Pde 1 & 2

T.Y.B.Sc.
Mathematics USMT 504

Partial Differential Equations
Dr. Abhaya Chitre

Head: Department of Mathematics
D.G.Ruparel College
September 13, 2022

Contents
1 Partial Differential Equations 1

1.1 Curves and Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Genesis of first Order P.D.E . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2.1 Classification of first order PDE : linear, semi-linear, quasi-linear and
non-linear equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3 Obtaining a partial differential equation . . . . . . . . . . . . . . . . . . . . . 10
1.3.1 Elimination of an arbitrary function φ from the equation φ(u, v) = 0
where u and v are functions of x, y and z. . . . . . . . . . . . . . . . 10
1.3.2 [4] Method to obtain a partial differential equation by the elimination of
arbitrary constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.4 Classification of Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.4.1 New Solutions from a Complete Integral . . . . . . . . . . . . 24
1.5 The Cauchy Problem (Initial value problem) . . . . . . . . . . . . . . . . . . . 35
1.6 Linear Equations of the First Order . . . . . . . . . . . . . . . . . . . . . . . . 35
1.6.1 Working rules for solving P p + Qq = R by Lagrange’s method. . . . . . 38
1.6.2 [4] Geometrical description of the solutions of P p + Qq = R and of the
dx dy dz
system of equations = = . . . . . . . . . . . . . . . . . . . 49
P Q R
1.7 Pfaffian Differential equations . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2 Unit II 59
2.1 Compatible system of first order Partial differential equations . . . . . . . . . . 59
2.2 [3] Charpits Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
2.2.1 [4] Working Rule while using Charpit’s Method . . . . . . . . . . . . . 78
2.2.2 Some standard types of partial differential equations . . . . . . . . . . 86
2.3 Jacobi’s Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
2.3.1 Jacobi’s method to find a complete integral for a first order p.d.e. in two
independent variables f (x, y, z, p, q) = 0. . . . . . . . . . . . . . . . . 120
2.4 Integral surfaces through a Given Curve: The Cauchy Problem . . . . . . . . . 125
3 Integral Transforms: Semester VI 127

Chapter 1
Partial Differential Equations
Course: Partial Differential Equations (Elective A)

Course Code: USMT5A4/UAMT5A4
Unit I: First Order Partial Differential Equations. (15L)
Curves and Surfaces, Genesis of first order PDE, Classification of first order PDE,
Classification of integrals, The Cauchy problem, Linear Equation of first order,
Lagrange’s equation, Pfaffian differential equations. (Ref Book: An Elementary
Course in Partial Differential Equations by T. Amaranath, 2nd edition, Chapter
1: 1.1, 1.2, 1.3, Lemma 1.3.1, 1.3.2, 1.3.3, 1.4, Theorem 1.4.1, 1.4.2, 1.5, Theorem
1.5.1, Lemma 1.5.1, Theorem 1.5.2, Lemma 1.5.2 and related examples)
Unit II: Compatible system of first order Partial Differential Equations.
(15L)
Definition, Necessary and sufficient condition for integrability, Charpit’s method,
Some standard types, Jacobi’s method, The Cauchy problem. (Ref Book: An El-
ementary Course in Partial Differential Equations by T. Amaranath, 2nd edition,
Chapter 1: 1.6, Theorem 1.6.1, 1.7, 1.8 Theorem 1.8.1, 1.9 and related examples)
Unit III: Quasi-Linear Partial Differential Equations. (15L)
Semi linear equations, Quasi-linear equations, first order quasi-linear PDE, Initial
value problem for quasi-linear equation, Non linear first order PDE, Monge cone,
Analytic expression for Monge’s cone, Characteristics strip, Initial strip. (Ref
Book: An Elementary Course in Partial Differential Equations by T. Amaranath,
2nd edition, Chapter 1: 1.10, Theorem 1.10.1, 1.11, Theorem 1.11.1, Preposition
1.11.1, 1.11.2 and related examples)
Reference Books
1. T. Amaranath; An Elementary Course in Partial Differential Equations; 2nd edition,

Narosa Publishing house.
2. Ian Sneddon; Elements of Partial Differential Equations; McGraw Hill book.
3. Ravi P. Agarwal and Donal O’Regan; Ordinary and Partial Differential Equations;
Springer, First Edition (2009).
4. W. E. Williams; Partial Differential Equations; Clarendon Press, Oxford, (1980).

2 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
5. K. Sankara Rao; Introduction to Partial Differential Equations; Third Edition, PHI.

Suggested Practicals on USMT5A4/UAMT5A4
1. Find general solution of Langrange’s equation.
2. Show that Pfaffian differential equation are exact and find corresponding integrals.
3. Find complete integral of first order PDE using Charpit’s Method.
4. Find complete integral using Jacobi’s Method.
5. Solve initial value problem for quasi-linear PDE.
6. Find the integral surface by the method of characteristics.
7. Miscellaneous Theory Questions.
1.1 Curves and Surfaces

(1) Plane curves of the form y = f (x) admit generalizations to two distinct classes of
plane curves. These are as follows.
(i) Parametrically Defined Curve is the plane curve Ç given by a function

f : D −→ R2 , f (t) = (x(t), y(t)) where D ⊆ R.
Usually, we express this by simply saying that C is the (parametrically defined)
curve given by (x(t), y(t)), t ∈ D. For example, the rectangular hyperbola is the
curve (t, 1t ), t ∈ R \ {0}.
(ii) Implicitly Defined Curve is the plane curve C given by an equation
F (x, y) = 0, where F is a real valued function defined on E ⊆ R2 and (x, y)
varies over the points of E.
Usually, we express this by simply saying that C is the (implicitly defined) curve
F (x, y) = 0, (x, y) ∈ E.
The reference to the domain E of F is skipped if E = R2 .
For example, the circle centered at the origin with unit radius is the curve x2 +
y 2 − 1 = 0.

(2) A parametrically defined curve C in R3 is given by x(t), y(t), z(t) , t ∈ D where
D ⊆ R.
Remarks 1.1.1.
• D is called as the parameter domain of the curve C .
• The variable t is the parameter for the curve. The functions x : D −→ R and
y : D −→ R are parametric equations for the curve.

• The pair x(t), y(t) for the curve in R2 or the triplet x(t), y(t), z(t) of real valued
functions is called as the parameterization of C .
1.1 Curves and Surfaces 3

• A parametrically defined curve C in R2 , given by x(t), y(t) , t ∈ [α, β], is said to be
smooth if the functions x and y are differentiable and their derivatives are continuous
on [α, β]. In this case, the arc length of C is defined to be
Z β p
l(C ) = x0 (t)2 + y 0 (t)2 dt. (1.1.1)
α

• A parametrically defined curve C in R3 , given by x(t), y(t), z(t) , t ∈ [α, β], is said
to be smooth if the functions x, y and z are differentiable and their derivatives are
continuous on [α, β]. In this case, the arc length of C is defined to be
Z β p
l(C ) = x0 (t)2 + y 0 (t)2 + z 0 (t)2 dt. (1.1.2)
α
• The curve C is determined by its parametrization that

nis the two functions x, y. o
The curve C should not be confused with its image x(t), y(t) ∈ R2 : t ∈ [α, β] .
For example, Consider the curves C1 in R2 given by (cos t, sin t), t ∈ [−π, π], and the
curve C2 given by (cos 2t, sin 2t), t ∈ [−π, π]. They both have the same domain [−π, π]
and the same image {(x, y) ∈ R2 : x2 + y 2 = 1}, but they are different curves, since
C1 winds around the origin (0, 0) once, while C2 winds around the origin (0, 0) twice.
• A standard parameter is the length of the curve of a point on the curve measured
from some fixed point on the curve. In such cases, at times, the symbol s is used
instead of t.
A parametrically defined curve C in R3 given by x(t), y(t) , t ∈ [α, β], gives the
position of a point in terms of the parameter t, which is often time, but need not be.
Suppose s is the distance of a point on the curve from some fixed point where distance
is measured along the curve. If we use s for the variable, the parametric equations
give the position in space in terms of distance along the curve. We might still imagine
that the curve represents the position of a moving object; now we get the position of
the object as a function of how far the object has traveled along the curve.
In general, if we have a parametric equations of a curve in terms of t and we want to
convert those to a parametric equations in terms of the arc length of a point from the
fixed point, we follow the following procedure.
Z tp
s(t) = x02 (t) + y 02 (t) dt
α
We get the value of s in terms of t.

We solve this to get t = g(s). !

Substitute this back into x(t), y(t) to get x g(s) , y g(s) .

• Suppose the parametric equations are in terms of s and let the equations be X(s), Y (s) .
0
What is X(s), Y (s) ?. It is the rate at which arc length is changing relative to

arc length. So, this must be 1.
• Consider a helix in R3 given by the parametric equations
x(t) = a cos t, y(t) = a sin t, z(t) = bt + c, t ∈ R,
where a, b, c ∈ R with a > 0 and b 6= 0. It lies on thecylinder givenby x2 + y 2 = a2 .

For α ∈ R let C denote a part of the helix given by x(t), y(t), z(t) , t ∈ [0, α, t ∈ R.
Then
Z tp
s= (−a sin t)2 + (a cos t)2 + b2 dt
0
Z tp
s= a2 + b2 dt
0
p
s = a2 + b2 t
s
=⇒ t = √
a2 + b2
Arc length parameterization is
s s sb

(x(s), y(s), z(s) = a cos √ , a sin √ ,√ +c
a2 + b2 a2 + b2 a2 + b2
0 a s a s b

(x(s), y(s), z(s) = − √ sin √ ,√ cos √ ,√
a2 + b2 a2 + b2 a2 + b2 a2 + b2 a2 + b2
s
0 a2 b2
k (x(s), y(s), z(s) k = +
a2 + b2 a2 + b2
=1
Definition 1.1.2. Parametrically Defined 3

Surface S in R is given by a function f :
D −→ R3 , f (u, v) = x(u, v), y(u, v), z(u, v) where x, y, z are real valued functions on D,
that is, x, y, z :−→ R.
Remarks 1.1.3.
(i) D is called as the parameter domain of the surface S .
(ii) The triplet (x, y, z) of real valued functions is called as the parameterization of S .
(iii) The surface S is determined by its parametrization (x(u, v), y(u, v), z(u, v)).
(iv) The surface S is not determined by the subset {(x(u, v), y(u, v), z(u, v)) : (u, v) ∈ D}.
(v) Consider the surface S1 in R3 given by the function f1 : [−π, π]×[0, 1] −→ R, f1 (u, v) =
(cos u, sin u, v) and the surface S2 in R3 given by the function f2 : [−π, π] × [0, 1] −→
R, f2 (u, v) = (cos 2u, sin 2u, v).
We can see that the functions f1 and f2 are distinct functions even though their ranges
are same. Hence these are different surfaces, since S1 goes around the z−axis once,
while S2 goes around the z axis twice.
Remark 1.1.4. [2] We will need the following Chain Rule.

(i) If z = f (x, y) and w = g(z), then w is a function of (x, y), and
∂w dw ∂z ∂w dw ∂z
= and = (1.1.3)
∂x dz ∂x ∂y dz ∂y
(ii) If z = f (x, y) and x = x(t), y = y(t), then z is a function of t, and
dz ∂z dx ∂z dy
= + (1.1.4)
dt ∂x dt ∂y dt
(iii) If z = f (u, v) and if u = u(x, y), v = v(x, y), then z is a function of (x, y), and
∂z ∂z ∂u ∂z ∂v ∂z ∂z ∂u ∂z ∂v
= + and = + (1.1.5)
∂x ∂u ∂x ∂v ∂x ∂y ∂u ∂y ∂v ∂y
∂w dw ∂z
z = f (x, y) and w = g(z) w is a function of x and y = ,
∂x dz ∂x
∂w dw ∂z
=
∂y dz ∂y
dz ∂z dx ∂z dy
z = f (x, y) and x = x(t), y = y(t) z is a function of t = +
dt ∂x dt ∂y dt
 
dx
 dt 
 
dz ∂z ∂z
= ∗ 
dt ∂x ∂y  dy 
dt
∂z ∂z ∂u ∂z ∂v
z is a function of (x, y) = +
∂x ∂u ∂x ∂v ∂x
∂z ∂z ∂u ∂z ∂v
= +
∂y ∂u ∂x ∂v ∂y
 ∂u ∂u 
∂z
 
 ∂x   ∂x ∂y 
∂z ∂z
 = ∗
   
z = f(u, )v and u = u(x, y), v= v(x, y) 
 ∂z  ∂u ∂v  ∂v ∂v 
∂y ∂x ∂y
 ∂u ∂u 
 ∂x ∂y 
 = Jacobian matrix of the functions u and v
 

 ∂v ∂v 
∂x ∂y
w.r.t. the variables x and y
Jacobian matrix of the transformation

φ : E −→ R2
defined by φ(x, y) = u(x, y), v(x, y) .
D ⊆ R2 ; f : D −→ R2 ; z, w : D −→ R.
f (u, v) = z(u, v), w(u, v) and u = u(x, y), v = v(x, y)
=⇒ z, w are functions of (x, y).
∂z ∂z ∂u ∂z ∂v
= +
∂x ∂u ∂x ∂v ∂x
∂z ∂z ∂u ∂z ∂v
= +
∂y ∂u ∂y ∂v ∂y
∂w ∂w ∂u ∂w ∂v
= +
∂x ∂u ∂x ∂v ∂x
∂w ∂w ∂u ∂w ∂v
= +
∂y ∂u ∂x ∂v ∂y
 ∂z ∂z     ∂u ∂u 
∂z ∂z
 ∂x ∂y   ∂u ∂v   ∂x ∂y 
= ∗
     
  
 ∂w ∂w   ∂w ∂w   ∂v ∂v 
∂x ∂y ∂u ∂v ∂x ∂y
 ∂u ∂u 
 ∂x ∂y 
Remark 1.1.5.   = Jacobian matrix of the functions u and v w.r.t. the variables x and y.
 
 ∂v ∂v 
 ∂u ∂u  ∂x ∂y
 ∂x ∂y 
det   is called Jacobian of u and v with respect to x and y and it is denoted by
 
 ∂v ∂v 
∂x ∂y
∂(u, v)
∂(x, y)
 ∂u ∂u 
∂(u, v)  ∂x ∂y 
= det 
 
∂(x, y)

 ∂v ∂v 
∂x ∂y
1.2 Genesis of first Order P.D.E

Definition 1.2.1. [4] An equation containing one or more partial derivatives of an unknown
function of two or more independent variables is known as a partial differential equation.
For example:
∂z ∂z 2
∂3z

(1) + = z + xy ∂z ∂z
∂x ∂y (4) + 3 = 2x
∂x ∂y ∂x

∂z ∂z ∂u ∂u ∂u
(2) z + =x (5) + + = xyz
∂x ∂y ∂x ∂y ∂z
" 2 #
1 ∂z 2

∂2z

∂z 3 ∂z ∂z
(3) = 1+ (6) y + =z
∂x2 ∂x ∂x ∂x ∂x
Definition 1.2.2. Order of a partial differential equation is defined as the order of the
highest order partial derivative occurring in the partial differential equation.
(1) order = 1 (3) order = 2 (5) order = 1
(2) order = 1 (4) order = 3 (6) order = 1
Definition 1.2.3. Degree of a partial differential equation is defined as the power

of the highest order partial derivative occurring in the partial differential equation after the
equation has been made free from radicals and fractions so far as derivatives are concerned.
(1) degree = 1 (3) degree = 3 (5) degree = 1
(2) degree = 1 (4) degree = 1 (6) degree = 2
Remark 1.2.4. Suppose x and y are independent variables and z is the dependent variable.
We use the following notations in PDE.
∂z ∂z ∂2z ∂2z ∂z
(i) p = (ii) q = (iii) r = (iv) s = (v) t =
∂x ∂y ∂x2 ∂x∂y ∂y 2
1.2.1 Classification of first order PDE : linear, semi-linear, quasi-linear and non-linear
equations
Definition 1.2.5. A first order equation f (x, y, z, p, q) = 0 is known as linear if
(a) it is linear in p, q and z and
(b) the coefficients of p and q are functions of x and y only.
Equations of the form:
P (x, y)p + Q(x, y)q = R(x, y)z + S(x, y).
are first order linear PDE.

1.2 Genesis of first Order P.D.E 9
For example:
First order linear PDE

∂z ∂z
+ = z + xy
∂x ∂y
∂z ∂z
yx2 + xy 2 = x2 y 3 + xyz
∂x ∂y
∂z ∂z
yx2 + xy 2 = x + y + x2 y 2 z
∂x ∂y
Definition 1.2.6. First order Semi-linear equation: A first order partial differential
equation f (x, y, z, p, q) = 0 is known as semi-linear equation if
(a) it is linear in p, q and
(b) the coefficients of p and q are functions of x and y only.
P (x, y)p + Q(x, y)q = R(x, y, z).
are first order semi-linear PDE.
∂z ∂z
For example: yx2 + xy 2 = x + y + x2 y 2 z 2 .
∂x ∂y
Definition 1.2.7. First order Quasi-linear equation: A first order partial differential
equation f (x, y, z, p, q) = 0 is known as quasi-linear equation if it is linear in p, q.
P (x, y, z)p + Q(x, y, z)q = R(x, y, z).
are first order quazi-linear PDE.
∂z ∂z
For example: yx2 z + xy 2 z = x + y, (x2 − yz)p + (y 2 − zx)q = z 2 − xy.
∂x ∂y
Types of first order PDE

A first order partial differential equation f (x, y, z, p, q) = 0 is known as
Linear eqns Semi-linear eqns Quasi-linear eqns Non-linear eqns
If it is linear in p, q linear in p, q linear in p, q not linear in

and z either p or q
and of P (x, y)p + Q(x, y)q P (x, y)p + Q(x, y)q P (x, y, z)p + Q(x, y, z)q
the form = R(x, y)z + S(x, y) = R(x, y, z) = R(x, y, z)
For yx2 p + xy 2 q yx2 p + xy 2 q yx2 z p + xy 2 z q p2 + q 2 = 1
example = x + y + x2 y 2 z = x + y + x2 y 2 z 2 =x+y
1.3 Obtaining a partial differential equation

1.3.1 Elimination of an arbitrary function φ from the equation φ(u, v) = 0 where u and
v are functions of x, y and z.
Theorem 1.3.1. The elimination of arbitrary function φ from the equation φ(u, v) = 0, where
u and v are functions of x, y and z (z is assumed to be a function of x and y), gives the partial
differential equation
∂(u, v) ∂(u, v) ∂(u, v)

∗p+ ∗q = .
∂(y, z) ∂(z, x) ∂(x, y)
Proof. Given
φ(u, v) = 0 (1.3.1)
We treat z as a function of x and y. (x and y as independent variables and z as dependent

variable).
∂y ∂x
So = 0 and =0
∂x ∂y
∂z ∂z
Let = p and = q.
∂x ∂y
Since φ is a function of (u, v) and u, v are functions of x, y and z where z is a function of x
and y, differentiating 1.3.1 partially with respect to x, we get

∂φ ∂φ ∂u ∂u ∂y ∂u ∂z ∂φ ∂v ∂v ∂y ∂v ∂z
= + + + + + =0
∂x ∂u ∂x ∂y ∂x ∂z ∂x ∂v ∂x ∂y ∂x ∂z ∂x

0 = φu ux + 0 + uz p + φv vx + 0 + vz p = 0
φu (ux + uz p) = −φv (vx + vz p)

1.3 Obtaining a partial differential equation 11
φu vx + vz p
=− (1.3.2)
φv ux + uz p
Similarly, differentiating 1.3.1 with resepct to y, we get

∂φ ∂φ ∂u ∂x ∂u ∂y ∂u ∂z ∂φ ∂v ∂x ∂v ∂y ∂v ∂z
= + + + + + =0
∂y ∂u ∂x ∂y ∂y ∂y ∂z ∂y ∂v ∂x ∂y ∂y ∂y ∂z ∂y

0 = φu 0 + uy ∗ 1 + uz q + φv 0 + vy ∗ 1 + vz q = 0
φu (uy + uz q) = −φv (vy + vz q)
φu vy + vz q
=− (1.3.3)
φv uy + uz q
Taking ratio of (2) and (3),

vx + vz p vy + vz q
− =−
ux + uz p uy + uz q
This implies
(vx + vz p)(uy + uz q) = (vy + vz q)(ux + uz p)
(uy vz − vy uz )p + (uz vx − ux vz )q + uz vz pq = ux vy − uy vx + uz vz qp
(uy vz − vy uz )p + (uz vx − ux vz )q = ux vy − uy vx

u uz u ux u uy
det y p + det z q = det x
vy vz vz vx vx vy
∂(u, v) ∂(u, v) ∂(u, v)

∗p+ ∗q = (1.3.4)
∂(y, z) ∂(z, x) ∂(x, y)
Examples 1.3.2. Eliminate the arbitrary function φ from each of the following equations and
find the corresponding p.d.e.
(i) [4] Obtain a partial differential equation by eliminating the arbitrary function φ from
φ(x + y + z, x2 + y 2 − z 2 ) = 0
Solution. We assume that z is a function of x and y.

∂z ∂z
Let = p and = q.
∂x ∂y
Let u(x, y, z) = x + y + z, v(x, y, z) = x2 + y 2 − z 2 .

From example 1.3.4, the differential equation is given by
∂(u, v) ∂(u, v) ∂(u, v)

∗p+ ∗q =
∂(y, z) ∂(z, x) ∂(x, y)
∂(u, v) ∂(u, v) ∂(u, v)

∂(y, z) ∂(z, x) ∂(x, y)

u uz u ux u uy
det y det z det x
vy vz vz vx vx vy

1 1 1 1 1 1
det det det
2y −2z −2z 2x 2x 2y
−2(y + z) 2(x + z) 2(y − x)
The required partial differential equation is
−(y + z) p + (x + z) q = y − x
(y + z) p − (x + z) q = x − y
(ii) φ(z − xy, x2 + y 2 + z 2 ) = 0

∂z ∂z
Let = p and = q.
∂x ∂y
Let u(x, y, z) = z − xy, v(x, y, z) = x2 + y 2 + z 2 .
From theorem 1.3.1, the differential equation is given by
∂(u, v) ∂(u, v) ∂(u, v)

∗p+ ∗q =
∂(y, z) ∂(z, x) ∂(x, y)
∂(u, v) ∂(u, v) ∂(u, v)

∂(y, z) ∂(z, x) ∂(x, y)

u uz u ux u uy
det y det z det x
vy vz vz vx vx vy

−x 1 1 −y −y −x
det det det
2y 2z 2z 2x 2x 2y
−2(xz + y) 2(x + zy) 2(−y 2 + x2 )

−(xz + y) p + (x + zy) q = (−y 2 + x2 )
(xz + y) p − (x + zy) q = (y 2 − x2 )

(iii) [3] φ(z − xy, x2 + y 2 ) = 0

∂z ∂z
Let = p and = q.
∂x ∂y
Let u(x, y, z) = z − xy, v(x, y, z) = x2 + y 2 .
∂(u, v) ∂(u, v) ∂(u, v)
∗p+ ∗q =
∂(y, z) ∂(z, x) ∂(x, y)
∂(u, v) ∂(u, v) ∂(u, v)

∂(y, z) ∂(z, x) ∂(x, y)

u uz u ux u uy
det y det z det x
vy vz vz vx vx vy

−x 1 1 −y −y −x
det det det
2y 0 0 2x 2x 2y
−2y 2x (−y 2 + x2 )
−y p + x q = (−y 2 + x2 )
y p − x q = (y 2 − x2 )
(iv) [3] φ(xy, x + y − z) = 0

∂z ∂z
Let = p and = q.
∂x ∂y
Let u(x, y, z) = xy, v(x, y, z) = x + y − z.
∂(u, v) ∂(u, v) ∂(u, v)
∗p+ ∗q =
∂(y, z) ∂(z, x) ∂(x, y)
∂(u, v) ∂(u, v) ∂(u, v)

∂(y, z) ∂(z, x) ∂(x, y)

uy uz uz ux ux uy
det det det
vy vz vz vx vx vy

x 0 0 y y x
det det det
1 −1 −1 1 1 1
−x y y−x
−x p + y q = y − x
x p−y q =x−y

√
(v) [3] φ(x + y, x − z) = 0

∂z ∂z
Let = p and = q.
∂x ∂y √
Let u(x, y, z) = x + y, v(x, y, z) = x − z.
∂(u, v) ∂(u, v) ∂(u, v)
∗p+ ∗q =
∂(y, z) ∂(z, x) ∂(x, y)
∂(u, v) ∂(u, v) ∂(u, v)

∂(y, z) ∂(z, x) ∂(x, y)

uy uz uz ux ux uy
det det det
vy vz vz vx vx vy
   
1 0 0 1
1 1
det  1  det  1  det
0 − √ − √ 1 1 0
2 z 2 z
1 1
− √ √ −1
2 z 2 z

1 1
− √ p + √ q = −1
2 z 2 z
1 1
√ p− √ q =1
2 z 2 z
√
p−q =2 z

xy x − y
(vi) [3] φ , =0
z z

∂z ∂z
Let = p and = q.
∂x ∂y
xy x−y
Let u(x, y, z) = , v(x, y, z) = .
z z
∂(u, v) ∂(u, v) ∂(u, v)
∗p+ ∗q =
∂(y, z) ∂(z, x) ∂(x, y)
xy x−y
u= v= −
z z
∂(u, v) ∂(u, v) ∂(u, v)

∂(y, z) ∂(z, x) ∂(x, y)

uy uz uz ux ux uy
det det det
vy vz vz vx vx vy
 x xy   xy y y x 
− −
 z z2   z2 z z z 
det   det   det  
1 y−x y−x 1 1 1
     
− −
z z2 z2 z z z
xy − x2 − xy −xy − y 2 + xy −y − x
z3 z3 z2
−x2 −y 2 −(x + y)
z3 z3 z2
x2 y2 −(x + y)
− 3
p − 3
q=
z z z2
x2 p + y 2 q = z(x + y)
(vii) [3] φ (xyz, x + y + z) = 0

∂z ∂z
Let = p and = q.
∂x ∂y
Let u(x, y, z) = xyz, v(x, y, z) = x + y + z.
∂(u, v) ∂(u, v) ∂(u, v)

∗p+ ∗q =
∂(y, z) ∂(z, x) ∂(x, y)
u = xyz v =x+y+z −
∂(u, v) ∂(u, v) ∂(u, v)

∂(y, z) ∂(z, x) ∂(x, y)

uy uz uz ux ux uy
det det det
vy vz vz vx vx vy

xz xy xy yz yz xz
det det det
1 1 1 1 1 1
x(z − y) y(x − z) z(y − x)
x(z − y) p + y(x − z) q = z(y − x)
x(y − z) p + y(z − x) q = z(x − y)
Example
p 1.3.3. Show that the partial differential equation corresponding to z = f (r) where
r = x2 + y 2 is y p − x q = 0.
Solution. z = F (r) where r = r(x, y)

So z is a function of x and y.
∂z ∂z ∂r ∂z ∂z ∂r
= ∗ and = ∗
∂x ∂r ∂x ∂y ∂r ∂y
∂r ∂r
This implies p = F 0 (r) ∗ and q = F 0 (r) ∗ .
∂x ∂y
∂r
p
Hence = ∂x ∂r
.
q ∂y
p ∂r 1 x ∂r 1 y
r = x2 + y 2 =⇒ = (2x) = and = (2y) = .
∂x 2r r ∂y 2r r
∂r
p x
So = ∂x ∂r
=
q ∂y
y
The required differential equation is y p − x q = 0.

∂(z, v)
Theorem 1.3.4. Let v = v(x, y) and z = f (v) then = 0 is a first order p.d.e. for z.
∂(x, y)
Solution.
∂z ∂z
= f 0 (v)vx , = f 0 (v)vy .
∂x ∂y
∂z ∂z
vy = f 0 (v)vx vy , vx = f 0 (v)vy vx .
∂x ∂y
∂z ∂z
=⇒ vy − vx = 0
∂x ∂y
=⇒ zx vy − zy vx = 0

z zy
=⇒ det x =0
vx vy
∂(z, v)
=0
∂(x, y)
That is vy p − vx q = 0 is the p.d.e.
We can also remember it in the form
p q
=0
vx vy

Examples 1.3.5. [4] Eliminate arbitrary function f from each of the following and obtain the
corresponding p.d.e.
(1) z = f (x2 − y 2 ) (4) lx + my + nz = f (x2 + y 2 + z 2 )

(2) z = f (x2 + y 2 )
y
(3) z = f (5) z = eax+by f (ax − by)
x
∂v ∂v
Solution. (1) z = f (x2 − y 2 ). Let v = x2 − y 2 . So = 2x, = −2y.
∂x ∂y
The partial differential is given by
∂(z, v)
=0
∂(x, y)
That is, vy p − vx q = 0
Substituting we get, (−2y) p − (2x) q = 0. The p.d.e. is y p + x q = 0
(2) z = f (x2 + y 2 ).
∂v ∂v
Let v = x2 + y 2 . So = 2x, = 2y.
∂x ∂y
∂(z, v)
=0
∂(x, y)
Substituting we get, (2y) p − (2x) q = 0. The p.d.e. is y p − x q = 0
y
(3) z = f .
x
y ∂v y ∂v 1
Let v = . So = − 2, = .
x ∂x x ∂y x
∂(z, v)
=0
∂(x, y)
1 y
Substituting we get, p − − 2 q = 0. The p.d.e. is x p + y q = 0.
x x
(4) lx + my + nz = f (x2 + y 2 + z 2 ). Let v = x2 + y 2 + z 2 .
This implies lx + my + nz = f (v)
Differentiating partially with respect to x and y, we get

∂z 0 ∂z
l+n = f (v) 2x + 2z
∂x ∂x

∂z 0 ∂z
m+n = f (v) 2y + 2z
∂y ∂y
∂z ∂z
l+n 2x + 2z
∂x = ∂x
∂z ∂z
m+n 2y + 2z
∂y ∂y
l + np x + zp
=
m + nq y + zq
(5) z = eax+by f (ax − by)

∂z
= aeax+by f (ax − by) + eax+by af 0 (ax − by)
∂x
∂z h i
= aeax+by f (ax − by) + f 0 (ax − by)
∂y
∂z h i
b = abeax+by f (ax − by) + f 0 (ax − by) · · · (1)
∂x
∂z
= beax+by f (ax − by) − eax+by bf 0 (ax − by)
∂y
∂z h i
= beax+by f (ax − by) − f 0 (ax − by)
∂y
∂z h i
a = abeax+by f (ax − by) − f 0 (ax − by) · · · (2)
∂y
Adding the two equations, we get
∂z ∂z
b +a = 2abeax+by f (ax − by)
∂x ∂y
The required p.d.e. is
∂z ∂z
b +a = 2abz
∂x ∂y
That is,
b p + a q = 2ab z
Theorem 1.3.6. Euler’s equation for a homogeneous function: Let z = f (x, y) be

a homogeneous function of x and y of degree n. Then, the function f (x, y) satisfies the first
order p.d.e.
x p + y q = nz
Proof. Since f (x, y) is a homogeneous function of x and y of degree n, we can write

y
f (x, y) = xn g
x
That is
y
z = xn g
x
∂z y y y ∂z 1 y
= −xn g ∗ 2 + nxn−1 g , = xn g 0 .
∂x x x x ∂y x x
∂z y y ∂z y

=⇒ = −xn−2 yg 0 + nxn−1 g , = xg n−1 g 0 .
∂x x x ∂y x
∂z ∂z y y y

x∗ +y∗ = −xn−1 yg 0 + nxn g + xn−1 yg 0 .
∂x ∂y x x x
∂z ∂z y
=⇒ x ∗ +y∗ = nxn g .
∂x ∂y x
∂z ∂z
=⇒ x ∗ +y∗ = nz
∂x ∂y
=⇒ x p + y q = nz.
The required p.d.e. is
x p + y q = nz
Equation p.d.e.
∂(u, v) ∂(u, v) ∂(u, v)

φ(u, v) = 0 ∗p+ ∗q = .
∂(y, z) ∂(z, x) ∂(x, y)
∂(z, v)
z = f (v), v = v(x, y) = 0, that is (vy p − vx q = 0)
∂(x, y)
z = f (x2 − y 2 ) y p+x p=0
z = f (x2 + y 2 ) y p−x q =0
l + np x + zp
lx + my + nz = f (x2 + y 2 + z 2 ) =
m + nq y + zq
z = eax+by f (ax − by) b p + a q = 2ab z
y
z = xn f x p + y q = nz
x
Euler’s homogeneous fn
of order n
1.3.2 [4] Method to obtain a partial differential equation by the elimination of arbitrary
constants
Consider the equation
F (x, y, z, a, b) = 0 (1.3.1)
where a and b denote arbitrary constats.

Let z be regarded as a function of two independent variables x and y.
Differentiating 1.3.1 with respect to x and y partially in turn, we get
∂F ∂F ∂z ∂F ∂F ∂z
+ =0 and + =0 (1.3.2)
∂x ∂z ∂x ∂y ∂z ∂y
Eliminating two constatns a and b from above two equations, we will get p.d.e.
f (x, y, z, p, q) = 0 (1.3.3)
Three situations may arise:
Case 1 The number of arbitrary constants is less than the number of independent variables,
then the elimination of arbitrary constants usually gives rise to more than one partial
differential equations of order one. For example, consider
z = ax + y (1)
where a is the only arbitrary constant and x, y are two independent variable.s Differ-
entiating (1) partially with respect to x and y, we get
∂z ∂z
=a and =1 (2)
∂x ∂y
Eliminating a between (1) and (2), we get
∂z
z= x+y
∂x
This is one p.d.e.
∂z ∂z
Since the equation = 1 does not contain arbitrary constant, so = 1 is also a
∂y ∂y
∂z
required partial differential equation. We get two partial equations z = x + y and
∂x
∂z
= 1.
∂y
Case 2 When the number of arbitrary constants is equal to the number of independent
variables, the the elimination of arbitrary constants give rise to a unique partial dif-
ferential equation of order one. For example, consider
az + b = a2 x + y (1)
where a is the only arbitrary constant and x, y are two independent variables. Differ-
entiating (1) partially with respect to x and y, we get
∂z ∂z
a = a2 and a =1 (2)
∂x ∂y
Taking product, we get

∂z ∂z
a2 = a2
∂x ∂y
∂z ∂z
=1
∂x ∂y
This is the unique partial differential equation.
Note that it is not a linear partial differential equation.
Case 3 When the number of arbitrary constants is greater than the number of independent
variables.
In this case, the elimination of arbitrary constants leads to a partial differential equa-
tion of order usually greater than one.
For example
z = ax + by + cxy (1)
1.4 Classification of Integrals 23
Differentiating partially with respect to x and y, we get,
∂z ∂z
= a + cy and = b + cx (2)
∂x ∂y
So,
∂2z ∂2z ∂2z

= 0, = 0, =c
∂x2 ∂y 2 ∂x∂y
Note that it is not a linear partial differential equation.
Examples 1.3.7. Form partial differential equations by eliminating arbitrary constants in each
of the following equation.
(i) z = (x − a)2 + (y − b)2 (iv) z = (x + a)(y + b)

(ii) z = a(x+y)+b
1
(iii) z = ax + a2 y 2 + b (v) z = axex + ∗ a2 e2y + b
2
Solution. (i) z = (x − a)2 + (y − b)2 (iii) z = ax + a2 y 2 + b
∂z ∂z ∂z
= 2(x − a), = 2(y − b) =a
∂x ∂y ∂x
2
1 ∂z 2
∂z
z=
1 ∂z
+ = 2a2 y
2 ∂x 2 ∂y ∂y
2
2 2 ∂z ∂z
∂z ∂z = 2y
4z = + ∂y ∂x
∂x ∂y
(ii) z = a(x+y)+b (iv) z = (x + a)(y + b)

∂z ∂z
= a and = (y + b)
∂x ∂x
∂z ∂z
=a = (x + a)
∂y ∂y
∂z ∂z ∂z ∂z
= z= ∗
∂x ∂y ∂x ∂y
Examples 1.3.8. Form partial differential equations by eliminating arbitrary constants a and
b in each of the following equation.
1 (v) z = ax + by. Show that the pde is

(i) z = axex + ∗ a2 e2y + b
2 z = x p + y q.
(ii) (x − a)2 + (y − b)2 + z 2 = λ2 , λ ∈ R
(vi) z 2 (1 + a3 ) = 8(x + ay + b)3 . Show that
(iii) z = x + ax2 y 2 + b the pde is p3 + q 3 = 27z.
(iv) 2z = (ax + y)2 + b Show that the pde (vii) z = (x + a)(y + b).Show that the pde is
is x p + y q = q 2 . pq = z.
1.4 Classification of Integrals

A first order partial differential equation in two unknowns x and y in its most general form
is given by
f (x, y, z, p, q) = 0 (1.4.1)
∂z ∂z
where p = and q = .
∂x ∂y
The solution z = F (x, y; a, b) of 1.4.1 represents a surface in (x, y, z) space. This surface is
called an integral surface of the partial differential equation 1.4.1.
Definition 1.4.1. A two parameter family of solutions z = F (x, y; a, b) of the equation
f (x, y, z, p, q) = 0.
is called a complete integral of the equation f (x, y, z, p, q) = 0 if the rank of the matrix

Fa Fxa Fya
M= is two.
Fb Fxb Fyb
1.4.1 New Solutions from a Complete Integral

[5] Let us try to generate some new solutions from a complete integral z = F (x, y; a, b) of
f (x, y, z, p, q) = 0 (p = zx = Fx , q = zy = Fy ).
Remark 1.4.2. Consider the surface which is obtained from the surface z = F (x, y; a, b) by
replacing the constants a and b by some functions of x and y.
So, this new surface is z = G(x, y) = F (x, y; a(x, y), b(x, y)).
Gx = Fx + Fa ∗ ax + Fb ∗ bx and Gy = Fy + Fa ∗ ay + Fb ∗ by
Since we denote Fx by p and Fy by q, we have,
Gx = p + Fa ∗ ax + Fb ∗ bx and Gy = q + Fa ∗ ay + Fb ∗ by
So z = F (x, y, a(x, y), b(x, y)) is a solution if and only if Gx = p and Gy = q.

Hence z = F (x, y, a(x, y), b(x, y)) is a solution f (x, y, z, p, q) = 0 if and only if
Fa ∗ ax + Fb ∗ bx = 0 and
Fa ∗ ay + Fb ∗ by = 0.
Definition 1.4.3. [9] Let Sa be a family of one parameter surfaces z = F (x, y; a) where a is
the parameter. Consider the following system of equations.
z = F (x, y; a),
0 = Fa (x, y; a).
The envelope E of the family of surfaces Sa , if exists, is defined as the set of all (x, y, z) ∈ R3
satisfying the above system of equations for some value of the parameter a.
For a fixed value of a, these two equations determine a curve Ca . The envelope E of the
family of surfaces Sa is the union of all these curves Ca .
The envelope E of the family of surfaces Sa , is obtained by eliminating a between

z = F (x, y; a), (1.4.2)
0 = Fa (x, y; a). (1.4.3)

The equation of E is found by solving Fa (x, y, a) = 0 for a as a function of x and y, that is
a = a(x, y) and then substituting into z = f (x, y, a).
Hence the envelope E will be of the form z = G(x, y) = F (x, y, a(x, y)).
Remark 1.4.4. Properties of an envelope.

(i) The envelope E of the family of surfaces Sa , touches every member of the family Sa along
the curve Ca .
Suppose (x, y, z) ∈ E. Then (x, y, z) satisfies 1.4.2 and 1.4.3 for some a. This implies
that (x, y, z) ∈ Ca for some a.
(ii) The normal to Sa at (x, y, z) has direction ratios (Fx , Fy , −1). The normal to E at
(x, y, z) has direction ratios (Fx + Fa ∗ ax , Fy + Fa ∗ ay , −1) = (Fx , Fy , −1) as Fa = 0.
Hence the envelope E and each member of the family of surfaces Sa have the same normal
direction ratios at every point on Ca . In other words, they have the same tangent plane
at every point on Ca .
Lemma 1.4.5. Let z = F (x, y; a) be a one-parameter family of solutions of
f (x, y, z, p, q) = 0 where p = zx = Fx , q = zy = Fy . Then the envelope of this family, if it
exists, is also a solution of f (x, y, z, p, q) = 0.
Proof. (Note: This is Lemma 1.3.1 from our syllabus.)
z = F (x, y; a) is a one-parameter family of solutions of f (x, y, z, p, q) = 0.
The envelope E of this family is obtained by eliminating a between
z = F (x, y; a), (1.4.4)
0 = Fa (x, y; a). (1.4.5)

The equation of E is found by solving Fa (x, y, a) = 0 for a as a function of x and y, a =
a(x, y) and then substituting into z = f (x, y; a).
Hence the envelope E will be of the form z = G(x, y) = F (x, y, a(x, y)).
We want to show that z = F (x, y, a(x, y)) is a solution of f (x, y, zp, q) = 0.
By remark 1.4.2, it is sufficient to show that p = Gx , q = Gy .
Now, G(x, y) = F (x, y, a(x, y)) =⇒ Gx = Fx + Fa ∗ ax = Fx + 0 ∗ ax = Fx = p and
Gy = Fy + Fa ∗ ay = Fx + 0 ∗ ay = Fy = q (as Fa = 0, Fb = 0).
Hence z = G(x, y) is also a solution of f (x, y, z, p, q) = 0
Definition 1.4.6. Let φ :R −→ R beany function. Let Sa,φ denote a one-parameter family
of surfaces given by z = F x, y, a, φ(a) . Consider the following system of equations.
z = F (x, y; a, φ(a)),
0 = Fa + Fb φ0 (a).
The envelope of Sa,φ , if exists, is defined as the set of all (x, y, z) ∈ R3 satisfying the above
system of equations for some value of the parameter a. The envelop of Sa is obtained by
eliminating a between the equations z = F (x, y; a, φ(a)), and Fa + Fb φ0 (a) = 0.
Definition 1.4.7. Let Sa,b be a two parameter family of complete solutions z = F (x, y, a, b)
of f (x, y, p, q) = 0 where a, b are the parameters. Let φ : R −→ R be any function. Let Sa,φ
the family of surfaces z = F (x, y; a, φ(a)). Then the envelope of Sa,φ is also a solution of
f (x, y, p, q) = 0. This solution is called a General integral of f (x, y, z, p, q) = 0.
When a particular function φ is used, we obtain a particular solution fo the partial differential
equation.
Different choices of φ may give different particular solutions of the partial differential equation.
Definition 1.4.8. Let Sa,b be a family of two parameter surfaces z = f (x, y; a, b) where a, b
are the parameters. Consider the following system of equations.
z = F (x, y; a, b),
0 = Fa (x, y; a, b),
0 = Fb (x, y; a, b).
The envelope E of Sa,b if exists, is defined as the set of all (x, y, z) ∈ R3 satisfying the above
system of equations for some values of the parameters a and b.
Lemma 1.4.9. Let Sa,b be a two parameter family of solutions z = F (x, y; a, b) of the partial
differential equation f (x, y, z, p, q) = 0 where a, b are the parameters. Then the envelope of
this family is also a solution of f (x, y, p, q) = 0.
(This lemma is lemma 1.3.2 as per our syllabus)
Proof. The envelope E of the family of surfaces Z = F (x, y; a, b) is obtained by eliminating

a and b from the equations
z = F (x, y; a, b),
Fa = 0,
Fb = 0.
We eliminate a and b from the equations Fa = 0 and Fb = 0.

We get a = a(x, y) and b = b(x, y).
Hence the envelope E will be of the form z = G(x, y) = F (x, y, a(x, y), b(x, y)).
We want to show that z = G(x, y) = F (x, y, a(x, y), b(x, y)) is a solution of f (x, y, z, p, q) =
0.
By remark 1.4.2, it is sufficient to show that p = Gx , q = Gy .
Now, G(x, y) = F (x, y, a(x, y), b(x, y).
Gx = Fx + Fa ∗ ax + Fb ∗ bx Gy = Fy + Fa ∗ ay + Fb ∗ by
= Fx + 0 ∗ ax + 0 ∗ bx = Fx + 0 ∗ ay + 0 ∗ by
= Fx (as Fa = 0, Fb = 0) = Fy (asFa = 0, Fb = 0)
=p =q
Hence z = G(x, y) is also a solution of f (x, y, z, p, q) = 0.
Definition 1.4.10. Let Sa,b be a two parameter family of complete integrals z = F (x, y; a, b)
of f (x, y, z, p, q) = 0 where a, b are the parameters. Then the envelope of Sa,b is also a solution
of f (x, y, z, p, q) = 0. This solution is called a singular integral of f (x, y, z, p, q) = 0.
Lemma 1.4.11. Let z = F (x, y; a, b) be a complete integral of f (x, y, z, p, q) = 0 and z =

F (x, y, a(x, y), b(x, y)) be the singular integral of f (x, y, z, p, q) = 0. Then the singular integral
satisfies the equations 
f (x, y, z, p, q) = 0, 
fp (x, y, z, p, q) = 0, (1.4.6)
fq (x, y, z, p, q) = 0.

Proof. This lemma is Lemma 1.3.3 in our syllabus.

Since z = F (x, y, a(x, y), b(x, y)) is the singular integral of f (x, y, z, p, q) = 0, we get
!

f x, y, F x, y, a(x, y), b(x, y) , Fx x, y, a(x, y), b(x, y) , Fy x, y, a(x, y), b(x, y) =0
That is,
!

f x, y, F x, y, a(x, y), b(x, y) , p x, y, a(x, y), b(x, y) , q x, y, a(x, y), b(x, y) =0
(1.4.7)
Differentiation partially w.r.t. a and b we get,
fz Fa + fp Fxa + fq Fya = 0 (p = Fx =⇒ pa = Fxa , q = Fy =⇒ qa = Fya )
and
fz Fb + fp Fxb + fq Fyb = 0 (p = Fx =⇒ pb = Fxb , q = Fy =⇒ qb = Fyb )
But z = F (x, y, a(x, y), b(x, y)) is the singular integral of f (x, y, z, p, q) = 0.
So Fa = 0 and Fb = 0.

fp Fxa + fq Fya = 0 ,
(1.4.8)
fp Fxb + fq Fyb = 0.
Now, z = F (x, y; a, b) be a complete integral of f (x, y, z, p, q) = 0.
Fa Fxa Fya
So rank of the matrix is 2.
Fb Fxb Fyb

0 Fxa Fya
As Fa = Fb = 0, rank of is 2.
0 Fxb Fyb

Fxa Fya
This means det 6= 0.
Fxb Fyb
Hence from 1.4.8 fp = 0 and fq = 0.
That is, fp (x, y, z, p, q) = 0 and fq (x, y, z, p, q) = 0.
Hence for z = F (x, y, a(x, y), b(x, y)) we have f (x, y, z, p, q) = 0, fp (x, y, z, p, q) = 0 and
fq (x, y, z, p, q) = 0.
Remark 1.4.12. Using the three equations


f (x, y, z, p, q) = 0, 
fp (x, y, z, p, q) = 0,
fq (x, y, z, p, q) = 0.

we can get the singular integral by eliminating p and q from the second and third equations and
then putting the values of p and q in the first one.
Remark 1.4.13. Let f (x, y, z, p, q) = 0 be a partial differential equation and z = F (x, y; a, b)

be its solution where a and b are the parameters.
Complete Integral General Integral Singular Integral

z = F (x, y; a, b) is called The envelope of the family The envelope of the family
a complete integral of surfaces z = F (x, y; a(x, y), φ(a(x, y)) of surfaces z = F (x, y; a, b)
if the rank of the obtained by eliminating a obtained by eliminating a
matrix between the equations and b between the eqns
Fa Fxa Fya
z = F (x, y; a(x, y), φ(a(x, y)) and z = F (x, y; a, b) and
Fb Fxb Fyb
is 2 0 = Fa + Fb φ0 (a) Fa (x, y; a, b) = 0 and
is called the general integral Fb (x, y; a, b) = 0
if φ is an arbitrary is called singular integral.
function.
If φ is a particular
function, then the envelope
is called the particular soln.
Special Integral Usually (but not always), the three classes namely, complete integral,
general integral and singular integral include all the integrals of the first order p.d.e.
f (x, y, z, p, q) = 0. However, there are some solutions of certain first order partial dif-
ferential equations which do not fall under any of the three classes. Such solutions are
called Special integrals.
√
Example 1.4.14. Consider the partial differential equation p − q = 2 z. Then
√ √
(i) show that F (x, y, x − z) = 0 is an integral of p − q = 2 z.
√
(ii) show that z = 0 is an integral of p − q = 2 z.
(iii) note that the integral z = 0 can not be obtained from the general integral.
√
(iv) this type of integrals are called as special integral of p − q = 2 z.
√
Solution. F (x + y, x − z) = 0
Differentiating partially w.r.t. x and y, we get
Example 1.4.15. Consider f (x, y, z, p, q) = z − px − qy − p2 − q 2 = 0.

(i) Show that the two parameter family of z = F (x, y; a, b) = ax + by + a2 + b2 is a complete
integral of the above partial differential equation.
√
(ii) Let b = 1 − a2 . Find the envelope of the family of surfaces z = F (x, y; a, b) (particular
integral).
(iii) Let b = a. Find the particular integral of z − px − qy − p2 − q 2 = 0.
(iv) Find the singular integral.
(v) Find the singular integral using the theorem 1.4.11.

Solution. (i)
z = F (x, y; a, b) = ax + by + a2 + b2
F (x, y; a, b) = ax + by + a2 + b2
p = Fx = a, q = Fy = b
Given partial differential equation is z − px − qy − p2 − q 2 = 0.
z − px − qy − p2 − q 2 = ax + by + a2 + b2 − ax − by − a2 − b2
=0
Hence z = ax + by + a2 + b2 is a solution of z − px − qy − p2 − q 2 = 0. (1)

To show that z = ax + by + a2 + b2 is a complete integral of z − px − qy − p2 − q 2 = 0.
F (x, y; a, b) = ax + by + a2 + b2

Fa Fxa Fya x + 2a 1 0
=
Fb Fxb Fyb y + 2b 0 1
Rank of the above matrix is 2. (2)

Hence from (1) and (2) z = ax + by + a2 + b2 is a complete integral of z − px − qy −
p2 − q 2 = 0.
√ √
(ii) b = 1 − a2 . Here φ(a) = 1 − a2
This means
p
F (x, y, a, φ(a)) = ax + ( 1 − a2 ) y + a2 + 1 − a2
p
z = ax + ( 1 − a2 ) y + 1. (∗)
The envelope E of the family z = F (x, y, a, φ(a)) is obtained by eliminating a between

z = F (x, y, a, φ(a)) and
Fa (x, y, a, b) + Fb (x, y, a, b)φ0 (a) = 0
and substituting this value of a in z = F (x, y, a, φ(a)) as shown below. φ(a) =

√ −a
1 − a2 . So φ0 (a) = √
1 − a2
−a
Fa (x, y, a, b) + Fb (x, y, a, b)φ0 (a) = (ax + by + a2 + b2 )a + (ax + by + a2 + b2 )b ∗ √
1 − a2
−a
= (x + 2a) + (y + 2b) ∗ √
1 − a2
p −a
= x + 2a + (y + 2 1 − a2 ) ∗ √
1 − a2
ay
= x + 2a − √ − 2a
1 − a2
ay
=x− √
1 − a2
ay
Fa (x, y, a, b) + Fb (x, y, a, b)φ0 (a) = 0 =⇒ x = √
1 − a2
a2 y 2
=⇒ x2 =
(1 − a2 )
=⇒ x − x2 a2 = a2 y 2
2
x
=⇒ a = p
x + y2
2
√
Substituting this in (∗), that is in z = ax + ( 1 − a2 ) y + 1 we get,
s
x2 y2
z=p + ∗ y+1
x2 + y 2 x2 + y 2
x2 y2
z=p +p +1
x2 + y 2 x2 + y 2
p
= x2 + y 2 + 1
p
Hence the envelope is given bypz = x2 + y 2 + 1.
The particular integral is z = x2 + y 2 + 1.
(iii) b = a. Here φ(a) = a. Hence φ0 (a) = 1.

This means
z = F (x, y, a, φ(a)) = ax + ay + 2a2 (∗∗)
Also,
Fa (x, y; a, b) + Fb (x, y; a, b) ∗ φ0 (a) = (ax + by + a2 + b2 )a + (ax + by + a2 + b2 )b ∗ 1

= (x + 2a) + (y + 2b)
0
Fa (x, y; a, b) + Fb (x, y; a, b) ∗ φ (a) = 0 =⇒ (x + 2a) + (y + 2b) = 0
=⇒ (x + 2a) + (y + 2a) = 0 (asb = φ(a) = a)
(x + y)
=⇒ a = − .
4
Substituting in (∗∗), that is inz = ax + ay + 2a2 , we get

−(x + y) −(x + y) (x + y)2
z= x+ y+
4 4 8
−(x + y)2
z=
8
−(x + y)2
Hence the envelope is given by z = .
8
−(x + y) 2
The particular integral is z = .
8
(iv) For singular integral, we take z = F (x, y; a, b) and the we eliminate a and b using the
equations Fa (x, y; a, b) = 0 and Fb (x, y; a, b) = 0.
z = F (x, y; a, b) = ax + by + a2 + b2
Fa = x + 2a
Fb = y + 2b
x
Fa = 0 =⇒ a = −
2
y
Fb = 0 =⇒ b = −
2
Substituting these values in z = F (x, y; a, b) = ax + by + a2 + b2 , we get
(x + y)2
z=−
4
That is,
4z = −(x2 + y 2 )
(v) Using lemma 1.4.11, we can find singular integral as follows.

We know that singular integral satisfies

f (x, y, z, p, q) = 0, 
fp (x, y, z, p, q) = 0,
fq (x, y, z, p, q) = 0.

z − px − qy − p2 − q 2 = 0, 

−x − 2p = 0,
−y − 2q = 0.

x y
This implies p = − , q = − .
2 2
Hence the singular solution is
z − px − qy − p2 − q 2 = 0
−x −y x2 y 2
=⇒ z − x −y − − =0
2 2 4 4
x2 y 2
=⇒ z + + =0
4 4
=⇒ 4z = −(x + y)2

Example 1.4.16. Consider the partial differential equation z 2 (1 + p2 + q 2 ) = 1.
(i) Show that (x − a)2 + (y − b)2 + z 2 = 1 is a complete integral of z 2 (1 + p2 + q 2 ) = 1.
(ii) Let b = 2a. Find the envelope of the family of surfaces z = F (x, y; a, 2a) (particular
integral).
(iii) Show that z = ±1 are the singular integrals.
Solution. The given partial differential equation is z 2 (1 + p2 + q 2 ) = 1. (1)
(i) We want to show that (x − a)2 + (y − b)2 + z 2 = 1 is a complete integral of (1).

First we will show that (x − a)2 + (y − b)2 + z 2 = 1 is a solution of (1).
(x − a)2 + (y − b)2 + z 2 = 1
Differentiation partially with respect to x and y we get
2(x − a) + 2zzx = 0 2(y − b) + 2zzy = 0

=⇒ zzx = −(x − a) =⇒ zzy = −(y − b)
=⇒ zp = −(x − a) =⇒ zq = −(y − b)
=⇒ z 2 p2 = (x − a)2 =⇒ z 2 q 2 = (y − b)2
z 2 (1 + p2 + q 2 ) = z 2 + z 2 p2 + z 2 q 2
= z 2 + (x − a)2 + (y − b)2
=1
Hence (x − a)2 + (y − b)2 + z 2 = 1 is a solution

of (1).
Fa Fxa Fya
Now we show that the matrix given by is of rank 2.
Fb Fxb Fyb
(x − a)2 + (y − b)2 + z 2 = 1 =⇒ z 2 = 1 − (x − a)2 − (y − b)2
p
z= 1 − (x − a)2 − (y − b)2
2(x − a)(−1) 2(y − b)(−1)

za = − p zb = − p
2 1 − (x − a)2 − (y − b)2 2 1 − (x − a)2 − (y − b)2
(x − a) (y − b)
=p =p
1 − (x − a)2 − (y − b)2 1 − (x − a)2 − (y − b)2
(x − a) (y − b)
= =
z z
x−a y−b
p = zx = − p q = zy = − p
1 − (x − a)2 + (y − b)2 1 − (x − a)2 + (y − b)2
x−a y−b
=− =−
z z
z(−1) − (x − a)za −(y − b)za
pa = zxa = − qa = zya = −
z2 z2
z + (x − a)za (y − b)(x − a)
= =
z2 z3
z 2 + (x − a)2 z(−1) − (y − b)zb
= qb = zyb = −
z3 z2
1 − (y − b)2 z + (y − b)zb
= =
z3 z2
z(0) − (x − a)zb z 2 + (y − b)2
pa = zxb = − =
z2 z3
(x − a)(y − b) 1 − (x − a)2
= =
z3 z3
x−a 1 − (y − b)2 (y − b)(x − a)

 
 
za zxa zya  z z3 z3 
 =



zb zxb zyb y−b (x − a)(y − b) 1 − (x − a)2
 
z z3 z3
Rank of the matrix is 2.

Hence (x − a)2 + (y − b)2 + z 2 = 1 is a complete integral of (1).
(ii) Let b = 2a So, φ(a) = 2a and φ0 (a) = 2. We want to find the envelope of the family
of surfaces z = F (x, y; a, 2a) (particular integral).
z = F (x, y; a, 2a)
That is
p
z= 1 − (x − a)2 − (y − 2a)2 (∗ ∗ ∗)
Fa (x, y; a, b) + Fb (x, y; a, b) ∗ φ0 (a) = ( 1 − (x − a)2 − (y − b)2 )a +

p
p
( 1 − (x − a)2 − (y − b)2 )b ∗ 2
1
= p ∗ 2(x − a)+
2 1 − (x − a)2 − (y − 2a)2
2(x − a) + 4(y − b)
= p
2 1 − (x − a)2 − (y − 2a)2
Fa (x, y; , a, b) + Fb (x, y; a, b) ∗ φ0 (a) = 0 =⇒ 2(x − a) + 4(y − b) = 0.

=⇒ (x − a) = −2(y − b)
=⇒ x − a = −2(y − 2a)
=⇒x − a = −2y + 4a
=⇒x + 2y = 5a
x + 2y
=⇒a =
5
Substituting this in (x − a)2 + (y − 2a)2 + z 2 = 1 we get,
x + 2y 2 2x + 4y 2

x− + y− + z2 = 1
5 5
4x − 2y 2 y − 2x 2

+ + z2 = 1
5 5
(4x − 2y)2 + (y − 2x)2 + 25z 2 = 25
16x2 − 16xy + 4y 2 + y 2 − 4xy + 4x2 + 25z 2 = 25
20x2 − 20xy + 5y 2 + 25z 2 = 25
4x2 − 4xy + y 2 + 5z 2 = 5
(2x − y)2 + 5z 2 = 5
Equation of the particular solution is (2x − y)2 + 5z 2 = 5.
(iii) By theorem 1.4.11, the singular integral satisfies the equations


f (x, y, z, p, q) = 0, 
fp (x, y, z, p, q) = 0,
fq (x, y, z, p, q) = 0.

Given partial differential equation is z 2 (1 + p2 + q 2 ) = 1

Hence f (x, y, z, p, q) = z 2 (1 + p2 + q 2 ) − 1.
fp (x, y, z, p, q) = z 2 2p
fq (x, y, z, p, q) = z 2 q
The three equations are z 2 (1 + p2 + q 2 ) − 1 = 0, z 2 p = 0, z 2 q = 0.

Since z 2 (1 + p2 + q 2 ) − 1 = 0 =⇒ z 2 (1 + p2 + q 2 ) = 1. Hence z 2 6= 0.
Hence p = 0 and q = 0.
Substituting this in f (x, y, z, p, q) = 0, that is, z 2 (1 + p2 + q 2 ) − 1 = 0, we get, z 2 = 1.
Hence z = ±1.
Thus singular integrals are z = 1 and z = −1.

y
Example 1.4.17. Show that z = ax + + b is a complete integral of pq = 1. Also show that
a
this partial differential equation does not have a singular integral. Find the particular solution
corresponding to the sub-family b = a
Example 1.4.18. Show that 2z = (ax + by)2 + b is a complete integral of px + qy − q 2 = 0
1.5 The Cauchy Problem (Initial value problem) 35
1.5 The Cauchy Problem (Initial value problem)

Given a first order partial differential equation and a curve in sspace, the Cauchy problem
is to find an integral surface (i.e. a solution) of the given partial differential equation which
contains the given curve.
In other words, given a partial differential equation (not necessarily non-linear)
f (x, y, z, p, q) = 0
and a curve x = x(s), y = y(s), z = z(s), s ∈ [a, b], the Cauchy problem is to find a solution
z = z(x, y) of the p.d.e. such that z(s) = z(x(s), y(s)) for all s ∈ [a, b].
(Note: We will be studying this in unit III in detail.)
1.6 Linear Equations of the First Order

This section describes a method of finding a general integral for a quasi-linear equation.
A quasi-linear partial differential equation of order one is of the form P (x, y, z)p+Q(x, y, z)q =
R(x, y, z). Such a partial differential equation is known as Lagrange equation.
Theorem 1.6.1. [4] The general solution of the Lagrange equation
P (x, y, z)p + Q(x, y, z)q = R(x, y, z), (1.6.1)
where P, Q and R are continuously differentiable functions on the domain D ⊆ R3 is
φ(u, v) = 0 (1.6.2)
where φ is an arbitrary function and
u(x, y, z) = c1 and v(x, y, z) = c2 (1.6.3)
are two independent solutions of

dx dy dz
= = . (1.6.4)
P Q R
Here, c1 and c2 are arbitrary constants and at least one of u, v must contain z.
Proof. This is theorem 1.4.1 in our syllabus.
Differentiating 1.6.2 partially with respect to x and y, we get

∂φ ∂u ∂u ∂z ∂φ ∂v ∂v ∂z
+ + + =0
∂u ∂x ∂z ∂x ∂v ∂x ∂z ∂x

∂φ ∂u ∂u ∂z ∂φ ∂v ∂v ∂z
+ + + =0
∂u ∂y ∂z ∂y ∂v ∂y ∂z ∂y
That is,

∂φ ∂u ∂u ∂φ ∂v ∂v
+p + +p =0
∂u ∂x ∂z ∂v ∂x ∂z

∂φ ∂u ∂u ∂φ ∂v ∂v
+q + +q =0
∂u ∂y ∂z ∂v ∂y ∂z
We can write this as:
φu (ux + puz ) + φv (vx + pvz ) = 0 (1.6.5)
φu (uy + quz ) + φv (vy + qvz ) = 0 (1.6.6)
Eliminating φu and φv between 1.6.5 and 1.6.6, we have

ux + puz vx + pvz
=
uy + quz vy + qvz
This means
(ux + puz )(vy + qvz ) = (vx + pvz )(uy + quz )
That is
ux vy + qux vz + puz vy + pquz vz = uy vx + quz vx + puy vz + pquz vz
ux vy + qux vz + puz vy = uy vx + puy vz + quz vx
ux vy − uy vx = p(uy vz − uz vy ) + q(uz vx − ux vz )
We can re-write this as
p(uy vz − uz vy ) + q(uz vx − ux vz ) = ux vy − uy vx
∂(u, v) ∂(u, v) ∂(u, v)

p +q = (1.6.7)
∂(y, z) ∂(z, x) ∂(x, y)
Now, we take the total derivatives of u(x, y, z) = c1 and v(x, y, z) = c2 , we get

∂u ∂u ∂u
dx + dy + dz = 0 

∂x ∂y ∂z

∂v ∂v ∂v
dx + dy + dz = 0 

∂x ∂y ∂z

That is,
ux dx + uy dy + uz dz = 0
(1.6.8)
vx dx + vy dy + vz dz = 0
Since u and v are independent functions, solving above two equations, we get
dx dy dz
= =
uy vz − uz vy uz vx − ux vz ux vy − uy vx
dx dy dz
= = (1.6.9)
∂(u, v) ∂(u, v) ∂(u, v)
∂(y, z) ∂(z, x) ∂(x, y)
1.6 Linear Equations of the First Order 37
Comparing 1.6.4 and 1.6.9,we obtain
P Q R
= = (1.6.10)
∂(u, v) ∂(u, v) ∂(u, v)
∂(y, z) ∂(y, z) ∂(z, x)
∂(u, v) ∂(u, v) ∂(u, v)

P =k ;Q = k ;R = k
∂(y, z) ∂(y, z) ∂(z, x)
Substituting in 1.6.7, we get
∂(u, v) ∂(u, v)
P p+Q q =p k +q k
∂(y, z) ∂(z, x)

∂(u, v) ∂(u, v)
=k p +q
∂(y, z) ∂(z, x)
∂(u, v)
=k ( From 1.6.7)
∂(x, y)
=R
Hence if u(x, y, z) = c1 and v(x, y, z) = c2 are two independent solutions of the system of
differential equations
dx dy dz
= = , (1.6.11)
P Q R
then φ(u, v) = 0 is a solution of P p + Qq = R, φ being an arbitrary function.
Remark 1.6.2.
(i) Equation 1.6.4 are called Lagrange’s auxiliary equations for 1.6.1.
(ii) Equation 1.6.4 is also known as characteristic equation associated with the equation
1.6.1.
Theorem 1.6.3. If u1 (x1 , x2 , . . . , xn , z) = c1 , u2 (x1 , x2 , . . . , xn , z) = c2 , . . . , un (x1 , x2 , . . . , xn , z) =

cn are independent solutions of the equations
dx1 dx2 dxn dz

= = ··· = = ,
P1 P2 Pn R
where P1 , P2 , . . . , Pn , R are coninuously differentiable functions of x1 , x2 , . . . , xn and z, not

simultaneously zero, then the relation φ(u1 , u2 , . . . , un ) = 0 where φ is an arbitrary differentiable
function is a general solution of the quasi-linear partial differential equation
∂z ∂z ∂z
P1 + P2 + · · · + Pn = R. (1.6.12)
∂x1 ∂x2 ∂xn
1.6.1 Working rules for solving P p + Qq = R by Lagrange’s method.

STEP 1 Write the given linear partial differential equation of the first order in the standard
form
P p + Qq = R (1)
STEP 2 Write down the Lagrange’s auxiliary equations for (1) namely
dx dy dz
= = (2)
P Q R
STEP 3 Solve (2). Let u(x, y, z) = c1 and v(x, y, z) = c2 be two linearly independent
solutions of (2).
STEP 4 The general solution (or integral) of (1) is written in one of the following three
equivalent forms:
φ(u, v) = 0, u = φ(v) or v = φ(u)
Examples 1.6.4. [4] Type I: Solve each of the following partial differential equations
y2z (iii) z p = −x
(i) p + xz q = y 2
x
(ii) tan x p + tan y q = tan z (iv) y 2 p − xy q = x(z − 2y)
y2z
Solution. (i) p + xz q = y 2
x
Lagrange’s auxiliary eqns are (ii) tan x p + tan y q = tan z
dx dy dz
= = . Lagrange’s auxiliary eqns are
P Q R dx dy dz
dx dy dz = = .
y2 z
= = 2 (1). P Q R
xz y dx dy dz
x
Taking the first two fractions, we get = = (1).
tan x tan y tan z
dx dy
y2 z
= .
xz Taking the first two fractions, we get
x
x2 dx = y 2 dy dx dy
= .
x3 = y 3 + c1 (2) tan x tan y
Taking the first and the last from (1). sin x
= c1 (2)
dx dz sin y
y2 z
= 2.
y
x
x dx = z dz. Taking the first and the last from (1).
x2 = z 2 + c2 (3). dx dz
= .
From (2) and (3), the required general tan x tan x
sin x
solution is = c2 (3)
sin z
φ(x3 − y 3 , x2 − z 2 ) = 0. From (2) and (3), the required general
Another form of the general solution
is
integral is sin x sin x
φ , =0
φ(x3 − y 3 ) = x2 − z 2 . sin y sin z
Another form of the general
integral is
sin x sin x
=φ Taking the second and the last from
sin z sin y (1).
dy dz
(iii) z p = −x = .
−xy x(z − 2y)
Lagrange’s auxiliary eqns are dy dz
dx dy dz =
= = . −y z − 2y
P Q R dy
dx dy dz (z − 2y) = −y
= = (1). dz
z 0 −x dz 1
=− z+2
dy y
(iv) y 2 p − xy q = x(z − 2y) dz 1
+ z+2
Lagrange’s auxiliary eqns are dy y
dx dy dz This is a linear equation in z and y.
= = . Its solution
P Q R Z is
dx
=
dy
=
dz
(1). ze ln y = 2eln y dy + c2
y2 −xy x(z − 2y)
Taking the first two fractions, we get =⇒ z y = y 2 + c2 z y + y 2 = c2 (3)
dx dy From (2) and (3), the required general
2
= . solution is
y −xy
φ x2 + y 2 , z y − y 2 = 0

x dx + y dy.
x2 + y 2 = c1 (2) where φ is an arbitrary function.

dx dy dz
Type II: Suppose one integral of = = is known but the another integral
P Q R
cannot be obtained. In this case, the one integral known to us is used to find
another integral as shown in the following examples
Examples 1.6.5. [4] Type II: Solve each of the following partial differential equations
(i) p + 3 q = 5z + tan(y − 3x) (iv) xz p + yz q = xy
(ii) z(z 2 + xy)(p x − q y) = x4 (v) p + 3 q = z + cot (y − 3x)
(iii) xy p + y 2 q = zxy − 2x2 (vi) x(z − 2y 2 ) p = (z − qy)(z − y 2 − 2x3 )
Solution. (i) p + 3 q = 5z + tan(y − 3x) dx dz

= .
Lagrange’s auxiliary eqns are 1 5z + tan c1
dx dy dz 1
= = . x = ln(5z + tan c1 ) = c2
P Q R 5
dx dy dz 5x − ln(5z + tan c1 ) = c2 (3).
= = (1). (c2 denotes a constant)
1 3 tan(y − 3x)
Taking the first two fractions, we get From (2) and (3), the required general
dx dy solution is
= . φ (y − 3x, 5x − ln(5z + tan c1 )) = 0
1 3
y − 3x = c1 (2). where φ is an arbitrary function.
Taking the first and the last from (1).
(ii) z(z 2 + xy)(p x − q y) = x4
xz(z 2 + xy) p − yz(z 2 + xy) q = x4 dx dz

= .
Lagrange’s auxiliary eqns are y zy − 2x
dx dy dz dx dz
= = . = .
P Q R y y z − 2 xy
dx dy
2
= = dx dz
xz(z + xy) −yz(z 2 + xy) = .
dz 1 z − 2c1
x4
(1). − 2c1 )+ c2 .
x = ln(z
Taking the first two fractions, we get x = ln z − 2 xy + c2 (3).
dx dy
2
=
xz(z + xy) −yz(z 2 + xy)
dx dy From (2) and (3), the required general
=⇒ = solution is
x −y
xy = c1 (2). φ xy , x − ln z − 2 xy = 0.
where φ is an arbitrary function.
Taking the first and the last from (1). (iv) xz p + yz q = xy
dx dz Lagrange’s auxiliary eqns are
2
= 4 .
xz(z + xy) x dx dy dz
dx dz = = .
= 3 . P Q R
z(z 2 + xy) x dx dy dz
dx dz = = (1).
= 3 . xz yz xy
z(z 2 + c1 ) x Taking the first two fractions, we get
x3 dx = z(z 2 + c1 ) dz . dx dy
=
x3 dx = 3
(z4 + zc 1 )dz . xz yz
x 4 z z2 dx dy
= + c1 = c2 . =⇒ =
4 4 2 x y
x
x4 − z 4 − 2z 2 c1 = c2 (3). = c1 =⇒ x = yc1 (2).
y
From (2) and (3), the required general Taking the first and the last from (1).
solution is dx dz
= .
φ xy, x4 − z 4 − 2z 2 c1 = 0 xz xy

where φ is an arbitrary function. dx dz

= .
z y
(iii) xy p + y 2 q = zxy − 2x2 y dx = z dz .
x
Lagrange’s auxiliary eqns are dx = z dz .
c1
dx dy dz x2 z2
= = . = + c2 .
P Q R 2c1 2
dx dy dz 2 2
= 2 = (1). x − c1 z
xy y zxy − 2x2 = c2 .
Taking the first two fractions, we get 2c1
dx dy y 2 c21 − c1 z 2
= 2 = c2 .
xy y 2c1
dx dy y 2 c1 − z 2 = c2 .
=⇒ = y 2 c1 − z 2 = c2 .
x y
x xy − z 2 = c2 (3).
= c1 =⇒ x = yc1 (2).
y
From (2) and (3), the required general
dx dz solution
is
= . φ xy , xy − z 2 = 0.
xy zxy − 2x2
where φ is an arbitrary function. dy dz

=
y z
(v) p + 3 q = z + cot (y − 3x) z
= c1 (2).
Lagrange’s auxiliary eqns are y
dx dy dz
= = .
P Q R Taking the first and second from (1).
dx dy dz
= = (1). dx dy
1 3 z + cot (y − 3x) = .
x(z − 2y 2 ) y(z − y 2 − 2x3 )
Taking the first two fractions, we get
dx dy
= y(z − y 2 − 2x3 ) dx − x(z − 2y 2 ) dy = 0
1 3 y(yc1 −y 2 −2x3 ) dx−x(yc1 −2y 2 ) dy =
=⇒ 3x = y + c1
=⇒ y − 3x = c1 (2) 0
(yc1 − y 2 − 2x3 ) dx − x(c1 − 2y) dy = 0
(yc1 − y 2 − 2x3 ) dx + x(2y − c1 ) dy = 0
dx dz M = yc1 − y 2 − 2x3 N = x(2y − c1 ).
= . ∂M ∂N
1 z + cot (y − 3x) = c1 − 2y and = 2y − c1
dx dz ∂y ∂x
= . ∂M ∂N
1 z + cot c1 − = 2(c1 − 2y).
x = ln (z + cot c1 ) + c2 x − ∂y ∂x
∂M ∂N
ln (z + cot (y − 3x)) = c2 (3) −
∂y ∂x 2(c1 − 2y) 2
From (2) and (3), the required general Hence = =−
N x(2y − c1 ) x
solution is which is a function of x alone.
φ (y − 3x, x − ln (z + cot (y − 3x))) = −2
−2
0. Hence I.F. is e x = ex = x−2 .
where φ is an arbitrary function. (yc1 − y 2 − 2x3 )x−2 dx + x(2y −
c1 )x−2 dy =Z 0
(vi) x(z − 2y 2 ) p = (z − q y)(z − y 2 − 2x3 )
x(z − 2y 2 ) p + y(z − y 2 − 2x3 ) q = Solution is (yc1 − y 2 − 2x3 )x−2 dx +
z(z − y 2 − 2x3 )
Z
Lagrange’s auxiliary eqns are terms not containing x dy = C
dx dy dz 1
= = . −(yc1 − y 2 ) − x2 = c2
P Q R x
dx dy y 2 − yc1 − x3
= = = c2 (3)
x(z − 2y 2 ) y(z − y 2 − 2x3 ) x
dz From (2) and (3), the required general
(1). solution is
z(z − y 2 − 2x3 )
z y 2 − yc1 − x3

Taking the last two fractions, we get φ , = 0.
dy dz y x
2 3
= where φ is an arbitrary function.
y(z − y − 2x ) z(z − y 2 − 2x3 )
[4] Type III: Let P1 , Q1 and R1 be functions of x, y and z. Then each fraction
dx dy dz
in Lagrange’s auxiliary eqns = = is equal to
P Q R
P1 dx + Q1 dy + R1 dz
(∗)
P1 P + Q1 Q + R1 R
If P1 P + Q1 Q + R1 R = 0, then we know that the numerator of (∗) is also 0. This

gives
P1 dx + Q1 dy + R1 dz = 0 (∗∗)
This can be integrated to get u1 (x, y, z) = c1 . This method may be repeated to

get another integral u2 (x, y, z) = c2 . P1 , Q1 , R1 are called multiplayers.
Examples 1.6.6. Solve the following equations
b−c c−a a−b (v) x(y 2 + z) p − y(x2 + z) q = z(x2 − y 2 )

(i) yz p + zx q = xy.
a b c
(vi) (x + 2z) p + (4xx − y) q = 2x2 + y
(ii) (mz − ny) p + (nx − lz) q = ly − mx
(vii) (z 2 −2yz −y 2 ) p+(xy +zx) q = xy −zx
(iii) x(y 2 − z 2 ) p − y(z 2 + x2 ) q = z(x2 + y 2 )
(viii) x(x2 + 3y 2 ) p − y(3x2 + y 2 ) q = 2z(y 2 −
(iv) (y − zx) p + (x + yz) q = x2 + y 2 x2 )
Solution. .
b−c c−a a−b a2 x dx + b2 y dy + c2 z dz

(i) yz p + zx q= xy. =
a b c 0
Lagrange’s auxiliary eqns are =⇒ a2 x dx + b2 y dy + c2 z dz = 0
dx dy dz
= = . Integrating,
P Q R =⇒ a2 x2 + b2 y 2 + c2 z 2 = c2 (3)
dx dy dz
= c − a = . From (2) and (3), the required general
b−c a−b
yz zx xy solution is
a b c
φ ax2 + by 2 + cz 2 , a2 x2 + b2 y 2 + c2 z 2 =

a dx b dy
= = 0.
yz(b − c) zx(c − a)
c dz
(1).
xy(a − b) (ii) (mz − ny) p + (nx − lz) q = ly − mx
Choosing x, y, z as multipliers, each Lagrange’s auxiliary eqns are
fraction equals dx dy dz
ax dx + by dy + cz dz = = .
= P Q R
xyz (b − c) + (c − a) + (a − b) dx dy dz
= = (1).
ax dx + by dy + cz dz mz − ny nx − lz ly − mx
= .
0
=⇒ ax dx + by dy + cz dz = 0. Choosing x, y, z as multipliers, each
Integrating fraction equals
x2 y2 z2 x dx + y dy + z dz
a + b + c = c1 =
2 2 2 x(mz − ny) + y(nx − lz) + z(ly − mx)
=⇒ ax2 + by 2 + cz 2 = c1 (2) x dx + y dy + z dz
= .
(c1 being arbitrary constant). 0
=⇒ x dx + y dy + z dz = 0.
Now, choosing ax, by and cz as multi-
Integrating
pliers for eqn (1), we get
a2 x dx + b2 y dy + c2 z dz =⇒ x2 + y 2 + z 2 = c1 (2)
= (c1 being arbitrary constant).
xyz a(b − c) + b(c − a) + c(a − b) Now, choosing l, m and n as multipli-
ers for eqn (1), we get
l dx + m dy + n dz Lagrange’s auxiliary eqns are

=
l(mz − ny) + m(nx − lz) + n(ly − mx) dx dy dz
l dx + m dy + n dz = = .
= P Q R
0 dx dy dz
= = 2 (1).
(y − zx) (x + yz) x + y2
Integrating,
=⇒ lx + my + nz = c2 (3) Choosing x, −y, z as multipliers, each
From (2) and (3), the required general fraction equals
solution is x dx − y dy + z dz
=
φ x2 + y 2 + z 2 , lx + my + nz = 0. xy − zx2 − xy − y 2 z + zx2 + zy 2

where φ is an arbitrary function. x dx − y dy + z dz

= .
0
(iii) x(y 2 − z 2 ) p − y(z 2 + x2 ) q = z(x2 + y 2 ) =⇒ x dx − y dy + z dz = 0.
Lagrange’s auxiliary eqns are Integrating
dx dy dz =⇒ x2 − y 2 + z 2 = c1 (2)
= = . (c1 being arbitrary constant).
P Q R
dx dy (Note that −x, y, −z also work as mul-
= =
2
x(y − z ) 2 −y(z 2 + x2 ) tipliers but
dz these multipliers are same as x, −y, z.
(1).
z(x2 + y 2 ) So, for
choosing the second set of multiplies,
Choosing x, y, z as multipliers, each we will not
fraction equals choose −x, y, −z as multipliers.
x dx + y dy + z dz Now, choosing −y, −x and 1 as multi-
= 2 2
x (y − z ) + y 2 (z 2 + x2 ) + z 2 (x2 + y 2 )
2
x dx + y dy + z dz −y dx − x dy + dz
= . =
0 −y + xyz − x2 − xyz + x2 + y 2
2
=⇒ x dx + y dy + z dz = 0.
= −y dx − x dy − +dz = 0
Integrating
= y dx + x dy − −dz = 0
=⇒ x2 + y 2 + z 2 = c1 (2)
= d(xy − dz = 0
(c1 being arbitrary constant).
1 1 1 Integrating,
Now, choosing , − and − as mul- xy − z = ln c2 (3)
x y z
tipliers for eqn (1), we get From (2) and (3), the required general
1 1 1 solution is
dx − dy − dz φ x2 − y 2 + z 2 , xy − z = 0.

x y z
= 2 where φ is an arbitrary function.
(y − z 2 ) + (z 2 + x2 ) − (x2 + y 2 )
1 1 1
dx − dy − dz (v) x(y 2 + z) p − y(x2 + z) q = z(x2 − y 2 )
x y z
= Lagrange’s auxiliary eqns are
0 dx dy dz
1 1 1 = = .
dx − dy − dz = 0 Integrating, P Q R
x y z
=⇒ ln x − ln y − ln z = ln c2 dx dy
= =
x 2
x(y + z) −y(x2 + z)
=⇒ = c2
yz dz
(1).
From (2) and (3), the required general z(x2 − y 2 )
solution is

x
1 1 1
φ x + y2 + z2,
2 = 0. Choosing , , as multipliers, each
yz x y z
where φ is an arbitrary function. fraction equals
1 1 1
x dx + y dy + z dz
(iv) (y − zx) p + (x + yz) q = x2 + y 2 = 2
(y + z) − (x2 + z) + (x2 − y 2 )
1 1 1
x dx + y dy + z dz solution is
= .
φ x2 − y − z, xy − z 2 = 0.

0
=⇒ x1 dx + y1 dy + z1 dz = 0. where φ is an arbitrary function.
Integrating
=⇒ xyz = c1 (2) (vii) (z 2 −2yz −y 2 ) p+(xy +zx) q = xy −zx
(c1 being arbitrary constant). Lagrange’s auxiliary eqns are
Now, choosing x, y and −1 as multipli- dx dy dz
ers for eqn (1), we get = = .
P Q R
x dx + y dy − dz dx dy
= 2 2 2 2 2 2 = =
x (y + z) − y (x + z) − z(x − y ) 2
(z − 2yz − y ) 2 (xy + zx)
x dx + y dy − dz dz
= (1).
0 xy − zx
= x dx + y dy − dz = 0
Integrating,
x2 + y 2 − 2z = c2 (3) From (2) Choosing x, y, z as multipliers, each
and (3), the required general fraction equals
x dx + y dy + z dz
solution is =
x(z − 2yz − y 2 ) + y(xy + zx) + z(xy − zx)
2
φ x2 + y 2 − 2z, xyz = 0.

x dx + y dy + z dz
where φ is an arbitrary function. = .
0
=⇒ x dx + y dy + z dz = 0.
(vi) (x + 2z) p + (4xz − y) q = 2x2 + y. Integrating
Lagrange’s auxiliary eqns are =⇒ x2 + y 2 + z 2 = c1 (2)
dx dy dz (c being arbitrary constant).
= = . 1
P Q R (note that finding second set of multi-
dx dy pliers is difficult)
= =
(x + 2z) (4xz − y) We will use the type I method.
dz
(1). From the last two equations of (1), we
(2x2 + y) get,
dy dz
Choosing 2x, −1, −1 as multipliers, =
(xy + zx) xy − zx
each fraction equals dy dz
2x dx − dy − dz =⇒ =
= (y + z) y −z
2x(x + 2z) − (4xz − y) − (2x2 + y) =⇒ (y − z) dy = (y + z) dz
2x dx − dy − dz =⇒ y dy − z dz = z dy + y dz
= .
0 =⇒ y dy − z dz = d(yz)
=⇒ 2x dx − dy − dz = 0.
Integrating Integrating,
=⇒ x2 − y − z = c1 (2) y 2 − z 2 − 2yz = c2 (3).
(c1 being arbitrary constant). From (2) and (3), the required general
Now, choosing y, x and −2z as multi- solution is
φ x2 + y 2 + z 2 , y 2 − z 2 − 2yz = 0.

y dx + x dy − 2z dz where φ is an arbitrary function.
=
y(x + 2z) + x(4xz − y) − 2z(2x2 + y)
y dx + x dy − 2z dz (viii) x(x2 + 3y 2 ) p − y(3x2 + y 2 ) q =
=
0 2z(y 2 − x2 ).
=⇒ y dx + x dy − 2z dz
Lagrange’s auxiliary eqns are
=⇒ d(xy) − 2z dz dx dy dz
Integrating, = = .
2
P Q R
xy − z = c2 (3) From (2) and dx dy
(3), the required general 2 2
= =
x(x + 3y ) −y(3x2 + y 2 )
dz dy y 3 + y 2
(1). =− x
2z(y 2 − x2 ) dx x 1+3 y 2

Choosing x1 , y1 , − z1 as multipliers, each x
dv 3 + v2
fraction equals v+x = −v
1 1 1 dx 1 + 3v 23
x dx + y dy − z dz

= 2 dv 3v + v
(x + 3y 2 ) − (3x2 + y 2 ) − 2(y 2 − x2 ) x =− 2
+ v
dx 1 + 3v 3
1 1 1
x dx + y dy − z dz v+v
= . = −4
0 1 + 3v 2
=⇒ x1 dx + y1 dy − z1 dz = 0. 1 + 3v 2 4
Integrating 3
dv = − dx
xy v+v x
=⇒ = c1 (2) 1 + 3v 2 4
z dv = − dx
(c1 being arbitrary constant). v(1 + v 2 ) x
1 2v 4
(note that finding second set of multi- + dv = − dx
v 1 + v2 x
pliers is difficult) ln v + ln(1 + v 2 ) + 4 ln x = ln c02
We will use the type I method. 2 )x4 = ln c0
From the first two equations of (1), we v(1 + v 2
y y 2

4
get, 1+ x = c2
x x
dx dy y 2
= x + y 2 x2 = c2

x(x2 + 3y 2 ) −y(3x2 + y 2 ) x
xy x2 + y 2 = c2

dy y(3x2 + y 2 )
=⇒ =− z x2 + y 2 = c2 (3)
dx x(x2 + 3y 2 )
From (2) and (3), the required general
solution
xy is
φ , z x2 + y 2 = 0.
z

dx dy dz
[4] RULE IV for solving = = (∗)
P Q R
Let P1 , Q1 and R1 be functions of x, y and z. Then all fractions in Lagrange
dx dy dz
auxiliary equations = = are equal to
P Q R
P1 dx + Q1 dy + R1 dz
(∗∗).
P1 P + Q1 Q + R1 R
Suppose the numerator is the exact differential of the denominator of (∗∗).

Then (∗∗) can be combined with a suitable fraction in (∗) to give an integral.
Examples 1.6.7. Solve the following partial differential equations
(i) (y + z) p + (z + x) q = x + y (vi) (x2 − yz) p + (y 2 − zx) q = z 2 − xy
(ii) y 2 (x − y) p + x2 (y − x) q = z(x2 + y 2 ) (vii) (x2 − y 2 − yz) p + (x2 − y 2 − zx) q =

z(x − y)
(iii) (x2 − y 2 − z 2 ) p + 2xy q = 2xz
(viii) (x2 + y 2 + yz) p + (x2 + y 2 − xz) q =
(iv) (1 + y) p + (1 + x) q = z z(x + y)
(v) xz p + yz q = xy (ix) cos(x + y) p + sin(x + y) q = z

1 (xiii) (2x2 + y 2 + z 2 − 2yz − zx − xy) p +

(x) cos(x + y) p + sin(x + y) q = z +
z (x2 + 2y 2 + z 2 − yz − 2zx − xy) q =
x2 + y 2 + 2z 2 − yz − zx − 2xy
p
(xi) x p + y q = z − a x2 + y 2 + z 2
(xii) p + q = x + y + z
Solution. (i) (y + z) p + (z + x) q = x + y ln(x + y + z) + 2 ln(x − y) = − ln C2

Lagrange’s auxiliary eqns are ln(x + y + z)(x − y)2 = ln c2
dx dy dz (x + y + z)(x − y)2 = c2 (7)
= = .
P Q R From (6) and (7), the required general
dx dy dz solution is
= = (1).
(y + z) (z + x) x+y x−y 2
φ , (x + y + z)(x − y) = 0.
y−z
Choosing 1, −1, 0 as multipliers, each where φ is an arbitrary function.
fraction equals
dx − dy
= (ii) y 2 (x − y) p + x2 (y − x) q = z(x2 + y 2 )
1(y + z) − (z + x) + 0(x + y)
dx − dy d(x − y) Lagrange’s auxiliary eqns are
= =− (2). dx dy dz
y−x x−y = = .
Choosing 0, 1, −1 as multipliers, each P Q R
fraction equals dx dy dz
= 2 = .
dy − dz dy − dz y 2 (x − y) x (y − x) z(x2 + y 2 )
= =− dx dy
z−y y−z = − 2 =
dy − dz d(y − z) 2
y (x − y) x (x − y)
= =− (3) dz
z−y y−z (1).
z(x + y 2 )
2
Choosing 1, 1, 1 as multipliers, each Taking the first two fractions of (1)

fraction equals dx dy
2
=− 2
dx + dy + dz y (x − y) x (x − y)
= = dx dy
y+z+z+x+x+y =⇒ 2 = ds 2 =⇒ x2 dx = −y 2 dy
dx + dz y x
(4)
2(x + y + z) Integrating, =⇒ x3 + y 3 = c1 (2)
From (2), (3) and (4), we have, Choosing 1, −1, 0 as multipliers, each
d(x − y) d(y − z) fraction equals
− = − =
x−y y−z dx − dy
dx + dy + dz = 2
(5) y (x − y) + x2 (x − y)
2(x + y + z) dx − dy
= (3)
Taking the first two fractions of (5) (x − y)(x2 + y 2 )
d(x − y) d(y − z) Combining the third fraction of (1)
=
x−y y−z with (3)
Integrating, ln(x−y) = ln(y−z)+ln c1 . dz dx − dy
x−y =
=⇒ = c1 (6). z(x2 + y 2 ) (x − y)(x2 + y 2 )
y−z dz dx − dy
Taking the first and third fractions of =
z (x − y)
(5) ln z − ln(x − y) = ln c2 .
d(x − y) dx + dy + dz z
− = = c2 . (4)
x−y 2(x + y + z) x−y
Integrating, −2 ln(x − y) = ln(x + y + From (2) and (4), the required general
z) + ln C2 solution is

z Choosing 1, 1, 0 as multipliers, each
φ x3 + y 3 , = 0.
x−y fraction equals
where φ is an arbitrary function. dx + dy
=
2+x+y
(iii) (x2 − y 2 − z 2 ) p + 2xy q = 2xz d(2 + x + y)
Lagrange’s auxiliary eqns are = (3)
2+x+y
dx dy dz
= = .
P Q R Combining third fraction of (1) with
dx dy dz
2 2 2
= = (1). (3), we get,
(x − y − z ) 2xy 2xz dz d(2 + x + y)
Taking the last two fractions of (1) = .
dy dy z 2+x+y
= 2+x+y
2xy 2xz = c2 (4)
dy dy z
= From (2) and (4), the required general
y z
y solution is
= c1 (2)
2+x+y

z 2
φ (1 + x) − (1 + y) , 2 =
Choosing x, y, z as multipliers, each z
0.
fraction equals
xdx + ydy + zdz where φ is an arbitrary function.
= 3
x − xy 2 − xz 2 ) + 2xy 2 + 2xz 2
xdx + ydy + zdz (v) xz p + yz q = xy
= 3 Lagrange’s auxiliary eqns are
x + xy 2 + xz 2 dx dy dz
xdx + ydy + zdz = = .
= (3) P Q R
x(x2 + y 2 + z 2 ) dx dy dz
Combining third fraction of (1) with = = (1).
(3), we get, xz yz xy
dz xdx + ydy + zdz Taking the first two fractions of (1)
= . dx dy
2xz x(x2 + y 2 + z 2 ) =
dz 2xdx + 2ydy + 2zdz xz yz
= . dx dy
z x2 + y 2 + z 2 =
x y
x2 + y 2 + z 2 x
= c2 (4) = c1 (2)
z y
From (2) and (4), the required general Taking the last two fractions of (1)
solution is dy dz
=
y x2 + y 2 + z 2

yz xy
φ , = 0. yc1 dy = z dz
z z
where φ is an arbitrary function. y 2 c1 z2
= + c2
2 2
yyc1 z 2
(iv) (1 + y) p + (1 + x) q = z
= + c2
Lagrange’s auxiliary eqns are 2 2
yx z 2
dx dy dz
= = . = + c2
P Q R 2 2
dx dy dz =⇒ y − z 2 2 = c2 (3)
= = (1). x
1+y 1+x z From (2) and (3), the required general
Taking the first two fractions of (1) solution is
dx dy

x
= (1 + x) dx = (1 + y) dy φ , xy − z 2 = 0
1+y 1+x y
(1 + x)2 = (1 + y)2 + c1
(1 + x)2 − (1 + y)2 = c1 ‘(2) (vi) (x2 − yz) p + (y 2 − zx) q = z 2 − xy
Lagrange’s auxiliary eqns are
dx dy dz 2xdx + 2ydy + 2zdz = 2(x + y + z)d(x +

= = .
P Q R y + z).
dx dy dz 2xdx + 2ydy + 2zdz = 2(x + y + z)d(x +
2
= 2 = 2 (1).
x − yz y − zx z − xy y + z).
Choosing 1, −1, 0 as multipliers, each
Integrating,
fraction equals
dx − dy (x2 + y 2 + z 2 ) = (x + y + z)2 + C2 .
= (x + y + z)2 − (x2 + y 2 + z 2 ) = −C2
1(x2 − yz) − (y 2 − zx) + 0(z 2 − xy)
dx − dy 2xy + 2yz + 2xz = −C2
= (2). xy + yz + xz = c2 (7)
(x − y) x + y + z From (6) and (7), the required general
Choosing 0, 1, −1 as multipliers, each solution
is
fraction equals x−y
dy − dz φ , xy + yz + xz = 0.
= = y−z
0(x2 − yz) + 1(y 2 − zx) − 1(z 2 − xy) where φ is an arbitrary function.
dy − dz
(3) (vii) (x2 − y 2 − yz) p + (x2 − y 2 − zx) q =
(y − z)(x + y + z)
Hence, from (2) and (3), z(x − y)
dx − dy dy − dz Lagrange’s auxiliary eqns are
= dx dy dz
(x − y)(x + y + z) (y − z)(x + y + z) = = .
dx − dy dy − dz P Q R
= dx dy
(x − y) (y − z) = =
2 2
(x − y − yz) (x − y 2 − zx)
2
d(x − y) d(y − z)
= dz
(x − y) (y − z) (1).
x−y z(x − y)
=⇒ = c1 (4) Choosing 1, −1, 0 as multipliers, each
y−z
Choosing x, y, z as multipliers, each fraction equals
fraction of (1) equals dx − dy
=
xdx + ydy + zdz 1(x − y − yz) − (x2 − y 2 − zx) + 0
2 2
x(x − yz) + y(y 2 − zx) + z(z 2 − xy)

2 dx − dy
= (2).
xdx + ydy + zdz z(x − y)
= 3 Combining this with the third fraction
(x + y 3 + z 3 − 3xyz)
xdx + ydy + zdz of (1),
(x + y + z)(x2 + y 2 + z 2 − xy − yz − xz) dz dx − dy
=
· · · (5) z(x − y) z(x − y)
=⇒ dz = dx − dy
Integrating,
Choosing 1, 1, 1 as multipliers, each
z − x + y = c1 . (3)
fraction of (1) equals
dx + dy + dz
= 2 (6)
(x − yz) + (y 2 − zx) + (z 2 − xy) Choosing x, −y, 0 as multipliers, each
From (5) and (6), fraction equals
xdx + ydy + zdz xdx − ydy
=
(x + y + z)(x2 + y 2 + z 2 − xy − yz − xz) x(x − y − yz) − y(x2 − y 2 − zx) + 0z(x − y)
2 2
dx + dy + dz xdx − ydy
= 2 . = 3
(x + y + z 2 − xy − yz − xz)
2 x − xy 2 − xyz − yx2 + y 3 + zxy
This means xdx − ydy
= 3
x + y 3 − xy 2 − yx2
xdx + ydy + zdz xdx − ydy
= dx + dy + dz. =
(x + y + z) (x + y)(x − xy + y 2 ) − xy(y + x)
2
xdx+ydy+zdz = (x+y+z)d(x+y+z). xdx − ydy
=
(x + y)[x2 − xy + y 2 − xy]
1.7 Pfaffian Differential equations 49
xdx − ydy general

=
(x + y)[x2 − 2xy + y 2 ] solution is
xdx − ydy x2 − y 2

= φ z − x + y, = 0.
(x + y)(x − y)2 z2
xdx − ydy where φ is an arbitrary function.
=
(x − y)(x2 − y 2 )
Combining this with the third fraction (viii) (x2 + y 2 + yz) p + (x2 + y 2 − xz) q =
of (1), z(x + y)
dz xdx − ydy
= (ix) cos(x + y) p + sin(x + y) q = z
z(x − y) (x − y)(x2 − y 2 )
dz xdx − ydy 1
= (x) cos(x + y) p + sin(x + y) q = z +
z (x2 − y 2 ) z
2dz 2xdx − 2ydy
= (xi) x p+y q =z−a
p
x2 + y 2 + z 2
z (x2 − y 2 )
2 2
ln(x − y ) − 2 ln z = ln c2 .
(xii) p+q =x+y+z
x2 − y 2
= c2 (4).
z2 (xiii) (2x2 + y 2 + z 2 − 2yz − zx − xy) p +
(x2 + 2y 2 + z 2 − yz − 2zx − xy) q =
Hence, from (3) and (4), the required x2 + y 2 + 2z 2 − yz − zx − 2xy
1.6.2 [4] Geometrical description of the solutions of P p + Qq = R and of the system

dx dy dz
of equations = = .
P Q R
Consider
P p + Qq = R
where P, Q and R are continuously differentiable functions of x, y, z and
dx dy dz
= =
P Q R
Let z = φ(x, y) be a solution of P p + Qq = R.
Then z = φ(x, y) represents a surface whose normal at any point (x, y, z) has direction
ratios p, q, −1.
dx dy dz
Also we know that the simultaneous equations = = represent a family of curves
P Q R
such that the tangent at any point has direction ratios P, Q, R.
Rewrite P p + Qq = R as P p + Qq + (−1)R = 0.
So, the normal to the surface z = φ(x, y) at any point is perpendicular to the member of
dx dy dz
family of curves = = through that point.
P Q R
dx dy dz
Since this holds for each point on z = φ(x, y), we conclude that the curves = =
P Q R
lie completely on the surface z = φ(x, y) whose differential equation is P p + Qq = R.
1.7 Pfaffian Differential equations

[3] By a Pfaffian differential equation, we mean a differential equation of the form
F1 (x1 , x2 , . . . , xn )dx1 + F2 (x1 , x2 , . . . , xn )dx2 + · · · + Fn (x1 , x2 , . . . , xn )dxn = 0 (1.7.1)

where Fi0 s, 1 ≤ i ≤ n are continuous functions.
The expression on the left hand side of 1.7.1 is called a Pfaffian differential form.
Definition 1.7.1. A Pfaffian differential form
F1 (x1 , x2 , . . . , xn )dx1 + F2 (x1 , x2 , . . . , xn )dx2 + · · · + Fn (x1 , x2 , . . . , xn )dxn
is said to be exact if we can find a continuously differentiable function u(x1 , x2 , . . . , xn ) such
that
du = F1 (x1 , x2 , . . . , xn )dx1 + F2 (x1 , x2 , . . . , xn )dx2 + · · · + Fn (x1 , x2 , . . . , xn )dxn . (1.7.2)
The function u(x1 , x2 , . . . , xn ) = c, where c is an arbitrary constant, is called the integral of
the corresponding Pfaffian differential form.
Definition 1.7.2. A Pfaffian differential equation
F1 (x1 , x2 , . . . , xn )dx1 + F2 (x1 , x2 , . . . , xn )dx2 + · · · + Fn (x1 , x2 , . . . , xn )dxn = 0
is said to be exact if the Pfaffian differential form on the left hand side of the equation is exact.
Definition 1.7.3. A Pfaffian differential equation
is said to be integrable if there exists a non-zero differentiable function µ(x1 , x2 , . . . , xn ) such
that the Pfaffian differential form
µ (F1 (x1 , x2 , . . . , xn )dx1 + F2 (x1 , x2 , . . . , xn )dx2 + · · · + Fn (x1 , x2 , . . . , xn )dxn )
is exact. The function µ(x1 , x2 , . . . , xn ) is called an integrating factor.
Remarks 1.7.4.
(i) u(x1 , x2 , . . . , xn ) = c is a surface in Rn such that at every point on it, the normal to
the surface has direction ratios (F1 , F2 , . . . , Fn ).
(ii) If a Pfaffian differential equation
is integrable, then there exist a non-zero differentiable function µ(x1 , x2 , . . . , xn ) and
a continuously differentiable function u(x1 , x2 , . . . , xn ) such that
h i
µ(x1 , x2 , . . . , xn ) F1 (x1 , x2 , . . . , xn )dx1 + F2 (x1 , x2 , . . . , xn )dx2 + · · · + Fn (x1 , x2 , . . . , xn )dxn
is exact. and
h i
du = µ(x1 , x2 , . . . , xn ) F1 (x1 , x2 , . . . , xn )dx1 + F2 (x1 , x2 , . . . , xn )dx2 + · · · + Fn (x1 , x2 , . . . , xn )dxn
and
u(x1 , x2 , . . . , xn ) = c
where c is an arbitrary constant is an integral of µ(F1 (x1 , x2 , . . . , xn )dx1 +F2 (x1 , x2 , . . . , xn )dx2 +
· · · + Fn (x1 , x2 , . . . , xn )dxn ) = 0.
Theorem 1.7.5. There always exists an integrating factor for a Pfaffian differential equation
in two variables.
Proof. This is theorem 1.5.1 in our syllabus.

Consider the Pfaffian differential equation in two variables
P (x, y) dx + Q(x, y) dy = 0 (1.7.3)
If Q(x, y) 6= 0, then
dy P (x, y)
=−
dx Q(x, y)
From the existence theorem for a first order ordinary differential equation, the above equa-
tion has a solution
F (x, y) = c1
Then the total derivative of F is

∂F ∂F
dx + dy = 0
∂x ∂y
We can see that this is an exact equation with solution F (x, y) = c.

This equation may differ form P dx + Q dy = 0 by an integrating factor but the two have
the same solution.
In general, a Pfaffian differential equation in more than two variables may not be inte-
grable. That is if the given Pfaffian differential equation in more than two variables is not
exact you may not be able to make it exact.
In the next theorem, we shall derive a necessary and sufficient condition for the integra-
bility (able to get converted into exact or not) of a Pfaffian differential equation in three
variables.
Remark 1.7.6. Let v = v(x, y) and let u = H(v) where H is continuously differentiable. (So
u is a function of x and y.) Then on differentiating the above equation partially with respect to
x and y, we get
ux = H 0 (v)vx , uy = H 0 (v)vy . (1.7.4)
On eliminating H 0 (v) from these equations obtain,
∂(u, v)
ux vy − uy vx = 0 that is = 0.
∂(x, y)
This is the first order partial differential equation.

This means if u = H(v) where the expression of H does contain x and y explicitly, then
∂(u, v)
= 0.
∂(x, y)
Lemma 1.7.7. Let u(x, y) = c1 and v(x, y) = c2 be two functions of x and y such that
∂v
6= 0 (1.7.5)
∂y
If, further
∂(u, v)
= 0, (1.7.6)
∂(x, y)
then there exists a relation
F (u, v) = 0 (1.7.7)
between u and v not involving x and y explicitly.
Proof. (Note: This lemma is Lemma 1.5.1 in our syllabus)
∂v
Since 6= 0, the function v is not independent of y.
∂y
We can get the value of y in terms of v and x and substitute it in the function u.
Thus we obtain the relation
F (u, v, x) = 0 (1.7.8)
We will now show that F does not depend on x.
Differentiating 1.7.8 partially with respect to x and y, we get,
∂F ∂u ∂F ∂v ∂F
+ + =0
∂u ∂x ∂v ∂x ∂x
and
∂F ∂u ∂F ∂v
+ =0
∂u ∂y ∂v ∂y
We can write this as
Fu ux + Fv vx + Fx = 0 (1.7.9)
and
Fu uy + Fv vy = 0 (1.7.10)
Case 1 vx 6= 0.
We eliminate Fv from the above equations

Fx + Fu ux
Fu ux + Fv vx + Fx = 0 =⇒ Fv = −
vx

Fu uy
Fu uy + Fv vy = 0 =⇒ Fv = −
vy

Fx + Fu ux Fu uy
So − =−
vx vy
=⇒ Fx vy + Fu ux vy − Fu uy vx = 0
=⇒ Fx vy + Fu (ux vy − uy vx ) = 0
∂(u, v)
=⇒ Fx vy + Fu =0
∂(x, y)
∂(u, v)
But = 0.
∂(x, y)
Hence Fx vy = 0.
As vy 6= 0, Fx = 0.
Thus F is independent of x.
Case 2 vx = 0.
∂(u, v)
Since = 0, ux vy − uy vx = 0.
∂(x, y)
This implies ux vy = 0.
As vy 6= 0, ux = 0.
Now, by 1.7.9, Fu ux + Fx = 0
Hence Fx = 0.
In any case, Fx = 0
Hence F is independent of x.
So, F (u, v, x) = F (u, v).
Hence there exists a relation F (u, v) = 0.
Definition 1.7.8. If X = (P, Q, R) is a vector field in R3 , and Py , Pz , Qx , Qz , Rx , Ry all exist,

then the curl of X is defined by
curl X = (Ry − Qz )î + (Pz − Rx )ĵ + (Qx − Py )k̂ (1.7.11)
Remarks 1.7.9.
(i) Note that the curl of a vector field is a vector field.
(ii) The definition of curl can be difficult to remember. To help with remembering, we
use the following determinant formula.
 
î ĵ k̂
 
 
∂ ∂ ∂ 
curl X = det 
 ∂x
 (1.7.12)
 ∂y ∂z 

 
P Q R
(iii) If X = (P, Q) is a vector field in R2 , then the curl of X = (P, Q) is a vector field in
R2 and is defined by
curl X = (Qx − Py )k̂ (1.7.13)

Lemma 1.7.10. If X = P (x, y, z), Q(x, y, z), R(x, y, z) and µ is an arbitrary nonzero
differentiable function of x, y and z then

X · curl X = 0 if and only if µX · curl µX = 0.
Proof. This is lemma 1.5.2 in the syllabus.

µX = µ(P î + Q ĵ + R k̂)
= (µ P î + µ Q ĵ + µ R k̂)
curl X = (Ry − Qz )î + (Pz − Rx )ĵ + (Qx − Py )k̂

curl µX = (µR)y − (µQ)z î + (µP )z − (µR)x )ĵ + (µQ)x − (µP )y k̂
h i h i
= µ(Ry − Qz ) + (µy R − µz Q) î + µ(Pz − Rx ) + (µz P − µx R) ĵ+
h i
µ(Qx − Py ) + (µx Q − µy P ) k̂
h i
= µ (Ry − Qz )î + (Pz − Rx )ĵ + (Qx − Py )k̂ + (µy R − µz Q)î + (µz P − µx R)ĵ+
‘ (µx Q − µy P )k̂
= µ curl X + (µy R − µz Q)î + (µz P − µx R)ĵ + (µx Q − µy P )k̂
h i h i
µX · curl µX = µ2 X · curl X + µX · (µy R − µz Q)î + (µz P − µx R)ĵ + (µx Q − µy P )k̂
h i h i
= µ2 X · curl X + µ P (µy R − µz Q) + Q(µz P − µx R) + R(µx Q − µy P )
h i
= µ2 X · curl X + 0
h i
= µ2 X · curl X

Since µ is a non-zero function, X · curl X = 0 if and only if µX · curl µX = 0
Theorem 1.7.11. A necessary and sufficient condition that the Pfaffian differential equation
X · dr = P (x, y, z)dx + Q(x, y, z)dy + R(x, y, z)dz = 0 (1.7.14)
be integrable is that
X · curl X = 0 (1.7.15)
Proof. This is theorem 1.5.2 in the syllabus.
Suppose the equation 1.7.14 is integrable.
So there exist a non-zero differentiable function µ(x, y, z) and a continuously differentiable
function u(x, y, z) such that
h i
du = µ(x, y, z) P (x, y, z)dx + Q(x, y, z)dy + R(x, y, z)dz
But,
∂u ∂u ∂u
du = dx + dy + dz (1.7.16)
∂x ∂y ∂z
Comparing 1.7.16 and 1.7.16, we get
∂u
µ(x, y, z)P (x, y, z) = ,
∂x
∂u
µ(x, y, z)Q(x, y, z) = ,
∂y
∂u
µ(x, y, z)R(x, y, z) = .
∂z

∂u ∂u ∂u
Since , , = gradiant u = ∇ u,
∂x
∂y ∂z
we have, µ(x, y, z)P (x, y, z), µ(x, y, z)Q(x, y, z), µ(x, y, z)R(x, y, z)) = ∇ u,

∇ u = µ(x, y, z)P (x, y, z), µ(x, y, z)Q(x, y, z), µ(x, y, z)R(x, y, z))

= µ(x, y, z) P (x, y, z), Q(x, y, z), R(x, y, z)
= µX
As curl (∇u) = 0 for all u, we have curl µX) = 0.
This implies µX · curl µX) = 0
Hence by 1.7.10 X · curl X = 0.

Conversely, suppose X · curl X = 0.
Suppose z is treated as a constant.
So the given Pfaffian equation becomes
P (x, y, z) dx + Q(x, y, z) dy = 0
By theorem 1.7.5, a Pfaffian differential equation in two variables is always integrable.

This means, there exists a non-zero differentiable function µ and a continuously differen-
tiable function U such that
∂U ∂U
= µ P, =µQ
∂x ∂y
and U (x, y, z) = c1 where c1 may contain z, is a solution (integral) of P (x, y, z) dx +
Q(x, y, z) dy = 0.
Multiplying the given Pfaffian equation by µ, we get
µ P (x, y, z)dx + µ Q(x, y, z)dy + µ R(x, y, z)dz = 0
So,
∂U ∂U
dx + dy + µ R dz = 0
∂x ∂y
∂U ∂U ∂U ∂U
dx + dy + dz − dz + µ R dz = 0
∂x ∂y ∂z ∂z

∂U ∂U ∂U ∂U
dx + dy + dz + µ R − dz = 0
∂x ∂y ∂z ∂z
∂U ∂U ∂U
As dU = dx + dy + dz, above equation becomes,
∂x ∂y ∂z

∂U
dU + µ R − dz = 0
∂z

∂U
Let K = µ R − .
∂z
Hence
dU + K dz = 0 (∗)
We are given that X · curl X

= 0.

Hence by 1.7.10, µX · curl µX = 0
µX = µ(P, Q, R)
= (µ P, µ Q, µ R)

∂U ∂U
= , , µR
∂x ∂y

∂U ∂U ∂U
= , , K+
∂x ∂y ∂z

∂U ∂U ∂U
µX · curl µX = , , K+ ·
∂x ∂y ∂z
∂2U ∂2U ∂2U ∂2U ∂2U ∂2U

∂K ∂K
+ − , − − , −
∂y ∂z∂y ∂y∂z ∂z∂x ∂z∂x ∂x ∂y∂x ∂x∂y

∂U ∂U ∂U ∂K ∂K
= , , K+ · , , 0
∂x ∂y ∂z ∂y ∂x
∂U ∂K ∂U ∂K
= −
∂x ∂y ∂y ∂x
∂(U, K)
= .
∂(x, y)
∂(U, K)
Hence .=0
∂(x, y)
So, by theorem 1.7.7, there is a relation between U and K independent of x and y.
Hence K can be expressed in terms U and z alone.
We can write the equation (∗) as
dU + K(U, z) dz = 0
dU
This is an ordinary differential equation = −K(U, z).
dz
This has a solution φ(U, z) = c where c is an arbitrary constant.
This solution can now be expressed in terms of x, y and z.
Therefore Pfaffian differential equation is integrable.
Remark 1.7.12. The Pfaffian differential equation P (x, y, z)dx+Q(x, y, z)dy+R(x, y, z)dz =
0 is exact if and only if curl X = 0 where X = P (x, y, z)î + Q(x, y, z)ĵ + R(x, y, z)k̂
Example 1.7.13. Show that the following Pfaffian differential equation is exact and find its
integral.
y dx + x dy + 2z dz = 0
Solution. Given Pfaffian equation y dx + x dy + 2z dz = 0
STEP 1 Write X.
Here, X = (P, Q, R) = (y, x, 2z).
STEP 2 Find curl X.

 
î ĵ k̂
 
 
∂ ∂ ∂ 
curl X = det 
 ∂x

 ∂y ∂z  
 
P Q R
 
î ĵ k̂
 
 
∂ ∂ ∂ 
= det 
 ∂x

 ∂y ∂z  
 
y x 2z
∂ ∂ ∂ ∂ ∂ ∂
= 2z − x î − 2z − y ĵ + x− y k̂
∂y ∂z ∂x ∂z ∂x ∂y
= 0î − 0ĵ + (1 − 1)k̂
= 0î + 0ĵ + 0k̂
= (0, 0, 0)
Hence the given Pfaffain equation is integrable.
STEP 3 Keep z as a constant and write the differential equation as follows:
y dx + x dy = 0
STEP 4 Find the solution of the above equation.

dx dy
y dx + x dy = 0 =⇒ =− .
x y
Solution is ln x + ln y = ln c1 where c1 is a constant and it may contain z.
This implies xy = c1 . So, U (x, y, z) = xy = c1 where c1 may contain z.
STEP 5 Find µ using the equations

∂U ∂U
= µ P, OR =µQ
∂x ∂y
Here y = µ y. So, µ(x, y, z) = 1.

∂U
STEP 6 Find K using the relation K = µ R− .
∂z
∂U
Since U = xy, = 0.
∂z
So, K = µ R.
Here K = 1 ∗ 2z.
STEP 7 Solve the differential equation dU + K dz = 0 and get U as a function of K and

z.
dU + 2z dz = 0. This means dU = −2z dz. So U = −z 2 + c.
Thus we have obtained U = xy = −z 2 + c.
Hence the integral (solution) is u(x, y, z) = xy + z 2 = c
[4] Special Method: Solution by inspection

Sometimes the given equation after making certain changes may contain parts which are
exact differentials. The following formulae my be useful to find the total derivative of the
exact parts.
xdy − ydx y

(7) x dy + y dx = d(xy)
(1) = d
x2 x
xdy − ydx y Z
df (x, y, z)
(2) = d ln (8) = ln f (x, y, z)
xy x f (x, y, z)
xdy − ydx
−1 y

(3) = d tan x dy + y dx

1

x2 + y 2 x (9) − =d
x2 y 2 xy
xdy + ydx
(4) = d (ln yx)
xy
(10) y 2 dx + 2xy dy = d(y 2 x)
xdy + ydx 1 2 2

(5) =d ln x + y
x2 + y 2 2 (11) 2(x dx + y dy + z dz) = d(x2 + y 2 + z 2 )
2xydy − y 2 dx
2
y
(6) 2
=d (12) 3x2 y dx + x3 dy = d(x3 y)
x x
Chapter 2
Unit II
Definition, Necessary and sufficient condition for integrability, Charpit’s method,

Some standard types, Jacobi’s method, The Cauchy problem. (Ref Book: An El-
ementary Course in Partial Differential Equations by T. Amaranath, 2nd edition,
Chapter 1: 1.6, Theorem 1.6.1, 1.7, 1.8 Theorem 1.8.1, 1.9 and related examples)
2.1 Compatible system of first order Partial differential equations

Definition 2.1.1. The equations

f (x, y, z, p, q) = 0, (1)
g(x, y, z, p, q) = 0. (2)
are said to the compatible on a domain D if
∂(f, g)
(i) 6= 0 on D,
∂(p, q)
(ii) If p = φ(x, y, z) and q = ψ(x, y, z) are obtained by solving equations (1) and (2) for p and
q then the Pfaffian differential equation dz = φ(x, y, z) dx + ψ(x, y, z) dy is integrable.
Theorem 2.1.2. A necessary ad sufficient condition for the integrability of
dz = φ(x, y, z) dx + ψ(x, y, z) dy
is
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

[f, g] = + p+ + q=0
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
Proof. dz = φ(x, y, z) dx + ψ(x, y, z) dy is integrable if and only if X · curlX = 0

60 CHAPTER 2. UNIT II
where X = (φ(x, y, z), ψ(x, y, z), −1).

 
î ĵ k̂
 
 
∂ ∂ ∂ 
curl X = det   
 ∂x ∂y ∂z 

 
P Q R
 
î ĵ k̂
 
 
∂ ∂ ∂ 
= det 
 ∂x

 ∂y ∂z 

 
φ ψ −1
∂ ∂ ∂ ∂ ∂ ∂
= (−1) − ψ î − (−1) − φ ĵ + ψ− φ k̂
∂y ∂z ∂x ∂z ∂x ∂y
= −ψz î + φz ĵ + (ψx − φy ) k̂
= (−ψz , φz , ψx − φy )
X · curl X = (φ, ψ, − 1) · (−ψz , φz , ψx − φy )
= −φ ψz + ψ φz − ψx + φy
Hence dz = φ(x, y, z) dx+ψ(x, y, z) dy is integrable if and only if −φ ψz +ψ φz −ψx +φy = 0.
That is, dz = φ(x, y, z) dx + ψ(x, y, z) dy is integrable if and only if
ψ φz + φy = φ ψz + ψx (2.1.1)
Substitue φ and ψ for p and q respectively in (1) and (2) and differentiate partially with
respect to x and z. So, we get
fx + fp ∗ φx + fq ∗ ψx = 0 (3) gx + gp ∗ φx + gq ∗ ψx = 0 (5)
fz + fp ∗ φz + fq ∗ ψz = 0 (4) gz + gp ∗ φz + gq ∗ ψz = 0 (6)
(3) + φ ∗ (4) =⇒ (5) + φ ∗ (6) =⇒
fx + φ ∗ fz + fp ∗ (φx + φ ∗ φz )+ gx + φ ∗ gz + gp ∗ (φx + φ ∗ φz )+
fq ∗ (ψx + φ ∗ ψz ) = 0 gq ∗ (ψx + φ ∗ ψz ) = 0
2.1 Compatible system of first order Partial differential equations 61
Thus we have the following two equations

fx + φ ∗ fz + fp ∗ (φx + φ ∗ φz ) + fq ∗ (ψx + φ ∗ ψz ) = 0 (7)
gx + φ ∗ gz + gp ∗ (φx + φ ∗ φz ) + gq ∗ (ψx + φ ∗ ψz ) = 0 (8)
Multiplying equation (7) by gp and equation (8) by fp and then subtracting the second from
the first, we get
fx gp − gx fp + φ(fz gp − gz fp ) + (fq gp − gq fp )(ψx + φ ∗ ψz ) = 0
That is
fx gp − gx fp fz gp − gz fp
(ψx + φ ∗ ψz ) = − −φ
fq gp − gq fp fq gp − gq fp
This means
1
(ψx + φ ∗ ψz ) = − [(fx gp − gx fp ) + φ (fz gp − gz fp )]
fq gp − gq fp
Hence
1
(ψx + φ ∗ ψz ) = [(fx gp − gx fp ) + φ (fz gp − gz fp )]
gq fp − fq gp
That is

1 ∂(f, g) ∂(f, g)
(ψx + φ ∗ ψz ) = +φ (2.1.2)
∂(f, g) ∂(x, p) ∂(z, p)
∂(p, q)
If we differentiate the equations (1) and (2) partially with respect to y and z, we get
fy + fp ∗ φy + fq ∗ ψy = 0 (3) gy + gp ∗ φy + gq ∗ ψy = 0 (5)
fz + fp ∗ φz + fq ∗ ψz = 0 (4) gz + gp ∗ φz + gq ∗ ψz = 0 (6)
(3) + ψ ∗ (4) =⇒ (5) + ψ ∗ (6) =⇒
fy + ψ ∗ fz + fp ∗ (φy + ψ ∗ φz )+ gy + ψ ∗ gz + gp ∗ (φy + ψ ∗ φz )+
fq ∗ (ψy + ψ ∗ ψz ) = 0 gq ∗ (ψy + ψ ∗ ψz ) = 0
Thus we have the following two equations

fy + ψ ∗ fz + fp ∗ (φy + ψ ∗ φz ) + fq ∗ (ψy + ψ ∗ ψz ) = 0 (7)
gy + ψ ∗ gz + gp ∗ (φy + ψ ∗ φz ) + gq ∗ (ψy + ψ ∗ ψz ) = 0 (8)
Multiplying equation (7) by gq and equation (8) by fq and then subtracting the second from
the first, we get
fy gq − gy fq + ψ(fz gq − gz fq ) + (fp gq − gp fq )(φy + ψ ∗ φz ) = 0

That is
fy gq − gy fq fz gq − gz fq
(φy + ψ ∗ φz ) = − −ψ
fp gq − gp fq fp gq − gp fq
This means
1
(φy + ψ ∗ φz ) = − [(fy gq − gy fq ) + ψ (fz gq − gz fq )]
fp gq − gp fq
Hence
1
(φy + ψ ∗ φz ) = − [(fy gq s − gy fq ) + ψ (fz gq − gz fq )]
gq fp − fq gp
That is

1 ∂(f, g) ∂(f, g)
(φy + ψ ∗ φz ) = − +ψ (2.1.3)
∂(f, g) ∂(y, q) ∂(z, q)
∂(p, q)
Hence ψ φz + φy = φ ψz + ψx if and only if

1 ∂(f, g) ∂(f, g) 1 ∂(f, g) ∂(f, g)
+φ =− +ψ
∂(f, g) ∂(x, p) ∂(z, p) ∂(f, g) ∂(y, q) ∂(z, q)
∂(p, q) ∂(p, q)
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

+φ + +ψ =0
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
By replacing φ by p and ψ by q we get
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

+p + +q =0
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
Remark 2.1.3. A solution of dz = φ(x, y, z) dx + ψ(x, y, z) dy is of the form
F (x, y, z, c) = 0 (2.1.4)
Hence if the equations f (x, y, z, p, q) = 0 and g(x, y, z, p, q) = 0 are compatible then they have
a one-parameter family of common solutions.
Example 2.1.4. Show that the equations
f = xp − yq − x = 0 (2.1.5)
g = x2 p + q − xz = 0 (2.1.6)
are compatible.
Solution.

∂(f, g) x −y 1
= det 2 = x + x2 y 6= 0 where D = (x, y) : x 6= 0, y 6= − (1).
∂(p, q) x 1 x
We will solve the equations 2.1.5 and 2.1.6 for p and q as follows:
p q 1
= =
x −y x x x −y
xz 1 x2 xz x2 1
p q 1
= 2 3
=
x + xyz x z−x x + x2 y
p q 1
= 2 =
x(1 + yz) x (z − xy) x(1 + xy)
1 + yz x(z − x)
p= , q= (2.1.7)
1 + xy 1 + xy
Given equations 2.1.5 and 2.1.6 are compatible if and only if 2.1.9 is completely integrable.
That is, the given system is compatible if and only if the differential equation
1 + yz x(z − x)
dz = dx + dy (2.1.8)
1 + xy 1 + xy
is integrable.
We know that the above equation is integrable if and only if
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

+p + +q =0
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)

∂(f, g) fx fp
=
∂(x, p) gx gp

p−1 x
=
2xp − z x2
= x2 p − x2 − 2x2 p + xz
= −x2 p − x2 + xz

∂(f, g) fz fp
=
∂(z, p) gz gp

0 x
=
−x x2
= x2

∂(f, g) fy fq
=
∂(y, q) gy gq

−q −y
=
0 1
= −q

∂(f, g) fz fq
=
∂(z, q) gz gq

0 −y
=
−x 1
= −xy
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

+p + +q = (−x2 p − x2 + xz) + x2 p − q − xy q
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
= −x2 + xz − q − xy q
= −x2 + x2 p − xyq (as g = 0 =⇒ x2 p = xz − q)
= x(−x + xp − yq)
=0 (as f = 0) (2)
Hence from (1) and (2), the given two equations are compatible.
1 + yz x(z − x)
Method 2 For proving that dz = dx + dy is integrable, we can show that
1 + xy 1 + xy
X · curlX = 0
STEP 1 Write X.
1 + yz x(z − x)
Here, X = (P, Q, R) = , , −1 .
1 + xy 1 + xy
STEP 2 Find curl X.

 
î ĵ k̂
 
 
∂ ∂ ∂ 
curl X = det 
 ∂x

 ∂y ∂z 

 
P Q R
 
î ĵ k̂
 
 
 ∂ ∂ ∂ 
 
= det 
 ∂x ∂y ∂z 

 
 
 1 + yz x(z − x) 
−1
1 + xy 1 + xy
∂ x(z − x) ∂ 1 + yz ∂ x(z − x) ∂ 1 + yz
= − î + ĵ + − k̂
∂z 1 + xy ∂z 1 + xy ∂x 1 + xy ∂y 1 + xy
x y (z − 2x)(1 + xy) − x(z − x)(y) z(1 + xy) − (1 + yz)x
= − î + ĵ + − k̂
1 + xy 1 + xy (1 + xy)2 (1 + xy)2
−xî + y ĵ z(1 + xy) − 2x − 2x2 y − xyz + x2 y − z(1 + xy) + x + xyz

= + k̂
(1 + xy) (1 + xy)2
−xî + y ĵ −x − x2 y
= + k̂
(1 + xy) (1 + xy)2
−xî + y ĵ x
= − k̂
(1 + xy) 1 + xy
−xî + y ĵ − xk̂
=
(1 + xy)
STEP 3 Find X · curl X.

1 + yz x(z − x) x y x
X · curl X = , , −1 · − , ,−
1 + xy 1 + xy 1 + xy 1 + xy 1 + xy
x + xyz 2
xyz − x y x(1 + xy)
=− + +
(1 + xy)2 (1 + xy)2 (1 + xy)2
=0 (3)
Hence the given Pfaffain equation is integrable.
Hence from (1) and (3), the given two equations are compatible.
Example 2.1.5. Given that
f = xp − yq − x = 0
g = x2 p + q − xz = 0
are compatible on D = (x, y) : x 6= 0, y 6= − x1 . Find a one-parameter family of common

solutions.
Solution. Since the two equations are compatible, the Pfaffian differential equation dz =
φ(x, y, z) dx + ψ(x, y, z) dy ( φ and ψ are obtained by solving the given equations for p and
q) is integrable and its solution is a common solution of the given two equations. .
We will solve the equations 2.1.5 and 2.1.6 for p and q as follows:
p q 1
= =
x −y x x x −y
xz 1 x2 xz x2 1
p q 1
= 2 3
=
x + xyz x z−x x + x2 y
p q 1
= 2 =
x(1 + yz) x (z − xy) x(1 + xy)
1 + yz x(z − x)
p= , q=
1 + xy 1 + xy
The required Pfaffian differential equation is
1 + yz x(z − x)
dz = dx + dy.
1 + xy 1 + xy
1 + yz x(z − x)
dz − dx = dx − dx + dy = 0.
1 + xy 1 + xy
1 + yz − 1 − xy x(z − x)
dz − dx = dx + dy = 0.
1 + xy 1 + xy
y(z − x) x(z − x)
dz − dx = dx + dy = 0.
1 + xy 1 + xy
dz − dx y dx + x dy
= = 0.
z−x 1 + xy
z−x
= c.
1 + xy
z−x
Hence = c is a one-parameter family of common-solutions.
1 + xy
Corollary 2.1.6. Show that the first order partial differential equations p = M (x, y) and
∂M ∂N
q = N (x, y) are compatible if and only if =
∂y ∂x
Proof. We write these equations in the form f = 0 and g = 0.

Therefore the two equations are f = p − M (x, y) = 0 and g = q − N (x, y) = 0.
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

First we will show that +p + +q = Nx − My
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)

∂(f, g) fx fp
=
∂(x, p) gx gp

Mx 1
=
−Nx 0
= Nx

∂(f, g) fz fp
=
∂(z, p) gz gp

0 1
=
0 0
=0

∂(f, g) fy fq
=
∂(y, q) gy gq

−My 0
=
−Ny 1
= −M y

∂(f, g) fz fq
=
∂(z, q) gz gq

0 0
=
0 1
=0
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

Hence +p + +q = Nx +p∗0−My +q∗0 = Nx −My (∗)
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
Now, Suppose the first order partial differential equations p = M (x, y) and q = N (x, y) are
compatible.
Therefore f = p − M (x, y) = 0 and g = q − N (x, y) = 0 are compatible.
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)
Hence +p + +q =0
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
∂M ∂N
Hence Nx − My = 0. Thus Nx = My . That is, Hence = .
∂y ∂x
∂M ∂N
Conversely, suppose = .
∂y ∂x
To show that f = p − M (x, y) = 0 and g = q − N (x, y) = 0 are compatible.
∂(f, g)
• To show that 6= 0.
∂(p, q)

∂(f, g) fp fq
=
∂(p, q) gp gq

1 0
=
0 1
= 1 6= 0 on any domain D.
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

• To show that +p + +q =0
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

+p + +q = Nx − My from ∗
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
∂N ∂M
=
∂x ∂y
=0
Hence f = 0 and g = 0 are compatible.
∂N ∂M
Remark 2.1.7. If p = M (x, y) and q = N(x, y) are two equations. If 6= then
∂x ∂y
f = p − M (x, y) = 0 and g = q − N (x, y) = 0 are not compatible.
∂z ∂z
Example 2.1.8. Show that the differrential equations = 2x − 3y and = 5x − 9y are
∂x ∂y
not compatible.
∂z ∂z
Solution. Given equations are = 2x − 3y and = 5x − 9y. That is, p = 2x − 3y and
∂x ∂y
q = 5x − 9y.
The given equations are of the form p = M (x, y) and q = N (x, y).
∂M ∂N
By 2.1.6, the two equations are compatible if and only if = .
∂y ∂x
∂M ∂N
Here M = 2x − 3y and N = 5x − 9y. So, = −3 and =2
∂y ∂x
∂M ∂N
Therefore 6= = 2 and hence the given equations are not compatible.
∂y ∂x
Example 2.1.9. [4]Show that the following partial differential equations are compatible and
solve them.
(1) p = x2 − ay, q = y 2 − ax.

y x
(2) p = x − ,q = y + 2
x2 +y 2 x + y2
x x

(3) p = 1 + e y , q = e y 1 − xy
Solution. (1) p = x2 − ay, q = y 2 − ax.

The given equations are of the form p = M (x, y) and q = N (x, y).
∂M ∂N
Hence, they are compatible if and only if = .
∂y ∂x
∂M ∂N
Here = −a, = −2a.
∂y ∂x
∂M ∂N
Hence = = −2a.
∂y ∂x
Therefore, the two equations are compatible.
Now we will find a common solution.
STEP 1 Solve the two equations for p and q.

Here p = x2 − ay = φ(x, y, z), q = y 2 − ax = ψ(x, y, z).
STEP 2 Construct the Pfaffian differential equation dz = φ dx + ψ dy.

That is, dz = (x2 − ay) dx + (y 2 − ax) dy.
dz = (x2 − ay) dx + (y 2 − ax) dy.

= x2 dx + y 2 dy − a(y dx + x dy).
= x2 dx + y 2 dy − ad(xy).
d(x3 + y 3 )
dz = − ad(xy).
3
On integrating, we get, =⇒ z = x3 + y 3 − axy + c.
STEP 4 z = x3 + y 3 − axy + c is the required solution
y x
(2) p = x − ,q = y + 2 .
x2 +y 2 x + y2
The given equations are of the form p = M (x, y) and q = N (x, y) where M =
y x
x− 2 2
,N = y + 2 .
x +y x + y2
∂M ∂N
∂y ∂x
∂M (x2 + y 2 ) − y(2y) ∂N (x2 + y 2 ) − x(2x)
Here =− , = .
∂y (x2 + y 2 )2 ∂x (x2 + y 2 )2
∂M (y 2 − x2 ) − y(2y) ∂N y 2 − x2
Hence = = = .
∂y (x2 + y 2 )2 ∂x (x2 + y 2 )2

y x
Here p = x − 2 2
and q = y + 2
x +y x + y2
STEP 2 Construct the Pfaffian differential

equation dz
= φ dx + ψ dy.
y x
That is, dz = x − 2 dx + y + 2 dy.
x + y2 x + y2

∂M ∂N
The right hand side is exact differential as = , there exists a function
∂y ∂x
∂F ∂F
F (x, y) such that = M and = N and this F can be found as follows:
Z ∂x ∂y Z
F (x, y) = M dx (treating y as constant) + N dy (only those terms from N
which do not contain x).
Once this F is obtained then we can see that the total derivative of F = dF =
∂F ∂F
dx + dy.
∂x ∂y
∂F ∂F
But dx + dy = M dx + N dy which is equal to dz.
∂x ∂y
That is, dF = dz.
Hence z = F + c and thus we will get the solution of the given equations.
So, we
Z will
find F using
Z
y x
F = x− 2 dx(treating yas constant) + y+ 2 dy (only
x + y2 x + y2
those terms which do not contain x).
x2 y2

−1 x
F (x, y) = − tan + + c.
2 y 2
x2 y2

x
=⇒ = − tan−1 + + c.
2 y 2
x2 + y 2

−1 x
STEP 4 z = − tan + c. is the required solution
2 y
x x

(3) p = 1 + e y , q = e y 1 − xy .
The given equations are
of the form p = M (x, y) and q = N (x, y) where M =
x x

x
1 + ey , N = ey 1 − y .
∂M ∂N
∂y ∂x
x
M = 1 + ey
x
∂M xe y
=− 2 ,
∂y y

x x
N = ey 1 −
y

∂N x x 1 1
=e y 1− −
∂x y y y
x
xe y
=− 2 .
y
∂M (y 2 − x2 ) − y(2y) ∂N y 2 − x2
Hence = 2 2 2
= = 2 .
∂y (x + y ) ∂x (x + y 2 )2
STEP 1 Solve the two equationsfor p and

q.
x x
x
Here p = 1 + e and q = e 1 − y
y y

x
x
x
That is, dz = 1 + e y dx + e y 1 − y dy.

∂M ∂N
The right hand side is exact differential as = , there exists a function
∂y ∂x
∂F ∂F
F (x, y) such that = M and = N and this F can be found as follows:
Z ∂x ∂y Z
F (x, y) = M dx (treating y as constant) + N dy (only those terms from N
which do not contain x).
Once this F is obtained then we can see that the total derivative of F = dF =
∂F ∂F
dx + dy.
∂x ∂y
∂F ∂F
But dx + dy = M dx + N dy which is equal to dz.
∂x ∂y
That is, dF = dz.
Hence z = F + c and thus we will get the solution of the given equations.
So, we willZfind F using Z
x
x x
F (x, y) = 1+e y dx(treating y as constant) + e y 1− dy (only
y
those terms which do not contain x).
x
F (x, y) = x + ye y + 0 + c.
x
STEP 4 z = x + ye y + c is the required solution.

Example 2.1.10. Show that the following partial differential equations are compatible and
solve them.
(1) xp = yq and z(xp + yq) = 2xy
(2) xp − yq = x and x2 p + q = xz
Solution. (1) xp = yq and z(xp + yq) = 2xy.

=⇒ f (x, y, z, p, q) = xp − yq = 0, g(x, y, z, p, q) = z(xp + yq) − 2xy = 0
∂(f, g)
∂(p, q)

∂(f, g) fp fq
=
∂(p, q) gp gq

x −y
=
zx zy
= 2xyz 6= 0 on D = {(x, y, z) ∈ R3 : xyz 6= 0}
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)

∂(f, g) fx fp
=
∂(x, p) gx gp

p x
=
zp − 2y zx
= zpx − zpx + 2xy = 2xy

∂(f, g) fz fp
=
∂(z, p) gz gp

0 x
=
xp + yq zx
= −x2 p − xyq

∂(f, g) fy fq
=
∂(y, q) gy gq

−q −y
=
zq − 2x zy
= −2xy

∂(f, g) fz fq
=
∂(z, q) gz gq

0 −y
=
(xp + yq) zy
= y(xp + yq)
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

Hence +p + +q
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
= 2xy + p ∗ (−x2 p − xyq) − 2xy + q ∗ y(xp + yq)
= −p2 x2 + y 2 q 2
= 0 as xp = yq
Hence the two equations are compatible.

Here xp = yq substitue in z(xp + yq)q = 2xy =⇒ z(2yq) = 2xy. This means
x
zq = x. So, q = .
z
y
This implies p = .
z

y x
That is, dz = dx + dy =⇒ z dz = d(xy) =⇒ z 2 = 2xy + c
z z
STEP 3 z 2 = 2xy + c is the required solution.
(2) xp − yq = x and x2 p + q = xz
=⇒ f (x, y, z, p, q) = xp − yq − x = 0, g(x, y, z, p, q) = x2 p + q − xz = 0
∂(f, g)
∂(p, q)

∂(f, g) fp fq
=
∂(p, q) gp gq

x −y
= 2
x 1
= x(1 + xy) 6= 0 on D = {(x, y, z) ∈ R3 : x 6= 0, xy 6= −1}
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)

∂(f, g) fx fp
=
∂(x, p) gx gp

p−1 x
=
2xp − z x2
= −x2 (p + 1) + xz

∂(f, g) fz fp
=
∂(z, p) gz gp

0 x
=
−x 1
= x2

∂(f, g) fy fq
=
∂(y, q) gy gq

−q −y
=
0 1
= −q

∂(f, g) fz fq
=
∂(z, q) gz gq

0 −y
=
−x 1
= −xy
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)
Hence +p + +q
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
= −x2 (p + 1) + xz + p ∗ (x2 ) − q + q ∗ (−xy)
= −x2 + xz − q − qxy
= −x2 + x2 p − qxy ( as x2 p + q = xz)
= x(−x + xp − qy)
= x(0) = 0 ( as xp − yq = x)
Hence the two equations are compatible.

Here xp = x + yq substitue in x2 p + q = xz =⇒ x(x + yq) + q = xz =⇒
x(z − x)
x2 + xyq + q = xz. This means q(xy + 1) = x(z − x). So, q = .
xy + 1
xy(z − x) y(z − x) 1 + zy
This implies xp = x + =⇒ p = 1 + =⇒ p =
xy + 1 xy + 1 1 + xy

1 + zy x(z − x)
That is, dz = dx + dy =⇒ z dz = d(xy) =⇒ z 2 = 2xy + c
1 + xy 1 + xy
(1 + xy) dz = (1 + zy) dx + x(z − x) dy

(1 + xy) dz = dx + zy dx + xz dy − x2 dy
(1 + xy) dz = dx + z(y dx + z dy) − x2 dy
(1 + xy) dz − z(y dx + z dy) = dx − x2 dy
(1 + xy) dz − z(y dx + z dy) dx − x2 dy
=
(1 + xy)2 (1 + xy)2
1
2 dx − dy

z
d = x 2
(1 + xy) 1
+ y
x
− x12 dx + dy

z
d =− 2
(1 + xy) 1
x +y
!
z 1
d =d 1
(1 + xy) x +y
On integrating, we get,
z 1
= 1 + c.
(1 + xy) x +y
z x
= + c.
(1 + xy) 1 + xy
z x
STEP 3 = + c is the required solution.
(1 + xy) 1 + xy
Example 2.1.11. [4] It is given that if f (x, y, z, p, q) = 0 which is homogeneous of degree n

∂f ∂f ∂f
in x, y, z then x +y + z = nf (x, y, z, p, q).
∂x ∂y ∂z
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)
Show that +p + +q =0
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
Solution. =⇒ f (x, y, z, p, q) = 0, px + qy − z = 0

∂(f, g) fx fp
=
∂(x, p) gx gp

fx fp
=
p x
= xfx − pfp

∂(f, g) fz fp
=
∂(z, p) gz gp

fz fp
=
−1 x
= xfz + fp

∂(f, g) fy fq
=
∂(y, q) gy gq

f f
= y q

q y
= y fy − q f q

∂(f, g) fz fq
=
∂(z, q) gz gq

0 −y
=
−1 y
= yfz + fq
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

Hence +p + +q
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
= xfx − pfp + p ∗ (xfz + fp ) + y fy − q fq + q ∗ (yfz + fq )
= xfx − pfp + p ∗ xfz + y fy − q fq + p ∗ xfz + q ∗ yfz
= xfx + y fy + (px + qy)fz
= xfx + y fy + zfz ( as px + qy = z)
= nf (x, y, z, p, q) as f homogeneous of degree n in x, y, z
=0
∂(f, g) ∂(f, g)
Example 2.1.12. If f (x, y, p, q) = 0, g(x, y, p, q) = 0 and + = 0 then [f, g] =
∂(x, p) ∂(y, q)
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)
+p + +q =0
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
Solution.

∂(f, g) fx fp
=
∂(x, p) gx gp

fx fp
=
p x
= fx gp − gx fp

∂(f, g) fz fp
=
∂(z, p) gz gp

0 fp
=
0 gp
=0

∂(f, g) fy fq
=
∂(y, q) gy gq

f fq
= y
gy gq
= gq fy − gy fq

∂(f, g) fz fq
=
∂(z, q) gz gq

0 fq
=
0 gq
=0
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

Hence +p + +q
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)
∂(f, g) ∂(f, g)
= +p∗0+ +q∗0
∂(x, p) ∂(y, q)
∂(f, g) ∂(f, g)
= +
∂(x, p) ∂(y, q)
=0 · · · given

Exercise 2.1.13. [4] Show that the following system of partial differential equations are com-
patibe and hence solve them.
(1) p = 6x + 3y, q = 3x − 4y, Ansz = 3x2 + 3xy − 2y 2 + c

ax2 + by 2
(2) p = ax + hy + g, q = hx + by + f, Ans : z = + hxy + gx + f y + c
2
x5
(3) p = x4 − 2xy 2 + y 4 , q = 4xy 3 − 2x2 y − sin y, Ans : z = − x2 y 2 + xy 4 + cos y + c
5
(4) p = (ey + 1) cos x, q = ey sin x, Ans : z = (ey + 1) sin x + c
(5) p = y 1 + x1 + cos y, q = x + log x − x sin y, z = y(x + logx) + x cos y + c

2 2 2
(6) p = y 2 exy + 4x3 , q = 2xyexy − 3y 2 , Ans : z = exy + x4 − y 3 + c
2.2 [3] Charpits Method 77
(7) p = sin x cos y + e3x , q = cos x sin y + tan y, Ans : x = 13 e3x − cos x cos y + log sec y + c
2.2 [3] Charpits Method

Charpit’s method finds a complete integral of a first order partial differential equation.
Let
f (x, y, z, p, q) = 0 (2.2.1)
be the partial differrential equation whose complete integral is being sought.
Definition 2.2.1. A family of partial differential equations
g(x, y, z, p, q, a) = 0 (2.2.2)
is said to be a one-parameter family of partial differential equations compatible with 2.2.1 if

2.2.2 is compatible with 2.2.1 for each value of a.
Remark 2.2.2. Suppose we are given a pde f (x, y, z, p, q) = 0. An integral of
dx dy dz dp dq
= = =− =− (2.2.3)
fp fq pfp + qfq fx + pfz fy + qfz
gives a partial differential equation g(x, y, z, p, q) = 0 which is compatible with f = 0.
We want g = 0 such that g = 0 is compatible with f = 0.
∂(f, g) ∂(f, g) ∂(f, g) ∂(f, g)

Hence +p + +q =0
∂(x, p) ∂(z, p) ∂(y, q) ∂(z, q)

fx fp fz fp fy fq fz fq

gx +p
+ +q
=0
gp gz gp gy gq gz gq
fx gp − fp gx + p(fz gp − fp gz ) + (fy gq − fq gy ) + q(fz gq − fq gz ) = 0
−fp gx − fq gy − (pfp + qfq )gz + (pfz + fx ) gp + (fy + qfz ) gq = 0.
=⇒ fp gx + fq gy + (pfp + qfq )gz − (pfz + fx ) gp − (fy + qfz ) gq = 0
∂g ∂g ∂g ∂g ∂g
fp + fq + (pfp + qfq ) − (pfz + fx ) − (fy + qfz ) = 0. (2.2.4)
∂x ∂y ∂z ∂p ∂q
This is a linear equation of the first order whose solution is our g.
We use the following theorem 1.6.3 to solve the above equation.

If u1 (x1 , x2 , . . . , xn , z) = c1 , u2 (x1 , x2 , . . . , xn , z) = c2 , . . . , un (x1 , x2 , . . . , xn , z) = cn are

independent solutions of the equations
dx1 dx2 dxn dz
= = ··· = = ,
P1 P2 Pn R
where P1 , P2 , . . . , Pn , R are coninuously differentiable functions of x1 , x2 , . . . , xn and z, not
simultaneously zero, then the relation φ(u1 , u2 , . . . , un ) = 0 where φ is an arbitrary differ-
entiable function is a general solution of the quasi-linear partial differential equation
∂z ∂z ∂z
P1 + P2 + · · · + Pn = R. (2.2.5)
∂x1 ∂x2 ∂xn
Hence, we consider the equations

dx dy dz dp dq dz dg
= = =− =− = = (2.2.6)
fp fq pfp + qfq fx + pfz fy + qfz pfp + qfq 0
Since any of the integrals of 2.2.6 will satisfy 2.2.4, we will find an integral of 2.2.6 which
involves p or q or both. Thus we will get the required equation g(x, y, z, p, q) = 0
The equation 2.2.6 is called Charpit’s Auxiliary equations.
Remark 2.2.3. Note that we want the functions p = φ and q = ψ. We can find p = φ (or
q = ψ) from the Charpit’s Auxiliary equations and then substitute the value of p = φ (or q = ψ)
in f = 0 to get q = ψ (or p = φ).
2.2.1 [4] Working Rule while using Charpit’s Method

STEP 1 Write the given partial differential equation in the form f = 0.
STEP 2 Find all the values required in Charpit’s Auxiliary equation.
STEP 3 Write the Charpits’ Auxiliary equation using the above values.
STEP 4 Select two proper fractions so that we can easily find the integral g = 0 involving
at least one of p and q.
STEP 5 Solve the equations f = 0 and g = 0 for p and q and get the functions p = φ and
q = ψ.
STEP 6 Write the Pfaffian differential equation dz = φ dx + ψ dy.
STEP 7 Integral of this equation is our required complete integral of f = 0.
Example 2.2.4. [3] Find a complete integral of each of the following partial differential equa-
tion by Charpit’s method.
(1) z 2 − xy pq = 0
(2) x2 p2 + y 2 q 2 − 4 = 0
(3) z 2 (1 + p2 + q 2 ) = 1
(4) px5 − 4q 2 x2 + 6x2 z − 2 = 0
(5) z 2 (p2 z 2 + q 2 ) − 1 = 0.
Solution. (1) f = z 2 − xy pq = 0
STEP 1 f (x, y, z, p, q) = z 2 − xy pq = 0
STEP 2 Charpit’s auxiliary eqns
dx dy dz dp dq
= = =− =−
fp fq (p fp + q fq ) (fx + pfz ) (fy + qfz )
dx dy dz dp dq
= = =− =−
−xyq −xyp (p (−xyq) + q (−xyp)) (2zp − ypq + 0) (2zq − xpq + 0)
dx dy dz dp dq
= = = = =
−xyq −xyp 2xypq (2zp − ypq) (2zq − xpq)
dx dy dz dp dq
= = = = =
xyq xyp 2xypq (2zp − ypq) (2zq − xpq)
STEP 3 Solving above equations.
Take p, q, 0, x, y as multipliers, each ratio equals
p dx + q dy + x dp + y dq
=
2xzp + 2yzq
dz p dx + q dy + x dp + y dq
This implies =
2xypq 2xzp + 2yzq
dz p dx + q dy + x dp + y dq
=⇒ 2
=
2z 2z(xp + yq)
dz p dx + q dy + x dp + y dq dz p dx + q dy + x dp + y dq
=⇒ = =⇒ =
z (xp + yq) z (xp + yq)
dz d(xp + yq)
=⇒ =
z (xp + yq)
=⇒ ln z = ln(xp + yq)a
Hence the required one parameter family of pde which is compatible with f = 0
is
g(x, y, z, p, q, a) = z − a(xp + yq) = 00.
STEP 4 Now we solve f = 0 and g = √0 simultaneously to get p and q.
zc 1 ± 1 − 4a2
So, q = where c =
y 2a

1
which is a solution of a c + =1
c
zc 1 hz i 1 hz i
Now, put q = in p = − yq =⇒ p = − cz
y x a x a

z 1
=⇒ p = −c
x a 1 1 1
Now, a c + = 1 =⇒ − c =
c a c
z
Hence p=
cx
STEP 5 Solve φ(x, y, z) dx + ψ(x, y, z) dy = dz
z zc
Here, dx + dy = dz
cx y
1 dx dy dz
=⇒ +c =
c x y z
1
=⇒ ln x + c ln y + ln b = ln z.
c
1
=⇒ z = bx c y c
1
Hence F (x, y, z, b, c) = z − bx c y c = 0 is an integral of f = 0.

Fb Fbx Fby
STEP 6 We can show that the matrix M = is of rank 2.
Fc Fcx Fc y
1
STEP 7 Conclusion: F (x, y, z, b, c) = z − bx c y c = 0 is a complete integral of f = 00.
(2) f = x2 p2 + y 2 q 2 − 4 = 0
We find all the values required in Charpit’s Auxiliary equation.
Charpit’s Auxiliary equations are
dx dy dz dp dq
= = =− =−
2x2 p 2y 2 q p(2x2 p) + q(2y 2 q) 2xp2 + p(0) 2yq 2 + q(0)
2x2 p 2y 2 q 2x2 p2 + 2y 2 q 2 2xp2 2yq 2
dx dy dz dp dq
2
= 2 = 2 2 2 2
=− 2
=−
2x p 2y q 2(x p + y q ) 2xp 2yq 2
Consider first and fourth ratio.
dx dp
2
=−
2x p 2xp2
dx dp
=−
x p
Integrating both sides, we get, ln x = − ln p + ln a. This means xp = a. That is,

a
p= .
x
a2
Substituting this in f = x2 p2 + y 2 q 2 − 4 = 0, we get, x2 2 + y 2 q 2 − 4 = 0.
x
a2
x2 2 + y 2 q 2 − 4 = 0
x
a2 + y 2 q 2 = 4√
4 − a2
q =
y
Consider the Pfaffian differential equation dz = p dx + q dy.
√
a 4 − a2
dz = dx + dy
x y
p
=⇒ z = a ln x + 4 − a2 ln y + b
√
z = F (x, y; a, b) = a ln x+ 4 − a2 ln y+b is a complete integral of f = x2 p2 +y 2 q 2 −4 =
0
dy dq
Note: If we consider second and fourth ratio, that is, 2
= − then we get
√ 2y q 2yq 2
z = c ln y + 4 − c2 ln x + d as a complete integral.
√
So, z = a ln x + √4 − a2 ln y + b
and z = c ln y + 4 − c2 ln x + d
are complete integrals of f = x2 p2 + y 2 q 2 − 4 = 0.
We can observe that one integral can be obtained from another merely by a change
in the choice of arbitrary constants.
Remark 2.2.5. A first order partial differential equation can have several complete
integrals.
(3) z 2 (1 + p2 + q 2 ) = 1. That is, f (x, y, z, p, q) = z 2 (1 + p2 + q 2 ) − 1 = 0.

dx dy dz dp dq
= = =− =−
2z 2 p 2z 2 q p ∗ 2z 2 p + q ∗ 2z 2 q 0 + p ∗ 2z(1 + p2 + q 2 ) 0 + q ∗ 2z(1 + p2 + q 2 )
2z 2 p 2z 2 q 2z 2 p2 + 2z 2 q 2 2zp(1 + p2 + q 2 ) 2zq(1 + p2 + q 2 )
dx dy dz dp dq
2
= 2 = 2 2 2
=− 2 2
=−
2z p 2z q 2z (p + q ) 2zp(1 + p + q ) 2zq(1 + p2 + q 2 )
Consider the fourth and the fifth ratio.

dp dq
− =−
2zp(1 + p2 + q 2 ) 2zq(1 + p2 + q 2 )
dp dq
=
p q
Integrating both sides, we get, ln p = ln q + ln a. This means p = qa.

Substituting this in f = z 2 (1 + p2 + q 2 ) − 1 = 0, we get, z 2 (1 + q 2 a2 + q 2 ) − 1 = 0.
= z 2 (1 + q 2 a2 + q 2 ) − 1 = 0
=⇒ z 2 q 2 (a2
√ + 1) = 1 − z
2
1 − z2
=⇒ q = √
z√ a2 + 1
a 1 − z2
So, p = √
z a2 + 1
Consider the Pfaffian differential equation dz = p dx + q dy.
√ √
a 1 − z2 z 1 − z2
dz = √ dx + √ dy
z a2 + 1 a2 + 1
z dz a dx dy
=⇒ √ =√ +√
Z 1 − z 2 Z a2 + 1 aZ2 + 1
z dz a dx dy
=⇒ √ = √ + √
1−z 2 2
a +1 a2 + 1
√
Z
2
z dz z
put t = 1 − z in √ . This implies dt = − √ .
2 1 − z2
Z Z1 − z
z dz p
Hence √ = − dt = −t = −( 1 − z 2 )
1 − z2
Z Z Z
z dz a dx dy
√ = √ + √
1−z 2 2
a +1 a2 + 1
p ax y
− 1 − z2 = √ +√ +b
a2 + 1 a2 + 1
c
Put a = −m, b = √ .
2
m +1
√ −mx y c
So, − 1 − z 2 = √ +√ +√
m2 + 1 m2 + 1 m2 + 1
√ −mx + y + c
=⇒ − 1 − z 2 = √
m2 + 1
=⇒ (1 − z 2 )(m2 + 1) = (−mx + y + c)2
(1 − z 2 )(m2 + 1) = (y − mx + c)2 is a complete integral of f = x2 p2 + y 2 q 2 − 4 = 0.
(4) px5 − 4q 2 x2 + 6x2 z − 2 = 0.

dx dy dz dp dq
= = =− =−
x5 − 8qx2 p(x5 ) + q(−8qx2 ) 5x4 p − 8q 2 x + 12xz + p6x2 0 + q6x2
x5 − 8x2 q x5 p − 8x2 q 2 5x4 p − 8xq 2 + 12xz + 6x2 p 6x2 q
dx dy dz dp dq
5
= 2
= 5 2 2
=− 4 2 2
=− 2
x −8x q x p − 8x q 5x p − 8xq + 12xz + 6x p 6x q
dy dq dy dq
Taking the second and the last ratio, = − 2 . This means = .
−8x2 q 6x q 4 3
3
That is, 4 dq = 3 y + 3a. So, q = (y + a).
4
3
Substituting q = (y + a) in px5 − 4q 2 x2 + 6x2 z − 2 = 0, we get
4
2
5 3
px − 4 (y + a) x2 + 6x2 z − 2 = 0
4
px5 − 9(y + a)2 x2 + 6x2 z − 2 = 0
9
4 (y + a)2 x2 + 6x2 z + 2
=p
x5
9(y + a)2 6z 2
− 3 + 5 =p
4x3 x x
We substitute the values of p and q in the Pfaffian differential equation dz = p dx+q dy.
9(y + a)2 6z 2 3(y + a)
dz = dx − 3 dx + 5 dx + dy
4x3 x x 4
6z 9(y + a)2 2 3(y + a)

dz + 3
dx = 3
dx + 5 dx + dy
x 4x x 4
6z 9(y + a)2 2 3(y + a)

3
dx + dz = 3
dx + 5 dx + dy (∗)
x 4x x 4
We will find an integrating factor of L.H.S. of the above equation.
6z ∂M 6 ∂N
M = 3 , N = 1. So, = 3 and = 0.
x ∂z x ∂x
∂M
∂z− f rac∂N ∂x 6
So, = 3 which is a function of x alone.
R 6N x
dx −3x−2
Hence e x 3 =e is an integrating factor.
Multiplying the both sides of the equation
−2 −2 −2 −2
6ze−3x −3x−2 9(y + a)2 e−3x 2e−3x 3(y + a)e−3x
dx + e dz = dx + dx + dy
x3 4x3 x5 4
−2 −2 −2
−3x−2 9(y + a)2 e−3x 2e−3x 3(y + a)e−3x
=⇒ d(ze )= dx + dx + dy
4x3 x5 4
−2 −2 −2
−3x−2 9(y + a)2 e−3x 3(y + a)e−3x 2e−3x
=⇒ d(ze )= dx + dy + dx
4x3 4 x5
−2
−2 3 −2 3 −2 2e−3x
=⇒ d(ze−3x ) = ∗ (y + a)2 ∗ e−3x ∗ 6x−3 dx + ∗ 2(y + a)e−3x dy + dx
8 8 x5
i 2e−3x −2
−2 3h −2 −2
=⇒ d(ze−3x ) = (y + a)2 ∗ e−3x ∗ 6x−3 dx + 2(y + a)e−3x dy + dx
8 x5
−2 3 −2
2e−3x−2
=⇒ d(ze−3x ) = d (y + a)2 ∗ e−3x + dx
8 x5
Integrating both sides
3 Z 2e−3x−2
−3x−2 2 −3x−2
ze = (y + a) ∗ e + dx (∗∗)
8 x5
−2
2e−3x
Z
Put t = x−2 in dx.
x5
−2
2e−3x
Z Z
−2
So, dt = −2x−3 dx and dx = − e−3x x−2 (−2x−3 ) dx.
x5
−2
2e−3x
Z Z
dx = − e−3t t dt.
x5
h e−3t t Z −3t
e i
=− − dt.
−3 −3
Z
1 h −3t i
= e t − e−3t dt.
3
1 h −3t e−3t i
= e t − . +b
3 −3
−2
1 h −3x−2 −2 e−3x i
= e x + +b
3 3
−2 3 −2 1 −2 1 −2
Hence from (∗∗), ze−3x = (y + a)2 ∗ e−3x + e−3x x−2 + e−3x + b
8 3 9
3 1 1 −2
z = (y + a)2 + 2 + + be3x
8 3x 9
This is a complete integral of px5 − 4q 2 x2 + 6x2 z − 2 = 0.
(5) z 2 (p2 z 2 + q 2 ) − 1 = 0.
dx dy dp dq dz dg
= =− =− = =
fp fq (fx + pfz ) (fy + qfz ) (p fp + q fq ) 0
fp fq fx + pfz fy + qfz pfp + qfq
2pz 4 2qz 2 0 + p(4p2 z 3 + 2zq 2 ) − 0 + q(4p2 z 3 + 2zq 2 ) p(2pz 4 ) + q(2qz 2 )
2pz 4 2qz 2 2pz(2p2 z 2 + q 2 ) − 2qz(2p2 z 2 + q 2 ) 2z 2 (p2 z 2 + q 2 )
dx dy dp dq dz dg
4
= 2
=− 2 2 2
=− 2 2 2
= 2 2 2 2
=
2pz 2qz 2pz(2p z + q ) 2qz(2p z + q ) 2z (p z + q ) 0
dp dq dp dq
2 2 2
= 2 2 2
=⇒ = −→ p = aq.
2pz(2p z + q ) 2qz(2p z + q ) p q
Substituting this in z 2 (p2 z 2 + q 2 ) − 1 = 0, we get, z 2 (a2 q 2 z 2 + q 2 ) − 1 = 0.

This means z 2 (a2 q 2 z 2 + q 2 ) = 1.
1 1
Hence q 2 = 2 2 2 and q = √ .
z (a z + 1) z a z2 + 1
2
a
So, p = √ .
z a z2 + 1
2
a 1
The Pfaffian differential equation is dz = √ dx + √ dy.
2 2
z a z +1 z a z2 + 1
2
√
That is, z a2 z 2 + 1 dz = a dx Z + dy.
p
Integrating both sides, we get, z a2 z 2 + 1 dz = ax + y + b. (∗)
dt
Let t = a2 z 2 + 1. So, dt = 2a2 z dz. This means = dz.
2 a2
Z p Z √
2 2
1
z a z + 1 dz = 2 t dt
2a
1 2 3
= 2 t2
2a 3
1 3
= 2 (a2 z 2 + 1) 2
3a
1 3
Substituting this in (∗), 2
(a2 z 2 + 1) 2 = ax + y + b.
3a
That is (a2 z 2 + 1)3 = 9a4 (ax + y + b)2 .
This is a complete integral.

Example 2.2.6. Find a complete integral of z = px + qy + p2 + q 2 by Charpit’s method and

also verify that it is a complete integral of z = px + qy + p2 + q 2 .
Solution. Given partial differential equation is f (x, y, z, p, q) = px + qy + p2 + q 2 − z = 0.

dx dy dz dp dq
= = =− =−
x + 2p y + 2q p(x + 2p) + q(y + 2q) p + p(−1) q + q(−1)
x + 2p y + 2q px + qy + 2(p2 + q 2 ) 0 0
dx dy dz dp dq
= = 2 2
=− =−
x + 2p y + 2q px + qy + 2(p + q ) 0 0
dp = 0 =⇒ p = a, dq = 0 =⇒ q = b
Substituting p = a and q = b in f = 0, we get ax + by + a2 + b2 − z. This implies
z = ax + by + a2 + b2 .
So, z = ax + by + a2 + b2 is a complete integral of px + qy + p2 + q 2 − z = 0.
We will verify this.

To show that z = ax + by + a2 + b2 is an integral of px + qy + p2 + q 2 − z = 0.
z = ax + by + a2 + b2 =⇒ p = zx = a, q = zy = b.
Substituting this in L.H.S. of px + qy + p2 + q 2 − z = 0, we get,
L.H.S.= ax + by + a2 + b2 − (ax + by + a2 + b2 ) = 0 = R.HS.
Hence z = ax + by + a2 + b2 is an integral of px + qy + p2 + q 2 − z = 0.
Now we wil show that this integral is complete.
If we writethis in the form z = 2 2
F (x, y; a, b) then F (x, y; a, b) = ax + by + a + b . To show
Fa Fax Fay
that M = where F (x, y, a, b) =is of rank 2.
Fb Fbx Fb y

Fa Fax Fay x + 2a 1 0
M= =
Fb Fbx Fb y y + 2b 0 1
Therefore, M is of rank 2.

2.2.2 Some standard types of partial differential equations

Type I
Example 2.2.7. If the partial differential equation is of the form f (p, q) = 0, that is, the
equation does not involve x, y and z explicitly then show that the Charpit’s Auxiliary equations
of f (p, q) = 0 are
dx dy dz dp dq
= = = = .
fp fq p f p + q fq 0 0
Also show that a complete integral of f (p, q) = 0 is z = F (x, y; a, b) where z = ax + Q(a)y + b

or z = P (a)x + ay + b where a, b are arbitrary constants and P and Q are functions.
Solution. f (p, q) = 0, the equation does not involve x, y and z explicitly. Hence fx = fy =
fz = 0.
dx dy dz dp dq
= = =− =−
fp fq p f p + q fq (fx + pfz ) (fy + qfz )
dx dy dz dp dq
= = =− =−
fp fq p f p + q fq (0 + p ∗ 0) (0 + q ∗ 0)
dx dy dz dp dq
= = = =
fp fq p f p + q fq 0 0
dp dq
Solving = , we get p = a (or q = a)
0 0
Substituting p = a (or q = a) in the given partial differential equation f (p, q) = 0, we get
f (a, q) = 0 (or f (p, a) = 0).
Using f (a, q) = 0, we get q = Q(a) for some function Q (or f (p, a) = 0 =⇒ p = P (a))
Hence the Pfaffian differential equation dz = p dx + q dy becomes dz = a dx + Q(a) dy (or
dz = P (a) dx + a dy).
Hence z = ax + Q(a) y + b (or z = P (a)x + a y + b) is a complete integral of f (p, q) = 0.
Example 2.2.8. Find a complete integral of each of the following partial differential equation
by Charpit’s method.
(1) q = 3p2
(2) p + q − pq = 0
(3) pq = c where c is a constant.
(4) p2 + q 2 = m2 where m is a constant.
(5) z 2 p2 y + 6zpxy + 2zqx2 + 4x2 y = 0 (Hint Divide by x2 y and put x dx = dX, y dy =

dY, z dz = dZ.)
1 1 1
(6) x2 p2 + y 2 q 2 = z (Hint dx = dX, dy = dY, √ dz = dZ).
x y z
1 1 1
(7) xp2 + yq 2 = 1 (Hint √ dx = dX, √ dy = dY, √ dz = dZ).
x y z
(8) x2 p2 + y 2 q 2 = z 2
1 1 1
(9) z 2 = pqxy (Hint dx = dX, dy = dY, dz = dZ)
x y z
Solution. (1) q = 3p2 . We write this as f (x, y, z, p, q) = 3p2 − q = 0.

Method 1
Since the equation is in the form f (p, q) = 0, we get p = a for some constant a and a
complete integral is in the form z = ax + Q(a)y + b.
We will find Q(a).
p = a. So, q = 3p2 =⇒ q = 3a2 .
So the Pfaffian differential equation dz = p dx + q dy becomes dz = a dx + 3a2 dy.
Integrating we get, z = ax + 3a2 y + b.
This is a complete integral of the equation q = 3p2 .
We can write the solution as z = F (x, y; a, b) = ax + 3a2 y + b.
method 2 We find all the values required in Charpit’s Auxiliary equation.
dx dy dz dp dq
= = =− =−
6p (−1) p(6p) + q(−1) 0 + p(0) 0 + q(0)
6p −1 6p2 − q 0 0
dx dy dz dp dq
= = 2 =− =−
6p −1 6p − q 0 0
dp = 0 =⇒ p = a. So, q = 3p2 =⇒ q = 3a2 .

So the Pfaffian differential equation dz = p dx + q dy becomes dz = a dx + 3a2 dy.
Integrating we get, z = ax + 3a2 y + b.
This is a complete integral of the equation q = 3p2 .
We can write the solution as z = F (x, y; a, b) = ax + 3a2 y + b.
(2) p + q − pq = 0. We write this as f (x, y, z, p, q) = p + q − pq = 0.

method 1
Since the equation is in the form f (p, q) = 0, we get p = a and a complete integral is
in the form z = ax + Q(a)y + b.
We will find Q(a).
Put p = a in q = 3p2 . So, q = 3a2 .
a
So the Pfaffian differential equation dz = p dx + q dy becomes dz = a dx + dy.
a−1
a
Integrating we get, z = ax + y + b.
a−1
This is a complete integral of the equation p + q − pq = 0.

a
We can write the solution as z = F (x, y; a, b) = ax + y + b.
a−1
method 2
dx dy dz dp dq
= = =− =−
1−q 1−p p(1 − q) + q(1 − p) 0 + p(0) 0 + q(0)
1−q 1−p p + q − 2pq 0 0
dx dy dz dp dq
= = =− =−
1−q 1−p p + q − 2pq 0 0
a
dp = 0 =⇒ p = a. So, a + q − aq = 0 =⇒ q = .
a−1
a
a−1
a
a−1
This is a complete integral of the equation p + q − pq = 0.
a
a−1
(3) pq = c We write this as f (x, y, z, p, q) = pq − c = 0.
This is in the form f (p, q) = 0.
We get, p = a and a complete integral is of the form z = ax + Q(a)y + b.
We will find Q(a).
Put p = a in pq − c = 0.
c
This implies q = .
a
c
a
c
a
This is a complete integral of the equation pq = c.
c
a
Method 2

dx dy dz dp dq
= = =− =−
q p p(q) + q(p) 0 + p(0) 0 + q(0)
q p 2pq 0 0
dx dy dz dp dq
= = =− =−
q p 2pq 0 0
c
dp = 0 =⇒ p = a. So, aq − c = 0 =⇒ q = .
a
c
a
c
a
This is a complete integral of the equation pq = c.
c
a
(4) p2 + q 2 = m2 . We write this as f (x, y, z, p, q) = p2 + q 2 − m = 0.
This is of the form f (p, q) = 0.
So, p = a and a complete integral is z = ax + Q(a)y + b.
We will √find Q(a) by putting p = a in p2 + q 2 = m2 .
So, q = m2 − a2 . √
So the Pfaffian differential equation
√ dz = p dx+q dy becomes dz = a dx+ m2 − a2 dy.
Integrating we get, z = ax + m2 − a2 y + b.
This is a complete integral of the equation p2 + q 2 = m2 . √
We can write the solution as z = F (x, y; a, b) = z = ax + m2 − a2 y + b.
Method 2
dx dy dz dp dq
= = =− =−
2p 2q p(2p) + q(2q) 0 + p(0) 0 + q(0)
2p 2q 2(p2 + q 2 ) 0 0
dx dy dz dp dq
= = =− =−
2p 2q 2(p2 + q 2 ) 0 0
√
dp = 0 =⇒ p = a. So, a2 + q 2 = m2 =⇒ q = m2 − a2 . √
So the Pfaffian differential equation
√ dz = p dx+q dy becomes dz = a dx+ m2 − a2 dy.
2
Integrating we get, z = ax + m − a y + b.2
This is a complete integral of the equation p2 + q 2 = √ m2 .

We can write the solution as z = F (x, y; a, b) = ax + m2 − a2 y + b.
(5) z 2 p2 y + 6zpxy + 2zqx2 + 4x2 y = 0
z2 z z
Dividing by x2 y, we get 2 p2 + 6 p + 2 q + 4 = 0.
x x y
z ∂z 2

z ∂z z ∂z
We will write this equation as +6 +2 +4=0 (1) .
x ∂x x ∂x y ∂y
Put x dx = dX, y dy = dY, z dz = dZ. (2)
x2 y2 z2
=⇒ = X, = Y, =Z (3)
2 2 2
Using (2), equation (1) becomes,
∂Z 2

∂Z ∂Z
+6 +2 + 4 = 0.
∂X ∂X ∂Y
=⇒ P 2 + 6P + 2Q + 4 = 0 (∗)
∂Z ∂Z
where P = ,Q = .
∂X ∂Y
This is of the form f (P, Q) = 0.
So, P = a and a complete integral is of the form Z = aX + H(a) y + b.
To find H(a) we have to first find Q. We substitute P = a in P 2 + 6P + 2Q + 4 = 0.
This implies a2 + 6a + 2Q + 4 = 0.
a2 + 6a + 4
=⇒ Q = − .
2
a2 + 6a + 4
The corresponding Pfaffian differential equation is dZ = a dX − dY .
2
2
a + 6a + 4
Hence a complete integral of (∗) is Z = aX − Y + b.
2
x2 y2 z2
We substitute X = ,Y = ,Z = .
2 2 2
z2 x2 a2 + 6a + 4 y 2
A complete integral of the given partial differential equation is =a − +
2 2 2 2
b.
a2 + 6a + 4 2
That is, z 2 = ax2 − y + 2b.
2
1 1 1
(6) x2 p2 + y 2 q 2 = z (Hint dx = dX, dy = dY, √ dz = dZ).
x y z
Dividing both sides by z, we get
x2 ∂z 2 y 2 ∂z 2

+ =1
z ∂x z ∂y
2 2 2 2
x ∂z y ∂z
√ + √ =1 (1)
z ∂x z ∂y
2
y ∂z 2

x ∂z
√ + √ =1
z ∂x z ∂y

1
Put x1 dx = dX, y1 dy = dY and √

dz = dZ ‘(2).
z
√
ln x = X, ln y = Y, 2 z = Z (3)
Substituting (2) in (1), we get

2 2
∂Z ∂Z
+ =1
∂X ∂Y
∂Z ∂Z
P 2 + Q2 = 1 where = P, =Q (∗)
∂X ∂Y
So, we get P = a and a complete integral of (∗) is Z = aX + H(a)Y + b.
To find H(a), we first find Q.
Substitute P = a in (∗). √
a2 + Q2 = 1. This implies Q = 1 − a2 . √
The corresponding Pfaffian equation
√ is dZ = a dX + 1 − a2 dY .
Integrating, we get Z = aX + 1 − a2 Y .
√
Hence a complete integral of the given partial differential equation is 2 z = a ln x +
√
1 − a2 ln y + b.
(7) xp2 + yq 2 = z.
Dividing both sides by z, we get
x ∂z 2 y ∂z 2

+ =1
z ∂x z ∂y
√ 2 2 √ 2 2
x ∂z y ∂z
√ + √ =1 (1)
z ∂x z ∂y
√ 2 √
y ∂z 2

x ∂z
√ + √ =1
z ∂x z ∂y

1

1 1
Put √x dx = dX, √y dy = dY and √ dz = dZ ‘(2).
z
√ √ √
2 x = X, 2 y = Y, 2 z = Z (3)

2 2
∂Z ∂Z
+ =1
∂X ∂Y
∂Z ∂Z
P 2 + Q2 = 1 where = P, =Q (∗)
∂X ∂Y

√ is dZ = a dX + 1 − a2 dY .
Integrating, we get Z = aX + 1 − a Y . 2
√ √
Hence
√ a complete integral of the given partial differential equation is 2 z = a2 x +
√
1 − a2 2 y + b.
(8) x2 p2 + y 2 q 2 = z 2 .
Dividing both sides by z 2 , we get
x2 ∂z 2 y 2 ∂z 2

+ =1
z 2 ∂x z 2 ∂y
x 2 ∂z 2 y 2 ∂z 2
+ =1 (1)
z ∂x z ∂y
2
y ∂z 2

x ∂z
+ =1
z ∂x
z ∂y

1

1 1
Put x dx = dX, y dy = dY and dz = dZ ‘(2).
z
ln x = X, ln y = Y, ln z = Z (3)

2 2
∂Z ∂Z
+ =1
∂X ∂Y
∂Z ∂Z
P 2 + Q2 = 1 where = P, =Q (∗)
∂X ∂Y
√ is dZ = a dX + 1 − a2 dY .
Integrating, we get Z = aX + 1 − a2 Y .
Hence a complete integral of the given partial differential equation is ln z = a ln x +
√
1 − a2 ln y + b.
xy ∂z ∂z
(9) z 2 = pqxy. We can re-write this as 2 = 1.
z ∂x ∂y
x ∂z y ∂z
That is =1 (1).
z ∂x z ∂y
1 1 1
Put dx = dX, dy = dY, dz = dZ. (2)
x y z
That is, ln x = X, ln y = Y, ln z = Z (3).
∂Z ∂Z
Using (2), equation (1) becomes, = 1.
∂X ∂Y
∂Z ∂Z
So, P Q = 1 (∗) where P = ,Q = .
∂X ∂Y
So, we get, P = a and a complete integral of (∗) is Z = aX + H(a)Y + b.
1
We put P = a in (∗). This gives Q = .
a
1
Hence the Pfaffian differential equation is dZ = a dX + dY .
a
1
Integrating both sides, Z = aX + Y + b.
a
1
Hence ln z = a ln x + ln y + b is a complete integral of z 2 = xypq.
a
1 ∂z 1 ∂z
(10) pq = 4xy. This means = 1.
2x∂x 2y ∂y
1 ∂z 1 ∂z
That is =1 (1).
2x ∂x 2y ∂y
Put 2x dx = dX, 2y dy = dY, 2z dz = dZ. (2)
2 2
That is, x = X, y = Y, z = Z 2 (3).
∂Z ∂Z
Using (2), equation (1) becomes, = 1.
∂X ∂Y
∂Z ∂Z
So, P Q = 1 (∗) where P = ,Q = .
∂X ∂Y
So, we get, P = a and a complete integral of (∗) is Z = aX + H(a)Y + b.
1
We put P = a in (∗). This gives Q = .
a
1
Hence the Pfaffian differential equation is dZ = a dX + dY .
a
1
Integrating both sides, Z = aX + Y + b.
a
1 2
Hence z = ax + y + b is a complete integral of z 2 = xypq.
2 2
a

Type II
Example 2.2.9. If the partial differential equation is of the form f (z, p, q) = 0, that is, the
equation does not involve x and yexplicitly then show that the Charpit’s Auxiliary equations of
f (z, p, q) = 0 are
dx dy dz dp dq dg
= = = = .
fp fq (p fp + q fq ) −pfz −qfz 0
Also show that the PfaffianZ differential equation is dz = Q(a, z)(a dx + dy) and a complete
dz
integral of f (z, p, q) = 0 is = ax + y + b where Q is a function.
Q(a, z)
Solution. The equation f (z, p, q) = 0, does not involve x and y explicitly. Hence fx = fy =
0.
dx dy dz dp dq
= = =− =−
dx dy dz dp dq
= = =− =−
fp fq (p fp + q fq ) (0 + p ∗ fz ) (0 + q ∗ fz )
dx dy dz dp dq
= = =− =−
fp fq (p fp + q fq ) p ∗ fz q ∗ fz
dp dq dp dq
Solving − =− , we get = .
p ∗ fz q ∗ fz p q
This means ln p − ln q = ln a (or ln q − ln p = ln a ).
p q
So, = a (or = a). That is, p = qa (or q = pa)
q p
Substituting p = qa (or q = pa) in the given partial differential equation f (z, aq, q) = 0, (or
f (z, p, ap)) we get q = Q(a, z) for some function Q (or p = P (a, z) for some function P ).
Hence the Pfaffian differential equation dz = p dx + q dy becomes dz = aQ(a, z) dx +

Q(a, z)Zdy (or dz = P (a, z) dx + aP (a,Zz)dy
dz dz
Hence dz = a x + y + b (or dz = x + a y + b) is a complete integral of
Q(a, z) P (a, z)
f (z, p, q) = 0.
Example 2.2.10. Find a complete integral for each of the following partial differential equation.
(1) p + q − zpq = 0 (7) pz = 1 + q 2 (13) p2 = z 2 (1 − pq)

(2) 9(p2 z + q 2 ) = 4 (8) 1 + p2 = qz (14) q 2 = z 2 p2 (1 − p2 )
(3) p2 = qz (9) p(z + p) + q = 0
(15) p3 + q 3 = 27z
(4) z = pq (10) p(1 + q) = qz
(16) z 2 (z 2 p2 + q 2 ) = 1
(5) pq = 4z (11) p3 + q3
− 3zpq = 0
z (17) 16z 2 p2 + 9z 2 q 2 + 4(z 2 −
(6) p(1 + q2) = q(z − α) (12) p + q = 1) = 0
c
Solution. (1) p + q − zpq = 0.
Method 1
This is in the form
Z f (z, p, q) = 0.
dz
So, p = aq and = ax + y + b is a complete integral of f (z, p, q) = 0.
Q(a, z)
Substitute p = aq in p + q − zpq = 0, we get aq + q − zaqq = 0. This means
q(a + 1 − zaq) = 0
a+1
Therefore, q = 0 or q = .
az
If q = 0, p = 0 and the corresponding Pfaffian equation becomes dz = 0 and hence
z =constant. But this is not a complete integral.
a+1 a+1
If q 6= 0 then q = and hence p = aq = .
az z
So, the Pfaffian differential equation dz = p dx + q dy becomes
a+1 a+1
dz = dx + dy.
z za
a+1
=⇒ z dz = (a + 1) dx + dy.
a
On integrating, we get
z2 a+1
=⇒ = (a + 1) x + y + b.
2 a
2(a + 1)
=⇒ z 2 = (a x + y) + b.
a
Method 2

dx dy dz dp dq
= = =− =− =
1 zq 1 − zp p(1 − zp) + q(1 − zp) 0 + p(−pq) 0 + q(−pq)
1 − zq 1 − zp − p + q − z(p2 + q 2 ) p2 q − pq 2
1 − zq 1 − zp − zpq − z(p2 + q 2 ) p2 q − pq 2
dx dy dz dp dq
= = 2 2
=− 2 =− 2
1 − zq 1 − zp zpq − z(p + q ) −p q −q p
dp dq dp dq
− 2
= − 2 =⇒ = =⇒ ln p = ln q + ln c1 . So, p = aq.
−p q −q p p q
Substituting in p + q − zpq = 0, we get aq + q − z(aq)q = 0. This means q = 0 or
a + 1 − zaq = 0.
a+1
Hence q = 0 or q = .
za
If q = 0 then p = 0. So, the Pfaffian differential equation is dz = 0 and integral is
z = c. But this is not a complete integral.
a+1 a+1 a+1
Suppose q = . This means p = a = .
za za z
a+1 a+1
So the Pfaffian differential equation dz = p dx+q dy becomes dz = dx+ dy.
z za
a+1
=⇒ z dz = (a + 1) dx + dy.
a
On integrating, we get
z2 a+1
=⇒ = (a + 1) x + y + b.
2 a
2(a + 1)
=⇒ z 2 = (a x + y) + b.
a
This is a complete integral of the equation p + q = zpq.

h 1
i1
2
We can write the solution as z = F (x, y; a, b) = 2(a + 1) x + y + b. .
a
(2) 9(p2 z + q 2 ) = 4.
This is of the form f (p, q, z) = 0.
We get p = aq. Substituting this in 9(p2 z + q 2 ) − 4 = 0, 9(a2 q 2 z + q 2 ) − 4 = 0
4
=⇒ q 2 (a2 z + 1) = .
9
2 1
So, q = √ .
3 a2 z + 1
Hence the Pfaffian differential equation dz = p dx + q dy becomes
2 a 2 1
dz = √ dx + √ dy.
3 √ 2
a z+1 3 2
a z+1
That is 3 a2 z + 1 dz = 2(a dx + dy)
1
=⇒ 3(a2 z + 1) 2 dz = 2(a dx + dy)
2 3
Integrating both sides, we get 3 ∗ 2 (a2 z + 1) 2 = 2(ax + y + b).
3a
1 2 3
(a z + 1) 2 = (ax + y + b).
a2
1 2 3
(a z + 1) 2 = (ax + y + b).
a2
(3) p2 = qz.
We get p = aq. Substituting this in p2 = qz, a2 q 2 = qz.
=⇒ q(a2 q − z) = 0.
z
So, q = 0 or q = 2 .
a
2 1
So, q = √ .
3 a2 z + 1
az z
dz = 2 dx + 2 dy.
a a
z z
dz = dx + 2 dy.
a a
dz 1 1
= dx + 2 dy.
z a a
2 dz
a = a dx + dy.
z
=⇒ a2 ln z = ax + y + b.
(4) z = pq.
We get p = aq. Substituting this in z = aq 2 .
z
=⇒ q 2 = .
a
√
z
So, q = ± √ .
a
Hence the Pfaffian
√ differential equation dz = p dx + q dy becomes
√ z
dz = az dx + √ dy.
a
√ dz
a √ = a dx + dy.
z
√ 1
2 az 2 = a x + y + b.
4az = (a x + y + b)2 .
(5) pq = 4z
z
We get p = aq. Substituting this in 4z = aq 2 , we get q 2 = .
√ 4a
z
So, q = ± √ .
2 a
Hence the√Pfaffian differential
√ equation dz = p dx + q dy becomes
z z
dz = ±a √ dx ± √ dy.
2 a 2 a
√ dz
±2 a √ = a dx + dy.
z
√ 1
±2 a ∗ 2z 2 = a x + y.
√ 1
±4 az 2 = a x + y + b.
16az = (a x + y + b)2 .
(6) p(1 + q 2 ) = q(z − α)

We get p = aq. Substituting this in p(1 + q 2 ) = q(z − α), we get aq(1 + q 2 ) = q(z − α).
=⇒ q = 0 or a(1 + q 2 ) = (z − α).
q = 0 =⇒ p = 0 and a solution is z = c. This is not a complete integral.
z−α
Suppose q 6= 0. Then a(1 + q 2 ) = (z − α). This implies q 2 = − 1.
√ a
z−α−a z−α−a
That is q 2 = . Hence q = ± √ .
a a
Thus the√ Pfaffian differential
√ equation dz = p dx + q dy becomes
z−α−a z−α−a
dz = a √ dx + √ dy.
a a
√ dz
a√ = a dx + dy.
z−α−a
Integrating√ both sides, we have
√
a ∗ 2 ∗ z − α − a = a x + y.
4a(z − α − a) = (a x + y + b)2 .
(7) pz = 1 + q 2
We get p = aq. Substituting this in pz = 1 + q 2 , we get aqz = 1 + q 2 .
This means q 2√− azq + 1 = 0
az ± a2 z 2 − 4
=⇒ q = .
2
Thus the Pfaffian
√ differential equation
√ dz = p dx + q dy becomes
az ± a2 z 2 − 4 az ± a2 z 2 − 4
dz = a dx + dy.
2 2
2dz
√ = a dx + dy.
az ± a2 z 2 − 4
√
2(az ∓ a2 z 2 − 4)
dx = a dx + dy.
√ 4
2
az ∓ a z − 42
dx = a dx + dy.
p2
az ∓ (a2 z 2 − 4) dz = 2(a dx + dy).
Integrating both sides, we have

" #
az 2 1 az p 2 2 4 p
∓ (a z − 4) − ln az + a2 z 2 − 4 = 2(ax + y) + b

2 a 2 2
" #
az 2 1 p p
∓ az (a2 z 2 − 4) − 4 ln az + a2 z 2 − 4 = 2(ax + y) + b

2 a
" #
p p
a2 z 2 ∓ az (a2 z 2 − 4) − 4 ln az + (a2 z 2 − 4) = 4a(ax + y) + ab

(8) 1 + p2 = qz
We get p = aq. Substituting this in qz = 1 + a2 q 2 , we get qz = 1 + a2 q 2 .
This means a2 q 2 − zq + 1 = 0
√
z± z 2 − 4a2
=⇒ q = .
2a2
Thus the Pfaffian differential equation dz = p dx + q dy becomes

√ √
z± z 2 − 4a2 z ± z 2 − 4a
dz = a dx + dy.
2a2 2a2

dz
2a2 √ = a dx + dy.
z ± z 2 − 4a2
√ !
2 z ∓ z 2 − 4a2
2a dz = a dx + dy.
4a2
√
z ∓ z 2 − 4a2
dz = a dx + dy.
2
√
z∓ z 2 − 4a2 dz = 2(a dx + dy).
Integrating both sides, we have

" #
z2 zp 2 2
4a2 p
2 2

∓ z − 4a − ln z + z − 4a = 2(ax + y) + b

2 2 2
" #
√ √
z 2 ∓ z z 2 − 4a2 − 4a2 ln z + z 2 − 4a2 = 4(ax + y) + b

(9) p(z + p) + q = 0
We get p = aq. Substituting this in p(z + q) + q = 0.
=⇒ aq(z + q) + q = 0. This means aq 2 + (az + 1)q = 0
az + 1
So, q = 0 or q = − .
a
az + 1 az + 1
dz = −a dx − dy.
a a
a
dz = −a dx − dy.
az + 1
ln |az + 1| = −ax − y + b.
(10) p(1 + q) = qz
We get p = aq. Substituting this in p(1 + q) − qz = 0.
=⇒ aq(1 + q) − qz = 0. This means aq 2 + (a − z)q = 0
a−z z−a
So, q = 0 or q = − = .
a a
z−a z−a
dz = a dx + dy.
a a
a
dz = a dx + dy.
z−a
a ln |z − a| = ax + y + b.
(11) p3 + q 3 − 3zpq = 0
We get p = aq. Substituting this in a3 q 3 + q 3 − 3azq 2 = 0.
3az
=⇒ q 2 (q(a3 + 1) − 3az) = 0. So, q = 0 or q = 3 .
a +1

3az 3az
dz = a dx + 3 dy.
a3+1 a +1
a3 + 1
dz = a dx + dy.
3az
a3 + 1
ln z = ax + y + b.
3a
3
(a + 1) ln z = 3a(ax + y + b).
z
(12) p + q =
c
z
We get p = aq. Substituting this in p + q =
c
z
=⇒ aq + q = .
c
z
=⇒ q = .
c(a + 1)
z z
dz = a dx + dy.
c(a + 1) c(a + 1)
c(a + 1)
dz = a dx + dy.
z
c(a + 1) ln z = ax + y + b.
(13) p2 = z 2 (1 − pq)
We get p = aq. Substituting this in p2 = z 2 (1 − pq)
=⇒ a2 q 2 = z 2 (1 − aq 2 ).
=⇒ q 2 (a2 + az 2 ) = z 2 .
z
=⇒ q = ± p .
a(a + z 2 )
az z
dz = ± p dx ± p dy.
2
a(a + z ) a(a + z 2 )
√
√ a + z2
a dz = ±a dx ± dy (∗).
z Z √
a + z2
We will find dz.
z
√
Put z = a tan θ
Z √ Z √
a + z2 a + a tan2 θ
dz = √ sec2 θ dθ
z a tan θ
Z √
1 + tan2 θ
= sec2 θ dθ
tan θ
Z
sec θ
= sec2 θ dθ
tan θ
Z
sec θ
= (1 + tan2 θ) dθ
tan θ
1 + tan2 θ
Z
= sec θ dθ
tan θ
Z
1
= sec θ + tan θ dθ
tan θ
Z
1 cos θ
= + tan θ dθ
cos θ sin θ
Z 1 1
= + tan θ dθ
sin θ cos θ
Z
= cosec θ + sec θ tan θ dθ
= ln |cosec θ − cot θ| + sec θ + c
√ sin θ z
a tan θ = z =⇒ =√ .
cos θ a
z 1
This means sin θ = √ and cos θ = √ .
a + z2 a + z2
Hence
Z √
a + z2
dz = ln |cosec θ − cot θ| + sec θ + c
z
Z √ √
a + z 2 √a r

a + z2 z2
=⇒ dz = ln − + 1+

z z z a
√ √ √
a + z2 a a + z2
= ln − + √

z z a
√ √ √
a + z2 − a a + z2
= ln + √

z a
Hence, integrating (∗), we get

√
a + z 2 − √a √a + z 2
!
√
a ln + √ = ax + y + b

z a
√
a + z 2 − √a p

√
a ln + a + z 2 = ax + y + b

z
(14) q 2 = z 2 p2 (1 − p2 )
We get q = ap. (Here we will use q = ap instead of p = aq)
Substituting this in q 2 = z 2 p2 (1 − p2 ), we get a2 p2 = z 2 p2 (1 − p2 ).
=⇒ p2 = 0 or a2 − z 2 (1 − p2 ) = 0
p2 = 0 =⇒ p = 0, q = 0 and z = c.
a2
Consider a2 − z 2 (1 − p2 ) = 0. This means p2 = 1 − 2
√ z
z 2 − a2
=⇒ p = .
z √
a z 2 − a2
So, q = 0 or q = .
z

√ √
z 2 − a2 a z 2 − a2
dz = dx + dy.
z z
z
√ dz = dx + a dy (∗).
z 2 − a2
√ z
Put t = z 2 − a2 . So, dt = √ dz.
z 2 − a2
Hence dt = dx + a dy.
tp= x + ay + b.
z 2 − a2 | = x + ay + b.
=⇒ z 2 − a2 = (x + ay + b)2
(15) p3 + q 3 = 27z.
We get p = aq.
Substituting this in p3 + q 3 = 27z, we get a3 q 3 + q 3 = 27z.
=⇒ q 3 (1 + a3 ) = 27z.
1
3z 3
This means q = 1
(1 + a3 ) 3
1 1
3az 3 3z 3
dz = 1 dx + 1 dy.
(1 + a3 ) 3 (1 + a3 ) 3
1 3a 3
1 dz = 1 dx + 1 dy (∗).
z3 3
(1 + a ) 3 (1 + a3 ) 3
Integrating both sides,
3 2 3a
z3 = 1 (ax + y + b).
2 (1 + a3 ) 3
8a3
z2 = (ax + y + b)3 .
(1 + a3 )
(16) z 2 (z 2 p2 + q 2 ) = 1.
We get q = ap. (Here we will use q = ap instead of p = aq)

Substituting this in z 2 (p2 z 2 + q 2 ) = 1, we get z 2 (p2 z 2 + a2 p2 ) = 1.
1 1
=⇒ p2 = 2 2 =⇒ p = ± √ .
z (z + a2 ) z z + a2
2
1 a
dz = ± √ dx + ± √ dy.
2
z z +a 2 z z + a2
2
√
z z 2 + a2 dz = ± dx ± a dy (∗).
Put t = z 2 + a2 . So, dt = 2z dz.

dt
Hence = z dz.
2
√ dt
Substituting in (∗), t = ± dx ± a dy (∗).
3
2
2 2
3 t = ±2(x + ay + b).
3
t 2 = ±3(x + ay + b).
t3 = 9(x + ay + b)2 .
(z 2 + a2 )3 = 9(x + ay + b)2 .

Type III h(x, p) = k(y, q).
Example 2.2.11. If the given partial differential equation

Z is in the formZh(x, p) = k(y, q),
then the equation has a complete integral of the form z = P (a, x) dx + Q(a, y) dy + b.
Solution. h(x, p) = k(y, q).

In this case f = h(x, p) − k(y, q).
fp = hp , fq = −kq , pfp + qfq = php − qkq , fx + pfz = hx + p ∗ 0, fy + qfz = −ky + q ∗ 0.
dx dy dz dp dq
= = =− =−
hp −kq php − qkq hx −ky
dx dy dz dp dq
=− = =− =
hp kq php − qkq hx ky
dx dp
=−
hp hx
hx dx + hp dp = 0
=⇒ dh = hx dx + hp dp = 0
Hence
h(x, p) = a
So,
h(x, p) = k(y, q) =⇒ k(y, q) = a
h(x, p) = a =⇒ p = P (a, x)
And
k(y, q) = a =⇒ q = Q(a, y)
Hence the Pfaffian equationZ dz = p dx + q Zdy becomes dz = P (a, x) dx + Q(a, y) dy.
A complete integral is z = P (a, x) dx + Q(a, y) dy + b.
Examples 2.2.12. [4] Find a complete integral for each of the following partial differential
equation.
(1) p2 + q 2 = x + y (6) z 2 (p2 + q 2 ) = x2 + y 2 (Hint z dz = dZ)

√
(2) x(1 + y)p = y(1 + x)q (7) z(p2 − q 2 ) = x − y (Hint z dz = dZ)
1
(3) p − 3x2 = q 2 − y (8) p2 + q 2 = z 2 (x + y) (Hint: dz = dZ)
z
(4) q = px + p2 (9) yzp2 = q (Hint z dz = dZ)
(5) py + qx + pq = 0 (10) p2 q 2 + x2 y 2 = x2 q 2 (x2 + y 2 )
Solution. (1) p2 + q 2 = x + y
We can write this as p2 − x = y − q 2 .
This is of the form h(x, p) = k(y, q).
We get, h(x, p) = a and k(y, q) = a.
Therefore,√ p2 − x = a and y − q 2 = a.
√
=⇒ p = ± a + x, q = ± y + a. √ √
Pfaffian differential equation is dz = a + x dx + y + a dy.
2h 3 3
i
A complete integral is z = (a + x) 2 + (a + y) 2 + b.
3
x y
(2) x(1 + y)p = y(1 + x)q. We write this as p= q
1+x 1+q
x y
Therefore, p = a and q = a.
1+x 1+y
a(1 + x) a(1 + y)
=⇒ p = ,q = .
x y
a(1 + x) a(1 + y)
Pfaffian differential equation is dz = dx + dy.
x y
a(1 + x) a(1 + y)
dz = dx + dy
x y

1 1
=a + 1 dx + a + 1 dy
x y
Integrating both sides, we get
z = a (ln x + x + ln y + y) + b
z = a (ln xy + x + y) + b
So, z = a (ln xy + x + y) + b is a complete integral of x(1 + y)p = y(1 + x)q.
(3) p − 3x2 = q 2 − y.
Therefore, p − 3x2 = a and q 2 − y = a.
√
=⇒ p = a + 3x2 , q = ± a + y.
√
Pfaffian differential equation is dz = a + 3x2 dx + a + y dy.
2 3
z = ax + x3 + (a + y) 2 + b
3
2 3
So, z = ax + x3 + (a + y) 2 + b is a complete integral of p − 3x2 = q 2 − y.
3
(4) q = px + p2 . We re-write this equation as px + p2 = q.
Therefore, px + p2 = a and q = a. √
−x ± x2 + 4a
Now, p2 + xp − a = 0 =⇒ p =
2
√
−x ± x2 + 4a
Pfaffian differential equation is dz = dx + a dy.
2
√
−x ± x2 + 4a
dz = dx + a dy
2 √
x x2 + 4a
= − dx + dx + a dy
2 p 2 √
x x2 + (2 a)2
= − dx + dx + a dy
2 2
" q √ 2 #
x2 1 x √ √

(2 a) q
z=− + x2 + (2 a)2 + ln x + x2 + (2 a)2 + ay + b
4 2 2 2
Formula used:
a2
Z p
xp 2 p
x2 + a2 dx = x + a2 + ln |x + x2 + a2 | + c
2 2
" #
x2 1 x p 2 (4a p
z=− + x + 4a + ln x + x2 + 4a + ay + b

4 2 2 2
" #
x2 1 x p 2 p
z=− + x + 4a + 2a ln x + x2 + 4a + ay + b

4 2 2
" #
x2 1 x p 2 p
So, z = − + x + 4a + 2a ln x + x2 + 4a + ay + b is a complete integral

4 2 2
of px + p2 = q.
p q
(5) py + qx + pq = 0. We re-write this as p(y + q) = −qx. That is = − .
x (y + q)
p q
Therefore, = a and − = a.
x (y + q)
ay
=⇒ p = ax and −q = a(y + q). So, q(a + 1) = −ay and hence q = − .
(a + 1)
ay
Pfaffian differential equation is dz = ax dx − dy.
(a + 1)
ay
dz = ax dx − dy
(a + 1)
(a + 1) dz = a(a + 1)x dx − ay dy.
x2 y2
(a + 1)z = a(a + 1) −a +b
2 2
2(a + 1)z = a(a + 1)x2 − ay 2 + b

So, 2(a + 1)z = a(a + 1)x2 − ay 2 + b is a complete integral of py + qx + pq = 0.
Second complete solution

p q
py + qx + pq = 0. We re-write this as py = −qx − pq. That is =− .
x+p y
p q
Therefore, = a and − = a.
x+p y
a
=⇒ p = a(x + p). This means p(1 − a) = ax. Hence p = x.
1−a
q
− = a =⇒ q = −ay.
y
ay
Pfaffian differential equation is dz = ax dx − dy.
(a + 1)
a
dz = x dx − ay dy
1−a
a x2 y2
z= − a + b.
1−a 2 2
a
2z = x2 − ay 2 + 2b
1−a
a
2z = x2 − ay 2 + b.
1−a
a
So, 2z = x2 − ay 2 + b is a complete integral of py + qx + pq = 0.
1−a
(6) z 2 (p2 + q 2 ) = x2 + y 2 (Hint z dz = dZ)

∂z 2 ∂z 2

We re-write this as z + z = x2 + y 2
∂x ∂y
Put z dz = dZ.
∂Z 2 ∂Z 2

This means + = x2 + y 2 .
∂x ∂y
=⇒ P 2 + Q2 = x2 + y 2 .
That is P 2 − x2 = y 2 − Q2 .
This is of the form h(x, P ) = k(y, Q).
We get, h(x, P ) = a2 and k(y, Q) = a2 .
(We will take the arbitrary constant as a2 instead of a)
Therefore, √ P 2 − x2 = a2 and y 2 p
− Q2 = a 2 .
=⇒ P = ± x2 + a2 and Q = ± y 2 − a2 .
√ p
Pfaffian differential equation is dZ = x 2 + a2 dx + y 2 − a2 dy.
√ p
2 2
That is z dz = x + a dx + y − a dy. 2 2
z2 xp 2 a2 p yp 2 a2 p
= x + a2 + ln |x + x2 + a2 | + y − a2 + ln |y + y 2 − a2 |.
2 2 2 2 2
√ √ p p
So, z 2 = x x2 + a2 + a2 ln |x + x2 + a2 | + y y 2 − a2 + a2 ln |y + y 2 − a2 | is a
complete integral of z 2 (p2 + q 2 ) = x2 + y 2 .
Formulae used:
a2
Z p
xp 2 p
x2 + a2 = x + a2 + ln |x + x2 + a2 |
2 2
Z p
x p a 2 p
x2 − a2 = x2 − a2 + ln |x + x2 − a2 |.
2 2
√
(7) z(p2 − q 2 ) = x − y (Hint z dz = dZ)
√ ∂z 2 √ ∂z 2

We re-write this as z − z =x−y
∂x ∂y
√ 2 3
Put z dz = dZ. This means z 2 = Z.
3
∂Z 2 ∂Z 2

Now, − = x − y.
∂x ∂y
=⇒ P 2 − Q2 = x − y.
That is P 2 − x = Q2 − y.
We get, h(x, P ) = a and k(y, Q) = a2 .
Therefore, √ P 2 − x = a and Q2 − y = a.
√
=⇒ P = ± x + a and Q = ± y + a.
√ √
Pfaffian differential
√ equation is dZ = x + a dx + y + a dy.
√ √
That is z dz = x + a dx + y + a dy.

3
2z 2 2h 3 3
i
= (x + a) 2 + (y + a) 2 + b
3
3
3 3 3
z 2 = (x + a) 2 + (y + a) 2 + b
1
(8) p2 + q 2 = z 2 (x + y) (Hint: dz = dZ)
z
2
1 ∂z 2

1 ∂z
We re-write this as − =x−y
z ∂x z ∂y
1
Put dz = dZ. This means ln z = Z.
z
∂Z 2 ∂Z 2

Now, + = x + y.
∂x ∂y
=⇒ P 2 + Q2 = x + y.
That is P 2 − x = y − Q2 .
We get, h(x, P ) = P 2 − x = a and k(y, Q) = y − Q2 = a.
Therefore, √P 2 = a + x and Q2 = y − a.
√
=⇒ P = ± x + a and Q = ± y − a.
√ √
Pfaffian differential equation is dZ = x + a dx + y − a dy.
1 √ √
That is dz = x + a dx + y − a dy.
z
2h 3 3
i
ln z = (x + a) + (y − a) + b
2 2
3
(9) yzp2 = q (Hint z dz = dZ)

=⇒ yz 2 p2 = zq
∂z 2

∂z
We re-write this as y z =z
∂x ∂y
z2
Put z dz = dZ. This means = Z.
2 2
∂Z ∂Z
Now, y = .
∂x ∂y
Q
=⇒ P 2 = .
y

Q
We get, h(x, P ) = P 2 and k(y, Q) = .
y
(We will take the constant as a2 instead of a)
Q
Therefore, P 2 = a2 and = a2 .
y
=⇒ P = a and Q = ay.
Pfaffian differential equation is dZ = a dx + a2 y dy.

That is z dz = a dx + a2 y dy.
z2 a2 y 2 b
=a x+ +
2 2 2
b
(We will take the constant of integration as instead of b so that we can cancel 2 in
2
the denominator). z 2 = 2a x + a2 y 2 + b
(10) p2 q 2 + x2 y 2 = x2 q 2 (x2 + y 2 )
p 2 q 2 + x2 y 2
= x2 + y 2
x2 q 2
p2 y2
+ = x2 + y 2
x2 q2
p2 2 2 y2
− x = y −
x2 q2

p2 y2
We get, h(x, p) = 2 − x2 = a and k(y, q) = y 2 − 2 = a.
x q
√ y
2 2
Therefore, p = ± x a + x and q = ± p .
y2 − a
√ √
Pfaffian differential equation is dz = x + a dx + y − a dy.
√ y
That is dz = x a2 + x2 dx + p dy.
y2 − a
Z p Z √ Z Z
dt y
x a2 + x2 dx = t p dy = dt
2 y2 − a
2 3 1 = tp
= t2
3 2 = y2 − a
1 3
= (a2 + x2 ) 2
3
1 3 p
z = (a2 + x2 ) 2 + y 2 − a + b
3
(11) p2 − y 2 q − y 2 + x2 = 0
We re-write this as p2 + x2 = y 2 q + y 2
We get, h(x, p) = p2 + x2 = a2 and k(y, q) = y 2 q + y 2 = a2 .
(We will take the constant as a2 instead of a)
√ a2 − y 2
Therefore, p = ± a2 − x2 and q = .
y2
√ a2 − y 2
Pfaffian differential equation is dz = a2 − x2 dx + dy.
y2
√
2
a
That is dz = a2 − x2 dx + 2 − 1 dy.
y

xp 2 a2 x a2
z= a − x2 + sin−1 − −y+b
2 2 a y

Type IV: [3] Clairaut Equation Method
Example 2.2.13. If the given partial differential equation is of the form z = xp + yq + h(p, q)
then show that z = ax + by + h(a, b) is its complete integral.
Solution. We will re-write the partial differential z = xp+yq+h(p, q) as xp+yq+h(p, q)−z =

0 (∗).
dx dy dz dp dq
= = =− =−
x + hp y + hq p(x + hp ) + q(y + hq ) − p + p(−1) q + q(−1)
x + hp y + hq xp + yq + php + qhq −0 0
dx dy dz dp dq
= = =− =−
x + hp y + hq xp + yq + php + qhq 0 0
Then the fourth and fifth fractions imply dp = 0 and dq = 0. This means p = a and q = b.
Substituting these values in (∗), z = a x + b y + h(a, b)
Hence z = ax + by + h(a, b) is a complete integral of z = x p + y q + h(p, q).

Examples 2.2.14. [4] Find a complete integral of each of the following partial differential
equation.
√
(1) z = px + qy + pq (5) z = px + qy − 2 pq
(2) (px + qy − z)2 = 1 + p2 + q 2 (6) (p + q)(z − px − qy) = 1
p (7) pqz = p2 (xq + p2 ) + q 2 (yp + q 2 )

(3) z = px + qy + c 1 + p2 + q 2
(8) Find a complete integral of
(4) z = px + qy + ln(pq) 2q(z − px − qy) = 1 = q 2
Examples 2.2.15. [4]

(1) Prove that a complete integral of z = px + qy − 2p − 3q represents a family of planes.

Also show that each member of the family passes through (2, 3, 0).

pq
(2) Prove that a complete integral of z = px + qy + represents a family
pq − p − q
of planes. Also show that for each member of the family, the algebraic sum of its
intercepts on the three coordinate axes is unity.
p
(3) Show that a complete integral of the equation z = px + qy + p2 + q 2 + 1 represents
all planes that are at unit distance form the origin.
Solution. (1) Prove that a complete integral of z = px + qy − 2p − 3q represents a family
of planes. Also show that each member of the family passes through (2, 3, 0).
z = px + qy − 2p − 3q is of the form z = px + qy + h(p, q) (Clairaut’s equation)

Hence z = pa + qa + h(a, b) is its complete integral.
So, z = ax + by − 2a − 3b is a complete integral of z = px + qy − 2p − 3q.
Since the equation z = ax + by − 2a − 3b is linear in x, y and z, the equation represents
a plane.
Since a and b are parameters, z = ax + by − 2a − 3b represents a family of planes.
We will show that each plane from the family z = ax + by − 2a − 3b passes through
(2, 3, 0).
We substitute (2, 3, 0) in the equation z = ax + by − 2a − 3b
L.H.S. = 0
R.H.S. = a(2) + b(3) − 2a − 3b
=0
Hence each member of the family z = ax + by − 2a − 3b passes through (2, 3, 0).

pq
(2) Prove that a complete integral of z = px + qy + represents a family
pq − p − q
of planes. Also show that for each member of the family, the algebraic sum of its
intercepts on three
coordinate
axes is unity.
pq
z = px + qy + is of the form z = px + qy + h(p, q) (Clairaut’s equation)
pq − p − q
Hence z = pa + qa+ h(a, b) is its complete integral.
ab pq
So, z = ax + by + is a complete integral of z = px + qy + .
ab − a − b pq − p − q
ab
Since the equation z = ax + by + is linear in x, y and z, the equation
ab − a − b
represents a plane.
ab
Since a and b are parameters, z = ax+by + represents a family of planes.
ab − a − b
We will find the intercepts on the three coordinateaxes of a member of this family.
ab ab
We re-write the equation z = ax + by + as ax + by − z = −
ab − a − b ab − a − b
We write the equation in the following form.
x y z
ab
+ ab
− ab
=1
− a(ab−a−b) − b(ab−a−b) − (ab−a−b)
2.3 Jacobi’s Method 113
x y z
b
+ a − ab
=1
− (ab−a−b) − (ab−a−b) − (ab−a−b)
b a
The plane makes x−intercept = − ab−a−b , y−intercept = − ab−a−b and z−intercept
ab
= ab−a−b .
Algebraic sum of all the three intercepts is
b a ab
− − +
ab − a − b ab − a − b ab − a − b
−b − a + ab
=1
ab − a − b
p
(3) Show that a complete integral of the equation z = px + qy + p2 + q 2 + 1 represents
represents a family of planes. Also show that for each member of the family, is at a
unit distance form
p the origin.
z = px + qy + p2 + q 2 + 1 is of the form z = px + qy + h(p, q) (Clairaut’s equation)
Hence z = pa + qa√+ h(a, b) is its complete integral. p
So, z = ax + by + a2 + b2 + 1 is a√complete integral of z = px + qy + p2 + q 2 + 1.
Since the equation z = ax + by + a2 + b2 + 1 is linear in x, y and z, the equation
represents a plane. √
Since a and b are parameters, z = ax+by + a2 + b2 + 1 represents a family of planes.
We will find the distance of√origin from each member of the family.
Distance of z = ax + by + a2 + b2 + 1 from the origin (0, 0, 0) is given by
a(0) + b(0) − (0) + √a2 + b2 + 1

=1

p
2 2
a + b + (−1) 2
2.3 Jacobi’s Method

This method is used for solving partial differential equations involving three or more inde-
pendent variables.
• Consider the pde f (x, y, z, ux, uy , uz ) = 0 (1)

(1).
• Note: x, y and z are independent variables and the dependent variable u does not
appear in the equation.
Definition 2.3.1. u = F (x, y, z, a, b, c) is said to be a complete integral of (1) if it satisfies

the pde (1) and the associated matrix
 
Fa Fax Fay Faz
 Fb Fbx Fby Fbz 
Fc Fcx Fcy Fcz
is of rank three.
Consider the following p.d.e. f (x, y, z, ux , uy , uz ) = 0 · · · (1)

(1).
We consider two one-parameter families of partial differential equations
h1 (x, y, z, ux , uy , uz , a) = 0 · · · (2)
h2 (x, y, z, ux , uy , uz , b) = 0 · · · (3)
Then h1 = 0, h2 = 0 are said to be compatible with f = 0 if
∂(f, h1 , h2 )
6= 0, · · · (4)
∂(ux , uy , uz )
and the Pfaffian form
du = ux dx + uy dy + uz dz · · · (4)
is integrable, where ux (x, y, z, a, b), uy (x, y, z, a, b) and uz (x, y, z, a, b) are obtained by solving
(1), (2) and (3).
Remark 2.3.2. Observe that the Pfaffian differential equation du = ux dx + uy dy + uz dz

is either exact or non integrable.
Hence the conditions for (4) to be exact are curl (ux , uy , uz ) = 0. That is, uxy = uyx , uyz =
uzy , uxz = uzx .
Theorem 2.3.3. If h1 (x, y, z, ux , uy , uz , a) = 0 and h2 (x, y, z, ux , uy , uz , b) = 0 are

compatible with f (x, y, z, ux , uy , uz ) = 0 then h1 and h2 satisfy
∂(f, h) ∂(f, h) ∂(f, h)

+ + =0
∂(x, ux ) ∂(y, uy ) ∂(z, uz )
where h = hi , i = 1, 2.
Proof. Differentiate f (x, y, z, ux , uy , uz ) = 0 partially w.r.t. x, y and z.
∂f ∂f ∂f ∂f
+ uxx + uyx + uzx = 0 · · · (1)
∂x ∂ux ∂uy ∂uz
∂f ∂f ∂f ∂f
+ uxy + uyy + uzy = 0 · · · (2)
∂y ∂ux ∂uy ∂uz
∂f ∂f ∂f ∂f
+ uxz + uyz + uzz = 0 · · · (3)
∂z ∂ux ∂uy ∂uz
Now, differentiating h1 (x, y, z, ux , uy , uz ) = 0 and h2 (x, y, z, ux , uy , uz ) = 0 partially w.r.t.

x, y and z, and representing the equations in terms of h where h = h1 or h = h2 , we get
∂h1 ∂h1 ∂h1 ∂h1 ∂h1 ∂h1 ∂h1 ∂h1

+ uxx + uyx + uzx = 0 + uxz + uyz + uzz = 0
∂x ∂ux ∂uy ∂uz ∂z ∂ux ∂uy ∂uz
∂h1 ∂h1 ∂h1 ∂h1
+ uxy + uyy + uzy = 0
∂h2 ∂h2 ∂h2 ∂h2 ∂h2 ∂h2 ∂h2 ∂h2

+ uxx + uyx + uzx = 0 + uxz + uyz + uzz = 0
∂x ∂ux ∂uy ∂uz ∂z ∂ux ∂uy ∂uz
∂h2 ∂h2 ∂h2 ∂h2
+ uxy + uyy + uzy = 0
We will represent h1 and h2 by h.

∂h ∂h ∂h ∂h
+ uxx + uyx + uzx = 0 · · · (4)
∂x ∂ux ∂uy ∂uz
∂h ∂h ∂h ∂h
+ uxy + uyy + uzy = 0 · · · (5)
∂h ∂h ∂h ∂h
+ uxz + uyz + uzz = 0 · · · (6)
∂z ∂ux ∂uy ∂uz
∂h ∂f
Now, ∗ (1) − ∗ (4) gives
∂ux ∂ux

∂h ∂f ∂f ∂f ∂f ∂f ∂h1 ∂h1 ∂h1 ∂h1
+ uxx + uyx + uzx − + uxx + uyx + uzx = 0
∂ux ∂x ∂ux ∂uy ∂uz ∂ux ∂x ∂ux ∂uy ∂uz

∂h ∂h ∂h ∂h ∂h
=⇒ fx + fux uxx + fuy uyx + fuz uzx − fux + uxx + uyx + uzx = 0.
∂ux ∂x ∂ux ∂uy ∂uz
Since uxy = uyx , uxz = uzx

∂h ∂h ∂h ∂h ∂h ∂h
=⇒ fx − fu + fu − fu uxy + fu − fux uzx = 0 (∗).
∂ux ∂x x ∂ux y ∂uy x ∂ux z ∂uz
∂h ∂f
Now, ∗ (2) − ∗ (5) gives
∂uy ∂uy
∂h ∂h ∂h ∂h ∂h
=⇒ fy + fux uxy + fuy uyy + fuz uzy − fuy + uxy + uyy + uzy = 0.
∂uy ∂y ∂ux ∂uy ∂uz

∂h ∂h ∂h ∂h ∂h ∂h
=⇒ fy − fu + fu − fu uxy + fu − fuy uzy = 0 (∗∗).
∂uy ∂y y ∂uy x ∂ux y ∂uy z ∂uz
∂h ∂f
Now, ∗ (3) − ∗ (6) gives
∂uz ∂uz
∂h ∂f ∂f ∂f ∂f ∂h ∂h ∂h ∂h
+ uxz + uyz + uzz − fuz + uxz + uyz + uzz = 0.
∂uz ∂z ∂ux ∂uy ∂uz ∂z ∂ux ∂uy ∂uz

∂h ∂h ∂h ∂h ∂h ∂h
=⇒ fz − fu + fu − fu uxz + fu − fuz uyz = 0 (∗ ∗ ∗).
∂uz ∂z z ∂uz x ∂ux z ∂uz y ∂uy
Addition (∗), (∗∗) and (∗ ∗ ∗), we get,
∂h ∂h ∂h ∂h ∂h ∂h
fx − fu + fy − fu + fz − fu = 0. (7)
∂ux ∂x x ∂uy ∂y y ∂uz ∂z z
The above equation can also be written as
∂h ∂h ∂h ∂h ∂h ∂h
fux + fuy + fuz − fx − fy − fz =0
∂x ∂y ∂z ∂ux ∂uy ∂uz
Hence the auxiliary equations for h1 and h2 are
dx dy dz dux duy duz

= = =− =− =−
fux fuy fuz fx fy fz
Example 2.3.4. By Jacobi’s method, solve the equation
z 2 + zuz − u2x − u2y = 0
Solution. .
STEP 1 z 2 + zuz − u2x − u2y = 0 is of the form f (x, y, z, ux , uy , uz ) = 0.
f (x, y, z, ux , uy , uz ) = z 2 + zuz − u2x − u2y .
fx = 0 fy = 0 fz = 2z + uz
fux = −2ux fuy = −2uy fuz = z
STEP 2 The auxiliary equations are

= = =− =− =−

Here = = =− =− =− .
−2ux −2uy z 0 0 2z + uz
That is, = = =− =− = .
−2ux −2uy z 0 0 −2z − uz
So, ux = a and uy = b.
Substituting this in the equation z 2 + zuz − u2x − u2y = 0, we get, z 2 + zuz − a2 − b2 = 0.
a2 + b2 − z 2
This means uz = .
z
STEP 3 The Pfaffian differential equation is du = ux dx + uy dy + uz dz.

a2 + b2 − z 2
du = a dx + b dy + dz.
z
z2
On integrating, u = ax + by + (a2 + b2 ) ln z − +c
2
This is a complete integral of the given p.d.e.

Example 2.3.5. Show that a complete integral of f (ux , uy , uz ) = 0 is u = ax + by + cz + d

where f (a, b, c) = 0. Hence find the complete integral of ux + uy + uz − ux uy uz = 0
Solution. The differential equation is f (ux , uy , uz ) = 0.

So, fx = 0, fy = 0, fz = 0.
Auxiliary equations are

= = =− =− =−
Here
= = = = =−
fux fuy fuz 0 0 0
Therefore ux = a, uy = b, uz = c.
Hence the Pfaffian differential equation is du = a dx + b dy + c dz.
Integrating,
u = ax + by + cz + d
This is a solution of f (ux , uy , uz ) = 0.

Here ux = a, uy = b, uz = c. Hence f (a, b, c) = 0.
This equation determines c in terms of a and b.
Hence the solution u = ax + by + cz + d has three constants and so it is a complete integral.
Now, we will find a complete integral of the differential equation ux + uy + uz − ux uy uz = 0.
Since the p.d.e. is of the form f (ux , uy , uz ) = 0, its complete solution is of the form
u = ax + by + cz + d, where ux = a, uy = b, uz = c.
Substituting in the given equation, we get a + b + c − abc = 0.
a+b
That is c = .
ab − 1
Example 2.3.6. (Objective question) Find a partial differential equation having a complete
integral as u = ax + by + (1 − a2 − b2 )z + d.
Solution. A trick for this type of questions is construct a pde of the form f (ux , uy , uz ) = 0.
We know that u = ax + by + cz + d is its complete integral where f (a, b, c) = 0.
That means ux = a, uy = b, uz = c and f (a, b, c) = 0.
Consider the given equation u = ax + by + (1 − a2 − b2 )z + d.
We will go in the reverse order.
Here c = (1 − a2 − b2 ).
=⇒ uz = 1 − u2x − u2y .
Hence the p.d.e. uz = 1 − u2x − u2y has u = ax + by + (1 − a2 − b2 )z + d as a complete integral.
We can verify also.
Examples 2.3.7. [3] Solve the following equations by Jacobi’s method.
(a + b)2 z z 2
(1) z + 2uz − (ux + uy )2 = 0. Ans: u = ax + by + − + c.
2 4
(a2 + αb2 )
(2) ux x2 − u2y − α u2z = 0, where α is a fixed constant. Ans: u = − + ay + bz + c.
x
(3) u2x + u2y + uz = 1, Ans: u = ax + by + (1 − a2 − b2 )z + c.
(4) xux + yuy = u2z . Ans: u = b ln x + (a2 − b) ln y + az + c.
Solution. (1) z + 2uz − (ux + uy )2 = 0 =⇒ f (x, y, z, ux , uy , uz ) = z + 2uz − (ux + uy )2 = 0
fx = 0 fy = 0 fz = 1 fux = −2(ux + uy ) fuy = −2(ux + uy ) fuz = 2
The auxiliary equations are

= = =− =− =−

Here = = =− =− =− .
−2(ux + uy ) −2(ux + uy ) 2 0 0 1
Substituting this in the equation z +2uz −(ux +uy )2 = 0, we get, z +2uz −(a+b)2 = 0.
(a + b)2 − z
This means uz = .
2
The Pfaffian differential equation is du = ux dx + uy dy + uz dz.
(a + b)2 − z
du = a dx + b dy + dz.
2
(a + b)2x z 2
On integrating, u = ax + by + − +c
2 4
(2) ux x2 − u2y − α u2z = 0, where α is a fixed constant.

=⇒ f (x, y, z, ux , uy , uz ) = ux x2 − u2y − α u2z = 0.
fx = 2ux x fy = 0 fz = 0 fux = x2 fuy = −2uy fuz = −2αuz

= = =− =− =−

Here 2
= = =− =− =− .
x −2uy −2αuz 2ux 0 0
So, uy = a and uz = b.
Substituting this in the equation ux x2 − u2y − α u2z = 0,, we get, ux x2 − a2 − αb2 = 0.
a2 + αb2
This means ux = .
x2
a2 + αb2
du = dx + a dy + b dz.
x2
a2 + αb2
On integrating, u = − + ay + bz + c
x
(3) u2x + u2y + uz = 1 =⇒ f (x, y, z, ux , uy , uz ) = u2x + u2y + uz − 1 = 0
fx = 0 fy = 0 fz = 0 fux = 2ux fuy = 2uy fuz = 1

= = =− =− =−

Here = = =− =− =− .
2ux 2uy 1 0 0 0
Substituting this in the equation u2x + u2y + uz − 1 = 0, we get, a2 + b2 + uz − 1 = 0.
This means uz = 1 − a2 − b2 .
du = a dx + b dy + (1 − a2 − b2 ) dz.
On integrating, u = ax + by + (1 − a2 − b2 )z + c
(4) xux + yuy = u2z . We write this as xux + yuy − u2z = 0.
fx = ux fy = uy fz = 0 fux = x fuy = y fuz = −2uz

= = =− =− =−

Here = = =− =− =− .
x y 0 ux uy 0
dx dux
This means uz = a and =− =⇒ ln x = − ln ux + ln b.
x ux
This means xux = b.
Substituting uz = a and xux = b in xux + yuy − u2z = 0, we get
b + yuy − a2 = 0
a2 − b
=⇒ uy =
y

b a2 − b
du = dx + dy + a dz.
x y
=⇒ u = b ln x + (a2 − b) ln y + az + c.

2.3.1 Jacobi’s method to find a complete integral for a first order p.d.e. in two
independent variables f (x, y, z, p, q) = 0.
Consider the following partial differential equation
f (x, y, z, p, q) = 0. (2.3.1)
The solution of 2.3.1 is a relation between x, y and z.

Suppose this relation is u(x, y, z) = 0. ‘(∗)
We treat u as a dependent variable and x, y, z as three independent variables ( z is a function
of x and y).
Differentiating (∗) partially w.r.t x and y respectively, we get
∂u ∂u ∂z
+ = 0.
∂x ∂z ∂x
and
∂u ∂u ∂z
+ = 0.
∂y ∂z ∂y
We write
∂u ∂u ∂u ∂z ∂z
= ux , = uy , = uz , = p, = q.
∂x ∂y ∂z ∂x ∂y
So,
ux + uz p = 0 and uy + uz q = 0
Hence
ux uy
p=− and q = − .
uz uz
On substituting these values in 2.3.1, we obtain a relation
g(x, y, z, ux , uy , uz ) = 0 (2.3.2)
Now, we apply Jacobi’s method that we have learnt above and find its complete integral.
This is a complete integral of the given p.d.e. f (x, y, z, p, q) = 0.
Examples 2.3.8. Solve the following equations by Jacobi’s method

(1) p2 x + q 2 y = z (4) z 3 = pqxy
(2) (p2 + q 2 )y = qz (5) p2 z + q 2 = 4
(3) p = (z + qy)2 (6) q − px − p2 .
Solution. (1) p2 x + q 2 y = z. We write this as p2 q + q 2 y − z = 0. Put
ux uy
p=− q=− .
uz uz
The given equation becomes
ux 2 uy 2

− x+ − y−z =0
uz uz
x u2x + y u2y − z u2z = 0
fx = u2x fy = u2y fz = −u2z fux = 2xux fuy = 2yuy fuz = −2zuz

= = =− =− =−

Here = = = − 2 = − 2 = − 2.
2xux 2yuy −2zuz ux uy −uz
dx dux dx dux 1 √ √ √
= − 2 =⇒ =− =⇒ ln x + ln ux = ln a =⇒ x ux = a
2xux ux 2x ux 2
Hence x u2x = a and
dy duy
= − 2 =⇒ yu2y = b
2yuy uy
Substituting these values in x u2x + y u2y − z u2z = 0, we get
a + b − z u2z = 0
√
a+b
=⇒ uz = √
z
√ √ √
a a a+b
The Pfaffian differential equation is du = √ dx + √ dy + √ dz.
x y z
√ √ √ √ √
On integrating, u = 2 a x+ y +2 a+b z+c
Writing u = c, we get a complete integral of the given p.d.e. as
√ √ √ √ √
2 a x+ y +2 a+b z =0
(2) (p2 + q 2 )y = qz.

We write this as (p2 + q 2 )y − qz = 0. Put
ux uy
p=− , q=− .
uz uz
ux 2 uy 2

uy
− y+ − y− − z=0
uz uz uz
y (u2x + u2y ) + z uy uz = 0 (∗)
fx = 0 fy = u2x + u2y fz = uy uz fux = 2 y ux fuy = 2 y uy + z uz fuz = z uy

= = =− =− =−
Here
= = =− =− 2 2
=− .
2yux 2yuy + zuz zuy 0 ux + uy uy uz
This means ux = a and

dz duz dz duz
=− =⇒ =− . That is, ln z = − ln uz + ln b. Hence z uz = b. So,
zuy uy uz z uz
b
uz = .
z
b b
Substituting ux = a and uz = in (∗), we get y a2 + y u2y + z uy = 0.
z z
That is, y u2y + b uy + y a2 = 0.
So,
p
−b ± b2 − 4a2 y 2
uy = .
2y
p
−b 1 b2 − 4a2 y 2
uy = ± .
2y 2 y
The Pfaffian differential equation is

 s 
−b 1 b2 b
du = a dx +  + 2
− 4a2  dy + dz. (∗∗)
2y 2 y z
p ! p
−b 1 b2 − 4a2 y 2 −b b2 − 4a2 y 2
Z Z Z
1
± . dy = dy + dy.
2y 2 y 2y 2 y
p
b2 − 4a2 y 2
Z
b 1
= − ln y + .
2 2 y
p
b2 − 4a2 y 2
Z
1
Consider dy
2 y
p
Put t = b2 − 4a2 y 2
=⇒ t2 = b2 − 4a2 y 2
b2 − t2
=⇒ y 2 = 2
√4a
b − t2
2
=⇒ y =
2a
1 1
=⇒ dy = √ (−2t) dt
2a 2 b2 − t2
1 −t
=⇒ dy = √ dt
2a b − t2 2
p
b2 − 4a2 y 2 −t
Z Z
1 1 t(2a) 1
Hence dy = √ √ dt
2 y 2 b − t 2a b − t2
2 2 2
−t2
Z
1
= dt
2 b − t2
2
t2
Z
1
= dt
2 t2 − b2
t2 − b2 + b2
Z
1
= dt
2 t2 − b2
b2
Z
1
= 1+ 2 dt
2 t − b2

1 2 1
t − b
= t+b ln
2 2b t + b

t b t − b
= + ln
2 4 t + b
p !
−b 1 b2 − 4a2 y 2 b t − b
Z
b t
Hence. ± . dy = − ln y + + ln
2y 2 y 2 2 4 t + b
p p
b b2 − 4a2 y 2 b b2 − 4a2 y 2 − b
= − ln y + + ln p
2 2 4 b2 − 4a2 y 2 + b

Now, we will solve the Pfaffian differential equation (∗∗).
p p
−b b2 − 4a2 y 2 b b2 − 4a2 y 2 − b
u=a x+ ln y + + ln p + b ln z + c
2 2 4 b2 − 4a2 y 2 + b
Writing u = c, we get a complete integral of the given p.d.e. as

p p
−b b2 − 4a2 y 2 b b2 − 4a2 y 2 − b
a x+ ln y + + ln p + b ln z = 0
2 2 4 b2 − 4a2 y 2 + b
(3) z 3 = pqxy We write this as xy pq − z 3 = 0. Put
ux uy
p=− q=− .
uz uz

ux uy
xy − − − z3 = 0
uz uz
xy ux uy − z 3 u2z = 0 (∗)
fx = y ux uy fy = x ux uy fz = −3z 2 u2z fux = xy uy fuy = xy ux fuz = −2z 3 zuz

= = =− =− =−
Here
= = 3
=− =− =− .
xy uy xy ux −2z zuz y ux uy x ux uy −3z 2 u2z
First and fourth ratio give

dx dux
=− .
xy uy y ux uy
dx dux
=⇒ =− .
x ux
=⇒ xux = a.
Hence
a
ux = .
x
Second and fifth ratio give
dy duy
=− .
xy ux x ux uy
2.4 Integral surfaces through a Given Curve: The Cauchy Problem 125
dy duy
=− .
y uy
=⇒ yuy = b.
Hence
b
uy = .
y
a b
Substituting ux = and uy = in (∗), we get
x y
a b
xy − z 3 u2z = 0
x y
ab − z 3 u2z = 0
√
ab
uz = 3
z2
√
a b ab
The Pfaffian differential equation is du = dx + dy + 3 dz.
x y z2
√
ab
On integrating, u = a ln x + b ln y − 2 √ + c.
z √
ab
Writing u = c, we get a complete integral of the given p.d.e. as a ln x+b ln y−2 √ =
z
0 This is a complete integral of the given p.d.e.
(4) p2 z + q 2 = 4.
(5) q − px − p2 .

2.4 Integral surfaces through a Given Curve: The Cauchy Problem

We want to find an integral surface passing through a given curve.
Let φ(u, v) = 0 or v = G(u) be the general integral of the particular differential equation
P p + Qq = R, where φ is an arbitrary function of uand v (or G is an arbitrary function
of u, ) where u(x, y, z) = c1 and v(x, y, z) = c2 are the solutions of the auxiliary equations
dx dy dz
= =
P Q R
Let C be the given curve whose parametric equations are given by
x = x0 (s), y = y0 (s), z = z0 (s)

where s is a parameter (not necessarily the arc length from a fixed point).
We want to find a particular surface F such that the surface F (u, v) = 0 contains the given
curve C.
This is done as follows. Consider the equations
u(x, y, z) = c1 , v(x, y, z) = c2
Substituting x = x0 (s), y = y0 (s), z = z0 (s) in these equations, we get
u(x0 (s), y0 (s), z0 (s)) = c1 , v(x0 (s), y0 (s), z0 (s)) = c2
Eliminating s between them, we get a relation between c1 and c2 .

Let that relation be given by F (c1 , c2 ) = 0.
Then F (u, v) = 0 is the requried solution as F (u, v) = 0
Example 2.4.1. Find the integral surface of the equation
(2xy − 1) p + (z − 2x2 ) q = 2(x − yz)
which passes through the line x0 (s) = 1, y0 (s) = 0, z0 (s) = s
Solution. The auxiliary equations are

dx dy dz
= = · · · (1)
2xy − 1 z − 2x2 2(x − yz)
Choose z, 1, x as multipliers.
zdx + dy + xdz zdx + dy + xdz
So each ratio in (1) equals = .
z(2xy − 1) + z − 2x2 + 2x(x − yz) 0
Hence zdx + dy + xdz = 0. This implies d(xz) + dy = 0.
Integrating xz + y = c1 . Hence u(x, y, z) = xz + y.
Now, choose 2x, 2y, 1 as multipliers.
So, each ratio in (1) equals
2x dx + 2y dy + dz
= .
0
Hence 2x dx + 2y dy + dz = 0. Integrating we get, x2 + y 2 + z = c2 .
Thus v(x, y, z) = x2 + y 2 + z.
Hence the general integral is v = G(u) where G is an arbitrary function.
That is, x2 + y 2 + z = G(xz + y).
Now, we want to choose G such that the given curve lies on the surface x2 +y 2 +z = G(xz+y).
(
That is, G should be such that G(s) = x0 s) + y02 (s) + z0 (s) = 1 + 0 + s = 1 + s.
2 2
Therefore the required integral is x + y + z = G(xz + y) = xz + y + 1.
That is x2 + y 2 + −xz + z − y − 1 = 0
Chapter 3
Integral Transforms: Semester VI
Course Code: USMT6A4/ UAMT6A4

Unit I: The Laplace Transform (15L)
Definition of Laplace Transform, theorem, Laplace transforms of some elementary
functions, Properties of Laplace transform, LT of derivatives and integrals, Initial
and final value theorem, Inverse Laplace Transform, Properties of Inverse Laplace
Transform, Convolution Theorem, Inverse LT by partial fraction method, Laplace
transform of special functions: Heaviside unit step function, Dirac-delta function
and Periodic function.
Unit II: The Fourier Transform
Fourier integral representation, Fourier integral theorem, Fourier Sine & Cosine
integral representation, Fourier Sine & Cosine transform pairs, Fourier transform
of elementary functions, Properties of Fourier Transform, Convolution Theorem,
Parseval’s Identity.
Unit III: Applications of Integral Transforms
Relation between the Fourier and Laplace Transform. Application of Laplace
transform to evaluation of integrals and solutions of higher order linear ODE.
Applications of LT to solution of one dimensional heat equation & wave equation.
Application of Fourier transforms to the solution of initial and boundary value
problems, Heat conduction in solids (one dimensional problems in infinite & semi
infinite domain).
Suggested Practicals based on USMT6A4 / UAMT6A4
1 Find the Laplace transform of differential and integral equations.
2 Find the inverse Laplace transform by the partial fraction method.
3 Find the Fourier integral representation of given functions.
4 Find the Fourier Sine / Cosine integral representation of given functions.
5 Solve higher order ODE using Laplace transform.

128 CHAPTER 3. INTEGRAL TRANSFORMS: SEMESTER VI
6 Solve one dimensional heat and wave equation using Laplace transform. Solve initial
and boundary value problems using Fourier transform.
7 Miscellaneous Theory Questions.
Reference Books:
1. Lokenath Debnath and Dambaru Bhatta, Integral Transforms and their Applications,
CRC Press Taylor & Francis.
2. I. N. Sneddon, Use of Integral Transforms, Tata-McGraw Hill.
3. L. Andrews and B. Shivamogg, Integral Transforms for Engineers, Prentice Hall of

India.
Bibliography
[1] Sudhir R. Ghorpade, Balmohan V. Limaye; A Course in Calculus and Real Analysis;
second edition, Springer.
[2] Sudhir R. Ghorpade, Balmohan V. Limaye; A Course in Multivariable Calculus and

Analysis; Springer International Edition.
[3] T. Amarantath; An Elementrary Course in Partial Differential Equations; Narosa.
[4] M. D. Raisinghania; Ordinary and Partial Differential Equations; 20th Edition, S.

Chand.
[5] Partial differential equations; Phoolan Prasad, Renuka Ravindrann; New Age, second
edition.
[6] George B. Thomas, Jr. , Ross L. Finney; Calculus and Analytic Geometry; 9th Edi-
tion, Pearson Education.
[7] Kendall E. and Atkinson; An Introduction to Numerical Analysis; Wiley.
[8] S.D. Conte, Carl de Boor; Elementary Numerical Analysis’ An Algorithmic Approach,
McGraw-Hill Book Company.
[9] http://www.math.iitb.ac.in/ siva/ma51515/AppEnvelopes.pdf

130 BIBLIOGRAPHY

Pde 1 & 2

Uploaded by

Copyright:

Available Formats

Pde 1 & 2

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Pde 1 & 2

Uploaded by

Copyright:

Available Formats

What topics are covered in Unit I of the course?

What topics are covered in Unit I of the course?

What method is described for solving the Cauchy problem?

What method is described for solving the Cauchy problem?

T.Y.B.Sc.

Mathematics USMT 504

Dr. Abhaya Chitre

September 13, 2022

1 Partial Differential Equations 1

3 Integral Transforms: Semester VI 127

Partial Differential Equations

Course: Partial Differential Equations (Elective A)

1. T. Amaranath; An Elementary Course in Partial Differential Equations; 2nd edition,

2. Ian Sneddon; Elements of Partial Differential Equations; McGraw Hill book.

4. W. E. Williams; Partial Differential Equations; Clarendon Press, Oxford, (1980).

5. K. Sankara Rao; Introduction to Partial Differential Equations; Third Edition, PHI.

1. Find general solution of Langrange’s equation.

3. Find complete integral of first order PDE using Charpit’s Method.

4. Find complete integral using Jacobi’s Method.

5. Solve initial value problem for quasi-linear PDE.

6. Find the integral surface by the method of characteristics.

7. Miscellaneous Theory Questions.

1.1 Curves and Surfaces

(i) Parametrically Defined Curve is the plane curve Ç given by a function

• D is called as the parameter domain of the curve C .

• The curve C is determined by its parametrization that

We get the value of s in terms of t.

• Consider a helix in R3 given by the parametric equations

x(t) = a cos t, y(t) = a sin t, z(t) = bt + c, t ∈ R,

where a, b, c ∈ R with a > 0 and b 6= 0. It lies on thecylinder givenby x2 + y 2 = a2 .

Definition 1.1.2. Parametrically Defined 3

(i) D is called as the parameter domain of the surface S .

Remark 1.1.4. [2] We will need the following Chain Rule.

(i) If z = f (x, y) and w = g(z), then w is a function of (x, y), and

(ii) If z = f (x, y) and x = x(t), y = y(t), then z is a function of t, and

Jacobian matrix of the transformation

1.2 Genesis of first Order P.D.E

(1) order = 1 (3) order = 2 (5) order = 1

(2) order = 1 (4) order = 3 (6) order = 1

Definition 1.2.3. Degree of a partial differential equation is defined as the power

(1) degree = 1 (3) degree = 3 (5) degree = 1

(2) degree = 1 (4) degree = 1 (6) degree = 2

(a) it is linear in p, q and z and

(b) the coefficients of p and q are functions of x and y only.

Equations of the form:

P (x, y)p + Q(x, y)q = R(x, y)z + S(x, y).

are first order linear PDE.

First order linear PDE

(a) it is linear in p, q and

(b) the coefficients of p and q are functions of x and y only.

Equations of the form:

P (x, y)p + Q(x, y)q = R(x, y, z).

are first order semi-linear PDE.

P (x, y, z)p + Q(x, y, z)q = R(x, y, z).

are first order quazi-linear PDE.

Types of first order PDE

Linear eqns Semi-linear eqns Quasi-linear eqns Non-linear eqns

If it is linear in p, q linear in p, q linear in p, q not linear in

the form = R(x, y)z + S(x, y) = R(x, y, z) = R(x, y, z)

where a, b, c ∈ R with a > 0 and b 6= 0. It lies on thecylinder givenby x2 + y 2 = a2 .

∂z y y ∂z y

∂z ∂z y y y