Functions of two variables:

Definitions:
a) A function of two variables is a rule g that assigns to each ordered pair ( x, y) in a
set R

a unique number g ( x, y ) . The set R is called the domain of the function

and the corresponding values of g ( x, y ) constitute the range of g . We refer to x and

y as the independent variables. We usually define g ( x, y ) as a formula and assume

that the domain is the largest set of points in the plane for which the formula is
defined and real-valued.
b) The graph of the function z g ( x, y ) , where ( x, y ) R 2 , is the set
S ( x, y, z )

| z g ( x, y ) and ( x, y ) R .

Geometrically, if z g (x , y ) measures the vertical distance from the point ( x, y) in

the xy -plane, then z describes the points ( x, y, g ( x, y )) that lie on a surface in space.
Thus, the graph of g is a surface in
domain R (see figure below).

whose projection onto the xy -plane is the

c) In an analogous manner, we define a function of three variables as a rule f that

assigns to each triple ( x, y, z ) in a set D

a unique number f ( x, y, z ) . We refer

to f ( x, y, z ) as a function of three independent variables x, y, z ; define it as a formula

and assume that its domain is the largest set of points in space
formula is defined and real-valued.

for which the

Example 1: Let T denote the temperature and D a metal plate in the xy -plane. The function
T ( x, y) then gives the temperature at a point ( x, y) D . If T varies not only on the plate but
also with time t , then T is a function of three independent variables x, y, t. Similarly, if the
metal plate lies in space, then T ( x, y, z, t ) is a function of four independent variables x, y, z, t.
Example 2: A polynomial function p( x, y) is a function of two independent variables x
and y. It is a sum of the functions of the form Cx m y n with nonnegative integers m , n and

C a constant; for instance p( x, y) 3x5 y3 7 x2 y 2 x 3 y 11 . A rational function is a

quotient of two polynomial functions.
Example 3: Let f ( x, y) 9 x 2 4 y 2 . Describe the domain and range of f .
Solution: The domain of f is the set of all ordered pairs ( x, y) for which f is defined and
real-valued. So, we must have 9 x 2 4 y 2 0 or equivalently, x 2 4 y 2 9 in order that
square root is defined and real. Since a simple closed curve in xy -plane divides the plane into
disjoint regions (inside and outside) with curve itself as the common boundary, and the point
(0, 0) satisfies x 2 4 y 2 9 , we get that the domain is the set of all points ( x, y) that is inside
or on the ellipse x 2 4 y 2 9 . We call it an elliptical region. See figure below.

Operations with functions of two variables:

If f ( x, y) and g ( x, y) are functions of two variables with domain D , then

( f g )( x, y ) f ( x, y ) g ( x, y ),
( f g )( x, y ) f ( x, y ) g ( x, y ),
( f g )( x, y ) f ( x, y ) g ( x, y ),
f
f ( x, y )
,
( x, y )
g
g
(
x
,
y
)

g ( x, y ) 0.

Trace and level curve of the graph of a function:

Definitions: Consider a function z f ( x, y) of two variables with domain R

and a

plane z c parallel to xy -plane. The equation f ( x, y) c is a curve, which represents the

intersection of the surface z f ( x, y) with the plane z c .

1) This cross-section of the surface z f ( x, y) in the plane z c is called the trace

or the contour curve of the graph of f in the plane z c .
2) The projection of the trace of f onto the xy -plane is called a level curve of
height c . That is, the set f 1 c x, y R | f x, y c, where c is a
constant, is a level curve of f .
Remarks:
1. As the constant c varies over the range of f , a family of level curves is generated. If

c is outside the range of f , then the trace is empty and hence no level curve is
obtained. We usually think of trace as a slice of the surface at a particular location
(height) and the level curve is simply the projection of this slice on to the xy -plane.
Note that the level curves lie in the domain of f .
2. Similarly, we obtain the trace of the graph of f in the plane x d (respectively,
y e ) and the corresponding level curve onto the yz -plane (respectively, the level

curve onto the xz -plane) by taking the intersection of the surface z f ( x, y) in the
plane x d (respectively, y e ) and their projection onto yz -plane (respectively,
onto the xz -plane).
3. The level curves are used to draw a 2-dimensional profile of the surface
z f ( x, y), such as a mountain range. Such a profile, called topographical or contour
map, is obtained by sketching the family of level curves in the xy -plane and labeling
each curve to show the elevation to which it corresponds.

4. Note that the regions on a topographical map where the level curves are crowded
together correspond to steeper portions of the surface. Isotherms and isobars are the
level curves used to indicate the places of same temperature and pressure,
respectively, on the weather report in the news.
Example 1: Sketch the graph of the function f ( x, y) x 2 y 2 .

Solution: We find the traces of the surface z x 2 y 2 in planes parallel to coordinate planes.
First, consider its trace in the plane z k : it is x 2 y 2 k . If k 0 this equation has no real
solution, so there is no trace. If k 0 then the graph of x 2 y 2 k is a circle of radius

centered at the point (0,0, k ) on the z -axis (figure (a) below). Thus, for nonnegative values
of k the traces parallel to the xy -plane form a family of circles, centered on the z-axis, whose
radii start at zero and increase with k. This suggests that the surface has the form shown in
figure (b) below:

To obtain more detailed information about the shape of this surface, we can examine the
traces of z x 2 y 2 in planes parallel to yz -planes. Such planes have equations of the form

x k so that we get z k 2 y 2 or y 2 z k 2 . For k 0 this equation reduces to y 2 z

which is a parabola in the plane x 0 that has its vertex at the origin, opens in the positive z
-direction and is symmetric about z -axis (the blue parabola in figure (a)).

Note that the effect of k 2 term in y 2 z k 2 is to translate the parabola y 2 z in the

positive z -direction so its new vertex in the plane x k is (k , 0, k 2 ) . This is the red parabola
in fig (a). Thus, the traces in planes parallel to the yz -plane form a family of parabolas
whose vertices move upward as k 2 increases (figure (b)). Similarly, the traces in planes
parallel to the xz -plane ( y k ) have equations of the form x2 z k 2 , which again is a
family of parabolas whose vertices move upward as k 2 increases (figure (c)). This graph of
z f x, y x 2 y 2 is called a circular paraboloid.
Example 2: Sketch the level curves of the surface z 100 x 2 y 2 .
Solution: The domain of f is the entire xy -plane and the range of f is the set of real
numbers less than or equal to 100. Its a circular paraboloid opening downwards and vertex at
(0, 0,100) . In the plane z 75 , the trace or the contour curve of f is the circle f ( x, y) 75
or 100 x2 y 2 75 . It is the circle in the plane z 75 with canter at (0,0,75) and radius 5 .
Therefore, the level curve is the circle x 2 y 2 25 whose center is at the origin and radius 5.

In the xy -plane ( z 0) , the level curve f ( x, y) 0 or equivalently x 2 y 2 100 is the circle

with center at origin and radius 10. Similarly, the level curve f ( x, y) 51 is the circle in the
xy -plane of radius 7 and center at the origin: 100 x2 y 2 51 or equivalently x2 y 2 49.

The level curve f ( x, y) 100 or equivalently x2 y 2 0 consists of the origin alone, i.e. it is
a point circle. The graph of the function and these few level curves of the surface
z 100 x 2 y 2 are shown in the figure below.

Traces of the surface z 100 x 2 y 2 in planes parallel to yz and zx planes are shown
below, and the corresponding level curves are parabolas opening downwards and with
vertices on z -axis.

Example 3: Describe the level curves of the function f ( x, y) x 2 y 2 .

Solution: The graph of the function f ( x, y) x 2 y 2 is a hyperbolic paraboloid (saddle
surface) as shown in the figure (a) below:

The level curves have equations of the form y 2 x 2 k where k is a constant. For k 0
these curves are hyperbolas opening along lines parallel to y -axis; for k 0 these curves are
hyperbolas opening along lines parallel to x -axis, and for k 0 the level curves consists of
the intersecting lines y x 0 and y x 0 . See figure (b).
Example 4: Describe the level surfaces of the function f ( x, y, z ) x 2 y 2 z 2 .
Solution: The value of f is the distance from the origin to the point ( x, y, z ) . Each level
surface

x 2 y 2 z 2 c, c 0 , is a sphere of radius c centered at the origin. For c 0 , the

level surface

x 2 y 2 z 2 0 consists of the origin alone. Of course, there is no level curve

for c 0. Thus, the level surfaces of the function are concentric spheres, as shown below.

Example 5: Describe the level surfaces of the function f ( x, y, z ) z 2 x 2 y 2 .

Solution: The level surfaces have equations of the form z 2 x 2 y 2 c . This equation
represents a cone if c 0 , a hyperboloid of two sheets of c 0 , and a hyperboloid of one
sheet if c 0 (see figure below). For detailed proof, see remarks below.

Remarks:
The level surfaces of a function of three independent variables x, y, z , namely

f ( x, y, z ) Ax 2 By 2 Cz 2 Dxy Exz Fyz Gx Hy Iz J

where A, B, C, D, E, F , G, H , I , J are all constants and at least one of A, B, C, D, E, F is not
zero, are called quadric surfaces or quadrics. Six common types of quadric surfaces are
1)
2)
3)
4)
5)
6)

Ellipsoids.
Hyperboloids of one sheet.
Hyperboloids of two sheets.
Elliptic cones.
Elliptic paraboloids.
Hyperbolic paraboloids

Their standard forms of equations, along with their graphs, are shown in the figures below.
Their traces in the planes parallel to the coordinate planes have also been indicated.
Special Cases:
a. If A B C D E F 0 , then f ( x, y, z) Gx Hy Iz J , whose level surface
represents a plane, with direction ratios of the normal to the plane as G, H , I .
b. In the case of ellipsoid, if a b c 1 , then the resulting surface represents a sphere
of radius 1 and centered at the origin.
c. In the case of elliptic cone, if a b 1 , then the resulting surface represents a
circular cone with axis as the z -axis and vertex at the origin.

 Note that there is no minus sign and constant term is 1.

Note that there is one minus sign and constant term is 1.

 Note that there are two minus signs and constant term is 1.

Note that there is no linear term and constant term is 0.

 Note that there is one linear term and two quadratic terms with the same
sign. Also, the constant term is 0.

Note that there is one linear term and two quadratic terms with opposite
signs. Also, the constant term is 0.

Limit of a function of two variables:

The definition of the limit of a function of two variables is completely analogous to
the definition for a function of a single variable. First, we define a two dimensional analog to
an open interval and a closed interval on the real line.
Open and closed disks: Using the formula for the distance between two points ( x, y) and

( x0 , y0 ) in the plane, an open disk D centered at ( x0 , y0 ) with radius 0 is defined by

D ( x, y)

| ( x x0 ) 2 ( y y0 ) 2

(1)

Thus, an open disk D is the set of all the points in the plane that are enclosed by the circle of
positive radius centered at ( x0 , y0 ) but do not lie on the circle. We call this open disk D
the -neighborhood of the point ( x0 , y0 ). The disk D is said to be a closed disk if the strict
inequality, in the set (1) is replaced by less than or equal to, . Thus, the set of points that
lie on the circle together with those enclosed by the circle is called the closed disk of radius
0 centered at ( x0 , y0 ) .

Interior and boundary points:

A point ( x0 , y0 ) is an interior point of a set R in the plane

if there is some open disk D

centered at ( x0 , y0 ) that is completely contained in R . If the set R is empty or if every point

of R is an interior point, then R is called an open subset of
boundary point of a set R in the plane

. A point ( x0 , y0 ) is called a

if every open disk centered at ( x0 , y0 ) contains

both points that belong to R and points that do not belong to R . The collection of all
boundary points of R is called the boundary of R . The set R is said to be closed subset of
2
if it contains all of its boundary points. See figures below.

For instance, the open unit disk, its boundary and closed unit disk as subsets of the plane are
illustrated in the figures below.

Note that the empty set and

are both open and closed as subsets of

Bounded and unbounded regions: A region (set) in the plane is bounded if it lies inside a
disk of fixed radius. A region is unbounded if it is not bounded. For example, bounded sets
in the plane include line segments, triangles, interiors of triangles, rectangles, circles, and
disks. Examples of unbounded sets in the plane include lines, coordinate axes, the graphs of
functions defined on infinite intervals, quadrants, half-planes, and the plane itself.

Definition: Let f be a function of two variables and assume that f is defined at all points
of some open disk centered at ( x0 , y0 ) , except possibly at ( x0 , y0 ) . The limit statement
lim

( x , y ) ( x0 , y0 )

f ( x, y) L

means that for every given number 0 , there exists a number 0 such that whenever
the distance between ( x, y) and ( x0 , y0 ) satisfies
0 ( x x0 )2 ( y y0 )2 ,

f ( x, y) satisfies
f ( x, y) L .

Note that the definition says that given any desired degree of closeness 0 , we must be
able to find another number 0 so that all the points lying within a distance of ( x0 , y0 )
are mapped by

to the points within distance

of L on the real line.

Remarks:
When we consider the lim f ( x) of a function of one variable, we need to examine the
x c

approach of x to c from two different directions, namely the left hand side and right hand
side of c on the real line. In fact, these two are the only possible directions in which x can
approach c on the real line (corresponding to left-hand limit and right-hand limit). However,
for a function of two variables, we write that ( x, y) ( x0 , y0 ) to mean that the point ( x, y) is
allowed to approach ( x0 , y0 ) along any of the infinitely many different curves or paths in the
plane

passing through ( x0 , y0 ) .

Existence of limit:
The limit of a function of two variables f ( x, y) is said to exist and equal L , as

( x, y) ( x0 , y0 ) , written symbolically as
lim

( x , y ) ( x0 , y0 )

f ( x, y) L ,

if the function f ( x, y) approaches L along every possible path that ( x, y) takes to approach

( x0 , y0 ) in the plane

within the domain of f .

Now, one obviously cannot check each path individually. This gives us a simple method for
determining that a limit does not exist.
Non-existence of limit:
If f ( x, y) approaches L1 as ( x, y) approaches ( x0 , y0 ) along a path P1 and f ( x, y)
approaches L2 L1 as ( x, y) approaches ( x0 , y0 ) along a path P2 , then the
lim

( x , y ) ( x0 , y0 )

f ( x, y) does not exist.

Our first objective is, therefore, to define the limit of f ( x, y) as ( x, y) ( x0 , y0 ) along a

path or a smooth curve C .
Limit along a curve:
Let C be a smooth parametric curve in

that is represented by the equations

x x(t ), y y(t ), t [a, b] .

If x0 x(t0 ), y0 y(t0 ) for some t0 [a, b] , then the

lim

( x , y ) ( x0 , y0 )

f ( x, y) along the curve C is

defined by

lim

( x , y ) ( x0 , y0 )
(along C )

f ( x, y) lim f ( x(t ), y(t )).

t t0

(2)

In the right hand side of the formula (2), the limit of the function of t must be treated as a
one sided limit if ( x0 , y0 ) is an end point of C . A geometric interpretation of the limit along
a curve for a function of two variables is depicted in the figure below.

As the point ( x(t ), y(t )) moves along the curve C in the xy -plane towards ( x0 , y0 ) , the point

( x(t ), y(t ), f ( x(t ), y(t ))) moves directly above (or below) it along the graph of z f ( x, y)
with f ( x(t ), y(t )) approaching the limiting value L.
Example 1: Consider a function of two variables defined by the formula

f ( x, y)

xy
,
x y2
2

( x, y) (0, 0) .

Find the limit of f ( x, y) as ( x, y) (0,0) along the following paths:

(a) the x -axis,
(b) the y -axis,
(c) the line y x ,
(d) the line y x ,
(e) the parabola y x 2 .
Solution: (a) The parametric equations of the x -axis are x t , y 0, t , with (0, 0)
corresponding to t 0 . So, we have

lim

( x , y ) (0,0)
(along x axis)

0
f ( x, y) lim f (t , 0) lim 2 lim 0 0.
t 0
t 0
t t 0

(b) The y -axis has parametric equations x 0, y t , t

, with (0, 0) corresponding to

t 0 . So, we have

lim

( x , y ) (0,0)
(along y axis)

0
f ( x, y) lim f (0, t ) lim 2 lim 0 0.
t 0
t 0
t t 0

y x has the parametric equations x t , y t , t , with (0, 0)

corresponding to t 0 . Thus, we get

(c) The line

lim

( x , y ) (0,0)
(along y=x )

t2
1
1
f ( x, y) lim f (t , t ) lim 2 lim .
t 0
t 0
t

0
2
2
2t

(d) The line y x has the parametric equations x t , y t , t , with (0, 0)

corresponding to t 0 . Thus, we get
lim

( x , y ) (0,0)
(along y=-x )

t2
1 1
f ( x, y) lim f (t , t ) lim 2 lim .
t 0
t 0 2t
t 0 2
2

(e) The parabola y x 2 has parametric equations x t , y t 2 , t , with (0, 0)

corresponding to t 0 . Thus, we have

lim

( x , y ) (0,0)
(along y x 2 )

t3
t
f ( x, y) lim f (t , t 2 ) lim 2 4 lim
0.
2
t 0
t 0 t t

t 0 1 t

We conclude therefore that the limit of the function f ( x, y )

xy
does not exist. See
x y2
2

figures below for geometrical view of these limits.

At (0, 0) , there is a sudden dip (hole) on the surface which supports our conclusion.

Example 2: Using the definition of limit, show that

lim

( x , y ) ( a , b )

y b and

lim

( x , y ) ( a , b )

x a.

Solution: We prove the first limit. The proof for the second limit is similar. Given any
number 0 , we must find another number 0 such that

y b whenever

0 ( x a)2 ( y b)2 . Note that

( x a)2 ( y b)2 ( y b)2 y b

so that taking , we have that

y b ( y b) 2 ( x a ) 2 ( y b ) 2

whenever 0 ( x a)2 ( y b)2 . This proves that

lim

( x , y ) ( a , b )

y b.

Algebra of limits:
Using the definition of limit, we can prove the following results:

lim

f ( x, y) g ( x, y) ( x, ylim
f ( x, y) ( x, ylim
g ( x, y).
) ( a ,b )
) ( a ,b )

lim

2)

( x , y ) ( a ,b )

f ( x, y) lim g ( x, y ) .
f ( x, y) g ( x, y) ( x, ylim
) ( a ,b )
( x, y )( a ,b )

3)

lim

f ( x, y)
f ( x, y ) ( x , ylim
) ( a ,b )
, provided that
g ( x, y )
lim g ( x, y)

1)

( x , y ) ( a ,b )

( x , y ) ( a ,b )

lim

( x , y ) ( a ,b )

g ( x, y) 0 .

( x , y ) ( a , b )

Remark:
Using Example 2 and the algebra of limits, we can easily prove that the limit of a polynomial
function in two variables always exists and is found simply by substitution.
Example 3: Evaluate

lim 5 x3 y 2 9 .

( x , y ) (1,4)

Solution: Using the preceding remark, we get

lim

( x , y ) (1,4)

5 x3 y 2 9 lim 5 x3 y 2 lim 9
( x , y ) (1,4)
( x , y ) (1,4)
3

5 lim x lim y 9
( x , y )(1,4) ( x , y )(1,4)
5 1 4 9 80 9 71.
3

Example 4: Evaluate

2 x 2 y 3xy
lim
.
2
( x , y ) (2,1)
5 xy 3 y

Solution: Note that it is the limit of the quotient of two polynomial functions, that is, a
rational function. Since the limit in the denominator is

lim

( x , y ) (2,1)

5xy

2 x 2 y 3xy
we have that lim

2
( x , y ) (2,1)
5 xy 3 y

3 y 10 3 13 0,

lim

( x , y ) (2,1)

lim

2 x y 3xy 14
.
5xy 3 y 13
2

( x , y ) (2,1)

x 2 x xy y
Example 5: Evaluate lim
.
( x , y ) (0,0)
x y

x 2 x xy y
Solution: Note that f x, y
is not defined for x y . So, the domain of
x y

f is

x, y | x y . Now, for x y , we have

x 2 x xy y ( x 1)( x y )
f x, y
x 1.

y
x

Therefore,

lim

x , y 0,0

f x, y

lim

x , y 0,0

x 1 x, ylim
x x, ylim
1 0 1 1.
0,0
0,0

2 xy
lim 2
does not exist by evaluating this limit along the
x , y 0,0 x y 2

x -axis, the y -axis and along the line y x .

Example 6: Show that the

Solution: Note that the denominator in the function f x, y

that f 0, 0 is not defined. Therefore, the domain of f is

2 xy
is zero at 0, 0 so
x y2
2

0,0 . Now, we evaluate

the limit along the given curves.

If we approach the origin along the x -axis, whose parametric equations are x t , y 0, t 0
and the point 0, 0 corresponds to t 0 , we find that

lim

x , y 0,0

f x, y

2t 0
2 xy
0
lim 2

lim

lim

0.
t 0 t 2 0
t 0 t 2
x , y 0,0 x y 2

If we approach the origin along the

y -axis, whose parametric equations are

x 0, y t , t 0 and the point 0, 0 corresponds to t 0 , we find that

lim

x , y 0,0

f x, y

2 0 t
2 xy
0
lim 2
lim
lim 2 0.

2
2
x , y 0,0 x y

t 0 0 t t 0 t

If we approach the origin along the x y , whose parametric equations are x t , y t , t 0

and the point 0, 0 corresponds to t 0 , we find that

lim

x , y 0,0

2t 2
2t 2
2 xy
f x, y lim 2
lim
lim
lim 1 1.
x , y 0,0 x y 2 t 0 t 2 t 2 t 0 2t 2 t 0

Note that since the limiting values are different when x, y 0,0 along different curves, it
follows that the limit of f does not exist at the origin.
Example 7: Show that the

x2 y
does not exist.
lim 4
x , y 0,0 x y 2

Solution: Note that the denominator in the function f x, y

that f 0, 0 is not defined. Therefore, the domain of f is

x2 y
is zero at 0, 0 so
x4 y2
0,0 . We prove that the

x2 y
lim
limit does not exist by showing that the limit
has different values as
x , y 0,0 x 4 y 2

x, y 0,0

along different curves in the plane. First, we let x, y approach the origin

along the line y mx . Note that the x -axis, the y -axis and the line y x are special cases
of the line y mx . The line y mx has parametric equations x t , y mt , t 0 and the
point 0, 0 corresponds to t 0 . So, we have

t 2 mt
x2 y
mt
lim 4
lim 4
0.
lim 2

2
2
2
x , y 0,0 x y

t 0 t mt t 0 t m
However, if we approach the origin along the parabola y x 2 , which has parametric
equations x t , y t 2 , t 0 and the point 0, 0 corresponds to t 0 , we find that

t 2 t 2
x2 y
t4
1 1

lim
lim 4 lim .
lim
t 0 4
t 0 2t
t 0 2
2 2
x , y 0,0 x 4 y 2
2

t t
This completes the proof.
Remark: It is often possible to show that a limit does not exist by the methods illustrated in
Examples 6 and 7. However, it is impossible to try to prove that lim f x, y exists by
x , y x0 , y0

showing that the limiting value of f x, y is the same along every curve that passes through

x0 , y0

since there are infinitely many such curves. However, one tool that can be used to

prove that a limit exists is the following generalization of the Squeeze theorem:

Squeeze Theorem:
Suppose that f x, y L g x, y for all x, y in the interior of some circle centered at

a, b ,

except possibly at

a, b .

equals L . To find L , we compute

If

lim

x , y a ,b

lim

x , y a ,b

g x, y 0 , then

lim

x , y a ,b

f x, y exists and

f x, y along a few paths from x, y to a, b .

x2 y
Example 8: Evaluate lim 2
x , y 0,0 x y 2

Solution: First, we find the limit along a few paths. Along the path x 0 , whose parametric
equations can be written as x 0, y y with 0,0 corresponding to y 0 , the given limit is

x2 y
02 y
0
lim 2

lim
lim 2 lim 0 0.

2
2
2
x , y 0,0 x y

y0 0 y y0 y y0
Next, along the path y 0 , whose parametric equations are x x, y 0 , we similarly find
that

x2 y
x2 0
0
lim 2

lim
lim 2 lim 0 0.

2
2
2
x , y 0,0 x y

x0 x 0 x0 x x0
Further, along the path x y , whose parametric equations are x t, y t , with

0,0

corresponding to t 0 , we have

x2 y
t2 t
t3
t
lim
lim 2 2 lim 2 lim 0.
t 0 t t
t 0 2t
t 0 2
x , y 0,0 x 2 y 2

Thus if the limit exists, then it must be equal to 0 . To show this, let f x, y

L 0 , and consider
f x, y L f x , y 0

x2 y
.
x2 y2

Since x 2 y 2 x 2 , we have for x 0 ,

f x, y L

Since

lim

x , y 0,0

x2 y
x2 y

y.
x2 y2
x2

x2 y
y 0 , the Squeeze Theorem gives us lim 2
0.
x , y 0,0 x y 2

x2 y
and
x2 y2

 x 12 ln x
Example 9: Prove that lim
exists and hence evaluate it.
x , y 1,0 x 12 y 2

Solution: Again, we find the limit along a few paths. Along the path x 1 , whose parametric
equations are x 1, y y with 1,0 corresponding to y 0 , the given limit is

x 12 ln x
02 ln x
0
lim

lim
lim 2 lim 0 0.

2
2
2
2
y 0 0 y
x , y 1,0 x 1 y

y 0 y y 0

Next, along the path y 0 , whose parametric equations are x x, y 0 , we similarly find
that

x 12 ln x
x 12 ln x
lim
lim
lim ln x 0.
2
x 1
x1
x , y 1,0 x 1 2 y 2

x 1
A third path through 1,0 is y x 1 , whose parametric equations are x t, y t 1 with

1,0

corresponding to t 1 . Along this path, we have

x 12 ln x
t 12 ln t
ln t
lim
0.
lim
lim
2
2
2
2
t 1
t 1 2
x , y 1,0 x 1 y
t

x 1 ln x
f x, y
2
x 1 y 2
2

Thus, if the limit exists, then it must be equal to 0 . To show this, let
and L 0 , and consider

f x, y L f x, y 0

x 1 ln x .
2
x 1 y 2
2

Since x 1 y 2 x 1 , we have for x 1 ,

2

f x, y L

x 1 ln x
2
x 1 y 2
2

x 1 ln x
2
x 1
2

ln x .

x 12 ln x
Since lim ln x 0 , the Squeeze Theorem gives us lim
0.
x , y 1,0
x , y 1,0 x 1 2 y 2

Note: Since ln x is a function of x alone, lim

x , y 1,0

ln x lim ln x 0 .
x 1

Example 10: Show that the

Solution: As

x, y

1
does not exist.
x , y 0,0 x y 2
lim

approaches

0,0

along any path, the values of f x, y

1
x y2
2

increase without bound or approach , that is there is no finite number to which f x, y

approaches when x, y 0,0 . Hence, the limit does not exist.
Example 11: Show that the following limit does not exist

Solution: The domain of the function f x, y

plane except for the point

0,0 .

x2 y2
.
lim 2
x , y 0,0 x y 2

x2 y2
consists of all points in the xy x2 y2

To show that the given limit does not exist as

x, y 0,0 , consider approaching 0,0 along two different paths in the domain of f
passing through origin. First, let x, y 0,0 along x -axis, whose parametric equations
are x x, y 0 with 0,0 corresponding to x 0 . The limit along this approach is
x2 y 2
x 2 02
lim 2

lim
lim 1 1.
x , y 0,0 x y 2 x 0 x 2 02

x0
Next, let x, y 0,0 along the line y x . Then, we obtain

x2 y2
x2 x2
0
lim 2

lim
lim 2 0.

2
2
2
x , y 0,0 x y

x , x 0,0 x x x0 2 x
Since the limit of f along two different paths is different, the limit of f x, y

x2 y2
x2 y2

does not exist as x, y 0,0 .

Exercises:

y
1) Prove that lim
does not exist.
x , y 1,0 x y 1

2) Prove that

3) Evaluate

xy 2
lim 2
does not exist.
x , y 0,0 x y 4

5x 2 y
lim 2
,
x , y 0,0 x y 2

x 2 xy
lim
and
x , y 0,0
x

lim

x , y 0,0

y 2 ln x 2 y 2 .

Continuity of a function of two variables:

Definition: The function f x, y of two variables is continuous at the point x0 , y0 if

f x0 , y0 is defined;

a)
b)
c)

lim

f x, y exists;

lim

f x, y f x0 , y0 .

x , y x0 , y0
x , y x0 , y0

The function f is continuous on a set S if it is continuous at each point in S .

Geometrically, this says that f is continuous if the surface z f x, y has no gaps or
holes.
Remark: A polynomial function p x, y in two variables is continuous throughout the
plane. A rational function

f x, y
of two variables is continuous wherever g x, y 0 .
g x, y

Example 1: Test the continuity of the function f x, y

x y
1
and g x, y
.
2
2
x y
y x2

Solution: The function f is a rational function of x and y , so it is discontinuous only

where it is undefined; namely at

0, 0 .

Similarly, g is a rational function and is

discontinuous only where it is undefined. That is, wherever y x 2 0 . Thus, the function g
is continuous at every point on the plane except those lying on the parabola y x 2 .
Example 2: Show that f is continuous at 0, 0 where
1

x0
y sin
f x, y
x

x 0.
0

Solution: In order to show that f is continuous at 0, 0 , we must show that for any 0 ,
there exists a 0 such that
f x, y f 0, 0 y sin

Now, since sin

1
whenever 0 x 2 y 2 .
x

1
1
1 for x 0 , we have y sin y for all x 0 . If x, y lies in the disk
x
x

of radius 0 centered at the origin, then x 2 y 2 2 so that the points 0, y satisfy

02 y 2 2 or y 2 2 or y . In other words, points satisfying y lie in the disk.

Hence,

if

we

let

it

follows

that

f x, y f 0,0 y

whenever

0 x 2 y 2 2 . This completes the proof.

Example 3: Find all the points where the function g x, y is continuous, where
x4
,

g x, y x 2 y 2
0,

x, y 0,0
x, y 0,0 .

Solution: The function g is a quotient of polynomials, except at the origin. Since the
denominator is never zero, g must be continuous at every point

x, y 0,0 .

Now, we

check the continuity at the origin. Notice that for all x, y 0,0 ,

g x, y 0 g x, y
Since

lim

x , y 0,0

x 2 0 , we have that

lim

x , y 0,0

x4
x4

x2.
2
2
2
x y
x

g x, y 0 , by using Sandwich theorem. This

shows that g is continuous at the origin also. Hence, g is continuous everywhere.

Theorem: (Continuity of certain types functions)
(a) Suppose that f x, y is continuous at a, b and g x is continuous at f a, b . Then the
composite function h x, y g f x, y g f x, y is continuous at a, b .
(b) If g x is continuous at x0 and h y is continuous at y0 , then the function

f x, y g x h y is continuous at x0 , y0 .
(c) If f x, y is continuous at

x0 , y0 and

if x t and y t are continuous at t0 with

x t0 x0 and y t0 y0 , then the composition f x t , y t is continuous at t0 .

Example 4: Determine where the functions f x, y 3x 2 y 5 , g x, y sin 3x 2 y 5 and
h x, y e x

are continuous.

Solution: Notice that h x, y g f x, y , where g t et and f x, y x 2 y . Since g is

continuous for all values of t and f is a polynomial in x and y (and hence continuous for
all x and y ), it follows from preceding theorem (a) part that h is continuous for all x, y .
Similarly, f and g are continuous at every point x, y in the xy plane.

Limit and Continuity for functions of three variables:

The concepts of limit and continuity for functions of two variables in
naturally to functions of three variables in 3 . In particular, the limit statement
lim

( x , y , z )( x0 , y0 , z0 )

extend

f ( x, y , z ) L

means that for every given number 0 , there exists a number 0 such that whenever
the distance between ( x, y, z ) and ( x0 , y0 , z0 ) satisfies

0 ( x x0 )2 ( y y0 )2 z z0 ,
2

f ( x, y, z ) satisfies
f ( x, y , z ) L .
The function f x, y, z is continuous at the point P0 x0 , y0 , z0 if
a)
b)
c)

f x0 , y0 , z0 is defined;

lim

f x, y, z exists;

lim

f x, y, z f x0 , y0 , z0 .

x , y , z x0 , y0 , z0
x , y , z x0 , y0 , z0

Most commonly considered functions of three variables are continuous wherever they are
defined. For example, the function

f x, y , z

3
x2 y2 2z

is continuous everywhere except where it is not defined. That is, f is discontinuous for all

1 2
x y 2 . Thus, f is continuous everywhere in
2
1
3
except at any point inside or on the paraboloid given by z x 2 y 2 . Similarly, the
2
function
points for which x 2 y 2 2 z 0 or z

g x, y , z

1
x y2 z
2

is continuous at each point in space except at the points on the paraboloid given by
z x2 y 2 .

Partial differentiation and Partial Derivatives:

If z f x, y then one can inquire how the value of z changes if y is held fixed and x
is allowed to vary, or if x is held fixed and y is allowed to vary. For example, the ideal gas
law in physics states that under appropriate conditions, the pressure P exerted by a gas is a
function of the volume V of the gas and its temperature T :

Pk

T
, where k is a constant.
V

Thus, a physicist studying gases might be interested in the rate of change of the pressure if
the volume is held fixed and the temperature is allowed to vary, or if the temperature is held
fixed and the volume is allowed to vary. We now define a derivative that describes such rates
of change.
Suppose that x0 , y0 is a point in the domain of a function f x, y . If we fix y y0 ,
then f x, y0 is a function of the variable x alone. The value of its derivative at x x0 , i.e.

d
f x, y0
dx
x x0
then gives us a measure of the instantaneous rate of change of f with respect to x at the point

x0 , y0 . Similarly, the value of the derivative of f x0 , y at

y y0 , that is

d
f x0 , y
dx
y y0
gives us a measure of the instantaneous rate of change of f with respect to y at the point

x0 , y0 . These derivatives are so basic to the study of differential calculus of multivariable

functions that they have their own name and notation: The process of differentiating a
function of several variables with respect to one of its variables while keeping the other
variable(s) fixed is called partial differentiation and the resulting derivative is called a
partial derivative of the function.
Definition: If z f x, y , then the partial derivative of f with respect to x is the function

f x x, y defined by

f x x, y lim
h0

f x h, y f x , y
h

provided the limit exists. Thus, we find the partial derivative of f with respect to x by
regarding y as a constant while differentiating the function f with respect to x .

Similarly, the partial derivative of f with respect to y is the function f y x, y defined by

f y x, y lim
h0

f x, y h f x, y
h

provided the limit exists. Thus, we find the partial derivative of f with respect to y by
regarding x as a constant while differentiating the function f with respect to y .
Notations: The partial derivatives f x x, y and f y x, y of the function z f x, y are
denoted by following symbols:

f z

f x, y
x x x
f z
f y x, y

f x, y .
y y y
f x x, y

The values of the partial derivatives of f x, y at the point a, b are denoted by

f
f
f x a, b and
f y a, b .
x a ,b
y a ,b
Example 1: Let f x, y x 2 sin 3x y 3 . Evaluate the partial derivatives (a) f x 3,0 and
(b) f y 1,1 .
Solution: (a) We have f x x, y 2 x sin 3x y 3 x 2 cos 3x y 3 3 , so that

f x 3,0 2 sin 3 cos .
3
3
3
2

(b) Again, we have f y x, y x 2 cos 3x y 3 3 y 2 so that

f y 1,1 3 1 1 cos 3 1 3cos 4 .
2

Example 2: Let f x, y, z x 2 2 xy 2 yz 3 . Determine f x , f y , f z .

Solution: For finding f x , treat y and z as constants and consider f as a function of x
alone. Thus, we have

f x x, y , z

2
x 2 xy 2 yz 3 2 x 2 y 2 .

Similarly, we obtain f y x, y, z 4 xy z 3 and f z x, y, z 3 yz 2 .

Example 3: Let z be defined implicitly as a function of x and y by the equation

x 2 z yz 3 x .

(1)

Determine z x and z y .
Solution: To find z x , first we differentiate equation (1) implicitly with respect to x,
treating y as a constant. So, we get

2 xz x 2

z
z
3 yz 2
1.
x
x

Then we solve the preceding equation for z x :

z
1 2 xz
2
.
x x 3 yz 2
Similarly, holding x constant and differentiating implicitly with respect to y , we obtain

x2

so that

z
z
z 3 3 yz 2
0
y
y

z
z3
. Note that both partial derivatives are rational functions of x, y, z .
2
y x 3 yz 2

Geometrical Interpretation: Partial derivative as the slope of a Tangent Line

A geometric interpretation of partial derivatives is indicated in figures below:

Fig.(a)

Fig.(b)

In the Fig. (a) above, the plane y y0 intersects the surface z f x, y in a curve C parallel
to the xz plane. That is, C is the trace of the surface in the plane y y0 . An equation for
this curve is z f x, y0 , and because y0 is fixed, the function depends only on x . Thus,
we can compute the slope of the tangent line to C at the point P x0 , y0 , f x0 , y0 in the
plane y y0 by differentiating f x, y0 with respect to x and evaluating the derivative at

x x0 . That is, the slope is f x x0 , y0 . We formally state this observation as follows:

Conclusion 1: f x x0 , y0 represents the slope of the line parallel to the xz plane and
tangent to the surface z f x, y at the point x0 , y0 , f x0 , y0 . That is, it gives the slope
of the surface in the x direction.
An analogous interpretation for f y x0 , y0 is shown in the Fig. (b) above. We state it as
follows:
Conclusion 2: f y x0 , y0 represents the slope of the line parallel to the yz plane and
tangent to the surface z f x, y at the point x0 , y0 , f x0 , y0 . That is, it gives the slope
of the surface in the y direction.
Hence, partial derivatives geometrically provide the slope of the tangent lines to the surface.
Figures (a) and (b) are combined below

to deduce the following assertion:

Conclusion 3: Conclusions 1 and 2 imply that we have two tangent lines associated with the

surface z f x, y at the point P x0 , y0 , f x0 , y0 . These two tangent lines to the surface

at the point x0 , y0 , f x0 , y0 are not collinear, and hence determine a unique plane that is
tangent to the surface z f x, y at the point x0 , y0 , f x0 , y0 .
Example 1: The plane x 1 intersects the paraboloid z x 2 y 2 in a parabola. Find the
slope of the tangent to the parabola at 1, 2,5 .
Solution: The curve of intersection of the paraboloid z x 2 y 2 and the plane x 1 is the
curve z 1 y 2 or y 2 z 1 , which is a parabola with vertex at 1,0,1 and opening upwards
in the direction of positive z axis. The tangent line to this parabola, that is, the curve of
intersection of the plane x 1 and the surface z x 2 y 2 at the point 1, 2,5 is shown in
the figure below.

The slope of the tangent is the value of the partial derivative z y at 1, 2 :

x2 y 2
2 y 1,2 2 2 4 .
y 1,2 y
1,2
As a check, we can use the equation of the parabola z 1 y 2 that lies in the plane x 1 and
ask for the slope at y 2 . The slope, now calculated as the ordinary derivative, is

dz
d
1 y 2
2 y 1,2 2 2 4 .
dy 1,2 dy
1,2
Example 2: Find the slopes in the x direction and in the y direction of the surface given
by f x, y

x2
25
y2
at the point
2
8

,1, 2 .
2

Solution: The tangent lines to the surface at the given point are shown below:

The partial derivatives of f with respect to x and y are

f x x, y x and f y x, y 2 y.
1
1
So, in the x direction, the slope is f x ,1 and in the y direction, the slope is
2
2
1
f y ,1 2 .
2

Example 3: Find the slopes of the surface given by f x, y 1 x 1 y 2 at the

2

point 1, 2,1 in the x direction and in the y direction.

Solution: The partial derivatives of f with respect to x and y are

f x x, y 2 x 1 and f y x, y 2 y 2 .
So, at the point 1, 2,1 , the slopes in the x direction is f x 1,2 2 1 1 0 and the
slope in the y direction is f y 1,2 2 2 2 0 . Note that the since values of the slopes
are zero, the tangent lines to the surface in the x direction and in the y direction are
parallel to x axis and y axis, respectively, as shown in the figure below.

Example 4: Let f x, y x 2 y 5 y 3 . Find the slope of the surface z f x, y at the point

1, 2

in the x direction and the y direction.

Solution: Differentiating f with respect to x and keeping y fixed yields

f x x, y 2 xy.
Thus, the slope in the x direction at 1, 2 is f x 1, 2 4 . That is, f is decreasing at
the rate of 4 units per unit increase in x or the slope of the surface z f x, y in the
increasing direction of x is decreasing. Similarly, differentiating f with respect to y and
keeping x fixed, we obtain

f y x, y x 2 15 y 2 .
Thus, the slope in the y direction at 1, 2 is f y 1, 2 61 . This implies that the slope of
the surface z f x, y is increasing in the increasing direction of x .
Example 5: Find the slope of the sphere x 2 y 2 z 2 1 in the y direction at the points
2 1 2
2 1 2
, , and , , .
3 3 3
3 3 3

Solution: We differentiate the equation x 2 y 2 z 2 1 implicitly with respect to y treating

z as a function of x and y taking x to be fixed. The computations give us

x y 2 z 2 1
y
y

0 2 y 2z

z
0
y

z
y
.
y
z

2 1 2
2 1 2
Substituting the y coordinate and z coordinate of the points , , and , , in
3 3 3
3 3 3
1
2 1 2
this expression, we find that the slope at the point , , is and the slope at
2
3 3 3

1
2 1 2
, , is . See figure below.
2
3 3 3

Note: In the preceding example, we can compute these slopes on the sphere in the y
direction by differentiating separately with respect to

the upper hemisphere

2 1 2
z 1 x 2 y 2 on which , , lies and lower hemisphere z 1 x 2 y 2 on which
3 3 3
2
1
2 1 2
, , lies and then evaluate the derivatives at x and y , respectively.
3
3
3 3 3

Tangent Plane and its normal line:

Suppose S is a surface with the equation z f x, y such that the partial derivatives

f x and f y are continuous, and let P0 x0 , y0 , z0 be any point on S . Let C1 be the curve of
intersection of S with the plane x x0 , and C2 be the curve of intersection of S with the
plane y y0 . Note that C1 and C2 intersect at P0 . Let T1 be the tangent to the curve C1 at

P0 and T2 be the tangent to the curve C2 at P0 .

Recall that the equation of C1 is f x0 , y and that of C2 is f x, y0 . Further, the slope of

the tangent line T1 to the curve C1 is

d
f y x0 , y0 ,
f x0 , y
dy
y y0
and the slope of the tangent line T2 to the curve C2 is

d
f x x0 , y0 .
f x, y0
dx
x x0
Since T1 and T2 are not collinear, these tangent lines determine a unique plane. This plane is
called the tangent plane to S at P0 . In fact, this tangent plane not only contains T1 and T2
but it contains every line that is tangent to any smooth curve on S that passes through P0 .

Equation for the tangent plane and its normal:

We know that the equation of a plane passing through a fixed point P0 x0 , y0 , z0 with
the vector N Ai Bj Ck as its normal is N QP0 0 , where Q x, y, z is any other point
on the plane, that is the equation of the plane is
A x x0 B y y0 C z z0 0,
or

C z z0 A x x0 B y y0 .

(1)

If C 0 , we divide both sides of (1) by C and write a A C and b B C to obtain the

equation for the tangent plane to S at P0 x0 , y0 , z0 as

z z0 a x x0 b y y0 .

(2)

The intersection of the plane (2) and the plane x x0 is the tangent line T1 , which we know
has slope f y x0 , y0 . Therefore, putting x x0 in (2), we find that the equation of T1 in the
point-slope form is

z z0 b y y0
so we must have b f y x0 , y0 . Similarly, by putting y y0 in (2), we obtain

z z0 a x x0
which represents the tangent line T2 with slope a f x x0 , y0 . Hence, substituting the values
of a and b in (2), we get

z z0 f x x0 , y0 x x0 f y x0 , y0 y y0

(3) ,

which is the required equation for the tangent plane to S at P0 x0 , y0 , z0 .

Equation (3) implies that the normal to the tangent plane has direction ratios
f x x0 , y0 , f y x0 , y0 ,1 . Thus, the equation of the normal line to the surface

z f x, y which passes through x0 , y0 , f x0 , y0 is

x x0 f x x0 , y0 t,

y y0 f y x0 , y0 t ,

z f x0 , y0 1 t ,

t .

y
Example 1: Find an equation for the tangent plane to the surface z tan 1 at the point
x

P0 1, 3, 3 .

Solution: We have
f x x, y

y
y
2 2
2
y x x y
1
x
1

f y x, y

and

x
1
,
2
2
y x x y
1
x
1

so that at the point P0 1, 3, 3 where x 1 and y 3 , we have

f x 1, 3

3
3

1 3
4

The equation of the tangent plane is, therefore, z

f y 1, 3

and

1
1
.
1 3 4

3
1

x 1 y 3 , or in the
3 4
4

standard form, it can rewritten as 3 3x 3 y 12 z 4 0 .

Physical Interpretation: Partial Derivative as the Rate of Change

Partial derivatives may be physically interpreted as the rates of change of function described
as follows:
1) As the point x, y moves from the fixed point x0 , y0 in the domain of the function

z f x, y , the value of the partial derivative f x at x0 , y0 , that is f x x0 , y0 gives

the rate of change of the function z f x, y in the direction of the positive x axis.
2) As the point x, y moves from the fixed point x0 , y0 in the domain of the function

z f x, y , the value of the partial derivative f y at x0 , y0 , that is f y x0 , y0 gives

the rate of change of the function z f x, y in the direction of the positive y axis.
Example 6: Let D x 2 y 2 be the length of the diagonal of a rectangle whose sides have
lengths x and y units where both x and y are allowed to vary. Find a formula for the rate
of change of D with respect to x if x varies with y held constant. Use this formula to find
the rate of change of D with respect to x at the point x 3 and y 4 .
Solution: Keeping y as a constant and differentiating implicitly both sides of the equation

D2 x 2 y 2 with respect to x , we get

2D

D
2x
x

Since D 5 at x 3 and y 4 , it follows that

D
x.
x

D
x

x 3, y 4

Thus, D is increasing at the rate of

D
x

x 3, y 4

3
.
5

3
unit per unit increase in x at point 3, 4 .
5

Example 7: In an electrical circuit with electromotive force (EMF) of E volts and resistance
R ohms, the current is I E R amperes. Find the partial derivatives I E and I R at the
instant when E 120 and R 15 , and interpret these derivatives as rates.
Solution: We have
I
E 1
I
E
E

and

2 .

E E R R
R R R
R

Thus,

when

E 120

and

R 15 ,

we

find

that

I E 1 15 0.0667

and

I R 120 15 0.5333 . This means that if the resistance is fixed at 15 ohms, the
current is increasing with respect to the voltage at the rate of 0.0667 ampere per unit volt
when the EMF is 120 volts. Similarly, with the same fixed EMF of 120 volts, the current is
decreasing with respect to the resistance at the rate of 0.5333 ampere per ohm when the
resistance is 15 ohms.
2

Partial derivative and continuity:

A function z f x, y can have partial derivatives with respect to both x and y at a
point without the function being continuous there.
Example 1: Consider the function

0 xy 0,
f x, y
1 xy 0.
Note that the graph of f consists of the lines L1 and L2 , and four open quadrants in the

xy plane (see figure below).

We claim that f is discontinuous at the origin but has first order partial derivatives at 0,0 .
Since the limit of f does not exist at the origin, f cannot be continuous. To see this, note
that along the x axis, f x, y 1 so that

lim

x , y 0,0

f x, y 1 . However, f x, y is

(along x axis)

constantly zero along the line y x (except at the origin), so that

lim

x , y 0,0

f x, y

yx

lim 0 0 .

x , y 0,0

x,0

Now, we discuss the partial derivative f x at

for any x . For this, we use the

geometrical concept of the partial derivative. To evaluate f x x,0 , hold y fixed at y 0 (the
xz plane). Then f x, y 1 for all x and the graph of f is the line L1 , which is parallel to

xy plane. The slope of this line at any x is 0 so that f x x,0 0 . In particular, f x 0,0 0
Similarly, to find f y 0, y for any y , note that it represents the slope of the graph of f , that
is the line L2 , which is also parallel to xy plane, so f y 0, y 0 for all y . In particular,

f y 0,0 0 . This proves our claim.

Example 2: Consider the function

xy
2
2
f x, y x y
0

x, y 0,0 ,
x, y 0,0 .

We claim that f x x, y and f y x, y exist at all points x, y but f is discontinuous at the

origin. Except at 0,0 , the partial derivatives of f are

f x x, y

fy

x
x, y

y 2 y xy 2 x

x2 y2

y 2 x xy 2 y

y2

x2 y y3

x2 y2

xy 2 x 3

y2

To find the partial derivatives of f at 0,0 , we use their definitions. Thus, we have
f x,0 f 0,0
00
lim
0,
x 0
x 0 x
x
f 0, y f 0,0
00
f y 0,0 lim
lim
0.
y 0

0
y
y
f x 0,0 lim

The continuity of f at 0,0 is not possible as the

lim

x , y 0,0

f x, y does not exist. This was

discussed in Example 1 in the section limit along a curve.

Higher Order Partial Derivatives:
The partial derivative of a function of two variables is again a function of two
variables. Similarly, the partial derivative of a function of three variables is again a function
of three variables, and so on. Therefore, it is possible to take the partial derivative of a partial
derivative. Also, we can take the partial derivative with respect to one variable and then take
a second partial derivative with respect to different variable, called a mixed order partial
derivative. Together, we call all such partial derivatives as higher order partial derivatives.
For example, the higher order partial derivatives for a function of two variables are as
follows:
Given z f x, y , the second order partial derivatives are
2 f
f
f x x f xx ,
2
x
x x
2 f
f
f y y f yy .
2
y
y y

The mixed second order partial derivatives are

2 f
f
f y x f yx ,
xy x y
2 f
f
f x y f xy .
yx y x

We can similarly write the higher order partial derivatives of a function of three or more
variables.
Note: The notation f xy means that we differentiate first with respect to x and then with
2 f
means just the opposite, that is we differentiate with respect to y
xy
first and the with respect to x .

respect to y , while

Theorem: If the function f x, y has mixed second order partial derivatives f xy and f yx
that are continuous in an open disk containing x0 , y0 , then

f xy x0 , y0 f yx x0 , y0 .

Example 1: For z f x, y 5x 2 2 xy 3 y 3 , determine the following higher order partial

derivatives:

2 z 2 f 2 z
,
,
, f xy 3, 2 .
xy yx x 2

Solution: (a) To compute

2 z
, first we differentiate z with respect to y , and then with
xy

respect to x :
z
2 x 9 y 2
y
2 z
z
2 x 9 y 2 2.
xy x y x

(b) For

2 f
, we differentiate f first with respect to x and then with respect to y :
yx

f
10 x 2 y
x
2 f
f
10 x 2 y 2.
yx y x y
(c) Differentiate with respect to x twice:
2 z z
10 x 2 y 10.
x 2 x x x

(d) We use part (b) to evaluate the mixed partial at the point 3, 2 :

f xy 3,2 2 .
Example 2: Determine f xy , f yx , f xx , f xxy , where f x, y x 2 ye y .
Solution: We have f x 2 xye y and f y x 2e y x 2 ye y . The mixed partial derivatives are
f xy f x y 2 xe y 2 xye y ,
f yx f y x 2 xe y 2 xye y .

Finally, we compute the second and higher order partial derivatives:

f xx f x x 2 ye y and

f xxy f xx y 2e y 2 ye y .

Example 3: By direct computations, show that

f xyz f yzx f zyx

for the function

f x, y, z xyz x 2 y 3 z 4 .
Solution: We have

f x x, y, z yz 2 xy 3z 4 ,

f y x, y, z xz 3x 2 y 2 z 4 ,

f z x, y, z xy 4 x 2 y 3z 3 .

f xy x, y, z z 6 xy 2 z 4 ,

f yz x, y, z x 12 x 2 y 2 z 3 ,

f zy x, y, z x 12 x 2 y 2 z 3 .

f yzx x, y, z 1 24 xy 2 z 3 ,

f zyx x, y, z 1 24 xy 2 z 3 .

f xyz x, y, z 1 24 xy 2 z 4 ,

Increment function and Differential function:

Let us recall the definition of increment of a function of one variable y f x . We
define the increment function y or f of the function f x at x a to be the change in

y that occurs when a is incremented by x (see Figure 1 below), that is

y f f a x f a .
Also, the differential function of y f x at x a , denoted by dy or df , is defined by the
equation

dy df y1 f a ,

Figure 1.
where a x, y1 is a point on the tangent line to the curve y f x at x a (see Figure
(1) above). Since the tangent line has the equation

y f a f a x a ,
we get y1 f a f a a x a f a f a x . Hence, the differential
dy y1 f a f a f a x f a
dy f a x.

or

Thus, the differential dy of y f x at x a is a linear function of the increment x .

Remark:
In general, the differential function of a function f x is the function of two independent
variables x and x defined as

df f x x .
Note that for f x x , the differential dx x , so that df f x dx for all x .
Theorem 1: If y f x is differentiable at x a , then

y f a x x

(1)

where 0 as x 0 .
Proof: Solving equation (1) for , we get

y f a x
. Thus, to prove the result, it is
x

sufficient to show that

y f a x
0 as x 0.
x

y f a x f a x f a f a x f a x f a

f a.
x
x
x
Since f is differentiable at x a , we have
We have

lim

x 0

f a x f a
f a.
x

f a x f a
Let x 0. Then lim lim
f a f a f a 0 .
x 0
x 0
x

Remarks: The identity in the preceding theorem shows that the differential dy f a x is
the principal (linear) part in the increment of a function. Also, for x small enough, one has

the approximate identity

y dy . We call

y f a f a x a

the linear

approximation to f x at x a . The factor is called the error in approximation which can

be made as small as desired relative to x by making x 0 .

Increments and differentials of a function of two variables:

Now, we discuss the increments and differentials of functions of two variables, which
are analogous to those for functions of one variable.
Definition: Let f x, y be a function of two independent variables x and y defined on
some open region R x, y | x0 x x1 , y0 y y1 in the xy plane. Let a, b R . We
define increment of f at

x, y a , b

to be the change in f that occurs when a is

incremented by x and b is incremented by y , that is

f f a x, b y f a, b .
Theorem 2: Suppose that f x, y and f x , f y are defined in an open rectangular region
R x, y | x0 x x1 , y0 y y1 containing the point a, b in the xy plane and f x , f y

are continuous at a, b R .Then for a x, b y R

f f x a, b x f y a, b y 1x 2y
where 1 and 2 are functions of x and y that both tend to zero as x, y 0,0 .
Proof: Let us write

f f a x, b y f a , b
f a x, b y f a, b y f a , b y f a , b
f a x, b y f a , b y f a , b y f a , b .
Using Lagranges Mean value theorem, since both f x and f y exist, there exist some

u a, a x and some v b, b y such that

f a x, b y f a, b y f x u, b y x,
f a, b y f a, b f y a, v y.

This gives us

f f x u, b y x f y a , v y
f x a, b f x u, b y f x a, b x f y a , b f y a , v f y a , b y
f x a, b x f x u, b y f x a , b x f y a , b y f y a , v f y a , b y
f x a, b x f y a, b y 1x 2 y ,

where 1 f x u, b y f x a, b and 2 f y a, v f y a, b . Since both f x and f y are

continuous, we have

lim

f x u, b y f x a, b 0,

lim

f y a, v f y a, b 0.

x ,y 0,0
x ,y 0,0

Let x, y 0,0 , so that 1 0 and 2 0 . This completes the proof.

Remarks:
a) Recall that the equation of tangent plane at a, b to the surface z f x, y is given
by

z f a, b f x a, b x a f y a , b y b .
The tangent plane stays close to the surface near the point of tangency a, b . This
implies that for x, y close to the point a, b , the z values on the tangent plane
should be close to the corresponding z values on the surface which are given by
z f x, y . So, we can approximate the value of a function of two variables

z f x, y near a given point of tangency using the tangent plane to the surface at
that point. The function defining z values on the tangent plane, namely,

L x, y f a , b f x a , b x a f y a , b y b
is therefore called the linear approximation of f x, y at a, b .
b) If we take x x a and y y b , then in the expression for f in the preceding
theorem, the first two terms correspond to the linear approximation of f x, y at

a, b ,

which is the differential of f . Note that for small enough dx x and

dy y , we have f df . We give a special name to functions that can be

approximated linearly in the above fashion: they are called differentiable functions.
Definition: Let f x, y be a function of two variables defined on some open region
R

. We say that f is differentiable at a, b R if the increment function of f can be

expressed as

f f x a, b x f y a, b y 1x 2y ,
where both 1 and 2 are functions of x and y , and 1 , 2 0 as x, y 0,0 . We
say that f is differentiable on R

whenever f is differentiable at every point in R .

Remarks: If f is differentiable at a, b , then we can write

f a x, b y f a, b f x a, b x f y a, b 1x 2y
so that f a x, b y f a, b f x a, b f y a, b , that is, the height of the surface
above the point a, b is approximately equal to the height of the tangent plane above a, b .
Theorem 3: (Sufficient condition for differentiability)
If f is a function of x and y defined on some open rectangle R containing the point a, b
and if f , f x and f y are continuous in a disk D centered at a, b , then f is differentiable
at a, b .
Proof: The proof follows by Theorem 2.
Theorem 4: (Differentiability implies Continuity)
If a function f x, y is differentiable at a, b , then it is continuous at a, b .
Proof: Since f x, y is differentiable at the point
increment function f

a, b , Theorem 2 shows that the

f a x, b y f a, b f x a, b x f y a, b y 1x 2y ,

where 1 , 2 0 as x, y 0,0 . This implies that

lim

x ,y 0,0

lim

f a x, b y f a, b

lim

f x a, b x

x ,y 0,0

x ,y 0,0

lim

x ,y 0,0

f y a, b y

lim

x ,y 0,0

1x

lim

x ,y 0,0

2 y

0 0 0 0 0.
Hence, we obtain that

lim

x ,y 0,0

f a x, b y f a, b , which shows that

is

continuous at a, b .
Remarks:
1) If a function is differentiable at a point, then the linear approximation (differential) at
that point provides a good approximation to the function near that point.
2) If a function has partial derivatives at a point, it need not be differentiable at that
point. We have seen in Example 1 and 2 in the section on partial derivative and
continuity that all the partial derivatives exist at every point but it is discontinuous at
the origin, so it cannot be differentiable at the origin, by Theorem 3. Here is another
example of a non-differentiable function for which f x and f y exist at the origin:

1 if x 0, y 0
f x, y
0 otherwise.
Example 1: (Using increments to estimate the change of a function)
An open box has length 3 ft , width 1 ft , and height 2 ft and is constructed from material that
costs $2 / ft 2 for the sides and $3 / ft 2 for the bottom. Compute the cost of constructing the
box, and then use the increments to estimate the change in cost if the length and width are
each increased by 3 in. and the height is decreased by 4 in.
Solution: An open box (with no top surface) having length x , width y , and height z has
surface area
S xy 2 xz 2 yz .

Since the sides cost $2 / ft 2 and the bottom $3 / ft 2 , the total cost is

C x, y, z 3xy 2 2 xz 2 yz .
The partial derivatives of C are Cx 3 y 4 z, C y 3x 4 z, Cz 4 x 4 y and the
dimensions of the box change by x 3 12, y 3 12, z 4 12 . Thus, the change in
total cost is approximated by

C Cx 3,1,2 x C y 3,1,2 y Cz 3,1,2 z 1.67.

Hence, the cost of constructing the box increases by approximately $1.67 .
Example 2: (Maximum percentage error using differentials)
The radius R and height H of a right circular cone are measured with errors of at most 3%
and 2% , respectively. Use increments to approximate the maximum possible percentage
error in computing the volume V of the cone using these measurements and the formula

V 1 3 R 2 H .
Solution: The radius R is measured with error at most 3% means that
Similarly, since the height H is measured with error at most 2% implies that

R
0.03 .
R

H
0.02 . It
H

is given that V 1 3 R 2 H , so that

2
VR RH
3

and

1
VH R 2 .
3

The change in V is approximated by

2

V RH R R 2 H .
3
3

Dividing this equation by the volume V 1 3 R 2 H , we obtain

V
R H
2

.
V
R H

Thus,

V V 2 R R H H 2 0.03 0.02 0.08.

Hence,

the

maximum

percentage error in computing the volume is approximately 8% .

Example 3: (Approximating the sag in a beam)
Suppose that the sag in a beam of length L, width w and height h is given by

S ( L, w, h) 0.0004

L4
,
wh 3

with all lengths measured in inches. A beam is supposed to measure L = 36, w = 2 and h = 6
with a corresponding sag of 1.5552 inches. Due to weathering and other factors, the
manufacturer
only
guarantees
measurements
with
error
tolerances
L 36 1, w 2 0.4, h 6 0.8 . Use a linear approximation to estimate the possible
range of sags in the beam.

Solution: The linear approximation of the sag at the point L, w, h is given by

L L, w, h S L, w, h SL L, w, h L 36 Sw L, w, h w 2 Sh L, w, h h 6 .
Now, SL 0.0016

L3
L4
L4
,
S

0.0004
,
S

0.0012
, so that at the point
w
h
wh 3
w2 h 3
wh 4

36,2,6 , we have
S 36,2,6 1.5552, SL 36,2,6 0.1728, Sw 36,2,6 0.0076, Sh 36,2,6 0.0076.
Thus, we have

L 36,2,6 1.5552 0.1728 L 36 0.7776 w 2 0.7776 h 6 .

From the stated tolerances, we have 1 L 36 1 , 0.4 w 2 0.4 , 0.8 h 6 0.8 .
Thus, the maximum sag occurs when L 36 1 , w 2 0.4 and h 6 0.8 . Therefore, the
linear approximation predicts that the maximum possible sag in the beam is

L 1.5552 0.1728 0.31104 0.62208 1.10592 .

Similarly, the minimum possible sag in the beam occurs when L 36 1 , w 2 0.4 and
h 6 0.8 , and the linear approximation predicts that it is given by

L 1.5552 0.1728 0.31104 0.62208 1.10592 .

Thus, based on linear approximations the sag is 1.5552 1.10592 or between 0.44928 and
2.66112 , which is the possible range of the sags in the beam. Note the uncertainty in the sag
of the beam is substantial.
Example 4: (Estimating the gauge of a sheet metal)
Manufacturing plants create rolls of metal of a desired gauge (thickness) by feeding the metal
through very large rollers. The thickness of the resulting metal depends on the gap between
the working rollers, the speed at which the rollers turn and the temperature of the metal.
Suppose that for a certain metal, a gauge of 4 mm is produced by a gap of 4 mm , a speed of
10m s and a temperature of 900 . Experiments show that an increase in speed of 0.2 m s

increases the gauge by 0.06mm and an increase in temperature of 10 decreases the gauge by
0.04 mm . Use a linear approximation to estimate the gauge at 10.1m s and 880 .

Solution: With no change in gap, we assume that the gauge is a function g s, t of the speed
s and temperature t . We are given that

g 0.06
g 0.04

0.3 and

0.004.
s 0.2
t
10

Hence the linear approximation of g s, t is given by

g s, t 4 0.3 s 10 0.004 t 900 .

With s 10.1m s and t 880 , we get the estimate

g 10.1,880 4 0.3 0.1 0.004 20 4.11.

Definition: Let x be incremented by dx x and y be incremented by dy y . We define
the differential or total differential of f x, y to be

df f x x, y dx f y x, y dy

(2) .

Similarly, for a function of three variables w f x, y, z , the total differential is

df
Remarks:

For

function

f
f
f
dx dy dz .
x
y
z

f x, y

of two

variables,

the

increment

function

f df 1x 2y so that f df . Similarly, for a function g x, y, z of three

variables, we can prove that

g g x dx g y dy g z dz 1x 2y 3z dg 1x 2y 3z ,
where i 0 i 1,2,3, as x, y, z 0,0,0 . Thus, again we obtain g dg .
Example 1: (Application of the total differential)
At a certain factory, the daily output is given by the function Q x, y 60

y units,

where x denotes the capital investment (in units of $1,000) and y is the size of the labor
force (in worker hours). The current capital investment is $900,000, and 1000 worker-hours
of labor are used each day. Estimate the change in output that will result if capital investment
is increased by $1,000 and labor is decreased by 2 worker-hours.
Solution: The change in output is estimated by the total differential dQ . We have
x 900, y 1,000, dx x 1, and dy y 2.

The total differential of Q x, y is

dQ

Q
Q
1 1
dx
dy 60
x
y
2 x

2 3
1
ydx 60 x y dy .
3

3 y
x
Thus, we obtain dQ 30
dy. Substituting for x, y, dx, dy we get
dx 20
23
x
y

3 1000
900
dQ 30
2 2.
1 20
23
1000
900

We conclude that the output decreases by approximately 2 units when the capital investment
is increased by $1,000 and labor is decreased by 2 worker hours.
Example 2: (Maximum percentage error in an electrical circuit)
When the two resistances R1 and R2 are connected in parallel, the total resistance R satisfies
1 1
1
.
R R1 R2

If R1 is measured as 300 ohms with a maximum error of 2% and R2 is measured as 500

ohms with a maximum error of 3%, what is the maximum percentage error in R ?
Solution: We are given that
value of

dR1
dR2
0.02 and
0.03. We have to find the maximum
R1
R2

R1R2
dR
. Since R
, we get
R
R1 R2

R
R2 2

R1 R1 R2 2
It follows that the total differential dR

dR

which implies that

R2 2

and

R
R12

.
R2 R1 R2 2

R
R
dR1
dR2 or by putting the above values,
R1
R2

R1 R2

dR1
2

R12

R1 R2

dR2 ,

dR
R2
dR
R1
dR

1
2 . Finally, we apply the triangle
R R1 R2 R1 R1 R2 R2

inequality to this equation to obtain

dR
R2
dR
R1
dR

1
2
R
R1 R2 R1
R1 R2 R2

500
300
0.02
0.03 0.02375.
300 500
300 500

Thus, the maximum percentage error in R is 2.4%.