Intuition What this page is for
The parent note taught you what level curves and level surfaces are . This page drills them across every scenario you could meet: positive/zero/negative level values, degenerate cases, unbounded shapes, a real-world map problem, and an exam-style twist. Before each example you will forecast the shape yourself, then we compute, then we verify by plugging back.
Every level-set problem falls into one of these cells. The worked examples below are tagged with the cell(s) they hit — together they cover the whole table.
#
Cell (what makes it special)
Example that hits it
A
c > 0 gives a nice bounded curve (circle/ellipse)
Ex 1, Ex 6
B
c = 0 degenerate (point / crossing lines)
Ex 1, Ex 2
C
c < 0 empty set (impossible value)
Ex 1, Ex 6
D
Sign of c flips the shape family (hyperbola flips axis)
Ex 2
E
Unbounded level curve (line, parabola, whole strip)
Ex 3
F
Domain restriction bites (log, root, division)
Ex 4, Ex 7
G
Level surface in 3D (n = 3 )
Ex 5
H
Real-world word problem (a contour map)
Ex 6
I
Exam twist — recognise shape without a clean formula
Ex 7
Definitions we lean on (from the parent):
Definition Level curve / level surface (recap)
For z = f ( x , y ) , the level curve of value c is {( x , y ) : f ( x , y ) = c } — all inputs giving the same output c , drawn in the flat x y -plane.
For w = f ( x , y , z ) , the level surface of value c is {( x , y , z ) : f ( x , y , z ) = c } , drawn in 3D space. Here c is a label , never a coordinate.
Common mistake Two facts we will lean on (stated here so this page is self-contained)
Crowding = steepness. When equal steps in c produce level curves that get closer and closer together, the surface is climbing faster there — the terrain is steeper. We use this in Ex 1.
Level-set dimension = (#variables) − 1. One equation removes one degree of freedom, so a level set of f ( x , y , z ) is a 2D surface , not a curve. We use this in Ex 5. (Both facts are the "common mistakes" warned about in the parent note.)
f ( x , y ) = x 2 + y 2 , sketch level curves for c = − 4 , 0 , 1 , 4 , 9 .
Forecast: you have seen this bowl. Guess: which of these five values give a real curve, which give a single point, which give nothing? Write your guess before reading on.
Step 1. Set f = c : x 2 + y 2 = c .
Why this step? By definition a level curve collects all inputs sharing the output c ; setting the formula equal to c is literally that condition.
Step 2. Compare with the circle template x 2 + y 2 = r 2 .
Why this step? We want the shape's name and size. Matching to a known template gives radius r = c instantly.
Step 3. Split by the sign of c — this is where "all cases" matters:
c = 9 : circle radius 9 = 3 . (cell A)
c = 4 : circle radius 2 . (cell A)
c = 1 : circle radius 1 . (cell A)
c = 0 : x 2 + y 2 = 0 forces x = y = 0 — a single point , the origin. (cell B, degenerate)
c = − 4 : a sum of two squares can never be negative, so no points — the empty set. (cell C)
Why split by sign? Because c is only real for c ≥ 0 , and c = 0 collapses the circle to a point. The formula secretly changes character at c = 0 .
Step 4. Notice the spacing: radii are 1 , 2 , 3 for c = 1 , 4 , 9 — equal jumps in c but the radii grow like c , so rings crowd outward.
Why note this? Crowded rings = steep terrain (the "crowding = steepness" fact stated at the top of this page).
Figure 1 — Concentric level circles of f ( x , y ) = x 2 + y 2 : green ring c = 1 (radius 1), blue ring c = 4 (radius 2), orange ring c = 9 (radius 3), the red centre dot for the degenerate c = 0 point, and a corner note that c = − 4 draws nothing.
Look at the figure: the three coloured rings (green c = 1 , blue c = 4 , orange c = 9 ) are the real circles; the red dot at the centre is the c = 0 point; and the note in the corner reminds you c = − 4 draws nothing. Trace outward and watch the rings bunch up — that visual crowding is the point of Step 4.
Verify: plug ( 3 , 0 ) into f : 3 2 + 0 2 = 9 ✓ (on the c = 9 ring). Plug ( 0 , 0 ) : 0 ✓ (the c = 0 point). And 4 = 2 ✓.
Why verify? Plugging a point back into f confirms it truly sits on the ring we claimed, catching any arithmetic slip in the radius.
f ( x , y ) = x 2 − y 2 , level curves for c = 4 , c = 0 , c = − 4 .
Forecast: one of these three is not a smooth curve. Which, and what does it become?
Step 1. c = 4 : x 2 − y 2 = 4 , i.e. 4 x 2 − 4 y 2 = 1 .
Why divide by 4? To reach the standard hyperbola form a 2 x 2 − b 2 y 2 = 1 , whose + sits on x 2 ⇒ it opens left–right (crosses the x -axis at x = ± 2 ). (cell D)
Step 2. c = − 4 : x 2 − y 2 = − 4 ⇒ 4 y 2 − 4 x 2 = 1 .
Why flip signs? Multiplying by − 1 moves the + onto y 2 ⇒ the hyperbola now opens up–down (crosses the y -axis at y = ± 2 ). The sign of c chose the axis. (cell D)
Step 3. c = 0 : x 2 − y 2 = 0 ⇒ ( x − y ) ( x + y ) = 0 ⇒ y = x or y = − x .
Why factor? A difference of squares factors; each factor = 0 is a straight line. Two crossing lines — the degenerate hyperbola (its own asymptotes). (cell B)
Figure 2 — Level curves of the saddle f ( x , y ) = x 2 − y 2 : the blue hyperbola (c = 4 ) opens left–right through x = ± 2 , the orange hyperbola (c = − 4 ) opens up–down through y = ± 2 , and the two dashed red lines y = x , y = − x are the degenerate c = 0 case, which are also the shared asymptotes.
Look at the figure: the blue hyperbola opens left–right (c = 4 ), the orange one opens up–down (c = − 4 ), and the two dashed red lines through the origin are the c = 0 case. Notice the blue and orange curves both hug the red dashed lines far from the origin — those lines are their shared asymptotes, which is exactly why c = 0 is those lines.
Verify: on c = 4 test ( 2 , 0 ) : 4 − 0 = 4 ✓. On c = − 4 test ( 0 , 2 ) : 0 − 4 = − 4 ✓. On c = 0 test ( 3 , 3 ) (line y = x ): 9 − 9 = 0 ✓.
Why verify? Testing one point per level value confirms each point lands on the correct branch (and that the c = 0 point really lies on a line, not a curve).
f ( x , y ) = y − x 2 , level curves for c = − 1 , 0 , 2 .
Forecast: solving for y should reveal a familiar 1-variable curve, just shifted. Guess the shape.
Step 1. Set y − x 2 = c and solve for y : y = x 2 + c .
Why solve for y ? Because then each level curve is the graph of a plain 1-variable function — instantly recognisable.
Step 2. Read off: every level curve is an upward parabola y = x 2 shifted vertically by c . These parabolas extend forever — they are unbounded (no largest x ). (cell E)
c = 2 : parabola with vertex ( 0 , 2 ) .
c = 0 : vertex at origin.
c = − 1 : vertex ( 0 , − 1 ) .
Step 3. Do any values of c give the empty set? No. For every real c , picking x = 0 , y = c already satisfies it. So there is no "impossible" cell here — unlike Ex 1.
Why check? A subtraction (not a sum of squares) can hit any real value, so no c is forbidden.
Figure 3 — Three identical upward parabolas y = x 2 + c stacked vertically: green (c = − 1 , vertex at ( 0 , − 1 ) ), blue (c = 0 , vertex at origin), orange (c = 2 , vertex at ( 0 , 2 ) ), with the red test point ( 1 , 3 ) sitting on the orange curve. The arms never turn back — the curves are unbounded.
Look at the figure: three parabolas of identical shape stacked vertically — green (c = − 1 ) lowest, blue (c = 0 ) in the middle, orange (c = 2 ) highest. The red dot marks the test point ( 1 , 3 ) sitting on the orange curve. See how none of the arms ever turn back: that is the "unbounded" claim made visible.
Verify: point ( 1 , 3 ) on the c = 2 curve: y = x 2 + 2 = 1 + 2 = 3 ✓, and f ( 1 , 3 ) = 3 − 1 = 2 ✓.
Why verify? Checking the same point two ways (from the solved form and from the original f ) confirms our algebra of solving for y was reversible.
f ( x , y ) = ln ( x 2 + y 2 ) . Find where f is defined, then the level curve f = 0 .
Forecast: the domain will have a hole. And f = 0 should be a circle of a specific radius — which one?
Step 1. Domain: ln needs a positive input, so require x 2 + y 2 > 0 .
Why strict > ? ln of 0 is undefined (and negatives too, but a sum of squares is never negative). So the only excluded point is x 2 + y 2 = 0 , i.e. the origin . Domain = R 2 ∖ {( 0 , 0 )} — the plane with a single puncture. (cell F)
Step 2. Level curve f = 0 : ln ( x 2 + y 2 ) = 0 .
Why is a log easy here? Apply e ( ⋅ ) to undo ln : x 2 + y 2 = e 0 = 1 .
Why e 0 ? ln and e x are inverses; e 0 = 1 .
Step 3. So f = 0 is the circle of radius 1 . And f = c generally gives x 2 + y 2 = e c , a circle of radius e c /2 — real for every c (since e c > 0 always). No empty cell, but the origin is never on any level curve.
Verify: ( 1 , 0 ) : ln ( 1 2 + 0 ) = ln 1 = 0 ✓. For c = 2 , radius = e 1 ≈ 2.718 ; test point ( e , 0 ) : ln ( e 2 ) = 2 ✓.
Why verify? Because a log/exp step is easy to invert wrongly; substituting the point confirms the circle radius really produces the intended level value.
f ( x , y , z ) = x 2 + y 2 − z . Describe the level surface f = 0 and f = 1 .
Forecast: with three inputs, one equation should give a surface (dimension 3 − 1 = 2 ), not a curve. Which quadric?
Step 1. f = 0 : x 2 + y 2 − z = 0 ⇒ z = x 2 + y 2 .
Why solve for z ? Then it reads as "height z over the point ( x , y ) " — this is a paraboloid (a bowl opening upward). (cell G)
Step 2. f = 1 : x 2 + y 2 − z = 1 ⇒ z = x 2 + y 2 − 1 .
Why? Same bowl, shifted down by 1 (vertex at z = − 1 ). Different c just slides the bowl vertically.
Step 3. Cross-check the dimension rule: 3 variables, 1 equation ⇒ level set is 3 − 1 = 2 dimensional ⇒ a surface. ✓ (This is the "level-set dimension = variables − 1" fact stated at the top of this page.)
Figure 4 — Two stacked paraboloid level surfaces of f ( x , y , z ) = x 2 + y 2 − z in 3D: the blue bowl is f = 0 (z = x 2 + y 2 ) with vertex at the origin, the orange bowl is f = 1 (z = x 2 + y 2 − 1 ) shifted down so its vertex sits at z = − 1 . Both are 2D sheets floating in 3D space.
Look at the figure: two stacked bowls in 3D — the blue one is f = 0 (z = x 2 + y 2 ) with its lowest point at the origin, and the orange one is f = 1 , the identical bowl slid down so its vertex sits at z = − 1 . Both are clearly 2D sheets (surfaces) floating in 3D space, confirming the dimension count.
These paraboloids are examples of Quadric surfaces .
Verify: on f = 0 , point ( 1 , 1 , 2 ) : 1 + 1 − 2 = 0 ✓. On f = 1 , point ( 1 , 1 , 1 ) : 1 + 1 − 1 = 1 ✓.
Why verify? A 3D point is easy to misplace; substituting confirms it lies on the correct shell rather than between the two bowls.
Worked example A hill's height is
H ( x , y ) = 100 − x 2 − 4 y 2 , where H is in metres. To keep the formula dimensionally clean we write it with an implicit length scale L = 1 m so that x , y are pure numbers (multiples of L ): H = ( 100 − x 2 − 4 y 2 ) m . Draw the "H = const" trails at H = 100 , 64 , 0 metres and interpret. Is H = 120 m possible?
Forecast: the x 2 and 4 y 2 have different weights — guess whether the contours are circles or squished ellipses, and which way squished.
Step 0 (units check). Inside the bracket, 100 , x 2 and 4 y 2 are all pure numbers (because x , y are measured as multiples of the unit length L = 1 m ), and the whole bracket is multiplied by 1 m to give a height in metres. So the addition 100 − x 2 − 4 y 2 is dimensionally consistent — we never add a length to an area.
Why this step? You can only add or subtract quantities with matching units; making x , y dimensionless (numbers of metres) keeps every term in the bracket a pure number and prevents the classic "length + area" blunder.
Step 1. Set H = c (in metres): 100 − x 2 − 4 y 2 = c ⇒ x 2 + 4 y 2 = 100 − c .
Why rearrange? To move all the square terms to one side, giving a sum of positive squares on the left — a recognisable ellipse form whose size is set by the single number 100 − c .
Step 2. H = 100 (the summit): x 2 + 4 y 2 = 100 − 100 = 0 ⇒ x = y = 0 — a single point , the top of the hill. (cell B/A boundary)
Why does it collapse to a point? A sum of positive squares equals 0 only when each square is 0 , forcing x = 0 and y = 0 .
Step 3. H = 64 : x 2 + 4 y 2 = 100 − 64 = 36 , i.e. 36 x 2 + 9 y 2 = 1 .
Why divide by 36? Dividing through by the right-hand number puts it in the standard ellipse form a 2 x 2 + b 2 y 2 = 1 with a = 6 , b = 3 : it reaches 6 along x but only 3 along y — squished in the y -direction because the 4 y 2 term grows four times faster than x 2 . (cell A, cell H)
Step 4. H = 0 (sea level): x 2 + 4 y 2 = 100 − 0 = 100 , so 100 x 2 + 25 y 2 = 1 , an ellipse with a = 10 , b = 5 — the widest contour.
Why is it the widest? Smaller c leaves a bigger number 100 − c on the right, and larger right-hand value means larger semi-axes, so lower trails are bigger loops.
Step 5. H = 120 : x 2 + 4 y 2 = 100 − 120 = − 20 < 0 — impossible , empty set. The hill never reaches 120 m; its max height is 100 m at the summit. (cell C)
Why does negativity mean empty? The left side x 2 + 4 y 2 is a sum of squares, so it can never equal a negative number — no point satisfies the equation.
Figure 5 — Contour map of the hill H = 100 − x 2 − 4 y 2 : the blue ellipse (H = 64 , semi-axes 6 and 3 ) nested inside the wider orange ellipse (H = 0 sea level, semi-axes 10 and 5 ), both stretched along x and pinched along y . The red centre dot is the summit (H = 100 ) and a corner note states H = 120 draws nothing (max height is 100 ).
Look at the figure: the blue ellipse (H = 64 ) is nested inside the wider orange ellipse (H = 0 , sea level), both stretched along x and pinched along y — that pinch is the 4 y 2 term at work. The red dot at the centre is the summit (H = 100 ), and the corner note reminds you H = 120 draws nothing at all.
Verify: ( 6 , 0 ) gives H = 100 − 36 − 0 = 64 ✓ (on the H = 64 trail). ( 10 , 0 ) : H = 100 − 100 = 0 ✓ (sea level). Summit ( 0 , 0 ) : H = 100 ✓, and no ( x , y ) makes H = 120 ✓.
Why verify? Substituting corner points of each ellipse confirms both the height labels and that the semi-axes (6 and 10 along x ) are correct.
f ( x , y ) = y x (with y = 0 ). What are the level curves f = 2 , f = 0 , and f = − 1 ? Where is f undefined?
Forecast: a ratio being constant — that should force a straight-line relationship. Guess through which point every level curve passes.
Step 1. Domain: y = 0 (no division by zero). So the entire x -axis is excluded — this is a domain-restriction case. (cell F)
Why the whole x -axis? The x -axis is exactly the set of points with y = 0 , and every such point makes the denominator zero, so all of them are thrown out.
Step 2. f = 2 : y x = 2 ⇒ x = 2 y .
Why multiply up? Clearing the fraction (multiplying both sides by y ) turns the ratio into a linear equation. But note y = 0 , so the origin is punctured out — it's the line x = 2 y minus the point ( 0 , 0 ) .
Why does the line pass through where the origin would be? Because a constant ratio means x and y scale together — all such points lie on a line through the origin. (cell I)
Step 3. f = 0 : y x = 0 ⇒ x = 0 (with y = 0 ) — the y -axis, minus the origin.
Why set x = 0 and not y = 0 ? A fraction is zero exactly when its numerator is zero while the denominator stays non-zero; so x = 0 , y = 0 , which is the y -axis with the origin removed.
Step 4. f = − 1 : y x = − 1 ⇒ x = − y , the line y = − x punctured at the origin.
Why? Clearing the fraction gives x = − y ; again y = 0 removes the origin. The negative constant produces a line with negative slope.
Figure 6 — Level curves of f ( x , y ) = x / y : three straight rays through (but excluding) the origin — blue line x = 2 y (f = 2 ), the orange vertical y -axis (f = 0 ), and the green line y = − x (f = − 1 ). The origin is drawn as a hollow circle to show it is removed from every level curve.
Look at the figure: three straight lines through the origin — blue (x = 2 y ), orange (the vertical y -axis for f = 0 ), and green (y = − x ). The origin is drawn as a hollow circle to signal it is removed from every level curve. All lines fan out from that one hole, which is the visual signature of a constant ratio.
So every level curve of x / y is a line through the origin (with the origin removed), and the slope of that line is set by c . The gradient idea (Gradient vector ) explains why these are all rays fanning out from one point.
Verify: ( 2 , 1 ) on f = 2 : 2/1 = 2 ✓. ( 0 , 5 ) on f = 0 : 0/5 = 0 ✓. ( 3 , − 3 ) on f = − 1 : 3/ ( − 3 ) = − 1 ✓.
Why verify? Substituting confirms each test point respects y = 0 and lands on the claimed line, guarding against accidentally including the forbidden origin.
Recall Quick self-test on the matrix
Which cell does "empty set" belong to, and give a function+value that hits it? ::: Cell C — e.g. x 2 + y 2 = − 4 (sum of squares can't be negative).
Sign of c flipping a hyperbola from left-right to up-down is which cell? ::: Cell D (Ex 2).
For f ( x , y , z ) = c , what dimension is the level set and why? ::: Dimension 2 (a surface); 3 variables minus 1 equation.
In Ex 6, why is H = 120 impossible? ::: It needs x 2 + 4 y 2 = − 20 < 0 ; the hill maxes out at 100 m.
"Sign, Zero, Squeeze." For any level-set problem check the Sign of c (which family / empty?), the Zero case (degenerate point or crossing lines?), and the Squeeze of the domain (roots, logs, division cutting holes).