Stress Concentrations in Bolt-thread Roots

Frozen-stress three-dimensional stress-analysis techniques

were used i n first phase of this investigation; t h e results
of these tests were augmented later by fatigue tests
a n d by metallographic inspection of cracked bolts

by James D. Chalupnik

~ s s r ~ ~ c r - ~ -allparent
The geometric origin of fatigue
fractures in oversize shank holts is considered in this
report. Bolt-nut interaction hetweeu the holt and both
a standard nut and a special nut are investigated. T h e
"frozen stress" method of three-dimensional stress
analysis was used to deter~ninethe magnitudes and loca-
tions of points of stress conretltrat.ion in plastic models
of both the oversize shank holt and a standard holt in
cornhinations with standard and special nuts,
T h e results indicate that. for i n s t a l l a t i o ~ ~ins which
the special nut is used, bolt geometry may he the cnll-
trolling factor in determining the f a ~ i g u elife of' the bolt
nut cornbillation, hrcausr the stress-cil>nce~rtratio~r factors
due to the nut loading are of the sameorder of magnitudeas
thestress-co~rcrntratio~if';~cLorsdue solely to holtgeo~netry.
On the other hand. fur install;rtionsi~lwhich standard nuts
are used, the nut-induced stress-cntrrrr~tratio~ifactors
are n ~ u c hlarger than those due In holt geometry; so.
holt gronretry will have less effect on the fatigue per-
form;mcr of the cornhination. If the n u t is threaded
onto the bolt so far that the first loaded thread of the
nut is in the t h r e a d - r u n m ~area
~ of the holt, fatigue per-
formance of the installation will ! w greatly reduced due
to the coml)ination of s~rt,ss-cil~ticetitratioli factors fronr
thread loadin# iirrcl section ch;tn#r and. again, 111)lt gewn-
etry may he a prominent causative factor.

Introduction
 I+
for 21 stLrn(lr~rd
bolt under thc same contiitions in tions. Not long there fter, HetBnyil pul)lishecl a
this rrrvJri. paper which contains i this author's opinion t h e
most information a v s lablr. in one sourt.e ahout
Previous Work stress concrntrations i nuts ancl holts. HeLenyi
invesLigntecl the t.ff(vt of six different n u t designs
&1~1cli h,rs heen ~vrittennlmut scrc:w fastmers and
on Lhe stress t l i s t r i l ) ~ion in bolts. This paper
~ l i e i rspplicutions I ~ u teven
, so, t h e volunie of these
points o u t the impor ance of investigaLing t h e
reports is stn;ill when one considers t h e massive
interaction of the n u t a d holt in tlireatlerl fasteners
~1st.of these devices in industry. Published re-
and has particular sign'ficance in the investigation
ports of actual stress analysis of screw fasteners
are few. no doubt due to t h e complicated geometry
of the screw fastener. Because of t h e geometrical
outlined in this report.
1
T h e work of HetBnyi,and Horger and Buckwalter
difficulties. many of t h e significant analyses of
stresses in screw fasteners have been experimental
in narurr and, generally, they utilize photoelastic
trchniqurs.
b
serves t o demonstrate t h e usefulness of three-di-
mensional frozen-stress techniques in the investiga-
tion of stress distrihutio s a n d stress concenLrations
in nut-bolt conil,inationb.
.An rurly attelnpt a t a purely analytical solution
~f stress distril~utionsin holts and n u t s was given
Frozen Stress
I
A hrief description cjf t h e frozen-stress method
I,y Strumeyer: ancl another was undertaken by
Den Harrog.' These analyses are interesting and
provitle 3x1 insight into t h e loading of threarled parts
nomcAnon.
in t l ~ a rtliry :lccount in a qualitative way for t h e
increasc in tlirei~tiIoi~dingfrom the top to the hot-
tom of' the nut. Both I)al)ers lack something in
~ o n l p l e t ~ n e s1)t~c;iusc
s. t l ~ ec ~ l c ~ t l a t i o arc
n s Imsed on
:I thre;+tl elasticit>- which is c:rlcul:ttetl for a two-
i i i n I ~ I nntl I must necessmily he in
t.rn,r Ir is true. i~~,n-cver. t l ~ a tsince the problem
is essvntially el;~stic.: u,waver. at elrv;~terl teni-
is unr in rlaoticir>-. t h r sh:il)tb of thc lo;rtl-distril)u-
~ ) e r ; ~ t u w sits
, viscr~sity is gre;~tlyretlit(wI and i t
tion ci1rl.r should I I Y the. s;Ilne. ; ~ n donly the scale of
magnituiir \\.oul(l Iv clinngetl I)y (:onsiritiring t h e
rnorc corrwt tl1rt.i-i!itn<,nsioni~ltlirc;~tl.
Ty!ric.ri cnrly rsiwrimc.nt;rl investigations used
two-rli~nrr.iionaiphotoel;~sticmodels; descriptions
of thesr rsperimrntnl investigations are given in
the r of Solakian,? Hall,' anrl Heywood."
Gootlier 'nnalyzrtl thread loarls using tlie unique
rr])erinirilt;~l~ e c l i n i ( ~ uofe measuring the external
(lisl>l:r\.r~lirnrs of n cylindrical n u t and relating these
tlisp1:rccnlcnts to the thread loatls and stresses.
Tile rr5ults of all these investignlivris show es-
sentially the satne thing, t h a t , in standard nuts, t h e
first thrr:ltls carry the IargesL portion of the load
ancl t h e r e l y impose large stress concentrations i n
the I~~oltin this region. T h e two-dimensional
photncl;~sti\~ studies indi~,ated t h a t the area of
Iiigheat stress concentrntion was located in the
root arc:r of tlie bolt threads, but there was dis-
itgrrenwnt ;IS to the magnitude and even some un-
threatls.
certainty ns to rsnctly which bolt t h e a d carries
t l ~ eIwa\ irsr loi~d.
Failure of two-dimensional photoelastic methods Scope of l n ~ e s t i ~ a t i o l
to yield itsefitl data can be attributed to two causes:
1 the rlifficulty involved in accurately machining
the two-cli~nensiondscrew threads, and (2) t h e
h s i c inability of the two-dimensional model t o
~luplicatethe three-dimensional stress distribution
existing in screw fasteners.
One of the firsr reports of three-cliniensional
stress an;\lysis of tlirenclecl fasteners in given by
Horger and Bucl;walter7 in a p a r w which also gives
results of fatigue tests on several bolt--nut comhina-

Experimental Mechanics 1 399

 again viewed in thc. 1rolariscope. T h e sections were
photographed m i l impected optically to tletermine
the rnagnitutles a n d locations of points of stress
concmtration. T h e results of these ohsen ations
are presented in subsequent section of this report.

Experimental Progkam
Prepuration of Specimens
Model nuts and bolts were prepared fvtin pre-
cast Hysol CP5-4290 Oirefringent plastk. Care
was exercised in machining the models from the
plastic stock to insure conformance with thread-
profile specificdtions and surface finish, and ro insure
t h a t the models were free of locked-in stresses some-
times caused hy heat generated during the ma-
chining operations. T h e machinist who per-
formed the machinihg operations on the niodels
is experienced in machining Hysol plastic m d is
aware of the special precautions which must be
taken in machining it. T h e quality of his work
on the nut and holt niurlels was generally high, a s
can he seen in t h r photogr,\phs in this report

Loading and Annealling

T h e models nere assembled with one standard
antl one special nut on each bolt, s~~spen(irrI inside
the annealing oven, and loaded with a de.3~1\\eight
of 20 lb. Srlf-alignlng harnesses were usrd for
hanging the I I I I ) L I C; ~~~n for
d applying the load to the
models. Wwshrm m x i e of blotting p'3prr were
usetl het\veen the harnesses and the nuts ro provide
for lliore uniform loarl transfer t o the nut models.
T h e loaded models are shown in the anneAing oven
in Fig. 2.
T h e room-temperature modulus of elasri~~iry for
the Hysol is 480,000 psi but, a t 270' F , the modulus
Frg. 2-Loaded models In oven
-
is only 2,190 mi. Tensile strength a t 2 i 0 F is
210 psi and Poisson's ratio is 0.5. T h e fringr \ d u e ,
which is a menliuru of thc. photoelastic sensirivity
of the materinl. is 1.35 psi fr,in. Fringe orders
for ,-in. sections of the shank portion of rlw model
bolts were calct~l:ated t o be 1.25 fringr for the
oversize and 1.51 fringe for the standard-shank bolt
for the 20-111 load. T h i s loading is equivnlenr to a
tensile stress in the prototype bolt of about 110.000
psi based on the pitch diameter of the threads or
115,000 psi in the standard shank and 9S 000 psi
in the oversize shimk.

Polurisope
All fringe photographs and all fringe-order counts
were taken a n a diffuped-light polariscopt.. Light
source is a bank of 17 15-w fluorescent tubes be-
hind a frosted Plesigllas diffuser. T h e polarizer,
analyzer and the two quarter-waves plates .Ire 10
in. in (lianwter nntl are inclcpenclently rotatable. tioning. T h e method of determining fringe order
Green fluorescent tuk)cs were usetl for illumina- in the unsectioned holt is described below.
tion for the black-and-while photographs. T h e Stress tlistril~uti~)n
in a cross section of the shank
grwn tuhes d o not provirle monot 11rom;ilic light, portion of a holt is essentially uniforrn; conse-
I ~ u their
t output covers n rrlntively narrow spectral quently, b y plotting the fringes in the photographs
Ijantl antl they are a satisf.~ctorylight source, pro- of the unsectionetl shank on 3 cross-section view of
vided the proper film is ustd. Korlax Co~nniercial the bolt, i t was possihle by interpolation to deter-
Protess Ortho film was usetl for the bldck-and- mine the fringe order in a slice c u t from this portion
\vhite photographs and the film was developed and of the holt. Calculations indicated t h a t the fringe
printed for contrast. Commercial Process Ortho level in the shank of the oversize holt would be
film is highly sensitive t o green light, so, the com- about 1.1 fringe'and t h a t in the standard bolt shank
hination provided good fringe photographs. Cool- would be about 1.4 fringe for a n 8-in.-thick section.
u h i t e fluorescent tuhes were used for niahing t h e Largest bending moment was found in the aver-
color photographs and for illun~inatingthe models size holt; however, the maximum bending stress
for the telescopr eunmination. Koclak Ektacolor was less t h a n 5 percent of the tensile stress. T h i s
Professional T y l ~ e"S" him was used for the color was cmisirlered to he satisfactory; so, the speci-
photographs mens were prepared for sectioning.

Preliminary Inspection of Fringes Sectioning of thp Models

Immediately following the annealing process, I t is considererl rlrsirable to section the Imlt a n d
tlie motlels were rr~novt?tlfrom the oven and placed nuts a s a unit, 1,ut it \\as not considered wise t o
in a small, squnre Plexiglas tank filled with mineral use normal shop methods for holding the models
oil and inspected ol)tit.nlly antl photographed in the (luring the sectioning ol)eriitions, because of the
~~olariscopr.Mineral oil has essc~ntiallythe same prohahility of loching in undesirahle stress patterns.
refractive index as Hysol and, consequently. per- A highly satisf.~ctor!, solution to this prol)letn was
mits the ohscrvation of fringes in tlir cylintlrically found, however Small woo~lenOowh n r r e fahri-
shaped bolt shank \vithout distortion. In this cater1 into which the bolt .issemi~liesw e n - ~~I.ut.tl,
case, the fringes could lw ol)servrd to \vithin ' :* in. coRin fashion T h e ~ p , \ v caround the bolt and nuts
or less from the edge of tlie bolt. was then filled n ~ t ha rigid, epoxy foam that was
These initial ol)servntions wcrr m:ttle t o cle- Fomrctl in place. T h e foani set u p and 1)onded t o
termine the fringe order of the st.rrss pattern frozen the test pieces in a matter of minutes and the
in the models and to determine the amount of wooden 1)oxrs pro\ ided solid m a t e r i d for clamping
Iwnrling introduced I)y the loading process prior t o in s l ~ o pmachines Both slices of the photoelastic
sectioning the models. Had a n untlesirahle loading plastic l)olts in their epoxy foam jackets are shown
condition been noted, the model could have been in Fig. 3.
reloaded and anneded iseveral times if necessary) T h e slices were rough c u t with a band sax\. then
until the desired loading conditions were obtained the faces were machined snrooth, flat and parallel
hefore the expensive models were destroyed by sec- with a fly cutter in a milling machine. T h e
thickness was held to 0.125 = 0.001 in. o \ e r the
section; so, no corrections for thickness \ ariations
were required.
A dull band-sat\ bl'tde was used acciclentally on
one of the roughing cuts, antl as a result, a number
of bolt threads were chipped o n the oversize bolt
section. The damage was more esthetic t h a n
functional, since the pertinent d a t a sought in this
investigation was contained ebsentially in the root
section of the Iiolt threads, and this part of the
model was not damaged

Detailed Inspection of Fringes

T h e photoelastic pictures of the sectioned nut and
bolt combinations are shown i n Fig. 4. Areas of
stress concentration in the bolts are readily identi-
fied as the roots of the threads. I n these areas, the
tensile bending stresses due t o t h e beam-type load-
ing of t h e bolt thrends combine with the stresses
due t o the tensile load on the bolt at a point where
t h e latter stresses are concentrated due to the
notch effect of the thread.
Fig. 3-Sectioned models T h e fringes in the photographs can be counted

Experimental Mechanics 1 401

 a. Standard bolt--standard n u t b. Standard bolt special n u t
Fig. 4-Fringe photograhs
of model slices

c Oversize bolt standard n u t d . Over>izr bolt specla! n u t

either from the end of 111~. 11~1lts. or f r l ~ ma n ;irljn- e;l(.h setting of tlie telescope. Errors in fringe
cent unloarlerl t h r r x l \vii~w,t l ~ fringe
, n ~ u n h e ris count c;lnsrd by aligning t h e tc.les~.o!x~\\-it11 the
zero. I n the area of ~ n : ~ s i i i ~ isit~ri ~i wconctsntril- Iwlis angle art> negligil~lc,since the ht~iix:~ngle$ is
tion, t h e fringes c o i w qiiit<~clc~st. together, hmv- ;ilioul 2.5 tleg am1 the errors go ;Is 1 - cws i l l , or
ever, and it \VHS f n u n ~ lto IN, iii111ossiI11~' to o l ~ t a i n allnLit 0.001.
tlie masiln~um fringe ~ ~ ~ fro111 u n l the I ~ l x k - i l n t l - Most of the fringe-order ~ ~ ~ i i xofi ~inrerest n ; ~ liere
white photographs. AIII<.II 11c1tc.rresults coultl I,(. occ~lron traction-frw surfilctts of lht. 11olt; con-
obtained from full-~.olor~!liotograpl~s takvn of 111~. s e q ~ ~ e n tthe
l ~ ~frirry?
. ortlrrs ilre ~ ~ r o l ~ ~ r t i oton :the
il
sectioned n~odels,Iiut evvn t l i ~ wresults wcre not ~ u x s i m u ~strcwras
n at L I I O S ~ points. Fringe order
conil~letelysatisfactory. TIN.Iiest results were 011- is propwtionnl to thc difference in the principal
tained by counting the fri11gc.s directly by nleans of stresses :md, on a tract.ion-free surface. the princi-
a short-focus telescope \\-it11 tlw models in thr. pal stresses art, normal and tangent to the surface,
polariscope. Using this n~t.tliotl. fringes could I x t h c normal vnml)rlnent being zero. Hence, for
counted quite accur~ttt%ly in !lie w r y root of' the locations of' intrresl, points of rn;lxirllr~n~frinyc
threads. I t was foilncl 1hit1 i( was necessxy to ortler are points of m;lsimum strew : ~ n dt l w fringe
line up t h e optical :>xis of tlw telescope with tllc order is proportion;rl to the stxrss :lt there points.
helix angle of t h e tlire:1<1root in order to rwtl thl. One point of n r ; ~ x i ~ ~ ~ slress
u l n nor f:~llingin the
highest fringe orders. 'I'l~is align~nent\vns c r i ~ i c a l , ;\l~o\.ecl;~ssiiic.:$tion that is used in later cirlcul:%tions
so a nlnsiniunl of t\vo thrtwtls c o i ~ l dbe read \vith in this p i l ~ e ris Ioc;ttr~l:tt Lhe cent~hrof' thc shank.
 I STANDARD
xOVERSIZE

SPECIAL NUT

Fig. %Fringe-order maxima for standard and oversize

bolts. (Standard nuts are at bottom)

Fortunately. this is a pain1 of uniaxial stress and

stress is l)roportional to fringe order here also. STANDARD NUT .
Test Results Fig. 6-Maximum fringe cobnts at bolt thread roots

Maximuni fringe orders obtninetl from the

telescopic investigation are sho.r\.n in Fig. 5. I n nhicli mate with the bpecial nuts are not loaded.
each case, tlie n ~ n s i m awere found in the thread- This has wliich will be
root area. Results from Fig. 5 are s l w v n graphi- discussed in a later
cally in Fig. 6 for the section of the bolt that mates
with the nut. Results from both standard and Analysis I
oversize bolts are plotted on the same graph. I t I

i
\rill he noted that the results lie very close t o each T h e inauiniun~frin e orders recorded for bolt
t~rlirrantl the cur\.es are drawn in t o "guide the threads loaded by the pecial nuts are all quite low.
c>-e through tlir points." I t can I)e seen iin- They are only slightly larger t h a n the fringe orders
nlecliately Lhat the special n u t offers a reduction in recorded for the threa -runout area. For e w n ~ p l e ,
n l n x i n ~ ~ ~stress
rn compared t o tlie stantlard nut. tlie four maxima in mlt threads lo,ided hy t h e
T h e reduction ainounts to approxi~iu~tely one- special nuts are 3.8, 3.8, 4.2 and 4.5 \.ihicli can be
third of the m a s i n ~ u mstress encuiinterrd with t h e conlpared to fringe ord r s of 3.2, 3.7. 3.8 and 2.4 i n
stantlard nut. This is acc:omplishcrl by provid- the runout areas of 111 overs&-shank Iwlt. T h i s

I
ing relirf in tlw root of the n u t thread which im- means t h a t t h e stress concentrations clue t o the
~)arLsa certain amount of flcxihility to the n u t reduction in area com ined with those due t o t h e
thread antl permits some of the upper threads t o presence of a notch t h e thread in the runout
s1l;lrc the load-carrying funclion. area give rise t o stress s of t h e same magnitude a s
I t would he wrong, on the basis of these results, those due t o the inte action of tlie I~olrand the
to say t h a t the special thread permits more threads
to carry tlie load, since i t will l ~ enoted t h a t t h e
first two "eng:~ged" thretids in the bolt ends
special nut.
1
T h e low fringe orde of 2.4 in one of the thread
roots is something of asurprise. It is not known if

Experimental .\I~char~ics 1 403

 i t is tlue to soni~: ~)ccuIii~rity of tlie S I I I V : ~ : ~n u t or signif~c.~r~<t. in this in\t~stig.ition. Sornt. concern
due to : ~ s y r n ~ ~ ~ eIo;~tling
t r i ( : of the spoci:ll nut lead- h ~ t l11ec.n v f ~ i w t lprior to the tc,sts t h ~ tthe nuts
ing to the low fringe o r d w on one sitlc o f tllc bolt. (onltl r.~silyI)t. 11ircwlt.tl too f.ir on Ihr Iwlt in nor-
A Iwntling ~nonicntis notcrl in tlw sc<.tioncontain- m , ~ lal~lllications Tlus contlitivn h , ~ s long been
ing the. lo^ reading ant1 tlic low re:~tlingis on the rc.cogni/vtl as tlrtrimt.nl.rl to oi~timurnservice life
sa~nt.side as the cumprCssi\-tsIwntling stress, so tlic in I ~ o l t t djoints sul)jw t to frlllgue Iv.iding. I)ec:~use
ol~serverlrlifferen(.r is consistt,nt nit11 the hending it places tlic first tl~rt.adof the. nut close to the
nioment in ~ h srh m k of tlir Irolt. hut it is tloul~tful t h r e d r u n o ~ ,irtvi,
~ t where tlie stress c-u~~ctmtrations
tlint the large rliKrrvnce in fringe ortlcr on the two tlue to nut loading are c oml)inerl w it11 concentra-
sides of the rrction can Iw attributed to the small tions due to notch effect. I n o v e r s i ~ el)olts, these
:tr!luunt of Iwnrling ol~servedin the sliimli of the also c o ~ n l ~ i nwith e ctrmentrations tlucz to section
I1olt. climge from the oversilr slinnk to thread-root
M;taimuni fringe ortlers in the threads loaded b y tliamrter. This means tli,it, in addition to the re-
the stantlard nuts are sul~stiintially higher t h a n duction in induc erl stress c v m ~ m t r a t i o noffered b y
those in any otlier part of the Imlts, and it is ap- the more flexilde tlix,arl~in the special n u t , this
p x e n t that thr first thrtrarl engaged I)y these nuts nut effe~.tivelyI ~ ~ a tthe l s Iwlt .i\%ayfrom ntlier arras
c:~rriestlie grr~ltestl o ~ l . Additionally. the maxi- of high strt.. '5s ( o n <rntr.tt on.
111um fringe orders rlue to the stanrl;lrtl nuts are T h e results indicate tlwt, for inst.dl,~tionsin
si~l~stantially higher tliim tllose dur t o tile special which t l ~ especi,d nut is i~sed,bolt g r o n r e t r will be
nut. a controlling f,u tor in tlt,t~.rrniningthc f,ltigue life
When the tllrrarl-root fringe, ortkr is 1)lotted of Llw Iwlt-nut c o m t h l t i o n 11t~:iusrl i l t , .;tress-
against the tlirexl position. as \v;tr clone in Fig. 6. ~omtwtr,itionfactors rluc. to thc nut art. of the same
an idc,a of the Io:trl tlistril)~~tion in the I~oltcan be r~i,~gnitude as tht. stress-( r~nwntralionf x t o r s due
olltainrd. It is ; I ~ I ~ ; I ~ P that
I I ~ the special I ~ L I ~ solelq to bolt gronlc.Lr\ On the other hand, for
distril~utrstlw !oxl in the Imlt " I w t r 4 ' tli:iii the instnll,ttiul~zIn nlliOi st.irirlnrtl nuts .wr uwtl, the
s t : n I r n i t F ( J ~tlw stt~n(l:irtl nut, thc first nut-inrl~~c rrl stw~.-c.orrc~ . n l r . ~ t mfnx t o r i ,Ire ~ n u c l i
Io:~rlr,rl thrc;~tlr.,~rriesmore Ioml than tlic set-ontl: lnrger t l i , ~ nthnv r l ~ w11) Iwlt grmnetr? ,in(l sv h l t
!vhrrtxs. tlw first t\vo lo:itlrtl tlirt.;rtls in the other geo~iirtrk shoi~ltl hnb e I ~ S S effect on t I N f . ~ igue t
<w~nl~in;tlion c x r y apl~roxim:~lely the s:lme load. p e r f o r ~ n , mt t of t h e c o n i l ~ ~ n . ~ t i o n
Thv gentk cur\-r\rure and t h e initial slopr of tlie
curve for tlir special nut n ~ o r (closrly
. rc-presents a n Conclusions
"itle;~l" loading cvrvcAthan tlr~esthe curve for the
stantlard nvr. 1. O ~ e r s i ~ c 4 . i nI~olts
L ;ire not rt.i r ~ ~ ~ ~ ! n r ~ r t l e t l
for fatigue al,plic,~tiuns. This is true of lwth ten-
Stress-cor~centratrmFactors sion and shear al~plications,since it apl)t3,irsthat,
I n calculating the stress con(.entrations, i t is in tlie latter, uneaplainecl mecl~anisms , ~ work
t
i m l ~ o r t a n tto specify the stress to which the local which cause the bolt t o trr p a r t i ~ l l ylontirtl in Len-
high stresses will he cornl)arctl. For instance, tlie sion.
stress can he cunipnretl to the nwnn shank stress, 2. Based on the results reported herein, i t
thcxstress 11:isrri on the pitch-diameter area, or even would appear that tlie sl)rri:~lnut offers clefinite im-
the stress m ; ~ r i ~ n ~inu nthe tlirend-runout region. provements in fatigue 1)erfvrm:mce of holted joints.
Each of these values offrrs useful information. T h e
particular value that suits a cert:iin application is a
matter to be deter~ninedby the user of the fastener. T h e author acknowledges the in.valu;hle assis-
Average stress-concentration factors based on tance rendered him in this project by R. L. K u r t a k
these three reference stresses are given for t h e four and R. E. Finigan and other members of the staff
tested combinations i n T a l ~ l e1. of the Technical Supliort Group of the Structural
T h e first two engaged threads in the special n u t s Research Laboratory, Commercial Airplane Divi-
were not loaded. This was a n unexpected phenom- sion, T h e Roeing Company.
enon, b t ~ tone tlint is of particularly noteworthy
References
TABLE 1-STRESSCONCENTRATION FACTORS 1. Slrutnu.vur, C . E . , "Stress lDistribulion in Roll$ and Nut.-." Trans.
FOR THE FOUR TEST COMBINATIONS AND T H R E E Inst. Naval Archit?cta, LX. 112-122, P l n c XI11 nnd X I V ,,1918).
REFERENCE STRESSES 2. Den Halortop, J. P., "The Mechanics of P l a b Rolors fur Turbo Gen-
erotors." Tmna. A S M E , APM-31-1, 51. 1-11 (1929).
Reference Stress - -. 3 . Solohinn, A . G., " 1 l . h ~ Threodrd Parts Foil," Pmdvct Ens.,
261-262 (July 19331.
Runout- 4 . Holl. S.G., '*Dcterrniwl?onof st re.^ Canrrntmtion in Screw Thrpnds
Average Pltch thread by Photwlnstie fithodr." Unicei.sity of Illinois Bulletin 24.5 (2932).
Nut Bolt shank diameter root 5. Hqvumd, X . B.. "Tensile Fillet S r w s s r a in Loaded Projrdio~r,'.
Pmc. Inrt. Merh. Gngrs. 159. 3831-391 (1948).
Special Oversize 3.2 2.1 1.2 6 . Gmdier. J. N.. "The ~ i s t r i t u t i o nof I ~ a on
d the Tkr<,ndrof Scrrru,~;~
Special Standard 2.7 2.9 1.4 J n l . Appl. Me&, A-20-16 (Morch 19-10).
Standard Oversize 4.0 2.1 1.3 7. Horger, 0. J . , and fiwkra,&r. T . V . . "Photoelasricit~or Applied tn
Design Problems." The Iron Age, 42-;TO (Mo-v 23. 1940).
Standard Standard 3.7 3.8 1.8 8 . Hrenyi, M.. " A Photoelnrtir Study of Roll and Nu1 Fnatmings."
Jnf. o f Appl. Mrch., A-9S-100 (.tun? 19510.

404 1 September 1968