)=x\Phi(0,x)=x, and by uniqueness (§ 3) Φ(t,⋅)=φtG1\Phi(t,\cdot)=\varphi_t^{G_1}, in particular φ1G1=g\varphi_1^{G_1}=g. Since s(j,1)=(−1)j−1(j−1)!s(j,1)=(-1)^{j-1}(j-1)!, G1G_1 is the stated series. Its jj-th term has valuation at least jc−v(j)jc-v(j), which is cc for j=1j=1 and at least c+1c+1 for j≥2j\geq2, because v(j)≤v(j!)≤(j−1)/(p−1)<(j−1)cv(j)\leq v(j!)\leq(j-1)/(p-1)<(j-1)c and v(j),(j−1)cv(j),(j-1)c are integers; this gives G1∈pcZp⟨x⟩dG_1\in p^c\mathbb{Z}_p\langle x\rangle^d and the congruence. Finally a field is determined by its flow, X=∂tφtX∣t=0X=\left.\partial_t\varphi_t^X\right|_{t=0}, and the flow by its time-one map, since the integer iterates determine φtX\varphi_t^X by continuity; so the map is injective. □\square

Proposition 3.8 is an analogue, for the group id⁡+pcZp⟨x⟩d\operatorname{id}+p^c\mathbb{Z}_p\langle x\rangle^d under composition, of the correspondence between a uniform pro-pp group and its Zp\mathbb{Z}_p-Lie algebra given by the exponential and the logarithm [13]. Here the Lie algebra is a module of vector fields and the exponential is the Lie series.

Remark 3.9 (Which maps are flows). The congruence g≡id⁡(modpc)g\equiv\operatorname{id}\pmod{p^c} is needed for the coefficients, not only at every point; Cantat and Jang make this point with x+(xq−x)ex+(x^q-x)e [7 Remark 2.2]. Pointwise congruence does not even give a time-one map. The map g(z)=zpg(z)=z^p satisfies g(z)≡z(modp)g(z)\equiv z\pmod p at every z∈Zpz\in\mathbb{Z}_p, but g−id⁡=zp−zg-\operatorname{id}=z^p-z is not divisible by pp. It is not a time-one map: for z≡w(modp)z\equiv w\pmod p the factor ∑iziwp−1−i\sum_i z^i w^{p-1-i} of zp−wpz^p-w^p is ≡pzp−1≡0(modp)\equiv pz^{p-1}\equiv0\pmod p, so gg contracts each residue class, while time-one maps are isometries. Near a fixed point, Jenkins and Spallone show that pp-adic analytic vector fields generate analytic flows and that formally equivalent maps tangent to the identity are analytically equivalent [17]. Crespo and Pelayo prove an existence lemma for time-dependent fields f(x,y)f(x,y) that vanish, together with their first derivatives in yy, along y=vy=v: the solution exists for all times in a fixed ball once the initial value is close enough to vv [10 Lemma 2.1]. In the language of this paper, centring at vv and rescaling y−v=rzy-v=rz multiplies the terms of degree ee in zz by re−1r^{e-1}, so for small rr the terms of degree at least two become as divisible as (1) requires. The linear terms are not changed by rescaling, which is why they have to vanish.

Thus Poonen’s interpolated orbits satisfy Theorem 3.2, with ∥X(x)∥=∥g(x)−x∥\lVert X(x)\rVert=\lVert g(x)-x\rVert by the theorem at t=1t=1, and the congruence Φ(t,x)≡x+t(g(x)−x)(modp2)\Phi(t,x)\equiv x+t(g(x)-x)\pmod{p^2} used by Jang [16] for c=1c=1 follows from the congruence for XX, since v(En)≥2v(E_n)\geq2 for n≥2n\geq2 by Lemma 2.1. Flow-box coordinates for several fields at once give a criterion for minimality of the group they generate, in any dimension and on balls of any radius. For d=2d=2 and k=1k=1 it is [16 Proposition 4], the step through which pp-adic flows enter the proof that residual transitivity implies minimality for Markoff surfaces.

Proposition 3.10 (Minimal actions on balls). Let g1,…,gd∈id⁡+pcZp⟨x⟩dg_1,\ldots,g_d\in\operatorname{id}+p^c\mathbb{Z}_p\langle x\rangle^d, x0∈Zpdx_0\in\mathbb{Z}_p^d and k≥1k\geq1, and suppose the matrix MM with columns (gi(x0)−x0)/pk(g_i(x_0)-x_0)/p^k lies in GL⁡d(Zp)\operatorname{GL}_d(\mathbb{Z}_p) (this forces k≥ck\geq c). Let XiX_i be the field of gig_i (Proposition 3.8) and φ(i)\varphi^{(i)} its flow. Then

F(t)=φt1(1)∘⋯∘φtd(d)(x0)F(t)=\varphi_{t_1}^{(1)}\circ\cdots\circ\varphi_{t_d}^{(d)}(x_0)

is a bijection from Zpd\mathbb{Z}_p^d onto B=B(x0,p−k)\mathbb{B}=\mathbb{B}(x_0,p^{-k}) with ∥F(t)−F(t′)∥=p−k∥t−t′∥\lVert F(t)-F(t')\rVert=p^{-k}\lVert t-t'\rVert, and the group generated by g1,…,gdg_1,\ldots,g_d maps B\mathbb{B} to itself and acts minimally on it.

Proof. The columns of MM are not divisible by pp, so v(gi(x0)−x0)=kv(g_i(x_0)-x_0)=k, and by Theorem 3.2 at t=1t=1, v(Xi(x0))=kv(X_i(x_0))=k. By Lemma 3.1 the terms n≥2n\geq2 of gi(x0)−x0=∑n≥1En(i)(x0)g_i(x_0)-x_0=\sum_{n\geq1}E_n^{(i)}(x_0) have valuation at least k+1k+1, so Xi(x0)≡gi(x0)−x0(modpk+1)X_i(x_0)\equiv g_i(x_0)-x_0\pmod{p^{k+1}}. For y∈By\in\mathbb{B}, v(Xi(y)−Xi(x0))≥c+k>kv(X_i(y)-X_i(x_0))\geq c+k>k, so v(Xi(y))=kv(X_i(y))=k on B\mathbb{B}, and the orbits of φ(i)\varphi^{(i)}, in particular of gi=φ1(i)=φ−1(i)g_i=\varphi_1^{(i)}=\varphi_{-1}^{(i)}, stay in B\mathbb{B}.

Let y,y′∈By,y'\in\mathbb{B} and s,s′∈Zps,s'\in\mathbb{Z}_p, and write

φs(i)(y)−φs(i)(y′)=(φs(i)(y)−φs(i)(y′))+∑n≥1(sn−s′n)En(i)(y′).\varphi_s^{(i)}(y)-\varphi_s^{(i)}(y')=\left(\varphi_s^{(i)}(y)-\varphi_s^{(i)}(y')\right)+\sum_{n\geq1}(s^n-s'^n)E_n^{(i)}(y').

The first bracket is (y−y′)+R(y-y')+R with ∥R∥≤p−1∥y−y′∥\lVert R\rVert\le p^{-1}\lVert y-y'\rVert, since φs(i)\varphi_s^{(i)} is the identity plus a contraction (§ 3). In the sum, the term n=1n=1 is (s−s′)Xi(x0)+(s−s′)(Xi(y′)−Xi(x0))(s-s')X_i(x_0)+(s-s')(X_i(y')-X_i(x_0)), and the second part has norm at most p−k−1∣s−s′∣p^{-k-1}|s-s'|. For n≥2n\ge2, ∣sn−s′n∣≤∣s−s′∣|s^n-s'{}^n|\le|s-s'|, and ∥En(i)(y′)∥≤p−k−1\lVert E_n^{(i)}(y')\rVert\le p^{-k-1} by Lemma 3.1 at y′y'. Hence

φs(i)(y)−φs′(i)(y′)=(y−y′)+(s−s′)Xi(x0)+e,∥e∥≤p−1max⁡(∥y−y′∥,p−k∣s−s′∣).\varphi_s^{(i)}(y)-\varphi_{s'}^{(i)}(y')=(y-y')+(s-s')X_i(x_0)+e,\qquad\lVert e\rVert\le p^{-1}\max(\lVert y-y'\rVert,p^{-k}|s-s'|).

Applying this to the dd factors of FF in turn, starting from y=y′=x0y=y'=x_0, gives F(t)−F(t′)=∑i(ti−ti′)Xi(x0)+eF(t)-F(t')=\sum_i(t_i-t_i')X_i(x_0)+e with ∥e∥≤p−k−1∥t−t′∥\lVert e\rVert\le p^{-k-1}\lVert t-t'\rVert. The matrix with columns Xi(x0)/pkX_i(x_0)/p^k is congruent to MM modulo pp, hence lies in GL⁡d(Zp)\operatorname{GL}_d(\mathbb{Z}_p), and so ∥F(t)−F(t′)∥=p−k∥t−t′∥\lVert F(t)-F(t')\rVert=p^{-k}\lVert t-t'\rVert. Thus FF induces an injective map from (Zp/pm)d(\mathbb{Z}_p/p^m)^d to the pmdp^{md} balls of radius p−k−mp^{-k-m} in BB, which is a bijection; as in the proof of Theorem 3.2, FF is onto BB.

For minimality let z,z′∈Bz,z'\in B, and choose t,t′t,t' with F(t)=zF(t)=z and F(t′)=z′F(t')=z'. Then

z′=φt1′(1)∘⋯∘φtd′(d)∘φ−td(d)∘⋯∘φ−t1(1)(z).z'=\varphi_{t'_1}^{(1)}\circ\cdots\circ\varphi_{t'_d}^{(d)}\circ\varphi_{-t_d}^{(d)}\circ\cdots\circ\varphi_{-t_1}^{(1)}(z).

Each φs(i)\varphi_s^{(i)} is an isometry and s↦φs(i)(y)s\mapsto\varphi_s^{(i)}(y) is Lipschitz uniformly in yy, so replacing the 2d2d times by integers close to them, that is φn(i)=gin\varphi_{n}^{(i)}=g_i^n, moves the result arbitrarily little. Hence an element of the group maps zz arbitrarily close to z′z'.

Relation to [7, 8, 16]. The map FF is the map (2.10) in the proof of [7 Theorem B], and the density argument is theirs. There the vectors Θf(x)=X(x)\Theta_f(x)=X(x) are only required to span the tangent space over Qp\mathbb{Q}_p, which makes FF a local diffeomorphism. Proposition 3.10 uses instead the Zp\mathbb{Z}_p-module spanned by the displacements g(x0)−x0g(x_0)-x_0. When this module is pkZpdp^k\mathbb{Z}_p^d (by Nakayama’s lemma, exactly when dd of the displacements form a basis of it), the minimal set is the whole ball of radius p−kp^{-k}, and FF is an explicit isometry onto it. When such a group is a finitely generated dense subgroup of a pp-adic analytic group GG, acting by analytic maps congruent to the identity modulo pp, its action extends to an analytic action of GG [8], and orbit closures become GG-orbits; Proposition 3.10 says which ball is a single orbit. Similarly, for character varieties of surface groups, Golsefidy and Tamam show that outside explicit exceptional sets, the closure in Zp\mathbb{Z}_p-points of an infinite orbit of the pure mapping class group contains an open set [15 Theorem K]; Proposition 3.10 names the ball once the displacements are known. Beyond the following numerical check I do not pursue the arithmetic of Markoff surfaces.

A numerical check on Markoff surfaces

I test the hypothesis of Proposition 3.10 on the surfaces Xθ:x2+y2+z2=xyz+θX_\theta:x^2+y^2+z^2=xyz+\theta of [5, 16]. Let sas_a be the involution that replaces xax_a by xbxc−xax_bx_c-x_a, where {a,b,c}={1,2,3}\{a,b,c\}=\{1,2,3\}, and let rot⁡a=sb∘sc\operatorname{rot}_a=s_b\circ s_c. It preserves the fibres xa=tx_a=t and acts on them linearly; for a point qq let nan_a be the order of this linear map modulo pp at t=qat=q_a. Near qq the map rot⁡ana\operatorname{rot}_a^{n_a} is congruent to the identity on each fibre, and Proposition 3.8 puts its fixed field there.

At a smooth point qq, two of the coordinates, centred at qq and divided by pp, give an analytic chart of the residue disc of qq onto Zp2\mathbb{Z}_p^2 by Hensel’s lemma. In this chart rot⁡ana\operatorname{rot}_a^{n_a} is only affine and unipotent modulo pp (compare [16 Lemma 31]; I checked this numerically at 180 points), but its pp-th power is congruent to the identity coefficientwise, and its field is pp times the field above. So if these fields and their conjugates by words in the involutions span pTqpT_q, where Tq=TqXθ∩Zp3T_q=T_qX_\theta\cap\mathbb{Z}_p^3 is the integral tangent lattice, then Proposition 3.10 with k=1k=1 in the chart shows that the group acts minimally on the ball of radius p−2p^{-2} about qq.

For p=5,7,11p=5,7,11 and a range of parameters θ\theta, each replaced by a random pp-adic integer congruent to it modulo p2p^2, I lifted every smooth point over Fp\mathbb{F}_p to random points qq and computed modulo p16p^{16} the Smith exponents k1≤k2k_1 \le k_2 of this span relative to TqT_q; the exact parameters are in the script (§ 8). Of the 7830 lifts, 7802 have (k1,k2)=(1,1)(k_1,k_2)=(1,1). The other 28 have (2,2)(2,2), and each is congruent modulo p2p^2 to a point of a finite orbit, where all the fields vanish: (0,0,±θ)(0,0,\pm\sqrt{\theta}) up to permutation (18 lifts), a point of the Cayley cubic X4X_4 (7), the orbit of (1,1,1)(1,1,1) on X2X_2 (2) and that of (1,2,0)(1,\sqrt{2},0) on X3X_3 (1). These are among the finite orbits of [16 Section 7.2], near which minimality fails. This is a computation at finitely many points, not a proof of minimality on the whole surface.

Flow-box coordinates by Newton’s method

Given xx and a point z∈B(x,p−k)z \in\mathrm{B}(x,p^{-k}), I compute (s,w)=Ψ−1(z)(s,w)=\Psi^{-1}(z) from Proposition 3.4. With ii as there, the iteration is

s0=0,w0=z,δj=(wj)i−xifi(wj),sj+1=sj+δj,wj+1=φ−δj(wj).(4)s_0=0,\qquad w_0=z,\qquad\delta_j=\frac{(w_j)_i-x_i}{f_i(w_j)},\qquad s_{j+1}=s_j+\delta_j,\qquad w_{j+1}=\varphi_{-\delta_j}(w_j). \tag*{(4)}

The division is exact: the numerator has valuation at least kk and the denominator exactly kk, so δj\delta_j is determined modulo pK−kp^{K-k}, which is the precision of the result.

Proposition 4.1. Let ej=s−sje_j=s-s_j. Then v(ej+1)≥2v(ej)+1v(e_{j+1})\ge2v(e_j)+1, hence v(ej)≥2j−1v(e_j)\ge2^j-1 and v(δj)≥2j−1v(\delta_j)\ge2^j-1. In particular ⌈log⁡2(K−k+1)⌉\lceil\log_2(K-k+1)\rceil steps give ss modulo pK−kp^{K-k}.

Proof. By construction wj=φ−sj(z)=φej(w)w_j=\varphi_{-s_j}(z)=\varphi_{e_j}(w) with wi=xiw_i=x_i. Expanding the Lie series at ww,

(wj)i−xi=ejfi(w)+ρj,ρj=∑n≥2ejnEn(w)i,v(ρj)≥k+1+2v(ej)(w_j)_i-x_i=e_jf_i(w)+\rho_j,\qquad\rho_j=\sum_{n\ge2}e_j^nE_n(w)_i,\qquad v(\rho_j)\ge k+1+2v(e_j)

by Lemma 3.1 at ww, where v(f(w))=kv(f(w))=k. Also ∥wj−w∥=p−k∣ej∣\lVert w_j-w\rVert=p^{-k}|e_j| by Theorem 3.2, so fi(wj)=fi(w)(1+ηj)f_i(w_j)=f_i(w)(1+\eta_j) with v(ηj)≥c+v(ej)≥1+v(ej)v(\eta_j)\ge c+v(e_j)\ge1+v(e_j). Therefore

δj=ej+ρj/fi(w)1+ηj=ej+O(p1+2v(ej)),\delta_j=\frac{e_j+\rho_j/f_i(w)}{1+\eta_j}=e_j+O\left(p^{1+2v(e_j)}\right),

and ej+1=ej−δje_{j+1}=e_j-\delta_j has valuation at least 2v(ej)+12v(e_j)+1. Since v(e0)≥0v(e_0)\ge0, induction gives v(ej)≥2j−1v(e_j)\ge2^j-1, and δj=ej−ej+1\delta_j=e_j-e_{j+1} inherits the bound.

Once ss and ww are computed, Corollary 3.5 gives first hitting times, first return times and orbit averages by closed formulas: an orbit average over balls of radius p−mp^{-m} costs pm−kp^{m-k} flow evaluations instead of a long ergodic sum.

The formal logarithm as a rectifying coordinate

The formal group of Gm\mathbb{G}_m or of an elliptic curve gives a natural example of an analytic field. Its formal logarithm is the rectifying coordinate of the flow, so Theorem 3.2 contains the local divisibility v([t]z)=v(z)+v(t)v([t]z)=v(z)+v(t) for z∈p1+cZpz\in p^{1+c}\mathbb{Z}_p, and Newton’s method of § 5 computes pp-adic logarithms with no loss of digits.

Let F(X,Y)∈Zp[[X,Y]]F(X,Y)\in\mathbb{Z}_p[[X,Y]] be a one-dimensional commutative formal group law and ω(z) dz\omega(z)\,dz its normalised invariant differential, so that ω∈1+zZp[[z]]\omega\in1+z\mathbb{Z}_p[[z]], ω(F(X,Y))∂YF(X,Y)=ω(Y)\omega(\mathcal{F}(X,Y))\partial_Y\mathcal{F}(X,Y)=\omega(Y), and

log⁡F(z)=∫0zω=∑n≥1ωn−1nzn,log⁡F(F(X,Y))=log⁡F(X)+log⁡F(Y),\log_{\mathcal{F}}(z)=\int_0^z\omega=\sum_{n\ge1}\frac{\omega_{n-1}}{n}z^n,\qquad\log_{\mathcal{F}}(\mathcal{F}(X,Y))=\log_{\mathcal{F}}(X)+\log_{\mathcal{F}}(Y),

and exp⁡F∈Qp[[z]]\exp_{\mathcal F} \in\mathbb{Q}_p[[z]] be the compositional inverse of log⁡F\log_{\mathcal F} [32 Sections IV.4–IV.5]. For Gm\mathbb{G}_m, F=X+Y+XY\mathcal F = X + Y + XY, ω=1/(1+z)\omega= 1/(1 + z) and log⁡F(z)=log⁡(1+z)\log_{\mathcal F}(z) = \log(1 + z). For a minimal Weierstrass equation over Zp\mathbb{Z}_p and the parameter z=−x/yz = -x/y, the points of E(Qp)E(\mathbb{Q}_p) with v(x)<0v(x) < 0, together with OO, form the kernel of reduction E(1)(Qp)E^{(1)}(\mathbb{Q}_p), and zz identifies it with pZpp\mathbb{Z}_p under the group law F\mathcal F [32 Propositions VII.2.1–VII.2.2]. Write 1/ω(z)=∑nwnzn1/\omega(z) = \sum_n w_n z^n, with wn∈Zpw_n \in\mathbb{Z}_p and w0=1w_0 = 1, and put

fF(u)=pcω(pu)=∑n≥0pc+nwnun,(5)f_{\mathcal F}(u) = \frac{p^c}{\omega(pu)} = \sum_{n\geq0} p^{c+n}w_nu^n, \tag*{(5)}

with cc as in (1). Its coefficients tend to zero and have minimum valuation cc, so fFf_{\mathcal F} is an analytic field in the sense of Section 2.1. For Gm\mathbb{G}_m it is the linear field pc(1+pu)p^c(1 + pu).

Proposition 5.1. (Formal groups as flows). For all u,s∈Zpu,s \in\mathbb{Z}_p, the flow of (5) satisfies

pφs(u)=F(pu,exp⁡F(p1+cs)),log⁡F(pφs(u))=log⁡F(pu)+p1+cs.(6)p\varphi_s(u) = \mathcal F\left(pu,\exp_{\mathcal F}\left(p^{1+c}s\right)\right), \qquad\log_{\mathcal F}\left(p\varphi_s(u)\right) = \log_{\mathcal F}(pu) + p^{1+c}s. \tag*{(6)}

Consequently:

(a) the orbit of 00 is pcZpp^c\mathbb{Z}_p, and s(u)=p−1−clog⁡F(pu)s(u) = p^{-1-c}\log_{\mathcal F}(pu) is its rectifying coordinate;

(b) exp⁡F\exp_{\mathcal F} and log⁡F\log_{\mathcal F} are mutually inverse isometries of p1+cZpp^{1+c}\mathbb{Z}_p;

(c) for z∈p1+cZpz \in p^{1+c}\mathbb{Z}_p and t∈Zpt \in\mathbb{Z}_p the multiple [t]z=exp⁡F(tlog⁡Fz)[t]z = \exp_{\mathcal F}(t\log_{\mathcal F}z) satisfies v([t]z)=v(z)+v(t)v([t]z) = v(z) + v(t).

Proof. Put e(s)=exp⁡F(p1+cs)e(s) = \exp_{\mathcal F}(p^{1+c}s). Its coefficients p(1+c)nanp^{(1+c)n}a_n, with v(an)≥−v(n!)v(a_n) \geq-v(n!) [32 Section IV.5], have valuation at least n(1+c−1p−1)n(1+c-\frac{1}{p-1}), so ee is a power series in ss converging on Zp\mathbb{Z}_p, with values in p1+cZpp^{1+c}\mathbb{Z}_p, and ω(e)e′=p1+c\omega(e)e' = p^{1+c} since log⁡F(e(s))=p1+cs\log_{\mathcal F}(e(s)) = p^{1+c}s. Let ψ(s)=p−1F(pu,e(s))\psi(s) = p^{-1}\mathcal F(pu,e(s)). All terms of F(pu,e)\mathcal F(pu,e) have positive valuation and those of degree at least two are divisible by p2p^2, so ψ\psi is a convergent power series in ss with values in Zp\mathbb{Z}_p and ψ(0)=u\psi(0) = u. By invariance of ω\omega,

ψ′(s)=p−1∂YF(pu,e)e′=p−1ω(e)ω(F(pu,e))⋅p1+cω(e)=pcω(pψ(s))=fF(ψ(s)).\psi'(s) = p^{-1}\partial_Y\mathcal F(pu,e)e' = p^{-1}\frac{\omega(e)}{\omega(\mathcal F(pu,e))}\cdot\frac{p^{1+c}}{\omega(e)} = \frac{p^c}{\omega(p\psi(s))} = f_{\mathcal F}(\psi(s)).

Power series solutions are unique (Section 2), so ψ(s)=φs(u)\psi(s) = \varphi_s(u), which is the first identity of (6). Applying log⁡F\log_{\mathcal F}, which converges on pZpp\mathbb{Z}_p, gives the second.

Since fF(0)=pcf_{\mathcal F}(0) = p^c, Theorem 3.2 shows that the orbit of 00 is pcZpp^c\mathbb{Z}_p and that ∣φt(0)−φs(0)∣=p−c∣t−s∣|\varphi_t(0)-\varphi_s(0)| = p^{-c}|t-s|. By (6) with u=0u = 0, pφs(0)=exp⁡F(p1+cs)p\varphi_s(0) = \exp_{\mathcal F}(p^{1+c}s) and log⁡F(pφs(0))=p1+cs\log_{\mathcal F}(p\varphi_s(0)) = p^{1+c}s. This gives (a). It also shows that exp⁡F\exp_{\mathcal F} maps p1+cZpp^{1+c}\mathbb{Z}_p isometrically onto p⋅pcZpp\cdot p^c\mathbb{Z}_p, with inverse log⁡F\log_{\mathcal F}, which is (b). For (c), write z=exp⁡F(w)z = \exp_{\mathcal F}(w) with v(w)=v(z)v(w) = v(z); then [t]z=exp⁡F(tw)[t]z = \exp_{\mathcal F}(tw) has valuation v(tw)=v(z)+v(t)v(tw) = v(z) + v(t) by (b). □

For a point QQ of an elliptic curve with v(z(Q))≥1+cv(z(Q)) \geq1 + c, (c) is the familiar statement that [t]Q[t]Q lies exactly v(t)v(t) steps deeper than QQ in the filtration E(n)(Qp)={v(z)≥n}∪{O}E^{(n)}(\mathbb{Q}_p) = \{v(z) \geq n\} \cup\{O\} (written En(Qp)E_n(\mathbb{Q}_p) in [32]). The classical proofs work in the formal group and give (b) on all of pZpp\mathbb{Z}_p for pp odd [32 Theorem IV.6.4]; here the weaker range p1+cZpp^{1+c}\mathbb{Z}_p follows from the metric statement of Theorem 3.2 alone. The same mechanism solves the discrete logarithm on anomalous elliptic curves: after lifting to Qp\mathbb{Q}_p and multiplying by pp into the kernel of reduction, the formal logarithm turns the group law into addition [33, 29]; Semaev’s version of the attack works in characteristic pp [31]. For Gm\mathbb{G}_m the flow is explicit, φs(u)=((1+pu)exp⁡(p1+cs)−1)/p\varphi_s(u) = ((1 + pu)\exp(p^{1+c}s) - 1)/p, and serves as a check.

By Proposition 5.1(a), Newton’s method of Section 4 at x=0x = 0 computes s(u)s(u) modulo pK−cp^{K-c} from u∈pcZpu \in p^c\mathbb{Z}_p modulo pKp^K, hence log⁡F(z)\log_{\mathcal F}(z) modulo pK+1p^{K+1} from zz modulo pK+1p^{K+1} for every z∈p1+cZpz \in p^{1+c}\mathbb{Z}_p, with no loss of digits.

Limitations and open questions

  • The proof of [24 Theorem B] appears to extend to every map in Zp⟨x1,…,xd⟩d\mathbb{Z}_p\langle x_1,\ldots,x_d\rangle^d, which would give Corollary 3.7 without flows; I have not checked this.

  • Proposition 3.10 gives minimality on one ball. Combining it with transitivity modulo pp requires transitivity on the fibres modulo pkp^k, which is related to [16 Question 32]. For the Markoff surface itself, transitivity modulo composite moduli is studied in [25], and for p>5p > 5 Golsefidy and Tamam bound the number of orbits modulo pℓp^\ell uniformly in ℓ\ell [15 Proposition N].

  • The computation on Markoff surfaces in § 4.3 concerns finitely many random lifts and is not a proof of minimality on the whole surface.

  • Proposition 5.1 treats one-dimensional formal groups only.

  • Crespo and Pelayo ask for a pp-adic version of the Eliasson–Vey linearization theorem near nondegenerate critical points of integrable systems [11 Question 5.14]. In the real case the flows of the quadratic parts g1,…,gng_1,\ldots,g_n of the Williamson normal form play a central role: in the elliptic case they generate a torus action, which the proofs use. Over Zp\mathbb{Z}_p the flow of a linear field x′=Axx'=Ax exists for all t∈Zpt \in\mathbb{Z}_p as soon as the entries of AA are divisible by pp (by 4 if p=2p=2), and this can be arranged by multiplying the gig_i by a power of pp, that is, by rescaling time. Since the gig_i Poisson commute, their Hamiltonian flows then give an action of the compact group Zpn\mathbb{Z}_p^n, which carries a Haar measure. I do not know whether these Zpn\mathbb{Z}_p^n-actions can play the role of the torus action.

Code availability

The flows, flow-box coordinates, time-one maps and minimal actions are implemented in the package galrax.padic of the open-source library Galrax, https://github.com/NotTruecCoder/galrax, written in JAX [6] on top of Diffrax [19]. The formal-group flows are in the file formal.py of that package. The numerical check on Markoff surfaces is the standalone script anc/markoff_smith.py in the arXiv source of this paper. From the anc directory it is run as python markoff_smith.py p depth lifts [θ_1,\ldots,θ_r], where the optional list of parameters defaults to 0,…,p2−10,\ldots,p^2-1; the three runs behind § 4.3 are listed in its docstring.

References

  1. [1]V. Anashin and A. Khrennikov. Applied Algebraic Dynamics. de Gruyter Expositions in Mathematics 49. Walter de Gruyter, Berlin, 2009.
  2. [2]J. P. Bell. A generalised Skolem–Mahler–Lech theorem for affine varieties. J. London Math. Soc., 73(2):367–379, 2006.
  3. [3]J. P. Bell, D. Ghioca, and T. J. Tucker. The Dynamical Mordell–Lang Conjecture. Mathematical Surveys and Monographs 210. American Mathematical Society, Providence, RI, 2016.arxiv.org/abs/0712.2344
  4. [4]R. L. Benedetto. Dynamics in One Non-Archimedean Variable. Graduate Studies in Mathematics 198. American Mathematical Society, Providence, RI, 2019.DOI
  5. [5]J. Bourgain, A. Gamburd, and P. Sarnak. Markoff triples and strong approximation. C. R. Math. Acad. Sci. Paris, 354(2):131–135, 2016. doi:10.1016/j.crma.2015.12.006.DOI
  6. [6]J. Bradbury, R. Frostig, P. Hawkins, M. J. Johnson, C. Leary, D. Maclaurin, G. Necula, A. Paszke, J. VanderPlas, S. Wanderman-Milne, and Q. Zhang. JAX: composable transformations of Python+NumPy programs, 2018. https://github.com/jax-ml/jax.
  7. [7]S. Cantat and S. U. Jang. Orbits of automorphism groups of affine surfaces over p-adic fields. Preprint, arXiv:2410.08579, 2024.arxiv.org/abs/2410.08579
  8. [8]S. Cantat and J. Xie. Algebraic actions of discrete groups: the p-adic method. Acta Math., 220(2):239–295, 2018. doi:10.4310/ACTA.2018.v220.n2.a2.DOI
  9. [9]W. Y. Chen. Nonabelian level structures, Nielsen equivalence, and Markoff triples. Ann. of Math. (2), 199(1):301–443, 2024. doi:10.4007/annals.2024.199.1.5.DOI
  10. [10]L. Crespo and Á. Pelayo. Darboux’s theorem in p-adic symplectic geometry. Preprint, arXiv:2508.15443, 2025.arxiv.org/abs/2508.15443
  11. [11]L. Crespo and Á. Pelayo. p-adic symplectic geometry of integrable systems and Weierstrass–Williamson theory I. Preprint, arXiv:2501.14444, 2025.arxiv.org/abs/2501.14444
  12. [12]D. L. desJardins and M. E. Zieve. Polynomial mappings mod pⁿ. Preprint, arXiv:math/0103046, 2001.
  13. [13]J. D. Dixon, M. P. F. du Sautoy, A. Mann, and D. Segal. Analytic Pro-p Groups. Cambridge Studies in Advanced Mathematics 61. Cambridge University Press, 2nd edition, 1999.DOI
  14. [14]A. Fan and L. Liao. On minimal decomposition of p-adic polynomial dynamical systems. Adv. Math., 228(4):2116–2144, 2011. doi:10.1016/j.aim.2011.06.032.DOI
  15. [15]A. Salehi Golsefidy and N. Tamam. Closure of orbits of the pure mapping class group in the character variety. Proc. Natl. Acad. Sci. USA, 122(15):e2416120122, 2025. doi:10.1073/pnas.2416120122.
  16. [16]S. U. Jang. Residual transitivity implies minimality for Markoff surfaces over p-adic integers, by means of p-adic flows. Preprint, arXiv:2502.18976, 2025.arxiv.org/abs/2502.18976
  17. [17]A. Jenkins and S. Spallone. A p-adic approach to local analytic dynamics: analytic conjugacy of analytic maps tangent to the identity. Ann. Fac. Sci. Toulouse Math. (6), 18(3):611–634, 2009. doi:10.5802/afst.1217.DOI
  18. [18]K. S. Kedlaya. p-adic Differential Equations. Cambridge Studies in Advanced Mathematics 125. Cambridge University Press, 2010.
  19. [19]P. Kidger. On Neural Differential Equations. PhD thesis, University of Oxford, 2021. arXiv:2202.02435.arxiv.org/abs/2202.02435
  20. [20]A. Klimov and A. Shamir. A new class of invertible mappings. In Cryptographic Hardware and Embedded Systems — CHES 2002, Lecture Notes in Computer Science 2523, pages 470–483. Springer, 2003.
  21. [21]M. V. Larin. Transitive polynomial transformations of residue class rings. Discrete Math. Appl., 12(2):127–140, 2002.DOI
  22. [22]J. Lubin. Nonarchimedean dynamical systems. Compositio Math., 94(3):321–346, 1994.
  23. [23]K. Mahler. An interpolation series for continuous functions of a $p$-adic variable. J. Reine Angew. Math., 199:23–34, 1958.DOI
  24. [24]K. Mallahi Karai and C. Weiß. Dimensions of orbit closures and discrepancy for dynamical $p$-adic sequences. Preprint, arXiv:2607.26897, 2026.arxiv.org/abs/2607.26897
  25. [25]C. Meiri and D. Puder. The Markoff group of transformations in prime and composite moduli (with an appendix by D. Carmon). Duke Math. J., 167(14):2679–2720, 2018. doi:10.1215/00127094-2018-0024.
  26. [26]F. Mukhamedov. Effective Strassmann certificates for local $p$-adic dynamical Modell–Lang interpolants. Preprint, arXiv:2607.14339, 2026.arxiv.org/abs/2607.14339
  27. [27]B. Poonen. $p$-adic interpolation of iterates. Bull. London Math. Soc., 46(3):525–527, 2014.DOI
  28. [28]J. Rivera-Letelier. Dynamique des fonctions rationnelles sur des corps locaux. Astérisque, 287:147–230, 2003.
  29. [29]T. Satoh and K. Araki. Fermat quotients and the polynomial time discrete log algorithm for anomalous elliptic curves. Comment. Math. Univ. St. Pauli, 47(1):81–92, 1998.
  30. [30]W. H. Schikhof. Ultrametric Calculus: An Introduction to $p$-Adic Analysis. Cambridge Studies in Advanced Mathematics 4. Cambridge University Press, 1984.
  31. [31]I. A. Semaev. Evaluation of discrete logarithms in a group of $p$-torsion points of an elliptic curve in characteristic $p$. Math. Comp., 67(221):353–356, 1998.DOI
  32. [32]J. H. Silverman. The Arithmetic of Elliptic Curves. Graduate Texts in Mathematics 106. Springer, 2nd edition, 2009.
  33. [33]N. P. Smart. The discrete logarithm problem on elliptic curves of trace one. J. Cryptology, 12(3):193–196, 1999.DOI
  34. [34]R. Strassmann. Über den Wertevorrat von Potenzreihen im Gebiet der $p$-adischen Zahlen. J. Reine Angew. Math., 159:13–28, 1928.DOI

Paper details

vlnc:W_6q9v4uvb
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents