MINISTRY OF AVIATION

AERONAUTICAL RESEARCH COUNCIL

CURRENT PAPERS

Wind Tunnel Measurements of the

Unsteady Pressures in and behind

a Bomb Bay (Canberra)

BY

J. E. Rossrter and A. G. Kurn

LONDON. HER MAJESTY’S STATIONERY OFFICE

1965

PRICE 4s 6d NET

U.D.C. No. A.I.(@) E.E.Canberrs : 533.695.9 : 533.6.013.2 : 533.6.011.3!~

C.P. No. 728

October, 1962

WIND TuNlTm MEAsmNTs OF Tm UNSTEADY FBEssms IN AND BEHIND A BOMB BAY (CANBERRA)

J. E. Rossiter and

A. G. Kurn

!lhe unsteady pressures acting in and behind the bomb bay of a l/20 so,: model of the Canberra aircraft have been measured in the 8 ft x 6 ft transonx tunnel. The unsteady pressures are described by their r.m.s. values and by amplitude and correlation frequency spectra. Time average values of the pressures were also neasured.

The results show that the Canberra bomb bay is of the "shallrm type" in which the airflow enters the bay and attaches to the roof before being deflected cut by the rear bulkhead. Pressure fluctuations are most intenso in the vicinity of the rear bulkhead and on the roof of the bay where the flow attaches. Changing the Mach number from O-3 to O-6 has in general little effect on the unsteady pressures (when expressed non-dimensionally) but xwxeasing incidence up to ten degrees produces some decrease in thex- magnitude.

Replaces R.,'t.E. Tech.Note NO. Aaro 2&+5 - A.R.C. 24,645.

LIST OF CONTJZNTS

1 INTRODUCTION

2 TEST DETAILS

&g

3

3

3 3 3

4

k 6

2.1 The model 2.2 Instrumentation 2.3 Range nf investigation

3 FfEsJL1s

3.1 Presentation of results 3.2 Datum conditions 3,3 Xean pressure distributions 3.4 Dxstrlbution cf the r.m.s. of the unstea&y pressures 3-5 Amplitude spectra of unsteady pressures 3.6 Correlation spectra of unsteady pressures

L CSIKLUSIONS

4X'? 07 SfxBOLS

i,:;'? 03 Rl5XiZWCES

LIST OF ILLUSTMTIONS

&A. of model end support

P~sii&ns "f pressure transducers

PhGt-.Graphs of mo&el mounted in 8 ft x 6 ft tunnel

Conversi'n frsm non-dimensiond,~requenoy to full scale frequency

Pxplitude spectra of pressures on fuselage with bay closed (Transducer H)

Vean pressure distribution m and behind bay

X.m.s. of unsteady pressures in an& behind bay

:r;:jlitude spectra of unsteady pressures I4 = O-3

Corlp"rison with tests of Ref.4

il"~'x: of Mach number on amplitude spectra of unsteady press-s

~x~e1ation spectra. 16 = 0.3, a = 5.6’

FiJp

1

2

3

4

5

6

7

0

9

10

11

-2-

I INTRODUCTION

As part 4-a prceamme of tests to investigate the effect of an open bmb-bay nn the vibratirn.Pf an aircraft, the unsteady prcssuresmeasured m.the bay of a Canberre are to be compared with measurements made on a wind tunnel."nrdel. This Note re

P 0rt.5 tie results nf the wind tunnel meas+u-ements~

The tests were made on a 1 20 scale model of the Canberra B7 in the 8 ft x 6 ft transonic tunnel during April 1962.

2 TEST DETAILS

2.1 The model

Fig.1 is a general afiangement dramng of th'e model used for the wx.nd +Jlnnel tests. The model 1s complete with fuselage, wings and tailplane but the fin and pilot's cabin falring have been omitted. A smell block fitted in the front of the bay represents an instrument tray which will be carried during the flight-tests.

The model was supported frnm above, by a short stream-lined strut attached to a sting which was held in the tunnel incidence mechanxsm (see Fig.3). This method of support was chosen as being the least likely tc affect the air flowing over the bomb bay, and the lower sde of the fuselz.gc.

The bomb bay has an inverted V-shaped roof, and both the roof and tl!e lower fuselage line are swept up slightly towards the rear (see Flg.1). S'h, doors retract partially into the bay when it 1s open. No attempt has been made on the model to represen-< structural detatis such as mngs and strlngsx The bay has a mean lengtudepth ratlo of approximately 8:l and a mean length; tidth ratio of approximately 5~1.

2.2 Instrumentation

Small capacity type pressure transducers' (I(AE‘ type IT&6-37) were used to measure the pressures actzng at 7 posItIons along the roof of the ‘my, and 10 posItions nn the fuselage surface behInd the bay (see F~g.2). The transducers were calibrated before each test period by applying a known steady pressure to each trsnsducer and observing the outputs from the trans- ducer smpltixers on a d.c. voltmeter. Varlatlons in the 0alLbretinn factors throughout the test were less than "5% During the test, the outputs from the transducer ampllflers were observed on a d.c. voltmeter to give time average pressures, and on a Dawes true r.m.s. meter (me 61%) to give the r.m.s. of the unsteady component of the pressuns. in addltlon the outputs from selected transducers were recorded in pairs on an Ampex Type 306-2 two channel tape recorder for subseql=nt frequency analysis using a Muzhead- Psnetrada WV frequency analyser -rmth a bandwdth ratlo of 15.9. The =cording and analysing system 1s described fully in Bef.2.

2.3 Range of investigation

The model was tested over s, Mach number range from O-3 to 0.6 and, 311 order to facilitate comparison mth the forthcoming flight tests, the modct incidence was adjusted at each speed to sxnulate flight at altitudes of %CQ and 20,000 ft. Model vibration limlted the kinetx pressure at whit!, the tunnel could be operated to 150 p.s.f. The corresponding unit Reynol? number together with model incidences are given m the table below.

The time average and the r.m.s. of the pressures at each transducer position were measured for 611 the test conditions. For M = O-3, a = 5*6”, amplitude spectra mere measured at al.1 the transducer positions. Correla-tion spectra were obtained for all transducers in the bay referred to transducer G (Fig.2), and for the transducers behind the bay on the starboard side of the fuselage referred to transducers J and N. In addition correlation spectra between transducers at the ssme longitudinal station but on opposite sides of the fuselage wers measured.

For the ether seven test conditions amplitude spectra were measured for transducer p*sitlons F,G,H and N.

In order to establish datum conditions, amplitude spectra were obtained for transducer H at Id = 0.3 and O-6 with the bomb bay filled in.

3 RESULTS

3.1 Fresentatlen of results

Time average pressures have been expressed in terms of the usual pressure coefficient Cp, and the r.m.s. of the unsteady pressures have been

made non-dimensional by dividing them by the tunnel kinetic pressure q.

A preliminary examination of the nature of the unsteady pressures showed that they were random in character. Such random functions can only be described by their average properties and it is convenient to use the nomenclature of spectral analysis.

A non-dimensional frequency parameter n, has been used in the frequency spectra where

n+ I

where f is measured frequency (0.p.s.)

U is tunnel speed (f.p.s.)

and L is the bomb bay length (ft)

Fig.4 gives conversion factors to enable n to be expressed in absolute frequency under full sosle conditions.

The ordinate of the amplitude spectra has been expressed as pe/q<E

-lL-

where p, is the r.m.s. of the pressure fluctuations within the bandudtn of the analyser used to obtain the spectra

E is the bandwidth ratio of the andyser (= 0*155)

and q is the tunnel kinetic pressure (p.s.f.1.

Previded the ordinate of

pdq

a spectrum varies slowly with n, then JE is independent of the value of E an2 in the limit, as E 1s made

smaller E + dr+/n, so that the total mean square of the pressure fluctuations is related to the spectral density by

if= Fl=CG

26

q* J ( ) d2 d(log n) .

n=O An important corollary is that if the pressures are acting on a smplz

single degree of freedom system, with a natural frequency f. and an

acceptance bandwidth Af, then the r.m.s. of the relevant part of the exclt? tion is given by

se'that the ordinate cf the spectrum may be used to make quantitative comparisons of buffet intensity.

The correlatxon between the pressures acting at twc' points xi ancl ?:

may be described by the correlation coeffuient

R... = P(xi't) p(x,,t) - _ -

It will be appreclated that RIJ may be evaluated either for the pressw.:

fluctuatioss within the whole frequency range ?r fnr the pressure fluctuatlont, within narrow frequency bands, in say. In the latter case, Rij 1s a fin&l .

of the rentre frequency, n, of the band and may be plotted as a "correlatlo:, spectrum".

qj( 1 n was cslculated. by fxst measuring the amplitude spectra of t+)s instadcaneous sum au3 CiifYerence cf p(xi,t) and p(xj,t) then

Rij(d =

(PC/q{+;, - (Pc/qJs);iff

.

4 [

(PLJqWx. (Pc/9J"),. 1 J

In.attempting to interpret the pwsical significance of the correla i,-,, cneffioient it must be borne in mind that it IS essentially a statistical quantity. The spatial distribution If the correlatirn c?efficlent gives :i3

-5-

indication of the size of the region over which disturbances retain their identity - on the average. In addition, the variation of the correlation coefficient along the direction in which the disturbances are travelling will contain a measure of their average speed. In this connection it 1s significant that for the special case of a pressure field which is entirely periodic in character, the correlation coefficient is equal to the cosine of the phase difference between the pressures at the twq measuring points. Typically, the variation of Rij(n) along the direction of travel of dis-

turbances superficlslly resembles a damped cosine wave. The distances between successive zero crossings may be interpreted loosely as ufd mea wavelength of the disturbances and the rate of decrease in successive peak values as a measure of the rate of decay of the disturbances.

3.2 Datum conditions

The amplitude spectra of the unsteady pressures at transducer position H with the bay filled in are given in Fig.5 for N = 0.3 an?. 0.6. At each speed the spectrum consists mainly of a peak which is at the fan blade frequency (i.e. fan speed in r.p.s. x number of fan blades) together with one or more secondary peaks at harmonics of the fan blade frequency. This type of spectrum has been obtained previously in tests in the 8 ft x 6 ft tunnel and the peaks are thought to be due to the setting up of standing wave patterns within the plenum chamber surrounding the working section. The spectral density for this datum case is small compared with the spctral density when the bay is open except at the frequency of the e peaks ad, for certain transducer positi*n,'at very high frequencibs (n s 5) (compare Fig.5 with Fig.8, x/L = O-27 and l-26, for example). The only correction which has been made to the frequency spectra obtained with the bay open is to remove any peaks which occurred at the fan blade frequency and its first and second overtones.

3.3 Mean pressure distributions

Fig.6 shows the dis'crlbution of the mean pressures along the roof of the bay and on the fuselage behind the bay for N = 0.3 and 0.6. The pressure distrlbutlons at low incidence are typical of a shallow DJpe bay In which the airflow enters the bay and attaches to the roof before being deflected out by the rear bulkhead (see for example Ref.3). The pressure rise associated with the flow attachment occurs at x/L + 0.45.

3.4 Distribution of the r.m.s. of the unsteady pressures *

The distribution of the r.m.s. of the unsteady pressures is shown in Flg.7. The pressure fluctuations are most intense in the viciniQ of the rear bulkhead and on the roof of the bay where the flow attaches. The intensity decreases rapidly with distance behind the bg.

The curves for N = 0.3, a = 5160 and 10*4O show that increaging incidence produces an appreciable reauctian in the magnitude of the fluctuations within the bay, and. suggests that the,hlgher values of the r.m.s. of the pressures measured at M = O-6 are probably mainly due to the lower inciaences tested at this spe&. The effect of the inoidence change on the pressures behind the bay ~WgXgible.

3.5 Amplitude spectra of unsteady pressures J-

Fig.8 shows the amplitude spectra of the un$,teady pressures ier 1.4 = 0.3, a = 5.6”. For all transducer pesitions the spectra are tolerabii smooth and cover a broad band of frequencies, indicating that the pressure fluctuations are random in character. The spectra are similar to those

-6-

measured by Fail and others 4 m a Canberra type bay formed in a cy1indrico.L fuselage. In Fig-y, the spectra for two points near the rear bulkhead are compared with the earlier measurements. The agreement between the two se:r: of results is reasonable bearing in mind the differences in the external shapes of the models and the slightly different positions of the transducers.

Included In Fig.8 are spectra for M = 0.3, a = IO*&' for trio pcsxticns in the bay and for two positions behind the bay. It can be seen that,although as has already been noted,inc.reased incdence produces an appreciable decrease in the magnitude of the unsteady pressures within the bay there 1s little change in the shape of the spectra.

The spectra obtained at different Mach numbers sre compared in Fig.10 for two positions v:ithin the bay and for two positions on the fuselage behind the bay. In looking at these spectra it should be remembered that the inculence of the model was ad.justed at each Mach number 'cc simulate flight at .59X3 and 20,OCC ft. It is therefore difficult to separate the effects of rncidence and Mach number. However since, as has already been noted, incidence has little effect on the unsteady Pre-, wes on the fuselage behind the bay (Fig.8) it follows from Fig.10 that Mao;. number also has little effect on these pressures. Within the bay, increasing incidence reduces the magnitde of The pressures (Fig.8) and hence the changes in the spectra shown in Fig.10 for x/L = O-95 are probably mainly due to the change of incidence. For x/L = 0*84, however, although the changes in magnitude are similar to those for x/L = 0*95* there are also changes in the frequency at which the maximum spectral dens>% occurs and these are probably due to the change in Mach number.

3.6 Correlation spectra of unsteady pressures

Selected correlation spectra for M = 0.3, a = 5.6' are shotm in Fig. il. Within the bay (Fi&Il(a)) the main features cf a spectrum is a trough followed at about ttice the frequency by a peak. The high values of the c-x-relation at the peaks suggests that dzsturbances retain their identity for qultt large distances as they travel along the roof of the bay. I& is therefote possible to calculate the time average speed of propagation u, of these disturbances. For, of it is assumed that tne frequency n' at which a peak occurs, may be associated with a disturbance whose wavelength X', is the seme as the distance between the measuring points, then

Values so calculated are given in the table below:

x/L

0.95 to 0.84 0.49 0.95 to 0.7-l 0.59 0.95 to o-58 0.59 0.95 to O-45 0.56

h (mean value over ‘given x/L range)

The mean value of G/U is O-57. It is not possible from the informat.on available to calculate the speed of the airflnw YE' just outside the bomb bay

but an idea of its magnitude can be estimated from the measured values of C given in Fig.6 by assuming P

a

C-7 L

u = (1 -cp) .

-7-

This gives a value for Yd U, meaned over the region of interest, of

0.73 so that -4% = O-78. That is, the mean propagation speed of dis-

turbances along the roof of the bay is about three quarters of the mean stream velocity just outside the bay, which is of the same order as the mean speed of disturbances in a turbulect boundary layer measured by Tack and otherd.

Behind the bay, the pressures on the side of the fuselage are well correlated in the region between the transducers at x/L = 1 *I5 and I.37 (Fig.ll(c)) but the pressures in this region are not well correlated titi the pressures just behind the Es,r bulkhead (x/L = 1.05, Fig.ll(b)), suggesting that there is a small region immediately behind the bay which is isolated in behaviour from the main stream. Fig.ll(d) shows that whereas immediately behind the bay (J ana K), in this smsll isolated region, the pressures ere positively correlated in a cross plane, further downstream the correlation is large but negative, suggesting that the wake from the bay is unstable and moves from side to side of the fuselage.

4 CONCLUSIONS

(I) The Canberra bomb bay is of the "shallow type" in which the airflow enters the bay and attaches to the roof before being deflected out by the rear bulkhead.

(2) The pressure fluctuations are most intense in the vicinity of the rear bulkhead and on the roof of the bay where the flow attaches.

(3) Changing Mach number from 0.3 to 0.6 has (in general) little effect on the pressure fluctuations (when expressed non-dimensionally) but increasing incidence up to ten degrees produces some decrease in their magnitude.

(4) A study of the correlation spectra suggasts that disturbances travel slang the reef of the bay at about three-quarters of the local airspeed ext=ernsl to the bomb bay or at about one half of the free stream speed.

cP

f

L

Id

n

P

ps

9

Rijb)

t

LIST OF SYMBOLS

pressure coefficient

frequency (c.p.s.)

length of bomb bay (= I*11 ft)

tunnel !6ach number

a non-dimensional frequency parameter (= fL/U)

pressure (p.s.?.)

r.m.s. of pressure in frequency band sf (p.s.f.)

tunnel kinetic pressure (p.s.f.)

correlation spectrum between pressures at points i and j

time (seconds)

-a-

u

;

YE 3

x

a

E

x

LIST OF SYI~OLS (Contd)

tunnel velocity (f.p.s.1

see para 3.6

longitudinal distance measured downstream from front lip of bomb bay (ft)

wing lnoidence

bandwidth ratio

a wavelength (ft)

A bar - above a quantity has been used to udlcate Its tune avjrago value.

LIST OF I;FFERENCES

gg.

1

2

3

4

5

Author(s)

hbd.0da, Vl.R., Cole, P.W.

Owen, T.B.

Boshko, A.

Fail, R., Low speed wind tunnel tests on the flos in Owen, T.B., bomb bays and its effect on drag and vibration. Eyre, R.C.W. Ur@dx.sho?--K.o.A-. &port.

Tack, D.H., Smith, M.-W., Lambed, R.F.

Title, etc

A sub-mln1atu-e alff'erentlal pressure transducer for use in wind tunnel moaals. Unpublished M:o.A, :Ropo&.

Techniques of pressure fluctuation measurer nts employed xn the R.A.E. low speed wind wnn~ s. A.G.A.R.D. Report 172. March, 1958.

Some measurements afflow In a rectangular cut out. N.A.C.A. T.N.3488. August, 1955.

Wall pressure correlations in turbulent 311‘ lw, J. ACOUS. SOC. a (4). April, 1961.

-9-

SIMULATED / SECTION XX

INSTRUMENT TRAY

SECTION Y Y (2X SCALES SHOWN)

L/; I-” 12 INS

MODEL SCALE

SECTION Z Z

FIG. I. G.A. OF MODEL AND SUPPORT.

VIEW IN DIRECTION X

K M 0 R

t 6 6 6

.~&-&-&&+@ fi

-8 8 8 8 r L N P

DIMENSIONS GIVEN AS PROPORTION OF BAY LENGTH (= 13 35 INS MOOEL SCALE)

8 TRANSOUCERS IN TUNNEL MODEL TRANSDUCER POSITIONS FOR FLIG#4T TESTS

FIG. 2. POSITIONS OF PRESSURE TRANSDUCERS.

I n= t

03 04 OS M 06

FIG. 4. CONVERSION FROM NON DIMENSIONAL FREQUENCY TO FULL SCALE FREQUENCY

01 FAN BLADE

FREQUENCY 0.05

I

0 01 0.05 0.1 0.5 I n 5 IO

FlG.5. AMPLITUDE SPECTRA OF PRESSURES ON FUSELAGE WITH BAY CLOSED (TRANSDUCER H)

ON & OF BAY ROOF ON FUSELAGE BEHINO BAY I 1

I(MEAN OF WRTAND 57-o TRANSWRS I 0

Cb

0-c

0

-05

-I 0 J 0 02 04 06 08 IO I2 I.4

I 1.6

XL

--X-- M= 06

--+--

0!.=.5.6~

d =10 4O

d- l2O

d = 2.20

FlG.6. MEAN PRESSURE DISTRIBUTIONS IN AND BEHIND BAY

ON $ OF BAY ROOF , ON FUSELACfE BEHIND BAY

I-

I I I I I I

j

77

A

X

+

M = 0.3 d= S6O

d = IO 4O

M- 06 d = l.t"

d= t2O

dAi4 STARBOARD (TRANS. J L N ANO P)

0 02 0.4 0.6 0.8 I.0 I.2 l-4 =eL I.6 PPP t PORT

(TRANS. K M 0 AN0 R)

FIG.7 r.m.s. OF UNSTEADY PRESSURES IN AND BEHIND BAY.

da 5-6O (CORRESPONDING To FLIG+I~- AT spoo ~7)

-- --- d = 10.4O (CORRESPONDING To F~1qiH-r AT 20,000 FT.)

01

0 I

0005 005 01 05 I n

FlC.8. AMPLITUDE SPECTRA OF UNSTEADY PRESSURES M=O.3 (a) IN BAY

d = 5 6O (CORRESP~~TN~ M FLIC+~T AT 5,000 FT)

-I

---- - d = 10 4O (CORRESPONDING To FLQHT AT 20,000 FT)

- ._.___ - ._ _-. _ -_,_

0.05 0 I 05 I 7-L 5 IO 20

FlG.8. (CONCL’D) AMPLITUDE SPECTRA OF UNSTEADY PRESSURES M=0-3 (b) BEHIND BAY

l/20 SCALE MODEL M + 03, d = !56-

----- I /6 5 (APPROX) SCALE CANBERRA TYPE BAY

IN CIRCULAR SECTION FUSELAGE (REF4)

i 0 01

o.oo5 \ 0 01 0.05 0 I 05 n 5 IO

REAR ’ BULKHEAD HMC

I .’ \I _c /

(4 IN BAY

0 01 0.05 0 I 0.5 I n 5 IO

(b) BEHIND BAY

FIG.9. COMPARISON WITH TESTS OF REF. 4.

001 005 01 05 I n 5 IO 20

,“:i5: 0 01 005 01 0-S I n 5 IO 30

0. I

w

0.01

I 0005

001 0.05 0 I 05 I n 5 IO 20

01

005

k

ijF

x L 5 I.26

I 05 I n 5 IO so

FGIO. EFFECT OF MACH NUMBER ON AMPLITUDE SPECTRA OF UNSTEADY PRESSURES (a) MODEL

INCIDENCES CORRESPONDING TO FLIGHT AT 5,000 FT.

0 01

0 00s <

01

0.1

00.5

i!$

0.01

I I I I I I 01 005 01 05 I l-L 5 IO i

I I I I I I 00s 01 05 I n 5 IO 4

es I 26

OGO5 0 01 005 01 OS I n S IO 20

FIG.10 (CONCiD) EFFECT OF MACH NUMBER ON AMPLITUDE SPECTRA OF UNSTEADY PRESSURES (b) MODEL INCIDENCES CORRESPONDING TO

FLIGHT AT 20,000 FT.

+10

Al/ \I\/-, I \

(a) IN BAY (RELATIVE TO mL=045j

+ I.0

0

-

/

/

I-- ,

/ F I.26

-1 o- 005 0-I OS I 5,lO 20

(b) BEHIND BAY (RELATIVE TO X/L=I.OS)

FIG.11. CORRELATION SPECTRA M=0.3 d=5-4

+ I.0

= (4

0

-I 0

005 01 OS I 5 * IO 20

(C) BEHIND BAY (RELATIVE TO x/~-I-26)

+ I.0

= (4

C

-10

=P,R

00.5 0.1 05 I 5nlO 20

(d) BEHIND BAY

FIG.II(coNc~D)CORRELATION SPECTRA M=O-3 d=S-b’

$72 c F No. 725 ..I,(&) E.E.CanbcrrB : . ..?.C. C.E. Ilo. 722 h.I.(L+z) c.E.Canberra :

533.6Y5.9 : 533.695.9 : 533.6.013.2 : 533.6.013.2 :

,‘i,!D ‘lWi,i‘L I,!&S”~,Ei,?S OF THE U~.STCIIDY 533.6.011.35 PESS”PLS 111 ,‘ND BCHMD I BOhB EAY (CANBCRPJ~).

WIND TUNNCL I@Is”N%E,ITS OF THE LNSTEADY 533.6.011.35 DRcSSURi;s IN ‘LND BEHIND ii BOELa BAY (C.J83E!-W).

Rosslter, J.E. ar1 Ku-n, 6.C. October, 196%

The unsteady ~resswes xtirg In and behind the bc;nb bay of a 1120 scale model of the Ca”berN aWWZft IZYC been nonswed In cne 8 it X 6 ft transonic tunnel. The unsteady ~csswrzs a-e described by their r.m.s.

“al”es and by amplftude and rorrcl~rion frcqucncy spectra, Tine a”ErS@ values Of the p-esswes wcrc r1sc nea.sl!red.

The results show that tne Ccnbcrra bcnb bay is of the “shallow type” In vmich the airflow! enters the bay and attaches to the roof before being deflected o”t by the rear bulMxnd. FrfSSure f1”ctc~tion.s are nest intense

(over ) --

-+I.(@) E.E.CcnbwW : 533.695.9 :

1” the vlclnity of the i-car bulkhad and on the rmf c,f the bay irhere the flow attaches. ChmelnC the Nxh nli~bw fro, 0.3 to 0.6 hm in Concral little effect cn the u”stm,y ~rcss”rcs (when cx~rcssed no”-dinensiorally) !,ut 1ncrcnsIn~ Incidcwe up t,, +,a, dq,mes ~rrod”cos scm (ICC~‘E~SC I” their i aC”ltude,

Wt 60 K 4

C.P. No. 728

Pubbsbed by

HER MAJESTY’S STATIONERY OFFICE

To be purchased from York House, Kmgsway, London, W.C 2

423 Oxford Street, London, w 1 13~ Castle Street, Edmbureh 2

109 St Mdry Street, Cardiff 39 Kmg Street, Manchester 2

50 Farrfdx Street, Bristol 1 35 Smallbrook, Rmgway, Bwmmgham 5

80 Chichester Street, Belfast 1 or through any bookseller

C.P. No. 728 SO Code No 23-9015-28