Professional Documents
Culture Documents
VectorTensorAnalysis PDF
VectorTensorAnalysis PDF
Complex Analysis
AIILFORS
BELLMAN Stability Theory of Differential Equations
Bucx Advanced Calculus
CODDnNUTON AND LEVINSON Theory of Ordinary Differential Equations
G01.011 13 AND SHANKS Elements of Ordinary Differential Equations
GRAVES The Theory of of Real Variables
GRIFFIN Elementary Theory of Numbers
IIILDEBRAND - rodurl ion to Numerical Analysis
Principles of Nunurriral Analysis
LAS b;leuu'nts of Pure aunt Applied Mathematics
LASS Vector and Tensor Analysis
LEIGIITON An Introduction to the Theory of Differential Equations
NEHAIU Conformal Mapping
NEWELL Vector Analysis
ROSSER Logic for 'Mathematicians
RUDIN Principles of Mathematical Analysis
SNEDDON Elciuents of Partial I)iiTerential Equations
SNEDDON Pourier Transforms
STOLL Linear Algebra and 'Matrix Theory
WEINSTOCK Calculus of Variations
VECTOR AND TENSOR
ANALYSIS
BY
HARRY LASS
JET PROPULSION LABORATORY
Calif. Institute of Tech.
4800 Oak Grove Drive
Pasadena, California
CHAPTER 2
DIFFERENTIAL VECTOR CALCULUS . . . . . . . . . . . . . 29
16. Differentiation of vectors 17. Differentiation rules 18. The
gradient 19. The vector operator del, V 20. The divergence of a
vector 21. The curl of a vector 22. Recapitulation 23. Curvi-
linear coordinates
CHAPTER 3
DIFFERENTIAL GEOMETRY . . . . . . . . . . . . . . . . . 58
24. Frenet-Serret formulas 25. Fundamental planes 26. In-
trinsic equations of a curve 27. Involutes 28. Evolutes 29.
Spherical indicatrices 30. Envelopes 31. Surfaces and curvi-
linear coordinates 32. Length of arc on a surface 33. Surface
curves 34. Normal to a surface 35. The second fundamental
form 36. Geometrical significance of the second fundamental
form 37. Principal directions 38. Conjugate directions 39.
Asymptotic lines 40. Geodesics
CHAPTER 4
INTEGRATION . . . . . . . . . . . . . . . . . . . . . . . . 89
41. Point-set theory 42. Uniform continuity 43. Some proper-
ties of continuous functions 44. Cauchy criterion for sequences
45. Regular area in the plane 46. Jordan curves 47. Functions
of bounded variation 48. Arc length 49. The Riemann integral
ix
x CONTENTS
50. Connected and simply connected regions 51. The line inte-
gral 52. Line integral (continued) 53. Stokes's theorem 54.
Examples of Stokes's theorem 55. The divergence theorem
(Gauss) 56. Conjugate functions
CHAPTER 5
STATIC AND DYNAMIC ELECTRICITY . . . . . . . . . . . 127
57. Electrostatic forces 58. Gauss's law 59. Poisson's formula
60. Dielectrics 61. Energy of the electrostatic field 62. Dis-
continuities of D and E 63. Green's reciprocity theorem 64.
Method of images 65. Conjugate harmonic functions 66. Inte-
gration of Laplace's equation 67. Solution of Laplace's equation
in spherical coordinates 68. Applications 69. Integration of
Poisson's equation 70. Decomposition of a vector into a sum of
solenoidal and irrotational vectors 71. Dipoles 72. Electric
polarization 73. Magnetostatics 74. Solid angle 75. Moving
charges, or currents 76. Magnetic effect of currents (Oersted)
77. Mutual induction and action of two circuits 78. Law of in-
duction (Faraday) 79. Maxwell's equations 80. Solution of
Maxwell's equations for electrically free space 81. Poynting's
theorem 82. Lorentz's electron theory 83. Retarded potentials
CHAPTER 6
MECHANICS . . . . . . . . . . . . . . . . . . . . . . . . . 184
84. Kinematics of a particle 85. Motion about a fixed axis 86.
Relative motion 87. Dynamics of a particle 88. Equations of
motion for a particle 89. System of particles 90. Momentum
and angular momentum 91. Torque, or force, moment 92. A
theorem relating angular momentum with torque 93. Moment of
momentum (continued) 94. Moment of relative momentum 95.
Kinetic energy 96. Work 97. Rigid bodies 98. Kinematics of
a rigid body 99. Relative time rate of change of vectors 100.
Velocity 101. Acceleration 102. Motion of a rigid body with
one point fixed 103. Applications 104. Euler's angular coordi-
nates 105. Motion of a free top about a fixed point 106. The
top (continued) 107. Inertia tensor
CHAPTER 7
HYDRODYNAMICS AND ELASTICITY . . . . . . . . . . . . 230
108. Pressure 109. The equation of continuity 110. Equations
of motion for a perfect fluid 111. Equations of motion for an
incompressible fluid under the action of a conservative field 112.
The general motion of a fluid 113. Vortex motion 114. Appli-
cations 115. Small displacements. Strain tensor 116. The
stress tensor 117. Relationship between the strain and stress
tensors 118. Navier-Stokes equation
CONTENTS xi
CHAPTER 8
TENSOR ANALYSIS AND RIEMANNIAN GEOMETRY. . . . . 259
119. Summation notation 120. The Kronecker deltas 121.
Determinants 122. Arithmetic, or vector, n-space 123. Contra-
variant vectors 124. Covariant vectors 125. Scalar product of
two vectors 126. Tensors 127. The line element 128. Geode-
sics in a Riemannian space 129. Law of transformation for the
Christoffel symbols 130. Covariant differentiation 131. Geode-
sic coordinates 132. The curvature tensor 133. Riemann-
Christoffel tensor 134. Euclidean space
CHAPTER 9
FURTHER APPLICATIONS OF TENSOR ANALYSIS . . . . . . 311
135. Frenet-Serret formulas 136. Parallel displacement of vectors
137. Parallelism in a subspace 138. Generalized covariant differ-
entiation 139. Riemannian curvature. Schur's theorem 140.
Lagrange's equations 141. Einstein's law of gravitation 142.
Two-point tensors
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . 339
INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
CHAPTER 1
THE ALGEBRA OF VECTORS
Fia. 2.
FIG. 3.
Z-a
Fm. 4.
and which has the same direction as a (see Fig. 3). We define
-a as the vector obtained from a by reversing its direction (see
Fig. 4).
We note that
x(ya) = (xy)a = xya
(x + y)a = xa + ya
Oa = 0 (zero vector)
It is immediately seen that two vectors are parallel if, and only
if, one of them can be written as a scalar multiple of the other.
4. Addition of Vectors. Let us suppose we have two vectors
given, say a and b. We form a third vector by constructing a
triangle with a and b forming two sides of the triangle, b adjoined
to a (see Fig. 5). The vector starting from the origin of a and
ending at the arrow of b is defined as the vector sum a + b.
We see that a + 0 = a, and if a = b, c = d, then
a+c=b+d
SEc. 6] THE ALGEBRA OF VECTORS 3
a+b+c
Fio. 5.
a-b
a-b
Flo. 6.
a
FIG. 8.
c=OB+BC
= b + x(a - b)
=xa+(1 - x)b
so that
c = xa + yb (5)
with l2+m2+n2p,, 0.
We shall, however, find (5) more useful for solving problems.
Example 1. Let us prove that the medians of a triangle meet
at a point P which divides each median in the ratio 1:2.
C
0
FIG. 9.
Let ABC be the given triangle and let A', B', C' be the mid-
points. Choose 0 anywhere in space and construct the vectors
from 0 to A, B, C, A', B', C', calling them a, b, c, a', b', c' (see
Fig. 9). From (5) we have
a'=4b+ic (8)
b' = Ja + 4c
Now P (the intersection of two of the medians) lies on the line
joining A and A' and on the line joining B and B'. We shall thus
find it expedient to find a relationship between the four vectors
a, b, a', b' associated with A, B, A', B'. From (8) we eliminate
the vector c and obtain
2a' + a = 2b' + b
or
*a' + '}a = *b' + b (9)
6 VECTOR AND TENSOR ANALYSIS [SFC. 6
FIG. 10.
b-a=2(d-c)
show that the intersection point of the two lines joining A and D
and B and C trisects these lines.
9. a, b, c, d are four vectors with a common origin. Find a
necessary and sufficient condition that their end points lie in a
plane.
10. What is the vector condition that the end points of the
vectors of Prob. 9 form the vertices of a parallelogram?
11. Show that the mid-points of the lines which join the mid-
points of the opposite sides of a quadrilateral coincide. The
four sides of the quadrilateral are not necessarily coplanar.
12. Show that the line which joins one vertex of a parallelogram
to the mid-point of an opposite side trisects the diagonal.
13. A line from a vertex of a triangle trisects the opposite side.
It intersects a similar line issuing from another vertex. In what
ratio do these lines intersect one another?
14. A line from a vertex of a triangle bisects the opposite side.
It is trisected by a similar line issuing from another vertex. How
does this latter line intersect the opposite side?
15. Show that the bisectors of a triangle meet in a point.
16. Show that if two triangles in space are so situated that the
three points of intersection of corresponding sides lie on a line,
then the lines joining the corresponding vertices pass through a
common point, and conversely. This is Desargues's theorem.
17. b = (sin t)a is a variable vector which always remains
fib?
parallel to the fixed vector a. What is Explain geomet-
r=xi+yj+zk (10)
and will be used quite frequently in what follows. The most gen-
eral space-time vector that we shall encounter will be of the form
u =u(x,y,z,t) =a(x,y,z,t)1+#(x,y,z,t)j
+ y(x, y, z, t)k (11)
It is of the utmost importance that the student understand the
meaning of (11). To be more specific, let us consider a fluid in
motion. At any time t the particle which happens to be at the
z
Fm. 11.
where 0 is the angle between the two vectors when drawn from a
common origin. It makes no difference whether we choose 0 or
- 0 since cos 0 = cos (- 0). This definition of the scalar product
arose in physics and will play a dominant role in the develop-
ment of the text.
Fm. 12.
a
a
ab then a b= a
b the a b
b
ab = (proj a)b (bI = (proj b)a (al (17)
SEC. 8J THE ALGEBRA OF VECTORS 11
[proj (b+c)]aIaF
f I
= (proj b)a Iai + (proj C)a Ial
Fm. 13.
so that
0
FIG. 14.
C2
C. Cone BC
C YECr?,oR
aright
_ _
Lau, AC
b2
cb _ -,
j'i yAC BC AND
angle
1'1 of
PIG. I° 0
-b2+a2_.
.]a
a -a 18.
'
c a
2ab (b $
+c
c
T'`iq°n°metr 0
_j.k:g=o
Ab _a)
. a
co$B
c,a
z [SEc.
s
SEC. 8] THE ALGEBRA OF VECTORS 13
Notice that p12 + q12 + r12 = 1. pi, q1, ri are called the direc-
tion cosines of the vector Y. Similarly, let j' and k' be unit
vectors with direction cosines p2, qa, r2 and pa, qa, r,. Thus
j' = p2i + q2j + r2k ()
k' = pai+qsj+rak
We also impose the condition that i', j', k' be mutually orthogonal,
so that the x', y', z' axes form a coordinate system similar to the
x-y-z coordinate system with common origin 0 (see Fig. 17).
We have r = r' so that xi + yj + zk = x'i' + y'j' + z'k',
where x, y, z are the coordinates of a point P as measured in the
x-y-z coordinate system and x', y', z' are the coordinates of the
same point P as measured in the x'-y'-z' coordinate system.
Making use of (21) and (22) and equating components, we find
that
x = p1x' + ply' + pgz'
y = q1x' + qty' + qaz' (23)
z = r1x' + r2y' + raz'
14 VECTOR AND TENSOR ANALYSIS 1Src. 8
FIG. 17.
Z' = 28. Now let a,a represent the cosine of the angle between
the xa and V axes. We can write (23) as
3
xa = I a0a.V, a = 1, 2, 3 (24)
a-1
By making use of the fact that i' j' = j' k' = k' i' = 0, we
can prove that
3
xa = I Asax#, a = 1, 2, 3 (25)
0-1
where Asa = a). We leave this as an exercise for the reader.
Let us notice that differentiating (24) yields
axa
- = a,a, of, or = 1, 2, 3 (26)
SEC. 91 THE ALGEBRA OF VECTORS 15
The triples (1, 0, 0), (0, 1, 0), (0, 0, 1) form a basis for our linear
vector space, that is, the space of number triples. We note that
the determinant formed from these triples, namely,
1 0 0
0 1 0 =1
00 1
does not vanish. Any three triples whose determinant does not
vanish can be used to form a basis. Let the reader prove this
result. We can define the scalar product (inner product) of
two triples by the law (a - b) = a'b' + a2b2 + a3b3.
9. Applications of the Scalar Product to Space Geometry
(a) We define a plane as the locus of lines passing through a
fixed point perpendicular to a fixed direction. Let the fixed
point be Po(xo, yo, zo) and let the fixed direction be given by the
vector N = Ai + Bj + Ck. Let r be the position vector to any
-4
point P(x, y, z) on the plane (Fig. 18). Now POP = r - ro is per-
pendicular to N so that
or
[(x - xo)i + (y - yo)j + (z - zo)kl - (Ai + Bj + Ck) = 0
and
A(x - xo) + B(y - yo) + C(z - zo) = 0 (27)
This is the equation of the plane. The point Po(xo, yo, zo) obvi-
ously lies in the plane since its coordinates satisfy (27). Equa-
tion (27) is linear in x, y, Z.
(b) Consider the surface Ax + By + Cz + D = 0. Let P(xo,
yo, zo) be any point on the surface. Of necessity,
Axo+Byo+Czo+D=0.
16 VECTOR AND TENSOR ANALYSIS [SEC.9
Subtracting we have
A(x - xo) + B(y - yo) + C(z - zo) = 0 (28)
Fia. 18.
Fia. 19.
d = JPoP NJ
= IAZ+B, +Cr+DI (29)
(A2 + B2 + C2)f
(d) Equation of a straight line through the point Po(xo, yo, zo)
parallel to the vector T = li + mj + A. From Fig. 20 it is
z
Fia. 21.
18 VECTOR AND TENSOR ANALYSIS [SEC. 9
Problems
1. Add and subtract the vectors a = 2i - 3j + 5k,
b = -2i+2j+2k
Show that the vectors are perpendicular.
2. Find the cosine of the angle between the two vectors
a = 2i - 3j + k and b = 3i - j - 2k.
3. If c is normal to a and b, show that c is normal to a + b,
a - b.
4. Let a and b be unit vectors in the x-y plane making angles
a and 6 with the x axis. Show that a = cos a i + sin a j,
b = cos 9 i + sin 3 j, and prove that
cos (a - 6) = cos a cos S + sin a sin 8
5. Find the equation of the cone whose generators make an
angle of 30° with the unit vector which makes equal angles with
the x, y, and z axes.
6. The position vectors of the foci of an ellipse are c and - c,
and the length of the major axis is 2a. Show that the equation
of the ellipse is a4 - a2(r2 + c2) + (c r)2 = O.
7. Prove that the altitudes of a triangle are concurrent.
8. Find the shortest distance from the point A(1, 0, 1) to the
line through the points B(2, 3, 4) and C(-1, 1, -2).
SEC. 91 THE ALGEBRA OF VECTORS 19
Show that
da
a= lal
dal
dt dt
19. Find the angle between the plane Ax + By + Cz + D = 0
and the plane ax + by + cz + d = 0.
20, a, b, c are coplanar. If a is not parallel to b, show that
ca ab aa ca b
c=
Icb bbi ab cb
aa ab
ab bb
21. For the an", Asa defined by (24) and (25), show that
a, -Ap r = Sp-
y-1
where &0°= Iifa=l4,Se=Oif aPd f.
22. If B1, B2, B3 are the components of a vector B, that is,
B = B'i + B2j + B3k (see Example 8 in regard to the super-
20 VECTOR AND TENSOR ANALYSIS [SEC. 10
Fia. 23.
a
13. The Triple Scalar Product. Let OP
%
Fia. 25.
Now
i j k
a (b x c) = (a,i + a2j + ask) . b, b2 bs
c1 C2 C3
= ai(b2ca - bac2) + a (baci - bic3) + as(bic2 - b2c1)
so that
a1 a2 as
a (b x c) = b, b2 bs (36)
C1 C2 C3
24 VECTOR AND TENSOR ANALYSIS [SEC. 14
x(b xc)]
-[(a xb)
x b (a
x b) x (a b)(b c)
implying X = 1. Thus
d) (b c) d) (40)
15. Applications to the Spherical Trigonometry. Consider the
spherical triangle ABC (sides are arcs of great circles) (see Fig.
26). Let the sphere be of radius 1. Now from (40) we see that
(a x b) (a x c) = (b c) - (a b) (a c)
FIG. 26.
Problems
1. Show by two methods that the vectors a = 2i - 3j - k,
b = -6i + 9j + 3k are parallel.
2. Find a unit vector perpendicular to the vectors
a=i-j+k
b=i+i - k
3. A particle has an angular speed of 2 radians per second, and
its axis of rotation passes through the points P(0, 1, 2),
Q(1, 3, -2)
Find the velocity of the particle when it is located at the point
R(3, 6, 4).
4. Find the equation of the plane passing through the end
points of the vectors a = a1i + a j + ask, b = b1i + bd + bsk,
c = c1i + cd + cak, all three vectors with origin at P(0, 0, 0).
5. Show that (a x b) x (c x d) = (acd)b - (bcd)a.
6. Prove that d x (a x b) (a x c) = (abc) (a d).
7. If a + b+ c= 0, prove that a x b= b x c= c x a,
and interpret this result trigonometrically.
SEC. 15] THE ALGEBRA OF VECTORS 21
Let a = aui + a2j + aak, etc. Write (41) as a single vector equa-
tion, and assuming (abc) F6 0, solve for x, y, z.
18. Find the shortest distance between two straight lines in
space.
28 VECTOR AND TENSOR ANALYSIS [Ssc. 15
du = dai.+d,6i+dyk (43)
as as as
du (-aa
dz + at
d) i
ax dx + ay dy + az
( -dz
+(axdx+aydy+a
atdt)j
dy -}- az dz }- at dt J k
+ ax dx +
For example, let r = xi + yj + zk be the position vector of a
moving particle P(x, y, z) in three-space. Then
dr = dx i + dy j + dz k
29
30 VECTOR AND TENSOR ANALYSIS [SEC. 16
and
v k (44)
dt dt 1 + dt + dt
d 2r z 2y , 2
a= j + dt2 k (45)
dt2 = dt2 I + dt2
a = -w2r (47)
Example 15. Let P be any point on the space curve (Fig. 29)
x=x(s)
y = y(s)
z = z(s)
c(t+
At At At At
_
d (u v)
dt
- udtdv + dudt'v (50)
Similarly
d(u x v) dv du
u v (51)
dt - x dt + dt
d(fu) - du d_f
u (52)
dt f dt + dt
u'du du (53)
auat
-
This result is not trivial, for FIG. 30.
Idul du.
Example 18. Motion in a Plane. Now r = rR, where R is a
unit vector (see Fig. 30). Hence
Now
R is perpendicular to R (see Example 16). Also
IdRI
= de
since R is a unit vector. We can easily verify this by differ-
entiating R = cos B i + sin 9 j. Hence v = R + r d8 P, where
Wt-
since
P= -d6R
(54)
34 VECTOR AND TENSOR ANALYSIS [Sic. 17
Thus
I d2r
- (dO)2]Rid(odO\p
dt r dt dt
(55)
Problems
1. Prove (51) and (52).
2. Prove (54).
/
3. Differentiate l r - dr1 with respect to t.
4. Expand dt [p x (q x r)].
d d2r
5. Show that r x dt) = r x
dr d2r
6. Find the first and second derivatives of (r dt dt2
7. r = a cos wt + b sin wt; a, b, w are constants. Prove that
r xdt = wa xb andd +w2r = 0.
RxdR= r x dr 7.2
10. If dt = w x a,
d = w x b, show that
dt(axb)=0x(axb)
11. If r = aew' + be-,", show that dt; -- w2r = 0. a, bare con-
start vectors.
Fia. 31.
(i) du+2Adu+Bu=0
d3u
dt3
a
d2u
te 2 d
du
0
du
such that u = i, = j, d = k for t = 0.
19. If u, is a solution of
d8u+Ad2u+Bdu+Cu=0
and if u2 is a solution of
d
dtu+Adtu+Bdt+Cu=F(t)
dV =
a dx + a dy + a dz (56)
ax ay az
SEC. 181 DIFFERENTIAL VECTOR CALCULUS 37
Now let r be the position vector to the point P(x, y, z).
r = xi + yj + zk
If we move to the point Q(x + dx, y + dy, z + dz) (Fig. 32),
dr = dxi + dyj + dzk
Now notice that (56) contains the terms dx, dy, dz and the terms
aP We define a new vector formed from gp by taking its
aP, aSP,
ax 8y az
three partial derivatives. Let del rp = gradient p be defined by
d(p = dr VV (58)
d(P
=u V(p (60)
ds
,P(x,y,z) =x2+y2-z
V(p=2xi+2yj-k
=2i-2j-katP(1,1,1)
Thus
N=2i-2j-k
3
Fio. 34.
But from Example 20, Vrj is a unit vector parallel to the vector
AP, and Vr2 is a unit vector parallel to the vector BP. Equation
40 VECTOR AND TENSOR ANALYSIS [SEC. 19
(63) shows that AP and BP make equal angles with the tangent
to the ellipse.
19. The Vector Operator V. We define
v - (1-+jay +kaz)'p
V(uv) = U Vv + v Vu (65)
Problems
1. Find the equation of the tangent plane to the surface
xy - z = 1 at the point (2, 1, 1).
Sac. 191 DIFFERENTIAL VECTOR CALCULUS 41
a = 1, 2, 3.
axa 9-1 ay8 axa
16. Apply the results of Prob. 15 above to the transformation
x = r cos 0
y = r sin 0
z=z
1 of , of
and show that af, are the components of Vf along the
Or r ae az
three mutually orthogonal unit vectors e, es, e. which occur in
cylindrical coordinates.
dr
17. If (_ v(x, y, z, t), show that d,P= ate + V(p.
A dy E
ry
Fia. 35.
so tha t
a(Pu)
ax + a(pv)
ay + a(Pw)
az
(66)
represents the loss of mass per unit time per unit volume. This
quantity is called the divergence of the vector pf. We see at
once that
. (pf ) = di v (pf) =
a(Pu) a(Pv) a(Pw)
= 1 dM (67 )
VV
ax + ay + az V dt
44 VECTOR AND TENSOR ANALYSIS [SEc. 20
V
a(q.u) + a(te) + a(cyw)
ax ay az
_ (ax
au av aw arp a'p sip
x +ay+ az + uax +v ay +u'aa
(68)
k/
axe+ayz+az,
SEc.21] DIFFERENTIAL VECTOR CALCULUS 45
a a a
curl f=vxf= ax ay az
u v w
Example 27
Vxr= =0
Example 28
1 i j k
a a a
v x (wf) =
ax ay az
cpu vv cv
[a(te) _ a()1 [a(sou) a(caw)
ay az JJ az ax
+ k [a(;) a(,Pu)]
ax ay
aw av au Ow
v x (cOf) = cO [i (ay az
+j az ax
U v w
Vx('vf)=(Pvxf+Via xf (72)
46 VECTOR AND TENSOR ANALYSIS [SEC. 21
av av av
ax ay az
a2,p 02,p C1 2,p
+ i
\ az ax ax az/ + k \ax ay
a2, ay ax/
Hence
Vxvio =0 (73)
V (v x f) =
a aw_av a au_aw' a(av_au)
ax ay az + ay az ax / + az \ax ay/
a2u a2ul a2V a2v a2w a2w
ay az az ay + az ax ax az + ax ay - ay ax
Thus
(74)
uxax+Uyay+u$az
Thus
df=axdx+af dy+azdz
y
=dxa+dya+dzaz
y
SFC. 211 DIFFERENTIAL VECTOR CALCULUS 47
and
since dr = dx i + dy j + dz k.
If f = f(x, y, z, t),
Example 32
+vsa
= vJ + vyj + vZk
(v - V)r = v (77)
u x (V x v) = Vti(u v) - (u V)v
Here we have applied the rule of the middle factor, noting also
that V operates only on v. Vn(u v) means that we keep the
components of u fixed and differentiate only the components of v.
Similarly, v x (V x u) = v) - (v V)u. Adding, we
obtain
Vu(u v) + Vn(u v) = u x (V x v) + v x u)
+ (u V)v + (v V)u
and
Example 34
Example 35
V. (u xv) = Vu (u xv) +Vv (u xv)
(V (V (80)
Example 36
V
Vxr
=0
13. df = (dr V)f + a dt
Sxc. 22] DIFFERENTIAL VECTOR CALCULUS 49
14. dco = dr VV + LP dt
15. V- (r 3r) =0
Problems
1. Show that V2(1/r) = 0 where r = (x2 + y2 + z2)}.
2. Compute V2 r, V2r2, V2(1/r2) where r = (x2 + y2 + z2)3.
3. Expand V(uvw).
4. Find the divergence and curl of (xi - yj)/(x + y); of
xcoszi+ylogxj -z2k.
5. If a = axi + fiyj + yzk, show that V(a r) = 2a.
6. Show that V x V(r)r] = 0 when r = (x2 + y2 + Z2) i and
r = xi + yj + zk.
7. Let u = u(x, y, z), v = v(x, y, z). Suppose u and v satisfy
an equation of the form f(u, v) = 0. Show that Vu x Vv = 0.
8. Assume Vu x Vv = 0 and assume that we move on the sur-
face u (x, y, z) = constant. Show that v remains constant and
hence v = f(u) or F(u, v) = 0.
9. Prove that a necessary and sufficient condition that u, v, w
satisfy an equation f(u, v, w) = 0 is that Vu Vv x Vw = 0, or
au au au
ax ay az
av ft av
0
ax ay az
aw aw aw
ax ay az
a=1 xa = a=1 x
so that V f is a scalar invariant under a rotation of axes. Also
see Prob. 21, Sec. 9.
15. Prove (79), (80), (81).
16. Let f = f1i + f2j + f3k and consider nine quantities
afi of,
9is i, j = 1, 2, 3
axi axis
Show that gi; = -gi and that three of the nine quantities yield
the three components of V x f. Use this result to show that
V x((pf) =('V xf+Vcpxf.
23. Curvilinear Coordinates. Often the mathematician, phys-
icist, or engineer finds it convenient to use a coordinate system
other than the familiar rectangular cartesian coordinate system.
If he is dealing with spheres, he will probably find it expedient to
describe the position of a point in space by the spherical coordi-
nates r, 8, sp (see Fig. 31). Let us note the following: The sphere
x2 + y2 + z2 = r2, the cone z/(x2 + y2 + z2)'1 = cos 8, and the
plane y/x = tan sp pass through the point P(r, 0, Sp). We may
consider the transformations
r = (x2+y2+Z2)}
z
a = cos- '
(x2 + y2 + z2)
tan-1 y
Sp =
X
We assume that the Jacobian J[(ul, u2, u3)/(x, y, z)] ;P,-, 0 so that
the transformation (82) is one to one in the neighborhood of a
point. A point in space is determined when x, y, z are known
and hence when ui, u2, u3 are known. By considering
ui(x, y, z) = ci
u2(x, y, z) = C2, u3(x, y, z) = es, we obtain a family of surfaces.
Through a point P(xo, yo, zo) will pass the three surfaces
FIG. 36.
u2(x, y) z) = u2(xo, yo, Zo), and u3(x, y, z) = u8(xo, yo, zo). Let us
assume that the three surfaces intersect one another orthogonally.
The surfaces will intersect in pairs, yielding three curves which
intersect orthogonally at the point P(xo, yo, zo). The curve of
intersection of the surfaces uI = ci and u2 = C2 we shall call the
us curve, since along this curve only the variable us is allowed to
change. Let u1, u2, us be three unit vectors issuing from P
tangent to the ui, us, us curves, respectively (see Fig. 36).
SEc. 231 DIFFERENTIAL VECTOR CALCULUS 53
Now Vu3 is perpendicular to the surface
u3(x, y, z) = u3(xo, yo, zo)
so that Vu3 is parallel to the unit vector u3. Hence u3 = h3 Vu3
where h3 is the scalar factor of proportionality between us and
Vu3. Now let dr3 be a tangent vector along the us curve,
dr3' = dss. Obviously dr3 us = dss, and
u1 U2 Us
Du1 Du2 X Vu3 = x = (h1h2ha)-1 (85)
h1 h g s
of of of
Vf = h1 au1 u1+----u2+-''U3
1 1 1
(87)
h2 49u2 h3 au3
In cylindrical coordinates
of
ar R+p+kr a9 az
a(flh2ha )
Now V(f1h2h3) VU2 X Vu3 = Vu1 Vu2 x Vu3, and
out
V (vu2 x vua) = 0
ao that (88) reduces to
l a a a 1a_
+az- r az-
a a
r (91)
V2V
r ar ar
-I-
aB r aB
r = (x2 + y2)i
From (91)
1 d (LV
r dr
r J =0
dr
or r ddV = c1
and
V =cllogr+C2
Finally we obtain the curl of f.
a(f1h1) a(f1h1)
V(f1h1) X VU, = Vul X Vul + Vu2 X Vu1
49U, 49U2
+ a(flhl
aus
Vus X Vul
us
Replacing Vu2 x Vul by -- hlh2, etc., we obtain
Vxf=
ul f a(hsf3)
h2h3 L au2
- a(h2f2 )1 u2 ra(hlfl) ^ a(hsfa)
aua J + hsh1 L aua aul
ua ` a(h1f1) (92)
La(h2f2)
Problems
1. For spherical coordinates, ds2 = dr2 + r2 d92 + r2 sine 9 d(p2
where B is the colatitude and (p the azimuthal angle. Show that
V Z V= 1 Ia I r2 sin B a
r2 sin 9 ar \\\ ar
}-
a9 `
a
(sin 6
aV
aB/J
+ a (sin 9
2. Solve V2V = 0 in spherical coordinates if V = V(r).
3. Express V - f and V x f in cylindrical coordinates.
4. Express V f and V x f in spherical coordinates by letting
a, b, c be unit vectors in the r, 9, (p directions, respectively.
5. Write Eq. (92) in terms of a determinant.
6. Show that V x [(r V9)/sin 0) = V(p where r, 0, (p are spher-
ical coordinates.
7. If a, b, c are the vectors of Prob. 4, show that
as
ar=0'
as
a9= b, as =sin9c
ab =0 ab = -a ab
= cos 9 c
ar a9 a(P
ac ac _ ac
= - sin0a - cos0b
ar=0' a9=0'
8. If x = r sin 9 cos (p, y = r sin 9 sin (p, z = r cos 9, then
the form ds2 = dx2 + dy2 + dz2 becomes
ds2 = dr2 + r2 d92 + r2 sin2 9 d(p2
3
X-f
SEC. 231 DIFFERENTIAL VECTOR CALCULUS 57
where
3
ax" axa
90Y = aI a" ay"
Check this result for the transformation to cylindrical coordi-
nates:
x = r cos 0
y = r sin 0
z=z
and obtain ds2 = dr2 + r2 d02 + dz2.
9. By making use of V2V = V(V V) - V x (V x V), find
V2V for V = v(r)e,, V being purely radial (spherical coordinates).
Find V2V for V = f(r)e, + (p(z)e, in cylindrical coordinates.
10. Find V IV if V = w(r)k x r.
11. Consider the equations
a2S
(X -f 'U)V(V s) + u V2s = p
ate
dt
= Kn (94)
ds
0 Y
x
Fia. 37.
dn
Let us now evaluate - -- and Since b is a unit vector, its
derivative is perpendicular to b and so lies in the plane of t and n.
Moreover, b t = 0 so that on differentiating we obtain
dd_b
t= 0. Hence ` is also perpendicular to t so that
dt
Kn
ds
do
- (Kt + Tb) 95)
ds
db
rn
ds
1=
(dl (a2sin2t+a2cos2t+b2)
2
(.)2
(a2 + b2)
Hence
t = (-a sin ti+acostj+bk)(a2+b2) 4
Now
Kn = d = (-a cos t i - a sin t j)(a2 + b2)-t
so that
K = a(a2 + b2)-i
Also
i j k
b = t x n = -a sin t a cost b (a2 + b2)-1
-cos t -sin t 0
= (b sin t i - b cos t j + ak) (a2 + b2)-+
SEC. 241 DIFFERENTIAL GEOMETRY 61
and
db
ds = rn = (b cos t i + b sin t j)(a2 + b2)-1
so that
T = b(a2 + b2)-1
Problems
1. Show that the radius of curvature of the twisted curve
x = log cos 0, y = log sin 0, z = V2 0 is p = csc 20.
2. Show that r = 0 is a necessary and sufficient condition that
a curve be a plane curve.
3. Prove that T = (r'r"r"').
K
define the unit tangent vector to this curve, this definition being a
generalization of the definition of the tangent vector for the case
d2xa
n = 3. Show that the vector ) a = 1, 2, . . . , n, is normal
ds2
to the tangent vector, and define the unit principal normal n,
and curvature K, by the equations
d2xa dta
Klnla, a = 1, 2, ..,n
d82 - ds =
a
Show that a = 1, 2, . . . , n, is normal to nl and that
ds
dnla =
to
!Is
- K1. Define the second curvature K2 and unit
a-1
la
normal n2 by th e equati ons d = -Klt a + K2n2 a, a = 1, 2,
d i
(s - r) b = 0 (96)
easily seen to be
(s - r) t = 0 (97)
(s - r) n = 0 (98)
Problems
1. Find the equations of the three fundamental planes for the
curve
x = at, y=bt2, z=cta
2. Show that the limiting position of the line of intersection
of two adjacent normal planes is given by (s - r) n = p where
s is the vector to any point on the line.
26. Intrinsic Equations of a Curve. The curvature and torsion
of a curve depend on the point P of the curve and consequently
on the are parameter s. Let is = f(s), r = F(s). These two
equations are called the intrinsic equations of the curve. They
owe their name to the fact that two curves with the same intrinsic
equations are identical except possibly for orientation in space.
Assume two curves with the same intrinsic equations. Let the
trihedrals at a corresponding point P coincide; this can be done
by a rigid motion.
Now
d
Adding, we obtain
0
s
64 VECTOR AND TENSOR ANALYSIS (SEC. 27
so that
constant = 3 (100)
since at P
tl=t2, n1=n2, b1=b2
Since (100) always maintains its maximum value, we must have
dr, _ dr2
tl = t2, n1 n2, bi = b2 so that or r1 = r2 locally.
ds ds
Hence the two curves are identical in a small neighborhood of
P. Since we have assumed analyticity of the curves, they are
identical everywhere.
Problems
1. Show that the intrinsic equations of x = a(9 - sin 8),
y = a(l - cos 8), z = 0 are p2 + s2 = 16a2, 7- = 0, where s is
measured from the top of the are of the cycloid.
2. Show that the intrinsic equation for the catenary
y=a'(ex/a+e-(sla))
2
Fla. 38.
Fia. 40.
r1=r+(c-s)t
ti = dr, dr
Ids +
/
lC - S)
dt ds
ds, - ds - t] dsl
ds
(C - s) K - n
ds1
r + r2
K12
K2 (105)
K2(C - 8)2
28. Evolutes. The curve t'
whose tangents are perpendicu-
lar to a given curve is called the
evolute of the curve. The tan-
gent to r' must lie in the plane
Fla. 41. of b and n of r since it is perpen-
dicular to t. Consequently
rl=r+un+vb
is the equation of the evolute. Differentiating, we obtain
dv) d8
tl=L(-ru+as , b+( +ds)n
Also t1 is parallel to r1 - r = un + vb (see Fig. 41). Therefore
(dv/ds) - UT (du/ds) + Pr
u V
SEC. 28] DIFFERENTIAL GEOMETRY 67
or
uv' - vu'
T=
u2+v2
= dsd- tan-' -vu)
Therefore
ra
= o
Tds=tan--'U -C
--
V
dr,
= = dt ds ds
t' ds1 ds dal - ds,
and
K2 + r2
K12 2
K
t,=--- -
entiating,
dr, A ds ds
= = re
ds, ds ds1 ds1
Therefore
rds=1 and t,=n
ds,
Differentiating,
dt, do ds _ 1
= x,n1 = `-Kt - rn)
dS1 A dS; r
and
K2 + r2
K12 =
T2
SEC. 30J DIFFERENTIAL. GEOMETRY 69
Problems
1. Show that the torsion of the tangent indicatrix is
T(dK/ds) - K(dr/ds)
Ti T2)
K(K2 +
Now
of
=
dr A = (s - r)
b + (s -r).- in, and setting
as ds ds
of
= 0, we obtain (s - r) n = 0. This locus is the rectifying
as
plane. The intersection of f = 0 and as = 0 obviously yields
the tangent lines which are the characteristics. Now
a If
as2
= -t n + (s - r) (-Kt - A) + (s - r) n da
a az
It is easy Y that s = r satisfies f =
Y to verify
as
= 4982 -
= 0, so that
the edge of regression is the original curve r = r(s).
A developable surface, by definition, is the envelope of a one-
parameter family of planes. The characteristics are straight
lines, called generators. We have seen that the envelope of the
osculating planes is the locus of the tangent line to the space
curve P. In general, a developable surface is the tangent surface
of a twisted curve. A contradiction to this is the case of a
cyiinder or cone.
31. Surfaces and Curvilinear Coordinates. Let us consider
the equations
x = x(u, v)
y = y(u, v) (110)
z = z(u, v)
SEc. 32] DIFFERENTIAL GEOMETRY 71
ds2 =
(arudu+-dvN 2
or 2
9r o9r
due + 2 du dv + (av)2 dv2
C_ J au . av
or
where
\12 ao9r .
2
E F G
au/ ' u av' (Ov)
Equation (112) is called the first fundamental form for the surface
r = r(u, v). In particular, along the u curve, dv = 0, so that
(ds) = 1'E du
and similarly (113)
(ds), = VG_ dv
Or
Now and av are tangent vectors to the u and v curves, so
that the parametric curves form an orthogonal system if and
Or Or
only
au av
72 VECTOR AND TENSOR ANALYSIS 1Szc. 33
Example 44. Consider the surface given by
r = r sin 8 cos v i+ r sin 8 sin V j+ r cos 8 k, r = constant
Differentiating,
ar = r cos 0
cos sP i + r cos 0 sin V j - r sin 8 k
d0
ar
= -r sin6sinpi+rsin8coscpj
a st
and
is 2
c1r c1r
F=-e = 0, G= r2 Sin2 0
E a8/ = r2,
so that ds2 = r2 d82 + r2 sin2 0 dcp2 and the 0-curves are orthog-
onal to the 9-curves. Of course the surface is a sphere.
33. Surface Curves. By letting u and v be functions of a
single variable t, we obtain
r = r[u(t), v(t)] (114)
which represents a curve on the surface (111). Along this curve,
dr = ` du + Wt dt. dr is completely determined when du
(arau ar dvl
dt av 1
and dv are specified, so that we will use the notation (du, dv) to
specify a given direction on the surface. Now consider another
curve such that ar =
+ av av, where su and av are the
su
au
differential changes of u(t) and v(t) for this new curve. Now
dr or = E du au + F(du av + dv au) + G dv av (115)
so that two curves are orthogonal if and only if
Edu su +F(duav +dv &u) +Gdv av = 0
or
dv
E+FCa6Vu+au/+Gd -=0 (116)
8u tvhl
that au need not be a unit tangent vector to the u curve since the
parameter u may not represent are length. Since
(ds),, = 1/E du
a necessary and sufficient condition for u to be arc length is that
E = 1. We define the unit normal to the surface as
n _ (Or/au) x (ar/av)
(118)
(ar/au) x (ar/av)
Kn =
edue+2fdudv+gdv2
ds2
edue+2fdudv+gdv2 (120)
Kn
Edue+2Fdudv+Gdv2
SEC. 361 DIFFERENTIAL GEOMETRY 75
where we define
a2r a2r
e=n-au2'
a2r
f
_n. auav' (121)
Fia. 43.
the normal curve in the same direction (du, dv) with normal n at
P (Fig. 43). We have
r"
K
=n r"-
K
since n rl" = r" n for two curves with the same (du, dv) [see
(119)]. Therefore
COS0= -
Kn
so that
K = K,, sec 0 (122)
2! -
(a2r
Au2 + 2
Consequently
a2r
au av Au AV
a2r
+ av2 Av2 +
1/ z z
D= Av2
Problems
1. For the paraboloid of revolution
r=ucosvi+usinvj+u2k
show that E = 1 + 4u2, F = 0, G = u2, e = 2(1 + 4u)`}, f = 0,
g = 2u2(1 + 4u2)-}, and find the normals to the surface and the
normal curvature for the direction (du, dv).
2. What are the normal curvatures for directions along the
parametric curves?
3. Find the second fundamental form for the sphere
r=rsin0coscpi+rsin6sincpj+rcosOk
r = constant.
4. Show that the curvature K at any point P of the curve of
intersection of two surfaces is given by
K2 Sln2 0 = K12 + K22 - 2K1K2 COS 0
where 9j, K2 are the normal curvatures of the surfaces in the direc-
tion of the curve at P, and 0 is the angle between their normals.
5. Let us make a change of variable u = u(u, v), v = v(u, 1).
Show that E, F, G transform according to the law
lz 2
E
Caul
+ 2F
au au
+G(av
\au
SEC. 37] DIFFERENTIAL GEOMETRY 77
au au au av av au av av
F
=Eaudo +F (aft do +audo )+Gag do
au\2
au av (avl 2
E Ii + 2F avav
UP- + G `avl
and that E due + 2F du dv + G dv2 = E due + 2F du do + G W.
Also show that
( au av (OV\2 1
e=± e
au +
Zf
au au + g au J
(au)2
au au au av av au av av
au av au av au av au CIO]
-± [e
(au)2 au av (av12
g av + 2f avav+ g av J
37. Principal Directions. From (120) we have
(K .E - e) due + f) du dv + g) dv2 = 0 (124)
or
A due + 2B du dv + C dv2 = 0
This quadratic equation has two directions (du, dv), (Su, Sv),
which give the same value for x,,. These two directions will
coincide if the quadratic equation (124) has a double root. This
is true if and only if
B2-AC= (K.F'-f)2- 0
or
, 2(F2 - EG) + gE - 2fF) + (f2 - eg) = 0 (125)
Moreover, we have
du
- B and d = - C if B2 - AC = 0,
so that A
(K. E - e) du + (x F - f) dv = 0
6126)
f) du + (xnG - g) dv = 0
The solutions of (124) give the two directions for a given x,,.
When x is eliminated between (124) and (125), the two directions
coincide and satisfy
HeG+gE-2fF (128)
2(EG - F2)
f2eg
K= F2EG (129)
r= ucospi+usine'j+cook
We have
Or
=cosSPi+sinrpj, = -u sins
au
2 zr
(ar)2 (Or\2
au- 1, F= au - am 0' G= am = u2 + c2
SEC. 371 DIFFERENTIAL GEOMETRY 79
Also
n = (ar/au) x (,9r/app) u2)-}
= (c sin p i - c cos v j + uk) (C2 +
(ar/au) x (ar/ap)j
82r f=n a2r = -c(c2 + u2)-#
au ago
2
f=F=0 (131)
Problems
1. Find the lines of curvature on the surface
x=a(u+v), y=b(u-v), z=uv
2. Show that the principal radii of curvature of the right conoid
x = u cos v, y = u sin v, z = {f (v) are given by the roots of
f'=K2 - of"(u2 + fF )iK - (u2 + ft )2 = 0
3. The surface generated by the binormals of the curve r = r(s)
is given by R = r + ub. Show that the Gauss curvature is
K = -.r2/(1 + r2u2)2. Also show that the differential equation
of the lines of curvature is
Or Or
normal at Q, where dr = PQ =
an
du +
- dv. Let the direc-
Or ar
tion of 1' be given by Sr = Su Sv. Since Sr lies in both
au + av
planes, we must have Sr n = 0 and Sr (n + dn) = 0. These
two equations imply Sr do = 0, or
=0
(au
Su
+ av
Sv) Can du +
a dv) (132)
Expanding, we obtain
far
(_au
-) du bu + I (-av -) Sv du + (auau
anav Or anau Or an
aU) au dvJ
ar an
Sv dv = 0 (133)
+ av av
which implies
an ar a2r
au au
-n- au,
- = -e
Similarly
an Or an Or -f
av au au av
an Or
9
We see that the asymptotic directions are those for which the
second fundamental form vanishes. Moreover, the normal curva-
ture rc, vanishes for this direction.
If e = g = 0, f 0, the solution of (135) is u = constant,
v = constant, so that the parametric curves are asymptotic lines
if and only if e = g = 0,f3PK 0.
Example 46. Let us find the lines of curvature and asymptotic
lines of the surface of revolution z = x2 + y2. Let x = u cos v,
y = u sin v, z = u2, and
r=ucosvi+usinvj+u2k
We obtain
Therefore
z
=2(1+4u)-}, f=n aua2rav =0
au2
z
g=n = 2u2(1 +4 U2)-f
2
SEC. 40] DIFFERENTIAL GEOMETRY 83
ar ar
Also F = -- - = 0, so that f = F = 0, and from (131) the
au av
parametric curves are the lines of curvature. The asymptotic
lines are given by due + u2 dv2 = 0. These are imaginary, so
that the surface possesses no asymptotic lines.
Problems
1. Show that the asymptotic lines of the hyperboloid
r=acos0seeipi+bsin0sec4, j+ctan 4k
are given by 0 ± ¢ = constant.
2. The parametric equations of the helicoid are
x=ucosv, y=usinv, z=cv
Show that the asymptotic lines are the parametric curves, and
that the lines of curvature are u + V 'u2+ c2 c2 = Ae}°. Show
that the principal radii of curvature are ± (u2 + c2)c-1.
3. Prove that, at any point of a surface, the sum of the normal
curvature in conjugate directions is constant.
4. Find the asymptotic lines on the surface z = y sin x.
40. Geodesics. The distance between two points on a surface
(we are allowed to move only on the surface) is given by
Among the many curves on the surface that join the two fixed
points will be those that make (136) an extremal. Such curves
are called geodesics. We wish now to determine the geodesics.
To do this, we require the use of the calculus of variations, and
so we say a few words about this important method.
Let us first consider the integral
fQ(xt,yi) (1
(x,, VS)
+ y'dx (137)
We might ask what must be the function y = y(x) joining the two
points P and Q which will make (137) a minimum. The reader
might be tempted to say, y' = 0 or y = constant, since the inte-
grand is then a minimum. But we find that y = constant will
84 VECTOR AND TENSOR ANALYSIS [SEC. 40
not, in general, pass through the two fixed points. Hence the
solution to this problem is not trivial. We now formulate a
y more general problem: to find
y = y(x) such that
j'f(x, y, y') dx (138)
dJ
is an extremal for a = 0. Consequently
da0 = 0 or
dJ Ib\ay,+a- (140)
da a-0
=
since
of of 81' of 81" of of
as aY as + aY' as if + C7F
'p,
(142)
dx (
f y yl ay
1J, + a
y
yy of
ay
d (af 1
y/ dx TO
y [Of _ d
(t\1 = 0 from (143)
lay dx ay'
f-y y,
aof = constant (144)
so that
J
(1 +y') - Y'(1 y'2), a
and
y'=
and finally
y= ± av-1x+ 0 (145)
d (ftaf of
axa = o (146)
dt a
X.
for a = 1, 2, . .. , n with is = The superscripts are not
powers but labels that enable us to distinguish between the vari-
ous variables. The formulas (146) are a consequence of the fact
that fo f dt must be an extremal when x`(1) is allowed to vary
while we keep all other x1 fixed, j = 1, 2, ... , i - 1, i + 1,
,n.
Let us now try to find the differential equations that u(t) and
v(t) must satisfy to make (136) an extremal. We write
s= f" (E,42 + 2Fuv + Gv2)} dt
dt \a4/ au - 0 (147)
d off` of
o (148)
dt avl av
where
f = (E,42+ 2Fuv + Gv2)i = dt' E = E(u, v), etc.
SEc. 401
Now
af
au
Eic + Fv
of
au
ice ---
CIE
au
+ 2uv
2f
OF
au
+ v2
aG
au
87
d OF -n/A
opu
(Eic + Fv) = 2( u2 aE + 2uv + 62
ds
while similarly (148) yields (150)
\
d (Fu+Gv) _2(u2av +24vav +v2av/
In Chap. 8 we shall derive by tensor methods a slightly differ-
ent system of differential equations.
Example 49. Consider the sphere given by
r = a sin 0cos(pi+a sin 0 sin cpj+acosOk
where ds2 = a2 do2 + a2 sin2 0 dp2 so that E = a, F = 0,
G = a2 Sin2 0,
and
OE _ aEE _ OF _ OF _ aG _ 8G _
80 ac av a9 ap 0' aB - a sin 29
2
ds ds
are a, 0, 0 and a, 0, 0. Hence sin' 8 = 0, and = 0, so that
0. Hence the geodesic is the are of the great circle joining
the two fixed points.
Example 50. Let us find y(x) which extremalizes
f y(1 + y")} dx
Since f = y(1 + y")# = f(y, y'), we can apply (144) to obtain
a first integral. We obtain y(1 + y'')t -- y''y(1 + y'')-l = a-1,
and simplifying this expression yields y' = ± (a2y2 - 1)I. A
further integration yields ay = cosh (0 ± ax). These are the
curves (catenoids) which have minimum surfaces of revolution.
Problems
1. Find the geodesics on the ellipsoid of revolution
x2+ z2 2
a2 + b2 = 1
Problems
1. What are the limit points of the set 0 < x < 1? Is the
set closed? Open? What are the boundary points?
2. Repeat Prob. 1 with the point x = removed.
3. Show that the set of all boundary points (the boundary)
of a set S is closed.
4. Prove that the set of all limit points of a set S is closed.
5. Prove that the complement of an open set is closed, and
conversely.
6. Why is every finite set closed?
7. Prove that the set of points common to two closed sets is
closed. The set of points belonging to both Si and S2 is called
the intersection of Sl and S2, written SI n82-
8. Prove that the set of points which belong to either S, or 82
is open if S, and S2 are open. This set is called the union of S,
and S2, written S, U S2-
9. An infinite union of closed sets is not necessarily closed.
Give an example which verifies this.
10. An infinite intersection of open sets is not necessarily open.
Give an example which verifies this.
Supremum. A number s is said to be the supremum of a set of
points S if
1. x in S implies x 5 s
2. t < s implies an x in S such that x > t
Example 51. Let S be the set of rationale less than 1. Then
1 is the supremum of S, for (1) obviously holds from the defini-
tion of S, and if t < 1, it is possible to find a rational r < 1 such
that t < r, so that (2) holds. We give a proof of this statement
in a later paragraph.
Example 52. Let S be the set of rationals whose squares are
less than 3, that is, SI x2 < 3. Certainly we expect the to
be the supremum of this set. However, we cannot prove this
without postulating the existence of irrationals. We overcome
92 VECTOR AND TENSOR ANALYSIS [SEC. 41
this by postulating
Every nonempty set of points has a supremum (152)
Hence the rationals whose squares are less than 3 have a supre-
mum. It is obvious that we should define this supremum as the
square root of 3.
The supremum of a set may be + oo as in the case of the set
of all integers, or it may be - oo as in the case of the null set.
The infemum of a set S is the number s such that
1. x in S implies s < x
2. t > s implies an x in S such that x < t
Example 53. Let a > 0 and consider the sequence a, 2a, 3a,
... , na, ... . If this set is bounded, there exists a finite
supremum s. Hence an integer r exists such that ra > s - (a/2)
so that (r + 1)a > s + (a/2) > s, a contradiction, since na 5 s
for all n. Hence the sequence Ina} is unbounded. This is the
Archimedean ordering postulate.
Example 54. To prove that a rational exists between any two
numbers a, b. Assumed: a > b > 0, so that a - b > 0. From
example 53, an integer q exists such that q(a - b) > 1, or
qa > qb + 1. Also an integer p exists such that p 1 = p > qb.
Choose the smallest p. Thus p > qb ? p - 1. Hence
qa>gb+1zp>qb,
and a > p/q > b. Q.E.D.
With the aid of (152) we are in a position to prove the well-
known Weierstrass-Bolzano theorem.
"Every infinite bounded set of points S has a limit point." The
proof proceeds as follows: We construct a new set T. Into T we
place all points which are less than an infinite number of S. T is
not empty since S is bounded below. From (152), T has a
supremum; call it s. We now show that s is a limit point of S.
Consider any neighborhood N of s. The points in N which are
less than s are less than an infinite number of points of S, whereas
those points in N which are greater than s are less than a finite
number of points of S. Hence N contains an infinite number of
S, so that the theorem is proved. We have at the same time
proved that s is the greatest limit point of S. A limit point may
or may not belong to the set.
SEC. 41) INTEGRATION 93
Problems
1. The set 1/2, 1/3,. . . , 1/n, . . ., 1, 2, 3, . . . is
unbounded. Does it have any limit points? Does this violate
the Weierstrass-Bolzano theorem?
2. Prove that every bounded monotonic (either decreasing or
increasing) sequence has a unique limit.
3. Prove that lim r" = 0 if Irl < 1. Hint: The sequence
r, r2, . . . , r", . . . is bounded and monotonic decreasing for
r > 0, and r"+1 = rr"
4. Show that if P is a limit point of a set S, we can pick out a
subsequence of 8 which converges to P.
5. Show that (152) implies that every set has an infemum.
6. Show that removing a finite number of elements from a set
cannot affect the limit points.
7. Prove the Weierstrass-Bolzano theorem for a bounded set
of points lying in a two-dimensional plane.
8. Let the sequence of numbers sl, 82, .. . , sn, . . . satisfy
the following criterion: for any e > 0 there exists an integer N
such that Is,,+F - snI < e for n ? N, p > 0. Prove that a unique
limit point exists for the sequence.
9. A set of numbers is said to be countable if they can be
written as a sequence, that is, if the set can be put into one-to-one
correspondence with the positive integers. Show that a count-
able collection of countable sets is countable. Prove that the
rationals are countable.
10. Show that the set S consisting of x satisfying 0 5 x 5 1
is uncountable by assuming that the set S is countable, the
numbers x being written in decimal form.
Theorem of Nested Sets. Consider an infinite sequence of
nonempty closed and bounded sets S1, S2, . . . , Sn, . such
that Sn contains that is, S1 M S2 D Sa M . There
exists a point P which belongs to every Si, i = 1, 2, 3, . . . .
The proof is easy. Let P1 be any point of S1, P2 any point of
S2j etc. Now consider the sequence of points P1, P2, ... ,
P,,
. . . . This infinite set belongs to S1 and has a limit point P
which belongs to 8, since Sl is closed. But P is also a limit of
P, Pn+1, ... , so that P belongs to Sn. Hence P is in all Sn,
n = 1, 2, . . . .
Diameter of a Set. The diameter of a set S is the supremum of
all distances between points of the set. For example, if S is the
94 VECTOR AND TENSOR ANALYSIS [SEc. 41
1
T,,I x such that Ix - 1
n < (n + 1)2
Problems
1. In the proof of (c) we exclude f(c) > s. Why?
2. If f (x) is continuous on (a, b), show that the set of values
{f(x) } is closed.
3. If f (x) has a derivative at every point of (a, b), show that
f (x) is continuous on (a, b).
4. If f (x) has a derivative at every point of (a, b), show that a
c exists such that f(c) = 0, a < c < b, when f (a) = f (b) = 0.
5. If f (x) has a derivative at every point of (a, b), show that
a c exists such that f(b) - f(a) _ (b - a)f'(c), a < c < b.
6. Show that if two continuous functions f(x), g(x) exist such
that f (x) = g(x) for the rationals on (a, b), then f (x) = g(x) on
(a,
7. Given the function f (x) = 0 when x is irrational, f (x) = 1/q
when x is rational and equal to p/q (p, q integers and relatively
prime), prove that f (x) is continuous at the irrational points of
(0, 1) and that f(x) is discontinuous at the rational points of
this interval.
44. Cauchy Criterion for Sequences. Let x1, x2, ... ,
xn, ... be a sequence of real numbers. We say that L is the
limit of this sequence, or that the sequence converges to L, if,
given any e > 0, there exists an integer N depending on a such
that IL - xnl < e whenever n > N(e). However, in most cases
we do not know L, so that we need the Cauchy convergence
criterion. This states that a necessary and sufficient condition
that a sequence converge to a limit L is that given any e > 0,
there exists an integer N such that Ixn - xml < e for n >_ N,
m N. That the condition is necessary is obvious, for
IL - xnl < e/2,
IL xn,l < e/2 for m, n > N implies Ix,, - xml < e for m, n >_ N.
The proof of the converse is not as trivial. Choose any e/2 > 0.
Then we assume an N exists such that Ix,, -- xml < e/2 for
m, n ? N. Hence Ixn+ < I xNl + (e/2), n N, so that the
sequence is bounded. We ignore xi, X2, . . . , xN_, since a finite
number of elements cannot affect a limit point. From the
Weierstrass-Bolzano theorem this infinite bounded set has at
least one limit point L. Hence, given an a/2, there exists an x.,
with n > N, such that IL - xAi < e/2. But we also have
98 VECTOR AND TENSOR ANALYSIS [SEC. 45
I xm - x,,l < e/2 for all m, n > N. Hence IL - xml < e for all
m > N. Q.E.D.
Problems
1. Show how the convergence of a series can be transformed
into a problem involving the convergence of a sequence.
2. Show that the Cauchy criterion implies that the nth term
of a convergent series must approach zero as n -- oo.
3. Show that the sequence 1, 1/2, 1/3, . . . , 1/n, . . . con-
verges by applying the Cauchy test.
45. Regular Arcs in the Plane. Consider the set of points
in the two-dimensional plane
P2 such that the set can be repre-
sented in some coordinate
system by x = f (t), y = sp(t),
a S t < 0, where f(t) and ap(t)
are continuous and have con-
tinuous first derivatives. Such
curves are called regular arcs.
A regular curve is a set of
points consisting of a finite
number of regular arcs joined
one after the other (see Fig.
P4 45).
Fia. 45. PoPI, P1P2, P2t 3, P3P4 are
the regular arcs joined at P1,
P2, P3. Notice that there are at most a finite number of dis-
continuities of the first derivatives. In Fig. 45 the derivatives
are discontinuous at P1j P2, P3.
46. Jordan Curves. The locus i x = f (t) a < t 5 fl, will be
y = P(t),
called a Jordan curve provided that f (t) and (p(t) are continuous
l curve correspond to two
and that two distinct points on the
distinct values of the parameter t (no multiple points).
A closed Jordan curve is a continuous curve having f(a)
,o(a) = 9(fl) but otherwise no multiple points.
From this we see that a Jordan curve is always "oriented,"
that is, it is always clear which part of the curve lies between two
points on the arc, and which points precede a given point. We
shall be interested in those curves which are rectifiable, or, in
Sr:c. 47] IN TkGRA TION 99
If the sums of this type for all possible finite subdivisions are
bounded, that is, if
n
I f(xi) - f(xc-1)I < A < -o (153)
s=x
f (O) = 0
Let us subdivide 0 5 x <_ 1 into the intervals
(n+1) Sx<n,
n = 1, 2, .. . , N. Now f(1 /n) = (1/n) sin (irn/2) so that
(I 2
=1 +3
+ 25 + N2
n
N
If \n/ f \n } )I
We cannot bound this sum for all finite N since the series diverges
as N -> w. N was chosen as an odd integer.
100 VECTOR AND TENSOR ANALYSIS [SEc.48
B = I IIf(t;) -
f-l
f(4--1)I + Ip(k) - v(4-1)I I
so that
A==<Bs /2A
Consequently if the curve is rectifiable, f (t) and '(t) are of
bounded variation, and conversely, if f (t) and ap(t) are of bounded
variation, A is bounded, and hence the curve is rectifiable.
If f(t) and V'(t) are continuous, then from the law of the mean,
I f(4) - f (ti--1) I
= I f (ii) (ti - ti-1) I < A I ti - ti-il, where A is the
supremum of If(t)l, a 5 t =< # and 4_1 5 Ti 5 ti.
SEC. 49] INTEGRA TION 101
n n
E If()
i-1
- f(t;-1)I < A 11(4-
i=1
- till = A(# - a)
Similarly, p(t) is of bounded variation, so that the curve is recti-
fiable. Under these conditions it can easily be shown that the
arc length is given by
J = K = fr f(x, y) ds = LfI(t), + * )} dt
R=R1+R2
(ii)
FIG. 47.
FIG. 48.
dsds = dr = dx i + dy j + dz k
we have
ds = fr X dx + Y dy + Z dz (156)
Jr (f ds/
ff dr = [x(t) dt (157)
dt + Y(t) dt + Z(t) dtJ
fo'
and
ft t of f
dr = f.' (t2 + 2t7) dt =
Jr JO
(x2 + 2x') dx = A
(c)
r (1, 1) ((1, 1)
x2 dx + y$ dy - 3
xa 1+y'1- 7
J(o, 0) f dr = J(0, 0) 0 4 (0 12
, o)
f(1'1)ydx-xdy=
ro, o)
f' x2dx-x(2xdx) _ -
Next we compute the integral by moving along the x axis from
x = 0 to x = 1 and then along the line x = 1 from y = 0 to
y = 1. We have
f
( o, o)
(0,0))
0, o)
f(" 1, o)
JAfdrJAVdrLd(B)(A) (158)
Our final result in (158) depends only on the value of (p(x, y, z)
when evaluated at the points A(xo, yo, zo), B(xi, yi, z,) and in no
way depends on the path of integration. If our path is closed,
then A = B and (p(B) = p(A), so that the line integral around
any closed path vanishes if f = Vip. However, the region for
which f = must be simply connected. Let us consider the
following example.
Example 57. Let
- yi xi
f LL
=x2+y2 -rx2+y2
Then f = VV where p = tan-' (y/x), and if we integrate f over
the unit circle with center at the origin, we have
d (tan_ixJ)fo2T dO 2r
so that our line integral does not vanish. The region for which
f = V(p is not simply connected since p is not defined at the
origin.
We now prove that if f f dr is independent of the path, then
f is the gradient of a scalar V. Let
JP(z,y.z) f dr JP(fdr)ds (159)
P(x, y, z) = o(xo, YO, zo)
Now
(P(x + Ax, y, z) - (P(x, y, z)
Ax
and since the line integral is independent of the path joining P
to Q, we choose the straight line from P to Q as our path of
integration, that is, dr = dx i. Thus
lim
cP(x+Ox,y,z) --P(x,y,z) =av
AX-40 Ax ax
x+Ax
iim f z
X (x, y, z) dx
= X (x, y, z)
AX-0 Ox
106 VECTOR AND TENSOR ANALYSIS [SEC. 52
Now
a'P
ax = X (x, y, z)
aq
_ -L.
dx+
f.. Y(xo,y,z)
ay -0Y
f ax dx + Y (x,), y, z)
= Y(x, y, z) - Y(xo, y, z) + Y(xo, y, z)
= Y(x, y, z)
Sic. 53] INTEGRATION 107
f axi+ayj+azk=Drp
V xf =0 (163)
and
JX
= 2xy9 dx + 02ez dy + 02. 0 . 9 dz
= x2yeZ
ABCD f dr
f(u,v+dv) =f(u,v)+(-dvv)f
Six. 53] INTEGRATION 109
f+
av - fi au} dude
(164)
Now
for ar an Or
(v x f)
(au x av = (° x f) x au av
J v)fI ar
C\au °/ av ^ L \av au
Hencee
dr = (v x f) a x
a du dv (165)
fiABCD f
ar ar
Now d x a du dv = area of sector ABCD in magnitude, and
au av
its direction is along the normal. We define
Or Or
dd =
aux& du dv (166)
so that
IABCDfdr= (V
We now sum over the entire network. Interior line integrals
will cancel out in pairs leaving only fir f dr. Also
I
over surface 8
(V f f (V
a
as the areas approach zero in size. We thus have Stokes's
theorem :
fir f dr = f f (V x f) dd (167)
8
110 VECTOR AND TENSOR ANALYSIS [SEC. 54
Comments
1. The reader may well be aware that (165) does not hold for
a mesh that has r as part of its boundary. This is true, but
fortunately we need not worry about the inequality. The line
integrals cancel out no matter what subdivisions we use, and for
a fine network the contributions of those areas next to r con-
tribute little to f f (V x f) dd. The limiting process takes care
of this apparent negligence.
2. We have proved Stokes's theorem for a surface of the type
r(u, v) discussed in Chap. 3. The theorem is easily seen to be
true if we have a finite number of these surfaces connected con-
tinuously (edges).
3. Stokes's theorem is also true for conical points, where no
dd can be defined. We just neglect to integrate over a small area
covering this point. Since the area can be made arbitrarily
small, it cannot affect the integral.
4. The reader is referred to the text of Kellogg, "Foundations
of Potential Theory," for a much more rigorous proof of Stokes's
theorem.
54. Examples of Stokes's Theorem
Example 59. Let r be a closed Jordan curve in the x-y plane.
Let f = -yi + xj. Applying Stokes's theorem, we have
ff
ff ax ay az
S
-y 0 x
=2ffdydx=2A
S
or
f- dr = f f V x f- dd > 0, a contradiction
s
Example 61. We see that an irrotational field is characterized
by any one of the three conditions:
(i) f = VP
(ii) Vxf=0 (169)
(iii) f dr = 0 for every closed path
ff x a) dd
s
ff
s
f f dd x
f= a a
we have
ff
s
[a(V g) - (a V)gl dd
=a- f f (V.g)dd-a f f
V dd (summed). There-
fore
xdr = a- f f (ddxv) xg
and
dr x g = f f (dd x V) x g (174)
S
We notice that in all cases
0 dr * f= Jf(doxv)*f (175)
S
Problems
1. Prove that fdr = 0 from (175).
2. Show that L, 'dr x ri taken around a curve in thex-y plane
is twice the area enclosed by the curve.
3. If f = cos y i + x(1 + sin y)j, find the value of ff - dr
around a circle of radius r in the x-y plane.
4. Prove that fr dr = 0.
5. Prove that ff dd x r = fir 2 dr.
s
6. Prove that f u Vv dr = - ,f v V u dr.
7. Prove that u Vv - dr = JJvu x Vv dd.
s
8. If a vector is normal to a surface at each point, show that
its curl either is zero or is tangent to the surface at each point.
9. If a vector is zero at each point of a surface, show that
its curl either is zero or is tangent to the surface.
10. Show that fi a x r dr = 2a. Jf dd, if a is constant.
S
11. If fi E dr = -cat Jf
s
B dd for all closed curves, show
that V x E =
---
C at
12. By Stokes's theorem prove that V x (V(p) = 0.
13. Show that fi dr/r = f f (r/r3) x dd where r = Id.
8
14. Find the vector f such that xy = f(O,(X'o, Y,)
o)
f dr.
15. If f = r/r3, show that V x f = 0 and find the potential yo
such that f = Vsp.
16. Show that the vector f = (- yi + xj)/(x2 + y2) is irrota-
tional and that (f r f dr = 2w, where r is a circle containing the
origin. Does this contradict Stokes's theorem? Explain.
17. Show that ff v x f dd = 0, where S is a closed surface.
s
18. Let C1 and C2 be two closed curves bounding the surfaces
S1 and S2. Show that
114 VECTOR AND TENSOR ANALYSIS [SEC. 55
JfJ(v.f)dr (176)
V
Jff.dd (177)
JJf.do=fJf(v.f)d.r (178)
8 V
SEC. 55] INTEGRATION 115
lira
fAs
Ar-.0 DT
Jff.dd
As
div f = lim AT
(179)
AT-40
We can write
tardivf= JJ (180)
Inthe derivation of (181) use has been made of the fact that for
each internal dd there is a -dd, so that all interior surface
integrals cancel in pairs, leaving only the boundary surface S
as a contributing factor. The sum of the ei Ar; vanishes in the
limit, for I Zet Aril 5 ZIel... OT S IEIo,.=V, and if div f is continu-
ous, IeI,. - 0 as 2r --i 0.
By choosing rectangular parallelepipeds and using the method
of Sec. 20, we can show that
fff.do
AS clu c3v aw
div f = slim (182)
AT ax + ay + 49Z
116 VECTOR AND TENSOR ANALYSIS [SFC. 55
ffJ V. () dr f8
r ad+J1
E
r"d (183)
fJ
S
f dd = 4irq (185)
x =--- ay
ay
a#
az
as ay
Y= (186)
az ax
Z =--- as
ax
as
ay
y=- xos
Ydx+r(y,z)
Hence
f aY aZ ," ar _ aa
ay az JZo ay + az / + ay az
fx ax
dx
+ ar - as
ax ay az
or ao
= X (x, y, z) - X (xo, y, z) +
ay - az
In general
g=[ faZdx+v(y,z)]j- foYdxk+Vp (187)
v arbitrary
and
g = (x2 + y2 + z2)#(y2 + (-zj + yk) + VSP
z2)
Subtracting, we obtain
JJJov2odr+JfJvo.vodr=JJove.do
R R
1/r, then V<p tends to zero like 1/r2, and f f cc V(p V. dd tends to
s
zero like 1 /r as r oo.
Example 71. Let f = f(x, y, z)a, where a is constant. Apply
ing (178), we obtain
120 VECTOR AND TENSOR ANALYSIS [SEC. 55
s
fffdd
= fff
V
7fdr (190)
Problems
1. Prove that f f dd x f = f f f (V x f) dr.
s y
2. Prove that ff dd = 0 over a closed surface S.
s
3. If f = axi + byj + czk, a, c constants, show that
b,
Jff.do = 4(a + b + c), where S is the surface of a unit
s
sphere.
fffdd
As
4. By defining grad f = Jim snow that
4,-+O OT
gradf = afi+
ax
afj+afk
ay az
fff.dd
5. By defining div f = lim °3 , show that
A?--+O GT
div f =
r2 sin 0 Lar
(r2 sin of,) + e (r sin 9 fe) + a (rf,)]
f fJ JfJ (P 0fdr.
SEC. 55] INTEGRATION 121
7. If w = V x v, v = V x u, show that
'9
v2dr = f f u xv.dd+ f f
R S R
8. Show that
f J f w Vu . Vv dr = f f uw Vv . dd - f if u V . (w Vv) dr
f J f V2dr = Jf pv.dd.
10. Show that f f IrJ2r . da = 5 fJJ r2 dr.
and that
dTl
1s= dd2 fs, log r ddl = - fv,
cIT2
f V1 r
20. IfV - f = V xf ='1i+¢2j+4,ak,f = Xi + Yj + Zk,
show that V2X =
49 O
- a3y - z2' and find similar expressions
ffv x w dd = ff
av au
w dr = ax ay
dydx
s s
(192)
{ w1 dr = jJv au av
=
Jf ax-
ay
dy dx
V2v=+a
a2v
axey =
a2v
2
(194)
and
au av au av _
ax ax + ay ay
If functions u(x, y), v(x, y) satisfy Eqs. (193) we say that
they are harmonic conjugates. The importance of such func-
y v
P (x, yl
Ix
Fia. 51.
tions is due to the fact that they satisfy the two-dimensional
Laplace's equation given by (194). If u satisfies V2u = 0, we
say that u(x, y) is harmonic.
Let us now consider two rectangular cartesian coordinate
systems, the x-y plane and the u-v plane (Fig. 51). Let
r=xi+yj.
Now to every point P(x, y) there corresponds a point Q(u, v)
given by the transformation u = u(x, y), v = v(x, y). Hence the
vector w = u(x, y)i + v(x, y)j corresponds to the vector
r -- xi+yj.
If now P(x, y) traverses a curve C in the x-y plane, Q(u, v) will
trace out a corresponding curve r in the u-v plane.
The curve u(x, y) = constant in the x-y plane transforms into
the straight line u = constant in the u-v plane. Similarly,
124 VECTOR AND TENSOR ANALYSIS [Ssc. 56
Y
I
0'
Fia. 52.
(4)
(3)
------------
Fra. 53.
au
dv=axdx+-dy=aydx - ax dy
a dx - jyOU dy + c
v = fa au (195)
1! 8x
Problems
1. Find the harmonic conjugate of xa - 3xy2, of ex coo y, of
x/(x2 + y2).
2. Show that u(x, y) = sin x cosh y and v(x, y) = cos x sinh y
are conjugate harmonics and that the curves u(x, y) = constant,
v(x, y) = constant are orthogonal. What do the straight lines
y = constant transform into?
3. If u(x, y), v(x, y) are conjugate harmonica, show that the
angle between any two curves in the x-y plane remains invariant
under the transformation u = u(x, y), v = v(x, y), that is, the
transformed curves have the same angle of intersection.
CHAPTER 5
STATIC AND DYNAMIC ELECTRICITY
E=- 4'a
s
ra (196)
a=1 ra
where ra represents the vector from P to the charge qa.
We have seen in Example 25 that V V. (r/r3) = 0. Conse-
quently,
V.E=0 (197)
- z,)2]1, and
l a _
V1
ri
ri
ot
r ri an
rr;3
(xi - )i -I- (yi- 7/)i + (zi -t)k
r;
127
128 VECTOR AND TENSOR ANALYSIS [SEC. 58
Jf.do=JfJv.()dr=0 (201)
JJE.dd=4irQ (202)
s
Fic. 54.
we call the body an insulator. Actually there is no sharp line
of demarcation between conductors and insulators. Every body
possesses some ability in conducting electrons.
At the surface of a conductor the field is normal to the surface,
for any component of the field tangent to the surface would cause
a flow of current in the conductor, this again being contrary to
the assumption that the field is static (no large-scale motion of
electrons occurring). Such a surface is called an equipotential
surface. The field is everywhere normal to an equipotential
surface, for the vector E = -V(p is normal everywhere to the
surface V(x, y, z) = constant.
Example 76. Consider a uniformly charged hollow sphere E.
We shall show that the field outside the sphere is the same as if
130 VECTOR AND TENSOR ANALYSIS [SEC. 58
Fic. 56.
From symmetry the field is normal to the plates. We apply
Gauss's law to the surface in Fig. 55 with unit cross-sectional
area.
f f E do = E = 4ara (204)
so that the field is uniform.
Example 78. We now determine the field in the neighborhood
of a conductor. We consider the cylindrical pillbox of Fig. 56
and apply Gauss's law to obtain
SEC. 58] STATIC AND DYNAMIC ELECTRICITY 131
EA = 4raA
or
E = 47ro N (205)
where a is the charge per unit area and N is the unit normal vector
to the surface of the conductor.
Example 79. Force on the surface of a conductor. We consider
a small area on the surface of the conductor. The field at a
point outside this area is due to (1) charges distributed on the
rest of the conductor (call this field E1), and (2) the field due to
the charge resting on the area in question, say E2 (see Fig. 57).
From Example 78, El + E2 = 41rv. Now the field inside the
Ei+E2
Fia. 57.
conductor at the point P' situated symmetrically opposite P is
zero from Example 75. The field at P' is E1 - E2 = 0. Thus
El = 2o per unit charge. For an area dS the force is
dE = (2av) (v dS) = 27rv2 dS (206)
xa x2 xl
x
q3 0 qa qi
I
II `
1
1
I
I
Fia. 58.
charges exist between the planes x = A, x = B, show that the
equation of a line of force is
n
qa(x - xa)[(x - xa)z + y2]i = constant
a-1
6. Point charges +q, -q are placed at the points A, B. The
line of force that leaves A making an angle a with AB meets the
plane that bisects AB at right angles in P. Show that
sin 2 = sin (+ PA B)
E_ -va-1 r,
SEC. 591 STATIC AND DYNAMIC ELECTRICITY 133
For a continuous distribution of charge density p, we postulate
that the potential is
= fff pdr
(207)
E = f f f Ta dr (208)
E(P) = lim f uJ -t dr
r +0
converges.
Now from Gauss's law
ff
8
f f f pdr
v
134 VECTOR AND TENSOR ANALYSIS [SEC. 59
V E = 4irp (210)
V 20
r L or
(r Or + e r t90/ + 8z \r 8z/ J
Consider an infinite cylinder of radius a and charge q per unit
length. At points where no charge exists, we have V2,p = 0.
Moreover, from symmetry, p depends only on r. Thus
r dr = constant = A
vAlogr+B
E_ -Vip= -A r, r=xi -f- yl
E=2r
r2
(212)
SEC. 601 STATIC AND DYNAMIC ELECTRICITY 135
Problems
1. Solve Laplace's equation in spherical coordinates assuming
the potential V = V(r).
2. Find the field due to a two-dimensional infinite slab, of
width 2a, uniformly charged. Here we have p = p(x) and must
solve Laplace's equation and Poisson's equation separately for
free space and for the slab, and we must satisfy the boundary
condition for the potential at the edge of the slab. The space
occupied by the slab is given by -a S x < a, - co < y < oo.
3. Solve Laplace's equation for two concentric spheres of radii
a, b, with b > a, with charges q, Q, and find the field.
4. Solve Laplace's equation and find the field due to an infinite
uniformly charged plane.
5. Prove that two-dimensional lines of force also satisfy
Laplace's equation.
6. Show that rp = (A cos nx + B sin nx) (Cell + De-Av) satis-
fies x+ a2 $
49Y2
= 0.
7. If Cpl and S02 satisfy Laplace's equation, show that cpl + 4o
and Ipl -- rp2 satisfy Laplace's equation. Does cl92 satisfy
Laplace's equation?
8. If (pi satisfies Laplace's equation and cp2 satisfies Poisson's
equation, show that cpi + 4p2 satisfies Poisson's equation.
60. Dielectrics. If charges reside in a medium other than a
vacuum, it is found that the inverse-square force needs readjust-
ment. That this is reasonable can be seen from the following
136 VECTOR AND TENSOR ANALYSIS [SEC. 61
/
of glass consists of positive
.+++++++++++.....++++ and negative particles, the
electrons being bound to the
0/0 Glass
---
-----j/ nucleus, we see that the field
due to the oppositely charged
Fm. 59 plates might well cause a dis-
placement of the electrons
away from the negative plate and toward the positive plate.
This tends to weaken the field, so that E = 47rv/x, where x > 1.
x is called the dielectric constant.
It is found experimentally that E -_ (qq'/Kr9)r for charges in a
dielectric. Applying this force, we see that Gauss's law is modi-
fied to read
fjE .d d= 4 Q
K
(213)
and if x is a constant,
f f 41rQ (214)
where D is defined as the displacement vector, D = xE _ -x V.
Poisson's equation becomes V D = -V (K Vg) = 4rp, and for
constantx
4rp
(215)
x
W =- - n q;q'i
ri,
(216)
W= jfff pv dr (217)
Now assume that all the charges are contained in some finite
sphere. We have V D = 41rp so that
W fff fff
8- f f f
Applying the divergence theorem,
W
87 S
ff ff
v
Now p,is of the order of 1/r for large r, and D is of the order of
1/r2, while do is of the order of r2. We may take our volume
of integration as large as we please, since p = 0 outside a fixed
sphere. Hence lim f f cpD dd = 0, so that
s
w= 8A f f f (E D) dr (218)
138 VECTOR AND TENSOR ANALYSIS [SEC. 62
=a
D = E = qr,r>>
r
and D=E=O,r<a
The total energy is
22x x
W=s- ffJ-sin O drdOdcp
q2
2a
so that D2 n2 = 0.
n2=_nl Si nce nl = - n2, we have
DN, = DN, (219)
FIG. 60.
We have taken the pillbox very
flat so that the sides contribute a negligible amount to the flux.
Equation (219) states that the normal component of the displace-
ment vector D is continuous across a surface of discontinuity
containing no charges.
We next consider a closed curve r with sides parallel to the
surface of discontinuity and ends negligible in size (Fig. 61).
Since the field is conservative,
Fio. 62.
for isotropic media. Hence
tan01_Kl
(221)
tan 02 K2
which reduces to
f if rpp' dr + ff ' ds = f f f v p dr + jJ p
dS (222)
0.1+vp-0Q'+q
a b
Problems
1. A conducting sphere of radius a is embedded in the center
of a sphere of radius b and dielectric constant K. The conductor
SFc.64] STATIC AND DYNAMIC ELECTRICITY 141
and
P= s,-I- q
t
=-t Cbq'+qJ
and p = 0 if we choose q' _ - (alb)q.
FIG. 63.
Er
8 q(r-bcos9)
or (r2 + b2 - 2rb cos B)'
(a/b)q[r - (a2/b) cos 0]
[r2 + (a4/b2) - (2a2/b)r cos B];
SEC. 651 STATIC AND DYNAMIC ELECTRICITY 143
and the surface distribution is given by
_ (Er)rs _ q b2 - a2
47r 4ir a(a2 + b2 - 2ab cos 9)'
Problems
1. A charge q is placed at a distance a from an infinite grounded
plane. Find the image point, the field, and the induced surface
density.
2. Two semiinfinite grounded planes intersect at right angles.
A charge q is placed on the bisector of the planes. What distri-
bution of charges is equivalent to this system? Find the field
and the surface distribution induced on the planes.
3. An infinite plate with a hemispherical boss of radius a is
at zero potential under the influence of a point charge q on the
axis of the boss at a distance f from the plate. Find the surface
density at any point of the plate, and show that the charge is
attracted toward the plate with a force
q2 4g2a3 fa
U(x, y) = A tan-' y
x
(224)
0= f8 f (9 V r r
V-r) - dd + f f (p V 1r - r Vp) - dd (225)
// B //
Now V(1/r) = -r/r', and on the sphere X,
r r8 r e$
Problems
1. If P is any point outside the closed surface S, show that
f f V(1/r)] dd = 0, where V2cp = 0 inside S and
r is the distance from P to any point of S.
146 VECTOR AND TENSOR ANALYSIS [SEC. 67
a2 2
aX2 +
2. Let rp satisfy V2cp =
1P
2 = 0. Let r be the closed
y°
boundary of a simply connected region in the x-y plane. If P
is an interior point of t, show that
1 1) app a[log (1/r)]
,P(P) - 2x ds
1(`log r an `0 an
where use is made of the fact that
ff A
(u V 2V - v V 2U) dA (u an -van) da
`
n being the normal to the curve.
3. Let (p be harmonic outside the closed surface S and assume
that ip --> 0 and rI V pl -> 0 as r -' oo. If P is a point outside 8,
show that
+ aB (sin o
0 (227)
+ 49(p (sin B
n(n + 1)
R dr (r2 dR)
or
r2 4+24+n(n+1)R=0 (230)
= 0 (234)
(235)
d
dIA
f 11 P,n(;&)Pn(p) dµ =0 if m : n (236)
J 11 P.'(µ) dµ = 2n + 1 (237)
[(1
- p!) dµn1 + n(n + 1)Pn = 0 (238)
dµ
(239)
+[n(n+1) -m(m+l)PPm=0
SFC. 671 STATIC AND DYNAMIC ELECTRICITY 149
or
-M2)(Pmddn-PndPm)]
µ[(1
+ [n(n + 1) - m(m + 1)]P,aPm = 0 (240;
Integrating between the limits -1 and + 1, we obtain
[n(n + 1) - m(m + 1)J f i P,nPn dµ = 0
and ifm -d n,
1 PmPn dµ = 0
Problems
1. Prove (237).
2. Solve V2V = 0 for rectangular coordinates by the method
of Sec. 67, assuming V = X(x)Y(y)Z(z).
3. Investigate the solution of VI V = 0 in cylindrical coordi-
nates.
68. Applications
Example 84. A dielectric sphere of radius a is placed in a uni-
form field Eo = Eok. We calculate the field inside the sphere.
The potential due to the uniform field is p = -Eoz = -Eor cos 0.
There will be an additional potential due to the presence of the
dielectric sphere. Assume it to be of the form ArPI = Ar cos 0
inside the sphere and Br-2PI ° Br-2 cos 0 outside the sphere.
We cannot have a term of the type Cr-2 cos 0 inside the sphere,
for at the origin we would have an infinite field caused by the
presence of the dielectric. Similarly, if a term of the type
Dr cos a occurred outside the sphere, we would have an infinite
field at infinity due to the presence of the sphere. If we let VI
be the potential inside and V11 the potential outside the sphere, we
have
VI = --Eor cos 0 + Ar cos 0
(243)
Vu = --Eor cos B + B cos 0
r2
VI= Vu at r=a
DN, = D.Y. or K aI = a II
at r = a (244)
A=K-'Eo, B_aaK-1Eo
x+2 K+ 2
so that
VI = -
I
-'V+2 Eor cos 0 = - K
2 Eoz (245)
+
SEC. 681 STATIC AND DYNAMIC ELECTRICITY 151
E_ -vVI=K+2Eo
and E is uniform of intensity less than E0 since K > 1. Outside
the sphere
K - 1 _ Eo
VII = -Eor cos B + cos 8 (246)
K + 2 r2
E,
aVII=Eocos0+2K + 2aa ocos9
FlG. 65.
Q +90C
VI = Q'
1
VII = (247)
K
Q = 4v Jf - avlr-a dS
so that
1 f2v ,
= (E cos 8 + a a2 sin 8 d8 dsp = B
Q
4 f-
and
V= -Eorcos8+Qr
For the charge distribution
aV
a- cIr
4"
4I(E0cos8+Q
V=b n= b ( 248)
I
n0 n0
P.(--)
The proof is omitted here that the Pn(p) are actually the Legendre
polynomials. However, we might expect this, since V satisfies
Laplace's equation and P,,(µ)rn is a solution of VI V = 0.
z
Fia. 66.
(b) r > b. In this case
/b n
V = q I Pn6u) (249)
rn=0 r
A. + b-(n+1) _
b-n-1(C2n+1
a2n+1 - C2n+l
- b2n+1)
Hence
V2(P) = q j
n
m
-0
b2n+1 - C2n+1
bn+l(a2n+l - C2n+1)
(rn
a2n+1
- rn+1 Pn(cos 0) (253)
Problems
1. Show that the force acting on the sphere of Example 86 is
F = kQEok.
2. A charge q is placed at a distance c from the center of a
spherical hollow of radius a in an infinite dielectric of constant K.
Show that the force acting on the charge is
SEC. 691 STATIC AND DYNAMIC ELECTRICITY 155
c2 ,410n+K(n+ 1)(a/
3. A point charge q is placed a distance c from the center of an
earthed conducting sphere of radius a, on which a dielectric
layer of outer radius b and constant K exists. Show that the
potential of this layer is
m
q (2n + 1 )b2n+1(rn - a2n+lr-n-1)
V = -c n c cn { [(K +1)n + 1]b 2n+1 + (n +1)(K -1)a 2n+1 } Pn(cos e)
0
n2b2)-e-a-'b-'(mta'+n$t)i*(s-ss) n1rx0
V = 8q to
I (m2a2 + sin
nil m-1 a
nrx mTryo miry
sin sin sin
a b b
f ffrldr
f2
W2 = f f f dr (257)
M
W3=fffr$dr
00
(259)
Now
V x(V xW) = --V2W
SEc. 711 STATIC AND DYNAMIC ELECTRICITY 157
so that
f =4-V X X -4
and hence
f = V x A + Vcp 260)
where
A=4-VxW, =
Problems
1. Show that (256) is a special case of (254).
2. Find an expression for (p(P) if V2(p = -41rp inside S and if
P is on the surface S.
3. f = yzi + xzj + (xy - xz)k. Express f as the sum of an
irrotational and a solenoidal vector.
71. Dipoles. Let us consider two neighboring charges -q and
+q situated at P(x, y, z) and Q(x + dx, y, z). The potential at
the origin 0(0, 0, 0) due to -q is -q/r, and that due to +q is
q/(r + dr), where r = (x2 + y2 + Z2)I and
r+dr= [(x + dx)2 + y2 + Z2]1
The potential at 0(0, 0, 0) due to both charges is
q q g2dr
r + dr r r2
i= i
(Mi.ri)
ri3
so that W = M as VV = M V(p.
72. Electric Polarization. Let us consider a volume filled
with dipoles. The potential due to any single dipole is given by
(261). If we let P be the dipole moment per unit volume, that is,
P = lim (AM/or), then the total potential due to the dipoles is
A"o
= I f f r ra- dr (263)
R
`P=fffv.(!)dr fff
R
SEC. 72] STATIC AND DYNAMIC ELECTRICITY 159
and
73. Magnetostatics. The same laws that have held for elec-
trostatics are true for magnetostatics with the exception that
OZcp.. = 0 always, since we cannot isolate a magnetic charge.
We make the following correspondences, since all the laws of
electrostatics were derived on the assumption of the inverse-
square force law, which applies equally well for stationary
magnets.
EH
Electrostatics
q qm
D ---) B (magnetic induction)
Magnetostatics
dct =
r3
P
Fia. 67. The total solid angle of a surface is
J'r.do
S r 3
(268)
12(P)=
4rdS=4u
fJr
r
ffpv.do_ Jfj.dd
Now the total charge inside a closed surface S is Q = f f f p dr.
R
If there are no sources or sinks inside S, then the loss of charge
per unit time is given by - aQ = - f f at dr.
R
Thus
ffpv.dd=- dr
162 VECTOR AND TENSOR ANALYSIS [SEC. 76
or (270)
= 0
at
P=RJ2 (272)
H=1fiJVI
c r c
x(Jdr
r
since V does not operate on dr and J is a constant. Thus
H = V x A, where A = (1/c) J dr/r = (11c) f f f i dr/r is
integrated over all space containing currents.
Now V A = fif V (j/r) dr f f (j/r) dd, so that if all
s
currents lie within a given sphere, we may push the boundary
of R to infinity, since nothing new will be added to the integral
yielding A. But when S is expanded to a great distance, j = 0
on S, so that that V A = 0. Also
VxH=47j
c
(275)
ff c
ff
S s
For an electric current J in a wire that loops r, we have
(276)
and
[Ml}[L]; [M]}[Ll'[K]i
[7'][µl} [c][TJ2
yielding
L/i [TL]
(1) (2)
Fca. 70.
77. Mutual Induction and Action of Two Circuits. Consider
two closed circuits with currents J, and J2 (Fig. 70). The
magnetic field at 0 due to J, is H, = V x A, where
A,=J, r dr
c (1)r
We define the mutual inductance of the two circuits as the mag-
netic flux through the surface B due to a unit current in (1).
This is
IM= f f L da= f f
a a
1(2) A1. dr = c drl) dr2
1(2) (1(1)
Hence
M= 1 dr, dr,
(277)
C f(2)J(1) r
field will act on J2 dr2 with an equal and opposite force. Thus
J,cJ2
df = J 2 drz x H, = (L) dr, x V r x dr2 (278)
Now
dr2 x (TI 1 x dr,) = V 1 (dr, dr2) - (dr2 "V 1) dr,
r r r
Jc 1) (dr,
f= 2 J (2)1(1) (C' dr2) (279)
v-=-V-1
r2,
1
r12
In (279), r = r2,.
Example 97. We find the force per unit length between two
long straight parallel wires carrying currents J, and J2. We use
(278) and the result of Example 95. We have H, = (2J,/cd)i
at right angles to the plane containing the wires. Hence
df=Jzdr2x 2J,
di. =
2J,J2
dr2xi
and the force per unit length is F = 2J,J2/cd. If the currents
are parallel, F is an attractive force; if the currents are opposite,
F is a repulsive force.
Problems
1. From (278) show that f = J2 f f dd2 x (V X H,).
a
2. Find the force between an infinite straight-line wire carrying
a current J, and a square loop of side a with current J2, the
SEC. 791 STATIC AND DYNAMIC ELECTRICITY 167
M=c f f dd
-ca f f
Applying Stokes's theorem, we have
_ - =VxE (280)
C at
VxH=4w(j+x) C
(281)
+--aD)
VxH= w- (i C
(283)
aD
We call - the displacement current.
t
We rewrite Maxwell's equations
V D = 4irp
(284)
V X
4c \1 + 41r aD1
t
SEC. 801 STATIC AND DYNAMIC ELECTRICITY 169
dt
Given that d = x = y = 0 when t = 0, show that
z = (E/wH) (1 - cos wt), y = (E/(X) (wt - sin wt), where
He
w = -
m
D) -
at
(i) V E=0
(ii) VH=0
aH
vxE= (286)
vxHcat
K aE
µK a2E
V2E = (287)
C2 at2
at= = V2f"(x - V0
so that
a2y 1 a2y
(288)
ax2 V2 at2
Vf = f'(r, t) Vr = f' r
V.(rr/
3f+rf" = rf V - r+V( )'r
V2f
=
f,
i j k
a a a AaH, i LaH,
ax ay az c at c at k
0 Ey E.
or
OE. ,.8H4
(i)
ax c at
(291)
aE, aH,
ax at
Similarly, on using (iv) of (286), we obtain
aH, K aE
(i)
8x c at (292)
aH, K aE,
ax c at
The four unknowns are E., E,, H,,, H,, which must satisfy (291)
and (292). If we choose H = E, = 0, we see that (i) of (291)
and (ii) of (292) are satisfied. Differentiating (ii) of (291) with
respect to x and (i) of (292) with respect to t, we obtain
a2Ey UK 61%
(293)
axe C2 at2
V.H = 0
OH
7E c at
VxH=4c1uE+4v a/
Assume a periodic solution of the form
z)e-;'e
E = Eo(x, y,
H = Ho(x, y, z)e-'"e
or
VxHo
4a/
=-lo-- Eo 2Kw
(296)
C 4r
174 VECTOR AND TENSOR ANALYSIS [SEC. 81
Problems
1. By letting E = f(x - Vt), Hz = F(x -- Vt), V = c/,
show that H. = VKIIA E.
2. Derive (287a).
a2y 1 a2y
3. Letr=x- Vt, s =x+ Vt, and show that-
aX2
=
V2 at2
2
reduces to -ay = 0. Integrate this equation and show that the
Or as
general solution of (288) is y = f(x - Vt) + F(x + Vt), where
f and F are arbitrary functions.
4. Prove that Maxwell's equations for insulators (a = 0) are
KaE
v x H= C at
and V x E_- c a (297)
H=-vx at
and V satisfies V2V =at2
c.
2W
6. Prove thatE=-VxeH=
c at
-V(V.W)+Kµa
C2 at2
isa
192W
solution of (297), provided that W satisfies V2W = e
at2
7. Derive (294).
8. Look up a proof of the laws of reflection and refraction.
9. By considering (i) and (iv) of Example 100, show that
Poe-4x'`i`
P=
81. Poynting's Theorem. Our starting point is Maxwell's
equations. Dot Eq. (iii) of (284) with H and Eq. (iv) with E
and subtract, obtaining
c(H-V xE - xH) = -H aB
at
ID
at
(298)
SEc. 821 STATIC AND DYNAMIC ELECTRICITY 175
Now from (218)
aw, 1 E aD aw,, _ H aB
1
at 4ir at at 4r at
Let us write j = jo + jc where jo represents the galvanic current
and j. = pv, the conduction current. Now
dr awmec,,,,a;c,, awM
at at
and from Sec. 75 it is easy to prove that E ja is Joule's power
loss = aQ Moreover, H V x E - E V x H = V (E x H), so
that we rewrite (298) as
aw, aw, awm aQl
c V (E X H) = -41r (299)
at + at + at + at
f if aQ dr + 4r f f E x H- dd f f at dr (300)
J
where w is the total energy density.
We define s = (c/4x)E x H as Poynting's vector. Equation
(300) states that to determine the time rate of energy loss in a
given volume V, we may find the flux through the boundary
surface of the vector s = (c/4T)E x H and add to this the rate
of generation of heat within the volume. It is natural to
interpret Poynting's vector as the density of energy flow.
Problems
1. Find the value of E and H on the surface of an infinite
cylindrical wire carrying a current. Show that Poynting's vector
represents a flow of energy into the wire, and show that this flow
is just enough to supply the energy which appears as heat.
2. Find the Poynting vector around a uniformly charged
sphere placed in a uniform magnetic field.
3. If E of Sec. 80 is sinusoidal, E = Eo sin co(x - Vt)k, find
the energy density after finding the magnetic wave H.
82. Lorentz's Electron Theory. For charges moving with
velocity v, j = pv, and Maxwell's equations become
176 VECTOR AND TENSOR ANALYSIS [SEc. 82
(i) 4irp
(ii)
(iii) (301)
(iv) V xH
= 4c
(pv + D)
These equations are due to Lorentz. From (ii) we can write
B = V x (Ao + Vx) = V x A. Substitute this value of B into
(iii) and obtain V x E C v x Af or its equivalent
vx(E+catA/ 0
I
Thus E + A is irrotational, so that E + 1 -vgp.
A
Let D= KE, B = pH, and substitute into (iv). We have
I
µ
V x (V xA) = - r
Ipv+ 4a( c
z
at2A-vaC1 SO
K(--AKa2S0/
= -K V2(p -
C at c at2
so that
_ Kp, 12,p 4rp _ 14 (303)
vg c2 812 K cat
SEC. 821 STATIC AND DYNAMIC ELECTRICITY 177
Equations (302) and (303) would be very much simplified if
we could make VA+
KA a __ O.
This is called the equa-
c- at
tion of gauge invariance.Let us see if this is possible.
i aA° =
Now B = V x A0 and E -v4p° so that
+ C at
I aAo l aA
E = -cat - cat - vp
where A = A0 + V. Thus
c1 (BA aA0 1a
at at -cat °x = v((P°
c°)
and
Iax _ _
(p + constant
C at
1 a2x asoo aso
C ate at at
Now we desire
a'
c2 at
ao z
C at C at21
or
vzx
2
-v A0 a°
(304)
cz atz cz
Problems
1. For the Lorentz transformations (see Prob. 11, Sec. 24),
show that
aH
yields V x E _ - and V H = 0.
4 4
c axa axs
3. If P;; show that for the Lorentz trans-
Ht
formations P12 = -171 = ll - (y2/C2)]l Complete the ma-
trix P 1.
83. Retarded Potentials. Kirchhoff's Solution of
2
1
72
V2 av - 47rF(x, y, z, t)
SEC. 83] STATIC AND DYNAMIC ELECTRICITY 179
Fia. 71.
radius e, and let S be the surface of a region R containing P (see
Fig. 71). We apply Green's formula to this region.
fJ f
R
V2,p dT = f f
B
Vp - V#) dd
+ f f (# V-p - p Vu') dd (306)
9
actd-r-4'rfff FOd-r= f f .. .
V2f t_V R B
+ fJ
a
. (307)
180 VECTOR AND TENSOR ANALYSIS [SEc. 83
= f dt (if ..
rr .+ f (308)
+ 91
Ef'(E + Vt)
E2
f(E + Vt) R]
JJ
, da .+ f f ...} (309)
s
Let us now return to a consideration of f(r + Vt). Since f
is arbitrary, let us choose f ° 0 for jr + VtI > 6, with the addi-
tional restriction that
as
f
f(r + Vt) d(r + Vt) = 1, where 6 is
arbitrary for the moment. Notice that f = 0 for Ir + Vti > a.
Now let us choose t2 > 0 and tl negatively large, so that for
all values of r in the region R, jr + Vt[ > a. Hence
[f(r + Vt) &p _ Vf'(r + Vt) _
L r at r Itt,
since Ir + V121 > a, I r + V4J > S.
Moreover
F f(r
ft + Vt) dt = 1 fttip f(r + Vt) d(r + Vt)
da F
=Vj f (x) dx (310)
1
dx = 1
V (F)
r z=o IL f(x) V (r t_-r/v
since dd is of the order E2, and f, f , tp, Dtp are bounded for a fixed a.
We also have that
lim -
t: 'pE,
f(E + Vt)
. dS = -
1
(312)
jI dt f f -
v 47rcp(P)
since f f dS = and for small 6,
E
Finally,
dtlf(r+ Vt)D (rf'
Dr] dd
fs f i:' L r2
2
dt{f(r+Vt)Dp-Vr].dd
= s ii fl' r r
f
[f(r + Vt)
- f f j,t dt r f'(Dr . dd) = ff f .t dt r
+ -f Dr] . dd + f f j i s rV f at dt (Dr dd) (313)
s
on integrating by parts and noticing that f} = 0. Finally the
right-hand side of (313) becomes equivalent to
r rV at t- -r/V
(314)
182 VECTOR AND TENSOR ANALYSIS [SEC. 83
V(P)
F= dr
` 1 JR J -r/ v
1
r)
- i fJ S (v , ~ V r + rV at
V
t= -r/'V
dd (315)
Ie r t-(r/V)
dr
317)
,v(P,t)'JLI x t - (r/V) dr
whe re V =
Finally,
B=VxA
l aA
E=-'cat-V (318)
Problem
1. A short length of wire carries an alternating current,
j = pv = Io (sin wt)k, -1/2 5 z 5 1/2.
(a) At distances far removed from the wire, show that
A= j-o1
cr
sin w (t - rc//J k
and that in spherical coordinates
Hw = cr sin 0 IC
cos w Ct - c) + r sin w t - -) 1
(c) Find p from the equation of gauge invariance, and then
E,, E8, E from E + c A - V o.
CHAPTER 6
MECHANICS
v' dt i +dt dt k
a=
d2x d2y j d zz
(319)
a=dt dtt+vdsdt =dtt+Kv2n
a
r
= [ dt
- r ` de]R -}- r d (2 de P
184
SEC. 84J MECHANICS 185
Example 101. Let us assume that a particle moves in a plane
and that its a cceleration is only radial. In this case we must
have r d [ r2 de1 = 0, and inte-
a v = vat so that as =
v
Also
axv=va,nxt= -vanb
and
an = I$)V,i°I
Problems
1. A particle moves in a plane with no radial acceleration and
constant angular speed wo. Show that r = Ae"o' + Be at.
2. A particle moves according to the law
r = cos t i + sin t j + t2k
Find the tangential and normal components of the acceleration.
186 VECTOR AND TENSOR ANALYST ° [SEc. 85
Fia. 74.
dr2 dr dri
dt - dt + dt
or
E
x.
S
Fia. 75. Fzo. 76.
T -dQ=(-a sin0dO+v)i+acos0doi
VQ(P) = v(1 - sin 0)i + v cos 6 j
Sec. 87] MECHANICS 189
a
1
m
(fl+f2+....+fn)_ 1I
mti_1
f,_ mff
v R
so that (325) becomes
\
d(vxh)=-GMrx (rxrR)
/
= -GMRx(RxdR)
-GM K R . ddR
R - R2 dRJ
t/
= GM ddR (326)
Problems
1. A particle of mass m is attracted toward the origin with the
force f = - (k2m/r6)r. If it starts from the point (a, 0) with the
speed vo = k/21a2 perpendicular to the x axis, show that the path
is given by r = a cos 0.
2. A bead of mass m slides along a smooth rod which is rotating
with constant angular speed w, the rod always lying in a hori-
zontal plane. Find the reaction between bead and rod.
3. A particle of mass m is attracted toward the origin with a
force - (mk2/r3)R._ If it starts from the point (a, 0) with velocity
vo > k/a perpendicular to the x axis, show that the equation of
the path is
r = a sec C
(OW k2)i - 0
avo I
4. In a uniform gravitational field (earth), a 16-pound shot
leaves the putter's fingers 7 feet from the ground. At what angle
should the shot leave to attain a maximum horizontal distance?
194 VECTOR AND TENSOR ANALYSIS [SEC. 89
f, l m, dt2
(332)
f f.(e)=M z
(334)
)ml
Fia. 79.
Example 107. If our system is composed of two particles in
free space and if they are originally at rest, then the center of
mass will always remain at rest, since f = 0 so that d2° = 0, and
r. = constant satisfies the equation of motion and the initial con-
dition 0. For the earth and sun we may choose the center
of mass as the origin of our coordinate system (Fig. 79). The
equations of motion for earth and sun are
M d2r2
m
d2r1
dt2
= - (rlGmMR
r2)2f
dt2
_ (rlGmMR
- 12)2
Problems
1. Show that the center of mass is independent of the origin
of our coordinate system.
2. Particles of masses 1, 2, 3, 4, 5, 6, 7, 8 are placed at the
corners of a unit cube. Find the center of mass.
3. Find the center of mass of a uniform hemisphere.
4. Find the force of attraction of a hemisphere on another
hemisphere, the two hemispheres forming a full sphere.
90. Momentum and Angular Momentum. The momentum of
a particle of mass m and velocity v is defined as M = mv. The
total momentum of a system of particles is given by M = j m;v1.
j-1
We have at once that
n
dM n dv1
I mj I f; (e) = f (335)
dt j-1
dt = j-1
Since
n
dri
H = Ir3xmjv r, x
m'dt
j=1 j=1
we have on differentiating
n n
dH _ d2r, dri dri
Z r, x m,
dt dt2 + j=1 dt x m' dt
n
I
j-1 r' x
(f1" + f, (') )
n
aii =
dt
I r; x
j=1
f;(e)
=L (338)
FIG. 82.
drQ n dr1 n
(r1 - rQ) x (f.(e) + f .(i))
dl xLm'dt+I
j=1 j-1
n
dr, n
Now Mr = m;r,, so that M dt = m1 -, and f,(i)=0,
j=1 l1 j=1
or
dtQa
LQ (e) -M Qx
dt (340)
dtc
= (r7 - rQ) x
/
(fj(ei
`
f3(i))
d2rQ
dt2
x I m, (rj - rQ)
j-1
We see that
n
dtiQr 'rQ
a, LQ(' + d12 x I mj (rj - rQ) (343)
js1
d2rQ n
3. r. - rQ is parallel to
dt'
2
Problems
n
a rQ d2rQ
1. Show that m4(r1 - rQ) x dt2 = M(r. - rQ) x
jet dt2
2. A system of particles lies in a plane, and each particle
remains at a fixed distance from a point 0 in this plane, each
SEC. 95] MECHANICS 201
+2 drd
dr; dr; I d.
(r; - rc) ]
2
_ (dro)2 (r, - rc) -{- L dt
di dt dt dt dt
Hence
1 (dr12 dr" n d
T=2M dtJ +dt
2
r`)]
[dt (r' -
(346)
+ j- 1
2 m'
n n n
T=IM
()2 -}- m; [-d a)(347)
(r; - r.)]2
-1 2 dt 1
of the system and moving with the center of mass, plus the
kinetic energy of the particles in their motion relative to the
center of mass.
96. Work. If a particle moves along a curve r with velocity v
under the action of a force f, we define the work done by this
force as
W=
fr f fr f t
f acts at right angles to the path, no work is done.
If the field is conservative, f = -V(p, the work done in taking
the particle from a point A to a point B is independent of the
path (see Sec. 52).
Now
dvi = fj(a) + fi(.)
m,
dt
dvi
mivi . = fi(e) . V1. + Vi
dt
and integrating and summing over all particles,
n
J`o
mivi dt' dt = La i:1
fi(a)
. vi dt + fee:' fi('i . v; dt
or
n
lmi[vi2(ti) - vi2(to)l = W(e) W° (349)
i=1
so that
dr,
(r, - r2)
di
- dr2dt/ 0 (350)
Also
W(o = f
ff
f is parallel to r, - r2, Ave have f = a(r, - r2) and
f (v, - v2) = a(r, - r2) (v, - v2) = 0
from (350). Thus W(° = 0.
Problems
1. A system of particles has an angular velocity w. Show that
n
T= Jm;lw x r,J2.
i-i
2. If to of Prob. 1 has a constant direction, show that T = }Iw2,
n
where I = md;2, d; being the shortest distance from m; to
line of w.
3. Show that dT = w L, by using the fact that T = 4-mv;2
ff pr dr
f f f p dr
R
dr, 2 ( 351)
J fR f f(e) dt2
H = f f f prxvdr
R
-it ff r x f(e) dr
where f(.) is the external force per unit volume.
98. Kinematics of a Rigid Body. Let 0 be a point of a rigid
body for which 0 happens to be fixed.
It is easy to prove that the velocity
of any other point P of the body must
f V (P1 = yr be perpendicular to the line joining 0
P to P, for if r is the position vector
from 0 to P, we have r - r = con-
stant throughout the motion so that
0. Q.E.D.
r dt =
We next prove that if two points
of a rigid body are fixed, then all
other particles of the body are rotat-
ing around the line joining these two
points. Let A and B be the fixed
points and P any other point of the
body. From above we have
FIG 85.'
If one point of a rigid body is fixed, we cannot, in general, hope
to find a fixed line about which the body is rotating. However,
there does exist a moving line passing through the fixed point so
that at any instant the body is actually rotating around this line.
The proof proceeds as follows: Let 0 be the fixed point of our
rigid body and let r` be the position vector to a point A. From
above we know that the velocity of A, VA, is perpendicular to
rA. Construct the plane through 0 and A perpendicular to VA
(Fig. 85). Now choose a point B not in the plane. We also
have that vB rB = 0, so that we can construct the plane through
0 and B perpendicular to vB. Both planes pass through 0, so
206 VECTOR AND TENSOR ANALYSIS [SEC. 98
FIG. 86.
v; =t=' + w,xr;
But
db da
+w x (b - a)
dt dt
and
ri = (a -- b) + ri Fio. 87.
Thus
vi = d +w x(b - a)+wi x (a -b)+w, xri (353)
and
dS DS
dt + w x
S (356)
dt
xr+Dr
v
dt
-- (357)
dt
v=ac A dt+wxr+ Dr
dt
D2r
d We easily explain the term as the acceleration of P
y-
d1x
_ -2 (w x
dj
t2 dt , + Cfm /
S Now f (force of attraction)
Fio. 90.
has no component eastward, so
that (fm/, = 0. We do not know at but to a first approxima-
d2x
2wgt cos A (362)
dt' =
If the particle remains in the vicinity of latitude A, we can keep A
constant, so that on integrating (362), we obtain
dx
=wgt2cosA
dt
dt +2w cos0dy =0
dtz-g-2w sin0dx =0
2
Fio. 91.
4. Using the coordinate system of Prob. 3, let us consider the
motion of the Foucault pendulum (see Fig. 91).
Let ii, i2, i$ be the unit tangent vectors to the spherical curves
r, 8, p. We leave it to the reader to show that the acceleration
214 VECTOR AND TENSOR ANALYSIS [SEC. 101
Dr
_ Bit + sin 6 rpis
dt
Also w = w(- cos A j - sin X k), so that we must find the rela-
tionship between i i2, is and i, j, k.
Now
r = i, = sin0cosci+sin Bsin cpj+cos0k
ail
i2 cos 0 cos v i + cos 0 sin sin 0 k
= aB =
_ ail
1s
1
H0? = f f f pr x (w x r) dr
R
The quantities
A = f f f p(y2 + Z2) dr
B=fffp(z2+x2)dr
C = fffp(x2 + y2) dr
(368)
D = f f fpyzdr
E = f f fpzxdr
F = f f f pxy dr
DRO,
Lo = dt + to x Hor
Hence
+ k \-E dt - D dt + C dta/
i k
+ wy WV wt
Aws - Fwv - Ewz, -Fws+Bwv - Dw=, -Ews - Dwv+Ccos
(369)
In the special case when the axes are so chosen that the
products of inertia vanish (see Sec. 107), we have Euler's cele-
brated equations of motion :
SEc. 1031 MECHANICS 217.
L. = A d[ + (C - B)wYwz
dwY
L = B + (A - C)w,wZ (370)
dt
dw,
L. = C + (B A)wzwy
dt
B dt +(A-C)w,wx=0 (371)
C d,+(B-A}wZwY=0
so that
A 2w12 + B2wV2 + C2w,2 =constant (373)
wz(0) = coo
0 (374)
ws(0) = 0
and we notice that (371) and the boundary condition (374) are
satisfied by
wt(t) = -coo
0
wZ(t) = 0
B dty + (A - C)wswo = 0
B z
dtzy + (- C)((A - B) w02w = 0 (376)
Problems
1. Solve the free body with A = B for wzj co,, co,.
2. A disk (B = C) rotates about its x axis (perpendicular to
the plane of the disk) with constant angular speed wo. A con-
stant torque Lo is applied constantly in the y direction. Find
w and W..
3. Show that a necessary and sufficient condition that a rigid
body be in static equilibrium is that the sum of the external
forces and external torques vanish.
4. A sphere rotates about its fixed center. If the only forces
acting on the sphere are applied at the center, show that the initial
motion continues.
5. In Prob. 2 a constant torque Lo is also applied in the z
direction. Find w,, and ws.
104. Euler's Angular Coordinates. More complicated prob-
lems can be solved by use of Euler's angular coordinates. Let
O-x'-y'-z' be a cartesian coordinate system fixed in space, and
let O-x-y-z be fixed in the moving body (Fig. 92).
The x-y plane will intersect the x'-y' plane in a line, called the
nodal line N. Let 0 be the angle between the z and z' axes,
L' the angle between the x' and N axes, and (p the angle between
the nodal line and the x axis. The positive directions of these
angles are indicated in the figure.
The three angles &, 0, rp completely specify the configuration
of the body. Now d represents the rotation of the O-z'-N-T'
(377)
d1 + dO + d
220 VECTOR AND TENSOR ANALYSIS [SEC. 104
w =
d k
wzi + w,j + wLk
wz'i' + wy j' + wZ k' (378)
Fia. 92.
sin 4, N+cos#T'
sin sin psin B
cosspsin 0
so that
ws
at at at
ki
sin p sin 0 d + cos s dB
day do
w = cos (p sin 8 --- - sin (p -
at at
day dp
W: = cos B d +
dt
do
sin 0 sin (P + cos <p
dt
d4,
wy = sin B cos - dt8 sin So (379)
wI = d cos 8 +- d
Also
wi=wss+w2+(O2=
(d)2
dd +
(dB 2
+Cdt
d2
+ 2 cos 8
dcp d#
(380)
dt dt
For the fixed frame
w=-
a
do dcp
wy sin>Gt - cos0sin8dt (381)
d +cos6d
105. Motion of a Free Top about a Fixed Point. Let us
assume that no torques exist and that the top is symmetric
(A = B). Euler's equations become
222 VECTOR AND TENSOR ANALYSIS [SEC. 106
(i) A
(iii) Cdts=O
Integrating (iii), we obtain wz = w, = constant. Multiply
by i = and add to (i). We obtain
A
d (co., + iwy) = iwo(C - A)(wx + i )
Integrating,
wz + 2Wy = ae<'(C-A)/ALot
so that
co. = a cos at
wy = a sin at (383)
where a = [(C - A)/A]wo and a is a constant of integration.
Now w2 = ws2 + Wye + ws2 = a2 + wp2 = constant, so that the
magnitude of the angular velocity remains constant during the
d
motion. Moreover, = 0, so that H is a constant vector in
fixed space. We choose the z' axis for the direction of H. Now
H = Awzi + Bw,j + Cwk
= Aa cos at i + Aa sin at j + Cwok (384)
This shows that H rotates around the z axis (of the body) with
constant angular speed a = [(C - A)/A]wo, and since H is fixed
in space, it is the z axis of the body which is rotating about the
fixed z' axis with constant angular speed -a = [(A - C)/A]wo.
Also H k = I HI cos 0 = Cwo, so that 9 is a constant since
I HI = constant. We say that the top precesses about the z' axis.
106. The Top (Continued). We have assumed above that the
weight of the top or gyroscope was negligible, or that the gyro-
scope was balanced, that is, suspended with its center of mass
at the point of support, so that no torques were produced. We
SEC. 106] MECHANICS 223
shall now assume that the center of mass, while still located on
the axis of symmetry, is not at the point of support. We now
have the following situation (Fig. 93) :
L = 1kx(.-Wk')
= WI sin ON
The three components of the torque are
Lz = WI sin 0 cos 'v
L" = - Wl sin 0 sin p
L. = 0
ZI
Fro. 93.
Euler's equations become
dw=
Wl sin 0 cos p = A + (C - A)w"ws
dt
dwdt"
- WI sin 8 sin (p = A + (A - C)wws (385)
0=CdsforA=B
Hence w: = coo. Multiplying Eqs. (385) by wx, w", Co., respec-
tively, and adding, we obtain
2
d (Awx2 + Bw"2 + Cw,2) = Wl sin 0(wy cos w" sin (p) (386)
224 VECTOR AND TENSOR ANALYSIS [SEC. 106
Fm. 94.
D2 = r2 -- (r r0)2
_ (x2 + y2 + z2) - (lx + my + nz)2
(12 + m2 + n2)(x2 + y2 + z2) - (lx + my + nz)2
_ 1l2(y2
= + z2) + m2(z2 + x2) + n2(x2 + y2) - 2mnyz
- 2lnzx - 2lmxy
Thus
I = f f fpD2dzdydx
= Ale + Bm2 + Cn2 - 2mnD - 2n1E - 2lmF
Let us replace 1, m, n by the variables x, y, z, and let us consider
the surface
,p(x, y, z) = Axe + By2 + Cz2 --- 2Dyz - 2Ezx - 2Fxy
= 1 (393)
A line L through the origin is given by the equation x = It,
y = mt, z = nt. This line intersects the ellipsoid rp(x, y, z) = 1
for t satisfying
(Al2 + Bm2 + Cn2 - 2Dmn - 2n1E - 2lmF)t2 = 1
or t2 = 1/I. The distance from the origin to this point of inter-
section is given by
d = (1212 + m2t2 + n2t2)} = t = I-f
so that
(394)
We know that a rotation of axes will keep I fixed, for the line
and the body will be similarly situated after the rotation. We
now attempt to simplify the equation of the quadric surface
,p(x, y, z) = 1. First, let us find a point P on this surface at
which the normal will be parallel to the radius vector to this
point. The normal to the surface is given by Vip, so that we
desire r parallel to Vv, which yields the equations
Ax - Ez - Fy By - Dz - Fx Cz - Dy - Ex
(395)
x y z
x2+y2+z2 = 1
such a point always exists. Equations (395) are then easily
deduced by Lagrange's method of multipliers.
We can arrange the constants of inertia into a square matrix
-E
I = -F B -D (398)
-E -D C
The elements of the matrix (an array of elements) are called the
components of I. Under a proper rotation we have shown that
we can write
A" 0 0
I= 0 B" 0 (399)
0 . 0 C"
228 VECTOR AND TENSOR ANALYSIS [SEC. 107
w = w1 + wyJ + w
111 121 I31
If we write (398) as 1112
122 I32 and
113 123 I3
Problems
1. Find the moments and products of inertia for a uniform
cube, taking the cube edges as axes.
2. Show that the moment of inertia of a body about any line
is equal to its moment of inertia about a parallel line through the
center of mass, plus the product of the total mass and the square
of the distance from the line to the center of mass.
3. Find the angular-momentum vector of a thin rectangular
sheet rotating about one of its diagonals with constant angular
speed we.
SEC. 1071 MECHANICS 229
4. If
3 3
Hp = Ipawa, Hp = I Ipacoa
a=1 a-1
3 3
lip = I ap'Ha, coo = apawa, 3 = 1, 2, 3
a-1 a-1
1 Ipaaa = I Ia°apa, 6, a = 1, 2, 3
a-1 amt
obtain the pressure on this face, P,, = f" The limit of this
AAy
quotient is called the pressure in the direction normal to the face
ORT. The y component of the pressure on the face RST is
P cos l4. Let f" be the y component of the external force per
unit volume, and let p be the density of the fluid. The equation
of motion in the y direction is given by
+f"oT=d
= p A zr y (403)
dt°
230
SEC. 109] HYDRODYNAMICS AND ELASTICITY 231
dm
since (p Ar) = = 0. Now AA = AA,. cos 0, so that (403)
dt dt
becomes
As A --- 0, we have
oAv dt Cp
i
0, so that if we assume f,,, d p
AA dt2
finite, we must have P = P.. Similarly, P = P. = P. = p.
Since the normal n for the tetrahedron can be chosen arbitrarily,
the pressure is the same in all directions and p is a point function,
p = p(x, y, z, t). We leave it to the student to prove that at the
boundary of two perfect fluids the pressure is continuous.
109. The Equation of Continuity. Consider a surface S bound-
ing a simply connected region lying entirely inside the liquid.
Let p be the density of the fluid, so that the total mass of the fluid
inside S is given by
M=
ff
R
p(x,y,z,t)d-r
Now there are only three ways in which the mass of the fluid
inside S can change: (1) fluid may be entering or leaving the
surface. The contribution due to this effect is JJ vp dd.
s
(2) matter may be created (source), or (3) matter may be
destroyed (sink). Let 4,(x, y, z, t) be the amount of matter
created or destroyed per unit volume. For a source, 4, > 0, and
for a sink, 0 < 0. The net gain of fluid is therefore
ffJ1dT_ jfpv.dd (406)
0 (408)
at + V - (Pv) =
= I tf
f pdT + f f f v(pdr)
since the volume V changes with time. However, p dr is the
mass of the volume dT, and this remains constant throughout
d
the motion, so that (p dT) = 0. Since F = dd , we obtain
fff(pf-vp)drfffpdr
This equation is true for all V, so that
pf - VP=p dv
d I
or
dt f - p VP (411)
dvav
From (76) we have that
dt at + (v V)v, so that an alterna-
tive form of (411) is
Vp (412)
at
+1VV2-Vx(VxV)=f-PVp
2
(413)
l 1
vX(Vxv)=VlX++1v2
P
v.[V(x+p+1v2)]=0
p 2
at r
(P Rdt-)
7. The energy equation. For a simply connected region R with
boundary S, the kinetic energy of R is
T=ifffpv2dr
R
Let the surface S move so that it always contains all the original
mass of R. Show that
236 VECTOR AND TENSOR ANALYSIS [SEc. 112
dT
dt
= ffJ R
-dv2
dt
= If! v.f!Vpdr
at =
- (x + P + 2 v2) + C(t),
and if p =
t
p(p),
-f- 2 + x +
z
f-D(t).
112. The General Motion of a Fluid. Let us consider the
velocities of the particles occupying an element of volume of a
fluid.Let P be a point of the
volume or region, and let vp
represent the velocity of the
fluid at P (Fig. 96). The veloc-
ity at a nearby point Q is
vQ = vp + dvp
= Vp + (dr . V)vp (416)
from (75). By (dr V)vp we
mean that after differentiation,
the partial derivatives of v are
calculated at P. We now re-
place dr by r for convenience,
Fio. 96.
so that r = xi .+ y} + zk if we
consider P as the origin and x,
y, z large in comparison with x2, y2, z2, zy, etc. Equation (416)
now becomes vQ = vp + (r V)vp. Now
r x (V x w) + (417)
from (9), (10), (12) of Sec. 22.
Now let
w = (r V)vp = x av + y av + z avl (418)
al p ay p azlp
SEC:. 1121 HYDRODYNAMICS AND ELASTICITY 237
and hence
-- P+yayay
ax av
ax ax
ay av
P +z azazP
az avI
=w
We did not differentiate the ' since they have
(' ayP
axP I IV P
been evaluated at P and so are constants for the moment. Thus,
using (417), we obtain
w = j V(r w) + J(V x w) x r (419)
Moreover, v =Vp+w, so thatV xv =V xw = (Vxv)p [see
(418)], and
and
IV(r w) = axi + byj + czk
We may now write (420) as
= T dt . ds
Er- ds + v. ds \dt/
ds (424)
dt
_ --0d(V-J2) =0 (425)
with constant velocity. Let the center of the sphere travel along
the z axis with velocity vo. We choose the center of the sphere
as the origin of our coordinate system. From Sec. 110, Prob. 6,
we have
f - pop= at
and
Dr
Hence = so that V2p = 0. Now at points on the surface
'P = (Ar +
! cos 8 (426)
r/J
(see Sec. 67). We need
B
-CA-
a3
cosB=0
so that B = a2A/2. Moreover, at infinity we expect the veloc-
ity of the fluid to be zero, so that the velocity relative to the
sphere should be -vo. Hence
a-
Vs
az z .= A = - vo
and
s
fluid
z+az?
0 (428)
°2`P-ax 2
aLT(X-'i)t.
We assume a solution of the form p = A (z)e Sub-
stituting into (428), we obtain
2w
tZ_voi f 41x2 d2A
- 2 A (z) +
e
dx2 I=0
so that
d2A _ 41r2
\2 A (429)
dz2
at -(x++2v2)+C(t)
and for a gravitational potential, x = gz, so that
at
=
- (gz + p
+ 2
v2) + C(t) (432)
a2V a_p dC
(434)
ate g az + dt
Substituting (431) into (434), we obtain
2
-v2 a2 A0 cosh zcosLA (x - Vt)
I2r
g 2w-
z cos (x - vt) ] + dt (435)
Ao (- v 227r2
cosh
27r
z+
g
sink
27r
z=0 (435a)
or
v2 = g tanh - z
In deep water z/X is large so that tank z 1, and the
velocity of the wave is v = (Ag/2,r)*.
Problems
1. Show that for steady motion of an incompressible fluid
under the action of conservative forces, (v O)w - (w V)v = 0,
wherew = D x v.
2. Show that dt (Pl = \P of v for a conservative system.
SEc. 1151 HYDRODYNAMICS AND ELASTICITY 243
3. If C is the circulation around any closed circuit moving
dC
with the fluid, prove that = p d (1) if the field is con-
servative and if the pressure depends only on the density.
4. Show that v = 2axyi + a(x2 - y2)j is a possible velocity
of an incompressible fluid.
5. Verify that the velocity potential (p = A[r + (a2/r)] cos 0
represents a stream motion past a fixed circular cylinder.
115. Small Displacements. Strain Tensor. In the absence
of external forces, a solid body remains in equilibrium and the
forces between the various particles of the solid are in equilibrium
because of the configuration of the particles. If external forces
are added, the particles (atoms, molecules) tend to redistribute
themselves so that equilibrium will occur again. Here we are
interested in the kinematic rela-
tionship between the old positions p° r p
of equilibrium and the new. We so Is
shall assume that the deforma- po r p
tions are small and continuous. Fia. 98 .
We expect, from Sec. 112, that
in the neighborhood of a given point Po, the remaining points will
be rotated about Po and will suffer a pure strain relative to Po.
Let r be the position vector of P relative to Po, and let s be the
displacement vector suffered by P, and so the displacement suf-
fered by Po (Fig. 98). Then
S = so + ds = so + (r V)so (436)
Let s = u(x, y, z, t)i + v(x, y, z, t)j + w(x, y, z, t)k. Since we
will be dealing with static conditions,
s = u(x, y, z)i + v(x, y, z)j + w(x, y, z)k
From (420),
S = so+J(V xs)P, (437)
where
w=x-asax PO
as
+yay Pc +z-
oz
as
PO
since s = v At.
We are interested in the position of P after the deformation
(now P') relative to the new position of Po (now Po ). This is
244 VECTOR AND TENSOR ANALYSIS [SEC. 115
auj a-l aw
+2 D Cx tax +xyax +xz at
1
Po P. 0
I r au av awl
D xy + yz - y2
+2 y Po + ay PO ay Po
aw
+ 2D zx azlPo + zy a P. az Po
and
r+ 1
i aw aul i law av
2 C 8x + azl 2 `ay + az
3
y' = Sa'xa (441)
a-1
We shall see in Chap. 8 that since r and r' are vectors, then, of
necessity, the s/ are the components of a tensor. Notice that
Sii = s; , so that the tensor is symmetric.
( au 1 au av 1 aw a_ul
i
- r1 = \1 + ax) 1 + 2 (ay + ax 1 + 2 ax + az/ k
(LU)
By neglecting higher terms such as 21 u-
y
we have
_ I
1+
au
. Similarly j --> r2, and lr2l = 1
I + avl
; k - rs,
Iril ax ay
246 VECTOR AND TENSOR ANALYSIS [SEC. 116
awl
and fr31 = 1 + . The angle between r1 and r2 is given by
I
au av
cos B = .
+r,jjr2+ v ay + ax'
The terms of the strain tensor are now fully understood. The
volume of the parallelepiped formed by r1, r2, r3 is
au av aw
V r2xrs = 1+ ax + ay + az
so that
V _ V _ au av aw
(443)
V ax+8y +az
The left-hand side of (443) is independent of the coordinate
system, so that V s is an invariant.
Finally, we see that the deformation tensor due to the tensor
- V (r w) has the components
Fia. 99.
and that fi = I tia dsa, where f, = f,, f2 = f,,, f3 = f., t12 = t,,,
a=1
. . . . We shall see later that this explains why the ti; are
dv called the components of a
tensor.
z Let us now consider the
resultant force acting on a
volume V with boundary 8
(see Fig. 100). Wehavefrom
(446)
fz = txx dsx + txv ds + tzx dst
so that
Y
F. = fz _ JJtxzdSx
- JIf Jr
acZy acz=
Fz 4-9t" dT (447)
V
ax + ay + az
E
_ E E(t1+t2+t3)
or 0,
e2 t2 - (t3 + tl) = 1 r t2 - E (t1 + t2 + t3) (449)
1 a 1-r-a-
e3 = 3 - - (11 + t2) t3 "- (t1 + t2 + t3)
E
The formulas for el, e2, e3 apply only in the immediate neighbor-
hood of a point P. Since points far removed from P will have
different stress ellipsoids, the principal directions will vary from
point to point. Hence no single coordinate system will exist
that would enable the stress and strain components to be related
by the simple law of (449). Let us therefore transform the
components of the stress and strain tensors so that they may be
referred to a single coordinate system. The reader should read
Chap. 8 to understand what follows. If he desires not to break
the continuity of the present paragraph, he may take formula
(456) with a grain of salt, at least for the present. Example 8,
Probs. 21 and 22 of Sec. 11, and Prob. 21 of Sec. 15 will aid the
reader in what follows.
If x1, x2, x3 are the coordinates above, and if we change to a
new coordinate system 11'. T2, 23 where
3
(.Ti), =
(xi)
Saisea# = I eaa
0-1 a-1 a-1
so that
e11 + 922+933 = e11 + e22 + e33 = el + e2 + e3 (453)
el -1+0
E t1+
e2=1 (454)
e3=1Eor
ts+
SEC. 117] HYDRODYN.4.1IICS AND ELASTICITY 251
eij = E aiaa,8eas
B=1 a-1
and since eap = 0 unless a = 6 [see (448)], we obtain
3 3
1 aiaaJaeaa = I aiaajaea
a=1 a=1
3
a -l
aiaaja (1 E to +)
3 3
1+ a
F, I aiaajala + Y' 1 aiaaja
a-1 a=1
and
1 + o
eij = E of + 4,5ij (455)
3 3 3
since 1j = I I aiaajalas = I aiaaala.
6=1 a -l a=1
Equation (455) is the relationship between the components of
the strain and stress tensors when referred to a single coordinate
system. We have
Q111
ell = 1 - E (tll + 122 + 133)
E
1 +Q_ oT
922 = E 122 -
- (111 + t22 + 138)
O
e33 = 1
+ 133 - E, (111 + E22 + 133)
(456)
1 +v_
912 = E 112
1+0.
923 = l23
E
1 +Q_
981 = E 131
252 VECTOR AND TENSOR ANALYSIS [SFC. 11
Solving Eels. (456) for the t;; and removing the bars, we obtain
01
t11 = [ell + 1
2-
(ell + e22 + e33)
1+
E au o
(au av aw
=i+o, axl-2vax+ay+az J
E rav v au av awt22
= 1+aLay1-2v (ax+ay+az)]
E taw v
(457)
tss = 1+aaz+1-2QV*s
E E (au av
t12=t21= 2(1+o) ay+ax
E av aw
t23 = t32 =
2(1 + v) \az + ay
E (aw au
= 2(1+Q)ax +az
t31 - t13 =
Equation (447) now becomes
E [d?u .
(a2u a2v a2w
F=
-}- v ax2 + 1 - 20 \49x2 + ax ay + ax az
E a2u a2v a2w a2u
+ 2(1 + u) aye ay ax + az ax + az2 1
_ E 1 a (au av aw
zu+1 -2aaxax+ay+ az J
-2(1+a)IV
2(1+o) [vu+ 1 1
The forces per unit volume in the y and z directions are
Fy
2(1+a)[Dzv+1 12vay(V - s)I
E 1 a
{V2W+ I-2o- az
(0 s)
so that
+
f= [V2S V(V s)] (458)
2(1 r) + 1 1 2cr
If we let R = RJ + R j + R2k be the external body force per
unit volume, p the density of the medium, then Newton's second
SEc. 117] HYDRODYNAMICS AND ELASTICITY 253
law of motion yields
R+ 2(1 +o,)[V2S+-Lv(V.S)] =
P
However,
V X (V X 62) = V(V S2) - V 'S2 = 0
so that
E(1 T a)
(1 +a)(1 - 2a)
V2S2 = pa2s2
-
at2
(462)
V X S2 = =0
axJ+-k=0
clx
254 VECTOR AND TENSOR ANALYSIS [SEC. 117
L'(1 - o.)
V,=
P(1 + v)(1 2v)
ap11
x1 = x
49p12 49p13
fl - ax, + axe + axi where x2 = y
x3=z
3
c3 apli
= jG
=1 ax
3
and
3 api; 3 a au; au; 2 a (V v) ap
1-1 ax? + axi/ 3' ax,
The equations of motion (465) are
2 a(V v)
p dt
dui
- = pFi + v I
3
1
a aui
ax' Cax'
+ au;
--
ax'
- -3
7 axi
- ap
axi
(466)
J
or
pd
Along with (467) we have the equation of continuity
at
=0
Problems
1. Derive (467) from (466).
2. Consider the steady flow of an incompressible fluid through
a small cylindrical tube of radius a in a nonexternal field. Let
v = vk and show that p = p(z) and +1 V2v = ap Show that the
boundary conditions are v = 0 when r = a, and v = v(r),
r2 = x2 + y2, and that = r dr (r dr) Hence show that
v = (A/4i7)(r2 - a2), where A is a constant and LP = A.
3. Consider a sphere moving with constant velocity vok (along
the z axis) in an infinite mass of incompressible fluid. Choose
the center of the sphere as the origin of our coordinate system.
Show that the equation of motion is p aE = q V2v - Vp and that
the boundary conditions are v = 0 for r = a, v = -vok at r = oo
258 VECTOR AND TENSOR ANALYSIS [SEC. 118
49Z
Assuming p = -,1 V2p and V2w1 = 0, show that (i) and (ii)
will be satisfied.
Let w, = 3voa/2r, rp = coz - (voa3z/4r3) + (3voaz/4r),
3,gvoaz
P= 2r3
and show that
V2w1 = 0, V2p = 0, p = -n V2rp
v= -V(p+wik=0forr=a
v = -vok for r = co
4. Solve for the steady motion of an incompressible viscous
fluid between two parallel plates, one of the plates fixed, the
other moving at a constant velocity, the distance between the
plates remaining constant.
5. Find the steady motion of an incompressible, viscous fluid
surrounding a sphere rotating about a diameter with constant
angular velocity. No external forces exist.
CHAPTER 8
TENSOR ANALYSIS AND RIEMANNIAN GEOMETRY
119. Summation Notation. We shall be interested in sums
of the type
S = alx, + a2x2 + . . . + anxn (468)
(471)
S = aixi
df = ax dxl + ax dx2 +
. + -CIx"dx"
of
_ dx'
ax'
= of dxa
axa
;.nd
df
afdJxa
dt - axa dt
S = Z I g-Oxaxs
0-1 a-1
We also notice that the gap can be thought of as elements of a
square matrix
gii 912 913
921 922 923
931 932 933
as
ax
-= as ga"
ax"
Problems
1. Write in full aaxa=bi,a,i= 1, 2, 3.
2. If aae7xax8xY = 0, show that ai;k + aki; + a; + aJik +
akii + aiki = 0.
3i,ti:'412 ...n i.,i,...i.
3. Show that 612..,, i. = 6;,i:...;..
4. If yi = aaxa, zi = by, show that zi = bas;x#.
5. Prove that
bas = are - a8r
baBratYaapi = arst + airs + astr - sari - aisr - arts
6. Show that the determinant I at at can be written
2 2
a1 a2
a11 a2'
= Etiiaia? = etiia1
atz a2 2
and that
at 1
a2 a3
at
2 z z
a1 a2 a3 = Eiika;a;ak = eiikai4a8
8 8
a1 a2 a$$
show that y= a
Given 9,, =
C) p
, show that
i,a1a2 an
n n n
a1 a2 an
The reader should note that this definition agrees with the
definition for the special case of second- and third-order deter-
minants which he has encountered in elementary algebra. The
definition of a determinant as given by (474) shows that it con-
sists of a sum of terms, nn in number. Of these, n! are, in gen-
eral, different from zero. Each term consists of a product of
elements, one element from each row and column. The sign
of the term a a'a . a depends on whether it takes an even
or odd permutation to regroup i1i2 in into 12 . . n.
Since i1 and i2 are dummy indices, we can interchange them
so that
fail = Ei,i,... ina,'a2 . . . ann = ei,i, ... inal=a2 . . , ann
Also
aaA7 = IaI (477a)
Example 117. Let us consider the n linear equations
yi = aQxa, a, i = 1, 2, .. , n (478)
Multiplying by Af, we obtain
A°y' =a' Asxa
so that summing on i, we have from (477)
A"y' = Ialaxa = Ialxx
If IaI 0 0, then
A°y' _ y'(cofactor of a' in IaI)
IaI - IaI (479)
J \xl,y2, xn) -
y 1 (480)
', x2, . . . ,,yry/Jfy',X2,
xn 1, y2, . . .
ay i
As a special case, suppose a = , where y' = y'(x', x2, , xn).
49x1
a yi axa
Now a`t = Let us consider j to be fixed for the moment.
ax- ay,
yt
Let Yi = 491, a , Xi = ax -, so that Yi = aQXa. If
ax y
l al 0
ax11
then
cix1
1 ay
(cofactor of Lyi in
8x
ax
and so
1 ay ay 8x8
(cofactor of a? in
ax TX 8y1
a,y
a
8x ay axa a2ya
a log la
laxI axa a2ya
ayi axa
Now S' = ax-
- -,a so
yi that upon differentiating with respect to
yk, we obtain
ayi a2xa axa a2yi axa
0 - axa (483)
ayk aye + ay; axa axa ayk
1+
Problems
i z i
al a2 a3
1. What is the cofactor of each term of a1
2 2
a2 a3
2
_ a2{ _ axa
8. If u' = ua axa, Vi = V. clt.
-, show that uaVU = uaVa, where
xi = x{(x1 x2 xn).
ax{ ax i
9. If u{ = ua , show that u{ = ua
ax"
axa a ax- 00
10. If gi; = gaa , show that gi; = 9a$
a2i
82i axi axi
a { , ui = ua axi azi
11. If u{ = ua , show that u{ = ua
axa axa axa
12. Apply (476) for the product of two third-order determi-
nants.
13. If A; = B,-s ayi ax , show that A{ = B, and that IAI = IBI,
A;A; = B )$B,!.
14. If X is a root of the equation Iai; -- Xbifl = 0, show that X
axa axa
is also a root of Id;; - Ail = 0 provided that B;; = aaa
aji
axa axa ax
baa 0.
!ii = a axi a
122. Arithmetic, or Vector, n-space. In the vector analysis
studied in the previous chapters, we set up a coordinate system
with three independent variables x, y, z. We chose three mutu-
ally perpendicular vectors i, j, k, and all other vectors could be
written as a linear combination of these three vectors. Any
vector could have been represented by the number triple (x, y, z),
where we imply that (x, y, z) -= xi + yj + zk. The unit vectors
could have been represented by (1, 0, 0), (0, 1, 0), and (0, 0, 1).
A system of mathematics could have been derived solely by
defining relationships and operations for these triplets, and we
need never have introduced a geometrical picture of a vector.
For example, two triplets (a, b, c), (a, #, y) are defined to be equal
SEC. 1221 TENSOR ANALYSIS 269
OX*
.-
Ai=Aai
a2a
Then
A' _ A8 ax a = Aa ax8 ax _Aaax
a28 axa ate axa
system, and let ds2 = dx2 + dy2 + dz2. The components of this
vector in a polar coordinate system -,re
R=X-ar+Yaary ar
+Zaz= cos 0X+ sin 0Y
0
ae ae CIO _ -sine X + cos e Y
X ax + Y ay + Z az^ r r
Z=Xa +Yaz+Zaz=Z y
Problems
1. If A'(x), B'(x) are components of two contravariant vec-
tors, show that C'"(x) = A'(x)B'(x) transforms according to the
ax' ar
law C*' = Cap axa ate, where C" =
2. Show that if the components of a contravariant vector
vanish in one coordinate system, they vanish in all coordinate
systems. What can be said of two contravariant vectors whose
components are equal in one coordinate system?
3. Show that the sum and difference of two contravariant
vectors of order n is another contravariant vector.
4. If X, Y, Z are the components of a contravariant vector in
an orthogonal coordinate system, find the components'in a spher-
ical coordinate system. By what must the 0 and Sp components
be multiplied so that we can obtain the projections of the vector
on the 0- and (p-directions?
'
5. If A' = Aa a ' show that A' = Aa a.
SEC. 125] TENSOR ANALYSIS 273
ax` ax'
6. Referring to Prob. 1, show that Cs, Caa
axa al$
N i N i
az
124. Covariant Vectors. We consider the scalar point func-
tion So = co(x', x2, . . . , x"), and form the n-tuple
(492)
(493)
ay' axa ay'
L'
J i=Aaa' (494)
so that
axa axQ
AaBa = AFB,
axp axa
AaBa = A"B,
a2s
= Asfl = AaBa
Hence AaBa is a scalar invariant under a coordinate trans-
formation. The product (AaBa) is called the scalar, or dot,
product, or inner product, of the two vectors.
Problems
1. If Ai and Bi are components of two covariant vectors, show
axa axe
that Ci; = A;B, transforms according to the law Oil = Cap
axa ax?
2. Show that C = AB; transforms according to the law
axp .
C"i
- C8 V axa
a
3. If Ai = A. axi show that A, = Aa
axi
4. If p and J, are scalar invariants, show that
grad (vO) = c grad P + 0 grad cp
grad [F(op)] = F'(,p) grad rp
ax N
1 b,b,... ax TaW*A:...,q,
alas
anal . .
axa* axle,
. aXa . . . axig
a-. (495)
alb,
SEC. 126] TENSOR ANALYSIS 275
AiE1 = AQBay -
CIO axy 492i
x6
axi ax; axa
axa axy
= A'Bay
ax axk
or
l
Ai ` $,k - Bay axp avk
axa axy
=0
axa axk
Since A' is arbitrary, we must have f k = Bay the desired
a2 ax
result.
Example 122. The Kronecker delta, S, is a mixed absolute
tensor, for
axi axa = axi axa = 80
= asa
a
' aga axi' axa axs
CO that
C AaB# Cs
axa axe = axa Ox'
lax ax
Igl =1g1 l ax or I#I} = IgIl a2
ax Of
B, = IgI}A' =
aax IgI'Aa axa
Ba2i
B.
ax ax-
axa a-
gap dxu dx = g,,, dxu dx'
axu ax,
or
°1°
a0 ^ 9u dxu dx' = 0 (496)
(9aO
axu ax
axa axs
9µY = 9a$ (496a)
axµ axY
Problems
axa ax0 axa axa
1. If A = Aao axe, show that A;; = Aaa
ai ax' axe
2. Show that A,, can be written as the sum of a symmetric
and a skew-symmetric component.
3. If A are the components of an absolute mixed tensor, show
that A; is a scalar invariant.
4. If Aa# are the components of an absolute covariant tensor,
and if Aa#A$,, = 8 , show that the Aal are the components of an
absolute contravariant tensor. The two tensors are said to be
reciprocal.
SEC. 1271 TENSOR ANALYSIS 279
an invariant?
7. From the relative tensor A of weight N, derive a relative
scalar of weight N.
8. If Ami is a mixed tensor of weight N, show that Amw is a
mixed tensor of weight N.
9. Show that the cofactors of the determinant lair, are the
components of a relative tensor of weight 2 if ail is an absolute
covariant tensor.
10. If Ai are the components of an absolute contravariant
vector, show that axe are not the components of a mixed tensor.
127. The Line Element. In the Euclidean space of three
dimensions we have assumed that
ds2 = dx2 + dye + dz2
In the Euclidean n-space we have
ds2 = (dx')2 + (dx2)2 + . .. + (dxn)2
= bo dx° dxa (497)
we have that dx' = axi d2a, so that (497) takes the form
ax° axe
ds2 = 6,0 dxµ dx
axµ aV
We may write ds2 = vr d2" dz', where
n
axa 8x# ax° ax°
9vr = a°a
a_ ax+' any
280 VECTOR AND TENSOR ANALYSIS [SEc. 127
Thus the most general form for the line element (ds)2 for a
Euclidean space is the quadratic form
The gap are the components of the metric tensor (see Example
125). The quadratic differential form (498) is called a Rie-
mannian metric. Any space characterized by such a metric is
called a Riemannian space. It does not follow that there exists
a coordinate transformation which reduces (498) to a sum of
squares. If there is a coordinate transformation
xi = xi(y', y2, . . . , y")
such that ds2 = Say dya dy5, we say that the Riemannian space is
Euclidean. The y's will be called the components of a Euclidean
coordinate system. Notice that gas = Sap. Any coordinate sys-
tem for which the g;; are constants is called a Cartesian coordinate
system.
We can choose the metric tensor symmetric, for
gii = Ij(gv + gig) + (gii - gii)
and the terms $(gt, - gii) dx' dx' contribute nothing to the sum
ds2. The terms J(g,i + gii) are symmetric in i and j.
Example 129. In a three-dimensional Euclidean space
ds2 = tax 1)2 + (dx2)2 + (dx3)2 for an orthogonal coordinate sys-
tem, so that
1 0 0
0 1 0
0 0 1
Let
x' = r sin 0 cos p = y' sin y2 cos y$
x2 = r sin 0 sin rp = y' sin y2 sin y2
x3=rcos9=y'cosy2
Now
axa 8x5
9ti(r, g, w) = gas ay; ayi
ax' ax' axe axe OxI 49x8
- ay: ayi + ay: ayi + ay' ayi
SEC. 127] TENSOR ANALYSIS 281
Hence
gll = (sin y2 cos y3)2 + (sin y2 sin y3)2 + (cos y2)2
Similarly
922 = (y') 2, 933 = (y')2(sin y2)2, gi,=0 for i5j
so that
ds2 = (dy')2 + (y')2(dy2)2 + (y' sin y2)2(dy3)2
= dr2 + r2 do2 + r2 sin2 0
is the line element in spherical coordinates. Since the g's are
not constants, a spherical coordinate system is not a Cartesian
coordinate system.
Example 130. We define gii as the reciprocal tensor to gi,,
that is, gi.2 gai = d,f (see Prob. 4, Sec. 126). The gii are the signed
minors of the g;i divided by the determinant of the gi;. For
spherical coordinates in a Euclidean space
1 0 0
0 0
1
0 r2 0
0 2
r2
0
0 0 r2 sin2 B 1
0 0
r2 sin2 9
L2 = g pA°A# (499)
two vectors by
cos 0 = A'Bi = Aig;;Bi = gi,AiB'
= giiA,B1 = gi'A;B, (500)
tive definite form, that is, gaaza# > 0 unless z' = 0, we have
gas(XAa + µB")(XAa + kB") > 0
or
axa axa
where hi, = gaa
au'aui
SEC. 127] TENSOR ANALYSIS 283
x = ,B(x - Vt)
y=y (502)
z = 2
xJ
t=fi -Ux
c
where
U` V-I-W ^ r1 _ U2)-1
- 1 + (VW/c):' `\ c I
We now assume that (x, y, z, t) represents an event in space
and time as observed by S and that (x, 9, 2, 1) represents the
same event observed by S (see Fig. 101).
284 VECTOR AND TENSOR ANALYSIS [SEC. 127
dx
The origin b has :f = y = 2 = 0' so that from (502) = - V'
dt
showing that S moves with a constant speed - V relative to S.
Similarly S moves with speed + V relative to S.
.7, 1,q
'P1:Mf,7
14
-V
S S
X X
z z
Fia. 101.
12
+ 22
-- 4
C
so that the speed of light is the same for both observers. This
is one of the postulates of the special theory of relativity. Start-
ing with this postulate and desiring the group property, we could
have shown that the transformations (502) are the only trans-
formations which keep dx2 + d y2 + dz2 - cz dt2 = 0 an invariant.
Let us now consider a clock fixed in the S frame. We have
x = constant, so that dx = 0, and from (502) dt = 0 dt. Hence
a unit of time as observed by S is not a unit of time as observed
by S because of the factor 0 5-6 1. 8 remarks that S's clock is
running slowly. The same is true for clocks fixed in the S frame.
SEC. 1271 TENSOR ANALYSIS 285
u2
()2 ()2
+ + \dt/2
as measured by S, then
p = [1 mo
- (u2/C2)]} -- m
d
dE _ (mxa) dxa
a-1
-V//-22aU2
Vh11aul
4. If w is the angle between the coordinate curves, show that
cos w = his/.
SEc. 127J TENSOR ANALYSIS 287
s r ( dxa dxO 4
- ` i \g°a dt dt dt (503)
of 1 (aga9 zaxs
ax' 2f \ ax'
and
(ga:xa +
2(ds/dt)2 dt2
2
d
If we choose s for the parameter t, s = t, dt = 1, dtQ = 0, and
use the fact that gii = gii, (504) reduces to
1 /agai ague agoal/ xa x
o =0
a
+ - -axti (505)
g'ax + 2 610 axa
SEC. 1281 TENSOR ANALYSIS 280
where
9r r° (.9g., ago
a9oa
(507)
ra$ 2 \ ax$ + axa axQ
The functions r" are called the Christoffel symbols of the second
kind. Equations (506) are the differential equations of the
geodesics or paths.
Example 135. For a Euclidean space using orthogonal coordi-
nates, we have ds2 = (dxl) 2 + + (dx") 2, so that gap = a.#
d1
and 0. Hence the geodesics are given by ds = 0 or
ft"
xr = ars + br, a linear path.
Example 136. Assume that we live in a space for which
dal = (dx')2 + [(x1)2 + c2](dx2)2, the surface of a right helicoid
immersed in a Euclidean three-space. We have
1 0 1 0
1
{1gi?il = 0
0 (x1)2 + c2 (x')2 .+ C2
Thus we have
rill =0,
x1
r21 = r14 = 0,
1 r212
(x') 2 + C2
1 -- -x1,
r22 r2
Problems
1. Derive the r;k of Example 136.
2. Find the differential equations of the geodesics for the line
element ds2 = (dxl)2 + (sin x')2(dx2)2.
3. Show that rqn = r'a.
4. For a Euclidean space using a cartesian coordinate system,
show that r, = 0.
5. Obtain the Christoffel symbols and the equations of the
geodesics for the surface
xl = ul cos u2
x2 = u' sin 412
x3 = 0
This surface is the plane x3 = 0, and the coordinates are polar
coordinates.
6. From (507) show that axµ = g,sroals + g,arp,,.
7. Obtain the Christoffel symbols for a Euclidean space using
cylindrical coordinates. Set up the equations of the geodesics.
Do the same for spherical coordinates.
8. Write out the explicit form for the Christoffel symbols of
the first kind : { i, jk I = g;,rlk.
9. Let ds2 = E due + 2F du dv + G dv2. Calculate IgJ, g'',
i, j = 1, 2. Write out the rk.
axk axe + a2x° axe
10. If r = r;7 axs
a. ax ax ax ax ax show that r;,y - rry
are the components of a tensor.
129. Law of Transformation for the Christoffel Symbols.
Let the equations of the geodesics be given by
d2xi d i J dxk _
ds2 + r "k ds ds ^ 0 (508)
and
AV
P 2k
d.V d2k
_- (5 0 9)
ds2 + ds ds °
for the two coordinate systems xi, xi in a Riemannian space. We
now find the relationship between the r k and r; . Now
SEc. 129J TENSOR ANALYSIS 291
dxi
__ a2' dxa
and
d2x` _ a2xi dxa dxa 9. d2x°
ds axa ds ds2 axa axa ds ds + axa ds2
Substituting into (509), we obtain
&V d2xa a22 i dxa dxa rik axi axk dxa dxa 0 (510)
axa ds' + ax# axa ds A + axe, ax" ds as
°
We multiply (510) by axi and sum on i to obtain
Comparing with (508), we see that (using the fact that rlk = rk;)
This is the law of transformation for the rk. We note that the
rk are not the components of a tensor, so that the rk may be
zero in one coordinate system but not in all coordinate systems.
Example 137. From (481) we have
algl
8 = Iglgae ax'
= a;rap + a1r00
=rap+r;p=2rµ
or
C log ra
_ p (512)
axp
292 VECTOR AND TENSOR ANALYSIS [SEC. 129
so that
n
I a,-,a; = Sas (515)
V-1
reduces to S;; =
axa axe
' =g"° ax. axe
a' axe
We multiply both sides by
amt
and
a:' CIO ax),
sum on i, so that
azf axp 8x8 8x'
axe
= aaa is a = #0 axe - iii (516)
A' = Aa _,
a2'
axa
A; = Aa
axa
82'
(517)
621
Replacing by a from (516), we see that
n axa
Aaa (518)
a-1
so that if we define
A;;=-Aa1
8x1
49A;
(522)
we have that
axa axa
A..s = Aa.a
axi axi
821 A
296 VECTOR AND TENSOR ANALYSIS (SEC. 130
so that
a
axlN aA axa ax N-1
ax 8xa 02'
+N ax
lax
aax'
ax _ laxe a2xs
and from (482), Hence
ax ax i ax
axa a x, axa
2
aaxA; aA1
Hence Ai,5 - A;,; =
- ax` is a covariant tensor of rank 2.
It is called the curl of the vector Ai. If the A; are the components
of the gradient of a scalar, Ai = a-P, then
axi
curl A; ==
a2ip
ax1 ax'
- ax' ax' = o621P
BA;
and write the intrinsic derivative of A; as
as
Example 144. The divergence of an absolute contravariant
vector is defined as the contraction of its covariant derivative.
298 VECTOR AND TENSOR ANALYSIS (SEC. 130
Hence
div A' = A' = ax -4- AarQ,
a logx IgI
Now rQ;
as = from (512), so that
_ 1 a
div A Aa
t
ax- + N/191 ax.
1
div As = aa
a (V f gj Aa) (528)
v 2(p 1,91
aaa (v' Fg19°° TX ) (529)
Sec. 1301 TENSOR ANALYSIS 299
0 r2 0 7 t g' l = 0
I
0 0 r2 sin2 B 0 0
r2 sin 9
so that
VF = 1
a
1 r2 sin B aF -1- a (sin B aF' + a / 1 aF'
{
r2 sin 0 ar \\\ ar 00 \ 00a l\sin 9 app/)
Example 146. In Example 144 we defined the divergence of
Ma
the vector A' as div A' = A + Aar' - For a Euclidean
Cti
= ff Svf.,)N; dv
S
41st -, 1
211 - -L
191
I
4132 =
1
411! = 0
f F.
a
ds ds = f do (531)
Problems
a2i
1. By starting with Ai = All , show that
Oza
so that
a2x' ax° axe
axk axj
= - (ray) 4 ask axe
Q Q
-(raa)Q6-50 _ -k
Substituting into (532), we obtain
(rij k)o = (rfY)Qaj kaa - (rk)Qaa
let Define
xi r
X' = Vs (535)
where s is arc length along the geodesic. Each V determines a
geodesic through P, and s determines a point on this geodesic.
Hence every point in the neighborhood of P has the definite
coordinate xi attached to it. The equations of the geodesics in
this coordinate system are
d2z1 1 d21 dxk
+ r'k ds ds = 0
dS2
2 i
gas s e + gaak° 0
as =
v`; =
a'
ax; + V ar°`'
Since V';k - Vik; and Vi are tensors, Va;k must be the components
of a tensor, from the quotient law (Sec. 126). It is called the
curvature tensor. We can obtain two new tensors of the second
order by contraction.
SEC. 133] TENSOR ANALYSIS 307
Let
ara ara
Ri'
Ba;
axa axa + r ara; - r%raa (539)
This tensor is called the Ricci tensor and plays an important role
in the theory of relativity.
We obtain another tensor by defining
ara
aa i ara
a;
Si; = Be.. (540)
ax1 axi
a log "ICI - ra
ax" 1aµ.
we have that
a2 log a2 log -VIgl 0
=
axi ax1 ax1 axi
azk
= f (z', z2, z", z"+1, y`,
k=1, 2, ..,n+I
ay'
. . . , yn)
j=1,2,...,n
(548)
We certainly must have aye ayj = ayj yi, and this implies
Problems
1. Show that Rhijk = - Rihik = - Rhikj and that
Riijk = Rhikk = 0
2. Show that Rhijk + Rhkij + Rhiki = 0-
3. If Rij = kgii, show that R = nk.
310 VECTOR AND TENSOR ANALYSIS [SEC. 134
911922
R:; = JRg.;.
5. Show that B k1 76 0 for a space whose line element is given
by ds2 = (dx')2 + (sin xl)2(dx2)2.
6. Derive (550) from (546), (547), (549).
1 49R
7. If R = gtiaR,, show that (R'),; =
2 ax;.
CHAPTER 9
FURTHER APPLICATIONS OF TENSOR ANALYSIS
i
135. Frenet-Serret Formulas. Let Ai = be the unit tangent
ds
vector to the space curve xi = xi(t), i = 1, 2, 3, in a Riemannian
space. In Example 148 we saw that the contravariant vector
t ° 0}
is normal to Xi. Let us define the curvature as K =
as g°0 88 aas
and the principal unit normal µi by
aai _ Kµi
= (551)
as
i
Since pi is a unit vector, we know that aS is normal to pi. Now
g°0A°,us = 0, so that the intrinsic derivative yields
Sµa bA°
g°'eA° + g°0 µB = 0
as bs
aµ0
g°a A° + Kg°aa°i8 = 0
as
or
(B68
µs
g°aa° + KAS =0 (552)
v'
- 1 Sµ + KAt
(553)
T (68
or
aIAi
-= -#c? - TY' (554)
as
(i)
where ,r is called the torsion and is the magnitude of
The vector rµe - as is thus normal to all three vectors V, Af, Pi.
Since we are dealing in three-space, this is possible only if
av
TEIZ = 0, or
as
avi
= (557)
as
Writing (551), (554), (557) in full, we have the Frenet-Serret
formulas
& dxP
KAi
ds ds
i
d+ µa s = - (Ki + Tv`) (558)
a
Pa = Tµ
ds + ds
SEC. 1361 FURTHER APPLICATIONS OF TENSOR ANALYSIS 313
dx
.
+ ra' ds
d a Q
dare the com
136. Parallel Displacement of Vectors. Consider an absolute
contravariant vector Ai(xl, x2, . . . , xn) in a cartesian coordi-
nate system. Let us assume that the components Ai are con-
stants. Now
a2' axi
A' = Aa A' = Aa
axa a2a
so that
822' axfi
dA: - _ Aa dxy
8xs axa (921y
where Rte,, is the curvature tensor (see Sees. 132, 133, 134).
137. Parallelism in a Subspace. We start with the Rie-
mannian space, V,,, ds2 = gab dxs dxs. If we consider the
transformation
xa = xa(ul, u2, . . . , u"), m<n (562)
Aa = aui ai
a = 1, 2, . .. , n (563)
Hence
ax" SAa _ I dal dull
9aa hlk
+ (r'i)ha ds
auk as ds y J
and
a
5. Show that ga$(x kxa + x {x k) + xaxIxk 0, where
ax, =
covariant differentiation is with respect to ui and h;;.
6. Show that A ; = x;ai + x°,as, for each a.
138. Generalized Covariant Differentiation. The quantities
aui are contravariant vectors if we consider i fixed; for if
xa = xa(ul u'")
aAuk° axe
Aff:k =a + A;r auk - Air?k (567)
Problems
1. Why is A ",
ill
a contravariant vector in V,i?
2. Show that
aA" ax?
Aa;:, = r,A,; ax*
au,
- r#rA - rAflk
aut
is a mixed tensor, by considering the scalar invariant b8c'daA';.
a
3. Show that x = x`; = aus, and that
320 VECTOR AND TENSOR ANALYSIS (SEC. 139
_ a2xa
44 = au' aui - r .xh + r;yx x
= xai + FRxx4.
aub
Moreover, ds2 = h;; du= du', and h;; = hab -- --- so that
aut alai
322 VECTOR AND TENSOR ANALYSIS [SEC. 139
R1212 R1212
(573)
Ih IhI
hit
or, = gaN y* (576)
R= Riiii
9ii9ii
l aR
5. If R; = gaiRii, show that R"a =
2 axi
6. If Rii = kgii (an Einstein space), show that R = giiRii = nk,
or Rii = (R/n)gii
7. Show that a space of constant curvature K is an Einstein
space and that R = Kn(1 - n).
140. Lagrange's Equations. Let L be any scalar invariant
function of the coordinates q', q2, . .. , qn, their time deriva-
tives 41, 42, ... , 4n, and the time t:
L = L(qi, 4i, t) = L(qi, qi, t)
If we perform a transformation of coordinates,
qa
= ga(gl, q2, . . . , qn)
aga
a = 1, 2, . . . , n, then qa = aqa
so that 4a is a function of
the Now
aL aL aqa aL 84a
aqi = aqa aqi + a4a aqi
aL aqa aL a2ga
= aqa aqi + a4a a4i 41 (581)
aq0
SEC. 140] FURTHER APPLICATIONS OF TENSOR ANALYSIS 325
:=1 2
ma
I/ i1
2
2 s 1 2 s s
V (1
xI x1 x1 x2, x2f x2f 7 xn)
_d aL _ aL -d(magma) _ n
_ m;
1 8ga0 x,a2 + aV
-.
dt ax; ax; dt 2 ax * ' ax;
=max;-(F,)a, ifgaQ=aae
and m,x; - (FT), = 0 for la Euclidean space and Newtonian
(;) _ aL
d aL
aq=
0, r= 1 ' 2' 'n (583)
Qr dx r = f a
49y
a
a
dxr = fa dya
and compute
d Wi
Qi =..L-lim -,
O AV
i not summed
mr" - mr02 = Q,
dt (mr2o) = Qe (586)
R(r de)
with Q, = 0, Qe = = rR. The solution to (586) is
do
Problems
1. A particle slides in a frictionless tube which rotates in a
vertical plane with constant angular speed w. Set up the equa-
tions of motion.
2. For a rigid body with one point fixed,
T =JA(w;+wy)+jCwa
Using Eulerian angles, show that if Q, = 0, then
C(O + cos 0 +') = R
A¢sin20+Rcos0=S
A#-A,2sinocoso+R,1 sino =Qe
where R, S are constants of integration.
aT
3. If T = q")gags, show that 2T =
aq-
q°°.
oL(q,
4. Define p, = and assuming we can solve for
and
aq
q* = gr(gl) . . , q^, p,, show that the Hamiltonian
328 VECTOR AND TENSOR ANALYSIS [SEC. 141
A= f`
e [(h - V )gas '-6 dxs
a aA ]
1
dX
i o e, we have
i not summed
rik n
Caxk + ax's - agi)
If i, j, k are different, then rk = 0. We also see that
rii
_ 1 agii
2 g ,i axi
2 1 agkk
rkk = - 2 9ii axi
Applying (590), we have
330 VECTOR AND TENSOR ANALYSIS [SEC. 141
1_ 1 ,tag, _ 1da
r11 2 dr
2g Or
1
r22
= _ 1
g11 a92z = -re a
2 or
1 _ 1 g11 agaa _= -r sine 0 e-A
2 Or
1441 = --g - _ 1
z
1 ag44
or
1
2
dv
dr
s
1
gs2 ag zz 1
112 = 2 Or r
(591)
raas= _ 1 22 a9aa
- sin 0 cos 0
29 00 =
r1aa= gsa2 an..
Or r
1 agsa
933 = cot 0
2 ae
8g 44 1 dv
rt = -219
4
4
44
Or 2 dr
and all other r;k vanish.
From (539)
ar:
R;; _
axe - axa + r
ar8, - r°raa
so that
R11 -
a
ar1R + art, + 4
ar14 +
ar Or Or
rilril + risrs1 + ri3r 1
+ ri4ril - r1i(ri1 + rig + ria + r44)
1 1 1 d2D 1 1 1(dv12 1 da
r2 r2+2dr2+r2+r2+4`drl 2r dr
l da l da dv
(592)
2r dr 4 dr dr
by making use of (591). Hence Einstein's law R;; = 0 yields
_ 1 d2v 1 (_f)2 1 A dv 1A
0
R11 2 dr2 + 4 Cdr 4 dr dr r dr -
SEC. 141] FURTHER APPLICATIONS OF TENSOR ANALYSIS 331
Similarly
R22 = e-a C 1 - 1= 0
2 r (dr dr) I-
1 (ddrv _ da
R33 = sine 8 C 1+
2r dr I - sin2 0 = 0 (593)
d +dr =o
or
X + v = constant = co
We desire the form of (588), as r --f oo, to approach that of
(587). This requires that X and v approach zero as r approaches
co.. Hence \X + v = 0 or X P. From R22 = 0 we have
ev C1 + r dr) = 1. Let y = so that dr = 1 dy and
'Y r
yCl+ydr)
or
dy dr
1-y r
and
2m
(594)
r
where 2m is a constant of integration.
The equations of the geodesics are
d2x' dx' dxk
r;k
a82 + d ds = 0
which yield
d20 dr d_9 CdLp)2
2 2
r33 =0
Z-2 + 2T E ds A + Vs J
332 VECTOR AND TENSOR ANALYSIS [SEC. 141
or
d20
+
2 dr db
- sin 0 cos 8
(j2 0 (595)
d,82 r ds ds ds
a dB a
If 0 = 2, = 0 initially, then 0 = satisfies (595) and the
ds 2
boundary conditions.
We also obtain
1'11(ds/2
()2
ds2+ ds+ r44(d- 2 =0
or
dt 1 da (dr)2 -x
1 e v-'
dv (dt)2
=0
ds2 + 2 dr ds - re a (thp)2
+2 dr ds/
(596)
d21 dt dr d2t dv dt dr
0 or - 0 (598)
ds2 + 2x14 ds ds - ds2 + dr ds ds
dt dt c
log d + v = log c or
ds = y
(600)
1 - 1
y \ds2 - r2 \ds2 + y
2
or
1
r4 + y
(l z
1
y (r d(p)2 - r2
SEc. 1411 FURTHER APPLICATIONS OF TENSOR ANALYSIS 333
or
Ch drl 2 2m
r2 dip/ + r2
h.2
- c2 - 1 r + 2m
r
h2
r2
d + u = h2 + 3mu2
d2
2U
(601)
T =u=-[l+eeos((p-w)j
where e, w are constants of integration. This is Newton's solu-
tion of planetary motion. We substitute this value of u in the
term 3mu2, and we obtain
d2u m 3m3 6m3
W,p2+u =h2+h; + h4 ecos(cp-w)
33
+ h4 [1 +
23 cos 2((p - co)]
41
_ J944+ constant
element.
fd sdt-
Assuming ds2 > 0 for a < t < 0, we can extremalize
fB(ga'odx" dx,
z: ra+.o,x-°ze1
1 + 1 + C..'6x1
a, axe2= 0
(607)
x$ + r: xax2 + CO, 10=0
where
dx;
9"'`9µ.i = a;7r x1 = -
ds
5. If
ds2 = -el, dr, dr2 - r,r2eM dpi er dt, dt2
e= - [1mM2
(m + L M)2 rl J
1 1 11 - (mm2M '
+ r' M)2 r2
el=(1+rl C1+m/
ex = e2µ-r
ar:'
9x° a,T
ax-1 2
rir2 = he-a
dt1 __ Cie_
ds
dt2
ds = C2e
d2v Mm M
v
d,P2 + L 1 + [1 + (m/M))'hi 11 + (m/M)]2h2l
+3Mv2r1+(mMM)J 2
L
Energy, equation for a fluid, 235-236 Gauss, divergence theorem of, 114-
of electromagnetic field, 175 120
of electrostatic field, 136-138 electrostatic law of, 128
kinetic, 201 Generalized force vector, 326
Envelopes, 69 momentum, 328
Equation, of continuity, 231-232 Geodesic coordinates, 303-305
of gauge invariance, 177 Geodesics, 83, 288-289
of motion for a fluid, 233-236, minimal, 294
302-303 Gradient, 36, 120, 273, 297
Equipotential surfaces, 129 Gravitation, Einstein's law of, 328-
Euclidean space, 8, 279, 308-309 329
Euler's angular coordinates, 219-221 Newton's law of, 190
equation of motion, for a fluid, 233, Green's formula, 118, 299-300
302-303 Green's reciprocity theorem, 139-
for a rigid body, 216-217 140
Euler-Lagrange equation, 85 Gyroscope, motion of, 222-225
Evolutes, 66
H
F
Faraday's law of induction, 167 Hamilton's equations of motion, 328
Field, 9 Harmonic conjugates, 123
conservative vector, 103 functions, 123, 143
nonconservative vector, 104 Heine-Borel theorem, 94
solenoidal vector, 117 Helix, 60
steady, 9 Hooke's law, 249
uniform, 10 Hypersurfaces, 282
Fluid, 230
I
general motion of, 236-238
Force moment, 196-200
Foucault pendulum, 213-215 Images, method of, 141-143
Frenet-Serret formulas, 60, 311-312 Induction, law of, 167
Functions of bounded variation, 99- Inertia, moment of, 216
100 product of, 216
conjugate, 122 tensor, 225-228
continuous, 95 Inertial frame, 211
properties of, 96 Infemum, 92
harmonic, 123 Inhomogeneous wave equation, 177
Fundamental forms, first, 71 solution of, 178-182
second, 74-75 Insulator, 129
Fundamental planes, 62-63 Integral, line, 101, 103-105
normal, 62 Riemann, 101
osculating, 62 Integrating factor, 111
rectifying, 63 Integration, of Laplace's equation,
145
G of Poisson's equation, 155
Interior point, 90
Gauge invariance, 177 Interval, closed, 89
Gauss, curvature, 78, 322 open, 89
344 VECTOR AND TENSOR ANALYSIS
P Rectifying plane, 63
Refraction, law of, 139
Parallel displacement, 313-315 Regions, connected, 102
Parallelism in a subspace, 315-317 simply connected, 102-103
Parametric lines or curves, 71 Regular arcs, 98
Particle, acceleration of, 30, 210 Relative motion, 187-188
angular momentum of, 196 time rate of change of vectors, 208
dynamics of, 189 Resistance, electric, 162
kinematics of, 184 Retarded potentials, 178-182
mass of, 189, 285 Ricci tensor, 307
momentum of, 196 Riemann integral, 101
Newton's laws of motion for, 189, Riemannian, coordinates, 305
211
curvature, 307, 320-323
rotation of, 22 metric, 280
velocity of, 30, 184, 209 space, 280
Particles, system of, 194 geodesics in, 288-289
Perihelion of Mercury, 334 hypersurface in, 282
Permeability, 160 Riemann-Christoffel tensor, 307-308
Planetary motion, 190-193 Rigid bodies, 203
Point, 89 motion of, 215-225
boundary, 90
interior, 90 S
limit, 90
neighborhood of, 90 Scalar, 1
set theory, 89 curvature, 307
Poisson's equation, 132-134 gradient of, 36, 273, 297
integration of, 155 Laplacian of, 45, 298-299
Poisson's ratio, 249 product of vectors, 10, 273-274
Polarization, 158-159 Schur's theorem, 323-324
Potential, of a dipole, 157 Schwarzchild line element, 329
electrostatic, 128 Second fundamental form, 74-75
vector, 117 geometrical significance, 75
velocity, 232 Sequence, 97
Power, 162 Cauchy criterion for convergence
Poynting's theorem, 174-175 of, 97
Poynting's vector, 175 Set, 89
Pressure, 230 bounded, 89
Principal directions, 77-78 closed, 91
Q complement of, 90
countable, 93
Quadratic differential form, 280 diameter of, 93
Quotient law of tensors, 276 infemum of, 92
R limit point of, 90
linear, 89
Radius of curvature, 24 open, 90
Recapitulation of differentiation supremum of, 91
formulas, 48 theorem of nested, 93
Reciprocal tensors, 281 Simply connected region, 102
346 VECTOR AND TENSOR ANALYSIS
Sink, 231 Tensor, components of, 274
Solenoidal vector, 117 contraction of, 275-276
Solid angle, 160 contravariant, 274-275
Source, 231 covariant, 274-275
Space, conformal, 294 curvature, 306-307
Space curve, 31 deformation, 246
arc length of, 100-101 inertia, 225-228
curvature of, 58, 311 mixed, 275
intrinsic equations of, 63 Ricci, 307
Jordan, 98 Riemann-Christoffel, 307-308
on a surface, 72 strain, 243-246
radius of curvature of, 58 stress, 246-248
tangent to, 31, 58, 311 two-point, 335
torsion of, 59, 312 weight of, 275
unit binormal of, 59, 312 Tensors, 274-278
unit principal normal of, 58, 311 absolute, 275
Space of n-dimensions, 268-269 addition of, 275
Special theory of relativity, 283-286 cross product of, 278
Spherical coordinates, 35, 50 outer product of, 278
indicatrix, 68 product of, 275
Steady field, 9 quotient law of, 276
Stokes's theorem, 107-112, 300-301 reciprocal, 281
Strain, ellipsoid, 245 relative, 275
tensor, 243-246 Theorem, of Ceva, 8
Streamline, 234 of Desargue, 7
Stress tensor, 246-248 of Menelaus, 8
Subtraction of vectors, 3 Top, motion of, 222-225
Summation convention, 259 Torque, 196-200
Superscripts, 14, 259 Torsion of a space curve, 59, 312
Supremum, 91-92 Transformation of coordinates, 269
Surface, 70 Trihedral, 59
are length on, 71 Triple scalar product, 23
asymptotic curves on, 81 Triple vector product, 24
average curvature of, 78 Two-point tensors, 335
conjugate directions on, 80
curves on, 72 U
developable, 70
first fundamental form of, 71 Uniform continuity, 95
Gauss curvature of, 78 Uniform vector field, 10
geodescis of, 83 Uniqueness theorems, 119
normal to, 73, 109 Unit charge, 127
principal directions on, 77 V
second fundamental form of, 74
Vector, associated, 281
T basis, 3, 8
center of mass, 194
Tangent to a space curve, 31, 58, 311 components of, 8-9
Tangential acceleration, 184 conservative field, 108
INDEX 347
Vector, contravariant, 270-272 Vectors, parallel, 2, 314
covariant, 273 parallel displacement of, 313-315
curl of, 45, 55, 297, 300 scalar, or dot, product of, 10, 273-
definition of, 1 274
differentiation of, 29 subtraction of, 3
displacement, 136 triple scalar product of, 23-24
divergence of, 42, 54, 120, 297-298 triple vector product of, 24-25
field, 9 vector, or cross, product of, 20-23
irrotational, 107, 111 Velocity, angular, 22-23
length of, 1, 281 linear, 30, 184, 209
operator del (v), 40 potential, 232
physical components of, 272 Vortex motion, 238-239
potential, 117
solenoidal, 117 W
space, 268-269
sum of a solenoidal and an irrota- Waves, equation of, 170
tional vector, 156-157 inhomogeneous equation of, 177
unit, 1, 281 longitudinal, 253-254
zero, 1 transverse, 172, 253
Vectors, addition of, 2, 275 IN'eierstrass-Bolzano theorem, 92
angle between two, 10, 281-282 Work, 103, 202
differentiation of, 29
equality of, 1 Y
fundamental unit, 8
linear combination of, 3 Young's modulus, 249