17 worked examples28 exercises6 figuresbölüm 11 / 14
11Volumes by shells and transcendental-function foundations
Here is a region you can draw in four seconds: the arch under $y=x-x^{2}$ between $0$ and $1$, resting on the horizontal axis. Spin it around the vertical axis and you get a solid you could hold in one hand. Now slice it the way we sliced everything last week — flat, across the axis — and the very first line of work is to solve $y=x-x^{2}$ for $x$, which hands you $x=\tfrac{1}{2}\bigl(1\pm\sqrt{1-4y}\bigr)$: two square roots and two boundary branches for one thin slab.
By the end of this section you can compute that volume with a single integral in $x$ that fits on one line, decide before writing anything which of the two slicing directions is the cheaper one, and say what $\ln x$ is without using the word exponent.
In 60 seconds
Slice parallel to the axis and each strip sweeps out a hollow cylinder of volume $2\pi\,(\text{distance to the axis})\times(\text{height})\times(\text{thickness})$; the same integral machinery, run on $1/t$, is what $\ln x$ is defined to be.
, axis = the vertical axis
$V=2\pi\int_a^b x\,f(x)\,dx$
the region is given as $y=f(x)\ge 0$ on $[a,b]$ and is turned about the vertical axis
Shell method, axis moved to $x=c$
$V=2\pi\int_a^b \lvert x-c\rvert\,h(x)\,dx$
the axis is a vertical line that misses the region; $h$ is top minus bottom
solving for an exponent, and differentiating or integrating anything with $e$ in it
Three most common mistakes
Dropping the radius factor and integrating $2\pi\int f(x)\,dx$. For the arch under $y=x-x^{2}$ that returns $\pi/3$, exactly twice the true $\pi/6$, because it pretends every strip travels the same distance per turn.
Using the coordinate as the radius after the axis has moved. About $x=3$ the radius of the strip at $x$ is $3-x$; writing $x-3$ gives $-8\pi$ for a solid that has volume $8\pi$.
Applying $\frac{d}{dx}e^{x}=e^{x}$ to $e^{2x}$. The rule is about the function $x\mapsto e^{x}$, not about anything containing an $e$; the chain rule makes it $2e^{2x}$.
Midterm 2 carries 28 percent of the course, the same as Midterm 1 and the Final, with quizzes at 10 and homework at 6. This week supplies two kinds of question that are cheap to grade and therefore common: a volume of revolution where the method choice is the whole difficulty, and a derivative that no standard rule touches until you take a logarithm first.
How much time do you have?
10 minutes
You leave able to set up and evaluate one shell integral about the vertical axis, which is the single most repeated task of the week.
In 60 seconds card, One strip, one hollow cylinder, Setting up any shell integral in four moves, Formula card
45 minutes
Add the moved axis, the rule that decides shells against washers, and the definition of the logarithm; you can now answer a full multi-part volume question and a from the definition question.
everything in the 10 minute path, When the axis moves, the radius is a distance, Which way to slice: the rule that decides, The logarithm is an area, Full exam-style question, Practice C
full read
The inverse construction of the exponential, , the scaffolded ladder and the interleaved set where the type of the question is hidden from you — this is the part that transfers to a question you have not seen.
all blocks in order, The exponential is the logarithm read backwards, Taking the logarithm first, Scaffolding comes off, Practice A to D, Mistake ledger, Check yourself
By the end of this section
Derive the shell $2\pi x f(x)\,dx$ by unrolling one strip, and use it to compute a volume about the vertical axis in a single integral.
Write the as a distance to the axis for an axis that is not a coordinate axis, and check the sign by evaluating it at an endpoint of the interval.
Decide between shells and disks or washers from the direction of the strip relative to the axis, and set up the same volume both ways when both are available.
Use $\ln x=\int_1^x dt/t$ to prove a logarithm law, to differentiate $\ln\lvert u(x)\rvert$, and to say why $\ln$ has no value at $0$ or below.
Explain why the exponential is defined as the inverse of the logarithm, and get $\frac{d}{dx}e^{x}=e^{x}$ out of that definition in one line instead of assuming it.
Apply logarithmic differentiation to a variable base with a variable exponent and to a long product or quotient, and multiply back by $y$ at the end.
Syllabus coverage
5.3
Volumes by
The volume element, the moved axis, horizontal axes and the choice between shells and washers are the first three blocks.
covered
6.2*
The natural logarithmic function, defined as an integral
The definition, the logarithm laws proved from it, the derivative from the Fundamental Theorem, and logarithmic differentiation.
covered
6.3*
The natural exponential function as the inverse of the logarithm
The inverse construction, the number $e$, the derivative and the integral of $e^{u}$.
covered
general bases $a^{x}$ and $\log_a x$
General exponential and logarithmic functions
Not on this week's syllabus line and not needed by anything here; the next section of the course opens with them, built on top of the two definitions given in this one.
deferred
Recall first
Disk and
Slicing across the axis at $x$ gives a disk of radius $R(x)$, or a washer if the solid has a hole: $V=\pi\int_a^b\bigl[R(x)^{2}-r(x)^{2}\bigr]dx$.
It is the method this section is measured against; half the work here is deciding which of the two to use, and the other half uses it as an independent check on a shell answer.
Substitution in a definite integral
If $u=g(t)$ is differentiable then $\int_{a}^{b}f(g(t))g'(t)\,dt=\int_{g(a)}^{g(b)}f(u)\,du$: the limits move with the substitution.
The proof that $\ln(ab)=\ln a+\ln b$ is one substitution, and it works only because the limits are carried along.
Fundamental Theorem of Calculus, Part 1
If $f$ is continuous on an interval containing $a$, then $\dfrac{d}{dx}\displaystyle\int_a^x f(t)\,dt=f(x)$.
It is the whole derivation of $(\ln x)'=1/x$. Without it the definition of $\ln$ would be a dead end.
Implicit differentiation
Differentiating $\ln y=g(x)$ with respect to $x$ treats $y$ as a function of $x$ and gives $\dfrac{y'}{y}=g'(x)$.
Logarithmic differentiation is exactly this step; the $y'/y$ on the left is where the method gets its leverage.
Derivative of an
If $f$ is differentiable and one to one with $f'\neq 0$, then $\bigl(f^{-1}\bigr)'(b)=\dfrac{1}{f'\bigl(f^{-1}(b)\bigr)}$: the slope of the mirror image is the reciprocal of the original slope at the matching point.
It turns the one line $(\ln x)'=1/x$ into the one line $(e^{x})'=e^{x}$. This section uses it once, and that once is the cornerstone.
Try it yourself first (3 questions)
1§11.0 — washers, from the previous section●●○○○
Three questions from earlier in the course, before anything new starts. Getting one wrong costs nothing here; it only tells you which of the recalls above to read twice.
Given
The region $R$ lies between $y=x$ (above) and $y=x^{2}$ (below) for $0\le x\le 1$
$R$ is rotated about the horizontal axis
Find
Which integral gives the volume of the resulting solid?
Hint 1/4
You are not asked to evaluate anything yet, only to say which two lengths are the radii of the ring and in which order they enter.
Hint 2/4
Washer method: $V=\pi\int\bigl[R(x)^{2}-r(x)^{2}\bigr]dx$, with $R$ the distance from the axis to the far boundary and $r$ the distance to the near one.
Hint 3/4
Here the axis is $y=0$, the far boundary is $y=x$ and the near one is $y=x^{2}$, on $0\le x\le 1$. So $R(x)=x$ and $r(x)=x^{2}$.
Hint 4/4
The integral is $\pi\int_0^1(x^{2}-x^{4})dx$, which equals $2\pi/15$.
Show solutionName the two radii
$R(x)=x,\qquad r(x)=x^{2}$
at each $x$ the slice runs from the axis out to the far curve and the hole reaches to the near curve
Order of magnitude: the solid sits inside the cylinder of radius $1$ and length $1$, volume $\pi\approx 3.14$, and it is a thin sliver of that. $2\pi/15\approx 0.42$ is comfortably inside.
Squares first, subtraction second. $\bigl(R-r\bigr)^{2}$ and $R^{2}-r^{2}$ are different numbers and only one of them is an area.
2§11.0 — one substitution, limits carried along●●○○○
The second recall in action. Nothing about this integral needs the new material; it is here because the proof of the first logarithm law is the same move.
Given
$\displaystyle\int_0^1\frac{2x}{x^{2}+1}\,dx$
Find
Evaluate the integral exactly.
Hint 1/4
Look at the numerator and the denominator together before reaching for any technique. Is one of them related to the other?
Hint 2/4
Substitution: if $u=g(x)$ then $\int_a^b f(g(x))g'(x)dx=\int_{g(a)}^{g(b)}f(u)\,du$, limits included.
Hint 3/4
With $u=x^{2}+1$ we get $du=2x\,dx$, and the given limits $x=0,1$ become $u=1,2$. The integral turns into $\int_1^2 du/u$.
Hint 4/4
That evaluates to $\ln 2-\ln 1=\ln 2$.
Show solutionChoose the substitution
$u=x^{2}+1,\qquad du=2x\,dx$
the factor $2x\,dx$ already sitting in the integrand is what makes this substitution free
$x=0\Rightarrow u=1,\qquad x=1\Rightarrow u=2$
moving the limits now means no switching back to $x$ at the end
Integrate in u
$\int_1^2\frac{du}{u}=\ln 2-\ln 1=\ln 2$
the antiderivative of $1/u$ on a positive interval is $\ln u$, and $\ln 1=0$
Answer $$\int_0^1\frac{2x}{x^{2}+1}dx=\ln 2$$
Check
Sanity bound: on $[0,1]$ the integrand runs from $0$ up to $1$, so the integral is between $0$ and $1$. $\ln 2\approx 0.69$ sits inside.
Whenever the top of a fraction is the derivative of the bottom, the integral is a logarithm of the bottom.
3§11.0 — the theorem that undoes an integral●●○○○
This one is a trap on purpose: the two plausible answers differ by whether you read the theorem or the antiderivative. Pick before you read the hints.
Given
$\displaystyle F(x)=\int_1^x\frac{1}{t}\,dt$ for $x>0$
Find
What is $F'(x)$?
Hint 1/4
You are being asked for a rate: how fast does the accumulated area grow as the right-hand end moves? No integral needs to be evaluated.
Hint 2/4
Fundamental Theorem of Calculus, Part 1: $\dfrac{d}{dx}\displaystyle\int_a^x f(t)\,dt=f(x)$ whenever $f$ is continuous there.
Hint 3/4
Here $a=1$ and $f(t)=1/t$, which is continuous for $t>0$. The theorem hands back $f$ evaluated at the moving limit.
Hint 4/4
So $F'(x)=1/x$.
Show solutionCheck the hypothesis
$f(t)=\tfrac1t \text{ is continuous on } (0,\infty)$
the closed interval from $1$ to $x$ stays inside $(0,\infty)$ for every $x>0$, so the theorem applies
the derivative of an accumulated area is the height of the curve at the moving end
Answer $$F'(x)=\frac1x$$
Check
Independent check by area: between $x$ and $x+h$ the extra area is roughly a rectangle of width $h$ and height $1/x$, so the rate of gain is $1/x$ per unit of $x$.
The theorem returns the integrand, not an antiderivative. If your answer still has an integral sign or a logarithm in it, you differentiated the wrong object.
Notation
symbol
reads as
means
watch out
$dV=2\pi r h\,dr$
d V equals two pi r h d r
the volume of one thin hollow cylinder of radius $r$, height $h$ and wall thickness $dr$
The three factors are circumference, height and thickness, in that order. If you cannot name which is which in your integral, you have not set it up yet.
$\lvert x-c\rvert$
the absolute value of x minus c
the distance from the point $x$ to the vertical line $x=c$
In practice the region sits on one side of the axis, so you drop the bars and write whichever of $x-c$ or $c-x$ is non-negative on the whole interval.
$h(x)$
h of x
the height of the strip at $x$, always measured top curve minus bottom curve
Not the value of one curve. Where the region is bounded above by $y=4$ and below by $y=x^{2}$, the height is $4-x^{2}$, never $x^{2}$.
$\ln x=\int_1^x\frac{dt}{t}$
natural log of x equals the integral from one to x of d t over t
the signed area under $y=1/t$ from $1$ to $x$
$t$ is a dummy variable and never appears in the answer; $x$ is the moving upper limit, and it must be positive.
$\exp=\ln^{-1}$
exp is the inverse of natural log
the function that undoes $\ln$; we write $e^{x}$ for $\exp x$
$\ln^{-1}$ here means the inverse function, not $1/\ln$. Those are different objects and only one of them is defined on all of $\mathbb{R}$.
$\frac{y'}{y}$
y prime over y
the derivative of $\ln y$ with respect to $x$, also called the
Getting $y'/y$ is not the answer. The last step of logarithmic differentiation is always multiplying back by $y$, written out in $x$.
Conventions used here
A radius is a distance
Every radius in this section is a non-negative number. We write the subtraction in the order that keeps it non-negative on the whole interval of integration, and we check that by putting one endpoint into the expression before integrating.
A signed radius produces a signed volume, and a negative volume is not a result you can repair at the end with an absolute value; it means the picture was wrong.
Height means top minus bottom
The height of a vertical strip is the upper boundary minus the lower boundary at that $x$. For a horizontal strip it is the right boundary minus the left boundary at that $y$.
The most common wrong shell integral of the week is the one that uses a single curve where the region actually sits between two.
Domain first for the logarithm
$\ln x$ is defined only for $x>0$. At $x=0$ and below there is no value to report, so the correct sentence is that $\ln$ is undefined there, together with $\lim_{x\to 0^{+}}\ln x=-\infty$ as a separate statement about behaviour.
Writing $\ln 0=-\infty$ treats a limit as a value, and that is how sign and domain errors get laundered into an answer.
Exact answers, with $\pi$ left standing
Volumes are reported in closed form with $\pi$ as a symbol, and logarithms are reported as $\ln 2$ rather than $0.693$, unless the question asks for a decimal.
A decimal hides whether the exact answer was reached, and it is the exact answer that the next part of the question usually needs.
Angles in radians
Every trigonometric function in this section takes radians, so $\frac{d}{dx}\sin x=\cos x$ holds as written.
One example here differentiates $(\sin x)^{x}$, and the derivative formula it leans on is false in degrees.
One strip, one hollow cylinder
Last section every slice ran across the axis and every slice was a disk. Here is the region where that choice starts charging you rent.
Solvable with what we have
Rotate the region under $y=\sqrt{x}$ on $[0,4]$ about the horizontal axis: each slice is a disk of radius $\sqrt{x}$, so $V=\pi\int_0^4 x\,dx$.
Rotate the region between $y=x^{2}$ and $y=4$ about the horizontal axis: washers, outer radius $4$, inner radius $x^{2}$.
Rotate the region under $y=x$ on $[0,3]$ about the vertical axis: the boundary $y=x$ inverts to $x=y$ in one step, so horizontal slices are washers.
Not solvable yet
Rotate the arch under $y=x-x^{2}$ on $[0,1]$ about the vertical axis: a horizontal slice needs $x$ written in terms of $y$, and the arch has two branches.
Rotate the region between $y=x$ and $y=x^{3}$ about the vertical axis: again two branches, and they swap roles at $x=1$.
Rotate the region under $y=x^{5}+x$ about the vertical axis: this boundary cannot be solved for $x$ in closed form at all.
Slice the first of these across the axis, at height $y$. The slab is a washer whose two edges are the two solutions of $y=x-x^{2}$, namely $x=\tfrac12\bigl(1\pm\sqrt{1-4y}\bigr)$. So
Nothing in that line is wrong, and it even collapses nicely: the bracket is $\sqrt{1-4y}$, so the answer is $\pi/6$. The cost is the first move. To write it down at all you had to invert the boundary, and the quadratic formula only obliged because the boundary happened to be a quadratic. Change $x-x^{2}$ to $x^{5}+x$ and the same first move has no output. A method whose opening step is solve the boundary for the other variable is a method that stops working as soon as the boundary is anything but simple.
RuleThe shell volume element
Conditions
$f$ is continuous and $f(x)\ge 0$ on $[a,b]$
$0\le a
the is the vertical axis $x=0$
$$\boxed{\;V=2\pi\int_a^b x\,f(x)\,dx\;}$$
Add up, over every strip from $a$ to $b$: how far that strip travels in one turn $\bigl(2\pi x\bigr)$, times how tall it is $\bigl(f(x)\bigr)$, times how thick it is $\bigl(dx\bigr)$.
Where the three factors come from
Take the strip between $x$ and $x+dx$: height $f(x)$, thickness $dx$. Turning it about the vertical axis sweeps out a hollow cylinder of radius $x$, height $f(x)$ and wall thickness $dx$. Cut that cylinder down one side and flatten it: it becomes a rectangular sheet of width $2\pi x$ — the circumference it used to be — height $f(x)$ and thickness $dx$. Hence $dV=2\pi x f(x)\,dx$, and adding over all strips is $\int_a^b$. The one approximation, that the inner and outer circumferences of a thin wall are the same, is exactly the one that vanishes in the limit.
One strip, unrolled. The strip at $\textcolor{#d1690a}{x}$ has to travel the whole circumference $\textcolor{#d1690a}{2\pi x}$ in one turn, which is why the radius appears as a factor and not as a limit.
Looks like this, but is not
$V=2\pi\displaystyle\int_0^1 f(x)\,dx$ for the same arch. The $2\pi$ is there, the height is there, the limits are the right ones, and it has the shape of every volume formula you have seen.
The factor $x$ is not decoration. The strip at $x=0.9$ travels nine times as far in one turn as the strip at $x=0.1$, so it sweeps nine times the volume for the same height and thickness. Dropping it gives $2\pi\int_0^1(x-x^{2})dx=\pi/3$, which is twice the true $\pi/6$ — and worse, it would return the same number if the region were slid to sit flush against the axis, where the true volume is different.
The arch under y = x − x² turned about the vertical axis
The problem from the opening, now with the strip running the other way.
expanding first was cheaper than any substitution here: two monomials, no rule needed
$=2\pi\cdot\frac{1}{12}=\frac{\pi}{6}$
$\frac13-\frac14=\frac{1}{12}$
Answer $$V=\frac{\pi}{6}$$
Check
Independent check by the other method: the washer set-up from the opening collapses, because $\bigl(\tfrac{1+s}{2}\bigr)^{2}-\bigl(\tfrac{1-s}{2}\bigr)^{2}=s$ with $s=\sqrt{1-4y}$, leaving $\pi\int_0^{1/4}\sqrt{1-4y}\,dy=\pi\cdot\frac16$. Two different slicings, same $\pi/6$.
One integral, two monomials, no substitution — against one quadratic-formula inversion and a square root inside a square in the other direction.
When the boundary is handed to you as $y=f(x)$, vertical strips keep it in that form. The method that needs no inversion is usually the one to reach for first.
The region under y = √x on [0, 4] turned about the vertical axis
A second one, chosen because it can be checked the other way exactly.
Given
Region: $0\le y\le\sqrt{x}$ for $0\le x\le 4$
Axis of rotation: the vertical axis
Find
the volume of the solid
SolutionBuild the element
$dV=2\pi\,x\,\sqrt{x}\,dx=2\pi\,x^{3/2}\,dx$
writing $x\sqrt x$ as a single power is what makes the antiderivative a one-liner
power rule with $n=3/2$, so the new exponent is $5/2$ and the constant is $2/5$
$4^{5/2}=\bigl(\sqrt4\bigr)^{5}=2^{5}=32$
half powers are cheaper through the square root than through a calculator, and this is the step people slip on
$V=2\pi\cdot\frac{2}{5}\cdot 32=\frac{128\pi}{5}$
collecting the constants
Answer $$V=\frac{128\pi}{5}$$
Check
Independent check by washers: at height $y\in[0,2]$ the region runs from $x=y^{2}$ out to $x=4$, so $V=\pi\int_0^2\bigl(16-y^{4}\bigr)dy=\pi\bigl(32-\tfrac{32}{5}\bigr)=\tfrac{128\pi}{5}$. Same number from a slicing in the other direction.
$4^{5/2}=32$, not $16$. Half powers are the single most reliable source of a wrong final number in this section; evaluate them through the square root, never by pattern.
Checkpoint
§11.1 — the three factors, and only those●●○○○
Thirty seconds, no evaluation. The region below is bounded above by a curve and below by the horizontal axis, and it is being turned about the vertical axis.
Given
Region: $0\le y\le x^{2}$ for $0\le x\le 3$
Axis of rotation: the vertical axis
Find
Which integral gives the volume?
Hint 1/4
Do not compute. Just name the radius, the height and the thickness of one vertical strip, and check which variable the thickness names.
Hint 2/4
Shell element: $dV=2\pi\,(\text{distance to the axis})\times(\text{height})\times(\text{thickness})$, integrated over the range of the thickness variable.
Hint 3/4
Here the strip at $x$ has distance $x$ to the vertical axis and height $x^{2}$, and the thickness is $dx$ with $x$ running from $0$ to $3$.
Hint 4/4
So the integral is $2\pi\int_0^3 x\cdot x^{2}dx=\dfrac{81\pi}{2}$.
the limits are the range of $x$, because $dx$ is the thickness
Answer $$V=\frac{81\pi}{2}$$
Check
Bound: the solid fits inside the cylinder of radius $3$ and height $9$, volume $81\pi$, and it fills exactly half of it. Half of a cylinder is a believable share for a region that hugs the outer wall.
The limits belong to the thickness variable. If your $dx$ integral runs to $9$, you have used the range of $y$.
⚠ Dropping the radius factor
the height and the $2\pi$ are the two visible parts of the formula, and $x$ looks like a variable of integration rather than a length
⚠ Integrating over the range of the wrong variable
the numbers $0$ and $9$ are both visible in the picture, and only the thickness says which pair is meant
wrong$V=2\pi\int_0^9 x\cdot x^{2}\,dx$
right$V=2\pi\int_0^3 x\cdot x^{2}\,dx$
When the axis moves, the radius is a distance
Everything so far quietly assumed the axis was the vertical axis, so distance to the axis and the coordinate $x$ were the same number. Move the axis three units right and they part company.
RuleShell method about a vertical line
Conditions
the axis of rotation is the vertical line $x=c$
the region lies entirely on one side of that line
$h(x)$ is the height of the strip at $x$, measured top boundary minus bottom boundary
Every strip still contributes circumference times height times thickness; only the circumference changed, because the strip now travels around a line that is $\lvert x-c\rvert$ away instead of $x$ away.
The strip at $\textcolor{#d1690a}{x}$ turning about $\textcolor{#6f42c1}{x=3}$. Its radius is the gap $\textcolor{#d1690a}{3-x}$, which shrinks to zero as the strip walks towards the axis — exactly what a radius should do.
Looks like this, but is not
For the region under $y=x^{2}$ on $[0,2]$ turned about $x=3$: $V=2\pi\displaystyle\int_0^2\bigl(x-3\bigr)x^{2}\,dx$. The $3$ is in there, the height is right, and signs are only bookkeeping.
Evaluate it: $2\pi\bigl[\tfrac{x^{4}}{4}-x^{3}\bigr]_0^2=2\pi(4-8)=-8\pi$. A negative volume. Signs are not bookkeeping when the quantity is a length: on $[0,2]$ the expression $x-3$ is negative everywhere, so it was never a radius. Put an endpoint in before you integrate — at $x=0$ the radius must be $3$, and $0-3$ is not $3$.
The region under y = x² on [0, 2] turned about the line x = 3
The axis has moved off the region entirely, which is the whole difficulty.
Given
Region: $0\le y\le x^{2}$ for $0\le x\le 2$
Axis of rotation: the vertical line $x=3$
Find
the volume of the solid
SolutionFix the radius, and test it
$\text{radius}(x)=3-x$
the region sits to the left of the axis, so the distance is axis minus coordinate
$\text{radius}(0)=3,\qquad \text{radius}(2)=1$
both positive, and both agree with what you can measure off the picture: this is the whole check
Independent check by washers in $y$: at height $y\in[0,4]$ the region runs from $x=\sqrt y$ to $x=2$, so about $x=3$ the outer radius is $3-\sqrt y$ and the inner radius is $1$, giving $\pi\int_0^4\bigl[(3-\sqrt y)^{2}-1\bigr]dy=\pi\bigl[8y-4y^{3/2}+\tfrac{y^{2}}{2}\bigr]_0^4=\pi(32-32+8)=8\pi$.
Test the radius at one endpoint before integrating. It costs three seconds and it catches the only error this set-up really invites.
The same region, turned about the line x = −1
One number changed, and with it the order of the subtraction.
Given
Region: $0\le y\le x^{2}$ for $0\le x\le 2$
Axis of rotation: the vertical line $x=-1$
Find
the volume, and one sentence on how it compares with the previous answer
SolutionFix the radius
$\text{radius}(x)=x-(-1)=x+1$
now the region is to the right of the axis, so it is coordinate minus axis
$\text{radius}(0)=1,\qquad \text{radius}(2)=3$
positive at both ends, and the near end of the region is now the far end of the previous problem
$4+\tfrac83=\tfrac{20}{3}$, and doubling with the $2$ gives $\tfrac{40}{3}$
Answer $$V=\frac{40\pi}{3}$$
Check
Comparison as a check: $\tfrac{40}{3}\approx 13.3$ against $8$ for the axis $x=3$. The tall part of the region, near $x=2$, is now the far part, and the far part is the part that sweeps the most volume — so the answer had to come out larger.
Which end of the region is far from the axis decides the size of the answer. Predicting that before computing turns the arithmetic into a check instead of a guess.
Checkpoint
§11.2 — naming the radius●●○○○
Thirty seconds. No integral to write, no arithmetic; the question is only what one length is.
Given
Region: $0\le y\le\sqrt{x}$ for $0\le x\le 4$
Axis of rotation: the vertical line $x=6$
Find
What is the shell radius of the strip at position $x$?
Hint 1/4
Forget the formula for a moment. Point at the strip at $x=1$ and at the line $x=6$, and say how far apart they are.
Hint 2/4
The shell radius is the distance from the strip to the axis of rotation, and a distance is never negative on the interval you are integrating over.
Hint 3/4
Here the region occupies $0\le x\le 4$ and the axis is at $6$, so every strip is to the left of it. Distance is the axis coordinate minus the strip coordinate.
Hint 4/4
The radius is $6-x$.
Show solutionWhich side is the region on?
$0\le x\le 4<6$
every point of the region is to the left of the axis, so the axis coordinate is the larger one
$\text{radius}=6-x$
distance is larger minus smaller, and this is the order that keeps it positive here
Test the ends
$6-0=6,\qquad 6-4=2$
both positive and both readable off the picture, which is the only check this step needs
Answer $$\text{radius}=6-x$$
Check
Extreme case: if the axis were moved to $x=4$, the formula would give a radius of $0$ at the right edge of the region, which is exactly what touching the axis should give.
Radius equals axis minus coordinate when the region is to the left, coordinate minus axis when it is to the right. Never memorise which — test an endpoint.
⚠ Using the coordinate as the radius after the axis has moved
every example up to this point had the axis through the origin, where the two happen to agree
Both methods are correct for every region, so the choice is never about being right; it is about which set-up leaves you an integral you can finish.
MethodChoosing the slice direction
Conditions
the strip meets the same pair of boundary curves along the whole interval; otherwise split the region first
the boundary is available as a function of the thickness variable
$$\boxed{\;\text{strip} \parallel \text{axis}\;\Rightarrow\;\text{shells};\qquad \text{strip} \perp \text{axis}\;\Rightarrow\;\text{disks or washers}\;}$$
If the thin rectangle you drew runs the same way as the axis, it wraps around the axis and becomes a shell; if it runs across the axis, it sweeps out a flat disk.
The whole decision in one place. Note what the diagram never asks: which axis it is. Only the direction of the strip relative to the axis matters, and the limits always belong to the thickness variable.
Looks like this, but is not
The summary that circulates before every midterm: rotate about the vertical axis and use shells, rotate about the horizontal axis and use washers.
It is only true for regions handed to you as $y=f(x)$, which is most but not all of them. Turn the region $0\le y\le 2$, $y^{2}\le x\le 4$ about the horizontal axis and horizontal strips are parallel to that axis, so they give shells about a horizontal axis. Turn the region under $y=\sqrt x$ about the vertical axis with horizontal strips and you get washers. The axis alone decides nothing; the axis together with the direction of your strip decides everything.
The region between y = x and y = x³ turned about each axis in turn
Same region, two rotations, and the cheaper method is a different one each time.
Given
Region: between $y=x$ (above) and $y=x^{3}$ (below), $0\le x\le 1$
Rotation (i): about the vertical axis. Rotation (ii): about the horizontal axis
Find
the volume in each case, using the cheaper method
SolutionRotation (i): vertical axis, so vertical strips are parallel
$\text{radius}=x,\qquad h(x)=x-x^{3}$
top minus bottom, and on $[0,1]$ the line is above the cubic
Consistency check without recomputing: both solids come from the same sliver of area, but the vertical-axis rotation moves that sliver on wider circles — its far edge sits a full unit from the axis, while for the horizontal axis the far edge is the same unit but the near edge is much closer. So $V_{1}>V_{2}$ was forced, and $\tfrac{4}{15}>\tfrac{4}{21}$ agrees.
Two set-ups, four monomials, no inversion of any boundary in either direction.
Neither method belongs to an axis. In both parts the strip was vertical; what changed was whether the axis ran along it or across it.
The region under y = x² on [0, 2] turned about the horizontal axis, using shells
Shells about a horizontal axis, on purpose, so that the answer can be checked against the method you would normally have used.
Given
Region: $0\le y\le x^{2}$ for $0\le x\le 2$
Axis of rotation: the horizontal axis
Find
the volume, by horizontal strips
SolutionTurn the region on its side
$\text{strip at height } y,\ 0\le y\le 4$
horizontal strips run the same way as the horizontal axis, so they are the ones that give shells
$\text{radius}=y,\qquad \text{height}=2-\sqrt{y}$
the strip runs from the parabola $x=\sqrt y$ across to the right edge $x=2$; right minus left
$4^{5/2}=32$, so the second term is $\tfrac{2}{5}\cdot 32=\tfrac{64}{5}$
$=2\pi\cdot\frac{16}{5}=\frac{32\pi}{5}$
$16-\tfrac{64}{5}=\tfrac{16}{5}$
Answer $$V=\frac{32\pi}{5}$$
Check
Independent check by the method you would normally use: vertical strips are across the horizontal axis, giving disks of radius $x^{2}$, so $V=\pi\int_0^2 x^{4}dx=\pi\cdot\tfrac{32}{5}$. The same $\tfrac{32\pi}{5}$ from a slicing in the perpendicular direction.
Here the disk route is plainly cheaper — one monomial against two and a half power. That is the point of running both: the rule tells you which is which before you commit.
Turning the picture ninety degrees turns shells into washers and back. The formulas do not know which axis is which; they only know parallel and across.
Checkpoint
§11.3 — which set-up avoids inverting the boundary●●●○○
Thirty seconds. The boundary below is a genuine obstacle: $y=x^{5}+x$ cannot be solved for $x$ in closed form, so any method that starts by doing so has already failed.
Given
Region: $0\le y\le x^{5}+x$ for $0\le x\le 1$
Axis of rotation: the vertical axis
Find
Which set-up never needs $x$ written in terms of $y$?
Hint 1/4
Do not evaluate anything. Ask only which direction of strip lets you use the boundary exactly as it is written.
Hint 2/4
A strip parallel to the axis gives shells; a strip across the axis gives disks or washers, and those need the boundary as a function of the other variable.
Hint 3/4
The axis here is vertical, so a vertical strip is parallel to it. The boundary is given as $y=x^{5}+x$, a function of $x$, and shells integrate in $x$.
Bound: the region sits under $y=2$ inside $0\le x\le 1$, so the solid fits in a cylinder of radius $1$ and height $2$, volume $2\pi\approx 6.28$. The answer $\tfrac{20\pi}{21}\approx 2.99$ is under half of that, which fits a region that thins out towards the axis.
The question which method is cheaper often has a one-word answer: whichever one keeps the boundary in the form it was given.
⚠ Mixing the two variables inside one integral
the radius is easy to read off the picture in one variable and the height in the other, and nothing in the picture objects
wrong$V=2\pi\int_0^4 y\bigl(2-x^{2}\bigr)dy$
right$V=2\pi\int_0^4 y\bigl(2-\sqrt{y}\bigr)dy$
⚠ Using one curve as the height when the region sits between two
the phrase the region under a curve is so common that the second boundary stops being read
wrong$h(x)=x^{2}\quad\text{for the region between } y=x^{2}\text{ and } y=4$
right$h(x)=4-x^{2}$
The logarithm is an area
The same integral machinery that just measured a solid is what this course uses to define the logarithm — and defining it that way is what turns its algebra from something remembered into something proved.
DefinitionDefinition: the natural logarithm
Conditions
$x>0$
$1/t$ is continuous on the closed interval between $1$ and $x$, which is what makes the integral exist
$\ln x$ is the signed area under $y=1/t$ between $t=1$ and $t=x$: positive when $x$ is to the right of $1$, negative when $x$ is to the left of it, and zero at $x=1$ because the area has not started.
The three consequences that do all the work
Derivative. $1/t$ is continuous on $(0,\infty)$, so the Fundamental Theorem, Part 1, gives $(\ln x)'=1/x$ with no computation. Monotone. That derivative is positive for every $x>0$, so $\ln$ is strictly increasing and therefore one to one. Range. By the multiplication law below, the value at $2^{n}$ is $n\ln 2$, which grows past every bound, and the value at $2^{-n}$ is $-n\ln 2$, which falls below every bound. A continuous strictly increasing function that goes from $-\infty$ to $+\infty$ takes every real value exactly once — which is precisely what the next block needs in order to invert it.
$\textcolor{#d1690a}{\ln 3}$ is not a rule about exponents here; it is the shaded area. Move the right-hand end and you move the value, which is what makes $\textcolor{#1f6feb}{1/t}$ the derivative.
Looks like this, but is not
$\ln 0$ should be the area from $1$ back down to $0$, so $\ln 0=-\infty$ is a value like any other, just an unusually large one.
There is no finite area to report: $\int_x^1 dt/t=-\ln x$ grows past every bound as $x\to 0^{+}$, so no number can be assigned. The correct pair of sentences is that $\ln$ is undefined at $0$, and separately that $\ln x\to-\infty$ as $x\to 0^{+}$. The first is about the domain, the second about behaviour, and writing $\ln 0=-\infty$ silently promotes a limit into a value.
$x$
$\ln x$
gain over the row above
$1$
$0$
—
$2$
$0.6931$
$0.6931$
$4$
$1.3863$
$0.6931$
$8$
$2.0794$
$0.6931$
$16$
$2.7726$
$0.6931$
Each doubling of $x$ adds the same $0.6931$ of area, never more and never less. That constant third column is $\ln(2x)=\ln x+\ln 2$ read off a table — the multiplication law is visible in the numbers before it is proved.
Proving ln(ax) = ln a + ln x from the definition
The one substitution that turns a statement about products into a statement about areas.
additivity of the integral over adjacent intervals; splitting at $a$ rather than anywhere else is the whole idea, because the first piece is already $\ln a$
the definition again, now on the transformed interval
Collect
$\ln(ax)=\ln a+\ln x$
adding the two pieces
Answer $$\ln(ax)=\ln a+\ln x\quad (a,x>0)$$
Check
Numerical check on a case the table above supplies: $a=x=2$ gives $\ln 4=1.3863$ against $\ln 2+\ln 2=2(0.6931)=1.3862$, agreeing to the digits shown.
Every other logarithm law comes from this one: put $x=1/a$ to get $\ln(1/a)=-\ln a$, then $\ln(a/b)=\ln a-\ln b$, then $\ln(x^{n})=n\ln x$ by repeating it.
Trapping ln 2 between two rectangles
The definition is an area, so it can be bounded by hand — no calculator, no series.
Given
$\ln 2=\displaystyle\int_1^2\frac{dt}{t}$
Two subintervals of width $\tfrac12$: $[1,\tfrac32]$ and $[\tfrac32,2]$
Independent check against the table: $\ln 2=0.6931$ does sit between the two bounds, and it is nearer the lower one — which is what a curve bending upwards should do.
Bounds like these are how you tell that an answer such as $\ln 2\approx 1.4$ is wrong without owning a calculator.
Differentiating ln of a square root, and the absolute value case
The derivative rule, and the small extension that makes it usable when the inside can be negative.
Given
$g(x)=\ln\sqrt{x^{2}+1}$
Find
$g'(x)$, and the general rule for $\ln\lvert u(x)\rvert$
$(\ln u)'=u'/u$, which is the Fundamental Theorem result composed with the chain rule
The absolute value version
$\frac{d}{dx}\ln\lvert u\rvert=\frac{u'}{u}$
for $u>0$ this is the rule just used; for $u<0$ write $\ln(-u)$, whose derivative is $(-u')/(-u)$, the same expression
$\int\frac{dx}{x}=\ln\lvert x\rvert+C$
reading the same fact backwards, and the bars are what make it valid on the negative side too
Answer $$g'(x)=\frac{x}{x^{2}+1}$$
Check
Independent check at a point: near $x=2$, $g$ changes by about $g(2.01)-g(1.99)=0.008$ over a step of $0.02$, a slope near $0.4$; the formula gives $2/5=0.4$.
Simplify with the logarithm laws before differentiating. It is not cosmetic — it is what removes one link from the chain, and every removed link is one fewer place to lose a factor.
Checkpoint
§11.4 — reading the definition backwards●●○○○
Thirty seconds. Nothing here needs a logarithm law; the definition on its own decides the answer, including its sign.
Given
$\ln x=\displaystyle\int_1^x\frac{dt}{t}$ for $x>0$
Find
What is $\displaystyle\int_1^{1/5}\frac{dt}{t}$?
Hint 1/4
Look at the two limits before anything else. Which one is bigger, and what does an integral do when you run it in that direction?
Hint 2/4
The definition is $\ln x=\int_1^x dt/t$ for every $x>0$, with no requirement that $x$ exceed $1$; and $\int_a^b=-\int_b^a$.
Hint 3/4
Here the upper limit is $x=\tfrac15$, which is positive, so the integral is exactly $\ln\tfrac15$. And $\ln\tfrac15=\ln 1-\ln 5$.
Hint 4/4
Since $\ln 1=0$, the value is $-\ln 5$.
Show solutionApply the definition as written
$\int_1^{1/5}\frac{dt}{t}=\ln\tfrac15$
the definition allows any positive upper limit, including one below $1$
$\ln\tfrac15=\ln 1-\ln 5=-\ln 5$
from $\ln(ax)=\ln a+\ln x$ with $a=5$, $x=\tfrac15$, together with $\ln 1=0$
Sign check without any law: the integral runs right to left, and $1/t$ is positive there, so the signed area has to be negative. $-1.61$ is negative; $+1.61$ would not be.
A logarithm of a number below $1$ is negative, always, and the definition says why: you are walking backwards under a positive curve.
⚠ Turning a sum inside the logarithm into a sum of logarithms
the multiplication law is remembered as splits into a sum, and the plus sign inside gets matched to the plus sign outside
wrong$\ln(a+b)=\ln a+\ln b$
right$\ln(ab)=\ln a+\ln b$
⚠ Reporting a value of the logarithm outside its domain
the graph appears to run off the bottom of the page near zero, and running off the page looks like arriving somewhere
right$\ln x \text{ is undefined for } x\le 0;\qquad \lim_{x\to 0^{+}}\ln x=-\infty$
The exponential is the logarithm read backwards
The last block ended with three facts about $\ln$: strictly increasing, continuous, and taking every real value. Together those say it has an inverse — and this course builds the exponential out of that inverse rather than the other way round.
DefinitionDefinition: the natural exponential, and its derivative
Conditions
$\ln$ is strictly increasing on $(0,\infty)$, because $(\ln x)'=1/x>0$ there
$\ln$ takes every real value, so its inverse is defined on all of $\mathbb{R}$
$(\ln)'$ is never zero, which is what lets the inverse be differentiated
Read the graph of $\ln$ backwards and you get a function defined for every real input; the number $e$ is the input whose logarithm is $1$; and because the slope of $\ln$ at a point of height $x$ is one over that point, the slope of the mirror image at $x$ is the height itself.
Why the derivative comes out as the function itself
Write $y=e^{x}$; by the inverse relation, $\ln y=x$. Differentiate both sides with respect to $x$. On the left the chain rule gives $y'/y$; on the right the derivative of $x$ is $1$. So $y'/y=1$, that is $y'=y$: the exponential is its own derivative. Nothing was assumed about $e^{x}$ here except that it undoes $\ln$ — the identity is a consequence of the definition, not a separate axiom, and the same one line run on $\ln$ is what produced $1/x$ a moment ago.
One curve and its reflection in $\textcolor{#6f42c1}{y=x}$. The point $\textcolor{#1f6feb}{(1,0)}$ on the logarithm is the point $\textcolor{#d1690a}{(0,1)}$ on the exponential — the same fact $\ln 1=0$ written twice.
Looks like this, but is not
$\dfrac{d}{dx}e^{2x}=e^{2x}$, because the exponential is its own derivative and there is an exponential right there.
The rule is about the function $x\mapsto e^{x}$, not about any expression containing an $e$. $e^{2x}$ is a composition, so the chain rule adds the derivative of the inside: $\frac{d}{dx}e^{2x}=2e^{2x}$. Check it at $x=0$: both $e^{x}$ and $e^{2x}$ pass through $1$, but $e^{2x}$ is $e^{x}$ compressed horizontally by a factor $2$, so it must climb twice as fast there — slope $2$, not $1$.
Getting the derivative of e^x out of the inverse rule
Two lines, and they are the reason this whole construction is worth the detour.
Given
$\exp=\ln^{-1}$, so $\ln\bigl(e^{x}\bigr)=x$ for every real $x$
$(\ln u)'=u'/u$ from the previous block
Find
$\dfrac{d}{dx}e^{x}$, from the definition rather than from memory
SolutionDifferentiate the inverse relation
$y=e^{x}\;\Longrightarrow\;\ln y=x$
this is the definition of the inverse, and it is the only property of $e^{x}$ used anywhere below
Independent check through the inverse-function rule instead: the slope of $\ln$ at the point $e^{x}$ is $1/e^{x}$, and the slope of an inverse is the reciprocal of that, namely $e^{x}$. Two routes, same answer.
The identity is earned, not assumed. That matters on an exam question worded from the definition, where quoting $(e^{x})'=e^{x}$ as a known fact is exactly what is being asked to be proved.
Evaluating the integral of e^(2x) from 0 to 1
The chain rule read backwards, with a bound to check the answer against.
Given
$\displaystyle\int_0^1 e^{2x}\,dx$
Find
the exact value
SolutionSubstitute
$u=2x,\qquad du=2\,dx,\qquad dx=\tfrac12\,du$
the inside of the exponential is the natural choice; the missing factor $2$ is supplied by the constant, which is what makes this cheaper than guessing an antiderivative
Independent bound: on $[0,1]$ the integrand runs from $1$ up to $e^{2}\approx 7.39$, so the integral lies between $1$ and $7.39$; and since $e^{2x}$ bends upwards, it must also come in below the trapezoid value $\tfrac{1+7.39}{2}\approx 4.19$. The answer $3.195$ satisfies both.
Bending upwards puts the true value below the trapezoid estimate and above the midpoint one. That single fact turns almost any exponential integral into something you can sanity check in ten seconds.
Checkpoint
§11.5 — what makes the cancellation legal●●●○○
Thirty seconds. Both of the identities below are true, but only one of them carries a restriction, and the question is which.
Which statement about the two identities is correct?
Hint 1/4
Do not check either identity numerically yet. Just ask, for each one, which of the two functions is applied first.
Hint 2/4
$\ln$ accepts only positive inputs and returns any real number; $\exp$ accepts any real number and returns only positive ones.
Hint 3/4
In $e^{\ln x}$ the first function applied is $\ln$, so the input $x$ must be positive. In $\ln(e^{x})$ the first is $\exp$, which accepts every real $x$ and hands $\ln$ a positive number.
Hint 4/4
So A needs $x>0$ and B holds everywhere.
Show solutionIdentity A starts with a logarithm
$x\xrightarrow{\ \ln\ } \ln x \xrightarrow{\ \exp\ } x$
the first step is $\ln$, whose domain is $(0,\infty)$, so $x$ must be positive before anything else happens
Take the logarithm of both sides, differentiate implicitly — which turns the left-hand side into $y'$ over $y$ — then multiply back by $y$ written out in $x$.
Three cases and only one of them is new. The two familiar rules each hold one of base and exponent fixed; when neither is fixed, no rule applies until a logarithm has been taken.
Looks like this, but is not
$\dfrac{d}{dx}x^{x}=x\cdot x^{x-1}=x^{x}$, by the power rule, since the exponent is right there and so is the base.
The power rule proves itself with the exponent held fixed, and the rule for a constant base holds the base fixed; here neither is fixed, so neither proof applies. Logarithmic differentiation gives $\frac{d}{dx}x^{x}=x^{x}(\ln x+1)$. The two answers happen to agree at $x=1$, which is why the error survives, so test at $x=2$: the true slope is $4(\ln 2+1)\approx 6.77$ and the fake one is $4$.
The derivative of (sin x)^x on (0, π)
Both the base and the exponent contain $x$, which is exactly the case with no rule.
Given
$y=(\sin x)^{x}$ for $00$
Find
$y'$ as a single expression in $x$, and its value at $x=\pi/2$
SolutionTake the logarithm of both sides
$\ln y=x\ln(\sin x)$
the power law moves the exponent down to the front, which is the move that makes the expression differentiable at all
$\sin x>0 \text{ on } (0,\pi)$
the restriction on the interval is what makes both logarithms defined; outside it the method needs $\ln\lvert y\rvert$
Independent check on the value at $\pi/2$, without differentiating anything: on $(0,\pi)$ we have $0<\sin x\le 1$ with equality only at $\pi/2$, and the exponent $x$ is positive, so $(\sin x)^{x}\le 1$ with equality only at $\pi/2$. That is an interior maximum of a differentiable function, and at an interior maximum the derivative must vanish.
One logarithm, one product rule, one chain rule, and one multiplication back — four steps, and the last one is the one that gets dropped.
A zero derivative arriving from two separate vanishing terms is worth a second look, and here the symmetry of the picture confirms it independently.
The derivative of a four-factor quotient, without the quotient rule
No variable exponent this time; the method earns its place by turning one messy quotient into three easy pieces.
Given
$\displaystyle y=\frac{x^{2}\sqrt{x^{2}+1}}{(x+3)^{5}}$ for $x>0$
Independent check at a point: at $x=1$ the formula gives $y(1)=\tfrac{\sqrt2}{1024}$ times $\bigl(2+\tfrac12-\tfrac54\bigr)=\tfrac54$, about $0.00173$; a difference quotient with a step of $10^{-3}$ on the original expression returns $0.00173$ as well.
One logarithm and three one-line derivatives, against a quotient rule wrapped around a product rule wrapped around a chain rule.
Every minus sign in the bracket comes from a factor that was in the denominator, and every coefficient from an exponent. That correspondence is the fastest way to check the bracket before multiplying back.
Checkpoint
§11.6 — the first line of the method●●○○○
Thirty seconds, one line of work. No derivative is required here, only the line that has to be written before one is possible.
Given
$y=x^{\sqrt{x}}$ for $x>0$
Find
What is $\ln y$?
Hint 1/4
You are not differentiating yet. The only question is what the logarithm law does to an exponent, and whether it cares that the exponent contains $x$.
Hint 2/4
Power law for logarithms: $\ln(x^{r})=r\ln x$ for $x>0$, for any exponent $r$, constant or not.
Hint 3/4
Here the base is $x$ and the exponent is $\sqrt{x}$, so the law brings $\sqrt{x}$ down in front of $\ln x$.
Hint 4/4
So $\ln y=\sqrt{x}\,\ln x$.
Show solutionApply the power law
$\ln\bigl(x^{r}\bigr)=r\ln x \text{ for } x>0$
the law does not care whether the exponent $r$ is a constant or an expression in $x$, because it was proved for a fixed value of $x$ at a time
$\ln y=\sqrt{x}\,\ln x$
with $r=\sqrt x$
Answer $$\ln y=\sqrt{x}\ln x$$
Check
Numerical check: at $x=4$, $y=4^{2}=16$ and $\ln 16\approx 2.7726$, while $\sqrt4\ln 4=2(1.3863)=2.7726$.
The exponent comes down as a multiplier, never as a second logarithm. That single move is the entire reason the method works.
⚠ Stopping at y prime over y
the hard part is finished by then, and the left-hand side looks like an answer because it contains $y'$
⚠ Turning a product inside the logarithm into a product of logarithms
the aim of the step is to break a product apart, and multiplying the pieces looks like the way to keep the equation true
wrong$\ln\bigl(f\cdot g\bigr)=\ln f\cdot\ln g$
right$\ln\bigl(f\cdot g\bigr)=\ln f+\ln g$
Setting up any shell integral in five moves
Every rotation problem where the strip you want to draw runs the same way as the axis — which is most of them when the boundary is given as $y=f(x)$ and the axis is vertical, or as $x=g(y)$ and the axis is horizontal.
Draw the region and the axis, and mark the side.
Shade the region and draw the axis of rotation as a line, even when it is one of the coordinate axes. Say out loud which side of that line the region is on; the next two steps both depend on the answer.
Draw one strip parallel to the axis.
Its thickness names your variable of integration. Vertical strip means $dx$; horizontal strip means $dy$. From here on, every quantity you write must be in that one variable.
Radius = distance from the strip to the axis.
Write it so that it is non-negative on the whole interval, then test it at one endpoint against the picture. For a vertical axis $x=c$ it is $c-x$ when the region is on the left and $x-c$ when it is on the right.
Height = far boundary minus near boundary.
Top minus bottom for a vertical strip, right minus left for a horizontal one. If the region is bounded by two curves, both of them appear here; if the boundary changes at some point, split the region there and add two integrals.
Limits = the range of the thickness variable.
Never the range of the other one. Then write $V=2\pi\int(\text{radius})(\text{height})\,d(\text{thickness})$ and expand the product before integrating: at this level it is almost always two or three monomials.
Where it goes wrong
The radius is written as the coordinate after the axis has moved, which is invisible until the answer comes out negative or too small.
The limits are taken from the other variable, usually because both ranges are visible in the picture and only the thickness says which is meant.
The height uses one curve where the region sits between two, which turns a correct method into a wrong region.
The region straddles the axis. Then the shell radius is not a single expression; split at the axis and treat each side separately.
Logarithmic differentiation, step by step
When the base and the exponent both contain the variable, and also whenever a derivative would otherwise need a quotient rule wrapped around a product rule wrapped around a chain rule.
Check positivity, or use bars.
Say on which interval $y>0$. If $y$ can be negative, work with $\ln\lvert y\rvert$ instead; the derivative formula that comes out is identical, because $\frac{d}{dx}\ln\lvert u\rvert=u'/u$ either way.
Take the logarithm and break it apart.
Apply the three laws once each: products become sums, quotients become differences, exponents come down as coefficients. The right-hand side should end up as a sum of separate terms with no fraction left inside a logarithm.
Differentiate both sides implicitly.
The left-hand side becomes $y'/y$ by the chain rule. Every term on the right is of the form $u'/u$, and none of them interacts with any other.
Multiply back by $y$, written out in $x$.
This is the step that gets dropped. $y'/y$ is a relative rate, not a derivative; the answer must contain the original expression as a factor.
Where it goes wrong
Stopping at $y'/y$ and calling it $y'$.
Splitting $\ln(f\cdot g)$ into $\ln f\cdot\ln g$, or $\ln(f+g)$ into anything at all.
Losing a minus sign on a factor that came from the denominator.
Using the method on an interval where the expression is negative without switching to $\ln\lvert y\rvert$.
Same region, axis at x = 5
The region lies entirely to the left of the axis.
Given
Region: $0\le y\le\sqrt{x}$ for $0\le x\le 4$
Axis: the vertical line $x=5$
Find
the volume
SolutionRadius, tested at both ends
$\text{radius}=5-x,\qquad 5-0=5,\ 5-4=1$
the region is to the left, so the axis coordinate is the larger one; both test values are positive
Independent bound: the region has area $\int_0^4\sqrt{x}\,dx=\tfrac{16}{3}$, and every strip of it is between $1$ and $5$ units from the axis, so the volume lies between $2\pi(1)\tfrac{16}{3}=\tfrac{32\pi}{3}\approx 33.5$ and $2\pi(5)\tfrac{16}{3}=\tfrac{160\pi}{3}\approx 167.6$. The answer $87.1$ is inside.
Region on the left: radius is axis minus coordinate.
Same region, axis at x = −1
One number changed, and the order of the subtraction changes with it.
Given
Region: $0\le y\le\sqrt{x}$ for $0\le x\le 4$
Axis: the vertical line $x=-1$
Find
the volume
SolutionRadius, tested at both ends
$\text{radius}=x+1,\qquad 0+1=1,\ 4+1=5$
now the region is to the right of the axis, so it is coordinate minus axis, and the two test values have swapped ends
$\tfrac25\cdot 32=\tfrac{64}{5}$ and $\tfrac23\cdot 8=\tfrac{16}{3}$
$=2\pi\cdot\frac{272}{15}=\frac{544\pi}{15}$
$\tfrac{64}{5}+\tfrac{16}{3}=\tfrac{192+80}{15}$
Answer $$V=\frac{544\pi}{15}\approx 113.9$$
Check
Two checks. The same bound applies, $\tfrac{32\pi}{3}\le V\le\tfrac{160\pi}{3}$, and $113.9$ is inside it. And this answer had to exceed the previous one: for every $x>2$ the new radius $x+1$ beats the old radius $5-x$, and $x>2$ is exactly where the region is tallest.
Region on the right: radius is coordinate minus axis.
Identical region, identical height function, identical limits — the only difference is which of the two numbers is subtracted from which, and that difference is worth $\tfrac{128\pi}{15}$ of volume.
How to tell them apart
Put an endpoint of the interval into your radius expression before integrating. If it returns the distance you can measure off the picture, the order is right; if it returns a negative number, reverse it. There is also a check that uses both answers at once: the two radii add to the constant $(5-x)+(x+1)=6$, so the two volumes must add to $2\pi\cdot 6\int_0^4\sqrt{x}\,dx=12\pi\cdot\tfrac{16}{3}=64\pi$ — and $\tfrac{416\pi}{15}+\tfrac{544\pi}{15}=\tfrac{960\pi}{15}=64\pi$.
Scaffolding comes off
The common skeleton
Draw the region and the axis; say which side of the axis the region is on.
Draw one strip parallel to the axis and name its thickness variable.
Write the radius as a distance to the axis, and test it at one endpoint.
Write the height as far boundary minus near boundary, in the same variable.
Integrate $2\pi(\text{radius})(\text{height})$ over the range of the thickness variable, expanding the product first.
1 · fully worked
The region under y = 4x − x² turned about the vertical axis
Every step written out, with the reason beside it. Read it once, then cover the right-hand column and read it again.
Given
Region: $0\le y\le 4x-x^{2}$
Axis: the vertical axis
Find
the volume of the solid
SolutionWhere does the region live
$4x-x^{2}=x(4-x)\ge 0 \iff 0\le x\le 4$
the limits are not given in the question; they come from where the curve is above the horizontal axis
Radius and height
$\text{radius}=x,\qquad \text{height}=4x-x^{2}$
the axis is the vertical axis itself, so distance to it is just the coordinate; the region is bounded below by $y=0$
$\text{radius}(0)=0,\ \text{radius}(4)=4$
the endpoint test: the strip at the left touches the axis and the strip at the right is four units out, both readable off the picture
$4^{3}=64$ so the first term is $\tfrac{256}{3}$; $4^{4}=256$ so the second is $64$
$=2\pi\cdot\frac{64}{3}=\frac{128\pi}{3}$
$\tfrac{256}{3}-64=\tfrac{256-192}{3}$
Answer $$V=\frac{128\pi}{3}\approx 134$$
Check
Bound as an independent check: the region fits inside the rectangle $0\le x\le 4$, $0\le y\le 4$, whose rotation about the vertical axis is a cylinder of volume $\pi(4)^{2}(4)=64\pi\approx 201$. The parabolic region fills about two thirds of that rectangle, and $\tfrac{128\pi}{3}\approx 134$ is two thirds of $201$.
The limits came from the region, not from the question. Whenever a rotation problem gives you a curve and the horizontal axis and nothing else, solving $f(x)=0$ is step zero.
2 · you write the reasoning
Easier on purpose: the numbers are smaller and there is nothing to factor. The steps are already written; your job is the reason column. Say why each line is allowed before opening the one below it. Turn the region bounded by $y=x$, $y=0$ and $x=3$ about the vertical axis.
$\text{radius}=x$, $\text{height}=x$, with $0\le x\le 3$
reasoning
The axis is the vertical axis, so the distance from the strip at $x$ to it is $x$ itself. The height is top minus bottom, and here the top is the line $y=x$ while the bottom is $y=0$, so it happens to equal $x$ as well — two different quantities that agree by accident, which is exactly the kind of coincidence to name out loud rather than rely on.
$V=2\pi\displaystyle\int_0^3 x\cdot x\,dx$
reasoning
The thickness is $dx$, so the limits must be the range of $x$, and the region exists for $0\le x\le 3$: at the left the triangle comes to a point, at the right it is cut off by the vertical line $x=3$.
$x\cdot x=x^{2}$, whose antiderivative is $x^{3}/3$ by the power rule; at $x=3$ that is $27/3=9$ and at $x=0$ it is $0$.
$V=18\pi$
reasoning
Multiplying out the constants. The answer keeps $\pi$ as a symbol because the question asked for a volume, not a decimal.
3 · find the buried error
Harder than the one above, and this time the work is done for you — badly. Exactly two of the four steps below contain an error. Find both. The problem: the region bounded by $y=x^{2}$, $y=4$ and the vertical axis (that is, the part of the strip between the parabola and the horizontal line $y=4$, with $x\ge 0$) is turned about the vertical axis. Find the volume.
Step 1. The strip at $x$ is vertical, so $\text{radius}=x$; the curve bounding the region is $y=x^{2}$, so $\text{height}=x^{2}$.
Step 2. The region runs from $0$ to $4$, so $V=2\pi\displaystyle\int_0^4 x\cdot x^{2}\,dx$.
The height is written as $x^{2}$. The region sits between the parabola and the line $y=4$, so the height of the strip at $x$ is top minus bottom, that is $4-x^{2}$.
The phrase the region bounded by $y=x^{2}$ trains the eye on the curve, and the horizontal boundary stops being read as a boundary at all. It is the single most repeated set-up error of the week.
right
$\text{height}=4-x^{2}$, which is $4$ at the axis and $0$ where the parabola meets the line.
⚠ step 2
The limits $0$ to $4$ are the range of $y$, not of $x$. The thickness is $dx$, and the strips exist only for $0\le x\le 2$, because the parabola reaches $y=4$ at $x=2$.
Both $4$ and $2$ are visible in the picture, and $4$ is the number the question mentioned out loud. Nothing objects until the answer is checked against a bound.
right
$0\le x\le 2$, from solving $x^{2}=4$ with $x\ge 0$.
4 · the bare problem
§11.1 — the bare problem●●●○○
No scaffolding this time. The same five moves, a region you have not seen before, and one number as the answer.
Given
Region: bounded above by $y=2x-x^{2}$ and below by $y=0$
Axis of rotation: the vertical axis
Find
Find the volume of the solid of revolution.
Hint 1/4
Before any formula: the question gives a curve and the horizontal axis but no interval. Where does this region actually start and stop?
Hint 2/4
Shell method about the vertical axis: $V=2\pi\int_a^b(\text{radius})(\text{height})\,dx$, with the radius being the distance from the strip to the axis and the height being top minus bottom.
Hint 3/4
Here $2x-x^{2}=x(2-x)$ is non-negative exactly on $[0,2]$; the radius is $x$ and the height is $2x-x^{2}$, so the integral is $2\pi\int_0^2 x(2x-x^{2})dx$.
Hint 4/4
That evaluates to $2\pi\bigl(\tfrac{16}{3}-4\bigr)=\tfrac{8\pi}{3}$.
Show solutionFind the interval
$2x-x^{2}=x(2-x)\ge 0\iff 0\le x\le 2$
the region is where the parabola is above the horizontal axis, and the question left this for you to find
Radius and height
$\text{radius}=x,\qquad \text{height}=2x-x^{2}$
the axis is the vertical axis, so distance to it is the coordinate; bottom boundary is $y=0$
Bound: the region fits inside the rectangle $0\le x\le 2$, $0\le y\le 1$, which sweeps a cylinder of volume $\pi(2)^{2}(1)=4\pi\approx 12.6$. The answer $\tfrac{8\pi}{3}\approx 8.4$ is two thirds of that, the usual share for a parabolic arch.
Two of the five moves — finding the interval and testing the radius — are the ones a bare problem still expects you to do without being asked.
Full exam-style question
One region, three axes: a full midterm-length volume questionexam format
The long question of this material in the shape it is usually asked: one region, several rotations, and a part that asks you to justify the method rather than to execute it. Work it on paper before opening the solution.
Given
Region $R$: bounded by $y=x^{3}$ on the right, $y=8$ above, and the vertical axis on the left
(a) the volume when $R$ is turned about the vertical axis; (b) the volume when $R$ is turned about the horizontal axis; (c) the volume when $R$ is turned about the line $x=2$; (d) one sentence for each part saying which method you chose and why
Solution(a) About the vertical axis — strips parallel, so shells
the boundary is given as $y=x^{3}$, a function of $x$; washers here would need $x=y^{1/3}$, which is available but adds a fractional power for no gain
$\text{(b) washers}$
vertical strips are across the horizontal axis, and the two radii are constants and a single power — shells would need $x$ in terms of $y$ and give a $y^{1/3}$ integrand
$\text{(c) shells}$
same strip direction as (a), only the radius changed; washers about $x=2$ would need both boundaries as functions of $y$ and a squared bracket
Independent check on (a) by the method that was not used: horizontal slices give disks of radius $x=y^{1/3}$, so $V_{a}=\pi\int_0^8 y^{2/3}dy=\pi\bigl[\tfrac35 y^{5/3}\bigr]_0^8=\pi\cdot\tfrac35\cdot 32=\tfrac{96\pi}{5}$, matching. And on (c) without integrating: the radii of (a) and (c) add to $x+(2-x)=2$, so $V_{a}+V_{c}$ must equal $2\pi\cdot 2\int_0^2(8-x^{3})dx=4\pi(16-4)=48\pi$; indeed $\tfrac{96\pi}{5}+\tfrac{144\pi}{5}=\tfrac{240\pi}{5}=48\pi$.
Three set-ups, eleven monomials, one method decision per part — and the only genuinely new work between (a) and (c) is a single subtraction.
Two rotations of the same region about parallel axes are linked: if the two radii add to a constant $k$, the two volumes add to $2\pi k$ times the area of the region. That gives you a free check on the second of any such pair.
Practice
A · concept 4 questions
1§11.3 — does the axis decide the method●●○○○
A sentence that circulates on every revision sheet before the second midterm. Decide whether it is true as stated, and be ready to say what it should say instead.
Given
Claim: rotating a region about the vertical axis always calls for the shell method.
Find
True or false, and give the corrected version in one sentence if it is false.
Hint 1/4
Do not look for a proof. A single region that can be handled both ways is enough to settle a claim that says always.
Hint 2/4
The deciding rule is that a strip parallel to the axis becomes a shell and a strip across the axis becomes a disk or a washer.
Hint 3/4
Take the region under $y=\sqrt{x}$ on $[0,4]$ and the vertical axis. Horizontal strips run across that axis, so they give washers with outer radius $4$ and inner radius $y^{2}$, over $0\le y\le 2$.
Hint 4/4
So the claim is false: with horizontal strips the same rotation is a washer problem.
Show solutionProduce a counterexample
$\text{region under } y=\sqrt{x},\ 0\le x\le 4,\ \text{vertical axis}$
pick a region whose boundary can be written both ways, so that both methods are genuinely available
the axis appears in this rule only in comparison with the strip, never on its own
Answer $$\text{False; the strip direction decides, not the axis.}$$
Check
Second counterexample in the other direction, to show the claim fails both ways: the region $0\le y\le 2$, $y^{2}\le x\le 4$ turned about the horizontal axis has horizontal strips parallel to that axis, so it gives shells about a horizontal axis.
Any claim of the form axis A implies method B is missing the variable that actually matters. Add and I slice this way to it and it becomes true.
2§11.1 — what the radius factor is doing there●●●○○
Two regions with exactly the same shape and area, placed at different distances from the axis. The question is what the shell formula predicts about their volumes, before any integral is written.
Given
Region $P$: the unit square $0\le x\le 1$, $0\le y\le 1$
Region $Q$: the unit square $10\le x\le 11$, $0\le y\le 1$
Both are turned about the vertical axis
Find
Which statement about the two volumes is correct?
Hint 1/4
No integration is needed to compare them. Ask how far each square travels in one full turn, and how that compares between the two.
Hint 2/4
The shell element is $2\pi(\text{radius})(\text{height})(\text{thickness})$, so at equal height and thickness the volume swept is proportional to the distance from the axis.
Hint 3/4
$P$ occupies $0\le x\le 1$ with average radius $\tfrac12$, while $Q$ occupies $10\le x\le 11$ with average radius $\tfrac{21}{2}$; heights and widths are both $1$.
the average radius went from $\tfrac12$ to $\tfrac{21}{2}$, a factor of $21$, and the volume followed it exactly
Answer $$V_{P}=\pi,\qquad V_{Q}=21\pi$$
Check
Independent check by geometry: $Q$ sweeps a cylindrical ring of inner radius $10$, outer radius $11$ and height $1$, whose volume is $\pi(11^{2}-10^{2})(1)=21\pi$. No integral needed.
Area alone tells you nothing about a volume of revolution. Distance from the axis is half the answer, and that is the whole reason the radius appears as a factor rather than as a limit.
3§11.4 — a logarithm law and its domain●●●○○
The law itself is not in doubt; what is in doubt is the set of numbers it is being asserted for. Read the claim slowly.
Given
Claim: $\ln\bigl(x^{2}\bigr)=2\ln x$ for every $x\neq 0$.
Find
True or false, and give the corrected version in one sentence if it is false.
Hint 1/4
Do not test with a positive number; the claim is obviously fine there. Ask where each side is defined, and try a number where the two answers differ.
Hint 2/4
The power law $\ln(x^{r})=r\ln x$ is proved from the definition, which requires the base to be positive; the definition of $\ln$ has domain $(0,\infty)$.
Hint 3/4
Put $x=-1$ into the claim: the left side is $\ln((-1)^{2})=\ln 1=0$, while the right side is $2\ln(-1)$, which does not exist.
Hint 4/4
So the claim is false as stated; with bars, $\ln(x^{2})=2\ln\lvert x\rvert$, it is true for every $x\neq 0$.
Show solutionTest at x = −1
$\text{left: } \ln\bigl((-1)^{2}\bigr)=\ln 1=0$
the square is taken first, so the input to the logarithm is positive and everything is defined
$\text{right: } 2\ln(-1) \text{ is undefined}$
the domain of $\ln$ is $(0,\infty)$, so the right-hand side has no value at all here
the same repair is why the antiderivative of $1/x$ carries bars
Answer $$\text{False; the correct identity is } \ln\bigl(x^{2}\bigr)=2\ln\lvert x\rvert.$$
Check
Independent check by differentiating both sides of the repaired identity: the left gives $2x/x^{2}=2/x$ and the right gives $2\cdot\tfrac1x$. Equal on both halves of the domain, which the unrepaired version could not manage.
Before accepting an identity, compare the two domains. Two expressions that agree wherever both are defined are still not the same statement if one of them is defined in more places.
4§11.5 — where the derivative identity comes from●●●○○
In this course the exponential was not introduced first and the logarithm second; it was the other way round. That order has a consequence for what counts as a proof.
Given
$\exp$ is defined as the inverse of $\ln$, where $\ln x=\int_1^x dt/t$
An exam question says: from the definition, show that $\frac{d}{dx}e^{x}=e^{x}$
Find
Which line of argument answers the question as asked?
Hint 1/4
The question restricts what you are allowed to use. List the facts about $e^{x}$ that exist at this point in the course; there is essentially one.
Hint 2/4
$\exp$ is the inverse of $\ln$, so $\ln(e^{x})=x$; and $\ln$ has derivative $1/x$, which came from the Fundamental Theorem applied to the defining integral.
Hint 3/4
Set $y=e^{x}$. The inverse relation gives $\ln y=x$, and differentiating both sides with respect to $x$ gives $y'/y=1$.
Hint 4/4
Multiplying by $y$ gives $y'=e^{x}$, which is the required identity.
Show solutionStart from the only property available
$y=e^{x}\iff \ln y=x$
this is what defining $\exp$ as an inverse means, and it is the only fact about $e^{x}$ that has been established
Differentiate and solve
$\frac{y'}{y}=1\;\Longrightarrow\;y'=y=e^{x}$
implicit differentiation on the left, and $y$ is never zero so the multiplication is safe
Answer $$\frac{d}{dx}e^{x}=e^{x}$$
Check
Independent route to the same line: the inverse-function rule says the slope of $\exp$ at $x$ is the reciprocal of the slope of $\ln$ at $e^{x}$, which is $1/(1/e^{x})=e^{x}$.
From the definition is a real constraint on the argument, not a decoration. It rules out every step that quotes a property of $e^{x}$ established later.
B · computation 6 questions
1§11.1 — shells about the vertical axis●●○○○
The plainest form of the week's main task: one curve, the horizontal axis, and a rotation about the vertical axis.
Given
Region: $0\le y\le x^{3}$ for $0\le x\le 2$
Axis of rotation: the vertical axis
Find
(a) Set up the shell integral.
(b) Evaluate it and report the volume.
Hint 1/4
Draw the strip before writing anything. Which way does it run relative to the axis, and what is its thickness called?
Hint 2/4
Shells about the vertical axis: $V=2\pi\int_a^b x\,f(x)\,dx$, where $f(x)$ is the height of the strip at $x$.
Hint 3/4
Here $f(x)=x^{3}$ and $x$ runs from $0$ to $2$, so the integrand is $x\cdot x^{3}=x^{4}$.
Hint 4/4
The integral is $2\pi\bigl[\tfrac{x^{5}}{5}\bigr]_0^2=\tfrac{64\pi}{5}$.
Independent check by washers: at height $y\in[0,8]$ the region runs from $x=y^{1/3}$ to $x=2$, so $V=\pi\int_0^8\bigl(4-y^{2/3}\bigr)dy=\pi\bigl[4y-\tfrac35y^{5/3}\bigr]_0^8=\pi\bigl(32-\tfrac{96}{5}\bigr)=\tfrac{64\pi}{5}$.
When radius times height collapses to one monomial, shells are almost always the cheaper of the two routes.
2§11.2 — the axis moves off the region●●●○○
Same curve as the previous problem, same interval; only the axis has moved, and with it the radius.
Given
Region: $0\le y\le x^{3}$ for $0\le x\le 2$
Axis of rotation: the vertical line $x=3$
Find
(a) Write the shell radius and check it at both ends of the interval.
(b) Evaluate the volume.
Hint 1/4
Do not write an integral yet. Point at the strip at $x=0$ and at the line $x=3$, and say how far apart they are; then do the same at $x=2$.
Hint 2/4
Shell method about a vertical line $x=c$: $V=2\pi\int_a^b(\text{distance to the axis})\times(\text{height})\,dx$, with the distance written so that it is non-negative on $[a,b]$.
Hint 3/4
The region occupies $0\le x\le 2$, entirely to the left of $x=3$, so the radius is $3-x$; the height is still $x^{3}$.
Hint 4/4
So $V=2\pi\int_0^2(3-x)x^{3}dx=\tfrac{56\pi}{5}$.
Show solutionRadius, and the endpoint test
$\text{radius}=3-x$
the whole region has $x\le 2<3$, so it lies to the left of the axis
$3-0=3,\qquad 3-2=1$
both positive, and both match distances you can measure off a sketch
$\tfrac34\cdot 16=12$ and $\tfrac{32}{5}$ from $2^{5}$
$=2\pi\cdot\frac{28}{5}=\frac{56\pi}{5}$
$12-\tfrac{32}{5}=\tfrac{60-32}{5}$
Answer $$V=\frac{56\pi}{5}$$
Check
Comparison check against the previous problem, whose answer was $\tfrac{64\pi}{5}$: the tall part of this region sits near $x=2$, which is the part closest to the new axis, so the volume had to come out smaller. It does, by $\tfrac{8\pi}{5}$.
Predict whether the answer should grow or shrink before you compute it. It costs one sentence and turns the arithmetic into a test.
3§11.1 — a region between two curves●●●○○
Now the strip has a curve at each end, so the height is a difference rather than a single value.
Given
Region: between $y=\sqrt{x}$ (above) and $y=x^{2}$ (below), for $0\le x\le 1$
Axis of rotation: the vertical axis
Find
(a) Write the height of the strip at $x$.
(b) Evaluate the volume.
Hint 1/4
Two things are missing before any integral can be written: where the region starts and stops, and which of the two curves is on top there.
Hint 2/4
Shells about the vertical axis with a region between two curves: $V=2\pi\int_a^b x\bigl[\text{top}(x)-\text{bottom}(x)\bigr]dx$.
Hint 3/4
The curves $y=\sqrt x$ and $y=x^{2}$ meet at $x=0$ and $x=1$, and on $(0,1)$ the root is above the square, so the height is $\sqrt x-x^{2}$.
Hint 4/4
So $V=2\pi\int_0^1 x(\sqrt x-x^{2})dx=\tfrac{3\pi}{10}$.
Show solutionHeight and limits
$\sqrt{x}=x^{2}\iff x=0 \text{ or } x=1$
raising both sides to the fourth power gives $x^{2}=x^{4}$, whose non-negative solutions are $0$ and $1$
$\text{height}=\sqrt{x}-x^{2}>0 \text{ on } (0,1)$
top minus bottom; testing at $x=\tfrac14$ gives $\tfrac12-\tfrac1{16}>0$, confirming which curve is on top
Order of magnitude: the lens has area $\int_0^1(\sqrt x-x^{2})dx=\tfrac23-\tfrac13=\tfrac13$, and it sits at an average distance of well under $1$ from the axis, so the volume must be under $2\pi\cdot 1\cdot\tfrac13\approx 2.1$. The answer $0.94$ is inside.
Whenever a region is described as between two curves, the limits usually come from where they cross, and that intersection is step one rather than an afterthought.
4§11.4 — an integral that produces a logarithm●●○○○
No volume here. The point is the pairing that turns up constantly once $\ln$ is available: a function and its derivative inside one integral.
Given
$\displaystyle\int_1^{e}\frac{\ln x}{x}\,dx$
Find
Evaluate the integral exactly.
Hint 1/4
Before choosing a method, look at whether one part of the integrand is the derivative of another part.
Hint 2/4
Substitution: with $u=g(x)$ you get $\int f(g(x))g'(x)dx=\int f(u)du$, and in a definite integral the limits move too.
Hint 3/4
Here $g(x)=\ln x$ has $g'(x)=1/x$, and the integrand is exactly $\ln x\cdot\frac1x$. The limits $x=1,e$ become $u=0,1$.
Hint 4/4
So the integral is $\int_0^1 u\,du=\tfrac12$.
Show solutionSubstitute
$u=\ln x,\qquad du=\frac{dx}{x}$
the whole factor $dx/x$ is already present, so this substitution costs nothing; trying $u=x$ or integration by parts would both be longer
the transformed integral is as simple as an integral gets
Answer $$\int_1^{e}\frac{\ln x}{x}dx=\frac12$$
Check
Independent check by differentiating an antiderivative: $\frac{d}{dx}\bigl[\tfrac12(\ln x)^{2}\bigr]=\ln x\cdot\tfrac1x$, and evaluating $\tfrac12(\ln x)^{2}$ between $1$ and $e$ gives $\tfrac12(1)-\tfrac12(0)=\tfrac12$.
Both the substitution and the check use the same fact from two directions: $(\ln x)'=1/x$. Getting an answer one way and confirming it the other is cheap here.
5§11.5 — an exponential integral with an exact limit●●○○○
The limit of integration is itself a logarithm, which is what makes the answer come out as a whole number.
Given
$\displaystyle\int_0^{\ln 3} e^{x}\,dx$
Find
Evaluate the integral exactly.
Hint 1/4
Nothing needs substituting. Write the antiderivative first and look at what the upper limit does to it.
Hint 2/4
The exponential is its own antiderivative: $\int e^{x}dx=e^{x}+C$; and $e^{\ln a}=a$ for every $a>0$.
Hint 3/4
Evaluating between the given limits $0$ and $\ln 3$ gives $e^{\ln 3}-e^{0}$, and $e^{\ln 3}=3$, $e^{0}=1$.
Hint 4/4
So the value is $3-1=2$.
Show solutionAntidifferentiate
$\int e^{x}dx=e^{x}+C$
the function is its own derivative, so it is also its own antiderivative
Evaluate and cancel
$\bigl[e^{x}\bigr]_0^{\ln 3}=e^{\ln 3}-e^{0}$
substituting the two limits
$=3-1=2$
$e^{\ln 3}=3$ because $3>0$, and $e^{0}=1$ because $\ln 1=0$
Answer $$\int_0^{\ln 3}e^{x}dx=2$$
Check
Sanity bound: on $[0,\ln 3]$, an interval of length $\ln 3\approx 1.10$, the integrand runs from $1$ to $3$, so the integral is between $1.10$ and $3.30$. The answer $2$ sits inside, and nearer the lower end as an upward-bending curve requires.
The cancellation $e^{\ln 3}=3$ needs $3>0$, which is free here. It is the same identity that fails at negative arguments, so it is worth naming even when it is obviously safe.
6§11.6 — a variable base with a variable exponent●●●●○
Neither the power rule nor the rule for a constant base applies to the function below. One tool does.
Given
$y=x^{\sqrt{x}}$ for $x>0$
Find
(a) Say in one sentence why neither standard rule applies.
(b) Find $y'$ as a single expression in $x$.
(c) Evaluate $y'$ at $x=4$.
Hint 1/4
Ask what each of the two standard rules assumes about the thing it differentiates, and check whether this function satisfies either assumption.
Hint 2/4
Logarithmic differentiation: take $\ln$ of both sides, differentiate implicitly so the left becomes $y'/y$, then multiply back by $y$ written in $x$.
Hint 3/4
Here $\ln y=\sqrt x\ln x$; differentiating the right-hand side by the product rule gives $\frac{\ln x}{2\sqrt x}+\frac{1}{\sqrt x}$.
Hint 4/4
So $y'=x^{\sqrt x}\cdot\frac{\ln x+2}{2\sqrt x}$, and at $x=4$ that is $4(\ln 4+2)\approx 13.5$.
Independent check by difference quotient at $x=4$: $y(4.001)-y(3.999)$ divided by $0.002$ returns $13.545$, against $4(\ln 4+2)=13.545$ from the formula.
If your final expression does not contain the original function as a factor, you stopped at $y'/y$.
C · exam level 4 questions
1§11.2 — picking the right set-up under time pressure●●●○○
Exam conditions: four integrals, one of them correct, and the difference between them is entirely in the radius and the limits.
Given
Region: $0\le y\le\sqrt{x}$ for $0\le x\le 4$
Axis of rotation: the vertical line $x=6$
Find
Which integral gives the volume of the resulting solid?
Hint 1/4
Do not evaluate any of the four. Decide two things first: what the radius of the strip at $x$ is, and which variable the limits belong to.
Hint 2/4
Shells about a vertical line $x=c$: $V=2\pi\int(\text{distance to axis})(\text{height})\,dx$, with the limits being the range of $x$ because $dx$ is the thickness.
Hint 3/4
The region occupies $0\le x\le 4$ and the axis is at $x=6$, so the radius is $6-x$ and the height is $\sqrt x$; the limits are $0$ and $4$.
Hint 4/4
So the integral is $2\pi\int_0^4(6-x)\sqrt x\,dx$, and it equals $\tfrac{192\pi}{5}$.
Show solutionRadius and limits
$\text{radius}=6-x,\qquad 6-0=6,\ 6-4=2$
the region is entirely left of the axis, and both endpoint values are positive
$\text{thickness}=dx \Rightarrow \text{limits } 0 \text{ to } 4$
the range of $y$ here is $[0,2]$, which is what the wrong answers use
$4^{3/2}=8$ so the first term is $32$; $4^{5/2}=32$ so the second is $\tfrac{64}{5}$
Answer $$V=\frac{192\pi}{5}\approx 120.6$$
Check
Bound as an independent check: every strip has radius between $2$ and $6$, and the region has area $\int_0^4\sqrt x\,dx=\tfrac{16}{3}$, so the volume lies between $2\pi(2)(\tfrac{16}{3})=\tfrac{64\pi}{3}\approx 67$ and $2\pi(6)(\tfrac{16}{3})=64\pi\approx 201$. The answer $120.6$ is inside.
The two things that separate a correct set-up from a plausible one are the order of the subtraction in the radius and which variable owns the limits.
2§11.3 — a horizontal axis that is not the coordinate axis●●●●○
The axis is horizontal, so vertical strips now cut across it — and it has been shifted down, so the two radii are not what they were.
Given
Region: $0\le y\le\sqrt{x}$ for $0\le x\le 4$
Axis of rotation: the horizontal line $y=-1$
Find
Which integral gives the volume of the resulting solid?
Hint 1/4
First decide the method by comparing the direction of your strip with the direction of the axis. Only then start measuring lengths.
Hint 2/4
Washers about a horizontal line $y=d$: $V=\pi\int\bigl[R(x)^{2}-r(x)^{2}\bigr]dx$, where $R$ and $r$ are distances from that line to the far and near boundaries.
Hint 3/4
Here the axis is $y=-1$, the far boundary is $y=\sqrt x$ and the near one is $y=0$, so $R=\sqrt x+1$ and $r=1$, with $x$ from $0$ to $4$.
Hint 4/4
So the integral is $\pi\int_0^4\bigl[(\sqrt x+1)^{2}-1\bigr]dx=\tfrac{56\pi}{3}$.
Show solutionThe two radii, measured from the new axis
$R(x)=\sqrt{x}-(-1)=\sqrt{x}+1$
distance from the axis $y=-1$ up to the top boundary $y=\sqrt x$
$r(x)=0-(-1)=1$
the hole reaches only down to the bottom boundary $y=0$, which is one unit above the axis; a shifted axis creates a hole where there was none
$4^{3/2}=8$, so the second term is $\tfrac43\cdot 8$
Answer $$V=\frac{56\pi}{3}\approx 58.6$$
Check
Independent check on the shape: with the axis at $y=0$ the same rotation is a solid disk stack of volume $\pi\int_0^4 x\,dx=8\pi\approx 25$. Lowering the axis by one unit both enlarges every outer radius and punches a cylindrical hole of radius $1$ and length $4$; the net effect must be an increase, and $58.6>25$ agrees.
Shifting a horizontal axis away from the region creates an inner radius even when the original solid had no hole. That new $r$ is the part most often left out.
3§11.6 — a volume that answers back with a logarithm●●●●○
Both halves of the section in one question: a shell integral whose integrand collapses, and a second one whose integrand does not.
Given
Region $R_b$: $0\le y\le \dfrac1x$ for $1\le x\le b$, where $b>1$
Find
(a) Show that turning $R_b$ about the vertical axis gives volume $2\pi(b-1)$.
(b) Find the value of $b$ for which that volume equals $10\pi$.
(c) With that value of $b$, find the volume when $R_b$ is turned about the line $x=1$ instead.
Hint 1/4
Take the parts in order and, for each one, write the radius before anything else. In part (a) watch what the radius does to the height.
Hint 2/4
Shells about a vertical line $x=c$: $V=2\pi\int(\text{distance to axis})(\text{height})\,dx$; and $\int\frac{dx}{x}=\ln\lvert x\rvert+C$.
Hint 3/4
For the region under $y=1/x$ on $[1,b]$: about the vertical axis the radius is $x$ and the height $1/x$, so the integrand is $1$. About $x=1$ the radius is $x-1$, so the integrand is $(x-1)/x=1-\frac1x$.
Hint 4/4
Hence $V=2\pi(b-1)$, then $b=6$, then $V=2\pi\bigl[x-\ln x\bigr]_1^6=2\pi(5-\ln 6)$.
Independent check on (c) by slicing the other way. Horizontal slices about $x=1$ are disks: for $0\le y\le\tfrac16$ the radius is $5$, and for $\tfrac16\le y\le 1$ it is $\tfrac1y-1$. That gives $\pi\cdot 25\cdot\tfrac16+\pi\int_{1/6}^{1}\bigl(\tfrac1y-1\bigr)^{2}dy=\pi\bigl(\tfrac{25}{6}+\tfrac{35}{6}-2\ln 6\bigr)=2\pi(5-\ln 6)$, the same value from a completely different set-up.
Compare (a) and (c): moving the axis by one unit turned an elementary integral into a logarithmic one. The logarithm is not decoration — it is what $\int dx/x$ is, and $\int dx/x$ is what a shifted radius leaves behind.
4§11.5 — a volume with e in the answer●●●○○
One line of set-up, one cancellation, and a limit of integration chosen so that the answer is exact.
Given
Region: $0\le y\le\dfrac1x$ for $1\le x\le e$
Axis of rotation: the vertical axis
Find
What is the volume of the resulting solid?
Hint 1/4
Write the product of radius and height before doing anything else, and simplify it.
Hint 2/4
Shells about the vertical axis: $V=2\pi\int_a^b x\,f(x)\,dx$, and here $f(x)=1/x$.
Hint 3/4
The integrand is $x\cdot\frac1x=1$, on the interval from $x=1$ to $x=e$.
radius times height is the constant $1$; if you reach for a logarithm here you have integrated the height instead of the product
$=2\pi\bigl[x\bigr]_1^{e}=2\pi(e-1)\approx 10.8$
$e\approx 2.718$
Answer $$V=2\pi(e-1)$$
Check
Independent check by sandwiching: the region has area $\int_1^{e}\frac{dx}{x}=\ln e=1$, and every strip of it sits at a distance between $1$ and $e$ from the axis, so the volume lies between $2\pi(1)(1)\approx 6.3$ and $2\pi(e)(1)\approx 17.1$. The answer $10.8$ is inside, and nearer the middle as a region spread across the whole range should be.
A logarithm appears in these problems only when the radius fails to cancel the height. Recognising the cancellation before integrating saves the most common wrong answer here.
D · interleaved 4 questions
1§11.3 — mixed with earlier material●●●●○
No hint here about which method or which section this belongs to; deciding that is the exercise. Read the region and the axis, then choose.
Given
Region: bounded by $y=x^{2}$ below, $y=4$ above, and the vertical axis on the left, with $x\ge 0$
Axis of rotation: the horizontal line $y=4$
Find
What is the volume of the resulting solid?
Hint 1/4
Two decisions before any formula: which way does your strip run compared with the axis, and does the resulting solid have a hole in it?
Hint 2/4
A strip across the axis gives disks or washers, $V=\pi\int\bigl[R^{2}-r^{2}\bigr]$, and $r=0$ exactly when the region touches the axis all along.
Hint 3/4
Here the axis $y=4$ is the top boundary of the region itself, so $r=0$ and $R(x)=4-x^{2}$, with $x$ running from $0$ to $2$.
Hint 4/4
So $V=\pi\int_0^2(4-x^{2})^{2}dx=\tfrac{256\pi}{15}$.
Bound as an independent check: the solid fits inside the cylinder of radius $4$ and length $2$, volume $\pi(16)(2)=32\pi\approx 100.5$, and it fills a bit more than half of it. $53.6$ is a bit more than half of $100.5$.
When the axis of rotation is one of the boundaries of the region, the solid has no hole and the washer formula loses its second term.
2§11.4 — mixed with earlier material●●●○○
Nothing announces which technique this needs. Look at the shape of the integrand and decide.
Given
$\displaystyle\int_0^1\frac{x}{x^{2}+1}\,dx$
Find
Evaluate the integral exactly.
Hint 1/4
Compare the numerator with the derivative of the denominator before choosing anything.
Hint 2/4
If $u=g(x)$ then $\int\frac{g'(x)}{g(x)}dx=\ln\lvert g(x)\rvert+C$, and a constant factor can always be moved outside.
Hint 3/4
Here $g(x)=x^{2}+1$ with $g'(x)=2x$, and the integrand has only $x$, so the integral is $\tfrac12\int_1^2\frac{du}{u}$ after the limits move.
Independent check by bounding: on $[0,1]$ the integrand rises from $0$ to $\tfrac12$, so the integral is between $0$ and $\tfrac12$; and $\tfrac12\ln 2\approx 0.347$ sits there. A second check: this integral is exactly half the one in the pretest, whose value was $\ln 2$.
Top is the derivative of the bottom, up to a constant, is the single most common signal for a logarithm. Spot the constant and the rest is bookkeeping.
3§11.5 — mixed with earlier material●●●●○
Two rules from earlier in the course meet one from this section. The order in which you apply them is the whole question.
Given
$g(x)=e^{x^{2}}\ln(3x)$ for $x>0$
Find
(a) Find $g'(x)$.
(b) Evaluate $g'(1)$ exactly.
Hint 1/4
Identify the outer structure before differentiating: is this a product, a quotient, or a composition? Name the pieces.
Hint 2/4
Product rule $(uv)'=u'v+uv'$, together with $\frac{d}{dx}e^{w}=e^{w}w'$ and $\frac{d}{dx}\ln w=w'/w$.
Hint 3/4
With $u=e^{x^{2}}$ and $v=\ln(3x)$: $u'=2x e^{x^{2}}$ and $v'=3/(3x)=1/x$.
Hint 4/4
So $g'(x)=2xe^{x^{2}}\ln(3x)+e^{x^{2}}/x$, and at $x=1$ this is $e(2\ln 3+1)\approx 8.69$.
Show solutionStructure first
$g=u\cdot v,\quad u=e^{x^{2}},\quad v=\ln(3x)$
naming the outer structure before differentiating anything stops the two chain rules from being confused with each other
Each factor separately
$u'=e^{x^{2}}\cdot 2x$
$\frac{d}{dx}e^{w}=e^{w}w'$ with $w=x^{2}$; the inner derivative is the part that is dropped when the rule is misremembered
$v'=\frac{3}{3x}=\frac1x$
$\frac{d}{dx}\ln w=w'/w$ with $w=3x$; the constant $3$ cancels, which is worth noticing before writing it
Independent check by a difference quotient at $x=1$: $\bigl(g(1.001)-g(0.999)\bigr)/0.002$ returns $8.691$, against $e(2\ln 3+1)=8.691$ from the formula.
$\ln(3x)$ and $\ln x$ differ by a constant, so they have the same derivative. Splitting the logarithm first is often faster than chaining through it.
4§11.5 — mixed with earlier material●●●●○
The function below combines a polynomial with an exponential. What is being asked is a question from earlier in the course, applied to a function that only became differentiable this week.
Given
$f(x)=x^{2}e^{-x}$ for every real $x$
Find
At which point does $f$ have a local maximum?
Hint 1/4
The exponential factor is never zero and never negative. What does that let you ignore when you look for sign changes?
Hint 2/4
A local maximum sits where $f'$ changes from positive to negative, and $f'$ can only change sign at a critical number.
Hint 3/4
Here $f'(x)=e^{-x}x(2-x)$, and since $e^{-x}>0$ the sign is that of $x(2-x)$, which is positive on $(0,2)$ and negative outside it.
Hint 4/4
So the sign goes $+\to-$ at $x=2$: a local maximum there.
the First Derivative Test; at $x=0$ the sign goes $-\to+$, which is a local minimum instead
Answer $$\text{local maximum at } x=2,\ f(2)=4e^{-2}$$
Check
Independent check by values: $f(1)=e^{-1}\approx 0.368$, $f(2)=4e^{-2}\approx 0.541$, $f(3)=9e^{-3}\approx 0.448$. The middle value is the largest of the three, as a maximum at $2$ requires.
Factoring out the exponential is the move that makes these sign lines short: a factor that is positive everywhere can be removed from a sign question entirely.
Mistake ledger (12 entries)
⚠ Dropping the radius factor
The strip at $x=0.9$ travels nine times as far in one turn as the strip at $x=0.1$. Without the factor $x$ the formula cannot tell them apart, and it returns exactly twice the truth for this region.
⚠ Integrating over the range of the wrong variable
Both ranges are visible in the picture; only the thickness of the strip says which one owns the limits. If the thickness is $dx$, the limits are the range of $x$.
wrong$V=2\pi\int_0^9 x\cdot x^{2}\,dx$
right$V=2\pi\int_0^3 x\cdot x^{2}\,dx$
⚠ Using the coordinate as the radius after the axis has moved
Every example before the axis moved had the two agreeing, so the habit forms invisibly. The radius is a distance to the axis, and it equals $x$ only when the axis is the vertical axis itself.
⚠ Subtracting in the wrong order and repairing it at the end
It happens to give the right size here, but only because the region lies on one side of the axis. Test the radius at an endpoint before integrating and the question never arises.
The radius is easy to read off the picture in one variable and the height in the other. Once the thickness is chosen, every length in the integrand must be written in that variable.
wrong$V=2\pi\int_0^4 y\bigl(2-x^{2}\bigr)dy$
right$V=2\pi\int_0^4 y\bigl(2-\sqrt{y}\bigr)dy$
⚠ Using one curve as the height when the region sits between two
The phrase the region under a curve is so common that a second boundary stops being read as a boundary. Height is always far minus near.
wrong$h(x)=x^{2}\quad\text{for the region between } y=x^{2} \text{ and } y=4$
right$h(x)=4-x^{2}$
⚠ Turning a sum inside the logarithm into a sum of logarithms
The law is remembered as splits into a sum, and the plus sign inside gets matched with the plus sign outside. The proof splits an integral at $t=a$, which only works for a product.
wrong$\ln(a+b)=\ln a+\ln b$
right$\ln(ab)=\ln a+\ln b$
⚠ Reporting a value of the logarithm outside its domain
The graph running off the bottom of the page looks like arriving somewhere. Undefined at a point and tending to minus infinity near it are two separate statements, and only the second is about behaviour.
right$\ln x\text{ is undefined for } x\le 0;\qquad \lim_{x\to 0^{+}}\ln x=-\infty$
⚠ Applying the derivative rule to a composition
The phrase it is its own derivative attaches itself to the letter $e$ instead of to the specific function $x\mapsto e^{x}$. Anything in the exponent brings a chain rule with it.
wrong$\frac{d}{dx}e^{-3x}=e^{-3x}$
right$\frac{d}{dx}e^{-3x}=-3e^{-3x}$
⚠ Cancelling in the direction that is not allowed
The two identities look symmetric on the page. The restriction belongs to whichever of the two functions acts first, and $\ln$ refuses negative inputs.
wrong$e^{\ln(-4)}=-4$
right$\ln\bigl(e^{-4}\bigr)=-4,\qquad e^{\ln x}=x\ \text{only for } x>0$
⚠ Stopping at y prime over y
The hard part is over by then and the left-hand side contains a $y'$, which reads like an answer. A relative rate is not a derivative until it is multiplied by the function.
⚠ Turning a product inside the logarithm into a product of logarithms
The aim of the step is to break a product apart, and keeping the multiplication looks like keeping the equation true. Test it once with $f=g=e$ and it collapses.
Close the page and write, from memory: the three factors of the shell volume element and what each one measures; the one rule that decides between shells and washers; the definition of $\ln x$ as an integral and the two things that follow from it in one line each; and the four steps of logarithmic differentiation, with a note on which one is usually dropped.
Explain, without integrating, why the factor $x$ in $2\pi\int x f(x)dx$ cannot be dropped, using two regions of the same shape at different distances from the axis?
c-shell-element
Write the shell radius for a region on $[0,2]$ turned about $x=5$ and about $x=-2$, and say how you checked the order of each subtraction?
c-shifted-axis
Give one region that gives shells about a horizontal axis, and one that gives washers about a vertical axis?
c-slice-direction
Prove $\ln(ab)=\ln a+\ln b$ from the integral definition, naming the substitution and saying where the limits went?
c-ln-integral
Derive $\frac{d}{dx}e^{x}=e^{x}$ in two lines from the inverse relation, and say which of $e^{\ln x}=x$ and $\ln(e^{x})=x$ carries a restriction?
c-exp-inverse
Differentiate $x^{x}$ and say, in one sentence, why neither the power rule nor the rule for a constant base applies to it?
c-log-diff
Glossary (11 terms)
cylindrical shellsilindirik kabuk
A thin hollow cylinder, described by an inner radius, a height and a wall thickness; it is what one strip of the region sweeps out when the strip runs parallel to the axis of rotation.
shell methodkabuk yöntemi
Computing a volume of revolution by integrating $2\pi\times(\text{radius})\times(\text{height})$ over the range of the thickness variable, with strips taken parallel to the axis.
axis of rotationdönme ekseni
The fixed line the planar region is turned around to produce a solid. It may be a coordinate axis or any line parallel to one.
shell radiuskabuk yarıçapı
The non-negative distance from a strip of the region to the axis of rotation. It equals the coordinate only when the axis passes through the origin along a coordinate direction.
washer methodhalka yöntemi
Computing a volume of revolution by integrating $\pi\bigl(R^{2}-r^{2}\bigr)$ over slices taken across the axis, where $R$ and $r$ are the distances from the axis to the far and near boundaries.
natural logarithmdoğal logaritma
The function on $(0,\infty)$ defined by $\ln x=\int_1^x dt/t$; equivalently, the signed area under $y=1/t$ from $1$ to $x$.
natural exponentialdoğal üstel fonksiyon
The inverse of the natural logarithm, written $\exp x$ or $e^{x}$; it is defined for every real number, takes only positive values, and is its own derivative.
inverse functionters fonksiyon
For a one to one function $f$ with range $J$, the function $f^{-1}$ on $J$ satisfying $f^{-1}(f(x))=x$; its graph is the graph of $f$ reflected in the line $y=x$.
logarithmic differentiationlogaritmik türev
Taking the logarithm of both sides of $y=f(x)$ before differentiating, so that products become sums and exponents become coefficients; used for a variable base with a variable exponent and for long products.
relative rate of change
The quantity $y'/y$, which is the derivative of $\ln y$; it measures change as a proportion of the current value rather than in absolute units.
volume elementhacim elemanı
The volume $dV$ contributed by one thin slice of the solid, written in terms of the thickness variable; integrating it over the range of that variable gives the total volume.
What comes next
§12 · Inverse trigonometric functions, hyperbolic functions, and indeterminate forms revisited
Two inverses were built here: the shell method as the inverse way of slicing, and the exponential as the inverse of the logarithm. The next section keeps only the second idea and pushes it further — inverting the trigonometric functions, meeting the hyperbolic ones, and returning to the indeterminate limits that now have $e^{x}$ and $\ln x$ in them.
Sources
James Stewart, Calculus, Ninth Edition — sections 5.3, 6.2* and 6.3* The definitions and the statement of the shell formula follow this book; the worked regions and functions here are different ones.
Course syllabus, week 11: Applications of Integration 5.3, 6.2*, 6.3* The three tokens covered above come from this line, and the assessment weights quoted on the card come from the same document.
The endpoint test on a shell radius, used throughout Not a theorem, just bookkeeping: substituting one endpoint into the radius expression before integrating catches the order-of-subtraction error that produces a negative volume.