-
I UNClk5SIFlEDLA-5658
Reporting Date June 1974
__ c.3 Issued September 1974
PUBLICLY RELEASiU3LEper ~. m~ 16 Date: 4dkL-
CIG74 REPoF‘TmUECTION BY M~u , CIC-14 Date:
f-/6-f5REPRO1)UCTION
COPY
PHERMEX Evaluation of Air Force
Tungsten and U-O.75 wt% Ti Penetrators (U)
J_==m:~FEQ
p:~_-K$—~= 0z~o
~~;—m.=q-.~ m—
1
t!)
:
10s alamosscientific laboratory
of she University of CaliforniaLOS ALAMOS, NEW MEXICO 87544
/\
=
— —z
by
L. W. HantelJ. W. Taylor
t!l?CHUNFJEDi
I
I
1
iIIIIIIIIIIIIIIIIIII
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
ABOUT THIS REPORTThis official electronic version was created by
scanningthe best available paper or microfiche copy of the original
report at a 300 dpi resolution. Original color illustrations appear
as black and white images.
For additional information or comments, contact: Library Without
Walls Project Los Alamos National Laboratory Research LibraryLos
Alamos, NM 87544 Phone: (505)667-4448 E-mail: [email protected]
-
. *
This report was prepared as an account of work sponsored by the
United
States Government. Neither the United States nor the United
States AtomicEnergy Commission, nor any of their employees, nor any
of their contrac-
tors, subcontractors, or their employees, makes any warranty,
express or im-
pliad, or assumes any legal liability or responsibility for the
accuracy, com-pleteness or usefulness of any information,
apparatus, product or process dis-closed, or represents that its
use would not infringe privately owned rights.
.
.
Work supported by Air Force Armament Laboratory, Air
ForceSystems Command, Eglin Air Force Base, FL.
.
.
..
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
~’ ‘NCLASSIFI
PHERMEX EVALUATION OF AIR FORCE TUNGSTENAND U-O.75 Wt% Ti
PENETRATORS
by
L. W. Hantel and J. W. Taylor
ABSTRACT (U)
The LASL PHERMEX machine was used to radiograph the in-teraction
of tungsten and two harnesses of uranium-0.75 wt~otitanium
penetrators with MIL-S-13812A armor plate at 45° obli-quity.
Multiple fixings permitted penetration-time histories ofboth
penetration and nonpenetration shots to be observed.
Penetration trajectory plots made from the radiographs andthe
detailed shapes of the penetrators midway through thepenetration
process lead to several interesting conclusions. Ingeneral, the
penetration process begins with a stage in whichhydrodynamic forces
govern and the penetrator is eroded rapid-ly, followed by a stage
during which the residual kinetic energyof the penetrator punches a
plug out of the armor.
The tungsten alloy’s relatively high ductility is
self-defeatingbecause the penetrator presents too broad a frontal
area to thearmor. The softer U-O.75 wt% Ti alloy produces a largex
fxagment
——— .. . behind the armor than the harder alloy, apparently
because the_
ml”““latter fragments catastrophically at late times.~_
-~ The uranium alloys, in contrast to the tungsten, bend
marked-~tigm~ ly during penetration, apparently because of their
lower elastic:= bl----—3~:1-
rno”dulus and Hugoniot elastic limits. These effects are
accen-Jm+-m’ tuated by the present penetrator design.~so
Metallography of these U-O.75 wt% Ti penetrators revealed._ol0Z ~
~-‘-- -–- centerline quenching voids.~~m?-g’—,
SFl: —————— __ —- ————. ————— ————.-
1. INTRODUCTION
The Los Alamos Scientific Laboratory (LASL), un-der an agreement
with the Air Force ArmamentLaboratory (AFATL), Air Force Systems
Command,Eglin Air Force Base, performed a detailed study
todetermine what uranium alloy would be the bestpenetrator material
for use in a proposed 30-mm gunsystem. 1 As part of that study, the
LASL high-
1. J. W. Hopson, L. W. Hantel, and D. J. Sandstrom,“Evaluation
of Depleted-Uranium Alloys for Use inArmor-Piercing Projectiles,”
Los Alamos ScientificLaboratory report LA-5238 (1973).
intensity 28-MeV PHERMEX flash x-ray facility wasused to
radiograph the penetrator-armor interactionthrough 152 mm of steel.
The radiographs showedthat during penetration, projectile
deformation wasconfined to a relatively narrow region next to
thepenetrator-armor interface. The deformation con-sisted of rapid
radial displacement of the penetratormaterial starting at the
moment of impact and con-tinuing until a plug was sheared out of
the armor orthe projectile was defeated. The displaced
penetratormaterial adhered to the walls of the hole, and the restof
the penetrator proceeded essentially undeformed.This unusual
behavior was attributed to the lowpropagation velocity of stress
waves in uranium
~wp -UNCIASS1-FliD
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
! .“
which requires far too long a time for significantpenetrator
deformation outside the interaction zone.
Since our first report, 1AFATL has proceeded witha vigorous
program to increase penetration byredesigning the penetrator, AFATL
has evaluatedseveral penetrator designs, and one of the
mostpromising is that examined in this program. The pro-jectile
tested is shown in Fig. 1. AFATL tests showedthat penetrators of
this design penetrated better thanother designs tested. However,
for the past fewmonths AFATL has experienced difficulty
inreproducing these results. Hence, AFATL asked us toexamine the
penetrator-armor interaction with our28-MeV PHERMEX machine. In
particular, wewould look for premature core failure and
examineentrance and exit hole angles, core deflection
duringpenetration, and core integrity behind the armor. Wewere also
asked for complete chemical andmetallurgical analysis of both
U-O.75 wtYOTi alloys.
II. EXPERIMENTAL
The experimental program consisted of firing thepenetrators into
50.4-mm-thick by 152-mm-wide by304-mm-long MIL-S-13812A armor plate
at 45°obliquity and radiographing the penetrator-armorinteraction.
Two U-O.75 wt~o Ti harnesses (46-47 R.and 53-54 R=) and W-2
Kennertium tungsten alloywere evaluated. All three materials were
fired at anincident velocity sufficient to penetrate the armor,and
the 53-54 R= uranium alloy was also fired at avelocity insufficient
to penetrate the armor, Severalshots were fired with each material,
allowingradiographs to be taken at various penetrationdepths. A
5-mm-thick 2024 T-3 aluminum witness
Fig. 1.Thirty-nnn HD API projectile (ditnensions
intnillitnelers).
plate 152 mm behind the armor verified penetrationor
nonpenetration by each shot.
The 30-mm rifle used for these tests was built byMathewson Tool
Company according to G. W.Amron drawings. Because our rifle barrel
has adifferent twist and chamber from that used byAFATL, we had to
fire a number of preliminary shotsto establish the powder loads
necessary to producethe penetration and nonpenetration shots and to
besure that this projectile was stable in flight whenlaunched from
our rifle. In all, we fired 22preliminary shots against
50.4-mm-thick by 304-mm-square MIL-S- 13812A armor plate at 45°
obliquity.We took orthogonal x rays of each shot to determinethe
yaw angle and projectile integrity at the target.As received, these
projectiles would not chamber inour rifle. However, by machining
about 4.75 mm offthe front of the nylon driving band, we could
makethe round chamber properly, and yaw at the targetwas held to
less than 1” in each plane. Table I liststhe loads and velocities
selected for the PHERMEXshots. Although these loads were
conservative on thebasis of the preliminary shots, when we began to
firein front of PHERMEX we found that the projectileswere not so
reproducible as we thought, so the armordefeated some projectiles
whose velocities were ex-pected to cause penetration. Because we
had only alimited number of each type of penetrator, we in-creased
the powder load to ensure obtaining ourpenetration sequence, The
velocity was increasedonly a few meters per second.
Timing for the radiographs was accomplished witha “make” foil
attached to the front of the armor plate
TABLE I
POWDER LOADS AND VELOCITIESFOR PHERMEX SHOTS
LoadMaterial (gcIb1379c)
FIardU-O.75 wt% TiSoft U-O.75 wt% TiW-2 Tungsten
Hard U-O.75 wt% Ti
Penetration Series
148148155
Nonpenemation series
Veloeity(m/see)
945
945975
140 899
;
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
target. When the projectile penetrated this foil, anelectrical
circuit was completed, giving us a “zerotime” signal. This signal
was then delayed to triggerPHERMEX at the desired time. The foil
was twolayers of O.127-mm-thick copper shim stockseparated by a
layer of tissue paper. Figure 2 showsthe circuit used.
RI. RESULTS
A. PHERMEX Shots
The penetration shots with tungsten alloypenetrators gave seven
radiographs at delays of 20,50, 100, 150,200, 275, and 325 W. These
radiographsare shown as Fig, 3.
The penetration shots with hard U-O.75 wtYo Tigave six
radiographs at delays of 50, 100, 150, 275,325, and 421 pa, shown
as Fig. 4. The nonpenetrationshots with this material gave four
radiographs atdelays of 20, 100, 150, and 250 ps, shown as Fig.
5.
The penetration shots with soft U-O.75 wt% Ti gavefive
radiographs at delays of 50, 100, 150, 275, and325 ~, shown as Fig.
6. In addition, a nonpenetrationshot at 100-&s delay is shown
as Fig. 7.
The most striking observation in these tests was thebending of
the uranium alloy penetrators. This isshown clearly in Fig. 4b, the
100-ps-delay penetrationshot with hard U-O.75 wtYOTi. This amount
of ben-ding was typical of the uranium alloy penetrators.The rear
of the penetrator is still at 45° obliquity tothe plate while the
part inside the armor is at about56° obliquity. This difference
ultimately results in thecurved penetration path shown in Fig. 6e,
the 325-ps-delay penetration shot with soft U-O.75 wt% Ti.
Incontrast, Fig. 3C shows the 100-pe-delay tungstenpenetration
shot. Here the penetrator is proceedingnormally with no tendency to
bend. This differencein behavior can be ascribed to the fact that
he
I meq 90 v
b
-11+
I 0.1MFD I
Projdctlle
Fig. 2.PHERA4EX trigger circuit.
tungsten alloy’s elastic modulus is about twice that ofthe
uranium alloy so it can resist the bending stressescaused by the
45° -obliquity striking angle. Adding tothis problem is the present
penetrator design thattends to increase the stress component
perpendicularto the penetrator axis. The first penetrator
designtested at LASL1 was a right circular cylinder exceptfor the
very tip, and PHERMEX radiographs show itpenetrating armor at 60°
obliquity without bending,despite this severer condition. It seems
appropriate tosuggest that the penetrator be redesigned with this
inmind.
Another interesting observation can be made bycomparing Figs. 3d
and 6c, which show tungsten andsoft U-O.75 wtYo Ti penetrators,
respectively, at 150-wdelay. Both penetrators had an incident
velocity of971 m/s. First, the amount of tungsten alloy adheringto
the front of the penetrator is significantly greater sothe armor
has a larger cross section to work againstand can offer more
resistance to penetration. Also,approximately 10% more of the
uranium alloypenetrator remains after an equal amount of
armorpenetration. Hence, the uranium alloy should have alower
protection ballistic limit (PBL), and it actuallydoes, by about 60
m/s. Careful comparison of thehard and soft uranium alloy
penetrators shows thatslightly less material may adhere to the
hardpenetrator on the average, but the difference is
quitesmall.
Figures 4e and 6d illustrate another interestingpoint. These
radiographs of hard and soft U-O.75 wt%Ti penetrators just after
penetration clearly show theadvantage of the softer, more ductile
material thatprovides a larger fragment behind the armor.
The PHERMEX radiographs provide further infor-mation on the
penetration process. In particular, onecan ascertain the trajectory
of the back of thepenetrator, the penetrator length, and the
penetra-tion depth as functions of time. If the
projectilevelocities in successive shots had been more ac-curately
controlled, these data would have lessscatter. However, the
reproducibility is good enoughto make the measurement worthwhile
and to il-lustrate the important features. We address thefollowing
questions.
How rapidly does the penetrator decelerate?What is the nature of
the “penetration” trajec-tory?What is the penetrator erosion
history?
Obviously, these phenomena are interrelated.The quantities
measured for each experiment
(within the precision of edge definition) were:
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
a. 20-w delay; v = 981.1 m/s. b. SO-W delay; v = 961.3 m/s.
c. 1OO-W delay; v = 971.6 tn/s. d. 150-w delay; v = 971.6
m/s.
e. 200-ps delay; v = 973.8 mjs. J 275-w delay; v = 969.5
m/s.
g. 325-w delay; v = 981.1 m/s.
Fig. 3.Penetration shots with tungsten penetrators.
4 .
.●
✎
✎
✎
✌
##HSHHWl ;
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
a. SO-W delay; v = 907.0 m/s. b: 100-I.Lsdelay; v = 945.4
m/s.
c. 150-ps delay; v = 967.7 m/s. d. 275-w delay; v = 945.4
m/s.
e. 325-ps delay; v = 946.7 m/s. f 421-w delay; v = 952.4
m/s.
Fig. 4.Penetration shots with hard U-Ti penetrators.
Xb, the position of the back of the penetrator,~, the position
of the front of the penetrator,andL, the penetrator length.
These data are presented, together with shotnumbers and initiul
projectile velocities, rounded offto three significant figures, in
Table II. These dataare also presented graphically in Fig. 8 in
which thefront and back positions of the penetrators and
theirlengths are plotted as functions of time. The datapoints are
connected by smooth curves.
The most striking feature of the curves is the factthat, within
the precision of the data, the velocity ofthe back of the
penetrator remains virtually constantduring approximately the first
150 ~ of penetrationand then drops very rapidly over the next 100
W.Furthermore, it is not coincidental that the plugsheared from the
armor forms during the rapiddeceleration of the penetrator. The
point is that if theplug were not forming, the penetrator would
bedefeated almost immediately.
5
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
.
a. 20-Ps delay; v = 896.4 m/s. b. 1OO-W delay; v = 893.0
m/s.
c. 150+s delay; v = 893.9 m/s. d. 250-P.s delay; v = 914.4
m/s.
Fig. 5.Nonpenetration shots with hard U-Ti penetralors.
Notice that during the first 50 to 100 PSthe projec-tile length
is being reduced at a rate that is greaterthan its penetration
velocity. During this time, theback end of the projectile proceeds
at nearly cons-tant velocity because only the yield strength of
thematerial can be transmitted to it, and that only at the“bar
sound speed” (the square root of Young’smodulus divided by
density). The stress at the projec-tile tip is, however, very great
at this time because itincludes the hydrodynamic contribution from
stop-ping or slowing the tip material. The situationchanges
relatively abruptly, because as the projec-tile length is reduced
the relative effect of the yieldstrength increases and the acoustic
signal transit timeis reduced rapidly. Very soon thereafter, the
projec-tile’s residual velocity becomes inadequate toproduce enough
hydrodynamic pressures to causeeither projectile or target material
to flow away fromthe collision zone against the restraining
stressessupplied by the armor’s strength. Unless the penetra-tion
at this time is sufficient to reduce the shearstrength of the
material around the potential “plug”of armor to a level such that
the residual kineticenergy can supply the required plastic work to
shearout the plug, the projectile is defeated. It is probably
6
at this stage of penetration that the differencebetween a
“mushroomed” projectile tip like that ofthe tungsten alloy and a
relatively smaller diametersuch as the U-Ti presents becomes most
important.(Refer again to Figs. 3d and 6c.) At this stage,
thediameter of the required plug in the armor is pittedagainst the
total energy remaining in the projectile.This diameter enters in
two ways. The shear energyhas already been mentioned. In addition,
the mass inthe plug must be accelerated so that the plug and
theresidual projectile material have a common velocity,In the
Appendix, we present an elementarymathematical model that
illustrates some of theabove points.
B. Chemical Analyses and Metallography
The LASL Analytical Chemistry Group analyzedthe two U-O.75 wt%
Ti alloys, using wet analyticalprocedures to determine the
concentration of theprincipal elements and a quantitative
spectrographicanalysis to determine trace-element
concentrations.Table III shows the results.
,
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
a. SO-W delay; v = 957.3 mjs. b. 1OO-PSdelay; v = 965.6 m/s.
c. 150-ps delay; v = 970.7 m/s. d. 275-I.is delay; v = 978.1
m/s.
e. 325-w delay; v = 984.2 m/s.
Fig. 6.Penetration shots with soft U-Ti penetrators.
Metallography of the hard and soft uranium allovpenetrators was-
performed by the LASL Materia~Technology Group. They sectioned one
sample ofeach penetrator material longitudinally andtransversely
and examined it metallographically.During metallographic
preparation of the samples,centerline quenching voids were
discovered,Micrographs of these voids are shown as Fig. 9. It
isknown that these voids can occur when the diameterof the quenched
section is approximately 25.4 mm orgreater.
We found the microstructure of the hard U-O.75wt% Ti to be
representative of U-O.75 wt% Ti aged topeak hardness. Figure 10
shows longitudinal andtransverse sections. We found the hardness of
thispenetrator to be 570 DPH. We also noted excessivecarbide
precipitate in this sample (Fig. 10 left).
The microstructure of the soft U-O.75 wt% Ti wastypical of
U-O.75 wt% Ti that has not been aged topeak hardness. Figure 11
shows longitudinal andtransverse sections. We found the hardness of
thismaterial to be 470 DPH.
wunllEH-m7
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
Fig. 7.Nonpenetration shot with soft U-Ti penetrator.100-W
delay; v = 940.6 m/s.
IV. CONCLUSIONS
One conclusion to be drawn from these obser-vations is that the
best penetrator should be a simpleright circular cylinder with the
maximum practicalaspect ratio for a given mass. The mass should
bechosen so that the velocity is maximum for a givenlaunch
capability. These considerations cannot beexpected to apply to
extreme cases, but they shouldsurely be applicable to variations of
about 50Y0around the length of the present projectiles. Evident-ly
if the projectile diameter is reduced too much, thebending moment
will be too small. One might alsosuspect that the hydrodynamic
penetrationmechanism might become less efficient, because
thearmor’s strength is fixed whereas the rest of thehydrodynamics
scales.
Our second conclusion is that there is apparentlyan optimum
combination of material properties thatwill give maximum
penetration capability and max-imum final fragment size. The
tungsten d_oy was tooductile, and the hard U-Ti alloy was too
brittle.Although the soft U-Ti alloy may not be ,the
ultimatechoice, it is certainly the best of the three.
Our third concision, which may ,be of con-siderable importance,
concerns the residual velocityof a projectile that has succeeded in
penetrating and,thus, the potential damage it can do to objects
behindthe armor. The data presented here were obtainedwith
projectiles fired at velocities near the PBL, so theresidual
velocity was small. Figures 8a and 8b, inparticular, show, however,
that the most rapid pro-jectile deceleration occurred just before
finalpenetration. One would therefore expect that if theinitial
velocity is increased slightly above the PBL,the residual velocity
after penetration will increase inmuch greater proportion. Put
another way, the
TABLE 11
PROJECTILE POSITION VS TIME
The time zero is the initial collision. The space zero is
theimpact surface of the plate, and the space coordinate ismeasured
positive into the plate along the direction ofpenetration.
shot
1546
155615401547154115451550
1551
1543156115481562
1552156315531554
155515571558
1559
xBack ‘Front LengthT (f&) (cm) (cm) (cm) VO (m/s)—. _
Tungsten Alloy Penetration Shots
20 –10.17 1.00 11.2550 – 7.66 2.25 9.75
100 – 3.42 3.84 7.17150 + 1.50 6.08 4.58200 + 4.38 7.29 2.92275
+ 5.50 8.58 2.75325 + 6.10 8.40 2.30
Hard U-Ti Penetration Shots
50 – 7.80 2.15 9.92100 – 2.67 4.75 7.25150 + 1.25 6.42 5.25275 +
7.92 10.58 2.67325 +11.00 12.50 1.50
Hard U-Ti NonPenetration Shots
20 –10.50 0.75 11.25100 – 3.33 4.50 7.75150 + 0.10 6.OO 5.6o250
+ 6.41 8.58 2.17
Soft U-Ti Penetration Shots
50 – 7.08 2.58 9.83100 – 2.50 5.00 7.50150 + 1.83 6.91 5.08275 +
9.25 12.50 3.25
981961
971971974970981
907
945968945
947
896
893894
914
957966971978
evidence indicates that the residual velocity is verysensitive
to the exact amount of armor penetrated (ina narrow range around
the PBL). This fact shouldhave military significance.
4
.8
*
,.
8
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
‘o- ‘\‘\ .“”------”
!3- .“;
-
TABLE HI
CHEMICAL ANALYSIS OF U-O.75 WT% TI
Element
TicNH202LiBeBNaMgAlSiP
Cav
CrMnFeNiCuZn
SrMo
4CdSnSbBaPdBi
Concentrationin Hard ~loy
0.85 wt%40’253265
-
>
Fig.Photomicrographs ofhard U-O.75 wt% Ti. 250X.tion.
10.Left: longitudinal section. Right: transverse see-
Fig. Il.Photomicrographs of soft U-O.75 w[% Ti. 250X. Left:
longitudinal section. Right: transverse see-tion.
APPENDIX
ELEMENTARY MODEL OF EARLY STAGE PENETRATION
Complete analytical modeling of the penetration8 process in
experiments like those discussed here is, of
course, very complicated. It is perhaps useful,however, to
consider a much simplified but
F. qualitatively fairly accurate model of the
importantfeatures.
Visualize a right circular cylinder of material ofdensity p ,
length L, dynamic yield strength a ~, andvelocity v, which strikes
another material whose mass
we consider infinite. We assume that the entirelength, L, is
subjected to the retarding stress, u ~. Thisis not a bad
approximation after one acoustic signaltransit time. The
deceleration of the cylinder is
pLdv
—=-UYdt(A-1)
except in a very narrow zone near the collisionplane. We shall
ignore this zone. If the instantaneous
11
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
I
velocity of the collision plane (because of To find the length
as a function of time, we rewritesimultaneous erosion of the
projectile and the target) Eq. (A-4) in the formis V, the time rate
of change of L is b
:=–(v-v) ,
when all ejected material flows out of the collisionzone.
In the early stages of penetration, before the armorbegins to
bulge and the plug begins to form, V = kv,where k is a constant
less than unity. In fact, Fig. 8shows that k = 0.5. Assume this to
be true indefinite-ly. We shall now calculate the trajectory of
such aprojectile into infinitely thick armor. Usingv – V = 0.5 v in
Eq. (A-2) and dividing Eq. (A-1) by(A-2), we have
0.5 pLvdv
dL ‘- ‘Y (A-3)
with the solution
()(v: - V2)L= LOe– —
4 (7Y(A4)
where
dLI
4 Uy(A-2) ! “=-2— v~+— In L/L. (A-6)
LO = initial length of projectile,vo = initial velocity of
projectile.
The first interesting result is that the final residuallength of
the stopped projectile is
Assuming that
(A-5)
dtw p
and integrate it numerically. Figure A-1 shows thelength vs time
for the set of parameters used in theabove calculation. Obviously,
the predictedbehavior is not only qualitatively correct; it b
semi-quantitatively correct in the sense that the
calculatedresidual length is within a factor of 2 of those
observ-ed experimentally, and the length vs time curvechanges slope
very abruptly between 150 and 3(XIPs.
The real questions that would have to be answeredto construct a
more accurate theory are, principally:
What determines the ratio of the projectile’s ero-sion rate to
that of the target?What are the details that finally cause a
finitelythick target to form a shear plug?
Further thought along these lines might lead to away of
selecting the combination of projectile aspectratio, mass, and
velocity that would give maximumpenetration with a maximum residual
fragment.
Lo = 11.4cm,P= 18 g/cm3,Vo = 9 x 104cm/s,UY= 20 kbar,
we calculate that L f = 1.84 cm, which is exactly thesort of
value we observe.
01 I I I 10 100 200 300Time(ps)
Fig. A-1.Calculated length vs titne for the projectile
describ-ed in the Appendix.
-1
-*.
KT:11O
12
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE
-
mmmmlmmmm
UNCLASSIFIED 11I1IIIIIIIIIIIIIIIIIIIII
Ul?CMWFIED :—
APPROVED FOR PUBLIC RELEASE
APPROVED FOR PUBLIC RELEASE