Chapter 8.3

Arithmetical Aspects of Power Series

[65/98%]
Show
LaTeX
Problem 8.130

The product of any mm consecutive integers is divisible by the product of the first mm positive integers.

?
Problem 8.131

The product of mm integers in arithmetic progression is divisible by m!m! if the common difference of the progression is prime to m!m!.

?
Problem 8.132

The product of the differences of any mm integers is divisible by the product of the differences of the first mm positive integers. (The product of the differences of mm variables x1,x2,…,xmx_1,x_2,\ldots ,x_m, ∏j<k1,2,…,m(xk−xj)\prod_{j<k}^{1,2,\ldots ,m}(x_k-x_j), consists of 12m(m−1)\tfrac 12m(m-1) factors.)

?
Remark.
?

Solution 96 of Part V.

Problem 8.133
?
(1)

For what integers nn is (n−1)!(n-1)! not divisible by nn?

(2)

For what integers nn is (n−1)!(n-1)! not divisible by n2n^2?

Problem 8.134

What is the exponent of the highest power of a prime number pp that is a divisor of n!n!?

?
Problem 8.135

In how many zeros does 1000! end when expressed in the decimal system?

?
Problem 8.136

Let aa and bb be positive integers. Then the integer 2a! 2b!2a!\, 2b! is divisible by a! (a+b)! b!a!\, (a+b)!\, b!.

?
Remark.
?

Cf. Problem 8.

Problem 8.137

If hh and nn are positive integers, then (hn)!/[(h!)nn!](hn)!/[(h!)^nn!] is an integer.

?
Problem 8.138

Let ss be an integer, tt a positive integer, (s,t)=1(s,t)=1. Show that only prime factors of tt divide the denominators of the coefficients of the power series with rational coefficients

(1+z)st=∑n=0∞st(st−1)⋯(st−n+1)znn!. (1+z)^{\frac st} = \sum _{n=0}^\infty \frac st\left(\frac st-1\right)\cdots \left(\frac st-n+1\right)\frac{z^n}{n!} .
?
Problem 8.139

Let pp be a prime factor of tt and let pα,pα0,pα1,…,pαn,…p^\alpha ,p^{\alpha_0},p^{\alpha_1},\ldots ,p^{\alpha_n},\ldots respectively be the highest powers of pp that divide t,t0,t1,…,tn,…t,t_0,t_1,\ldots ,t_n,\ldots, where tnt_n denotes the denominator of the nnth coefficient in the power series expansion of (1+z)s/t(1+z)^{s/t}; (s,t)=1(s,t)=1. Calculate lim⁡n→∞(αn/n)\lim_{n\to \infty }(\alpha_n/n).

?
Remark.
?

We term the function f(z)f(z) an algebraic function, if it satisfies an equation of the form P0(z)[f(z)]l+P1(z)[f(z)]l−1+⋯+Pl−1(z)f(z)+Pl(z)=0P_0(z)[f(z)]^l+P_1(z)[f(z)]^{l-1}+\cdots +P_{l-1}(z)f(z)+P_l(z)=0 where P0(z),P1(z),…,Pl(z)P_0(z),P_1(z),\ldots ,P_l(z) are polynomials, P0(z)≠0P_0(z)\neq 0.

The proper perspective for evaluating the particular result contained in 138, 139 is given by the following general theorem due to Eisenstein: ``If the power series with rational coefficients a1z+a2z2+a3z3+⋯+anzn+⋯a_1z+a_2z^2+a_3z^3+\cdots +a_nz^n+\cdots represents an algebraic function of zz, then there exists an integer TT, such that a1Tz+a2T2z2+a3T3z3+⋯+anTnzn⋯a_1Tz+a_2T^2z^2+a_3T^3z^3+\cdots +a_nT^nz^n\cdots has integral coefficients.'' [153.]

Problem 8.140

Find the smallest integer TT with the property that all the coefficients of the power series expansion (1+Tz)s/t(1+Tz)^{s/t} are integers.

?
Problem 8.141

Prove Eisenstein's theorem in the simplest special case, namely for rational functions.

?
Problem 8.142

We say that the power series

s0t0+s1t1z+s2t2z2+⋯+sntnzn+⋯ \frac{s_0}{t_0}+\frac{s_1}{t_1}z+\frac{s_2}{t_2}z^2+\cdots +\frac{s_n}{t_n}z^n+\cdots

with rational coefficients (sn,tns_n,t_n integers, tn≥1t_n\geq 1, (sn,tn)=1(s_n,t_n)=1) satisfies the Eisenstein condition, if there exists an integer TT such that TnT^n is divisible by tnt_n for n=1,2,3,…n=1,2,3,\ldots. If two power series f(z)f(z) and g(z)g(z) with rational coefficients satisfy the Eisenstein condition, then the series f(z)+g(z)f(z)+g(z), f(z)−g(z)f(z)-g(z), f(z)g(z)f(z)g(z), also satisfy the condition, and also f(z)/g(z)f(z)/g(z) (assuming g(0)⋛0g(0)\gtreqless 0) and f(g(z))f(g(z)) (assuming g(0)=0g(0)=0).

?
Problem 8.143

If the two power series with rational coefficients

a0+a1z+a2z2+⋯+anzn+⋯ ,b0+b1z+b2z2+⋯+bnzn+⋯ a_0+a_1z+a_2z^2+\cdots +a_nz^n+\cdots , \qquad b_0+b_1z+b_2z^2+\cdots +b_nz^n+\cdots

satisfy the Eisenstein condition, then the series

a0b0+a1b1z+a2b2z2+⋯+anbnzn+⋯ a_0b_0+a_1b_1z+a_2b_2z^2+\cdots +a_nb_nz^n+\cdots

also satisfies the condition.

?
Problem 8.144

Show that the Eisenstein condition is equivalent to the following pair of simultaneous conditions:

?
(1)

The denominators t1,t2,t3,…,tn,…t_1,t_2,t_3,\ldots ,t_n,\ldots contain only finitely many different prime numbers.

(2)

(log⁡tn)/n(\log t_n)/n is bounded, n=1,2,3,…n=1,2,3,\ldots.

Problem 8.145

Expand the two functions

11−z+12−z+14−z+⋯+12n−z+⋯ , \frac{1}{1-z}+\frac{1}{2-z}+\frac{1}{4-z}+\cdots +\frac{1}{2^n-z}+\cdots , z2−z+z224−z2+z329−z3+⋯+zn2n2−zn+⋯ \frac{z}{2-z}+\frac{z^2}{2^4-z^2}+\frac{z^3}{2^9-z^3}+\cdots +\frac{z^n}{2^{n^2}-z^n}+\cdots

in powers of zz. Do the power series obtained satisfy the two conditions (1), (2) in 144? Are the functions represented algebraic?

?
Problem 8.146

By a hypergeometric series we mean a series of the form

F(α,β,γ;z)=∑n=0∞α(α+1)⋯(α+n−1)⋅β(β+1)⋯(β+n−1)1⋅2⋯n⋅γ(γ+1)⋯(γ+n−1)zn. F(\alpha ,\beta ,\gamma ;z) = \sum _{n=0}^\infty \frac{\alpha (\alpha +1)\cdots (\alpha +n-1)\cdot \beta (\beta +1)\cdots (\beta +n-1)}{1\cdot 2\cdots n\cdot \gamma (\gamma +1)\cdots (\gamma +n-1)}z^n .

If α\alpha and β\beta are rational numbers but not both integers, and if γ\gamma is a positive integer, then it is possible to choose a positive integer TT such that all the coefficients in the hypergeometric series F(α,β,γ;Tz)F(\alpha ,\beta ,\gamma ;Tz) become integers. (The function

2π∫01dx(1−x2)(1−zx2)=∑n=0∞(12(12+1)⋯(12+n−1)1⋅2⋯n)2zn \frac2\pi \int _0^1 \frac{dx}{\sqrt{(1-x^2)(1-zx^2)}} = \sum _{n=0}^\infty \left(\frac{\tfrac 12(\tfrac 12+1)\cdots (\tfrac 12+n-1)}{1\cdot 2\cdots n}\right)^2 z^n

for example is not an algebraic function [Problem 90 of Part I]: the Eisenstein condition is necessary but not sufficient for a function to be algebraic.)

?
Problem 8.147

Assume that α≠γ\alpha \neq \gamma, β≠γ\beta \neq \gamma and α,β,γ\alpha ,\beta ,\gamma are not all rational, but all the coefficients in the hypergeometric series F(α,β,γ;z)F(\alpha ,\beta ,\gamma ;z) are rational. Then α\alpha and β\beta are the roots of an irreducible quadratic equation with rational coefficients.

?
Problem 8.148

In the case discussed in 147, F(α,β,γ;z)F(\alpha ,\beta ,\gamma ;z) does not represent an algebraic function of zz. Prove this from Eisenstein's theorem!

?
Problem 8.149

A rational function (quotient of two polynomials), whose expansion in ascending powers of zz has only rational numbers as coefficients, is the quotient of two polynomials with rational coefficients.

?
Problem 8.150

Let the power series f(z)f(z) with rational coefficients satisfy an equation of the form

P0(z)[f(z)]l+P1(z)[f(z)]l−1+⋯+Pl−1(z)f(z)+Pl(z)=0; P_0(z)[f(z)]^l+P_1(z)[f(z)]^{l-1}+\cdots +P_{l-1}(z)f(z)+P_l(z)=0 ;

where P0(z),P1(z),…,Pl(z)P_0(z),P_1(z),\ldots ,P_l(z) are polynomials, P0(z)≠0P_0(z)\neq 0. Show that f(z)f(z) also satisfies an equation of the same form, where the polynomials occurring in it have coefficients that are rational integers.

?
Problem 8.151

Let the power series with rational coefficients y=a0+a1z+a2z2+⋯+anzn+⋯y=a_0+a_1z+a_2z^2+\cdots +a_nz^n+\cdots satisfy an algebraic differential equation, i.e. an equation of the form

R(z,y,y′,y′′,…,y(r))=0, R(z,y,y',y'',\ldots ,y^{(r)}) = 0 ,

where RR denotes a rational entire function (multinomial) of its r+2r+2 arguments. Show that yy must satisfy an equation of the form

R∗(z,y,y′,y′′,…,y(r))=0, R^*(z,y,y',y'',\ldots ,y^{(r)}) = 0 ,

where R∗R^* denotes a rational entire function with integral coefficients.

?
Problem 8.152

Let the functions f(z)=c0+c1z+c2z2+⋯f(z)=c_0+c_1z+c_2z^2+\cdots, F0(z),F1(z),F2(z),…,Fl(z)F_0(z),F_1(z),F_2(z),\ldots ,F_l(z) (F0(z)≠0F_0(z)\neq 0) be regular in a neighbourhood of the origin and related by the equation

F0(z)[f(z)]l+F1(z)[f(z)]l−1+⋯+Fl−1(z)f(z)+Fl(z)=0. F_0(z)[f(z)]^l+F_1(z)[f(z)]^{l-1}+\cdots +F_{l-1}(z)f(z)+F_l(z)=0 .

Then the power series

f∗(z)=cm+cm+1z+cm+2z2+⋯ , f^*(z) = c_m+c_{m+1}z+c_{m+2}z^2+\cdots ,

if mm is suitably chosen, satisfies an equation of the form

G0(z)[f∗(z)]k+G1(z)[f∗(z)]k−1+⋯+Gk−1(z)f∗(z)+Gk(z)=0, G_0(z)[f^*(z)]^k+G_1(z)[f^*(z)]^{k-1}+\cdots +G_{k-1}(z)f^*(z)+G_k(z)=0 ,

where k≤lk\leq l, G0(0)=G1(0)=⋯=Gk−2(0)=0G_0(0)=G_1(0)=\cdots =G_{k-2}(0)=0, Gk−1(0)≠0G_{k-1}(0)\neq 0, G0(z)≢0G_0(z)\not\equiv 0 and Gκ(z)G_\kappa (z), κ=0,1,2,…,k\kappa =0,1,2,\ldots ,k, is a rational entire expression with integral coefficients in F0(z),F1(z),…,Fl(z),z,c0,c1,…,cm−1F_0(z),F_1(z),\ldots ,F_l(z),z,c_0,c_1,\ldots ,c_{m-1}, possibly divided by a power of zz. (Function-theoretic preparation for 153, 154.)

?
Problem 8.153

If the power series P0(z),P1(z),…,Pl(z)P_0(z),P_1(z),\ldots ,P_l(z) with rational coefficients satisfy the Eisenstein condition, and if the power series f(z)f(z) with rational coefficients satisfies the equation

P0(z)[f(z)]l+P1(z)[f(z)]l−1+⋯+Pl−1(z)f(z)+Pl(z)=0, P_0(z)[f(z)]^l+P_1(z)[f(z)]^{l-1}+\cdots +P_{l-1}(z)f(z)+P_l(z)=0 ,

then f(z)f(z) also satisfies the Eisenstein condition.

?
Problem 8.154

Let sns_n and tnt_n be integers, (sn,tn)=1(s_n,t_n)=1, tn≥1t_n\geq 1, and denote the greatest prime factor that divides tnt_n by PnP_n. We say that the power series

s0t0+s1t1z+s2t2z2+⋯+sntnzn+⋯ \frac{s_0}{t_0}+\frac{s_1}{t_1}z+\frac{s_2}{t_2}z^2+\cdots +\frac{s_n}{t_n}z^n+\cdots

satisfies the Tchebychev condition if Pn/nP_n/n is bounded, and the Hurwitz condition if log⁡Pn/log⁡n\log P_n/\log n is bounded, n=2,3,4,…n=2,3,4,\ldots. For the Tchebychev condition as well as for the Hurwitz condition analogous theorems to 142, 143, and 153 hold.

?
Problem 8.154.1

Let P(z),Q(z),Q2(z),Q3(z),…P(z),Q(z),Q_2(z),Q_3(z),\ldots be power series with rational coefficients, and set y=c0+c1z+c2z2+⋯y=c_0+c_1z+c_2z^2+\cdots. If one of the following six equations holds

y=P(z)+Q(z),y=P(z)−Q(z),y=P(z)Q(z), y=P(z)+Q(z) , \qquad y=P(z)-Q(z) , \qquad y=P(z)Q(z) , y=P(z)Q(z),assuming Q(0)=1[142], y=\frac{P(z)}{Q(z)} , \quad \text{assuming } Q(0)=1 \qquad \text{[\textbf{142}]} , y=P[Q(z)],assuming Q(0)=0, y=P[Q(z)] , \quad \text{assuming } Q(0)=0 , y=Q(z)+Q2(z)y2+Q3(z)y3+⋯+Ql(z)yl,assuming Q2(0)=Q3(0)=⋯=Ql(0)=0[152], y=Q(z)+Q_2(z)y^2+Q_3(z)y^3+\cdots +Q_l(z)y^l , \quad \text{assuming } Q_2(0)=Q_3(0)=\cdots =Q_l(0)=0 \qquad \text{[\textbf{152}]} ,

then cnc_n is a uniquely determined rational number, given as a rational and entire function of the coefficients of 1,z,z2,…,zn1,z,z^2,\ldots ,z^n in the expansions of P(z),Q(z),Q2(z),…P(z),Q(z),Q_2(z),\ldots, which therefore can have only such prime numbers in its denominator as occur also in the denominators of the coefficients mentioned above. [Cf. op. cit. 144.]

?
Problem 8.154.1

Let P(z),Q(z),Q2(z),Q3(z),…P(z),Q(z),Q_2(z),Q_3(z),\ldots be power series with rational coefficients, and set y=c0+c1z+c2z2+⋯y=c_0+c_1z+c_2z^2+\cdots. If one of the following six equations holds

y=P(z)+Q(z),y=P(z)−Q(z),y=P(z)Q(z), y=P(z)+Q(z) , \qquad y=P(z)-Q(z) , \qquad y=P(z)Q(z) , y=P(z)Q(z),assuming Q(0)=1[142], y=\frac{P(z)}{Q(z)} , \quad \text{assuming } Q(0)=1 \qquad \text{[\textbf{142}]} , y=P[Q(z)],assuming Q(0)=0, y=P[Q(z)] , \quad \text{assuming } Q(0)=0 , y=Q(z)+Q2(z)y2+Q3(z)y3+⋯+Ql(z)yl,assuming Q2(0)=Q3(0)=⋯=Ql(0)=0[152], y=Q(z)+Q_2(z)y^2+Q_3(z)y^3+\cdots +Q_l(z)y^l , \quad \text{assuming } Q_2(0)=Q_3(0)=\cdots =Q_l(0)=0 \qquad \text{[\textbf{152}]} ,

then cnc_n is a uniquely determined rational number, given as a rational and entire function of the coefficients of 1,z,z2,…,zn1,z,z^2,\ldots ,z^n in the expansions of P(z),Q(z),Q2(z),…P(z),Q(z),Q_2(z),\ldots, which therefore can have only such prime numbers in its denominator as occur also in the denominators of the coefficients mentioned above. [Cf. op. cit. 144.]

?
Problem 8.155

We term the power series with integral coefficients a0+a1z+a2z2+⋯+anzn+⋯a_0+a_1z+a_2z^2+\cdots +a_nz^n+\cdots primitive if the numbers a0,a1,a2,…,an,…a_0,a_1,a_2,\ldots ,a_n,\ldots have no common divisor apart from ±1\pm 1. Prove that the product of two primitive power series is a primitive power series.

?
Problem 8.156

If the power series with integral coefficients a0+a1z+a2z2+⋯+anzn+⋯a_0+a_1z+a_2z^2+\cdots +a_nz^n+\cdots represents a rational function, then it can be expressed in the form P(z)/Q(z)P(z)/Q(z), where P(z)P(z) and Q(z)Q(z) are polynomials with integral coefficients and Q(0)=1Q(0)=1. [155.]

?
Problem 8.157

Let

ϑ=0.a1a2a3…an…=a110+a2102+a3103+⋯+an10n+⋯ \vartheta = 0.a_1a_2a_3\ldots a_n\ldots = \frac{a_1}{10}+\frac{a_2}{10^2}+\frac{a_3}{10^3}+\cdots +\frac{a_n}{10^n}+\cdots

be a positive number ≤1\leq 1 expressed in the decimal system, 0≤aν≤90\leq a_\nu \leq 9, ν=1,2,3,…\nu =1,2,3,\ldots. The power series with integral coefficients a1z+a2z2+a3z3+⋯+anzn+⋯a_1z+a_2z^2+a_3z^3+\cdots +a_nz^n+\cdots represents a rational function if ϑ\vartheta is rational, and does not represent a rational function if ϑ\vartheta is irrational.

?
Problem 8.158

Let there be only finitely many distinct numbers in the infinite sequence of numbers a0,a1,a2,…,an,…a_0,a_1,a_2,\ldots ,a_n,\ldots. Then the power series a0+a1z+a2z2+⋯+anzn+⋯a_0+a_1z+a_2z^2+\cdots +a_nz^n+\cdots represents a rational function if and only if the sequence of coefficients after a certain number of terms is periodic.

?
Problem 8.159

Let ll be an integer, l≥0l\geq 0, and P(z)P(z) a polynomial with integral coefficients. In the power series expansion of P(z)(1−z)−l−1P(z)(1-z)^{-l-1} the coefficients are integers that form a periodic sequence modulo any prime number pp; the length of the period is a power of pp.

?
Problem 8.160

Let DD be an integer and not divisible by the odd prime number pp. The coefficients in the power series expansion of [(D−1)z]/[(1−Dz)(1−z)][(D-1)z]/[(1-Dz)(1-z)] are periodic mod pp, where the length of the period is a proper or an improper divisor of p−1p-1.

?
Problem 8.161

Let the integers DD and pp be as in 160. In the power series expansion of (1−Dz2)−1(1-Dz^2)^{-1} the coefficients are periodic mod pp. The length of the period is always a divisor of 2(p−1)2(p-1). It is or is not a divisor of p−1p-1 according to (Dp)=+1\left(\dfrac Dp\right)=+1 or −1-1.

?
Problem 8.162

The sequence of Fibonacci numbers (cf. Problem 53 of Part VII) is periodic mod mm, where mm is any integer. Calculate the length of the period for all prime numbers less than 30.

?
Problem 8.163

In a power series expansion with integral coefficients that represents a rational function the coefficients, after a certain number of terms, form a periodic sequence mod mm (mm arbitrary).

?
Problem 8.164

The algebraic function 1/1−4z1/\sqrt{1-4z} yields on expansion in powers of zz a power series with integral coefficients. These coefficients are for no odd prime number pp periodic mod pp. (163 cannot be extended to algebraic functions.)

?
Problem 8.165

The radius of convergence of a non-terminating power series with integral coefficients is ≤1\leq 1.

?
Remark.
?

For the background of 165--173 see G. Pólya: Math. Ann. 77, pp. 497--513 (1916), also Proc. Lond. Math. Soc., Ser. 2, 21, pp. 22--38 (1923), and F. Carlson: Math. Z. 9, pp. 1--13 (1921). For heuristic considerations see also MPR, 2, p. 46, ex. 14.

Problem 8.166

A non-terminating power series with integral coefficients that is convergent in the interior of the unit circle cannot represent a bounded function there.

?
Problem 8.167

If a power series with integral coefficients represents an algebraic function that is not a rational function, then its radius of convergence is <1<1.

?
Problem 8.168

In 167 the upper bound 1 is the best possible. In other words: If ε>0\varepsilon >0, then there exists a power series with integral coefficients that represents an algebraic function that is not a rational function and whose radius of convergence is >1−ε>1-\varepsilon.

?
Problem 8.169

Let the coefficients of the polynomial

P(x)=a0xr+a1xr−1+⋯+ar−1x+ar,a0≠0,r≥1 P(x) = a_0x^r+a_1x^{r-1}+\cdots +a_{r-1}x+a_r , \qquad a_0\neq 0, \quad r\geq 1

be real. The analytic function defined by the power series

[P(0)]+[P(1)]z+[P(2)]z2+⋯+[P(n)]zn+⋯ [P(0)]+[P(1)]z+[P(2)]z^2+\cdots +[P(n)]z^n+\cdots

is

  1. a rational function, if the rr numbers a0,a1,…,ar−1a_0,a_1,\ldots ,a_{r-1} are all rational,

  2. not a rational function, if the rr numbers a0,a1,…,ar−1a_0,a_1,\ldots ,a_{r-1} are not all rational. (Cf. Problem 168 of Part II.) [Problem 85 of Part I.]

?
Problem 8.170

If the sequence of non-negative integers a1,a2,a3,…,an,…a_1,a_2,a_3,\ldots ,a_n,\ldots is bounded and contains infinitely many terms that are different from 0, then the series

a1z1−z+a2z21−z2+a3z31−z3+⋯+anzn1−zn+⋯ a_1\frac{z}{1-z}+a_2\frac{z^2}{1-z^2}+a_3\frac{z^3}{1-z^3}+\cdots +a_n\frac{z^n}{1-z^n}+\cdots

does not represent a rational function.

?
Problem 8.171

If the sequence a1,a2,a3,…,an,…a_1,a_2,a_3,\ldots ,a_n,\ldots satisfies the same assumptions as in 170, then the series

a1z1+z+a2z21+z2+a3z31+z3+⋯+anzn1+zn+⋯ a_1\frac{z}{1+z}+a_2\frac{z^2}{1+z^2}+a_3\frac{z^3}{1+z^3}+\cdots +a_n\frac{z^n}{1+z^n}+\cdots

does not represent a rational function.

?
Problem 8.172

Denote by QnQ_n the ``cross sum'' of the number nn, i.e. the sum of its digits. E.g. we have Q137=1+3+7=11Q_{137}=1+3+7=11. The power series

Q1z+Q2z2+Q3z3+⋯+Qnzn+⋯ Q_1z+Q_2z^2+Q_3z^3+\cdots +Q_nz^n+\cdots

has the unit circle as its natural boundary.

?
Problem 8.173

In an infinite table of logarithms let the logarithms of all the positive integers 1,2,3,…1,2,3,\ldots be written underneath each other in their natural order, in such a fashion that decimal places of the same order appear in the same vertical column. If we interpret the digits in any vertical column as the coefficients of a power series, then we obtain a power series that cannot be analytically continued.

In a different formulation we are required to prove the following theorem: Let jj be an integer and in the decimal representation of log⁡1,log⁡2,log⁡3,…,log⁡n,…\log 1,\log 2,\log 3,\ldots ,\log n,\ldots let the jjth digit after the decimal point be denoted by d1,d2,d3,…,dn,…d_1,d_2,d_3,\ldots ,d_n,\ldots respectively. Then the analytic function defined by the power series

d1z+d2z2+d3z3+⋯+dnzn+⋯ d_1z+d_2z^2+d_3z^3+\cdots +d_nz^n+\cdots

has as its natural boundary the circle ∣z∣=1\left|z\right|=1.

?
Problem 8.174

If f(z)f(z) is a power series with H-integral coefficients, then

df(z)dzand∫0zf(z) dz \frac{df(z)}{dz} \qquad \text{and} \qquad \int _0^z f(z)\, dz

are also power series with H-integral coefficients.

?
Problem 8.175

If f(z)f(z) and g(z)g(z) are power series with H-integral coefficients, then

f(z)+g(z),f(z)−g(z),f(z)g(z) f(z)+g(z) , \qquad f(z)-g(z) , \qquad f(z)g(z)

are also power series with H-integral coefficients, and under the further assumption that g(0)=±1g(0)=\pm 1 also

f(z)g(z). \frac{f(z)}{g(z)} .
?
Problem 8.176

If f(z)f(z) is a power series with H-integral coefficients and f(0)=0f(0)=0, then

[f(z)]mm! \frac{[f(z)]^m}{m!}

is also a power series with H-integral coefficients, n=1,2,3,…n=1,2,3,\ldots [174].

?
Problem 8.177

If f(z)f(z) and g(z)g(z) are power series with H-integral coefficients and f(0)=0f(0)=0, then g[f(z)]g[f(z)] is also a power series with H-integral coefficients.

?
Problem 8.178

The function φ(z)\varphi (z), well defined by the differential equation

(dφ(z)dz)2=1−[φ(z)]4 \left(\frac{d\varphi (z)}{dz}\right)^2 = 1-[\varphi (z)]^4

and the initial conditions φ(0)=0\varphi (0)=0, φ′(0)>0\varphi '(0)>0, may be expanded in a power series with H-integral coefficients.

?
Problem 8.179
(ez−1)3≡2(z33!+z55!+z77!+⋯ )(mod4). (e^z-1)^3 \equiv 2\left(\frac{z^3}{3!}+\frac{z^5}{5!}+\frac{z^7}{7!}+\cdots \right) \pmod4 .
?
Problem 8.180

For every prime number pp, we have

(ez−1)p−1≡−(zp−1(p−1)!+z2(p−1)(2p−2)!+z3(p−1)(3p−3)!+⋯ )(modp). (e^z-1)^{p-1} \equiv -\left(\frac{z^{p-1}}{(p-1)!}+\frac{z^{2(p-1)}}{(2p-2)!}+\frac{z^{3(p-1)}}{(3p-3)!}+\cdots \right) \pmod p .
?
Problem 8.181

For every composite number mm greater than 4 we have

(ez−1)m−1≡0(modm). (e^z-1)^{m-1} \equiv 0 \pmod m .
?
Problem 8.182

The Bernoulli numbers BnB_n are defined as the coefficients in the power series expansion

zez−1=1−z2+∑n=1∞(−1)n−1Bn(2n)!z2n \frac{z}{e^z-1} = 1-\frac z2+\sum _{n=1}^\infty \frac{(-1)^{n-1}B_n}{(2n)!}z^{2n}

(Problem 154 of Part I). We have

B1=16=1−12−13,B2=130=−1+12+13+15, B_1=\frac16 = 1-\frac12-\frac13 , \qquad B_2=\frac{1}{30} = -1+\frac12+\frac13+\frac15 , B3=142=1−12−13−17,B4=130=−1+12+13+15, B_3=\frac{1}{42} = 1-\frac12-\frac13-\frac17 , \qquad B_4=\frac{1}{30} = -1+\frac12+\frac13+\frac15 , B5=566=1−12−13−111,B6=6912730=−1+12+13+15+17+113, B_5=\frac{5}{66} = 1-\frac12-\frac13-\frac{1}{11} , \qquad B_6=\frac{691}{2730} = -1+\frac12+\frac13+\frac15+\frac17+\frac{1}{13} , B7=76=2−12−13,B8=3617510=6+12+13+15+117. B_7=\frac76 = 2-\frac12-\frac13 , \qquad B_8=\frac{3617}{510} = 6+\frac12+\frac13+\frac15+\frac{1}{17} .

In general we have

Bn=Gn+(−1)n(12+13+1α+1β+1γ+⋯ ), B_n = G_n+(-1)^n\left(\frac12+\frac13+\frac1\alpha +\frac1\beta +\frac1\gamma +\cdots \right) ,

where GnG_n is an integer and 2,3,α,β,γ,…2,3,\alpha ,\beta ,\gamma ,\ldots denote those prime numbers that exceed by one the divisors of 2n2n. [Expand first zz and then z/(ez−1)z/(e^z-1) in ascending powers of (ez−1)(e^z-1) and apply 179--181.]

?
Problem 8.183

The coefficients CnC_n in the series

zφ(z)=∑n=0∞Cn(4n)!z4n, \frac{z}{\varphi (z)} = \sum _{n=0}^\infty \frac{C_n}{(4n)!}z^{4n} ,

where φ(z)\varphi (z) denotes the function defined in 178, are rational numbers and moreover the denominator of CnC_n is not divisible by any square other than 1 and is divisible only by such prime numbers as are ≡1(mod4)\equiv 1 \pmod4. [Expand first zz and then z/φ(z)z/\varphi (z) in ascending powers of φ(z)\varphi (z).]

?
Problem 8.184

The differential equation

(d℘(z)dz)2=4[℘(z)]3−4℘(z) \left(\frac{d\wp (z)}{dz}\right)^2 = 4[\wp (z)]^3-4\wp (z)

has a well-defined solution of the form

℘(z)=1z2+∑n=1∞Dn(4n)!z4n−2. \wp (z) = \frac{1}{z^2}+\sum _{n=1}^\infty \frac{D_n}{(4n)!}z^{4n-2} .

(℘(z)\wp (z) is a particular elliptic function, termed a ``lemniscate function''; its period-parallelogram is a square.) The numbers DnD_n are rational, and moreover the denominator of DnD_n is not divisible by any square other than 1, and is divisible only by such prime numbers as are ≡1(mod4)\equiv 1 \pmod4. [Show that ℘(z)=[φ(z)]−2\wp (z)=[\varphi (z)]^{-2}, cf. 183. Expand first z2z^2 and then z2[φ(z)]−2z^2[\varphi (z)]^{-2} in ascending powers of φ(z)\varphi (z); z2z^2 as a function of φ(z)\varphi (z) satisfies a linear differential equation of second order.]

?
Problem 8.185

If a power series with H-integral coefficients ∑n=0∞(an/n!)zn\sum_{n=0}^\infty (a_n/n!)z^n satisfies a homogeneous linear differential equation with constant coefficients, then the sequence of coefficients a0,a1,a2,…,an,…a_0,a_1,a_2,\ldots ,a_n,\ldots is periodic mod mm after a certain number of terms, where mm is an arbitrary integer. [Examples of this are found already in 179, 180.]

?
Problem 8.186

The power series

y=1+2x1!+6x22!+20x33!+⋯+(2nn)xnn!+⋯ y = 1+\frac{2x}{1!}+\frac{6x^2}{2!}+\frac{20x^3}{3!}+\cdots +\binom {2n}{n}\frac{x^n}{n!}+\cdots

shows that theorem 185 cannot without further assumptions be extended to homogeneous linear differential equations whose coefficients are polynomials.

?
Problem 8.187

If a transcendental entire function g(z)g(z) has a power series expansion with H-integral coefficients, then we have for the maximum modulus (Chapter 1 of Part IV) of g(z)g(z)

lim sup⁡r→∞M(r)e−rr≥12π. \limsup _{r\to \infty } M(r)e^{-r}\sqrt r \geq \frac{1}{\sqrt{2\pi }} .
?
Problem 8.188

Let the power series in descending powers of zz

bmzm+bm−1zm−1+⋯+b1z+b0+b−1z+b−2z2+b−3z3+⋯ b_mz^m+b_{m-1}z^{m-1}+\cdots +b_1z+b_0+\frac{b_{-1}}{z}+\frac{b_{-2}}{z^2}+\frac{b_{-3}}{z^3}+\cdots

converge outside some circle. Let the numbers b1,b2,…,bmb_1,b_2,\ldots ,b_m be rational. If the series assumes integral values for infinitely many integral values of zz, then b0b_0 is rational and

b−1=b−2=b−3=⋯=0. b_{-1}=b_{-2}=b_{-3}=\cdots =0 .

(Generalization of 93.) [Hurwitz--Courant, p. 30.]

?
Problem 8.189

The series

2z2+1=2z+24z−232z3+2128z5−522048z7+⋯ \sqrt{2z^2+1} = \sqrt2z+\frac{\sqrt2}{4z}-\frac{\sqrt2}{32z^3}+\frac{\sqrt2}{128z^5}-\frac{5\sqrt2}{2048z^7}+\cdots

represents an integer for infinitely many integral values of zz.

?
Problem 8.190

Deduce 114 from 188.

?
Problem 8.191

Let the power series

F(z)=bmzm+bm−1zm−1+⋯+b1z+b0+b−1z+b−2z2+⋯ F(z) = b_mz^m+b_{m-1}z^{m-1}+\cdots +b_1z+b_0+\frac{b_{-1}}{z}+\frac{b_{-2}}{z^2}+\cdots

that is not everywhere divergent represent integers for all sufficiently large integral values of zz. Then F(z)F(z) is an integral-valued polynomial.

?
Problem 8.192

If two polynomials are such that they assume integral values at the same points of the complex plane, then either their sum or their difference is a constant. (This constant is naturally an integer.)

?
Problem 8.193

If mm is the degree of y=f(x)y=f(x) and nn the degree of z=g(x)z=g(x), then we have for sufficiently large ∣y∣\left|y\right|

x=a1y1m+a0+a−1y1m+a−2y2m+⋯ , x = a_1y^{\frac1m}+a_0+\frac{a_{-1}}{y^{\frac1m}}+\frac{a_{-2}}{y^{\frac2m}}+\cdots , z=bnynm+bn−1yn−1m+⋯+b0+b−1y1m+b−2y2m+⋯ , z = b_ny^{\frac nm}+b_{n-1}y^{\frac{n-1}{m}}+\cdots +b_0+\frac{b_{-1}}{y^{\frac1m}}+\frac{b_{-2}}{y^{\frac2m}}+\cdots ,

where all mm determinations are admissible for y1my^{\frac1m}.

?