Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
.
.
A TREATISE ON BESSEL FUNCTIONS AND THEIR APPLICATIONS TO PHYSICS. .
McLeod wm^^mmm: . J.Presented to the LIBRARY of the UNIVERSITY OF TORONTO by Mr. R.
'GRAY. 1895 [All Rights reserved./ AND M."A TEEATISE ON BESSEL FUNCTIONS// AND THEIR APPLICATIONS TO PHYSICS. ... CAMBRIDGE PROFESSOR OF MATHEMATICS IN THE UNIVERSITY COLLEGE OF NORTH WALES. M." Bacon. as Nature growes furder disclosed. Hontron : MACMILLAN AND AND NEW YOEK. that more kindes of them. BY ANDREW G.A. PROFESSOR OP PHYSICS IN THE UNIVERSITY COLLEGE OF NORTH WALES MATHEWS. FELLOW OF ST JOHN'S COLLEGE.] CO.A. B. " And as for the there cannot faile to bee Mixt Mathematikes I may onely make this prediction.
. «& C. AT THE UNIVEESITY PRESS. CLAY.C^f'% CTambrilrge : PRINTED BY J. F.
and growing almost every branch of necessary for their practical application. and of the theory of a complex And even from the purely physical point of view it is variable. if any. but experience has repeatedly shown that the most abstract analysis may unexpectedly prove to be of the highest importance in mathematical physics. waste of to that part of the . on the one hand. book which deals with physical our aim has been to avoid. Some readers may be inclined to think that the earlier chapters analysis. it may be so now. impossible to say that an analytical formula is useless for practical purposes. and that they afford excellent illustrations of the more recent theory of differential equations. This book has been written in view of the great importance of the Bessel functions in mathematical physics and its principal object is to supply in a convenient form so much of the theory of the functions as is . that space would have allowed of a more extended treatment. that we could almost wish anything superfluous . worked out in some detail. of the analytical theory included in the present work has failed to be of some use or other in and we are so far from thinking that the later chapters has been inserted. With regard applications.PREFACE. but it contain a needless amount of tedious must be remembered that the properties of the Bessel functions are not without an interest of their own on purely mathematical grounds. especially in the chapters on the complex theory and on definite integrals. and to illustrate their use by a selection of physical problems. As a matter of fact it will be found that little.
to Professor A. and III. any pretension of writing an elaborate physical treatise. Jn {^) = also especially for the roots Our thanks are due to the proof sheets. M. J. and.A. and Professor J. have been extracted... has been that the chapter on diffraction long. for his care in reading Finally we wish to acknowledge our sense of Mr The bibliographical list as anything but a on pp. time and space in the discussion of trivialities. We have endeavoured to choose problems of real importance which naturally require the use of the Bessel functions. Dr Meissel has also very generously placed at our disposal the materials for Tables II. the former in manuscript. so as to bring out clearly the direct physical One result of this course significance of the analysis employed. . V. on the other. VI. the help and this treatise.. McMahon has of very kindly communicated to us his formulae and other transcendental equations... from which the substance of that attention in this country to the valuable chapter is mainly derived. the accuracy with which the text has been set up in type by the workmen of the Cambridge University Press. A. Gibson. Thomson for permission to make free use of their researches on fluid motion and electrical oscillations respectively. Lodge for copies of the British Association tables from which our tables IV. 289 291 must not be regarded of treatises and memoirs which have been list — consulted during the composition of this work. and to treat them in considerable detail. but is proportionately rather attract we hope that this section may more general and interesting results contained in Lommel's memoirs. and to the Berlin Academy of Sciences and Dr Meissel for permission to reprint the tables of Jq and Jj which appeared in the Ahhandlungen for 1888.VI PREFACE. much pleasure that we acknowledge encouragement we have received while composing It is with We are indebted to Lord Kelvin and Professor J. G.
CONTENTS.
CHAPTER
I.
Introductory.
Bernoulli's problem of the oscillating chain,
solid
1
;
conduction of heat in a
Bessel's differential
cylinder,
5.
2;
Bessel's
astronomical problem, 3;
equation,
CHAPTER
II.
Solution of the Differential Equation.
Solution by series when n is not an integer, 7 ; and when n is an integer or zero, 8; definition of Jn{^\ H; elementary properties, 13; definition and explicit expression of Yn{x\ 14; expressions for Jn{^) and r'„(^) as integrals, 15.
CHAPTER
III.
Functions op Integral Order. Expansions in Series of Bessel Functions.
Proof of the theorem exp ^
(if
 i~i) = 2
00
Jn{x)t'^i
with
corollaries,
17;
Bessel's expressions of Jn{x) as a definite integral, 18; transformation of powerseries into series of Bessel functions, 19; Neumann's expression for F„, 23; the addition theorem, 24; Neumann's extension thereof, 25; expan
sions in series of squares
and products of Bessel functions, 29
;
Schlomilch's
theorem, 30.
CHAPTER
IV.
Semiconvergent Expansions.
Solution of Bessel's equation by successive approximation, 34;
new
ex
pression for Jn{x) as a definite integral, 38; the semiconvergent series for Jn (x) and Yn (x), 40 numerical value of log 2  y, 41 ; table of J^^, (^), 42.
;
Vlll
CONTENTS.
CHAPTER
Bessel's proof that
V.
Functions.
The Zeroes of the Bessel
J„(^)=0 has an
calculation of the roots, 46 ; Stokes's
p. 241.]
infinite
number
and McMahon's
formulae, 49.
of real roots, 44; [See also
CHAPTER
Rayleigh's appUcation of Green's
YI.
FourierBessel Expansions.
theorem, 51;
value of the integral
Jn{Kr)Jni\r)rdr, 53; application to the conduction of heat, 54; formulae /: for FourierBessel expressions, 55, 56.
CHAPTER
VII.
Complex Theory.
Hankel's integrals, 59—65; the functions /„, Kn, 66 proof of the semiconvergent series, 69.
—
8.
Lipschitz's
CHAPTER
CHAPTER
VIII.
Definite Integrals involving Bessel Functions.
IX.
The Relation of the Bessel Functions to Spherical
Harmonics.
CHAPTER
X.
Vibrations of Membranes.
Equation of motion, 95; solution by Bessel functions, 96; interpretation
of the result, 97; case of an annular
membrane,
99.
CHAPTER XL
Hydrodynamics.
Rotational motion in a cylinder, 101
;
oscillations of a cyUndrical vortex
about a state of steady motion, 104; hollow irrotational vortex, 106; vortex surrounded by liquid moving irrotationally, 109; waves in a tank, 110; rotating basin, 113; motion of a viscous liquid, 116; Stokes's problem of the
cylindrical
pendulum, 118.
CONTENTS.
IX
CHAPTER
XII.
Steady Flow of Electricity or of Heat in Uniform Isotropic Media.
Differential equations, 122; potential of electricity from disksource in
due to
electrified circular disk,
flow
infinite
124; conducting medium, 126;
medium between two parallel planes: problem of Nobili's flow in 128; rings, cylindrical conductor with source and sink at extremities of axis, 131; flow in infinite medium with one plane face separated from infinite plate electrode by film of slightly conducting material, 132; case in which
flow in infinite
the
medium has another
situated, 136;
parallel plane face in which the second electrode is further limitation of conductor by cylindrical surface, 139;
flow in cylindrical conductor with electrodes on the
139.
same generating
line,
CHAPTER
XIII.
Propagation of Electromagnetic Waves along Wires.
Equations of electromagnetic field, 142 ; modification for symmetry round an 144 solution of the equations, 146 ; solution adapted to long cylindrical
;
axis,
conductor in insulating medimn, 147; case of slow signalling along a cable, 150; electric and magnetic forces in this case, 151 ; expansion of xJq{x)IJ'q {x) in ascending powers of x, 153; approximative values of Jo{r})/Jl{rj) and
^0 ('?)/^o ('?) for large values of
effective resistance
rj,
and
effective selfinductance of cable,
;
155; case of more rapid oscillations, 157; 157 ; Hertz's solution
Hertz's solution for a wire, 162.
for
a timeperiodic electric doublet, 161
CHAPTER
XIV.
Diffraction of Light.
calculation of intensity of Bessel functions, 167 discussion by of Lommel's U, V functions and their relations to Bessel functions, 170 ; apphcation of results to Fraunhofer's diffraction phenomena, 178; graphical expression of intensity, 179; apphcation to Fresnel's more general case, 181;
;
Diffraction through a circular orifice, 165;
illiunination, 166; expression of intensity
graphical discussion, 182; positions of
184; diffraction
positions of
maxima and minima of illumination, when the orifice is replaced by an opaque circular disk, 192; maxima and minima, 194; intensity due to infinite linear source,
198; Struve's discussion of this case, 201; application to determination of spacepenetrating power of telescope, 205; diffraction through narrow slit with parallel edges, 208 ; expression of Fresnel's integrals in terms of Bessel
functions, 209.
I
X
CONTENTS.
CHAPTER
XV.
Miscellaneous Applications.
Small vibrations of a gas, and variable flow of heat in a sphere, 212 stability of vertical wire, 215; torsional vibrations of a solid cylinder, 218 Bernoulli's problem for a chain of varying density, 221 differential equatio:
;
reducible to Bessel's standard form, 222.
Note on the Second Solution of
VANISHES at Infinity
Bessel's Equation
which
22
22
Examples
FORMULiE FOR RoOTS OF BeSSEL AND RELATED FUNCTIONS
24
Explanation of Tables
Tables
.
.
.
.
24
24
I.
Tablk
Jo{x), Jj^{x), for
x = 0,
is
'01,
...,
155
.
247
values
Table
II.
Jnix) for
of
x=0,
1, ..., 24,
and
not
all integral
n
for
which J^ix)
less
than 10^^
266 280
281
Table in. The
Table IV.
roots of J^{x) = 0, with the corre spending maximum or minimum values of Jq {x
first fifty
a;
Jo{Xy/i) for
Ii{x) for
= 0,
•01,
1,
2, ...,
,
6
.
.
.
.
Table V.
Table VI.
x = 0,
51
.
.
.
282 285
In{x) for
w = 0,
11,
and x = 0,
'2,
Bibliography
28
Graph of
J.
and
J^.
COEEIGENDUM.
The analysis on p. 39 But it holds good if n
for
if w >  between  1 and ^ and the semiconvergent expressio (asymptotically) the relations (i6) and (19) on p. 13, so that th
is
invalid
.
lies
,
J^ satisfies formula is applicable for
all real
values of n.
CHAPTER
I.
INTKODUCTORY.
Bessel's functions, like so
many
others, first presented
;
them
selves in connexion with physical investigations it may be well, therefore, before entering upon a discussion of their properties, to give a brief account of the three independent problems which led
to their introduction into analysis.
The first of these is the problem of the small oscillations of a uniform heavy flexible chain, fixed at the upper end, and free at the lower, when it is slightly disturbed, in a vertical plane, from
its
position of stable equilibrium. It is assumed that each element of the string may be regarded as oscillating in a horizontal straight is the mass of the chain per unit of Then if line. length, I the
m
length of the chain, y the horizontal displacement, at time t, of an element of the chain whose distance from the point of suspension is X, and if T, T\ dT are the tensions at the ends of the element,
we
find,
by resolving
horizontally,
mdx^^(T^\dx
'^'^'^dt^dxVdx)'^'''
or
m
dt'
dx
V
dx)
Now,
to the degree of approximation
we
are adopting,
T = mg {l — x))
and hence
If
g=
write z for
{I
^(i
_^)g_^.
for
we
which y = ue'^^^,
and consider a mode of vibration u being a function of z, we shall have
d^u
dz^
— x),
du
dz
n^
„
g
'
G. M,
2
INTRODUCTORY.
Let US put
K
[l.
= n^lg, and assume i^ = + a^z + a^z"^ 4tto
a solution of the form
= ^ayZ'^
;
then
^(2ao
+ 3.2.a3^+
4(ai
+(r +
1) m^+i^r^^
+
)
)
+
2a,v2r+
+ (r +
1) a^+i^''+
)
and therefore
+ K''(ao\(hz+ +arZ'' ^ ai + /cX = 0, 4ao + K'a^ = 0.
= 0,
^ic'z^
^" 2^32+2^3^T2~
is
j
= ao<^ (/c,
say.
z\
The
series
^ (/c,
^r),
as will be seen presently,
it is
a special case
of a Bessel function;
absolutely convergent, and therefore for all finite values of k and z. arithmetically intelligible,
The
by
fact that the
upper end of the chain
is
fixed is expressed
the condition
which,
when
I
is
given,
is
which comes to the same thing,
(^ (k, Z)
a transcendental equation to find k, or, n. In other words, the equation
expresses the influence of the physical data upon the the normal vibrations of the type considered. It will of periods be shown analytically hereafter that the equation (p (k, Z) = has
=
always an infinite number of real roots so that there will be an infinite number of possible normal vibrations. This may be thought intuitively evident, on account of the perfect flexibility of
;
the chain
but arguments of this kind, however specious, are always untrustworthy, and in fact do not prove anything at all. The oscillations of a uniform chain were considered by Daniel Bernoulli and Euler {Gomm. Act. Petr. tt. vi, vii, and Acta Acad, Petr.t. v.); the next appearance of a Bessel function is in Fourier's
;
Theorie Analytique de la Chaleur (Chap, the motion of heat in a solid cylinder.
vi.) in
connexion with
It is supposed that a circular cylinder of infinite length is heated in such a way that the temperature at any point within it
by comparison of the two values of dH. coaxial with the given cylinder. Bessel was originally led to the discovery of the functions which bear his name by the investigation of a problem connected with elliptic motion. say. at time then v is a function of x and t. ot \px' X ox) function of x only Fourier writes k for KjGI). . by find The boundary condition leads to a transcendental equation to g but this is not the place to consider the problem in detail. if we put t = 5^. 1—2 . the rise of temperature dv = ^ dt. where CD .¥ axis : INTRODUCTORY. Hence. we find there is a solution which is substantially the same function as that obtained Bernoulli. so that if D is the density. depends only upon the distance of that point from the axis of the The cylinder is then placed in a medium which is kept cylinder. at a distance x from the Take a portion of the it cylinder of unit length. is C the specific heat. If the excess of the amount of heat which enters the part considered K above that which leaves ^irx it in the interval 27rx (t. . which may be stated as follows.. The volume and ^irxdx.dx. dhi dx^ 1 du X dx n k __ ~ ^ and now. and consider that part of which is bounded by cylindrical surfaces. and assumes v = '^^~"^ u being a this leads to the differential equation . t. Let V be the temperature. at zero temperature and it is required to find the distribution of temperature in the cylinder after the lapse of a time t. +K 1 ?r \ ox + ^^ dx\ ) ( ^tt^ — dx) ) dx / dt: dH = ^irK of the part is ^ ( ^+^ dxdt.x ^.^dt ot = dH. and of radii w. is the conductivity of the cylinder. t + dt) is j'^S or. except that we have \gx' instead of k^z. bv 27rxdx .
<^ which vanishes at A and A' that is. of which a focus. yu. Moreover. We may — (ft /JL therefore assume 00 = XAr sin 1 TfjL.INTRODUCTORY. defined by the relation _ ' area of elliptic sector A SP ' area of semiellipse APA' (By Kepler's second law of planetary motion. P A ing the auxiliary circle in Q. : </>) (j)) : where e is the eccentricity. coefficients Ar are functions of e which have to be deter . when <^ . e. and G the centre. Draw the ordinate NP(^ meet(7Q. or.) Now by area of ASP orthogonal projection area of APA' = area oi : ASQ : area of AQA' = (ACQCSQ):AQA' = ( Ja^</) — Jea^ sin ^ira^ = ((!>— e sin tt. thing. and join ^P. 4 and the mined. which is the same it is </>. is the Then in the language of astronomy. Hence (f) /jl. supposing that is A S is proportional the centre of attraction. where ^ It is _ ' area of sector ACQ AQA'' area of semicircle found convenient to introduce a quantity called the mean anomaly. /SfQ. — /x is a vary while e remains constant. and periodic function of fi /A is a multiple of tt. to the time of passage from to P. the eccentric anomaly of P number of radians in the angle A CQ. </> are connected relation = fi if /i — e sin by the 3 <!>. Let axis. B A' is the major be a point on an ellipse.
be obtained directly from the Ar can be expressed integral. ( Idu a? /. d^ 1 du ~ X J. we have = ~ — l. in fact. obtain ^irrAr = Jo j cos r/jb d(f>. = 1^(^1) _r# Now <^ = when // = 0. • Write X for e. iTTT^. after partial integration of du p with respect I (jy. ifFerentiating 4 with respect l^rAr cos ra to /x. but merely show that Ar satisfies a linear differential equation which is analogous to those of Bernoulli and Fourier. follow up the investigation here. = I cos r{(f) — e sin </>) cZ<^. cos r/. the expansion may. however. O/X =— \ '^^^ J cos r((j) — x sin r((j> <^) d<f> — I XJQ COS <f) cos — x sin (</) (j>) d<j) ^2 rn rhi — I X'J Q {(1 — ^cos<^) — 1} cosr r(6 — x sin 0)1 + TT — 1^ ojsin ^)c?(^ — r^i^ 7* r . 1\ x^J aa? \ . we find that dhi . and <^ = tt when fi fi = we tt] so that by .hanging the independent variable from to <^. dfjb id therefore. then. multiplying by cos r/i 'd<t> and integrating.INTRODUCTORY. (/>) c?<^.^/. in a series of positive powers of e. We shall not. and put 7rr  Ar= COS Jo r{<i> — x sin 6) d6 to . [sin — 7" = or finally '^"^^1^*. and which is 4y = — I rTTJo cos r((f> — e sin 5 function and The Bessel's expression for Ar as a definite integral.
If we put rco = z. The differential equation is.INTRODUCTORY. d^u da? 1 du X dx f^ n S)» of n is is placed upon the value the most important case for physical applications is when n zero or a positive integer. Moreover when n is integral the . the most convenient foundation upon which to base the theory of the functions we shall therefore define a Bessel function to be a solution of the differential equation . dhi dx^ 1 du X dx nu _ ~ K ' we put the transformed equation d'^u ^/!= is 1 du which is a special case of Bessel's standard form with ?^ = 0. . In the general theory no restriction analytical theory presents some spacial features so that for both reasons this case will have to be considered separately. for many reasons. this becomes 1 d?u dz^ du z dz +'+(i5)"=0' and this is what is now considered to be the standard form of Bessel's equation. If in Fourier's equation. .
] ^^+«. {r~ substitute this expression in the lefthand side of the and observe that ^x^^ = mx'^. or r=^±n.n"] a. . r=^.CHAPTER II. Now there is no first loss of generality in supposing that a^ condition to be satisfied is T^ is not zero. if we differential equation.x'+' + I [{(r + sf . since in this case when n=\. the result is {(r . and all the as with odd suffixes must be zero. with . but this does not we still have the distinct solutions is require separate y. SOLUTION OF THE DIFFERENTIAL EQUATION.n^} as + a^. * An exception occurs discussion. 2 The equation will be fortnally satisfied if the coefficient of every power of X in this expression can be made to vanish.) then. neither of these values of r will make (r + iy — n^ vanish*.2X^ + . the differential equation = + ^§ + 4! dx (^^»=)2/ dx^ may be written in the form Assume that there is a solution of the form y = x^ (ao + ai^ + a. consequently a^ = 0. The only peculiarity that r~ \ y^ and leads to both of these solutions. hence the — n^ = 0. .^ n=^.nr) a.x' + + 1)^ . we denote the operation /r ^ by ^. In general.
we take r = n. except a real : integer. . if ii is a positive integer y. A formal " V^ solution of the differential equation is there fore obtained by putting 2 {2n yy^^^' + 2) "^ 2 . therefore.' by the change which.4.(2n2)(2n4)"^'". . on account of the coefficients in the series becoming infinite.2s. to y^ is not constant hence (with the same ^/j the of the solution differential reservation) general equation is The ratio of y = Ay^ + By^. [ll. 4 (2n . are identical . it will be sufficient to suppose that u is zero or a positive integer.. Similarly when n is a negative integer y^ is still available..' 2F7¥''W7¥7%'^ is '" the rapid convergence of which obvious. y^ If n = 0. + 2) {2n + " •" 4) ^2.2 becomes unintelligible. 4 (2?i + 4) 2 . 2(271+2) a.(2n + 2s)^ r= —n we obtain the / formal y. it is necessary to discover a second integral and since n appears only in the form of a square in the differential equation.. and y.^a^ ^i + 2(27i2)"*"2. as might be anticipated. the infinite series which occur in y^ and y^.(27i solution In a similar way by putting + 2)(2ri + 4). (2?i + 2) {2n + 4) ' and so on.8 If SOLUTION OF THE DIFFERENTIAL EQUATION. but y^ is unintelligible.4. In each of these cases.we have s and hence aa =— = (2n +s)as n + a^a = 0. If n is any finite real or complex quantity. . are absolutely convergent and intelligible for all finite values of x each series in fact ultimately behaves like 2.. 4 .2 A and B being arbitrary the integrals constants. only differs from y^ of n into — n.
..] SOLUTION OF THE DIFFERENTIAL EQUATION. ao = a.(271 + The reader will a few of the first terms of the perhaps follow the argument more easily if he will write down 8um» and also a few in the neighbourhood of s = 2n.4. ^'''~' 2n ' 6'2JH3 02n Cl/2)l—2 2 (2n H. 9 quations.2n + 1) (a. * 271 (2^ + 2). 62.2n + 5) (as + 6s log ^) + (as_2 + 6s_2 log x) + (2n+ 2s) bs] x^+^ The expression y if will be a formal solution of the differential the coefficient of every term a. .~*^"^^ or d?~**+* log x on the equation righthand side of this identity can be made to vanish. 2n {2n + 2) ^''^^^ 25)' hn+. 4 . ...n^) y = — 2nhoos^ + {(.2) 2 .. and making a few easy reductions.2^1 + 2) 61) a. In order that this may be the case it will be found that the following conditions are necessary*: (i) (ii) (iii) — The coefficients bo. 4 (2n ... . + 61 log x) x''+' + ..4) 2 .. ..2) (2n .4) ' up to Oo (iv) "'*''"2. 61.^ is indeterminate . 2(2n2) a. (271 . All the as with odd suffixes must vanish.11. .**+i +X 2 {5 (. .= and so on. In accordance with the general theory of linear differential we assume a solution y = (do + bo log oc) OS'' + (a. The coefficient Uo is indeterminate.2/..(2w2)(27i2)(2n4). ^sni must all vanish.. we find that this form of y gives ^y + (x. observing that ^•^ (iC'"' log W) = 7?lV"' log X + 2w«*"'. I m Ben. a.^( y '2.2* The coefficient a. + 61 log x) + (. 4.
(2/1 1 + 2)(2?i + 4) 1 1 1^ i2"*"4^2n + and..s.. with odd suffixes must vanish. ..4. ^^ ^ ^ ^!!! 2)(27i 2. .2s (2k . All the coefficients which y= where 3/x ci2nyi + b^y. ^ (..~"^^+2 + .A 4)  M^g*' /v. and {l y^ ^^  ^~^^ + .4. y assumes the form expressed in terms of .4)+ .(2m + 2s) do not vanish may therefore be two of them if we choose a^n and h^n for these two. + 2)(2» + 4).— n+4 ^"'"^''("^)'r'^2(2. /o^ .32)+ 2...) 7 [ = iv'' ll ( .. 2) . ^__l (2. in general. (the solution previously obtained). 2'^27i + 4r^' when s > 0. all [ll. and the coefficients bon+i. a 'hn+2s ti 2 4. (v) Finally Cttvtt+a 2 (271 12) 2 {2n "^ + '''+' 2) 2 (2^1 ""^"^^ + 2) "~ 0^2n+2 4(271 + 4) + 2)l2^27H2r^"' + ^ " (2m 8) ^^^^^ 4 (271 + 4) 2 (2i 7i 2 (1 .(2n2)(2n."^ox/o.(2?i + 2)(2n + 4) 4. 2?i2)^J + "* +2(2»i + 2)2 + 2n 4) J2 + 2j "^ ~27^..... /^^ (. 2 (2?n.(27i + 2s)7 V2s +1^ + 2?^ 25>/ ..4.10 SOLUTION OF THE DIFFERENTIAL EQUATION.6.72wT2y(2?i + 4 "^ 2/1+ 2 : "^ '^"' 2^H^4 + (_).. 2.2) (2?i /2. ^ . hn+s^ etc..o/or o x+ 2 + (2?i ( o 4 .4. .25(2n + i (1 + 4). + 8 . ..
and the coefficient of x^ infinite. and F„ {x) = J". its general value 1 . ! interpreted to mean 1).n)=.II. B being arbitrary constants. proceeding by ascend It becomes any other finite value of X it is finite and calculable.] SOLUTION OF THE DIFFERENTIAL EQUATION.. ^ f.^_3)!^n+4 2! 2^^1(711)!) The when n is definition of is Jn (x) may be extended to include the case not integral by means of Gauss's function Tin. 11 The the sum characteristic properties of the integral y^ are that it is of ^/i log x and a convergent series. ing powers of x. . on iccount of the logarithm which it involves.n\. I The general and solution of the differential equation 2/ is IA I = ^2/1 + %2. is zero. y. but the quotients of these by 't^^ . for reasons which will appear as we fundamental integrals.2 When n = 0. as usual.. when k is infinite. 4.1) ^ ^"" + —XT ^n2 ) 2w5(. '" AC 2. (2/2 + 2) (271 + 4) ^ _ ] •••) =V ^ 10 is (where. its value may be more simply does not involve any negative powers of x. Tl(K. z AC) K^\ 12 .^^!^ 2 (271 +2) "^2. Un is the of same as 7i!. is the limit. «^"W2«. When n is a positive integer. 8. . and written It is proceed. {x) log X  Y'' ('^^ . not y^ found convenient. . ^ ^ (?1 + ^ 1)(71 + 2). in which only a limited powers of x occur. to take as the integer.. number for of negative when a? = 0.. but not onevalued. when n is a positive and y^ special integrals will . after the manner of x~^^ .(?1 + . These be denoted by Jn ix) and TT^ (^)j so that ^^ . which also denoted by T(n + 1).
[ll. Forsyth. Un is intelligible is and finite for all real finite n. 1 5 When to the argument write Jn instead of Jn (x). J. the function J_n (^).»*+2si 2''+^ UsU{n + s) 7 2"+^^ ri (s ^ ''  1) n (/i + s) ' ^ * It ^x'^'^~^ if 2»^+2s+i n^n (ti+s + 1) (14') is may be proved that e is •/_(h+c) when infinitesimal. 196 .. 144. precisely the limiting value of . the general definition of J^ (oc) relation the by is given "^^^^^^ 2+^UsU(n and in like + sy '"^ manner. when Tin becomes It will be well define IIO to be 1.)"«/„ so that ( : the convention is necessary in order to secure the continuity of J^ . and so on. ing. The function values of infinite . Treatise on Differential Equations (1885) Gauss. except when n by a convention we a negative integer. however. properly speaknot does it is. the convention* expressed by the formula Jn (^) = (— cc )** Jn (^)> [n integral]. if n is not a positive integer.n (^) = If n X 2^. Werke IIL p. is complex. shall not require this extension of its This being premised. thus mean —^^ . J^^ and indicate will remains the same throughout we may differentiation with respect X by accents . /_„ is defined as in and n a positive integer. convenient in this case to adopt exist.12 SOLUTION OF THE DIFFERENTIAL EQUATION.) p. to recall the properties of Un which are expressed by the formulae n (— n) li (n — 1) = IT cosec nw n(i) = jv^ Un = nU(nl) j > I 13 (Cf. By that differentiating the general expression for Jn{x) we find j:=^ ()n^+2s)a.^. ^^ U{n + s)' '4' is a positive integer. The function Iln may but we also be interpreted when n meaning.
If Jn+1 is eliminated by combining 16 and the formula is obtained . .1) n (71 + s) It will be found that with the convention 1 5 these formulae are true for result all values of n.II. It is worth while to notice the special 18 17. we have /p?l+2S— 1 / « = ^ 2^^+28 /pn+2S— 00 /y 2"+^' ns n Oi + 5 . J.] SOLUTION OF THE DIFFERENTIAL EQUATION.n + *Jn+i J. It follows from 17 that = Jn—2 and it ^^n + "n+2> may be proved by n induction that ^ ^ ~ ^ n—s ~ *" n—s+2 "1 n " w— s+4 — • • • "T y— ) *Jn+Si the coefficients being those of the binomial theorem for the expo nent s. which is the same thing.^ it will be found that 2n —J.9) . and similarly by eliminating Jni J.1) "^7 n (s . UC Again. Jo=Ji. and are continually required in applications. that is. writing n 00 + 25 \S in the form {n\s)\1 s. —^ ^ 20 The formulae 16 — 20 are very important. or. 13 (on writing s+1 for . J'.
1 . and II = — "" JLn ~ '^4. and may be function of the if" order. 31 referred to as the Yn was discovered by C. ill. F„+. The analogous precisely similar. —^W ^ n 1 ^\ "•^' 22  '^"J ^^—^^ .= function i (1 + ^^) 1 1 ..«. 18. is distinct F^ is which moreover a solution of Bessel's equation which satisfies the relations * from /„. formulae for Wn. 20. Ann. «. (1871). The Neumann . + Wn^ "'„+. 30 where A. 143). ^. 23 Now if we put 2o^ \ =X"5 24 with the convention <Tq = 0. are Tr>F„_. process. which may be obtained by a although more tedious. 16. who obtained it by a different process from that here employed.14 SOLUTION OF THE DIFFERENTIAL EQUATION.. X Y'.F„+F„_. [ll.. The explicit form 30 was given by Schlafli {Math. 20 CC * 7^ * «— 1 ~ ^n> 27 = 0.^F„ + — W 4W W. 19. Neumann. T.= 28 29 Y.3)! r. p. which are of exactly the same form as The explicit form of Y^ (n is Yn = Jn log ^   J2«i 1)! ^^ + 2^3 (^_ ^ J 2)1 2*^5 (w.
= 2^111(711)^^^+. put y = uJn{so). Taking the general value of Jn (x).i(a. is AX n(^)j ^ dx The leading term of J_n (^) 2n yfT \ ^~^ ^^^d ^y making thi lagree with the preceding..... n not being an xJ^. d'^u du dx rhence X Jn {so) 0.(x) r^ integer.1) ^" + ..) where u is a new dependent variable is an integral.. = 1 22n (^7^)2 „2«+l + . By properly choosing bo the constants A and B it make y identical with JL^ (x) or Y^ possible (x) according as n is not or must be an integer. we find that : then. that J dx xJ^(x) = _ 22^1 nwn(n . we 2 obtain 2 n (n — 1) n (— n) The value of B will depend 33 TT upon the lower dx limit of the integral p J xJi{xy ..SOLUTION OF THE DIFFERENTIAL EQUATION. There is 15 another way of deducing a second solution of the In that equation differential equation which deserves notice. observing that J.. ^•^nW^=^ A ['^ = dx „ 32 the complete solution of the equation.
[ll. We are thus led to the curious result that "?i+] * n "~ ^nJ^n+i =^ ~ • 3^ 30. the value of B depends upon the lower limit of the integral. i^« find.16 SOLUTION OF THE DIFFERENTIAL EQUATION.Jn {^Yn n\l ~ Yn+\ {j^n ~ ^^n+i) Yn+i) = identically. . 34 to the form With the help of i6 and 19 this may be reduced 2 JnJn+i — sin mr. when n is an integer. 19. 27 that + Jn+i {nYn . that. Hence u is independent of x : and by making use of the explicit forms of Jn and Yn it is easily found that the value of u is unity. is Returning to 32. + ^^« 37 As in the other case.xYn+i) . we a positive integer. If u^^ X (t/^i+^y^j. by comparison with when n = ^« ^. to X Divide both sides of 32 by Jn{x) and differentiate with respect then in the case when y = J^n (^) we have : d [Jn\ dx \ Jn ) or 2 sin nir ' TTXJI JnJn — JnJn = 2 sin TTX TlTT. — J ^y n+\) we find with the help of 16. + JnJni = irx Neumann we write 35 This suggests a similar formula involving functions. 26.
it will be supposed. or is J_n (a?). is a positive integer.CHAPTER III. so that we have + 00 _oo identically X £t G. unless the the that parameter n. contrary The exdefinition of the Bessel functions. exp 2 (^ . Throughout is this chapter pression for the Bessel function of the rC'^ order is which may be written in the form hence we conclude that Jn {x) of exp is the coefficient of ^'^ in the expansion 9(^ — to powers of t 7) according In fact. EXPANSIONS IN SERIES OF BESSEL FUNCTIONS. obtained by putting r — s (_)«:pn+2s = n.^~') = exp  2" exp — ^— "0 2^r! 2^s! and the coefficient of t^. I . FUNCTIONS OF INTEGRAL ORDER. M. which occurs in the expressed. is In the same way the coefficient of t~^ in the expansion (—)**/„ {x).
. Multiplying the first of these four formulae by cos to tt. 44 Similarly / sin ncf) sin (j? sm <p)d(f> = ~^ {1 ~ ( l)"") J» (•^) 45 By addition. 39 and hence cos (x sin ^) sin (x sin <^) = Jo (x) =s \ 2J2 (^) cos 2(f) + 2/4 (x) cos 4<^ + sin 3</> . A r sin I r/jb. . . n<l) ncj) and cos {x sin <^) dcj) = ttJ^ (x). . . . I cos /. in a single formula. . r Jo COS {nj) — X sin <^) d^ = TTt/ii (^r). 40 2 Ji (^r) sin ^+ 2 Jg (a?) + 2 J5 (a?) sin 5<^ + . W 42 . n<j> cos (x sin (j>) d(f> =^ [1 + (— 1)**} J^ {x). . if n is odd. [ill.3(/) 43 These formulae are true for all finite values of (p. — or. It will be remembered that an integral of this form presented itself in connexion with Bessel's astronomical problem in fact we .r into thus ^ —(pi = Jo (^) * 2/2 cos 2(^ + 2/4 (x) cos 4</) — — 2 J3 (a. arrived at the result that fi if = (j) — e sin 1 (j). 0. . . if n is even (or zero). 41 Change cos (. . where Ar= — cos r((j>—e sin 6) d6. 46 which holds good for all positive integral values of n. we obtain integrating from cos /. ^ j^ (^^ ^ 2i Ji (x) sin + 2J2 (^) cos 2<t> + 2iJs (x) sin 3<^ + 2 J4 (x) cos 4<^ + (f) .) cos sin (a? cos 0) = 2 Ji (^) cos <^ + 2 Jg (^) cos 5^ — cos (f)) . The absolute value of Jn+i{x)IJn{x) decreases without limit when n becomes infinite hence the series on the right hand is : absolutely convergent for all finite values of x and t Suppose that t = e^'. then the identity becomes ^ixsin^. then = (j> fjL + 1. .18 FUNCTIONS OF INTEGRAL ORDER.
should be observed that the if 47 is absolutely convergent.... .2). + 2^2« + .... + l) ^ . 19 We DOW see that Ar may be written in the form Ar = Jr(re). H . [ . and let the expressions to powers of w cos then (/> : transformation be applied to the identities 42 and on the left hand be expanded according by equating the coefficients of cos*^ (f> on both sides we find x" = 2" »! . + (« + 2) J„« + (" {j„ (n + 4)(« + T) ^^^^ + 6)(»Mj)(« + l) ...... and this decreases without limit when s increases indefinitely.. ^^_^. 2—2 I .l)(...2. The first three cases are 1 n. . + V/^s +..(. + 6/3 + IOJ5 + . In confirmation of these results series in brackets in it . and in this notation = </> /ti + 2 \ji(e) sin /i + 2 ^2 (2e) sin 2fi ^^Js (3e) sin ~ ^^ 1 3/^ + . x = 2J. . ...+ l) ^g^^^^^./o + 2J2 + 2/4 + ... (?i + 2s)(ri4sl)(7i4.(..92). and also for .(/7._..EXPANSIONS IN SERIES OF BESSEL FUNCTIONS..^ _ Now let this 43. for we write it Xcg Jn+28 we have Cm ^n+2. + 2 (25 + 1) /2m + x' = 2 (4/2 + I6J4 + 36J6 4.+2 ~ ^ Cs Jn+28 4 + s)x^ (5 + l)(n + 2s){n + 2s + l) (n 2(2?i * + 4g + 6) ^ 2(2/1 + 45 + 2) + .). It is known that 2 cos n(t> = (2 cos ^)«  ^ (2 cos ^y' + J ^^ ^'^ Ji (2 cos ^)'^*  . ^?i+2s ^ + 1 p 4/ ^^ which holds good for all positive integral values of n = 0...
. I ^2«(sl)(52)(s3)the 8! ^•••J" ^ . .20 Moreover if FUNCTIONS OF INTEGRAL ORDER. ^ ^8 ^ n+28 becomes ultimately infinitesimal when h increases indefinitely. rearranging the terms. .^. the expression on be made as near to x^ as we please by taking may h large enough moreover the series . sum it can be proved by induction that ^ 2n. in general. . OS ^*~^ 7.. we obtain the expression 2 where and.n(n ^ .. 2'^+'{n + 2h + l)l(h+l)\ r Since the series in brackets the right hand 2'(h is + 2){n + 2h{2).(rk) 2^^(n + r + /0!r!/i! ' or. . = ao. = 2ai. for each power of x its expression in Bessel and functions. Now suppose we have an 00 infinite series X agx" = (Xo + a^x + a^^ + then on substituting 00 .t/ + mrl)(r2). therefore the relation 00 X is = Z .n^'l^_y.. Sh denotes the [ill. ^ 1 ^g2 L ^4 1! ^2'{sl){s2y 2! ^'r^ 2^51)^^ I of I I . bo hgJg — hoJo + 6i biJi + b^2 + • •. b^ = Sag + 2ao.nlS. .o sum within brackets ending with a term in ax or «« according as s is odd or even. which "" is the same thing.2\'^"'\' convergent. . . Ul 2^ CgJn+28 all finite true and arithmetically intelligible for values of x.
we may put = V)Js. «S1 when for all where « is a definitely In this case the assigned proper fraction. and the series XhgJg is absolutely convergent. andhence ^^Y'l ^^Y'.iauchy's first test is when the . If is ultimately equal to unity have to be we are assured of the possibility of an expansion such as that here considered. positive integral values of s above a certain limit. may be found by any method As an illustration of this.) . limit of bsJs/bsiJai is agoo/asi. / . If the series SugX^ 21 and XbsJg are both absolutely convergent. (JC = ^ . nd the arithmetical truth of this relation 49 may be verified by a aethod similar to that employed above. the coefficients which proves convenient. </c.. series ^agsf satisfies of convergence that is to say. by 1 8. . Then Y'. Now by repeated application of 20 xJi = 4 Jg = 2(2/24/4+ eJe.I rj EXPANSIONS IN SERIES OF BESSEL FUNCTIONS.. let us assume Fo = Jolog^ + Sc^/^. 1 Therefore Xn^CnJn = — ^^J'o = 2^Ji — ^«^3 . = J[ log ^ + ^ + Sc^ J^.„ „ ^+2 Cn(jn+Jn^Jn) X 2j.^Y. Cases in which asxjas^^ examined separately. IXagOf An important special case .
6 which may be transformed into F2 4.y x.^'^ T 2. and therefore — t/ —= 12 2 1«/2+^4 + J  1^1 Consequently 2. Now we have identically. Differentiate both sides with regard to and apply thus that is...3 ^ 2.. Hence and finally tn'CnJn = 4 (2/2 ..10 ^^ •••}' = J2 log ^ ^ 2 ^0 5 ^2 . 18.2) (4s + 4)^^+' + 6)^+^"^" ^^ + ^2. by 40 or 47..22 FUNCTIONS OF INTEGRAL ORDER.5 ^ 6..4J4 ^QJ^ .5 V n 1 ^ T . [ill. Fo = /ologilJ+4^2«^2J4+/e. 50 17.7 ^ 6(4^ + 3) 2(4^45) (45 44)(4s ^(45 Similarly 4. l=/o+2e72+2J4+..). 29.8+6.
will be found in the .H:^^^^" ^ may be proved by is induction with the help of 20 and 28. + l).+a. 71 "^""^^ +4 fn(n + l) J + n+6 f^i(!L+JiK!i±?) + 9I r 3! 2^. Other illustrations of these expansions examples at the end of the book.. If There write another interesting way of expressing F^. + s) ( (. then we have ?? +2 Fn= /« log .or again into jz n /^2 1\ 3 ^ 2. + "^(251) (25 + 1)*^^ +"25 (25 + 2) l) ^ ^^«+^+. 5 ! + 5l) _^ 1^^ J and the corresponding expression for can at once be written down.(25 + 2) (2.l) (2.4 ^ 1. 6FT3)1 .r^ + 2(n4l) '^ "^ ^' a. . we Tn = 2«i (7^  1) 1 ^« + ^ fp2"°(yi3)l 2! ^^'''^' ^ Wn and "^ _^^^ "^••• + ^"^ 2^(^^ry)l' 57 a polynomial which has already appeared in the expressions for Yr^.3 ^ B H 2. 55 Neumann has given a general formula for F„ which written in the form may be "2 This r ^5W r!"7^"^57... 2s (n n+25 fa(.
... (V) + Jn+4 (U)J. [ill.. . + Jn+2 (u) J2 (V) + Jn+4 (U) J. {u) J^^s (v) + Jn+s {^l) Js {v)] find that ..Js (u) Jn+s  W .V)} = /i (U) Jni (v) + Js(u)Jn3(v){.t = = ex^^{t\ exp(^l)exp(^l) = ^ijn{u)t'''^ijn{v)t\ 00 —00 t Multiply out..Ji (u) Jn+i (v) . 60' . By 38 we have ''ij^{u^v). A some of these will now be given.. + Jn+2 (U) J..... ..+Jn (u) Jo (v) + J2 (U) Jn+2 (V) + J4 (u) Jn+i {v)+ ...{V)+. 59 n Observing that Jn (— is odd 2 [Jn v) = {—YJ{v). and equate the coefficients of the powers of both sides thus : on J. {Jn{u)Jo{v) . and in general n + i {y 1 if [J.. and if 60 n is even 2 {Jn(u + V) + Jn{uV)] = Jo W Jn{v) + J^ (u) Jn2 {v)+. great number of valuable and interesting results are connected with a proposition which may be called the addition theorem for Bessel functions .{u + v) = Jo{u)J^{v)^J^{u)J^{v)v2J^{u)J^{v).24 FUNCTIONS OF INTEGRAL ORDER. we (^^ + V) + Jn (U .{V) + .
It should be noticed that Jn (yi) immediately presents according as n is even or odd.Jn+i (u) Ji (v) .Jn{uV)} = Jo(u) Jn(v) I + J. (u) Jn+2 (V) + J4 (U) Jn+i {V)+. . +00 —00 Neumann only considers the case when n = 0'^ but the itself. . (v) .] EXPANSIONS IN SERIES OF BESSEL FUNCTIONS. . k = then t> sin fl + 00 — 00 * Strictly speaking.. By putting i^ a? and we obtain from these formulae expressions imaginary parts of Jn {oc = v= are real quantities. . . .. generalisation immediately suggests . .. 61 f is even (W 2 {«^» + ^) . for the real and + yi)..V)} = /i (^0 /„_! (v) + J3 (u) j„_3 (v) + + j"„_i (u) j.. By 38 we have exp I (kt  ^) =3*^Vn W t^ .Jn (U .Js (u) Jn+3 (V) . 61' where a. r2 Jn+i(f^)J'i(v)Jn+s(u)J'3(v) . .III.) <". — 00 0) 2 k"Jn e^^ : {so) «" = e^ ''"« ^ 2 /„ (ia. now therefore I {^lyii^iHH)]' that is. .Ji (u) Jn+i (V) . —* 62 Put = r.Jn+z (u) J3 (v) 71 . Similarly 25 u if n is odd {Jn (U^V). and y 2/^. itself as a real or purely imaginary quantity We will now consider a remarkable extension of the addition theorem which is due to Neumann*.
let ^ = TT ^ .(r) + 3)^J"„+3W+. [ill. thus (ri) = ^« /„ (r) + rJ^+. 64' ^ cos(w 3! As a Jn special case. we have /„ (re«) = eVn (r)  ""^^ ei»+"«J„+.26 FUNCTIONS OF INTEGRAL ORDER. J^+. 6 successively equal to 7 and multiply the results together thus . an expression for Jn (6e^* + ce^*) in the Jn (he^' + ce^O = Co  Cii (6 sin /9 ( + c sin 7) 72i^(6sinyQcsin7y .. Returning to 63 let us put r. (r) + l". /3 and —00 —00 = e" i(6sin/3+csinY) +00 t +<» — 00 1 e^^^Jn (h) V' S —00 e^y'Jn (c) <". • . ^ . 66 Now the lefthand expression is equal to 2Jn(heP'\cey^)P'. (r) where f„ = J„ (r) cos nd + r sin 2! ^ sin (w+ 1) ^ J„+i (r) ^ cos(?z r^ sin2 + 2)^J'„+3(r') /I r^sin'^ ^yand rjn ON r / \ sm (71 + 3) eJn+3 (r) + . — 00 and form if the righthand side the coefficient of t^ gives is expanded according to powers of t. ... . c. h. (r) + 1^ J^+s (r) + . Equating the coefficients of P. . / 64 = «/% (r) sin n& — r sin ^ cos (w f 2! 7^ sin^ 1) ^*/»+i {r) ''sin(7z + 2)^/«^. . 65 .
(c) cos a + 2J^{h)J.EXPANSIONS IN SERIES OF BESSEL FUNCTIONS. 0. 1 69 we change a into tt — a.2 J. (6) J. etc. then we have {J¥^) = Jo (6) Jo (c) . where Co 27 = e^^'Jn (b) Jo (c) + e{^'^ ^+y}^ Jn^{b) Mc) +. 70 and hence. a2 = (6 cos y8 + c cos 7)2 + (6 sin /? + c sin 7)^ =^b^ and also Let us put /3 + 2bcco&(l3y) + c^ 6 sin /8 + c sin 7 = a . 0^. Moreover we shall suppose in the first instance that n If = 0. . we put we have be^^ + ce^^ = a... by supposing J. a real quantity.ei^^^^) ^'y)'Jn+. then we are brought back to the addition formulae. however. 1 69' and this is Neumann's result already referred to. 71 . Suppose a Jo =^ . . .. (c) c = 6.{c) .2J! (b) + 2J!(b). (c) = Jo {b) Jo If + 2i (yJs(b) Js(c) cos 5a.2 J.. . put a = . .. (b)Jn{c) Py}iJ^^^ (6) J^ (c) 4. (6 V2) = J^ (6) .^ (c) cos 2a . 68 manner for Cj.. we shall only consider in detail the case when be^^ + ce^^ is a real quantity. (c) cos 5a.. (c)  . this formula is too complicated for practical purposes. + ey'J. By way of verification. (b) J. Since. keUn+i) and in like e{(n+x) y^}i J^ Q)) J^^^ (c) + e{ («+2) y^?]ij^ (b)Jn+. (c) + 2J4 (b) J... the formula becomes Jo {j¥ 2bc cos + QL c^) = Jo (6) Jo (c) + 2S J* (6) J. . —7= then the general formula 6y becomes in this special case Jo (Jb' + 26c cos a + c^) = Jo (b) Jo (c) . (b) J..
. 4. (b) J3 (c) Substitute h out.. thus 2 J" (h) cos 2a (^) cos (26 sin I) = JS(b) . In order to obtain another special case of 67 let us suppose that n = 1.}. Then if b.2Jl (6) cos a + + . Since the four quantities 6. /3. and .28 FUNCTIONS OF INTEGRAL ORDER. (c)}  {bJ. retaining the condition 6 sin y3 + c sin 7 = 0. (c) . we will write a. etc. + ce«* for ae^* on the righthand side. the formula 6y gives aJi (a) = ae^* J^ (6) Jo (c) + e«^ Jo (6) Ji (c) { eJ. (b) J.Jnx (C)] Jn {b)] COS wa. Let us choose — 7).bJ. .cJ^ (b) Jo (c) + cJ. (c) . (b) Jo (c) + cJo (b) J. c. 7 are connected by this single relation. . This result 73 may also be obtained from 69 by applying to both sides the operation ^ = ^db^'do^ means of obtaining in fact this operation affords the easiest the formulae for n = 2.(b) J. (b) J. it may be verified. (c) 4. In 69 and 69' suppose 6 Jo cos [ill =c .Jnl m Jn (c) + C [Jn+i (C) . (b) J. . as before. (c) + e'' J.(6)Ji(c) + e^^J3(6)J2(c). c and (/3 a has the meaning already assigned to geometrically or otherwise.{b) J. (c) cJo(6)Ji(c) [Jn+i Q>) . (b) J. for which.e''J.^ 72 J"o (26 I) = Jo'(^) + ^Ji (jb) cos a + 2/ 2a . . that it. (b) J.cJ.cJ. (a) = {bJ. (c)] cos 2a + [bJ. multiply and equate the real part of the result to aJi (a). there still are three independent quantities at our disposal. 8. Now when n = l. thus a J. . (b) J. . (c) = 6Ji(6)Jo(c) + + S ("T [^ 1 bJ. . . (c)} cos a .
.] EXPANSIONS IN SERIES OF BESSEL FUNCTIONS.. Q>) cos . and equate coefficients thus 1 = e7J(^) 4 2Ji(x) + 2/ (o^)  ..2 J.III. may b. 29 Special results may be from 69.3 = = 16 2/0^1 + 6/1J2+ IOJ2/3+ f . + 1)(2. that /. (2^ + 52) ^ . + 2 (2s + 1) JsJs+i + . (h) cos I ^ 74 + This 2Mb)J..(b)co8^„. • .. cos ^26 2) = 2/0 (b) J..... we find. i ^(. • r . and in general + + + 2(7i— g)(2n . putting obtained from 73 like those deduced 6 = c.. after a little re B duction. ] 76 .(2n + gl) «/^+s+. jji/2 5/2/3 + 14/3/4 + .K 75 ^2 ^ ) ^^ while for the odd powers of x X a. (b) J. — l)(2?i ^j + 2 ).. + g 1) ^s^s+i + ) . expand both . and in general ^1 = 22«in ! (n  1)1 /«_i/^ + (2n + 1) JnJni /o ox r r + e3) + n yln '/n+ie^n+2 H 2n(2n + o)(2nhl) 0~j J j e^n+2e^«+3 + .... • 2n(2M + 25l)(2/i + l)(2n + 2).. also be deduced from 72 by differentiating with respect to In 72 and 74 write x for 6. For example. sides according to powers of cos ^ ..
any function /(cc) can be expanded in the form f(x) = ^ao+ a^Jo (x) + a^Jo (2^) + I . ^^ 79 The lemma may be established as follows. 5. We have therefore V n "'J 2^+1 5! (5 + 1)!* 1. . 64= 64^4+ lOdg + 2ao.3. ' I — rrJo To prove this.. This result is also due to Neumann (Leipzig Berichte 1869). By means of when and so on. to prove that The arguments employed already we may write 00 00 in a similar case will suffice ^a^ = % {b^Jl + hs+iJsJs+i).(25+1) =t (_)» n^+i u^+i (25 1 _ cos nu + 2)! WW .'o Jo Jl — and. . 62 = 4a2 + 2ao. 75 and 'J^ the series '^a^ may be transformed into 60 = «o> ^1 = 2ai.30 FUNCTIONS OF INTEGRAL ORDER. 7S TT. + an Jo (n^) + ^^ ) . We will conclude this chapter by a proof of Schlomilch's theorem. ^n =2 r i< cos ^ nu i{rrO(^)d^]y da. we IT shall Vo v/l?M require the lemma • P 1 r / ±\ JA Jo Ji{nusm(p)a(j) =1 — cos?m . if ?i > 0. which will have to be examined. . 77 and that this is arithmetically intelligible so long as x remains inside the circle for which Sa^^ is absolutely convergent. . &3=16a3 + 6ai. . . that under certain conditions. [ill.
r
^ft
we
EXPANSIONS IN SERIES OF BESSEL FUNCTIONS.
if
31
Now N(
we assume
1
/(^)
shall
==
^
^0
+ cixJo
(.2:)
+
...
+
anJo{nx)
+
...
have
/' (x)
=—
(f>
(Zi
Ji
(a;)
— 2a2Ji (2x) — ... — na„ Ji (nx) —
TT
x
;
....
Write u sin
for
^z;,
and integrate both
thus
sides with regard to
</>
jetween the limits
and
Jo
I
/' (u sin ^)d(l)
— —z nan
1
Jo
\
Ji (^^ sin
<^) d(f)
_^an(cosnw
W
id therefore
— 1)
tt
Jo
I
/' {it
sin
^)
c?(^
= ]£ an COS
1
?llt
—S
f
ttr
— 2 anCosniA+
Hence
TT
(Xo
—/(O))
.
So

ao
/(O)
=^
j
1'^
j
/' (^
si^^
^) ^*^
j
^^
',
which
is
the same thing,
Oq— id,
\
wl /' ]/(0) +
(i*
sin
(^)
c?</)
[
c?t*,
when
ri
>
0,
IT
Gtn
=—
I
It
COS
nu\\
f {u
sin
<f>)
d<j)
[
du.
putting sin (^ = f we obtain the coefficients in the form given by Schlomilch but it must be observed that the theorem has not yet been proved. All that has been effected is the deter
On
,
;
mination of the coefficients, assuming the possibility of the expansion 78 and also assuming that the result of differentiating the expansion term by term
is
f
{x).
32
FUNCTIONS OF INTEGRAL ORDER.
[ill.
tti,
In order to verify the result a posteriori, let the coefficients Oo, eta, etc. have the values above assigned to them, and let us
write
i/r
(;r)
=
2
flo
+ (h Jo («) +
u
Ota */o
(2a;)
+
. .
.
.=/(0) +
f
u U\
Jo(£c) COS n
+ Jo{2x) COS
2u
+
...j
F(u)du,
where
F{u)=\
Jo
f (u sin ^)
IT
d<l>.
We
have
^ ^
Jo (w^)
=
I
Jo
cos (nx sin
cos
?ii/
<^) c?<^
_ p
on putting x
sin
lies
(j>
du
Jo \/af^u^
= u.
and
tt,
Now
if
a;
between
we
are entitled to assume an
expansion
'6(u)
= Co + Ci cos U + C2 cos 2?^ +
from
. .
.,
— u^)~^ ^(i*) being equal to (x^ = tt the zero from u = x to i^
;
u=
isolated
to u = x, and values u = 0, x,
equal to
tt
being
left
out of consideration.
By
the usual process,
therefore
while
if
n>
IIT.]
EXPANSIONS IN SERIES OF BESSEL FUNCTIONS.
This
is
33
the series which appears in
>r
(x),
and since the value
to
tt
of 6 (u) is known for the whole range from = tion of an isolated discontinuity when u
with the excep
x,
we have
=
Put
it
^
f(0)
+
ttJo
^r Jo ffi'^l^^M^^t sIx'u'
sin
(l)
— ^,u cos
(f)
=
v''>
^^^n this becomes
= /(0) + {/(«) /(O)}
This shows that Schlomilch's expansion is valid, provided that between and tt, and that the double integral
X
lies
Jj V^2
_
f2
_ ^2
0,
7;
taken over the quadrant of a circle bounded by f =
= 0,
is
p + »?^«^ =
admits of reduction to ^
{/W ~/W}
which
certainly the case, for instance, if
f(x)
is finite
and continuous
over the whole quadrant.
CHAPTER
IV.
SEMICONVERGENT EXPANSIONS.
The powerseries which have been obtained for J^ and Yn — Jn^og X are convergent for all finite values of x, but they become practically useless for numerical calculation when the modulus of X is even moderately large it is therefore desirable to find expressions which approximate to the true values of Jn and Yn when X is large, and which admit of easy computation. The expressions which we shall actually obtain are of the form
;
Jn
=
a/
— {P sin
a?
+ Q cos x]
cos
x],
Yn = a/
where P,
Q,
^ {Rsinx + S
R,
S
ing to descending
are series, in general infinite, proceeding accordpowers of x. It will appear that these series are
ultimately divergent, and the sense in which the equations just written are to be understood is that by taking a suitable number of terms of the expressions on the right we obtain, when x is large,
the approximate numerical values of Jn and F„. The approximate value of Jo when x is large was discussed by Poisson *, but not in a
in this chapter we shall very detailed or satisfactory manner follow the method of Stokes f, which is important as being capable of application to a large number of functions of a kind which
;
In a later chapter frequently occurs in physical investigations. the question will be discussed in a different manner, depending on
the theory of a complex variable.
Sur la distribution de la chaleur dans les corps solides. Joum. de I'Ecole Polyt. cah. 19 (1823), p. 349. t On the numerical calculation of a class of definite integrals and infinite series (Camb. Phil. Trans, ix. (1856 read March 11, 1850), p. 166 or Collected Papers,
; ;
*
II.
p. 329).
On
the effect of the internal friction of fluids on the motion of pendulums
;
(Camb.
Phil. Trans, ix. (1856
read Dec.
9,
1850), p. [8]).
IV.]
SEMICONVERGENT EXPANSIONS.
35
iix~^\
In Bessel's differential equation put Jn{oG) be found that u is a solution of
=
then
it will
Now when
is
x
if,
small
;
and
— large compared with n, the value of (n^ l)/x^ after the analogy of the process employed in the
is
expansion of an implicit function defined by an algebraical equation f{u, x) — 0, we omit the term {n^ — J) iila? in the differential
equation,
we obtain
of which the complete solution is
II
= ?^ = ul
sin
X
h
B cos x,
are justified a posteriori in regardthis as an solution of 8i because this value of u ing approximate does in fact make (n^ — l)u/x^ small in comparison both with u and
where A,
B are
constants.
We
.
,
d"u
^r„.
with
Let us now try to obtain a closer approximation by putting
II
=
i(,.2
= IAq
^
\
j
sin 57+ ^0 H
(
—
]
cos x,
where Ao, Aj, Bq, B^ are constants.
dx"
a^ x^
This value of u gives
a?
\
I
\
]
ay^
a?
)
The expression on the
right becomes comparable with
x~^, if
we assume
2A,
= {n^\)B,,
2B,=
The value
Uo
of Wo thus
\
(n^i)A,. becomes

= Aq Ksm X
——
^
cos
d?^
+
Bo
Icos
x
^ ^— sm xh
and we have
It will be observed that we thus obtain not only an approximate solution when x is large, but an exact solution when n= ±^
or
I
±
.
We
shall return to this second point presently.
3—2
I
36
SEMICONVERGENT EXPANSIONS.
Let us now assume
[iV.
u = u^\then
it will
J.2^ sin x\ Bo. cos
—
x
;
be found that
dx^
M (.iK.!
)a.naj1i;
^^
or
If,
—^ (J.
2
sin
X
+ B2 cos
a?).
then,
we put
2.4
the value of u becomes
U ==u,
= A,
jsm
^+
*
^2^
cos
X

^
sm
^^^^^ ^
+ Bq ^cos X
and we have
s—^ sm
57
—
^
^^——^ cos x]
S.(.^*)..J i)(.J>(.Y) (^.^,,^...,,
Proceeding in this way, we find by induction that,
Ur
if
we put
82
= AoUr + BoVr
cos
iz;
where
Ur = sm a;
+
_
^
^
zx
\ ^ ^ 2 4a^
.
sm a? 
. . .
+
2.4.6...(2r2)a?i
Vr
sm
(^^r
+
+
. . .
^
>
^^^
«3
= COS
a?
^r—
i sm X —
i^Vi;

COS a?
+
cos
2. 4. 6... (2, 2) a;'
(^a;
+
^
'"j.
»4
. less than 2{r—l)a. and the analysis we have employed shows the exact is the approximate value of a solution.I IV. that it is not to be inferred that the expression above given for Ur reflects in any adequate way the functional properties of an exact solution and . do not debar us from employing the expressions Vr up to a certain point for approximate numerical that as r is calculation.6..»'+i . however. which is Ur constants: this x. It is to be observed that the closeness of the approximation has been estimated throughout by the numerical smallness of the ex pression the value of which is zero for an exact solution . independent of Then we have S^ = C {Pr COS (a — x) + Q.l SEMICONVERGENT EXPANSIONS. . small and J iir is a closer approximation than Uri so long as n" — ( —^— is numerically. It is convenient to alter the notation by putting Ao sin 0C + Bo cos os = G cos (a — x).x)}. 37 1 + The expression ^^ is 5oCOS^^+^^^7rjL 85 therefore an approximate solution so long I is 2.(2r2)a. 86 because C and a being new G' = Al + Bl. These considerations. sin (a .. taken larger and larger the expression Ur finally. deviates more and more (in every sense) from a proper solution. sense in which %. and the degree of the approximation obtained.4. is legitimate.. Aq cos x — BQ^mx = C sin (a — x).
by 14. positive. _^1^ 8a.. x~'.l .a. . a so that What remains to be done is to determine the constants G and %^~^ may be an approximation to the special solutions by Jn and J_n or? of Bessel's equation which are denoted is integral.3(8^)^ ^^ x^.49) 1.1) (47^^  9) (4?i^ ..1) (4n^ . (2s 1) (2s .2.38 SEMICONVERGENT EXPANSIONS.^J„(a^). and J^ cos cos ^) sin^^ <i>d<f> = ^irU = {n  J) S ^^^ n\ 11 (71 + s ) 2V7rn(7ii). (4n^ . x~^..25)(4?i. n(t) = n(J)^ V^ 2'' Us 2^ Us* n(2s) (a. ^ (4n^ ./^ . involving ^r+i respectively. ^ t^ n(25)n(9i+s) • Now since s is a positive integer and therefore n (2s) = 2'U (s) . cos <^) sin" </> c?(/). [iV. This proves the proposition. when n Jn in and by Jn and Y^shall require For this purpose we a new expression 7i for the form of a definite integral. cos <^) sin^'^ </> (Z</> =2 1 cos (a. Qr between them contain r terms...9) (4?i.25) ^•• 1.2.i) = i( y^....4 {^xy 88 n ^^~ and P. cos ^= 1 .3.. it may be proved that "^^ ^"^^ Supposing that C^ +^ is real " 1 • x"" V^ 2^^n(n^ 1) j we observe that fn J ^^^ ^^ ^^^ "^^ ^^^'" "^ ^'^' ^° To show fir this. ( \S rfps COS^' di Jo AA \^S) n(ni)n(. consider the integral cos I Now U= and put (a. where 1 r — ^ (4w2l)(4?i29) 1 2 {Sxf .3).
it 11. (2n + l)7r .] SEMICONVERGENT EXPANSIONS.^ 1 fjLX .. (2/jb — ^y^ d/ju. a^y^ dfi. that the approximate value is when x is very large and positive 2«+^^. Jo (2fjb In the integi'al Ui put fix = t: then assumes the form 2 2x) 1 ^^^HJo' "^ where V=\ xUl^T (Y llp''^costdt to ^ a? By {^' breaking up the interval from ' ' ' • into the intervals I) (I '2) t) • • • '''^'^''^ (^'"2" '''"')' (* ^) '^ ^' the greatest multiple of tt contained in . on integral ^ dcj)— writing cos ^' cos /i^* + sin ic sin /^' Uii for cos (^z. =2 = 2 Jo sin j cos fiw . 89 then and. V is finite. it may be proved that hence the most important term in Ui x'^^Jo 1 Qn+h fx /»*• COS tdt. the becomes U—Ui cos x+ where 17^ sin x. X becomes.IV. cos <^). it In the same way of f/o may be proved . . however large x may be is x. and This represents the true value more and more nearly the larger at the same time approaches the value (supposing that x is positive).
i. when a? 94 then the approximate value of Yn quantity.. (1869).40 SEMICONVERGENT EXPANSIONS. If we by . 93 is and 2 the natural logarithm of the value of which log2 = 69314 71805 59945 30941 72. 471. pp.1) (4?i^  9) (4?i^ . P {(2n + l)7r ]1 .yjr(0) if log write 7 for Euler's constant.25 ) For the sake of reference it is convenient to give in this place the corresponding expression for Yn which may be deduced from a result of Hankel's (Math. is. 92 ( 4y^2 _ 1) (4^2 _ 9) (4^^2 _ 25) (4n^  49) ^_ 47z2 ^ _ (4w^ . is a large real positive F. which is otherwise denoted or n^O) h n(0). 494) to be considered later on. and the value of which is ry = _^(0) = is 57721 56649 01532 86060 65. to the [iV. same degree of approximation... Ann.. ^ l)7r n ooY .^1 .. and hence. 2. (CO) = (log 2y)J^ (X) + y^ [q cos _ jl?!^+i)^ . Comparing this with 87 we see that and hence that ^^"^^ = V i [^ '^' f"^^^^ ^ ^ f(2n + Qsmj^ .
— Hence and by means of the relation 2^^ _ — "~7 Jn — oTn—\ d T o n+1 X . 97 rela It is interesting to confirm these results by means of the tion t/ft+i In — dn t n+i • The approximate formulae for J^ and F„ hold good for all values of x with a large modulus. Thus to the first degree of approximation = (log 27) Jn + The value 2 Jn+i. a? X = A // 2 sm X. provided the real part of x is positive. of log 2 —7 is. log 2 7 = 11593 =" CD 15156 58412 44881 07. By the general definition of Jn *^^ {x).i) f irx " 2TT ~ 2 . 41 the values of Jn P> and Q on the right hand being those given above. V irx . already alluded to. is a pure imaginary will have to be considered separately in a subsequent chapter.] SEMICONVERGENT EXPANSIONS. {  and Ji x" x^ _ = ^~2^n(i)l273'^2.3. We n is will now briefly consider the case. ' 4 . 1 . we have *' ~ 2^ n (. (^).4. when the half of an odd integer. The case when a. 96 to twentytwo places. 3 2 — cos X. and that value of \/x is taken which passes continuously into a real positive value when x is real and positive.IV.5"*'*j iff^ \ — = A //2^ y TT sin .
tance in certain physical applications.+^ where k is any positive or negative The functions thus obtained are of imporinteger. we may calculate the expression for Jj.42 SEMICONVERGENT EXPANSIONS. so that the following short table may be useful. 2n . [IV.
Theory of Linear (1889).] SEMICONVERGENT EXPANSIONS. ceptional character of the solutions. 48 ^^brresponding to I 18 d^U ax . differ by an integer and it is this circumstance that gives rise to the exso that . 5. I. = .. (Cf. 1 du X ax r^ — \ 71^ x^. Differential Equations. vol. chap. when 2n is an odd integer we have the case when the roots of the indicial equation.) . Craig. although not themselves integral.
or zero. ^ . . this being by far the most for physical applications. THE ZEROES OF THE BESSEL FUNCTIONS.. It has been shown (see 46 above) that f'^ Jo (x) 1 = TTJO cos (x sin <^) d(f> == Suppose that x rti cos (xt) . that is to say. or minimum {x) is It has already been observed that the function Jn and continuous for all finite values of we will now proceed to investigate the zeroes of the function. we reproduce here Bessel's original proof of has an infinite number of the theorem that the equation J^ (x) = real roots*. For the purpose of realising the general behaviour of a transcendental function it is important to discover. and a positive proper fraction cos j. the values of the independent variable which cause the function to vanish or become infinite or to assume a value. if possible. 14. but of its directness and simplicity.(^)=f IT J then H (2m  m) irt ^ 2 Vl^^ ^y _2 2m+m' n ^y cos 2 ^/{(2m + myv"} Berlin Abhandlungen (1824). important case On account not only of its historical interest. It will be supposed in the first place of X which make Jn {x) = that n is a positive integer. finite maximum a. Art. tt where m is a dt positive integer. . the values 0.CHAPTER V.
on writing 2m + TTiy Jo (. . and must therefore vanish for some value of x in the interval. . 1^ ^_ + 7^=. ^^ 45 Now .... u^. + mj . yfj. —h— 3.. 5. <Um.^3 + . It has been shown in the is last chapter that the asymptotic value of Jo (x) \/i*=°K^3' . . = 2 C^' ^ j_^ sm . is 2 _ (2m + w)2 \//x2 / evidently positive. ^r. + ^02 integral TTU Sin ^ 2 ^ ^J^^. . TTV dv COS J a 2 ^/{{2m\myv'] _ ~ f^^ ] ak (hir "^ iru\ du """"H^ "^/ V'{(2m l.. + i)"^ Um.. is positive or negative according as m is even and consequently as x increases from kir to (A? 4.(^ + uy] fi ' and hence.=^ (ly _ (2m V/^2 + . we \.^(h^uy ^/^^(huy) and is therefore positive moreover TTU cos TT I..(ir^ The +1 I. Therefore Jo (a?) or odd. .i 2 ^===== du —  = \ du. TTU  1 j^^^^^ { — .v.] THE ZEROES OF THE BESSEL FUNCTIONS. are a series of positive quantities and Ui<U2<Us. by taking a sively equal to 1.U2. =h+ — (2771 1). where Va. Jo (x) Therefore Jo (os) has been reduced to the form = i/i + ^t2 .) ^ . _= l+uY) du IV" ' Jo cos TT 2 \//A2_(2m 7=^^. This proves that the equation Jo (x) = has an infinite number of real positive roots the negative roots are equal and opposite to the positive roots. 1 .1) tt where k is any positive integer (or zero) Jo (x) changes sign.(h + uy ^ du h I sm . b and putting h succesfor obtain.
. that the equation Jn (x) = roots. We between any two consecutive roots of Rn = there must be at least one root of Rn+i = and hence. + )7r=306305. it existence of the real roots of Jn(x) having thus been remains for us to devise a method of calculating them. where k. we infer from the asymptotic value of Jj (x) that the large roots are approximately given by x = (k\l)'Tr. The proved.. Let f and R^ have the meanings just given to them. as before. although of little or no practical value. or a minimum. SO that the large roots of /„ (^) are approximately given x=(k + ^)7r.. begin by explaining a process which. be any positive integer. as before. give an example of the degree of the approximation. To k=9 . where A. by [v.. is very interesting theoretically. suppose then is (A. is let 71 Now any large integer. and put will We .w 4 then it ^' may be verified from i6 or lo that P ^''+'~ so that Rn+i vanishes infer. 306346.. the true value of the corresponding root being It has been proved that ^^^^'^" dx ' now between every two successive roots and therefore between function J^ of must be at least one roots there Jo(oo) every pair of adjacent = has also an infinite Thus the equation Jj (x) root of Ji (x). W of Jo(x) the derived must vanish at least once.. that _ dRn ' d^ when R^ is a maximum . a large positive or negative integer. number of real roots .46 THE ZEROES OF THE BESSEL FUNCTIONS. and put (D'v. by induchas an infinite number of real tion.
. if /3 is a root of must also be very small for approximations from /§ (f) ^g (^) is = 0.32.. in fact.. The points belonging to a particular cluster are. of course. each in the neighbourhood of a root of Rn = 0. and therefore. 144678.22. 2 142999. fs{&) a sufiiciently large value of s.. so that X= 00^ to three places of decimals. ^ suppose n = . 1 2. 542.. cluster " " condensing = As an illustration.42 and the equations fs = are fl = p4f+4 = f9f2 + 36f36 = f^16fhl44f257Gf + 576 = 0.22. equationys(f) and even for moderate values of s. = 24048... derived from different equations y*g 0... For large values of s. and (f) is the Then when s is very large (f) is very small (f>s (j)s Rn... 144678.. we may expect to obtain for all finite values of f.\'. there will be real roots of the algebraical = which are approximations to the roots of i?„=0. If we put = 144678. and so on. in order.32 12. then '"' ^0 ^"^12. therefore. is 47 where /s(f) the rest of the series. sum of the first s terms of Rn.. The real roots of these are.22 12. and the least positive root of Jo (^) a?i = is. and we already see the beginning of a condensation in the succession 142999. it will fall into found that be they ultimately groups or clusters.] THE ZEROES OF THE BESSEL FUNCTIONS. If the real roots of the equations /« = are plotted off. but we must be careful to see that small as well as/g (/3) = 0. ^ 4 we find ^= 2405.
Sl{8xf 5! (8^)5 P = if cos if=Vi^ + Q2 it is yjr Q=Msmy}r. Trans./r. the results. + ir. of course. which depends upon the semiconvergent expression for To fix the ideas. suppose n = 0. for the moment. Write. 182. = tan~^ that Now may not difficult to see when x is so large that we for use a few terms of P and Q to find an approximate value Q^ Jo (^). we are justified in calculating . p. Phil.+ 16^2^ 512^ ^fr= '^ tan~M. . ^ . [V. * Camb.8^ 33 3417 ' 98 512^ 16384^ )) The value of Jq (x) is (approximately) 99 which vanishes when k being any integer. VP^ + and tan~^ ^ as if P and Q were convergent series same kind of meaning as the derived. Then we have approximately where wnere i _ J^g^ p^^ j ^ \o 1^ 3^ 5^ 7_ 4!(8^y 21(80))'^ hJ~0^ Sx Put then o. will have the series P and Q from which they are We thus obtain the semiconvergent expressions i/=i=4. (1856). ix.48 THE ZEROES OF THE BESSEL FUNCTIONS. The best practical method of calculating the roots is that of Stokes*.
31 ^^ oo4 » .. / 1 33 H 3417 \ I V8^ </> 512^1^^16384^ large.] THE ZEROES OF THE BESSEL FUNCTIONS. a = ^tt (2?i . for the roots of Jj (w) loo The corresponding formula = is x_ The same method orders. . 6 =— (A. m= 4^i'^. .(/^  J + 2^2 (4^ _ 1) .1 + 4A.th root of Jn (x) = is \ m1 4(ml)(7m31) 8a 2 (Say 32 (m . ~3T5i2V^3and therefore Substituting for . 49 then we have to solve the transcendental equation a.. tan~^ .). 050661 x_ /c'Zo+ ^^_^ ___^4_^___. '") on the supposition that and x are both Assume oo = cf> + ^^ ~+^^+ „. reducing to decimals. ^ 33 / 1 \ *+8U"^^^"V"512i^3j+.3779) '^•••' lOO' 15(8ay where G. a b c =^ 1 . then. 245270 + iy"(4FTiy'^' '^ applicable to Bessel functions of higher The general formula '''' for the A. we have finally  .1) (83m'^  982m 4. M. etc. with the help of Gregory's series.. . + ••• a 1 =^ o . =^+ .. 151982 7r"'^"^''^^"^4F:rr is + 015399 (4A. <j> its value — J) tt.6^4(4^ _ 1)3 + '053041 2 62051 •••> or. A « .v.
the error has somehow arisen in the reduction of ll79/57r^. for the present to state that if integral. functions J© (^) and J^ {x) extending over a sufficient interval to show how they behave when x It is seems probable that between every pair of successive real Jn (x) there is exactly one real root of Jn+i (^). The reader will find at the end of the book a graph of the comparatively large. The values of the roots may also be obtained by interpolation is from a table of the functions. It does not appear that this has been strictly proved. which is the exact value. and was kindly communicated to the authors by Lord Rayleigh. . Griffith. M'^Mahon. provided the tabular difference sufficiently small. [v. there must in any case be an odd number of roots in the interval. It has been worked out independently by Mr W.50 THE ZEROES OF THE BESSEL FUNCTIONS. roots of With regard sufficient to values of n which are not n is . St B. the incorrect value '245835 for the numerator of Stokes gives the last term on the righthand side of loi'. to a decimal. so that there is no It may be remarked that reasonable doubt of its correctness. This formula is due to Prof. all the roots of the equation are real. it will be real the equation J^ (x) = has an infinite number of real roots and if ti >— 1.
and u <j) independent of and z. adopting. November 1872). FOURIERBESSEL EXPANSIONS. — 0. Then will satisfy Laplace's equation if dhi .> («r) = AJn(/cr) is according as n is not or an integer. If V"(f) we use cylindrical coordinates satisfied r. method of Lord Rayleigh {Phil.CHAPTER VI. in fact. already appeared in the introduction (p. Laplace's equation potential function <^. Consider the two particular solutions ^= ylr ef« cos nOJn (/cr) 105 = e~^^ cos n6Jn (Xr) } 42 . 3). z. 1 du whence or u u = AJn (/cr) + £/_. which must be by a 6. One of the most natural ways in which the Bessel functions present themselves is in connexion with the theory of the potential. This has. the consider this part of the theory in some detail. Mag. is I02 is a real positive quantity. be comes Assume where /c (/> = loe""^ cos n9. we will now in the main.
r = a. we obtain X JO I Jn (fcr) Jn (Xr) rdr] K K J„ (ko) J^ (Xa) X +^ \ Jo I Jn (Kr) Jn (Xr) rdr K + X J^ (ko) Jn (Xa) . is and the part of the integral on the lefthand side of io6 which derived from the flat circular end bounded by 2^ = 0. [VI. ^=X J^ (Kr) cos 710J^ (Xr). and nothing is contributed to For the curved surface r = a. Working out the other side of 106 in a similar way. When z = + the integral. as usual. ^ i/r and t— denote the space rates of variation and along the outward normal to the element dS. = and V^^]r = 0. and that k and X are integrate over the surface bounded by is and the planes ^ = 0. 7i First.52 then since V^if) FOURIEEBESSEL EXPANSIONS. an integer. is so that the corresponding part of the integral r2n •277 r'" /•" XJ^ (ko) J^ (Xa) adedze^''+^^^ cos'nO Jo Jo 0^0 irXa K\X Jn{ica)J'n{Xa). <j>^ vanishes. (j> = cos 710Jn (kv). ^ = + oo . suppose that positive. oc . When z = 0. and one valued of <l> . (j) = e~'"' cos nOJn (fca)y ^ = \e~^^ cos nOJ!^ (Xa). is r2n ra \\ Jo Jo \ rdOdr cos^ ndJn (kv) = ttX ra I /„ Jo (/cr) Jn (Xr) rdr. Then we may the cylinder r = a. continuous. Green's theorem gives the surface integration being taken over the boundary of a closed space throughout which </> and yjr are finite.
finally (k — X) I Jo Jn (kv) Jn (\r) r dr = —TT (^^« ('^*) *^« (^^) ~ '^^^ (f^^) *^n (\a)}. All these cases occur in physical applications. and then = 27r. 108' The condition 108' if K. make h decrease indefinitely. of J^ {ax) = 0. {Xa )]. = 1 08 provided that XJn (ko) J'n (Xa) — fcJ^ (ko) Jn (Xo) 0. . (i) (ii) (iii) X are different roots of Jn (aos) they are different roots = 0. we J Vj (\r) rdr =  (\o) {j^^ + (l  j^) Jli^a)] 109 When n is not an integer. and find. The very important conclusion different. provided that in addition to the cylindrical surface already considered we construct a diaphragm extending from the axis to the circumference in the plane ^ = 0.VI. 53 or.] FOURIERBESSEL EXPANSIONS.Jn {Xa) Ji {Xa) . if if is satisfied. we may still apply Green's theorem. among other ways. 1 107 follows that if k and X are Jn{Kr)Jn(\r)rdr = 0. first for ^ = 0.XaJn {Xa) /. with the help of Jo rj'. that then we divide both sides by h. 0. In the formula 107 put k =\ + h. Reducing obtain finally with the help of the differential equation. they are different roots of AxJn {ax) + BJn {ax) = where A and B are independent of x. for and consider 6 this as a double boundary. Taylor's theorem.{Xr)rdr = ^ [\a this [J"^ {\a)f .
symmetrical in and X. dV or where k is the conductivity of the material of the h is what Fourier calls the " external conductivity. To show the function application of these results. IK Suppose. medium of temperature zero. Jo' (X) + A Jo (X) = 0.64 FOURIERBESSEL EXPANSIONS. 2 = z = + CO and suppose that its convex surface is surrounded by . = 6'"^ COS 2n7rJn (kv). if XA. therefore the additional part contributed to the lefthand side o io6 is — n sin 2w7r cos Inir Now this is I I Jo Jo ^ tc J^ (kv) J^ (Xr) drdjz. and consequently the formula 107 remains true for all real values of n. then. In th( same way the formulae 108 and 109 are true for all values of n. that the base of the cylinder is permanently heated s( . and when 6 = (J) 27r. Then when the flow of heat ha{ in the cylindei become steady. therefore the same ex pression will occur on the righthand side of io6. and suppose moreover. [V] When6> = 0." If cylinder. and the second will also be satisfied. and we put V= <f>. ~ = e~^sm ^nirJn (Xr) . when r = l. that X is any root of this equation. the temperature must satisfy the equation V at any point and moreover. we will employ th( to obtain the solution of a problem in the conduction of heat i Consider the solid cylinder bounded by the surfaces r = 1. the first condition is satisfied .
and no J a Jo f Vo ( V)/(r) rdr = aJ' J^ (X. because this satisfies all the conditions of the problem. 5. lat 55 Then the temperature the temperature at a distance r from the centre is Jo(\r). to the form is expressed by 1 12.FOURIERBESSEL EXPANSIONS. In particular we may suppose put the distribution symmetrical about the centre. Assuming that this is so. roots \i. Xg.r) rdr and therefore . at any point within the cylinder is F=e^2Jo(Xr)./o(V) that follows by 107. we put 113 /(r) it = 5:^. 109. for the range considered. 1 112 Now there does not appear to be any physical objection to supposing an arbitrary distribution of temperature over the base of the cylinder. finite and to r = 1. The question is whether this continuous from r = function can be reduced. The equation no has an infinite number of real so that we can construct a more general function </> =1^.6^*2 Jo (>^s^) 1 III id this will represent len subject to the same the temperature of the same cylinder conditions. we can at once obtain the coefficients for if A 8 in the form of definite integrals. provided the temperature varies continuously from point to point and is everywhere finite. except that the temperature any point of the base is now given by <f)o = XAsJo(\sr). and </>o=/(r) where f(r) any function of r which is onevalued.
as before. obtained 116 the coefficients Ag being given by the formula 114. Whenever the transformation <l> 1 1 3 is legitimate. nddd = 27r^„ s sJ'n (Xs^) and ri r27r JoJo /(r. s cos n0 + J5„. is coefficients. when ^ = is ^0 = S (^ cj) cos n6 + B sin nd) Jn (X?). with a more complete f(r. the function = 2Ase^'Jo(Xsr) 115 gives the temperature at any point of the cylinder. which vanishes when ^ = + 00 and also when r = 1. when its convex surface. f{r.valued circle r = 1. one. it Thus if. we have a potential which remains unaltered when 6 is changed into 6 + 27r. Whenever this reduction is possible. where the summation If refers to we restrict the quantities X the quantities n to integral values and take foi the positive roots of Jn (X) = 0. . can be reduced to the form member of 117. 1 1 . sin 716) Jn (Xgr) we find successively •277 f{r. The function may be point in a solid cylinder when interpreted as the temperature at anj the flow of heat is steady. 6) cos 7i6Jn (Xgr) rdddr — ttA^^ Jn rl ^ \ ' ^0 Jl (V^) rdr g (Xs) Aji^ . 6) cos /. and continuous over the of the righthand subject only to the conditions of being finite. we have 1 1 ^ = SS {A^. [VI.56 FOURIERBESSEL EXPANSIONS. 6). 6) easy to obtain the notation. the convex surface maintained at a constant temperature zero. is surrounded by a medium of zero temperature. A much more (l) general form of potential function cos is by putting = X(A ne + B sin n6) e'^ Jn(\r) n and \ independently. and the base of the cylinder heated according to the law expressed We are led to inquire whether an arbitrary function by 117. and the circular base ing to the law <l>o=fir) is permanently heated accord = tAsJo(\ry. The value .
] FOURIERBESSEL EXPANSIONS. some of these analogous alytical expansions are given in the Examples. Similar considerations apply to the expansion 118. 471). consider the expression on the righthand side of 113.. Other physical problems We do not propose to discuss the validity of the expansions obtained in this chapter. Id may be constructed which suggest to 113 and 118. pp. and others same kind. 89 (1880). and require a disproportionate discussions to amount of space. applied are connected with the amount analytical investigation usually . of restriction order that in it which must be applied to an arbitrary function in may admit of expansion in the required way. . Crelle. ii. 2nd edition. and applications the physical these restrictions are generally Such a work as Fourier's TJi^orie Analytique de la Chaleur is sufficient to show that the instinct of a competent physicist preserves him from mistakes in analysis. SO that 57 ^"' in the ' " ^'HXs) Jo L ^^^' ^^ ^^^ ^^^"^ ^^'^^ rdedr\ I same way s j ig' Bn. 362). and to a paper by the same author. Ann. then if the infinite series XAsJ. 19 —39. = ^j''2(x) J J ^(^^ ^^ ^^^ ^^"^" ^^«^) rdOdr. Weber (ibid. And after all. Vol. 146). iv. entitled Einige Anwendungen der Residuen rechnung von Gauchy. For further information the reader is referred to Heine's Kugelfunctionen. the value of these the practical physicist. even when he employs functions of the most complicated and peculiar description. See also the papers by du BoisReymond (Math. 109. to do so in a satisfactory way would involve a great many delicate considerations. and Hankel (ibid. 210. satisfied from the nature of the case.{\sr) is uniformly convergent. There are a few elementary arguments. in the present stage of is not very great for the difficulties of the mathematics. which. where the coefficients As are determined by 1 14 . its value is/(r). viii.* As an example of the conclusions to which Dr Heine is led. p. t. of the * vi.
It will be understood that this does not in any way amount to a proof of the proposition: but it shows how. . etc.. ^ may be reduced to the form Vg is </> = + (pi COS ^ + sin &) + (pa cos 2^ + og sin 2^) + = S (/)s cos sO + og sin s&). . the expansion may be regarded as a straightforward algebraical transformation verifiable a posteriori. Now if Xi. By expressing v^ in terms of sines and cosines of multiples of and rearranging the terms. the series for if and we put = A[^^Js {\r) + A[^^Js (\. . = 00 Sv^r*. we may and in legitimately write ps = lA\t!Js(\mr). and admits of an finite. it may be possible to determine the constants A^i\ A'^^\ etc. this is transformed into <^ = Vo + ^1^ + ^^2^2 + . [vi. for instance. generally speaking. which is equivalent to 118. =Z CO Us . r^. X2. this comparison indefinitely leads to a convergent series A^{Us{\r) + AfJs{\. /3o <7i . so as to make the coefficients of the same power of r on both sides agi^ee If the result of carrying out to any extent that may be desired.r) + Ps . absolutely convergent expansion = (l> Uq + Ui { U2 + . where in Ug is a homogeneous Cartesian coordinates rational integral function of degree s Then by putting x — r cos 0. and continuous all over the circle r = 1. .58 FOURIERBESSEL EXPANSIONS. (X. where Q. are the roots of Jg(X) Js (^i^)j Js (^2^)> etc. in a particular case of the kind considered. an integral homogeneous function of cos Q and sin 6. . where p^ and og are. that ^ is a function which is onevalued. .r)^. . Suppose. although not amounting to a demonstration.r) + sin 5(91^!?/. infinite powerseries in r. w.r)] . . each begin with a term in = 0. . arrive at a valid formula like manner we may and then <t> = ^ [cos s(9X^S/« (X. each beginning with a term in r^. may help to explain the possibility of these expansions in some simple cases.. y—T sin ^. y.
unless the expressly stated. properties of the Bessel functions may be proved or by means of the theory of a complex variable.zY~^ dz = 2" VttII {n .CHAPTER VIL COMPLEX THEORY. Now I cos {x cos <f>) sin^'* c^cZt/) =  cos {xt) (1 . plained. is unnecessary complication. . in Forsyth's Theory of Functions. that the real part of x is also Then. We may therefore write x^ j e""'^ (1 .ty'^ dt r+i on the supposition that the integral is taken along the axis of real quantities from — 1 to + 1. it will be supposed that n\\ is real and positive. and. as exfew for instance.J) Jn {x\ 1 1 9 and this will remain true for any finite path of integration from ~ 1 to + 1 which is reconcilable with the simple straight path. as already proved (p. 88). To avoid throughout contrary positive. of the most obvious of these applications will be given in the illustrated Many A present chapter.
=a paths of integration. Each solution must. the determination of the constants is is not always easy. Then if t 0. and in fact this the principal difficulty that has to be overcome. minate until we of the path. of course. jB are independent of It will be found that d^u 1 du (3 =^'£ a. + 001.z'f^ dz j^ii^'H^^r^^dz. Under the restrictions imposed upon x and n. [VII.60 COMPLEX THEORY. 120 where a. It must be carefully borne in mind that. in calculating the value of u for any particular path. by a straight line from . 121 as the path of integration a straight line from followed to oci. the function (1 — z^)'^~i must The value of u is not deterbe taken to vary continuously. be expressible in the form AJn + BJ_n or AJn + BYn'. therefore u will be a solution of Bessel's equation €f^^^ if (1 . where k is any finite real quantity. the admissible values of a and /3 are + 1. We are thus naturally led to consider the function J a {lz^idz w. the is general value for the same (or an equivalent) path where k is any real integer. By assigning these special values to a and ^ and choosing different vanishes when 2. Consider the function u^^x^l we may choose 1 to '^'^''(lz'^y^dz.^2)«+J and also when z = ^.z^) 4 (2n + 1) iz] ^'' (1 . If fix U is the value of (1 — z^y^~^ at some one point one value of u for a particular path. (1 . we obtain a large number of solutions in the form of definite integrals. —1 and A.
if 61 denotes a real variable. = X'' I [ V«'' (1 . the general value being ^. and one value of U. and which reduces to + 1 when z we take . that value of (1 — z^y~^ = 0. is any real integer.] COMPLEX THEORY. cos^ (j)d(j) I22 then x^ 1 Jo cos (x sin (p) = ^G^Jn > and if we put 2 sin {x sin cos^* " x" Jo I <f>) 6 d6 e^sinh<^ x'' I cosh^" (f) dcj) = iO„T«.1")"^ dt+i[ e^« (1 + t'y^ dt\ = — x^\ JO COS (a^sin 0) IT cos^**</)c?</>  ix"" \ I sin (x sin (j)) cos** (fxicj) e"* j «"^^ "^ cosh^" </> c?</) .=iCn(Jn + iTn). let us write Cn n = 2y7rU(ni). if °° In the same way.z'^y^ dz. u^ = x^\ V^* (1 . 1 26 taken along a straight line from — 1 to line from to oci. [ For convenience.VII.J I24 = e(2«1)fc^ if^ 125 where A. one value of u^ is followed by a straight and the general value is . 11^ 123 is Tn is also a solution of Bessel's equation. tt. 0.
of z. One The value of the result is Jo.1) Jo  e'"'^ ^2«26i fjn = ()« 2ie^ sin 2n.(ns\i)U{2n2sl) Us = and that 1 2^ Tl{2nl)U(ns'l ) U(nl)n(s) (— )' TT COSCC ' n (tI —S— 1) n (5 — W) = TT COSeC (71 — 5) TT = tlTT.^ 5^?^^> . or modulus. . 128 taken along a path inclosing the points — 1 and + 1.. Then. by the binomial theorem. u.r Therefore one value of u. and then from ei to + ooi. then round the origin. Observing that (ni)(n^). we may put and therefore one value of u^ is given by 8= 5 J +oDi integral on the righthand side may be evaluated by the path along the axis of imaginary quantities from + ooi taking to ei.sin 2n7r ^(^i)0^f)(»g + ) ^ ^^n 251) ^"+^. and such that throughout the integration ^ > 1.62 COMPLEX THEORY. [VII. Jo which = (y e"'^^ (e'"'''' . Now consider the integral u^ = x^\f + ooi e"^^ (1  2'2)«i dz. where \z\ means the absolute value.. is given by = 2e^^.
^2)«i dt = 2Cn cos TITT ( J„ COS 71 TT — T.VII. «j.1 cos riTT J_w (a?). we argument of it (1 —z'^y^i is must take along FG. and by Jn {^^ putting A.i siu 7l7r).(^^^) + e^^'^ U^ + x^'j e**^ (1 . S"   >1 throughout the integration but the value of u^ w^ill not be altered if we suppose the integration taken along the path represented in the fiofure below.] COMPLEX THEORY.(2n '^^ GH. and the general value 129 It has been convenient to suppose that . 63 ue have u^ = 47r cos nir e^"*"* ^.v S On+i n{n\) on reduction. ABCDEFGH oo« e If the 3taken to be zero along DE.f . Hence we C/2 along AB. BC. 1 3O . = 2 = :P7^ n(«i) 7^ = 2(7. and + (2n have one value of u^ given by 1) tt '^*  1) tt = e. Us to be . The general value is we have the special value ^^3 1^3 = e(2^*~^)*'^*'it3.
()'' COS eirJne = ()'' sin eTrT^e e • Divide both sides by (— )^ e and make vanish . if we like. COMPLEX THEORY.) = ilogn(. Then 123 : still remains valid so long as the integrals are finite. To find its real value.) = then the limit of /(e)/€ when e is g:g. write being a small positive quantity. thus The form expression on the righthand side may be written in the n+e n1 /^x (_)n+s \2j r n(5) fxy (xV^ ^ . however. this with 20n cos nir J_.64 Since n is. [Vll.j. 131 becomes an identity.y ^ €U(n + 6 + s)\2j "^V27 t^ 1 f(e) € f^Y [2) ' where ' *^^'^"'n(n + 5)n(sf €)V2j ~U(s)Il(n + s€)\2) Let us use the notation ^(. 132 zero is . regard 131 as a definition of T^ for all values of x. e J{ne) . that Jn — Jn COS nir — Tn sin mr. and n being supposed an integer. but the value n not really indeterminate. 1 3 1 The formula 123. by which T^ was defined. is real. is intelligible only so long as the real part of x is positive we may. must agree with the real value Hence we infer that of u^. When of n is an integer. then T^ is —e for w.
S)^ Comparing 11 this with the formula 30 (p. ) a very large positive quantity. 5 G.e — s)7r sin 67r ' Lt en ( e=o ?z = TT/ + 6 + s) ' is a. pp. approximate value of Yn given on p. . * volume of the Mathematische Annalen we Hankel writes Yn for a Or. the last integral on the righthand side may be neglected and then. memoir in the first have been following in this chapter. by applying d? When is . 71 65 Again n (— + e + s) n therefore (?i — e — s .] COMPLEX THEORY. 40. 14) we see that. when is a positive integer. I—«— 1 1 . 69 • —71. Jo sin (^ sin <f)) cos^* 6 d<b 135 ttI Jo e'^^'^'^'f' cosh^cj) d<l>\. 40 above. whose positive.}. then. Stokes's to method* we obtain the There would be a certain advantage in taking J^ and T^ as the fundamental solutions of Bessel's equation in all cases when n is This course is practically adopted by Hankel. when n positive integer. . explained later on. 2 + i/r (0)) J. M. (p. F„ = (log27)J„ + T„ + TT r^ I . 7rr^ = 2{Yn. preferably. that of Lipschitz. 12).VII. with the notation of and on substituting T„ its value as a definite integral.1) = TT cosec {n.(log p. . U{nsl) Finally. the other two integrals. for 134 or.
II. will Hitherto the real part of x has been supposed positive we now suppose that « is a pure imaginary.Un + iV. it becomes 1 d^u du (^ n\ proceeding as in Chap. 2[Yn  7) /.l . function we may denote. expression on the lefthand side of that equation by which T„ was originally defined. Since. it is easily found that. = when n (log 2 is a positive integer. If in the ti. = ttT. however. we shall continue to use F^ is now being generally in the sense previously de It may be remarked F. sm .). by F^ . and there will be another solution obtained by taking the real part of But it is found that both these functions (and in like i~^ Yniit). this is defined for all values of n by the formula J n— vre. and it is im certain t applications to discover a solution which when is infinite. fined. = ti. while the other becomes infinite.6() COMPLEX THEORY. we put x = it becomes iH^l. When n is an integer we have the solution In as before. By I_n (0' defined by 00 m+28 ^^2"+^Ii{s)TV{n\s) 139 /_ {t) = i. the equation has two independent solutions I^ (t). for the moment. notation adopted..J_n (it) = one of which vanishes when ^ Z2n4. manner I^ and J_^) become portant vanishes for infinite when t = oc . 137 . Neumann s that. [VII. n (5)"n"("¥T^)J = 0. We shall effect this by returning to the equation 123. when n is not an integer. ^r ZiiTT 136 Tre' COSWTT (see 131 above). If in Bessel's equa tion we put a.
and which are the most convenient solutions to take as the fundamental ones. when n is an integer. it suggests that we should try the function {t sinh <^) d(i> Jo ^osh2"</) n. to Then In and Kn are solutions of 138 which are available for all positive values of n. J rt. Vn The but if = Jo I sinh (t sin <^) cos^** </> c?</> + I sin Jo (t sinh 6)coshr^ 6 d(l>. verified that it satisfies the equation 138.] COMPLEX THEORY. 141 ^n+i ^ T ^^ T T Jn—\ T ^n~ —f\ —" 5—2 . f Un=t''\ is cos Jo {t d^ this does not give us what positive which shall vanish a solution of Bessel's transformed equation 138. whether or not n is an integer. 67 where Un =  cos n Jo {t sinh (f>) cosh^ 6 dS. function Vn is infinite 2n + 1 is negative Un is finite.x 2^V7r^^ f* cos (i sinh (^)cZ<f) reducing. It will be found convenient to write rr /. which tion Un by changing n into — = oo and it is easily obviously vanishes when t is obtained from t^ This func .VII.—^ . Now although we want. it may be proved that J„ — t 'n. so that it is the solution we require. namely a solution when n is when cos t is infinite. As in Chapter II. whether n is positive or negative and it may be verified that sinh 0) cosh^*^ .
.9) (4n^ . since the formula for /„ works out verj the formula fairly. iii. (4^^^ . This may be used for approximate calculation wher t is large.7"1~ . cos nir.  f .1) (4^^ . p. so that the semiconvergent expression Kn TT is R r Ji .9) (4n^ . 14) leads us to conclude that Rn{t) is th( Yn {it). leads to the conclusioi foi B = VJtt . A somewhat trouble some and not very d' Analyse. satisfactory process. IT.9) \ There does not appear to be much reason to doubt the correctness of these results. p. Kn (t) must agree. as they should be. When u = Kn.25) 8! (80' ] J which terminates. and are consistent... and gives an exact solution.(0. The function here called Kn (t) is 2" times Mr. vol.68 COMPLEX THEORY.25 ) Sl(SJf ] +. It is in order to preserve this analogy that it has been thought desirable to modify the notation proposed in Basset's Hydrodynamics.J ^^ ^ f + 8^ ^ "^ 2! (80^ (4yi^ . with In+iKn . 14^ if The formula 30 real part of i~^ (p. 19. [VII and that the functions Kn are connected by relations of precisel}/ the same form. Laurent's Tiaite that it is evident that ^ = 0. even for comparatively small values of t. By applying Stokes's method to the we obtain the semiconvergent expansion differential equation 13? i^n' .1) (4n^ . .InKn+i = j COS UTT. suggested by a formula ii t.1) (in^ . 255. up to a numerical factor with iJ„(0(log27)/. 1) . (4?i^l) (4?i2l)(4?i29) ) The corresponding expression for In (^^^ is . Basset's K^. when n is half ai odd integer.
integral : The first is GnX~^^Jn.] COMPLEX THEORY. treated. Camb. relation 0= ^^'(l. tained in Chap. Trans. along the contour of the rectangle whose vertices are at the points corresponding to the quantities —1.t^yi dt{ij ^(^+it)i {1  (1 + ity}''^ dt i where t is. taken. throughout. and when h=oo. and indicates how the general case may be Let us suppose that n\^ and the real part of x are both Consider the integral positive. in the positive direction. where h is Then the it total value of the integral is zero. and by expressing as the sum we obtain the of four parts arising from the sides of the rectangle. +1.\[[. this is IX. above. (Crelle lvi. y 69 vol. confirmed by Stokes's investigation. the third integral vanishes hence CnX''Jn {cc) = ie'"" Jo (2it \ tj'^ e""' dt ie^'^j (2it\t^y'^e''^dt Jo IT I45 The argument vanishes : of 2it \. The method is that of Lipschitz who works out in detail the case when w = 0. p. would be easy to multiply these applications of the complex theory to any extent we will conclude this chapter by giving an alternative proof of the semiconvergent expansion of /„ obIt . a real variable. p. 1+hi.1^ must be taken to be ^ when t hence . real and positive. —1+hi.. Phil. [38]. lY. rh h e^(^+it)i f 1 Jo _ (_ 1+ ity]nh fit. 189).
and the argument of to [VII where ^"~^ is real. Substitute these expressions in preceding equation. .^ =f . when ^ = Applying a similar transformation (2ri (— %t + and putting + l)7r we have I dt 2i) Now put a..(n«+)/f (s1)! s! V2ia.(« s + i)/.. and make use of the formula in th( f"e^f^+«c^f then. we obtain (to s terms) 146 + R. . = 2'V7rn(7ii).o^^)(ftf)r .. 1 +— . we may write f Y'i_.g V2W and y/ V 0f 2m:.(»^)f .. («i)(«). . vanishes i^)""^.70 COMPLEX THEORY. (1— r^l nmy be ex panded a similar way. then By (i Maclaurin's theorem. (H^)(«f).. if = n(ni + s) we further observe that c. »i— S where ^ is some proper fraction.
the values are still : further increased by replacing .] COMPLEX THEORY.VII. '^'^ UsU(ni) W ' Jo ^ ^ UsU (ni)* (2a)« The expression on the right hand is precisely the absolute value of the coefficient of cos yjr or sin ^fr. 71 where 2n (n i)R = ("i)(»t). in the (5 + l)th term of the series so that the semiconvergence of this expansion. „od(ia) each by unity. — Jn (00) are fully established. and its approximate representation of a/ / TTOC .g. if we put \x\=a.od(i. when s >7i — ^. .("^+i) Each of the value if integrals on the right hand is increased in absolute the quantities under the integral sign are replaced by their moduli moreover. A fortiori. as the case may be.
and formula ""^ we may change make use of the Qmx gQg nxdx= I. the formula 42 (p. 38) ["" we have d(f>. DEFINITE INTEGRALS INVOLVING BESSEL FUNCTIONS. By 18) or 90 (p. Jq (hx) 1 = cos (bx cos S) let TT. shall give a selection of these integrals. Now if a = di + 6 = 61 + bji.'o Suppose that then I h is real. In the present chapter we Sonine. m w? + „ n^ thus / e"^^ Jo Jo {hx) dx = i . and a be a real positive quantity : e'"'^ Jo Jo (bx) dx =TT Jo I dxl Jo e~^^ cos {bx cos </>) d<f>.CHAPTER VIII. arranged in as natural an order as the circumstances seem to admit.. Hankel. Many definite integrals involving Bessel functions have been evaluated by different mathematicians. the order of integration on the right hand.ad<f> TTJo a^ + This result will still be true is for complex values of a and b pro vided that the integral convergent. Under the conditions imposed upon a and b. . in the Others will be found examples at the end of the book. more especially by Weber. and Gegenbauer. ttsi.
1 \/a? — b\ We thus obtain Weber's results Joibx) cos ax dx 1 = . expression e~^^ Jo {hx). if 6^ < a we must put \l}p _ a^ = 4. Again being real let b be real and positive.DEFINITE INTEGRALS. a 1 [ e^'^Joibx) dx 'Jb' Jo If b^ > a^ the positive value of ^/b^ . and put ai instead of and positive thus . under this condition we shall 00 still have ^ e*^^ Jo (bx) dx = Jo that value of Va2 + 62 . = 0. 149 a. and thus obtain i Jo{bx)dx = jy 148 and as a special case 1 Jo Jo {x) dx = l. ai>6. .o? must be taken . hence the integral P (x) is mvergent a trigonometrical function of x if . When b is real and we may put a = 0._ \ I 62 f > a\ 150 ) Jo Jo (bx) sha. 147 Va^ + 6^ being taken which reduces to a when 6 positive. 73 when x cosh ^/x is j very large.axdx=^0 cos axdx = a^ Jo ( > b\ 151 Jo Jo{bx) sin axdx = ^a'bV . behaves like QaiX 62a? P{x).
and on changing the order of integration rtiTT becomes 2n(ml)cos2^2ira"' ^^ cos^c/) [ f J o Jo Jo II (m — rriTT 1) cos "^ _m+l which Pt ^' {lt)dt . this To prove .74 DEFINITE INTEGRALS.) another due to the same author . sin mir. cos (i) d<i>.00 I we observe 2 = that n /*°° ^•^"'1 Jo («*') c?a:5 c^^ p «*"~^ cos Jo this Jo '^Jo (ow. [VII When a^ = hr the integrals become divergent. due to Lipschil is r I af^~^ Jo (ax) dx — n(— i)n(^)ngi) . the reason i that they ultimately behave like Jo V^ fjx is instead of like ' ^0 The formula 147 with which we started (Grelle lvi. 1 5 where m is a positive proper fraction.
2tt^ value given above. to 75 kt is. n(™i)n("i±i)n(i)^ sm mTT. DEFINITE INTEGRALS. The :e result may be written in various other forms : thus if we use of the formula n(»i)n(i+^) )ecomes cos = g ^ ~ '^'^^{2 ) 2ml x'"'^ j Jq (ax) dx = .Jll.
because ^rj and w. 2: cos ^ = ytx.are onevalued. If where A and B are independent of r. and therefore „ I I dfjLd<l>= 47rFo sm mr © = 47rFo m . which is the same thing. r+i +1 r+ r+n (Oo=Vo J ^. =(A sin mr + . [VIII. we have 09 O/JL — IWC0 = /»j2 " dr V r^ drj or. 155 Then we introduce polar coordinates by writing a — x = r sin 6 cos h — y = r sin c (f). the transformed expression for . and coo « is finite when r=0 we sin mr where of (o^ is the value of o) when r = 0. 1 r 54 Now and put consider F as a function of rectangular coordinates a. we have where ^' is H= Jo re*"^"^' c?r 4>'c^yLt c?</). sin 0.76 DEFINITE INTEGRALS. ^ whence co 8^ (ro)) + m¥a) = . J 1 J 7T 4>. Now from the definition (o it is clear that. c moreover . 6. — = r cos 6. let us write J _oo J —00 J —00 poo r+co^ if J(^_^)2+(^.B cos mr). if Fo is the value of V when r = 0. must have B = 0.H(c^)=«} ^^^ ^^ ^^ ^_^.+«._y. dadbdc=n.
mil] definite integrals. = pcos ^./ . after integrating with respect to re^'*" Vo (mr) rfr = ^2 e ^\ 1 ^ 158 More generally. dr = p sin ft (^ cos n0 we obtain gpV / reP ^ J. /•<» then the formula becomes. then since ^ —00 —00 we have ^(a.^ ^^^ oo (oa:)"} ^a dl df. y. finally. the preceding theorem this fl is 77 By = 47r O' — /'* re^"^'^ sin mr dr Hence. z).J— p. ^ 2/). b = r sin ^.. 59 . 157 the equation satisfied by <l> being 3^ + In particular we ^+ 6^ '™^* = *^ may put ^ = Jo (wir). J —00 J 00 6)e^M(«^)^+(^2/)2}c^ad6 = ^e"4^^4)(a?. 00 =— Suppose now that e 42)2 ^ (a.«_^)2+ poo J —00 J — p. 4> by putting = J„ (mr) (A x cos n6 y + B sin n^).i (7nr) I e^^^V^o^ (e^) \ B sin w^) c^^ 1 = —e 4j)8 J"„ (mp) (A cos n^ + B sin n/3). where r^ = a^ + a = r cos ^. 1 56 ^ is independent of c . (6_j„a+ ^^ ^^ g_^2. and suppose that x=y = 0] 6.
then the integral with respect to ^ r+n I is ^pr sin e+n&i^0 COS {2ippr sin 6 = 21 Jo — nO) d6 = 2'iri''In(2p^fyr). . putting \ for Qn. and division of reP^ Jn (mr) /„ (2p^pr) dr = ^^ 6"^+^'^ J^ (mp). after substitution. '^P" and fi m more symmetrically. known that i— = n/ ^+2—9 1 1 1 r* X cc^'^^e^'^da. jB . In this formula put [VIII. evaluate the integral Weber proceeds to Jo which is convergent so long as < It is g <n+ . j\e»'^ J„ (Xr) I„ (^r) dr = ifi.. A = l. After this singularly beautiful analysis. ~ e'^^Vn (^J i" or again. which does not affect the conver By making fi infinitesimal. changing /jl into gence of the integral. B = i ^ = i7r. both sides by 27ri^e^*f^. The formula thus becomes. Jo or.78 DEFINITE INTEGRALS. for 2p'p. In all these formulas the real part of p^ must be positive in order to secure the convergence of the integrals. we obtain the additional result 162 [^^'e^''Jn(^r)dr^^^^^e^.
) contains a large number of Some of these may be /erified .<zni j^^ Q (xr) dr= ^. and make use of 162. we obtain /.— /. f^ is 164 is by putting q = n. Conine's original memoir.A dr{ ^0 x^'^^r'^+i J^(\r) e"'"^ dx./o X^ 2n+i n (n . We :o will now is consider a remarkable formula which is analogous It that which generally know^n as Fourier's Theorem..^ LL (n is 2?V . : de stated as follows — may . xvi.e. ddered general special artifices adapted to the particular cases conbut to do this would convey no idea of the author's method of procedure. We are therefore reluctantly compelled to content ourselves with referring the reader to M. find Ij^^dr^^. lonble integral if (I. so that an intelligible account of jhem would involve the reproduction of a great part of the memoir. 230) that the value of the not affected by changing the order of integrawe do this.^l^_.^q) 2_9 ' X> n(ki) 163 25nixng^(ill) U(n. which is based upon the theory of complex integration. Ann. (Xr) dr= ^^. and /ery elegant ' remarkable formulae. r n i6s sed by Weber's results have been independently confirmed and generalGegenbauer and Sonine. whose elaborate memoir (Math. especially by the latter.7111. we obtain it may be shown /\. p. 79 therefore C"^. iiind .J x^ el ar^^ dx ^^ 2«+in(ni^). by .] DEFINITE INTEGRALS. Now tion .ig) By putting q =n+ l. which permissible so long as n not zero.ve Jn{^r)dr^. md for this we have not space. The formulae obtained are connected by a 7ery close chain of deduction.
Taking cylindrical coordinates p.<» dp drjl e'^^ pcf) Jo Jo (p. — 2rp cos — R"\ p^ rj p<l>{p) cos then _ fP / 2*^ p(f) (p) {cos 710 COS nrj — sin nO sin nr)} dp dr) dr) J Jo rp r2n J (z^TB^^ pfL ^p^ COS nrj COS Q dp Now when z is positive (see p. II. hence the potential on the positive side of the rp J q p! = is Vi = COS n6 f2n .6)}^ [z' + e'e = v. z) is r^+ST JQ J gJe Put r^ \ r' + p'.80 DEFINITE INTEGRALS. 72) <e^ . any point nd'dpde' ' Then the potential of the lamina at 'p (r. 1 1 By Neumann's Jo O^R) formula 27) we have 00 = Jo O^r) Jo (\p) + 2 S Js (^^') Js (V) ^^s ^V. 6. 15 . that each elemenl attracts according to the law of gravitation. method of discussing namics. this integral is given in Basset's Hydrodywe will begin by giving this analysis in s slightly modified form.2rp cos (6' . (p) cos nr) /« O^R) d\. p. and if n is any real positive integer. and that the surface density at any point is expressed by the formula cr = <l>(p) cos nO. 6. although perhaps not altogether rigorous vol. then A very instructive. moreover. z let us suppose that wc a thin material lamina bounded by the have in the plane 2: = circles p — q and p=p\ suppose. [VIII If ^ (r) IS a function which is finite and continuous for ali values of r between the limits p and q (p and q being real and positive). .
if we = g or r = 'p. M. except possibly when r admit the validity of the process by which i68 has it is zero.ii)(l hence. Jq Assuming that the values of differentiating valid dV ^ dV and ~^ can be found by when 2 = 0. the proposition immediately follows.] DEFINITE INTEGRALS. we under the integral sign. we change the F. With regard to the case when r = g' or p it may be observed that just as the equation lis the attraction of a uniform really connected with the fact that is altered (algebraically) by a of surface circular disc density pass from a point close to the centre on one side to a point close to the centre on the other. 81 . = cos nd Jo I d\j J dp q I e~^^ p(f) Jo (p) cos nrj J^ (\R) drj = 27r cos nO p d\ Jo rp I dpe~^^ p<f) J q (p) Jn (Xr) J^ (\p). I in all other cases iThus. been deduced from 167. i68 Now. of a uniform semicircle as we G. if lue assume that V^ remains the same if order of integration. normal to its plane. so we may infer that when r = q or p an amount 47rcr as we \ dz dz J 2=0 because the expression on the left is ultimately the difference in the attraction. if r lies between p and the left hand is 47rcr I q.Viri. the value of the expression on = 47r(/) (r) cos nQ . and that the results are have ( _ ^p) f ^Jj OZ J z=0 \\0Z I = 47r cos nO d\ Jo J q r dp \p 6(p)Jn (\r)Jn (Xp). of its bounding pass from a point close to the middle point 6 • . on By making a similar assumption we is find that the potential the negative side of the plane ^ = Vo = 27r cos nd I d\ I Jo dpe^^ p<f> (p) Jn (Xr) Jn (Xp).
471) and du BoisReymond Consider the integral rh I rp d\ I Jo J r \p Jn O^r) Jn (Xp) dp r. diameter on one side of other*.pJ: (ph) Jn {rh)] by 107. = i<^ (r)\ J ^^^. to the corresponding point on the Thus we are led to the conclusion that j if d\j \p(l> (p) Jn (\p) Jn O^r) dp r=q or p.82 DEFINITE INTEGRALS. p>r>0. from the memoirs of Hankel (Math. give : The assumptions involved in substance. p is If increased by a small is amount dp. it [VIII. where with rt is a positive integer. (Grelle LXix. ["^Jn . derived. in the foregoing analysis are not to we will therefore very easy justify. This is what the analogy of the ordinary Fourier series would lead us to expect. Now suppose that h is very large. especially the first one the of a more outline satisfactory demonstration. . argument applies is to any disc at a point on the circumference where the curvature continuous. the corresponding change in w rh du= I Jo dX {\p Jn (Xr) Jn (Xp) rh I dp] = pdp = Jo \Jn (rX) Jn (pX) dX (ph) J: (rh) F^l. Ann. and (27z let + l)7r 4 . and p. I then we may put Jn (ph) * This = ^— ^ sin (a  ph). 82). viii. h are positive quantities.
Jn (Xp) dxj Xp Jn (Xr) 6—2 .r) h)\ P+r cos(p ] ^ p^dp 7rr^ sin (p .] . dx .r) h /_x^+i pr of + r)/4 + p r ]' given by Hence the value u when h is infinite is h=. I That \ is to say dx it \p Jn (Xr) Jn (Xp) dp = 4. may be thought I paradoxical that the value of the integral dp.^ . therefore ultimately when h is du = ^r^^ • P'r' nrhjrp —^ . pcos(« Lt + ^— ^ r)/^ . '^^'* ^= 2 and therefore C/= ^. < ^ In the same way I may be shown that if <r dx\ \p Jn (Xr) Jn (Xp) dp = i. and T.. h=ooJr pr pi dp=Lt ^ "^ t.VIII.. . for J^ (rh). = ^fP+'' Lt ^— (oc ry doc . \ = consequently tt?* 0.rh) 1= W cos (a ph) ^ ^ sin (a v / ^ —p sin (a — ph) cos (a — rh)} _p^ dp ~ irr^ — \sm(p r)h sin (2a \~^^^ ( I — (p 4. P~''sin/ia.oo [Jr pr In virtue of two theorems due to Dirichlet T. p^ dp cos hoc. Jo 00 (x{rf . DEFINITE INTEGRALS. It has been shown by Hank el that this last formula holds good even when q It = 0. .. {rh) . J'^ 83 and similarly very large. Lt psin(«r)A ^^ .
Then by a lemma due to du BoisReymond . = r. The fact is that this should be independent of the upper statement is true only on the sup position that p and r are separated by a fiiiite interval. h)\ follows from what has just been proved that Lt 1 n = h c^{pji)dp \ if „ r ^<r<^. noticing that these results are analogous to the reduction of the effective portion of a plane wave of light to that of a part of the first Huygens zone. Thus it may be verified that if the positive proper fraction e be so determined that sin a. may be worth For convenience. /. let us write = \p Jn (Xr) Jn (Xp) dX I J then it ^ (/3. r + eirjh). [VIII. SO that the value of the whole integral It should be observed that the expression rh I rr+e J Q \p Jn dXj J r (X?0 *^n (V) ^P no particular when h becomes meaning which h and It infinite and is e infinitesimal has unless a relation e respectively assigned connecting the ways in infinite become and infinitesimal. h=(x>J q [0 „ r<q or r>p. . = q or r = p. is obtained by confining the variation of p to the infinitesimal range (r. case of this proposition the inequalities q<r<p Now suppose that f{p) is any function of p which throughout the interval {q. In the first must be understood to mean that r — q and p — r are finite positive quantities and in like manner with regard to the other inequalities which occur in this connexion.84 DEFINITE INTEGRALS. p) remains finite and continuous and is always increasing or always decreasing as p goes from q to p. . which vanishes when p limit p. then Lt I d\ \ ^ Xp Jn (Xr) Jn {\p)dp=^\. . however small that may be.
Now by 169. 169 Next suppose r j'fip) *(p. Lt h= co J ffip) r $ (p. some quantity between q and Suppose that r is outside the interval {q. r<p and that r and p are separated by a in the may be proved Lt same way that ff{p)<i>{p. when h becomes infinite. h) dp = i/(r) + M . 85 I' J q Zip) * (p.it [ f(p) f(r)}. Therefore Lt provided It f'f{p)<t>(p.] DEFINITE INTEGRALS. h) dp =/(3) r * (p. if Hs a positive quantity greater than p. because finite interval. p). and consequently must M be zero. both the integrals on the right hand vanish. Lt ff(p)^(p. h) dp + {f{p) J(r)}jy> (p. by adding the last = 0. p. h) dp = 4/(r) + M [fip) f{r)]. the same must be true of that on the right. Since the expression on the left hand is independent of p. is zero if is fi ultimately separated from r by a and it is J when = r. h) dp J q + where is /i {fip)f{q)]\''^{p>K)d<j. h) dp = q.h)dp = if{r) 170 finite interval. then by the same lemma =fir)[ * (p.h)dp hoo J q = if(r) 170' when r exceeds 5 by a finite quantity.VIII. fju M is certainly finite.. .h)dp = [r>p or r<q]. and therefore Lt / ^=00 J q f(p)<^(p. two formulae.h)dp and hence. then. h) dp and therefore Lt where fV(p) * (p.
if we like. finally. The function f{p) may even present any finite number of isolated discontinuities in the interval.86 Hence. [VIII. we may. and it may become infinite for isolated values of p provided that is J Q is finite. integration. DEFINITE INTEGRALS. It is upon the formula 171 that Hankel bases of the validity his proof of the FourierBessel Expansions briefly discussed in Chap. f{r) if > Lt h = oo J q I dpi \pf(p)JnOy^r)Jn(>^p)dX Jo \f{'^) q<r<p. VI. r<q or r>p. . r = q or r=p. change the order been supposed that f(p) continually increases or condecreases throughout the interval (q. removed. It has . 171 of and. as in the other case. p) but it may be tinually was done as shown. by Dirichlet in the analogous case of the that the formula is still valid if this restriction Fourier integrals. .
p.CHAPTER IX. Pn (oo) denotes the zonal harmonic of the nth. 2' ^ tan^ 6 Put mains = x/n. and suppose that n becomes infinite while x re finite then since . id spherical If. (1868). There )vered is a remarkable connexion between the Bessel functions harmonics which appears to have been first disMehler {Grelle Lxviii. and Hobson Heine. (1872). Ann. 134. others. 141). . it is known 6) that Pn (cos = cos^^ e\l ^ '"^"l. order. v. as usual. THE RELATION OF THE BESSEL FUNCTIONS TO SPHERICAL HARMONICS. by His results have since been developed by 135. pp. and Math.
Changing x into i^. and hence Lt Pn as before. Thus every theorem in zonal harmonics may be expected to : yield a corresponding theorem in Bessel functions of zero order it may. the limit. and proceed Lt Qn as before.^^. we find = (cosh ^) N ^" err cosh<^ ^^ (see p.(cosh^) = 1 r+°° d<i> ^J (cosh ^ 4 sinh 6 cosh 0)^+^ ' and if we put 6 — x/n. 184. Kugelfmictionen that I.C.7) /o (^)  Yo {x) {Kugelf. X cos ^Y J the same as that of / 0Y^ and this is e^^cos*^ . g^ i\^^^ " TT Lt Pn = ('cos e''' "^^ "^ ) [ c^</> = 7rJo(^). of . 74 Lt Q^ (^cos ^j = (log 2 . happen in particular cases that the resulting theorem is a mere identity.. p. of course. 165) is Q. when n is infinite. cos ( j—Jo (x) Another known theorem (Heine. [iX.245*). or presents form which has to be evaluated.88 for SPHERICAL HAilMONlCS.. we obtain Lt Pn i cosh  J = /o (^). . 67). * itself in an indeterminate There is a misprint on p. f cos X n \ is + * sm n ix cos .i. 245 : in line 8 read C instead of .
Ji .+. 207) 7rPUcos6/) = 2^^ "^'' ^ nl2"r~^^ / ^''''^ ^ "^ ' ^"^ ^ ''''^ "^^^ ''''^ ^* ^"^^ and therefore TrLtP^fcos) = Lt ^=«. Then {Kugelf.cos. Now when t is very large we have asymptotically and therefore to the same degree of approximation. p. In the same way p. when n large.. = gs Consequently . es Lt n=oo (^TITT [^ Pi fcos ll \ ^/j = f "e^^cos.) ] j^^ A +MVi _£)"* = 1.) n 0^ . 282).] SPHERICAL HARMONICS. may be inferred from the formula (Z. is Jl{ns) n ill + g) _ n(2n) 2^^ . i. 223) Q^( urnVTn^cosn n(27i + l)_^ r (cosh ^ cosh scl>d4> 6' ^)2nU(n~\^s)U(ns)}o + sinh cosh ' «^)«+i . U" ( (^^ +.IX. For the sake of a consistent notation we shall write Pi {x) instead of Heine's PI {x). p.g)^^+^ y27r2^+^e ^2^2ri+i ~"2^l "^J hence we infer that [ n) Lt 2^ n(». eos s<i> #. = i'Js{x). i.^) l n(27i) jjo [%.. 89 Heine has shown that the Bessel functions of higher orders may be regarded as limiting cases of the associated functions which he denotes by P'l {x).^ cos s(f) TTJo # 175 c. c. which is Heine's formula it (l. 2TTe^'' (n + 8)''+'+^ 2?4 (n .)n(.
— Jtt < a < K. K. {ti) = i {Ks (ti + 0) + iT. latter = {yKs{xoi). K.A cosh Jo scf) d(f> 17 with = a (cos a + 1 sin a). is a pui imaginary and if we define Kg (ti). Ks{x) has been supposed that the argument of th( is a real quantity. — J7r<a<:7r.(0). 17 we write iBts («) for Heine's Ks{^ symbol for an extended definition of th function previously denoted by it.90 that SPHERICAL HARMONICS.)F. I i present point of view. it functions Is{oc). give a meaning to Ks (x) in all cases except when a. we find that Ks (ti) = ^« {(log 2 . VII.(^).{x).0)}. {x + Oi) = (. we may write and use the K^ (x) = (y X putting [ e^cosh. when it is ma( . into ix we may obtain similar formulae for Is {x (log27)J". x \Oi = a (sin a + i cos a).J By we i. by the formu .^# = {yK. [iX. The function i~''^ Kn (i From our has in fact already appeared in Chap.. VII.H'' 3} =/V^"** cosh s. it {2»+' J\ ^.7) /. (ti . if. were expressly introduced to meet the difficulties arising fron the values of Jg {x).(a. not J^ and Y but Jn and i~'^ Kn (ioo) as above defined. to avoid confusion. In the Kugelfunctionen Heine practically defines his Kg the formula (^) b. in fact the function.iy e^'^'^osh I ^osh s^ d<j> where hence if.(x) = (yKs(x) Jtt. Yg (x) when ^ is a pure imaginary. (0  F. the most proper course is take as the standard solutions of Bessel's equation. lyi By changing x and In Chap. where t is real.
1 cos (/)) =S s. To Heine : is also due the following excellent illustration of the nethods of this chapter.cos <f) ¥— 26c cos 2n' (f> + c^ '^"' a = cosn . Then rnd Vd?^ — 1 = z sin n . Pn {xy _ V^2 . radicals (see Kugelf.3.X.Pl{x) Pi (y) cos scj^. ^2/2 .5.Gn (0. Vw^ ^ — 1 = i sin  n xy — \/x^ — 1 . 82 •ileine's fact. I For any point on the axis of real quantities Yn (t) = (log 2y)Jn (t) . Gn{x) is identical with in a different sense in have Kn(x) employed {x).sin .5. on the whole. 91 )nevalued as explained above. c V = COS  . («).iri Jn (t)..Oi) = . h — cos  .] SPHERICAL HARMONICS. It is t.. with a proper determination of the signs of I.he known that.3. now the current in England with but generally usage jonformity t must be admitted that Heine's notation is. p.(2nl)}2 (nlY ' _^  { 1.. in Kn We : )referable one. 'Jy'^ — 1 cos </) = cos _ ~ cos + sin . the The reader will observe that.(2/tl )] {ns)\{n^s)\ [8>0l Now suppose that X Ahere b. n {ya. we call it Gn )ecomes a onevalued function defined by then Y^ {x) also i8i yn{^)={\og2r^)Jn{x)Gn{x) ohe formula vith a discontinuity along the axis of real quantities expressed by Yn {t + Oi)  Yn (t . 49). and that n is very large.1 . 183 ^vhere the summation refers to and ttn = 1. c are finite real quantities..
1 ] supposed that 1 are real and that infer that b> c. p. (c) cos 50. (n ^ 2^ {1.{J27re.n^^* ' and similarly 2^^(?z!)2(?i+s)!(w5)! thus finally we arrive at Neumann's formula Jo (v 6 . + s)l{nsy. [i: to the second order.(2. for the extension of spherical harmonics and Be .92 SPHERICAL HARMONICS. 175. and employing we ha) Jo J(h^ .26c cos 00 (/) + c^) = Jo (6) Jo (c) + 2t J^ (h) Jg (c) cos 1 5</). we find ' _ ~ J.. c + 2S Gs (b) J.J{a?l){fl) cos c^} = I' PU2/) «^„ I.^bc cos <^ 4. W cos5</) i {Kugelf. A.3.26c cos (f) + c\ 183. from the expansion (n + i) Qn [xy .^l))^ n7r '~n=l p. Making n increase indefinitely. a result originally given by p. By combining Fo {j¥ 86. 60). 333) we derive the formula c') G.. Neumann {BesseVsche Functioi We shall not proceed any further with the analytical pari the theory. if a = j¥ . and 181 we 2bc cos (l> + c') = Yo (b) Jo (c) + 2£ 1 F. i In the same way.c") = S AJs (6) /« (c) cos 5^. 89.5. 1 84. Proceeding as on . (6) J.= ^ where Lt^ ^. in {Jb'2bccos^ + which it is = Go (6) Jo (c) h. (2 ?i!y?Z7r _J " {J§Tre'^(2nf'+iYn7r _ " n=l 2^^^(^ 2. (c) cos s(t>.
p. XVI.0 Lommel's treatise. Jour. and Greenhill. of Math. 93 unctions to a pdimensional geometry the reader may consult leine {Kugelf. Trans. i. in particular. 49). — xxv. Quart. M. 1881. see rlaisher in Phil. and ential equations by means of Bessel functions he may be referred and the papers by the same author in the \fatJiematische Annalen. while for the solution of ordinary differ). (On Kiccati's equation. o\. S. .) . 449 479) and Hobson {Proc. 431. xxii. pp. L.LxJ SPHERICAL HARMONICS.
y) it defines the form of the membrane at any instant. may be neglected. under certain assumptions made for the purpose of simplifying the analysis. and z = (t>{x. which for simplicity we may suppose bounded by lines o curvature. shall attend only to the transverse vibrations. VIBRATIONS OF MEMBRANES. am let Tds be the tension across a straight line of length ds drawi anywhere upon the membrane moreover let dS be an elemen . rg are the principal radii of curvature. By the term membrane we shall understand a thin. perfectly flexible. th applied force on the element is dS . Then if ri. of area. of suitable constraints applied at one or more all situated in the same plane. We = if its z represents the plane which contains the if membrane ir undisturbed position. I : One of the simplest applications of the Bessel functions occurs in the theory of the transverse vibrations of a plane circular membrane.CHAPTER X. and assume that the tension remains unaltered during the motion moreovei . When the membrane slightly displaced from its position of stable equi librium. material lamina. d(\>ldx will b< supposed that and d(f>ldy are so small that their square. of uniform density throughout and we shall suppose that it is maintained in a state of uniform tension by means is closed boundaries. and then left to itself. Let a be the mass of the membrane per unit of area. it will execute small oscillations the nature of which we shall proceed to consider.
d'z^ 1 _a^ = 1 . vhere ^ is the small angle which the inwarddrawn normal vith the axis of z. 10 satisfy the initial conditions. becomes /d"z /a^ d'^z' d^z\ vith c T =— fl . By changing from rectangular to '•quation i may be transformed into d'^z _ Jd^ Idz l_dy\ suppose that the membrane is circular. then we have to find a solution of 3 so as *y = 0. and bounded circle the r = a. suppose that the element is concave to the positive then the equation of motion is lirection of the axis of z : ocZ^ ^= 2^ f + i") cZ>Sf cos ^/r.:. bod cos ^/r lence the equation of motion d'^z . when r = a. and such that z tor all Now values of t. makes Now.nd its line of action is along the For •learness. ^he for all boundary conditions. at points on the fixed boundaries of cylindrical coordinates the general to these are that z membrane. normal to the element.] VIBRATIONS OF MEMBRANES. . and that prescribed values when ^ values of t. neglecting squares of small i_ quantities. 95 . then putting 4 u is independent of '> has to satisfy the equation d^u Idu 1 d^u V . Assume /here z t . 2 of i it latisfy remains to find a solution the initial and sufficientl}^ . = u cos pt.nd dzjdt may have = 0. = 0.
to each of these corresponds a normal vibration of the type 9. /c^"^ ct + Dns sin nO sin /cj*^ ct) Jn (4"' 0> 1 1 where Ans. derived from are enabled to construct the more general solution 1. If the initial configuration is defined by z=f(r.. value of n the associated quantities Jn (fca) = 0. etc. etc. we 10 must have Jn (ko) and this is = 0. k^. = v cos nd. ^mt J^m denote arbitrary constants. this will be a solution provided 1 dv drIt will is r dr («S)o. be sufficient for . 6) we must have /(?'. if [x. the solution of 8 is = AJn (kt) + J5Fn (kv). that this equation has an infinite number of real roots Kj. From r = : hence the conditions of the problem v must be finite when = 0. y is a function of ?' only.^ sin nO) Jn (4"^ r). !=» By assigning to n the values 0. i : ^ . and we further assume that u where that (Pv ?. z = AJn (tcr) cos nO. a transcendental equation to find k. and we have a solution of 3 in the form B z = AJn (kv) cos n0 cos pt — AJn (fcr) cos nO cos Kct 9 In order that the boundary condition may be satisfied. e) = t {Ans cos nd + B. we z = X (Ans cos n6 cos + Gns COS and taking with each 4"\ k^'\ 4"'. The initial conditions this particular type of vibration « which result in and no others are that when = 0. /cj*' ct + Bns sin nO cos icf*' ct nd sin Ons. 2.96 VIBRATIONS OF MEMBRANES. our present purpose to suppose that n a positive integer V this being so. etc. V. k^. It has been proved in Chap.
^) = tto + tti cos ^ sin ^ + da cos 2^ F . V. 6) sin nOJ^ {k. it follows from the formula 19 (p. for conthe in integrals. Dm naay be determined. </> (r. 5G. — and f(r. 14 from which the coefficients Cns. venience. bs being functions of r. we must have (f> (r. 13) that Jn (f^s^) so that = Jni (f^sa). the period being 27r or an aliquot part of 27r in like manner </> (r. 7 . . . In order to realise more clearly the character of the solution thus obtained. Dns we suppose. and VIII. 0). thus 6).X] •md whenever VIBRATIONS OF MEMBRANES. 6) = X/c?) c {Gns cos nd + D^ sin nd) Jn (/c?> r). the coefficients Gns. 0). and are periodic in 0. writing Kg. VI. by the equation If the all membrane are zero. /(?'.v) rdr. ^«« ^2 y>o/ \ f (r. 6) and continuous. motion is defined that the initial generality. let us return to the normal oscillation corre sponding to Z = Jn (fCgid COS n6 cos KgCt. If. h 61 + 62 sin 2^ + . . in the expressions for we may put J. = 0. finite. for the sake of greater however. Since Jn(Ksa)= 0.. From the nature are onevalued. . instead of k^^\ "^ns = = 2 r27r fa \ ::rff^^\ '\ j /O^ ^^j^ d6\ ^) ^os neJn {ksv) rdr. 6) 97 admits of an expansion of this form the Bns are determined as in Chap. M. 1 5 Ks being the sth root of Jn (/c^a) G. of the case the functions f(r. . the quantities ««. "fficients A^is. 57) form of definite In fact.f_i (fCsa) for J^^^ (ksO) starts from rest. /(?% —may be expanded in the form . The possibility of exthese functions in form of the series panding ^Aj^JniiCsr) has been already considered in Chaps. (pp.
98
VIBRATIONS OF MEMBRANES.
[X.
Each element
of the
membrane executes
a simple harmonic
oscillation of period
and of amplitude
J^ {ksv) cos nO.
The amplitude
at rest,
if
vanishes, and the element accordingly remains
Jn{icsr)
or if
= 0, cos nO = 0.
not only
The
first
equation
is
satisfied,
when r = a,
that
is
at the boundary, but also
when
r
r— — a,
Kg
= — a,
Kg
...
r
=
Kg
a
;
consequently there exists a series of with the fixed boundary.
(5
— 1)
nodal circles concentric
The second
equation, cos n^
TT
= 0,
is satisfied
when
Stt
/,_(4^l)7r.
therefore there
a system of n nodal diameters dividing the 2n equal segments every one of which vibrates in precisely the same way. It should be observed, however, that at any particular instant two adjacent segments are in opposite
is
membrane
into
phases.
The normal vibration considered is a possible form of oscillation not only for the complete circle but also for a membrane bounded by portions of the nodal circles and nodal diameters.
It is instructive to notice the dimensions of the quantities
which occur in the equations.
or
The dimensional formula
for
is
that for
T is
c is
M=
[?']
[MI.«];
= =
[MT='];
[I.T'].
hence by 2 that of
Since Kg
is
[c]
found from
J^, {fCgO)
= 0,
Kga
is
an abstract number,
and
X.]
VIBRATIONS OF MEMBRANES.
99
Thus the period
dimension [T].
If
^ttJksC
comes
out, as of course it should, of
we
write
jjls
for
/Cga,
so that ^s is the sth root of
Jn (cc)
—
0,
the period
may be written in the form
27ra
This shows very by increasing the mass of clearly the membrane, or diminishing the tension to which it is subis
where
M
the mass of the whole membrane. the period
is
how
increased
jected.
particular case, suppose then smallest root of J^ (^) =
;
As a
n
— 0,
and
let
ytti
= 2*4048,
mode
the
of
we have
the gravest
vibration which
is
symmetrical about the centre, and
its fre
quency
is
Thus, for instance, if a circular membrane 10 cm, in diameter and weighing '006 grm. per square cm. vibrates in its gravest mode with a frequency 220, corresponding to the standard A
adopted by Lord Rayleigh, the tension
T is
determined by
4{.
whence
in
of force this is
dynes per centimetre, approximately. In gravitational units about 50 grams per centimetre, or, roughly, 3'4 lb.
foot.
per
r
= a and
In the case of an annular membrane bounded by the circles r = 6, the normal type of vibration will generally in
volve both Bessel and
Neumann
functions.
Thus
if
we put
1
z=A
~r^
—
i

T7T
—
^r
cos n6 cos KCt,
7
Jn{Ka)
Yn{Ka))
this will correspond to a possible K is determined so as to satisfy
mode
of vibration provided that
Jn {Ka) \\
(kI))
 Jn (fcb)
Yn
(fca)
=
0.
1
8
7—2
100
It
VIBRATIONS OF MEMBRANES.
[X.
may be
Yn that
this equation has
inferred from the asymptotic values of an infinite number of real roots
;
J^ and and it
seems probable that the solution
SS
is sufficiently
{A cos nO
+ B sin
n6}
\*^^,
^
Yn {ko)
cos KCt
19
starts
general to meet the case
when the membrane
from rest in the configuration defined by
Assuming that
this is so, the coefficients
A,
B
can be exprecisely
pressed in the form of definite integrals by a similar to that explained in Chap. VI. Thus if
method we write
u
it
=
Jn (kv)
Jn(/ca)
Yn
(/cr)
Yn(Ka)'
\
will
be found that
/•27r
rb
Ja
dd
Jo
I
f{r, 0) ur cos
nOdr = LA nOdr
,
20
f{r, 6) ur
sin
'^dO
i
= LB,
where
21
This value of
L may
perhaps be reducible to a simpler form
in virtue of the condition
Jn (fCa) Yn
the
(Kb)
 Jn (fcb) Yn (Ka) = 0.
For a more detailed treatment of the subject of this chapter reader is referred to Riemann's Partielle Differential
gleichungen and Lord Rayleigh's Theory of Sound,
CHAPTER
XI.
HYDRODYNAMICS.
In Chapter VI.
(f)
it
has been shown that the expression
= l,(A cos nO + Bsin n6) e~^ t/n(^^)
= 0,
satisfies
Laplace's equation
V2</,
and some physical applications of this result have been already considered. In the theory of fluid motion may be interpreted
<j>
as a velocitypotential defining a form of steady irrotational motion of an incompressible fluid, and is a proper form to assume when
we have
to deal with cylindrical boundaries.
shall not stay to discuss
any of the special problems thus suggested, but proceed to consider some in which the method of
procedure
is less
We
obvious.
Let there be a mass of incompressible fluid of unit density moving in such a way that the path of each element lies in a
plane containing the axis of z, and that the molecular rotation is equal to &>, the axis of rotation for any element being perpendicular to the plane which contains its path.
Then, taking cylindrical coordinates r, 6, z as usual, and denoting by u, V the component velocities along r and parallel to the axis of z respectively,
and
or
dz
= 2a).
2
102
Equation
I
HYDRODYNAMICS.
shows that we
[xi.
may put
where
yjr
is
Stokes's current function
;
thus equation 2 becomes
""
'
dr \r dr I
dz \r dz j
i/r
^
;
When
we put
the motion
is
steady,
is
a function of r and z
and
if
q^
= u^ \ v^,
may
so that q is the resultant velocity, the dynamical equations
be written in the form
14(4^^)^ "^=°l
dz^dz^^"^^
whence
it
r dz
"'
as a function of
i/r.
follows that (ojr
is
must be expressible
ft)
The
simplest hypothesis
= fr,
6
where
J*
is
a constant
;
on this assumption, 4 becomes
dz \r dz J
dr \r dr )
Now
the ordinary differential equation
dr \r dr
is satisfied
)
by
where
A
and
B
are arbitrary constants
yjr
;
and
if
we assume
=
'x^\
pr cos nz,
find from 7 that
1 \
where p
is
a function of r only,
d^p
I
we
(
,
dp
the solution of which
is
p^CnUinr) ^BJL^^nr),
Finally, then,
>/r
= Jfr^ + ^72 + 5 4 r%
{(7„7i (nr)
+ BJ<^x {nr)]
cos nz,
8
XI.]
HrDRODYNAMlCS.
103
where the values of n and of the other constants have to be
determined so as to meet the requirements of the boundary
conditions.
Suppose, for instance, that the fluid fills the finite space inThen closed by the cylinders r = a, r — b and the planes z = ±h.
the boundary conditions are
1^ dz
when r = a
or
b,
=
;
for all values of z
and
dr
when z = ±h,
for all values of
r such that
a^r^b.
One way
we put
of satisfying these conditions
is
to
make
yjr
constant
and equal to zero at every point on the boundary.
Now
if
this is
of the right form,
and vanishes
for r
= a.
;
It vanishes
when
r
= b,
provided the values of n are chosen so as to satisfy
I, (no)
K^ (nb)  1^
{nb)
K^
{no)
=
lO
(7„
and, finally,
it
vanishes
when
z=±h
if
the coefficients
are
determined so that
n
for all
[Ii(na)
K^(na))
values of r such that
a>r^b.
Assuming the possibility of this expansion, the coefficients are found in the usual way by integration.
The stream lines are defined by
i/r
= const.,
6
= const.,
that the outern^ost particles of fluid remain, throughout the motion, in contact with the containing vessel.
(The above solution was given in the Mathematical Tripos,
Jan. 1884.)
dp .
we obtain P . during the steady motion. in the second case w= boundary conditions.SO)) icJ[ (kt) (71 Js (Krn = m (4a). n. 250) ] m {4a)^ — (n — scoy] . (Xry 22 initial or The constants must be determined by appropriate For instance. . If the quantity m''{4fco^{nSQ)y} (n is positive. 105 From equations 17 . is — pQ cos mz sin s6.^ —w)^ 19 div s(n — SO)) m reduction. suppose 711.HYDRODYNAMICS. after and on substituting these expressions in 18 we d^w dr^ a little 1 dw ' {rn? [^(o^ Jm2_{4ft)2_ r dr 1 I (n sayy] — (n (^ scoY — — s = 0. in the disturbed motion.— r — 23 5ft) yj By and 16 the corresponding radial velocity is = p cos Tnz sin {nt — sd) initial velocity if p^ is the value of /> when r=a. small it is necessary that J5 = To fix the ideas. s. the whole interior of the cylinder r = a. if it is — \l and negative. . (o assigned^ . the along the radius. w may be everywhere in 20 and D = in 21.dw = {nsw)^nscD)^ . {4ft)2 — {n — sayy] find. . suppose the fluid to occupy. let it — soyy be called /c^. and that 4ft)2 >{n — s(oy then by 19 and 21 A n '^ {n — sco) \{n ^ . Then in order that. 20 r^ j^(. let it be denoted by 21 Then in the first case w = AJg (kv) + BYs {kt). CIs (\r) + DK.{. for r = a.
s. i=0. = 0. given. the velocitypotential is and the velocity at any point r If a is the radius of the free surface. supposing that its this is prescribed." The general solution involves both the is J and the / functions. we can obtain a very general solution by the different disturbances of the type considered compounding which arise when we take dififerent values of m. is cd. Now by 23. {4ft)^ — (n — smY} mpo Jg (fca) [ (n — sod) \ {n — SO)) kJ'^ (ko) — K^Po/m kJ'. value being. the constant A is determined. — There is no difficulty in realising the general nature of the disturbance represented by the equations 16. the pressure for the unis disturbed motion . n. po may have any (small) constant value . any arbitrary periodic function of the it is Lord Kelvin time. r that of a hollow irrotational is This 72^ obtained by putting = 0.106 HYDRODYNAMICS. according to (o When is " possible to construct in this way the solution for distribution of the generative disturbance over the any arbitrary and for cylindric surface. it evidently travels round the axis of the cylinder with constant angular velocity n/s.Js (ko) course. [xi. . Another case of steady motion vortex in a fixed cylindrical tube. where c is a constant. the other initial component velocities and the pressure must be adjusted so as to be consistent with the equations 16 19. (ko) Of initial 24 ^.
If the fixed when r =6 that boundary is defined by r = is.XL] HYDRODYNAMICS. we corresponding to 1 7 and 1 8 dt^T ( cs\ 2cu ' ^ r' S'UT ( CS\ 27 dp dr p r . are arbitrary constants. must find a first approximation to the form of the free surface. much from its mean and if we take the equation r = p cos mj2 sin (nt — s6) . _ 1 dw dr ^ m 1 Sna therefore the differential equation satisfied by w is d^w dr^ dw rdr (^^'+3^=0. 6. the coordinates remain invariable and r. Is(mr) (mb) Thus w still i Ks(mr)] ' 31 K'. by 16 and 29 . w = AIs (mr) + BKs (mr).(mh)] j) We have to express the condition that = ITo at every To do this we point on the free surface for the disturbed motion. su r 1 / = ~[ncs\ \ 28 sw Lence mr ^ w.^ we must have r = = dr when r = b. B . ^ of a particle of fluid 6 — ^7 r^ whence 6= • r In the disturbed motion r does not differ value ?'o. 107 Putting these values of U and TI in equations i6 and proceedfind for the approximate equations ing as before. 30 Consequently where A. In the steady motion.
108 HYDRODYNAMICS. = Putting Vo a. and writing pa for the corresponding value the approximate equation of the free surface is r c =a n ^^^ cos mz cos {nt .s6) and the condition p gives.s6). . Po. ct =— [X we obtain a first approximation by putting ^ giving p it mean value po and neglecting the variation of z : thus r = po cos m^ sin ( n — ^ I ^ and therefore r =n n cos — cs mz cos (nt — sd). sc 3 Now by 16 and 26 p= TIo + ^c^l— = Uo 1 a' — 2) + ^ cos mz cos (nt . with the help of 33.
\ . N is are real an abstract number. been anticipated. travelling with angular velocities (=?)'(•?) respectively about the axis of the vortex.] HYDRODYNAMICS. In this case we w — AKs (mr) and 37 reduces to N=. which is positive whenever a. for the undisturbed motion.ma The third ' Ks {ma) by Lord Kelvin is that of a cylindrical core rotating like a solid body and surrounded by liquid which extends to infinity and moves irrotationally. we write surface in the steady motion). from other considerations. Thus if a is case considered the radius of the core.) 39 J j^ = U — r ^ f when r>a. This 109 may be regarded as an equation to find n when the other If quantities are given. with no slip at the interface between it and the core. s there are two oscillations of the type 16. b. we have JJ = cor when a>a^ r < a. to disturbances The interpretation of 38 is that corresponding to each set of values m. Thus the steady motion is stable in relation This might have of the type here considered.^XI. m and h>a. (the angular velocity at the free and AT _ ^n I^^^ _ ^'» (^^)) ^ ^'"^""Vlimh) K\{mh)] ' \^s(nia) \l[{mh) _ Ks(ma) Kimh)]^ ) 37 the roots of the equation 35 are given by 38 n=(D{s±sJN). A special must put case worth noticing is when Z) = oo . .
a transcendental equation to find k when the other constants When k is known. (fca)> ' = . and 4 A(n — sco) \{n — soy) kJ^ (ko) "^ Jg m . / v « 4 t Eliminating A/B. we and z = ir/m are like. ma fixe boundaries of the We = will now consider the irrotational wavemotion of hom< geneous liquid contained in a cylindrical tank of radius a ac depth h. referred to the original paper abo\ if we Since the expression for i involves the factor sin mz.. suppose that the planes z = fluid. The upper surface is supposed free. For a proof that the equation 43 has an infinite number and for a more complete discussion of the three problen is ( in question. we obtain. [X i( For the disturbed motion we start as before with equations and by precisely the same analysis we find w — AJs (kv) w = BKs (mr) with when r < a. both sides. .sayf interface p.^mMY(»M{n. and in the plai z when undisturbed. .. A Js (/ca) = BKs (ma). the reader cited.' when r>a. P /— „ mKaJ'AKa) Js (/ca) ^ Ks (ma) ai given. K^aKg (ma) 25ft) mKg (ma) K^aK'(ma) n— _ ' S(o which is the same thing. and ct must have the same value o and 27 it follows that the values of 7 by ly are the same when those of w agree hence the two conditions t be satisfied are.. {4ft)2  . {n . n is given by ft) 5 + 2m \ Jk^ + my * real roots. by 23 and 29. At the w Now .swy] . ^.110 HYDRODYNAMICS.BK: {ma). on reduction. KaJg(Ka) Jg (/ca) or.
all fulfilled if or These conditions are (f) we assume 46 = AJn (fcr) sin nd cosh is K(z{h) cos mt. The equation 47 has an infinite number write. so that for each value of n we may more generally.. D are arbitrary constants. B.] HYDRODYNAMICS. C. 47 we have. trigonometrical factor Moreover instead of the single A cos mt sin nO in the typical term we may put {A cos mt\B sin mt) sin nO \(C cos mt + B sin mt) cos nOy where A. 49 The equations 46. Ill The velocitypotential </> must satisfy the equation and also the boundary conditions = ^ cz T when z — — h. . — m^ or cosh Kh + g/c sinh = m^ gK tanh «. 45 = when r = a. . 50 and by compounding the solutions which arise from different which contains integral values of n we obtain an expression for (j> a doubly infinite number of terms. 49 give a form of ^ corresponding to a normal type of oscillation when the liquid occupies the whole interior of the tank. (t> = XAsJn (/cfr) cosh 4^^) (z + s h) cos m^^H sin nO. a free surface. If gravity is the only force acting. ! provided that /c chosen so that J^(/ca)=0. n must be a whole number in order that (^ may be onevalued.XI. 47./t. as the condition for when ^ = therefore 0. of roots K^^\ K^2\ 6^c. neglecting small quantities of the second order . Kh = 0.
the period of the corresponding oscillation is a m A specially V X^ case a occurs interesting when a rigid vertical diaphragm. may be adapted By the methods of Chap. whose thickness may be neglected. VI..A) sin if. so that \ if. It will be observed that in 49 Kh is an abstract number . [XI. the solution to suit a prescribed form of initial free surface defined by the equation V=f{r). and mt put . extends from the If the position of the axis of the tank to its circumference. and since this vanishes rest. we have a new set which are ob tained by supposing where k is any integer. we have a 51 Integrating free form of the surface initial defined by possible start = 0. is defined by ^ = 0. This excludes some. moment any elevation of the free surface at *=©. = ^AJq (kv) cosh /c (^ 4. let us take </> n = 0. ^kAJq when ^ (kv) sinh kIi sin mt.112 HYDRODYNAMICS. but not all. from 7] = — 2 — A sinh KhJo (kvX tc < 2 the summation referring to the roots of j. = 1^ when (9 = 0. 53 . 5 1 the liquid must be supposed to with regard to t. of the normal oscillations which are possible in the absence of the barrier but besides . and is = \. as usual. Thus in the simplest case.{Ka) = 0. As a simple then illustration. those which can be retained. we write ij for the above the mean level. we put Ka root of Jq{X) — 0. we may put <j> = AJj. in the special case last considered.(Kr) cos ^ cosh k(z + h) cos mt. in addition to the diaphragm other conditions. we must have. when k = 0.
277).\ = gK tanh kJi ) ' The equation tan x'X has an infinite = and to each of these corre number of real roots. Similar considerations apply to the vibrations of a circular ofiembrane with one radius fixed. and of a membrane in the shape of a sector of a circle (see Rayleigh's Theory of Sound. . if 7 of these assumptions is secured if we suppose the radius of the basin. M. the liquid and basin together rotated like a rigid body and \irther that the velocity is always equal for particles in the 5ame vertical.] HYDRODYNAMICS. Another instructive problem. Mag. it is required to investigate the oscillations of the liquid on the assumptions that the motion of each particle learly horizontal. {Phil.XI. (f> — Ar~^ sin {icr) P. and only deviates slightly from what f would be . 109). which the same thing. I. is 118 or. G. sponds an oscillation of the type represented by 53. containing uniform angular velocity heavy homogeneous liquid. for simplicity. p. (o^a is small in comparison with and that the greatest depth of the liquid is small in comparison The legitimacy a is with a. (1880). ohat. the function = (^ cos mt + B sin m^) J^ C^^^) cosh k{z \h) sin 7i^. where k </> is any integer. if we put (2k + l)7r . defines a normal type of a tank of depth h bounded by the cylinder r — a and the planes 6 = ±a. may be stated as follows. p. due to Lord Kelvin '5) X. oscillation in with the conditions 47 and 49 as before. and equal to . with the conditions tan m? Ka— Ka = 0. that the h. More generally. We shall suppose. A kvith circular basin. 1 . mean depth 8 is constant.OS cos ^ cosh 2 k(z + h) cos mt. rotates w about the vertical through its is infinitely it i3entre.
5< where u. v . v now denote the component it. y. [XL Let the motion be referred to horizontal rectangular axes which meet on the axis of rotation. these axes. and are rigidly connected with the basin. v are the component velocities. parallel to w. velocities along the radiu vector and perpendicular to The equation of continuity. Then if u. the equation of continuity is li z is HM)*I=»' while the condition for a free surface leads to the equations 56 dx dp ^^ dx ' 5/ _ dz If w^e eliminate p from du di 55 by means of 57 and change to polai coordinates. in the h new dt notation. ft is (— ^ _ /du \dr dv _ u\ '^\ dz — rdO r) 5 From the equations 58 we obtain dz_ rdd d'z rdddt hence by operating on 59 with f ^ + 4^^ j and eliminating i^.114 HYDRODYNAMICS. of a particle whose coordinates are mate equations of motion are the approxi dt p dx 55 dt p oy the depth of the liquid in the vertical through the 1i\point considered. we obtain — 2q}V + g d_z dr 0. i.
115 is obtain a differential equation in [. = f cos {mQ is — wi). we have approximately 65 is small in comparison with gh. after reduction.nt). drcular pond with a vertical bank. 66 67 n^—^(o^\ dr r ji 8—2 . vhere /c^ = . u=() when r = a 64 The boundary jives. the equation 2ma) If o) Jm (ko) — iiKa J'^ {ko) = 0. : n are constants. for condition the determination of n. u=U sin {md ..l.o which. found be Let us assume ^ .] HYDRODYNAMICS.lf=o. z. = Fcos (md — nt). (?)t=». md put z — Jm (tcr) COS (md — nt). and f a function of r only then on jubstitution in 6i we find f^i«. — .XI.vhere m. gh nd 65 becomes In the general case )0 it will be found that the equations 58 and are satisfied by putting i. that of an open lidered. we take m to be a real integer. gh 6^ ^ The work now proceeds as in other similar cases already conThus for instance in the simplest case.
the equations of motion are . an forces act. 7 of the lake or sea.116 HYDRODYNAMICS. be remarked that the problem was suggested to Lor( Kelvin by Laplace's dynamical theory of the tides the solution i applicable to waves in a shallow lake or inland sea. and \ the latitud is supposed to be of comparatively smal the chapter with a brief account of th dimensions. 244. that if that no we suppose the liquid to be of unit density. [X^ By assuming we for the solution of 62 obtain a value for z which circular may be adapted to the case of a with circular island in the middle. as in Basset's Hydrodynamics viscous liquid. p. II. We will conclude application of Bessel functions to the twodimensional motion of It may be shown. if we pu It should ft) = 7 sin X. which being the earth's angular velocity. pond : .
1) Ka] + (1+ i) (a . /c !nd therefore that In order to obtain this explicitly in a real form. This leads to a function of r only. md vhere "^ is %= '. eliminating 117 p drV dr)'' V^X = g that rm~^' 0. . xf'^ = ^ ' Lnd if we write ^_ :)ne 2 value of "^ is We obtain a real function for '^) f<:^} ^/r by putting = (a + y8i) e^*^ Jo {(1 ^ :. 771 being any real constants.^i) e^«*' Ji {(1 +i) Ka]. + (« .i) ^ ««) f + W». = ^e»"<». let us write «/. 72 comparatively simple solution may be constructed by sup IEence. cribed to be aco sin a(o sin mt when r = a.i) (a + /30 e'^'^'Ji {(1 .A') e"**^ e/o {(1 + ^) /cr].r 0. d^^r H dr^ — dr r 1 d'lr mi fjL . then mt = — l^] \drJr=a = k(1. Suppose the velocity is 73 pre )9.^.HYDRODYNAMICS. {(1 .
Co n. Jan. n^. (This example Tripos. . surrounded by viscous liquid 0] density p. velocity V at any instant is expressed by the formula where v=fjLlp.118 HYDRODYNAMICS. = 0' 7 and. /n being the coefficient of viscosity. c. The current equation function yjr must vanish at infinity. carrying The boundary condition may be with it the particles of liquid which are in contact with it. for vol. P and 'v/r Q being real functions : then = law {(P + Qi) (cos mt + i sin mt) + (P — Qi) (cos mi — i sin mi)) — = lao) (P cos mt Q sin mt). we shall endeavour to give an outline The practical problem is that of taking into account the viscosity of the air in considering the small oscillations. of a cylindrical pendulum. in addition. and satisfy th ^^(^^^1)^ ^^ = Va cos 6. are con ^ jugate complex constants. Wednesday afternoon. [XI. Trans.) very important application of the theory is contained in Stokes's memoir " On the effect of the internal friction of fluids A on the motion of pendulums" (Camb. are conjugate complex constant. is taken from the paper set in the Mathematical 3.): the details of the investigation the reader should consult the original paper. we begin by supposing that we have an infinite cylinder of radius a. when r = a. realised 75 by supposing the to fill the interior of an infinite liquid cylinder of radius a. under the action of gravity. but of the analysis. In order tc simplify the analysis.. Phil. ix. the boundary conditions od = ^ or F sin 6. which is constrained to move with angular velocity (o sin mt about its axis. also extending to infinity and we proceed to construci a possible state of twodimensional motion in which the cylinde] moves to and fro along the initial line ^ = in such a way that it? . 1883. and of small absolute value.
= iril(i^)W}.{\r)=P + iQ. %„ These equations are satisfied by X Xo =K.+^%'W = ^. . 119 Now if we assume ^vnti ^_ + j5^ (^)l 1 4. if The boundary conditions are satisfied — + %(a)=ca. whence 81 B= and 2ca X (^) + ^X («) i A„ B^ are obtained from these by changing into i .{(l\i)^/nr}. Q2ynoti l^' + BoxAr) sin 6 79 part of this expression. HYDRODYNAMICS. namely the sum of the first and third terms.XI.\ where 80 P and Q are real functions of r. X=K. satisfies the equation and the remaining part satisfies are chosen so that provided the functions ^. (1 4.i) \ln Put =\ (li) then \/no=\o'.
8 1 may be regarded as giving the motion when the cylinder is constrained to move according to the law expressed by y6. J Now let us suppose that we have a pendulum consisting oi a heavy cylindrical bob suspended by a fine wire and making We shall small oscillations in air under the action of gravity. [xi. of the fluid The equations 79. to amounts. assume that when the amplitude of the oscillation is sufficiently small. must act at time be the density of the cylinder: then the force which t upon each unit length of it.e^'^''^ L=^Bx(a) ^ ca{ax'(a)Sx(a)] and Lo Let is cr g conjugate to L. per unit 82 Z = 2TTpvnia {Le"^'^^ where L.Noe^'^'<^% with 84 ^ iVo= the conjugate quantity. the motion will be approximately of the same type as that which has just been workec out for an infinite cylinder. . we shal have I = F = ce^*' + Cog'^"**. and the period sufficiently great. By proceeding shown as in Basset's that the resistance to the Hydrodynamics. 8( The force arising pei unit of length. 80. 7^(o/))a^'f.120 HYDRODYNAMICS. F=jr<ra''^]^+Z dt = 27riva^ (N'e^^^ . il. arising liquid. and to the order of small quantities. so that if f is the horizontal displacement of the bob at time t from its mean position. from gravity which acts upon the bob first is. in order to mainis tain the prescribed motion. 280 it may be motion of the cylinder. from the viscosity of the surrounding length of the cylinder.
which satisfied identically if (o we put . and substituting for f its value in terms of the time.p)gc = ^nvHN.vHN^ + {(Tp) gc^ e^"^^** = 0. I V %(«) + «% (a) V This.4in. by the f and f together with the equations 86 and 8/.] HYDRODYNAMICS. e^"^^' we obtain { ^nvHN +{(Tp) gc] is + (. is an equation to find n which must be solved by approximation: since the motion is actually retarded. which is the same thing. or. Equating this to the value of ditional equation F given above. the proper value of n must have a positive imaginary part. Doing this. we have the which must hold at every instant. with ^ (a) defined by So above. comparison with a.XI. I 121 where is the distance of the centre of mass of the pendulum from con the point of suspension. As might be expected. and may therefore be differsntiated with regard to the time. c Co . The constants and are determined . 4:P'7l^ when p is very small in = g/l initial values of approximately.
STEADY FLOW OF ELECTRICITY OR OF HEAT IN UNIFORM ISOTROPIC MEDIA. 6. required for the application of Bessel Functions to problems regarding the distribution of potential. but the problems solved may be regarded as problems in the theory of the steady flux of heat or incompressible fluid moving irrotationally. which deals with FourierBessel Excontains all that is pansions. Chapter but it may be discussion by a few examples advisable to supplement that theoretical take here fully worked out. z. Other with notes as to their solution in certain cases will problems be found in the collection of Examples at the end of the book. The potential the flux becomes the temperature in the thermal analogue. while the conductivities and strength of source (or sink) involve no change of nomenclature. The method in of translation of electricity is well under stood. the sources and sinks in the former become positive and negative charges in the latter. or even of the distribution of potential and force in an electrostatic field. In the discussions in this chapter we speak of the flow as electric. then in all the problems here considered the differential equation which holds throughout the dx^ ' medium is dy^ r. above. We a few cases of electric flow of some physical interest. the potential in the flux theory and that in the electrostatic theory coincide.CHAPTER XII. dz^ or in cylindrical coordinates rr laF la^F a^^^ . while specific inductive capacity takes the place of conductivity. VII. If V be the potential.
and consider the electricity delivered or drawn off uniformly over a small spherical electrode of perfectly conducting substance (of radius r) buried in the medium at a distance great in comparison with r from any part of the bounding surface. and 7 are the conditions to be fulfilled the problems which we now proceed to give examples of. the equation of condition is led into or Let us define a source or sink as a place where electricity is drawn off from the medium.] STEADY FLOW OF ELECTRICITY OR OF HEAT. >Sf 4>7rkr ^ The quantity on the right is half the resistance between a source and a sink thus buried in the medium and kept at a difference of potential 2F. so that say k2=0. and the resistance will be double the former amount. n. V^ are the potentials in the two media infinitely near that point.^ an insulator. 2. k^ the condition which holds is At the where into the respective media. 123 surface of separation of two media of different conductivities ki. Let it be kept at potential V. then since F = constant/r. If the electrode is on the surface (supposed of continuous curvature) of the medium the electrode must be considered as a hemisphere. 3. If one of the media is n^.XII. and deliver or withdraw a total quantity S per unit of time. and therefore in the two cases just specified 4}7rk in Equations i. denote normals drawn from a point of the surface and Fj. In this case Z= S L_ 2irkr ' 6 When r is made infinitely small we must have rV finite. Those we here choose are taken from a very instructive paper . ETC.
JB r = ^re^ sin (\n) J„ (\r) ^ ." Grelle.124 STEADY FLOW OF ELECTRICITY Oil OF HEAT [XIL by Weber ("Ueber Bessel'sche Functionen und ihre Anwendung auf die Theorie der elektrischen Strome. This can be written ^ J e^^ [sin {X (n a + r sin 6)} + sin {X (ri . then if z be taken along the axis of the disk. and gives the proper value of the electric density solution. l . Bd. 1873). and the addition of some explanatory analysis.r sin 0)}] — X But we know that if > . 18 above. If be the shall prove first the following proposition. 8 where the upper sign is to be taken for positive values of z ana the lower for negative values. X since changing the order of integration permissible here. Consider the integral /7"\ f Jo e~^^ sin (XV]) cos (Xr sin 6) — A. we have verified the dX I Jo e'^sm(Xri)Jo{Xr) =I e~^^ sin (Xvj) \ I cos (Xr sin 6) dS d\ =  I dd I e~^^ sin Xr^ cos TTJo Jo is (Xr sin 6) — . if then when shall z = 0. r" I e~^^ sin w dx = Jo a — + 1 V . if e > 0. and the origin at the centre. . 75. and are given with only some changes in notation to suit that adopted in the present treatise. and c is the potential at the disk. By 46. In the first we can prove place this expression for that it reduces to a constant V satisfies 2 . p. potential due to a circular disk of radius r^ on which there is a We V charge of electricity in equilibrium unaffected by the action of electricity external to the disk.
"j^^:^^) I'm tan. ^ r ^ B 2Jo sm ^ . sin^ . so that the integral sought in that case 77^2. down on the when it vanishes. ^ ri € . is — dO. c?^tani ^—^—nsm ^ n —r 10 The e integral on the left 0. Similarly except just at the beginning each element of the second is — ^irdd. ^\nx —=o — tan~^ 2 dx ^ TT . and the integral requires discussion. 125 Multiplying this equation by da and integrating both sides from a = e to a= CO (e > 0). + rsin^ —A — 1 € 7^ tan 2 nrsin^' and 3 e~^^ sin (\ri) cos (\r sin 6) — ^v Jo J i = ?ol 2 2J (^<9tani 7*1 —+ . first The second is of the two integrals on the right vanishes if e be very also vanishes when r^ > r. In doing value of the integral on the right this there is no difficulty if 7\> r\ but if r^ < r. Hence for r > rj is when ^ = sin~^ (n/r). for which ^ = sin~^ r^jr. e r Each element of the right since e is first just written ^7rd6 except just not zero. Now Jo tan. The small. Thus the TT integral considered has the value 1 . tan 2 2 _.r sm ^ . d6.] IN UNIFORM ISOTROPIC MEDIA. the element if Hence we evaluate the of the last integral in lo. .2 2 rsin^ ^ tan^ gij^ ^^ fi Jo e d6—2\ J integral _^nr tan^ . — = tt sin^ — r .r.r^rsmd = 4'de{le / Jo \2 i^ = ^2 ^ . is convergent for all positive values of equivalent expression on including of e we shall obtain the for a small value the right very positive of 8 when z — O. we obtain e~^^ /. u d6 tan~^ Tj ^ .XII. .
as well as dV/dz. c.. and be immersed in an unlimited medium of conductivity k. for c £^ =+ 1 dv dz c .. and we have = — when rj = r. and let be composed of perfectly conducting material. I 4>7r we have . otherwise obtained. Or the whole is cJit^Jt^ density. — sm — ^ if TT r r > ri. r" I 27rVo sin (\ri) i/ Jo(Xr) v ov / . This is is a result which can be Hence the solution completely verified.. gives a constant potential at every point of the disk of radius r^.126 STEADY FLOW OF ELECTRICITY OR OF HEAT finally for e [XII. We in the can flow now of electricity. The rate of flow from the disk to the medium is —kdV/dz per unit of area at each point of the electrode. . and is. At the edge by 1 3 the flow will be infinite if the disk is a very thin oblate ellipsoid of revolution as here supposed to be but in this.)Jo(\r) \ the differential equation.. convert this result into the solution of a problem Let us suppose that the electrode supplying electricity is the disk we have just imagined. 73 above.... and is of course in the direction of the normal. the two results coincide. Then to a constant it the potential at any point of the electrode is F= — TT 1 sin (Xrj) Jo (^0 J ^r A • 14 The sink or sinks may be supposed at a very great distance so that they do not disturb the flow in the neighbourhood of this diskshaped source. taking the faces of the disk together.. d\ = — rl ~ ^^^ ^^_^ two x^ by 151. continuous when z = 0. for all values of r. and in any actual case the total flow from the vicinity of the edge can obviously be made it is . if r <n II . Lastly to find the distribution. Thus the expression 9r V=— satisfies f "^ J Q e^^ sin (\r. p. / TT J sin (Xri) Jo (^r) —= A.
and we have S If the disk laid is on the bounding surface of a conductor the flow will take place only from one face to the conducting Then mass. a source and a sink. 127 please in comparison with the total flow elsewhere disk. c = ^ i6 holds all over the surface In this case the condition 8F/3?i = of course 2 holds within the and the at diskelectrode. /)2. except conductor.XII. placed anywhere on the surface at such a distance apart that the streamlines from or to either of them are not in its neighbourhood disturbed by the position of the other.] IN UNIFORM ISOTROPIC MEDIA. c^. At any point of the disk distant r from the centre 1 dV^2c dz TT S 27rkrj Vr^ 1 ^^ — r^ — r^ ' We can now find the resistance of the conducting mass between two such conducting electrodes.and have their farther extremities . then if R the resistance between them R Ci Co S If the wires leading the current up to and away from the electrodes have resistances pi. by increasing the radius of the as small as we The total flow from the disk to the medium is thus Putting in this for dV/dz its value we get TT Jo = ScJcTi. The whole current up to the disk by which the current enters is S. time Thus the amount supplied by each side of the disk per unit is 4cA. and S has only half of its value in the other case. For distinction let the potentials be and sink disks be denoted by Ci.ri. and we have seen that c is of the source the potential of that disk.
c.(Ci . and along the outgoing are V^c^ = Spj. and R = l{V. For the sinkdisk the outward current >Sf in like manner is = — 4c2A. Fron the distribution of potential the streamlines can of course found also.S(pi + p. so that ^^ n + ^2 ^ 4A?rir2 1 ^ 1 ' ^kr^ ^kr^. Ij^^kr^ the part due to This result is of great importance. have seen that the potential at the sourcedisk is d also that from one side of the disk for conduction S= 4!cjcri.^ri). bt . means of calculating an inferior limit to the correction to be made on the resistance of a cylindrical wdre in consequence of being joined to a large mass of metal. We Another expression for the resistance can be found as follows. its From this problem we can proceed to another which is identical with that of Nobili's rings solved first by Riemann.V. and two disk electrodes are applied to these planes. Hence S=2k (CiVi — and CiVi = — c^r^. It is required to find the potential at each point oi the conductor and the resistance between the electrodes. for it gives a the second. V^. An infinite conductor is bounded by two parallel planes z = ±a. so that their centres lie m the axis of z.Cg) = ^ (pi + p^). (at the generator or battery) potentials Fj.128 STEADY FLOW OF ELECTRICITY OR OF HEAT at [XII. the falls of potential along the inleading electrode. From the latter form of the result we infer that Ij^kri is the part of the resistance due to the first disk.r2. so that Co.)}. .F2 = %. Fi  Fa .
fulfil 129 : The must the following conditions a^ + . Without loss of generality and hence we must put /•oo =— <^ (\) V may be \/r (\). to the other two conditions. fora<. s z= ±a. p._ ^ AfTrkr^ sinXri 1 cosh (\a) \ be satisfied. \^^ 9 G.XII.^ . 73. M. \= S . supposed zero when ^ Thus 20 becomes = 0. The . dV According to the . Hence if we take 2<j) (\) cosh (Ka) . when r > n . last condition the current is supposed to flow along the axis in the direction of z decreasing.< + a. sin (Xn). The first condition is satisfied by assuming h where </> (X).dX r/A ^ . 21 With regard above.] IN solution UNIFORM ISOTROPIC MEDIA. ^ )oth conditions will .:a7+a. r> r^. F= Jo 2(f> (\) sinh {Xz) /« (Xr) cZ\. . 151. X . . tor . solution of the problem is herefore rr S ^ [ (X^) r°°sinh(X^) Smh /^ ... by 150. y\r (X) are arbitrary functions of fulfil \ which render the lintegral convergent and the other necessary conditions.5— oz = 0. Jo I sin (\ri) Jo (^^) d\ = 0. when r<ri.0..
nirz '''' ^ ' • ^^ Hence niTZ nTT . 200 below. we have to a first approximation. p. is the result ^^ and we obtain from 23 * T^ 1 2kri log2 irka If the electrodes are extremely small we may put Xri for sin Xri S_ 27rk >3 r sinh (X^) r /^ \ jN /: J cosh(Xa) z= ±a._ nirz S ^ „. ^ (^Y \2a) . 7? — ^1 ~ ^2 __ I This of course could have been obtained at once from 19 by = r^. 1 ^. that 2 if x be positive r sin iix) d^ '°° ^«(^>4L Hence putting j k^ for (7Z7r/2a)2 we have r Jo(xr)xcZx ^2 /t^ + X^ r_<^i_^ r X sin (frX) cZX /c^ + X^ "ttJi \/p=li( . ^ . This expression applies to the space between the two planes Hence expanding by Fourier's method we obtain sinh (X^) cosh (Xa) _ " 2X a 1 .C2 = 7— I tanh (Xa) Jo (^'^) sin (Xrj) — If the second term in brackets be neglected in the last expression. Vi. an approximation to the For we have from 23 .„ WTT F= yZ sm ^ sm^r— 2 2a rta Tr >Sf r°° C t/o (Xr) XcZX nrry 28 1 j Hii Now it will be proved. since by 1 1 the integral is equal to 7r/2.130 STEADY FLOW OF ELECTRICITY OR OF HEAT [XII. To obtain a nearer approximation the simply putting r^ and expression on the right in 24 may be expanded in powers of r If terms of the order rl/a^ and upwards be neglected. From resistance this we can R between the Ci easily obtain electrodes. /mrV ^'"^ T mr .
^ = S . 2a ilZaCc) 2 irka 1 For consider the series 6. bounded by nonconducting matter. the problem becomes more complicated. 131 But it can be shown that according as ^r > or < 0.+ 6381112^+ 9—2 34 . &c. 95. ^ (1) / 9r2 o^ + .] IN UNIFORM ISOTROPIC MEDIA.^. Thus since f and r are both here positive Jo /^^ + x^ Ji VF^^n: ii ^\^i' ^^ Substituting in 28 K we obtain . 58. To solve it in this case a must be added to V fulfilling the following conditions part V . for ^ < Cj a<z<{a. Ann.+ r or ^rv dz^ = ^> = 0. the conditions stated are by 33 ^ = . or Pogg.7. nirz ^ . p. 2 sm^r— 2a 31 which agrees with the solution of the problem given by Riemann (Werke.sm.XII. March. (2) ^ for ^ = ± a.— S sm nrka i nir ^r . Bd. dV dV If we write 32 Vl\2 and denote by L^ (c).^ sm — sm ^— . Z2 (c). 1855). the same quantities with found to be fulfilled c sub •stituted for r. /. If the conducting mass instead of being infinite be a circular cylinder of axis z and radius c./ff ^2 (^).
Hence the resistance P_JL ~2Ati log 2 7rA. L2 (r) = ir. can be so taken as to first the differential equation in the medium.a 2 A. three conditions are fulfilled. z = ± a. 2 ka if ~ ^M^(c)_ n 2 ^ is ^2 (c j ka approximately .I sin '^ '^ 2 2a 1 _ M^ic) ^^ If in 37 we were to put r = 0. and could evaluate the integrals we should obtain Ci— Cg.(c)L.. The the .132 STEADY FLOW OF ELECTRICITY OR OF HEAT coefficients 61.(c)L. the resistance. Thus By the general differential equation we have of which there are two known V\2for solutions h If the first is ' 1 iiVlX2 X^ and integrated it is expanded infinite in powers of r found that it becomes r = 0.(r) r = irka sin I. We obtain in the notation of 32 is and the total potential at any point is F+ M. (3). so that the change in the resistance due to the limitation of the flow to the finite cylinder is When r = 0. [XII.. which are functions of r.(r)M.. thit may be a film of gas separating an electrode of metal from a con . For example.a ^ Weber extent We now pass on to another problem also considered by A plane metal plate which may be regarded as of infinite is separated from a conductor of relatively smaller conductivity b} a thin stratum of slightly conducting material. We therefore take as the solution of 35 J iVlV where ^^ a constant to be determined by the remaining equation of condition. the difference of potential now This divided by S would give existing for the given total flow S.be very small. fulfil 63.
and the origin on the surface of the small and conductor close to the plate. where Let p. at the point z Thus the point electrode is applied We further suppose that there is a difference of potential w between the surface of the conductor and the metal plate on the This will give a slope of potential through other side of the film. the second is the increased potential at each point in consequence of the rise in crossing the film from the plate.XII. Thus w be the positive difference between the plate and the conductor surface. solution we should have had if the film had not existed.r=^. be the distances of any point r from the electrode and from image in the surface respectively. the film of amount w/8 if 8 be the film thickness. h= p. z. and thus the flow per unit of area across the film is wh^\Z. = 0. +r + 7r„ ^ dr dz^ 1 dw d^w „ 40 The first term on the right is the throughout the conductor. Then the differential and the other conditions laid down are satisfied by equation its provided that w fulfils the equations dz at the surface. A value of w which satisfies 40 is given by w Jo . — a. Take the axis of z along the line through the point electrode perpendicular to the metal plate. calculate the resistance for the case in which the electrode is is applied at a point within the conducting mass. This must be equal if to the rate at which electricity is conducted up to is the surface of the conductor from within. 1^+^ dz 8k/ki. A.] IN UNIFORM ISOTROPIC MEDIA. If the con ductivity of the film be k^ the resistance for unit of area will be S/A^i. the condition holds when z = 0. and d^w ^7 dr^ = 0. which kdV/dz. 133 We shall ducting liquid as in cases of polarization in cells.
134 STEADY FLOW OF ELECTRICITY OR OF HEAT cf) where (\) is an arbitrary function of X to be determined. .
Thus we obtain for the potential at any r z+a Take a new variable f given by the solution becomes V=S. were there a combination of two equal positive sources of strength Sj^irk.2« \rj j2(l)''11.. as we suppose (z + a)/h may If h/r be small of t we can expand (hH^ + r^)~^ in ascending powers the binomial theorem and integrate term by term. parts. Avith a linear source extending along the axis of z from the image to . enough p — mass be of small thickness then nearly p —r. at distance f from the point If the conducting —{z + a). Hence value of the integral may be calculated if r be very great r be not 27r/kr" I . and z\a = a.(2«l)pg)™ 47 if R'o'/hH^ vhich the ^ too small. 135 by integration by point z.')' \p p p) pj J^ _^2£=. e h 27rkh per unit of length.4.3.x and of intensity . + 2'nkh 2'jrkh 45 Jo ^{^^^ z\ay r^ meaning of film this solution is that the introduction of the renders the distribution of potential that nonconducting which would exist for the same total flow ^Sf.. The »4i7rk ('.. by We thus get + r^ = rJo 2.. Thus we obtain ^irkXr if Jo^/WTrV be neglected.nm: ] IN UNIFORM ISOTROPIC MEDIA. at the electrode and its image.
Thus We by the solution out above for the infinite stratum with disk electrode worked we have The potential w must fulfil the conditions tor ^oz = 0. and the source is a disk electrode of radius r. h^ .136 STEADY FLOW OF ELECTRICITY OR OF HEAT [XII. z=a (since the flow from the sourceelectrode is supposed unaffected by ^) h dV ^ oz w==0. A We may mass is solve similarly the problem in which the conducting z = 0. simply add a quantity w. bounded by two parallel infinite planes. = 0. to the distribution of potential which could have existed if there had been no film. ZTrArJo cosh (\a) r*sin(Xn) rr^ ^^^ \^ \ Jo (Xr) dX .dV dz w= ^ 8h . sin (Xrj) "^ 4i7rkri cosh (Xa) [cosh (Xa) ^^^^ h hX sinh (Xa)} ' ' _ '^^ r°° X(z 27rkriJ cosh (Xa) {cosh (Xa) — a) sin (Xr^) J^ (Xr) dX + hX sinh (Xa)} 49 . As before a feebly conducting film is supposed to exist between the metal plate and the conducting substance. . (Xr) dX. with its centre on the axis of z applied to the latter.2h 2 1 Jo I sinh (Xa) X<t> (X) /„ (Xr) dX cosh (Xa) (f> (X) Jo (Xr) dX = 0. Sh ^ . besides of course the differential equation for points within the medium. The first condition is satisfied if we take w= Also when z Jo \ 2 cosh \(z — a)<j) (X) J. for z = 0. as before. the metal plate and the plane z = a. or the potential at a great distance from the electrode varies inversely as the cube of the distance.
is bhe apparent resistance of polarization. the resistance of the film between the plate and the con ductor.{\r)d\ + hX sinh * (Xa)j is The resistance of the compound stratum now considered bherefore ^^^_l__i!og2^ ^ irka 4A. Its approximate value. kept from complete contact with the plate by the disengagement of gas. ^^''^ '^^' ^ ^ This is the difference of potential between the electrodes. if and %/h is capable of being taken as infinitely small. = Jq (0) nearly. that the disk and the metal plate. that is with half the total difference of electrodes at distance 2a. and we put unity is term The last for Jo(Xr) in the expansion just found. Jo (Xr) between the plate and the electrode. 137 8 If rsmh(Xz)\h\cosh(\z) • /^ ^ x ' r /^ ^ 27r^TJo cosh(Xci)+/i\smh(Xtt) \^^ ^ X we denote by Va the potential at the disk electrode . we have ^'^ and if \ r ^\ \^^ psinh (Xa)f ^Xcosh (\a) ^^^^^^«^^^^ 27r^Jo ^sh(Xa) + AXsinh(Xa)'''' . Comparing it with the difference is.xil] so that in uniform isotropic media. S Y ' the area of the electrode be very small ^^ = S /""sinh (Xa) 2^ j cosh(Xa) + h\ cosh (Xa) + /iXrsinh~(X^) ^ '^'^ . potential given by 23 for the two which is A it tanh (Xa) Jq (^^) '^\ me see that exceeds the latter by Sh r 27rk J Jo cosh (Xa) {cosh (Xa) f \J. 1 is 27rka h ° a * The value ^series of F in 50 can be so as to enable comparisons of the value of expanded in a trigonometrical V to be made . in the case of a liquid in a voltaic cell.ri A 27rk ] ^oi\r)d\ f cosh (Xa) (cosh (Xa) + hX sinh (Xa)} is * Since the resistance which 80 is taken as of very small radius. of potential for the same flow through the stratum of the conductor without the plate.
4i^. the second ro( Thus if r be fairly great the first ten of the series jnst written down will suffice for V.(. = V . V r ' ^r Tr_ T(\ N ^^ _ ^''~27rk]o'''' '^^'^^'^^~'7rk{a^ is ^^ (^' + ^V) Jo{\r)\d\ + ha^hy^)}o /i^+W The expansion thus a' ^ V + hY za rej'^dk ^~7rka^a'\ha + h^fM'''''^^ by 30 above.) by the relation 54. To we put sinh (\^) cosh (\a) 4. we obtain the condition This transcendental equation has an infinite number of roots positive which are the values of find the expansion //. ha .h\ cosh (Xz) _ „ ~ + h\ sinh (ka) z— a ^ ^ ' a ( — a)/a} and Multiplying both sides by cos {/x {z integrating from to a we obtain from the lefthand side and from the right sin 2fjb\ _ 1 ^ _ /.+. ^Ji ^/}T^ ^ The is first root of 54 is always greater than tt. smaller the greater h is.138 STEADY FLOW OF ELECTRICITY OR OF HEAT [xil for different values of r. Thus __ Hence h 2a>^ g^ + /j.^^')=K. A suitable form to assume is a •' From this by the equation h^ oz — wy) which holds for 2^ = 0. Hence for z has a considerable value at a distance from the axis.. K.
We have which fulfils the conditions in this case a function to add to V V ^r— dz = 0. for K . verify that if in the quantities Li (r). 139 The solution can be modified by a like process to that used above to suit the case of a cylinder of finite radius c.. the general differential equation. tor ^ = 0. (c) A (r) . and the radius be unity.XII. at equal distances from the middle crosssection of the cylinder. for = z=a. having their sides parallel to generating lines and ends of the cylinder. = ±a. ()S + ^B) and . h .] IN UNIFORM ISOTROPIC MEDIA..^ = *. and satisfies of 32 above. of the centres of the electrodes from the central crosssection are here ± is by their breadth the angle subtended at the while the 20. the surface conditions to be satisfied are summed up in the equations 1^=0. (c) L. Li{c).. </> '</)< ^ ^=+c = 0.M.Sf ^ g^ + ZiV M. where for..::— cz + ^— = dr 0. . We shall merely sketch the solution. The reader may . be replaced by fi/a. leaving the reader to calculation.. height of the cylinder is 2a. The distances axis for  < + + I3 < z<^tB ^>z>(l3 + B) for all other points.(r) ^d ^ As a final and very instructive example of the use of Fourier Bessel expansions we take the problem of the flow of electricity in a right cylindrical conductor when the electrodes are placed on the same generating line of the cylindrical surface. 5— r dr = c. fill in the details of The differential equation to be satisfied by the potential in this case is 8^ laF ^ 18^ a^^ dr^ ^r dr r" dO^ dz' ^^ If the electrodes be supposed to be small equal rectangular disks.. k ^ .
in which when n = 0. if ^ (r) 58 differential equation 5 7 will be satisfied function of r which satisfies the equation be a dHi . 1 Idu du (n^ fn^ .. (2m + l)7r5^ for the cosines in the value A^. (2m+l)7r. s ^s M + 2zsinn<f)C0S7i^^ 2 2m + f . </> is to be put instead of 2 sin To find the effect of making the . . . This can be obtained by Fourier's method and the <!> 4c = — si.^ „. j (27?i {cos + — 2^^ l)7r„ /3 . 1) 7r/4a Remembering that and therefore putting 4c</)S/7r2=l.^ . '^' '^ _ ~ 4c 1 2 sin nylr' ^ 2m +1 n<l> (1) n f (2m + l)7r^ ^ 2^— ''' (2m+l)7r.) (27?i + l)7r .140 STEADY FLOW OF ELECTRICITY OR OF HEAT [xil. and </)S for sin n(\> sin j(2m +1) 7rS/4a} 71 (2m + C(^S is finite.ni^(r)sm^ and the ^— 1 N^) y] cos n^.^ + .) ) 1 1 . . electrodes very small w« substitute ..cos (2m+l)7r.J (^ 2a ^^ n(f>ln. .. 1 . /2m + Hence we put To complete the is solution the constant A^. clearly done by writing .J.^ (^ + S) sm ^— [ —^^ ^^ Now assume V=^ZZAm. n must be chosen This so as to ensure the fulfilment of the surface condition. We result is have first to find an expression for <I> which fulfils these conditions.
Ann. Several of his results are included in the »lxamples at the end of this volume... ^ a 2a A. 141 we get the solution where ^0 = 1. ei= eg = 63 = . nalysis dealing for the »essel Each of Weber's papers contains a very valuable introductory most part with definite integrals involving Functions. 64. for . Bd. 3e applied The reader may = bhe potential is given by F = z — sm Tia sm nd\. IN UNIFORM ISOTROPIC MEDIA. some more complicated problems of example a conducting cylinder covered with a oaxial shell of relatively badly conducting fluid. solution For an putting in infinitely long cylinder we can obtain the by 60 IT (2m + l)7r = ^ j_ — = a\.. any part of a circular disk with a circumference. =e„= . refer to another paper The reader may 6. =2. . • 1 r/ /•'^7r\ m sm na sm no cos f a be infinitely small the second part of this expression vanishes 1 bnd the first term can be written Fil ^^ fhich agrees with 8d..  2r cos (g t ^) + r^ r2' ^^l2rcos(a^) + 1845) urce and a sink in for the potential at its an expression given by Kirchhoff (Pogg. the two electrodes eing in the fluid and core respectively and a cylindrical core overed with a coaxial cylinder of material of conductivity comlectric flow. ind replacing summation by 71(9 integration. 1873) for the solution of by Weber {Crelle. Thus we obtain F=P6„ cos as before being 1. [ ^""i^^rl sin (\/3) sin (\z) ^0 *aJ^ {%\) all d\ 61 and the others 2.XII. parable with that of the core. verify as another example that if the electrodes central crosssection at points for which ^ at the + a.2z z in 1 ..
They have since been used in a somewhat modified form with great effect by Hertz and by Heaviside in their researches on the propagation of electromagnetic waves. a. PROPAGATION OF ELECTROMAGNETIC WAVES ALONG WIRES. The modification used by these writers reciprocal relation is important as showing the electric and the and enables the function called the auxiliary magnetic to be in with most dispensed vectorpotential investigations ol which exists between the force. R. If P. 0. The Maxwell equations of the electromagnetic field were first given by in 1865*. this nature. Q.CHAPTER XIII. of electric and magnetic capacity forces in a k. 7 denote the components //<. medium of conductivity k. electric inductive and magnetic inductive capacity the equations referred to are .
one P. •omponents of electric force are also continuous. propagation with a straight wire as guide an isotropic medium. these 143 From may be derived the equations dP_ dQ dy dx ~ dR^^ dz ' ^ ' dx The first dy dz ^ Doint X. rcles The lines of magnetic force are therefore round the wire as axis. that the magnetic at any point in a plane coinciding with the axis is at right igles to that plane.XIII. Therefore there is no component of electric force at for the case of 1 Now jght angles to a plane coinciding >llows 4 with the axis. id consider ' Thus we may choose the axis of x as the axis of symmetry.] PROPAGATION OF ELECTROMAGNETIC WAVES. the field is symmetrical round the wire at very instant. These By a like process we obtain the equations of propagation of lectric force &c. Eliminating and R by means of the first of (2). the second and the third of and (4). y. >f } From the equations given above the equations of propagation an electromagnetic wave can be at once derived. At the surface of separation between two media the normal 'omponents of the magnetic induction. &c. and the tangential comThe tangential )onents of the magnetic force. except course at the origin of the disturbance. ^ and 7. From this it >rce by the equations connecting the forces. z. of these expresses that there is no electrification at the and the second that the magnetic force. i)f the solenoidal condition at every point. i) we get md similarly two equations of the same form for re the equations of propagation of magnetic force. from the axis in a plane passing through We shall denote the distance of the point considered from the . being purely fulfils nductive. are continuous. parallel the axis. and another R. only two components of electric force.
the wire.^^=^+^^^R.144 PEOPAGATION OF ELECTROMAGNETIC [XIII. differential equation satisfied . regions of the field to be considered. > The appropriate A: \ = 0. and is surrounded at a distance b} a coaxial conducting tube which may be supposed to extend tc There will therefore be three infinity in the radial direction. The differential equation. namely. to each region. the outside conduct ing tube. k.  dH dH 8 . found above are perfectly general and apply. and 2 we get for From our special case the equations r» ^. ap and d'P dP ^''^^dt^^f'W=d^^dp^'pTp' Eliminating H .k^+. we easily find in the same way that H satisfies differential equation precisely the a same as ii. fi. If the electric and magnetic each will forces be simply periodic witl respect to x and t. thus correspond to the /3 of equations i and 2. . . o Taking suppose putting it first filled the space between the two conductors we shal with a perfectly insulating isotropic substance differential equations are therefore obtained b.. and the space between them. In dealing with the problem we shall suppose at first that the wire has a certain finite radius. Finally. ^^f. and its distance from the axis by p. origin along the axis shall use for the which will I by x. and magnetic force at the same point the symbol H.kR + 7 . P we see that R must be taken so as to satisfy a slightly different equation.=^^ dp' ~ ^ ^_dR_dP dt da^ Eliminating first ^ for the H and R from these equations we find by P d'P d'P 1 . be of the form . dR d'R d'R d'R IdR 1 . in lo and 1 1. with proper values the quantities k.
find Idu . 145 Let p^^Q(mxnt)x^ Jl = ^Q(mxnt)i^ where u. In precisely the same way we get from 1 1 the equation the differential equation of the Bessel function of order 1. been proved to be that of light. medium of capacities k. namely An equation of the same form as 14 is obtained in a similar manner for H. so that we may neglect the displacement currents represented by the second terms on the left in equations i.] WAVES ALONG WIRES. Substituting in lo.Xlll. If we denote m^ — k/jlti^ by p^ and write f 1 for ppi. remembering that k d^u = 0. We . Turning now to the conductors we suppose that in them k 1 is small in comparison with k. /jl is according to theory Jl/Kfi and this velocity has. we . 1 2 becomes d^u du which is the differential equation of the Bessel function of zero order J^ (f ). The velocity of propagation of an electromagnetic disturbance in a . 10 . In ordinary conductors K/k is about 10~^' in order of magnitude. We shall denote this by q^ and an. If the the dielectric m^ wave were not controlled by the wire we should have in — k/jl7i^ = 0.^irk^ni substituting m^ write the equations thus obtain the proper differential equations by for p^. v denote the values of f(p) for these two quantities. The quantity m^ — essentially real being 27r times the frequency of the vibration. „ K/jun^ is in general complex since mi includes a real factor which gives the alteration of amplitude with distance On the other hand n is travelled by the wave along the wire. for air at least.
since 77 there vanishes. t^j. is very It is clear from the value of Y^ix) given at p.(?) + 6F<. that this condition can only be a = 6 (7 it — log 2). B. Two values of q and 97 will be required. the wire. . C. ^ = {^^0 (?) + CFo (f e(«^^»^) ^ 19 = 20 P i) {(7 log 2) Jo (^2) + Y. is positive. We thus get for the value of in the three regions. respectively. 17 M = a/. The solution of 1 5 has of course the same form. drj^ rjdrj \ rf I where 77 = qpi = pi Nw? — ^irh^ini. For very large values of 7) both Jo ij]) and Fo (97) become infinite.().146 PROPAGATION OF ELECTROMAGNETIC [XIII. e<^"*> \ which A. '?72. for large values of fulfilled if the argument. where 7 of the is Euler's constant. p. with 97 substituted for ^. and therefore 77. and the outer conductor. 11. for small values of the argument Fo becomes very so that in the wire gi^eat. D are constants to be determined by means of )) (. g'2. The general solution of 13 is as we have seen above. we must put 6 = 0. the dielectric. one for the wire and the other for the outer conductor. 40 above great. we shall denote these by ^i. and we have to choose the arbitrary constants so that u may vanish in the outer conductor when p. where a and h are arbitrary constants to be determined to suit the conditions of the problem. being understood that the real part argument Again.72)1 P in the conditions which hold at the surfaces of separation between the adjacent regions. the equations 18 P = ^J"o(%)e<"^^*>*.
we get from we put k = 0. periodic 8 and 9 may be written in the form (4!7rk From the P we R — Kui) R = miH — Eliminating first LuniH = miR — ^rdp . H.{^) + CY'm — ?. 10—2 .] WAVES ALONG WIRES...m]. namely that the and the tangential magnetic force are it From the latter condition circles follows that the lines of I magnetic force.. for m^ — fjuKti^. writing. value of 147 can easily obtain the at component Since all the quantities are right angles to the axis. e(— ^) ^ 24 In the wire on the other hand where q^pi [q^ = m^ — 47ryLtA. we have i?= ^ = ^J^(9. as at p. . 91. q^drj^ q^ ^2 dv2 q2 We i now introduce the boundary electric tangential continuous. being round the axis of the wire in the dielectric. gi 25 ^^_47r^8P^_4^ ^ 9^1 qi ^^ Lastly in the outer conductor G'ofe) for we have. _ — ^Kn^ — ^^TTixkni dp ^ — Kui) dP {^7rk m^ — 4f7rfjLkni — fiKU^ dp * ' Thus if in the dielectric 19. {(7log2)Jo(^2)+Fo(7.0e<^^"*>^ _. then R. = j [BJ'. are so also in the wire and also in the outer conductor. between these equations we obtain m^ jj.i.)l. force conditions. and write ^^ remembering that ppi= f.XIII.
a^. 0. Also if ag the radius of the insulating cylinder.M Fo(a  ^ 'cnqfi.(?)=!. Yo{^) for p = ch. o^ag^ in comparison with others involving the factor aia^. (?) + =0F„'(?)} ^' i)G. a^.. 30 Denoting the values of Jo(^). 5^o(?)i. are small the approximate values of the functions at the cylindrical boundaries are /oW=i.fcnq^Go M Jq ( i\ ' 4>7rk. small. find after a little reduction we . is moderately see that is p reduces to m where X . for the ratios and «!. we find 4<7rk2piGo (V2) M^)2 . the surface of the wire give for and for p = (i^ ^ W„' A.. pa^ is also small.(f) = iog?. and /? = a2 by *^o(f)i.and neglecting JoMl^oiViX (^o(v2)IOqM. . r. and the real part of m^ is 47rYV the wavelength. and eliminating the four constants D by means of the equations just written down. factors putting for brevity <^.2.) Y^{^% Considering that KjuL first is 1/V^ where long waves of low frequency and remembering is the velocity of light in the dielectric. Now when /. terms involving the Oi^aa. Thus if ai is not large pai is very small. fla for ^ttA^i. Using these values for the J and Y functions in equation 31. (?„« = . Jo(^)2> ^0(^)2.log ^''.iv. 47rZ.^. V we nearly. B.piG. w.148 PROPAGATION OF ELECTROMAGNETIC These conditions expressed for fxill.
may be neglected in comsmall and frequency and second. of the very great. and and it is we get or.\ . so k and for The first of the other two terms within the brackets is small values of ai. rr>?=tjl + „—J— . unless the frequency. 35 . since p'^ = mP — Kfin. the values of cj). «2/ a.^. Hence the modulus of the vibrations is third term. are large. ^ ' H'V^ VStt Wk^ «i \/k.] WAVES ALONG WIRES.XIII. is large in comparison with that of the The same thing can be proved of the second term first Hence the third term. on the supposition of low values of aj. a^. ^. Equation 32 thus reduces to parison with the first and the third. very we may neglect the second term in comparison with the first. term. all 149 In j cases which occur in practice it may be assumed that last that the approximately ^irixhi. nj^ir. term within the brackets in the preceding expression is ^is Further Kti 1 = n/fiV^. clear that the second term of 33 bears to the first only a In this case then small ratio unless 02 be very great indeed.2. a. J ^^ Let now the frequency be so small that q^a^ iThen we have is very small.= m^ — n^j V^.
This corresponds to a wave travelling with velocity \/2n/V'rc. r c 1 the the capacity of LUU = I/ttOi^Aji. c = /c/(2 log a^/oi) (where k is taken in electromagnetic ^^ units). Thus we write instead ^_ n i 1 But if cable.150 PKOPAGATION OF ELECTROMAGNETIC [XIII. and with increasing amplitude. each r be the resistance of the wire. and hence th( damped relative strengths of the notes of the piece would be altered. and hence we we used only the approximation to of 35 real part. and having its in travelling amplitude damped down a distance ^^Nnrc. and the result agrees with that founc We by a direct solution of this simple case of the general problem. the higher notes of i root square of music would be transmitted faster than the lower. is great in take only the imaginary part of w? as given by 35 we shall get a value of m. velocity of phase propagation being proportional to tht of the frequency of vibration. that is we shall make the first m which 35 affords. have thus fallen upon the ordinary case of slow signalling a submarine or telephone cable. and taken per unit of length. . higher notes being weakened relatively to the lower. It is therefore left out of account. V2 I 36' taking the positive sign. the real part of which is great in comparison with that which we should obtain if The modulus of the second term in the brackets if comparison with unity. anc piece The the harmony might if the distance were great enough be disturbec Further these higher notes are more rapidh out with distance travelled than the lower. and we have = 7^ ^/nrc. th( from this cause. in which the electromagnetic along induction may be neglected. to l/e of its initial amount The other root of m^ would give a wave travelling with the same speed but in the opposite direction.
very is the same over any cross1> ^^d the value of P section of the wire. We now find the electric is motive intensity in the wire and magnetic given by The electro vvhere 771 = qipi. 40 therefore vanishes at the axis of the wire.^ 7o^ki l^ / 'nrc.] WAVES ALONG WIRES. so that R= —1= realizing y^rp'\/nrce^'^^~''^^^^. radial electromotive intensity in the wire is given But ^ 4/0X^1) = . there along the axis. JoC'?!) But = the suffix denoting that p is less than the radius of if the wire is.rf).. as we here suppose it to be. wire at Hence if 70 denote the = 0. 39 Again we find _ ^jrvrc * 2 R = ijo'^p sinrc e R intensity is I qqs / ^ /five /^ _ ^^ _ TTN j . when ^ total current in the and therefore Hence realizing we obtain P= The which is ^„. by 38 25. and at the surface the ratio of amplitude to that of P is Jaj "Jnrc. the wire. The magnetic above] force in the wire is given by the equation [26 which by what has gone before reduces to I . we have = x the plane 0.Vf»cos(y'p^. its Elsewhere and the electromotive R is sensible.S:iII. thin. can 151 forces.
log ^) of which the realized form is ^ P = 7or ^1 .pe 2 cos [^^ jwntj.152 PROPAGATION OF ELECTROMAGNETIC [xill The realized form of this is H=. Hence 6j(»"a.2fic V' log e "^^'^ cos I) L/"^ x nt^ .2flC V. fulfils the required the Hence results are so far verified.— By ^ i. we find by the approximate We Pop for values of the functions given at p. where r denotes Thus before the resistance of the wire per unit length.Clog. and at time ^ = 0.ne) i^ P = fy^r (1 . 5 + (71og(^xtii) = 7or. Hence subtracting the former equation we Pop find = yor. 4^ a wellknown theorem we ought to have numerically 27rpH= where 7 is 47r —^ 7.. G can be found from 29 by putting (by 18) A = y^r and elimi nating B. Thus we obtain C= 270 /xcrF^ very approximately. But at the surface of the wire Pop = 7o^. the total current at any crosssection. 45 . For 7 we have here the equation y = ry^e^rnxnt)i^ which when compared with 42 obviously tion numerically. Putting the electromotive intensity at distance p from the axis in the plane ^ = 0. ^^ rela shall now calculate the forces in the dielectric. 148 above B + Clog(ppi) = Po. a.
xJq/Jq) by relations proved at p. important memoirs on the same Vols.. i. 49 completed for slow vibrations in a cable of So far we have followed with certain modifications the analysis Thomson. he reader may refer for details of other applications to Elec rical Oscillations. To that work f Prof.^/f ^ cos (^^ n^) .WAVES ALONG WIRES. as set forth in his Recent Researches in and Magnetism (the Supplementary Volume to his 'electricity Edition of Maxwell's Electricity and Magnetism). subject. i^ r 153 and '^6' y since p— m nearly. and selfinductance in the case of a cable carrying and which is also useful in other pplications when)ai. . ^« Thus the solution mall radius. is ^= . J. 47 Finally from 24 and 36 we have 48 Iir=2^e<"^^^^«>. 13 above Jo U=Xy. .2 . 44. q^a^./^e cos (^^ — wnt + ^y . Electrical of xJq{x)IJq{x) in ascending in of use the discussion of the effective owers of x which will be shall We now obtain an expansion jsistance ipidly alternating currents. we have by the Denoting the function xJo(x)/Jq(x) (or. J. q^a^ are not very small. we get 46 ^ = (1 + /^ F^7o a/^ \ n p e^'"^'^ \ retaining only the real part R= ^ 2 '2fiV'y. for brevity. 23. and ii. passim. lence "7. Keference should also be made to Mr Oliver leaviside's ^apers.
. lOoio = Gdg Hence a4 = 1 x" .154 Therefore PROPAGATION OF ELECTROMAGNETIC [xil ^^ = ^0^ + 2) + Now zt *l 5 guided by the value in brackets in 50 we assume =— 2+ ..)('2 + ' + . Sag = — 2aio + io^s + 2a4a6.. — 2as + ha^ + a^.)=^2^(^ + a4a.V Multiplying these expressions out and equating coefficients w find = — 2a6 +Ja4.^+.... . Then by a?( 5 1 + 4a4«^+6aea^+.. + a^cc!^\ aaX^\ .
following table of values 4i7rfi^n(h% is 155 J. given by Prof. J. Thomson : .WAVES ALONG WIRES.
p. In the same circumstances and further V when the modulus of zirx ^ x is very great. e~^ 5 modulus of w be very large. semiconvergent expansion = and th Mv)_. any value of x with ret 90 = i" A / 9. if cos (/i . for p. from the relation Jn(ioo) i''^ In(x) (139. ' Joiv)~ when 7} = iw. 68.COS rnr by 142. p. 66). 143. Similarly = ^''~^ a/ ~ approximately. we may write by Again that. and a? is a complex quantity of which the and of which the modulus is large.1) TT . Gn(ix)___ Thus we get of which Go{v)l(^o(v) = h is a particular case. that approximately 68. if the . e* 5 x have a very large modulus. positive.] . p. it is real part i worth noticing part positive.156 PROPAGATION OF ELECTROMAGNETIC [xil prove. Jn{ix) Thus we have the result _ 5 JUix)' already stated above for n = 0.
5 m in the case of a cable will surrounded by sea water. k . If the value of P at = the surface of the wire be denoted by P^^. Everything is supposed symmetrical about the axis of By t 1 8 we have for the axially directed electromotive intensity a point in the wire distant p{= 'r^jiq) from the axis P = ^Jo(7. 6i Pa^ tti).i)e(^^"*>*. gives an ex ression for the current density parallel to the axis of the wire at istance p from the axis. lS 157 out the case of a further illustration the reader rapid that both q^Ui and may work ^cillations so q^ai are very large. le wire. The damping in this case is is of much smaller modulus ow. and nothing will be ained by using copper in preference to an inferior metal. = ^Jo(%)e<"**^*>*. is the product of n by the reciprocal of the velocity of )efficient of i within the brackets. its propagation is thus V. Here if aik.WAVES ALONG WIRES. mductor shall now calculate the current density at different distances the axis in a wire carrying a simply periodic current. 'his 6o multiplied by h^. and the rfective resistance and selfinductance of a given length I of the We om mductor. the outside mainly control the damping.2 be small compared with a^k^. Here fat by 33 ^ V^ {qj^Oa qM 1 ^ Iras m id approximately m The f ^ ^ JM'Kn \y aA V aSJ j^^ "^ ' 59 avelled. and the distance while the amplitude is diminishing to the fraction 1/e original value. the conductivity of the wire. since the imaginary part of lan the real part.
and will therefore be determined if But since the induced electromotive find it for the surface. and so {p . 12 . Now the impress( by subtracting from electromotive force is the same all over any crosssection of tl P P wire at a given instant. the total current in the wire we ha The electromotive intensity is the resultant parallel to tl axis of the impressed and induced electromotive intensities. Thus parallel to the where A'T (A' induced intensi E=A U(^i) . denote tl surface ^ impressed electromotive intensity = a.n ^^ AVi^l we get e<"^^^) *• Putting r for the resistance (= llira^^k^ of unit length of t wire and using the expansion above. \ i tensity due to any part of the current is directly proportional to i timerate of variation. constant) is put for the axis at the surface.^.^^n?P ^"1"^^ +'^(248 E^ + 8640 TF~—Jr M „ . [ solve the problem proposed we must separate the part impress* the induced part.. the induced electromotive intensity at tl must be directly proportional to the timerate of variati* Hence by 6i if of the whole current in the wire.(248 Or taking the impressed two ends of a length we have I V R' +8640 VJrV difference of potential between t of the wire the resistance of which is *^~^V . R' fl 180 +"j^ 13 ui^n'i' f 1 u.158 PROPAGATION OF ELECTROMAGNETIC [x] The magnetic force at the surface is = Oi)' Therefore if V be 47rr = — ^irajla. since q^ =— ^ir^k^ V 12 r^ 180 r' •^'"J^ '""r^ +/.
If there be no dislacement current in the dielectric comparable with that in the nre. The inductive electromotive force er unit length in the conductor at any point is then equal to the xte of variation of the surface integral of magnetic force taken er unit length in the dielectric at that place. magnetic return current be in the following [' manner.(^^ + ___/^_..^^^^^^' AJ^ M log ^ e(^) \ is 67 ut this last expression ' by what has been stated above AT.' = 21og^ 68 ad X=2Zlog+Z. oint. Z X\ F. a supposition sufficiently nearly in accordance with the fact )r all capable of practical purposes. L' we get F=ET + ZT. ^Bius f 66 Bl and L' are the effective resistance and selfinductance I the length It of the wire... then by 63 we easily find ^. if there is will be in circles round the axis of displacement current. If 159 we denote the by series in brackets in the first and second terms espectively R'. dXjdx. and is given by 62. IfJoix'Jiybe denoted by Yi. Now. since 27rrir=47rr. dYjdy by a. . = 2 \l ^L^nfr. 'hus if Hr be the magnetic force at distance r from the axis of ae wire ^1 r ad ['ilrdv = .. and will be inversely as the radius of the circle at any H le wire.j. Thus we obtain J. ad x XY'X 'Y^ . ielectric. remains to determine the constant A'.WAVES ALONG WIRES. and the skin on the outside of the as in a highly conducting eing regarded force is no we can there find that so outside.
1892. in an unlimited dielectri* medium.. . The problem of electrical oscillations has been treated some what differently by Hertz in his various memoirs written ii connection with his very remarkable experimental researches* He discussed first the propagation. London. and is now insensible except in ar L stratum at the surface of the wire. and. at Equation 65 shows the effect of jjlj on R' and L' at different If however the frequency be very great. or Electric Waves (the English Translation of the sai die See Hertz's work. Barth.160 PROPAGATION OF ELECTROMAGNETIC [XIII. we may take th vibrator as an electric doublet. Jones). an ( the origin midway between them. Taking the simple. between the two conductors. ilber I Untersuchungen Ausbreitung der elektrischen Kraj J. and that the The equations * of motion are those given on p. 142 above. put in 63 Jo (rjOl^o (Vi) = '^* We find for this case Thus in the limit R' is constant term infinitely thin — lA' . that is as consisting electrically two equal and opposite pointcharges at an infinitesimal distanc apart. case first as an introduction to the secon( which we wish to give some account of here. a form in which the values of R' and L' are easily calculated for any given values of x and n from the Table of Jo(x\/i) given the end of the book. E. secondly. of electric and magnetic disturbances from a vibrato consisting of two equal plates or balls connected by a straigh wire with a sparkgap in the middle. force are circles round the wire. everything symmetrical about the axis of that the electr lines of magnet: forces lie in planes through the axis. is It is clear in this case tht x. A. Macmillan and Co. set up by an initially impressed difference of potentia guided by a long straight wire. The current reduces to the indefinitely great. and having the line joining them along the axis of x. the propagatioi in the same medium of disturbances generated by such a vibrato The action of the vibrator simpl consisted in a flow of electricity alternately from one plate or bal to the other. by Mr D. we must frequencies. Leipzig. 1893.
z. Hertz's notation we take this function as dU/dt. a of magnetic force in the 161 syinmetiy the component and the equation medium is dy dz holds. This shows that is a complete differential of some function of In y. 8dz — ydy connecting the other two components. so that 1 .m I WAVES ALONG WIRES.
z. Here p~ = y. from whicl it may be regarded as built up. We H TT = ^ ^'^ mpWe 7^ 77 of this solution are very interesting but. as they do not involve any applications of Bessel functions. we must in calculating them from TI use the formulae Kfi \dx^ ^^ P docdp take the meridian plane as plane of os.+ z^. p. we have referred to them do not consider them in detail. foi the solution of which the use of Bessel functions is requisite. so that the which is at right angles to the meridian plane is magnetic force Thus identical with yS. and hence if we put now P and i? for the axial and radial components of electric force. Thus we write It is easy to see that we may put f{x. From by becomes this solution the electric differentiation. ^) without affecting the electric and magnetic fields. PROPAGATION OF ELECTROMAGNETIC [xill. The fully worked out results inasmuch as the case of the propagation of waves along a wire.162 only. and of the electric doublet. where r <I> 74 is is the maximum moment the distance of the point considered from the origin. and the equation of propagation is = A solution adapted to the vibrator we have supposed is n = — sin {mr — nt). and magnetic forces are found In cylindrical coordinates x. dt" /an "^ a^ '^ i8n\ dp' dp) since 11 is independent of 0. the surface of 11 at each point the conductor a simpK In the problem of the wire we have the medium close to . 6 the equation 8^^ J. ma} be very instructively compared with this simple case.
e\ or J 00 cospf d^ . ^If we exclude any damping out of see that or ^. 78 Substitution in the differential equation which holds for the medium ay 18/ dp^^^ptp^"^' nV)/=0. We denote m^ — u^k/jl by p^ and suppose that p^ is positive. p. that is is less that the velocity of propagation tion in the dielectric. We have therefore in the insulating medium n = 2(76^0 (w) sin {mx nt^ e). [ence at any point just outside the surface I we of p. 67 above. as where pp is real. n = 2C r^^L^ d^ + hjp' i' f+QO sin (mx nt\. sin (ma? —nt + e). than that of free propaga We have therefore instead of 79 dp'^pdp This is satisfied ^J '^' by Jo iwp) ^^^ ^J ^0 (w) The latter solution only is applicable outside the wire. 81 h 11—2 . and is infinitely extended in at least one way so that there is no reflection to be taken into account.^ sin (mx — + e). get . We shall suppose that the wire is very thin and lies along the axis of x.WAVES ALONG WIRES. Putting p sinh (^ = f we . are constants. 163 harmonic function of the distance of the point from a chosen origin. n = ^ sin {trx — nt\A cannot involve a? e). the wave or change of form it is therefore a function Thus n =f{p) sin (mo) — nt^ gives for /the equation e). 79 shall Here if/m.is the square of the velocity of propagation. / must be 80 zero at infinity. this solution may be written nt = 2C •! cos {pp sinh (f>) defy . where G and e Now by IT 140.
82 p=0. may be compared it is with that obtained above. that is is if the velocity of propagation is that of light the solution Yl = C log p sin (mx — nt+ e). as may very easily be seen. cases the wave at any instant in the wire may be dividec up into half wavelengths. we have so that neglecting the imaginary part we have <?. 74. and each curve then consists of a pair h lines. such that for each lines of force start oui from the wire and return in closed curves which do not intersect and are symmetrically arranged round the wire. When ipp is small. to the wire. the electric force. 162. from which of course capable of being derived. When p = normal parallel 0. one passing out straight to infinity. This result [XIII.log ^^ sin (mx nt+ If e). . p. the othe returning to the wire. The direction the force in the curves is reversed for each successive halfwave.(w) = (7 + logf). Hence at the surface of the wire U =2gU\.164 PROPAGATION OF ELECTROMAGNETIC WAVES. 83 for this particulai as may easily be verified by solving directly In all case.
and consider the portion of the wavefront of radius a which fills the orifice. a sin 6ddd(h p — — sm (mf  ^^ nt). ^1. Let the distance of any point of the edge of the orifice from the source be a. and its longitude be (b. a secondary source of light. n— ^ttJT. and at a fixed distance from the latter. regarding the element as formula. The problem in spherical here considered is the diffraction produced by a small circular opening in a screen falls light propagated waves from a point source.CHAPTER XIV. If the angular polar distance of an element of this part of the wavefront be 6. X and T being the length and period of the wave. where m = 27r/\. and using the ordinary fundamental we obtain for the disturbance (displacement or velocity of an ether particle) produced at P by • this source the expression . and it is required to find the intensity of illumination on which at any point P of a plane screen parallel to the plane of the opening. . P. Case of Symmetry round an Axis. Thus. We take as the axis of symmetry the line drawn from the source to the centre of the opening. the area of the element may be written a^ sin 0d6d(j). Putting f for at the distance of this element from the point. «. DIFFRACTION OF LIGHT. if the angular polar distance of the edge of the orifice be the whole disturbance at is P a  r27r rOi 1 T^ sin ^ sin L XJo Jo g (m^  nt) dddcj). of the screen which the illumination is to be found.
. Then the distance P f 2 = {J + is ct (1 _ cos B)Y + (a sin ^  ^cos <f>f + ^ sin^ <^.166 DIFFRACTION OF LIGHT. and denoting them by tn. and the distance of the screen from the spherical wave of radius a. fj write the expression in the form X r27r ab\.F'=°^* + 2^'''. we obtain for the total disturbance the expression Separating now those terms of the argument within the large brackets which do not depend upon p from the others. Hence finally if the opening be of so small radius r. h P from the axis of symmetry. so that •07 = ^(6 we may rr + )_„i. from the element is a (1 — cos 6) + 6. and P be so near the axis that we may substitute 1/6 for the factor 1/f. Because of the symmetry of the illumination we may suppose without loss of generality that the longitude of the point is given by f from the element to the nearest point or pole of so that the distance of the screen P is zero. If now we sin'id = p^l4<a% write p for a sin so that 6. we have approximately a+b ^ ?P ^ *+2l. or 7— (C sin 'ST + S cos ot). Let f be the distance of [XIV. Since 6 small this reduces to 2 or i = h+ —\ ^ sm2  _ > sm ^ cos or a6. </) + 1^ .
= r27r I cos Jo ( Ji/r — a. k. . cos (^) d<f) Jo 27r cos ^yjr Jq (x). Ahh.] DIFFKACTION OB^ LIGHT. 27r ? we have p27r fa+b . r2iT I sin (« cos cf)) d<j> = 0. This can be done by the following process due to Lommel * depending and it upon the properties of Bessel Functions. Hence where j denotes the value of x when p * = r. 167 where ^= The r/o '^" ¥ (Sr ''^ p °°« a) p'^'"^* intensity of illumination at P is thus proportional to only remains to calculate the integrals G and S. d. xv. Akad. 1886. ? . by 44. (a.\ . .. cos <^) (a. C^(/) = cos ^ilr since sin ^yjr Jo I cos cos 6) d(l>. Bayer.xrv. Wissensch. 18 above. p. d. But cos ^yfr Jo I cos {x cos 0) cZ<^ = = 2 cos Ji/r I COS . Changing the order of integration in G we have Now considering the inner integral and writing 2'Tra +b ^ .
13 above. xJi (x) The same process may now be repeated on the integral of the second term on the right and so on. value of ^i/r when x=^z. we get and hence by integration Jo / X^Jni{x)dx Integrating by parts and using this result cos ^slr . [XIV. putting 47r2fVV^^X^ for z\ can thus be found by evaluating the series C/j. The values G and S The series t/j. This can be done easily the tables of Functions numerical Bessel by given at the end of of this volume. and writing Thus putting ^y for the we obtain finally. by x^. Series proceeding by ascending powers of zjy can easily be found by a process similar to that used above. First multiplying 19. 2 These integrals can be expanded in series of Bessel Functions in the following manner. and then continuing . U<^ proceed by ascending powers of yjz. = X''Jn{x). . p. We begin by performiog the partial integration first upon cos Ji/r xdx. U^ for the given value of z. xJq (x) dx = cos ^yjr . we can show that S= 62\2 rz 1 g—^2 sin Ja/t . and rearranging we obtain X^Jn^ {x) = nx^'^Jn {x) + ^V^ {x). ocJo {oc) dx.168 Similarly DIFFRACTION OF LIGHT.
I . Similarly we obtain I? cos 1^ S = Trr^ b 2y .=2(l)''gJV^(^).VIV. p. 169 the process. making use of the equation ^{x^Jn{x))=XJn+. .. 12/ 2y i2/ i2/ ) where Fo = J. 13. ' which is in fact the relation ^n {X) = Jn {x) . 1 ~^i (^) sin fix"^ .Jn+i (x).] DIFFRACTION OF LIGHT. sin fjLZ^ . G= ^ —^ I Jo (x) cos { f xdx 'Jo b'\^ < 27rf2 ^ 1 . xdx ''" 2W' Ik '"'^ "^° ^^^ ~ 4^' "^ •^' ^^^ °°' ''^' 1 1 Proceeding in this way we obtain + 2 2/ \2y 3! l2^. Vh stated in i6. Jo (^) + 9~ 1 PI I .°°^ ^y F.{x). above.(^)(J)V. Thus remembering that we have as the first step in the process /JLX'^ ..(^)+. +•••].'^iy y V.
29 above we put x = n we get l=Jl{z) 2Jl{z) + 2Jl(z)\. \ where n may be any positive or negative First all of all it z. Ux sin Jy — U^ cos I?/ = cos ^r Vq cos J?/ V^ sin J^/.170 DIFFRACTION OF LIGHT. integer. 'I .(^) ) + . Jn^... = sinj(y + J). Consider the more general functions Un = (f)" Jn (^) (f )"^' /»« (^) + • • • = 2(l)i>g)""'*/„+^(^).+ F. adopting however following somewhat different analysis. Squaring 4 and [/"a 4 Fo = cos i U+\. these of short discussion properties. [XI Comparing 4 and £/i 5 with 6 and 8 we get cos \y + 6^2 sin Jy = sin z — + F^ sin \ij  Fj cos \y.p{^). which give [^. F„ = (J)" J„(^) g)''^V„+. 5 and 6 and 8 we obtain equivalent expressic on the scree the intensity of illumination at the point = 1 if irr^ thus for P '^i The calculation of these ZJand F functions by We means of tab^ of Bessel Functions will be facilitated by taking advantage follow Lommel in t certain properties which they possess. is clear that the series are convergent if in values of y and Now + 75 p.
. SSa s generally m„ = F. f7.^^.i(^). find cosz= Jo(z)2J^(z) + 2J...{z)  ..2Js(z) + 2J.+. than the geometric series more is for Un convergent icries ^Sc^nfly < 1. putting the second and first of 4 successively. 171 Hence we see that since J. 13 U^. y = z But putting in 42 and 43. U^— Fi = J sin z.DIFFRACTION OF LIGHT.. sin ^ J" S"\we Returning now to 1 1 and 1 2 easily find Un+Un^. sin^ Therefore p.1)"+^ J^ l)»+i> J.)I Un + Un+.). . = (ir(^n+^n+. 18 above. = = tiZ V. (z) («).4(<^) = (— l)'*/. when z =y Uo=Vo = i{Jo{2)\COSz}. we . = t^' ^ [ /„ (^) + cos z] p= T\. Vn is more convergent than the geometric series It is therefore in more convenient in the former case to use Z7„.. the latter to use If Vn for purposes of calculation.+. Fn+F„+.) = 2/^" ( Vn + 5 Also since /_. = and therefore Z^'.(z) = 2J^(z) .).. and also z^y 1 —n for n in Un + f^n+2 =(!)"( F_n + F_. </> = 0. (z) < 1 each of the other Bessel It follows that if y/z < 1 the Injunctions must be less than 1/V2.. = i[J"oWcos^l..= F.^.. we find.= (^^Jjn(zl 14 Vn^Vn+.( (~Jjn(z) F.
we get But by 16 and 18 this becomes ^2+Fo = cosi^2/+ V Dififerentiating again we obtain or by 16 and i8 « . _z d^ i. This gives by successive differentiation the equation d'^Un __ m1 3"^^ ^j.. p..w+. Similarly [19. Using in the second above] z line of this result the relation [i6. p.^... Differentiating ii.(«)(0'". [XIV.172 DIFFRACTION OF LIGHT. 13 we get =f dz QJ ^'^^^ ^'^ "^ (f)"^' ^^^^^ ^'^ l6 =  Un+i. we find +(f)". we obtain by differentiating 12 and using the relation 13 above] J'n {2) = Jn1 (^) Jn (^) Z dz ^^^^^' and therefore Again differentiating the first of 9.'.
for which y = z. 13 above. {y. z + h)=U„ + h^^ + ^^?'^+ . and then deducing the others by ^nij ^n2.ly^ sin i (2/ + ^) ' 20 If in = (..««. z + h). Vni... cases.. those namely. nJn (z) = ^Z Jni {z) + ^Z Jn+i (2). we have By Taylor's theorem U^y. Un+2..2 + n)z pi(^2yy '^^' ^ These expansions are highly convergent and permit of easy calculation of ^^n+s> •• . z + h). Differentiating 1 1 and 1 2 with respect to y. repeating this process ^2n+i it is 173 By clear that we shall obtain + ^2n+i = (.TV.«"»t(s+j) which include the equations las particular = 2/1 (z) . cos z= Jo (z) — 2/2 (^) + 2/4 (z)— .. by using 16 and 18 to calculate Un+i.ir COS i f^2n+2 + V_^ (2/ + I) flatting in 0. The functions Un+i. we fall back upon 9. equations we put 9 the values of the functions as given in the defining equations 11 and 12 we obtain the theorems these n= 21 + S(l)'j(f)*" sin ^ ()''*")j. by means Calculating the successive differential coefficients of 16.'. p.. F„ ••• . and using in the resulting expressions the relation 20.] DIFFRACTION OF LIGHT. . Un {y..2Js(z) + 2J. Can be found from Un. and rearranging the terms we obtain n+2 P(2z + h) P = %{iy h "}rZZ u„.. Vn. successive applications of 14. namely. (z) . ~pU2yy Similarly we can prove that VnKy.
and application we obtain d^h = i{c'Un.2U.174 DIFFRACTION OF LIGHT. [xiv. of this result By successive differentiation.n + 3 ^ n+2 dz^ and generally ) d^Un dz' m+2p' 27 .. we find i.„ by 16 and 24 above. = y^^^ifU„ 24 Now if It be a function of y we have Using of 24 this theorem we find by successive differentiation 9^ 25 26 If we consider 2/ as a function of z then ' dz If dz dy dz y ==cz f^ = i(cfr„_.^Un^.lr.
i^c^w =y Ui 32 Jo 1: (s^^O • sin ^2/ (1 . 29 V') conclude this analytical discussion with some theorems in wliich definite integrals involving Bessel Functions are expressed in We terms of the U and V functions.v. 31 But equations 4 and (7 cos 5 give iy + S sin ^y = ^ Trv" irr' Ui. ^n J. .a7Jo(^)^^.. rz By I above we have 52^2 ^^27rt'J cos^>/r. "Z^t^ = 112 if . Jo Jo (zu) cos ^y (1 — u^) — W) . . 30 Similarly we obtain B = 27rr2 J j sin {^yi^) nJ^ (zic) du. G sin iy—S cos l^/ = j— and these by 30 and 3 1 f/g.+ F. Now let X = 2u. then therefore since z"^ = ^ir^^r"jXJf G — lirr^ •I I cos i^yu^) . it 175 Similarly can be shown that dz' The )ut calculation of the differential coefficients can be carried by these formulae with the assistance of 14 which now become F.] DIFFRACTION OF LIGHT. »i+2 ~ 1 ^n . give the equations . vJq (zu) du.
1 sin (pr^) Fj ^ .u") . we have instead 35.( ) Un+l Thus for ?i if + 1.COS ^ y ^y ^ Vo siu iv ^ rr Fj. the theorem 33 hold for any integral value of w it holds But as we have seen above it holds for n = 1. and write y \ use the variable p (= ?(r) where r is th( = kr^.176 DIFFRACTION OF LIGHT. it therefore integral values of n. I Ji (zu) sin \y{\ 2 — u'^). u^du = — t/ U^. u^^^du =.C0siy(lw').u''du=(^ and differentiate. Un 33 making use of the (2:11) relation Jni =n — ZU 1 Jni {zy) .Jn {ZU). since [XIV I and similarly J I (zu) cos ^2/ (1 — ti^) . Differentiating with respect to z we get. Jo ^+ 2/22/2/ 3 1 1 z^ sin Jv ~ Tr cos Vo \y ^ ^^ V^. z = Ir. Now if we assume ' / Jn^i2u).u^) W^^du =  y \y/ [~\ Un. u^du = '[ 7)TT ^ 1 . we easily obtain I Jn {zu) cos \y (1 . holds for all Similarly we obtain Jo \ Jn^ {zu) sin iy (1 .36 P gill ZifC J a I' Jo (^/>) cos (P/)2) p^p == IC + \ sin (pr2) fC V^t cos (JAr^) Fi 1 rr I Jo (/p) sin (p/90 . r / uJo {zu) 2* 1 Z^ du = .pdp = 1 ^2 1 j^cos^j^ cos (JA^r^) Fo . 34 The r^ values of 1 G r in / 6 and 30 give X cos hyu^ uJo {zu) ^^ . 35 y Similarly those of ^ ^ X in 8 7 and give cos iv Tr / sm ^w^ 1 • o .sm 7 . „ du = . 3^ y y If instead of we radius of the orifice.
:IV. Of this problem there are two cases which = 0. nena. 12 G. and then introducing the orifice between the lens and the screen. When negative the orifice is to be supposed illuminated by light converging to the pointsource. as Lommel realised when the interference phenomena are observed with a spectrometer. When a — — a is a may be either positive or negative. We .pdp = j^ cos ^^ brmulae which will be found useful in what follows. angles to the line joining the source with the centre of the orifice. :^reat distance from the points out. and afterwards pass a on the more general 3ase (2). with b negative. and the screen is there situated with its plane at right This can angles to the axis of symmetry.. on which the light from the orifice. the telescope and collimator of which are adjusted for parallel rays. is positive. ipecial cases of They are more general theorems which can easily be obtained oy successive differentiation. 1Y7 If Tj vanish.] DIFFRACTION OF LIGHT. be realised at once by producing a converging beam of light by means of a convex lens. We to shall consider it here. The orifice is placed between the collimator and the telescope at right angles to the parallel beam produced by the b. When a falls and therefore its on the orifice. is orifice falls is at a very This arrangement. and has received specially problem may be much attention. M. the lightwave front convex towards the direction The interference is then to be considered as on a screen produced passing through the source. . (1) that in which y The first case is that of Fraunhofer's diffraction pheno/anish. which now coincides with the focal plane of the lens. and at right of propagation. 6.tated come now to the application of these results to the above. (2) that in which y does not distinguished. now r be made infinite while and we have roc I and k do not vanish. either = oo and 6 = oo . Fq and j^ Jo ilp) cos {\kp') 'pdp =1 ^ Z^ ^ sin ^^ 37 j^ Jo {Ip) sin (P^2) . When 2/ = 0. or a=is former case the wave incident on the orifice parallel screen In the and the plane. foraier.
magnifying If a magnifying lens is . p. Now when J^ (z)/z is a maximum or a minimum * Camb. writing M' for O+ jSf^ we obtain 38 M^=\IMz)\\ Airy gave* notation. of Tables of Bessel Functions the value of By means be found wdth the greatest ease.source. Putting = in 3. by simply doubling the value of Ji for any given argument.4. The result is shown graphically in the adjoining diagram. in the present \ 274 "^2. Phil. Trajis. Or. . for the same quantity the expression.6. This case can be virtually realised by receiving the light from the opening by an eye focused on the source.4. The maxima J^{z)lz is a maximum of light intensity are at those points for which or a minimum. or through a used the arrangement is equivalent to a telescope focused upon the point. 1834. and the screen in the focal plane of the lens.178 DIFFRACTION OF LIGHT.6 2.8"^ "V M can which is simply the quantity on the right of 38. The minima are those points for which J^ {z) = 0. and dividing the result by the argument. The diffraction pattern is then produced on the retina. [XIV. we have so that.4. with the opening in front of the objectglass. a convex lens may be placed at a greater distance from the source than the principal focal distance of the lens.6. This is Fraunhofer's arrangement and we shall obtain the theory of the phenomena observed The screen may be examined by the naked eye lens. so as to receive the light after having passed the orifice.4. 283. by him if we put 3/ 1/ = in the above theoretical investigation.
p.(^)J. 179 But z 30 that Z^ z the condition becomes which (20.DIFFRACTION OF LIGHT. 13 above) is equivalent to ^J. z 18 .(^)=0.
I) IT. 2. maximum and following Table gives the values of z corresponding to zero values of 22r^Ji_{z).180 is DIFFRACTION OF LIGHT. is (m + f ) tt. [xiv. where or m is the number between is of the root. 3 contains the corresponding values of if I z . approximately a zero and the next following maximum The rings are thus ultimately equidistant. that is in distance The is therefore or wavelengths Xzlir. for maxima minima The difference of path of the rays from opposite extremities is 2r tan~^ of a diameter of the orifice to the point 5/6. which are contained in Col. and the difference is Jtt. zjir. Hence for great values of z the difference the values of z for successive TT. (m + and of Jg (^) — 0. 2rf/6 P ^ The col.
Phil. '062. The curves are drawn with values of ^ as abscissae. with the expansions of Vq C7"i. March. so that the fraction of the whole light The values of this fraction for outside any dark ring is Jl (z). Mag. the successive roots of Ji (z) = 0. equation 3 being used if z>y. and of if2 as ordinates. 1881. so that more than f^ of the whole light is received within •047. the fraction of the total illumination outside any value of z is Jl (z) + J\ {z). or 'Wave Theory of Light. becomes 39 If z is made infinite the expression in the brackets 1. .XIV. and we have and U^ can be calculated by the formulae given above from the Tables of Bessel Functions at the end of the present volume. as has been pointed out by Lord Rayleigh*.'zdz = 4 r z^J\ (z) dz. . are approximately "IGl.] DIFFRACTION OF LIGHT. . But at a dark ring J^ {z) = 0. y is In the more general case of diffraction. . of M^ are given in the Table below for the values of y stated. '090.(z)J[(z)}J. not zero. and 9. 181 The whole portional to light received within a circle of radius z is pro r M. . We also give here some diagrams showing the forms of the intensity curve for the same The maximum and minimum values values of y.. and F. the second dark ring.(z) z Hence J rM^zdz = 2{lJliz)Jliz)}.' Encyc. p. 433. if ^ < y. contemplated by Fresnel. But ^^{J. 9th Edition. ' Hence.. Brit.
182 DIFFRACTION OF LIGHT. 0005490 Max. Min. 10173467 0000269 0001858 11578479 . 4715350 7015587 8306007 . [xiv.0040631 + 0016225 0026305 Min. 2 u. Min. y 3831706 y 0122609 + 0106159 . 3030827 3625773 3831706 7015587 9440724 10173467 .0178789 + 0013239 + 0074093 0002313 0031966 0001826 Max.0043104 Max.
y 0004513 Min. Min.] DIFFRACTION OF LIGHT.:iv. 4431978 7015587 + 0067178 + 0068485 + 0068964 + 0045384 0017711 . Min.0010782 + 0076624 + 0048204 0004806 0004756 0007931 0002637 Max. 10173467 . Max. 183 ^ y 2649454 8831706 = 97r.
This is shown in the next diagram. Lines parallel to the axis of y are drawn for the values of z which satisfy Jj (z) = 0. the tabular value of U^{z) nearest to a zero value is taken. for different values of y. are given by Lommel in his memoir. The axis of ordinates is that of y.184 DIFFRACTION OF LIGHT. is very convergent only a few terms need be \ and the value of h {2z h)/2y found. The squares C/g of these are the values of M^ which correspond to the roots of = 0. [XIV. . We conclude the discussion of this case of diffraction with an account of an interesting graphical method of finding. those of 2UJy for the same arguments are calculated. '^UJy which correspond to these values of z. each accompanied by a graphical representation of the results. It the curves. and the axis of abscissae that of z. and therefore that Values of z which render U^ = 0. These are transcendental curves having 0. as will be seen from the short axis of z Let now the edge of a sheet of paper be kept parallel to the and be moved along the diagram from bottom to top. double points on the axis ^ = discussion below. The short Tables with the illustrative diagrams given above will serve as a specimen. values of ^U^jy. to zero. the values of z which give maxima or minima. for the value of y for that position . These are called the lines /i {z) — 0. TJ^ or Vq. being thus found. V^ for the values of z for which Tables of Bessel Functions are available. with that is from the equation Since the series retained of h. and the value oi z + h which causes U^ to vanish is found by means of the expression on the right of 22 equated 2 put for n. Supposing required to find the roots of U^. will intersect all The distances from the axis of y along the edge of the paper in any of its positions to the points ol intersection are values of z. The formulae of interpolation are 22. are obtained by interpolation from those of Ui. 23 above. On the same diagram are drawn the curves U^/y^ = 0. Elaborate Tables. The are values of z which give a maximum or minimum of illumination. arise The maxima and minima which TJ2 through the vanishing it is of are found in a similar manner.
] DIFFRACTION OF LIGHT. and are therefore values of z which with that or maximum lOTT minimum values of M"^.XIV. 185 which satisfy 41 value of y give . .
whenever cos ^y = 1.. CTi = 1 — cos ^y.zly} zero. tan — ^mir = — V2/. TJ^Iy"^ 0.zly{=J.186 DIFFRACTION OF LIGHT. = sin iy. equations y^ equivalent to U2 = splits into two straight lines coincident with the axis of z and the curve represented by Where the curve meets the I . ^ = ±54°44'8"2. [XIV. so that The value y is of c?y/(?2^ is therefore indeterminate at the points on the of axis of y. that is where the curve meets the axis of y. that is when ^y = 2m7r. If y\ z be current coordinates the equation of the pair of tangents at a double point is It is very easy to verify by differentiation and use of the properties and y = 4m7r. But this value of y makes Ui and likewise U.4^m7ry = 2z\ y' Thus the equations of the tangents are y' — 4m7r = ^2z'. axis of z we may regard CTj = a — = so the two that the curv< 0. C/g = Thus U^ 0. Fi = 0. The curve iJa = therefore meets the axis of y at every multiple of 47r. When z = 0. that is each point in which the curve meets the axis a double point.. or y = 4m7r. so that by 9. by and their inclinations to the axis of z are given that is ^ = ± ^2.(z)^U. Vo = 1. of the functions that when ^r = ' dz" "2' ^' ^ybz dy' so that the equation of the tangents reduces to {y' .
or as /?< or >.ee. within the region of the curve corresponding to the geometrical But at points along a line Ji(z) = 0.UiJi. To settle where the maxima and minima are we have to calculate d^Mydz\ Now d^^\y)d'z^'^^^'^ Thus considering first points upon the lines Jj is = 0. that is also within the region of shadow. On the are other hand. Hence no intersection of a line Ji{z) = the geometrical with the other curves can exist in the region of the diagram correpositive so long as sponding to the geometrical shadow. dy I ind is y^ z>y.] DIFFRACTION OF LIGHT. touch the curves J^ (z) = 0. 185 shall see below that this intersection takes place at . have by 3. that is. It will )f be seen from the curves in the diagram that the value dy/dz is negative so long at least as y <z. Thus the curves U^/y^. Thus the branches of the curve — cut the of abscissae at right angles. y .bove.xis 2/ = 0.IV. shadow. or maxima when U^ — O the points on the curves t/g = minima according as J^ t/3 zjy {= J^ {J^ — Uiz/y)] is  negative or positive. We )oints satisfying vhen y = 0. we have a maximum or a minimum according as Jo ^2 positive or negative. 187 We lence for the last curve dy dz\f^V dy \y^ ^^^' 7 U^ly"^ v^hen . as we shall . the equation J^ {z) = 0. as shown in the diagram on p.
that is wherever a iin satisfying the equations Ji (z) . that is M^ l when z=0. now to the diagram. and so on. from the axis of Jo(z) U^ is ings being 0. and so all points of a portion of a curve /i (z) of the other branches between curve.. Also. The vanish first three successive differential coefficients of 4m7r.2 = for a branch of t curve = UJy'^ changes sign while Jx the value of dy/dz is zero. when the process descril above of using the diagram is carried out. and because of U^ = 0. At these points therefore d^M^/dz^ changes sign.. and y = at the double poir . = 0. z as abscissae. wherever the curve or U'2/y^ = has a maximu minimum Referring ordinate.188 DIFFRACTION OF LIGHT. [XP Calculating d^M^/dz^ we see that this does not vanish for point = 0. = until a branch of the curve U^/y^ = along a line Jj For pa is crosst clearly U^ changes sign. the number of cro. intersect there is a point of inflexio ' \ drawn with M^ as ordinates and values ] It follows by the statement above as to the inclination of tl ] curve within the region corresponding to the geometrical shadow that within that region there can be no point of inflexion on tl intensitycurve. we pass along a curve Uy. a hence they also separate regions of the curve IT^/y^ = which g maxima from those which give minima. z. there are points of inflexion \ ] the intensitycurve. as can easily be verified. since U. Thus Jo{z) = changes sign. U^jy'^ = and cross Jx {z) = 0. 2. . positive for the second. and a curve U^ Ji (z) = of the curve of intensity. If 1. so that Ux is a maximum 01 minimum. tTg = 0. according to the number of branches of the which have been crossed to reach that portion by proceeding J^(z) = latt aloi negative for the fii portion. we can see how to i i dicate the points where there are maxima and minima. p. Us =— it But C/3 and C/i must be further noticed that when but not so U^\ for by 14 when Ji{z) = is a maximum or a minimum. that is also for U^ {z) does not .. th Jx{z) changes sign. maxima. d Ux/dz = 0. when U^z/y = = 0. give intercepted Here give minima. while Jo(z) does not. 185.
On :hen the other hand. or to a point of crossing of Ji(z) = 0.(^)+. are ruled ght. z. from the first points of crossing to the second. whichever comes first. from the axis of z the points of meeting with U^/y^ = 0.. lence as y increases in comparison with ZJa the equation = Jo (^) . are ruled heavy. = 0. and The lower regions of the curves Jj (z) = 0.[V. and on alternately. 3 by if y be made smaller we have more and more nearly z. be easily identified as regions which give maxima minima when the diagram is used in the manner described r bove. are ruled heavy the diagram the other regions.cos iy The reader may verify that the nore and more nearly holds. . Thus the whole diagram is filled in. The regions of the curves UJy^ = can now.. namely U 0''^ . As we have seen = 73=Focos(i2/+j) /o(^)g)V. 50 that the branches of Uz/y"^ = bo p.cos(jy+g. . and has there the same double tangents. then the region from that oint to the next point at 3 on. the reader 189 id as there. light. d'M^ _S 1 le double points are places of minimum (zero) value of M^. which give maxima. 1 To mark regions which give minima they . starting from the ouble points.] DIFFRACTION OF LIGHT. are ruled heavy the next egions. as may verify. which Us changes sign is ruled light. Thus we approach more and more nearly verify the statement made at The value of M^.. Thus the first regions from the double points to a maximum or linimum of the curve. the lines J2(«) 187 above. vtl . mrve Jo (^) — cos Jy = meets the axis of y at the same points as in comparison with he exact curve.
190
DIFFRACTION OF LIGHT.
[XIV.
with increasing z and stationary y, that is with increasing obliquity of the rays, approaches zero. Hence at a great distance from the
geometrical image of the orifice the illumination
zero.
is
practically
y=cz.
Consider a line drawn in the diagram to fulfil the equation A line making the same angle with the axis of y would
have the equation y
= z.
Let us consider the intensities
for
points on these two lines.
Since ylz
line
=c
for
the
first
line,
then
for
any point on that
=
sin{j.(cli)}(i/,l/3
+
^/a...)
47
For the other
tinction,
line
we
have, accenting the functions for dis
= sin hz (c + U(cJ,c=J3+c»J5 ...
),
4/
u:=Ij.Ij^+—
Therefore
cos
JS:
fc
+
)
+ Jo 0^/2+0^/4I
...
48
Now
and
if
the radius of the geometrical shadow be
fo>
then
——
n
— C.
XIV.]
If
DIFFRACTION OF LIGHT.
191
f
the other line y
be the distance of a point of the illuminated area upon — zjc we have evidently
I
of
?r'=?j
As
2/j
= 0, or the axis special cases of these lines we have ^ = = or ^^is of and z. ^^^ The last is dotted in z, y 3/
the diagiam, and by the result just stated corresponds to the edge of the geometrical shadow.
The
at the axis of
intensities for points along the first line are the intensities symmetry for different radii of the orifice, or with
h,
constant radius for different values of
the distance of the screen
points along the second line are the intensities for the case of Fraunhofer, already fully considered.
orifice.
from the
Those
for
In the
first
192
DIFFRACTION OF LIGHT.
[XIV.
Some
values are given in the
XIV.]
DIFFRACTION OF LIGHT.
193
to pass unimpeded. Going back to the at above for 167 the intensity, we p. original expressions, obtained see that for the total effect of the uninterrupted wave we have
of the
wave being allowed
by
I
and
2
(7oo
=
Stt
I
Jo {Ip) COS
^^^^ ^^^
(PpO P(^P "" "^
^^^^^
^
2
^^^
2^
51
l^
IB
^'^
^ ^'^
/"°
'
*^"
P^P^'^l
cos
2^
Thus we get
IppEis at p. 170 above, tt?*^ be taken as unity. This is as it ought to be, as it leads to the expression l/(a46)'" for the intensity at thus verify the the point in which the axis meets the screen.
We
statement, made on the
last page, that the
maximum
illumination
is
at the centre of the geometrical image of the orifice
four times
that due to the uninterrupted wave.
It
might be objected that the
original expressions obtained,
which are here extended to the whole wavefront, had reference
only to a small part of the wavefront, namely that filling the It is to be observed however that the effects of those orifice.
elements of the wavefront, which
lie at
a distance from the axis,
are very small compared with those of the elements near the axis, and so the integrals can be extended as above without error.
To
on
find the illumination with the
opaque disk we have simply
the values of Cr, Sr, given
differences
to subtract
p.
from the values of
the
orifice.
Coo,
Sjo
169
for
Thus denoting the
2 
by
Cj, Si
we get
Ci
= Coo  Cr = 
(Fosin2/

"^
Fj cos ly)
53
S, = S.Sr=
(V,cosy{V,smy)
y
'2\2
and
M\=Q\Vl^V\\
M\=i^'jU+U\+m2U,mx,\(^y
+
G. M.
54
or
+
'^)
W,oo^\{y+~j^.
13
55
194
DIFFRACTION OF LIGHT.
[XIV.
Comparmg
and U^ by Fq.
If
see that they are the
these with the expressions on p. 170 for M^ we same except that now Ui is replaced by F,
= 0,
that
is if
the point considered be at the centre of the
geometrical shadow, Fo
=1
,
Fj
= 0,
and
'(!)
the brightness there is always the same, exactly, as if the opaque disk did not exist. This is the wellknown theoretical result first pointed out by Poisson, and since verified by experiment.
that
is
For any given values of y and z those of Ui, U^ by the equations g
Vo
M^
is easily
calculated from
=
cosi(^y
+
~j\
U„
F,
=
sini(yfQfr,.
A
valuable set of numerical Tables of
Mi
all
fully illustrated
by
curves will be found in Lommel's memoir.
When
z
is
continually increased in value the equations
Fo
=
cosi(^2/+j, 2/'
Fi
=
sin
1
(y
+
y,
more and more nearly
zero.
by ii U^ and Ui continually thus becomes 4/y^ at a great disMl approach tance from the shadow, as in the uninterrupted wave.
hold, since,
The value
of
As mum.
before
we can
find the conditions for a
maximum
14,
or mini
Differentiating, and reducing by i8 and
we
obtain
The maxima and minima have
J,(z)
place therefore
or Fo
when
= 0,
= 0.
which
satisfy!^
The
roots of these equations are the values of z
.]
DIFFRACTION OF LIGHT.
195
d Kd
or
are, therefore, values of z
minimum.
;
The
roots of J^ {z)
which make J^ {z) and Y^ a maximum — are given at the end of this
book
those of Fo
to,
=
ynilar
and obtained
can be found by a formula of interpolation in the same way as, 43 above.
The tangent
of z
is
of the inclination of the curves Fq
=
to the axis
given according to 24
by
By
using in this the values
we
see that \i
y
= 00
^
,
=x
,
that
is
the curves are for great
lues of
y
parallel to the axis of y.
Also since
f
^.=^«gy^=+
the asymptotes of these curves are the lines
table of the roots of this equaparallel to the axis of y. tion is given at the end of this book, and as has been seen above values are given approximately by the formula (p. 46) their large
drawn
A
(m + f) IT.
Writing now
=
cosi(,
and making the values of
disappear,
+
i^)
y,
+
(f)V.(f)V.4...=
0,
z small, the terms after the
first all
and we are
left
with
cos
i(y+3
= o,
59
that
is
2/2
+ ^2^ (2m +1)773/.
13—2
We infer that the branches of the curve Fq = become near the origin arcs of circles all touching the axis of z at the origin. The curves Fq = 1&. [XIV. . This equation represents a circle passing through the origin.196 DIFFRACTION OF LIGHT.
Jj (z) = 0. and Fo = the value of ^^ vanishes. 197 Where both Ji (2) = .M'^) = I J. 62 ^) . but there are no others. we can easily prove that Vl + Vl + V:^ + V'. Hence there can be no the line J^ (z) = with Fo is = 0.if^) + i (M[^ . {z) cos \z + (c . points of inflexion of the intensity curve can only exist outside the shadow region of the diagram. But by the diagram intersection of positive everywhere. for the sake of comparing further the case of the disk with that of the orifice.] DIFFRACTION OF LIGHT.^ = Ul + Ul + U'. let us contrast the intensity along a line y = cz with that along the line y = z/c. but not so that of ^g— Hence at such points the curves of in tensity have points of inflexion. Thus the state ment just made proved. negative if y>z. Lastly. Accenting the quantities for the second line. 58 becomes z For since ^y dz  9_i_ ^ /zY which —is is positive if y< z. As in the other case.XIV. Now we have and for the disk ^'^^ M^.'^U'^ + for the orifice 6i 2Jo(z)cosiz(^c^^). except when y < z. Thus by 6 1 e (M\ .Fi. F_i = .
Bayer. The discussion given above is in great part an account of this memoir. 77 be drawn on the screen. From the result obtained above for Fraunhofer's interference is phenomena. Akademie der Wissenschaften contains another most elaborate memoir by Lommel on the diffraction of a screen bounded by straight edges. A little consideration shows that the illumination at anj . If the source is at a great distance from the telescope we may suppose with sufficient accuracy that the plane of the orifice is at right angles to the ray point of the linear source. and is therefore by the line y=^z. so that 62 becomes On this line M=M\ M^^Mi bounded and 2 {Mi It is clear . The shadow region is that for which y > z. The same volume of the Ahlmndlungen der Konigl. namely that the intensity of illumination 4 — portional to z JT (z). = .198 DIFFRACTION OF LIGHT. which contains. and given above.Jo {2) cos z. refer for further information on these cases of Lommel's paper. and a few other results. Let the circular orifice be the opening of the objectglass of the telescope which in Fraunhofer's experiments is supposed focused on the source of light. c = 1. can only here find space for some particular applications therein made of Bessel We Functions to the calcul€ition of Fresnel's integrals. \ coming from any \ i Let rectangular axes of f. z dark rings which from the diagram that as y increases the number of fall within the shadow also increases. and making use of the properties of Bessel The reader must diffraction to ^ ^ functions established in the earlier chapters of this book. profind asj the source of light being a point. a wealth of numerical and graphical results of great value. in which the analysis is in many respects similar to that used in the first paper.M%^. we can the intensity at any point of the screen when the source is uniform straight line arrangement of independent pointsources. with deviations here and there from the original in the proofs of various theorems. [XIV. as we have indicated. and let the line of sources be parallel to the axis of 77 and in the plane f= 0.
16. p. p. and is due to Dr H.. Struve*.. The integral is therefore 1 r Jljz) dz This integral may be transformed in various ways into a form suitable for numerical calculation. Before. we can give three lemmas on which first is Struve's analysis his process depends. and f the distance of point considered from the axis of the telescope. 27rr ^ z dz jnce dT] _z dz fjb'^'.. + 2 J4 (2c sin <^) cos 2a * j dcf) Wied. however.^(c)cos a + 2J^(c) cos 2a+ cZ<^ . The process here adopted depends on the properties of Bessel functions. Brit. r be the radius of the objectglass. 433. Ann. 28 above. 63 By 72. I . 9th Ed.^^oint I ^] DIFFRACTION OF LIGHT.. we have Jo f 2c sin Ij = J^(c) + 1 2J. . is we have to prove The a theorem of Neumann's and expressed by the equation Jl {z) =TTJo J^ {2z sin a) da. (1882). 1008. But Jo (2c sin 1 =1 = I cos (2c sin sin ^ <^J f^ {Jo (2c sin (j)) + 2 J2 (2c sm <^) cos a I. p. + Encyc. of the screen 199 factors must depend upon f and (constant omitted) be represented by f But if z^ dr). . Another method of obtaining the same result will be found in Lord Rayleigh's Wave Theory of Light\.
p.>0. (ff/n.200 DIFFRACTION OF LIGHT. and y^ TT xdz for cZ</> and 65 becomes r / P sin (xz) dz which is the second lemma stated above. ^(j) sin ?i<^ . X for ^. j=^=^dz.. J. [XIV. is the equation . ioV2(cos</)cos^) iflV2(cos(ycos<^) first Subtracting the second expression on the right from the we obtain ^ V 2 (cos <^ <^ — cos ^) d/n. J e V2 let (cos ^ — cos </>) For ^ and last put and n be made very great . The second lemma r . first 'TTj if we write 2 for c and a for <^. Hence ^ Write ocz rsm^^ for 6.(vx) 2 = °° f sin {xz) . If Fn denote the zonal harmonic of the ?ith order. The quantity on the left is ^7rJo(0). . (see p. Identifying terms in the two equations we obtain Jl (c) =1 2 f'" Jm (2c sin 6) J. 64 ^ !. the equation becomes r^cos 4>d<l> _ r°° sin (fydcf) Jo \/e^6^ ^2 ~J9 Je V(t>^6^ V(l>^6^ 32 above). 40. Thus the lemma is estab lished.n d6. then it is a theorem of Dirichlet's that p„ (cos e) = I ^ ^ r_^M^^^ — V 2 (cos </> d<j> + j i cos 6) tkJ) J r^M^i±^ — V 2 (cos 6 f^ d<t> cos </>) _ ~ r^ sin ^(f) sin cos . or f^"" Jl {z)= {2z sin a) da. by 18 above.
_. . let {iisy\2j {y (zY 22«(ns) n(25 + i) ^ 2 n (2s + 1) Returning now to the integral Jl{z)dz I J V us denote it Z V^^ — V'^ by Z. p. == z'^z^ v'^Jo tt Ja (2^ sm 'rrJv 13 (2sr sin a) = z sm ^ I = — {J^ (2z sin a) + Jg (2^ sin a)). Z=— But by Ja 20.. ooN^or°°sin(2^sinasiny3)(Z5 Jv \Z^ TT J — V^ lemma But if we put r 2 sin a sin L Hence — — /3 = ^ we get by the second tt sin (2z sm /3) a^ (22^ sin a sin p) c?^ ^ . .„. vF^i? =2^»(^*> Z——\ (sin yS + sin 3/3) c^y8 Jq {2v sin a sin fi) sin acZa. and by 45. 1 f^'' • ^ f/ Jo • o • . 18 fir Ji (2z sin a) sin (2z sin a sin yS) sin ^ d^. lemma I is expressed by the equation Jo (^ sin a) sin ada = . We \ have by the first lemma a) eta. Using the general definition of integral order (p. TTJo Js (2z sin a) so that 'z 1 = f"" I TTJo sin(2^sinasin)8)sin3/5cZy8. p.. 12 above) of a Bessel function we get Jo {z sin a) sin a c?a =2 Ayr^ _ ( « ) sin^^+i a da / I which was to be proved. The third . 66 I:..
sin /5) cos.{z)=rff.^«<^>we find Multiplying by zdz and integrating Now by 68 z T^' =— / cos iz sm c/) sm Hence H^i^z) =— l [ {1 "cos (^sin^)sin^rf(9 — TT I — COS (^ sm (9) d)\ sm Ode 6>cZl9. Jo = ^[ TT Jo sin2(i^ sin sin .202 DIFFRACTION OF LIGHT. ^ 2 z dz (^S^«(^)=^. Jo then by the series in 69 ^^(^> = XT / X 2 f ^3 z' z' \ ^{l^73lX3^T+1^3r5r7jthat ^^ We shall now prove ^1 It (2^) = — TT 2^ r^'^ sin {z sin ^) cos.ddQ. 6S Expanding sin {z sin ^) and integrating we obtain Now let jETi (z) be another function defined by H.(z)zdz.y5 c?/3. 71 jo can be verified by differentiating that \ d ( d\ ^^ . 2v sm ^ J^^ d^ ^ 67 2 =— sin (2?. Let now H^^ (z) be a function defined by the equation "^ Ho(z) = [ TTJo sin (z sin (9) dO. which the third lemma becomes Z=TT j (sm^ + smSyS) f^'^ I — ' . b}^ [XIV.
. The series used in Chap. We must then have recourse to a semiconvergent series. K^{z) discussed in the Theory of Sound. See Theory of Sound. to which the reader is referred for further details. process*. Jo term vanishes at both If we write 2v for z and jS for 6 the equation becomes cos^ I H^ (2v) = 2 ^^ [ "sin (2v sin ^) TT Jo ^d^. It is 2 ri^.^ ^ ^^g. but when z is large this series is not convenient. can be found easily by the method of Lipschitz.^ ^. § 302. we denote the is by §i (z\ the relation between the two functions The value of H^ (z) can be calculated when z is not too great by the series in 70. H^ {z) given in the first of the three equations just written that H^ {z) approximates when z is large to Integrating 74 by parts we I obtain H^ (^z) = ce the integrated — IT sin {z sin 6) cos. dv. It 203 Ih s may be noted clear from the form of that every element of this integral is positive. iv. jETi The function H^(z) If differs however from the latter function (z) used by Struve. similar to that established in Chap.OdOy limits.DIFFRACTION OF LIGHT. 75 Hence EA2v) ^ by 67. already vii. above for the Bessel functions. ^^^ ^ ^^ to be observed that the functions here denoted by Hq{z\ Hi{z) are the same as Lord Rayleigh's K(z).{z)iH. § 302.{z)^TT I J ^iz sin 0(10 ( _2 p * e*'^' . The following is a brief outline of the By the definitions of the functions we have J.
. iv.4) (32 4) (52 4) (72 4) 1.Z' + 1^ Ho (z) = TT .(8^)3 I r.3.iv) i which are 0.) + A /— TTZ V [P sin {z i'7T)Q cos {z p.j7r){^gj (P.(8^)^ ) •••} y^^%in(. h.Itt) 1 i *"T 1.2. 48 above being written for From 73 and is this the value of H^ {z) is at once found by the relation H^ {z) = TT ^{z + Z' .. in as namely Jo (2) Chap.(82)2 "* (P .204 DIFFRACTION OF LIGHT.L .i. Z'z'^ . we get the expansions required.4. is zero.(z' . // .) 2z Vtt^"^'^ (P4)(3^4) . . 3'^ . some reduction h This integral Expanding the binomials and integrating.. and if A = oo round the where h is gives after rectangle. h \.^z' + 12 cos . . and ..2. the angular points of real it and positive. Now take the integral (in which w = u\.3. 6z'  .j. making use of the theorem ^0 and equating the real part of the result to ^ttJo {z) and the to imaginary part —^i7rHQ(z). 76 {z where P and Q have the values stated on x). [XIV.2..P 3^ b^z' + . (z 1 .4) (32 4) (52 4 ) 1. Jtt)!.
In this way can be determined the minimum distance apart at which the luminous lines may be placed and Now sources.5 The two ratio of the intensity of illumination midway between the lines to that at either is therefore 2L{h^)_ L(0) + L{7ry This has been calculated by Lord Rayleigh (to comparison is due) with the following results whom this Z(0)=3333. so that X(7r) = 0164. let there be two parallel and equally luminous linewhose images in the focal plane are at a distance apart = 7r//t. still be separated by the telescope.52. 4J per cent.35^1 (2v) = 0. = 2X(j7r) 955. only about line.32. We shall take the image of one as corresponding to ?.) . 32. Thus the intensity at any distance corresponding to ing to V ?• = is proportional to —v^^ . L (^7r) = '167l.XIV. be noticed that H^{2v) is nowhere and that Hi (2v)lv^ has maxima and minima equation values at points satisfying the » > d dv H. It is of great importance to compare the inv//Lt tensity at the image of either line with the intensity halfway between them. 3 12. for the distance stated. less than at the image of either I . and that of the other as correspond= TT. 7 P.(2v)^ 4>v^Ho(2v)SH. L(0)+L(7r) The intensity is 79 therefore. Putting 'rrHi{2v) ^^'^=2lW we have by JO L{v) = 1 2^v'^ 2V P. 0.(2v ) if _^ ' V* ' The corresponding values of v are therefore the roots of 4^2 ITo (2t.] It is \' DIFFRACTION OF LIGHT. to 205 zero. say.
By multiplying 2 IT H.206 DIFFRACTION OF LIGHT. to a constant factor. The point will be at a distance ^ from the edge of the geometrical shadow. Now which gives h 2r' Since h is the focal length of the objectglass. represents the whole get illumination received by the screen from a single luminous point fxd^. it plainly may be regarded by the latter point from an infinite illuminated area in front of the objectglass. Or. the two lines are. the illumination received by the same point from an area bounded by the straight line parallel to tj corresponding to the constant value of f. at an angular distance apart equal to that subtended by the wavelength of light at a distance equal to the diameter of the objectglass. called This result shows that the resolving (or as it is sometimes the spacepenetrating) power of a telescope is directly proportional to the diameter of the objectglass. the image of which is at the centre of the focal plane. We have by 71 Jo (2vf irJo Jo V = ij^ cos^^c^/3 = . and will be inside or outside the shadow according as f is positive or negative. Two lines unless at a greater angular distance could therefore hardly be separated. by the mode in which H^ (2v)/v^ was obtained. {2v) (2vy by that is by dv. uniformly as the illumination received If the integral is taken from f to I oo it will represent on the same scale. by this result. [XIV. and integrating from f = — xto f = oowe an expression which. Now '*'' '*" "''• j„ (2vy i„ (ivY i„ (2vy .
3 . 1^ .] DIFFRACTION OF LIGHT. 27rr ^ .XIV. shadow is therefore half the full intensity. and therefore enables the enlargement of the image produced by the diffraction of the objectglass to be estimated. 207 The second term on the right can be calculated by means Hence we get of the ascending series 70. Struve's result we write when v is positive "*" . 3 1' 8^ . 5 . 32 . 7 } 80 the expression given by Struve for uniform plane source. 32 . the image a the intensity produced by 00 For the sake of agreement with to from v + of which extends This multiplied by 4/7r is . 7(+^)=i Hence from if 81 source extends 12 . 52 . 3 5 — 00 to I be the illumination when the plane + QO we have I(+v) + I(v)=I=l. 5 2V 12 . of the edge The reader may verify that when v is great the semiconvergent series gives approximately TTVv 12yV 27r^ v^ following Table (abridged from Struve's paper) gives the = hXvl^irr intensity within the geometrical shadow at a distance f The from the edge. The intensity at the are together equal to the full intensity. but on opposite sides of it. This states that the intensities at two points equally distant from the edge of the geometrical shadow.
slit bounded by parallel edge^ shall suppose that the diffraction may be taken as the same ii every plane at right angles to the slit. we get The disturbance cos at P is therefore 27r (t v'^\ ™ — ^ = cos ^TTV^ COS J ^ + sin ^irv^ sin 27r ^ is t . pole so that its distance from the source is a\b. P from th element and from the The disturbance at P produced wi be proportional to cos ^ir {^—^] &3 ds s. t . where ab\ The whole intensity is thus proportional to \\ j \co^\Trv^. so that the problem is on of light passing through a narrow We in only two dimensions. Let a then be the radius of a circula wave that has just reached the gap. and consider an element of th Let also b be the distance of P from th wavefront in the gap. shall [XI\ We now consider very briefly the theory of diffractio. being proportional to cos27r?/^ therefo + sin2 ^irv^. . .dv\ {^m^irv^.dvY . then it is very easy to show that ''ifWriting as usual \ir^ for 27rB/\. the integrals being taken over the whole arc of the wave the slit.208 DIFFRACTION OF LIGHT. and 8 the retardation of the secondar wave (that is the difference between the distances of pole). The intensity of illumination due to the element constant. If the distance of the element from the pole be and s I small in comparison with h. ds the length ( the element of the wave.
Thus we R'lave i [ obtain V .. We shall write (7=1 cos^TTV^dv. .3)/2 sufficiently great the integral we Thus of 84 may be made as small as we please. by simple addition of the values of the G. 83 Let us now consider the Bessel function integrals on the right.n+z (^).(^)+ . 'he 209 it problem is thus reduced to quadratures. 14 . 'arious ised .{z) + . 82 ^ = 1 f A /— sin zdz=i fV*(^) V TT^ j Jo' dz..n+i{^) + i f Jm+s (^) d^' 84 on the By taking (4n 4.+ ^J^n+Az) + J. and remains to (uate the integrals.. then C=J [' A /— cos zd^ = il' J k (^) d^. 86 These series are convergent. lation is by means of Bessel functions.DIFFRACTION OF LIGHT. methods of calculating these integrals have been but the simplest of all for purposes of numerical calcuavailable. {z) dz = J. and give the numerical value of the integrals to any degree of accuracy from Tables of Bessel functions of order (2w + 1)/2.. Using the relation J. Jo id S=j Jo sin ^irv'^dv. +J.. {z) + J^{z) + J^. get right •I Similarly we find I ^= J r J^{z)dz = J. {z) + J. M.. when Tables are hTTV Let = 2.^(z) = 2Ji(z) + Ji(z) = 2Jl (Z) \2JL(z)+. S are known as Fresnel's integrals.
Vo on the spiral and draw the chord. [xiv. expressed long ago in series of ascending powers in terms of definite integrals by latter semiconvergent series suitable for use by Knochenhauer. 198. are apparently due to Lommel. 203 above. we have only to mark the two point Vi . is wrappec more and more round the poles A The origin of the curve corresponds to the pole of the poin considered. shows graphically how the in the diagram. and that the curvature there is irv. C and 8 were Gilbert.Q cos \z) J where The proof of V is left to the reader. p. so that if Vj. successive alternate functions for the given argument. v^ correspond to the distances from thpole to the edges of the slit.210 DIFFRACTION OF LIGHT. and From the when at p. that the inclination of the tangent to the axis of abscissae is ^irv^. The abscissae of the curve are values of G and the ordi nates values of S. It can be shown that the distance along the curve from the origin to any point is the value of V for that point. The square of the lengti . and are stated in The series the second memoir referred to above. V is large are obtainable by a process similar to that sketched It is not necessary however to pursue the matter is The very elegant construction shown known as Cornu's spiral. which value of C^ + S"^ varies. here. As V varies from closely to x and from to — x the curve and B. He (P gives also the series COS = 1 r J_ Jo I (z) dz = ^/2 iz+Q sin ^z) 87 S=ifji (z) dz = V2 (P sin \z .
1 I As DIFFRACTION OF LIGHT. 211 of this chord will represent the intensity of illumination at the The square of the length of the chord from the origin point.(^). V varies it will be seen that the value of this sum oscillates and more rapidly while approaching more and more nearly more to the value ^. 14—2 . to any point v is the value of C^ 8'\ that is of + '».y_i(z)d«j%ijjv.
6. although difficult as of not so those already considered. ^ _du 1 . as a particular where v is a function of r only.CHAPTER XV. such as the small vibrations of a gas. r. so that sin ^ 9^ V dd J sm2 6 (2). function of t. and Sn is a surface spherical harmonic of order n. We will begin with the equation which occurs in various physical problems. it ^ d(l>^ Then after substitution in equation appears that v must satis ^ dr^ if 2 ^ f 2 _ n(n + l) r^ \ r dr ^ ^^ ^ ' \ y and now we put . such that ' „ dhi . Using polar coordinates. appear too important or too interesting to be passed over entirely or simply placed in the collection of examples. ^du" ^ do) "^ sin^ 6 d<f>'] Assume. u du di a^ r^ ( is a. d /_. V 9^ dr'^ sin 6 BO V^"" solution. </>. In this concluding chapter we propose to give a short account some special applications of the Bessel functions which. or the variable flow of heat in a solid sphere. MISCELLANEOUS APPLICATIONS.
thus with a simplified notation. (See Riemann's Partielle Differ entialgleichungen. xvii. 176 and Rayleigh's Theory of Sound. the function Sn reduces to a constant.. we have a solution which 1886): may be adapted to the following problem (Math. a particular solution of the equation i or 2. 213 find that w satisfies the equation Hence ^ = ^^«+^ ('^^) + ^Jn^ ('^^O* finally . M+ ^j_. = . In practice n a whole number._ . B the function is adapted in the usual way to suit the particular conditions of the problem. and by a proper determination of the constants K. Prove that.. pp. m . IS is n. and J^+Af^'^) = \/ — sm sii Kr (see p. chap. The whole is time t.^^^'^ {^/„. at temperature zero. Tripos. after the temperature at a point distant r from the centre is y cool in a medium _ ^^ ° 4 sin K — — 4>Kb /c6 /e6 cos kI) sin Kr ^hh ' sin 4/c6 r . uniform homogeneous sphere of radius h is at uniform temperature Vq. 42) .) If in the above — 189 we suppose n = 0.. and is surrounded by a spherical shell of the same substance left to " A of thickness h at temperature zero. A.] MISCELLANEOUS APPLICATIONS.
= b. [XV. when t = 0.. =j^^.6 "' which agrees with the result above given.v^ d /sin Kr\ h sin fcr _ ' dr\ r J r when r = 2b: that /c is.6 cos . . VI." ratio of the surface Here the conditions to be satisfied are to r V= Vq from r = V=0 r=b .214 MISCELLANEOUS APPLICATIONS.. above. Proceeding as in Chap. and when r for all values of t Now. Jo Jo A^lb 4.6 as above stated. we infer that r2b rib A^ that is. conductivity to the internal h being the conductivity. \ sin^ Krdr — \ Vr sin Kvdr. . if cos 2/c6 26^ leading to ~ sin 2/c6 "^ /t sin 2/c6 _ ' ~46^~ ~26 tan 2.6 =^2^^... = 26.6 — . assuming a solution of the form sinA:r j^^vj[^ the last condition is satisfied if ^_.6 ^ 4.6— sin 4. i = ) ^0 1 ^ sin Krdr 6 cos . where the values of k are given by the equation tan 2.r = 2h.6 _ and hence /sin sin «6 Kb j^ _ " 4 k' sin .
such as a knittingneedle. Let w wire above the clamped end. is stable when placed vertically with its lower end clamped in a vice whereas it . vertical in the same way a very long stability.to be a pure imaginary ± iol/k. we imposed must take a. When there is a forced vibration on the system. that of thestability of an isotropic circular cylinder of small crosssection held in a vertical position with its lower end first is clamped and upper end free. as when a spherical bell vibrates in air. 215 observe that to cases Returning to the solution given by equation 4 above. tion of this will be found at the We The will is then An illustra27r//ca. and y its horizontal displacement from the vertical through that end. Differentiate with respect to x . y) I being the whole length of the wire. be the weight of the wire per unit of length. we obtain by taking moments for the part of the wire above (x. we may when a^ is real and positive.XV. so as to obtain a there is when as in the timeperiodic solution. It is a matter of common observation that a comparatively short piece of steel wire. To find the greatest length consistent with we consider the possibility of a position of equilibrium which only deviates slightly from the vertical. end of the book. now proceed to consider two problems suggested by The period the theory of elasticity. problem just discussed. /3 its Then if x is the height of any point on the flexural rigidity. then or . would be impossible to keep piece of the same wire. the solution is applicable a "damping" of the phenomenon considered.] MISCELLANEOUS APPLICATIONS.
216 .
) I . (See Greenhill. Theory of Elasticity. Soc. and therefore z. . ii. Therefore ^ = 0. . The critical least value of I obtained from this equation gives the first is in length of the wire when it the vertical position and if I . and Love. the second is convenient if we wish to know whether a given piece of wire will be stable if placed in a vertical position with its lower end clamped. Again. Hence. = 0. p. Math. 297.] MISCELLANEOUS APPLICATIONS. and it is only the second of these that satisfies when r = 0. that is. Proc. or 1996 \^JIw.XV. Of the two formulae given the first is the proper one for determining the critical length for a given kind of wire. 1881. the whole weight of the wire. when r = l^y p. Ji (kv) 217 = ar^ + /3r^ + . the vertical position will be stable. IV. Phil. less shows signs of instability than this. It is found that the least root of is approximately 1'88: so that the critical length is about C?)'. . Camh. To the degree of approximation adopted we may put or in terms of /3 and W. approximately. when x must be zero. in order that the assumed form of equilibrium may be possible.
dw ^^'r[dd du 2^2 d^J' dw du\ y di'd?' 1 fd(rv) 2'STs i( dr "dd. As another simple illustration derived from the theory of elasticity. The stresses across a cylindrical surface r = constant ^ are Prr = XA + 2yLt du dr y fdu dw\ We where solution of the type proceed to construct a particular ^ = 0.= (X + 2f. 115). and 7. z) are the coordinates of any point of the cylinder. In orde to obtain a periodic vibration with no damping.)^^ dt' d'v dz r dr dd where _ * 1 9 (ru) 1 dv dw dz r dr rdO d {rv)\ 1 . XV. v=yi'^''''^'\ r only. V. Vf corresponding displacements. p. F is a function of u. the equations of motion for small vibrations are If {r. we suppose that j The torsional character of the oscillation is clear from th is real. form of .218 MISCELLANEOUS APPLICATIONS. a short discussion of the torsional vibration of an isotropic solid circular cylinder of radius c. after Pochhammer and Love {I. w dt^ dr dd dz 9^3 p. we will give. w. and the u. 'p are constants.e. w=o. 6.
and Pzr vanish Prr^ Per. 219 If we put. = 0. If /c = 0. if KcJ^ (kc) (see pp. Thus V must satisfy the equation ^ dr^'^rdr^X and since ~ "' is /. for the moment. which is the same thing. the proper solution where A:2=^j92_y If the curved surface of the cylinder is free. then the stresses c. Pzr niust vauish when r = Now P^ identically .^V. iiid the equations of motion reduce to two identities and pfVZ = ^r. the differential equation to find V is ' dr'^ r dr r^ of which the solution is V=Ar + B r I . v. 179).^VZ + ^Z^^{^^^{rV)). P^ will vanish if dr\ when r = c : r \~^' that is. r^J V must be finite when r = 0. 2^^ = 0.o have A = 0. 13. if KcJ[ (kc) or. — Ji (kc) = 0.] MISCELLANEOUS APPLICATIONS.
is negative the type of vibration is altered thei a damping of the vibration as we go in one direction alon the axis of the cylinder. by the equation t*' we have a solution Unless some further conditions are assigned. Special solutions may be constructed to suit special boundar conditions thus for instance if we put : \ . But we may of which there also take for any one of the real roots of is any real constant. i . k. . with and in particular we have as a special case when ?i = 0. If pp^ Jfijp so long as /x/c. vibrations of an : period are possible. When the period is lirj'p the velocity of propagation parallel the axis of the cylinder is 'P t ^ W/^ Jpp'IikV 7 which is approximately equal to compared with pp^. This leads to a solution of the original problem in the shape u = 0. : now . is sma i is — /jlk.. 1.. and Z is a constant. If and 7 determined. our present purpose [XV V=Ar. p is arbitrary whatever value of k^ is taken that is to say. 2. w = 0. for MISCELLANEOUS APPLICATIONS. an infinite number. . real or pure imaginar constant.220 or. as a k^ is any one of these.
Grelle. as we have begun.XV. /c.m —=^ . the memoir of Pochhammer. . we have du ^ qx du „ „ or dH daf n\l X du dx K^ _ ' X where K = 2irpJ(n+l)lg. with the oscillations of a modify Bernoulli's problem by supposing that the density at any point of the chain varies as the nth power of its distance from the lower end. . Love.] MISCELLANEOUS APPLICATIONS.a^x + a^"^ . For the discussion of the extensional and flexural vibrations and Many < other illustrations of the use of Bessel functions in the theory of elasticity will be found in recent memoirs by Chree. equation of motion is 1. i . Ixxxi. chain. . the dt" dx\n + ldx)' and if we put y = u cos ^irpt. the reader should consult Love's treatise already referred to. parallel planes. mode of vibration for a cylinder of radius the circular ends of which are glued to fixed The doubly the curved surface of the cylinder being left free. let us Proceeding as on p. Rayleigh and others. + Assuming u=aQ\. when m is a real integer. To conclude.n COS pt + Bsm sm pt) Ji (ksV) sin s. p. but measuring x from the free end. 324. is where u a function of x. 221 v = X (As. infinite number of constants Agm* J^mi have to be determined by suitable initial conditions. Lamb.^ any root of J^ (Kgc) = 0. and this gives a possible c and length 2^.
216. I that case u=p. 3(n+l)(7i + 2)(n+3) "*"•••}* which is the same thing. \ = h n=±J.222 MISCELLANEOUS APPLICATIONS. [XV we find that the difierential equation is satisfied by 2) ) '^^Y n + l"^2(n+l)(n + 1^ 2. to extension of Bernoulli* problem problem is due. a blin( composed of a very large number of small uniform horizontal rods the shape of the blind being defined by the curves with X positive. remarks that to realise the conditions of practically. th« Professor Greenhill. instead of the chain. y = Ax^J^ J4^y/(^+i)^l whom this cos 2^. the differential equation of the form can be reduced to the standard form Sj£('S)»». therefore. at p. In connection with the reduction of the differential equatio it may be pointed out here. Thus n —\ gives a triangular blind. (See p. that. SO that it is integrable by Bessel functions. by Finally. /i = l. we should take. which This gives a general rule for the transformation in cast it is not so obvious as in the case considered above. 233 below for other examples. and so on. K^ = 4^l9fi. for M.) . if yr~^ be substitute* 1 and sc^k^ for r. or.
be found that the quantity on the left equation reduces to which vanishes when A. part of iz must be negative. that such that the real part of zi is negative a" e^^i z be complex. the latter condition does not hold for real values of solution z. and I (1 .= 1. Tn{x)^ Jn{^) ^^ ^^^^ solution of the general differential equation which vanishes at infinity is not without difficulty. i Further the investigation given in Chapter vii. . 60 above. The solution then in the absence of fulfilment of complex.NOTE. We can find a which holds for both real and complex values of z as follows. if It has been proved. so that. the following explanation of Weber's treat the determination of the coefficients of nd IAs ment of the problem (Crellej Bd. 1873) may not be superfluous. 75. 59. For by differentiation in the differential it will /''(X2l)'*ic?X. = 2''J^U{n J) J^ (z). above shows that the differential equation dw dz^ is satisfied 1 dw z dz also by taking =/. These two definite integral solutions of the differential equation seem to be due to Riemann. pp. and when X= oo . In the' second of these of integration is solutions. either z must be essentially or n<0.X2)«i ^A. and the solution is of great importance for physical applications. on the supposition that the path the axis of real quantity from 1 to oo the real along . who gave them in his memoir on Nobili's Rings referred to on p. 128 above.
.. where ^~ (2nl)(2n3). and we have putting in the second integral for A. = 0. and putting x for the second term on the right.xY A. we have zero 5 multiplied by £b" vanishes when the sign of integration in the 'under the quantity Expanding second term by Taylor's theorem we obtain the expression The 0. Integrated this takes the "^ form \ x^ ^ 2 or^i r is (2w).(2n  1) ! .xf  T e""^ (1 + }^f~^dX Thus if u It is a solution of the general difierential equation. = A^T^{x) + BJ^(x\ This can be done as B^ are constants to be determined. = x.3. first term on the right of a.{2wl) where 4„. where X^)"^ real 6^'^ (1 w= Since (. 1 chosen.T . It will hold also for complex values of x provided that the real part of x is positive. that and the limit when £c = of x'^u is therefore .3 2»^0il)! (0 ^\^ 1 • V AV A2(«i) 1^ <^< 1). and we have Real part oi w={. 4 is real in the first integral we may use X as variable in the second part also.5. Multiplying both sides by a.224 NOTE. = r^n r^"''^ j\2«i + ^!^ X2H3+ / + \^^r\ dX.. Let a new path of integration from + We have to 0. /x. We now write u 1. for the moment /^^ let z be real.". is say. and therefore follows.} ..=)"^ cZ/. e''^ i JJ^ (1  d\ + J^" + /. and from to oo i be Now.
. 14 above. and 229 7 below) we get for the values ^ = ylog2. ^"^ ('^1) 6 ''riting 2^.. Thus so 'W„^i is derived from u^^ coefficient..= C{Y. we have r°o /•I 2»l „t 2nl ^:= ^cos(a^^). where 8 C is a constant. M. or x^'^^ X dx x^ without any change of r?..1)A„ = l. (2. 15 ..(l^)^^cZ^+ Jo ^e''^(l+e)~2di. Hence finally . = {nl)\. of Thus x specified the solution 7.{x)}.1)— 1. p. 225 Now the limit of from the explicit form of Yn{x). 30. 19. (2.3.. we see that is .^^. 3.{x) + {y\og2)J. of the general differential equation may be written by 6 and u.I NOTE. p. comparing with the case of = (see Examples 15. Ja [ntegrating by parts we get ' dx ^»2^12^^2^^)o""("')(^^) C X 2n ^^ 2n+l rr ^t rt = a.2"i x"" r„ (x) when a.4 for what we have called u. G.
. 4.Jn'+IY) J.. (cc) Jiv" + 1f + (n . = 22 ()*'*"'> where cf> = cosh~^n. [. If. the radius veeto SF is given by the equation a r 1 1 e + cos <^ = (see p. = J.. as on p.+ . we put fL = <^  e sin ^.}.%J. 19). (iv) Y. n being supposed Q 1 greater than 1. Verify the following expansions (i) : — (x). in the problem of elliptic motion. sin na. Prove that.EXAMPLES. Bessel's expression for </> in terms of ft. where (iii) — sinh"^ n. (a^). 1 2 {Jj (e) cos fi + J2 (2e) cos 2/a +. 3..{x) + ^{{n+ = Jj.= nx = <^ 2'% sinh s^^/g (ic). 1. e"^ = J. ^2(25+l)(2s+2)'^^^^ 4s(25 + 3)'^^^^^ L^  '' ' ' . cos nx = Jq(x) + 2'X (— ) cosh cosh s<f>Ja s<^ /« (x). prove that de^ e Be e" dfi? and hence obtain 2. (ii) cosh nx sinh + 2^ cosh s<f>Jg (x).\ogxiJ.
(c) + 5 J. .3. 227 Prove that l=Jl+2Jl + 2Jl+..3. ^ . [. ' ^«^^^"2».3^3 (b) J. [" e'^ K..J.n!l^ ^'2(2^ + 2)^ 2. Jl = 2 (JoJ.EXAMPLES. first complete elliptic integral to modulus k..J.J. + J^Js + JoJ.)..{ax)Jo{bx)dx = {Jo {bx)Y •'0 {a' + b'y^F{b{a' + bY^]. vol. Prove that Yq{x) = 2 f°^ I cos (x cosh 0) dO J """Jo 15—2 .c^£c = Att (a^ 6^) "*. + . = {a" by^ tanh1 "^^Iz^ . {b) J. + J.4. II.4. Show ( that ^WV^) . [F (k) denotes the 8. (bx) dx = Jo {b'  a^)* tan'^ n/^J^ b^a. (c) Prove that ). Jo r K.. (c) .) 7.).2 V6^T? {J. 4) J results (taken from Basset's Hydro " ^»^^^Jo I 1 + + ' a^ Kq (ax) cos 5a.5... b^a. + J. J.3(2n + 2)^ (2n45)(27^+7) 2. + 2J. dx = 27ri (a^ + 46^)^ F \ 2b {a^ + 4bT^}. (b) J.(2ri + 2)(2n + 4) (2n + 7){2n + 9) 1 ' 2. ~Jl=2 (J.(27^+2)(27^ + Verify the following dynamics.
— 19) are taken from Weber's paper in Prove that T^ Jo 10. integral on Or. and integrate first with Prove that /•°°ee^sin(Xa. Hence show by integrating first with respect to X on the right and having regard to 151.) Jq [ax) dx = 1.'\ with respect to x then with respect to . p. sin (Xx) dx. that 2 rsm{Xx) s/A^l and therefore r'^^dx=r'':'j^dx. Jo last example. 1873. 75. the value the right can be evaluated by know: of log x given by the equation \ogx= Jo u du may be first substituted in the given integral.r) J^ (r) c?r. (seep. [The following Examples (9 Grelle.228 EXAMPLES. and the second theorems. more simply. (a^ > 1) = 0.] 9. e~^^ log cct/^ {x) dx = — 1 r^ \ — = dX c^ I e~*^ logo. Bd. respect to cc. Prove that I Jj^ (a. {a^>iy [Substitute the value of /j (x) from Ex. 73.^ r* sin (Xic) e j dX I sin (A.JoW^a:(y + log2). 1 where [By the r°° I y =— g^log ccc^ic. 9. sin( Prove that Jo I loga. 12. and the integration performe u.] 11. 40 above).) ~ I — TTg^ — dX= /•« _^.
[Use Neumann's series for F^. 15) 2 C 1 dK s/a^i x. ^ say) . of iz is negative) always such that the real both satisfy the differential equation of the Bessel function of zero order. prove that for z real and positive (— . 22 above..f <^^^X.T > a) (a. Prove that the definite integrals +1 Ji pizK /•'» piz\ Jix' A Jk^i J^' d\ (in which z may be real or imaginary.\) pAsi tt /ta.  and for z real and negative (= 2 x). TT ia.^ . with the result of the preceding example.EXAMPLES. if any.kx\ c/A (. but is part. 229 Prove that Jo p. p.. (^) I 'l^. 73.] A. * 1" cos (xX) . 16. < a). 149. 13. 2 /°°cos(icA) . If (see Ex.i^(^^) Use the theorems /^^ COSXX J. r / \ I . n [15. Establish the equations sin (a\) Jo (\x) /l^ dk _ sinh ka e^^^ f^ f 00 + X2 g . and the theorem. Asin(a.
15 we are entitled to do. Prove that if is a function of x and y which satisfies the equation + dx" ^2 dy^ + K^w = 0.{ ^0 (»^«) + w (y .)} + Jy. f(z) = A. as from Ex.).{ r. so that [" 19.230 17. 13 and the theorem 149. i. p. am (Weber. . If we write. Math. (z) + {y. ^^ dX = . (. 9.] 18.log 2) /.(z)^nj.Y. 16. Ann. prove that f{^)f (^) = ^ttiJ^ (a.i«g 2) J. circle. Prove that I = f{x) dx i. as well as ' its derivatives x— dx . p. Prove that '^^ r "^^^ 20. A of Ex. 73. then I ud<j> — ? ' 2'7ru^J^ (kv).. .— and which. (ikz) + i J.). Jo and hence by Ex. when t6 the integral is is the value of u taken along the circumference of the at the origin. (z).(z). is finite and continuous dy for all points within and upon the x' r 27r circle + y^r^^O. . = . that B = i{ylog2). EXAMPLES. and therefore 2 TT by the theorem [Change from +x to —x along a semicircle round the origin and have regard to the term J^ [x) log x in Yq (a. (ife).
) If rj = (1 + e^Y sin ^ be the equation to a curve referred to inclined to one another at an angle cot"^ e. ^ ' 23. p. where k zero or a real integer.EXAMPLES. is a rational integral function of x~^. F=Ji(Xct) when z z = = and and ot < > c. 2 — Jn (^e) sin nx. 11. ^— = when tzr c. Hydrodynamics. then ~^ 24. show that the axes oblique equation of the curve referred to rectangular axes. 2 ^ TTX (Lommel. 21. 1^ so on. 231 Prove that if V= then 27r~' I dfji I e~M2 cos \v cos (xvJq (jJioj) dvj Jo Jo V= Jq (Act) when z = and and ct < c. 0^  6^ = %LJ^ (nx). 5how also that if ^^H then F= 27r~^ I cl/x e j '^^ sin Xv sin /xv J^ (/w^) c?'^. is 00 y ^ If = '^ i^ ()"^^ ' ne. ~ ^Ji"(n6)}2 if ?^ b n'J^(nb)' is Prove that =^+ J. (Basset. with the axis of x coinciding with that of $. n given by Jq (nb) the summation extending to all values of — 0.) . 33. l^^M dV ~ "a~ ^ when z = and ct > c. 22.
(27il)(2n3) 4 Express u^ as a Bessel function. Prove that the equation is satisfied by F= (il„. = (i)™ ».. then V„ Ux)} = (.3.(2^l) aj»+i r ^ I 1. and show that I '") ..(2n+3) 1 1 .) 27. sin ^. Show that the equation ^ is satisfied 2dR :>^ (/^+l) by either of the series t^« = + 1)1 i 1. cos mcf> + B^ sin m<^) (where x = p cos (f).™ Ux).. Prove that D'. for {J. Prove that (Lommel.) 26.J)»»«»''"V„.232 25.(2n l. Show also that and obtain a corresponding expression 28. (2n + 3) (2w + 1 5) 4 } (l)M.(2711)2 1 ^ '^ '^1.„ (Jl — /u.2..^* if D denote r ax .3. EXAMPLES.(2?^+l) + l.„ i>» {«*V„ ( Vx)} Ux).^)}. 2p) y=p ^^ ."»'«'/„. 2 2 .{. (Lommel.
j. da? (a form of Riccati's equation) id solve SxV^o. .')]. Deduce from (iv) that if ^'y^x'^'y=. then (iv) If then y = x^''''[.. a^.tum ^^^ *^^ roots of a^'"+^ = + ^..EXAMPLES.0. (Lommel.„(ra.) I . (ii) If (Py 2nldy I thfti then y=:x''[AJ. C^m are arbitrary constants. 29. 233 Verify the following solutions of differential equations by of Bessel functions If : means (i) — en 2m+l 2m ^:^raen ao.. and Cq. Ci..{x) + BJ_^{x)].l/„(yx^) + (v) i.
where F„ ^ {nJMY.(npJfjims). {nPJM)sl^gJ^ fjt. and of the chain. where 32.0. IT + r —+ dw dr^ (Pr^ n^)w. there will be no semidiurnal tide at about 300 miles from the end of the estuary.000 ^ 322 ft. (3 {nJMJ. (npJM)} J. Prove that the solution of Riccati's equation ccx can be made to depend upon the solution of Bessel's equation d'^w r'^ . =24. If a bead of mass M be attached to the lowest end of a uniform flexible chain hanging vertically. and if Cda {dx + 6. show that in a Vshaped estuary 53 fathoms (10. then the complete integral of dy + IS ^ 2 {^^^A^{^^'i)r'])y =o (Lommel.{x) vanishes when a. Prove that if u is any integral of when X is a function of ic. which communicates with the ocean. where m is the How Assuming that Jf. where a. (npJM)  sjmg Y.) 31. 113 6^ seq.234 30. (n^JfT^^) {ii/3jM)}Y. b are constants. denoting are the mass of unit length values of n to be determined ? 33. 2/\/m^. (See p. EXAMPLES.) . being the total mass of the chain and bead. ^ (J„ n cos nt + Bn sin nt) F„. then the displacement at a point of the chain distant s from the fixed end is. for the small oscillations about the vertical.) deep. oi = a/p.
h being the internal and surface conductivities. due to the disturbance propagated in the is air outwards. A sector of an infinitely long circular cylinder is two rigid planes inclined at an angle 2 a. at time t. 235 The is initial temperature of a homogeneous solid sphere of radius a given by Vq = Ar~^ its cos 6 (sin mr — mr cos mr) is : prove that at time t temperature provided that m is a root of the equation {ah — 2k) (ma cot ma — 1) = mW^. resultant pressure of the air on the bell zero position of vibrating in a given manner as a rigid body about a equilibrium which is at a given distance from a large perfectly rigid obstacle whose surface is plane . k. determine the motion at any point in the air. where aS'^ is a spherical surface harmonic of degree n. 34. ^ I . i being an integer and c the velocity of propagation of plane waves in air. kz]. z) of the air in the cylinder t/ '2qpaP cos pO'^ J k ^ (nr) 2\ (n^a^p^)Jp(na) ^^ r / / \ or cosnct{e'^~ * sin. the velocity potential at any point is 1 (r. Prove that. whose coordinates. point. referred to the centre as origin and the bisector of the angle of the sector as initial line. is qr^ cos p9 cos net. are r. and a is air. 6. and the surrounding at zero temperature.) vibrating in such a manner that the normal component of the velocity at any point of its surface is Sn cos kat. where pa = iTr. A sphere is 36. and is closed at one end bounded by by a flexible membrane which is so that the velocity at any forced to perform small normal oscillations.EXAMPLES. medium being 35. 0. is the real part of the expression r (1 + ikc)fn (ike) — ikcf (ike) ' where Pn denoting the zonal harmonic of degree n. is except Show that the when n = l. (Weber. A spherical bell of radius c the velocity of transmission of vibrations through the surrounding Prove that the velocity potential at any point outside the bell at a distance r from the centre.
Jo Jo <^ by properly determining (u) so as to satisfy the boundary conditions. a being the radius J!p (no) the of cylinder. . a general form of the current function is where An. = 0. Show integral also that if the cylinder were surrounded by viscous liquid the solution of the problem might be obtained from the definite " e Xu(fi (u) t/j (Xu) Ji (Xr) du. air. with respect to the axis r = 0. cylinder of radius a is filled with viscous and made to rotate with uniform angular velocity w about liquid at time t is its axis. the upper and lower sign before l(^ corresponding to the first and second term in the bracket respectively. which initially at rest. A force suddenly cease to act. quantities.236 where EXAMPLES. Prove that the velocity of the Ae/i (Aa) where the difierent values of A. p. prove that the frequency sO. is /> of the electrical oscillations about the distribution of electricity where the surface given by the equation density proportional to cos ^8 {P<^h) = 0. given mass of air is at rest in a circular cylinder of radius c of a constant force to the axis. then the velocity function sequent time varies as at any sub where a is is the velocity of sound in values of k satisfying J^ (kc) neglected. 38. = gives the requisite values of n. are the roots of the equation J^ (Xa) = 0. 116. (Cf. where v is the velocity of propagation of electromagnetic action through the dielectric inside the cavity. n are arbitrary complex 40. the summation extends to all and the square of the condensation A right is circular liquid. and where ^ is a real quantity given by the equation n'^ — n^ ± k"^.) is A is right circular cylindrical cavity whose radius a is made in an infinite conductor . In twodimensional motion of a viscous fluid. symmetrical 39. the action Show that if the under 37.
if/ A distant r from the axis. moving forwards with velocity V at right angles to its axis. r the distance of the point illuminated from the geometrical image of the source. is suddenly stopped the velocity potential inside the cylinder at a point prove that . Prove that the integral of the expression in the preceding example taken for a linesource involves the evaluation of an integral of the form 45. 41. — 0. be the origin plate of magnetic permeability /x and thickness c : A m of rectangular coordinates x. and x be perpendicular and y is parallel show that O the potential behind the plate given by the equation ^ where Jo 1p'e /x1 f^ + 1 43. Ji(ax)J^(bx) ^^ /: I . if the opening of the objectglass of the telescope problem considered at p. and the summation Jj' (ko) is taken for values of k which satisfy the equation 44. pE the inner radius of the opening.EXAMPLES. may be expressed in the form mra Jo the upper or lower sign being taken according as z positive. magnetic pole of strength if w. y.TT~\ cos Kat. 237 of radius c Prove that the current function due to a fine circular vortex. to the plate. the direction in which the cylinder was moving. . 178 above be ringshaped the intensity of illumination produced by a single pointsource at any Prove that in the diffraction point of the focal plane is proportional to (1 \i pY is z" z= 27rJirlXfj where B the outer radius. Jj^ /» i/r ^ Tr =2 K cos *^l ('«*') ' (ko) where a all is the velocity of sound in air. and strength m. ^ and where the radius makes an angle with is given by the equation uj^. — z' is negative or is placed in front of an iron 42. and/ the focal length of the objectglass. right circular cylinder of radius a containing air.
. 53 above].2a6 (Struve. A ±1L_ A jj^iMVrdr 49.J) yo Va2 + 52 _ 2(j5 cos <^ Hence prove that .) 48. and [using 108. jjJ^(v.. ru be <^ (r). ij/ (r). Show p.+^zr. . EXAMPLES. and p the density. if the initial values of ru. r JJax)JJbx) 7« ah f^ sin (^ Ja"" + 6^  2«6 cos cos <35) <i)) . xJx^$^ ^^io a^ + 6^ .:n+9^nO df dt dt 0. Show that 2" ^TT 47. . prove that. that for the motion specified Bp = Bp = . the equation of motion d^u is _k + ^n p /d^u fo^u 4 du\ du'' df with the surface condition \ ^ ' dr r is Prove that the complete solution subject to the condition stated u= 2^J^ {rjp) {Ap sin Cpt + Ap cos c^t} +2 where sin Cpt T «^_ fe) {3 i^P Vp + B'p cos Cpt)^ Cp being the pth root of the equation [r}^ = c^pa^/{k + n)] which holds at the surface r == a of the sphere. n (?i .)rrdr Obtain the equation of motion of a simple pendulum of variable length in the form d^ "^ d^dO = ^:jr.. radial direction that if strained symmetrically in the perform radial oscillations show u be the strain at distance r from the centre. k and n the solid A isotropic is sphere is and then left to : bulk and rigidity moduli.238 46.
) 51. where a and b are constants. C.6)]^ ^TT . "^ . . . this is not the case the value of the integral a+b—c all positive and that if (Sonine. u by means of Bessel functions.J) 2=*"^! n (7^1 are is zero. Jan.^ (cos <^) a* (l2aCOS</) + a2)"= prove that Jn {si a? \h^ . Solve the equation in that x = gl/h^ when bjjga is small. 15. is The problem (See Lecornu. p and (0 being arbitrary constants. the equation of motion for the small oscillations may be written where u = W.2ab cos <^) = 2m in . provided that 6 + ca.EXAMPLES.^3 ^y (Gegenbauer.) 50. suggested by If the functions C" (cos <j>) are defined by the identity i"" C. the swaying of a heavy body let down by a crane. . and show that ii l 239 =a+ ht. R.) I . 1894.) %^ J^ « b' c^ (. . c + a— 6. .1) '1 8=0 {n . . a^b'^c'^ . Prove that if 9^ >m> — 1 I a a>b 52. J^ (bx) Jn (ax) x"'''+'^dx = — b"" {a" ^„ J. by^/ if .) If 1 j J^ (ax) J„. and prove we have approximately '('©'"(Vl) ^~7fe(''?)~(v'f'). (Sonine. and that the value 7^1 >  of the integral is zero a< 6. (bx) J^ (ex) x''^dx _ [{a + b\c){a + bc){b + ca){G + a.
P. 55.n (i) /^^^Vo(rsin^) = :Sf^. with the notation of example (Hobson.. L.' Prove that / e/^ (r sin 0) sm'^^'OdO = ^+r = . S. . Froc. J\ fj J«_i /" (r sin 0) sin»+ierfe '^»).i__^ log ^±^^?±Z . M. (ibid. (Xp) <^X  . 7.) 54. Verify the following expansions 00 : — . .240 53. xxv. Prove that 2 r* sin (w + r) Jo u+r ^ ' (Sonine and Hobson.(cos^). Wi (r sin ^)^ ^^ 50. EXAMPLES.
m = 4w. in order of magnitude. of the equation ^s)_ m+3 4 (7m^ + 82 m .153855^^ + 185743m .9) 32 (83m^ + 2075771^ . m = 4?^^ 16 M. M^Mahon obtains the following important results. in order of magnitude. shortly to be published. Jtt (2n + 4s + 1). The sth root. [n a paper. in order of magnitude. of the equation («) m+7 ^ 8y 4(7m2+ 154m + 3(8y)3 95) " 32 (83m^ 4.6277237) 105 {s^y where (ii) i^ = J tt {2n +451 ). Professor J. y= G. 7n=4:n^ The sth root." for a MS.'ORMUL^ FOR CALCULATION OF THE ROOTS OF BESSEL FUNCTIONS.3535m^ + 3561m + 6133) 15(8y)3 where.3039m + 3537) 15 (8y)^ (where (iii) y = Jtt (2n + 4s + 1). as above. copy of which we are indebted to the author. "On the roots of the Bessel and certain related functions. . The sth root.982m + 3779) 15 (s^y 64 (m  1) (6949m^ . of the equation (a) IS aj„ P m1 8/3 4(ml)(7m31) 3 (8l3f 32 (m  1) (83m^ . the (i) first of which has already been given in part.
(vi) The sth root. p>l. . of the equation Y. Y. in order of magnitude.{x)^{y\og2)J.{x) J^ipx) 2) where . x^^^ in (i) if (3  ^tt be therein substitute( for (v) The sth root. in order of magnitude. which is the same.log Y^jx) J^{X) J^ (x). in order of magnitude. but with ' I _ ^~ m+3 8p '^ _ 4(m^ + 46m63)( p^l) 3(8p)^(pf) • _ "* 32(m^+ 185m^.4 (m Pi'^~ 8p 32(ml)K114m 5(8p3)(pl)  1) (m25) (p^. in order of magnitude.1 )  HSpfipi) + 1073)(pl) .2053m + 1899)(p^l) 5{8pf{pl) [Of course here also the G functions functions without altering the equation.] (viii) may be used instead of th' The sth root.m4n.{x) = is given by the series for /3. (vii) The sth root.242 (iv) FORMULA FOR CALCULATION OF THE The sth root. of the equation — — = 0. of the equation ^' P '' J.(px) Jnipx)""' where 8^ v"^^ '^.G^ (x) = F„ (x) + (y . of the equation ^{r„(a=) is + (ylog2)/„(x)l=0  given by the series for £c^*' in (ii) if y Jtt be therein substituted for in here the differentiated is of course Euler's constan expression [The y is and not to be confounded with the y in the expression for the root. or. of the equation Kjx) is ¥'„{px) given by the same formula as in (vi). in order of magnitude.
it is equivalent to Sx ^^^^^31:2^' and other equivalent equations can be obtained by means of the Table on p. and x = 231. 42 above. 56. Vol. II. 3. The equation is the expression of the surface condition which the motion must fulfil. 'hen =# the equation is equivalent to tan x = Xj nch occurs in many problems (see pp. The equation of which the roots are given in (ii) is also of great importance for physical applications. Sound. 219. It expresses the condition that there cylindrical boundary. kinds of Examples sical the equation in are found in applications. fore give the possible values of seq. 113. but with ^' _m+7 Sp _4 (. is no motion of the gas across the [See Lord Rayleigh's Theory of Sound. The roots of this equation can therefore be calculated by the formula in (i). 16—2 I .] all 'he following notes on these equations (i) may be useful. et kct.>t^)70m199)(p^l) '^3(8p)3(pl) (p^ _ 32 (m^ + 245m^ 3693m + 4471) 5(8p)=(pl) [As before the  1) G of functions may here replace the Y functions. Vol. above). The roots of the equation given in (iii) are required for the problem of waves in a fluid contained within a rigid spherical envelope. 178. p.] When n = ^.. see pp.ROOTS OF BESSEL FUNCTIONS. ii. the equation equivalent to tan x^x. given also by the equation in equation is equivalent to (i) when w = f = 2x 2ar^' . Again when ri = the taxix which gives the spherical nodes of a gas vibrating within a spherical envelope. = . When n = \. 191. 96. ?i and elsewhere above. for example it gives the wave lengths of the vibrations of a fluid within a right cylindrical envelope. 265—269. pp.) is The roots thereLord Rayleigh's Theory of (See is the radius. where a k. 30 243 given by the formula in (vi). 2. and when n the equation is equivalent to tan X = 2x.
x and px are the values of xa.S 8 while. k6.h. The values of x and px are as before those of xa. of K for the problem. In conclusion. . equation thus give the possible values of k for the problem. it may be preferable to interpolate the values from Tables of the numerical values of the functions. k6. 5. 7. In (viii) the equation given is derived from the conditions 6. are k^ h ^3 = 240482 55577 = 552007 81103 = 865372 79129 =1179153 44391 k^ ks kg k^ = 1493091 77086 = 1807106 39679 =2121163 66299 ks k. where a. The roots of the equation in (vi) are required for many physical problems. It is the proper extension of (iii) for this annular space. where a. k6.S' +h. for larger values of n. or the vibrations of an annular membrane. where = :r . for example the problem of the cooling of a body bounded by two coaxial right cylindrical surfaces. h are the internal and external radii. if these are available. . (See p. ^„ z= (n  1) TT k. where a. 99 above. it may be stated that the ten first roots of J^ {x) = 0.S^ + h. which must hold at the internal and external surfaces of a fluid vibrating in the space between two concentric and fixed spherical surfaces.244 ROOTS OF BESSEL FUNCTIONS. = 2435247 15308 =2749347 91320 ^10= 3063460 64684 + h. As before.) The values of x and px are those of The roots of the KCt. The ation of the wave lengths roots of the equation in (vii) are required for the determinof the vibrations of a fluid contained between two coaxial right cylindrical surfaces. as calculated by Dr Meissel and given in the paper referred to below. . LogAi = 859976 01403 Log 7^2 = 7 41 558 08514 Log 7*3=690532 68488 Log 7*4= 678108 01829 Log 7*5 =692939 63062 and LogAj means the common logarithm of h^ increased by 10.S' . The roots thus give the possible values If for low values of s the formulae for the roots are any of them not very convergent. 4. h are the internal and external radii. 8^  . h are the internal and external radii..
placed at our disposal by its author. x. for integral values of x' n and less the values of J^ from x—1 positive to The table may be used. when x is given. all Dr Meissel. Table II.. with the corre(x) I . only change that has been made is to write J^ (x) and Jj (x) instead of I^ and Ij^. among other purposes. with ever increasing rapidity. there there a special value of n for which the function J^ (x) changes sign for the last time from negative to positive. and then diminishes as n increases. that. y + \ we may put x — y + h. and the value of Jo(l7l) has been altered from •3932.. We are indebted to Dr Meissel and the Berlin Academy of The Sciences for permission to include this table in the present work. Three obvious misprints in the column of arguments have been corrected. [Table loneii I^ I. III. is a reprint of /. which is taken from the first of gives the first 50 roots of the equation J^ the papers just referred = 0. Table to. = which are not than 10~^^ Jn {^) = Jn (y) + ^^Jn {v) + 2I ^n (2/) + • • • = Jn (y) + J^ {^ Jn (y)  Jn^l (?/)} We functions take this opportunity of referring to two papers on the Bessel by Dr Meissel contained in the annual reports on the OberIt is is Bealschule at Kiel for the years 1889—90 and 1891—2... shown.EXPLANATION OF THE TABLES. is derived from an unpublished MS. that the function then increases to its absolute maximum.. and then 24. (x). very kindly It gives. to 3922. Thus if X lies between two consecutive integers y. among other things.. in accordance with a communication from Dr Meissel. for the calculation of J^ (x) when x is not integral.V' Dr Meissel's "Tafel der Bessel'schen Func und originally published in the Berlin Ahhandlungen for 1888.
and YI.. of course. (x). : These tables do not require any special explanation the functions /„ are the same as those denoted by that symbol in the present work. so the tabular difference has been taken to be "01 instead of "001. . maximum or minimum (x) according as they are positive or negative.246 EXPLANATION OF THE TABLES. V. was thought too long to reprint. are extracted from the Reports of the British Association for the years 1889 and 1893. which are. Tables IV. The Association table corresponding to Y.. spending values of Jq values of Jj.
.
0440050585745 0760780977632 0538274I033C 0540I305444S 0572265544440 0566855120374 054I9477I39: 0561426721439 0550517577543 02 103 I 04 0443285761209 0446488193730 0751851323654 0449657657556 0747339037965 0452793929666 054372550001 0756332080477 0555980742014 054546379432 .0426787063833 0430150899695  053444386841 0778251432583 053637850525 098 099 I'OO ioi I 0773933236862 0433482922506 0769581880965 0436782895795 0765197686558 .0547 i6249i6c^ 054882 1 4901. 05656003207. (continued). 05579365079: 05593150345' 05619508257! 09 0467969801675 iio I'll 0719622018528 0714898476008 .{x) o8o 08 1 0846287352750 0842579716344 0838833832154 0368842046094 0372680644052 0671132744264 049828905756 06661 37 I 20084  050082967264 082 0376491556779 0661116273214 0656070571706 0651000385275 0645906085271 0640788044651 050333356702 083 0835049978414 0831228436109 084 085 086 087 088 0380274508136 .248 ^o(^) Table J^{x) I.0384029224303 0387755433798 0391452867506 0395 1 21258696 050580057262 050823052439 051062326032 0827369488950 0823473423352 0819540528409 05I297862I46 0635646637944 05 1 529645 197 051757659906 089 090 o'9i 0815571095868 0398760343044 081 1 5654201 10 0402369858653 0807523798123 0630482241224 0625295232074 0620085989562 0614854894203 0609602327933 05I98I89I306 052202324741 0405949546079 0409499148347 O413O184IO976 0803446529473 0799333916288 052418945868 052631740655 092 093 0795186263226 094 095 0416507081996 0791003877452 0419964911971 0786787068613 0782536148813 0604328674074 052840695388 0599034317304 053045796666 0593719643626 0588385040333 0583030895983 0577657600358 0423391654020 053247031406 096 097 . 0567829793905688833061: ri6 II7 0690855713099 0681046860871 0487762857858 0493098352841 049571 1892924 0685964625798 0490448614448 0676102779403 0671 132744264 0472452557702 0466779697485 0461096097935 ii8 II9 [•20 0498289057567 0455402167639 j 056989593521 . 0495028451994 0489402624312 0483764428365 04781 14270507 070536532581 1 0700556419592 0695719763505 0479491324496 0482284025373 0485041 266 1 54 II4 115  o5632079488( 05644244679.05550585709' 05565176603. 05667354480. «^o(^) J.0470902394866 0473800312980 0476663355426 II2 II3 0710146128520 0500641505700 05606531646. 05504406911^ 05520199991: 05535593220: I 0545037624510 0539541280398 105 I I I I 0742795556434 0455896789778 0738221214269 0458966020374 0733616348841 04620014067 1 5 0728981299655 0465002736858 0724316408322 06 07 0534028943664 0528501013700 08 0522957890804 0517399976146 0511827671736 050624 1 38039 1 .
0579446250290 0247007069667 0530467081267 0528480052675 0578982519892 236 237 238 0023482827990 0018208107961 0241219731308 0235437635298 0229661 184046 2*00 0578478431892 0577934063221 0012953780380 0007720190934 0002507683297 0526458158577 05244015961 19  0577349494047 0576724807757 239 2*40 0522310565146 0223890779141 0520I85268I82 .Table J. {continued).{x) I. 249 J.392204369660 2IO 2.1 1 0568292135757 0160929324324 0155262458341 0567232577628 0566134151091 0564996980564 0386418479668 0380627601627 212 0374832147732 0369032530185 0579323466925 0579760442028 213 214 215 0149606770449 0143962647452 0138330474865 0132710636881 0563821193544 0562606920596 0580156197639 0580510738087 0580824071043 0363229161163 216 217 2i8 0357422452782 0127103516344 0121509494713 0115928952037 0110362266922 0104809816503 0561354295339 0560063454436 0558734537577 0351612817064 05810962075 15 0345800665906 0581327161851 0339986411043 03 34 1 219 220 221 0557367687469 0555963049819 0554520773326 0553041009659 055I5239I345I 0581516951731 704640 1 6 0328353236143 0581665598167 0581773125501 0581839561397 0581864936842 0581849286141 222 223 0099271976413 0322535138478 0093749120752 0088241622061 03167165S1784 0310897976496 0305079732690 0299262260050 0293445967833 224 225 0549969642278 0082749851289 0548378356647 0077274177765 0581792646910 0581695060074 0581556569863 0581377223803 226 227 0546750219981 0071814969172 0066372591512 0545085398603 0543384061721 228 229 0286631264839 0060947409082 0055539784446 0541646381412  0281818559374 0276008259222 0270200771606  0581 157072713 230 231 0539872532604 0538062693065 0580896170703 0050150078400 0044778649952 0039425856288 0264396503162 0258595859901 0580594575158 0580252346743 0579869549389 332 233 234 235 0034092052749 0028777592796 0536217043381 0534335766941 0532419049921 0252799247180 .{x) M^) 20O 2OI J^{x) 0576724807757 0455402167639 0569895935262 0223890779141 0449698315660 0443984951500 0438262485071 0570867635566 0571798364542 0572688083032 0573536755217 0218126821326 0212369710458 0200877624399 0195143444226 0189417700977 0183700789621 202 203 0576060090955 0575355433450 0206619845483 0574610928248  0432531326660 0426791886896 042 10445767 1 5 204 2*05 0573826671543 0573002762707 0574344348624 0575110834122 0575836185927 0576520381599 2o6 207 208 209 0572139304279 0571236401957 0570294164587 0569312704151 0415289807326 0409527990183 0403759536945 0177993104055 0577163402048 0577765231529 0578325857645 0578845271345 0172295037073 0166606980332 0397984859446 o.
250 .
Table i I. 251 . (continued).
252 .
.
{x) I. + 0197686482 3 592 5"93 5 94 + 0124976133333 + 0127901450011 + 01 308090501 95 + 01 3369868 1 5 24 + 0293409028587 + 0291650066443 + 0289865764324 632 6'33 :  + 0288056330291 634 635 1 595 + 0136570093728 + 0286221974417 + 0139423038646 + 0142257270250 + 0145072544661 + 0147868620168 + 0284362908764 + 0282479347366 + 0280571506204 + 0278639603186 + 0233891994542 + 0195047873 596 636 637 5*97 598 5*99 638 639 640 + 0235829218223 + 0192394409' +0237739834141 + 0189726366 + 0239623697870 + 01 870440 1 7 + 0241480667734 + 0184347639 + 0243310604823 + 0181637509 i i 600 + 01 5064525725 1 + 0276683858128 . + 0202918049 + 0200309965.{x) + 0026970884685 + 0334332836291 + 0030309828079 + 0033639796739 + 0036960484490 + 0040271586530 + 0333450716975 + 0332537912108 + 0331594557948 + 0330620793320 + 0150645257251 + 027668385812 562 563 602 603 564 565 604 605 + 0153402218596 + 027470449272 + 0156139269116 + 02727017305.{x) 600 601 J.254 J. J. 624 625 i + 0107066577404 + 0303419269333 + 01 10092778957 + 0113102820941 + 0300187387793 + 0116096438881 + 0298531825185  586 587 626 627 588 589 + 0119073370272 + 02968501 10895 + 01 1 2033354593 + 0295142444729 + 0215284812471 + 0217454557531 62S + 02195991 16876 629 + 0221718321770 630 631 ' i 590 591 + 0223812006132 + o2o8o8694oH + 0225880006549 + 0227922162289 + 0229938315309 + 0231928310269 + 0205510464. + 0158856175969 + o27o6757969e + 0161552708575 +02686269192: 40164228638636 + 0043572799459 + 0329616759609 + 0046863821304 + 0050144351544 + 0053414091135 + 0056672742533 + 0328582600738 + 03275 184631 59 + 0326424495830 + 0325300850207 + 02665553263: 566 567 6o6 607 6o8 568 5*69 609 + 0166883740153 + 02644612490 + 01 695 1 7789443 + 02623449199 + 0172130565159 + 02602065 732< + 0174721848302 + o258o464448< 570 + 0059920009724 + 0324147680223 571 610 6II + 01 77291422243 + 02558647725 + 0179839072737 + 02536617955 + 0182364587942 + 02514377548 + 0184867758430 + 02491928929 + 0187348377209 02469274539 \ + 0063155598244 + 0322965142271 572 + 0066379215205 + 0321753395193 573 + 0069590569321 + 0320512600255 574 + 0072789370930 + 0319242921139 575 +0075975332017 612 613 614 615 + 0317944523919 + 03 1 66 1 75 77048 + 0315262251336 + 0313878719939 + 0312467158333 + 0189806239737 + 02446416835 576 577 578 579 580 581 + 0079148166242 + 0082307588961 + 0085453317250 + 0088585069926 616 617 618 619 620 621 + 01 9224 1 143934 + 0194652890201 + 0197041281434 + 01 99406 1 23040 + 02423358290 + 02400101392 + 02376648642 + 02353002557' + 0091702567575 + 0311027744304 + 0309560657922 + 0308066081529 + 0306544199716 + 0304995199305 + 0201747222949 + 02329I6567C + 0204064391629 + 02305140526 + 0206357442103 + 02280929684 + 0208626189957 + 0225653571C + 0210870453362 + 02231961215 + 0213090053077 + 02207208775 + 0218228101C + o2i57i8o54( + 0213191002: + o2io6472o8( 582 + 0094805532571 + 0097893689100 583 + 0100966763183 584 + 0104024482698 585 622 623 .(x) 5*60 56i Table J. {continued).
.
256 .
.
1 14239232683 . 0111907971956 0233817816088 o.0223664044201 956 957 9i8 0130154834519 0222 1 3042 1 159 0132368386105 0220576532901 958 0206143674127 0207572313841 919 0134566298203 0219002552542 9'20 959 0136748370765 0*217408654960 960 0208978718369 0139524811J I . (continued).0097639022336 024I68I558953 0100049580330 02404263031 1 1 942 943 0178578063718 0180364787772 0182131589336 0183878314938 944 9'45 946 947 0102447473906 0239148699952 0104832480333 0237848898088 0107204378374 02365270481 19 01095629483 10 I833026I2 0235 0185604813228 0187310934989 0188996533147 906 9*07 oi675287869( 908 909 948 9*49 0190661462784 0192305581154 0193928747687 016545476 1 3.1 18859606752 910 9II 950 01612644307 01591485641 0231022247743 0229592484581 0228I4I6I7686 951 952 953 912 913 0195530824010 0197111673948 0198671 163543 01570191907 01548765325 01527208 1 25 0121148294781 914 915 0123422369306 0226669811094 0I2568I62I757 9*54 955 0200209161060 0201725537001 02032201641 14 0204692917400 0225177230692 01505522548 01483710843 01461775272 01439718106 0141 7541625 916 917 0127925845233 .0259467700807 0057534966296 0258613087962 .258 Jq{x) 88o 88i Table J.0232430745006 o.1695884292c 902 903 904 905 0092780811380 02441 2444426 1 0095216024131 0242914320868 .{x) ^1 0039233803177 0041871036007 .00601 16721752 0257733920049 0062689563221 0256830309615 0065253246851  0143197577219 0145314326565 0147414262841 924 9*25 0212509233706 02 1 083 74506 1 2 0209146650470 0207437018341 886 887 926 927 0149497 196801 015 1562941057 0205708740917 0203962006501 888 889 928 929 015361 1310096 0155642120296 0157655189943 0202197004987 0200413927844 •0198612968091 890 891 0255902371444 930 931 892 893 894 895 0254950222539 0253973982110 0252973771561 0075411566939 0251949714476 0070352170997 00728869296S9 0067807529947 0159650339244 0161627390345 932 0196794320281 •0194958180481 • 933 0163586167343 934 935 0165526496306 0167448205283 0169351124322 0171235085481 019310474624^ 0077925844909 0250901936605 0080429527028 019123421661 896 897  0249830565850 936 937 938 018934679206 0082922378016 0248735732253 0085404163904 0247617567976 0087874652054 0246476207294 009033361 183 1 018744267450 018552206727 898 899 0173099922846 0174945472543 0176771572752 939 018358517507 900 901 0245311786573 940 941 018163220400 017966336149 017767885629 017567889848 017366369941 017 163347 17c o.0260297648278 0054944541843 .{x) I.1 1 65 565 1 5436 o.0264073703240 0263368657691 0262638429381 9*20 921 0136748370765 01389 14405500 0217408654960 0215795016778 0214161816342 882 883 0044501092388 0047123720982 922 923 0141064205893 884 885 026I883II7I96 0049738671456 0261 102822332 0052345694498 . J. 01633665697.
0048474475654 0245488525942 0045974634906 0245935764451 0245846278846 + 0043478298146 037 0245399424757 0244928473884 038 0244433496098 039 0243914563561 036 +0045891393496 + 0048297616575 +0050696737897 + 0053088528877 0043472746169 040 0243371750714 + 0055472761849 17—2 .1 1 89722444 1 7 0221795482032 o.I J.{x) J. 259 J.0075766924729 0*26 027 028 0248773226477 + 0021489469834 0248546005495 +002395395I2I7 0238924318032 0073305907338 0239645052073 0070840 1 2 783 1 0240341 105535 024IOI 2434487 0241658997463 029 030 031 0248294163353 + 0026413662567 0248017748933 +0028868367285  0068369832284 0247716813482 +003 13 1 7829476 0065895266972 0063416678354 0060934313045 0242280755465 0242877671958 0058448417794 0247391410602 +00337618 13968 0247041596243 + 0036200086339 033 0246667428695 +0038632412933 034 0246268968580 + 0041058560885 032 035 0243449712877 0055959239457 0243996846626 0053467024979 0244519044079 0050972021363 0245016278580 .1 35031922668 0213063373585 0132768846695 0130494992737 02I437970I667 02I567323829I 0246357974862 0040969056455 0246755140400 0038463812722 003 0247127246760 0035957261846 0*04 0247474282103 0033449650599 002 005 0128210594048  0247796237059 0030941225625 02I6943879I79 0I259I5884679 02I8I9I522398 0123611099451 0219416068367 o.{x) 0208978718369 0139524811741 QOG ooi 0245935764451  0043472746169 0210362771833 0I372839882I5 021 1724360660 o.{x) I.0010874068044 0008369385737 0005866343931 0107215354950 0*14 015 0249564887171 0227333299512 0I0483850I258 0228369767107 01024537 15952 0229382348196 0100061240280 0230370967084 0097661316004 0249636064351 016 017 018 0249682220330 0003365I863I4 0249703375168 0000866 1 56 1 65 0249699551355 + 0001630503669 0231335550495 0095254185376 019 0249670773804 + 0004124550795 0249617069854 +0006615743298 0232276027579 0092840091 1 13 0233192329916 0090419276375 0087991984743 0234084391517 020 02I 0234952148834 0085558460188 0235795540759 0236614508629  0249538469258 + 00091038 39761 0249435004182 + 001 1 588599292 023 0249306709197 + 0OI406978I546 022  0083118947058 0080673690044 0*24 025 0249153621275 +OOI6547I46743 0248975779783 + 0019020455697 0237408996230 0078222934162 0238178949800 . (continued).1 21296473933 0220617419863 o.1 16638647900 006 007 008 0248093104724 0028432233416 0248364880658 .0025922920290 024861 1 562881 0023413532364 024883315 1876 0020904315537 0249029650581 009 oio oii 00183955 15458 0222950162390 0114295922054 0224081370836 oi 1 1944305207 0225 I890I9654 0I095840363I7 0226273023521  0249201064392 012 0249347401 155 013 0249468671 167 0015887377509 0013380146780 .{x) Table J.
{x) I. (continued).1 7678529895' 0227635047621 + 0IOI228662586 022661 1876971 +0103403435462 0225567019886 + OI055659I5987 0224500600296 +0107715903254 [063 [O64 0223412744130 +0109853197747 [062 [065 102 103 0I694I4723099 + 017832699423. 0137045758956 + o20i5i49622< 0135024693407 + 02026946301.1 9585943813 1 + 015 1466762702 [046 1047 0239617840978 0238179235177 0237425221101 + 0069606813400 0238910139979 +007I93I847339 +0074247537568 086 087 088 0194335703428 + 0153277381926 o 192793946683 + 0155071144022 0191234337275 + 0156847888004 0189657046181 [048 1049 +0076553665638 089 + 0158607454682 10*50 [051  0236648194462 + 0078850014227 090 091 0188062245963 + 01 6034968668 1 0186450 1 10748 0235848254136 + 008 1 1 36367 1 58 0235025501155 + 008341 2509421 0234180038696 + 0085678227I9I 023331 1972068 +0087933307849 + 0162074428448 1052 092 093 1053 01848208 1 6208 + 0163781526274 0I83I 74539542 + 0165470828298 0I8I5II45946I to54 1055 [056 1057 094 095 + 016714218452^ 0232421408701 + 0090177540002 0I7983I756I65  + 0. 10 '80 0203201967112 + 0142166568299 120 0132991936860 + 0203853 1458< . 0171190300407 + o.260 Jq{x) [0'40 [0*41 Table J.0207380739286 + 0136394852837 [078 .0206007080560 + 0138334322500 [079 0204614104523 + 0140258285937 [076 ii6 II7 OI4IO5 1973900 + 01990925621: ii8 II9 0139054922470 + 02003142397. 0167623824113 + 01 798496 1 246 0I658I 7794883 + 018135302100 0163996828 161 104 105 + 018283708920 0222303579310 + 0] [977601366 0I62I6III7996 + 018430168840 01603 108597 1 2 + 018574669196 015 844624989 1 [066 0221173235728 + 0114088917441 0220021845238 + 01 16 1 86950748 0218849541635 106 107 [067 [068 + 018717197526 + 0118271507531 108 109 0156567486350 + 018857741568 0154674768122 + 018996289269 [069 0217656460650 + 01 203423955 1 5 [O70 [07I 0216442739924 + 0122399423927 0215208519001 iio III 0152768295436 + 019132828777 0150848269694 + 01926734844^ OI48914893455 + 01939983684: 0146968370410 + 01953028273: 0145008905360 + 01965867509J 0143036704202 [072 0213953939309 [073 0212679144146 [074 021 1384278663 [075 + 0124442403513 + 01 2647 1 146550 + 01 2848546687 1 + 01 30485 1 79874 1*12 113 I 14 0210069489850 + 0132470102543 II5 + 01978500312^ 0208734926518 + 0134440053463 [077 .16879544685c 0231508458131 + 00924I07I3500 0230573231989 + 0094632619458 to58 0229615843992 + 0096843050272 [059 0228636409922 + 0099041799642 [o6o [061 0I78I356II325 + 0170430469041 0176423208066 + 017204710678^ 098 0174694730946 + 01 736452 1 767: 096 097 099 I'OO I'OI 0172950365937 + 01 7522466 1 24. J^oC^) Jiix) 0243371750714 +0055472761849 0242805134273 + 0057849210087 0242214793214 + 0060217647828 o8o 081 0203201967112 +0142166568299 0201770826005 + 01 44 05 89964 1 5 0200320840603 + 0145935398812 0198852172014 + 0147795605727 0197364983034 + 01 49639449 1 22 [0*42 082 ' [0'43 fo'44 10 '45 0241600808767 +0062577850293 0240963264405 +0064929593703 0240302245833 083 084 085 + 0067272655308 o.
261 Joi^) J. (continued).Table Jo{x) I.1 034261 3 1 706 + 0217801937572 001200062838 1  +0233170730054 +0233080588274 0101243938632 + 0218632946448  0009669352567 0099053547496 + 02 1 9441 539632 0096855182759 + 0220227656988 0094649069469 + 0220991240623 0092435433245 + 0221732234896 0090214500248 + 0*222450586415 0087986497163 + 0223146244045 008575 165 1 176 0081262341601  0007339093458 +0232967393973 00050 1 008 1 428 + 023283II776I9 0*002682546537 + 0232671971904  0000356718505 + 023248981 1 743 + 0001967173307 + 0232284734267 + 0004288899920 + 0232056778820 + 0006608232761 + 0231805986948 + 0008924943683 + 023 532402401 + 001 1238804987 + 0231236071121 1 + 0223819158911 + 0225096576153 + 0225700992096 •82 0083510189950 + 0224469284397 •83 84 •85 0079008334679 + 0013549589443 + 023091 7041237 0076748398145 + 0226282492413 0074482761342 + 0226841039560 007221 1653982 + 0227376598268 86 •87 88 89 90 91 0069935306115 00676539481 12 + 0227889135543 + 0228378620665 + 0228845025194 + 0229288322968 + 0229708490101 + 0230105504990 + 0015857070317 + 0018161021385 + 0020461 2 6961 + 0022757431916 1 + 0230575363062 + 0230211089083 + 0229824273953 + 0229414974489 + 0025049441700 + 0228983249662 + 0027337022362 + 0029619950574 + 0031898003653 + 0034170959578 + 0228529160587 + 0228052770520 + 0227554144849 + 0227033351083 0065367810637 0063077124631 00607821 21 280 92 •93 .0058483032003 •94 •95 0056180088419 + 0230479348310 0053873522332 + 0230830003018 + 0036438597013 + 0226490458847 + 0038700695332 + 0040957034634 + 0043207395768 + 0045451560353 + 0225925539874 + 0225338667993 + 0224729919124 + 0224099371266 96 '97 0051563565704 + 0231 157454348 0049250450632 + 0231461689817 0046934409328 + 0231742699216 0044615674094 + 0232000474620 98 99 200 + 0047689310797 + 0223447 104491 .1 142063 14208 + 0213312919188 01120684561 10 + 0214255026208 0032989723038 +0232940723529 0030659704782 +0233059055883 0028328619340 +0233154162418 0025996698919 +0233226053376 0023664175616 +0233274741260 002 1 33 1 28 1 388 +0233300240831 0018998248037 +0233302569105 00 1 6665 307 1 80 +0233281745349 0014332690232 +0233237791079 010992 1287289 + 0215175051739 0I077650289I8 + 0216072924488 0105599902872 + 0216948575381 o.{x) J^(x) 0132991936860 +0203853145865 0130947701315 + 0204990414012 0128892199715 + 0206I0634I4I6 0126825645926 + 0207200837037 0044615674094 + 0232000474620 0042294477301 +0232235010376 0039971051364 +0232446303109 0037645628720 +023263435 1 719 00353 1 8441 806 +0232799157379  0124748254710 +0208273812006 OI2266024I7II +0209325179625 0120561823424 +0210354855380 0II84532I7I84 +0211362756947' 0II633464II33 + 02 1 2348804 1 93 o.
12*36 12*37 12*38 12*39 12*40 + 0*122219116616 + 0*186339651802 +0*124075619437 + 0*184958024768 + 0*125918219608 + 0*183559146848 +0*127746745377 + 0*182143169785 + 0*129561026518 + 0*180710246883 + 0*184087205211 + 0*185270010670 2*78 + 0*186433370658 279 + 0*187577184813 2*76 2*77 + 0*11924942413 + 0*11730996174. {continued).66 2*67 268 2*69 12*29 12*30 12*31 + 0*194258848041 + 0192983823702 + 0*191690671617 + 0190379533851 + 0*189050554121 2*70 2*71 + 0*176587888562 + 0*13066222900.106887699579 + 0108849062765 +0*110797950308 + 0*200357198756 + 0199174885273 + 0197973759015 + 0*196753953565 + 0*195515604234 + 0169823630622 +0*139858114750 + 0171213147086 + 0172584452006 + 0173937423535 + O175271941729 + 0*1 38043 II 5 84^ + 0*136215837361 + o*i3437646923c^ + 0*13252520245^ 12*26 12*27 12*28 2. + 0*11 340082859" 2*80 + 0*188701354781 + 0*11143155927. .262 Jq(x) I2'00 I2'OI Table J^{x) I. + 0*11536036312. '^o(^) J^{x) + 0*047689310797 +0*223447104491 + 0*049920430320 + 0*052144702973 + 0*054361913660 + 0*056571848157 + 0*222773200930 + 0222077744768 + 0221360822234 +0*220622521586 2*40 2*41 + 0*129561026518 + 0*180710246883 + 0*131360894344 + 0133146181728 + 0134916723111 + 0*136672354521 +0179260532985 12*02 2*42 + 0177794184461 +017631I359I92 1203 1 2 '04 12*05 1 243 2*44 245 + 0*174812216550 + 0*058774293132 + 0219862933107 + 0*060969036167 + 0*063155865777 + 0*065334571427 + 0*067504943560 + 0*219082149091 + 0218280263834 +0*217457373624 + 0*216613576726 + 0*138412913587 +0173296917383 + 0*140138239554 + 0*141848173298 + 0143542557339 + 0145221235856 + 0*171765624000 +0170218500152 2 •06 2*46 2*47 12*07 1 2 08 2*48 2*49 + 0*168655711017 +0*167077423179 12*09 12*10 12*11 + 0*069666773607 +0*215748973377 + 0*071819854013 + 0*073963978255 + 0*076098940860 + 0*078224537427 + 0*214863665770 +0*213957758045 +02I303I356277 2*50 + 0146884054700 + 0*165483804615 12*12 12*13 251 + 0*148530861410 +0*163875024675 2*52 + 0*150161505225 + 0*162251254066 253 + 0*151775837096 + 0160612664833 2 '54 12*14 12*15 + 0*212084568463 + 0*153373709704 +0*158959430343 + 0*154954977468 + 0*157291725265 + 0*156519496560 + 0*158067124921 + 0*159597722266 + 0*161111150104 +0155609725554 +0153913608430 +0152203552365 +0*150479737058 + 0*080340564642 + 0*211117504511 + 0082446820302 + 0*084543103331 + 0*086629213798 + 0*088704952938 + 0*210130276228 + 0209122997309 + 0*208095783320 + 02070487 5 1 69 1 255 12*16 12*17 2*56 2'57 12*18 12*19 1 2*58 259 2 '20 + 0*090770123171 + 0*205982021700 + 0092824528115 + 0*204895714458 + 0*094867972612 + 0*203789952902 + 0*096900262741 + 0202664861776 + 0*098921205837 + 0201520567620 260 2*61 + 0162607271746 +0148742343422 + 0164085952318 + 0*165547058774 + 0166990459905 + 0168416026353 +0146991553564 12*21 12*22 12*23 262 263 +0145227550765 +0143450519461 12*24 12*25 264 2*65 + 0*14166064522^ + 0*100930610511 + 0*102928286663 + 0104914045507 + o. + 0*177885147930 + 0*179163605667 + 0180423149549 + 0181663669309 + 0*12878774189: + 0.12690193509c + 012500500357. + 012309714321 12*32 +0*112734178832 +0*114657566356 12*33 +0*116567932311 12*34 1235 2*72 2*73 +0*118465097559 +0*120348884405 2*74 + 0*187703877780 275 + 0182885056640 + 0*12117855082.
0048586609352 + 0035805572692 +0215775543638 +0033623165893 + 0216100856151 +0031439073935 358 3 '59 + 0216404320223 +0029253513878 + 0216685922259 + 0027066702765 + 0212326339882 + 0*211809163903 + 0*211271195961 + 0*210712504851 0050675756773 0052758325976 0054834115851 0056902926099 360 + 0210133161369 0*058964557249 .{x) J.x) I.0025230095486 + 0216625355135 0027378117768 + 0068192057526 + 0*066060920168 +0063924854454 3'4i 342 343 3 "44 + 0061784075111 +02IOI75525783 +0059638797173 345 3*46 347 + 0210761 169428 +0057489235957 +0211325296943 +0055335607039 + 0211867868729 + 0053178126239 + 0212388847348 + 0051017009592 + 0216340851750 0029521821957 + 0216034933785 0*031660999316 + 0215707647547 0033795441703 348 3 '49 + 0215359041426 0035924941590 + 0214989165880 .Table J. 263 J.{.0023077964423 + 0216888399712 . (continued).{x) +0188701354781 +0111431559278 [320 321 + 0*216685922259 + 0*027066702765 +0189805784222 + 0109452757129 + 0190890378823 + 0107464624869 + 0191955046298 +0105467365986 + 01 9299969640 1 + 0103461184712 3 "22 + 02x6945650832 + 0024878857605 + 0217183496687 + 0022690195350 + 0217399452738 + 0020500932874 + 0217593514066 + 001 83 II 286951 323 3 '24 + 0194024240934 + 01 01 446286001 +0195028593748 + 0099422875508 + 0196012670759 +0097391159571 +0196976389945 +0095351345187 +0I979I967I360 +0093303639994 325 + 0*217765677921 + 0016121474234 + 02 1 79 1 59437 1 7 + 0218044313033 + 0218150789610 + 0218235379352 326 3"27 328 329 3 '30 + 0013931711237 + 0011742214308 + 0009553199615 + 0*007364883118 +0198842437136 + 0091248252250 +0199744611493 + 0089185390809 + 02 1 82980903 1 9 + 0005177480555 + 009299 1 2074 1 4 + 0000806278917 0001377090000 0003558684713 + 0200626120738 + 0087115265106 + 0201486893280 + 0085038085131 + 0202326859628 + 0082954061409 +0203145952399 + 0218338932728 + 0218357918950 333 + 0218355063505 3*34 + 0218330383064 331 332 + 0080863404982 335 + 0218283896439 0005738290927 + 0218215624587 + 0218125590599 + 0218013819702 + 0217880339252 0007915694697 +0203944106324 +0078766327385 336 3'37 + 0204721258250 + 0076663040627 +0205477347147 +0074553757168 0010090682449 0012263041002 338 + 02062I23I4II4 + 0072438689899 + 0206926102377 + 0070318052122 + 0207618657300 + 0208289926385 + 0208939859276 + 0209568407762 339 3 '40 0014432557586 + 0*217725178732 0*016599019864 + 0217548369742 0*018762215954 + 0217349946004 0020921934445 + 0217129943348 .0038049292086 + 0214598073436 + 0214185818679 + 0213752458244 + 0*213298050815 0*040168286951 0042281720622 + 0212888197522 +0048852473334 +0213365886137 +0046684733877 + 0213821882244 + 0044514007788 + 0214256157060 + 0042340511767 350 351 352 3'53 3 '54 + 0214668683969 + 0040164462629 +0215059438525 +0215428398451 0044389388228 0046491085613 + 0037986077278 3*55 356 357 + 0212822657111 .{x) J.
264 .
0205905422669 1477 +0033273346769 0206108933120 1478 +0031211325913 0206291796530 + 0029147579677 0206454008627 0044563402147 046547 I 8876 I 0198670276496 01980838 1 88 1 8 5i6 517 5i8 519 5 '20 048525014094 0197478032331 050496685220 0196852989694 052462009949 0196208765420 054420796844 + 0C27082314586 0206595567180 0195545435866 .Table Joix) I.0202307366980 + 0055782708050 + 0053753273205 + 005 1 7 19847828 + 0049682636966 0202737 2941 1 505 506 507 0203457124785 1 0203146426455 026473238057 028501882349  0203067538060 0202657999596 0203535231400 02039035 1957 1 508 5*09 030526331722 0202228563094 032546387470 020I779284I82 03456185 1456 + 0047641845902 0204251268330 + 0045597680133 + 0043550345355 + 0041500047438 + 0039446992407 0204578457081 5IO 020I3I0220408 0204885067267 0205 1 7 1082373 512 513 036572526126 020082 143 1 239 038578214533 0200312978045 040578720351 0199784924098 042573847897 0205436487924 514 5*15 0199237334565 + 0037391386420 0205681271486 + 0035333435752 . (continued). 265 Jo(^) J.{x) JAx) + 0106484118490 0i8503i66i6i5 +0104629151946 0185958463314 +0I027650I3033 0186866122350 + 0I0089I893564 0187754562014 + 0099009986107 0188623707542 +0097119483970 0189473486119 4 So +0027082314586 0206595567180 481 +00250I5737I79 0206716471994 482 483 484 485 + 0022948053986 0206816724913 + 0020879471508 0206896329814 + 0OI88IOI96I97 .0206955292607 +0016740434436 0206993621235 + 0095220581 177 0190303826889 +0093313472454 0191 1 14660960 +0091398353204 0191905921406 + 0089475419488 0192677543276 486 487 +0014670392520 0*207011325670 + 0012600276630 020700841 7910 488 + 0010530292822 0206984911980 + 489 0008460646998 0206940823925 490 491 + 0087544S680IO 0193429463596 + 0085606896092 + 0083661 701655 + 0081709483202 +0079750439794 +0077784771035 [94161621377 + 000639 1 54489 1 0206876I7I8IO + 0004323192042 0206790975716 +0002255793783 0206685257736 + 0OOOI895552I4 0206559041974 0OOI8753I88I7 0206412354539 194873957618 [95566415311 492 493 [96238939443 4 '94 495 019689 47 7005 1  0003938623732 0206245223541 0206057679091 0205849753289 + 0075812677046 0197523976991 +0073834358450 0198 1 36390405 + 007I8500I6350 0198728670261 496 006000155243 008059709376 oioi 17082484 4 "97 498 499 5'oo 5OI +0069859852307 019930077 1 592 OI2I7207I276 0205621480228  0205372895984 + 0067864068323 019985 265 447 1 0014224472827 0205104038614 016274084604 02048 1 4948 1 48 0204505666588 018320704486 020364130779 0204176237900 0203826708006 022404162240 0024440598094 +0065862866820  0200384268898 + 0063856450617 0200895585039 + 006 1 84502 29 1 3 0201386562994 + 0059828787267 0201857167913 5*02 503 5 "04 + 0057807947575 .
266 X Table I. {continued). .
267 f . (continued).Table II.
n .268 Table II. [continued).
{continued). 269 Jn{7) n 19555 597 + 030007 92705 .000468 28234 82345 833 030141 72200 85940 120 ^ 016755 55879 95334 236 + 015779 + 034789 + 033919 + 023358 81446 61367 918 63247 51183 285 66049 83179 632 35695 05696 084 + 012797 05340 28212 537 73075 428 + 005892 05082 lO II 12 13 14 15 + 002353 + 000833 + 000265 + 000077 + 000020 + 000005 93443 88267 135 47614 07687 815 56200 35894 568 02215 72522 133 52029 47759 069 05902 18514 143 24944 64660 269 16 17 18 + 000001 16122 74444 403 05036 96762 619 00959 75833 201 00173 14903 330 00029 66471 543 00004 83925 930 75348 588 11221 932 01601 804 19 20 21 22 23 24 25 26 27 28 00219 522 00028 933 00003 673 29 450 30 31 + 0 053 006 001 32 .Table II.
n . (continued).270 Table II.
271 n . {continued).Table II.
{continued). n .272 Table II.
(continued).Table n II. 273 .
274 Table II. n . (continued).
.
{continued).276 Table II. n .
(contmued). 277 m .Table II.
n .278 Table II. {coiUiuued).
Table II. 279 n . (continued).
roots of jT (^) = 0. maximum or minimum .280 Table The first fifty III. with the corresponding values of Jq{x).
Table IV. 281 X .
.282 X Table V.
283 .Table X V^. {continued).
(continued). .284 X Table V.
Table VI. 285 x .
(continued).286 Table VI. X .
Table VI.(x) hi^) Nil Nil Nil •o'=24829 1 584037 •08139087425642 •0789398097 1 214 •o'°i9866o852ii9 •0^255240920874 •07438834749717 •0633 I 6390536 I 5 •OS 1 5992 1823 1 20 •o'°637748i 54995 •03102559132723 •05582022868887 •0422488661477 I •o*682o8563ii42 •08164357788982 •07165452506 I 06 •07996062403333 •0^433537513798 •0515095405 1 219 •05580928790861 •03175196213558 •04173686673046 •044505989 I 30 I 2 •03398740613950 •03827978932673 •0216001 7336352 •05446656506452 •04 1 •03104953102941 •03224639142001 16770209099 •04276993695123 •04607607604085 •03124988823 I 59 •022919467 1786 •0^5081 367 I 5570 I •03449225284743 •03849664857007 •0^850453706344 •02153415828186 •0=266357538382 •03243684776533 •0137719020155 •03454025096400 •03813702326455 •0216835897328 •0333248823452 •0501531656813 •074 I 0887 38 I 66 •107756685981 •I •0=447211872992 •0=729479022559 •01 •0=141017510822 •0=237340311951 16036566222 •0180554571973 •0275537875687 •041 329963501 2 •0=389320693838 •0=624273178058 5446479987 I •0=980992761666 •218632053769 •305975090770 •423890764347 •0610477626605 •OI51395II5677 •0889386166028 •127975549614 •0229885833970 •03439996 1 1 745 •581912714514 •792285668997 1^07068675643 I •182096322090 •256488941728 •05079844175 19 •0741 16632 1 596 •357956089960 •495379239735 •106958821921 •I ^437 1 302 1 8 10 528 I 3670643 1^9 I 7 10069457 •680308520630 •927710973612 i^2569i8o48ii •216329392995 2^542971 13760 •303668787505 335577484714 •422966068203 . (continued). 287 ^6{^) I.
288 Table VI. X . (continued).
etc. Ermittelung der absoluten Storungen in Ellipsen. ToDHUNTER (I. pp. Petrop. etc. 264). (Berlin Abh. Untersuchung des Theils der planetarischen Storungen. 157. Lame's. Trans. M. G. i. und Cylinderfunctionen fortschreitenden Entwickelungen. Kugel.). 162). i. Trans. Theorie der Bessel'schen Functionen. Vol.). RiEMANN. (Comm. 1824). Bessel. vii. and Rayleigh. TREATISES.). Basset. 19 . IL Bernoulli (Dan. Neumann J. Ac. (Ibid. cap. etc. de oscillationibus etc. Treatise on Hydrodynamics. R. Cylindrical. On Fluctuating Functions (Irish Acad. Jacobi. (1843) p. Functions. Hansen. 1843)). Theoremata flexili corporum filo connexorum.) Sur la distribution de la Chaleur dans les corps solides PoissoN. „ „ EuLER. etc. v.). and Bessel's Byerley. etc.). Studien tiber die Bessel'schen Functionen. Heine (E. Hamilton (W. (Phil. (Journ. Hargreaves. Theory of Sound. Petrop. (Crelle XV. Ellipsoidal Harmonics. (C). 19). 1848). Handbuch der Kugelfunctionen. VL 108. Fourier. LoMMEL (E. de I'Ecole Polyt. Formula transformationis integralium definitorum. Ueber die nach Kreis. VII. Partielle Differentialgleichungen. (Other references will he found in the text. 178. XIX. On a general method of Integration. Treatise on Laplace's.) I. pt.BIBLIOGRAPHY. (Schriften der Stern warte Seeburg (Gotha. pt. 13). Treatise on Fourier Series and Spherical. 99 and Acta Acad.) Demonstratio theorematum. MEMOIRS. (ibid. Theorie Analytique de la Chaleur. ii.
). iv.). x. . (2) v. (Danzig.). Ueber die Anwendung der Bessel'schen Functionen in der Theorie der Beugung (Zeitsch. 1869).). Ann.). Die Fourier'schen Beihen functionen (ibid. Hankel „ „ (H. Die Fourier'schen Integrale (P. (2) vi. d. Cylinderfunctionen u. Math.). Heine (E. xv.). xcv.). xvi. Weber (H. iii.290 Anger. fiir (ibid. iii. Integrale Cylinder Neumann (C). „ etc. Suir uso delle linee lungo le quali il valore assoluto di una DU BoisReymond funzione e costante (Ann. „ Ueber Ueber (ibid. Ueber die Bessel'schen Functionen u. Functionen LOMMEL. Die Cylinderfunctionen erster u. Ann.).).). Zur Theorie der Bessel'schen Functionen (ibid. „ „ Zur Theorie der Bessel'schen Functionen (ibid. Math.). Phys. Bestimmte Integrale mit VIII.). u. (Crelle lxix. ix.). Ueber die Bessel'sche Transcendente I. Untersuchungen iiber die Function /J.). gen „ iiber die Bessel'schen Functionen (Math.). Ann.). d. lxxvi. vi. Ann. LiPSCHiTZ. Formeln (Math. Ueber Entwickelung einer arbitraren Function zweier Variabeln nach den Bessel'schen Func die Convergenz der „ „ tionen.). 1855). ScHLOMiLCH. lxxv. di Mat. Ueber einige bestimmte Integrale. „ die stationaren Stromungen der Elektricitat in Cylindern (ibid. (Leipzig Ber. u. u. VIII. Integration der Gleichung „ „ durch Bessel'sche Functionen (Math. Ann.). 137). Ueber eine mit den Bessel'schen Functionen verwandte Function (ibid. (Math. Ann. BIBLIOGRAPHY. etc. Sopra un Teorema di Jacobi. Phys.). Ueber die Nobili'schen Farbenringe (Pogg. RiEMANN.). Einige Bemerkungen zu Herrn Neumann's Untersuchun ScHLAFLi. (Crelle lvi. Ann. xiv. eine Darstellung willkiirlicher Functionen durch Bessel'sche Functionen (Math.).. zweiter Art (Math. II. ntwickelung einer Function nach Produkten der FourierBessel'schen Quadraten u. I.). etc. etc..). Ueber die Bessel'sche Function (Zeitsch. (Brioschi Ann. ii. ihre Anwendung. Die FourierBessel'sche Function (Crelle lxix. iv.
xvi. Bryan. Jan. The equations Stability of a Plane Plate. 310). LXXIL.). xxv. Ueber die Fortpflanzungsgeschwindigkeit kleiner Schwingungen. Ann. . L. (J.). p. Various papers in the Wiener Berichte. XVII. Rayleigh. and relations connecting them with Hyperspherical and Spherical Harmonics (ibid.. (Proc. XVI.). 1881).). M. LXXIV. Trans. HoBSON. Notes on Bessel Functions (Phil. LXIX.). On Bessel's Functions. On Oscillations in Cylindrical Conductors L. Journ. etc.).. solid 49). (Proc. Glaisher Bessel (Proc. L.BIBLIOGRAPHY. 291 Mehler.). L. Electrical McMahon (J. M. Ueber die Darstellung einer IH^ ^F „ willkiirlichen Function zweier Variabeln durch Cylinderfunctionen (Math. Soc. xxii. On Riccati's Riccati's „ On Height Proc. XCV. 1881). M. v. lxv. (Math. Phil. Equation and Bessel's Equation (Quart. of an isotropic elastic cylindrical coordinates (Proc. 250). Thomson (J. Greenhill. . xxii. IV.).). etc. S. J.). M. p. Soc. W.. (Crelle lxxxi. p.).. S. 1872). Gegenbauer (L. p. etc. ix. of Math. the roots of the Bessel and certain related func tions (Annals of Math. in polar and Camb. Systems of Spherical Harmonics (Proc. Phil. 1895).). Ann. xiv. Notiz iiber die Functionen P^ (cosO) und J (x) (ibid. S. S. 431). Mag.. consistent with Stability (Camb. LXVI. SoNiNE (N. Recherches sur les fonctions cylindriques.. On Equation (Phil. On the relation between the functions of Laplace and ochhammer. Strutt [Lord Rayleigh]. L. Vols. Chree.
/ .
.
.
DEC 1 5 1983 PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET .
This action might not be possible to undo. Are you sure you want to continue?