I Subjeot I Report on Atomic Bomb Damage Evaluation Memo to Lt Col J A Derry 1• Copy Number l of the subje ot report is transmitted for delivery to Maj Gen L R Groves This report prepared by Dr TT G Penney is entitled A Report On the Pressure Wave Caused By the Atomio Bomb Explosion in Hiroshima and Nagasaki Included with the report are two sets of photographs showing tm damage caused in each oity as well as certain specific objects and structures whioh were used as a basis for calculating the hydrostatic pressure and wind velocity resulting from the explosion A key explaining the photographs is also included as well as maps of eaoh city -•as I 2 Copy number 2 of this report with one set of photographs has been transmitted to Sir John Anderson in London and Copy number 3 is in the possession of Sir James Chadwick in Washingtcn The final draft of this report was oanpleted on the night before Dr Penney I s s oheduled departure fran London for Washingix n Copy number 1 was proofread subsequent to release of oopies number 2 and 3 A nwnber of additions were made to Copy number l at that time so that it is the only canplete oopy in existence 3 The British government would like to release this report to certain of its high policy officials but have agreed at Dr Penney's request to limit availability of the report to Sir John Anderson Sir James Chadwick and Dr Penney until suoh time as General Groves grants approval to further distribution It is Dr Penney's opinion that access to the canplete report should be limited to the British Chiefs of Staff The report is arranged in such a fonn that further distribution of limited portions of the report could be made Dr Penney has requested for Sir Jo m Anderson that General Groves grant formal approval for further distribution of the report and in form Dr Penney 01' the limitations which he wishes placed on such distribution Dr Penney has also requested that a oopy of the Master report in General Groves possession be furnished to the A group at the time approval for further distribution is granted 4 Nl D G Stur1es Capt C E T DECLASSIFIED Authority IND 7 7 'fV6 59 I DH LA' ' lflf D utbontY fJtiP J'fi6$'1 SECRET General atatetlent t J iong the obaorvers of and Hagaaaki durin bl at efteota ' 'Ile Dr R Serber ra tner thot feature waa notioed Uain resul ta · The followin n are considered to • ' r 1 Tke 2 The - - • ' 0 • 3 'nae hei t• deatruotion l' •'fbe aotual to11n ge ot ' • · ' t• • • i _ _Pr· 1 ·r - · - t T n T ' I · which would have oa u sed the aame bla at dli e waa 500Q in Hiroahima o nd '20 000 in Nagasaki - - · - flt• ' J - • exaotl - that predicted • - 0 illt'0ma n oolleoted wo tld e ble r onabl accurate_ pr edi ctiona to be_Dad e ot the bl at damage likel to be u sed in any oi ty where T • • I' tvmic boab e loaioa wu oonoeivo ble• SECRET DECLASSIFIED _utbont NND 1 Cfo-6 Ser General com o rison of J iroshiL- a and Even ·e n intellii ent observ_e r ltf i saki and Hiroshioa r hic '-l of roints of much core rein1'orced oonc rete a c esert of c l ezlJ swept c 8 rred remaina there were r o l trona Arlns Works and the Torpedo was sometluna lef't to see and the One could stand iD the cl ty I but thi 'was not Hei ht of Burst Size ot Ball Japanese army observers bservers 'H l tohing ohing of burst ih ·Hiroshi m I as An independent ometry ot some H ·94 H 95 show the bu s v An unburt1 t wa clearly visible of the horizontal the height ot burst were ple ced From the ge metry of height of burst was 1800 feet The shadows had tinot penumbra tni measured t irly well ot fire when it was 300 teet than the · maximum diameter is cn1y ot the order one millisecond -1 ___ _g -' · __' y - C • -4· ' · Telegrap i poles very near the point on the gro unier t nt · _ot the explosion were often lett • ta nding r _- ' ' • • l l ash Bums The side e aed to ' ' •_ i _ - severely charred w - gt it ''- •_ • · h fi sh · · · By marking on a map the direo o or_ i and -fdo ni thi• tor about five or six oa• es the •ntre §'fl t' cj d - t 0 r pli oec within about 100 feet ·- Confirmation of the ·centre - the direction in which At Jliro• hia t'la• h burns on to a rad 1 ua- - - DECLASSlFIED DECLA SIFIED Authortl ' N tJD 7 i -r 5 l - cl to the W l S fi 19- C since it as th elves th i n of the bomb • -r 1 ·_ ' • '• 'l ro rty more of t l e • I No 19 rge fires were ' started inshntly by the borr b grass mats which and no dvubt were - J riing out iide SOl A Straw ind IUUlj t o a ir r ere it niteu i im ediate 1 fires viere shrted in this wey A more likely oause were the ch l rco l l braziers one or which vi9 s to be found i n nettly every hvu ie - Sorre trees were ignited by the heat of the explosion 11 0 Pioture H 93 shows ol_e ar_case of the nesh ignition of a tree J any peculie r effects due to fire winds were bserved by the J l panese but were not recorded very acc 1 rately C s t e r i i ps rticul ir · ·-il people into one of the rivers at Hiroshilll l Wind Telocities up to 50 f sec a low figure· probably often exceec ed were record coi ip I E d with 6 f s before the explosion Heavy fire d lmage extended about 8000 ft S of X in Nagasaki to ·about 6000 ft c ircularly about X i n Hiroshim l Several isolated fires were found in both cities up to 9COO ft where the cause W l s clearly a bre zier or something similar i nside the building l hill protected the building from fl l sh burn Omet i mes l Solll tiraes large building c - a l nd t h e could not h l ve been flash ignition of s ome tinderlike material r The most rem l rk ble e C 1 0 from fi re d lll' £e 'fas the Torpedo 'l'IOrks at Na 11saki hctory The whole f nmework -ns 'l'IOod but t re W s no fi re i n the lll lin Sevenl s I 9 ll sheds the aint shed further evide nce of da nsge The most impressive long range thn ge Y - as· the collapse displaced Possibly t_hese were the results of the typhoon and not tho at omic bomb The Maoh Y Only by l coincidence would it have been possible to detect Z Y evidence ot the variation ot pressure up the stem ot the Y There was a very large chimney stack in the Mitsubishi rms Y or 3 500 ft S of X The cbil IMy was badly cr9 Cked but more interesting to us N s l steel l dder running frol top to bottom The stack was at le l st 120 ft l igh The ladder was bom d between the supports which were at about 8 rt centres wJl height in tho bowing oould be detected h l d bent as f ar as it could without stretching No v I I iation The reason s th 9 t the ladder ' be soock l t the rivets the air flow ms very l l rgely controlled by the chimney and there seem d ·no aecurat way of llowinz for this • A t lloii a distar ce ot seven miles over steep hills at l ea st 600fedt about 10' of t he il ass cai ie out b nec rer seques tered l ocalities only lPL hiih 4 ci e3 _ '- An i n tiresti _li effoc t described· as follows 1 inda· v d oom lete to •12 000 ft soc e d ana e P toljo ooo ft or _ even more • 2 Pbster d ge -' ceiling o r a · a lb ver- - r eo vy 'l t 9 000 ft lllt'de te t 12 000 ft lllC et 15 000 ft - 3 R ot d ge to sl te or tile he 1vy o t 10 JOO tt light o t 1 0C ft '3a SJ eet meta l roofs le errin3 o c'ILtli' y e x ori e r to t i on - -J '-' Dx ir u ly the e as for s19 t - tilt He avy 9 t 6 000 tt light at 12 0IJC ft L ll'indow frames o nd doors 5 Nine i nch brick 1 9 11 surbllrb UJ r ou f• ll s c nclce i l euil · r t 5 000 tt some cracks at G ooc ft re cracks st 7 COO f t Untouched t 8 000 tt 6 Re i nf orced c oncrete b•1Udin • 10 'lftJ la 6 floors u- d l root but • till st'U'ldina up to 2 l'lCO Minor struot ur-1 dama ·• at t lrO tt 4 0CO ft 7 High 1 ielit y • t- • l traM t-uiU · No daaage to fru but aneb bl own in up to 1500 tt H iroshirna in remarkable C ntraat with the scarcity in Nagdald • In the latter city tor example we were unable to find a single flagpole or lightning •- conductor just bent over by the wind of the exploaion in Hiroshima had the choice of about twenty reverse order If our visits to the two' cities had been in the might have miased a few of the observations actually made 11e ·in Hiroshima but thi• loss we could well afford on the other hand we might have gained two or three more reliable figurea tar Nagasald and thereby improve our _estimates of the tonnage in thia city retations of the Observations We explain the principles underlying the interpretations of obaervationa which led to estimates of the peak pressure The i c a l rHults are given later in this seation Crushed Ketal Cans e of the sjmpleat methods of estimating the peak prHsure was f'rom the arusqing of oil drU1Ds gaaoline oana er any· other e114 ty thin metal vessel -with a small opening The assU1Dption hare iqde ia that the bldt wave pressure came on instantaneoualy the rHulting preHure on the can was more than tiw caae could withstand and t walls oollapsed inward•• 1 The air inside was oompre• -ed adiabatically to auch a point that the pressure inaide wu lHs by a certain amount that the prHsure outside thi• emount being the pr•••ur• dift'erenoe out • ide and in that the wall• could • tand in th tr crl Dpled oonclition The ·unoer ainti•• involved are DECLASSlFJEO Authonty NJ JD t 10o5't_ SECRET· Let p • i t oi - V-0 ·-- -----· - W• ·•' meohanioal atrength against crushing -' t J •ur• the ale- F P 1• taken IVO of vol'I ml• 1• 100 1 - V and u H ' 1 thia report we have plottod in Figs and and 2 tho rise in fLir 0 • - -' 1 pE e sura insida a can namely Po Vc V0 4 - 1 in pound • to the square inoh aga inat tho porcantage loas 1n voluno ratio Vc V0 Also given is the compresaion For example trom Fig I i t is soen that the riao in air • preasure· corresponding with a 20 per oent loss of volune is 5 -3 psi It 1a worth p inting out that a drun ar tin can can withstand for a • hart time oonaidorably moro than it• critioal statio crushing pressure Tlio· ·•idas give way bacau• o of in• tability and the ill8tability requires time to e ow Conaaquently one might sanetimas find cans intact at pla ces where the peak pr• •ur• was highc- than the static strength of the can If however the can did give w93 t h e amount of distortion woul d ooirwl l ly b8 conaiderable cauae the strength crop• onoe dents h ave appeared e oan easily ••e by a •int l• ca1culation that the time required f or a •- l l dent t o e ow to a large ono ia vary amal l For ax le if' a dent has bee n astablisb d and tbore i• an unba1anoed pE e• sura on the t wo aid•• of t he dent of 1 psi t he aooalara tion is vary larg If the t hiokJwH of • • ta1 1a 0 020-inch t hin the pre• sure 1 p • i move• the meta1 2 inohH tr011 re• t in five llillieeoonda can 1a tha cruahed meta1 tuba Kany examples o f thi a w re fowid but not a • ingle one of - ioh quantitative aignitioa noe · Th• bea t perha p• • waa a reatangular oopper drain pipe near the Shin Aioi Baahi i n Hiroah lma However even for thia· t he •ample baok tor tHt PipH autfer rx-o two diaad te g•• equally Alona ita lenath unleaa one _w hat the 'PE'H• ure inaide wu • orten' the pipe w a not - - ·f aey _ 1-Hihlo ' tii' • nd aa-antaa• ia that the pi 2 ·'· - · n ··_ SECJRET '1 '• - - DECLASSlFl E D ' Authonty NND qoo5j DECLASSIFIED Authonty tJtJD I qOo5j and then whipped about in the wind Viaual Estimates It proved possible to make one or two absolute estir ates of the peak pressure as 11ell as aevere l co arisona between Nagase lei and Hiroshima •iJlt ly by inspection of the nature of the dEllllage Canpari• on wu also made with the damage at the M8 cdo d Raneh in Nn Kexico caused by the 'lrinity lcploaion ' l mbranea Plates and Panel• Pa er Panals Knowing that thl Japanoae 1'0rur• home• uae paper panel• as partitiona much was hoped far in the way of getting the critical distance at which these panels feilod However our luck in this respect was poor At Nagasaki we thought we ha d found a beautiful ex1111ple at the go be rracks 26 CCX feet to S of X but on checking with the householder we discovered that the damage• was caused by a •small l oy9 The panel had been exposed side on to the bl t · The bottan panels were gone these of oourse were ths ones which the child oould 7' reach The upper panel• were intact 'nle panals could stand about 0 5 psi see later in the Tests Seotion and hence the peak pressure was not IOOro than this From the collapse of a wall supported by rather flimsy wooden uprights we estimated that the peak pressure wa s greater than 0 4- psi Hence we have a good 'traoketing of the peak pressure hm-e • One alcx st perfect eq le of the failure 12000 rt 3 fran X in Hiroahima within 1° Hs was indoor• pie The rHt were intfot and half n dozen pieces wm-e brought back for teat SECRET r W DECLASSlFIED I AUthonty J qoo5_j_ · ·· -· ·1_•· · · • -_ - ' L -- · • ·• ' ''V· - · - ·• _- · - • ''-•' ' '' ••• t ·• ·•• ' • - -l ' • · · • · ·· - 3100 feet fran· X 4 Wall panels of Tra nsforme - 30° Panel Panel 3 was valuable not value but ' -- it was very similar to Panels 5 The pressure on 3 'J V't _ i d be very _c losely deterudned from evi dence of other t ypes most precisely · '• from the bending of flagpoles and thus a simple and small correction gave a most valuable point for panel 5 in Nagasaki The detai ls of 1 - 5 are gi ven l ater in the Tests Section Wooden Floor · Another variation on the failure of a panel was foWld in the telephone exchange 800 feet 1' of X in Hiroshima floor cover d by mats failed noted The joints in one panel of a wooden The floor was t a ken up and its construction Quite a good value was obtained for the peak hydrostatic pressure in his building Barrack lfalls a nd Wall of a Timber Stare Barn The wooden supports of a practically a irt ight barn in Hiro hima end the wooden supports of the walls of two buildi ngs i n t he Kmsigo barraclcJ at · Nagasaki broke undar t he l ydrostat ic pressure of t he blast From the details of tho cons truction estimate o f the peak pre ssure were made Lr Problems An interes ting s ori os of observations wa s made on the bending of s t eel flagpoles and lightning conductors in tbs wind from t he expl osi on Similar problems were the bending of steel l adders t he snapping of telegraph or powsr lino pole• and the failure of smoke sta cks 'lb• drag on an object of area A in an air stream of denaity - · · - velocity Al is by detini tion FA wh • F and Co 1e the drag ooefficient ·- l J f Cor u2 L ' c _ - '-• -- · For a pipe of diameter D and length· L exposed II 5 SECRF T DECLASSIFIED 4uwont y tJtJD q0o5j perpendicularly to the wind '- ua W · I• • _drag ooetfioient is a funotion of the Reynold's number R for the particular value• ot D and u •- - R f· ooetficient prev µ ing - 1ft' Values of found that 0n have been msesured in various wind tunnels It has been Oz depends to aome nt on tht roughness• of he surface A p pe 1-inoh diemeter whose surface has been covered with oarborund Ull powder somet iJ les gives - uuoh os twice the drag as one whose surfaoe has been hi y polished Of course these are extreme limits and in the exSIJl le• with which we are oonoerned the drag ooefficient oe n usually bo estimated within 10 per cent The deduction ot the value of u and hence ot the peak shook wave pressure leads -to a value with an c-rar of not more than per cent due to the uncertainty in On• The bending moment at the base of a vm-tioal pole subject to tbs horiwntal air stream V is This is to be equated to the yielding bending manent of the pole If the pole is a tube of outer and inner radii a and b and T is the yield ng tension of the JD terial then the yielding bonding moment K is M 4T aJ - b h Alternatively K can be meas red by experiment Thus a piece of a steel tube flagpole fran Hiroahima was lrought baok and K measured directly agreement ri th formula S Th wu very olose as n ani ng a value fer T of 20 tona sq inoh In the cue or telegraph poles aamples of the wood from pole• at Naguaki SECRET ' I ·t - the moment of inertia about the base the equation of motfon If I h aoceleraf j ng ie The bending lllOme nt · d ue to tne weight of the pole has been negl_ected This is a satisfaotcry aprro dmation · · ainoe it is a small ooITeotion to·- oorrection - Since we are concerned only with the motion near the peak of the pressure Tra ve v e may Vlrite F t F0 -kt With this simplification the accelerating phase and the decelereting phase ·iu-e symmetrical and the displacemont at the end of the accelerating position Solving the equation for the acc elerating phase subject to the conditions t 0 Q 0 ff O and ffl'iting for the observed final displacement reached at time T we find T 2Fo k - WDL2k • Fo 2l VDL2 12r d DL2r2 F0 1 - - We write F I where bu 2Su i s the change of mus hlooity in the time t which has elapsed since the shock wave struck the pole NOl'I' insert the Riemann rela tionship far the motion beh ind a plil le shock 6 u S o - C there o is the velocity of sound at the pole end C ths DECLASSIFIED Authority N1J D_ 1 qCojj_ 0fCLASSlrlrO • uthont f JJ T t LClo' c C-ve rt -nin - of ·_· 0 - L LIU SECjET -ton ' cs 3tones or a ' most the s ame dimension ll s the ones ere o v erturni d f or1t u·ds · r r obe b y the stone nearl y toppl ed s ung bo c lc a nd the r e vene rind of the sue ti on p haAe t oppl e d it t he otl- er ··· Y• • Just as wi th telegraph polea chimney st· cks a nd· v t hcr objects ffectcd l y the wind there 11us a rei _ ion near X · abc ut 2 000 f t rl diua tihere the blas t wave wns no t a simp le ver tical one runnin h ori zontal ly One can do little rith tlteory in the i entral r e v on and · e t h erefore c onfine ourselves to observati ons further out 'r he theory of the ov ert urni n oi · o recta n ul a r stone i n a b l a st ave l eads to so ie d iff e ren J 11 e quations · hich could only Qe a cc· v·r tely s olved by nune rica1 r ethods Eo eve r the data do notj ust fy a n e lab o r ate investi at j on and we t io de s on e 1·easonable fPrOri 11R t ions Let 0 be the anisle bet e en tlte v rt i co l o nd the li ne joini n 6 th c entre o f gravity to the t urni n i e dge Let I be t he 1 01 1ent of i n ert i a abou t t his ed e t he mass of the tonf Z t he hei ga' c nd the lcn th of the sid e of t h • t se 1 ot the t urning e d_ e The 1 idth of tlle r t on e does no t e nter e cep t f i nall y in the Reyn ol ds nunber and e 01nit this ass 'lC 1in effectively uni t idth U1Jin g a bsolu t e uhi t s the e 1tiat h n of o t ion we app rox te by 0 where ' ia the dra 1 p ree aw -e pe r _ nit tu·cn s fu no t t on vf t ·i t 1n u we did bef ore in the e uat _o n of r o t i • n of the 1 l1 s t ic yield 1 f i r od F 'o - t' tho equati on or lllOtion tlilen integi·ates • - • _ z Z t - _ ·O The aolution ·f a ·V Tl-le X 1 o - r ' aimh11t i 1 - t- 'i' -_lltr_- i ' _ 9'· t •m DECLASSLFIED Authoritv J JT 7 qcoS_j_ Th value of' Fo obeys the relationship llgX z2 riting M f XZ and Allising the expression '41- f'or A we have f'in ally t l This 11a y be used as the basi3 of' an iterative process to detennine Fo The procedure we adopt is to 11se the knovm tonnage to select t he pro er v l uee of' P and f'or a shrine in Hiroshim 1 ooet't'icient t'rom the above equatioM Then vie m oal culate the drag Using thia ooet'f'ioient we then p roceed to the N ass ki results to ed -' the peak pressure •• • i fj series of observations in both Hiroshima and Nacasald l 'fhs loss of volume due to canpression was measured for the blue-print container• - · _ 1 ' rith much greater accur-acy than m possible in the other cases because this particular dr 111 was m-ought back for test In other cases the l ss of volume was estimated with good accuracy f'ran measurements ID3 do on the dt-l 111 or can --- - - --i' Thia was possible for containers ot rectangular aross section but tailed far thooe of_circular cr-oss section However r e discovered by actually trying out tho idea that visual estimates could be made with surprising accuracy Bight various drums of circular cr-oss section ware a-ushed by reducing t he pressure in3ide until oollapse occurred and in cost 09-ses we made a visual estial te of the loss of volume Tho loss of vol1 m1e ranged frcn 15 per cent to 35 per cont there is a remarkable difference in the appearance of drums at these two extremes Drums wm-e divided into ola sses according t o whether tho loss of volume W'M 10-15 per oent 15-20 per cent 20-25 per cont JO per oent J0-35 per cent 25- Tlwre wu no difficulty in placi ng the druns in tho right class• especially after had moasured two or three Two of the obssrvation given below aro · bued on a visual Htimate of theloss of volwno t'rom photographs of drums of oiroular croas Hotion Tsst of tho thssea The hypotheses were that on collapse the air was compressed adiabatically and that the iz-essure di 'fsrenoo between outside am insids 11as • c• essure of tpe drum in the collapsed position the other W9re 111 1 de to see il' these 'll'l re 1 F ••· The l cruca used were • thicknea 0 048 · There r e two main ridges with 1'ost of the Jai-neso drums were exact reapecta A hyvao auction puizp lonred the prs3sure inside slowly with paus s for equalisation of tempor ture at collapse the pressure inside II 10 ffa J road by a morCUl'y manometer oonneoted to the• inside m UUL •t fi h · - j • oond collapa The vol lDe in the oolla -ed poaition -a rurther wit 1 - n _oop 4 -' '- gra dual -in'-oontra st -to the tir• t which wu audderi • The r r oouai • e · 1 1 - - 1t - a -rough gue• s the oollapse took one quarter of a ·• - nd Ll _o- Ji 'J 752 im _Fir• t collapse· when pressure inaide wu 295 11m Pressure 1ust after oollapH 505 IIDo Volume before collapse 206 400 c c Volume after coll pa• 143 100 c c Atmoapherio pressure Second collap_s _at 483 nm It 1r11S of oourse not possible to get -i he manometer readin instantly att - oollapse be • the mercury bounced about the prHa1re inside waa in the Ngion ot 30 an but wa s falling • lowly due to the oooling of the a ir inaide In • i of this slight uncertainty wa nay say that 1 th• c IDJresaion ratio was 1 44 and the pressure inside at oallapae namely 295 mn • hould have ri• en to 495 mn oompared with 505 ou c-nd 2 thll ir•aaure outai de aft er t'irat oollapae lZ'H • ure Htim ted inside namaly 495 plua the • tre h in the oollapnd poait i cn namely 269 • one deaaribed aboYe ant are not q110hd here Bl ue- i nt oontainer s-a ieai _111 '4 - 4 - i Nt _ kft - 1n1M 1D1 contain_ ahown in H 7 7 a blw-print ot a oellent oonatruot1011 bMi a nioely titting in the text because the papers would act aa practically The metal lever catch was badly bent due to the compressed air inside out at the • Im• time The results were u follows Criginal volume 57125 o c Vol llle in collapsed state 51525 c o Los• in volume 9 s Adiabatic pressm-e developed 2 28 p s i per cent Further-oollapse in collapsed state at ·4 2 an Hg 0 81 p s i 3 1 p s i Peak pressm-e in blaat Hence we have P 3 1 p s i at R 5000 ft Comparison with the peak p -e usure e•timated at di• tanoes 4000 - 5000 ft shows R Mle that the present value i • low 1 The container was The follO'lfing are posaible reasons - in a •trong bu ilding and the pressure inside may not have reached the outside value at the plaoe where the containar wu situated unlikely 2 The wall• of the container may have partly blown out• betcre the lid gave ny mo•t probable 3 The oontainer may have oontained nuneroua papers at explosion poaaible n 12 reachld inside_ the can was 13 p s 1 _·· v· - --- -i ' we Adding 1 p s i fer the strength of the have 14 p s i P at R 3800 ft Crushed can Hirosh i l la The can shown in H 68 4100 ft s of X in Hiroshima 1'laS estimated fran measurements on the can to have a compression ratio 1 22 or the loss of volume '- cent • The adiabatic pressure was 4 65 p s i Adding 0 85 p s i far he strength of the can in the canpressed state the metal was v«ry thin we have P Crushed asoline dru 5 5 p s i at R 4100 ft Hiroshima The crushed gasoline druc appearing in H 46 and H 25 was es ti ted visually to have suffered aloes of volume 5 per cent caniression rati o 1 53 The adiabatic overpressure was the-refore 13 p s i This d1'lm was similar to those tested in the laboratory except that the druns tested had four small ridges in each of the three sections separated by the large ridges - e fowid that drums with a c0111pression similar to the one now being considered were able to withstand about 9 p s i and that in the crushed state - ere able to withstand 5 5 p s i Reducing this to 5 p s i to allow for the sanewhst sker construction we have _P Critioal limits 18 p s i at R tor the o ruahi There mire very aany four-gallon gasoline cans 9 x an fiiroshima I agx-eed at this value I DECLASSlFJE O I Authority AUJ1 7fC o5'7 oonstruotion •Some are oe de of copper and some of galvanised sheet iron One f the better type of copper pipes was found on the reinfca-ced concrete building next to the Shin Aioi Bashi in Hiroshima 1000 f t •• of X Th• particular building appears in H 26 where the drain pipe is also seen J section of this pipe was taken is a SeJit le the distortion of the cross section was not quite the same over the whole length of the pipe but it we thought that the sample was a fair e verage • Figure •1 gives a traoing of the arosa Hction 'rhe tru of section was meMured with the aid of a pla nimeter and found to be 1 9 in 2 • n original dimensions were found by straightening the Ntal and maauring them the internal dim naions were 1 78 x 2 8 in 2 20 H1nc1 tba cxnpresaion ratio The llir pressure inside the pipe frcia 11'1 iur• l therefca-e rose to O p s i above atmospheric Now we must assure ourselvH that the wind pressure • not enoogh to crush n- the _pipe after the pressure inside am out had beooae _ xi - - y t ma e1 y equal By tea_ting the section we estimated that about 5 p s i · -°-' ther diatca-ticn • • • - re necessary to cauae This is far in ucesa of the wind ir•ulre hydrostatic pressure in the blast a ocordir g to this obs P i · •· I '·- 5 p s i at R 1000 ft ti 'l'he t --- -· t erva on is 5 P - i - - ' ' ··- h explained above this is an incorrect result· • the cause of the •rrcairuat be that the pipe was not 'criished unifm- w y over its whole length n 14 SE RET tailed but about 20 per cent were intact Sane of the paper was lrought bao tor teat It was not posaible to make a s a tiatactory panal on the full aoa1e and therefore half acal e wu choaen e h with a hole 3• x 1¾ The paper wu mounted between two wooden olampa ne ot the clamp• was· the end wall ot a box f'ran which •re two tubes one to a Jzyvao pt lllp and the other to a am-our wa ngw ter When the piinp was ate rted a partial vactnm1 was created in the box and the paper panel pulled imfflrds However the paper did not lreak beoauae the paper wu not air tight and a tr rushed through the bolH The P BP waa not able to burat th9 p aper• When a Jr•ssure pump we put on the box the paper bowed outwar d s but again the leakage of air waa aerious and the reading nre not oonai dared reliable third poaaibility was frif'd and f ound to be auooeaatul The paper pa ml meraury waa meaaured and t hat head which oauaed rupti re waa noted did not leak through tlw hole• in the papc- they nre auoh too tine Th• curvature ot tlw paper pane l Ttry aaall aDII the bowina wu one to two milliaefre5 'lb• •liaht ditferenoe of bead betwe11 the oentro ot the panel and the edge• wu ooneidered unM111 J1 K tant ' 'be head 1raa Nuured to the centre ot the panel Tbl head of merQll y to oauao Id lure in tbNe aep -ate teata w z-e fallont- of aome glaaa - r is chute went fro11 the top floor to the growd floo '· •J l _arg ' lrl lgi1_1g 'SE of I on the SI -c rne of the T j ction of t tie street car ' - ·· -' - ·- - _ _ _ J_ines ·1'here were three panels on e h f k p t a11 't elow the posting orifice The '# ··--· paneus were 2 1 6 1 1 0 end 1 1 0n re ii iveiy _in this order descend ing · The disu nce 0 between the supporticg teel edges ' n ither side was o• ill the edges were imet into a metal sheet slot 3 4 deep but the support against the large force of fracture could o ly be· considered simple and first floors were frectured One of the two that escaped was a 2 1t• nel ill but one of the panels on the higher floors were intact It appeans that the peak pressure near the growxi was greater than it -was hither up No doubt the panels in the upper floor s escaped fracture to some extent because b y tia the reflected ahock from the ground reached t hem some a ir had ent -red the postine orifice arxi built up a supporting pressure of a few p s i on th _iosi c No d oubt also the peak pressure does drop in the conditiom pr•nilinc here at t he rate of perhaps l p s 1 per 20 ft ris e o bovo crourd levol Tt-b point n eds t heor ct i cnl study Specimens of the gla• a wero aubr i ttod to tochnolocist of 1• essrs Pilkin ton Ltc a r H M N the El•ss He atl't«i that U-e tonuil• ie'1 against bending fracture of the glass wot id be nry closo to 9000 lb in 2 stre t h r e also the opinion that class was ude by t he Anerican Plate Gle s Company - in The thickne3 of t he class was 0 2 - J '- lj Considering tl e panel as a beu we ve c usins fracture 11as 4 5f lb in p s i Correctin fer t he Therefore we estiaete P 22 p s i • f is condderec a fair estbate Perhaps 2· or J p s i srculc be adde to illC lf for -tho suri rts being slightly better than si• e a t c allow for the fact t hat the loed was not xa tly criticztl • F owt vcr ·1t seecs safe to say t t the peak pres iur• •t the br-rourd l e vel •s between 20 and 30 p s 1 ard tl ro t the best esti N te is P 25 p a i • at R 600 tt• - JI· Thi H•ur-• wu tb eretore 0 96 p a f tor ttii bill ioale pinil i l l tar Uii taot ibii on itii full • • ttii p WOW cl 1- rilaUTilyj on bal 1' tbi thiakniu w 1n eetiJDat e ot the the crecking - ' t ank uilding ·SE of l on the SW corner of - ·• '- ' ' l'here were three panels on each ·rioor all be low t he posting orifice y ·' ' were 2 1 6 11 l'0 end 1 1 0 8 re · ·iv ly in this order descending Th The disunce ill the edges were inset between the supporting' s teel edges on either side was 6• into a metal sheet slot 3 4 deep but the support against the large force of F'our of the panels on the ground fracture could o nly be considered simple One of the two that scaped was a 2 1 tn nel and first noors were fractured All but one of the panels on the higher noors were intact It appears that the peak pressure near the gro ll'Xl was greeter then it · was hicher up ho doubt the panels in the upper noors escapee fracture to some extent because by t iae the refle cted s hock from the ground re-ached t hem soo e air had ent red the poetinl orifi ce arxl built up e supporting pressure of a few p s i on th _in si e No d oubt also the penk pressure does drop in th conditiorw preffilin here at the rate of perhaps 1 p s i per 20 ft ris e above grourxl level n-1 int n•eds t heorcticcl study specimens of the glasa were aubaitted to tcchnolocis t of tessrs Pilkington Lt u- H 1 oore th• t l • ss He atated that th tensile 1• strength agains t bending fracture of the gless would b• v ry close to 9000 lb in F e els o e x ressed the opinion that class was aede by t e Aaerican Plate Glass Company i n tu11erical The t hickness of the class was 0 23 - Jr · · Consic ering the panel as a beaa we r ve that the bending soaent I c· · c «1_ 2 6 lb · i n ·i re c e -· · ' bis equals 9000 x t 23 'I causing fracture 1Y8S 4 5F lb l n ' P • 17 5 p s i Correctin fer t he f inite le th of the panel -re_qu 'e the ·· · - addition of 25 t o tris value of P i co ridcerec a fair es t bat e Therefore we estitete P 22 p s i Perrcrs 2 or 3 This p s 1 srculc be addec to il lMi for tho surports being slightly better t han sia i E e ld tc allow for tl e f act that the locd was not cxertly criti cal Howt ver it see s safe to say th t U-e peak prescur e at the grourd level r es between 20 and 30 p s 1 estimate is at R 600 tt em tl-pt the best - ---- - -·- -_-_ - -- ___ - - a e ' in Hiroah ima _ oo • tt W N 11 ot X had small wooden -- t · ' · ' a • aratruaturi on par t at thi main oun P oh · • --_of Ofet •· t- '•'i ' ' ---t · tff u r - t i oodin tlocr na COflZ ed liy na ta a·t ot ttii blaat WU9 tm timi ot thi loaion and wu aitfioioiit to lrii Jc the joists or one ot- ttiii tio realc••t to pieoea t f • setlcxr i -- • '• • · - - Jibe atrongth ot the panel whi_c h Jrok• • ••• • was •ntirely that • 1 I • ot t _ Joists the boorcla i avo no reoistaiioe to tho dowimn1rd thrust of · the pr•ssure 4' •• and boards- were tongued aa a in Pigur• • The the space low th• floor was to ' It is interesting to note that -the panel corresponding to tbs one tha l roke at the other • 1 of the floor did not fail Hence 0 we may have SCIH contidenco au 'ficient to OBUSe failure of the end panel• Tb9 panel whioh gav• way was 43• x 136 It was- supported by 9 equidi stant Joints at 17 centres • The ends of the Joists nearHt to the wall rested on the brick party wall and were therefore only simply supported The joist • ran trom the party wall over a beam 4 x 6 and than ovsr two more similar beams each spaoed at 43 centres and finally on to a mn l 'trick wll based on the cowoto Eight ot the nine Joiata under tha panel -nr• snapped The one that did not tail was the ocrMr one •• The one at the other corner wu different trcin the rest it was • ly supported at both enda Apparently that earner of the floor had boon taken up at some ti me and the Joiat out through or r e plaoed by a short piece DECLASSlFlED Authonty tJtJD 7 qooSj ' SECRET in very good oondition - there was not the slightest trace of inches 2 22 2 2 14- 21 16 4- 2 1 16 15 5 2 14- 6 2 1i16 12 2 1 16 2 14- 2 1 16 2 14- 21 2 •1 16 · • J· ' ' • · '· · · · t t · - ' 7 8 9 not lroken Henoe the average dim•nsions v ere 2 1 16• x 2 1 16• and the point ot lreak - s 14• t'rOII the simply supported end ucept far the end joist whioh was simply aupported both enda and broke in the middle To repreaent joiata 1 - 8 we oonaider a beam on three • pen-ta all at the aame level the aeparation between the beam• beina L L • 43• Th• left-hand aupport and the central aupport are aiq le while the r14lht on• 1a clamped · horizontally wt w be the load per 'inoh run and meaaure z t'rom th• lett-ham aupport Writing y for th• detleotion we have the following oondi tiona - t X 2L DECLASSIFIED Authonl AWJl 71 Then the load per inch run strength of · the beam probl 1111 above wa get aa the load tor the single joist w 136 P 8 28 l ccording to the Forest Products Laboratory the max imun bending moment to a ii t - ' is the average of 14050 and 17280 i e 15660 lb in i i - · leaa 10 per cent to the tact that the wood was cross-grained 1 in 10 at the ' #-i·' ' · · ·- · - - - • failure Therefci-e the bending manent to snap a 2• x section is 14100 lb in f2'- - ' · The ccrrespoiding figure fer a section 2 1 16• x 2 1 16 is 15500 lb in Hence -- t··v · • o on1 136 P s 29 x 1 3 2 15 500 p 6 6 p s i There still remains open the question whetb r the failure of the J9ists It does not see was irogreasive and of course the time taken fer failure profitable to elaborate th• calculation• t'urth_ since soae of the ccrrectiona ' are po itive and others negati ve P ft have as our best eat imate 6 6 p a i at R • 3800 tt It should be noted that there wre two other W00den panels S t ported by Joista in another part of the ain noer 'lb• Joi ah spanned supported at both ends and spaoed at 17 centres horizontally by 2• vertically 31 • ware simply The diasntions ·wsre The Joists had tailed badly 1 • The critical peak iressure is between 5 _and 6 p s i but these panels were not as good fer our purposes as the one described above 1'oolchHt N An excellent· example of the dishin of a netal plate aaohine ahop 2300 ft • _from Ht of ¾- x wu found in a be_aty The plate waa 16 x 18 and there was a P• rit at the centre · No distortion of the supporting edges ccw d be ' ' observed even when a atraight edge was held against them • II 18 SECRET DECLASSlFIED Authonty tJtJD I qoc-51_ i DECLASSLFIED Authonty N JD_ - qec5j SECRET Henoe L • ot tne A 2050 lb tour ·• l1i ------- ----- 1 Sa te door Hiroahima A aa te of old and cheap design 200 t't 1'ran X was d ubad by tile blast• The two mi ld-eteel door• Syll' llletrioal about their central free edges were 72 l x 2 4 - • Th8 tbiclaless wa• 0 363 ' I be aet at the centre of the 1'reo edges was 4- • The two short • ides ot' each door nre aimply suppcrted the long central edge was of course tree and tho outside long edge was • imply supported except for two hinge• each 10• long the centres of the binges being separated by 20• A model ' IS made of half' the door and an approximately uniform load was applied by means of a testing machine acting through wooden blocks a rd wet sand The results or the test and the deduction s tharef'rom are na ll' al '-• given below Door model 14- 8 x 4- 95• model• Inaid• dimensiona of box 6 x 15 r Length and width to • cal a 1 5 ThiclcnHa on one'1'ifth aoale would be 0 0726• Actual thickneaa 0 05611 • lb Load Deflection 0 200 300 4-00 500 600 700 800 900 1000 inchea 0 0 20 0 25 0 30 0 35 0 39 0 4-2 0 4-7 0 54 0 58 0 61 0 62 0 65 o 68 0 70 0 72 · 0 74·0 76 0 79 0 1340 90 0 95 1 02 25 100 1100 1200 1300 14-00 1500 1600 1700 1800 1900 2000 2100 Ht a tt - r- inin 2100 lb load waa 0 23• • Not an exact II 20 _ • 0 41 21 35 0 47 249 49 0 62 ·WI 63 0 78 447' 77 0 87 547 91 0 95 6 4 0 36 1 07 743 0 40 843 0 46 992 0 58 o 61 0 73 0 81 Did not give a very well shaped permanent set The tree edge was not bond very regularl y but the set was approximately 0 5• It will be seen by a ca aparieon of tlie la st tl'lo oal umns with the pr vious table that the ordinary l aws of elastic deformation hold tairly -well According to these laws if the t lickness is increased n times and the load n 2 then the detleotion is n times less From Figure is 2700 l b Figure S shows the results graphically the load required to produce the scale Scaling this back to a thiokna s 0 072611 l -600 lb The pr essure p is therefor• 4600 ' 6 X 15• 7 49 p s i SEC T DECLASSlFJEO 1Authonty A tJD 710c 2j_ This figure perhapaneed• correction to allow tor th9 tact that the mild An of'f'ice cabinet in the Conmunications Bureau in Hiro• hima 5000 tt B N E ot X wa s dished by the bla• t The top metal plate 231 x 15• and the thick t hioknes• was 0 01 B• The edge• wr• aimped ¾ over the • idH centre waa 1 3• and the two- long edges ware bent in i a Ji Tht set at the 3 16• respectively The short edgH were straight It was considered that the tHts on the Nagasaki tool oti • t trovided sufficient data for evaluating this o'beervation Aoocrdi ng to the test • on th9 tool ohel t a_ load 4-1 p s i on a panel freely supported at the edgH 0 021• thiok and 6 x 6¾• • ides gave a· set 0 35• This is ju at the amount required according to the scaling msthod u• ed below We define ths effective • i ze of the plate• int era of tlw au ot the i mrerH squares of the sides This method of procedure i • reasonable u long aa th• edgH are not widely different in length - I t the edge• ware very di N'erent the whole problem would change in ohara oter For uuiple it one edge were infinite th• meta l plate would yield 11110h more euily only a bendina Ht however would · be involved and not a • tretohing of the central part• ot th• di• h oan euily be atraightened type w• A bending Ht out with a ma ll ett and anvil b lt ttw Hts of the are con• i dering cannot beoa u• e the central pa rts of the metal have been • tretohed Thu• tor th• oftioe oabi Mt L1 2 1 15 2 1 23 6 2 1 • 2 1 6 2 ··• 17 9• 1 6 75 21 i • fl6 25 and tar the plat reprn•nt -• he tool chHt L J½• ·• ' 1 -- - Th• ratio ot metal thioki leH to L tor th• oftio• cabinet ·'· • the ID04el tool obe• t 1• 0 00336 • • A •• st - ' • l - - •- • Tho pi-eaaui-H are 1n 9 thi H •r • • • Authontv Ali IT' 7 Qrr t q DECLASSlFIEO Authority i tJt ID 10oa Thia obaerva t1 r 1a uaure ••t ma ted here and that estimated at 12 000 ft fran the paper panels ia F Ob b y real' and illustrates the local variations in the peak pressure at e' • t g••diatances I Cbe panel of the end rail of a large timber storage barn 15 000 ft S E of X in iN ahima had tailed under the blast The barn was practially airtight there WC'e no YindOffs and only anall doors both of which the owner of the barn stnte d hatical'l y ware closed at the time of the explosion cracks about 1 16 between the planks of which the wall was made panel which bad tailed wu Dade of vertical planks o 5 There were The particular thick nailed on to snen horizontal pieces of wood 1 4• horizontally x 1 85 depth average values point of failure averaged 51• from one end 12' high The planks ware about The 11 wide and The 1f0od was red pine aimilar to the wood of the Joists of the · telephone exchange The clearanoe a t the bottom above ground was p l were f'ree top and bottom The at OSII 1• and the piecea of which there were 7 each 110 long were nailed at each end to sturdy circular Ufright pieces buried in The end wall ooneiated of two panel• of the type dHcribed and the the ground aide walls were aimilar except that the apana were about 85 instead of 110- The inverted V apaoH at the two end W1lll• were filled in by planka mounted horizontally TheH W'C'e undiaturbed as were all panel• except the one dHaribed Uaing the meuured yield atrenath of tha timber the hydroatatio i c-eHure is 0 23 p a 1 preaaure Thia we eatinate mu at ·be inareued to 0 28 p a i to gi Te the peak the blut P Therefor• 0 28 p a i at R • 15000 tt • i DECLASSlFl ED Authoricy tJtJD 710o5j_ sUlllJlY 'Pbrhd t·ci n 7hey ere apa d 8 f aet u art fhe _ o a - _- • ' _ ' Tile 10 id per squ ire inch on the · -all that would caua the inap •ing of' 'the vert ical sup 1ort s is c 4-0 p s i finite yiel d t lll e i -1 rot si inificant The correction t J be r 1ade for the 1 ear to the barracks were so ne X•per Xlnel screens expose·d aideon to 1he blast 'i'h se 7 e r r ot dL '1 ed al t cu 'l they • o•i c l ve at 0 5 p i Her ce _we me y ' OJ° i fa moderate c cc u cl cy t 'iat P 0 45 p s i at R 23 JC0 ft SECRITT concrete Titis fli'jfr ·Af n sOn ble eetim te of the peak pressure lll 9 Y be theory of the dhhing of a me -ibr e by pressure The nel wi o 8' x 13' 9J'ld the dish 7'3 s 22 The oonorete l l s 4¥' thiok · but this t a t is not• required since he· concrete oted only as a medium tor ap plyinz the load to the stet l According to the theory by Kirlorood see report The reint'o rcements -ere 5 16 11 rods on 611 centres ot dishing given by Taylor s 24 equation 44 ' P the sides being 2a ind 2b a b ii nnd 1 similcu- theory the dis ing H produced by a pressure given by 1 sech 27 2b -- 520 lb in whel ¥ 1 T is the yielding tension per inch run In our c a se assuming a 20 ton steel T y fo nul9Y it is f ound that P 20 p s i Substituting into The loc eleI' tions o the -0 f'ini te time DECLASSIFIE D AUthonty tJJJD I 10c5j_ Pailure ot Reinforced Concrete P nele N l saki 11ld Hiroshina the Tot'F do i'Torb in shelter in th the windo 1s which in any The only opening of the gro nd floor t Thus when the press pressure the building the pressure in the shelter did no r -• other hand the large vtindow space on the ground floor level · · --' - alloYted the pressure inside to re lch ' ' th 'l t of the outeide 1 lmost inshnta neousl y The et't'eot of the preosure difference on the two oides of the rein forced concrete n or ws s to dish in i ost or the p t nels A oaretul 3tudy al lll de of the const1' lotic- j or the floor in Qrder to estilll lte the peak pressure in the blast Practically the s Ulle situation to th t just dtscribed was round in the Radio Building in Hiroshi m 3100 rt N E of X Again details were t aken ot the floor construotion The problem of the Nagasaki psnel was riut to the Co n orete Section of the Road Research LaboratoryI Ha nnondsworth I t o estimate the unifora static load th t 110uld cause failure t o the extent observed It eeems howe-rer that not sut'ficient is known about oono te panels to make DY absolut figures cerh in The edge onditions ot the panel are important for assessing the load and all that the Ro9d Research Laboratory could say after eon• idering the pn-blem in some detail was th3 t the load w l• more t 10 p • i and les• than 20 p s i • Muoh closer limit• to the peak pre •• ure uy be had by comparing the floor t in Nag • aki with the one in Hiroahima Since the pressure acting on the latter is knOl' 11 with some 5acuraoy t'rom other obaervations it is possible to obtain trom the compariaon a reliable estim te tor Nagasaki There were three similar panels in agas lki e lch N 51 · The p l lels were continuous on t1-ir 23' 15' x 23' • see Pictures sides running over heavy beams brought up over the supports and o riapped · the next 9 5 9 811 and abo ut The efleotion at the centres were - · third panel could mt be' me -sured accurat ly t ecause an eno noous pile of rubble Md been throffll on it A fourth panel -i th a little extra support in one Orner due to the stairs had not dished The floor panel in Hiroshima was 20' x 20' centres A second similar panel hsd not failed as ell as the one in Hiroshi ms ws J li the rods were ½ at 6 11 The panel was not supported • Two contiguous edges were fixed to the outside one of the other edges was continuous over a heavy be m undenieath am then on to the f s r ws 11 The fourth edge was al so continuous to a neighbouring room but the support was a vertical au across the J uil d ing side of this W ll l was a similar panel which had not fciArled n throu to the b sement On the other The -1 l d id not There appeared to be no failure of this wall but a curious pattern was seen on the wall itsel ' 10 centres slx wed clearly in the mortar Vertical g s about 3 w ide and No doubt the steel reinforcements caused the p_ J aster to craok a nd fall away near the steel The steel as exposed in a few places The · noor iv l 13 also attached to a aeoond ffrtical wall whi ch did not go through to the basement 11 - i Thia waas was 5' 611 f rom one side and made the fourth of a r oom 20' x 14' 6 11 • This 'lf U contained • door and a window •aidej _t the wall was the passage leading from the t ron door -·1 had parted t rom tblt noor i • _ ••• -z - u i • · · · refntorc--nta were • _•ri -_i c ''_ l _ 51 ••• ' - T _ ••• ·W · 1'he1 d efieotion - V • On the other The passage wall there n a a 3 gap at the centre and the steel 15' x 23' dished 9 8 9 lld the Hirosh ir e p i nel 14' 611 x 20' di5hed 7 Since the p reasure g 5 p si on the N g l s9 ki panel • e · within 1 p s i P 12 5 p s i R l 05O ft in U a gasal i D£CL '5SlFl£ D 1 utboril N Ji 7' Cc-5j_ arthemrare drain _ - 1 This is an int esting example because it demonstrates the great pressure d9veloped on the ground in Nagasaki Untortunately the observation does not permit more than rough acouraoy A TO rtica J • arthemrare era in pipe halt embedded on the sioping bank of the approach to tha Shiroyama Primary School a few yards from the place where picture N was taken 11 50 tt 1' ot X 'lroke under the pressure of the blast · Th• embedded t was very firmly supportad and the whole pipe was held in position by a layer of rtar ¾ thiok The pipe was only JO long the mernal diame er of the earthenware was 10 and the tlli ckness broken to pieces ½ The entire exposed half was most of the pieoe were awept away by ttie blast b- it a few were trapped inside There seemed no doubt that the pipe had tailed by crushing and not by flexure Professor A J S Pippard of Imperial College London kindly a de a test described in h1 e letter which is printed below Usi bis data neglecting the strength ot the mortar and scaling down fran a thicknesa i to ½- the crushing pressure was 1 50 p s i P 1 50 p s -i at R 11 50 ft This estimate must be qualified to some extent as follows It is not clear how the estilmt• of peak pi -essure should be modified to allow tor the contours of the ·bank and tb 1 s obtain an es imate - tor level ground The pulse that 'lroke th •'#' ·• ' j ' -$ - 7 -- pipe was probably the reflected wave fran the path just in front of the drain pi · The peak IrHSure or course is llllltiplied by a taotor at lea st four on thia ·'-- - tleotion • ' o Wi '• • • · • • '1f When this ret'leoted pulse strikes the bank still a ttrthe r llllltiplif oation is introduced In other wcrds the incident downward pulse bu bel n DECLASSlFI ED 11 Uthont tJtJD 7 qec 51_ JI - f· • funneled into a ndge shaped angle about 120° one side of the wedge • uwJ tiplioation produoed_ in this way Dear Penney I have nO 't ma de the test on the earthenware pipe which I think is as near as we am hope to get to -the oonditions you described to me · The pipe was 10i outside diameter and 9 inside diameter and a section a pprox imately 4• wide waa out from this and bedded in s and i n a specially a de box Care - t _a ken to prevent the sand from getting inside the pipe by puttying the edges of the specimen to the sides of the box - A st it't bar was placed on the top surface of the sand and a load applied to its centre thus approximating to a uniformly distributed load along the s d resti ofj t he t 09 aeotion of the specimen The t e st went quite stea dily until a load of 4 75 tona •u Nac hed when a cra ck waa heard and on di smantling t he test w found that tbD specime i had cra cked i nto four e qual quadra nt • the oraok• being at the endJa of the · diameters t hrough the loa d and at th• end• of the cUametor at rilht-an H to this I t looks theref ore n• t hough thi• pip• ff0' ud take frm 1 - 1¼ t ons per inch l ength before failure oc curred provided th• l oad oan be oon• idered reasonably uniform · likely i n your caae Your• • incerol C• d· A J s Pippa rd l' S Hem• ' i ·J i The dif ficulty in intarpreting the failure a pole ia of course the uncertain pull from thB wires · fY The IIIOl St f avourabl e nr polea in a line running to the Mitsubishi Torpedo Works One pole a little off the main line i resumably caITying a side line at 400 f eet from measured the wood was in good condition there appeared to be only one ii cross bnr and the blast ran in the direction of the wires The pole was snapped at ground level Diameter at baa B• Wood Di11111eter et top 7 Height 27' Thuyopsis dolobrata or Conninghamia lanceolata Bending ltoment to snaps• section according to Forest Products Laboratory Prince• Risborough 1 125 000 lb in •• Bsuming 11 The dr8g pressure was 24 p s i • dr coefficient 0 5 apparently the proper value for a smooth 16 p s i cylinder of such large diameter a peak pressure we a Now we have to oonsider two factors influencing our opinion in opposite d ireotiona First that the pole snapped •so that the estimate must be increased Second that the pole carried wires Si nce the wires wore not much affected by the wind of the e cploaion it is submitted that a peak pressure 16 p s i at 400 ft is apProx tely cOXTeot P 16 P•••i• at R 400 1't An interesting observation but not one fran which we have attBIIt ted a qualitative estimate related to a pole near the Shiroyama Primary School This pole anapp• 11 feet fran th• tops the diameter wu uniformly - and the surface was very emooth 1 The diatance from X wu 1100 teet App •ntly this pole snapped in ti» incident pulH before th• reflected shock reached it Other examples in • ·•· the range 1000 t't - 1 500 t't 111r• found where the rather heavy and complicated top• of ponr polH mapped the pole near the top _ See-1 1 ot re N Lt -· •• __ ' • · - -- - _ II 26 DECLASSlFJED Authority lJ Jt 1Co5j_ the reason for the very strong riethod of fixing the base Pei lure of I bee m lliroshima A rolled mild · teel I beam 1900 ft W of X in lliroshima failed under the high it was used to c Ty other side The cross wire 'lbere appeared to be no need to al low - ' ' J- • 'i ' for the drag of the wire on the beam and there were no complications involved in _ r ' -- · the nature of cross arms or other objects at the top parallel to the centre web Figure 1 of the beam gives a tracing of the shape of the beam copied from a piece of paper pressed against the o 1d of the beam one other similar beam which had collapsed completely was The bla st was There was The be8111 whic l stood up partly shielded from tho blast by a building and therefore the estimate _b aaed on this beam must be considered a lower limit M Let S be the 4rag pressure per square inch on the beaa _ i - ent j •' r Then the bending about the be 3e · is 242000 S lb in Lot Y bo tho yiold otrongth Thon tho yiolding bonding ' is 16 9 Y 2 t jwith _Y 25 tons in tho most pi·obable value fer this steel the bending · · ' OJ' imanent i f is 95000 lb in S • _Th• · Hence drag coefficient is about 2 56 ••••1 0 8 and the peak pressure in the shock just · autticient to 011-use failure is 12 p s i Al lowing tor the tact that one_ I beam was irostrate while the partly 4 · _ ' _ - - t - one waa only slightly bent we estimate that the peak iressure wa s • P 15 p s i at R 1900 ft SECR'· - · ·• i ' J • · ·· · Fl le3 in Hiroshima Two observations on flagpoles · i Hiroshi JM now to be described are · i e • 1 • c • ngton ' J -- tt · - 1 - a le ngth of one of the poles including the ben part 1 and Their shock pressure that wc uld just OJ l8e yield ' - q t I' '-· The first fla grole from the vt 200 ft E of X was of leng h2t9 11 ' de to al l01r for the- inertia of the lee and the finite lllgle of bend outside diameter 2 4-0 11 thickness 0 140 la-'6n 1Ilgl e 0 114 radians According to N P L the drag pressure N Ls 0 50 p s i and the _ s rook velocity to cause yield W'ls 1 186 air _e whar- e is the und ve ocity in the original The drag coefficient we s 0 1 5 The shook pressure lffis 7 0 ·p s i The moment of iner i is 8850 lb ft 2 and the yielding her di ng moment in the set c ondition Using fonnuls e ' 2 x 10 4- W9 3 1 poundal s rt If and knowing by the meth d ot tri e l and 9 Jld error that the f inal c hook pressure i3 about 10 s i we have o 2 44 seo-1 £ a d 0 23 p s i Fe Finding a self-consistent s olution using the V Lriation of drag coeffi cient with Reynolds number given in the N P L diagnm Figure l of his Report we get that the drag coefficient was 0 2 the time of yi ld was 6 28 sec and P 9 2 e t p s i R 3 200 feet - The error in thia estimtite is considered to be lc s than 1Q% he second pole on the Electric C ci 1t a ny Buildine u ist Moe 2 1 0 0 ft from x e6i length 114- outside di Dleter 2 4-0 thioknes 0 140'' yielded an llgle radian• Accord ing to N P L the s hook vclooi ty was 1 342 shook pressure WS '5 13 6 p s 1 c - tor he obHrved' yield V 0' lt- '' oond · · · ft2 · ' feet 3 1 aeo- 1 we find t t in o er tc aoobunt - the ·peak drag preaaure a• 3 4L p e i'• 'ffii ti iii of 0 waa 15 l p a 1 960 - and the yielding bending moment as before wu 9 2 x - P Un o • Md t he r-eak The moment of jnertia w l s ' '_ Uaing the value of 0 075 The drag pressure to cause yield was 2 2 p s i • and the draze oosfficient was o _6 55 e o- - -· tho oootr - Int drag 0 69 I - · t · - t•-r --- · r · ' · · i •t i DECLASSLFIE D Authority 1$ NND I q00 @ 2100 rt ·t• Homlcins N P L '- 'll_ _ - 'f her to your ieit of 23rd November the calculations sugsested The tube lett wi th 11s was measured o nd a ppe l red to be straight down to a point lbout 18 inches t'rom the lower end A Ortion 18 inches long was cut to include the straight Ortion llld tested in bending using four Oint loading in a testing machine The bending moment appl ied hen the strain-novement curve ceased to be linear as detennined by electric str in was taken as being e qual to tlut a lied 9 t the point of junction gauges of the distorted and undistorted parts ot the tube during the e osion From this the unttonnly distributed pressure which might have been 'Plied to the tube was calculated using conventionl l tonnub e Results u-e as follows Length ot Mast 239 inches Effective Pressure Length ot llast E f fective Pressure 0 5 • 0 12 lb s q in 114- inches • 2 62 • 18 lb sq i n In calculating the strength ot the blas t waves neoess ry to roduce these pressures the s implifyi ng a ssumr tion has been 111 d e t hat the wave is ot' · the torm ot a no nnal shook wave t ravell i ng perpendic ularly t o t he lll9 St a nd t he etteot ot heat radi ated t' rom t he e xplos ion h s f en negl ec ted _ _ j bo ' inte out - hat while I t shoul d the pressure d ifferenoes ssoc i ated_ with the w9 ve • _ tront mq be oonaiderable the ti ml taken for the wsve to pass the mast lesa • - t- #•'· ' ' it han - • o· 2 milliseos - - u ½• ae - 1a small oomrared with the nat ral period ot the mast · • r the 20 ' - rt st It 11 therefore ass • tha t - I ·' t- - ' Y tho pre••ures doveloped by tho air • t touow n ' At u tho· n - • The drag ooetti9ienttor tho mast bu been t J -en M that oorreapond _ng to · -' a airly amoo h oylind r the results for which are available trom N P L ' -• l Compn11Hed Air 'l'unnel teat • J i • - i ' ' '• ft i •i -'• · • Pig 1 Over the r i nge ot Reynolds Number R - SECRET DECLASSLFIED Authonl - llhlJ __7 qco5Cf R· 5 x 105 the_ dr l g coefficient is 0 3 coc ficient of 0 9 ·The ·•- and for the atmosr heric conditions nnst to give the r ca J oulll_ted ressures is 9 bout 1350 ft sec for the 20 ft mn it md 1530 ft sec • • ·• for the ff 1 ft lll l st The conditions behind the w9 ves may be re9 d off f'rom Fig 1 It r 11 be notic d th9 t for the short lll lst the air following the ve is mov ng recent work in the N P L High Speed Tunnels on a l ach number or 9 bout 0 45 c linders indicates th t compressibility effe cts 9 t 1cm Reynolds numbers in genenJ low critical Reynolds N1 111ber incre l se the d coefficient lt libch numbers l bove 0 3 Further int'om l tion on this subject is given in l J -per by A Ferri The Influence of Reynolds Number at high M l ch Numbers Atti di Guidoni No 67 69 which indiC ltes tbt l s the Ms ch number is increased the· Critical Reynolds n1 E1ber for cylinders is lso inore l sed given in this paper it ppears th t at the cL- ag coefficient is I bout 0 9 l By extI'9 pobtin the curv-es Reynolds Number of bout 5 x 1 r ith this val e of the drag ooe ffic ent for the shorter st the W ve velocity is l bout 1470 rt The f i gure s on which the curves or Fig rcr eo 1 arc based are given on t accompanying sheet · Yuurs fa thi'ully C A Hankirus i I NATIONAL F tiS C lL LABJ T0RY i I Dh i s ion Basis of Calculation Calcul l t ion The atmospheric oondi tions MY - been tf i- en as Tessure p1 Tempera ture T1 25°C n iity r1 0739 Jbs cu ft · The ir h s been assumed dry the influence of h idity is unlikely to C U O nore th o J l rer gent e -ror The velocity or sound at 25°c · a - 1138 rt sec Viscosity or air at T3°A given by SECRET DECLASSlFIED Authont 71Cc5j_ itr'here u D •r Jr3 1r3 u re velocity density tempenture l nd visoo11ity of air_ behind 7 l ve D is dbmeter of IMst Drag ooeffioient CD Pressure on JMst Ma NI Gber of W l ve Ji4 0 2 rt ½ f 3 u 2 3 U 91 where U velocity of n re • lli oh NI Uber of 3 ir behind -nve u 3 Uya3 Telocify of IOUnd behin_d wave 113 _ 'l' where a3 · tt sec 298 The conditions mhind tm nve given in Pig 1 m ft • oaleulated uaiain tlw well known Rankine-R lyleigh rehtions throUE h a perpemiow ar sboet n The ''ataiJ ation temperature rise plotted i • tm e ro•aa waperature oTer the ambient temperature T1 ua1aini that the air behind tbe • 1a brouJht to reat aaiabatioally 'l'he er era diaaipation baa been olloulatec ha tba expreHi on inera di• aipation por aq ft ot per eeo r1•• • Sta nation te• perature • Speoitio heat of air at oonatant pruaure • C H U lb l°c taken aa 0 2413 Wr ht iron ladder• on the roof of Cit lliro• hima Office• of the c ity Offices in of X nre affected by the wind of the explosion e -10 ft hig her getting fron one region of some 10 feet h ·· apet wal' Ji an then down about ho feet illu• trat - - t pp_e d to be Pictures H oon• truotion of the• e ladder• ic and H over 1 r The wind of t he expiosion parallel to the parapet and caused the tops of the ladders to' bond over SuppoH that the peak drag pressure was S p s i manent of tl1-e dra g about the point of failure Now n iron 1rll5 4340 Then the bending S p s i • ll118t H imate the bending moment that Clluaed the failure of the It waa not clear to the writer whether the stress distribution 11Cross the ••9tion was t1IO aynmietrioal parts one in tension and the other COIIPt'tssion the magnitude of the atrHs •rywhere being the yield value or mether tlw atrua waa mer• like that prevailing while the atrain wu • till elastic Th• ' bending 1110ment per unit width in the two oaaH art Ta2 4 and Ta2 6 where a is the '_tMolmeaa and T is the yield strength J coordingly a piece of l' t'ought iron was olanl ed as a cantilever the loaded tree length was 27 the width wu 1 45• a rn the thicknes• - • 0 266 The load deflection resul ts were as follows beginning with halt the weight of the bar plus the weight of the pan Load _ _· I r • I lb Deflection 2 0 7 0 65 12 1 34 17 1 97 22 2 60 27 30 3 27 3 80 32 ol 11 Tt ins 4 8 oreeping wu plotted The yield load 1'Ba 30 lb and the SECRET DECLASSIFIED Authomy tJ J 71Co5j_ 7 5 at R 3600 rt pal't cubr of the overturned stones is gnnites 1 Number I granite with bold m9rkings and very highly Poli shed surface• Density 2 57 E J l• per oc 2 Number II granite dull grey stone rough surf l ce Density 2 j8 Keasurements were usu illy confined to I The blast as perpendicular to one t l ce X 9 8 Z 28 5 Stones overturned five eX5D1ples 9 hlost ident io l G ite I Gnnite I I 29 SECRET DECLASSLFIED Authonty J NND - qco5j - - - · CFC 1 u 1 T X 6 5 _ - 17 5 one exs mple over f 1ard 2 Granite I of these the second is the most nearly critical nd re tulve x z 8-55 om T t · 1e know t hat P is approximately 7 p s 1 and vc - i sZ O aec- 1 • Solving p e i Notice how the inertia of the stone put up the It' the wind had been of inf inite duration a value Po 0 318 p s i · • ould have overtu ed the stone compared with 0 92 p s i actually necessary in this explosion The time taken to reach the topp ing posi'tion T aS 0 2 secs Shrine in N as9 k i Kany stones were overturned in a · shrine 4700 t't Not' X in Nagasaki near to the To do Works The blast was not quite nonnal to the stones the angle of ine dence being about 70° The oritioal stones whioh overturned were 14xf6x38 1Dd the turning edge ws s the 1611 side A prelimimuzy solution of the problem shol °s that -roeAz-i we c t 1 36 seo- 1 P is abo ut 8 p s i • Solving Fo 1 16 p s 1 • compare with a v lue 0 49 p s i i n a steady wind Using the dr g ooeft'ioient -10 82 f o und in Hirdlahima we find that the p ak pressure w l s 7 9p s i time to reach the ·toppling position was 0 4 seas To oorreot t or the blast not being incident no lly we divide Fo · by s 70 l nd h l ve for these stones at normal incidence Fo 1 21 to t'ind the peak pressure we take the drq o 77 Hence our t'inal figure is -1 coet't'icient l e 0 8 2 seo 70° 6 i p s i DECLASSLFIED uthonty tiJlJ l _ 1 1 f'c'_2j_ SECR Vi •ltl Estitnates of the Pe 'lk Pressu -e fro the Severi t scientific accuracy of D 1 19 e Nevertreless notes the d UD 'lge e gr e exactly with the results in this RciJC rt even though considered to dis9 0 ree Tni teatur es of interest were the size of the g as I fr9-EPcer its the d9 lll l ge - to win d ow f ames l nd doors the 1 pearance of cracked plaster thews yin which he root structure - - · · ' and oorrtg l ted iron of the outhouses had been q_ffect ed · • 1 -' l nd the g e ner9 l l y oat Mlce of the building f'rom outside and w ithin An l 1 tempt was made t o eotimate where the d 'll l9 ee in N l ges'lki snd in Hiroshima was equal to th t i n the Macdonald Ranch The estimates ·nere 12 000 feet S of X in-Nagas l ki and 7 500 ft in Hiroshill 9 In Nagasaki the dam9 ge to the Prefeatl•nl Office 1 0 800 ft was distinctly more severe then at the odonald R 'lnoh while in the American Consulate 13 000 ft it 9 S slightly less In Hiroshima the chm9 ge on the west side of the Mijuk i Bashi W ls dist ctly more severe th n at the M acdonsld Ranch T1hile 500 ft to S E ot this bridge the chmage W l s distinctly less severe · The pe k pressure l t • he Jlacdom J d Ranch was 1 4- p s i was 10 200 ft from X The distr 1ce Hence thts visutl estimates on the d9111 1 ge 1 t the 1 4- p s i level ouggest th l t the tonns ge in Nagasaki was 1 J 8 gre1 ter th l n -Trinity ' and the tonnage in Hiroshima w i e 0 4-0 th9 t of 'lrinity blast tonnage tor Trinity is 13 000 Thus we have frolll visua l • stims tes alone the tonnaee at Hag 'lld W ls 19 000 The -Ccept ed and s t HiroshiJD 9 a 5 000 A tu her oomp lrison 9 s attempted to 1 ssess the d iatanoe from X in H iroshim l at which the d I IIl l ge was the same as at Dijina Tharf in Nagas9 ki di• tanoe _10 500 tt • The place elected was the Ga a Works distance 6 500 f t ·Thi• oompa ison gives the tonnage in N saki ls h 2 time s th9 t in l iroshi Il _ a in l s isingly a oour Lte estimate 0 ' i ' ti Sui arising 1 4 at R • 12 000 tt in Nagas ki 1 4 at R 500 tt in HiroshiDa ' SECR T not too near he -centre shJulft be titted to that centro sho11ld be fitted to a theoretical pressure curve obtained by the most complete calculations yet perfonoed The disadvant l ge of 1 is th9 t 'the questiont of ling small charges to large charges is involved Also the experimental results on aotual bare charges are variable ll'ld nothing better tru r 10 per cent aocuracy can be given for the peak pressu adius curve tor a iven tonnage or a 20% accuracy in esti ting the tonnage from a known pressure r9 dius curve The disadvantage of 2 is th t the thematical ditt'iculties are very great snd colll lete confidence cannot be placed i n the results so tar obtained Further the results apply only to free-airn Tne complicated shock wave patterns caused by the proximity to tbs £ l OUnd cannot be introduced into the theory The tonnage corresp nding with the theoretical curve i s found by dividing tbs energy relea sed by the detonstion energy of T N T Hvwever the energy released by a T N T e t losit'n in air is to so • e extent augmented by after burning ct the products No doubt this is the principal reason tor the divergence be en the two methxls The simplest procedure and the one actually adopted_ here is to use the e rimental results on l l charges A re X rt O S R D 4076 by A H Taub gives a valuable s 13es ot curves f or the peak pressure on thl gro tor expl sion ot unit weight of T N T above the ground a 'YC'ro1 riate cul •es were selected and scaled up to tit as well as pcssible t o servations in Hiroshi ma and Na gasak i - 5 000 t ons tor Hiroshilll3 l lld 6 The best tit a ' pe _ 20 000 tons for Nagasald• since '-l the h81gh%ot detonation 1850 ll'ld 1800 tt respecti vely we have in the I II 1 c- cR•- r ✓ E two cases 8 3 Hiroshina 5 1 Nagasaki we h ve the following expected valuos · ' _ ___ ··• ' I 900 Rft p p ••• 1 I 22 6 I J 1120 j 1570 · 2240 3140 381 0 4 4130 6100 I • 13 2 9 5 7 2 5 5 4 J 20 6 1 16 5 Ne e se ki I 2500 IRt't 2830 3540 Li240 4950 6000 20 7 14 5 10 5 8 1 5 7 7080 18500 I 4 4 I 3 3 I I P I p s i 1 2 3 ' • I I I The observations to be fitted are e s follovm I J DECLASSlFIEO Authority NND '7 qCc cf SEC RP T ot Results lliroshilll I Method of Determination 1000 Dishing of Reinforced Concrete F9 11el 1900 15 Yielding of ve1·tioal I beam 2100• 15 4- Yielding of steel f lag pole 320QK 9 2 Yielding of steel flag p0le 3600 7 5 Yielding of a wrotJcht i ron l adder 3800K 6 6 Failure ot 'IOOden floor 4100 5 5 Crushing of a tin c lll 1 13 0hine oil can 5000 3 7 Dishing of top of office c bi net 5000 3 2 Crushing of a blue print drum 6500 1 4- Visual est i mate 12oo lil 0 5 Failure of side-on paper p l nel s 0 28 Failure of vrul of timber storage b l ni • 15000 lif- · V l lues marked with an asterisk re cons idered t he most reliable v 6 w Q_ u ____ - - ' ' ' • - J ' f t t 1fF SECRET I SECRF 'm 2300- 1 I 100 Collopse of ea rthenw i dr i in pipe Dishing of top of tool ohest ' 34-00 r 3800 14 405oK ' Cru shed C UI 12 5 Dishing of reinforced oonorete t'lo r p II el Overturni ng of enoritl stones 3- 3 5 Collspse of 4 gulon gasoline csns Visusl es tiir lte 'Hs 11 of bsrr cks st K U igo Vslues rurked with a n ssterisk sre considered the roost reliable • 4 1 1' 'l - _ J J k -• _ · ti% ' ' '' · - 1- ·- • f · -_ SECRE1 III 4 - DECLASSIFIED -- Authority rJ Jl 7 qcc 5'f •u•Ec- ET J ' _ in Tthich these expected curYes fit to the aotus J observatiotu1 Ii • It will be notice J tret the observed f l ' i 1e y re 'luiolr ly than the expected ou s r-¥ OUI Vf a ThiB effect is 'be l ev d to be real repregents tho absorption of energy from the blast wave in causing destruction Furthermore the fit ' t the high ressure end i not go and the writer is not uite certain how the observed curve should be drs wn error Possibly also Taub's curves are in the experimenbl result are not al1r9 Ys reliable near the flame Using Taub' s Figure 11 the f ree air pres sure at 1850 rt from 5000 tons is 11 3 p s i sing a better experimental experiments on the safe amngement for applying a uniform lo In Yet 15and is not very satisfactory air balloon woul d be_muoh better 4-J l b a - - 0-- - '1' ' ' --- 1 _ _ t- RJ - n Nagasaki the preesure on tM_ grourn undoubtedl f -reached 150 - 200 p s i • According to Taub' s Figure 11 the preuure at 1800 t't in tree air from 20 000 tons is 50 p a i On impact at a igid W l ll the ground this wuld rise to 150· - 200 p s i Thia agreea with the o baerva t ione o n the earthenware drain pipe and on the dished too l chsat To aW1111 riae we m - • ay t hat the observations show that the blast equivalent in Hiroshima W ls 5000 tons a nd in Nagasaki was 20 000 tens Theae figures may be oolllp ired with toose obtained by comparing the I B W c9 loulationa with the r- gauge ue l aur imenta of AlvareJ quoted are 5000 and 15 000 tons reapectivel y agree•nt is aatiafaotory es ecial l y when the oomidored t f••t of cloud tore reaching''the - • ' '- i - - • ' · • It _ ' · daoreu• ti- izaitjit 1 • ·or tha •i · - · agree very well with Figure 11 of is fair at large distances Hoy ever taken much further th i n Taub' s curves and one I B M run really represents high explosive cur-res b sed on figures in the high pressure region 20 - 2 p s i nd the agreement here is not good Y T l 1'tmm HM ing to the I B M run tons the over pressure is 1 1526 ft I B M Hence necess 9 ey Until is performed we prefer only experimental curves on high explosive in t'onni ng our estimates tonna se blast equivalent • • ---- - ------------ SECRET III 6 DECLASSIFIED AuthonC II tJtJD I qOo5_CJ_