w

AD-773 960

TENSILE (COMPRESSIVE) PROPERTIES OF GLASS-EPOXY COMPOSITES AS A FUNCTION OF VOLUME FRACTION

M. J. Michno, Jr. , et al

Monsanto Research Corporation

>

Prepared for:

Office of Naval Research Advanced Research Projects Agency

December 1973

DiSTRIBSiED BY:

U^ilT National Technical Information Service U. S. DEPARTMENT OF COMMERCE 5285 Port Royal Road, Springfield Va. 22151 n-

WT

.... „

Si'ciintv Cl.issifiriTti'in 04 yzs^t? DOCUMENT CONTROL DATA -R&D

f Security clas^ificntion nt tiU\ hmly of »h<!f,icl untl int/i-xin^ .mnottttion rmi.^t he enferrd when thv avcratl rfport is chissilied)

OHIGINATING ACTIVITY (Cnrporut^ author)

Monsanto Research Corporation

2«. REPORT SECURITY CLASSIFICATION

Unclassified 2b. CROUP

J REPORT TITLE

Tensile (Compressive) Properties of Glass-Epoxy Composites as a Function of Volume Fraction

4 DESCRIPTIVE HO J FS (Typv ol rnporl .mil inctustvv liatcs)

*. AU T HORlSI fFiri? name, mid'llv ini'utl, last nitmf)

M. J. Michno, Jr., and F. J. Shea

RLPOR T DATE

December 1973 9tf. CONTRMCT OR GRANT NO

N0001'4-67-C-02l8 h. PROJEC T NO

7«. TOTAL NO. OF PAGES

^ZL 7b. NO. OF REFS

8a. ORIGINATOR'S REPORT NUMBERIS)

HPC 73-163

9b. OTHER REPOR i NO(S) (Any other numbers that may be assigned thta report)

0 DISTRIBUTION STATEMENT

Approved for public release; distribution unlimited

II SUPPLEMENTARY NOTES 12 SPONSORING MILI TARY ACTIVITY

Office of Naval Research Washington, D. C.

I i AHSTKAC T

The mechanical performance of unidirectionally reinforced glass epoxy composites was studied as volume fraction filler was varied over the range 30-70 percent. Experimentally determined quantitites included: longitudinal and transverse tensile strength, ultimate strain, modulus; and Polsson's ratio; and longitudinal and tr^.isverse compressive strength. Data analysis Indicated that all properties varied linearly with volume fraction. Longitudinal properties and Polsson's ratio were interpreted in terms of the rule of mixtures.

Reproduced by

NATIONAL TECHNICAL INFORMATION SERVICE

U S Oeparlment of Commerc» Sprlngf.eld VA22151

DD .,2rr..1473 (PAGE 1)

S/N 010 1.807.680 1 Seruritv Clastification

N.

^

*

-

*

S^t urity dttft^Iliratior

Mechan ics Volume fraction Reinforcsd Plastics Tensile Strength Compressive Strength Moduli U;timate Stra ins Poisson's Ratio

R o L r. W T

L* DO .?oRvM473 IB*™ lfA( ?) ^4/

r< O L E V T

»

J Security Cla;sifici.t.}n ,1

i.

^^ 9" ■

HPC 73-163 ITEM A002

TENSILE (COMPRESSIVE) PROPERTIES OF GLASS-EPOXY COMPOSITES

AS A FUNCTION OF VOLUME FRACTION

by

«

M. J. Michno, Jr. and F. J, Shea

December 1973

D D C ^n?\

Ffn U 1ST4

E Monsanto/Washington University Association

High Performance Composites Program Sponsored by ONR and ARPA

Contract No. N000U-67-C-0218, ARPA Order 876

Approved for Public Release: Distribution Unlimited.

The views and conclusions contained In this document are those of the authors and should not be Interpreted as necessarily representing the official policies either expressed or Implied, of the Advanced Research Projects

Agency or the U. S. Government.

ii M-

rr ;

FOREWORD

The research reported herein was conducted by the staff of

Monsanto/Washington University Association under the sponsorship of

the Advanced Research Projects Agency, Department of Defense, through

a contract with the Office of Naval Research, NOOOIi»-67-C-02l8 (formerly

NOOOU-Se-C-OO^K ARPA Order No. 876, ONR contract authority NR-SSö-MV

^-13-66, entitled "Development of High Performance Composites."

The prime contractor Is Monsanto Research Corporation. The

Program Manager Is Dr. Rolf Buchdahl (Phone 311»-694-l»721).

The contract Is funded for $7,000,000 and expires 30 June, 197'*.

icy

W

1

^

■

"*

TENSILE (COMPRESSIVE) PROPERTIES OF GLASS-EPOXY COMPOSITES

AS A FUNCTION OF VOLUME FRACTION

INTRODUCTION

Filament winding is a ccnmonly employed viable technique for

producing glass-reinforced plastic comr^ents. Processing variables

which affect the mechanical performance of these components include

widing tension, mandrel temperature, and winding speed [1]. Winding

tension directly controls the volume fraction filler (v^). Herein we

consider the effect of variable vf on the strength of glass-epoxy

composites.

A general tensor polynomial strength criterion proposed by Tsal

and Wu [2] appears to have wide applicability and for specially

orthotropic material under plane biaxial stresses (In the absence of

shear stress) assumes the form:

F.a, + F.o. + F-.a. + I n 2^2 in r22CF2 + 2F12al02 " , (0

o. ■ longitudinal stress

o« ■ transverse stress

F., F - 2nd rank strength tensors (contracted notation)

F.., F.-, F.. - '♦th rank strength tensors

The normal stress Interaction tensor F.2 can be determined only from

tests under combined states of stress. However, the non-Interacting

components of the strength tensors can be experimentally de^rmined from

simple states of stress:

Fi -x J_ X' F2 " Y ' Y'

11 1

XX' 22 I

YY'

(2)

(3)

^

'

'&

u

. nr

'Hü. I ili.H'iptB—M—i

where X, X', Y, Y' are uniaxial tensile and compressive strengths along

longitudinal and transverse directions respectively. D-termination of

X, X', Y, Y' as a function of v^. would allow calculation of F,, F„, F,.. f 12 11'

and F2 as a function of vf. If additional determinations of ^«(v,)

were made the applicability of Eq. (1) for a range of vf could be

ascertained. Other commonly employed failure criteria such as maximum

strain require knowledge of the ultimate strains at failure and the

elastic constants. Hence, in addition to strength data we seek to

determine as a function of vf longitudinal and transverse moduli

(E.., E ); longitudinal and transverse ultimate strains (e,." .

e2 ); and Poisson's ratio v.«.

MATERIALS AND SPECIMENS

Compos'tes were prepared using ehe same materials and filament

winding techniques as reported in detail elsewhere [3]. The matrix

resin was Epon 828 plus Shell Curing Agent Z(20 phr*). The filler was

PPG 1062-T6** E glass roving. Unidirectional composites were prepared

by filament winding on a rectangular mandrel with four 6 In. x 8 in.

faces. After winding the mandrel was placed in a vacuum chamber for

45 min. to remove any entrapped air from the samples. The mandrel was

then remounted on the filament winder and the final volume fraction was

established by clamping steel plates to the mandrel faces, the final

thickness of the sample being cortrolled by contact with preselected

shims. Winding stresses in the glass were then relieved by cutting the

glass tape across two corners of the mandrel. The composites were then

B-staged while rotating on the mandrel under heat lamps. Final curing

was at 100oC for 30 mln. followed by 2 hours at 150oC.

* Parts per hundred resin

** Manufactured by PPG Industries, Inc.

2

^

w

r»» v.

TENSILE SPECIMENS

Straight sided specimens 1/2 in. wide and 6 In. long were cut

from the plates at 0° and 90° to the fiber direction. These were

tabbed with glass fabric reinfo.ced laminates to establish a gage

length of 3 in.

Several hoop wound tubular samples at selected volume fractions

were also tested In tension.

COMPRESSION SPECIMENS

Compression samples were prepared by adhesive bonding unidirectional

composite sheets to either side of a 1/2 in. thick a'uminum honeycojib.

The resultant sandwich plate was then cut Into strips 1 in. wide by

1 1/2 in. long with the fiber direction either parallel or perpendicular

to the length. The honeycomb of the rectangular sandwiches was then

machined away (slotted) at the mldsection. The slotting produced an

unsupported mid-length of approximately 0.3 In. Care was taken to

eliminate adhesive from the composite faces along this unsupported

length. As In the case of tensile samples, several hoop wound tubular

samples at selected volume fractions were tested to determine transverse

compress Ive strength.

EXPERIMENTAL DETAILS

The tensile specimens were tested In an Instron tester at a

cross-head speed of 0.2 in/min. An extensometer was used to continuously

record strains to complement the continuous load record. All tests were

conducted at room temperature (750F). Composite stress and modulus

were calculated based upon the original cross sectional area and the

recorded loads.

i

X «I

^

I ': » "- -, , ■1. II Hi, HWWW—1

Compressive samples were also tested in an Instron tester at a

crosshead rate of 0.2 in/min. Details of the testing followed those

recommended in [k] and included the use of a lightly clamped lateral

guide (outside the unsupported length). This method of testing resulted

In the desired compressive failure mode within the unsupported length

of the specimen and avoided end brooming or crushing and premature

buckling failure. The ultimate load at failure was reduced to the

ultimate stresses at failure based on the original cross sectional area

of each face plate and the assumption that each face plate carried one

half of the total load.

All tests on tubular samples were conducted on an MTS closed-loop

servo control system. The tubes were Instrumented with strain gages to

determine ultimate strains and Polsson's ratio.

TEST RESULTS AND DISCUSSION

Stress-strain curves for both 0° and 90° composites were essentially

linear to failure. The resulting graphs are shown in [3] and are not

duplicated here.

Tabulated test results appear In the Appendix. Data gathered

includes: longitudinal propcties X, e^ , E^ (Table Al) and,

X' (Table A2); transverse properties Y, e " , E«. (Table A3) and Y',

-e«.ult (Table AM; Polsson's ratio v. (Table A5). C )mpresslve modul I 22 2]

and ultimate strains were not determined for longitudinal samples.

Compressive moduli for transverse samples were approximately equal to

the transverse tensile moduli.

H

^ twmmmm

■5"

-

:

Data include the observed values of the various quantities and

when appropriate the average values of several observations at a given vf.

Clearly scatter exists and a statistical treatment of the experimental

data Is In order.

In all cases a simple linear relationship was sought and data

were fit to the best straight line employing the least square technique.

The resulting relations

X ■ 225.^

ult 11

11

X'

0.97v

S.O'JV

Y = 5-6^

.u,t - - 1.12v L

E22 - 6.09v

v - is.yAv

ips and associated standard deviations (SD) are:

CO

(5)

(6)

(7)

(8)

(9)

(10)

(ID

+ 6.86; SD = 10.59 (Ksi)

+ 1.62; SD = .Zk {%)

+ 1.70; SD - .50 (Msi)

+68.8C.; SD «= 7.28 (Ksi)

+ 4.13; SD - 0.95 (Ksi)

+ 1.09; SD •= .08 it)

- I.M; SD - 0.26 (Msi)

+13.71; SD ■ 1.6A (Ksi)

These relationships are depicted graphically in Figure 1-8. The solid

lines and shaded region represent the appropriate linear dependency

upon volume fraction plus and minus one standard deviation. The open

circles Indicate a single experimental observation and an open circle

with a scatter band represents the actual experimental scatter about

the average valje. In the case of ultimate longitudinal strain e,,

(Fig. 2) an attempt was made to fit the best horizontal line i.e.

e ult - const, however such a fit resulted in a standard deviatior. which

exceeded the standard derivation of a linear relationship.

if one considers the longitudinal data (Figs. 1-3) the product of the

ultimate strain and modulus for a given vf accurately predict the ultimate

L

• i

w

ti i Wmijil

tensile strength at that v. I.e.

vf " " vf (12)

This result is to be expected in view of the fact that behavior was

nearly linear to failure.

Examination of the transverse tensile data (Figs. 5-7) discloses

the transverse modulus E . (Fig. 7) increases with v, at a greater rate

than does the transverse tensile strength Y (Fig. 5). Hence, the

transverse ultimate strain e (Fig. 6) must decrease with Increasing v.

If for any given v, strength and strain are to be nearly linearly

related, I.e.

■ (e"1' E„) vf v, 22 22 f (13)

A simple computation based on Eq. (8)-(10) verifies that Eq. 13 Is

essentially correct.

A point worth noting is that the transverse strength data Y (Fig. 5)

and Y' (Fig. 8) Include several test results from tubular samples (see

Appendix 1). Thene results agree favorably with the results of tests

on tensile strip or sandwich strip samples Indicating that free edge

and/or size effects were accounted for within observed scatter.

RULE OF HIXTURCS

Longitudinal Properties

The linear form for longitudinal strength and moduli as a function

of v. Indicates that Interpretation in terms of the rule of mixture

Is possible. Consider longitudinal tensile strength X (Eq. b). We wish

>

]

-; <mmm

to express this equation in the form:

X - X-v, + X (1 - vj f f m f

where,

X = composite strength

X, ■ strength of the fiber

X ■ strength of the matrix at composite ultimate strain

v, ■ volume fraction filler.

(IM

Us ing Eq. (5) and say v, ■ 0.50 we determine e ult 11 2.11%. We know [3]

that the modulus of the matrix is E « O.hk Msi hence it follows the m

Xm ■ 3.21* Ksi. Composite strength at v » .50 is determined from Eq. {k)

to be X ■ 119.60 Ksi. We can then incorporate Eq. (U) and back

calculate Xf. The result is Xf - 230 Ksi and the r.jle of mixtures Eq. (14)

can be written:

X - 230vf + 9-24(1 - vf) Ksi (15)

X, based upon the above calculation is a little lower than the value one

would anticipate for virgin E-glass fibers; however, it may be representative

of the strength of a typical fiber after filament winding.

A similar calculation for longitudinal compresslve strength results in

*■

X' - 170.3vf + 9-24 (1 - vf) Ksi (16)

Of course, this computation assumes the tensile and compresslve moduli and

ultimate strains ar« equal, a fact not verified.

Rule of mixtures Interpretation of the linear longitudinal modulus

vs. v, (Eq. 6) results In

E11 - ll.lvf + .44 (1 - vf) (Msi) (17) ht

A

The back calculated E^ - 11.1 (Ms!) agrees favorably with commonly

accepted valu«s of modulus for E-glass.

Poisson's Ratio

Poisson's ratio v.« can also be closely approximated by a rule of

mixture formulation [5]. The appropriate relationship is:

v 12 vfvf

vfvf + v-. (1 " V,) m (18)

where

v, - Poisson's ratio of the fiber

v - Poisson's ratio of the matrix

v, ■ volume fraction fiber

Using the values of v, - .25 and v ■ .35 reported In [6] the rule

of mixture prediction for v.. was calculated and is presented In Fig. 9.

This figure also depicts the experimentally determined values v9 .

These values of v. were used along with knowledge of E,, to calculate *! "

S.» ■ S2. (compliance coef.). The corresponding value of v.« was then

calculated from knowledge of E.. and the relationship -v.» = S.-E...

Clearly a great deal of scatter exists but the rule of mixtures prediction

appears to apply to the range 58% 1 Vr £ 75*

CONCLUSIONS

Longitudinal and transverse tensile strength, ultimate strain, and

moduli data can be well correlated by linear relationships over a volume

fraction range 30-70 percent. Longitudinal and transverse compresstve

strength data are also linearly related to vf In this range. In particular

'rule of mixtures' models for longitudinal strength, and modulus are

appropriate. Knowledge of X(v,), X^v,), Y(vr). and Y'(vr) allows

strength tensor components F., F.., F., F«, to be determined as a function

8

*)

of v,. Reported values of ultimate strains and Polsson's ratio would

a'low computation of the maximum strain failure criterion predictions

as a function of v.. Data available limit this prediction to

58^ £ vf <^ 75% and to quadrants one and four of either principal strain

or stress space.

^ l

I REFERENCES

1. Oesai, R. R. and Kalnin, I, L., "The Effect of Filament Winding Process Variables on the Performance of Carbon or Fiberglass Reinforced NOL Rings," 24th Annual Technical Conference, Reinforced Plastics/Composites Division, SPI, 1969-

2. Tsai, S. W. and Wu, E. M., "A General Theory os Strength for Anisotropie Materials," J. Composite Materials, Vol. 5, p. 58, 1971.

3. Lavengood, R. E, and Ishai, 0., "The Mechanical Performance of Cross-Plied Composites," Polymer Engineering and Science, Vol. 11, No. 3. p. 226, 1971.

k. Structural Design Guide for Advanced Composite Application, Advanced Composites Division, ARML, Chap. 7, Test Methods, 2nd Edition, January 1971.

5. Ashton, J. E. Halpin, J. C, and Petit, P. H. , Primer on Composite Materials: Analysis, Technomic, I969.

6. Lavengood, R. E. and Ishai, 0., "Deformational Characteristics of Unidirectional Glass Epoxy Corrposites In flexure", Jour, of Matls., JMLSA, Vol. 5, No. 3, p. öM, 1970.

/*?

li

t .v.- «Mi —in "mmmi

1

180 [ J2I60 —

^140 c

-

^ 120 CO

■"■

Tens

ile

00

o

o

o

-

° 60 '•o

-

| 40 o>

o 20

X 50

Vf (%)

60

Fig. 1. Longitudinal Tensile Strength vs vf

70

.

!!-

J

. _ -

r ^ ■w Ü

B

O Ik

^m

CO - 3 — o c

"O T 3 "^S—^»*^B"—^^

^— -zrri

r-ts _I \ x

0) 1 ^— o E 1 —

■♦-

3 h- _i *

Ui o 1

40 50 60 70 Vf(%)

Fig. 2. Ultimate Longitudinal Strain vs vf

^

M

r <

!-

e

o

■HMMH

8

7 o .E 6 ■o

I 5 C o -J 4

LÜ 40 50 60 70

Vf (%)

Fig. 3. Longitudinal Modulus vs vf

1

1

!

c a;

>

V) CO

no

2! 100 E o o

c •a 3

90

80

S 70 c o 60

x 50 30 40 50 60

Vf (%)

70

Fig. <♦. Longitudinal Compressive Strength vs v^

»I» i "i i "rvmrntm

I ^

in j*

10 •♦- D) C a; 9 k.

•♦- (^ 8 if c 7 <u h-

6 Q> V) k- Q> 5 > V) a. o 4

30 40 50 60 Vf (%)

70

Fig. 5. Transverse Tensile Strength vs v-

80

r »

■■■■■■■

g .8 c 'ä .7

S 6 ifi

<ü .5 >

§ 4

0) .O

o i .2

-I (Si 3 (M

w 0 40 50 60

Vf (%)

70 80

Fig. 6. Ultimate Transverse Strain vs v.

,

r

^v"

I i mmi m\ t mmmmmmlmmmim ■■■■■■a.

o

£

O

0)

g

It CM

UJ 0 40 50 60 70

Vf(0/o)

Fig. 7. Transverse Modulus vs v^

1

i

"^^ wmm i

>i I I II—I I M ' " ■

■

to

S 30

0) 28

26 > (/) V) Q> k-

Q.

E o o

k_

> CO c

24

22

20

18

16

30 40

1

—

o // — o

ft // ■■

■

rfL %

— 2 rß • - •

—

^

/ /^s*—

—

1 1 i 1 1 1 50

Vf (%)

60 70 80

Fig. 8. Transverse Compresslve Strength vs v-

^

(

\V

w

* ""»'■ »I«I,M

< '

^

^

o oc jm

o en

O CL

0 .2 4 .6 .8

Vf Volume Fraction Filler

Fig. 9. Poisson's Ratio vs v.

1.0

. (

ih

a

■•"■^

■ m

«S C<4 N N OllcN «M CM

IA IT»

I I I I I <S CM CM CM|CM CM CM CM

O ff\ — lo — c^1 ■"; (M — CM|CM CM «M CM

CMICM

i i i i i \r\\o 3 u>|m in ir>>o »O

1« oo O^CM — r^ r~~ r^ c — <r> c o — — o

oofoo-» cMJ;>orA rnjo ^ -^ —. ^ ^ "".

„O — O CM CM MCM r^vU- rom^ ^ £ ^ x ^ — — —

0> r^ j ff« m

C ««•■ oo f.-3- mcM — CM CM

O NO CM OO

0*1 r-* o> trv CM — CM CM

IA t^ J- CM

2 g u. < -I o — -I z I p z ■ d E i ae — Ul o E 2 x m •it —

CM IA IA IA

o z

a. <

i

< UJ

I I

CM IAOOOOIOI.» «r><r>oo >ol m t^la> <n so CM o -»

u

jS »»\ 00 00 "-lO CM >0 0> «^ ^_ I I I I I ' ^ u^^ tAllÄxÖ JT -» lA jr -»Ur ^ lA

<r> r~ CA

1 ^^ * IAl

r^cMk r^ \ö \o ^- oo iA

CM IA Ö lÄloN O "»•»tiDS „ ft: r^r^lr^ oo oo<r>ooooo^

IA CM IA i— (A CM (ANO

VA 1*1 «A rAI« ^5 *S CAOO oo a\a> o o

^ \ß CM <M00<M— llArAtAOClCM

O J vO vO vOl— IA f* IA r^ CC* a^3^a^o^|(^la^oo o o^o ff*

u- o IA

CA ON CA

IA

-^ fM CM

m 8 (A

3

l] u- — u ID 01 *< "O

2 5

-^^r—^ rr .

C

C O o

CO s^^

5 *J ►-

3 — O CJ O — 'O 00 r->N\o CM CM«"» eo

co j- \o mi 1 1 ... ... u

M CS <M «M .>< — — —. CM «M CM CM CM CM CM CM CM

^U» \o i«i —

j|lM CM CM CM

X 00 Ö iv-, tsi vn - ~ r^ CM

HI I I I I

CM j-\o oo — !<M IA f-ooa» —

r« r^ »o \o r^|r~« r» >o r~ vo r~ . MD vr> r-~

—» CONACO —cMCTkoo vr\ \o vo <r\ mirir«. ir>

<r> r~ to <»MA o — ■* -a- J- -»«• JT ro H5

o -3- CM

<»MA — . CM M) r»

fwO CM O (M CMKO 0O — vO

i^ <j> co ro ro <r> r»- CM p«- r»- \0 \& tr\oooONOr».r^iAi/\

S — j- m \o \D \Q \D \0 2 o

r-.

^

C I

m

I 3

;

1

L

, -T"

^i

^

(

in o\ ^- *— CO -T «^ h» VO r~~ r^ \0 <M CMJOO CM CM r^i cr. o IA r-» j- JT vo en o^ &\oo oo oo oo

-3" CM (TV oo

CM o-i

I^OC 00 oo O"* O"* o^ c*

O — — O

o o

rrwO o oo

P^iriOI-T f*>— •—l^lA--r^COCM<M

J- J- J- VD CM J- CM, oo oo oo «Ti o oo tr>i

(»^ CMT» CM O O 00 OO O O O O

CQ <

CM -T O r«-. in r» J- -3- in IA in m

mtr»\oocM—ocsiAi^m _ \o oor^oo-3-ot~.\ootn—«>j>-vor*\Of*.(r>m

vo —oomj-r»«cM\otnmcM rv. oooo r^ r«oo r->|ao r-.f«~ f^ r-.

CM 00 ool<

iA»~-Too*^' oo ^.OICM ^- m— r^o oo a>oo CM co<T\<no>oooooooolo^ooooooco cnoooooo <r\

00 CM

CM en

IA IA

5

_

^ -

,K LOvD in oo i^ trv o o III I

j- CM r~ oo mko o vo «^ rM^ <M 01 i" <*>-T r>"> rA

\0 \A

CM — — CM

vO OO (Tt O r^i

— — — CM|<M i I I I I CO — vO Ol CTi

CM e^» CO CM CM

UA00

CM CM

« cn

r in V

ta

crioo ^ in o «_ so vO vo LA CO v£ v£ >•

coo-.T^i-a- M c^ r». ^C;K "j '^■a; "^^ °r?0? **. t ivoKo r^. cn \ö vö r~|r^ oo oo o~ioo oo onlcncn »* r*

< H < Q UJ *0 m ^J

1*- o •— CM vD m ;> in m in in z UI 1-

Ul Irt oc UJ >■ u>

in in in o ,- m -a- \0 \D NO >o R »A m

3

3

ra

0)

l- o a. a;

o

<■ « — m mKo o^ m i^ o inlr^ o -a- q^loo p o r-in \opJvo r>»uj vovoKo vc ß^K r

CM I I CM

mho en m r^ o min» «J -J "'rs

i ■! I • • » •! • i

— oo mu JT j- j-ta- in rB- in

I I ml i. o— oo —

CM I I • • I jH Jill'

00 mcMrn m m CM

%o in-a- CM -r

ii ii

sol 0O| -3" j-^ <r> <ri vo in in a» CMOO—vol»^o mo in

JE *-^ mm NO mso so >-

so r» »A vo vß |r^ vo mr^l

b, o o^ <r>

so r^ r»

m CM CMICM j- — CM oo

vö vo t^-po so en NO vo

3 I^LDOCD

3 No|mvo

en

1- * * * *

m in

-T -3- o m

o m

if ,

w

— ^- O «M J- — 3 — CM ^ r«« ^•

CM .... . «M f»> O «"■> tsl .— u

I

r«-\ CNJ CN

r~. — — n f*> oo o -t CN C-J

O O CM CM

5-a-kD<n cr\ CM r^ crilin JJ-kOJ- 'A -3- p^ l.^La- .l «M|(M CM CM CM CM CM |,-1

a. a. o

> z <

< < 2

U- > r^ oo o LTl LA vT)

■I--

00 — -3-

I i-

I

2

•— r>.r».\Oroj-in\o rAcr\j-r~Jloooor^m cvoon .— \oo — oor^tri»- Jf o^o^o^ooc^\o^a^ .— c^^ooolo^o^oo .— oo CMCM-^^CM.— oo^- *-' — — — — ——— CM— CMCMCMl— — CMCM CMCMCM CMCMCMCMCMCMCMCMCM

\

- -

-p^v

■

vf(*)

ii8

58

59

61

63

75

TABLE A5 - POISSON'S RATIO

^1

.133

.148

.081

.120

.103

.112

.098

.106

.109

.112

.077

.068

• 117

'

ii