\documentclass[11pt]{article} \usepackage[T1]{fontenc} \usepackage{lmodern} \usepackage[hmargin=1in,vmargin=0.88in]{geometry} \usepackage{amsmath,amssymb} \usepackage{fancyvrb} \usepackage{xurl} \usepackage[hidelinks]{hyperref} \urlstyle{same} \setlength{\emergencystretch}{2em} \setcounter{secnumdepth}{0} \providecommand{\tightlist}{\setlength{\itemsep}{0pt}\setlength{\parskip}{0pt}} \title{Harmonic dyadic shifts and short sums of factorial divisors} \author{Will Howe, Grant Huffman, Jensen Kohlmeyer, Liam Kruer} \date{Unpublished draft - 1 November 2026} \begin{document} \maketitle \begin{abstract} Let \(h(N)\) be the maximum, over \(1\le R\le N\), of the least number of distinct positive divisors of a practical integer \(N\) whose sum is \(R\). We prove \[h(n!)=O((\log n)^9).\] The main ingredient is a finite entropy lemma: a law modulo \(2^k\) with entropy at least \(gL\log 2\) in every \(L\)-bit lower prefix becomes supported on more than half the group after \(O(g^{-2}(\log k)^2)\) independent copies, each multiplied by an independently chosen power of two. The shift distribution is harmonic. A divisor-digit construction supplies \(g\asymp1/\log k\) at linear factorial index. Positive quotient recursion, binary carrying, and a factorial radix then give the stated bound. \end{abstract} \section{1. Statement, conventions, and source overlap}\label{statement-conventions-and-source-overlap} All unmarked logarithms are natural; \(\log_2\) has base two. A sum of no terms is zero. Local modular representations may repeat divisors. The final representations use distinct positive divisors. For a practical \(N\ge1\), define \[h(N)=\max_{1\le R\le N} \min\{|D|:D\text{ is a set of positive divisors of }N,\ \sum_{d\in D}d=R\}.\] The original question often uses \(R0\) and \(n_0\) such that \[h(n!)\le C_h(\log n)^9\qquad(n\ge n_0).\] Consequently the strict-power version of Erd\H{o}s 18(c) holds with exponent \(10\). The divisor-block construction and the positive quotient/radix framework are adapted from \href{https://jenwin.io/papers/erdos18b-factorial-divisors.pdf}{Kohlmeyer-Kruer, Section 3 and Sections 5-6}. Plasma AI's author-recorded \href{https://erdosproblems.ai/wiki/theory/number_theory/L63_polylogarithmic_dyadic_representatives}{L63} already gives an overlapping quantitative interface from short dyadic representatives to factorial divisor sums. We give the complete transfer argument for the parameters used here, and credit those earlier transfer mechanisms. The separated-digit entropy obstruction \href{https://erdosproblems.ai/wiki/theory/number_theory/L64_factorial_digit_entropy}{L64} excludes a \textbf{fixed positive} entropy rate at this factorial scale. Our input rate tends to zero, and the harmonic \textbf{additive} shift lemma has polynomial cost in its reciprocal. The final law sums overlapping shifted inputs; it is not a single separated-digit family covered by that support bound. The source's displayed multiplicative block expansion is also not used. The related \href{https://erdosproblems.ai/wiki/theory/number_theory/L66_divisor_residue_spread}{L66} records a vanishing-rate residue estimate for the full odd-divisor law and a conditional multiplicative mixing criterion. That multiplicative criterion is not assumed here. Novelty of the harmonic additive argument remains to be assessed externally. The use of entropy in additive combinatorics has a substantial prior history; see \href{https://arxiv.org/abs/0906.4387}{Tao}, \href{https://arxiv.org/abs/2306.13403}{Green-Manners-Tao}, and the multiscale convolution theorems of \href{https://annals.math.princeton.edu/wp-content/uploads/annals-v180-n2-p07-p.pdf}{Hochman}. Weighted sums over cyclic groups also appear in \href{https://arxiv.org/abs/1512.00135}{Abbe-Li-Madiman}. We do not claim these general ideas as new. The proposed contribution here is the specific harmonic-shift estimate in Lemma 2, uniform as the lower-prefix entropy rate tends to zero, and its application to factorial divisors. A separate novelty report records the scope of the source comparisons; no global priority certification is claimed. \section{2. Harmonic shifts from lower-prefix entropy}\label{harmonic-shifts-from-lower-prefix-entropy} For a finite random variable \(Y\), let \(H(Y)\) denote Shannon entropy in natural units. Conditional probabilities on zero-probability events can be chosen arbitrarily. \textbf{Lemma 2 (harmonic dyadic shifts).} Let \(k,r\ge1\), let \(02^{k-1}\), and \(B+B=\mathbb Z/2^k\mathbb Z\). \textbf{Proof.} Write residues using their canonical \(k\) binary digits, indexed from position zero. Given the lower \(t\) bits of \(X\), let \(p_t\) be the conditional probability that bit \(t\) is one. Define \[u_t=\mathbb E[2\min(p_t,1-p_t)].\] For binary entropy \(\mathcal H(p)=-p\log p-(1-p)\log(1-p)\), \[\mathcal H(p)\le2\sqrt{\min(p,1-p)}.\] Indeed, for \(0\le p\le1/2\), \[-p\log p\le(2/e)\sqrt p,\qquad -(1-p)\log(1-p)\le p\le\sqrt p,\] and \(1+2/e<2\); use symmetry on the other half. Thus \[\mathbb E\mathcal H(p_t)\le\sqrt{2u_t}.\] The entropy chain rule and Cauchy-Schwarz give \[gL\log 2\le\sum_{tj\), it is zero. Thus, writing \(\gamma_{i,j}\) for the conditional sign-bias of that bit given its own lower prefix, and setting \(\gamma_{i,j}=1\) when \(A_i>j\), \[\mathbb E[(-1)^{S_j}\mid\mathcal G_j] =(-1)^{C_j}\prod_i\gamma_{i,j}. \tag{2.5}\] No independence of carries or output bits is asserted. Unconditionally, each \(\gamma_{i,j}\) is a function of its own independent pair. Therefore \[\mathbb E\left|\mathbb E[(-1)^{S_j}\mid\mathcal G_j]\right| =(1-\beta_j)^r, \quad \beta_j=\frac1{\mathcal L_k}\sum_{a=0}^j\frac{u_{j-a}}{a+1}.\] Equation (2.4) gives \[\beta_j\ge \frac{\sum_{t=0}^ju_t}{(j+1)\mathcal L_k} \ge\frac{c_g}{\mathcal L_k}. \tag{2.6}\] A binary law with absolute sign-bias \(z\) has entropy at least \((1-z)\log 2\), by concavity. The output's lower \(j\) bits and the full shift vector are measurable under \(\mathcal G_j\), so \[\begin{aligned} H(S_j\mid S\bmod2^j,\mathbf A) &\ge H(S_j\mid\mathcal G_j)\\ &\ge\log 2\,\bigl[1-(1-\beta_j)^r\bigr]\\ &\ge\log 2\,\bigl[1-e^{-rc_g/\mathcal L_k}\bigr]. \end{aligned}\] This entropy comparison does not assume independence after conditioning only on the output prefix. Summing over \(j\), using the chain rule conditioned on the \textbf{same full vector} \(\mathbf A\), and then \(H(S)\ge H(S\mid\mathbf A)\), proves (2.2). Under (2.3), \(H(S)\ge(k-\tfrac12)\log 2\). Entropy is at most log support size, hence \[|B|\ge2^k/\sqrt2>2^{k-1}.\] For every residue \(x\), the sets \(B\) and \(x-B\) intersect; thus \(B+B\) is the whole group. \(\square\) The proof first bounds \(H(S\mid\mathbf A)\). Randomness in the unobserved shifts is therefore not being used as a hidden entropy source. \section{3. Divisor digits at linear factorial index}\label{divisor-digits-at-linear-factorial-index} Fix \[C_0=e^{100},\qquad k_0=\lceil e^{1000}\rceil.\] For every integer \(k\ge k_0\), put \[m=\lceil C_0k\rceil,\quad c=\frac8{\log m},\quad g=\frac c4=\frac2{\log m},\quad b_0=\lfloor\log_2m\rfloor.\] These choices satisfy \[c\le\tfrac13,\quad g\le\tfrac12,\quad 2\le b_02/3\) and \(m/6>\log(m+1)\) for \(m\ge100\), this proves (3.3). Write \(a=\log_2m\). If \(\ell\le2a+2\), all integers through \(m\) divide \(m!\), so \[|D_\ell|\ge\min(m+1,2^\ell)\ge2^{\ell/3}\ge2^{c\ell};\] here \(a\ge2\) ensures \(2a+2\le3a\). Otherwise put \(v=\lfloor(\ell-1)/a\rfloor\). Then \[\ell/(2a)\le v\le k/a.\] Every product of \(v\) distinct primes at most \(m\) divides \(m!\) and is at most \(m^v\le2^{\ell-1}\). Distinct prime subsets give distinct integers. With \(P=\pi(m)\), (3.3) yields \[P/v\ge C_0/(2\log2),\qquad\log(P/v)>40.\] In particular \(P\ge v\). Since \({P\choose v}\ge(P/v)^v\), \[\log_2|D_\ell| \ge v\log_2(P/v)\ge\frac{20\ell}{\log m}\ge c\ell.\] \(\square\) \textbf{Digit convention.} Start at endpoint \(b_0\). Define successive endpoints \[b_{s+1}=\min(2b_s,k)\] until \(b_L=k\), and let \(\ell_s=b_{s+1}-b_s\). Choose an initial digit \(d_0\) uniformly among the odd integers in \([1,2^{b_0})\). Independently choose \(d_{s+1}\) uniformly from \(D_{\ell_s}\), and set \[U=d_0+\sum_{s=0}^{L-1}2^{b_s}d_{s+1}.\] The number \(J=L+1\) \textbf{includes the initial digit}. A zero digit contributes no divisor term. Before the final block the endpoint doubles, so \(J\le2+\log_2 k\), giving the count below. Because all digits are below their block radices, there are no carries between blocks and \[1\le U<2^k,\qquad U\text{ is odd},\qquad J+1\le3\log k. \tag{3.4}\] Each initial digit divides \(m!\). Every later nonzero term divides \(2^km!\), and \[2^km!\mid(m+2k)!.\] Thus \(U\) is a positive sum of at most \(J\) divisors of \(M!\), where \(M=m+2k\). For \(2\le j\le b_0\), \(U\bmod2^j\) is uniform among odd residues, giving maximum mass \(2^{1-j}\le2^{-gj}\). For \(j>b_0\), let \(b\) be the last completed block endpoint at or below \(j\). If \(j\) cuts a block, that block has length at most \(b\), so \(b\ge j/2\); at \(j=k\), take \(b=k\). Fixing a residue fixes every completed digit. Independence and (3.2) give \[\begin{aligned} \max_a\mathbb P(U=a\bmod2^j) &\le2^{-(b_0-1)}\prod_{\text{completed appended blocks}}|D_\ell|^{-1}\\ &\le2^{1-cb} \le2^{1-cj/2} \le2^{-cj/4}=2^{-gj}. \end{aligned}\tag{3.5}\] The last inequality uses \(j>b_0\ge4/c\). Take an independent uniform \(E\in\{0,1\}\), and let \(V=U+E\) be the \textbf{actual positive integer}; let \(X=V\bmod2^k\) be its residue. Then \[1\le V\le2^k,\] and \(V\) is a sum of at most \(J+1\) divisors of \(M!\). The possible integer value \(V=2^k\) supplies a positive witness for residue zero. Parity of an output residue determines the unique compatible value of \(E\), since \(U\) is odd. Equation (3.5), and uniform parity at \(j=1\), therefore imply \[\max_a\mathbb P(X=a\bmod2^j)\le2^{-gj}\qquad(1\le j\le k). \tag{3.6}\] In particular \(H(X\bmod2^j)\ge gj\log2\), as required in Lemma 2. \section{4. Short positive representatives of every dyadic residue}\label{short-positive-representatives-of-every-dyadic-residue} Apply Lemma 2 to this \(X\), taking \[r=\left\lceil20(\log k)^4\right\rceil.\] This satisfies (2.3): \(\mathcal L_k\le2\log k\), \(\log(2k)\le2\log k\), \(g^{-1}\le\log k\), and \(\log2>2/3\), so the right side of (2.3) is at most \(18(\log k)^4\). Each residue in the support \(B\) of the shifted sum has an actual positive witness formed from \(r\) shifted integers \(V_i\). Expanding their divisor representations gives at most \(r(J+1)\) terms. If \(d\mid M!\) and \(0\le a0\), put \(T=\lceil\log_2(R+1)\rceil\). Then \(R\) is a sum of \(O((\log T)^7)\) distinct divisors of \((4T)!\). \textbf{Proof.} Fix an absolute \(D\ge16A_0/\log2\). For a current positive integer \(R_i\) with bit length \(T_i\), take \[k_i=\left\lfloor\frac{T_i}{D\log(T_i+2)}\right\rfloor.\] Fix a threshold \(T_0\) so that, for every \(T_i\ge T_0\), \begin{itemize} \tightlist \item \(k_i\ge k_0\) and \(k_i\ge T_i/[2D\log(T_i+2)]\ge2\); \item \(F(k_i)\le T_i\); \item the upper bound in (4.3) is at most \(2^{T_i/2}\). \end{itemize} These are compatible eventual conditions. For the height condition, use \[\log(F(k_i)!)\le A_0k_i\log(A_0k_i)\] and, eventually, \(\log(A_0k_i)\le2\log T_i\). The factorial contribution is at most \(2A_0T_i/D\); the logarithm of the count in (4.3) is \(O(\log\log T_i)\). Our fixed choice of \(D\), followed by a fixed sufficiently large \(T_0\), leaves ample margin below \((T_i/2)\log2\). All constants and thresholds are independent of the target \(R\). Choose the positive representative \(S_i\equiv R_i\pmod{2^{k_i}}\) from Section 4. Since \(R_i\ge2^{T_i-1}\), \[00\), the integer \(r\) itself divides \(n!\) and is smaller than all scaled terms. Thus distinctness is preserved, and, for example, \(\mathsf H(n)\le n\) is a valid finite base bound. Fix \[c_0=\frac{\log2}{16},\qquad \ell=\left\lfloor\frac{c_0n}{\log n}\right\rfloor,\qquad m'=n-\ell,\qquad Q=\frac{n!}{m'!}.\] Choose a fixed \(n_*\) so that \(\ell\ge c_0n/(2\log n)\ge1\), \(m'\ge1\), and all following eventual bounds hold for \(n\ge n_*\). For any \(0\le R\le n!\), write \[R=aQ+r,\qquad 0\le rC_h\), and consequently \[\operatorname{practicalH}(n!) \le C_h(\log n)^9<(\log n)^{10}.\] Thus \(C=10>0\) gives the formal strict-power quantifiers, including the endpoint \(n!\). Reading the statement for this comparison is not a formalization or a Lean verification. \textbf{Corollary 6 (the three questions together).} Theorem 1 implies affirmative answers to all three questions of Erd\H{o}s Problem 18: (c) with strict exponent 10, (b) for every positive exponent, and (a) with strict exponent 10 along the factorials. \textbf{Proof.} The implication for (c) was proved above. For (b), fix \(\varepsilon>0\) and write \(x=\log n\). The exponential series gives \[n^\varepsilon=e^{\varepsilon x}\ge \frac{(\varepsilon x)^{10}}{10!}>C_hx^9\] once \(x>10!C_h\varepsilon^{-10}\). For (a), the elementary factorial estimate \((n+1)!\ge2^n\) gives \[\log\log((n+1)!)\ge\log n+\log\log2.\] Consequently \(\log(n+1)\le4\log\log((n+1)!)\) for all sufficiently large \(n\). Set \(y=\log\log((n+1)!)\). Theorem 1 yields \[h((n+1)!)\le C_h4^9y^9C_h4^9\). Factorials are practical and form an unbounded sequence, so this supplies infinitely many witnesses for (a). \(\square\) Part (b) was proved earlier by Kohlmeyer and Kruer \href{https://jenwin.io/papers/erdos18b-factorial-divisors.pdf}{KK}; see also the \href{https://conjectures.io/results/e93a2766-4c70-4564-b565-d0c556f35929}{accepted part-(b) source record}. The implication above does not claim a new discovery of (b). For (a), the earlier work of GPT-5.6 Sol Pro, prompted and posted by Liam Price, and the explicit account of Wouter van Doorn and GPT-6 Astra Pro \href{https://github.com/Woett/ChatGPT-s-note-on-Erdos18/blob/56b455ae6949b76f41292d35829034ebb8663241/Practical\%20numbers\%20and\%20Egyptian\%20fractions.tex}{VD} give the stronger square-log-log bound on a suitable sequence of practical numbers. Our factorial corollary does not improve that general-sequence bound. The proposed new result is the bound at every large factorial in (c); it is not obtained by logically combining the statements of (a) and (b). The earlier greedy factorial bound of Hughes \href{https://arxiv.org/abs/2609.10902v1}{Hughes} also does not supply a fixed polylogarithm. \section{8. Acknowledgments and AI assistance}\label{acknowledgments-and-ai-assistance} The author list is Will Howe, Grant Huffman, Jensen Kohlmeyer, and Liam Kruer. No institutional affiliations are asserted. This manuscript was developed with Codex assistance. AI systems proposed mathematical arguments and performed informal checks, including separate checks of the entropy and arithmetic parts. AI review is not represented as external peer review or formal verification. The accompanying Lean development independently proves the quantitative bound and the three original statements. Its compiler receipts and axiom audits are provided with this draft; this is local verification, not sponsor acceptance. We credit the divisor-digit construction and the positive lifting and factorial radix methods to Kohlmeyer and Kruer \href{https://jenwin.io/papers/erdos18b-factorial-divisors.pdf}{KK}. The overlapping quantitative transfer criterion is recorded by Plasma AI \href{https://erdosproblems.ai/wiki/theory/number_theory/L63_polylogarithmic_dyadic_representatives}{L63}. Its separated-digit obstruction \href{https://erdosproblems.ai/wiki/theory/number_theory/L64_factorial_digit_entropy}{L64} and multiplicative criterion \href{https://erdosproblems.ai/wiki/theory/number_theory/L66_divisor_residue_spread}{L66} delimit, rather than supply, the new mixing step. The arguments required here are proved in the text. \section{9. Formal verification}\label{formal-verification} The accompanying Lean 4.34.0 project proves the unconditional theorem \nolinkurl{Erdos18c.ninthPowerBound} and the exact original statements \nolinkurl{Erdos18.erdos_18a}, \nolinkurl{Erdos18.erdos_18b}, and \nolinkurl{Erdos18.erdos_18c}. It uses mathlib commit \nolinkurl{5ed2965256430c3649e86755f9576b54eca72435}. The printed axiom dependencies are \texttt{propext}, \texttt{Classical.choice}, and \texttt{Quot.sound}; no unproved problem-specific assumptions are introduced. The formal proof realizes the harmonic mixing estimate using convolution and order-two translation differences. It then constructs the divisor digits and positive modular representatives. For the integer lifting and factorial recurrence it uses logarithmic potentials, with different absolute constants from the exposition. It obtains the same ninth-power conclusion. The code adapts credited infrastructure from Kohlmeyer and Kruer's 18(b) source, whose SHA-256 is \nolinkurl{4d6ee89617415288e9f127beb3c0e0211c011f891889fcb43d0e16373bd14088}. The final statement file reproduces only the definitions and target types from the pinned Formal Conjectures source, and proves their equivalence to the internal definitions. It does not import the unproved conjectures. Build instructions, the source correspondence, and limitations of the local verification are recorded in the accompanying verification report. \section{References}\label{references} \begin{enumerate} \def\labelenumi{\arabic{enumi}.} \tightlist \item J. Kohlmeyer and L. Kruer, \emph{Short sums of factorial divisors and Erd\H{o}s Problem 18(b)}, manuscript dated 16 September 2026. \href{https://jenwin.io/papers/erdos18b-factorial-divisors.pdf}{Author-hosted PDF}. The cited version has SHA-256 \nolinkurl{f89f4f46b9535c5c990a29bf0ddf1eb305123429a3b89dfb508e95c8cc9b2d6a}. \item Plasma AI, \emph{Polylogarithmic dyadic representatives imply short factorial sums}, L63, updated 14 October 2026. \href{https://erdosproblems.ai/wiki/theory/number_theory/L63_polylogarithmic_dyadic_representatives}{Statement}; \href{https://erdosproblems.ai/wiki/theory/number_theory/L63_polylogarithmic_dyadic_representatives/_proof}{proof}. Author-recorded conditional result; accessed 1 November 2026. \item Plasma AI, \emph{Subexponential support of short factorial divisor blocks}, L64, updated 14 October 2026. \href{https://erdosproblems.ai/wiki/theory/number_theory/L64_factorial_digit_entropy}{Statement}; \href{https://erdosproblems.ai/wiki/theory/number_theory/L64_factorial_digit_entropy/_proof}{proof}. Accessed 1 November 2026. \item Plasma AI, \emph{Full-divisor residue spread and a quantitative mixing criterion}, L66, updated 14 October 2026. \href{https://erdosproblems.ai/wiki/theory/number_theory/L66_divisor_residue_spread}{Statement}; \href{https://erdosproblems.ai/wiki/theory/number_theory/L66_divisor_residue_spread/_proof}{proof}. Accessed 1 November 2026. \item T. F. Bloom, \emph{Erd\H{o}s Problem 18}, \href{https://www.erdosproblems.com/18}{problem record} and \href{https://www.erdosproblems.com/forum/thread/18/proof-claims}{proof-claim record}, accessed 1 November 2026. These are problem/status sources, not proof inputs. \item P. Erd\H{o}s and R. L. Graham, \emph{Old and New Problems and Results in Combinatorial Number Theory}, Monographies de L'Enseignement Mathématique 28, Université de Genève, 1980. Historical problem reference; no theorem from this book is used without proof here. \item The Formal Conjectures Authors, \emph{Erd\H{o}s Problem 18}, \href{https://github.com/google-deepmind/formal-conjectures/blob/a64e959c32c270df4dd0d96d2027481f15750534/FormalConjectures/ErdosProblems/18.lean}{Lean statements}, commit \nolinkurl{a64e959c32c270df4dd0d96d2027481f15750534}, Apache-2.0. \item T. Tao, \emph{Sumset and inverse sumset theory for Shannon entropy}, Combinatorics, Probability and Computing 19 (2010), 603-639, DOI \href{https://doi.org/10.1017/S0963548309990642}{10.1017/S0963548309990642}. \href{https://arxiv.org/abs/0906.4387}{arXiv:0906.4387}. Background, not a quantitative input to Lemma 2. \item B. Green, F. Manners and T. Tao, \emph{Sumsets and entropy revisited}, Random Structures \& Algorithms 66 (2025), e21252, DOI \href{https://doi.org/10.1002/rsa.21252}{10.1002/rsa.21252}. \href{https://arxiv.org/abs/2306.13403}{arXiv:2306.13403}. Background on entropic doubling and homomorphisms. \item E. Abbe, J. Li and M. Madiman, \emph{Entropies of weighted sums in cyclic groups and an application to polar codes}, Entropy 19 (2017), 235, DOI \href{https://doi.org/10.3390/e19090235}{10.3390/e19090235}. \href{https://arxiv.org/abs/1512.00135}{arXiv:1512.00135}. Related weighted-sum and source-polarization context. \item M. Hochman, \emph{On self-similar sets with overlaps and inverse theorems for entropy}, Annals of Mathematics 180 (2014), 773-822, DOI \href{https://doi.org/10.4007/annals.2014.180.2.7}{10.4007/annals.2014.180.2.7}. \href{https://annals.math.princeton.edu/wp-content/uploads/annals-v180-n2-p07-p.pdf}{Published paper}, especially Theorems 2.7-2.9. Background, not an input to Lemma 2. \item W. van Doorn and GPT-6 Astra Pro, \emph{Practical numbers and Egyptian fractions}, \href{https://github.com/Woett/ChatGPT-s-note-on-Erdos18/blob/56b455ae6949b76f41292d35829034ebb8663241/Practical\%20numbers\%20and\%20Egyptian\%20fractions.tex}{source manuscript}, commit \nolinkurl{56b455ae6949b76f41292d35829034ebb8663241}, September 2026. The source credits \emph{Sparse Divisor Sums} by GPT-5.6 Sol Pro, prompted and posted by Liam Price. No new priority for (a) is claimed here. \item S. D. Hughes, \emph{Sums of distinct divisors of factorials}, \href{https://arxiv.org/abs/2609.10902v1}{arXiv:2609.10902v1}, 9 September 2026. \item Conjectures.io, \emph{Erd\H{o}s 18(b)}, \href{https://conjectures.io/results/e93a2766-4c70-4564-b565-d0c556f35929}{accepted result record}, accessed 1 November 2026. Records acceptance for (b); no local replay is implied by citing the record. \end{enumerate} \end{document}