Derivatives of inverse trigonometric functions sit at the intersection of three AP Calculus skill rows: implicit differentiation, the chain rule, and the ability to recognise which inverse function a problem is actually asking about. On the AP Calculus AB and BC exams this topic appears as a stand-alone derivative problem, as a chain-rule wrapper over a non-linear inner function, and — in BC — as a building block inside an integration-by-parts setup or a differential-equation model. Mastering the six core derivative rules is therefore not optional; the rubric has a dedicated row for the formula and a second row for the substituted value, and candidates who stumble on either row lose a point that is otherwise free.
The six derivative rules AP Calculus tests verbatim
The exam-writing team treats inverse trigonometric derivatives as a closed list of six formulas. If a student walks into the testing room remembering only the arcsine and arctangent rules, the harder questions will still get partial credit, but the cleanest score on a stand-alone derivative problem comes from writing the correct rule, identifying the inner function, applying the chain rule, and simplifying. The six rules, in the order most AP problem sets present them, are: d/dx arcsin(u) = u′/√(1 − u²); d/dx arccos(u) = −u′/√(1 − u²); d/dx arctan(u) = u′/(1 + u²); d/dx arccot(u) = −u′/(1 + u²); d/dx arcsec(u) = u′/(|u|√(u² − 1)); d/dx arccsc(u) = −u′/(|u|√(u² − 1)). Each of these is a one-line memory task, but each carries a sign and a denominator pattern that the rubric checks separately.
Two design choices in these formulas are worth pausing on because they shape the rubric rows. First, the arcsine and arccosine derivatives are negatives of each other; the rubric never gives both a full point if the student confuses which numerator to use. Second, the arcsecant and arccosecant rules use |u| in the denominator — a feature that catches students who learned a simplified form from an older textbook. AP graders mark the absolute value as part of the formula row, so omitting it costs the same as using the wrong sign. For most candidates reading this, the highest-leverage study move is to write all six formulas in one column and then cover them up and reproduce them cold, because the rubric will not give partial credit for "I knew the formula but wrote the wrong one".
AB versus BC coverage: the College Board lists arcsin, arccos, arctan, and arccot as part of AB Topic 2.5, Differentiation: Composite, Implicit, and Inverse Functions. Arcsecant and arccosecant are explicitly BC-only — they appear in BC Topic 2.7 and are not assessed on the AB exam. A student preparing for AB can therefore safely defer the secant and cosecant rules to a later review cycle, but a BC candidate should treat them as part of the same memorised list rather than a separate topic. In my experience tutoring BC students, the most common early-October mistake is assuming that arcsec and arccsc are "extra credit" formulas; they are not, and they appear inside BC free-response problems on parametric and polar topics more often than students expect.
How the rubric scores a stand-alone inverse trig derivative
A typical stand-alone derivative question on the AP exam reads something like: "Let f(x) = arctan(3x). Find f′(x)." The rubric, even when compressed into a single line, reads in three discrete rows. The first row is the formula: the student must show the structure u′/(1 + u²) with u = 3x correctly identified. The second row is the substituted value of u′: with u = 3x, u′ = 3, and that 3 must appear in the numerator. The third row is the simplified answer: f′(x) = 3/(1 + 9x²). A student who writes the correct structure but forgets the inner derivative loses the second row; a student who writes the structure with the inner derivative but does not square the 3 inside the denominator loses the third row. Three points, three rows, no slack.
For a question that asks for a derivative at a point — "Find f′(2) where f(x) = arcsin(x/5)" — the rubric adds a fourth row: the evaluation. The student must show the substitution x = 2, the resulting argument 2/5, and the arithmetic that turns √(1 − 4/25) into √(21/25) and then into √21/5. The cleaned answer is f′(2) = (1/5) / (√21/5) = 1/√21 = √21/21. Notice that the chain rule has already been applied: the inner derivative is 1/5, which is what the rubric's "u′" row is looking for. A common pitfall here is cancelling the 5's before writing the formula down; the rubric wants to see the 1/5 explicitly, otherwise the chain-rule row is scored as missing even though the final number is correct.
The arcsecant and arccosecant versions add a fifth row on BC exams: the absolute value. The formula is u′/(|u|√(u² − 1)), and the rubric reads the absolute value as a separate notational requirement. A common error on BC papers is to write u′/(u√(u² − 1)) without the bars; the rubric scores this as a missing notation row, and the student loses a point that has nothing to do with whether the answer is mathematically defensible. For most candidates, the cleanest habit is to wrap every u in |·| the moment the problem says arcsec or arccsc, even when u is provably positive. The grader does not infer positivity from context; the symbol is the row.
Chain rule wrapping: where the topic gets tested on FRQs
On the free-response section, derivatives of inverse trigonometric functions almost never appear as a bare formula. They appear wrapped in a chain rule, often with two layers of substitution. A representative BC problem is: "Let g(x) = arctan(√(1 + x²)). Find g′(x) at x = 0." The rubric unfolds this in five rows. The first row is the outer formula: u′/(1 + u²) with u = √(1 + x²). The second row is u′ itself, which requires another chain rule: u = (1 + x²)^(1/2) so u′ = (1/2)(1 + x²)^(−1/2) · 2x = x/√(1 + x²). The third row is the substitution of u and u′ into the outer formula, giving g′(x) = (x/√(1 + x²)) / (1 + (√(1 + x²))²). The fourth row is the algebraic simplification of the denominator: 1 + (1 + x²) = 2 + x². The fifth row, on this particular problem, is the evaluation at x = 0, which collapses g′(0) to 0.
The trap on this style of question is the algebraic simplification row, not the formula row. Students who can write the structure u′/(1 + u²) but then cannot simplify 1 + (√(1 + x²))² into 2 + x² lose a point that has nothing to do with inverse trig knowledge. The rubric reads simplification as a separate scored line, and a candidate who leaves the answer in unsimplified form is marked down even when the structure is correct. In practice, the cleanest habit is to simplify the squared radical immediately: (√(1 + x²))² = 1 + x². Doing this in one line keeps the work readable and gives the grader the simplification row on a silver platter.
Chain-rule wrappers also appear inside integration problems on the BC exam. A standard reverse-chain problem reads: "Evaluate ∫ 1/(1 + 4x²) dx." The student is expected to recognise u = 2x, du = 2 dx, and the integrand as (1/2) · (1/u) · du / (1 + u²) = (1/2) arctan(2x) + C. The derivative rule for arctan is therefore not just a differentiator's tool; it is the integration-table entry that lets the student undo a derivative of the form 1/(1 + u²). A student who memorises the derivative rule without ever reading it backwards as an antiderivative template will spend twice as long on every integration problem of this shape. For most AP candidates, drilling the derivative in the form d/dx arctan(u) = u′/(1 + u²) and then reading it as ∫ u′/(1 + u²) du = arctan(u) + C is the single most efficient study move for this topic.
Implicit differentiation problems that hide an inverse trig derivative
Implicit differentiation is where the topic stops looking like a memorisation task and starts testing whether the student can recognise an inverse trig function in a different costume. The classic example is: "If y = arctan(x/y), find dy/dx." The student must differentiate both sides with respect to x, apply the chain rule to the right-hand side, and solve for dy/dx. The first derivative row on the rubric is the formula u′/(1 + u²) applied to u = x/y. The second row is the quotient rule applied to u itself, which produces (y − x·y′)/y². The third row is the algebraic solve for y′. The fourth row, when the problem asks for it, is a simplification that produces a clean expression in x and y only.
The most common error on this style of problem is mis-identifying the outer function. A student who reads the right-hand side as "arctan of (x/y)" and then differentiates as if arctan were a constant coefficient will produce a nonsense answer that loses the formula row. The rubric marks the first row on whether the student wrote d/dx arctan(u) = u′/(1 + u²) in some recognisable form, even if u is itself a function of x. A useful diagnostic: if the right-hand side contains a fraction with 1 + (something)² in the denominator, the outer function is arctan; if it contains √(1 − (something)²) in the denominator, the outer function is arcsin or arccos. The rubric's first row rewards this recognition, not blind formula application.
Inverse trig derivatives also show up in related-rates problems. A representative AB problem: "A camera is placed at ground level and rotates upward to track a balloon. The angle of elevation is θ(t) = arctan(t/100). Find dθ/dt at t = 0." The rubric reads the formula row, the inner-derivative row, and the evaluation row in the usual way. The pedagogical point is that inverse trig is a compact way for the problem-writer to package a related-rates setup; the student who has drilled the derivative on plain arctan(t/100) will get a full score without needing to interpret the physics at all. In my experience, students who skip the bare-formula drilling in favour of "context" problems end up losing points on the standalone derivative row inside the context problem. Both drills are necessary.
Domain and range restrictions the rubric expects
Every inverse trigonometric function has a principal-value domain and range, and the AP rubric occasionally tests whether the student is aware of them. The standard ranges are: arcsin : [−1, 1] → [−π/2, π/2]; arccos : [−1, 1] → [0, π]; arctan : ℝ → (−π/2, π/2); arccot : ℝ → (0, π); arcsec : |x| ≥ 1 → [0, π] excluding π/2; arccsc : |x| ≥ 1 → [−π/2, π/2] excluding 0. The derivative rules are valid only on the interior of these domains. A student who differentiates arcsin(x) at x = 2 will produce a real-number answer of 1/√(1 − 4) = 1/(i√3), which is not a real number and signals to the grader that the domain row was not checked.
The most common domain trap on the AP exam is the arcsecant and arccosecant rule combined with a u that crosses zero. If u = x, then |x| appears in the denominator; if the problem asks for the derivative at x = 0, the answer is undefined and the rubric expects the student to say so. A common error is to write the formula, plug in x = 0, and report 0/|0|·(something) = 0, which is wrong on two rows: the |u| row and the evaluation row. The grader will mark the answer incorrect. The defensive habit is to scan the argument of the inverse trig function before differentiating: if it is 0, the derivative of arcsec or arccsc does not exist there.
Domain restrictions also appear implicitly when the inner function of arcsin or arccos is bounded. If f(x) = arcsin(sin(x)) on the interval [π/2, 3π/2], the derivative is not simply 1; it flips sign because arcsin is locally equal to π − x on that interval. AP problems occasionally test this, and the rubric marks the first row on whether the student recognised the piecewise nature of arcsin(sin(x)). A safe study move is to graph y = arcsin(sin(x)) once and see the triangular wave; the derivative is a square wave with period 2π, and the rubric's simplification row tests for that.
Common pitfalls and how to avoid them
The first pitfall is the sign of arccos and arccot. The derivative of arccos(u) is negative; the derivative of arccot(u) is negative. Students who learn the six rules in the order arcsin, arccos, arctan, arccot, arcsec, arccsc sometimes write the sign for arccos and then pattern-match the rest of the list as positive, which costs them on every arccot and arccos problem. A diagnostic drill: list the six formulas and circle the negative ones; arccos and arccot are negative, and so are the arcsec and arccsc rules' companion pair (arccsc). The remaining two — arcsin and arctan — are positive. The rubric does not give partial credit for the right structure with the wrong sign.
The second pitfall is the chain rule on a quadratic inner function. Students often write d/dx arctan(x²) = 1/(1 + x⁴), which is wrong. The correct answer is 2x/(1 + x⁴). The rubric scores the inner-derivative row separately, and a student who skips it loses the row even when the outer structure is correct. The defensive habit is to underline u and write u′ next to it before differentiating. If u = x², then u′ = 2x has to appear in the numerator; if u = 3x + 1, then u′ = 3 has to appear; if u = sin(x), then u′ = cos(x) has to appear.
The third pitfall is forgetting the absolute value on arcsec and arccsc. The rubric marks the |u| row as a separate notation point, and a missing pair of vertical bars is a one-point deduction even when the rest of the answer is correct. The defensive habit is to wrap the argument in |·| the moment the problem mentions arcsec or arccsc, regardless of whether the inner function is positive. A student who wrote d/dx arcsec(x) = 1/(x√(x² − 1)) on a BC FRQ would lose the absolute-value row; a student who wrote d/dx arcsec(x) = 1/(|x|√(x² − 1)) would not. The two answers are mathematically equivalent on the parts of the domain where x > 0, but the rubric is a notation rubric, not a mathematical-truth rubric.
Worked example: a BC-style two-part FRQ
Consider the following problem, which is representative of how the topic shows up on a BC free-response. Let f(x) = arctan(eˣ). Part (a): find f′(x). Part (b): find the linear approximation of f(x) at x = 0, and use it to estimate f(0.1). The rubric for part (a) is three rows. The first row is the outer formula u′/(1 + u²) with u = eˣ. The second row is u′ = eˣ. The third row is the simplified answer f′(x) = eˣ/(1 + e^(2x)). A student who writes the structure with u = eˣ but forgets to differentiate the inner function loses the second row; a student who writes u′ = eˣ in the numerator but does not square eˣ in the denominator loses the third row.
Part (b) tests the local-linearisation formula L(x) = f(a) + f′(a)(x − a) at a = 0. The rubric has three rows for this part as well. The first row is f(0) = arctan(e⁰) = arctan(1) = π/4. The second row is f′(0) = e⁰/(1 + e⁰) = 1/2. The third row is L(x) = π/4 + (1/2)(x − 0) = π/4 + x/2, and the estimate is L(0.1) = π/4 + 0.05. A student who writes the linearisation formula correctly but cannot evaluate f′(0) loses the second row even though the chain rule on part (a) was applied correctly. The two parts of the problem are scored independently; the rubric does not give a "credit by reference" for the part (a) answer inside part (b).
The pedagogical point of this two-part problem is that the derivative rule feeds the linearisation step, but the linearisation step is scored on its own rows. A student who treats part (b) as a separate, smaller problem and gives it its own three lines of work will score higher than a student who tries to compress the work into one paragraph. The rubric rewards each line of correct reasoning with one point; the goal is to make every line count by writing it on its own.
Inverse trig derivatives inside integration on the BC exam
The BC exam's Unit 8 covers applications of integration, and inverse trig derivatives show up as antiderivative templates. The integration-by-parts setup ∫ arcsin(x) dx is a classic: let u = arcsin(x), dv = dx, so du = 1/√(1 − x²) dx and v = x. The result is x·arcsin(x) − ∫ x/√(1 − x²) dx, and the remaining integral is a u-substitution: let w = 1 − x², dw = −2x dx, so the integral becomes −(1/2)∫ w^(−1/2) dw = −√w + C = −√(1 − x²) + C. The full antiderivative is x·arcsin(x) + √(1 − x²) + C. The rubric for this problem has four rows: the IBP setup, the integral of dv, the substitution on the remaining integral, and the final answer.
Notice that the inverse trig derivative rule appears twice in this single problem. The first appearance is du = 1/√(1 − x²) dx, which is the derivative of arcsin(x) read in differential form. The second appearance is implicit: the substitution w = 1 − x² is set up precisely because the denominator √(1 − x²) is the radical that defines the arcsin derivative. A student who has drilled the derivative in the form d/dx arcsin(x) = 1/√(1 − x²) will recognise the radical and the substitution; a student who has only memorised the formula without understanding the radical structure will fumble the IBP step. The rubric marks the IBP setup on its own row, so the recognition matters even if the eventual algebra is correct.
Inverse trig antiderivatives also appear in differential-equations models on the BC exam. A standard logistic-style model is dP/dt = k·P·(1 − P/M), and the separation-of-variables step produces an integral on each side that resolves to an arctanh or an arctan depending on the algebraic rearrangement. A student who can identify the antiderivative of 1/(1 + u²) as arctan(u) + C without going back to the derivative table will save time on every problem of this shape. The rubric marks the antiderivative row separately, so this recognition is a directly scored skill, not a time-saving nicety.
Comparison of the six inverse trig derivative rules at a glance
The table below summarises the six rules in the format the rubric reads them. Use it as a one-page study card before any practice test; the goal is to reproduce every row from memory in under three minutes.
| Function | Derivative rule | Sign | Denominator form | AP coverage |
|---|---|---|---|---|
| arcsin(u) | u′/√(1 − u²) | + | √(1 − u²) | AB (Topic 2.5) |
| arccos(u) | −u′/√(1 − u²) | − | √(1 − u²) | AB (Topic 2.5) |
| arctan(u) | u′/(1 + u²) | + | (1 + u²) | AB (Topic 2.5) |
| arccot(u) | −u′/(1 + u²) | − | (1 + u²) | AB (Topic 2.5) |
| arcsec(u) | u′/(|u|√(u² − 1)) | + | |u|√(u² − 1) | BC only (Topic 2.7) |
| arccsc(u) | −u′/(|u|√(u² − 1)) | − | |u|√(u² − 1) | BC only (Topic 2.7) |
Two structural observations from the table. The first is that the sign pattern is consistent: the first and third functions in the canonical list (arcsin, arctan) are positive, and the second, fourth, fifth, and sixth (arccos, arccot, arcsec, arccsc) are negative. A student who remembers this pattern can recover the sign even if a specific rule is forgotten. The second is that the denominators come in two flavours: the arcsin/arccos pair has √(1 − u²), and the arctan/arccot pair has (1 + u²). The arcsec/arccsc pair uses √(u² − 1), which is the same radical as the first pair but with the sign flipped inside. A useful mnemonic: "minus one" goes with arcsin and arccos, "plus one" goes with arctan and arccot, and the secant/cosecant pair uses the same radical as the sine pair but without the leading 1.
Preparation strategy: how to drill this topic for a 5
The first preparation move is a one-week memorisation cycle on the six rules. Write the rules from memory each morning, then apply each to a one-line derivative problem: d/dx arcsin(5x), d/dx arccos(2x + 1), d/dx arctan(x³), d/dx arccot(√x), and for BC, d/dx arcsec(x²), d/dx arccsc(1/x). The first two days of the cycle will produce sign errors and missing-chain-rule errors; by day five the answers should be coming out cleanly. The rubric rewards fluency: a student who can write the answer in under 90 seconds has time to spend on the chain-rule and simplification rows, which is where the partial-credit points live.
The second preparation move is a chain-rule drill. Take ten chain-rule problems from past FRQs and solve each one in two lines: identify u, identify u′, write the derivative, simplify. The goal is to get to a state where writing the chain rule is automatic and the only thinking happens at the simplification step. A student who spends 40% of the exam time on the chain-rule row is leaving points on the table; the chain rule should be the row that takes 10 seconds, not the row that takes a minute. For most candidates, the chain-rule drill is the single highest-leverage study activity in the three weeks before the exam.
The third preparation move is an integration drill in the reverse direction. Take ten integration problems of the form ∫ u′/(1 + u²) du and ∫ u′/√(1 − u²) du, and solve them by reading the derivative rule backwards. The goal is to internalise the dual role of the inverse trig derivative rule: it differentiates on the way forward and integrates on the way back. A student who has done this drill will recognise a 1/(1 + 9x²) integrand as a u = 3x, du = 3 dx setup, and will produce arctan(3x)/3 + C without consulting a table. The rubric marks the antiderivative row on its own, so this recognition is a directly scored skill.
Conclusion and next steps
Derivatives of inverse trigonometric functions are a high-density topic on the AP Calculus exam: six formulas, three chain-rule rows, two sign rows, and a notational absolute-value row that BC candidates must defend. A student who has memorised the six rules, drilled the chain rule, and read the formulas backwards as antiderivative templates can expect to score full marks on the standalone derivative questions and most of the wrapper questions on the FRQ. The most efficient next step is a structured one-week cycle that pairs cold-recall drills with chain-rule and integration drills; this is the cycle that converts the topic from a memorisation chore into a scored, repeatable skill. AP Courses' AP Calculus BC one-to-one programme maps each student's chain-rule and absolute-value errors against the rubric rows for inverse trig derivatives, then turns that diagnostic into a focused drill plan.