4.1.12 · D4Calculus I — Limits & Derivatives

Exercises — Power rule — proof for integer, rational exponents

2,222 words10 min readBack to topic

The one rule under test everywhere below:

The two engines behind it (from the parent note): the Limit definition of the derivative and, for whole-number powers, the Binomial theorem.


Level 1 — Recognition

You just apply "drop and drop": slide to the front, then subtract one from the exponent.

Q1. Differentiate .

Q2. Differentiate .

Q3. Differentiate (that is, ).

Recall Solution — Q1, Q2, Q3

Q1. Here . WHAT: apply . WHY: is a variable base with a constant exponent — exactly the shape the power rule handles.

Q2. Here (a negative integer — Stage 3 of the proof covers this). WHY the extra minus: subtracting one from a negative exponent makes it more negative.

Q3. Rewrite the root as a power first: , so . WHY rewrite: the power rule needs an exponent, not a root symbol — same number, cleaner form.


Level 2 — Application

Now the exponents are messier, and you combine the power rule with the fact that constants multiply through and sums differentiate term-by-term.

Q4. Differentiate .

Q5. Differentiate . (Hint: split the fraction first.)

Q6. Find the slope of at .

Recall Solution — Q4, Q5, Q6

Q4. WHAT: rewrite each piece as a clean power. and . Now differentiate term by term (a constant multiple just rides along): WHY the middle term turned positive: — two minus signs.

Q5. WHAT: split before differentiating (you cannot power-rule a quotient directly here). WHY split: dividing powers of the same base subtracts exponents — turning one messy fraction into two simple powers. Now differentiate:

Q6. WHAT: differentiate, then plug in. : At : , so WHY : the exponent means "cube the square root" — take , then cube.


Level 3 — Analysis

Here you reason about the rule, not just crank it — reversing it, using it geometrically, and reading off where a derivative behaves strangely.

Q7. A curve has slope everywhere and passes through the origin's height rule -type family. Which power function has derivative exactly ? (Find and the leading coefficient if .)

Q8. Find the equation of the tangent line to at the point where .

Q9. For , explain — using the derivative — why the graph has a vertical tangent at . (See the step figure.)

Recall Solution — Q7, Q8, Q9

Q7. WHAT: run the power rule backward. If then . We need this to equal . WHY match pieces: two power functions are equal only if their exponents match AND their coefficients match.

  • Exponent: .
  • Coefficient: .

So (check: ✓).

Q8. WHAT: get the slope, get the point, assemble the line .

  • Point: . So .
  • Slope: , so . WHY point-slope: it is the fastest way to build a line from one point and one slope.

Q9. WHAT: compute and watch it as . WHY it blows up: is a positive number that shrinks to as , so grows without bound: . An infinite slope means the tangent line stands straight up — vertical. Both sides ( and ) give , so the slope rockets to from both sides. See the near-vertical tangent in the figure below.

Figure — Power rule — proof for integer, rational exponents

Level 4 — Synthesis

Now you rebuild the rule itself — proving special cases from the Limit definition of the derivative and stitching in the Chain rule / Implicit differentiation.

Q10. Prove from the limit definition (Stage-1 style) that , showing every cancellation.

Q11. Prove from first principles that using the common-denominator trick (Stage 3).

Q12. Use implicit differentiation (Stage 4) to prove .

Recall Solution — Q10, Q11, Q12

Q10. WHAT: write the limit-definition quotient and expand by the Binomial theorem. Expand: . Subtract (the leading term cancels): WHY divide by : every surviving term carries at least one , so dividing is legal and exposes which term has no left: Send : the last two terms each contain and vanish.

Q11. WHAT: . Form the difference quotient and combine over a common denominator.

= \lim_{h\to0}\frac{1}{h}\cdot\frac{x - (x+h)}{(x+h)x}.$$ WHY common denominator: it collapses two fractions into one whose top simplifies. The top is $x-(x+h) = -h$: $$= \lim_{h\to0}\frac{-h}{h\,(x+h)x} = \lim_{h\to0}\frac{-1}{(x+h)x}.$$ WHY the $h$ cancels: it was the only thing preventing us from setting $h=0$. Now let $h\to0$: $$f'(x) = \frac{-1}{x\cdot x} = -\frac{1}{x^2} = -x^{-2}. \checkmark$$ **Q12.** WHAT: set $y = x^{2/3}$ and cube both sides to clear the fraction (here $p=2,\,q=3$): $$y^3 = x^2.$$ WHY cube: it turns a fractional exponent into two *integer* powers we already proved. Differentiate both sides with respect to $x$; the left needs the [[Chain rule]] because $y$ depends on $x$: $$3y^2\,\frac{dy}{dx} = 2x.$$ Solve for $\frac{dy}{dx}$: $$\frac{dy}{dx} = \frac{2x}{3y^2}.$$ Substitute $y = x^{2/3}$, so $y^2 = x^{4/3}$: $$\frac{dy}{dx} = \frac{2x}{3\,x^{4/3}} = \tfrac23\,x^{1 - 4/3} = \tfrac23\,x^{-1/3}. \checkmark$$

Level 5 — Mastery

Boundary cases, wrong-rule detection, and a limit-of-the-rule sanity check — where deep understanding separates from mechanical use.

Q13. Two look-alikes: differentiate (a) and (b) . Which one uses the power rule, which does not, and why? Give both derivatives.

Q14. Zero/degenerate case. Consider (defined for ). Compute two ways: (i) directly, and (ii) by the power-rule formula with . Confirm they agree, and explain what the formula's factor of is "protecting" us from.

Recall Solution — Q13, Q14

Q13. The deciding question: is the exponent constant with a variable base, or is the base constant with a variable exponent?

  • (a) . The base varies; the exponent is just a constant number. This is the power rule — nothing special about the number sitting up top:
  • (b) . The base is constant; the exponent varies. This is an exponential, not a power. Its rule (see Derivative of exponential functions) is , and since : WHY they differ: the power rule was proven by expanding in ; that machinery cannot touch a variable exponent. Different structure ⇒ different tool.

Q14. (i) Directly. For , , a constant. The derivative of any constant is (from the limit definition, for all ): (ii) By the formula. gives They agree. WHAT the factor of protects: even though is undefined at , the leading factor multiplies it to before that ever matters. The formula quietly encodes "a constant has zero slope" — the front factor is exactly what kills the otherwise awkward . WHY this is the promised consistency: Stage 2 of the proof and the general formula give the same answer, so the four stages knit together seamlessly.


Recall Self-test scoreboard

Nailed L1–L2 unaided ::: You can apply the rule reliably. Nailed L3 unaided ::: You can reason about slopes and tangents, including vertical ones. Nailed L4 unaided ::: You can rebuild the rule from the limit definition and implicitly. Nailed L5 unaided ::: You spot the power-vs-exponential trap and handle degenerate cases — mastery.

Connections