Intuition What this page is for
The parent note gave you the rules. Here we run them through every situation the topic can throw at you — every sign, every degree comparison, the sneaky 0 0 hole, the x 2 sign flip, a slant asymptote, a word problem, and an exam trap.
The plan: first a matrix listing every case class, then one worked example per class. When you finish, no exam question can surprise you, because you'll have seen the cell it lives in.
Definition "Degree" (deg) of a polynomial
The degree of a polynomial is ==the highest power of x that appears with a nonzero coefficient==. For example 2 x 3 − x + 5 has degree 3 (the x 3 term wins), and a constant like 7 has degree 0 . Throughout this page we write deg top for the degree of the numerator and deg bottom for the degree of the denominator. The whole "limit at infinity" story is decided by comparing these two numbers , because the highest power is the term that dominates when x gets huge.
Every problem in this topic is one of these cells. The last column says which worked example below hits it.
#
Case class
What makes it distinct
Answer type
Example
A
Nonzero ÷ zero, sides disagree
left limit and right limit have opposite signs
VA, limit DNE
Ex 1
B
Nonzero ÷ zero, sides agree (squared denom)
denominator → 0 + both sides
VA, ± ∞ both sides
Ex 2
C
Zero ÷ zero — removable
top and bottom share a factor
hole, NOT a VA
Ex 3
D
Rational at infinity, deg top < deg bottom
numerator dies faster
HA y = 0
Ex 4
E
Rational at infinity, deg top = deg bottom
leading terms survive
HA = ratio of leaders
Ex 4
F
Rational at infinity, deg top > deg bottom
numerator wins
no HA (± ∞ )
Ex 4
G
Root + sign care , x → − ∞
$\sqrt{x^2}=
x
=-x$
H
Transcendental (e − x , arctan )
not rational, still settles
HA from known limits
Ex 6
I
Word problem (real-world settling)
asymptote = long-run value
HA with units
Ex 7
J
Exam twist — VA cancels but another survives
mixed hole + real asymptote
one hole, one VA
Ex 8
K
Slant asymptote (deg top = deg bottom + 1 )
numerator exactly one degree higher
oblique line y = m x + b
Ex 9
We'll walk them in order.
x → 1 lim x − 1 x + 2 , and both one-sided limits.
Forecast: Numerator at x = 1 is 3 (nonzero), denominator → 0 . So it blows up — but does it go up or down? Guess the sign on each side before reading on.
Step 1 — Check the numerator's sign near x = 1 . Why this step? If the top were also heading to 0 we'd have a 0 0 hole (cell C) and could NOT conclude an asymptote. Here x + 2 → 3 > 0 , safely nonzero and positive. Good — this is a genuine "nonzero ÷ zero" (cell A/B).
Step 2 — Right side x → 1 + . Why? The sign of the answer is decided entirely by the denominator's sign, and that sign differs left vs right. Take x = 1.001 : x − 1 = 0.001 > 0 , a tiny positive . So 0 + + 3 = + ∞ .
Step 3 — Left side x → 1 − . Take x = 0.999 : x − 1 = − 0.001 < 0 , a tiny negative . So 0 − + 3 = − ∞ .
Step 4 — Combine. The two sides disagree (+ ∞ vs − ∞ ), so lim x → 1 x − 1 x + 2 does not exist . But x = 1 is still a vertical asymptote (a one-sided infinite limit is enough).
Verify: Plug x = 1.0001 : 0.0001 3.0001 = 30001 — huge positive ✓. Plug x = 0.9999 : − 0.0001 2.9999 = − 29999 — huge negative ✓. Signs match Steps 2–3.
The figure below plots this function: watch the cyan curve rocket up just right of the amber line x = 1 and plunge down just left of it — that visible disagreement is exactly the "+ ∞ vs − ∞ " of Steps 2–3. The dotted white line is the horizontal asymptote y = 1 (equal degrees, ratio 1/1 ).
x → − 2 lim ( x + 2 ) 2 5 − x
Forecast: Denominator is a square , so it can't go negative. What does that force about the two sides? Guess before reading.
Step 1 — Numerator sign near x = − 2 . 5 − x → 5 − ( − 2 ) = 7 > 0 , nonzero positive.
Step 2 — Denominator behaviour. Why a square matters: ( x + 2 ) 2 ≥ 0 always, and it hits 0 only at x = − 2 . Near − 2 it is a tiny positive — from both sides, because squaring kills the sign. So ( x + 2 ) 2 → 0 + left and right alike.
Step 3 — Combine. 0 + + 7 = + ∞ from both sides. Because the sides agree , the two-sided limit exists as + ∞ :
lim x → − 2 ( x + 2 ) 2 5 − x = + ∞. Vertical asymptote x = − 2 .
Verify: x = − 1.99 : ( 0.01 ) 2 = 0.0001 , 0.0001 6.99 = 69900 ✓. x = − 2.01 : ( − 0.01 ) 2 = 0.0001 , 0.0001 7.01 = 70100 ✓. Both large positive — agreement confirmed.
Common mistake Assuming every VA has disagreeing sides
Cell A (odd power in denominator) flips sign; cell B (even power) does not . Always look at the power of the vanishing factor.
x → 4 lim x − 4 x 2 − 16 — asymptote or hole?
Forecast: Denominator → 0 . Does it explode? Check the numerator first before guessing.
Step 1 — Test the numerator at x = 4 . 4 2 − 16 = 0 . Why this changes everything: both top and bottom → 0 , so this is the indeterminate form 0 0 — NOT automatically an asymptote. The vanishing factor might cancel.
Step 2 — Factor and cancel. Why factor? To expose and remove the shared ( x − 4 ) . x − 4 x 2 − 16 = x − 4 ( x − 4 ) ( x + 4 ) = x + 4 ( x = 4 ) . The cancellation is legal because in a limit x = 4 , so we never divide by an actual zero.
Step 3 — Take the limit of the simplified form. lim x → 4 ( x + 4 ) = 8 . Finite! So there is no vertical asymptote — just a removable hole at the point ( 4 , 8 ) . (See Continuity and removable discontinuities .)
Verify: x = 4.001 ⇒ x + 4 = 8.001 ≈ 8 ✓. Original at x = 3.999 : here 3.99 9 2 = 15.992001 , so − 0.001 15.992001 − 16 = − 0.001 − 0.007999 = 7.999 ✓ — no blow-up, confirming a hole not a wall.
Worked example Evaluate all three at
x → ∞ :
(D) x 2 + 3 2 x + 1 , (E) 2 x 2 + 9 4 x 2 − x , (F) x 2 x 3 + 1 .
Forecast: One dies to 0 , one lands on a nonzero number, one runs off to ∞ . Match each to its fate before reading.
The master move — divide by the highest denominator power. Why: this turns every term into a constant times x k 1 , and we know lim x → ∞ x k 1 = 0 for k > 0 . That single fact does all the work.
(D) deg top 1 < deg bottom 2 . Divide by x 2 : x 2 + 3 2 x + 1 = 1 + x 2 3 x 2 + x 2 1 → 1 + 0 0 + 0 = 0. HA y = 0 . Why: the numerator only has negative powers of x — it dies.
(E) deg top = deg bottom = 2 . Divide by x 2 : 2 x 2 + 9 4 x 2 − x = 2 + x 2 9 4 − x 1 → 2 + 0 4 − 0 = 2. HA y = 2 — the ratio of leading coefficients 4/2 .
(F) deg top 3 > deg bottom 2 . Divide by x 2 : x 2 x 3 + 1 = x + x 2 1 → + ∞. No horizontal asymptote. (Since deg top exceeds deg bottom by exactly 1 this is also a slant case — worked fully in Ex 9; general theory in Slant (oblique) asymptotes via polynomial division .)
Verify (plug x = 1000 ):
(D) 1000003 2001 ≈ 0.0020 → 0 ✓.
(E) 2000009 3999000 ≈ 1.9995 → 2 ✓.
(F) 1 0 6 1 0 9 + 1 = 1000000.000001 → ∞ ✓.
both horizontal asymptotes of f ( x ) = 2 x − 1 9 x 2 + 4 .
Forecast: A square root of x 2 hides a ∣ x ∣ . Do you expect the same limit on both ends, or two different ones? Guess.
Step 1 — Pull x 2 out of the root. Why: to isolate a x 2 1 term we can send to 0 . 9 x 2 + 4 = x 2 9 + x 2 4 = ∣ x ∣ 9 + x 2 4 . The key fact: x 2 = ∣ x ∣ , not x .
Step 2 — x → + ∞ . Here x > 0 , so ∣ x ∣ = x . Divide top and bottom by x : 2 x − 1 x 9 + 4/ x 2 = 2 − 1/ x 9 + 4/ x 2 → 2 9 = 2 3 . HA y = 2 3 on the right.
Step 3 — x → − ∞ . Why the sign flips: now x < 0 , so ∣ x ∣ = − x . That minus sign is the whole exam trap. 2 x − 1 − x 9 + 4/ x 2 = 2 − 1/ x − 9 + 4/ x 2 → 2 − 3 = − 2 3 . HA y = − 2 3 on the left.
Step 4 — Conclusion. Two different horizontal asymptotes: y = 2 3 (right) and y = − 2 3 (left).
Verify: x = 1000 : 1999 9 , 000 , 004 ≈ 1999 3000.0007 ≈ 1.5008 → 1.5 ✓. x = − 1000 : − 2001 3000.0007 ≈ − 1.4993 → − 1.5 ✓.
The figure makes the sign flip visible: the cyan curve levels off toward the amber line y = 2 3 as you travel right, but toward the white line y = − 2 3 as you travel left. That mirror-image split is entirely due to ∣ x ∣ = − x appearing on the left.
x 2 = x for x → − ∞
This is the #1 sign error in the whole topic. On the left half of the graph x is negative, so x 2 = ∣ x ∣ = − x . Skip the minus and you'll report y = 2 3 on both sides — wrong.
Worked example Horizontal asymptotes of
g ( x ) = e x + 1 e x − 1 , both ends.
Forecast: As x grows, e x dwarfs the constants; as x → − ∞ , e x → 0 . Two different ceilings? Guess.
Step 1 — x → + ∞ . Why divide by e x : it's the fastest-growing term, the analogue of "highest power." e x + 1 e x − 1 = 1 + e − x 1 − e − x . Since e − x = 1/ e x → 0 : → 1 + 0 1 − 0 = 1. HA y = 1 on the right.
Step 2 — x → − ∞ . Why NOT divide by e x here: as x → − ∞ , e x → 0 , so plug it in directly. e x + 1 e x − 1 → 0 + 1 0 − 1 = − 1. HA y = − 1 on the left.
Step 3 — Sanity on arctan (parent's forecast drill). For comparison: lim x → ∞ arctan x = 2 π and lim x → − ∞ arctan x = − 2 π — same two-ceiling pattern.
Verify: g ( 5 ) = e 5 + 1 e 5 − 1 = 149.41 147.41 ≈ 0.9866 → 1 ✓. g ( − 5 ) = e − 5 + 1 e − 5 − 1 = 1.0067 − 0.9933 ≈ − 0.9866 → − 1 ✓.
Worked example A drug's concentration in the blood (mg/L) after
t hours is C ( t ) = t + 8 40 t for t ≥ 0 . What concentration does the patient approach after a very long time, and what does that number mean?
Forecast: As hours pile up, does C keep climbing forever or level off? Guess the ceiling.
Step 1 — Recognise the case. C is a rational function, deg top = deg bottom = 1 → this is cell E . Expect a finite horizontal asymptote = ratio of leading coefficients.
Step 2 — Take the limit at infinity. Divide by t : C ( t ) = t + 8 40 t = 1 + t 8 40 → 1 + 0 40 = 40.
Step 3 — Interpret with units. The horizontal asymptote is C = 40 mg/L . Physically: the concentration rises toward a plateau of 40 mg/L and never exceeds it — the long-run steady level. It never quite reaches 40 (asymptotes are approached, not touched), matching how the graph hugs the ceiling.
Verify: C ( 100 ) = 108 4000 ≈ 37.04 ; C ( 10000 ) = 10008 400000 ≈ 39.97 → 40 ✓. Rising toward 40 , staying below it ✓. Units: h mg/L ⋅ h = mg/L ✓.
The figure shows C ( t ) climbing steeply at first, then bending to run flat just under the amber ceiling C = 40 — the curve approaches it but the amber arrow marks that it is never reached.
h ( x ) = x 2 − 9 x 2 − 2 x − 3 , find every vertical asymptote and every hole.
Forecast: The denominator x 2 − 9 = ( x − 3 ) ( x + 3 ) vanishes at x = 3 and x = − 3 . Are both vertical asymptotes? Careful — check the numerator at each.
Step 1 — Factor both. Why: to see which zeros of the bottom are shared by the top (holes) and which are not (true asymptotes). h ( x ) = ( x − 3 ) ( x + 3 ) ( x − 3 ) ( x + 1 ) .
Step 2 — At x = 3 : shared factor ( x − 3 ) cancels. Why this means a hole: both top and bottom hit 0 there — a 0 0 form that cancels, so no blow-up. Simplified: h ( x ) = x + 3 x + 1 for x = 3 . The hole's height: plug x = 3 into the simplified form, 3 + 3 3 + 1 = 6 4 = 3 2 . Hole at ( 3 , 3 2 ) .
Step 3 — At x = − 3 : factor does NOT cancel. In x + 3 x + 1 the denominator still vanishes at x = − 3 while the numerator gives − 3 + 1 = − 2 = 0 . Nonzero ÷ zero → vertical asymptote x = − 3 . (Odd power → sides will disagree: left is 0 − − 2 = + ∞ , right is 0 + − 2 = − ∞ .)
Step 4 — Bonus horizontal asymptote. Equal degrees 2 = 2 , leading coefficients 1/1 → HA y = 1 .
Verify: hole height x + 3 x + 1 at x = 3 gives 6 4 = 0.6667 ✓. Near x = − 3 : x = − 2.99 ⇒ 0.01 − 1.99 = − 199 (right side, − ∞ ) ✓; x = − 3.01 ⇒ − 0.01 − 2.01 = 201 (left side, + ∞ ) ✓. HA: h ( 1000 ) = 999991 997997 ≈ 0.998 → 1 ✓.
Worked example Find the slant asymptote of
f ( x ) = x − 1 x 2 + 3 x + 2 .
Forecast: deg top = 2 , deg bottom = 1 , so deg top is exactly one more than deg bottom. There is no horizontal asymptote (top bigger → bust), but the curve does line up with a slanted straight line far out. Guess whether that line rises or falls.
Step 1 — Why polynomial division? Why this tool: when deg top = deg bottom + 1 , dividing the polynomials rewrites the fraction as ( line ) + ( small remainder ) . The remainder is a proper fraction that → 0 at infinity, so the line is exactly what the curve hugs. Division is the tool that peels the line off the top.
Step 2 — Do the long division of x 2 + 3 x + 2 by x − 1 : x − 1 x 2 + 3 x + 2 = x + 4 + x − 1 6 . (Check: ( x + 4 ) ( x − 1 ) = x 2 + 3 x − 4 , plus remainder 6 gives x 2 + 3 x + 2 ✓.)
Step 3 — Send x → ± ∞ . Why: the term x − 1 6 is nonzero over a huge number → 0 . What survives is the line y = x + 4 . So the slant asymptote is y = x + 4 on both ends (it rises, slope + 1 ).
Step 4 — Note the vertical asymptote too. The denominator x − 1 = 0 at x = 1 with numerator 1 + 3 + 2 = 6 = 0 : nonzero ÷ zero → VA x = 1 . So this function has one slant and one vertical asymptote, no horizontal one.
Verify: at x = 1000 , f ( 1000 ) = 999 1003002 ≈ 1004.006 , while x + 4 = 1004 — the gap is ≈ 0.006 → 0 ✓. At x = − 1000 , f = − 1001 997002 ≈ − 996.006 and x + 4 = − 996 , gap ≈ − 0.006 → 0 ✓. Remainder identity ( x + 4 ) ( x − 1 ) + 6 = x 2 + 3 x + 2 ✓.
Recall Cover-the-matrix self-test
One-sided signs when denominator has an odd-power zero ::: opposite (limit DNE, still a VA)
Even-power zero in denominator ::: same sign both sides
Both top and bottom → 0 ::: possible removable hole — factor first
deg top < deg bottom at infinity ::: HA y = 0
deg top = deg bottom ::: HA = ratio of leading coefficients
deg top > deg bottom ::: no HA
deg top = deg bottom + 1 ::: slant asymptote (found by polynomial division)
x 2 as x → − ∞ ::: − x , flips the sign — two HAs possible
x finite or x to infinity
Cell C factor cancel hole