Reference

Theorem index

Named results and key definitions, in the order Hardy proves them.

Chapter I: Real variables

§ResultStatement
§1 Example
Listing the positive rationals (Examples I, 4)
Listing p/qp/q along the diagonals p+q=2,3,…p + q = 2, 3, \dots, the fraction p/qp/q is term 12(p+q−1)(p+q−2)+q\tfrac12(p+q-1)(p+q-2) + q.
§1 Definition
Rational number
p/qp/q with integers pp, q≠0q \ne 0; in lowest terms when q>0q > 0 and pp, qq have no common factor.
§1 Theorem
Uniqueness of lowest terms
Every rational number has exactly one representation p/qp/q in lowest terms.
§3 Theorem
Density of the rationals
Every segment of the line, however short, contains infinitely many rational points.
§3 Theorem
Gauss's theorem on rational roots (Examples II, 3)
A rational root of xn+p1xn−1+⋯+pn=0x^n + p_1x^{n-1} + \dots + p_n = 0 with integer coefficients is an integer dividing pnp_n.
§3 Theorem
Rational squares
A positive fraction m/nm/n in lowest terms is the square of a rational only if mm and nn are perfect squares; in particular no rational squares to 22.
§4 Theorem
No greatest member of L
In the section of the positive rationals by x2<2x^2 < 2, x2>2x^2 > 2, the lower class has no greatest and the upper class no least member.
§6 Definition
Section of the rationals
A division of the rationals into non-empty classes LL, RR, every member of LL less than every member of RR.
§6 Theorem
The three kinds of section
In a section of the rationals either LL has a greatest member, or RR a least member, or neither; never both.
§8 Definition
Real number (Definitions, §8)
A section of the rationals in which both classes exist and the lower class has no greatest member; irrational if it corresponds to no rational.
§8 Theorem
Sign of a real number
Every real number is exactly one of positive (00 in the lower class), zero, or negative (some negative rational in the upper class).
§9 Definition
Order of real numbers
α<β\alpha < \beta when the lower class of α\alpha is a proper part of the lower class of β\beta.
§9 Theorem
Trichotomy
For real α\alpha, β\beta exactly one of α<β\alpha < \beta, α=β\alpha = \beta, α>β\alpha > \beta holds, since any two lower classes are nested.
§10 Theorem
Sum of real numbers
The classes of all sums a+ba + b and all sums A+BA + B leave at most one rational unclassified, and so define a real number α+β\alpha + \beta.
§11 Definition
Product and reciprocal of real numbers
For positive α\alpha, β\beta, αβ\alpha\beta has classes (ab)(ab), (AB)(AB), and 1/α1/\alpha has classes (1/A)(1/A), (1/a)(1/a); signs by the usual rules.
§12 Theorem
√2 squares to 2
2⋅2=2\sqrt2 \cdot \sqrt2 = 2, and x2=2x^2 = 2 has exactly the two roots ±2\pm\sqrt2.
§13 Theorem
Repeated square roots are constructible
Every number built from rationals by +,−,×,÷+, -, \times, \div and square roots is a length constructible with ruler and compasses.
§13 Theorem
Roots of a quadratic
For rational pp, qq with p2>qp^2 > q, the roots of x2+2px+q=0x^2 + 2px + q = 0 are −p±p2−q-p \pm \sqrt{p^2 - q}.
§14 Corollary
Conjugate surds (Corollary, §14)
If A+B=C+DA + \sqrt B = C + \sqrt D then A−B=C−DA - \sqrt B = C - \sqrt D, unless B\sqrt B and D\sqrt D are rational.
§14 Theorem
Dissimilar surds
If MM, NN are coprime integers, neither a perfect square, then M\sqrt M and N\sqrt N are not rational multiples of each other.
§14 Theorem
Equality of mixed surds (Theorem, §14)
If A+B=C+DA + \sqrt B = C + \sqrt D with A,B,C,DA, B, C, D rational, then A=CA = C and B=DB = D, or BB and DD are both squares of rationals.
§15 Example
The real root of z³ = 3z + 8
The equation has exactly one real root, positive and irrational, defined by the section x3<3x+8x^3 < 3x + 8, x3>3x+8x^3 > 3x + 8.
§17 Theorem
Dedekind's theorem
For every section (L,R)(L, R) of the real numbers there is a real α\alpha that is either the greatest member of LL or the least member of RR.
§18 Theorem
Infinitely many points near a point of accumulation
Every interval round a point of accumulation of SS contains infinitely many points of SS.
§18 Definition
Point of accumulation
ξ\xi is a point of accumulation of SS if every interval (ξ−δ,ξ+δ)(\xi - \delta, \xi + \delta) contains a point of SS other than ξ\xi.
§19 Theorem
Weierstrass's theorem (Theorem, §19)
An infinite set contained in a bounded interval has at least one point of accumulation in that interval.
Misc. Theorem
Arithmetic–geometric mean inequality (Misc. Examples I, 9)
For non-negative a1,…,ana_1, \dots, a_n: (a1+⋯+an)/n≥(a1⋯an)1/n(a_1 + \dots + a_n)/n \ge (a_1 \cdots a_n)^{1/n}.
Misc. Theorem
Cauchy's inequality (Misc. Examples I, 10)
(∑arbr)2≤∑ar2∑br2(\sum a_rb_r)^2 \le \sum a_r^2 \sum b_r^2, from Lagrange's identity.

Chapter II: Functions of real variables

§ResultStatement
§20 Definition
Function
yy is a function of xx if to some values of xx there correspond one or more values of yy; being defined everywhere, one-valued, or given by a formula is not required.
§20 Theorem
The vertical line test
A rule is one-valued wherever defined if and only if every vertical line meets its graph at most once; it is defined at aa exactly when x=ax = a meets the graph.
§21 Theorem
First-degree equations are straight lines
If AA, BB are not both zero, Ax+By+C=0Ax + By + C = 0 is a straight line, and every straight line has such an equation.
§21 Definition
Graph and locus
The graph of a function is the set of points (a,b)(a, b) with bb a value at aa; the locus of ϕ(x,y)=0\phi(x, y) = 0 is the set of points satisfying it.
§21 Theorem
When x² + y² + 2Gx + 2Fy + C = 0 is a circle
It is the circle with centre (−G,−F)(-G, -F) and radius G2+F2−C\sqrt{G^2 + F^2 - C} if G2+F2−C>0G^2 + F^2 - C > 0, a point if it is 00, and empty if negative.
§22 Definition
Polar coordinates
x=rcos⁡θx = r\cos\theta, y=rsin⁡θy = r\sin\theta, with r=x2+y2>0r = \sqrt{x^2 + y^2} > 0 and 0≤θ<2π0 \le \theta < 2\pi.
§22 Theorem
Polar equations of a line and a circle
A line at distance pp from OO in direction α\alpha is rcos⁡(θ−α)=pr\cos(\theta - \alpha) = p; the circle of radius AA with centre (c,α)(c, \alpha) is r2+c2−2rccos⁡(θ−α)=A2r^2 + c^2 - 2rc\cos(\theta - \alpha) = A^2.
§23 Definition
Polynomial
A function a0xm+a1xm−1+⋯+ama_0x^m + a_1x^{m-1} + \dots + a_m with constant coefficients.
§23 Theorem
The leading term of a polynomial dominates
If ∣x∣≥1|x| \ge 1, the lower terms of a0xm+⋯+ama_0x^m + \dots + a_m total at most K/(∣a0∣∣x∣)K/(|a_0||x|) times the leading term, where K=∣a1∣+⋯+∣am∣K = |a_1| + \dots + |a_m|.
§23 Theorem
The parabola y = x² is convex
For λ,μ>0\lambda, \mu > 0, λ+μ=1\lambda + \mu = 1: (λx02+μx12)−(λx0+μx1)2=λμ(x1−x0)2≥0(\lambda x_0^2 + \mu x_1^2) - (\lambda x_0 + \mu x_1)^2 = \lambda\mu(x_1 - x_0)^2 \ge 0, so every chord lies above the curve.
§24 Definition
Rational function
A quotient P(x)/Q(x)P(x)/Q(x) of polynomials; removing a common factor changes the function at that factor’s zeros.
§24 Theorem
Rational functions are closed under the rational operations
Sums, products, quotients and compositions of rational functions are rational functions (exceptional points set aside).
§25 Theorem
The polynomial part is the asymptote
If R=P/Q=S+T/QR = P/Q = S + T/Q with deg⁡T<deg⁡Q\deg T < \deg Q, then ∣R(x)−S(x)∣≤C/∣x∣|R(x) - S(x)| \le C/|x| for all large ∣x∣|x|.
§26 Definition
Explicit algebraical function
A function generated from xx by finitely many rational operations and root extractions.
§27 Definition
Algebraical function
A root yy of ym+R1(x)ym−1+⋯+Rm(x)=0y^m + R_1(x)y^{m-1} + \dots + R_m(x) = 0 with R1,…,RmR_1, \dots, R_m rational functions; it has at most mm values at each xx.
§27 Theorem
Explicit algebraical functions are algebraical
Every explicit algebraical function satisfies a polynomial equation in yy with coefficients rational in xx; the converse fails from degree 5 on.
§28 Theorem
A periodic algebraical function consists of constants (Examples XV, 15)
The values of a periodic algebraical function are the same finite set of constants at every xx; hence sin⁡x\sin x, cos⁡x\cos x are transcendental.
§28 Theorem
A periodic rational function is constant (Examples XV, 14)
If a rational function satisfies f(x+a)=f(x)f(x + a) = f(x) with a≠0a \neq 0, it is constant; so sin⁡x\sin x and cos⁡x\cos x are not rational.
§28 Corollary
The inverse circular functions are transcendental (Examples XV, 16)
arcsin⁡x\arcsin x and arccos⁡x\arccos x are infinitely many-valued on (−1,1)(-1, 1), so not algebraical.
§28 Definition
Transcendental function
A function that is not algebraical.
§30 Principle
Graphical solution of equations
The real roots of f(x)=ϕ(x)f(x) = \phi(x) are the abscissae of the intersections of y=f(x)y = f(x) and y=ϕ(x)y = \phi(x).
§30 Theorem
sin x = x only at 0 (Examples XVII, 5)
∣sin⁡x∣<∣x∣|\sin x| < |x| for x≠0x \neq 0, so sin⁡x=x\sin x = x has the single root 00.
§30 Theorem
The quadratic by a parabola and a line (Examples XVII, 1)
The roots of ax2+2bx+c=0ax^2 + 2bx + c = 0 are where y=x2y = x^2 meets y=−(2bx+c)/ay = -(2bx + c)/a: two, one or none as b2−acb^2 - ac is positive, zero or negative.
§31 Definition
Function of two variables
z=f(x,y)z = f(x, y) if values of zz correspond to pairs (x,y)(x, y); its graph is the surface of points (x,y,z)(x, y, z).
§31 Theorem
Planes and spheres
Ax+By+Cz+D=0Ax + By + Cz + D = 0 is a plane; x2+y2+z2+2Fx+2Gy+2Hz+C=0x^2 + y^2 + z^2 + 2Fx + 2Gy + 2Hz + C = 0 is a sphere of centre (−F,−G,−H)(-F, -G, -H) when F2+G2+H2−C>0F^2 + G^2 + H^2 - C > 0.
§32 Theorem
Curves through the intersections of two curves (Examples XVIII, 3)
For every λ\lambda, f+λϕ=0f + \lambda\phi = 0 passes through every common point of f=0f = 0 and ϕ=0\phi = 0.
§32 Theorem
Rational parametrisation of the ellipse (Examples XVIII, 4)
x/a=2t/(1+t2)x/a = 2t/(1 + t^2), y/b=(1−t2)/(1+t2)y/b = (1 - t^2)/(1 + t^2) gives the ellipse x2/a2+y2/b2=1x^2/a^2 + y^2/b^2 = 1 minus the point (0,−b)(0, -b), each other point once.
§33 Definition
Contour map (Examples XIX, 5)
The curves f(x,y)=cf(x, y) = c for a selection of heights cc, each the projection of the section of z=f(x,y)z = f(x, y) by the plane z=cz = c.
§33 Theorem
Cylinders, cones and surfaces of revolution (Examples XIX, 4, 7–9)
In space f(x,y)=0f(x, y) = 0 is a cylinder; rotating x=f(z)x = f(z) about OZOZ gives x2+y2=f(z)\sqrt{x^2 + y^2} = f(z); x2+y2=z2tan⁡2αx^2 + y^2 = z^2\tan^2\alpha is a right circular cone.
§33 Definition
Surfaces and curves in space
In space one equation f(x,y,z)=0f(x, y, z) = 0 represents a surface and two equations a curve; three generally represent isolated points.
§33 Example
Twisted rods form a hyperboloid (Examples XIX, 10)
Rods joining two circles of radius aa, twisted through α\alpha, form a hyperboloid of one sheet with waist radius acos⁡12αa\cos\frac12\alpha: a curved surface made of straight lines.

Chapter III: Complex numbers

§ResultStatement
§36 Definition
Addition of displacements
[x,y]+[x′,y′]=[x+x′,y+y′][x, y] + [x', y'] = [x + x', y + y']: the parallelogram rule.
§38 Definition
Product of displacements ((6))
[x,y][x′,y′]=[xx′−yy′,xy′+yx′][x, y][x', y'] = [xx' - yy', xy' + yx'], obtained from triangles similar in the same sense.
§39 Definition
Complex number
A pair (x,y)(x, y) written x+yix + yi, added and multiplied as the displacement [x,y][x, y].
§39 Theorem
Division of complex numbers
If x′+y′i≠0x' + y'i \ne 0 then (x+yi)/(x′+y′i)={(xx′+yy′)+(yx′−xy′)i}/(x′2+y′2)(x + yi)/(x' + y'i) = \{(xx' + yy') + (yx' - xy')i\}/(x'^2 + y'^2), and this quotient is unique.
§40 Theorem
Factor theorem for complex numbers
A product of two complex numbers vanishes only if one of the factors vanishes.
§42 Theorem
Multiplication by i
i2=−1i^2 = -1, and multiplying by ii turns a displacement through a positive right angle.
§43 Theorem
At most n roots
A polynomial equation of degree nn, with real or complex coefficients, has at most nn roots; with the fundamental theorem of algebra, exactly nn counted with multiplicity.
§43 Theorem
The real quadratic
az2+2bz+c=0az^2 + 2bz + c = 0 (real coefficients) has two real roots, two equal roots, or two conjugate complex roots (−b±iac−b2)/a(-b \pm i\sqrt{ac - b^2})/a, and no others.
§44 Definition
Modulus and amplitude
z=r(cos⁡θ+isin⁡θ)z = r(\cos\theta + i\sin\theta) with r=∣z∣r = |z|; the amplitude θ\theta is many-valued, its principal value lies in (−π,π](-\pi, \pi].
§45 Theorem
De Moivre's theorem
(cos⁡θ+isin⁡θ)n=cos⁡nθ+isin⁡nθ(\cos\theta + i\sin\theta)^n = \cos n\theta + i\sin n\theta for every integer nn.
§45 Theorem
Modulus and amplitude of a product ((2), (3))
∣zz′∣=∣z∣∣z′∣|zz'| = |z||z'|, and am⁡(zz′)\operatorname{am}(zz') equals am⁡z+am⁡z′\operatorname{am} z + \operatorname{am} z' or differs from it by 2π2\pi.
§45 Theorem
Triangle inequality ((6))
∣z1+⋯+zn∣≤∣z1∣+⋯+∣zn∣|z_1 + \dots + z_n| \le |z_1| + \dots + |z_n|, with equality only when all the zkz_k have the same amplitude.
§46 Theorem
Bilinear transformations map circles to circles (Examples XXI, 22)
w=(az+b)/(cz+d)w = (az + b)/(cz + d), ad−bc≠0ad - bc \ne 0, maps circles and lines to circles and lines; those through the pole −d/c-d/c go to lines.
§46 Example
Circles of Apollonius (Examples XXI, 20)
For λ≠1\lambda \ne 1, ∣z−b∣=λ∣z−a∣|z - b| = \lambda|z - a| is a circle with centre (b−λ2a)/(1−λ2)(b - \lambda^2 a)/(1 - \lambda^2), orthogonal to every circle through aa and bb.
§46 Example
Concyclic points and the cross ratio (Examples XXI, 25, 28)
Four distinct points lie on a circle or a line iff their cross ratio is real.
§46 Theorem
Conjugate roots (Theorem 3)
The non-real roots of an equation with real coefficients occur in conjugate pairs.
§46 Example
Perpendicular and parallel lines (Examples XXI, 9)
The lines abab and cdcd are perpendicular iff (a−b)/(c−d)(a - b)/(c - d) is purely imaginary, and parallel iff it is real.
§46 Theorem
Rational functions of a complex variable (Theorems 1 and 2)
A rational function R(x+yi)R(x + yi) equals X+YiX + Yi with X,YX, Y real rational functions of x,yx, y; if RR has real coefficients then R(x−yi)=X−YiR(x - yi) = X - Yi.
§46 Example
Similar triangles (Examples XXI, 11)
Triangles abcabc and xyzxyz are similar in the same sense iff (b−a)/(c−a)=(y−x)/(z−x)(b - a)/(c - a) = (y - x)/(z - x).
§47 Definition
Roots of a complex number
a1/na^{1/n} is any zz with zn=az^n = a, and am/n=(a1/n)ma^{m/n} = (a^{1/n})^m.
§48 Theorem
Roots of unity
The nn-th roots of unity are 1,ωn,…,ωnn−11, \omega_n, \dots, \omega_n^{n-1} with ωn=cos⁡2πn+isin⁡2πn\omega_n = \cos\frac{2\pi}{n} + i\sin\frac{2\pi}{n}, and they sum to 00 for n≥2n \ge 2.
§48 Theorem
Solution of z^n = a
For a=ρ(cos⁡ϕ+isin⁡ϕ)≠0a = \rho(\cos\phi + i\sin\phi) \ne 0, zn=az^n = a has exactly nn roots ρn {cos⁡ϕ+2kπn+isin⁡ϕ+2kπn}\sqrt[n]{\rho}\,\{\cos\frac{\phi + 2k\pi}{n} + i\sin\frac{\phi + 2k\pi}{n}\}, k=0,…,n−1k = 0, \dots, n - 1.
§49 Theorem
General form of De Moivre's theorem
If α\alpha is rational, one of the values of (cos⁡θ+isin⁡θ)α(\cos\theta + i\sin\theta)^\alpha is cos⁡αθ+isin⁡αθ\cos\alpha\theta + i\sin\alpha\theta.

Chapter IV: Limits of functions of a positive integral variable

§ResultStatement
§50 Theorem
Every sequence extends to infinitely many functions of x
A function of nn is the restriction to the positive integers of infinitely many functions of xx, e.g. F(x)+csin⁡πxF(x) + c\sin\pi x.
§51 Theorem
All interpolations of a function of n
If FF interpolates ϕ\phi, then GG does exactly when G−FG - F vanishes at every positive integer.
§52 Theorem
Finite sets of positive integers
A set of positive integers is finite if and only if it is empty or has a largest member.
§52 Theorem
There are infinitely many primes (Euclid; Examples XXIII, 2)
No finite list p1,…,pNp_1, \dots, p_N of primes is complete: every prime factor of p1⋯pN+1p_1 \cdots p_N + 1 is new.
§54 Definition
For large n
A property holds for large nn if it holds for every n≥n0n \ge n_0, for some n0n_0; equivalently, it fails for only finitely many nn.
§54 Lemma
Negation of “for large n”
A property fails to hold for large nn exactly when it fails for infinitely many nn.
§54 Lemma
Two properties for large n hold together
If PP holds for n≥n1n \ge n_1 and QQ for n≥n2n \ge n_2, both hold for n≥max⁡(n1,n2)n \ge \max(n_1, n_2).
§58 Definition
Limit of a sequence (Definition I)
ϕ(n)→l\phi(n) \to l if for every δ>0\delta > 0 there is n0n_0 with ∣ϕ(n)−l∣<δ|\phi(n) - l| < \delta for all n≥n0n \ge n_0.
§60 Definition
Tending to +∞+\infty (Definition II)
ϕ(n)→+∞\phi(n) \to +\infty if for every Δ\Delta there is n0n_0 with ϕ(n)>Δ\phi(n) > \Delta for all n≥n0n \ge n_0.
§61 Theorem
Finitely many terms do not matter
Altering finitely many terms of a sequence changes neither its limit nor its tending to ±∞\pm\infty.
§62 Theorem
A convergent sequence is bounded
If ϕ(n)→l\phi(n) \to l there is KK with ∣ϕ(n)∣<K|\phi(n)| < K for every nn; so every sequence converges, tends to ±∞\pm\infty, or oscillates finitely or infinitely.
§62 Definition
Oscillation, finite and infinite (Definitions, §62)
ϕ(n)\phi(n) oscillates if it tends neither to a limit nor to ±∞\pm\infty; finitely if ∣ϕ(n)∣<K|\phi(n)| < K for all nn, infinitely otherwise.
§63 Theorem
Limit of a sum (Theorem I)
If ϕ(n)→a\phi(n) \to a and ψ(n)→b\psi(n) \to b then ϕ(n)+ψ(n)→a+b\phi(n) + \psi(n) \to a + b.
§65 Theorem
Limit of a product (Theorem II)
If ϕ(n)→a\phi(n) \to a and ψ(n)→b\psi(n) \to b then ϕ(n)ψ(n)→ab\phi(n)\psi(n) \to ab.
§66 Theorem
Limit of a quotient (Theorem IV)
If ϕ(n)→a\phi(n) \to a and ψ(n)→b≠0\psi(n) \to b \neq 0 then ϕ(n)/ψ(n)→a/b\phi(n)/\psi(n) \to a/b.
§66 Theorem
Limit of a reciprocal (Theorem III)
If ϕ(n)→a≠0\phi(n) \to a \neq 0 then 1/ϕ(n)→1/a1/\phi(n) \to 1/a.
§67 Theorem
Rational functions of limits (Theorem V)
If ϕ(n)→a\phi(n) \to a, ψ(n)→b,…\psi(n) \to b, \dots and R=P/QR = P/Q with Q(a,b,… )≠0Q(a, b, \dots) \neq 0, then R(ϕ(n),ψ(n),… )→R(a,b,… )R(\phi(n), \psi(n), \dots) \to R(a, b, \dots).
§68 Theorem
Limit of a rational function of n
(a0np+… )/(b0nq+… )(a_0n^p + \dots)/(b_0n^q + \dots) tends to 00, to a0/b0a_0/b_0, or to ±∞\pm\infty (sign of a0/b0a_0/b_0) as p<qp < q, p=qp = q or p>qp > q.
§69 Theorem
Monotone sequences (Theorem, §69)
An increasing sequence tends either to a limit or to +∞+\infty; to a limit exactly when it is bounded above (Cor. 1), and then the limit is at least every term and at most any bound (Cor. 2, 3).
§71 Lemma
Nested intervals
If [pk,qk][p_k, q_k] are nested closed intervals with qk−pk→0q_k - p_k \to 0, then pkp_k and qkq_k tend to a common limit lying in every interval.
§71 Theorem
Weierstrass’s theorem (proof by bisection)
A bounded set with infinitely many points has a point of accumulation; repeated bisection gives monotone end-point sequences with a common limit.
§72 Lemma
Bernoulli’s inequality
(1+δ)n≥1+nδ(1 + \delta)^n \ge 1 + n\delta for δ>0\delta > 0 and every positive integer nn.
§72 Lemma
Shifting does not change a limit (Examples XXV, 7)
If ϕ(n)→l\phi(n) \to l then ϕ(n+p)→l\phi(n + p) \to l for any fixed integer pp.
§72 Theorem
The limit of xnx^n
xn→+∞x^n \to +\infty (x>1x > 1), →1\to 1 (x=1x = 1), →0\to 0 (∣x∣<1|x| < 1); it oscillates finitely for x=−1x = -1 and infinitely for x<−1x < -1.
§73 Theorem
The number e as the limit of (1+1/n)n(1 + 1/n)^n
(1+1/n)n(1 + 1/n)^n increases strictly and is less than 33, so it tends to a limit ee with 2<e≤32 < e \le 3.
§74 Lemma
Inequalities for powers (§74 (1)–(10))
For rational r>0r > 0 and α>0\alpha > 0, α≠1\alpha \neq 1, (αr−1)/r(\alpha^r - 1)/r increases with rr; hence rxr−1(x−y)>xr−yr>ryr−1(x−y)rx^{r-1}(x - y) > x^r - y^r > ry^{r-1}(x - y) for x>y>0x > y > 0, r>1r > 1, reversed for 0<r<10 < r < 1.
§75 Theorem
The limit of n(xn−1)n(\sqrt[n]{x} - 1)
For every x>0x > 0, n(xn−1)n(\sqrt[n]{x} - 1) tends to a limit f(x)f(x), with f(1)=0f(1) = 0, f(1/x)=−f(x)f(1/x) = -f(x), f(x)≥1−1/xf(x) \ge 1 - 1/x and f(xy)=f(x)+f(y)f(xy) = f(x) + f(y).
§76 Definition
Convergent series and its sum
∑un\sum u_n converges to ss if the partial sums u1+⋯+unu_1 + \cdots + u_n tend to ss; it diverges or oscillates as the partial sums do.
§77 Theorem
Brackets may be inserted in a convergent series (§77 (8))
Grouping the terms of a convergent series in brackets gives a convergent series with the same sum; the converse is false.
§77 Theorem
Comparison test (§77 (11))
If 0≤un≤Kvn0 \le u_n \le K v_n and ∑vn\sum v_n converges, then ∑un\sum u_n converges, with sum at most K∑vnK\sum v_n.
§77 Example
Divergence of the harmonic series
1+12+13+⋯1 + \tfrac12 + \tfrac13 + \cdots diverges to +∞+\infty: the first 2K2^K terms sum to at least 1+K/21 + K/2.
§77 Theorem
Series of positive terms (§77 (9), (10))
A series of terms ≥0\ge 0 converges or diverges to +∞+\infty, and converges if and only if its partial sums are bounded.
§77 Theorem
The terms of a convergent series tend to zero (§77 (7))
If ∑un\sum u_n converges then un→0u_n \to 0; the converse is false.
§78 Theorem
Recurring decimals are exactly the rationals (Examples XXIX)
A decimal terminates or recurs if and only if it represents a rational number; so a decimal that does neither is irrational.
§78 Theorem
The geometric series
1+r+r2+⋯1 + r + r^2 + \cdots converges to 1/(1−r)1/(1-r) iff −1<r<1-1 < r < 1; diverges to +∞+\infty if r≥1r \ge 1; oscillates finitely if r=−1r = -1 and infinitely if r<−1r < -1.
§79 Definition
Function defined as a limit
ϕ(x)=lim⁡n→∞ϕn(x)\phi(x) = \lim_{n\to\infty}\phi_n(x) at each xx where the limit exists; e.g. lim⁡xn/(xn+1)\lim x^n/(x^n+1) is 00, 12\tfrac12, 11 for ∣x∣<1|x| < 1, x=1x = 1, ∣x∣>1|x| > 1.
§80 Theorem
Existence of the upper bound (supremum)
Every non-empty set of real numbers bounded above has a least upper bound MM: no member exceeds MM, and every number less than MM is exceeded by some member.
§81 Theorem
An increasing bounded sequence tends to its upper bound
If ϕ(n)\phi(n) increases and is bounded above, then ϕ(n)→sup⁡nϕ(n)\phi(n) \to \sup_n \phi(n).
§81 Definition
Bounds of a sequence
The upper bound MM of ϕ(n)\phi(n) satisfies ϕ(n)≤M\phi(n) \le M for all nn, and ϕ(n)>M−δ\phi(n) > M - \delta for some nn, for every δ>0\delta > 0; similarly the lower bound mm.
§82 Theorem
Convergence and the upper and lower limits (§82 (5))
A bounded sequence tends to a limit if and only if λ=Λ\lambda = \Lambda, and the limit is then their common value.
§82 Definition
Upper and lower limits
Λ=lim⁡‾ ϕ(n)\Lambda = \overline{\lim}\,\phi(n) is the lower bound of the superior numbers and λ=lim⁡‾ ϕ(n)\lambda = \underline{\lim}\,\phi(n) the upper bound of the inferior numbers; m≤λ≤Λ≤Mm \le \lambda \le \Lambda \le M.
§83 Principle
General principle of convergence (Theorem 1)
ϕ(n)\phi(n) tends to a limit iff for every δ>0\delta > 0 there is n0n_0 with ∣ϕ(n2)−ϕ(n1)∣<δ|\phi(n_2) - \phi(n_1)| < \delta for all n2>n1≥n0n_2 > n_1 \ge n_0.
§84 Principle
Cauchy’s criterion for series (Theorem 2)
∑un\sum u_n converges iff for every δ>0\delta > 0 there is n0n_0 with ∣un1+1+⋯+un2∣<δ|u_{n_1+1} + \cdots + u_{n_2}| < \delta for all n2>n1≥n0n_2 > n_1 \ge n_0.
§85 Definition
Limit of a complex sequence
ρ(n)+iσ(n)→r+is\rho(n) + i\sigma(n) \to r + is means ρ(n)→r\rho(n) \to r and σ(n)→s\sigma(n) \to s; a complex series converges when its real and imaginary parts do.
§86 Theorem
Complex convergence through the modulus (§86 (10), (11))
ϕ(n)→l\phi(n) \to l if and only if ∣ϕ(n)−l∣→0|\phi(n) - l| \to 0.
§87 Theorem
The limit of znz^n
zn→0z^n \to 0 if ∣z∣<1|z| < 1, zn→1z^n \to 1 if z=1z = 1, and znz^n has no limit for any other complex zz.
§88 Corollary
Sums of rncos⁡nθr^n\cos n\theta and rnsin⁡nθr^n\sin n\theta
For ∣r∣<1|r| < 1: ∑n≥0rncos⁡nθ=1−rcos⁡θ1−2rcos⁡θ+r2\sum_{n\ge0} r^n\cos n\theta = \frac{1 - r\cos\theta}{1 - 2r\cos\theta + r^2} and ∑n≥1rnsin⁡nθ=rsin⁡θ1−2rcos⁡θ+r2\sum_{n\ge1} r^n\sin n\theta = \frac{r\sin\theta}{1 - 2r\cos\theta + r^2}.
§88 Theorem
The complex geometric series
1+z+z2+⋯1 + z + z^2 + \cdots converges if and only if ∣z∣<1|z| < 1, and its sum is then 1/(1−z)1/(1-z).
§89 Theorem
Rules for O and o (§89 (1)–(4))
O(ϕ)+O(ψ)=O(ϕ+ψ)O(\phi) + O(\psi) = O(\phi + \psi), O(ϕ)O(ψ)=O(ϕψ)O(\phi)O(\psi) = O(\phi\psi), O(ϕ)o(ψ)=o(ϕψ)O(\phi)o(\psi) = o(\phi\psi), and f∼ϕf \sim \phi implies f+o(ϕ)∼ϕf + o(\phi) \sim \phi.
§89 Definition
The symbols O, o and ~
f=O(ϕ)f = O(\phi) if ∣f∣≤Kϕ|f| \le K\phi for large nn; f=o(ϕ)f = o(\phi) if f/ϕ→0f/\phi \to 0; f∼lϕf \sim l\phi if f/ϕ→l≠0f/\phi \to l \ne 0.

Chapter V: Limits of functions of a continuous variable. Continuous and discontinuous functions

§ResultStatement
§90 Definition
Limit as x→∞x \to \infty (Definition 1)
ϕ(x)→l\phi(x) \to l if for every δ>0\delta > 0 there is x0x_0 with ∣ϕ(x)−l∣<δ|\phi(x) - l| < \delta for all real x≥x0x \ge x_0.
§90 Definition
Oscillation as x→∞x \to \infty (Definition 3)
If ϕ(x)\phi(x) tends neither to a limit nor to ±∞\pm\infty it oscillates, finitely if ∣ϕ(x)∣<K|\phi(x)| < K for x≥x0x \ge x_0, infinitely otherwise.
§90 Definition
Tending to ∞\infty as x→∞x \to \infty (Definition 2)
ϕ(x)→∞\phi(x) \to \infty if for every Δ\Delta there is x0x_0 with ϕ(x)>Δ\phi(x) > \Delta for all x≥x0x \ge x_0.
§92 Theorem
Monotone functions have limits at infinity
A function increasing steadily for large xx tends to its upper bound if bounded above, and to ∞\infty otherwise.
§94 Definition
Limit as x→ax \to a
ϕ(x)→l\phi(x) \to l as x→ax \to a if for every ε>0\varepsilon > 0 there is δ>0\delta > 0 with ∣ϕ(x)−l∣<ε|\phi(x) - l| < \varepsilon whenever 0<∣x−a∣<δ0 < |x - a| < \delta.
§94 Definition
One-sided limits
ϕ(a+0)\phi(a + 0) and ϕ(a−0)\phi(a - 0) are the limits as x→ax \to a through x>ax > a and x<ax < a; the two-sided limit exists exactly when both exist and agree.
§95 Theorem
Monotone functions have one-sided limits
If ϕ\phi increases steadily near aa, then ϕ(a−0)\phi(a - 0) and ϕ(a+0)\phi(a + 0) exist and ϕ(a−0)≤ϕ(a)≤ϕ(a+0)\phi(a - 0) \le \phi(a) \le \phi(a + 0).
§96 Theorem
Algebra of limits for x→ax \to a (Examples XXXV, 1)
If ϕ→l\phi \to l and ψ→l′\psi \to l' as x→ax \to a then ϕ+ψ→l+l′\phi + \psi \to l + l', ϕψ→ll′\phi\psi \to ll', and ϕ/ψ→l/l′\phi/\psi \to l/l' if l′≠0l' \ne 0.
§96 Principle
Cauchy’s principle for x→ax \to a
ϕ(x)\phi(x) tends to a limit as x→ax \to a iff for every ε\varepsilon there is δ\delta with ∣ϕ(x2)−ϕ(x1)∣<ε|\phi(x_2) - \phi(x_1)| < \varepsilon whenever 0<∣x1−a∣,∣x2−a∣<δ0 < |x_1 - a|, |x_2 - a| < \delta.
§97 Theorem
(sin⁡x)/x→1(\sin x)/x \to 1 (Examples XXXVI, 13)
From sin⁡x<x<tan⁡x\sin x < x < \tan x for 0<x<12π0 < x < \tfrac12\pi: cos⁡x<(sin⁡x)/x<1\cos x < (\sin x)/x < 1, so (sin⁡x)/x→1(\sin x)/x \to 1 as x→0x \to 0.
§98 Definition
Orders of smallness
ϕ\phi is of the mmth order of smallness as x→0x \to 0 if ϕ(x)/xm\phi(x)/x^m tends to a limit l≠0l \ne 0; OO, oo and ∼\sim are defined as for sequences.
§99 Definition
Continuity at a point
ϕ\phi is continuous at ξ\xi if ϕ(ξ−0)=ϕ(ξ)=ϕ(ξ+0)\phi(\xi - 0) = \phi(\xi) = \phi(\xi + 0), i.e. for every ε\varepsilon there is δ\delta with ∣ϕ(x)−ϕ(ξ)∣<ε|\phi(x) - \phi(\xi)| < \varepsilon for ∣x−ξ∣<δ|x - \xi| < \delta.
§100 Definition
Classification of discontinuities (Examples XXXVII, 18)
A discontinuity is simple (both one-sided limits exist), an infinity (tends to ±∞\pm\infty on at least one side), or oscillatory (anything else).
§100 Theorem
Continuity of sums, products, quotients and compositions (Examples XXXVII, 1, 2, 11)
Sums, products, quotients (nonzero denominator) and compositions of continuous functions are continuous; so polynomials and rational functions are continuous where defined.
§101 Theorem
Intermediate value theorem
If ϕ\phi is continuous on [x0,x1][x_0, x_1] and η\eta lies between ϕ(x0)\phi(x_0) and ϕ(x1)\phi(x_1), then ϕ(x)=η\phi(x) = \eta for some xx between them (and there is a least such xx).
§101 Theorem
Monotone functions with the intermediate value property are continuous
A steadily increasing (or decreasing) function that takes every value between its values at the ends of an interval is continuous there.
§102 Theorem
Persistence of sign (Theorem 1)
If ϕ\phi is continuous at ξ\xi and ϕ(ξ)>0\phi(\xi) > 0, then ϕ>0\phi > 0 throughout some interval (ξ−δ,ξ+δ)(\xi - \delta, \xi + \delta).
§102 Corollary
Values of both signs nearby force a zero (Theorem 2)
If ϕ\phi is continuous at ξ\xi and vanishes, or takes both signs, at points arbitrarily near ξ\xi, then ϕ(ξ)=0\phi(\xi) = 0.
§103 Theorem
Continuous functions attain their bounds (Theorem 2)
A function continuous throughout [a,b][a, b] takes its upper bound MM and lower bound mm as values.
§103 Theorem
Continuous functions on a closed interval are bounded (Theorem 1)
A function continuous throughout [a,b][a, b] is bounded on [a,b][a, b].
§104 Definition
Oscillation in an interval
O(a,b)=M(a,b)−m(a,b)O(a, b) = M(a, b) - m(a, b); it increases with the interval and O(a,b)≤O(a,c)+O(c,b)O(a, b) \le O(a, c) + O(c, b) for a≤c≤ba \le c \le b.
§106 Theorem
Heine–Borel theorem
If every point of [a,b][a, b] lies inside an interval of a set II (the end points at an end), then finitely many intervals of II have the same property.
§107 Theorem
Finitely many pieces of small oscillation (Theorem I)
If ϕ\phi is continuous on [a,b][a, b], then [a,b][a, b] can be divided into finitely many sub-intervals in each of which the oscillation of ϕ\phi is less than a given ε\varepsilon.
§107 Theorem
Uniform continuity (Theorem II)
If ϕ\phi is continuous on [a,b][a, b], for each ε>0\varepsilon > 0 there is δ>0\delta > 0 such that every sub-interval of length less than δ\delta has oscillation less than ε\varepsilon.
§108 Definition
Continuity of a function of two variables
ϕ(x,y)\phi(x, y) is continuous at (ξ,η)(\xi, \eta) if for every ε\varepsilon some square ∣x−ξ∣,∣y−η∣≤δ|x - \xi|, |y - \eta| \le \delta has ∣ϕ(x,y)−ϕ(ξ,η)∣<ε|\phi(x, y) - \phi(\xi, \eta)| < \varepsilon throughout.
§108 Example
Separate continuity does not imply continuity
2xy/(x2+y2)2xy/(x^2 + y^2) (with value 00 at the origin) is continuous in xx and in yy separately but not in (x,y)(x, y) at the origin.
§109 Theorem
Implicit function theorem
If ff is continuous near (a,b)(a, b), f(a,b)=0f(a, b) = 0 and ff is strictly monotone in yy, then f(x,y)=0f(x, y) = 0 defines a unique continuous y=ϕ(x)y = \phi(x) near aa with ϕ(a)=b\phi(a) = b.
§110 Theorem
Inverse function theorem
If FF is continuous and strictly increasing on [a,b][a, b], it has a unique inverse on [F(a),F(b)][F(a), F(b)], which is continuous and strictly increasing.

Chapter VI: Derivatives and integrals

§ResultStatement
§112 Theorem
A differentiable function is continuous
If ϕ′(x)\phi'(x) exists then ϕ\phi is continuous at xx; the converse is false (∣x∣|x|, xsin⁡(1/x)x\sin(1/x), x3\sqrt[3]{x} at 00).
§112 Definition
Derivative
ϕ′(x)=lim⁡h→0{ϕ(x+h)−ϕ(x)}/h\phi'(x) = \lim_{h \to 0} \{\phi(x + h) - \phi(x)\}/h, the limit being finite and the same from both sides.
§114 Theorem
Chain rule (§114 (6))
If ff is differentiable at xx and FF at f(x)f(x), then {F(f(x))}′=F′{f(x)}f′(x)\{F(f(x))\}' = F'\{f(x)\}f'(x).
§114 Theorem
Derivative of an inverse function (§114 (7))
If ψ\psi is continuous and strictly monotonic with inverse ϕ\phi, and ψ′(y)≠0\psi'(y) \ne 0, then ϕ′(x)=1/ψ′(y)\phi'(x) = 1/\psi'(y) where x=ψ(y)x = \psi(y).
§114 Theorem
Sum, multiple, product, reciprocal and quotient rules (§114 (1)–(5))
(f+F)′=f′+F′(f + F)' = f' + F', (kf)′=kf′(kf)' = kf', (fF)′=f′F+fF′(fF)' = f'F + fF', (1/f)′=−f′/f2(1/f)' = -f'/f^2, (f/F)′=(f′F−fF′)/F2(f/F)' = (f'F - fF')/F^2.
§115 Definition
Derivative of a complex function of a real variable
(ϕ+iψ)′=ϕ′+iψ′(\phi + i\psi)' = \phi' + i\psi'; rules (1)–(5) of §114 remain valid.
§117 Theorem
Repeated roots and the derivative (Examples XLI, 2–3)
A root of multiplicity mm of a polynomial ϕ\phi is a root of multiplicity m−1m - 1 of ϕ′\phi'; so α\alpha is a repeated root exactly when ϕ(α)=ϕ′(α)=0\phi(\alpha) = \phi'(\alpha) = 0.
§117 Theorem
Rolle's theorem for polynomials (Examples XLI, 10)
Between two real roots of a polynomial lies a root of its derivative (proved algebraically).
§118 Theorem
Derivative of a rational function
By partial fractions, {A/(x−α)p}′=−pA/(x−α)p+1\{A/(x - \alpha)^p\}' = -pA/(x - \alpha)^{p+1}; hence (xm)′=mxm−1(x^m)' = mx^{m-1} for every integer mm, and 1/x1/x is not the derivative of a rational function.
§119 Theorem
Derivative of a rational power
(xm)′=mxm−1(x^m)' = mx^{m-1} for every rational mm (with x≠0x \ne 0).
§120 Theorem
Derivatives of the trigonometrical functions and their inverses
sin⁡′=cos⁡\sin' = \cos, cos⁡′=−sin⁡\cos' = -\sin, tan⁡′=sec⁡2\tan' = \sec^2; (arctan⁡x)′=1/(1+x2)(\arctan x)' = 1/(1 + x^2), (arcsin⁡x)′=±1/1−x2(\arcsin x)' = \pm 1/\sqrt{1 - x^2} (++ for the principal value).
§121 Theorem
Leibniz's theorem
(uv)(n)=∑r=0n(nr)u(n−r)v(r)(uv)^{(n)} = \sum_{r=0}^{n} \binom{n}{r} u^{(n-r)} v^{(r)}.
§122 Theorem
A positive derivative at a point (Theorem A) (Theorem A)
If ϕ′(x0)>0\phi'(x_0) > 0 then ϕ(x)<ϕ(x0)\phi(x) < \phi(x_0) for xx just below x0x_0 and ϕ(x)>ϕ(x0)\phi(x) > \phi(x_0) for xx just above it.
§122 Corollary
A positive derivative on an interval gives an increasing function (Theorem B, Cor. 1)
If ϕ\phi is continuous on [a,b][a, b] and ϕ′(x)>0\phi'(x) > 0 in (a,b)(a, b), then ϕ\phi is strictly increasing on [a,b][a, b].
§122 Corollary
Positivity from a positive derivative (Theorem B, Cor. 2)
If ϕ′(x)>0\phi'(x) > 0 throughout (a,b)(a, b) and ϕ(a)≥0\phi(a) \ge 0 (with ϕ\phi continuous at aa), then ϕ(x)>0\phi(x) > 0 for a<x≤ba < x \le b.
§122 Theorem
Rolle's theorem (Theorem B)
If ϕ\phi is continuous on [a,b][a, b], differentiable in (a,b)(a, b), and ϕ(a)=ϕ(b)\phi(a) = \phi(b), then ϕ′(ξ)=0\phi'(\xi) = 0 for some ξ\xi in (a,b)(a, b).
§123 Theorem
First derivative test
If ϕ′\phi' changes sign from ++ to −- at ξ\xi (with ϕ′(ξ)=0\phi'(\xi) = 0), then ϕ\phi has a maximum at ξ\xi; from −- to ++, a minimum.
§123 Definition
Maximum and minimum
ϕ(ξ)\phi(\xi) is a maximum if ϕ(ξ)>ϕ(x)\phi(\xi) > \phi(x) for all x≠ξx \ne \xi in some interval (ξ−ε,ξ+ε)(\xi - \varepsilon, \xi + \varepsilon); similarly for a minimum.
§123 Theorem
Necessary condition for a maximum or minimum (Theorem C) (Theorem C)
If a differentiable ϕ\phi has a maximum or minimum at ξ\xi, then ϕ′(ξ)=0\phi'(\xi) = 0; the condition is not sufficient (x3x^3 at 00).
§124 Theorem
Second derivative test (Theorem D) (Theorem D)
If ϕ′(ξ)=0\phi'(\xi) = 0 and ϕ′′(ξ)<0\phi''(\xi) < 0 then ξ\xi gives a maximum; if ϕ′′(ξ)>0\phi''(\xi) > 0, a minimum.
§125 Example
A derivative that exists everywhere but is discontinuous
x2sin⁡(1/x)x^2\sin(1/x) (with value 00 at 00) has a derivative at every point, but ϕ′\phi' oscillates between −1-1 and 11 as x→0x \to 0.
§126 Theorem
Mean value theorem
If ϕ\phi is continuous on [a,b][a, b] and differentiable in (a,b)(a, b), then ϕ(b)−ϕ(a)=(b−a)ϕ′(ξ)\phi(b) - \phi(a) = (b - a)\phi'(\xi) for some ξ\xi with a<ξ<ba < \xi < b.
§127 Theorem
Zero derivative means constant
If ϕ′(x)=0\phi'(x) = 0 throughout an interval, then ϕ\phi is constant there; two functions with equal derivatives on an interval differ by a constant.
§128 Theorem
Cauchy's mean value theorem
If ϕ,ψ\phi, \psi are continuous on [a,b][a, b] and differentiable in (a,b)(a, b), ψ(b)≠ψ(a)\psi(b) \ne \psi(a), and ϕ′\phi', ψ′\psi' never vanish together, then {ϕ(b)−ϕ(a)}/{ψ(b)−ψ(a)}=ϕ′(ξ)/ψ′(ξ)\{\phi(b) - \phi(a)\}/\{\psi(b) - \psi(a)\} = \phi'(\xi)/\psi'(\xi) for some ξ\xi in (a,b)(a, b).
§129 Theorem
Darboux's theorem
If ϕ\phi is differentiable on [a,b][a, b] and γ\gamma lies between ϕ′(a)\phi'(a) and ϕ′(b)\phi'(b), then ϕ′(ξ)=γ\phi'(\xi) = \gamma for some ξ\xi in (a,b)(a, b): a derivative has the intermediate value property, continuous or not.
§130 Definition
Integral (integral function) (§130)
If ϕ′(x)=ψ(x)\phi'(x) = \psi(x), then ϕ\phi is an integral of ψ\psi, written ϕ(x)=∫ψ(x) dx\phi(x) = \int\psi(x)\,dx.
§130 Theorem
Integrals differ by constants (§130 (2))
On an interval, if ϕ\phi is one integral of ψ\psi then the integrals of ψ\psi are exactly the functions ϕ+C\phi + C.
§131 Theorem
The integral of 1/x (§131 (4), (5))
∫dx/x=log⁡∣x∣\int dx/x = \log|x| for x≠0x \ne 0, where log⁡x\log x (for x>0x > 0) is the integral of 1/x1/x vanishing at 11; 1/x1/x is not the derivative of any rational function.
§132 Theorem
Linearity of integration (§132 (1), (2))
∫{f+F}=∫f+∫F\int\{f + F\} = \int f + \int F and ∫kf=k∫f\int kf = k\int f, the constants being adjusted properly; hence every polynomial can be integrated term by term.
§133 Corollary
Integral of ψ(ax + b) (§133 (4))
If ϕ\phi is an integral of ψ\psi and a≠0a \ne 0, then ∫ψ(ax+b) dx=ϕ(ax+b)/a\int\psi(ax+b)\,dx = \phi(ax+b)/a.
§133 Theorem
Integration of rational functions (§133)
If the denominator of a real rational function can be factored, its integral is a polynomial plus rational functions plus logarithms plus inverse tangents, found by partial fractions.
§133 Theorem
The chain rule backwards (§133 (3), (5))
∫F′{f(x)}f′(x) dx=F{f(x)}\int F'\{f(x)\}f'(x)\,dx = F\{f(x)\}; in particular ∫f′(x)/f(x) dx=log⁡∣f(x)∣\int f'(x)/f(x)\,dx = \log|f(x)|.
§136 Theorem
Integration by substitution (§136 (1))
If ∫ψ(x) dx=ϕ(x)\int\psi(x)\,dx = \phi(x) and x=f(t)x = f(t), then ∫ψ{f(t)}f′(t) dt=ϕ{f(t)}\int\psi\{f(t)\}f'(t)\,dt = \phi\{f(t)\}.
§136 Theorem
Rationalisation (§136)
If x=R1(t)x = R_1(t) and y=R2(t)y = R_2(t) are rational in tt, then ∫R(x,y) dx=∫R{R1(t),R2(t)}R1′(t) dt\int R(x, y)\,dx = \int R\{R_1(t), R_2(t)\}R_1'(t)\,dt is the integral of a rational function.
§137 Theorem
Rational parametrisation of a conic (§137)
A line of slope tt through a point of a conic meets it in one more point whose coordinates are rational in tt; so ∫R(x,ax2+2bx+c) dx\int R(x, \sqrt{ax^2+2bx+c})\,dx can always be rationalised.
§138 Theorem
Integrals of 1/√(x² ± α²) and 1/√(α² − x²) (§138)
∫dx/x2±α2=log⁡∣x+x2±α2∣\int dx/\sqrt{x^2\pm\alpha^2} = \log|x + \sqrt{x^2\pm\alpha^2}| and ∫dx/α2−x2=arcsin⁡(x/α)\int dx/\sqrt{\alpha^2 - x^2} = \arcsin(x/\alpha) for α>0\alpha > 0.
§139 Theorem
Linear numerator over √X (§139)
With X=ax2+2bx+cX = ax^2 + 2bx + c: ∫(λx+μ)/X dx=(λ/a)X+(μ−λb/a)∫dx/X\int(\lambda x + \mu)/\sqrt X\,dx = (\lambda/a)\sqrt X + (\mu - \lambda b/a)\int dx/\sqrt X.
§141 Theorem
Integral of √(ax² + 2bx + c) (§141)
With y=ax2+2bx+cy = \sqrt{ax^2+2bx+c}: ∫y dx=(ax+b)y/2a+{(ac−b2)/2a}∫dx/y\int y\,dx = (ax+b)y/2a + \{(ac - b^2)/2a\}\int dx/y.
§141 Theorem
Integration by parts (§141)
∫f′(x)F(x) dx=f(x)F(x)−∫f(x)F′(x) dx\int f'(x)F(x)\,dx = f(x)F(x) - \int f(x)F'(x)\,dx.
§142 Theorem
The general integral ∫R(x, √X) dx (§142)
Any such integral reduces to integrals of xm/Xx^m/\sqrt X (by a reduction formula), of 1/{(x−p)mX}1/\{(x-p)^m\sqrt X\} (by x−p=1/tx - p = 1/t), and of quadratic-denominator terms (by Hardy's substitution), all elementary.
§144 Theorem
Products of sines and cosines (§144)
Any polynomial in cos⁡ax,sin⁡ax,cos⁡bx,sin⁡bx,…\cos ax, \sin ax, \cos bx, \sin bx, \dots is a finite sum of terms αcos⁡kx\alpha\cos kx, βsin⁡kx\beta\sin kx, and so can be integrated.
§146 Theorem
The half-angle substitution t = tan ½x (§146)
With t=tan⁡12xt = \tan\frac12x: cos⁡x=(1−t2)/(1+t2)\cos x = (1-t^2)/(1+t^2), sin⁡x=2t/(1+t2)\sin x = 2t/(1+t^2), dx/dt=2/(1+t2)dx/dt = 2/(1+t^2), so every rational function of cos⁡x\cos x and sin⁡x\sin x can be integrated.
§147 Theorem
Integral of an inverse function (§147)
If ϕ\phi is the inverse of ff and y=f(x)y = f(x), then ∫ϕ(y) dy=xf(x)−∫f(x) dx\int\phi(y)\,dy = xf(x) - \int f(x)\,dx; e.g. ∫log⁡x dx=x(log⁡x−1)\int\log x\,dx = x(\log x - 1).
§148 Theorem
Area of a loop (Examples LIV, 9, 18)
A loop traced once anticlockwise by (x(t),y(t))(x(t), y(t)) has area equal to the increase of 12∫(xy˙−yx˙) dt\frac12\int(x\dot y - y\dot x)\,dt; in polar coordinates, of 12∫r2 dθ\frac12\int r^2\,d\theta.
§148 Theorem
The ordinate is the derivative of the area (§148)
If ϕ\phi is continuous and positive and Φ(x)\Phi(x) is the area under y=ϕ(x)y = \phi(x) from 00 to xx, then Φ′(x)=ϕ(x)\Phi'(x) = \phi(x) (assuming areas exist).
§149 Theorem
Length of a plane curve (§149)
If arcs have lengths and arc/chord →1\to 1, the length S(x)S(x) of y=ϕ(x)y = \phi(x) satisfies S′(x)=1+ϕ′(x)2S'(x) = \sqrt{1 + \phi'(x)^2}.

Chapter VII: Additional theorems in the differential and integral calculus

§ResultStatement
§150 Theorem
Mean value theorem of the second order (§150 (2))
If f′f' is continuous on [a,a+h][a, a+h] and f′′f'' exists inside, then f(a+h)=f(a)+hf′(a)+12h2f′′(a+θ2h)f(a+h) = f(a) + hf'(a) + \tfrac12 h^2 f''(a + \theta_2 h), 0<θ2<10 < \theta_2 < 1.
§150 Theorem
Taylor's theorem with Lagrange's remainder (§150, general mean value theorem)
If f(n−1)f^{(n-1)} is continuous on [a,a+h][a, a+h] and f(n)f^{(n)} exists inside, then f(a+h)=∑k<nhkk!f(k)(a)+hnn!f(n)(a+θnh)f(a+h) = \sum_{k<n} \frac{h^k}{k!}f^{(k)}(a) + \frac{h^n}{n!}f^{(n)}(a + \theta_n h), 0<θn<10 < \theta_n < 1.
§151 Theorem
Taylor's theorem with remainder o(hⁿ) (§151 (1), (2))
If ff has an nn-th derivative at aa, then f(a+h)=∑k≤nhkk!f(k)(a)+o(hn)f(a+h) = \sum_{k \le n} \frac{h^k}{k!}f^{(k)}(a) + o(h^n) as h→0h \to 0.
§151 Theorem
Uniqueness of expansions (Examples LVI, 1)
If ∑k≤nakxk+o(xn)=∑k≤nbkxk+o(xn)\sum_{k\le n} a_k x^k + o(x^n) = \sum_{k \le n} b_k x^k + o(x^n) as x→0x \to 0, then ak=bka_k = b_k for all kk.
§152 Theorem
Binomial series for 0 ≤ x < 1 (§152 (2))
(1+x)m=∑(mn)xn(1 + x)^m = \sum \binom{m}{n} x^n for rational mm and 0≤x<10 \le x < 1, by Lagrange's remainder; the case −1<x<0-1 < x < 0 needs Cauchy's form (§168).
§152 Theorem
Sine and cosine series (§152 (1))
sin⁡x=x−x3/3!+⋯\sin x = x - x^3/3! + \cdots and cos⁡x=1−x2/2!+⋯\cos x = 1 - x^2/2! + \cdots for all real xx, since ∣Rn∣≤∣x∣n/n!|R_n| \le |x|^n/n!.
§152 Definition
Taylor's series (§152)
The series ∑hnn!f(n)(a)\sum \frac{h^n}{n!} f^{(n)}(a) (Maclaurin's series when a=0a = 0); it converges to f(a+h)f(a+h) if and only if the remainder Rn→0R_n \to 0.
§153 Theorem
Test of the first non-vanishing derivative (§153)
If ϕ(k)(ξ)=0\phi^{(k)}(\xi) = 0 for 1≤k<n1 \le k < n and ϕ(n)(ξ)≠0\phi^{(n)}(\xi) \ne 0: for even nn, a maximum if ϕ(n)(ξ)<0\phi^{(n)}(\xi) < 0 and a minimum if >0> 0; for odd nn, neither.
§154 Theorem
L'Hôpital's rule (§154 (b), (3))
If f(0)=ϕ(0)=0f(0) = \phi(0) = 0 and f′(x)/ϕ′(x)→lf'(x)/\phi'(x) \to l as x→+0x \to +0, then f(x)/ϕ(x)→lf(x)/\phi(x) \to l; the converse is false.
§154 Theorem
Limits 0/0 by first non-vanishing derivatives (§154 (a))
If f,ϕf, \phi and their first n−1n-1 derivatives vanish at 00 and ϕ(n)(0)≠0\phi^{(n)}(0) \ne 0, then f(x)/ϕ(x)→f(n)(0)/ϕ(n)(0)f(x)/\phi(x) \to f^{(n)}(0)/\phi^{(n)}(0).
§155 Definition
Circle of curvature (Examples LIX, 8)
The circle with second-order contact at (ξ,η)(\xi, \eta): centre (ξ−η1(1+η12)/η2, η+(1+η12)/η2)\bigl(\xi - \eta_1(1+\eta_1^2)/\eta_2,\ \eta + (1+\eta_1^2)/\eta_2\bigr), radius (1+η12)3/2/∣η2∣(1+\eta_1^2)^{3/2}/|\eta_2|.
§155 Definition
Contact of the nth order (§155)
Curves y=f(x)y = f(x), y=ϕ(x)y = \phi(x) have contact of order nn at ξ\xi if f(k)(ξ)=ϕ(k)(ξ)f^{(k)}(\xi) = \phi^{(k)}(\xi) for k≤nk \le n but not for k=n+1k = n + 1; then {ϕ(ξ+h)−f(ξ+h)}/hn+1\{\phi(\xi + h) - f(\xi + h)\}/h^{n+1} tends to a non-zero limit.
§156 Definition
Partial derivatives (§156)
fx(x,y)=lim⁡h→0{f(x+h,y)−f(x,y)}/hf_x(x, y) = \lim_{h\to0}\{f(x+h, y) - f(x, y)\}/h and similarly fyf_y: ordinary derivatives with the other variable held fixed.
§157 Theorem
Chain rule for f(φ(t), ψ(t)) (§157, theorem of the total differential coefficient)
If fx,fyf_x, f_y are continuous and ϕ′,ψ′\phi', \psi' continuous, then ddtf{ϕ(t),ψ(t)}=fxϕ′(t)+fyψ′(t)\frac{d}{dt}f\{\phi(t), \psi(t)\} = f_x\phi'(t) + f_y\psi'(t).
§158 Example
Partial derivatives without continuity (§158)
2xy/(x2+y2)2xy/(x^2+y^2) has both partial derivatives everywhere but is discontinuous at 00; 2xy(x+y)/(x2+y2)2xy(x+y)/(x^2+y^2) is continuous but violates the chain rule at 00.
§159 Corollary
Approximate linearity of the increment (§159 (1))
δz=(fx+ε)δx+(fy+η)δy\delta z = (f_x + \varepsilon)\delta x + (f_y + \eta)\delta y with ε,η→0\varepsilon, \eta \to 0; and if δz=λδx+μδy\delta z = \lambda\delta x + \mu\delta y is approximately true then λ=fx\lambda = f_x, μ=fy\mu = f_y.
§159 Definition
Differentiable function of two variables (§159)
f(x+h,y+k)−f(x,y)=Ah+Bk+εh+ηkf(x+h, y+k) - f(x, y) = Ah + Bk + \varepsilon h + \eta k with ε,η→0\varepsilon, \eta \to 0; intermediate between existence and continuity of fx,fyf_x, f_y.
§159 Theorem
Mean value theorem for two variables (§159)
If fx,fyf_x, f_y are continuous, f(x+h,y+k)−f(x,y)=hfx(x+θh,y+θk)+kfy(x+θh,y+θk)f(x+h, y+k) - f(x, y) = hf_x(x+\theta h, y+\theta k) + kf_y(x+\theta h, y+\theta k) for some 0<θ<10 < \theta < 1.
§160 Theorem
Total differential and its invariance (§160)
dz=zx dx+zy dydz = z_x\,dx + z_y\,dy, and this formula holds whether x,yx, y are independent or functions of another variable.
§161 Theorem
Existence of the area (the definite integral) (§161)
If ff is continuous on [a,b][a, b] then j=Jj = J, and ss, SS and every Riemann sum ∑fvδv\sum f_v\delta_v tend to this common value as the longest piece tends to 00.
§161 Theorem
No lower sum exceeds any upper sum (§161)
For any two partitions, s≤S′s \le S'; hence j=sup⁡s≤J=inf⁡Sj = \sup s \le J = \inf S.
§161 Lemma
Refinement of a partition (§161 (1))
Adding points to a partition does not decrease the lower sum ss and does not increase the upper sum SS.
§162 Definition
Definite integral (§§161–162)
For continuous ff, ∫abf(x) dx\int_a^b f(x)\,dx is the common limit JJ of the lower and upper sums, the area under the graph; it equals F(b)−F(a)F(b) - F(a) for any FF with F′=fF' = f.
§163 Definition
Measure of an angle (§163)
The measure of the angle AOPAOP is twice the area of the sector AOPAOP of the unit circle with centre OO.
§163 Theorem
Sector area and the inverse tangent (§163)
The sector of the unit circle between the xx-axis and the ray y=mxy = mx has area 12∫0mdt/(1+t2)\tfrac12\int_0^m dt/(1+t^2); so arctan⁡m=∫0mdt/(1+t2)\arctan m = \int_0^m dt/(1+t^2).
§164 Theorem
Integrals as limits of sums with any partitions (§164)
For continuous ff, 1n∑r=0n−1f(r/n)→∫01f\frac1n\sum_{r=0}^{n-1} f(r/n) \to \int_0^1 f, and more generally any Riemann sums whose longest piece tends to 00 tend to the integral.
§165 Theorem
Additivity and linearity of the integral (§165 (1)–(5))
∫abf+∫bcf=∫acf\int_a^b f + \int_b^c f = \int_a^c f for any a,b,ca, b, c (with ∫ab=−∫ba\int_a^b = -\int_b^a, ∫aa=0\int_a^a = 0), and ∫(kf+ϕ)=k∫f+∫ϕ\int(kf + \phi) = k\int f + \int\phi.
§165 Theorem
First mean value theorem for integrals (§165 (8))
If ff is continuous on [a,b][a, b], then ∫abf=(b−a)f(ξ)\int_a^b f = (b - a)f(\xi) for some ξ\xi in [a,b][a, b].
§165 Theorem
Fundamental theorem of the integral calculus (§165 (10))
If ff is continuous, F(x)=∫axf(t) dtF(x) = \int_a^x f(t)\,dt has derivative f(x)f(x), and is therefore continuous.
§165 Theorem
Generalised mean value theorem for integrals (§165 (9))
If f,ϕf, \phi are continuous and ϕ≥0\phi \ge 0 on [a,b][a, b], then ∫abfϕ=f(ξ)∫abϕ\int_a^b f\phi = f(\xi)\int_a^b\phi for some ξ\xi in [a,b][a, b].
§165 Theorem
Positivity and bounds for integrals (§165 (6), (7))
If a<ba < b and H≤f(x)≤KH \le f(x) \le K on [a,b][a, b], then H(b−a)≤∫abf≤K(b−a)H(b-a) \le \int_a^b f \le K(b-a); in particular f≥0f \ge 0 implies ∫abf≥0\int_a^b f \ge 0.
§166 Theorem
Integration by parts for definite integrals (§166)
If f′f' and ϕ′\phi' are continuous on [a,b][a, b], then ∫abfϕ′=f(b)ϕ(b)−f(a)ϕ(a)−∫abf′ϕ\int_a^b f\phi' = f(b)\phi(b) - f(a)\phi(a) - \int_a^b f'\phi.
§166 Theorem
Substitution in a definite integral (§166)
If ϕ′\phi' is continuous on [a,b][a, b] and ff is continuous on the range of ϕ\phi, then ∫ϕ(a)ϕ(b)f(t) dt=∫abf{ϕ(x)}ϕ′(x) dx\int_{\phi(a)}^{\phi(b)} f(t)\,dt = \int_a^b f\{\phi(x)\}\phi'(x)\,dx.
§166 Example
Wallis's integrals (Examples LXVI, 9–10)
un=∫0π/2sin⁡nx dxu_n = \int_0^{\pi/2}\sin^n x\,dx satisfies nun=(n−1)un−2nu_n = (n-1)u_{n-2}, with u0=12πu_0 = \tfrac12\pi, u1=1u_1 = 1.
§167 Corollary
Schlömilch's, Lagrange's and Cauchy's forms of the remainder (§167 (3), (4))
For 1≤p≤n1 \le p \le n, Rn=(1−θ)n−pf(n)(a+θh)hnp (n−1)!R_n = \frac{(1-\theta)^{n-p}f^{(n)}(a+\theta h)h^n}{p\,(n-1)!} for some θ\theta; p=np = n gives Lagrange's form, p=1p = 1 Cauchy's.
§167 Theorem
Taylor's theorem with integral remainder (§167 (1), (2))
If f(n)f^{(n)} is continuous, f(a+h)=∑k<nhkk!f(k)(a)+Rnf(a+h) = \sum_{k<n}\frac{h^k}{k!}f^{(k)}(a) + R_n with Rn=hn(n−1)!∫01(1−t)n−1f(n)(a+th) dtR_n = \frac{h^n}{(n-1)!}\int_0^1(1-t)^{n-1}f^{(n)}(a+th)\,dt exactly.
§168 Theorem
Binomial series for −1 < x < 1 (§168)
For rational mm and −1<x<1-1 < x < 1, (1+x)m=∑(mn)xn(1+x)^m = \sum\binom{m}{n}x^n; Cauchy's remainder satisfies ∣Rn∣≤∣m∣(1±∣x∣)m−1∣(m−1n−1)∣∣x∣n→0|R_n| \le |m|(1\pm|x|)^{m-1}\left|\binom{m-1}{n-1}\right||x|^n \to 0.
§169 Theorem
Error of the trapezium rule (§169 (1); Examples LXVII, 2)
∫abf=12(b−a){f(a)+f(b)}−112(b−a)3f′′(ξ)\int_a^b f = \tfrac12(b-a)\{f(a) + f(b)\} - \tfrac1{12}(b-a)^3f''(\xi) for some ξ\xi in (a,b)(a, b).
§169 Theorem
Simpson's rule with its error (§169 (2))
∫abf=16(b−a){f(a)+4f(a+b2)+f(b)}−12880(b−a)5f(4)(ξ)\int_a^b f = \tfrac16(b-a)\{f(a) + 4f(\tfrac{a+b}2) + f(b)\} - \tfrac1{2880}(b-a)^5f^{(4)}(\xi); the rule is exact for cubics.
§170 Definition
Integral of a complex function (§170)
For f=ϕ+iψf = \phi + i\psi with ϕ,ψ\phi, \psi real and continuous, ∫abf=∫abϕ+i∫abψ\int_a^b f = \int_a^b\phi + i\int_a^b\psi.
§170 Theorem
Modulus inequality for integrals (§170 (1); Examples LXV, 12)
For continuous complex (or real) ff and a<ba < b, ∣∫abf∣≤∫ab∣f∣\left|\int_a^b f\right| \le \int_a^b|f|.

Chapter VIII: The convergence of infinite series and infinite integrals

§ResultStatement
§173 Theorem
A positive series cannot oscillate (A, §173)
A series of positive terms either converges or diverges to +∞+\infty.
§173 Theorem
Bounded partial sums criterion (B, §173)
A series of positive terms converges if and only if its partial sums are bounded above.
§173 Theorem
Comparison theorem (C, §173)
If ∑un\sum u_n converges and 0≤vn≤Kun0 \le v_n \le K u_n, then ∑vn\sum v_n converges and ∑vn≤K∑un\sum v_n \le K\sum u_n; if ∑un\sum u_n diverges and vn≥Kunv_n \ge K u_n with K>0K > 0, then ∑vn\sum v_n diverges.
§173 Corollary
Limit comparison test (D, §173)
If un,vn>0u_n, v_n > 0 and un/vnu_n/v_n tends to a non-zero limit, then ∑un\sum u_n and ∑vn\sum v_n converge or diverge together.
§174 Theorem
Cauchy's (root) test (§174, 2)
If vn≥0v_n \ge 0 and vn1/n≤r<1v_n^{1/n} \le r < 1 for all large nn, then ∑vn\sum v_n converges; if vn1/n≥1v_n^{1/n} \ge 1 for infinitely many nn, it diverges.
§175 Theorem
Comparison of ratios (§175, 4 and 5)
If un,vn>0u_n, v_n > 0 and vn+1/vn≤un+1/unv_{n+1}/v_n \le u_{n+1}/u_n for large nn, then vn≤Kunv_n \le Ku_n; so convergence of ∑un\sum u_n gives convergence of ∑vn\sum v_n (and dually for divergence).
§175 Theorem
d'Alembert's (ratio) test (§175, 6 and 7)
If vn+1/vn≤r<1v_{n+1}/v_n \le r < 1 for large nn then ∑vn\sum v_n converges. If the ratio or the root tends to ll, the series converges for l<1l < 1 and diverges for l>1l > 1.
§176 Theorem
Dirichlet's theorem on rearrangement (§176)
A convergent series of positive terms converges to the same sum whatever the order of its terms.
§177 Theorem
Multiplication of positive series (§177)
If ∑un=s\sum u_n = s and ∑vn=t\sum v_n = t are convergent series of positive terms, then ∑k∑m+n=kumvn\sum_k \sum_{m+n=k} u_m v_n converges to stst.
§179 Theorem
Abel's (Pringsheim's) theorem (§179)
If ∑un\sum u_n is a convergent series of positive decreasing terms, then nun→0nu_n \to 0.
§180 Theorem
Maclaurin's integral test (§180)
If ϕ\phi is positive, continuous and decreasing for x≥1x \ge 1, then ∑ϕ(n)\sum\phi(n) converges iff ∫1ξϕ\int_1^\xi\phi tends to a limit ll; the sum is then less than ϕ(1)+l\phi(1) + l.
§181 Theorem
Convergence of Σn^(−s) (§181)
∑n−s\sum n^{-s} converges if s>1s > 1, with sum less than s/(s−1)s/(s-1), and diverges if s≤1s \le 1.
§182 Theorem
Cauchy's condensation test (§182)
If ϕ(n)\phi(n) is positive and decreasing, ∑ϕ(n)\sum\phi(n) and ∑2nϕ(2n)\sum 2^n\phi(2^n) converge or diverge together.
§183 Corollary
Convergence of the hypergeometric series (§183)
The series 1+αβ1⋅γ+⋯1 + \frac{\alpha\beta}{1\cdot\gamma} + \cdots converges if γ>α+β\gamma > \alpha + \beta and diverges if γ<α+β\gamma < \alpha + \beta.
§183 Theorem
Gauss's test (finer ratio test) (§183)
If vn>0v_n > 0 and vn+1/vn=1−a/n+O(1/n2)v_{n+1}/v_n = 1 - a/n + O(1/n^2), then ∑vn\sum v_n converges for a>1a > 1 and diverges for a<1a < 1 (and, by Ch. IX, for a=1a = 1).
§184 Definition
Infinite integral (§184)
∫a∞ϕ(x) dx=lim⁡X→∞∫aXϕ(x) dx\int_a^\infty\phi(x)\,dx = \lim_{X\to\infty}\int_a^X\phi(x)\,dx when the limit exists.
§185 Theorem
Comparison test for infinite integrals (§185)
For positive continuous integrands, ∫a∞ϕ\int_a^\infty\phi converges iff bounded; ψ≤Kϕ\psi \le K\phi transfers convergence; if xsϕ(x)→l>0x^s\phi(x) \to l > 0 the integral converges iff s>1s > 1.
§186 Theorem
Substitution and integration by parts in infinite integrals (§186)
The rules of §166 extend to infinite integrals by letting the upper limit tend to infinity; for parts, if two of the three limits exist so does the third.
§187 Definition
Infinite integral of the second kind (§187)
If ϕ→±∞\phi \to \pm\infty as x→a+0x \to a+0, ∫aAϕ=lim⁡ε→+0∫a+εAϕ\int_a^A\phi = \lim_{\varepsilon\to+0}\int_{a+\varepsilon}^A\phi; ∫aA(x−a)−sdx\int_a^A (x-a)^{-s}dx converges iff s<1s < 1.
§189 Principle
Substitution with a non-monotone relation (§189)
When transforming an integral by y=g(x)y = g(x), split the range where gg turns and use the correct branch of the inverse on each piece.
§190 Theorem
Series whose terms are eventually of one sign (§190)
If unu_n has constant sign from some n0n_0 on, ∑un\sum u_n converges or diverges (it cannot oscillate), and the tests for positive series apply to its tail.
§191 Definition
Absolute convergence (§191)
∑un\sum u_n is absolutely convergent if ∑∣un∣\sum |u_n| converges.
§191 Theorem
Absolutely convergent series converge (§191)
If ∑∣un∣\sum |u_n| converges, then ∑vn\sum v_n, ∑wn\sum w_n and ∑un\sum u_n converge, ∑un=∑vn−∑wn\sum u_n = \sum v_n - \sum w_n, and ∣∑un∣≤∑∣un∣|\sum u_n| \le \sum |u_n|.
§192 Theorem
Rearrangement of absolutely convergent series (§192)
Every rearrangement of an absolutely convergent series is absolutely convergent with the same sum.
§193 Definition
Conditional convergence (§193)
∑un\sum u_n is conditionally convergent if it converges but ∑∣un∣\sum |u_n| diverges; then its positive and negative parts both diverge to ∞\infty.
§193 Theorem
Riemann's rearrangement theorem
A conditionally convergent real series can be rearranged to converge to any given real number, or to diverge to +∞+\infty or −∞-\infty.
§195 Theorem
Alternating series test (Leibniz) (§195)
If φn>0\varphi_n > 0 decreases steadily to 00, then φ0−φ1+φ2−⋯\varphi_0 - \varphi_1 + \varphi_2 - \cdots converges, its sum lies between any two consecutive partial sums, and ∣s−sn∣≤φn+1|s - s_n| \le \varphi_{n+1}.
§196 Theorem
Abel's test (§196)
If ∑an\sum a_n converges and φn\varphi_n is positive and decreasing, then ∑anφn\sum a_n\varphi_n converges.
§196 Corollary
Convergence of ∑φncos⁡nθ\sum \varphi_n\cos n\theta and ∑φnsin⁡nθ\sum\varphi_n\sin n\theta (Examples LXXIX, 2)
The partial sums of ∑cos⁡nθ\sum\cos n\theta, ∑sin⁡nθ\sum\sin n\theta are at most 1/∣sin⁡12θ∣1/|\sin\frac12\theta|; so if φn↓0\varphi_n \downarrow 0 both series converge, except the cosine series when θ\theta is a multiple of 2π2\pi.
§196 Theorem
Dirichlet's test (§196)
If the partial sums of ∑an\sum a_n are bounded and φn\varphi_n decreases steadily to 00, then ∑anφn\sum a_n\varphi_n converges.
§197 Theorem
Absolute convergence of complex series (§197)
∑(vn+iwn)\sum (v_n + iw_n) is absolutely convergent (both ∑vn\sum v_n, ∑wn\sum w_n absolutely convergent) if and only if ∑∣un∣\sum |u_n| converges; it then converges.
§198 Theorem
Power series: three kinds of behaviour (§198, A)
A power series may converge for all zz (∑zn/n!\sum z^n/n!), only for z=0z = 0 (∑n! zn\sum n!\,z^n), or in a region (∑zn\sum z^n, for ∣z∣<1|z| < 1).
§199 Theorem
Convergence spreads inwards (Abel's lemma) (§199, B)
If ∑anzn\sum a_n z^n converges (or merely has bounded terms) at z1z_1, it converges absolutely for every ∣z∣<∣z1∣|z| < |z_1|.
§200 Theorem
Radius of convergence by the ratio or root test (Examples LXXX, 3)
If ∣an+1/an∣→λ|a_{n+1}/a_n| \to \lambda or ∣an∣1/n→λ|a_n|^{1/n} \to \lambda, the radius of convergence of ∑anzn\sum a_n z^n is 1/λ1/\lambda.
§200 Theorem
The circle of convergence (§200)
A power series converges only at 00, or absolutely everywhere, or absolutely for ∣z∣<R|z| < R and nowhere for ∣z∣>R|z| > R; nothing general can be said on ∣z∣=R|z| = R.
§201 Theorem
Uniqueness of power series (§201)
f(z)=∑n≤manzn+O(∣z∣m+1)f(z) = \sum_{n \le m} a_n z^n + O(|z|^{m+1}) as z→0z \to 0; hence if ∑anzn=∑bnzn\sum a_n z^n = \sum b_n z^n near 00 then an=bna_n = b_n for all nn.
§202 Theorem
Multiplication of absolutely convergent series (§202)
If ∑un=s\sum u_n = s and ∑vn=t\sum v_n = t absolutely, the Cauchy product ∑wn\sum w_n, wn=∑m=0numvn−mw_n = \sum_{m=0}^{n} u_m v_{n-m}, converges absolutely to stst.
§203 Theorem
Absolutely convergent integrals converge (§203)
If ∫a∞∣f∣\int_a^\infty |f| converges, so do the integrals of the positive and negative parts of ff, and ∫a∞f\int_a^\infty f.
§203 Theorem
Convergence of ∫ϕ(x)cos⁡tx dx\int \phi(x)\cos tx\,dx and ∫ϕ(x)sin⁡tx dx\int\phi(x)\sin tx\,dx (§203)
If ϕ≥0\phi \ge 0, ϕ′≤0\phi' \le 0 is continuous and ϕ→0\phi \to 0, then ∫a∞ϕ(x)cos⁡tx dx\int_a^\infty\phi(x)\cos tx\,dx and ∫a∞ϕ(x)sin⁡tx dx\int_a^\infty\phi(x)\sin tx\,dx converge (t≠0t \ne 0); for ϕ=x−s\phi = x^{-s} absolutely iff s>1s > 1.

Chapter IX: The logarithmic, exponential and circular functions of a real variable

§ResultStatement
§205 Theorem
Basic logarithmic inequality (Examples LXXXIII, 1–2)
(x−1)/x<log⁡x<x−1(x-1)/x < \log x < x - 1 for x>0x > 0, x≠1x \ne 1; equivalently h/(1+h)<log⁡(1+h)<hh/(1+h) < \log(1+h) < h.
§205 Theorem
First properties of the logarithm
log⁡1=0\log 1 = 0; log⁡\log is continuous and strictly increasing with derivative 1/x1/x; log⁡x→∞\log x \to \infty as x→∞x \to \infty; log⁡(1/x)=−log⁡x\log(1/x) = -\log x, so log⁡x→−∞\log x \to -\infty as x→+0x \to +0.
§205 Definition
The logarithm
For x>0x > 0, log⁡x=∫1xdt/t\log x = \int_1^x dt/t.
§206 Theorem
Functional equation of the logarithm (§206 (1))
log⁡xy=log⁡x+log⁡y\log xy = \log x + \log y for all positive xx, yy.
§206 Theorem
Uniqueness of the logarithm (Examples LXXXIV, 1)
A differentiable ff on (0,∞)(0, \infty) with f(xy)=f(x)+f(y)f(xy) = f(x) + f(y) is Clog⁡xC\log x.
§208 Theorem
The logarithm is slower than every power
x−αlog⁡x→0x^{-\alpha}\log x \to 0 as x→∞x \to \infty for every α>0\alpha > 0; indeed log⁡x<xβ/β\log x < x^\beta/\beta for x>1x > 1, β>0\beta > 0.
§209 Theorem
Behaviour of the logarithm at 0
xαlog⁡x→0x^\alpha\log x \to 0 as x→+0x \to +0 for every α>0\alpha > 0.
§210 Theorem
Incompleteness of scales of infinity (Examples LXXXV, 1, 3)
If f/F→0f/F \to 0 with both tending to ∞\infty, then fF\sqrt{fF} lies strictly between them on the scale.
§210 Theorem
The logarithmic scale of infinity
log⁡k+1x\log_{k+1} x tends to infinity more slowly than every positive power of log⁡kx\log_k x; the scale x1/nx^{1/n}, (log⁡x)1/n(\log x)^{1/n}, (log⁡log⁡x)1/n(\log\log x)^{1/n}, … is ordered by growth.
§211 Definition
The number e defined by log⁡e=1\log e = 1
ee is the unique number with log⁡e=1\log e = 1; 2<e<32 < e < 3, and log⁡ey=y\log e^y = y for rational yy.
§212 Definition
The exponential function
For real yy, eye^y is the positive xx with log⁡x=y\log x = y; it is continuous, positive and strictly increasing from 00 to ∞\infty.
§213 Theorem
Derivative and addition formula of the exponential (§213 (1), (2))
ddyey=ey\frac{d}{dy}e^y = e^y and ey+z=eyeze^{y+z} = e^y e^z.
§213 Theorem
The exponential beats every power (§213 (3))
y−αey→∞y^{-\alpha}e^{y} \to \infty as y→∞y \to \infty, for every α\alpha.
§213 Theorem
Uniqueness of the exponential (Examples LXXXVI, 2)
A differentiable solution of f(y+z)=f(y)f(z)f(y+z) = f(y)f(z) is 00 or eaye^{ay}.
§214 Definition
The general power
For a>0a > 0 and real xx, ax=exlog⁡aa^x = e^{x\log a}; the laws of indices hold, ddxax=axlog⁡a\frac{d}{dx}a^x = a^x\log a and ddxxm=mxm−1\frac{d}{dx}x^m = mx^{m-1} for real mm.
§215 Theorem
The exponential limit (§215 (1))
(1+x/n)n→ex(1 + x/n)^n \to e^x and (1−x/n)−n→ex(1 - x/n)^{-n} \to e^x; for ∣y∣<n|y| < n, (1+y/n)n<ey<(1−y/n)−n(1 + y/n)^n < e^y < (1 - y/n)^{-n}.
§216 Theorem
The logarithmic limit
n(x1/n−1)→log⁡xn(x^{1/n} - 1) \to \log x and n(1−x−1/n)→log⁡xn(1 - x^{-1/n}) \to \log x, with n(1−x−1/n)<log⁡x<n(x1/n−1)n(1 - x^{-1/n}) < \log x < n(x^{1/n} - 1).
§217 Theorem
Change of base
log⁡ax=log⁡x/log⁡a\log_a x = \log x/\log a; in particular log⁡10x=0.4343…×log⁡x\log_{10} x = 0.4343\ldots\times\log x.
§218 Theorem
Euler's constant (Examples XC, 1)
1+12+⋯+1n−1−log⁡n1 + \frac12 + \dots + \frac{1}{n-1} - \log n increases to a limit γ=0.5772…\gamma = 0.5772\ldots with 0<γ≤10 < \gamma \le 1.
§218 Theorem
Logarithmic test of convergence
∑1/(n(log⁡n)s)\sum 1/(n(\log n)^s) and ∫∞dx/(x(log⁡x)s)\int^\infty dx/(x(\log x)^s) converge iff s>1s > 1; likewise with further iterated logarithms.
§219 Example
ee is irrational (Examples XCI, 6)
ee is not a rational number.
§219 Corollary
Series for axa^x (§219 (4))
For a>0a > 0, ax=∑n≥0(xlog⁡a)n/n!a^x = \sum_{n\ge0} (x\log a)^n/n! for every real xx.
§219 Theorem
The exponential series (§219 (1))
ex=∑n≥0xn/n!e^x = \sum_{n \ge 0} x^n/n! for every real xx; in particular e=∑1/n!e = \sum 1/n!.
§219 Theorem
The exponential theorem (§219 (3))
∑xnn!⋅∑ynn!=∑(x+y)nn!\sum \frac{x^n}{n!} \cdot \sum\frac{y^n}{n!} = \sum \frac{(x+y)^n}{n!} for all real x,yx, y.
§220 Theorem
The logarithmic series
log⁡(1+x)=x−x22+x33−…\log(1+x) = x - \frac{x^2}{2} + \frac{x^3}{3} - \dots for −1<x≤1-1 < x \le 1, and the series diverges otherwise; in particular log⁡2=1−12+13−…\log 2 = 1 - \frac12 + \frac13 - \dots.
§221 Example
Machin's formula (Examples XCII, 12)
14π=4arctan⁡15−arctan⁡1239\frac14\pi = 4\arctan\frac15 - \arctan\frac1{239}, which with the arctan series gives π\pi to six decimals from a few terms.
§221 Theorem
The series for the inverse tangent
arctan⁡x=x−x33+x55−…\arctan x = x - \frac{x^3}{3} + \frac{x^5}{5} - \dots (principal value) for −1≤x≤1-1 \le x \le 1; in particular 14π=1−13+15−…\frac14\pi = 1 - \frac13 + \frac15 - \dots.
§222 Lemma
Size of the binomial coefficients
If mm is not a non-negative integer, (mn)→0\binom{m}{n} \to 0 if m>−1m > -1, while ∣(mn)∣|\binom{m}{n}| is eventually non-decreasing if m≤−1m \le -1.
§222 Theorem
The binomial series for any real exponent
For real mm (not a non-negative integer), (1+x)m=∑(mn)xn(1+x)^m = \sum\binom{m}{n}x^n for ∣x∣<1|x| < 1; at x=1x = 1 the series converges iff m>−1m > -1 (sum 2m2^m); at x=−1x = -1 iff m>0m > 0 (sum 00).
§223 Definition
The exponential function defined by its series (§223 (1))
exp⁡x=∑n≥0xn/n!\exp x = \sum_{n\ge0} x^n/n!; it satisfies exp⁡x exp⁡y=exp⁡(x+y)\exp x\,\exp y = \exp(x+y) and exp⁡′=exp⁡\exp' = \exp, and exp⁡x=ex\exp x = e^x for rational xx (for irrational xx this defines exe^x).
§224 Definition
Analytical definition of the circular functions (§224 (1)–(6))
arctan⁡x=∫0xdt1+t2\arctan x = \int_0^x \frac{dt}{1+t^2}, 12π=∫0∞dt1+t2\frac12\pi = \int_0^\infty\frac{dt}{1+t^2}, tan⁡\tan its inverse, cos⁡y=(1+tan⁡2y)−1/2\cos y = (1+\tan^2 y)^{-1/2}, sin⁡y=tan⁡ycos⁡y\sin y = \tan y\cos y on (−12π,12π)(-\frac12\pi, \frac12\pi), extended by tan⁡(y+π)=tan⁡y\tan(y+\pi) = \tan y, cos⁡(y+π)=−cos⁡y\cos(y+\pi) = -\cos y, sin⁡(y+π)=−sin⁡y\sin(y+\pi) = -\sin y.
§225 Theorem
Addition formula for the tangent (§225 (1))
tan⁡(y1+y2)=tan⁡y1+tan⁡y21−tan⁡y1tan⁡y2\tan(y_1+y_2) = \frac{\tan y_1 + \tan y_2}{1 - \tan y_1\tan y_2} whenever both sides are defined.
§225 Theorem
Addition of inverse tangents
arctan⁡x1+arctan⁡x2=arctan⁡x1+x21−x1x2\arctan x_1 + \arctan x_2 = \arctan\frac{x_1+x_2}{1-x_1x_2} if x1x2<1x_1x_2 < 1, with +π+\pi added if x1x2>1x_1x_2 > 1, x1>0x_1 > 0 and −π-\pi if x1x2>1x_1x_2 > 1, x1<0x_1 < 0.
§226 Theorem
Addition formulae for cosine and sine (§226 (2))
cos⁡(y1+y2)=cos⁡y1cos⁡y2−sin⁡y1sin⁡y2\cos(y_1+y_2) = \cos y_1\cos y_2 - \sin y_1\sin y_2 and sin⁡(y1+y2)=sin⁡y1cos⁡y2+cos⁡y1sin⁡y2\sin(y_1+y_2) = \sin y_1\cos y_2 + \cos y_1\sin y_2 for all real y1,y2y_1, y_2.

Chapter X: The general theory of the logarithmic, exponential and circular functions

§ResultStatement
§229 Definition
Complex curvilinear integral
∫Cf(z) dz=∫C(u dx−v dy)+i∫C(v dx+u dy)=∫t0t1f(z(t)) z′(t) dt\int_C f(z)\,dz = \int_C (u\,dx - v\,dy) + i\int_C (v\,dx + u\,dy) = \int_{t_0}^{t_1} f(z(t))\,z'(t)\,dt for a rational function ff and a smooth path CC.
§229 Theorem
Integrals of derivatives depend only on the end-points
If FF is rational with no pole on CC, then ∫CF′(z) dz=F(B)−F(A)\int_C F'(z)\,dz = F(B) - F(A) for every path CC from AA to BB.
§230 Definition
General logarithm
Log⁡ζ=∫Cdz/z\operatorname{Log}\zeta = \int_C dz/z, where CC is a path from 11 to ζ\zeta not passing through 00.
§231 Theorem
Functional equation of the complex logarithm (Examples XCIV, 3)
Log⁡z1z2=Log⁡z1+Log⁡z2\operatorname{Log} z_1 z_2 = \operatorname{Log} z_1 + \operatorname{Log} z_2 is completely true; with principal values it holds only when the principal amplitudes sum to a number in (−π,π](-\pi, \pi].
§231 Definition
Principal value of the logarithm
log⁡ζ=log⁡∣ζ∣+iϕ\log\zeta = \log|\zeta| + i\phi with −π<ϕ≤π-\pi < \phi \le \pi; given by the straight path when ζ\zeta is not real and negative.
§231 Theorem
The values of Log ζ
The values of Log⁡ζ\operatorname{Log}\zeta are exactly log⁡∣ζ∣+i(ϕ+2kπ)\log|\zeta| + i(\phi + 2k\pi), k∈Zk \in \mathbb Z, where kk is the number of times the path (closed by the segment back to 11) winds round 00.
§232 Theorem
The exponential function is one-valued (Theorem)
If z=exp⁡ζz = \exp\zeta means that ζ\zeta is a value of Log⁡z\operatorname{Log} z, then each ζ\zeta determines exactly one zz; for real ζ\zeta, exp⁡ζ=eζ\exp\zeta = e^\zeta.
§233 Theorem
Value of exp ζ
exp⁡(ξ+iη)=eξ(cos⁡η+isin⁡η)\exp(\xi + i\eta) = e^\xi(\cos\eta + i\sin\eta).
§234 Theorem
Addition theorem for exp
exp⁡(ζ1+ζ2)=exp⁡ζ1exp⁡ζ2\exp(\zeta_1 + \zeta_2) = \exp\zeta_1\exp\zeta_2 for all complex ζ1,ζ2\zeta_1, \zeta_2.
§235 Definition
General power
aζ=exp⁡(ζLog⁡a)a^\zeta = \exp(\zeta\operatorname{Log} a) for a≠0a \ne 0; the principal value uses log⁡a\log a.
§236 Theorem
Number of values of a^ζ
aζa^\zeta has infinitely many values unless ζ\zeta is real and rational; for ζ=p/q\zeta = p/q in lowest terms it has exactly qq values. All values lie on an equiangular spiral.
§236 Example
Values of i^i (Examples XCV, 1)
ii=e−(2k+12)πi^i = e^{-(2k + \frac12)\pi}, all real and positive; principal value e−π/2e^{-\pi/2}.
§238 Definition
Circular functions of a complex variable
cos⁡ζ=12{exp⁡(iζ)+exp⁡(−iζ)}\cos\zeta = \frac12\{\exp(i\zeta) + \exp(-i\zeta)\}, sin⁡ζ=−12i{exp⁡(iζ)−exp⁡(−iζ)}\sin\zeta = -\frac12 i\{\exp(i\zeta) - \exp(-i\zeta)\}; every identity of elementary trigonometry remains true.
§239 Theorem
Circular and hyperbolic functions
cos⁡iζ=cosh⁡ζ\cos i\zeta = \cosh\zeta, sin⁡iζ=isinh⁡ζ\sin i\zeta = i\sinh\zeta, cosh⁡iζ=cos⁡ζ\cosh i\zeta = \cos\zeta, sinh⁡iζ=isin⁡ζ\sinh i\zeta = i\sin\zeta.
§240 Theorem
Real and imaginary parts of cos and sin
cos⁡(ξ+iη)=cos⁡ξcosh⁡η−isin⁡ξsinh⁡η\cos(\xi + i\eta) = \cos\xi\cosh\eta - i\sin\xi\sinh\eta, sin⁡(ξ+iη)=sin⁡ξcosh⁡η+icos⁡ξsinh⁡η\sin(\xi + i\eta) = \sin\xi\cosh\eta + i\cos\xi\sinh\eta.
§241 Theorem
Inverse tangent as a logarithm
For real xx, arctan⁡x=12iLog⁡1+ix1−ix\arctan x = \frac{1}{2i}\operatorname{Log}\frac{1 + ix}{1 - ix}, completely true; similarly arccos⁡x=−iLog⁡{x±i1−x2}\arccos x = -i\operatorname{Log}\{x \pm i\sqrt{1 - x^2}\}.
§242 Theorem
Exponential series
exp⁡z=∑n≥0zn/n!\exp z = \sum_{n \ge 0} z^n/n! for every complex zz.
§243 Theorem
Cosine and sine series
cos⁡z=∑(−1)mz2m/(2m)!\cos z = \sum (-1)^m z^{2m}/(2m)! and sin⁡z=∑(−1)mz2m+1/(2m+1)!\sin z = \sum (-1)^m z^{2m+1}/(2m+1)! for every complex zz.
§244 Theorem
Logarithmic series
log⁡(1+z)=z−z2/2+z3/3−⋯\log(1 + z) = z - z^2/2 + z^3/3 - \cdots (principal value) for ∣z∣≤1|z| \le 1, z≠−1z \ne -1.
§245 Theorem
Logarithmic series on the unit circle
For −π<θ<π-\pi < \theta < \pi: ∑(−1)n−1cos⁡nθ/n=12log⁡(4cos⁡212θ)\sum (-1)^{n-1}\cos n\theta/n = \frac12\log(4\cos^2\frac12\theta) and ∑(−1)n−1sin⁡nθ/n=12θ\sum (-1)^{n-1}\sin n\theta/n = \frac12\theta.
§246 Theorem
Exponential limit
(1+z/n)n→exp⁡z(1 + z/n)^n \to \exp z for every complex zz.
§247 Theorem
General binomial theorem (Theorem)
For all complex mm and ∣z∣<1|z| < 1, ∑n(mn)zn=exp⁡{mlog⁡(1+z)}\sum_n \binom mn z^n = \exp\{m\log(1 + z)\}, the principal value of (1+z)m(1 + z)^m.

Appendices: Appendices

§ResultStatement
App. II Theorem
Winding number counts the roots (Examples on Appendix II, 1)
If ff has no root on a simple closed path γ\gamma described once anticlockwise, then the change of arg⁡f\arg f round γ\gamma, divided by 2π2\pi, is the number of roots of ff inside γ\gamma.
App. II Lemma
Winding round a large square (Appendix II (A))
If ∑j≥1∣aj∣/(∣a0∣Rj)<12\sum_{j\ge1} |a_j|/(|a_0|R^j) < \tfrac12, then on the square ∣x∣,∣y∣≤R|x|, |y| \le R the polynomial f=a0zn+⋯+anf = a_0z^n + \dots + a_n has no root on the boundary, and arg⁡f\arg f increases by 2nπ2n\pi round it.
App. I Corollary
Companion to Minkowski's inequality (Appendix I (17))
For k>1k > 1 and positive am,bma_m, b_m: ∑(am+bm)k>∑amk+∑bmk\sum (a_m + b_m)^k > \sum a_m^k + \sum b_m^k.
App. II Lemma
d'Alembert's lemma (Appendix II (B))
If ff is a polynomial and f(z0)≠0f(z_0) \ne 0, there are points zz arbitrarily close to z0z_0 with ∣f(z)∣<∣f(z0)∣|f(z)| < |f(z_0)|.
App. II Theorem
Fundamental theorem of algebra (proof by the least value of |f|) (Appendix II (B))
∣f(z)∣|f(z)| attains a least value over the plane, and by d'Alembert's lemma that value is 00.
App. II Theorem
Fundamental theorem of algebra (proof by winding) (Appendix II (A))
Every polynomial of degree n≥1n \ge 1 with complex coefficients has a root; proved by nested squares on which the winding number of the image is nonzero.
App. I Theorem
General theorem of the means (weighted AM–GM) (G, Appendix I)
If q1+⋯+qm=1q_1 + \dots + q_m = 1 with all qv>0q_v > 0, then a1q1⋯amqm≤q1a1+⋯+qmama_1^{q_1}\cdots a_m^{q_m} \le q_1 a_1 + \dots + q_m a_m, with equality only when all aa's are equal.
App. I Theorem
Hölder's inequality (H, Appendix I)
For k>1k > 1 and positive am,bma_m, b_m: ∑ambm≤(∑amk)1/k(∑bmk′)1/k′\sum a_m b_m \le (\sum a_m^k)^{1/k}(\sum b_m^{k'})^{1/k'}, with equality only when amk/bmk′a_m^k / b_m^{k'} is independent of mm.
App. IV Theorem
Correlation of homogeneous with common Cartesian geometry (Appendix IV)
(x,y)↦(x:y:1)(x, y) \mapsto (x : y : 1) matches the ordinary plane with the points having z≠0z \ne 0, and the line ax+by+c=0ax + by + c = 0 with the line ax+by+cz=0ax + by + cz = 0 minus its point at infinity (b:−a:0)(b : -a : 0).
App. III Theorem
Interchanging limits under uniform convergence
If a(m,n)→bma(m, n) \to b_m as n→∞n \to \infty uniformly in mm, and a(m,n)→cna(m, n) \to c_n as m→∞m \to \infty for each nn, then lim⁡mbm\lim_m b_m and lim⁡ncn\lim_n c_n both exist and are equal.
App. I Theorem
Minkowski's inequality (M, Appendix I)
For k>1k > 1 and positive am,bma_m, b_m: (∑(am+bm)k)1/k≤(∑amk)1/k+(∑bmk)1/k(\sum (a_m + b_m)^k)^{1/k} \le (\sum a_m^k)^{1/k} + (\sum b_m^k)^{1/k}, with equality only when the sets are proportional.
App. III Example
Repeated limits may differ (Appendix III (3), (4))
lim⁡x→0lim⁡y→0(x−y)/(x+y)=1\lim_{x\to0}\lim_{y\to0}(x-y)/(x+y) = 1 but lim⁡y→0lim⁡x→0(x−y)/(x+y)=−1\lim_{y\to0}\lim_{x\to0}(x-y)/(x+y) = -1; and lim⁡x→1−∑(xn−1−xn)=1\lim_{x\to1-}\sum(x^{n-1}-x^n) = 1 while ∑lim⁡x→1−(xn−1−xn)=0\sum\lim_{x\to1-}(x^{n-1}-x^n) = 0.
App. IV Theorem
Two lines of homogeneous geometry always meet
Distinct lines ax+by+cz=0ax + by + cz = 0 and a′x+b′y+c′z=0a'x + b'y + c'z = 0 meet in exactly one point, the class of (bc′−cb′,ca′−ac′,ab′−ba′)(bc' - cb', ca' - ac', ab' - ba'); parallel lines meet on the line at infinity.
App. I Corollary
Young's inequality
For x,y>0x, y > 0, k>1k > 1 and k′=k/(k−1)k' = k/(k-1): xy≤xk/k+yk′/k′xy \le x^k/k + y^{k'}/k', with equality only when y=xk−1y = x^{k-1}.