What exactly is an asymptote?

An asymptote is a straight line that a curve gets arbitrarily close to but never actually reaches as x or y goes to infinity (or, for vertical asymptotes, as x approaches a specific forbidden value). The curve and the asymptote are separate objects — the asymptote is a reference line, not part of the graph of the function.

What is the difference between an asymptote and a hole?

A vertical asymptote is where the function blows up to ±∞ as x approaches some value — think 1/(x−2) at x=2. A hole is a single missing point on an otherwise normal curve, where the function is undefined (0/0) but the surrounding graph behaves fine. (x²−1)/(x−1) has a hole at x=1: the expression simplifies to x+1 everywhere except at x=1, where it's undefined.

Can a function cross its asymptote?

A function can absolutely cross its horizontal or slant asymptote — sometimes many times — in the middle of the graph. The asymptote only describes the end behavior as x goes to ±∞. A vertical asymptote, however, can never be crossed, because the function is not defined at that x-value in the first place.

How do I find asymptotes algebraically?

For a rational function p(x)/q(x), vertical asymptotes are the x-values where q(x)=0 but p(x)≠0. Horizontal asymptotes depend on degrees: if deg(p) deg(q), there is no horizontal asymptote. When deg(p)=deg(q)+1, do polynomial long division — the non-remainder part is the slant asymptote.

Why does my graphing calculator not show the hole?

A hole is infinitely thin — a single point — and a screen is made of finite pixels. The calculator samples the function at, say, a few hundred x-values across the window. Unless one of those samples lands exactly on the undefined x, the hole is invisible. The calculator draws a continuous curve through its samples and the gap vanishes at screen resolution.

Home › Articles › Reading a Graph

Reading a Graph: Asymptotes, Holes, and What That Gap Means

You plot 1/(x−2) and the curve rockets toward a vertical wall at x=2 without ever touching it. You plot (x²−1)/(x−1) and get a straight line — but somewhere on that line there's supposed to be a hole. Here's what those features actually are, how to find them with algebra before you trust the screen, and why your calculator sometimes hides the hole completely.

Last updated: May 4, 2026.

An asymptote is a line, not part of the curve

An asymptote is a straight line that a curve gets arbitrarily close to but never actually reaches. The key word is reference. The asymptote is not part of the function's graph — it's a guide line that describes where the curve is heading. Think of it like the center line on a highway you're drifting toward but never cross.

There are three kinds, and each tells you something different about the function:

Vertical asymptote — a line x = a where the function blows up to +∞ or −∞ as x approaches a. Caused by dividing by zero when the numerator isn't zero.
Horizontal asymptote — a line y = b that the function approaches as x goes to +∞ or −∞. Describes end behavior, not local behavior.
Slant (oblique) asymptote — a diagonal line y = mx + b the function approaches at the far edges. Shows up when the numerator's degree is exactly one higher than the denominator's.

Worked example: f(x) = 1/(x−2)

This is the cleanest rational function on Earth. Two asymptotes, no holes, no surprises.

Vertical asymptote. The denominator x − 2 is zero when x=2. The numerator is 1 (never zero). So the function is undefined at x=2, and as x sneaks toward 2 the output explodes:

f(1.9) = 1/(−0.1) = −10
f(1.99) = 1/(−0.01) = −100
f(2.01) = 1/(0.01) = 100
f(2.1) = 1/(0.1) = 10

From the left the function dives to −∞; from the right it shoots up to +∞. That's the signature of a vertical asymptote at x = 2.

Horizontal asymptote. What does 1/(x−2) do as x gets huge? At x=1002, f = 1/1000 = 0.001. At x=1,000,002, f = 0.000001. The output squeezes down toward zero without ever reaching it. So y = 0 is the horizontal asymptote.

Plot it on the graphing calculator and you'll see two branches hugging two lines they never touch — a textbook rational function.

Worked example: f(x) = (x²−1)/(x−1) and the disappearing hole

This one looks like it should have a vertical asymptote at x=1 — the denominator is zero there, right? Watch what happens when you factor:

(x² − 1)/(x − 1) = (x−1)(x+1)/(x−1) = x + 1, as long as x ≠ 1.

The (x−1) cancels. So the function is algebraically identical to x + 1 everywhere except at x=1, where the original expression is 0/0 — undefined. The graph is a straight line of slope 1 through the y-axis at 1… with a single missing point at (1, 2).

That missing point is a hole, also called a removable discontinuity. Removable because if you just defined f(1) = 2 by hand, the function would be continuous. Compare that to x=2 in the previous example, where no amount of patching saves you — the function is genuinely infinite there.

Why your graphing calculator may not show the hole

Pull up (x²−1)/(x−1) on any graphing tool and you'll almost always see what appears to be a perfect straight line. No gap. The hole vanishes. Why?

A hole is a single point — mathematically, a set of measure zero. The calculator samples the function at a finite number of x-values across the viewing window — typically a few hundred, one per pixel column. Unless one of those sample points lands exactly on x=1, the renderer never evaluates 0/0 and never notices anything is wrong. It just draws a line through all the points it did compute.

Even when a sample does hit the bad x, the gap is one pixel wide. At normal zoom, one pixel is indistinguishable from a continuous line. You have to zoom in hard — or algebraically know where to look — to see the hole at all.

Moral: the screen is a lie by omission. Always check the denominator algebraically for values that make it zero, and factor the numerator to see whether those zeros are asymptotes or holes.

Finding asymptotes algebraically (faster than the graph)

For a rational function f(x) = p(x)/q(x) in lowest terms:

Vertical asymptotes: solve q(x) = 0. Each root that doesn't also kill the numerator gives a vertical asymptote.
Holes: values that make both p(x) and q(x) zero simultaneously — they cancel out.
Horizontal asymptotes:
- If deg(p) < deg(q): y = 0.
- If deg(p) = deg(q): y = leading coefficient of p ÷ leading coefficient of q.
- If deg(p) > deg(q): no horizontal asymptote.
Slant asymptotes: when deg(p) = deg(q) + 1, do polynomial long division. p/q = (linear quotient) + (remainder/q). The linear quotient is the slant asymptote.

Example: (x² + 1)/(x − 1). Long division gives x + 1 + 2/(x−1). The remainder vanishes as x → ∞, so the slant asymptote is y = x + 1. There's also a vertical asymptote at x=1.

Most people get this wrong

Two mistakes show up constantly on homework and exams.

Mistake #1: treating a hole like a vertical asymptote. If the factor that makes the denominator zero also appears in the numerator, it cancels. That's a hole, not a wall. (x²−1)/(x−1) does not have a vertical asymptote at x=1 — it has a hole, and the rest of the function is a perfectly tame line.

Mistake #2: thinking a horizontal asymptote can't be crossed. It absolutely can. The asymptote only describes what happens at the edges, as x → ±∞. In the middle, the curve can weave across y = b as many times as it wants. The function f(x) = sin(x)/x, for instance, has horizontal asymptote y = 0 and crosses it infinitely many times.

What you cannot cross is a vertical asymptote, because the function isn't defined there. Different rule entirely — vertical means "forbidden x-value," horizontal means "long-run trend." Mixing those up is where students lose points.

Domain gotchas

The domain of a rational function is "all real numbers except where the denominator is zero." That set always includes every vertical asymptote and every hole. Write it as an explicit exclusion:

f(x) = 1/(x−2) — domain: x ≠ 2.
f(x) = (x²−1)/(x−1) — domain: x ≠ 1, even though the graph looks like the full line y = x+1.
f(x) = 1/((x−3)(x+4)) — domain: x ≠ 3, x ≠ −4, with vertical asymptotes at both.

Watch for composed functions too. √(x−5) has domain x ≥ 5; ln(x) has domain x > 0. The graph just stops at the edge, and that edge is part of the function's behavior — not an asymptote, not a hole, just "undefined beyond here."

Frequently asked questions

Is the asymptote part of the graph of the function?

No. The asymptote is a reference line — a straight line drawn as a guide. The function's actual graph is separate. Some textbooks draw asymptotes as dashed lines precisely to signal "this is not part of f, it's a landmark." A function never equals its asymptote at any finite x (for vertical and for strict horizontal asymptotes; the curve can touch or cross a horizontal asymptote at finite x, it just can't settle on it at infinity).

Can a function have more than one horizontal asymptote?

Yes — one for x → +∞ and a different one for x → −∞. A classic example is f(x) = x/√(x²+1), which approaches y = 1 as x → +∞ and y = −1 as x → −∞. Rational functions are nicer: if they have a horizontal asymptote, it's the same on both sides.

How do I tell a hole from an asymptote just from the formula?

Factor top and bottom. If a factor cancels completely, the x-value where it was zero is a hole. If the denominator still has a factor that makes it zero after cancellation, that's a vertical asymptote. If both — say (x−1)² on the bottom and (x−1) on top — you cancel one copy and the leftover (x−1) gives you an asymptote.

Do all functions have asymptotes?

No. Polynomials don't — they go to ±∞ without approaching any line. sin(x) and cos(x) don't either; they oscillate forever without converging. Asymptotes mostly show up in rational functions, exponentials (eˣ has horizontal asymptote y=0 as x → −∞), logarithms (vertical asymptote at x=0), and a handful of other cases.

What's a "removable" discontinuity versus a non-removable one?

A removable discontinuity is a hole — you can "remove" it by defining the function's value at that point to equal the limit. A non-removable discontinuity is a vertical asymptote or a jump; no single value you could plug in would make the function continuous, because the left and right limits disagree (or are infinite).

Try it yourself

Plot 1/(x−2) and (x²−1)/(x−1) side by side on the graphing calculator. Zoom into x=1 on the second one — see if you can catch the missing pixel. Then try (x² + 1)/(x − 1) and watch the slant asymptote y = x + 1 emerge as you zoom out. Features you can predict with algebra beat features you hope the screen will show.

This article is for general education. See our Terms.