Search Knowledge

© 2026 LIBREUNI PROJECT

Introduction to Calculus/Limits and derivatives
Page 2 of 12

Derivatives: Rates of Change

Understanding the derivative as a limit, proving fundamental differentiation rules, and examining linear approximations.

Lesson briefLimits and derivatives · 3 h · 3 outcomes

Required knowledge

  • Epsilon-delta limit laws and continuity on closed intervals
  • Algebraic factorization, difference quotients, and trigonometric identities

Learning outcomes

  • Derive instantaneous rates of change and tangent slopes from the difference-quotient limit
  • Prove and apply the product, quotient, chain, and inverse-function differentiation rules
  • Quantify local linear approximation error using the tangent-line remainder

Assessment target

Compute derivatives from first principles and via the chain/inverse rules, and bound linear approximation error near a base point.

Scope boundary

  • Fréchet and Gateaux derivatives in higher dimensions and measure-theoretic differentiability.

A particle’s position along a line is modeled as s(t)=ts(t)=\sqrt{t} metres after tt seconds. For h≠0h\ne0 with 4+h≥04+h\ge0 its average velocity between t=4t=4 and t=4+ht=4+h is

s(4+h)−s(4)h=4+h−2h.\frac{s(4+h)-s(4)}{h}=\frac{\sqrt{4+h}-2}{h}.

At h=0h=0 this expression becomes 0/00/0 so substituting zero does not give an instantaneous velocity. Instead, rationalizing for h≠0h\ne0 gives 1/(4+h+2)1/(\sqrt{4+h}+2) This approaches 1/41/4 as hh approaches zero from either side: the limiting rate is 1/41/4 m/s. The derivative formalizes this limiting process and, geometrically, gives the slope of a tangent line.

Differentiability and the Derivative

We formally define the derivative of a function at a point as a limit (Definition 1.1.1) of the difference quotient.

Definition 1.2.1The Derivative

Let f:I→Rf: I \to \mathbb{R} be defined on an open interval II containing cc The function ff is said to be differentiable at cc if the limit

f′(c)=lim⁡h→0f(c+h)−f(c)hf'(c) = \lim_{h \to 0} \frac{f(c+h) - f(c)}{h}

exists. If this limit exists, we denote it by f′(c)f'(c) or dfdx(c)\frac{df}{dx}(c)

Secant lines of $y = x^2$ at $a = 1$ converge to the tangent line of slope $f'(1) = 2$ as $h \to 0$. Secant lines with diminishing step sizes h = 1, 0.5, and 0.25 converging to the tangent line with slope 2 at x = 1 for y = x squared.
Figure 1.2.1: Secant lines of y=x2y = x^2 at a=1a = 1 converge to the tangent line of slope f′(1)=2f'(1) = 2 as h→0h \to 0.

Differentiability is a stronger condition than continuity (Definition 1.1.4). A function can be continuous but not differentiable, but the converse is always true.

Theorem 1.2.2Differentiability Implies Continuity

Builds on:Definition 1.2.1

If ff is differentiable at cc then ff is continuous at cc

ProofExpand proof

To prove lim⁡x→cf(x)=f(c)\lim_{x \to c} f(x) = f(c) it suffices to show that lim⁡x→c(f(x)−f(c))=0\lim_{x \to c} (f(x) - f(c)) = 0 For x≠cx \neq c we can write:

f(x)−f(c)=f(x)−f(c)x−c⋅(x−c).f(x) - f(c) = \frac{f(x) - f(c)}{x - c} \cdot (x - c).

Taking the limit as x→cx \to c

lim⁡x→c(f(x)−f(c))=lim⁡x→c(f(x)−f(c)x−c⋅(x−c)).\lim_{x \to c} (f(x) - f(c)) = \lim_{x \to c} \left( \frac{f(x) - f(c)}{x - c} \cdot (x - c) \right).

Since ff is differentiable at cc the limit of the first term is f′(c)f'(c) Since lim⁡x→c(x−c)=0\lim_{x \to c} (x - c) = 0 we have:

lim⁡x→c(f(x)−f(c))=f′(c)⋅0=0.\lim_{x \to c} (f(x) - f(c)) = f'(c) \cdot 0 = 0.

Therefore, lim⁡x→cf(x)=f(c)\lim_{x \to c} f(x) = f(c) and ff is continuous at cc

∎

A classic counterexample of a function that is continuous but not differentiable is f(x)=∣x∣f(x) = |x| at x=0x = 0 The difference quotient as h→0+h \to 0^+ is 11 but as h→0−h \to 0^- is −1-1 so the limit does not exist.

Non-differentiability at a corner: left secant slope is $-1$ while right secant slope is $+1$, so no tangent exists at $x = 0$. Graph of the absolute value function f(x) = |x| at the sharp corner x = 0 with opposing left and right difference quotients.
Figure 1.2.2: Non-differentiability at a corner: left secant slope is −1-1 while right secant slope is +1+1, so no tangent exists at x=0x = 0.

Worked trace: derivative from the definition via rationalization. Let f(x)=xf(x) = \sqrt{x} Find f′(4)f'(4) using the limit definition.

  1. Form difference quotient: 4+h−4h=4+h−2h\frac{\sqrt{4+h} - \sqrt{4}}{h} = \frac{\sqrt{4+h} - 2}{h}
  2. Multiply numerator and denominator by conjugate (4+h+2)(\sqrt{4+h} + 2) (4+h−2)(4+h+2)h(4+h+2)=(4+h)−4h(4+h+2)=hh(4+h+2).\frac{(\sqrt{4+h} - 2)(\sqrt{4+h} + 2)}{h(\sqrt{4+h} + 2)} = \frac{(4+h) - 4}{h(\sqrt{4+h} + 2)} = \frac{h}{h(\sqrt{4+h} + 2)}.
  3. Cancel h≠0h \neq 0 14+h+2.\frac{1}{\sqrt{4+h} + 2}.
  4. Evaluate limit as h→0h \to 0 f′(4)=lim⁡h→014+h+2=14+2=14.f'(4) = \lim_{h \to 0} \frac{1}{\sqrt{4+h} + 2} = \frac{1}{\sqrt{4} + 2} = \frac{1}{4}.

The calculation confirms the particle’s instantaneous velocity at t=4t=4 1/41/4 m/s. Cancel the factor hh only for h≠0h\ne0 before taking the limit; the quotient is not defined at h=0h=0

Worked trace: a reciprocal function and a domain boundary. Let g(x)=1/xg(x)=1/x for x≠0x\ne 0 At x=2x=2

g′(2)=lim⁡h→012+h−12h=lim⁡h→02−(2+h)2(2+h)h=lim⁡h→0−h2h(2+h)=lim⁡h→0−12(2+h)=−14.g'(2)=\lim_{h\to0}\frac{\frac{1}{2+h}-\frac{1}{2}}{h} =\lim_{h\to0}\frac{\frac{2-(2+h)}{2(2+h)}}{h} =\lim_{h\to0}\frac{-h}{2h(2+h)} =\lim_{h\to0}\frac{-1}{2(2+h)}=-\frac14.

Two features of this trace generalize. First, the same cancellation pattern appears whenever the numerator has a factor hh the algebra exposes the removable discontinuity rather than hiding it. Second, the computation is valid only at points where gg is defined on a two-sided neighbourhood. At x=0x=0 the function is undefined, and no difference quotient exists, so g′(0)g'(0) is not a number to be found—it is a question with no domain.

Worked trace: a continuous function that fails differentiability. Let

f(x)={xsin⁡(1/x),x≠0,0,x=0.f(x)=\begin{cases}x\sin(1/x), & x\ne 0,\\ 0, & x=0.\end{cases}

Continuity at 00 follows from ∣xsin⁡(1/x)∣≤∣x∣|x\sin(1/x)|\le |x| The difference quotient is

f(h)−f(0)h=sin⁡(1/h).\frac{f(h)-f(0)}{h}=\sin(1/h).

As h→0h\to0 sin⁡(1/h)\sin(1/h) oscillates between −1-1 and 11 and has no limit. The function is continuous at 00 but not differentiable there. The counterexample shows that the converse of Theorem 1.2.2 is false and that a derivative can fail to exist even when the graph has no jump or corner in the usual pictorial sense.

Verification Check

0 / 4 answered

Consider f(x)=x2sin⁡(1/x)f(x)=x^2\sin(1/x) for x≠0x\ne0, with f(0)=0f(0)=0. Evaluate each statement regarding differentiability and derivative continuity at x=0x=0:

Classify each option as correct or incorrect, or leave uncertain options unanswered (“I don't know”) and verify at any time.

1
The function fails to be differentiable at x=0x = 0 because sin⁡(1/x)\sin(1/x) oscillates rapidly without limit.
2
The function is differentiable at x=0x = 0 with derivative value f′(0)=0f'(0) = 0 by the squeeze theorem.
3
The derivative f′(x)f'(x) exists everywhere but fails to be continuous at x=0x = 0 due to oscillation.
4
Differentiability at a point implies that the derivative function must be continuous at that point.

Fundamental Rules of Differentiation

Pause & Predict: Tangent Slopes Along a Cubic Curve. In the interactive playground below, you can move the point of tangency aa along the curve f(x)=0.16x3−0.5xf(x) = 0.16x^3 - 0.5x At what values of aa does the tangent line become horizontal f′(a)=0f'(a) = 0 Where is the slope positive f′(a)>0f'(a) > 0 and where is it negative f′(a)<0f'(a) < 0 Predict the coordinates of the horizontal tangents before adjusting the slider.

DerivationExpand

Differentiating f(x)=0.16x3−0.5xf(x) = 0.16x^3 - 0.5x using the power and sum rules: f′(x)=3(0.16)x2−0.5=0.48x2−0.5.f'(x) = 3(0.16)x^2 - 0.5 = 0.48x^2 - 0.5. Setting f′(a)=0f'(a) = 0 0.48a2−0.5=0  ⟹  a2=0.50.48≈1.0417  ⟹  a≈±1.02.0.48a^2 - 0.5 = 0 \implies a^2 = \frac{0.5}{0.48} \approx 1.0417 \implies a \approx \pm 1.02.

  • For ∣a∣>1.02|a| > 1.02 f′(a)>0f'(a) > 0 (the curve is strictly increasing).
  • For ∣a∣<1.02|a| < 1.02 f′(a)<0f'(a) < 0 (the curve is strictly decreasing).
  • At a≈−1.02a \approx -1.02 (local maximum) and a≈1.02a \approx 1.02 (local minimum), the tangent line is horizontal.
Interactive exploration

Slide the point of tangency

Move a along the curve and compare the tangent slope with the local shape.

a = 1.10
secant limit → tangenta = 1.10, f′(a) = 0.08
Slide the point of tangency
secant limit → tangenta = 1.10, f′(a) = 0.08

A tangent line records the limiting slope at a selected point.

To compute derivatives efficiently, we establish algebraic rules derived from the definition.

Theorem 1.2.3Differentiation Rules

Builds on:Definition 1.2.1

Let ff and gg be differentiable at xx

  1. Sum Rule: (f+g)′(x)=f′(x)+g′(x)(f + g)'(x) = f'(x) + g'(x)
  2. Product Rule: (fg)′(x)=f′(x)g(x)+f(x)g′(x)(fg)'(x) = f'(x)g(x) + f(x)g'(x)
  3. Quotient Rule: If g(x)≠0g(x) \neq 0 (fg)′(x)=f′(x)g(x)−f(x)g′(x)g(x)2\left(\frac{f}{g}\right)'(x) = \frac{f'(x)g(x) - f(x)g'(x)}{g(x)^2}
ProofExpand proof

We prove the Product Rule. Let F(x)=f(x)g(x)F(x) = f(x)g(x) Using the definition:

F′(x)=lim⁡h→0f(x+h)g(x+h)−f(x)g(x)h.F'(x) = \lim_{h \to 0} \frac{f(x+h)g(x+h) - f(x)g(x)}{h}.

Subtract and add the term f(x)g(x+h)f(x)g(x+h) in the numerator:

F′(x)=lim⁡h→0f(x+h)g(x+h)−f(x)g(x+h)+f(x)g(x+h)−f(x)g(x)hF'(x) = \lim_{h \to 0} \frac{f(x+h)g(x+h) - f(x)g(x+h) + f(x)g(x+h) - f(x)g(x)}{h} =lim⁡h→0(f(x+h)−f(x)h⋅g(x+h)+f(x)⋅g(x+h)−g(x)h).= \lim_{h \to 0} \left( \frac{f(x+h) - f(x)}{h} \cdot g(x+h) + f(x) \cdot \frac{g(x+h) - g(x)}{h} \right).

Since ff and gg are differentiable, their individual limits exist. Since gg is differentiable, it is continuous, so lim⁡h→0g(x+h)=g(x)\lim_{h \to 0} g(x+h) = g(x) Applying limit laws:

F′(x)=f′(x)g(x)+f(x)g′(x).F'(x) = f'(x)g(x) + f(x)g'(x).

∎

The Chain Rule

The Chain Rule allows us to differentiate composite functions. It is the mathematical engine behind backpropagation in deep neural networks.

Theorem 1.2.4The Chain Rule

Builds on:Definition 1.2.1

If gg is differentiable at xx and ff is differentiable at g(x)g(x) then the composite function y=(f∘g)(x)=f(g(x))y = (f \circ g)(x) = f(g(x)) is differentiable at xx and:

dydx=f′(g(x))⋅g′(x).\frac{dy}{dx} = f'(g(x)) \cdot g'(x).

Linear Approximation

The derivative provides the best linear approximation of a function near a point. Geometrically, this is the tangent line equation.

f(x)≈f(a)+f′(a)(x−a).f(x) \approx f(a) + f'(a)(x - a).

For example, linearising f(x)=xf(x) = \sqrt{x} at a=4a = 4 gives L(x)=2+14(x−4)L(x) = 2 + \frac{1}{4}(x - 4) which closely tracks x\sqrt{x} near x=4x = 4 while gradually diverging further away:

Linear approximation $L(x) = 2 + \frac{x-4}{4}$ to $\sqrt{x}$ at $a = 4$: at $x = 4.4$, $L(4.4) = 2.100$ exceeds $\sqrt{4.4} \approx 2.098$ by $0.0024$. Linear approximation of the square root function near x = 4, showing the tangent line L(x) and the small truncation error at x = 4.4. a=4 x=4.4
Figure 1.2.3: Linear approximation L(x)=2+x−44L(x) = 2 + \frac{x-4}{4} to x\sqrt{x} at a=4a = 4: at x=4.4x = 4.4, L(4.4)=2.100L(4.4) = 2.100 exceeds 4.4≈2.098\sqrt{4.4} \approx 2.098 by 0.00240.0024.

Let us visualize how the linear approximation matches f(x)=sin⁡(x)f(x) = \sin(x) around a=1.0a = 1.0 and verify the error increases as we move away from aa

python

Interactive Lab

1import numpy as np
2import matplotlib.pyplot as plt
3 
4def f(x): return np.sin(x)
5def df(x): return np.cos(x)
6 
7a = 1.0
8x = np.linspace(0.0, 2.0, 200)
9l_approx = f(a) + df(a) * (x - a)
10 
11plt.plot(x, f(x), label='f(x) = sin(x)', color='#3b82f6', linewidth=2)
12plt.plot(x, l_approx, '--', label='Linear Approx at x=1', color='r')
13plt.scatter([a], [f(a)], color='black', zorder=5, label='Point of Tangency')
14plt.legend()
15plt.title("Tangent Line (Linear Approximation)")
16plt.grid(True)
17plt.show()
18 
19# Compute error at x = 1.1 and x = 1.8
20print(f"Error at x = 1.1: {abs(f(1.1) - (f(a) + df(a)*(1.1-a))):.6f}")
21print(f"Error at x = 1.8: {abs(f(1.8) - (f(a) + df(a)*(1.8-a))):.6f}")
22 

Implicit Differentiation

When variables are defined implicitly (e.g., x2+y2=1x^2 + y^2 = 1 we differentiate both sides with respect to xx applying the Chain Rule to yy (treating yy locally as a differentiable function of xx

For the unit circle x2+y2=1x^2 + y^2 = 1 differentiating yields:

2x+2y⋅dydx=0  ⟹  dydx=−xy(y≠0).2x + 2y \cdot \frac{dy}{dx} = 0 \implies \frac{dy}{dx} = -\frac{x}{y} \quad (y \neq 0).

Worked trace: Folium of Descartes (x3+y3=6xyx^3 + y^3 = 6xy at (3,3)(3, 3))Expand
  1. Differentiate both sides with respect to xx using the Product Rule on 6xy6xy 3x2+3y2dydx=6y+6xdydx.3x^2 + 3y^2 \frac{dy}{dx} = 6y + 6x \frac{dy}{dx}.
  2. Collect dydx\frac{dy}{dx} terms: (3y2−6x)dydx=6y−3x2  ⟹  dydx=2y−x2y2−2x.(3y^2 - 6x)\frac{dy}{dx} = 6y - 3x^2 \implies \frac{dy}{dx} = \frac{2y - x^2}{y^2 - 2x}.
  3. Evaluate at (3,3)(3, 3) dydx∣(3,3)=2(3)−3232−2(3)=6−99−6=−1.\frac{dy}{dx}\Big|_{(3,3)} = \frac{2(3) - 3^2}{3^2 - 2(3)} = \frac{6 - 9}{9 - 6} = -1.
Section Exercises

Deep Practice

Apply differentiation rules, evaluate differentiability contracts, and explore the boundary between derivative existence and derivative continuity.

Exercise 1.2.1: Derivative from First Principles

Verificationapplication

Using Definition 1.2.1, compute the derivative of f(x)=1xf(x) = \frac{1}{\sqrt{x}} at an arbitrary point x>0x > 0

HintExpand

Form the difference quotient 1/x+h−1/xh\frac{1/\sqrt{x+h} - 1/\sqrt{x}}{h}, find a common denominator, and rationalize the numerator.

Worked solutionExpand
  1. Form the difference quotient: 1x+h−1xh=x−x+hhxx+h.\frac{\frac{1}{\sqrt{x+h}} - \frac{1}{\sqrt{x}}}{h} = \frac{\sqrt{x} - \sqrt{x+h}}{h \sqrt{x}\sqrt{x+h}}.
  2. Multiply numerator and denominator by the conjugate (x+x+h)(\sqrt{x} + \sqrt{x+h}) (x−x+h)(x+x+h)hxx+h(x+x+h)=x−(x+h)hxx+h(x+x+h)=−hhxx+h(x+x+h).\frac{(\sqrt{x} - \sqrt{x+h})(\sqrt{x} + \sqrt{x+h})}{h \sqrt{x}\sqrt{x+h}(\sqrt{x} + \sqrt{x+h})} = \frac{x - (x+h)}{h \sqrt{x}\sqrt{x+h}(\sqrt{x} + \sqrt{x+h})} = \frac{-h}{h \sqrt{x}\sqrt{x+h}(\sqrt{x} + \sqrt{x+h})}.
  3. Cancel h≠0h \ne 0 −1xx+h(x+x+h).\frac{-1}{\sqrt{x}\sqrt{x+h}(\sqrt{x} + \sqrt{x+h})}.
  4. Take the limit as h→0h \to 0 using limit laws: f′(x)=−1xx(x+x)=−1x(2x)=−12x3/2.f'(x) = \frac{-1}{\sqrt{x}\sqrt{x}(\sqrt{x} + \sqrt{x})} = \frac{-1}{x(2\sqrt{x})} = -\frac{1}{2x^{3/2}}.

Verification criteria. Requires: (a) explicit difference quotient; (b) algebraic conjugate rationalization; (c) valid cancellation of hh for h≠0h \ne 0 (d) correct final algebraic term −1/(2x3/2)-1/(2x^{3/2})

Exercise 1.2.2: Continuity Without Differentiability

Negationrecall

Show that f(x)=∣x∣f(x) = |x| is continuous at x=0x = 0 by Theorem 1.2.2‘s converse, but not differentiable at x=0x = 0 by Definition 1.2.1.

HintExpand

Evaluate the left-hand and right-hand limits of the difference quotient (f(0+h)−f(0))/h(f(0+h) - f(0))/h.

Worked solutionExpand
Show corner diagram(Click to reveal)
Non-differentiability at a corner: left secant slope is $-1$ while right secant slope is $+1$, so no tangent exists at $x = 0$. Graph of the absolute value function f(x) = |x| at the sharp corner x = 0 with opposing left and right difference quotients.
Figure 1.2.2: Non-differentiability at a corner: left secant slope is −1-1 while right secant slope is +1+1, so no tangent exists at x=0x = 0.

Continuity: f(0)=0f(0) = 0 Since ∣f(x)−f(0)∣=∣∣x∣−0∣=∣x∣<ϵ|f(x) - f(0)| = ||x| - 0| = |x| < \epsilon whenever ∣x−0∣<δ=ϵ|x - 0| < \delta = \epsilon lim⁡x→0f(x)=f(0)=0\lim_{x \to 0} f(x) = f(0) = 0 Thus ff is continuous at 00

Non-Differentiability: The difference quotient at c=0c = 0 is: f(0+h)−f(0)h=∣h∣−0h=∣h∣h.\frac{f(0+h) - f(0)}{h} = \frac{|h| - 0}{h} = \frac{|h|}{h}. As h→0+h \to 0^+ ∣h∣=h|h| = h so lim⁡h→0+hh=1\lim_{h \to 0^+} \frac{h}{h} = 1 As h→0−h \to 0^- ∣h∣=−h|h| = -h so lim⁡h→0−−hh=−1\lim_{h \to 0^-} \frac{-h}{h} = -1 Since the one-sided limits disagree 1≠−11 \ne -1 the two-sided limit does not exist. Hence ff is not differentiable at 00 This confirms the converse of Theorem 1.2.2 is false.

Verification criteria. Requires computing both one-sided difference quotient limits and concluding divergence.

Exercise 1.2.3: Product and Quotient Rules on Algebraic Forms

Constructionapplication

Using Theorem 1.2.3, find f′(x)f'(x) for f(x)=x2sin⁡xx2+1f(x) = \frac{x^2 \sin x}{x^2 + 1}

HintExpand

Apply the Quotient Rule with numerator u(x)=x2sin⁡xu(x) = x^2 \sin x (which requires the Product Rule) and denominator v(x)=x2+1v(x) = x^2 + 1.

Worked solutionExpand
  1. Let u(x)=x2sin⁡xu(x) = x^2 \sin x By the Product Rule: u′(x)=(x2)′sin⁡x+x2(sin⁡x)′=2xsin⁡x+x2cos⁡x.u'(x) = (x^2)' \sin x + x^2 (\sin x)' = 2x \sin x + x^2 \cos x.
  2. Let v(x)=x2+1v(x) = x^2 + 1 Its derivative is v′(x)=2xv'(x) = 2x
  3. Applying the Quotient Rule: f′(x)=u′(x)v(x)−u(x)v′(x)v(x)2=(2xsin⁡x+x2cos⁡x)(x2+1)−(x2sin⁡x)(2x)(x2+1)2.f'(x) = \frac{u'(x)v(x) - u(x)v'(x)}{v(x)^2} = \frac{(2x \sin x + x^2 \cos x)(x^2 + 1) - (x^2 \sin x)(2x)}{(x^2 + 1)^2}.
  4. Expanding the numerator: (2x3sin⁡x+2xsin⁡x+x4cos⁡x+x2cos⁡x)−2x3sin⁡x=x2(x2+1)cos⁡x+2xsin⁡x.(2x^3 \sin x + 2x \sin x + x^4 \cos x + x^2 \cos x) - 2x^3 \sin x = x^2(x^2 + 1)\cos x + 2x \sin x. Thus: f′(x)=x(x(x2+1)cos⁡x+2sin⁡x)(x2+1)2.f'(x) = \frac{x(x(x^2+1)\cos x + 2\sin x)}{(x^2 + 1)^2}.

Verification criteria. A complete solution requires correct composition of product and quotient rules and simplification.

Exercise 1.2.4: Linear Approximation Error Bound

Constructionapplication

Let f(x)=xf(x) = \sqrt{x} Find the linear approximation L(x)L(x) of ff at a=25a = 25 Use L(x)L(x) to estimate 26\sqrt{26} and show without a calculator that the error is less than 1/5001/500

HintExpand

Recall L(x)=f(a)+f′(a)(x−a)L(x) = f(a) + f'(a)(x - a). For f(x)=xf(x) = \sqrt{x}, f′(25)=1/(225)=1/10f'(25) = 1/(2\sqrt{25}) = 1/10.

Worked solutionExpand
Show linearisation diagram(Click to reveal)
Linear approximation $L(x) = 2 + \frac{x-4}{4}$ to $\sqrt{x}$ at $a = 4$: at $x = 4.4$, $L(4.4) = 2.100$ exceeds $\sqrt{4.4} \approx 2.098$ by $0.0024$. Linear approximation of the square root function near x = 4, showing the tangent line L(x) and the small truncation error at x = 4.4. a=4 x=4.4
Figure 1.2.3: Linear approximation L(x)=2+x−44L(x) = 2 + \frac{x-4}{4} to x\sqrt{x} at a=4a = 4: at x=4.4x = 4.4, L(4.4)=2.100L(4.4) = 2.100 exceeds 4.4≈2.098\sqrt{4.4} \approx 2.098 by 0.00240.0024.
  1. At a=25a = 25 f(25)=5f(25) = 5 and f′(x)=12xf'(x) = \frac{1}{2\sqrt{x}} so f′(25)=110=0.1f'(25) = \frac{1}{10} = 0.1
  2. The linear approximation is: L(x)=5+110(x−25).L(x) = 5 + \frac{1}{10}(x - 25).
  3. At x=26x = 26 L(26)=5+110(26−25)=5+0.1=5.1.L(26) = 5 + \frac{1}{10}(26 - 25) = 5 + 0.1 = 5.1.
  4. Error analysis: Squaring 5.15.1 gives (5.1)2=26.01(5.1)^2 = 26.01 Since 26<26.0126 < 26.01 26<5.1\sqrt{26} < 5.1 Also (5.099)2=(5.1−0.001)2=26.01−0.0102+0.000001=25.999801<26(5.099)^2 = (5.1 - 0.001)^2 = 26.01 - 0.0102 + 0.000001 = 25.999801 < 26 Thus 5.099<26<5.1005.099 < \sqrt{26} < 5.100 confirming ∣26−5.1∣<0.001<1500=0.002|\sqrt{26} - 5.1| < 0.001 < \frac{1}{500} = 0.002

Verification criteria. Requires the explicit tangent line equation, the estimate 5.15.1 and the rigorous square comparison verifying the error bound.

Exercise 1.2.5: Implicit Differentiation and Normal Slopes

Constructionapplication

Consider the Folium of Descartes x3+y3=6xyx^3 + y^3 = 6xy Find the equations of all horizontal and vertical tangent lines on the curve where x≠0x \ne 0 and y≠0y \ne 0

HintExpand

From the worked trace, dydx=2y−x2y2−2x\frac{dy}{dx} = \frac{2y - x^2}{y^2 - 2x}. Horizontal tangents have numerator =0= 0; vertical tangents have denominator =0= 0.

Worked solutionExpand
  1. Horizontal Tangents: dydx=0  ⟹  2y−x2=0  ⟹  y=x22\frac{dy}{dx} = 0 \implies 2y - x^2 = 0 \implies y = \frac{x^2}{2} Substitute into x3+y3=6xyx^3 + y^3 = 6xy x3+(x22)3=6x(x22)  ⟹  x3+x68=3x3  ⟹  x68=2x3.x^3 + \left(\frac{x^2}{2}\right)^3 = 6x\left(\frac{x^2}{2}\right) \implies x^3 + \frac{x^6}{8} = 3x^3 \implies \frac{x^6}{8} = 2x^3. Since x≠0x \ne 0 divide by x3x^3 x3=16  ⟹  x=24/3=223.x^3 = 16 \implies x = 2^{4/3} = 2\sqrt[3]{2}. Then y=x22=28/32=25/3=243=42/3y = \frac{x^2}{2} = \frac{2^{8/3}}{2} = 2^{5/3} = 2\sqrt[3]{4} = 4^{2/3} The horizontal tangent line is y=243y = 2\sqrt[3]{4}

  2. Vertical Tangents: By symmetry (swapping xx and yy leaves x3+y3=6xyx^3 + y^3 = 6xy unchanged), the denominator vanishes when y2−2x=0y^2 - 2x = 0 giving the vertical tangent at x=243x = 2\sqrt[3]{4}

Verification criteria. Requires setting numerator/denominator to zero, solving with the curve equation, and identifying the coordinates.

Exercise 1.2.6: Derivation of the Inverse Function Rule

Generalisationsynthesis

Let ff be strictly monotonic and differentiable on an interval II with f′(x)≠0f'(x) \ne 0 Let g=f−1g = f^{-1} be its inverse. Use Theorem 1.2.4 on the identity f(g(y))=yf(g(y)) = y to prove that g′(y)=1f′(g(y))g'(y) = \frac{1}{f'(g(y))}

HintExpand

Differentiate both sides of f(g(y))=yf(g(y)) = y with respect to yy, assuming gg is differentiable.

Worked solutionExpand
  1. The inverse function identity states: f(g(y))=yfor all y in the range of f.f(g(y)) = y \quad \text{for all } y \text{ in the range of } f.
  2. Differentiate both sides with respect to yy applying the Chain Rule (Theorem 1.2.4) to the composite left-hand side: ddy[f(g(y))]=ddy[y].\frac{d}{dy}[f(g(y))] = \frac{d}{dy}[y].
  3. By Theorem 1.2.4: f′(g(y))⋅g′(y)=1.f'(g(y)) \cdot g'(y) = 1.
  4. Since f′(x)≠0f'(x) \ne 0 for all xx f′(g(y))≠0f'(g(y)) \ne 0 Dividing gives: g′(y)=1f′(g(y)).g'(y) = \frac{1}{f'(g(y))}. Equivalently, setting y=f(x)y = f(x) so g(y)=xg(y) = x (f−1)′(f(x))=1f′(x).(f^{-1})'(f(x)) = \frac{1}{f'(x)}.

Verification criteria. A complete derivation requires correct application of chain rule to identity and explicitly noting f′(g(y))≠0f'(g(y)) \ne 0

References & Further Reading

  • Michael Spivak (2008). Calculus (4th ed.), Chapter 9 (“Derivatives”, pp. 143–162) and Chapter 10 (“Differentiation”, pp. 165–181); for the construction of a continuous nowhere-differentiable function, see Chapter 24. Publish or Perish.
  • Tom M. Apostol (1967). Calculus, Volume 1: One-Variable Calculus with an Introduction to Linear Algebra (2nd ed.), Chapter 4 (“Differential Calculus”, pp. 155–182). John Wiley & Sons.
  • Walter Rudin (1976). Principles of Mathematical Analysis (3rd ed.), Chapter 5 (“Differentiation”, pp. 103–118), and Chapter 9 (“Functions of Several Variables”) for the total derivative Df(c)Df(c) on Rn\mathbb{R}^n McGraw-Hill.