Continuity and Differentiability
Continuity
Informally, a function is continuous if you can draw its graph without lifting your pen. The precise definition pins this idea to limits.
Continuity at a point
A function $f$ is continuous at $x=c$ if all three conditions hold:
- $f(c)$ is defined,
- $\displaystyle\lim_{x\to c} f(x)$ exists (left-hand limit $=$ right-hand limit), and
- $\displaystyle\lim_{x\to c} f(x)=f(c)$.
If any one fails, $f$ is discontinuous at $c$. For piecewise functions, the usual exam task is to match the left-hand limit, right-hand limit and function value at the join.
Continuity on an interval
$f$ is continuous on an interval if it is continuous at every point of it. Polynomials, $\sin x$, $\cos x$, $e^{x}$ are continuous everywhere; $\tan x$ is continuous except where $\cos x=0$.
Algebra of continuous functions
If $f$ and $g$ are continuous at $c$, so are $f\pm g$, $fg$, and $\dfrac{f}{g}$ (provided $g(c)\ne0$). Compositions of continuous functions are continuous. This lets you certify most expressions as continuous without epsilon-delta work.
$f(2)=2+1=3$. LHL $=\lim_{x\to2^-}(x+1)=3$. RHL $=\lim_{x\to2^+}(2x-1)=3$. All three agree, so $f$ is continuous at $x=2$.
RHL $=4$ and $f(1)=k$. LHL $=\lim_{x\to1^-}kx^2=k$. For continuity $k=4$.
It is a quotient of continuous functions, continuous wherever the denominator is non-zero, i.e. for all $x\ne3$. At $x=3$ the function is undefined, so it is discontinuous only there.
$f(0)=0$. LHL $=\lim_{x\to0^-}(-x)=0$, RHL $=\lim_{x\to0^+}(x)=0$. All equal $0$, so $|x|$ is continuous at $0$ (though, as the next page shows, not differentiable there).
- Continuous at $c$ $\iff$ $f(c)$ defined, $\lim_{x\to c}f(x)$ exists, and the two are equal.
- For piecewise functions, equate LHL, RHL and $f(c)$ at each join.
- Polynomials, $\sin,\cos,e^{x}$ are continuous everywhere; rational functions wherever the denominator is non-zero.
- Sums, products, quotients (non-zero denom) and compositions of continuous functions are continuous.
Differentiability and the Chain Rule
The derivative of $f$ at $x=c$ is the limit of the average rate of change:
$$f'(c)=\lim_{h\to 0}\frac{f(c+h)-f(c)}{h},$$
provided this limit exists. Geometrically it is the slope of the tangent at $\big(c,f(c)\big)$.
Differentiability implies continuity
If $f$ is differentiable at $c$, then it is continuous at $c$. The converse is false: $f(x)=|x|$ is continuous at $0$ but has a corner there, so $f'(0)$ does not exist (left slope $-1$, right slope $+1$). Corners and sharp points break differentiability.
Standard derivatives
| $f(x)$ | $f'(x)$ | $f(x)$ | $f'(x)$ |
|---|---|---|---|
| $x^{n}$ | $nx^{n-1}$ | $\sin x$ | $\cos x$ |
| $\cos x$ | $-\sin x$ | $\tan x$ | $\sec^2 x$ |
| $e^{x}$ | $e^{x}$ | $\ln x$ | $\tfrac1x$ |
The chain rule
For a composite $y=f(g(x))$,
$$\frac{dy}{dx}=f'\big(g(x)\big)\cdot g'(x).$$
Differentiate the outer function, keep the inside, then multiply by the derivative of the inside. Combined with the product rule $(uv)'=u'v+uv'$ and quotient rule, this handles almost every function in the chapter.
Chain rule with outer $u^5$, inner $u=3x^2+1$: $\dfrac{dy}{dx}=5(3x^2+1)^4\cdot 6x = 30x(3x^2+1)^4.$
$\dfrac{dy}{dx}=\cos(x^2)\cdot 2x = 2x\cos(x^2).$
Product rule: $\dfrac{dy}{dx}=2x\,e^{x}+x^2 e^{x}=e^{x}(x^2+2x).$
The left-hand derivative is $\lim_{h\to0^-}\dfrac{|h|}{h}=-1$ and the right-hand derivative is $\lim_{h\to0^+}\dfrac{|h|}{h}=+1$. Since $-1\ne+1$, $f'(0)$ does not exist.
- $f'(c)=\lim_{h\to0}\dfrac{f(c+h)-f(c)}{h}$ = slope of the tangent.
- Differentiable $\Rightarrow$ continuous; continuous does not imply differentiable (e.g. $|x|$ at $0$).
- Chain rule: $\dfrac{d}{dx}f(g(x))=f'(g(x))\,g'(x)$.
- Product rule $(uv)'=u'v+uv'$; quotient rule $\left(\tfrac uv\right)'=\dfrac{u'v-uv'}{v^2}$.
Logarithmic, Implicit and Parametric
Some relations are not given as $y=f(x)$. Three techniques extend differentiation to these cases.
Implicit differentiation
When $x$ and $y$ are tangled in one equation, differentiate both sides with respect to $x$, treating $y$ as a function of $x$ (so $\tfrac{d}{dx}y^n = ny^{n-1}\tfrac{dy}{dx}$), then solve for $\tfrac{dy}{dx}$.
Logarithmic differentiation
For products/quotients of many factors, or for a variable raised to a variable power such as $y=x^{x}$, take $\ln$ of both sides first. This converts powers into products and products into sums, which are far easier to differentiate. For $y=u(x)^{v(x)}$ this is the only elementary route.
Parametric differentiation
If $x=f(t)$ and $y=g(t)$, then
$$\frac{dy}{dx}=\frac{dy/dt}{dx/dt}=\frac{g'(t)}{f'(t)},\qquad f'(t)\ne0.$$
Second-order derivatives
Differentiating $\tfrac{dy}{dx}$ again gives $\tfrac{d^2y}{dx^2}$, the rate of change of the slope — needed for concavity and the second-derivative test in the next chapter.
Differentiate implicitly: $2x+2y\dfrac{dy}{dx}=0 \Rightarrow \dfrac{dy}{dx}=-\dfrac{x}{y}.$
Take logs: $\ln y=x\ln x$. Differentiate: $\dfrac{1}{y}\dfrac{dy}{dx}=\ln x+1$. So $\dfrac{dy}{dx}=x^{x}(1+\ln x).$
$\dfrac{dx}{dt}=2at,\ \dfrac{dy}{dt}=2a$. Hence $\dfrac{dy}{dx}=\dfrac{2a}{2at}=\dfrac{1}{t}.$
$\dfrac{dy}{dx}=3x^2$, then $\dfrac{d^2y}{dx^2}=6x.$
- Implicit: differentiate both sides w.r.t. $x$ (use chain rule on $y$), then solve for $\tfrac{dy}{dx}$.
- Logarithmic differentiation is essential for $y=u(x)^{v(x)}$; take $\ln$ first.
- Parametric: $\dfrac{dy}{dx}=\dfrac{dy/dt}{dx/dt}$.
- Second derivative $\tfrac{d^2y}{dx^2}$ is the derivative of the first derivative.