1 Calculus 4/14/03 4:08 PM Page 1

SPARKCHARTSTM
“BU

CALCULUS I SPARK
CHARTS TM
THE
DIF

BACKGROUND AND FUNCTIONS TYPES

WHAT IS CALCULUS? WRITING FUNCTIONS DOWN represented horizontally, and the dependent variable verti- TRIGONO
A table is a list that keeps track of input values (such as ages cally. A graph represents a function as long as it passes the
TM

Calculus is the study of “nice”—smoothly changing—functions. Sum and

SPARKCHARTS
of a cactus) and corresponding output values (such as the vertical line test: for every x-value, there is at most one y -
• Differential calculus studies how quickly functions are
number of needles on the cactus) of a function. There may value. sin(A ±
changing at particular points.
• Integral calculus studies areas enclosed by curves.
not be a universal equation that describes such a function. cos(A ±
• The Fundamental Theorem of Calculus connects the two. An equation such as f (x) = x2 + 1 describes how to tan(A ±
numerically manipulate the independent variable (often x) Double-a
FUNCTIONS
to get the output value f (x). If y = x2 + 1, then y is a
WHAT IS A FUNCTION? sin(2A)
“function of x,” and it is the dependent variable.
A function is a rule for churning out values: for every value cos(2A)
you plug in, there’s a unique value that comes out. A graph represents a�function� visually. If y = f (x), then
• The set of all the values that can be plugged in is the domain. plotting many points x, f (x) on the plane will give a pic- Half-angl
Function Not a function
• The set of all the values that can be output is the range. ture of the function. Usually, the independent variable is
A
sin =
TYPES OF FUNCTIONS For more on functions, see the SparkChart on Pre-Calculus. 2
CONSTANT FUNCTIONS—Horizontal lines The graph of a quadratic func- RULES FOR EXPONENTS: RULES FOR LOGARITHIMS:
A constant function y = c has only one output value. Its tion is a parabola. The general
aloga b = b loga an = n
graph is a horizontal line at height c. quadratic can be expressed in the Squared
form f (x) = a(x − h)2 + k, am+n = am an loga (bc) = loga b + loga c
LINEAR FUNCTIONS—Straight lines
where h = − 2a b
and k = f (h). am sin2 A =
A linear function can be expressed in the easy-to-graph
am−n = loga cb = loga b − loga c
� �
• The vertex is at (h, k).
slope-intercept form y = mx + b, where b is the an
• If a > 0, the parabola opens Pythagore
y -intercept (the value of f (0)) and m is the slope. The a−n = 1
loga 1
= − loga b
up. If a < 0, the parabola opens an b sin2 A +
slope of a straight line measures how steep it is; if (x1 , y1 ) Parabola. Here a > 0.
down.
and (x2 , y2 ) are two points on the line, then the slope is amn = (am )n loga bn = n loga b tan2 A +
• If |a| is large, the parabola is narrow. If |a| is small, the
change in y y2 − y1 1 + cot2
m= = . parabola is wide. 1 √ √
change in x x2 − x1 an = n
a loga n b = n1 loga b Special tr
RATIONAL FUNCTIONS m √ � √ �m √ θ (deg)
an = am = n a
n
loga n bm = m n loga b
A rational function is a quotient of two polynomials:
0◦
f (x) = p(x)
q(x) . It is defined everywhere except at the roots of
� a �n
= an
q(x). The zeroes of f (x) are those roots of p(x) that are b bn
30◦
not also roots of q(x).
(ab) = a b
n n n 45◦
• If the degree of p(x) is greater than the degree of q(x)
(deg p(x) ≥ deg q(x) ), then at points where |x| is very � a �−n � b �n 60◦
=
large, f (x) will behave like a polynomial of degree b a
90◦
deg p(x) − deg q(x).
• If deg p(x) < deg q(x) , then when |x| is very large, TRIGONOMETRIC FUNCTIONS—Sinusoidal waves
f (x) will approach 0. See Limits and Continuity. • Angle measure: Angles can be measured in degrees or
50495

POLYNOMIAL FUNCTIONS in radians: θ (in degrees) COMBININ

A general polynomial function can be written in the form 180 = θ (in radians)
π . • Two fun
EXPONENTIAL AND LOGARITHMIC FUNCTIONS—Very fast
f (x) = an xn + an−1 xn−1 + · · · + a1 x + a0 . • The unit circle is the cir- subtracte
or very slow growth
cle of radius 1 centered at function i
The polynomial has degree n if an �= 0. All polynomials Simple exponential functions can be written in the form the origin O = (0, 0). original f
are defined for all real numbers. If n is odd, then the range y = ax , where the base a is positive (and a �= 1). The
ISBN 1-58663-623-5

• If A = (1, 0), then any undefined

9 781586 636234

is all real numbers, too. If n is even, then the polynomial function is always increasing if a > 1 and always decreas- point P = (x, y) on the
reaches some maximum if an < 0 or some minimum if
an > 0. The y -intercept is the constant term a0 . A poly-
ing if a < 1. The domain is all the reals; the range is the
positive reals. Exponential functions grow extremely fast—
circle defines the angle
� AOP of measure θ
LIM
nomial of degree n has at most n roots or zeroes—val- faster than any polynomial. The basic shape of the graph is (counterclockwise). Since
ues of x where the graph crosses the x-axis—and at most always the same, no matter the value of a. LIMIT OF A
θ and θ + 2π define the
n − 1 “turns” (peaks or valleys) in its graph. If function
same angle, all trig functions satisfy f (x) = f (x + 2π); they
Logarithmic functions have the form y = loga x. The num- x gets clos
are periodic with period 2π (or π).
ber loga b is “the power to which you raise a to get b”: approache
loga x = y if and only if ay = x. Sine: sin θ = y, the y -coordinate of P = (x, y). or the value
For all θ, −1 ≤ sin θ ≤ 1. Sine is an odd function. the existen
REMEMBER: Logarithms are exponents. limx→a f (
Cosine: cos θ = x, the x-coordinate of P = (x, y).
$7.95 CAN

Logarithmic functions are defined for the positive reals only; For all θ, −1 ≤ cos θ ≤ 1. Cosine is an even function. type of disc
SparkCharts is a registered trademark
Copyright © 2002 by SparkNotes LLC.

they have the same basic shape as exponential functions but a • One-side
Tangent: tan θ = x , the slope of the line
y
different orientation.
OP .
A Barnes & Noble Publication

approache
Changing the base of a logarithm is the same thing as Secant: sec θ = cos
1
θ = x
1
limx→a− f
10 9 8 7 6 5 4 3 2
Printed in the USA $4.95

Polynomial of degree 5
Roots: −2, −1, − 2 , 3 , 2
1 2
mulplying the logarithm by a constant: Cosecant: csc θ = sin1 θ = y1 when x is c
4 “turns”
loga x = (loga b) (logb x) . limx→a+ f
All rights reserved.

of SparkNotes LLC.

Cotangent: cot θ = tan 1

θ = sin θ = y
cos θ x
to and larg
SPECIAL CASES OF POLYNOMIALS—Lines and parabolas Also, loga b = log1 a .
A polynomial of degree 0 is the constant function f (x) = a
b
Sinusoidal functions can be written in the form • If the lim
for a �= 0. The number e is a special real sided limi
number (approximately
y = A sin B(x − h) + k.
Contrariwi
A polynomial of degree 1 is a linear function. 2.71828), often used as a base • |A| is the amplitude. its of f (x
for exponetial functions. The
A polynomial of degree 2 is a quadratic; it can be written in • k is the average value: halfway between the maximum limx→a f (
logarithm base e is called the
the form f (x) = ax2 + bx + c . It has 0, 1, or 2 roots, natural logarithm and is writ- and the minimum value of the function. sided limit
which are found using the Quadratic Formula: ten loge x = ln x. The natu-
• 2π A NOTE ON
√ ral log follows all logarithm B is the period. A larger B
−b ± b2 − 4ac means more cycles in a given 1. The limi
x= . rules. Any logarithmic expres-
2a sion can be written in terms of natural logarithms using the interval. Ex: lim 1
2
√ x→0 x
• If √b2 − 4ac < 0, then the quadratic has no real roots. change of base formula: the function
• h is phase shift, or how far
• If √b2 − 4ac = 0, then the quadratic has 1 real root. ln x 2. We can
the beginning of the cycle is
loga x = . limit at infi
• If b2 − 4ac > 0, then the quadratic has 2 real roots. ln a from the y -axis.

This downloadable PDF copyright © 2004 by SparkNotes LLC. SPARKCHARTS™ C++ page 1 of 4
 “BUT HOW IS ONE TO MAKE A SCIENTIST UNDERSTAND THAT
THERE IS SOMETHING UNALTERABLY DERANGED ABOUT
SPARK DIFFERENTIAL CALCULUS.…” ANTONIN ARTAUD
CHARTS TM

TYPES OF FUNCTIONS (CONTINUED)

d the dependent variable verti- TRIGONOMETRIC IDENTITIES • Two functions can be composed: f (g(x)) is a function if the y -axis. Equivalently, rotating f (x) 180◦ around
unction as long as it passes the the domain of f (x) contains the range of g(x). This is the origin leaves f (x) unchanged.
Sum and difference formulas
x-value, there is at most one y - sometimes denoted by (f ◦ g)(x). Ex: sin x is an odd function.
sin(A ± B) = sin A cos B ± cos A sin B
NOTE: In (f ◦ g), g is applied first and written second.
cos(A ± B) = cos A cos B ∓ sin A sin B
Ex: If f (x) = x + 2 and g(x) = 4x, then
tan(A ± B) = 1∓tan
tan A±tan B
A tan B (f ◦ g)(x) = 4x + 2.
Double-angle formulas
CHANGING A SINGLE FUNCTION
sin(2A) = 2 sin A cos A
1. Vertical translation: Adding a constant c will translate
cos(2A) = cos2 A − sin2 A
the function vertically c units (up if c is positive, down if c
= 2 cos2 A − 1 = 1 − 2 sin2 A
is negative). The new function y = f (x) + c has the same
Half-angle formulas Odd Function Even Function
Not a function shape and the same domain as the original function.
1 − cos A 1 + cos A 6. The Inverse function, or reflection over the diagonal
� �
A A 2. Horizontal translation: The function y = f (x − c) is a
sin = ± cos = ± y = x: Suppose the function f (x) passes the “horizontal
2 2 2 2 shift of the original function c units horizontally (to the right
if c is positive, left if c is negative). The new function has the line test” in its domain: f (x) never takes the same value
RULES FOR LOGARITHIMS: A 1 − cos A twice. Such a function has a unique inverse f −1 (x) whose
tan = same shape and the same range as the original function.
2 sin A domain is the range of f (x), and vice versa.
loga an = n 3. Vertical stretching and compressing: If c > 1, then the
Squared function formulas • The inverse function has the properties that
function y = cf (x) is a vertical stretch of the original
loga (bc) = loga b + loga c
1 − cos 2A 1 + cos 2A function by a factor of c. If c < 1, then y = cf (x)is a � (f (x))� = x for all x in the domain of f (x) and
f −1
sin2 A = cos2 A = compression of the original function by a factor of c. The f f −1 (x) = x for all x in the domain of f −1 (x).
loga cb = loga b − loga c 2 2
� �
horizontal distances remain unchanged. • The inverse �of−1the inverse function is the original
Pythagorean Identities function: f −1 (x) = f (x).
�
loga 1
= − loga b 4. Horizontal
� � stretching and compressing: The function
b sin2 A + cos2 A = 1 y = f xc is a horizontal stretch of the original function if • Graphically, y = f −1 (x) has the same shape as the
loga bn = n loga b tan2 A + 1 = sec2 A c > 1 (a compression if c < 1) by a factor of c. Vertical orginal function, but is reflected over the slanted line
1 + cot2 A = csc2 A distances remain the same. y = x. Ex: y = ex and y = ln x are inverse functions.
√ See graph in Exponential and Logarithmic Functions.
loga n b = n1 loga b Special trigonometric values
• If f (x) takes the same value more than once, we
√ θ (deg) θ (rad) sin θ cos θ tan θ
loga n bm = m restrict the domain before taking the inverse.
n loga b √
0◦ 0 0
=0 1 0 Ex: y = x2 on the whole real line has no inverse, but
2
√ √ √ the function y = x2 on the positive reals only has the
30◦ π 1
= 1 3 3 √
6 2
√
2 2
√
3 inverse y = x.
45◦ π 2 2
1 Inverse trigonometric functions: sin x, cos x, and tan x
4 2 2
√ √ are periodic and take on the same value many times. To
60◦ π
3 2
3 1
2 3 Vertical Stretch Horizontal Stretch
construct inverse functions, we restrict the domain of the
√
90◦ π
2 2
4
=1 0 undefined 5. Reflection over the axes: The function y = −f (x) is a trigonometric functions to a single cycle.
reflection of the original function over the y -axis. The Arcsine, the principal inverse of sine, is often denoted
S—Sinusoidal waves sin−1 x (not to be confused with sin1 x = csc x). It is
function y = f (−x) is a reflection of the original function
an be measured in degrees or
COMBINING FUNCTIONS over the x-axis. defined� on the�interval [−1, 1] and takes values in the
= θ (in radians)
. • If f (x) = f (−x), then f (x) is called even; it range − π2 , π2 .
π • Two functions can be combined arithmetically—added,
remains unchanged when reflected over the x-axis. Arccosine, or cos−1 x, is defined on [−1, 1] and takes
• The unit circle is the cir- subtracted, multiplied, or divided. The domain of the new
Ex: cos x is an even function. values in the range [0, π] .
cle of radius 1 centered at function includes only points that are in the domain of both
• If f (x) = −f (−x), then f (x) is called odd. A Arctangent, or tan−1 x, is defined on the � whole� real
the origin O = (0, 0). original functions. (A quotient function will always be
reflection over the x-axis is the same as a reflection over line and takes values in the open interval − π2 , π2 .
• If A = (1, 0), then any undefined at the zeroes of the denominator.)
point P = (x, y) on the
circle defines the angle
� AOP of measure θ
LIMITS AND CONTINUITY
gets very large, positively or negatively. Limits at infinity are where the two one-sided limits tend to infinities of opposite
(counterclockwise). Since
LIMIT OF A FUNCTION one-sided. If limx→±∞ f (x) = L exists and is finite, the sign.
θ and θ + 2π define the
If function f (x) comes infinitely close to some value L as line y = L is a horizontal asymptote to the graph of f (x). Ex: The function f (x) = x1 has limx→0+ f (x) = +∞,
atisfy f (x) = f (x + 2π); they
x gets close to a, we say that “L is the limit of f (x) as x Ex: lim ex = 0 . The line y = 0 is a horizontal asymp- limx→0− f (x) = −∞, and a vertical asymptote at x = 0.
r π). x→−∞
approaches a” and write limx→a f (x) = L. The existence tote to the function y = ex .
inate of P = (x, y). or the value of limx→a f (x) by itself says nothing at all about CONTINUITY OF BASIC FUNCTIONS
NOTE: The statement “f (a) exists” implies that f (a) is a finite
Sine is an odd function. the existence or the value of f (a). Rather, comparing the real value. However, the statement “limx→a f (x) exists” • All linear, polynomial, exponential, logarithmic, and
ordinate of P = (x, y).
limx→a f (x) and f (a) tells about the continuity or the could mean that this limit is infinite. We’ll say “limx→a f (x) sinusoidal functions are continuous.
p(x)
Cosine is an even function. type of discontinuity of f (x) at x = a. exists and is finite” to mean that the limit exists and is not infinity. • Rational functions q(x) are continuous everywhere
except at the roots of q(x). At those roots of q(x) which are
pe of the line OP . • One-sided limits: We can look at the limit of f (x) as x CONTINUITY AND DISCONTINUITY also roots of p(x), they have removable discontinuities if
approaches a from one side only. The left-hand limit, • If f (a) exists and is equal to limx→a f (x), we say that the multiplicity of the root in p(x) is at least as great as the
limx→a− f (x) exists if the f (x) is close to some value f (x) is continuous at x = a. If f (x) is continuous at every multiplicity of the root in q(x). All other roots of q(x) give
when x is close to and smaller than a. The right-hand limit a ≤ x ≤ b, we say that f (x) is continuous on the interval vertical asymptotes. See graph on page 4 in Sketching
limx→a+ f (x) depends on values of f (x) when x is close [a, b]. If f (x) is continuous at every real x, we say that f (x)
cos θ
= x Graphs: Summary.
sin θ y to and larger than a. is continuous on the whole real line or simply continuous.
• If the limit of f (x) as x → a exists, then so do both one- • The trigonometric functions tan x = cossin x
x and
written in the form
sided limits, and the three limits have the same value. • If limx→a f (x) exists and is finite but is not equal to sec x = 1
cos x are continuous everywhere except at the
− h) + k. f (a) (which may or may not exist), we say that f (x) has a
Contrariwise, if both the right-hand and the left-hand lim- zeroes of cos x, which are odd-integer multiples of π2 . The
removable discontinuity at x = a . Ex: the function trigonometric functions cot x = cos
its of f (x) as x → a± exist and are equal, then the sin x and csc x = sin x
x 1
2
limx→a f (x) exists and is equal to the value of the one- y = xx−2 −4
is indistinguishable from y = x + 2 every- are continuous everwhere except at the zeroes of sin x—
alfway between the maximum
sided limits. where except when x = 2, where it is undefined. But the
he function. integer multiples of π . All discontinuities for these four
discontinuity at x = 2 could easily be “removed” by insert-
A NOTE ON INFINITY functions are vertical asymptotes.
ing a point to make the graph continuous.
1. The limit limx→a f (x) can be infinite. Vertical asymptote: If either of the one-sided limits
Ex: lim 1
2 = +∞. Here, the line x = 0 is a vertical asymptote: limx→a− f (x) or limx→a+ f (x) exists and is infinite, then TIP: In practice, functions encountered in the classroom are
x→0 x
the function tends towards the line but never quite reaches it. f (x) has a (possibly one-sided) vertical asymptote at discontinuous only at isolated points.
2. We can also look at the limit of f (x) at +∞ or −∞. A x = a. The two-sided limx→a f (x) may or may not exist.
limit at infinity, if it exists, is what f (x) tends toward as |x| Often, functions teachers use will have vertical asymptotes
CONTINUED ON OTHER SIDE

This downloadable PDF copyright © 2004 by SparkNotes LLC. SPARKCHARTS™ C++ page 2 of 4
 1 Calculus 4/14/03 4:08 PM Page 2

LIMITS AND CONTINUITY (CONTINUED) TAK

Trigo
THE EPSILON–DELTA (ε − δ ) DEFINITIONS OF LIMITS AND CONTINUITY
� �� �
Product: lim f (x)g(x) = lim f (x) lim g(x) know
x→a x→a x→a
Limits: limx→a f (x) = L if and only if for every ε > 0, there exists some δ > 0 such that the tr
f (x) limx→a f (x)
whenever x is within δ of a, f (x) is within ε of L (that is, |x − a| < δ implies that Quotient: If lim g(x) �= 0 , then lim = . d(
|f (x) − L| < ε ).
x→a x→a g(x) limx→a g(x)
Continuity: Function f (x) is said to be continuous at x = a if and only if for every ε > 0, The Squeeze Theorem: d(
there exists some δ > 0 such that whenever x0 is within δ of a, f (x0 ) is within ε of f (a) If f (x) ≤ g(x) ≤ h(x) near x = a, and lim f (x) = lim h(x) = L,
(that is, |x0 − a| < δ implies that |f (x0 ) − f (a)| < ε ). x→a x→a
d(
then lim g(x) exists and is equal to L.
Equivalently, f (x) is continuous at x = a if limx→a f (x) exists and is equal to f (a). x→a
LIMIT LAWS The classic application of this theorem establishes that limx→0 sinx x = 1.
Suppose f (x) and g(x) are two functions, a is a point (possibly ±∞) near which both
f (x) and g(x) are defined. (These are only true if both limx→a f (x) and limx→a g(x) exist
Frequently-encountered limits:
US
lim sin x
x =1
and at least one of them is finite!) x→0
THE TAN
Sum: lim f (x) ± g(x) = lim f (x) ± lim g(x)
� � lim cos x−1
x =0
�The ta
x→0
x→a x→a x→a
Scalar multiple: lim cf (x) = c lim f (x)
� �
Here, c is any real. lim xn e−x = 0 for all n a, f (a
x→a x→a x→∞

The tan
f (x):
TAKING DERIVATIVES Useful

INTUITION METHODS AND TRICKS

If a and b are two points in the domain of f (x) then the average rate of change of f (x) Assume that f (x) and g(x) are two differentiable functions.
f (b)−f (a)
on the interval [a, b] is , a measure of how fast f (x) has increased
� or decreased
d �
f (x) ± g(x) = f � (x) ± g � (x)
�
b−a Sum and Difference:
is also the slope of the line through the points a, f (a) and
�
over the interval. This dx
b, f (b) on the graph of f (x).
� �
d �
cf (x) = cf � (x)
�
Scalar Multiple:
The derivative of f (x) at a point x = a is the instantaneous rate of change, a measure dx
d �
of how fast f (x) is increasing or decreasing at a. Equivalently, the derivative is the slope f (x)g(x) = f � (x)g(x) + f (x)g � (x)
�
Product:
of the�tangent�line to the graph of f (x) at the point x = a—the unique line through the dx
point a, f (a) that touches the graph at only that point near x = a. MNEMONIC: If f is “hi” and g is “ho,” then the product rule is “ho d hi plus hi d ho.” DISPLA
We compute the derivative f � (a) by looking at the average rate of change of f (x) on the f (x) f � (x)g(x) − f (x)g � (x) MOTIO
� �
d
Quotient: =
interval [a, a + h] and taking the limit�as h goes
� to 0.�Equivalently, f (a)
�
� is the limit as dx g(x) g 2 (x) Suppos
h → 0 of the slope of the line through a, f (a) and a + h, f (a + h) . MNEMONIC: “Ho d hi minus hi d ho over ho ho.” seconds
• The fi
DEFINITION The Chain Rule takes the derivative of composite functions. Here are two ways of writing it:
� instanta
1. (f ◦ g) (x) = f � (g(x)) g � (x).
If the limit limh→0 f (a+h)−f (a)
exists, we say that f (x) is differentiable at x = a and dy dy du particle
2. If u = g(x) and y = f (u) = f g(x) , then =
� �
h . for the
the limit is the derivative of f (x) at x = a, denoted by f � (a). dx du dx
• The fi
f (x+h)−f (x) IMPLICIT DIFFERENTIATION
The function f � (x) = limh→0 h is the derivative function of f (x). If it is the inst
defined whenever f (x) is defined, then f (x) is called differentiable. Implicit differentiation uses the product and chain rules to find slopes of curves when it is tion: a(
difficult or impossible to express y as a function of x. Leibniz notation may be easiest “meters
If f (x) is differentiable at x = a, then f (x) is continuous at a. The converse is not true:
when differentiating implicitly. Take the derivative of each term in the equation with
a function can be continuous but not differentiable. There are two cases where this occurs: MAXIM
respect to x. Then rewrite dx = y � and dx
dy dx
= 1 and solve for y � .
1. No tangent: Ex: f (x) = |x|. The Ex 1: x + y = 1
2 2 A local�
function is continuous at x = 0 since Implicitly differentiating with respect to x gives the expression 2x dxdx dy
+ 2y dx = 0, which
point c
limx→0+ |x| = limx→0− |x| = 0, simplifies to 2x + 2yy � = 0 or y � = − x
least (o
y . The derivative can now be found for any point on
but the derivative f � (0) is undefined the curve, even
tion in s
since the left-hand slope limit, √ though it is not actually a function. You will get the same result if you first solve minimu
|h| for y = ± 1 − x2 and keep track of the ± signs in different quadrants.
limh→0− a relativ
h = −1 , does not equal
the right-hand slope limit, Ex 2: x cos y − y 2 = 3x
Vertical tangent at x =0 The glo
No tangent at x =0 |h|
limh→0+ h = 1 . Differentiate to obtain first dx dx
cos y + x d(cos y) dy
− 2y dx = 3 dx
dx
, and then
dx f (x) a
cos y − x sin yy � − 2yy � = 3 . Finally, solve for y � = xcos y−3
sin y+2y . conside
2. Vertical tangent: The slope of a the abs
vertical line is “undefined.” If f (x) DERIVATIVES OF BASIC FUNCTIONS If th
has a vertical tangent at x = a, then inclu
the derivative f � (a) is undefined and d(c)
Constants: =0 A constant function is always flat. alwa
the graph of f � (x) will have a verti- dx poss
cal asymptote at x = a . Ex: Valu
√
f (x) = 3 x has a vertical tangent at d(mx + b)
Linear: =m The line y = mx + b has slope m. The wo
the point (0, 0). The derivative func- dx
f � (0) undefined f � (0) undefined or maxi
tion, f � (x) = √ 3
1
, goes to infinity
3 x2
at 0. d(xn ) Critical
Powers: = nxn−1 True for all real n �= 0.
dx All ext
either a
NOTATION d(an xn + · · · + a2 x2 + a1 x + a0 )
Polynomial: = nan xn−1 + · · · + 2a2 x + a1 How to
dx 1. Chec
Different notations for the derivative function are useful in different contexts. The most
d(ex ) defined
common notations in calculus are f � (x), y � , dx f (x), and dx
d dy
. The last two are in Leibniz
�y Exponential: = ex This is why e is called the “natural” f (x) =
notation; dx evolved from �x = change in x , or slope. The expressions dy and dx rep-
dy change in y dx logarithm base: Aex are the only func- it may b
resent infinitesmal changes in y and x. tions that are their own derivatives.
2. Chec
• Higher-order derivatives can be written in “prime” notation: f � (x), f �� (x), f ��� (x), d(a )
x
then of
f (4) (x), or in Leibniz notation: dy d2 y d3 y d4 y = ax ln a When in doubt, convert ax to ex ln a .
dx , dx2 , dx3 , dx4 . � dx extremu
• The derivative at a particular point a is most often expressed as f (a) or
� dy �
. • If
dx �
x=a d(ln x) 1
Logarithmic: = Found using implicit differentiation. then
dx x • If
When in doubt, convert
d(loga x) 1 then
= loga x toln x
ln a . • If
dx x ln a
then

This downloadable PDF copyright © 2004 by SparkNotes LLC. SPARKCHARTS™ C++ page 3 of 4
 TAKING DERIVATIVES (CONTINUED)
Trigonometric: Found using the definition of derivative and the Squeeze Theorem. If you Inverse Trigonometric: A pain. Found by implicit differentiation.
know the derivatives of sin x and cos x, you can find all the rest using the definitions of
d(sin−1 x) 1 d(cos−1 x) 1
the trigonometric functions and the quotient rule. =√ = −√
dx 1 − x2 dx 1 − x2
d(sin x) d(cos x)
= cos x = − sin x
dx dx d(tan−1 x) 1 d(cot−1 x) 1
= =−
d(tan x) d(cot x) dx 1 + x2 dx 1 + x2
= sec2 x = − csc2 x
dx dx
d(sec x) d(csc x) d(sec−1 x) 1 d(csc−1 x) 1
= sec x tan x = − csc x cot x = √ =− √

TM
dx x x2 − 1 dx x x2 − 1

SPARKCHARTS
dx dx

USING DERIVATIVES
THE TANGENT LINE: APPROXIMATING f (x) NEAR x = a Alternatively, you can use the second derivative test: Mean Value Theorem (MVT):
• If f �� (a) < 0, then f (a) is a local maximum. A generalization of Rolle’s
�The tangent line to a curve y = f (x) at the point
• If f �� (a) > 0, then f (a) is a local minimum. Theorem. If f (x) is continu-
a, f (a) is given by the equation
�
• If f �� (a) = 0, then you must check whether f � (x) ous on the closed interval
y = f (a) + f � (a)(x − a) .
switches sign around x = a. [a, b] and differentiable on
The tangent line gives a (very) crude approximation to the open interval (a, b),
TIP: Often, but not always, f �� (a) = 0 means that f (a) is
f (x): If h is small, then f (a + h) ≈ f � (a)h + f (a) . then there exists a point
neither a minimum nor a maximum. Counterexample:
Useful when f � (a) is known and f (x) is hard to compute. c ∈ (a, b) such that the
f (x) = x4 has f � (0) = f �� (0) = 0. The second derivative
slope of the tangent to f (x)
test tells you nothing, but the changing sign of the first deriva-
at x = c is the same as the
tive indicates that f (0) = 0 is a local minimum.
slope of the secant line f (b)−f (a)
MVT: f � (c) = b−a
3. Check endpoints: If the domain is a closed interval [a, b], �through � the �two points f (b)−f (a)
a, f (a) and b, f (b) : that is, f (c) = b−a .
�
�
always check f (a) and f (b) when looking for extrema. Also
check any boundary points that are included in the domain. Two functions having the same derivative differ by a con-
stant: If f � (x) = g � (x), then f (x) = g(x) + C, for some
Approximating f (x) near a: For small h, TIP: Non-included boundary points (including points where
)g (x)
� real C . Equivalently, a function has only one family of
f (a + h) ≈ f � (a)h + f (a) f (x) is not defined, especially at vertical asymptotes) and
antiderivatives. This theorem follows from the MVT. See
behavior at ±∞ (including horizontal asymptotes) may affect
DISPLACEMENT, VELOCITY, ACCELERATION: 2 the Calculus II SparkChart for more on antiderivatives.
existence of global extrema. Ex: f (x) = (x2 + 1)e−x ,
�
(x) MOTION IN ONE DIMENSION graphed under “Maxima and Minima”, has a local minimum at Extreme Value Theorem: A function f (x) continuous on
Suppose a particle’s position on a line in meters at time t x = 0 but no global minimum, though f (x) > 0 for all real x. the closed inteval [a, b] will assume a global maximum and
seconds is determined by the function s(t). a global minimum somewhere on [a, b].
SKETCHING GRAPHS: SUMMARY
• The first derivative s� (t) of the position function gives the L’HÔPITAL’S RULE
riting it:
instantaneous rate of change of motion; in other words, the 1. Endpoints: If the domain is an interval, evaluate the
to evaluate indeterminate form limits: 0 and ±∞ .
0 ±∞
particle’s instantaneous velocity v(t) = s� (t). The units function at the endpoints. If the domain is the whole real Used
for the first derivative are “meters per second,” m/s. Suppose both f (x) and g(x) are differentiable around a
line, establish what happens at ±∞. Horizontal asymp-
• The first derivative v � (t) of the velocity function will give and g � (x) �= 0 on an interval near a (except perhaps at a).
totes will appear if limx→±∞ f (x) is finite. Evaluate f (0)
the instantaneous rate of change of velocity, or accelera- to find the y -intercept. If lim f (x) = 0 and lim g(x) = 0
x→a x→a
hen it is tion: a(t) = v � (t) = s�� (t). The units for acceleration are 2. Gaps: Find all isolated points x = a where f (a) is not OR
“meters per second per second,” m/s2 . If lim f (x) = ±∞ and lim g(x) = ±∞ ,
easiest defined. For each point a, look at limx→a± f (x). A verti- x→a x→a
on with cal asymptote will appear if limx→a± f (x) = ±∞. A f (x)
then limx→a g(x) = limx→a g� (x) .
f � (x)
MAXIMA AND MINIMA

Series Editors: Sarah Friedberg, Justin Kestler

removable discontinuity (hole in the graph) will appear if
A local� minimum • L’Hôpital’s Rule can also be applied if the limit is one-
� (or maximum) is a limx→a f (x)exists and is finite.
point c, f (c) such that f (c) is the sided (x → a± ).
, which 3. x-intercepts: If it is easy to determine when f (x) = 0,
• L’Hôpital’s Rule can also be applied if the limit is taken as

Anna Medvedovsky
least (or greatest) value of the func- do so. If not, evaluating the function at the critical points
point on x approaches infinity (x → ±∞).

Contributors: Jacob Kaufman,

tion in some interval around c. A local and the endpoints will
rst solve • If f � (x) and g � (x) also satisfy the conditions for
minimum or maximum is also called indicate where to look for

Illustration: Matt Daniels

L’Hôpital’s Rule, higher derivatives can be taken until the

Design: Dan O. Williams

a relative minimum or maximum. No global minumum zeroes. limit is well-defined.

$7.95 CAN
The global minimum (or maximum) is the point where 4. Rise and fall: Determine • L’Hôpital’s Rule cannot be applied to a fraction if the top
then the intervals where the limit is infinite and the bottom limit is zero, or vice versa.
f (x) assumes its least (or greatest) value in the domain
considered. The global minimum or maximum is also called function is increasing and Ex: lim x−1ln x
. Since lim ln x = 0 and lim x − 1 = 0 , use
the absolute minimum or maximum. decreasing by looking at x→1 x→1 x→1

3 9$4.95
the sign of f (x). If
� L’Hôpital’s Rule:
If the domain is a closed interval (an interval that d(ln x)

f � (x) > 0 , then f (x) is lim ln x = lim d(x−1) dx

= lim 1/x = 1.
includes its endpoints), then a continuous function will x−1 x−1 x→1 x−1 1
increasing. If f � (x) < 0,
dx

flat. always have global minimum and maximum points, • L’Hôpital’s Rule can be used to evaluate other indetermi-
possibly at one of the endpoints. (This is the Extreme then f (x) is decreasing. Horizontal asymptote: y = 2
1
nate forms, such as ±∞ · 0. The key is to convert the www.sparknotes.com/errors

20593 36235
Value Theorem.) 5. Local extrema: Find all Vertical asymptote: x = −1 and x = 2 expression to 0 or ±∞ .
Report errors at

0 ±∞
local extrema by looking at Removable discontinuity at x = 1
pe m. The word extremum can be used to mean either minimum Ex: limx→−∞ xex . Convert to the expression limx→−∞ e−x x
,
the critical points where
−∞
or maximum. The plural of extremum is extrema. f (x) = 0 or where f (x) is not defined.
� � which is an indeterminate form ∞ . Applying L’Hôpital’s Rule,
6. Concavity: Determine when the function cups up or convert to limx→−∞ −e−x = 0.1
Critical points are points where f � (x) is zero or undefined.
All extrema—that is, all minima and maxima—happen down by looking at the sign of f �� (x). If f �� (x) > 0, the
either at endpoints or at critical points. function is concave up; if f �� (x) < 0, then f (x) is concave SKETCH OF A FUNCTION AND TWO OF ITS DERIVATIVES
down. If f �� (a) = 0, then the function is temporarily not
+ a1 How to find extremum points:
curving at x = a; if f (x) is changing sign near x = a,
��
7

A: Vertical tangent
1. Check critical points where f (x) exists but f � (x) is not B: Local minimum
then this is a point of inflection (change in concavity).
defined. Such a point may be a local extremum, as in C: Point of Inflection
ral” f (x) = |x| at x = 0. It may be a point of discontinuity. Or THEOREM HIGHLIGHTS D: Point of Inflection
ly func- E: Point of Inflection
it may be neither. Intermediate Value Theorem: If f (x) is continuous in an inter-
ives. F: Local maximum
2. Check critical points where f � (x) = 0. If f � (a) = 0, val [a, b], then somewhere on the interval it will achieve every • The function is decreasing
then often, but not always, the function will have a local value between f (a) and f (b): if f (a) ≤ M ≤ f (b), then from the y -axis to B,
ex ln a .
extremum at x = a. there exists some c in the interval [a, b] (notation: c ∈ [a, b])
increasing from B to F, and
then decreasing from F on.
• If the sign of f � (x) switches from + to − at x = a, such that f (c) = M. This is a completely intuitive statement! • The function is concave
iation. then f (a) is a local maximum. down from the y -axis to A,
• If the sign of f � (x) switches from − to + at x = a, Rolle’s Theorem: If f (x) is continuous on the closed interval concave up from A to C,
then f (a) is a local minimum. [a, b], differentiable on the open interval (a, b), and satisfies concave down from C to D,
• If the sign of f � (x) does not switch around x = a, f (a) = f (b), then for some c in the interval (a, b), we have concave up from D to E,
concave down from E to G,
then f (a) is neither a maximum nor a minimum. f � (c) = 0.
and not curving from G on.

This downloadable PDF copyright © 2004 by SparkNotes LLC. SPARKCHARTS™ C++ page 4 of 4