1. Computing a rate upside down
the whole part, and the unit markWhat not to write
“The average rate of change of on is .”
What to write
“.”
Why: A rate is output change per unit of input: . The units say it too: cells per HOUR, metres per SECOND.
MATH 203 Calculus I • Concordia University, Montreal
This sheet is not a summary of sections 2.1 and 2.2 of Thomas' Calculus: you have the textbook. It answers one question only, what makes students lose marks on rates of change and the limit laws in MATH 203 at Concordia University, and which precise gesture avoids each loss.
The limits of this chapter are rarely hard; the algebra around them is. A fraction inside a fraction, a negative exponent, a conjugate multiplied on one side only, a factorization that forgets its leading coefficient: these cost more marks than any limit law. Your scientific calculator helps with the tables of secant slopes, and every trap below says when it can be trusted and when it cannot.
The thread of the chapter
The secant is the only slope you can compute and the tangent is what the secants approach; reaching it means a limit whose substitution gives , an order to rewrite, and the marks are lost in the ALGEBRA of that rewriting, not in the limit.
On a calculator table, keep every digit until the last division: , and rounding to first destroys the estimate.
Name the law on the line where you use it, and check its permit first: a quotient law used on a denominator that tends to earns nothing, even with the right number.
Each row reads from left to right: the assumptions, then the result. A red cell is not an answer, it is the finding that the form settles nothing and the instruction to rewrite it. Every case is followed by a worked example.
Read a line as: this expression may be rewritten this way, with this result. The red lines are not answers: the first two are rewrites that are FALSE, the last is a form that decides nothing until the expression is rewritten.
| Expression | Rewrite | Result |
|---|---|---|
| common denominator | ||
| Example: , then divided by : the rate of on is . | ||
| , | reciprocal | , |
| Example: at ; with the top does not even vanish. | ||
| times | ||
| Example: at . | ||
| no such rule no such rule | ||
| Example: while . What to do: Keep the root whole; to get rid of it in a limit, multiply top and bottom by the conjugate. | ||
| no such rule no such rule | ||
| Example: while . What to do: Rewrite over a common denominator. | ||
| at | substitution | indeterminate settles nothing |
| Example: at . Same form, other result: At , while : same form, two answers. What to do: Factor out of top and bottom (or conjugate, or common denominator), cancel it for , substitute again. | ||
Every false rewrite in this table turns a limit that exists into one that seems not to, or the reverse. Test any rewrite at one simple value, such as or , before building on it.
These are the errors I correct most often in session. Each one costs marks on a paper, even when the reasoning behind it is right.
What not to write
“The average rate of change of on is .”
What to write
“.”
Why: A rate is output change per unit of input: . The units say it too: cells per HOUR, metres per SECOND.
What not to write
“With , the slope of at is .”
What to write
“Right side: ; left side: . The slope is between and , about .”
Why: A curve that bends upward makes every right-hand secant too steep and every left-hand one too flat. Two sides give a bracket, one side gives a biased guess.
What not to write
“Substitution gives , so does not exist.”
What to write
“The form is . For , , so the limit is .”
Why: The zeros on top and bottom say that is a common factor. After cancelling, the function is a line with a HOLE, and the limit is the height of the hole.
What not to write
“, which is at , so the limit of does not exist.”
What to write
“, so the quotient is for , and the limit is .”
Why: : a negative exponent means a reciprocal. Rewrite every negative or fractional exponent BEFORE doing anything else.
What not to write
“, so .”
What to write
“With the conjugate, for , so the limit at is .”
Why: The root of a sum is not the sum of the roots: , . Only the conjugate removes a root legally.
What not to write
“, so the limit at of is .”
What to write
“, so the quotient is for , and the limit is .”
Why: A quadratic with leading coefficient factors as . Expanding the product back takes ten seconds and catches the missing .
What not to write
“.”
What to write
“, so the quotient law does not apply; the expression must be studied another way.”
Why: Every limit law has a permit: the limits of the pieces must exist, and for a quotient the limit of the denominator must not be . Write the permit before the law.
What not to write
“ for , so the limit at is .”
What to write
“ but , arbitrarily close to : the limit does not exist.”
Why: A table only sees the points you chose; the points all make an even multiple of . And a table that jumps to for tiny , as does, is round-off.
Substitute c first, on the side. What you get picks the tool
If a number, and the function is a polynomial or a rational function whose denominator is not at → done: the limit is the value; name the theorem
Example:
If with polynomials → factor out of both: roots, sum or difference of cubes, grouping, synthetic division
Example: at
expand the factorization back before cancelling
If with a square root → multiply top AND bottom by the conjugate
Example: at (two conjugates)
If with fractions inside a fraction, or negative exponents → rewrite the exponents, common denominator, then multiply by the reciprocal
Example: at
If a bounded factor ( or of something wild) times a factor tending to , or known only by inequalities → Sandwich Theorem, with two bounds that share their limit
Example: , so the limit at is
If a number over something tending to 0 → the quotient law has no permit; say so, the study of such limits comes with infinite limits
Example: at with and
No branch fits, or the algebra stalls? Estimate with a table of values on both sides, to know what you are aiming for, then return to the algebra: a table suggests, only the rewriting proves.
A marker ticks steps. Here they are in order, with the concluding sentence expected word for word.
When to use it: The question says estimate, or the exact slope needs a rule not yet available (an exponential, a data table)
Concluding sentence
“The slopes of the secants to the right of decrease to , those to the left increase to ; the slope of the tangent at lies between them, so it is to two decimals.”
The trap: A single value of , or a single side: the estimate is then biased and nothing on the page shows it.
Marking: Usually 1 mark for the table on each side and 1 for the justified estimate.
When to use it: A factor with no limit (, ) that stays bounded, or a function given only through inequalities
Concluding sentence
“For all , . Since , the Sandwich Theorem gives .”
The trap: Writing for all : false for , where the bounds swap. Use .
Marking: Typically 1 mark for the bounds, 1 for their common limit, 1 for naming the theorem.
Five minutes of checking recover more marks than one more problem started in a hurry.
Test a rewrite at one simple value
Before cancelling or building on an identity, put x = 1 or x = 9 into both sides. A false rule fails at once.
against at : against , the leading coefficient is missing.
A table on both sides, as a target
Evaluate the expression at c plus and minus 0.01 with the calculator. The exact answer must be close to both values.
For : gives , gives , so is plausible and is not.
The sign of the answer
Take x just beside c and read the sign of every factor, or read whether the curve rises or falls. The limit or the slope cannot have the opposite sign.
decreases on , so its average rate there is negative: , never .
Radian mode
Any sine or cosine in a rate or a limit is in radians. Check the mode before the first keystroke.
The rate of on is ; degree mode returns .
Let and . Estimate the slope of the tangent at with two secants, then find it exactly as the limit of the secant slopes, and write the equation of the tangent.
Step 1
Secant slopes with the calculator: at gives , at gives . The slope is about .
Why
Two sides give a bracket and a target: the exact answer must land between and .
Step 2
The slope of for is , . At it gives .
Why
Writing the form shows the marker why a rewriting follows; the tangent slope is .
Step 3
Common denominator on top: .
Why
This is the step where the marks go: the parentheses around carry the minus sign to both terms.
Step 4
Divide by , and use : for .
Why
The factor appears only after turning around; cancelling it is legal because .
Step 5
is rational with denominator at , so . Tangent: , that is .
Why
Naming the theorem on rational functions justifies the substitution that was forbidden two lines earlier.
Step 6
Check: , inside the bracket of the first step.
Why
A lost minus sign would give , outside the bracket, and the check catches it for free.
The conclusion, written out
“The slopes of the secants tend to as , so the tangent at has slope and equation .”
The classic mistake on this problem: Writing , or without parentheses, which gives and no common factor.
The average rate of change is the slope of a secant line through two points of the graph: change in output divided by change in input. The slope of the tangent is what those secant slopes approach as the second point slides toward the first. It is a limit, estimated with a table of secant slopes and found exactly by rewriting the quotient.
Compute the slope of the secant from the point to a second point at distance h, for h equal to 0.1, 0.01 and 0.001, and also for the negative values. Keep all the digits until the final division. The two lists close in on the tangent slope from opposite sides, and the bracket tells you how many decimals your estimate deserves.
Do not stop and do not say the limit does not exist. Zero over zero means the top and bottom share a factor. Factor the polynomials, multiply by the conjugate if there is a square root, or combine fractions over a common denominator after rewriting any negative exponent. Cancel the common factor, allowed because x is not equal to c, then substitute again.
A table suggests a limit, it never proves one. The calculator can round a difference of nearly equal numbers to zero, and badly chosen points can hide an oscillation, as with the cosine of pi over x at x equal to 0.1, 0.01 and 0.001. Use the table as a target, then decide the limit with algebra and the limit laws.
Use it when one factor has no limit but stays bounded, typically the sine or cosine of one over x, and the rest tends to zero, or when the function is only given through inequalities. Build two bounds valid for every x near the point, check that they have the same limit, and name the theorem in your conclusion.
© Ahmed Squalli Houssaini. Revision sheet published at www.letuteurscientifique.ca/en/fiches/math203-tangent-lines-limit-laws. Free for personal and classroom use; republishing it elsewhere requires written permission (legal notice).