← Back to the practice exams

Midterm 3 Crash Course

From zero to exam-ready in six lessons. Every worked example is a real past-exam problem, checked against the official key. Tonight 6:45 PM. No calculator.

Run the circuit →Drill the 6 memorize-templates on the practice circuit

Memorize these 6

These six problem types show up on every past exam (2025, 2024, 2019, 2018). Learn the template, then fill in the blanks. Short on time? Memorize this card, then do the "Now try" problems.

① Unit square → parallelogram (linear map)on 4/4 past exams
T(u,v) = P + u·edge₁ + v·edge₂ P = one corner, edges = (neighbor corner) − P CHECK: T(0,0), T(1,0), T(0,1), T(1,1) = the 4 corners ∬ y dA sign: region mostly above x-axis → +, symmetric across x-axis → 0, mostly below → −
Exactly this type: 2025 #4 · 2024 #3 · 2019 #8 · 2018 #8
Lesson 3 →
② Green's theoremon 4/4 past exams
∮ P dx + Q dy = ∬ (Q_x − P_y) dA (counterclockwise) If Q_x − P_y = constant k → answer = k · Area Clockwise curve → flip the sign Ring/hole: outer CCW, inner CW
Exactly this type: 2025 #8 · 2024 #5 · 2019 #7 · 2018 #5
Lesson 4 →
③ Match a triple integral to a solidon 4/4 past exams
Pair ∫ signs with d's INSIDE-OUT (first d ↔ innermost ∫) Outer limits = numbers = slicing range Inner limits = floor → ceiling Constant limit (z=0, y=1) = flat face Linear limit (1−x) = slanted face Slices shrink to a point = cone / pyramid
Exactly this type: 2025 #1 · 2024 #1, #2 · 2019 #5 · 2018 #1, #3
Lesson 1 →
④ Cylindrical / spherical setupon 4/4 past exams
Cyl: x=r cosθ y=r sinθ z=z dV = r dz dr dθ Sph: x=ρ sinφ cosθ y=ρ sinφ sinθ z=ρ cosφ dV = ρ² sinφ dρ dφ dθ φ measured DOWN from +z axis, 0 ≤ φ ≤ π cone z=√(x²+y²) ↔ φ = π/4 xy-plane ↔ φ = π/2 integrand = (f converted) × (r or ρ² sinφ)
Exactly this type: 2025 #2, #3 · 2024 #4 · 2019 #3 · 2018 #4
Lesson 2 →
⑤ Parametrize a surface + sign of ∬ f dSon 4/4 past exams
graph z=f(x,y): r = ⟨u, v, f(u,v)⟩ solved for one var (y = g(x,z)): r = ⟨u, g(u,v), v⟩ cylinder x²+y²=a²: r = ⟨a cos u, a sin u, v⟩ sphere radius a: r = ⟨a sinφ cosθ, a sinφ sinθ, a cosφ⟩ Area = ∬ |r_u × r_v| du dv ∬ f dS = ∬ f(r(u,v)) |r_u × r_v| du dv Sign of ∬ f dS = sign of f on the surface (f odd over a symmetric surface → 0)
Lesson 5 →
⑥ Nonlinear change of variableson 4/4 past exams
J = det [ x_u x_v ] = x_u·y_v − x_v·y_u [ y_u y_v ] ∬_R f dA = ∬_S f(x(u,v), y(u,v)) · |J| du dv Bounds come from S (the uv box), NOT from R
Exactly this type: 2025 #5 · 2024 #6 · 2019 #6 · 2018 #6
Lesson 3 →

Links open the practice-exam page. Tap that year's tab if it isn't already showing; the step-by-step explanation is on the right side, next to the problem.

1/6

Reading iterated integral limits → match the solid

The one idea

The limits of a triple integral are a recipe for building the solid: the outer variable says where you slice, and the inner limits say where each slice starts and stops. Read them from the inside out and the shape falls out.

The recipe

  1. Pair the ∫ signs with the d's inside-out. In ∫∫∫ … dz dy dx, the innermost ∫ goes with dz, the middle with dy, the outer with dx.
  2. Outer limits are plain numbers. That's the range you slice along.
  3. Middle limits (can depend on the outer variable) give the shadow / 2D slice.
  4. Inner limits are floor → ceiling (or left wall → right wall).
  5. Read each limit as a face: a constant (z = 0, y = 1) is a flat face; a linear limit (z = 1 − x) is a slanted plane.
  6. Plug in the endpoints of the outer variable. If the slice shrinks to a point, it's a cone or pyramid. If it stays the same, it's a prism/wedge.
xz ceiling z = 1 − x floor z = 0 walk up: 0 → 1−x

Formula box

∫a..b ∫g₁(x)..g₂(x) ∫floor..ceiling f dz dy dx outer=numbers middle=shadow inner Volume = ∭ 1 dV

Worked example · 2025 #1 (A) and (B)

(A) ∫₀¹ ∫z−11−z ∫z−11−z f dx dy dz

1. Pair them. Order is dx dy dz: inner = x from z−1 to 1−z, middle = y from z−1 to 1−z, outer = z from 0 to 1.Why: the first d belongs to the innermost ∫.
2. Slice at a fixed height z. Both x and y run over [−(1−z), 1−z]: a square centered on the z-axis with half-width 1 − z.
3. Watch the slice change. z = 0: square −1 ≤ x,y ≤ 1. z = ½: half-size. z = 1: shrinks to the point (0,0,1).Why: a square that shrinks linearly to a point = pyramid.
(A) = square pyramid, base [−1,1]×[−1,1] on the xy-plane, apex (0,0,1).

(B) ∫₀¹ ∫₀¹ ∫₀1−x g dz dy dx

1. Pair them. Inner z: 0 → 1−x. Middle y: 0 → 1. Outer x: 0 → 1.
2. Shadow. The outer two give the unit square 0 ≤ x,y ≤ 1.
3. Floor & ceiling. Floor z = 0 (flat). Ceiling z = 1 − x (slanted plane): height 1 at x = 0, height 0 at x = 1. No y in it, so every y-slice looks the same.
(B) = a wedge / ramp: tall face on the yz-plane, sloping down to the line x = 1, z = 0.

Top traps

  • Pairing the d's backwards. The first differential goes with the innermost ∫.
  • Using a vertical plane (like y = x) as a z floor/ceiling. A face that contains the z-direction bounds the shadow, never z. (2024 #1)
  • Integrand for volume is just 1.

Now try

On the practice page, the step-by-step explanation is on the right side of each problem. Try first, then open it.

2/6

Cylindrical & spherical setup

The one idea

Round solids are ugly in x, y, z but have constant limits in r, θ (cylindrical) or ρ, φ, θ (spherical). You convert the boundaries, convert the function, and multiply by the extra stretch factor (r or ρ² sin φ), which is where most points get lost.

The recipe

  1. Pick the system. Axis-symmetric with z-tops/bottoms (paraboloid, cylinder, cone + paraboloid) → cylindrical. Spheres, shells, cones from the origin → spherical.
  2. Convert every surface. x² + y² = r². x² + y² + z² = ρ².
  3. Cylindrical bounds: z from floor to ceiling (in r). r: 0 to where floor = ceiling. θ: which quadrants (full circle = 0 to 2π).
  4. Spherical bounds: ρ from the spheres. φ from cones / the xy-plane (φ = π/2) / z ≤ 0 (φ from π/2 to π). θ from the signs of x and y.
  5. Integrand = (function converted) × (r for cylindrical, ρ² sin φ for spherical).
+z (φ = 0) xy-plane (φ = π/2) ρ φ cone z=√(x²+y²): φ = π/4 −z (φ = π)

Formula box

CYLINDRICAL x = r cosθ y = r sinθ z = z dV = r dz dr dθ SPHERICAL x = ρ sinφ cosθ y = ρ sinφ sinθ z = ρ cosφ dV = ρ² sinφ dρ dφ dθ 0 ≤ φ ≤ π (from +z axis) 0 ≤ θ ≤ 2π cone z = √(x²+y²) ↔ z = r ↔ φ = π/4 sphere x²+y²+z² = a² ↔ ρ = a θ: Q1 0..π/2 Q2 π/2..π Q3 π..3π/2 Q4 3π/2..2π

Worked example · 2025 #2 (cylindrical)

Set up the mass of the solid between the cone z = √(x²+y²) (below) and the paraboloid z = 6 − x² − y² (above), density y², order dz dr dθ.

1. Convert the surfaces. Cone: z = r (floor). Paraboloid: z = 6 − r² (ceiling).Why: x² + y² = r², so √(x²+y²) = r.
2. Inner (z): r ≤ z ≤ 6 − r².
3. Find r max: set floor = ceiling. r = 6 − r² → r² + r − 6 = 0 → (r+3)(r−2) = 0 → r = 2. So 0 ≤ r ≤ 2.Why: the solid ends where the two surfaces meet. Not r = √6 (where the paraboloid hits z = 0).
4. θ: nothing cuts it, so 0 ≤ θ ≤ 2π.
5. Integrand: y² = r² sin²θ, times the extra r → r³ sin²θ.
Mass = ∫₀^2π ∫₀² ∫_r^(6−r²) r³ sin²θ dz dr dθ (= 104π/15, not asked)

Quick spherical one · 2025 #3

∭ y dV over 1 ≤ x²+y²+z² ≤ 4, x ≤ 0, y ≥ 0, z ≤ 0, order dφ dθ dρ.

Integrand: y = ρ sinφ sinθ, times ρ² sinφ → ρ³ sin²φ sinθ.
ρ: 1 ≤ ρ ≤ 2. θ: x ≤ 0, y ≥ 0 is quadrant II → π/2 ≤ θ ≤ π. φ: z ≤ 0 is below the xy-plane → π/2 ≤ φ ≤ π.
∫₁² ∫_(π/2)^π ∫_(π/2)^π ρ³ sin²φ sinθ dφ dθ dρ (= 15π/16 > 0, makes sense since y ≥ 0)

Top traps

  • Forgetting r or ρ² sin φ. This is the #1 point-loser. The integrand box includes it.
  • Mixing up the angles: z conditions → φ; x/y sign conditions → θ. Lower half is φ from π/2 to π, not 0 to π/2.
  • Swapping floor/ceiling, or using where one surface hits z = 0 instead of the intersection.

Now try

The step-by-step explanation is on the right side of each problem on that page.

3/6

Change of variables / Jacobian

The one idea

A map T(u,v) bends a simple uv box into a weird xy region, so you integrate over the easy box instead. The price is the stretch factor |J|, the absolute value of the 2×2 determinant of partial derivatives.

The recipe

  1. Linear map, square → parallelogram: pick a corner P of the parallelogram. edge₁ and edge₂ = the two neighboring corners minus P. T(u,v) = P + u·edge₁ + v·edge₂.
  2. Check by plugging in all 4 corners (0,0), (1,0), (0,1), (1,1). They must hit all 4 corners of the parallelogram.
  3. Jacobian: J = x_u·y_v − x_v·y_u. Use |J|.
  4. Rewrite: ∬_R f dA = ∬_S f(x(u,v), y(u,v)) |J| du dv. Bounds come from the uv region S.
  5. Nonlinear map: same steps. Map the corners/edges to identify the picture (edges may curve).
(0,0)u=1v=1S T P = T(0,0)u·edge₁v·edge₂

Formula box

T(u,v) = P + u·edge₁ + v·edge₂ J = det [ x_u x_v ] = x_u·y_v − x_v·y_u [ y_u y_v ] dA = |J| du dv ∬_R f dA = ∬_S f(T(u,v)) |J| du dv

Worked example · 2024 #3 (linear + shift)

Find T sending the unit square onto the parallelogram with corners (0,1), (−1,2), (1,3), (2,2).

1. Pick P. P = (0,1), the bottom corner.Why: any corner works; its two neighbors give the edges.
2. Edges from P. to (−1,2): ⟨−1, 1⟩. To (2,2): ⟨2, 1⟩. Check: (0,1) + ⟨−1,1⟩ + ⟨2,1⟩ = (1,3), the 4th corner ✓.
3. Assemble. T = (0,1) + u⟨−1,1⟩ + v⟨2,1⟩ = (−u + 2v, u + v + 1).
4. Check all 4 corners. T(0,0) = (0,1), T(1,0) = (−1,2), T(0,1) = (2,2), T(1,1) = (1,3) ✓.
T(u,v) = (−u + 2v, u + v + 1). (Key also accepts (2u − v, u + v + 1).) Here |J| = |(−1)(1) − (2)(1)| = 3.

Quick nonlinear one · 2025 #5

∬_R y dA with T(u,v) = (u + ½v², uv), S: 1 ≤ u ≤ 2, 0 ≤ v ≤ 1, order du dv.

x_u = 1, x_v = v, y_u = v, y_v = u → J = u − v². On S, u ≥ 1 ≥ v², so |J| = u − v². Function y = uv.
∫₀¹ ∫₁² uv(u − v²) du dv (= 19/24, not asked)

Top traps

  • Forgetting the shift P. A purely linear map always sends (0,0) to (0,0).
  • Using corners as edge vectors. Edges are differences of corners.
  • Forgetting |J|, or using R's bounds instead of S's.

Sign of ∬ y dA (2025 #4b, 2019 #8): look at the parallelogram. Symmetric across the x-axis → 0. Mostly above → positive.

Now try

The step-by-step explanation is on the right side of each problem on that page.

4/6

Green's theorem

The one idea

A line integral around a closed curve equals a double integral of Q_x − P_y over the region inside. On the exam the messy terms usually cancel and Q_x − P_y is a constant, so the answer is just that constant × the area.

The recipe

  1. Name P and Q. F = ⟨P, Q⟩, so ∮ F·dr = ∮ P dx + Q dy.
  2. Compute Q_x − P_y (always Q with x, minus P with y).
  3. Check the orientation. Counterclockwise (region on your left) = +. Clockwise = flip the sign.
  4. Integrate over D. Constant k → k · Area(D).
  5. Hole? The positive boundary is outer CCW + inner CW. If the inner curve is drawn CCW, it enters with a minus.
Dcounterclockwise= positiveregion on your leftas you walk

Formula box

∮_C P dx + Q dy = ∬_D ( ∂Q/∂x − ∂P/∂y ) dA C counterclockwise (positive). Clockwise → × (−1) Q_x − P_y = k (constant) → answer = k · Area(D) Circle area πa² Ring π(R² − r²) Triangle ½bh

Worked example · 2025 #8

∮_C F·dr, C = unit circle (arrows counterclockwise), F = ⟨−y + y cos x, x + sin x⟩.

1. Orientation. Arrows: up on the right, left on top → counterclockwise → + sign.
2. Name them. P = −y + y cos x, Q = x + sin x.
3. Partials. Q_x = 1 + cos x. P_y = −1 + cos x.
4. Subtract. Q_x − P_y = (1 + cos x) − (−1 + cos x) = 2.Why: the cos x terms cancel. That's the hint they want Green's.
5. Area. ∬_D 2 dA = 2 · π(1)² = 2π.
∮_C F·dr = 2π

Top traps

  • Ignoring orientation. Clockwise would give −2π.
  • Computing P_y − Q_x or P_x − Q_y. It is always Q_x − P_y.
  • Ring (2024 #5): outer CCW, inner CW. With both drawn CCW, boundary = A − B, so −3·3π = 10 − ∮_B → ∮_B = 10 + 9π.

Now try

The step-by-step explanation is on the right side of each problem on that page.

5/6

Parametric surfaces & surface integrals

The one idea

A surface is a flat uv region bent into 3D by r(u,v). The little patch du dv gets stretched into area |r_u × r_v| du dv, so surface area and ∬ f dS are just double integrals with that factor attached.

The recipe

  1. Parametrize. Pick two free variables (u, v), solve for the third. Graph z = f(x,y) → ⟨u, v, f(u,v)⟩. y = g(x,z) → ⟨u, g(u,v), v⟩. Plane: solve the plane's equation for one variable. Cylinder/sphere/cone: use angles (table below).
  2. Domain D: read the u, v ranges off the picture (edges of the piece).
  3. Partials r_u and r_v.
  4. Cross product r_u × r_v (i, j, k determinant; middle term gets a minus).
  5. Length |r_u × r_v| = dS factor.
  6. Area = ∬_D |r_u × r_v| du dv. ∬ f dS = ∬_D f(r(u,v)) |r_u × r_v| du dv.
  7. Sign questions: dS is always ≥ 0, so the sign of ∬ f dS = the sign of f on the surface. If f is odd and the surface is symmetric (and dS matches), it's 0.
  8. Identify a picture: eliminate u, v (e.g. x² + z² = y⁴ → circles of radius y² around the y-axis).

Formula box

graph z=f(x,y) ⟨u, v, f(u,v)⟩ y = g(x,z) ⟨u, g(u,v), v⟩ cylinder x²+y²=a² ⟨a cos u, a sin u, v⟩ sphere ρ = a ⟨a sinφ cosθ, a sinφ sinθ, a cosφ⟩ cone z=√(x²+y²) ⟨v cos u, v sin u, v⟩ revolve about y, radius R(y): ⟨R(u) sin v, u, R(u) cos v⟩ r_u × r_v = ⟨ y_u z_v − z_u y_v, −(x_u z_v − z_u x_v), x_u y_v − y_u x_v ⟩ dS = |r_u × r_v| du dv Area = ∬ |r_u × r_v| dA ∬ f dS = ∬ f(r(u,v)) |r_u × r_v| dA

Worked example · 2019 #9

r(u,v) = ⟨u² sin v, u, u² cos v⟩, 0 ≤ u ≤ 1, 0 ≤ v ≤ 2π.

1. Identify. y = u, and x² + z² = u⁴(sin²v + cos²v) = y⁴. So at each y, a full circle of radius y² around the y-axis.Why: radius ∝ y² starts pinched and flares out → a horn along +y (4th picture).
2. Partials. r_u = ⟨2u sin v, 1, 2u cos v⟩. r_v = ⟨u² cos v, 0, −u² sin v⟩.
3. Cross. i: (1)(−u² sin v) − (2u cos v)(0) = −u² sin v.
j: −[(2u sin v)(−u² sin v) − (2u cos v)(u² cos v)] = 2u³.
k: (2u sin v)(0) − (1)(u² cos v) = −u² cos v.
r_u × r_v = ⟨−u² sin v, 2u³, −u² cos v⟩ |r_u × r_v| = √(u⁴ sin²v + 4u⁶ + u⁴ cos²v) = √(u⁴ + 4u⁶)
4. Area integrand g(u,v) = √(u⁴ + 4u⁶).
5. Sign of ∬ xz dS. xz = u⁴ sin v cos v. ∫₀^2π sin v cos v dv = 0, so the whole thing is 0.Why: the horn is symmetric about the y-axis; xz > 0 and xz < 0 parts cancel.
Horn (4th picture) · ⟨−u² sin v, 2u³, −u² cos v⟩ · √(u⁴ + 4u⁶) · ∬ xz dS = 0

Top traps

  • Dropping the minus on the j-component of the cross product.
  • Writing dS = du dv (forgetting |r_u × r_v|), or leaving f in x, y, z instead of u, v.
  • "z ≥ 0 so it's positive" when f is z − 1 (2024 #8 → zero). Check where f is positive vs negative.

Now try

The step-by-step explanation is on the right side of each problem on that page.

6/6

Double integrals: reversing order & polar

The one idea

The limits describe a 2D region; sketch it, then slice it the other way to get new limits. When the inner integral is impossible (sin(x²), e^(x²)), reversing is the whole trick; when the region is round, polar with dA = r dr dθ is.

The recipe

  1. Read the region. Inner limits are the curves; outer limits are the numbers. dx inner = horizontal strips (Type II); dy inner = vertical strips (Type I).
  2. Sketch it and label the corners.
  3. Re-slice. New outer = numbers covering the region; new inner = bottom curve → top curve (or left → right), read off the sketch.
  4. Evaluate. The new inner integral usually produces exactly the factor needed for a u-substitution.
  5. Polar: x = r cosθ, y = r sinθ, x² + y² = r², dA = r dr dθ. r = cos θ is a circle of radius ½ centered at (½, 0); r = a cos θ is a circle of diameter a.
xy x: y → √π y: 0 → x y = x√π

Formula box

Type I (vertical strips): ∫_a^b ∫_g₁(x)^g₂(x) f dy dx Type II (horizontal): ∫_c^d ∫_h₁(y)^h₂(y) f dx dy Polar: dA = r dr dθ x² + y² = r² r = cos θ ↔ (x − ½)² + y² = ¼ (θ from −π/2 to π/2)

Worked example · 2025 #9

∫₀^√π ∫_y^√π 2 sin(x²) dx dy

1. Why reverse? sin(x²) has no elementary antiderivative in x, so the dx-first order is impossible.
2. Read the region. Horizontal strips: at height y, x goes from the line x = y to the wall x = √π; y from 0 to √π. Triangle with corners (0,0), (√π,0), (√π,√π), below y = x.
3. Re-slice vertically. x from 0 to √π; at fixed x, y from 0 (floor) to x (the diagonal).Why: new limits come from the picture, not by swapping numbers.
4. Inner. ∫₀^x 2 sin(x²) dy = 2x sin(x²).
5. Outer. ∫₀^√π 2x sin(x²) dx = [−cos(x²)]₀^√π = −cos π + cos 0 = 2.
Region = triangle (0,0), (√π,0), (√π,√π) · Value = 2

Top traps

  • Reversing by swapping numbers instead of reading the sketch (gives the mirror triangle above y = x).
  • Polar without the extra r (2019 #2: (2r² + r sinθ)·r, not 2r² + r sinθ).
  • Misreading r = cos θ as a half-disk at the origin. It's a small circle through the origin to the right.

Now try

The step-by-step explanation is on the right side of each problem on that page.

Copied. Paste into any AI chat.