The ε–δ definition, now as a tool you prove with, not just apply. In analysis every limit law is a theorem with a proof — this is where you learn to build those proofs yourself, including the standard tricks every analyst carries around.
For two centuries calculus rested on infinitesimals nobody could define. The ε–δ limit answered "what does approaches mean?" and made the whole subject rigorous.
Every instantaneous rate is a limit, and the convergence of every numerical algorithm is a limit. The ε–δ habit of "given a target tolerance, find the needed precision" is exactly how error analysis works.
Berkeley mocked infinitesimals in 1734 as "the ghosts of departed quantities." Cauchy and Weierstrass made ε–δ rigorous a century later; Weierstrass then built a function continuous everywhere yet differentiable nowhere.
$\displaystyle\lim_{x\to a}f(x)=L$ means: for every $\varepsilon>0$ there exists $\delta>0$ such that, for all $x$, $$0<|x-a|<\delta\ \Longrightarrow\ |f(x)-L|<\varepsilon.$$
Read it as a challenge–response game: the challenger picks a tolerance $\varepsilon$ around $L$, and you must produce a radius $\delta$ around $a$ that forces every $x$ (except possibly $a$ itself) to land within $\varepsilon$ of $L$. The "$0<|x-a|$" clause makes the value $f(a)$ — even its existence — irrelevant. Negation, for proving a limit is not $L$: there is an $\varepsilon>0$ such that every $\delta>0$ admits some $x$ with $0<|x-a|<\delta$ but $|f(x)-L|\ge\varepsilon$.
Prove $\lim_{x\to2}x^2=4$.
Prove $\lim_{x\to3}\dfrac1x=\dfrac13$.
A function has at most one limit at $a$: if $\lim_{x\to a}f=L$ and $\lim_{x\to a}f=M$, then $L=M$.
If $\lim_{x\to a}f=L$ and $\lim_{x\to a}g=M$, then $\lim(f+g)=L+M$, $\lim(fg)=LM$, and if $M\ne0$, $\lim\dfrac fg=\dfrac LM$.
If $g(x)\le f(x)\le h(x)$ for all $x$ near $a$ (except possibly $a$), and $\lim_{x\to a}g=\lim_{x\to a}h=L$, then $\lim_{x\to a}f=L$.
$\lim_{x\to0}x^2\sin(1/x)=0$: since $|\sin|\le1$, $-x^2\le x^2\sin(1/x)\le x^2$, and both bounds tend to $0$.
$\lim_{x\to0}\sin(1/x)$ does not exist.
$\lim_{x\to a^+}f=L$ uses $a<x<a+\delta$ in place of $0<|x-a|<\delta$ (similarly from the left); the two-sided limit exists iff both one-sided limits exist and agree. $\lim_{x\to\infty}f=L$ means: for every $\varepsilon>0$ there is $N$ with $x>N\Rightarrow|f(x)-L|<\varepsilon$. All the laws above carry over word for word.
• $\delta$ may depend on $\varepsilon$ and $a$ — never on $x$.
• Bound the uncontrolled factor (like $|x+2|$) by restricting $\delta\le1$
before solving for $\delta$.
• The limit ignores $f(a)$ entirely — that is the role of $0<|x-a|$.
• To disprove a limit, you get to pick $\varepsilon$; the opponent picks
$\delta$; you must beat every $\delta$.
Original problems written for this guide in the style of Spivak's exercises. Write a full solution before opening the answer.
Prove $\lim_{x\to1}(3x+2)=5$ from the definition.
Prove $\lim_{x\to2}(x^2+x)=6$ from the definition.
Find $\lim_{x\to0}x\cos(1/x)$ with justification.
Prove $\lim_{x\to9}\sqrt{x}=3$ from the definition.
Write out the full proof that a limit, if it exists, is unique.
Prove the product law $\lim(fg)=LM$ in full.
Prove: if $\lim_{x\to a}g=M\ne0$, then $\lim_{x\to a}\dfrac1{g}= \dfrac1M$.
Prove that $\lim_{x\to0}\dfrac{|x|}{x}$ does not exist.
Course coverage. This unit is part of the complete MAT158H1 topic map. It should cover the following items before you move on:
Use the current UofT MAT158H1 calendar topics as the coverage spine, and pair lecture notes with a Spivak-style proof-calculus text for extra reading. The exact assigned textbook should still be checked against the instructor syllabus.
The limit spine is: quantify closeness in the range, choose closeness in the domain, then verify implication under punctured-neighborhood hypotheses.
Use these after the main practice set. They are designed to force definition-level reasoning, theorem selection, and mixed application rather than routine calculation only.
Prove $\lim_{x o1}(2x+3)=5$.
State the main bounding idea behind proving the product law for limits.
Use sequences to show $\lim_{x o0}\sin(1/x)$ does not exist.
Why does $\lim_{x o0} |x|/x$ not exist?