A mass-distribution criterion for predicting the effect of control manipulation on the recovery from a spin

MISSING IMAGE

Material Information

Title:
A mass-distribution criterion for predicting the effect of control manipulation on the recovery from a spin
Alternate Title:
NACA wartime reports
Physical Description:
11, 3 p. : ; 28 cm.
Language:
English
Creator:
Neihouse, A. I
Langley Aeronautical Laboratory
United States -- National Advisory Committee for Aeronautics
Publisher:
Langley Memorial Aeronautical Laboratory
Place of Publication:
Langley Field, VA
Publication Date:

Subjects

Subjects / Keywords:
Spin (Aerodynamics)   ( lcsh )
Aeronautics -- Research   ( lcsh )
Genre:
federal government publication   ( marcgt )
technical report   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
Summary: Results of spin-tunnel tests of 65 models indicated that when the airplane design simulated that of the earlier single-engine type, with mass distributed chiefly along the fuselage, aileron-with and elevator-up settings aided recovery, and the rudder was the predominant control for recovery. When the design approached the design of multiengine airplanes (or the more recent single-engine airplanes with wing tanks and wing armament) with the mass distributed chiefly along the wings, however, aileron-against and elevator-down settings were conducive to the most rapid recovery and the elevator was the predominant control. The primary importance of the mass distribution of an airplane in determining its spinning characteristics is demonstrated and a useful criterion for predicting the optimum control manipulation for recovery, based on a non-dimensional mass-distribution parameter, is presented. Charts that should be useful for such predictions to both the pilot and the designer are included.
Statement of Responsibility:
A.I. Neihouse.
General Note:
"Report no. L-168."
General Note:
"Originally issued August 1942 as Advance Restricted Report."
General Note:
"Report date August 1942."
General Note:
"NACA WARTIME REPORTS are reprints of papers originally issued to provide rapid distribution of advance research results to an authorized group requiring them for the war effort. They were previously held under a security status but are now unclassified. Some of these reports were not technically edited. All have been reproduced without change in order to expedite general distribution."

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 003594848
oclc - 71002941
System ID:
AA00009459:00001

Full Text
k/P 6' I~


NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS





WARTIME REPORT
ORIGINALLY ISSUED
August 1942 as
Advance Restricted Report

A MASS-DISTRIBUTION CRITERION FOR PREDICTING
THE EFFECT OF CONTROL MANIPULATION ON
THE RECOVERY FROM A SPIN
By A. I. Neihouse

Langley Memorial Aeronautical Laboratory
Langley Field, Va.








NACA


WASHINGTON
NACA WARTIME REPORTS are reprints of papers originally issued to provide rapid distribution of
advance research results to an authorized group requiring them for the war effort. They were pre-
viously held under a security status but are now unclassified. Some of these reports were not tech-
nically edited. All have been reproduced without change in order to expedite general distribution.

L 168
DOCUMENTS DEPARTMENT



































Digitized by the Internet Archive
in 2011 with funding from
University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation


http://www.archive.org/details/massdistribution001ang









NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS


ADVANCE A7STRICTED REPORT


A MASS-DISTRIBUTION CRITERION FOR PREDICTING

THE EFFECT OF CONTROL MAITFULATION ON

THE REC)V7RY FROM A SPIN

By A. I. Neihouse


SUMMARY


Results of spin-tunnel tests of 65 models indicated
that when the airplane design simulated that of the ear-
lier single-eng-ine type, with mass distributed chiefly
along the fuselage, aileron-with and elsvator-up settings
aided recovery, and the rudder was the predominant con-
trol for recovery. When the design approached the design
of multiengine airplanes (or the more recent single-engine
airplanes with wing tanks and wing armament) with the mass
distributed chiefly along the wings, however, aileron-
against and elevator-down settir.s were conducive to the
most rapid recovery and the elevator was the predominant
control.

The primary importance of the mass distribution of
an airplane in determining its spinning characteristics
is demonstrated and a useful criterion for predicting the
optimum control manipulation for recovery, based on a non-
dimensional mass-distribution parameter, is presented.
Charts that should be useful for such predictions to both
the pilot and the designer are included.


INTRODUCTION


During the past 5 years, 65 models, representing air-
planes covering a wide range of dimensional and mass de-
sign characteristics, have been tested in the NACA free-
spinning wind tunnel. As is to be expected, these models
have shown varied spin and recovery characteristics, re-
flectin- the differences in the proportions and mass dis-
tribution of the models. A consistent difference, however,
in spin and recovery characteristics was early apparent














between models heavily loaded along the fuselage and those
lightly loaded alonigu the fuselage, or he.vil; loader. along
the wi:r.s. In an effort to establish mass distribution,
and not aerodynamic characteristics, as the prim.ar: fac-
tor causing this difference, a series of speci I1 tests was
u:.-ertaken for many of the models, and results of such
tests have been accu.mulated for 19 representative dosi i.s.
For these tests, the mass distribution of each model was
varied and models hose mass distribution was originally
chiefly along the fuselage were reloaded until the moss
was distributed chiefly along the -ings. Models loac:d
chiefly alone; the wings liko'-ise had their mass distribu-
tion reversed.

A qualitative analysis of the results was obtained
for 65 models tested in the spin tunnel, as well as of
the results of special tests for 19 of these models.
Definite rules have been formulated concerning the effects
of control manipulation on the recovery from the spin, as
influenced by the airlane mass distribution. A criterion
based on a nonditenrion~i mas- distribution parameter has
been established for predicting these effects.


AprARATUS -AID TESTS


The spin-testil,: tech-nique in the i. CA free-spinnin-.
wind tunnel and the construction of spin models are de-
scribed in detail ir reference 1. The models, constructed
of balsa, are tballrnstc- for d. :- ic similarity to the cor-
respoading airplaiir 'b7 istallttion of Roperr heights at
s'itnble lo cot ionGs. An auto:at ic clockwork delayed-action
amchanisnBm or a man.ietic remote-control mechanism is in-
stalled in the model to nctunte the controls for recovery.

The model '-ith the rrdder set 'ith the srin is launched
in the -in <- nand into the vertical upDrard air stream of
the tu.nel. -'.e airspeed is adjusted to equal thi vertical
rate of cdesceit o& the model er.d 'he model is thus k.;t at
a fixed hci--ht until rocover;- is at cat d. Recover: is
general': atter.oted by' re-versal )f the rudder alone from
f'll with to fJull iavist tho s:', alt nox-:h the mechanism
an; be arra-i.cd to .-o- anyr or all cf the controls. T-ox
rccove:y is .iudgod y tl.e ni. c of turns from the move-
mrnt of the rudder to tw. c cs-3ation of the spinni-, rsta-
tio.. The effect of ailiroa setti on the srinnr..
characteristics is usual. ec' luit.d b; a co:-.rarison of













the number of turns necessary for recovery by rudder re-
versal alone from spins for '-.ich, for example, the ailer-
ons are set (not moved) with the spin (right aileron up
in a right spin) and the number of turns necessary for
recovery from spins for which the ailerons are set against
the spin. Results of spins in which the elevator is full
up are compared with results obtained for spins with ele-
vator neutral or full down. In a fer instances, for the
special tests, the effects of aileron and elevator set-
tings have been based on a comparison of the vertical
speed ian the attitude of ths steady spin.

The models tested in the spin tunnel have covered a
wide range of dimensional and mass characteristics and
include seaplane and landplane, biplane and high- and low-
wing monoplane types, and multienrine and single-engine
designs. The 19 models used in the special tests repre-
sent different ty;es. For the special tests, the mass
characteristics were varied by moving ballast weights
from either the wing tips or the fuselage extremities to
the center of gravity or by moving ballast weights to
either the wing tips or the fuselage extremities from the
center of gravity, the position of the center of gravity
being kept constant.


RESULTS


The data analyzed are presented in figures 1, 2, and
3. These figures are an attempt to represent graphically,
by a single point, the important mass-distrib"tion char-
acteristics of each model. In table I the models are
given numerical designations to permit tLeir identifica-
tion in the figures.

In the Euler equations of motion, the influence of
the mass distribution depends on three factors: IA Iy,
ly Ig, and Ig Ix, where Ix, Iy, and IZ are the
moments of inertia about the X, Y, and Z body axes,
respectively. For presentation in the figures, these
factors have been -ade nondimensional by dividing by mb ,
where m is the mass and b is the span of the airplane.
Iz Ix
The parameter -- was taken as the ordinate for the
mbe
figures. This parameter is a factor affecting the inertia















pitching moment and increases when mass is added along the

fuselage. The abscissa y-5Z is the factor affecting
mbL
the inertia rollin0 moment and the neat ive values numer-
ically increase as weight is added along tle wings. Inas-
mu.ch as the sum of the three mass parameters is equal to

zero, the value of the third paramot r, L--- nlay be
mb
indicated by a third scale at 4b0 to the ordinate and ab-
scissa scales. Tnis tnird parameter is a factor affect-
ini, th,: inertia yawiing mnor.ent, the large positive values
iviicati.g6 that the mess distribution is chiefly alor.E the
wings and the large ne;'tive values indicating that the
mass distribution is chiefly alon,; the fuselage. The three
k- k kZ
partz meters :n y also be w7ritton as -----, ------- ,
kR k a b b
y- ~ ay
and "-- respectively, here ky, ky, and kz are

the radii of ryr.tion abort the X, Y, and Z axes,
rear ect ively.

Figure 1 sho'-s th effect of eileron setting on the
recovery characteristics .,s indic- trd b,- routiLe tests.
Aileron d.at were -vfil ole for only 53 of the models.
The ty;nu of points used to des4_in-tc the different models
indic'+,tos whether :ctti;:i,: the rilerons with the spin or
against the s'.in r '.,auc: the .urns for recovery. Figure
2 gives sjmil-Ir infcr-fr tion for tnri cle-'tor, dlti lbein-
-vilablc for 60 of the models. The points indicate
t'ahtther ele'r:tor-ur settir's or elevator-down settings
iru moru favorable for recovery. Fi ire 3 presents the
r sul:ts of s cil. tests of 19 models -ith almer ,d mass
district ion. In this fjuru, differoat r:.'ss rrnnGemont s
of tn r mre od el are "re re ser3 tel I'y tnc same number and
thu lot ,,r "a" is employed to denote the altered or ebnor-
mael loadiia., condition. The syrmbols indlicate the effects
of both ailerons nrd elcvetor setti.rgs.


DXSCUSICN

C i-t_.'r.x- :.r_[ r '"'.t '.i f .o :_tr l___- D ff .cts. An in-
spection of the fi-.ures sho-s a distinct gro"r i:- of the
points rcpresenti'-. tn different effects of control













settings. partial separation of the effects is obtained
by independent consideration of each of the three mass
parameters. The most complete separation, however, appears
to bc given t.,1 consideration of the inertia yavirg-rnoce:..t

parrTeter -----,
mb2
Examination of figure 1 indicates that at a value of
IX ly -4
the inertia yawing-moment parameter b--- -- of -50 X 10 ,
mb
almost complete separation of the aileron effects takes
place. For larger negative values, ailerons rith the spin
usually had a favorable effect on the recovery character-
istics .:'d ailerons against the stin had an unfavorable
off ect. As the rameter value of -50 X 10-4 was ap-
proached, instances were observed 7-here aileron setting
had. no noticeable effect on the recovery characteristics.
For negative values of the Tarameter numerically smaller
than -50 X 104 and for positive values, the aileron
effect reversed so that aileron settii.-,s against the srin
had a favorable effect on recovery: whereas aileron set-
ti.,-'s with the spin were detrimental. In the vicinity of
this reversal value, a critical region existed for which
it appears that only slight variations in mass distribu-
tion may completely reverse the aileron effect. An excep-
tion to the general rule -as obtained in this region in
only one instance.

The effect of elevator settings, according to the
data of fi -ire 2, tends to reverse in the neighborhood of
a value of the yawing-moment parameter of zero. There ap-
pears to be a critical region between the values of
2C X 10~4 in which the effect of elevator settings may
be in either direction. For negative values of the param-
eter numerically greater than -20 X 10 elevator-up
settings were usually conducive to most rTpid recovery.
In several instances, however, for models that gave either
very flat or very steep spins, the elevator setting had
little or no effect. For positive values of the parameter
-4
greater than 20 X 10 on the other hand, elevator-down
settings were very definitely instrumental in effecting
satisfactory recovery. In an extreme case, no recovery
could be obtained from the elevator-up, aileron-neutral
spin by full rudder reversal alone; whereas movement of
the elevator alone from the full-up to the full-down posi-
tion gave satisfactory recovery.















The data from the special tests for 19 models, given
in figure 3, appear to prove that the separation indicated
for elevator and aileron effects in figures 1 and 2 depend
predominantly on the mass distribution of the models rath-
er than on aerodynamic factors. The 19 models tested are
believed sufficiently representative of different airplane
types to permit a generalization of the conclusion. Iodel
15, for example, represents a lightly loaded, sin le-en-
gine reconnaissance monoplane whereas model 6 represents
a high-speed, heavily loaded, twin-engine attack airplane.
It must be appreciated that aerodynamic factors may modi-
fy the results for some combinations of mass arrangement
and extreme aerodynamic design to the extent that the con-
trol effects may be dictated by the aerodynamic character-
istics.

Sequence of control manirulation for recovery..- The
conclusions drawn from the figures are particularly sig-
nificant in that they indicate that the relative imrnor-
tance of the different airplane controls for recovery from
the spin may change radically between airTlanes of differ-
ent types, Prior to the recent extended application of
wing armament for combat tvres, airplane structural desi-n
procedure was such that the airplane was characterized
structurally by relatively lieht wings. Practically all
the disposable load wa.s carried in the fusclare, althou-h
some gasolia r might be carried near the center of the
wi.ngs. These characteristics still apply to the private-
owner class of airplanes. This structural arra.ngemcnt of
.the airplane results in a mass loadi.'. chiefly alo.. the
I-- Ix
fuselage and the value of -t-_- will tend to be lar.-;
mb'
and positive, while the value of the inertia yawing-moment
IX Iy
parameter --- is nec.-,tive. The installation of
mb"
wing er.jines tends to increase the eight along the wings
and it can therefore generally be assumed that multiengine
airplanes have high nce-tive values of the rameter

y IZ IX ly
---, and positive values of the parameter mb *
mb mb
Present-day military desi;: i of single-engine airplanes is
also toward heavy wines. Tne desire for increased ra ge
has increased the amonuit of gasoline carried in the wings.
Guns and ammunition nre carried oiutbtord of the propeller,
and the metal wi ..s with the mechanism for retracting the













landing gear are inherently heavier than in older designs.

The results of the model tests sho- that, for the
earlier si..gle-engine military design and the present-day
privately owned airplanes, the rudder is generally the
predominant control for recovery from the spin and that
full rudder reversal is the most effective control ranip-
ulation. Movement of the elevator to the down position
before the reversal of the rudder tends to shield the rud-
der and retard recovery; whereas, movement of the elevator
after the rudder has been completely reversed and rotation
has be..un to slow up may offer a favorable pitching romont,
tending to aid recovery without adversely.affecting the
rudder action. Movement of the elevator alone rarely yives
recovery. Because high rates of descent will probably be
associated with recovery with full-down elevator, the amount
the elevator is moved down will deCeLd on how rnuch assist-
ance is needed from the elevator to produce a satisfactory
recovery. The effect of ailerons -ill be contrary to the
effects expected in normal flight and holding the ailerons
against the spin -ill retard recovery; -hereas holding the
ailerons 'ith the spin will assist recovery.

For multiergine airplanes and for the more recent
single-engine military desi-..s, the elevator tends to be-
come the predominant control for recovery. The movement
of the elevator down is essential to a rr-id recovery.
Rudder reversal, although of less importance than eleva-
tor reversal, will generally improve recovery. Aileron
position is critical and ailercn settings -ith the srin
may greatly retard recovery; whereas ailercn-aiPinst set-
tin s will be favorable. All controls for airplanes of
these types have the effects tl-t -ould be expected of
tem in normal flight.

It may be said in summarizing that, for airplanes of
relatively lihit loading alo-.: the wings, full rudder re-
versal before moving the elevator down is imperative; mov-
ing the elevator down after the rudder reversal is desir-
able. Fc-r airplanes heavily loaded along the wings, mov-
i :. the elevator down is imperative; full rudder reversal
is desirable.

Ar-l1icatioe to fli :.t.- T..e values of the criterion
at which the aileroi. and elevator effects in the spin
reverse, as shown by the figures, ajtly strictly to mod-
els only. The general conclusions, however, should be
apTlicable to fli.::-.t, although, because of possible scale














effects, the reversals may occur in flight at somewhat
different values of the criterion than are indicated "y
the tunnel data.

The meager cor arative flight data available indicate
that the values for the reversal of aileron and elevator
effects will probably be changed somewhat but there are
not enough full-scale data available to fix the flight val-
ues. It is desirable that more flight data be obtained in
an effort to establish definitely the values in flight at
which the aileron and elevator effects reverse.

Explanation of mass effects.- A possible explanation
of the dependence of the effectiveness of the elevator
and ailerons on the mass distribution is presented briefly.

The application of Euler's d.-r.r.ziceal. equations to the
case of an -irplane in a steady spin :ives, for the iner-
tia yawing moment about the body axis, the expression

(IX Iy) sinlQ cos aO"

whcre

is the angle of wing tilt to the horizontal,
positive when right wing is down

a angle of attack

2 angular velocity ?bout spin axis

For a spin in any given direction, the algebraic sign of
the inertia yawing moment depends only on the al, -braic
signs of IX Iy and the :.. le Q In a right spin,
the tunnel results indicate that setting the ailerons with
the spin leads to a positive value for 4Q; whereas setting
the ailerons against the spin leads to a neL,?tive value
of 0 For models loaded so that IX Iy is negative,
setting the ailerons with the spin will produce a favor-
able effect in that the inertia yawing moment will be neg-
ative and will act to turn the airplane away from the di-
rection of rotation (against the spin). Conversely, for
desi, ns where Iy Iy is positive, ailerons set with the
spin will produce an inertia yaring moment in the direction
of the spin. The fact t'it, for the results presented in
chart 1, the reversal of aileron effect does not take place
when IX Iy is zero can be attributed to secondary














aerodynamic factors. A similar explanation may be applied
to the elevator effect, as the model results indicate that
setting the elevator up usually leads to a positive value
of QP and elevator down to a negative value of (Q for
a right spin.

For the present, only the qualitative effects of the
controls are considered. No attempt is made to predict
the magnitudes of these effects which are probably influ-
enced by many secondary factors, such as the autorotation
characteristics of the wings or the yawing moment due to
sideslip. The values of the inertia pitching and rolli.
moments also undoubtedly influence the spin and recovery
characteristics, although on the basis of existing data
they do not appear to be of primary importance in the pro-
diction of the direction of the control effects.


CONCLUDIUhJ REMARKS


Data presented indicate that mass distribution is a
primary factor in determining the direction of aileron
and elevator effects in recovery from the spin and that
the directions of the effects and the optimum control pro-
cedure for recovery, therefore, may be predicted qualita-
tively on the basis of the mass-distribution parameter.

When the airplane design simulates that of the ear-
lier single-engine airplane, with mass distributed chiefly
along the fuselage, aileron-with and elevator-up settings
can be expected to aid recovery. When the design approaches
that of a multiengine airplane (or the newer single-e::gine
airplane with wing tanks and wing armament), with mass dis-
tributed chiefly along the wings, aileron-against and ele-
vator-down settings will be conducive to the most rapid
recovery.

From the normal control configur-ation for spinning
(rudder full with, elevators full up, and ailerons neutral),
the most rapid recovery for any airplane will generally be
obtained by full, rapid rudder reversal followed immediate-
ly by redid movement of the elevators to the full-down po-
sition and of the ailerons in the direction determined by
the mass criterion. For airplanes loaded chiefly along














the fuselage, the rudder movement is essential, whereas
for the airplanes loaded chiefly along the wings; the
elevator movement is essential.



Lant-ley YMemorial Aeronautical Laboratory,
Nutio al Advisory Committce for Aeronautics,
Langley Field, Va.



REFERENCE


1. Zimmerman, C. H.: Preliminary Tests in the N. A. C. A.
Free-Spinning Wind Tunnel. Rep. No. 557, NACA, 1936.











TABLE I

KEY TO :."0'ZI DESIGNATION ON FI-'y^ES
[Unless otherwise indicated, comparative recovery data
available for both ailerons and elevators]


Model Airplane IModel Airplane
desig.et ion represented desig.:.t ion represented
(a)
- - -- -


aLetter "al after model number
normal as indicated.


bSeaplane.


34
35
36
37
38
39
40
41
42
43
44
45
c46
C47
48
49
50
c51
c52
53
54
55
C-56
57
58
59
g60
g61


64
65


P3-2A
XF-46
X1,I17- 1
S3B- 1
bXO S N-1

XF3F-2
P-36A
XF4F-3
I7ACA CANARD
p-39
XF4F-2
P-26A
eXOSN-1
b0X S i- 1
XS32U-1
bXOSS-1
:,del 159
X3F B-1
bXOS2U-1
XP- 40
Yp-37
XS 2A-1
eXOS2U-1
CXOSS-1
bxs0o- 1
CXS03C-1
XN2 Y-1
F4B-2
F2F-1
S3D-1
Yp-43
XP-47B


indicates loading varied from


cComparative recovery data for elevators only.


dComparative recovery data for ailerons only.


eLAndplane.


gNo comparative recovery data.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Cl 7
18
19
20
21
d22
23
d24
25
26
27
28
29
30
C31
32
C33


XP-50
XI-2
XB-A3-3
XF 5- 1
XP-38
YF;.- 1
a-20
NB-1
BT-14
XF4J-1
B-26
bH3y--3

BT- 9
0-52
XSE-1
V-143
PB-2
P-35
A-17
XET-13
XF2A-2
XSB2 -1
F2A- 1
XET-11
XI 3F-2

XFL- 1
XF2A-1
XBT-12
fV-143
p-44
XF3 2-1


fLengthened empennage.









NACA Fil








I
'-q^ 4 a- a





























E 4E
X -, --4...


























'0
/,r n al



.o o:_,!'!- \\N\ i \ \ o 1

,\\, \ \,tt\ t a\ I \ I a





-,o ~_ i i ....\ o_, _____
,4 _'. t_ .
&CO ^,Q/'/- -. 4- 4.- -4-.--





S 4 I --




















0: 70
t I \ 0"7 0\i \0' 0i \ Ol \ Oi \0\\ \\l\oE
\ __ ___ ,' (I



N ~~ .:- -. -t,



Cl-__ Y-.^" K *, ,

S, '-- -










4. 1 '

,-4,4
-i ^ +t-, -' -

'[ 4 4 ; '- ; v



U^.Oir Ot' OS OOQ Ogz uue Osi 001 OS 0
a&D~WsnJ 6u~ pap SS gPA^rJT




44-
4-. 4f



-0
42 _
.'4- >l
o' CT)
-g -a-
o; 4o o



C' 0
o, E
il)














4 <' +










1 \ \\

'o \\000 ,\
^o '^vX\\

^0 *^\^'4,



&y^ $ 0'Sf -- -


w
41 t
o- "
c


o ,

0 A 0
U, U, Z

o 0 0)


\\1\-\\\


-4--
4,


0*
0

0







-o

0s 0

0
0
-o
'? 0


0
0
0J


0

-- 3



o i
-o jE


0i


,-01 xOOtr OW OO0 O9Z OO OI 001 O0 0

iboinj 6uolo poppo sev w '


0
> 0



-~- 0.
-n0 0-O
U,
L.


0-"
: E


iu J
L-
:.3.


NACA


Fig.2


r----^d------fc-----li


\ \NN\\ \\\\N.\ .\N\\\\











FI.r C


C'Zk


0o


I


- \. o

0
0 0













V o





I : I, ... -4- -


i [ j'l 0I \1 o


T T ,
f


OO 09 00e 091 001 OS


a Pjosnj buo|D pappo ssoDN


(c0








+E E
U)+- c0


4- -
01 >

>
+- 0



.0) (.-
b- 4) L


I[ACA


42


c


c-
*0
a

2)

*0

0D


c

-7



0

C





0

o












0






a
4-














I1)
c
0












0
>t-c
0






C+->

CO
LV
-Q)
C- 0


(0

(-4-
0
0 '





- )



(0'


E


09s


t,-Ol X







UNIVERSITY OF FLORIDA

3 1262 08103 280 6






UNIVERSITY OF FLORIDA
DOCUMENTS DEPARTr'NT
120 MARGTOCN SC. ENrE LIBRARY
P.O. BOX 117011
GC-iNrJESVILLE, FL 32611 -7011 USA