Quick Read: A graph is a picture of relationships. Intercepts tell you where the graph meets an axis; stationary points occur where the derivative is zero; turning points are places where the graph changes from increasing to decreasing or vice versa; asymptotes describe lines the graph approaches in a particular limiting sense; and end behaviour describes what happens as the input becomes very large positive or negative. These ideas overlap, but they are not interchangeable.
One-sentence answer: to read a graph well, ask where it is defined, where it crosses or touches the axes, where its direction changes, how steep it is, what it approaches, and what happens at the edges of its domain.
Why this 2015 Mathematics note needed an upgrade
The original article gave a short checklist for axial intercepts, stationary points and asymptotes. The core ideas were useful, but several statements were too compressed.
- An x-intercept is related to a real zero of a function, but “roots of an equation” is broader than simply saying x-intercepts.
- A stationary point is not always a turning point.
- A turning point need not always be stationary if the graph has a corner or cusp.
- An asymptote is better defined through limiting behaviour than as a line the graph simply “tends towards at infinity”.
- The numerator-degree rule for oblique asymptotes is a useful rational-function shortcut, not the general definition.
This page now develops a durable graph-reading framework suitable from upper-secondary Mathematics into Additional Mathematics, A-Level, IGCSE and IB contexts.
Start with the domain
Before asking where a graph goes, ask where the function is defined.
For example:
- y = 1/x is not defined at x = 0.
- y = √x over the real numbers is defined only for x ≥ 0.
- y = ln x is defined only for x > 0.
Domain restrictions can create endpoints, gaps or vertical asymptotic behaviour. They are part of the graph, not a technical detail added afterward.
Axis intercepts
The y-intercept
For a graph of y = f(x), the y-intercept occurs where x = 0, provided 0 belongs to the domain.
So compute f(0).
If f(0) = 5, the graph meets the y-axis at (0, 5).
If f(0) does not exist, the graph has no y-intercept.
The x-intercepts
An x-intercept occurs where y = 0. For y = f(x), solve:
f(x) = 0.
The real solutions that lie in the domain give the x-coordinates of the x-intercepts.
Are x-intercepts the same as roots?
For a real-valued graph y = f(x), a real zero or root of f corresponds to an x-intercept. But the language should still be used carefully.
- An equation can have complex roots that do not appear as x-intercepts on the real Cartesian plane.
- A repeated real root may make the graph touch the x-axis and turn back rather than cross it.
- The point must lie in the function’s domain.
So “x-intercept” is a geometric statement; “root” or “zero” is an algebraic statement.
Crossing versus touching the x-axis
The behaviour near a root depends on the function.
- y = x crosses the x-axis at 0.
- y = x² touches the x-axis at 0 and turns upward.
- y = x³ crosses the x-axis at 0 but flattens there.
These three graphs all have a zero at x = 0, but their local shapes are different.
Gradient: the local direction of travel
For a differentiable function, the derivative f′(x) gives the gradient of the tangent.
- f′(x) > 0 → the function is locally increasing.
- f′(x) < 0 → the function is locally decreasing.
- f′(x) = 0 → the tangent is horizontal at that point, assuming the derivative exists.
This leads to the idea of a stationary point.
Stationary points
A stationary point occurs where:
f′(x) = 0.
Its tangent is horizontal.
Common stationary-point types include:
- local maximum;
- local minimum;
- stationary point of inflection.
Turning points are not identical to stationary points
A turning point is a point where the graph changes direction from increasing to decreasing or from decreasing to increasing.
For a smooth differentiable graph, a local maximum or minimum commonly has f′(x) = 0 and is therefore stationary.
But consider:
- y = x³ has f′(0) = 0, but the graph continues increasing through x = 0. It is a stationary point of inflection, not a turning point.
- y = |x| turns at x = 0, but the derivative does not exist there. It is a turning point but not a stationary point.
That distinction prevents a common examination error.
Classifying a stationary point using the derivative sign
One robust method is to check the sign of f′(x) on either side.
- positive → zero → negative: local maximum;
- negative → zero → positive: local minimum;
- same sign on both sides: not a local maximum or minimum.
This first-derivative sign test directly checks whether the graph changes direction.
The second derivative test
If f′(a) = 0:
- f″(a) > 0 suggests a local minimum;
- f″(a) < 0 suggests a local maximum;
- f″(a) = 0 is inconclusive.
For y = x⁴ at x = 0, for example, f″(0) = 0 even though the origin is a local minimum. So “second derivative equals zero” does not by itself mean point of inflection.
Points of inflection
A point of inflection is associated with a change in concavity.
Informally:
- concave up: slope is generally increasing;
- concave down: slope is generally decreasing.
A point where f″(x) = 0 is only a candidate. The concavity must actually change.
Asymptotes: use limits, not the “never touches” myth
Students are sometimes told that an asymptote is a line the graph approaches but never touches. That is not a safe general definition.
A graph can cross a horizontal or oblique asymptote. What matters is the limiting relationship.
Vertical asymptote
The line x = a is a vertical asymptote if f(x) becomes unbounded as x approaches a from at least one side:
f(x) → ±∞ as x → a⁺ or x → a⁻.
Example: y = 1/x has the vertical asymptote x = 0.
Horizontal asymptote
The line y = L is a horizontal asymptote if:
f(x) → L as x → ∞ or x → −∞.
Example: y = 1 + 1/x approaches y = 1 as |x| becomes large.
Oblique or slant asymptote
The line y = mx + c is an oblique asymptote if:
f(x) − (mx + c) → 0 as x → ∞ or x → −∞.
Example:
y = (x² + 1)/x = x + 1/x.
As |x| becomes large, 1/x tends to 0, so the graph approaches y = x.
The rational-function degree shortcut
For rational functions, polynomial division gives useful quick rules.
- degree of numerator < degree of denominator → often y = 0 is the horizontal asymptote;
- equal degrees → horizontal asymptote given by the ratio of leading coefficients;
- numerator degree exactly one greater → polynomial division gives a linear quotient, producing an oblique asymptote.
These are rational-function shortcuts. They are not the definition of an asymptote.
A hole is not a vertical asymptote
Consider:
f(x) = (x² − 1)/(x − 1).
For x ≠ 1, this simplifies to x + 1. At x = 1, the original expression is undefined.
The graph therefore has a removable hole at (1, 2), not a vertical asymptote, because the limit remains finite.
End behaviour
End behaviour asks what happens as x becomes very large positive or very large negative.
For a polynomial, the leading term often determines the end behaviour.
- y = x² rises at both ends;
- y = −x² falls at both ends;
- y = x³ falls to the left and rises to the right;
- y = −x³ rises to the left and falls to the right.
End behaviour provides the large-scale shape before smaller features are added.
Asymptotic behaviour and end behaviour are related but different
A function may have end behaviour that approaches a line, giving a horizontal or oblique asymptote. A polynomial such as x³ has clear end behaviour but no horizontal or oblique asymptote.
Increasing and decreasing intervals
Once critical points are known, divide the domain into intervals and inspect the sign of f′(x).
For example, if f′(x) is:
- positive on x < −1;
- negative on −1 < x < 2;
- positive on x > 2;
then the graph rises, falls, then rises, suggesting a local maximum at x = −1 and a local minimum at x = 2 if the function is differentiable there.
Graph sketching: a reliable order
- Find the domain.
- Find intercepts.
- Check symmetry if useful.
- Find discontinuities and asymptotes.
- Differentiate.
- Find stationary and other critical points.
- Classify maxima/minima/inflexions.
- Determine increasing/decreasing intervals.
- Determine concavity if required.
- Check end behaviour.
- Join the information into a coherent sketch.
The order matters because each piece constrains the possible shape.
Example: y = x³ − 3x
Intercepts:
x³ − 3x = x(x² − 3) = 0, so x = 0, ±√3. The y-intercept is also (0, 0).
Stationary points:
f′(x) = 3x² − 3 = 3(x² − 1), so f′(x) = 0 at x = ±1.
f(−1) = 2 and f(1) = −2.
The derivative changes + → − at x = −1, so (−1, 2) is a local maximum. It changes − → + at x = 1, so (1, −2) is a local minimum.
End behaviour: the leading term x³ dominates, so the graph falls left and rises right.
Those facts already determine most of the sketch.
Example: y = x³ and the stationary inflection trap
f′(x) = 3x², so f′(0) = 0.
But f′(x) is positive on both sides of 0. The function is increasing before and after the origin.
So (0, 0) is stationary but not a turning point. Its concavity changes, making it a stationary point of inflection.
Example: y = |x| and the non-stationary turning point
The graph decreases toward x = 0 and increases after x = 0, so the origin is a local minimum and a turning point.
But the left derivative is −1 and the right derivative is +1. There is no single derivative at 0.
This is why “turning point = dy/dx = 0” is not a universal rule.
Common graph-reading mistakes
- Calling every stationary point a maximum or minimum.
- Assuming f″(x) = 0 automatically gives an inflection.
- Calling every denominator zero a vertical asymptote. Cancellation may produce a hole instead.
- Believing a graph can never cross an asymptote. Horizontal and oblique asymptotes can be crossed.
- Ignoring the domain.
- Finding coordinates without interpreting the shape.
- Sketching from a calculator image without explaining why the graph has that form.
Calculator graph versus mathematical graph
A graphing calculator or software window shows only a selected viewing rectangle. A feature can disappear if the scale is poor.
- A very narrow turning point can look like a corner.
- A distant intercept can sit outside the window.
- A near-vertical curve can be mistaken for an asymptote.
- Two branches can appear connected because of plotting resolution.
Technology is excellent for exploration and checking, but algebra and calculus explain the structure.
Why graph interpretation matters beyond sketching questions
Graphs are representations of relationships. The same reasoning appears in:
- motion graphs in Physics;
- rate and concentration graphs in Chemistry;
- population curves in Biology;
- economic cost and revenue models;
- probability distributions;
- data visualisation;
- optimisation problems.
The important habit is to translate visual shape back into a statement about the underlying quantities.
A graph-reading checklist
- What are the axes and units?
- What is the domain?
- Where are the intercepts?
- Where is the function undefined?
- Where is it increasing or decreasing?
- Where are the stationary points?
- Which stationary points actually turn?
- Are there corners or cusps?
- What asymptotes exist?
- What is the end behaviour?
- What does each feature mean in the problem context?
For students moving into calculus
Do not learn differentiation only as a procedure for producing f′(x). Keep the geometric meaning attached:
- f′ tells you local slope;
- the sign of f′ tells you increasing/decreasing behaviour;
- zeros of f′ identify stationary candidates;
- changes in the sign of f′ identify turns;
- f″ helps describe how the slope itself changes.
Then calculus becomes a language for graph shape rather than another collection of algebraic rules.
Updated from eduKatePunggol’s June 2015 “Graphs Axis, Stationary Points and Asymtotes”. The original purpose as a quick reference is preserved and expanded into a precise graph-reading framework covering domain, intercepts, stationary versus turning points, inflection, asymptotes, limits and end behaviour.
