R. & M. No.

2655
(11,559)

A.R.C. Technical Report

'. ".

~, <l~.

"':1.,.

MINISTRY OF SUPPLY
AERONAUTICAL RESEARCH COUNCIL
REPORTS AND MEMORANDA

Supersonic Theory for Oscillating

Wings of Any Plan Form
By
W. P. JONES, M.A., A.F.R.Ae.S.,
of the Aerodynamics Division, N.P.L.

Crown Copyright Reserved

:LONDON:

HER MAJESTY'S

STATIONERY

1953

PRICE

3s 6d

NET

OFFICE

Supersonic Theory for Oscillating Wings of

Any Plan Form
By

W. P.

JONES,

M.A., A.F.R.Ae.S.,

of the Aerodynamics Division, N.P.L.

Reports and Memoranda No. 2655 *

June, 1948

11aJ Aircraft Est~tt~sttJlt}n

(

2. 5MAR 1954
liBRARY

Summary.-A theory for thin wings of any plan form describing simple harmonic oscillations of small amplitude
in a supersonic air stream is developed. It is based on the use of the generalised Green's Theorem in conjunction
with particular solutions which vanish over the characteristic cone with vertex at any point in the field of flow.
The theory can be used to calculate the aerodynamic forces acting on fluttering wings when the modes of distortion
are known.

1. I ntroduction.-The problem of determining the aerodynamic forces acting on oscillating

wings of any plan form in a uniform supersonic stream is considered. Formal solutions are
obtained for the case of thin wings on the basis of linearised theory. In two dimensions, experimental evidence' has shown that the linear theory" is inadequate and that the thickness of the
aerofoil should be taken into account". This is probably true also in three dimensions but, as a
first step towards the solution of the thick wing problem, the linearised theory for thin wings
is developed in this paper. For wings with leading and trailing edges inclined at all points at
angles to the direction of flow greater than the Mach angle, the solution is readily derived. In
the more difficult case of wings with edges inclined to the stream at angles less than the Mach
angle, the problem is reduced to one of finding the solution of an integral equation of a type
similar to that which arises in incompressible flow.
In the limiting case when the frequency of oscillation is zero, the formulae derived in this
paper lead to those already assumed for steady motion': 5.
2. Basic Equations.-The velocity potential f due to a small disturbance caused by a wmg
oscillating in an air stream V satisfies the equation

where V o is the velocity of sound, and where Xl , Yl , Zl define the position of any moving point
of the stream at time t l Let c represent a unit of length and assume that
Xl

eX cot .

Yl

cY,

Zl =

cZ, . tl

cT

V '

* Published with the permission of the Director, National Physical Laboratory.

1

where M
VIVo and sin,u = lilY!.
rewritten as
occce'

In the new co-ordinates X, Y, Z, T, the equation can be

(1)

For simple harmonic motion, let eP= eP'ei U

cp'

where A = pc/V, and use the transformation

=-=- q>e-i)' sec- Ji . cot J.l

(2)

Substitution in equation (I) then yields, when the exponential factor is suppressed,
(3)
where k

=:-

A sec,u.

For steady motion, k-= O.

The velocity potential eP of the disturbance must be zero at the wave front*, and for the
problems considered in this paper 0eP!oZ, the normal velocity, is known at the wing surface.
Hence the boundary conditions associated with equation (3) are (fJ = 0 over the wave front, and
o(fJ!oZ known over the wing.
Let X o, Yu , Zo be the co-ordinates of a particular point in the field, and let x - Xu - X,
y ~_ Y, - Y, z
Z; - Z, r'2 .... ),'2.). Z2. Then the characteristic cone r with vertex at X o , Yo,
Zo is defined by
(4)

The solution of equation (3) with the boundary conditions stated is derived by the application
of the generalised Green's Theorem":" to the volume enclosed by T and the wave front (see
Figs, I, 2). When (fJ and hence eP are determined, the lift distribution l(x l , Yl) is given by the
usual formula
(5)

where Po is the air density of the undisturbed stream,

3. Generalised Green's Theorem.-Let

(6)

and let P denote a particular solution of equation (3). The theorem then gives the following
integral relation between the required solution (fJ and P, namely

r [PL((fJ)

Jv

- (fJL(P)] d. =

r [(fJ0P
OV

Js

* The wave front is regarded as the envelope of the Mach

2

P o(fJ JdS .
OV

cones with vertices on the leading edge.

(7)

The integral on the left is taken over the volume enclosed by the surface S excluding, however,
any points of singularity which may be within the region. The integral. on the right is taken
.over the external surface S and the surfaces enclosing the singularities. In equation (7), the
symbol v refers to the co-normal which has direction cosines -l, m, n, where l, m, n are the
direction cosines of the inward drawn normal at any point of the surface. It then follows that
the operator
(8)

The normal and the co-normal have the same projection on the X

plane.

Since wand P are solutions of equation (3), L(w) = L(P) = 0, and equation (7) reduces to

r (w oP _

Js

a'/!

pOW) dS

ov

(9}

For the problems considered in this paper the surface S includes parts of r and the wave front
as shown in Figs. 1,2. Now on the wave front, w = 0, and it can also be shown that ow/a'/! = 0.
This simplifies equation (9) and indicates that further simplification might be possible by the
use of a particular solution P which vanishes on I', In such a case oP/o v = 0 also, since the
co-normal at any point on r coincides with a generator. This procedure was used by Volterra"
to o"Qtain unique solutions of problems connected with the propagation of electric waves in a.
conducting medium, and is to a certain extent followed in this report.
4. Particular Solutions.-Since r is defined by x 2 - r2 = 0, the required particular solutions
which vanish on the cone are assumed to be functions of x and r only. They must therefore
satisfy
(1O)
Assume a solution of the form
(11)

It is readily verified that equation (11) will satisfy equation (10) provided

0'1 + 2(n + 1) of + f
oq2
q
oq

(0

1) :~~

+ (0

- 2;) ~~ -

(12)

n(n

0-; l)g =

Solutions of equations (12) and (13) for particular values of n are then obtained.
equation (12) is satisfied by

A cos q + B sin q ,
q
3

(13)
When n

0,

(14)

where A and B are arbitrary constants. When () = 1, however, q = 0 and the first term in
equation (14) becomes infinite. To avoid singularities, therefore, the particular solution is
chosen to be

I,
Similarly, when n

smq.
q

(15)

1, a suitable solution of equation (12) is

f2 =

4q (sin q -

(16)

q cos q) .

The required particular solutions of equation (13) must vanish when ()

to be
and

1, and they are found

gl = loge [() + y({)2 - I)J,

(17)

_
2
y({)2 - 1)
g2 - loge [() + y({) - I)J - ---fJ---~

(18)

for n = 0 and n = 1 respectively. Hence, particular solutions of equation (10) (and equation
(3)) which vanish on the characteristic cone can be taken to be
lJf1 -_ smq
-- -- loge [()

and
lJf = 9

-x [ -

q)] [log [() +

a (sin
~-

q oq

+ y({) 2
- )
1J,

y({)~9

(19)
-

- ~-1)]
I)J - -y({)2
--()

(20)

When r = 0, () ....,.. 00 , and both lJf1 and lJf2 become infinite along a line through X o , Yo, Zo parallel
to the X-axis. These singularities are excluded from the field of integration by a narrow
cylinder as shown in Figs. 1, 2.
With the aid of Green's Theorem and the use of equations (19) and (20) it is then possible to
obtain formal solutions of equation (3) for wings of any plan form in oscillatory or steady motion.

z
5. Wing of Infinite Span.-First,
consider the relatively simple case of the
oscillating flat plate of infinite span with
leading edge along the OY-axis at right
angles to the direction of flow. The
characteristic cone r with vertex at
X o, Yo, Zo will intersect the wave front
represented in this case by the planes
Z = X as shown in the diagram. It
should be noted that the enclosed
volume is cut in two by the wing, which
is assumed to lie in the plane Z = o.
The singularities within the volume
which arise from the particular solutions
lJfl and lJf2 are enclosed in the thin
cylinder marked C. If Green's Theorem
is now applied to the volume PADEB,
the relation

FIG.

1.

_
Jsr (rJ> OP
ov

pOrJ
ov

dS

_
Jcr (rJ> OP
ov

pOrJ
ov

dS

(21)

is obtained, where the first integral is taken over the external surface of the volume, and the
second integral is taken over the surface of the thin cylinder. Over the area ADE of the wave
front rJ> = 0, and orJ>/ov = since the co-normal lies in the plane Z = X. Over the cone, P = 0,
and oP/o'/! = since the co-normal is a generator. It then follows, since %v = %Z at the
wing surface, that

dS =

( (rJ> OP _ pOrJ
OV
ov

Js

r (rJ> OPoZ _ p~rJoZ

dX dY

Jw

(22)

taken over the area W representing the part BDE of the wing cut off by r. The remaining
integral depends on the limiting form of the particular solution chosen as r ~ 0. Over the surface
of the cylinder %v = %r, and it can be shown that, as r ~ 0, formulae (19) and (20) yield
PI ~ -

oP I
or

sinkx

~loge

r,

sin kx
kxr'

-~-~-

lTf

r 2 ~ -

[sin kx - kx cos kx] I g

k 3x 2
0 er ,
(23)

For the first particular solution

I
j" [rJ> oP I _ P orJ>JdS
~~"xo ~"2n(rJ> oPor _
c

P OrJ>I) r de dX
I

or

_ 2 ~,xo msink(Xo - X)dX

n
'
k(Xo - X)
,

(24)

Xs

where X, is the co-ordinate of the base of the cylinder.

yields

In this case, therefore, equation (21)

(25)

The suffix (a' is introduced to indicate that the value of rJ> immediately above the surface is used.
Similarly, if Green's Theorem is applied to the volume GBDE below the wing which contains
no singularities, equation (21) yields
(26)

where ([\ refers to the value of (fJ immediately below the wing. Integration over the image of
the volume GBDE in the plane Z = 0 enclosed by the characteristic cone with vertex at X o ,
Yo, -e; gives
(27)
since PI involves terms in (Zo - Z)2 only. For the particular problems considered in this paper
o(fJ,,/2Z = o(fJb/2Z, and addition of equations (26) and (27) leads to the relation (fJ" --= - (fJb'
It then follows from equations (25) and (27) that

n r~o (fJ~~n_~(~o ~_X) dX = r ,(fJa O~l dX dY

JX s
k(XIJ - X)
JH oZ

(28)
Equations (26) and (27), and hence (28), are only valid when the wing divides the volume enclosed
by the characteristic cone and the wave front into separate parts. In the case of the second
particular solution P 2 , the required solution (fJ is given similarly by

_":JXo (fJ_2_ sin ~(Xo_~ X)dX =

k:!. JX s

oX o k(Xo - X)

Jw

Jw

(fJ ~~:!dX dY
a

(1Z

Poo(fJa dX dY .
~ oZ

(29)

From equation (28), by differentiation with respect to X o , it follows that

(30)

Then, by the use of equation (29), the above formula reduces to

(31)

I t can also be shown that

(32)

Since, if

(X~.

v![(X o

X)2 - Z 02], it can be proved that

(33)

it follows from equations (31) and (32) that

n(jJ(Xo, Yo, Zo) =

rxo-zo rYo+a (j(jJa cosk V[(Xo - X)2 - (Yo - Y)2 - Z 02J

ho-a 3Z V[(Xo _ X)2 _ (Yo _ Y)2 _ Z 02J dX dY.

Jo

Without loss of generality, it can be assumed that Yo

(34)

0, and if the substitution

(35)

is made, formula (34) reduces to

rxo -

n(jJ = - 2 Jo

Zo

r"/2 (j(jJ

Jo

(36)

(j cos (krt. cos f3) d(3 dX .

Now, in terms of Bessel Functions,

cos (krt. cos (3)

]o(krt.) - 212(krt.) cos 2(3

and (j(jJa!3Z is a function of X only.

(jJ(Xo , Zo) = -

+ 2]4(krt.}

cos 4(3 - , etc.,

(37)

Hence, on integration, equation (36) yields

[0-Zo 33~a lo(k V[(X

o-

X)2 - Z 2)J dX .

(38)

When Zo = 0, equation (38) reduces to the formula

(jJ(Xo)

XO

(j(jJ
._a 10[k(Xo - X)J dX ,
3Z

(39)

which, when expressed in terms of the original co-ordinates, leads to the Temple and j ahn"
solution, namely,
(40)

6~

Wings of Finite Aspect Ratio.-For the purposes of this report, wings of finite aspect ratio

are classified into two types A and B.

A plan form is of type A if the tangent at any point of its outer edge (leading edge wing tips
and trailing edge) is inclined to the direction of flow at an angle y greater than the Mach angle
p. Such wings will cut the volume enclosed by the wave front and r into two separate parts
such as P ABDE and BDEG in the infinite wing case. Then, as in the previous section, it follows
that (jJ is given by formula (34), namely

(jJ (X Y Z)
no, 0 ,0

Jrw

(j (jJa(k cos
3Z
q

q) dS ,

(41)

where q k V[(Xo - X)2 - (Yo - Y)2 - Z 0 2J and W is the part of the wing cut off by r. By
the use of equations (41), (2) and (5), the aerodynamic forces on any fluttering wing of type A
for which (j(jJa!3Z is known can be calculated. Triangular wings with vertex angles greater
than f-t, and trapezoidal wings with wing tips inclined to the stream at angles greater than f-t are
of this class.
7

A plan form is of type B when y < fl at one

or more positions along the leading edge and wing
tips. To avoid consideration of flow conditions
in the wake at this stage, it is assumed that y > fl
for all points along the trailing edge as for wings
of type A. Particular examples of type B plan
forms are triangular wings with vertex angles less
than u, and rectangular wings. For the rectangular
wing, the flows above and below are not everywhere
independent since for points near the tip the wing
does not divide the volume enclosed by the wave
front and the characteristic cone with vertex at a
particular point into two separate parts. For the
triangular wing with y < fl, the wave front is taken
to be the Mach cone with vertex at 0 as indicated
in Fig. 2.

H
FIG.

2.

In this case, the volume PGEHFO is also not divided into two separate parts
OTR. However, Green's Theorem can be applied to the volume as a whole to give

by~the

wing

(42)

where the integral is taken over the surfaces of the Mach cone with vertex at 0, the characteristic
cone r with vertex at P, the upper and lower surfaces \V of the part OMBN of the wing cut
off by r, and over the cylinder 0 enclosing the singularities which arise from the chosen particular
solutions. Over the Mach cone, (/) = 0 = 0(/)/01', while over r, PI = 0 = oPt/01', and the
corresponding integrals vanish. At the wing surface, 0/01' = %Z according as points on
the upper or lower surfaces are considered. It should also be remembered that o(/)a/oZ =
P(/)b!(1Z. Integration over the wing surface VV then gives

J ((/) - (/) )oPoZ

TV

dS

(43)

The contribution due to the surface C has already been given in section 5.
comparison with equations (28), (29) and (30) that

I t then follows by

(44)

where K ---: (/)" - c[Jb' Along OM and ON, K = 0, and along MN, PI = P 2 = O. Then, since K
is independent of Zo and oP!oZo = - 2:p!aZ, equation (44) can be re-written in the form

= _ ~_ J

azo

IV

k V[(Xo ----=-_ X? -jYo =-Y? ~ 2'o~J dX dY

V'[(X o - X)2 - (Yo - Y)2 - Z o2J

KCOS

(45)

Finally, 8<P/8Z is known over the wing surface, and by differentiating equation (45) the following
integral equation for <P a is derived, namely,
'Jt

8<P =
8Zo

_ ~

8Z02

f
W

<P cos k V[(Xo - X)2 - (Yo - Y)2 - Z o2J dX dY

a
V[(Xo - X)2 - (Yo - Y? - Z o2J
'

(46)

where Zo ~ O. It should be remembered that <Pa = - <P b , and that, by equations (28) and
(29), equation (44) is equivalent to equation (31) for wings of type A.
The form of equations (41) and (45) indicates that the particular solution which is infinite
everywhere on the characteristic cone, namely,

= k cosq

(47)

could have been used in the preceding analysis. Formulae (41) and (45) then follow from (7),
if the integral over the surface of the characteristic cone is neglected, and if the finite parts
only of certain integrals are considered, as suggested by Hadamard. Such a method is used
by Olga Todd" to solve the problem of a two-dimensional aerofoil in non-uniform motion. In
the finite wing case, however, the justification of such a procedure would probably be rather
complicated.
For wings of type A, the required solutions are given directly by equation (41). In general
however, both equations (41) and (46) will have to be used. For a wing of type B as shown
below,

;'

,,"

.,

,,

/5

,I..y.

,, """'- J.'
FIG.

3.

the velocity potential at any point of the unshaded part is given by equation (41), but for a
point P in a shaded part equation (46) must be used. Over triangle TSR, for instance, the function
<P a must satisfy equation (46), and it must also vanish along the wing tip TS. Along TR, <P a is
given directly by equation (41). Solutions have already been given for problems of this type
in the steady motion casev". With the aid of such solutions, the calculation of derivatives for
oscillating wings of type B should not be very difficult, at least for low values of k. The
numerical application of the theory does, however, require further consideration.

7. Steady Motion.-In this case, k

0, q = 0, and the particular solution PI reduces to

PI(O), where

P(O) = I X o -X
I
oge

+ V[(Xo -X)2

- (Yo - Y)2 - (Zo -Z)2J

V[(Yo _ Y)2 + (ZO - Z)2J
.
9

(48)

Substitution in equation (31) yields the following formulae

(49}
(50}

(51}
Equation (51) is also given directly by (41), when k = 0, and this is the form generally used for
wings of type A in the steady case. I t gives the potential due to a distribution of sources of
strength (j(pJoZ over the part Vif of the wing cut off by the characteristic cone.
For wings of type B, equation (44) gives
(52}
This can be written alternatively in the forms
(53)

(54)
(55)

where, as usual, the integral is taken over the part W of the wing cut off by the characteristic
cone with vertex at XI), YI), Zo. Formula (55) may be regarded as giving the potential due to
a distribution of doublets of strength rfJa -- rfJ b equal to the discontinuity in the velocity potential
at the wing surface. Another alternative form can be derived from equation (54) if use is
made of the fact that Pl(O)[rfJ a - rfJ b] = 0 round the edges of the area "V. Then, since
ap 1(0)/axo = - 3P1(0)/ax, integration by parts gives
(56)
=

Jw

a~JQ)

azo axa (rfJ

- rfJ b ) dX dY

(57)

where differentiation under the integral sign does not introduce extra terms, since P 1(0) = 0
along the boundary whenever the corresponding limits of the integral are not independent
of ZI)' By the use of equation (48), it follows that, when Z = 0,
aP1(0) _
Zo(Xo - X)
aZtI - [(y;;=-yt-+Z~~J V[(Xo ----=-Xr=-(Yo--=-Y)2 -=--Z-;;-~J

_ a

- aY tan

_l(ZO v[(Xo - X)2 - (Yo - Y)2 - Z02])

--- -- (X
X) (Yo - Y) --~-- .

-=-

10

(58)

(59)

The integral equation to be solved for wings of type B can then be written in the form
(60)
where '2J(P!3Z o is known over the wing, and Zo ~ O. When tJJ a - tJJ b is independent of Y, equation
(60) can be reduced to a single integral with respect to X by the use of equation (59).
8. Concluding Remarks.-The formulae given in this report provide a basis for the calculation
of the aerodynamic forces acting on oscillating wings of any plan form for any mode of distortion.
In the first place, however, it would be interesting to see whether the aerodynamic damping
for pure pitching oscillations about certain axis positions is negative as in two-dimensional
theory". Calculations for triangular wings could be carried out and experiments on half-wing
rigid models could possibly be made for comparison. The models however, would have to be
thin as otherwise the theory might not apply",

REFERENCES
No.

Title, etc.

Author

J. B. Bratt and A. Chinneck

2

G. Temple and H. A. Jahn

W. P. Jones

M. J. Lighthill

D. R Taunt and G. N. Ward ..

Olga Todd

A. G. Webster

A. Robinson

Information Report on Measurements of Mid-chord Pitching Moment

Derivatives at High Speeds. A.R.C. 10,710. July, 1947.
(Unpublished.)
Flutter at Supersonic Speeds. Derivative Coefficients for a Thin
Aerofoil at Zero Incidence. R & M. 2140. April, 1945.
The Influence of Thickness/Chord Ratio on Supersonic Derivatives
for Oscillating Aerofoils. R. & M. 2679. September, 1947.
The Supersonic Theory of Wings of Finite Span. R. & M. 2001.
October, 1944.
Wings of Finite Aspect Ratio at Supersonic Velocities. R. & M.
2421. January, 1946.
A Boundary Value Problem for a Hyperbolic Differential Equation
arising in the Theory of the Non-uniform Supersonic Motion of an
Aerofoil. A.RC. 9660. May, 1946. Published in Courant
Anniversary Volume. Interscience Publishers Inc., New York,
1949.
Partial Differential Equations of Mathematical Physics. Hafner
Publishing Co., New York.
Aerofoil Theory of a Flat Delta Wing at Supersonic Speeds. R. & M.
2548. September, 1946.

11
(1875)

Wt.13/806

K.9

12/53

H.P.Co.

34-261

PRINTED

IN

GREAT

BInTAJ~

R. & M. No. 2(

Publications of the
}~eronautical Research Council

*
*
,~

.j<

ANNUAL TECHNICAL REPORTS OF THE AERONAUTICAL RESEARCH COUNCIL

(BOUND VOLUMES)1936 Vol. I. Aerodynamics General, Performance, Airscrews, Flutter and Spinning.
40s. (40s. 9d.)
Vol. II. Stability and Control, Srructures, Seaplanes, Engines, etc. 50s. (50s. 10d.)
1937 Vol. I. Aerodynamics General, Performance, Airscrews, Flutter and Spinning.
(Os. (40s. 10d.)
Vol. II. Stability and Control, Structures, Seaplanes, Engines, etc. 60s. (61s.)
1938 Vol. I. Aerodynamics General, Performance, Airscrews. 50s. (5Is.)
Vol. II. Stability and Control, Flutter, Structures, Seaplanes, Wind Tunnels,
Materials. 30s. (30s. 9d.)
19~~9 Vol. 1. Aerodynamics General, Performance, Airscrews, Engines. 50s. (50s. Ud.)
Vol. 11. Stability and Control, Flutter and Vibration, Instruments, Structures,
Seaplanes, etc. 63s. (64s. 2d.)
1940 Aero and Hydrodynamics, Aerofoils, Airscrews, Engines, Flutter, Icing, Stability
and Control, Structures, and a miscellaneous section. 50s. (51s.)
1941 Aero and Hydrodynamics, Aerofoils, Airscrews, Engines, Flutter, Stability and
Control, Structures. 63s. (64s. 2d.)
1942 Vol 1. Aero and Hydrodynamics, Aerofoils, Airscrews, Engines. 75s. (76s.3d.)
Vol. II. NOIse, Parachutes, Stability and Control, Structures, Vibration, Wind
Tunnels. 47s. Bd. (48s. 5d.)
1943 Vol. I. (In the press).
* Certain. other reports proper to these nolumes will
Vol. II. (In the press).
subsequently be included in a separate volume.
ANNUAL REPORTS OF THE AERONAUTICAL
1933-34
Is.6d. (ls.8d.)
1934-35
Is.6d. (Is. Sd.)
Apnl, 19:J5 to Dec. 31, 1936. 4s. (4s. 4d.)

RJ~SEARCH

1937
1938
1939-48

COUNCIlL2s. (2s.2d.)
Is. Sd. (Is. Sd.)
3s. (3s. 2d.)

INDEX TO ALL. REPORTS AND )1[EMORANDA PUB:f.,][SHED IN THE ANNUAL

TECHNICAL REPORTS, AND SEPARATELYApril, 1950
R. & M. No. 2600. 2s. 6d. (2s. 7id.)
AUTHOR INDEX TO ALL REPORTS AND MlEl\1IOJRANDA OF THE AERONAUTICAL
RESEARCH COUNCIL1909-194~1
R. & M. No. 2570. ISs. (ISs. 3d.)
I

INDEXES TO THE TECHNICAL REPORTS OF THE AERONAUTICAL RESEARCH

COUNCILDecember 1, 1936 - June 30, 1939. R & M. No. 1850. Is. 3d. (Is. 4!d.)
:'uly 1, 1939 - :'une 30, 1945.
R. & M. No. 1950. Is. (Is. l!d.)
:'uly 1, 1945 - :'une 30, 1946.
R. & M. No. 2050. Is. (Is. lid.)
~'uly 1, 1146 - December 31, 1946.
R. & M. No. 2150. Is. 3d. (Is. 4!d.)
>nuary 1, 1947 - June 30, 1947.
R. & M. No. 2250. Is. 3d. (Is. 4!d.)
July, 195 ~.
R. & M. No. 2350. Is. 9d. (Is. lO~d.)
Prices in brackets include postage.

Obtainable from

HER MAJESTY'S ST A TIGNER Y OFFICE

York Honse, Kmgsway, Lancon, W.C.2; 423 Oxford Street, London, WI.
(Post Orclers, POBox 569, London, S.E.l); 1301 Castle Street, Edinburgh 2,
39 King Street, Manchester 2; 2 Edmund Street, BIrmingham 3; 1 St.
Andrew's Crescent, Cardiff, Tower Lane, Bristol 1 ; 80 Chichester Street,
Belfast OR THROUGH ANY BOOKSELLER

- J . . . . . . . . . . . . . _ ~ ~ ~ _~.

._~_w_,_,

__.__

__

~.~_~.

'"~_.

__w__

S.O. Code No. 232655

R. &. M. No. 26: