History

CHAPTER IX

Survey of Battle Casualties, Eighth Air Force,June, July, and August 1944

Allan Palmer, M. D.

The need for information regarding causes ofdeath in KIA (killed in action) battle casualties resulted in the organizationof the Medical ORS (Operational Research Section), Professional ServicesDivision, Office of the Chief Surgeon, ETOUSA (European Theater of Operations,U.S. Army). Maj. Allan Palmer,¹MC,was appointed chief of the section on 1 June 1944.

The purpose of the Medical ORS was to investigate battlecasualties from an operational point of view in order to evaluate moreaccurately the wounding power of various weapons and the effectiveness ofprotective measures. At the start, it also seemed that information would beobtained which would be of value and interest to officers, not only in themedical but also in the other services. It was postulated that the machinery forcollecting data provided adequate liaison with the various branches of the ArmedForces.

Aside from occasional accidents or special incidents(appendix G, p. 827) in the United Kingdom, which were investigated by theMedical ORS, the first operational project to be dealt with was a survey ofbattle casualties sustained by the heavy bombardment groups of the Eighth AirForce. The survey was carried out with the cooperation of Brig. Gen. Malcolm C.Grow, USSTAF (U.S. Strategic Air Forces) surgeon, the Eighth Air ForceOperational Analysis Section, the Ordnance Department, the Royal Air Force, andthe Royal Canadian Air Force. A period of 3 months was taken as the time duringwhich the day operations of the Eighth Air Force might yield a satisfactorysample of casualty data for study. The 3 months chosen were June, July, andAugust 1944 (D-5 to D+86). In September 1944, additional personnel were providedby the Army Air Forces, Air Technical Service Command, for a 3-monthcontinuation of the study which was to include examination of KIA casualtiesfrom the Ninth Air Force and Troop Carrier Command.

An ideal casualty survey would provide complete informationabout all battle casualties and about all individuals exposed to risk,irrespective of the

¹Following the completion of the Eighth Air Force casualty survey, Major Palmer participated in and reported on a ground force battle casualty survey. This survey was conducted with the Third U.S. Army's XII Corps during the last 2 days of the war in Europe.
Based on the experience he had gained in collecting missile casualty data, Major Palmer prepared a draft of a proposed table of organization and equipment for the establishment of a casualty research detachment.-J. C. B.

548

severity of injury. The points which pertain to such a surveyof aircrew battle casualties follow.

1. Strength of forces engaged in operations for the surveyperiod.-(1) Bombardment divisions and groups taking part in operations; (2)types and number of aircraft and combat personnel involved, such as"man-combat missions" carried out; and (3) hospitals serving theEighth Air Force.

2. Losses.-(1) Aircraft and personnel about which noinformation could be secured because of failure to return from enemy territory;and (2) casualty data, including the causes of death and regional distributionof wounds, fractures, and amputations in the personnel killed or wounded byenemy gunfire and returned to the United Kingdom in aircraft, which could beobtained from post mortem examination of the killed, interrogation and X-rayexamination of the wounded in hospital, survey of "Care of Flyer"reports and group operations at AAF (Army Air Force) stations, andidentification of missiles.

3. Battle damage data pertaining to aircraft in whichcasualties were sustained.-This information would be of the greatestimportance for the identification of the weapon causing wounds in cases wherethe responsible missile was not retained by the casualty. As far as protectionto personnel is concerned, a knowledge of the relative frequencies of hits byenemy missiles on aircraft bearing casualties, from various directions, wouldenable one to place protective armor more advantageously in the aircraft or onaircrew personnel.

4. Flight formations.-The formation of heavy bombers inflight should be studied from the point of view of risk to combat crewpersonnel.

COLLECTION OF DATA

Strength of Forces

The 40 heavy bombardment groups of the Eighth Air Force,divided into 3 divisions, are listed in table 174.

All the groups of the 1st Division were composed of B-17aircraft and those of the 2d Division, of B-24 aircraft. Five groups of the 3dDivision (34th, 486th, 487th, 490th, and 493d) were originally composed of B-24aircraft but were changed to B-17's on 24 August, in mid-July, on 1 August, on18 August, and on 18 August, respectively. Thus, the 3d Division also consistedentirely of B-17 aircraft on and after 24 August. Tables 175 and 176 give asummary of Eighth Air Force heavy bomber day operations, by divisions andman-combat missions, for the 3 months' period of this survey. These data wereobtained through the Office of the Surgeon, USSTAF. The total number of aircrewpersonnel taking part in Eighth Air Force heavy bomber day operations is givenin terms of "man-combat missions," the

549

average crew of a B-17 being 9 and the average crew of a B-24being 10. Thus, a total of 69,682 sorties corresponds to a total of 657,096man-combat missions for the 3 months' period.

TABLE 174.- Eighth Air Force heavybombardment groups, by divisions

Group 1st Division	Group 2d Division	Group 3d Division
91st	44th	34th
92d	93d	94th
303d	389th	95th
305th	392d	96th
306th	445th	100th
351st	446th	385th
379th	448th	388th
381st	453d	390th
384th	458th	447th
398th	466th	452d
401st	467th	486th
457th	489th	487th

	491st	490th
	492d	493d

TABLE 175.-Distribution of 69,682 EighthAir Force heavy bomber (B-17, B-24) day operations for the period 1 June to 31August 1944, inclusive, by divisions

Division	Number of sorties	Aircraft missing in action¹		Aircraft returning from enemy territory²
Division	Number of sorties	Number	Percent	Number	Percent
	B-17's
1st	23,488	237	1.01	23,251	98.99
3d	16,236	153	.94	16,083	99.06
Total	39,724	390	.98	39,334	99.02
	B-24's
3d	5,510	31	.56	5,479	99.44
2d	24,448	272	1.11	24,176	98.89
Total	29,958	303	1.01	29,655	98.99
Grand total	69,682	693	.99	68,989	99.01

¹Casualty data are not available.
²Casualty data are available.

550

TABLE 176.-Distribution of 657,096man-combat missions, 1 June to 31 August 1944, inclusive,by Eighth Air Force heavy bomber (B-17, B-24) day operations

Aircraft	Sorties		Man-combat missions
Aircraft	Number	Percent	Number	Percent
B-17:
Missing in action¹	390	0.98	3,510	0.98
Returning from enemy territory²	39,334	99.02	354,006	98.99
Total	39,724	57.00	357,516	54.4
B-24:
Missing in action¹	303	1.01	3,030	1.01
Returning from enemy territory²	29,655	98.99	296,550	98.99
Total	29,958	43.00	299,580	45.6
Grand total	69,682	100.00	657,096	100.0

¹Casualty data are not available.
²Casualty data are available.

The medical installations which served the Eighth Air Forceduring the period, in addition to sick quarters at each AAF station, were the1st, 7th, 65th, 91st, 97th, and 184th General Hospitals and the 49th, 121st,136th, 231st, 280th, and 303d Station Hospitals.

Losses

Of the 69,682 sorties in which 657,096 man-combat missions wereaccomplished, 693 aircraft (0.99 percent) and 6,540 aircrew personnel (1.00percent or 10.0 per 1,000) were MIA (missing in action), leaving a balance of650,556 man-combat missions in 68,989 aircraft, from which battle casualty datawere available for survey. The casualty survey study pertained specifically tobattle casualties resulting from enemy gunfire, sustained by the personnelcarrying out and returning from a total of 650,556 man-combat missions.

A few incidental facts were collected in relation to the6,540 aircrew personnel MIA during the 3 months. Followup records at Eighth AirForce headquarters showed that, for the first 8 months of 1944, 2 out of 5 (40percent) MIA personnel were possibly KIA and 3 out of 5 (60 percent) were knownto be WIA (wounded in action), prisoners of war, or evaders.

By arrangements through official channels, all KIA aircrewbattle casualties returning to the United Kingdom in heavy bombers, as well asall those dying in hospital within 24 hours of entry or before surgicaltreatment, were brought for examination to the Medical ORS laboratory, locatedon the grounds of the

551

FIGURE274.-Operational ResearchSection, Office of the Chief Surgeon, ETOUSA, located on the grounds of theCambridge American Military Cemetery, Cambridge, England A. Workroom. B.Laboratory.

552

Cambridge American Military Cemetery (fig. 274). Examinationsfor the missile and injuries caused by the missile were made. No search was madein any part of the body where it was obvious a missile had not penetrated. TheGraves Registration Service of the Quartermaster Corps at the Cambridge AmericanMilitary Cemetery was very cooperative in notifying the Medical ORS when aircrewbattle casualties were received for burial, so that the bodies could be broughtimmediately to the laboratory for examination.

Questionnaire forms, requesting such data as thecircumstances of death, combat crew position, and altitude, were completed andforwarded by the group surgeons of the AAF stations from which the KIAcasualties came. Of the 110 KIA casualties for the 3 months' period, 89 (81percent) were examined at the Medical ORS. Of the 21 not examined, 7 werecasualties who died in a hospital 1 to 6 days after being wounded.

Daily admission and disposition reports were received fromthe 12 hospitals serving the Eighth Air Force. From these, the entries ofaircrew battle casualties were noted. There were 1,007 WIA battle casualties forthe 3 months' period. Since ORS consisted of only one medical officer and oneenlisted man during the first 2 months of the survey, it was impossible to visitand interrogate all battle casualties in hospitals before they were discharged.However, during the third month of the survey, interrogations were accomplishedwith the aid of additional enlisted men, and a total of 434 (43 percent) of theWIA were seen. X-ray records of the majority of the remaining WIA casualtieswere examined for missile size, number, and location, and for fractures. Thehospital admission and disposition reports further served the purpose ofdetermining the time spent in hospitals by WIA casualties and theirredisposition to duty, to the Zone of Interior, or to a Detachment of Patients.

Further checking of the completeness of the casualty surveywas accomplished by a medical officer of ORS visiting each of the 40 heavybomber stations. The purpose of these visits was to verify the battle casualtystatus of aircrew members from a perusal of "Care of Flyer" reports ofthose patients missed in hospitals and to obtain information on battlecasualties whose injuries were so slight as not to require hospitalization. TheCare of Flyer reports also provided more accurate information on the finaldisposition of WIA casualties and on the time lost from flying status.

Missiles were identified from British and U.S.²ordnance publications and when necessary by consultation with a member ofthe Ordnance Office, USSTAF. Photographic records of missiles were madeperiodically by a photographer supplied by the Army Pictorial Service. It wasnot possible to have photographic equipment issued to the Medical ORS³for making photographic rec-

²USSTAF Ordnance Memorandum No.5-2, 24 Mar. 1944, subject: German Ammunition.
³It is a sad reflection upon the imaginative foresight of theresponsible officers to see this reluctance to furnish essential equipment. Oneof the most important aspects of any casualty survey is the photographic andX-ray record of the casualties. For the immediate purpose of the survey and tofurnish a permanent record for future study and teaching, photographs of theexternal wounds and internal wound track with X-rays of the body regionsinvolved are of paramount interest and value.-J. C. B.

553

ords of fatal wounds. The only such records available werethose photographs taken at the AAF stations for the USSTAF surgeon's"Body Armor Reports." These photographs were borrowed and photostaticcopies of them obtained. A 39-percent photographic coverage of KIA casualtiesfrom the 1st and 2d Divisions was achieved in this way. No photographs wereavailable of casualties sustained by the 3d Division.

Aircraft Battle Damage

Group and squadron operations offices at each of the heavybomber stations provided the identification number of each aircraft in which aKIA or a WIA battle casualty occurred, so that a report on the damage to theplane might be secured.

Permission was obtained from the Commanding General, EighthAir Force, for the Medical ORS to borrow, for photostatic copying, the battledamage reports of heavy bomber aircraft in which casualties had occurred. Thesereports were prepared by the AAF station engineers on most damaged B-17 and B-24aircraft and were forwarded to the Operational Research Section of the EighthAir Force. The serial numbers of aircraft in which casualties occurred wereobtained from every AAF station operations office and submitted to the EighthAir Force Operational Research Section. Since only one copy of these battledamage reports was prepared, it was necessary to obtain, deliver forphotostating, and return personally the reports at regular intervals. Battledamage reports pertaining to a total of 656 aircraft, in which there were 771casualties, were obtained. This represents a 70-percent coverage of aircraftbattle damage data associated with 70 percent of the casualties in the survey.Tables 177, 178, 179, and 180 give the number and types of the aircraft examinedin this way by divisions, the cause of the damage, and the number of casualtiesper aircraft.

TABLE 177.- Distributionof 541 casualties in 461 flak-damaged B-17 aircraft examined

1st Division			3d Division
Casualties in each aircraft	Aircraft examined	Total casualties	Casualties in each aircraft	Aircraft examined	Total casualties
1	221	221	1	176	176
2	33	66	2	18	36
3	4	12	3	6	18
4	2	8	4	1	4
	260	307		201	¹234

¹Two of these casualties were hit with 20 mm.cannon shell fragments.

554

TABLE 178.- Distribution of 193 casualtiesin 172 flak-damaged B-24 aircraft examined

3d Division			2d Division
Casualties in each aircraft	Aircraft examined	Total casualties	Casualties in each aircraft	Aircraft examined	Total casualties
1	33	33	1	122	122
2	1	2	2	12	24
3	1	3	3	3	9
	35	38		137	155

TABLE 179.- Distribution of 28 casualtiesin 19 examined B-17 aircraft damaged by missiles from fighter aircraft

1st Division			3d Division
Casualties in each aircraft	Aircraft examined	Total casualties	Casualties in each aircraft	Aircraft examined	Total casualties
1	10	10	1	3	3
2	2	4	2	1	2
3	3	9
	15	23		4	5

TABLE 180.- Distribution of 9 casualties in4 examined B-24 aircraft damaged by missiles from fighter aircraft

2d Division
Casualties in each aircraft	Aircraft examined	Total casualties
1	1	1
2	1	2
3	2	6
	4	9

Flight Formations

The flight formations of heavy bombers over enemy territoryon missions, during which there were battle casualties, were studied.Diagrammatic flight formations were obtained from the Operations Office,Operational Research Section, Eighth Air Force. A total of 288 complete groupflight formation plans of Eighth Air Force heavy bombers was available. Theseformations were selected on the basis that in each one there was at least one

555

casualty in one aircraft. Thus, the relationship betweencasualties and flight formations for 539 battle casualties (48 percent of thetotal sample) was observed.

ANALYSIS OF BATTLE CASUALTIES

General

During the 3 months in which this survey of battlecasualties returning to the United Kingdom was conducted, there were 1,117battle casualties of whom 110 were killed and 1,007 wounded by enemy fire. Table181 shows their distribution between the heavy bombardment groups. Thisrepresents a casualty rate of 0.172 percent (1.72 per 1,000) in terms ofman-combat missions about which casualty data were available and 0.170 percent(1.70 per 1,000) of all man-combat missions. The ratio of MIA personnel to knownbattle casualties is approximately 6 to 1 (5.85 percent).

These figures may be compared with those given by theSurgeon, USSTAF, in his "Annual Report of Health" for 1943-44. For 1year ending with the first month of the present survey, the battle casualty ratefor the Eighth Air Force is reported as 0.201 percent (2.01 per 1,000) ofman-combat missions credited. The MIA rate for the same period was 1.95 percent(19.5 per 1,000). The ratio of MIA personnel to known battle casualties was 9.7percent or

TABLE 181.- Distribution of 1,117 aircrewbattle casualties of 1st, 2d, and 3d Divisions, by heavy bombardment group

1st Division B-17's			2d Division B-24's			3d Division B-17's
Group	Men wounded	Men killed	Group	Men wounded	Men killed	Group	Men wounded	Men killed
91st	20	2	44th	23	1	34th¹	19	3
92d	49	2	93d	24	2	94th	56	5
303d	38	4	389th	17	4	95th	17	2
305th	44	7	392d	28	5	96th	32	2
306th	33	3	445th	18	---	100th	25	1
351st	29	1	446th	12	4	385th	20	1
379th	35	7	448th	23	1	388th	20	4
381st	28	4	453d	12	1	390th	33	7
384th	32	5	458th	13	1	447th	27	1
398th	27	---	466th	27	5	452d	34	4
401st	15	5	467th	15	3	486th¹	25	2
457th	36	1	489th	36	3	487th¹	9	2
			491st	19	3	490th¹	9	2
			492d	16	---	493d¹	12	---
	386	41		283	33		338	36

¹Changed from B-24 aircraft to B-17 aircraftduring July and August 1944.

556

nearly 10 to 1. Thus, the casualty rate and MIA rate for the3 months which are the subject of this report are, respectively, 15.4 and 48.7percent less than the corresponding figures for 1943-44.

In the present survey, the 1,117 battle casualties occurredin a total of 944 aircraft. Table 182 gives their frequency and distribution inthe two types of heavy bombers. Approximately 72 percent of the casualtiesoccurred with a distribution of one per aircraft. Multiple casualties peraircraft in the two types of heavy bombers did not differ significantly in theiroccurrence.

From the data in tables 176 and 182, it can be seen that theaircrew battle casualty rate is 2.10 per 1,000 man-combat missions in B-17'sand 1.26 per 1,000 in B-24's. Thus, the risk of becoming a battle casualty wasapproximately two-thirds (67 percent) greater to B-17 personnel than it was toB-24 aircrew personnel. Since this conclusion is derived from an analysis ofonly those casualties who were brought back to the United Kingdom, it cannot beassumed that the real risk rates in the two types of bombers are as representedby the cited figures. If, for example, there was a higher casualty rate inmissing B-24 personnel than in missing B-17 personnel, the figures could changesignificantly. However, with the available information, one must take account ofthe apparent difference which is very clearly significant. The reasons or causesfor the difference merit further investigation. A fuller analysis of flak hitson aircraft in which casualties occurred appears in another chapter (p. 620).

TABLE 182.- Frequencydistribution of 1,117 battle casualties, by category, in 944 heavy bombers (B-17'sand B-24's)

Number of aircraft	Casualties per aircraft	Category
Number of aircraft	Casualties per aircraft	WIA	KIA	Total casualties
B-17:
533	1	482	51	533
70	2	130	10	140
16	3	44	4	48
4	4	14	2	16
1	5	3	2	5
624		673	69	742
B-24:
277	1	256	21	277
34	2	54	14	68
8	3	20	4	24
1	6	4	2	6
320		334	41	375
944		1,007	110	1,117

557

Causes of Casualties

Table 183 gives the causes of the wounds sustained by the1,117 casualties. Approximately 86 percent of the casualties were hit by flakfragments. Less than 4 percent were hit by shells or shell fragments fired fromenemy fighter planes. Practically all of the 7.8 percent of casualties hit bysecondary missiles were the result of flak hits on the aircraft. Secondarymissiles include fragments of Plexiglas; pieces of dural from the skin of, orobjects in, the plane; bulletproof glass; brass fittings; and parts ofelectrical heating and radio equipment and .50 caliber machinegun ammunition.

Two unidentified missiles causing one KIA and one WIAcasualty were found to be pieces of commercially pure zinc, the origin of whichwas not ascertained. There were three individuals who sustained injuries duringmore than one mission. Two were struck by flak fragments on two differentmissions and one was struck by Plexiglas on one mission and by flak on another.One of the two hit twice by flak was killed as a result of the second hit.

TABLE 183.- Distributionof 1,117 aircrew battle casualties, by category and causative agent of wounds

Causative agent	Category of casualty		Total casualty
Causative agent	WIA		KIA		Number	Percent
Flak fragments	865	98	963	86.2
Shells:
20 mm.	37	7	44	3.9
13 mm.	3	1	4	.4
7.92 mm.	2	---	2	.2
Secondary missiles	84	3	87	7.8
Unknown	16	1	17	1.5
Total	1,007	110	1,117	100.0

Distribution of Casualties According to CombatPosition

Table 184 shows the frequency with which aircrew personnel in differentcombat positions became missile casualties. In this and similar tables andfigures, the positions of bombardier, togglier, and nose gunner, like those ofthe top turret gunner and engineer, are regarded as the same.

The high casualty rate for waist gunners was partially due to the fact thatheavy bombers frequently carried two waist gunners. This practice wasdiscontinued to a large extent, but accurate information as to the frequencywith which aircrews included two waist gunners during the survey was not known.

The high casualty rates for navigators and bombardiers was to be expectedfrom their positions in the nose of the aircraft. They lacked the protectionprovided by other personnel and portions of the ship's structure and by being

558

in the most forward compartments of the aircraft; theywere exposed to the greatest density of flak. The leading edges of the wingsand other parts of aircraft are known to receive a greater density of flak hitsthan trailing edges. The lowest incidence of casualties appears to occur in theball turret gunner's position. This was partially due to the fact that onlyone of the two types of aircraft (B-17) carried a man in that combat position.

TABLE 184.-Distribution of 1,117 battle casualties due toall missiles, by category and combat position

Position	WIA casualties		KIA casualties		Total casualties		Case fatality rate (percent)
Position	Number	Percent	Number	Percent	Number	Percent	Case fatality rate (percent)
Pilot	75	7.4	8	7.3	83	7.4	9.6
Copilot	68	6.8	6	5.5	74	6.6	8.1
Navigator	123	12.2	13	11.8	136	12.2	9.6
Bombardier	178	17.7	18	16.3	196	17.6	9.1
Radio operator	87	8.6	8	7.3	95	8.5	8.4
Waist gunner	212	21.1	21	19.0	233	20.9	9.0
Ball turret gunner	59	5.9	7	6.4	66	5.9	10.6
Top turret gunner	84	8.3	10	9.1	94	8.4	10.6
Tail gunner	121	12.0	19	17.3	140	12.5	13.6
Total	1,007	100.0	110	100.0	1,117	100.0	9.9

The Demarcation of Body Regions

The lack of a standardized method of demarcation of theregions of the body makes it impossible to compare accurately thedistribution of wounds in any two or more collections of casualty data. Most ofthe available information pertains only to the wounded so that the truedistribution of wounds or, more accurately, hits by bullets and fragments fromhigh explosive shells has not been recorded. The military surgeons who caredfor WIA casualties were the ones who recorded the locations of wounds. Theirrecords have been the source of material from which casualty statistics havebeen compiled, and little information was obtained from their recordsconcerning the difference between the wounds of entrance and of exit.Frequently, two or more wounds may have been produced by a single missile, butthe surgeon was more concerned about the treatment of wounds than about theeffectiveness of weapons or the development and use of body armor.Consequently, he usually failed to record the information which wouldenable those interested in body armor to compile accurate data pertaining tothe regional distribution of hits and the type of causative agent.

In the Medical and Surgical History of the War of the Rebellion,⁴ref-

⁴The Medical and Surgical History of the War of the Rebellion. Surgical History. Washington: Government Printing Office, 1883, vol. II, pt. III, p. 691.

559

erence is made to the relative amounts of superficial areapresented by the principal divisions of the human body from Longmore.⁵However, the relative percentages given for the different body regionsare at variance with more recently determined measurements, particularly thatfor the head and neck region. Longmore's figure for the head region was8.51 percent as compared with Burns and Zuckerman's figure of 12 percent.⁶In view of the fact that all complete casualty samples studied duringWorld War II show wounds of the head and neck to be even in excess of 12percent, it is likely that the measurement of Burns and Zuckerman is themore accurate one.

In this survey, the lines of demarcation between the bodyregions (fig. 275) were uniformly adhered to in accordance with those recordedby Burns and Zuckerman. The following is quoted from the British report:

There are no agreed surface markings by which one region ofthe body can be definitely demarcated from another-e.g., the thorax from theneck. No absolute demarcation is possible because a line projected from thesurface through the body would, in certain places, pass through two or moreregions. For example a shot penetrating the body horizontally just below theribs might easily pass both through the liver (an abdominal organ) and throughthe lower part of the lung (a thoracic organ). Again, shots in the buttockregion could be regarded as wounds of the lower limb as well as wounds of thepelvis-if they penetrated deep enough. It was therefore necessary to definecertain arbitrary lines of demarcation. The following were chosen as a faircompromise.

Head and Neck Region (Symbol H)

The demarcation of the head and neck region from the thoraxpresents difficulties in view of the fact that the domes of the pleurae extendan inch or so above the clavicles. A line passing transversely across the bodyan inch above the clavicles would, however, cut out a large part of the neck,and the part which lies below the cricoid cartilage. A compromise is suggestedin a line which passes transversely one inch above the upper limit of themanubrium sterni. At the back the line chosen passes immediately below the spine of the seventh cervical vertebra (vertebra prominens). This point caneasily be found as it lies immediately above the most prominent vertebra whenthe neck is slightly bent. In side view the lower limit of the neck is taken asa line which passes immediately above the line of the clavicle.

Chest Region (Symbol C)

The shoulders, both in front and behind, are most readilydemarcated from the thorax and neck by a line which joins the upper limits ofthe anterior and posterior axillary folds with the point where the vertical neckjoins the sloping shoulders.

The demarcation of the thoracic cage from the abdomenpresents difficulties. A compromise has been effected in a simple line whichapproximately demarcates the lower limit of the pleural cavities. In frontthis line is taken as passing from the lower end of the sternum obliquelydownward and laterally to the eighth intercostal space. Where this line meets theanterior end of the eighth space a horizontal line is carrieddirectly around the body, approximately across the spine of the first lumbarvertebra, to meet the anterior end of the eighth intercostal space onthe opposite side.

Consideration of the relations given in several worksonsurface anatomy suggests that this line, though not entirely accurate, is a faircompromise between the various statements that are made about the inferiormargin of the pleural cavities.

⁵Longmore, T.: Gunshot Injuries: Their History,Characteristics, Features, Complications, and General Treatment. London:Longmans, Green and Co., 1887, p. 595.
⁶Burns, B. D., and Zuckerman, S.: The Wounding Power ofSmall Bomb and Shell Fragments. R. C. No. 350 of the Research and ExperimentsDepartment of the Ministry of Home Security.

560

FIGURE 275.-Outline form with demarcations of body regions.

It does not, however, clearly demarcate the thorax from theabdomen in so far as penetrating wounds of a lower part of the thoracic cagemight very well enter the upper abdominal organs, and especially the liver.Furthermore, some shots through the subcostal angle might hit the posterior andinferior parts of the lungs.

Abdominal Region (Symbol A)

Lines chosen to demarcate the lower limits of the abdomen infront are those passing along the inguinal lines and meeting over the pubes. Atthe back the line curves from the region of the anterior superior spine, uplaterally and then downward to just below the middle of the natal cleft. Thelatter point is approximately at the level of the ischial tuberosities.

Upper Extremity Region (Symbol U)

The lines of demarcation between the upper extremityregionand head-and-neck and chest regions are as described under Chest Region.

561

Lower Extremity Region (Symbol L)

The lines of demarcation between the lower extremityandabdominal regions are as described under Abdominal Region. Any hit from thefront or back above the level of these lines would penetrate into the lower partof the abdominal cavity. They therefore provide an adequate demarcation betweenthe abdomen and lower limbs from these two aspects. Laterally, however, they aredeficient as landmarks, since a shot could smash below their level through thehip region into the pelvis.

It should be noted again that the demarcations suggestedabove represent useful and practical compromises, and not absolute anatomicallines of limitation. Even were the latter definable, it is doubtful whether theresults obtained would have been materially affected by their use, or indeed byany other set of practical lines of demarcation which might be suggested.

A great many difficulties were encountered in the analysisof extensive and multiple wounds, especially those observed in KIA casualties.The following criteria were adhered to closely:

1. Only wounds of entrance located in the region wheremissiles first struck the body were recorded as hits regardless of otherregions traversed by the missile.

2. A wound or hit which appeared to be located on a line ofdemarcation between two regions was regarded as occurring in that region inwhich the missile track extended beyond the point of entry. For example, awound at the junction of the chest and upper extremity would be a wound of thechest if the missile entered the chest or would be an upper extremity wound ifit was confined to the shoulder or other part of the upper extremity.

3. Bruises and abrasions were disregarded, but all missilewounds in which the skin's surface continuity was interrupted were recorded.

4. Although it is possible for a soldier who has been killedby fragments from a bursting projectile to be struck after death by other shellfragments, an effort was made to disregard such secondary hits when it wasobvious that they had occurred. For example, when there were wounds of entranceon opposite surfaces which could not have occurred from the burst of a singleshell, only those hits thought to be inflicted primarily were recorded.

5. When there was doubt as to which was an entrance and which an exit wound, thechoice depended arbitrarily on the missile trackbeing directed above the horizontal plane, with the man standing erect.

6. In the case of extensive mutilating wounds of one orseveral regions of the body-obviously due either to very large fragments ordirect hits by projectiles at the instant of or just before detonation-thelocation of such hits was regarded as being in the region located nearest thecenter of the area of mutilation.

7. When a mutilating wound involved an extremity and thearea of mutilation was not continuous with the mutilation of the torso, anadditional hit was recorded for the extremity, and its location was regarded asbeing at the most proximal level of traumatic amputation.

8. Wounds resulting from an injury in addition to woundscaused by missiles were disregarded unless they could have been the cause ofdeath.

562

Regional Distribution of Wounds Due to All Missiles

A distinction is made between the regional distribution andthe regional frequency of wounds. In the former, a large number of wounds,considerably in excess of the number of casualties, may be distributed over the various regions ofthe bodies in a sample of casualties. Percentages in the tables that pertain to the regional distribution of wounds refer to thetotal number of wounds. In the case of regional frequency, one is concernedonly with the frequency with which the various regions of the body arewounded regardless of the number of wounds in each body region. Percentages inthe tables which pertain to the regional frequency of wounds refer to the number of casualties.From such tabulations, casualties who sustainedwounds in more than one region of the body must either be excluded or anadditional entry made for them; that is, for those hit or wounded in more than one region (Multiple,Symbol = M). Regional frequency tabulations in this survey include an additional entry for casualties wounded in more thanone region of the body.

Table 185 shows the regional distribution of all wounds dueto all missiles in 1,115 aircrew battle casualties. Two casualties, not includedin table 185, were known to have been killed by flak before thecrashlanding of their aircraft. They were badly burned, and the location of alltheir wounds could not be determined.

TABLE 185.-Distribution of 1,553 wounds due to all missiles in 1,115 aircrew battle casualties, by category of casualty and by anatomic location (regional distribution) of wounds

Anatomic location	Wounds in WIA casualties		Wounds in KIA casualties		Total wounds
Anatomic location	Number	Percent	Number	Percent	Number	Percent
Head	258	19.8	69	27.1	327	21.1
Chest	63	4.9	37	14.5	100	6.4
Abdomen	28	2.2	23	9.0	51	3.3
Extremities:
Upper	382	29.4	65	25.5	447	28.8
Lower	567	43.7	61	23.9	628	40.4
Total	1,298	100.0	255	100.0	1,553	100.0

Onthe basis of surface area presented by the differentregions of the body, the wounds of the head and neck in all casualties (21.1percent) are more frequent than would be expected. The projected surfacearea of this region is approximately 12 percent of the entire body (table 199).On the other hand, the wounds of the chest and abdomen, whose mean projectedsurface area is approximately 27 percent of the entire body, are far below theexpected number.

563

This and other indications of the protective value of bodyarmor will be discussed in the section on flak casualties (p. 570). In table185 are included all wounds even though two or more may have been produced by asingle missile following a straight path. The regional distribution of woundsor, more accurately, of hits caused by flak fragments only is given in table200.

Regional Frequency of Wounds Due to All Missiles

Table 186 shows the regional frequency of wounds due to allmissiles in 1,117 casualties. A comparison of the total percentages in tables185 and 186 reveals the slight differences between regional distribution andregional frequency tabulations of wounds in the same complete sample of WIA andKIA casualties. If the regional frequency percentages had been calculated on thebasis of casualties wounded in single regions only and if those wounded inmultiple regions had been omitted from the sample, the differences in values bythe two methods of presentation would have been even less significant.

TABLE 186.-Distribution of 1,117 aircrew battle casualties, by category and by anatomic location (regional frequency) of wounds due to all missiles

Anatomic location	WIA casualties		KIA casualties		Total casualties
Anatomic location	Number	Percent	Number	Percent	Number	Percent
Head	182	18.1	39	35.5	221	19.8
Chest	27	2.7	11	10.0	38	3.4
Abdomen	10	1.0	7	6.4	17	1.5
Extremities:
Upper	246	24.4	1	.9	247	22.1
Lower	419	41.6	9	8.2	428	38.3
Multiple regions	123	12.2	43	39.0	166	14.9
Total	1,007	100.0	110	100.0	1,117	100.0

On the other hand, a comparison of the values fromtables 185 and 186 demonstrates the relative differences in regionaldistribution and regional frequency tabulations of wounds that exist whensamples of casualty data are broken down into WIA only and KIA only. Thedifferences are only very slight in the case of the wounded but are very markedand of a different order entirely in the case of the killed. The reason forthese differences is that the killed were much more frequently struck inmultiple regions or rather that those casualties with multiple wounds were much more apt to die. Thus,head-and-neck and multiple regions were woundedin 74.5 percent of the KIA as compared with 30.3 percent in WIA casualties.

564

Fatality Rates

A survey of ground force casualties at Bougainville, S.I.,from 15 February to 21 April 1944 (p. 281), was the first of its kind preparedby U.S. Army medical personnel that included an evaluation of wounds in bothfatal and nonfatal American battle casualties. The overall case fatality rateamong the ground force casualties in this sample was approximately twice asgreat as in the casualties of Eighth Air Force personnel. Probably, the mainreason for this difference is the speed with which the air force casualtiesreceived adequate medical care. Table 187 compares the regional frequency ofwounds and the case fatality rates in the Eighth Air Force in Europe andin the ground forces at Bougainville.

The marked difference in fatality rates in wounds of the headand neck (17.6 percent for Eighth Air Force casualties as compared with 37.5percent at Bougainville) is at least partly due to the fact that a large numberof the wounds of the head in Eighth Air Force casualties were mild lacerationsdue to Plexiglas fragments. Casualties due to flak fragments only in the EighthAir Force and the casualties at Bougainville are compared in table 202.

TABLE 187.-Case fatality rates in the 1,117 battlecasualties of the Eighth Air Force and in the 1,788 casualties of the ground forces at Bougainville

Anatomic location	Eighth Air Force		Ground forces at Bougainville
Anatomic location	Number of casualties	Percent KIA	Number of casualties	Percent KIA
Head	221	17.6	384	37.5
Chest	38	28.9	231	37.6
Abdomen	17	41.1	114	42.1
Extremities:
Upper	247	.4	320	.3
Lower	428	2.1	407	3.4
Multiple regions	166	25.9	332	30.4
Total	1,117	9.8	1,788	22.1

Single and Multiple Wounds

Table 188 gives the incidence of wounds occurring in one ormore than one region of the body. The head and neck are regarded as one region.Wounds of both upper or both lower extremities are regarded as occurring in oneregion, whereas wounds in one or both upper and one or both lower extremitiesare regarded as occurring in two regions. Thus, 39.1 percent of the KIAcasualties as compared with 12.2 percent of the WIA casualties sustained woundsin more than one region.

565

TABLE 188.-Distribution of 1,117 battle casualties due to all missiles, by category and by number of regions wounded

Regions wounded	WIA casualties		KIA casualties		Total casualties
Regions wounded	Number	Percent	Number	Percent	Number	Percent
Number:
1	884	87.8	67	60.9	951	85.1
2	97	9.6	29	26.4	126	11.3
3	25	2.5	9	8.2	34	3.0
4	1	.1	3	2.7	4	.4
5	---	.0	2	1.8	2	.2
Total	1,007	100.0	110	100.0	1,117	100.0

Fractures and Amputations

The regional distribution of fractures due to all missiles is shown in table189. In this analysis, fractures of both bones of the leg or forearm and ofany number of bones of the shoulder, hip, wrist, or ankle joints and of the handor foot or of the cranium are regarded as single fractures when produced at thesame instant by one missile. When it was apparent that one missile had producedfractures in more than one region or, for example, when one missile had made anentrance and an exit wound of the same region, such as the chest or skull, andhad fractured skeletal structures at both wounds, two fractures were counted.Fractures due to all missiles occurred in 31.8 percent of the casualties (26.3percent of the wounded and 85.3 percent of the killed). In the study offractures, the total 1,007 WIA casualties were available for study.

TABLE 189.- Regional distribution of fracture wounds in 1,109 aircrew battle casualties due to all missiles

Region	1,007 WIA casualties (265 or 26.3 percent with fractures)		102 KIA casualties¹ (87 or 85.3 percent with fractures)		Total casualties² (31.8 percent with fractures)
Region	Number of fractures	Percent	Number of fractures	Percent	Number of fractures	Percent
Head	28	10.0	48	45.7	76	19.8
Chest	12	4.3	22	21.0	34	8.9
Abdomen	---	.0	1	1.0	1	.3
Extremities:
Upper	111	39.8	16	15.2	127	33.0
Lower	128	45.9	18	17.1	146	38.0
Total	279	100.0	105	100.0	384	100.0

¹Of the total 110 KIA casualties, only 102 were available for study in regard to fracture occurrence.²Of the total 1,117 WIA and KIA casualties, 1,109 were available for study in regard to fracture occurrence.

566

However, only 102 of the 110 KIA casualties were included in the fracture survey. This explains the variation in the figures for KIAcasualties and wounds and total casualties as seen in tables 185, 189, and 190.

The relationship between wounds and fractures is given intable 190. Thus, for all wounds, the incidence of fractures was 24.8 percent.In the WIA casualties, 21.5 percent of the wounds were complicated by fractures,and in the KIA, 42.3 percent of the wounds were complicated by fractures.(Compare with table 189.)

Amputations are included as fractures in tables 189 and 190and are shown separately in table 191. There was only one instance of asurgical amputation necessary following a soft-tissue wound. The injury was athrough-and-through wound of the soft tissues anterior to the left femur,necessitating amputation of the left thigh. This amputation is not included intable 191.

TABLE 190.-Relationship between wounds and fractures in1,109 aircrew battle casualties dueto all missiles, by anatomic location

Anatomic location	WIA casualties			KIA casualties¹			WIA and KIA casualties
Anatomic location	Number of wounds	Number of fractures	Percent	Number of wounds	Number of fractures	Percent	Number of wounds	Number of fractures	Percent
Head	258	28	10.9	67	48	71.6	325	76	23.4
Chest	63	12	19.0	34	22	64.7	97	34	35.1
Abdomen	28	---	.0	23	1	4.3	51	1	.2
Extremities:
Upper	382	111	29.1	64	16	25.0	446	127	28.5
Lower	567	128	22.6	60	18	30.0	627	146	23.3
Total	1,298	279	21.5	248	105	42.3	1,546	384	24.8

¹Data are for 102 of the total 110 KIA casualties.

TABLE 191.-Distribution of 34 traumaticamputations in 32 battle casualties, by anatomic location

Anatomic location	Amputations
Anatomic location	In WIA casualties	In KIA casualties	Total
	Number	Number	Number
Thigh	6	6	12
Leg	6	---	6
Foot	5	1	6
Hand	6	---	6
Arm	---	4	4
Total	23	11	34

567

Of the WIA battle casualties, 23 (2.3 percent) sustainedtraumatic amputations; 20 of these amputations were due to flak. Of theremaining three, two were due to 20 mm. shells. The missile responsible for theamputation in the third case was not discovered. Two casualties had twoamputations, one of both thighs and the other of one thigh and one arm. Inthe KIA group, all but one arm amputation, for which a 20 mm. cannon shell wasresponsible, were due to flak.

Return to Duty Status of Aircrew Battle Casualties

The severity of wounds sustained by aircrew battle casualtiesmay be evaluated on the basis of time lost from flying status. For thisanalysis, casualties are regarded as KIA, permanently grounded, or grounded forperiods of less than 24 hours, for 1 to 7 days, for 7 to 30 days, or for 30 to90 days. The period of observation after injury was limited to 90 days. Thus,any casualty still in hospital or who had not returned to his organizationafter 90 days was regarded as permanently grounded. Tables 192 through 197 showthe relative severity of wounds of different regions on the basis of time lostfrom flying status.

By regarding KIA casualties and those permanently groundedtogether, it is noted that 378 (33.8 percent) of the 1,117 battle casualtieswere permanently lost from flying status. Of the 739 casualties who returned toair-combat duty 99 (13.4 percent) lost less than a day; 256 (34.6 percent), lessthan a week; and

TABLE 192.-Distribution of 221 aircrewbattle casualtieswith wounds of the head and neck, due to all missiles, by disposition

Disposition	Missile causing casualty						Total casualties
	Flak		From fighter aircraft		Secondary or unknown		Total casualties
	Number of casualties	Percent	Number of casualties	Percent	Number of casualties	Percent	Number	Percent
Returned to flying status after-
0-1 day	11	8.3	---	0.0	27	32.1	38	17.2
1-7 days	21	15.9	3	60.0	36	42.9	60	27.1
7-30 days	32	24.2	---	.0	12	14.3	44	19.9
30-90 days	10	7.6	---	.0	---	---	10	4.6
Total	74	56.0	3	60.0	75	89.3	152	68.8
Permanently grounded	22	16.6	---	.0	8	9.5	30	13.6
Killed in action	36	27.4	2	40.0	1	1.2	39	17.6
Total	58	44.0	2	40.0	9	10.7	69	31.2
Grand total	132	100.0	5	100.0	84	100.0	221	100.0

568

563 (76.1 percent), less than a month from flying status.Wounds of the abdomen and of more than one body region accounted for thegreatest relative loss of men from air-combat duty, the rates being 55.6 and54.5 percent, respectively. Of those whose wounds were confined to theupper extremity, only 25.5 percent were permanently lost to air-combat duty.

TABLE 193.-Distribution of 38 aircrew battle casualties with chest wounds due to flak fragments,¹ by disposition

Disposition	Flak fragments
Disposition	Number of casualties	Percent of casualties
Returned to flying status after-
0-1 day	2	5.2
1-7 days	5	13.2
7-30 days	8	21.1
30-90 days	6	15.8
Total	21	55.3
Permanently grounded	6	15.8
Killed in action	11	28.9
Total	38	100.0

¹During the survey period, flak fragments were responsible for all chest wounds.

TABLE 194.-Distribution of 18 aircrew battle casualties with abdominal wounds due to all missiles, by disposition

Disposition	Missile causing casualty						Total casualties
	Flak		From fighter aircraft		Secondary or unknown		Total casualties
	Number of casualties	Percent	Number of casualties	Percent	Number of casualties	Percent	Number	Percent
Returned to flying status after-
0-1 day	---	0	---	0	---	0	---	0
1-7 days	4	26.7	---	0	---	0	4	22.2
7-30 days	3	20.0	---	0	---	0	3	16.7
30-90 days	1	6.6	---	0	---	0	1	5.5
Total	8	53.3	---	0	---	0	8	44.4
Permanently grounded	3	20.0	---	0	---	0	3	16.7
Killed in action	4	26.7	2	100.0	1	100.0	7	38.9
Total	15	100.0	2	100.0	1	100.0	18	100.0

569

TABLE 195.- Distribution of 243 aircrew battle casualties with upper extremity wounds due to all missiles, by disposition

Disposition	Missile causing casualty							Total casualties
	Flak		From fighter aircraft		Secondary or unknown			Total casualties
	Number of casualties	Percent	Number of casualties	Percent	Number of casualties	Percent	Number	Percent
Returned to flying status after-
0-1 day	23	9.7	---	0	2	66.7	25	10.3
1-7 days	26	11.0	---	0	---	0	26	10.7
7-30 days	87	36.9	3	75.0	1	33.3	91	37.5
30-90 days	38	16.1	1	25.0	---	0	39	16.0
Total	174	73.7	4	100.0	3	100.0	181	74.5
Permanently grounded	61	25.9	---	0	---	0	61	25.1
Killed in action	1	.4	---	0	---	0	1	.4
Total	236	100.0	4	100.0	3	100.0	243	100.0

TABLE 196.-Distribution of 421 aircrew battle casualties with lower extremity wounds due to all missiles, by disposition

Disposition	Missile causing casualty							Total casualties
	Flak		From fighter aircraft		Secondary or unknown			Total casualties
	Number of casualties	Percent	Number of casualties	Percent	Number of casualties	Percent	Number	Percent
Returned to flying status after-
0-1 day	30	7.6	1	5.5	---	0	31	7.4
1-7 days	47	11.8	2	11.1	---	0	49	11.6
7-30 days	105	26.5	8	44.5	1	16.7	114	27.1
30-90 days	99	24.9	2	11.1	2	33.3	103	24.5
Total	281	70.8	13	72.2	3	50.0	297	70.6
Permanently grounded	107	27.0	5	27.8	3	50.0	115	27.3
Killed in action	9	2.2	---	0	---	0	9	2.1
Total	397	100.0	18	100.0	6	100.0	421	100.0

570

TABLE 197.-Distribution of 176 aircrew battle casualties with multiple wounds due to all missiles, by disposition

Disposition	Missile causing casualty							Total casualties
	Flak		From fighter aircraft		Secondary or unknown			Total casualties
	Number of casualties	Percent	Number of casualties	Percent	Number of casualties	Percent	Number	Percent
Returned to flying status after-
0-1 day	3	2.1	---	0	---	0	3	1.7
1-7 days	9	6.2	3	14.3	1	10.0	13	7.4
7-30 days	35	24.2	9	42.9	3	30.0	47	26.7
30-90 days	15	10.3	1	4.8	1	10.0	17	9.7
Total	62	42.8	13	62.0	5	50.0	80	45.5
Permanently grounded	46	31.7	4	19.0	3	30.0	53	30.1
Killed in action	37	25.5	4	19.0	2	20.0	43	24.4
Total	145	100.0	21	100.0	10	100.0	176	100.0

Blast Injury

Otoscopic examination of eardrums failed to reveal any caseof blast injury in any of the 434 WIA battle casualties who were interrogated.The only instances of ruptured membrana tympani in the 89 KIA casualtiesexamined were those that were torn as a result of basal skull fractures. Itis known that rupture of the eardrums occurs at very much lower blast pressurethan does lung damage and the absence of the former probably precludes theoccurrence of any blast injury of the lungs.

CASUALTIES DUE TO FLAK

Distribution of Flak Casualties According to Combat Position

Because of the relatively large proportion (86.2 percent)of casualties due to flak, it was thought desirable to analyze them separately.Data pertaining to protection by body armor, altitude at which injurieswere sustained, time interval between injury and adequate surgical treatment,time lost from flying status, sizes of fragments causing wounds, and therelative vulnerability of different parts of the body in different aircrewcombat positions, will be discussed in this section. The frequency with whichaircrew personnel in different combat positions became casualties dueto flak is shown in table 198.

The fact that heavy bombers occasionally carried two waistgunners probably accounts for the highest flak casualty rate for thatcombat position

571

TABLE 198.-Distribution of 963 casualties due to flak, by category and by combat position in heavy bombers

Position	WIA casualties		KIA casualties		Total casualties		Case fatality rate (percent)
Position	Number	Percent	Number	Percent	Number	Percent	Case fatality rate (percent)
Pilot	67	7.7	6	6.1	73	7.7	8.2
Copilot	58	6.7	6	6.1	64	6.6	9.4
Navigator	114	13.2	13	13.2	127	13.2	10.2
Bombardier	136	15.7	16	16.4	152	15.8	10.5
Radio operator	78	9.0	6	6.1	84	8.7	7.1
Waist gunner	191	22.2	18	18.5	209	21.6	8.6
Ball turret gunner	46	5.3	7	7.1	53	5.5	13.2
Top turret gunner	71	8.2	9	9.2	80	8.3	11.3
Tail gunner	104	12.0	17	17.3	121	12.6	14.0
Total	865	100.0	98	100.0	963	100.0	10.1

Also, the reasons for the relatively high casualty ratesfor bombardiers and navigators and the low casualty rate for ball turret gunnershave been discussed previously.

Projected Body Surface Area

It was to be expected and it had been observed that highexplosive shell fragments hit the body more at random than the"aimed" fire of bullets. Thus, it was to be expected that an analysisof wound distribution in a complete sample of WIA and KIA casualties due onlyto flak fragments might best reveal information pertaining to the relativedegree of protection or lack of protection to the various body regions. Inorder to determine the mean projected area of the body and to make a correctestimate of the proportions of its different parts, it was necessary to weightobservations according to the probable frequency of different positions of thebody in actual operations. Unfortunately, there was no information on which to basean estimate of the correct weighting values. An arbitrarymean figure was obtained for the present study from the three views of thestanding and kneeling figures and from the photograph taken from the front ofthe prone position (Burns and Zuckerman). By including the two other views ofthe prone position, the average value derived for the size of the human targetmay be slightly greater than the true mean projected service position. It ishardly likely that the error is as much as 5 percent. It should be noted thatvariations in the weighting factor would have a far greater influence onestimates of the mean projected area of the body and its parts than wouldalterations in the lines of regional demarcation discussed earlier. Seven menwere measured, and, in spite of the differences in their size, the measurementsshowed a remarkable

572

similarity in the proportions of the mean projected area ofeach part of the body. Their heights and body weights were as follows:

Subject:	Height (inches)	Body weight (pounds)
1	72	144
2	72	182
3	72	146
4	70?	168
5	70	182
6	69	140
7	67	126

Table 199 gives the mean smoothed values for the actualprojected surface areas in square feet and percent as determined frommeasurements of subjects 1, 2, 3, and 4. These percentage values may be regardedas the relative proportion of hits or wounds expected to be present in thevarious regions of the body in a random complete sample of casualties due onlyto high explosive shell fragments. Less than the expected number of woundsobserved in any region would be a reflection of the protection of that region,while more than the expected number of wounds observed in any region would bedue to a lack of protection to that region.

TABLE 199.-Mean projected area of body

Region	Square feet	Percent
Head and neck	0.50	12
Chest	.67	16
Abdomen	.46	11
Extremities:
Upper	.92	22
Lower	1.65	39
Total	4.20	100

The Regional Distribution of Hits Due to Flak

Table 200 shows the location of 1,222 flak hits sustained by961 battle casualties. Table 200 also shows the relationship between the woundsexpected and the wounds observed for each of the body regions on the basis ofthe projected surface area of each of the regions. The lower incidence of woundsin the thoracic and abdominal regions protected by body armor is quite marked.Secondary wounds due to flak fragments traversing more than one region of thebody were not counted here but were included in table 185, which lists allwounds due to all missiles. The rather noticeable decrease in the incidence ofwounds of the head and neck in the flak casualties (15.6 percent) as comparedwith those due to all missiles (21.1 percent) was due to the frequency ofPlexiglas wounds in the unprotected area of the head in bombardiers andnavigators.

573

Surgeons' records frequently gave diagnoses of wounds ofthe face or head due to "Flak," when in reality interrogation of thecasualties revealed that fragments of flak had penetrated Plexiglas coveredareas of the noses of aircraft, dispersing myriads of Plexiglas fragments. Thelatter, however, seldom caused wounds of any part of the body other than theeyes, the circumorbital regions, and the neck. These wounds were usually mildand caused very little loss of time from flying status. The instances in whichPlexiglas was found in other parts of the body usually occurred when fragmentslodged in soft tissues after being driven ahead or along the track of the shellor other metallic fragments.

TABLE 200.-Distribution of 1,222 flak hits on 961 aircrew battle casualties, by anatomic location

Anatomic location	Hits in WIA casualties		Hits in KIA casualties		Total hits		Wounds expected (body surface area) (percent)
Anatomic location	Number	Percent	Number	Percent	Number	Percent	Wounds expected (body surface area) (percent)
Head	137	13.1	54	30.3	191	15.6	12
Chest	51	4.9	26	14.6	77	6.3	16
Abdomen	19	1.8	10	5.6	29	2.4	11
Extremities:
Upper	330	31.6	52	29.3	382	31.3	22
Lower	507	48.6	36	20.2	543	44.4	39
Total	1,044	100.0	178	100.0	1,222	100.0	100

The Regional Frequency of Hits Due to Flak

Table 201 shows the regional frequency of flak hits sustained by 963 flak casualties. Again, the marked differences in the regional frequency of wounds compared with KIA casualties are shown in this table.

TABLE 201.-Distribution of 963 casualties, by category and by anatomic location (regional frequency) of flak hits

Anatomic location	WIA casualties		KIA casualties		Total casualties
Anatomic location	Number	Percent	Number	Percent	Number	Percent
Head	96	11.1	36	36.7	132	13.8
Chest	27	3.1	11	11.2	38	3.9
Abdomen	11	1.2	4	4.1	15	1.6
Extremities:
Upper	235	27.2	1	1.0	236	24.5
Lower	388	44.9	9	9.2	397	41.1
Multiple regions	108	12.5	37	37.8	145	15.1
Total	865	100.0	98	100.0	963	100.0

574

Fatality Rates

A comparison of fatality rates is shown in table 202. Thefirst two columns compare the case fatality rates of wounds due to all missileswith those due to flak only in aircrew personnel. The larger percentage offatal wounds of the head due to flak (27.4 percent as compared with 17.6 percentfor all missiles) is explained by the relative mildness of Plexiglas wounds ofthe face which are included in the first column of the table. The lower fatalityrate for flak wounds of the abdomen (26.7 percent as compared with 41.1 percentfor all missiles) is explained by the severity of abdominal wounds due tomissiles from enemy fighter aircraft. The third column shows the case fatalityrates for ground force casualties at Bougainville. The higher case fatalityrates which occurred in every region of the body in the series of Bougainvillecasualties must clearly be due in a large part to the fact that the woundssustained by the ground forces concerned were more severe than the

TABLE 202.-Comparison of case fatality rates ofwounds due to all missiles with those due to flak in Eighth Air Force and in theground forces in Bougainville, by anatomic location of wounds

Anatomic location	Eighth Air Force		Ground forces at Bougainville (all missiles)
Anatomic location	All missiles	Flak	Ground forces at Bougainville (all missiles)
	Percent	Percent	Percent
Head	17.6	27.4	37.5
Chest	28.9	28.9	37.6
Abdomen	41.1	26.7	42.1
Extremities:
Upper	.4	.4	.3
Lower	2.1	2.2	3.4
Multiple regions	25.9	25.5	30.4
Average	9.8	10.2	22.1

flak injuries received by the aircrews, which are thesubject of the present report. It is known too that in jungle warfareas fought at Bougainville there was a preponderance of small arms or "aimed"fire, and it is known that a bullet is relatively much more lethal than ashell fragment. Another possible reason for the difference in case fatalityrates in the Ground Forces and the Air Forces may be the greater speed withwhich air force casualties received surgical treatment. (See appendix H, page843, for a detailed comparison of World War II missile casualty data.)

In general, the number of wounding missiles per casualty inthe Air Forces was lower than in the ground force casualties. A possibleexplanation may

575

be the proximity of ground force casualties to bursts ofexploding projectiles and to a greater degree of fragmentation of mortar andartillery shells as compared with antiaircraft shells.

Altitudes at Which Casualties Sustained Wounds

The altitude at which 441 (386 WIA and 55 KIA) casualties dueto flak sustained wounds was known. Table 203 shows the manner in which thecasualties were distributed between the two types of heavy bombers.Approximately 70 percent of casualties due to flak in B-17's were wounded atan altitude of 24,000 feet or above, whereas 92 percent in B-24's werewounded at 23,000 feet or below. This difference in altitude may in some wayaccount for the difference in casualty rates in the two types of aircraft.

TABLE 203.-Altitude of B-17 and B-24 aircraft at which 441 casualties due to flak sustained wounds

Altitude in feet	Casualties in B-17's			Casualties in B-24's
Altitude in feet	WIA	KIA	Total	WIA	KIA	Total
Above 26,000	42	3	45	---	---	---
24,000-26,000	149	18	167	8	3	11
22,000-24,000	35	2	37	45	9	54
20,000-22,000	26	5	31	34	---	34
18,000-20,000	2	3	5	12	1	13
Below 18,000	12	9	21	21	2	23
Total	266	40	306	120	15	135

Time Interval Between Injury and Surgical Treatment

The time interval between injury and adequate surgicaltreatment in hospital was recorded for 375 WIA casualties due to flak. Table204 shows the period of time which elapsed between injury and surgical treatmentfor casualties in the two types of aircraft. Approximately 90 percent of all WIAbattle casualties were adequately treated in hospital within 4 hours after theywere wounded.

Return to Flying Status

The relative severity of wounds due to all missiles, asjudged by time lost from flying status, was shown in tables 192, 193, 194, 195,196, and 197. Table 205 shows the time lost from flying status by the 963casualties due to flak in the two types of aircraft. Thus, 64.3 percent of allcasualties due to flak were returned to flying status within 3 months afterbeing wounded.

576

TABLE 204.-Distribution of 375 WIA casualties due to flak in B-17 and B-24 aircraft, by time interval between injury and surgical treatment

Time interval	Number of casualties
Time interval	In B-17's	In B-24's	Total
Hours:
1	18	24	42
2	52	37	89
3	113	29	142
4	50	11	61
5	23	6	29
6	2	2	4
More than 6	6	2	8
Total	264	111	375

TABLE 205.-Distribution of 634 and 329casualties due to flak in B-17 and B-24 aircraft, respectively, by disposition

Disposition	B-17		B-24
Disposition	Number of casualties	Percent	Numbers of casualties	Percent
Returned to flying status after loss of-
0-1 day	56	8.8	13	4.0
1-7 days	69	10.9	43	13.3
7-30 days	184	28.9	86	26.5
30-90 days	108	17.0	61	18.8
Total	417	65.6	203	62.6
Permanently grounded	157	24.7	88	27.2
Killed in action	60	9.7	38	10.2
Total	217	34.4	126	37.4
Grand Total	634	100.0	329	100.0

Relationship Between Flak Fragments and Disability

If the purpose of wounding enemy personnel is to causemilitary loss, then it is apparent that some means must be devised forevaluating that loss on the basis of severity of the wounds. Lamport⁷ hasstated: "If the tactical value of causing a casualty is considered asdirectly proportional to the days lost from full service, an incongruous resultarises with a single casualty causing

⁷Lamport, H.: Missile Casualties Report No. 15, Office ofScientific Research and Development, Washington, D.C., September 1945.

577

100 days lost from duty being presumably equivalent in themilitary sense to putting each of ten men out of action for 10 days."Lamport has developed two hypotheses which may be used to demonstrate gradationof disability produced by wounds. For both of his methods, it was necessary tochoose some period of disability, in days lost from active duty, that wouldcorrespond to the total military loss of a man. By the means described in hisreport, he arrived at the conclusion that a wound causing a man to lose 45 daysamounts to a 100 percent military loss for that man and that a wound causing 6days' loss amounts to a 50 percent military loss. Figure 276 is the curverepresenting Lamport's second hypothesis and shows the relationship betweendays lost from active duty and the percent tactical military loss.

The severity of wounds has been evaluated in terms ofmilitary losses ranging from 1 to 100 percent, and the values for these losseshave been interpolated from a table which is contained in Lamport's report andreproduced here (table 206).

TABLE 206.-Relationship between days lost fromactive duty by a casualty and the resulting military loss

Days lost (T) (number)	Military loss (L)¹ (percent)	Grouping (percent)	Days lost (T) (number)	Military loss (L)¹ (percent)	Grouping (percent)
0	0	0	21	91	91-95
1	11	1-10	22	92
2	21	11-20	23	93
3	29?	21-30	24	94
4	37	31-40	25	95
5	44	41-50	26	95
6	50	51-60	27	96	96-100
7	56	51-60	28	96
8	61	61-70	29	97
9	65		30	97
10	69		31	97
11	72	71-80	32	98
12	75		33	98
13	78		34	98
14	80		35	98
15	83	81-90	36	98
16	84?		37	99
17	86		38	99
18	88		39	99.0
19	89		40	99.1
20	90		41	99.2
			42	99.3
			43	99.4
			44	99.4
			45	99.45

¹Maximal loss (L) is 100 percent forT=45days, when T=days lost from active duty. Method II (Annuity law) L=100- 100 x 0.5066(T) / (6).

578

FIGURE 276.-Lamport's curve II, showing relationship between days lost from active duty and percent tactical military loss.

There were 376 instances where the complete flak fragmentscausing the wound were recovered. Added to these were casualties withthrough-and throughfatal wounds with either no fragment or only part of a fragment retained, fatalavulsions, amputations, and decapitations. For reasons to be given later, all ofthe latter fatalities were regarded as being due to fragments heavier than 20grams. The total sample of data numbers 443 observations. These include theslightly and severely wounded, as well as those who were permanently disabled orkilled. In calculating the correlation coefficient, the two variables taken intoaccount were the days lost from flying status, inter-

579

FIGURE 277.-Chartshowing relationship between "military loss" and weight of flakfragments causing wounds in aircrew personnel.

polated from Lamport's table into percent military loss,and the size of the fragment in grams.

For the purpose of analysis, all casualties who were lost toflying status longer than 26 days, or who were permanently disabled or killed,were grouped together with those who lost from 27 to 45 days, or in terms ofmilitary loss, from 95 to 100 percent. The correlation coefficient for the twovariables was found to be 0.288?0.047. The test of significance (t) forthe coefficient was found to be 6.13 (P=less than 0.01). The mean value formilitary loss per casualty was 90.0 percent (which corresponds to a loss of 21days) and the mean fragment weight was 10.07 grams. A regression equation wascalculated which was found to be:

x=0.44y+86.5

where x=percent military loss and y=fragmentweight in grams. It may be observed, for example, from this equation thatfragments weighing 1 gram generally may be expected to produce casualties, theaverage of which may be regarded as a military loss of 87 percent. From table206, this is seen to

580

correspond to the loss of about 17 days. Figure 277 shows therelationship between "military loss" and fragment weight in the formof a curve plotted for the 443 observations upon which this report is based.

Despite all of the other variables that must be present, theobserved correlation coefficient as calculated is statistically significant andmay be regarded as real. It should be pointed out, as can be seen from the tableof values for the two variables x and y, that fragments weighingmore than 20 gm. produced the greatest number of permanently disabled or killedcasualties. It may be assumed that fragments weighing more than 20 gm. wereprobably so damaging to personnel, as well as to aircraft, that they wereresponsible for casualties in aircraft that were shot down in enemy territory,and thus the casualties could not be included in the sample of data undersurvey. Otherwise, the number of observations in this group might have beengreater still.

Correlation Between Wound Size and Fragment Size

It seems advisable to examine the relationship between the sizeof a wound and the size of the fragment causing it. It should be pointed outagain that the lack of available information on the velocity of fragmentsprevents taking that important variable into consideration here. Accuratesurface areas of wounds were not determined in most instances because of theirregular outline of the area margins. Instead, the surface area of a wound hasbeen arbitrarily regarded as the product of two dimensions in centimeters. Therewere 75 instances in KIA flak casualties in which both the area of the wound ofentrance and the fragment size were fairly accurately determined. In all of theobservations, the flak fragments were completely retained along the wound track.The wounds ranged in size from 1 cm.²to50 cm.². The flak fragments ranged from 1 to100 gm. in weight and from 1 to 9 cm. in their greatest dimension.

Correlation coefficients were calculated to show therelationship between (1) wound size and fragment weight and (2) wound size andgreatest dimension of fragment. In the case of the former, the correlationcoefficient was found to be 0.49?0.12 (t=4.1; P=less than 0.01).This correlation is highly significant. The degree of correlation between thesize of the wound and the greatest dimension of the flak fragment is evengreater as shown by the correlation coefficient of 0.63?0.12 (t=5.3; P=lessthan 0.01).

A further correlation coefficient was calculated; namely,that for the area of the wound against the product of the weight and maximumdimension of fragment. Although it was found to be significant, it was less sothan either of the coefficients just given.

For this purpose, 36 fatal wounds due to flak fragments wereavailable for study. Decapitations, avulsions, and amputations, obviously due tovery large flak fragments, were deleted. Of the 36 fatal wounds, 6 retained partof the fragment along the wound track. The identity of the missile causing thewound in the other cases was confirmed by the knowledge that it was flak

581

that had damaged the aircraft in which the casualty occurred.The range of sizes in the fatal wounds in this group was from 1 cm.²to 108 cm.². The mean wound size was 20cm.², for this sample. The difference betweenthe means of the sizes of through-and-through wounds and the sizes of woundswith retained fragments is not statistically significant. However, if only thosewounds due to retained fragments weighing less than 20 gm. (54 observations) arecompared with through-and-through fatal wounds, the difference in theirrespective mean sizes is significant-the difference being 13.80 cm.²(standard error?5.14). Therefore, through-and-through fatal wounds in thisseries of observations may be regarded generally as being caused by fragmentsweighing more than 20 grams.

Sizes of Fragments Causing Wounds

The sizes of flak fragments responsible for wounds weredetermined by weighing those recovered from the dead and estimating the weightsof others from their X-ray silhouettes. In the case of the latter, the fragmentswere estimated in grams from their linear dimensions. A large series of X-raysof fragments of known weight were available as a standard. A total of 505 flakfragments seen in X-ray films or recovered from 438 (361 WIA and 77 KIA)casualties were available for study.

The KIA casualties from whom flak fragments were availablewere only those examined in the Medical ORS laboratory during the 6 months'period, June through November 1944. There were 164 bodies examined from theEighth Air Force and Ninth Air Force and Troop Carrier Command. Of the total,144 (87.8 percent) were flak casualties. The 81 fragments causing fatal woundsin 77 casualties represent the recovery of 54.0 percent of flak fragmentscausing fatal wounds. Although some flak fragments were recovered from the otherKIA casualties, they were not included in this analysis because there wasevidence that the fragments found were only portions of the fragments causingthe fatal wounds. In several instances, fragments smaller than expected werefound along a missile track having both entrance and exit wounds.

In general, it may be assumed that those fragments which werecompletely recovered were of lower velocity and of smaller size than those whichcaused fatal wounds in the remaining 67 (46.5 percent) KIA casualties, from whomnone or only partial fragments were recovered. In instances where more than onefragment was found in a fatal wound with one point of entrance, it was assumedthat refragmentation of the primary fragment had occurred. In nearly all suchcases, the refragmented fragments could be fitted together. The weight creditedfor such multiple fragments was the total weight of the pieces. Fragments whichcaused secondary wounds in KIA casualties are included in the group of fragmentscausing nonfatal wounds. Table 207 shows the weight distribution of flakfragments according to nonfatal and fatal wounds.

In three KIA casualties, there were two fragments, both ofwhich caused fatal wounds. Four fragments are credited with having caused fatalwounds,

582

TABLE 207.-Weightdistribution of 505 flak fragments recovered from nonfatal and fatalwoundsin 438 casualties

Type of wound	Number of casualties	Number and weight of fragments						Total fragments
Type of wound	Number of casualties	0-50	50-250	?-1	1-5	5-20	>20	Total fragments
		Mg.	Mg.	Gm.	Gm.	Gm.	Gm.	Number
Nonfatal	361	56	64	46	100	100	58	424
Fatal	77	---	---	---	6	43	32	81
Total	438	56	64	46	106	143	90	505

although they were recovered from the extremities. Two ofthese were recovered from the knee joint of one casualty and weighed 15.29 and6.10 gm., respectively. The other two were removed from the thighs of twocasualties and weighed 12.04 and 31.74 gm., respectively. In each of thesecases, the actual cause of death was attributed to hemorrhage, shock, andanoxia.

A series of flak fragments of varying sizes are seen infigure 278.

Sizes of Fragments Causing Fracture Wounds

Thirty flak fragments were recovered which had caused thirtyfracture wounds of the skull. Twenty-three (77 percent) of these produced fatalwounds. The 30 fragments were distributed according to weight as shown in table208. Again, it should be stated that many fatal skull fracture wounds wereobserved from which no fragments were recovered. Thus, it may be concluded thatflak fragments weighing more than 5 gm. are much more likely to cause fatalpenetrating wounds of the skull than fragments weighing less than 5 grams.

TABLE 208.-Weightdistribution of 30 flak fragments recovered from nonfatal and fatal fracturewounds of the skull

Type of wound	Number and weight of fragments			Total fragments
Type of wound	1-5	5-20	20	Total fragments
	Gm.	Gm.	Gm.	Number
Nonfatal	3	3	1	7
Fatal	1	16	6	23
Total	4	19	7	30

Fifty-six flak fragments were recovered which had caused fracture wounds ofthe extremities. The distribution of these fracture wounds is shown in table209. Only complete fractures of the bones or joints listed, in which the mis-

583

FIGURE 278.-Primarymissiles (flak). Each of these fragments caused a fatal heavy bomber aircrewcasualty. They range in weight from 1 to 106 gm. Some of them are broken-offretained portions of larger fragments that produced fatal through-and-throughwounds. It was observed that no flak fragment weighing less than 1 gm. had beenfound to produce a fatal wound and that probably no fragment weighing less than20 gm. was capable of producing a through-and-through fatal wound.

584

siles were retained at the site of the fracture, were included. Thus again,in general, only fragments weighing more than 5 gm. produced fatal fracturewounds of the extremities.

TABLE 209.-Weightdistribution of 56 flak fragments recovered from fracture wounds of the extremities

Fracture wound of-	Weight and number of fragments					Total fragments
Fracture wound of-	50-250	?-1	1-5	5-20	>20	Total fragments
	Mg.	Gm.	Gm.	Gm.	Gm.	Number
Humerus	---	---	---	5	1	6
Elbow	---	1	1	2	1	5
Radius and/or ulna	---	---	1	2	1	4
Wrist	---	1	---	1	---	2
Femur	---	---	2	4	4	10
Knee	---	---	---	6	1	7
Tibia and/or fibula	1	2	2	6	4	15
Ankle	---	---	2	2	3	7
Total	1	4	8	¹28	²15	¹56

¹Includes 6 fatal fracture wounds; 1 of the humerus,2 of the femur, 2 of the knee, and 1 of the tibia and/or fibula.
²Includes 1 fatal fracture wound of the femur.

Flak, Mortar, and Artillery Shell Fragments

In a report by the Bombing Survey Unit on American ground force casualtiessustained in the Cassino area (p. 541) are some data on the sizes of mortar andartillery shell fragments causing nonfatal wounds. The sizes of fragmentscausing fatal wounds were not determined. Table 210 compares the sizes anddistribution of 424 flak fragments which caused nonfatal wounds in 361 aircrewcasualties with 157 mortar and artillery shell fragments which caused wounds in27 ground force casualties.

TABLE 210.-Comparisonof flak fragments in nonfatal wounds in 361 aircrew casualties with mortar andartillery shell fragments in 27 ground force casualties in the Cassino area

Flak fragments in 361 aircrew casualties			Mortar and shell fragments in 27 ground force casualties
Weight	Number	Percent of total fragments	Weight	Number	Percent of total fragments
Mg.			Mg.
0-50	56	13. 2	0-50	77	49. 0
50-250	64	15. 1	50-250	29	18. 5
Gm.			Gm.
?-1	46	10. 8	?-1	33	21. 0
1-5	100	23.6	1-5	18	11.5
5-20	100	23. 6	5-20	---	---
>20	58	13.7	>20	---	---

585

The increased average number of fragments per casualty in the ground forces(5.8 percent as compared with 1.2 percent per aircrew casualty) may be partiallyexplained by the breakup of fragments after hitting. The lower carbon content ofthe steel from which mortar and artillery shells are made would account for thefiner breakup and the greater irregularity in the shape of their fragments. Thepreponderance of wounds in ground force casualties associated with fragmentssmaller than 1 gm. in weight (88.5 percent as compared with 39.1 percent foraircrew casualties) attests to the greater vulnerability of ground force troopsand to the greater protection of aircrew personnel against small low-velocityfragments. The casualty risk rate for troops from mortar and artillery fire inthe Cassino area was estimated to be approximately 27 percent. This was theestimated casualty rate for two infantry regiments exposed to enemy mortar andartillery fire during 7 days of combat. Even with all MIA aircrew personnelincluded as battle casualties, the casualty rate for the Eighth Air Force for 3months was only 1.2 percent.

The Effect of Body Armor on the Distribution of Flak Wounds

It has been impossible to collect accurate data to show theincidence of personnel hit but uninjured by flak in the regions of the bodyprotected by armor. Records in the Office of the Surgeon, USSTAF, showed only 15such instances for the 3 months' period of this survey. An evaluation of theprotection afforded by armor may be obtained from a study of the quantitativerelationship between flak hits and projected body surface areas. Thisrelationship is shown in table 211. The values in the table are based on the1,222 flak hits observed in the 961 casualties referred to in table 200 and maybe regarded as indices of vulnerability. The mean areas of the different regionsof the body used in the calculations are the same as those referred to in table199 and are shown in table 211 as "Hits expected." The application ofmean projected surface areas as measured to a man usually seated in a heavybomber is a purely arbitrary one. The presence of combat equipment and thestructure of the aircraft in addition to the wearing of body armor probablyinfluenced the regional distribution of flak hits materially. A purely randomdistribution of hits on unprotected individuals would cause all the indices inthe table to be 1.00. The effective protection to the chest and abdomen isapparent as indicated by indices of 0.39 and 0.22 for these regions,respectively. An index of 1.42 for the upper extremities reveals this region tobe most vulnerable to hits in aircrew personnel. It should be pointed out thatthe relatively high vulnerability index of 1.30 for the head-and-neck region isdue largely to the relatively greater vulnerability of the neck rather than ofthe head itself. When one compares the vulnerability of different body regionsas demonstrated in table 211 with the regional distribution of hits (table 200),the point that is demonstrated in the former is the relatively highvulnerability of the head and neck and upper extremities in proportion to thesurface area projected by these regions.

586

TABLE 211.-Relativevulnerability of different body regions as compared with the 1,222 flak hits observed on body surface area of961 casualties

Body region	Body surface area or hits expected	961 WIA and KIA aircrew casualties (body armor worn)
Body region	Body surface area or hits expected	Hits observed	Index¹
	Percent	Percent
Head	12.0	15.6	1.30
Chest	16.0	6.3	.39
Abdomen	11.0	2.4	.22
Extremities:
Upper	22.0	31.3	1.42
Lower	39.0	44.4	1.14
Total	100.0	100.0

¹Index = (Percent of hitsobserved) / (Percent of hits expected)

Table 212 is a breakdown of the regional distribution of flakhits given in table 200 according to combat position. The values are expressedas percentages of total hits received in each combat position.

Table 213 shows the quantitative relationship between flakhits and body surface areas in terms of vulnerability indices for each of thecombat positions in U.S. heavy bomber aircraft. Here again, it may be said thata purely random distribution of hits on unprotected individuals would cause allthe values in the table to be 1.00. Thus, for example, the ball turret gunnerappears to have the greatest protection from flak hits of the abdomen. Thecopilot appears to have the least chest protection, or at least he sustains thegreatest number of chest hits. However, it is clearly apparent that throughoutthe

TABLE 212.-Regionaldistribution of flak hits according to combat position
[Values expressed as percentages of total hits received in each position]

Combat position	Head	Chest	Abdomen	Upper extremity	Lower extremity

Pilot		20.2	5.3	2.1	28.7	43.7

Copilot		15.2	9.4	1.1	30.6	43.7

Navigator	18.9	9.1	3.2	32.0	36.8
Bombardier		19.7	5.7	1.5	37.9	35.2
Radio operator	9.7	2.9	1.9	29.1	56.4
Waist gunner	14.9	6.0	2.6	30.6	45.9
Ball turret gunner	9.2	4.6	0	27.7	58.5
Top turret gunner	20.2	5.7	2.9	18.3	52.9
Tail gunner	9.5	7.0	3.8	36.9	42.8
Average	15.6	6.3	2.4	31.3	44.4

587

different combat positions the number of flak hits of the chest and abdomenin proportion to the areas these regions present is relatively low. It is alsoapparent that one of the vital regions of the body, that is, the head and neck,of a man in any combat position is relatively poorly protected from flak. Thehead and neck are most vulnerable in the pilot's and top turret gunner'spositions.

TABLE 213.-Relativevulnerability (index of vulnerability) of body regions in different combat

Combat position	Head	Chest	Abdomen	Upper extremity	Lower extremity
Pilot	1.68	0.33	0.19	1.31	1.12
Copilot	1.27	.59	.11	1.39	1.12
Navigator	1.58	.57	.30	1.46	.94
Bombardier	1.64	.36	.14	1.72	.90
Radio operator	.81	.18	.18	1.32	1.44
Waist gunner	1.24	.37	.24	1.39	1.18
Ball turret gunner	.77	.29	0	1.26	1.50
Top turret gunner	1.68	.36	.26	.83	1.36
Tail gunner	.80	.44	.35	1.68	1.09

The choice of the measurements for projected surface areaswas purely arbitrary and since the areas were measured for a man not in anaircraft are likely to be quite different for a man in his aircrew combatposition. Perhaps a more accurate estimate of the projected surface areas ofbody regions for aircrew personnel and the effectiveness of body armor could beobtained from an analysis of flak wound distribution in samples of armored andunarmored aircrew personnel. The available material for such an analysisconsisted of 104 of the flak casualties in the present survey (88 WIA and 16 KIA)who were known not to be wearing body armor. In addition to these, a perusal ofEighth Air Force records before the introduction of body armor revealed 307known flak casualties (294 WIA and 13 KIA) which were sustained during theperiod August 1942 to December 1943.

Table 214 shows the distribution of flak wounds in unarmoredand armored aircrew personnel. The figures in column 3 for armored individualsare those given in table 200 for all casualties due to flak, less those known tohave been sustained in unarmored personnel.

Flak wounds of the chest and abdomen before the use of bodyarmor accounted for 13.3 percent of the casualties as compared with 8.2 percentsince the use of body armor. Chi-square test of the significance of thedifferences between these figures gives the value x²=9.70(n=1, P less than 0.005). From the numbers available, thisdifference is very highly significant.

A number of unarmored KIA casualties were omitted fromcolumns 1 and 2 of table 214 because the identity of the missiles causing thefatal wounds could not be ascertained from the old records. Could these havebeen in-

588

TABLE 214.-Regionaldistribution of flak wounds in unarmored and armored aircrew personnel

Region	Wounds in unarmored casualties¹		Wounds in armored casualties²
Region	Number	Percent	Number	Percent
Head	92	16.1	173	16.1
Chest	56	9.8	59	5.5
Abdomen	20	3.5	29	2.7
Extremities:
Upper	151	26.4	349	32.5
Lower	252	44.2	465	43.2
Total	571	100.0	1,075	100.0

¹382 WIA; 29 KIA.
²777 WIA; 80 KIA.

cluded, the incidence of wounds of the head, chest, and abdomenundoubtedly would have been greater and the apparent beneficial effect of bodyarmor marked. Protective steel helmets were generally worn by aircrew personnelboth before and since the introduction of body armor and, as might be expected,the incidence of flak wounds of the head remained unchanged.

An analysis was made of the incidence and case fatality ratesof flak injuries of the head sustained by men wearing and not wearing steelhelmets. Information was obtained from 458 aircrew casualties, 369 of whom worehelmets and 89 of whom did not wear helmets (table 215). Only those regions ofthe head normally protected by a steel helmet were considered in this analysis.

TABLE 215.-Distributionof 458 casualties with cranial injuries due to flak, by protected or unprotectedhelmet area

Helmet area	Number of casualties	Total wounds		Fatal wounds
Helmet area	Number of casualties	Number	Percent	Number	Percent fatality
Protected	369	33	8.9	19	58
Unprotected	89	13	14.6	10	77
Total	458	46	10.1	29	63

It would appear that both the incidence and case fatalityrate of injuries of the head due to flak were decreased by the wearing of thesteel helmet. However, the sample of data was not sufficient to give statisticalsignificance to the differences between either the incidence or the casefatality rate for cranial wounds of protected and unprotected individuals.

589

CASUALTIES DUE TO SECONDARY MISSILES

The Distribution of Casualties Due to Secondary Missiles

A total of 104 aircrew battle casualties (100 WIA and 4 KIA)were due to secondary or unknown missiles. This represents 9.3 percent of thetotal of 1,117 casualties in the 3 months' survey. As stated previously,secondary missiles include fragments of Plexiglas, pieces of dural from the skinof or objects in the plane, bulletproof glass, brass fittings, parts ofelectrical heating and radio equipment, and .50 caliber machinegun ammunition(figs. 279, 280, 281, and 282). In this analysis, however, secondary missilesleft in wounds along

FIGURE 279.-Secondarymissiles (dural). The most common of secondary missiles in aircrew casualtiesare the aluminum alloy fragments from the skin and other parts of aircraft knownas dural. Only the top left fragment, the tip of the throttle of a B-17,produced a fatal wound by transecting a man's trachea and lodging in his neckalongside his vertebral column. The other pieces were found along the fatalwound tracks caused by primary missiles.

the path of enemy missiles were not included. Only thosemissiles which alone were responsible for wounds, secondary to hits elsewhere byflak or fire from enemy aircraft, were included. Table 216 shows thedistribution of battle casualties due to secondary and unknown missiles.

Plexiglas Wounds

All of the WIA casualties due only to Plexiglas sustainedinjuries from fragments of Plexiglas set in motion by flak. In one instance, afragment of a 20 mm. cannon shell, in addition to a fracture wound of theright arm, caused

590

FIGURE 280.-Secondarymissiles (body armor). The second most common of secondary missiles found in thetracks of fatal wounds in aircrew personnel are plates and fragments of bodyarmor and helmet. These are most readily recognized by their uniform thicknessof approximately 1 millimeter and by the fact that they are nonmagnetic thoughobviously of greater density than fragments of dural.

a secondary Plexiglas wound of the face. In 73 of thecasualties, the wounds produced by the fragments of Plexiglas were the only onessustained, and all of these occurred in the unprotected area of the face andneck with the exception of one, which was in the forearm. Two others sustainedtwo wounds from Plexiglas fragments, one each of the head and forearm. Out ofthe total of 88 Plexiglas wounds in 76 individuals, there were only 3 thatoccurred on protected parts of the body; that is, parts of the body protected byas little as the sleeves of the man's uniform. These three wounds occurred onthe forearms of three individuals. Figure 283 shows diagrammatically the loca-

591

tion of the 85 (97 percent) Plexiglas wounds that occurred on the unprotectedpart of the body. Plexiglas wounds were sustained by men in all combatpositions. However, bombardiers, navigators, and top turret gunners accountedfor 68 percent of them.

TABLE 216.-Distributionof 104 aircrew battle casualties, by category and by causative agent(secondarymissile)

Causative agent	WIA casualties	KIA casualties	Total casualties
Plexiglas	76	---	76
Dural	4	1	5
Bulletproof glass	3	---	3
Motor bearing	---	1	1
Zinc fragment	1	1	2
.50 caliber machinegun ammunition	---	1	1
Unknown	16	---	16
Total	100	4	104

FIGURE 281.-Secondarymissiles (Plexiglas). The third most common of secondary missiles causing woundsin aircrew personnel are fragments of Plexiglas. These fragments do not producefatal wounds and never penetrate deeply into tissue except when driven in by aheavier primary missile. Plexiglas fragments produce only slight superficialwounds and lacerations by themselves.

592

FIGURE 282.-Secondarymissiles (miscellaneous). These consist of a bearing from an aircraft'sengine, parts of electrical apparatus, clothing, personal equipment, oxygenline, rubber, zipper, and "dog tag" chain. With the exception of thebearing which by itself produced a fatal head wound, all of these were found inaircrew personnel along the fatal wound tracks caused by primary missiles.

Return to Flying Status of Casualties Caused by SecondaryMissiles

Table 217 shows the relative severity of wounds due to Plexiglasand other secondary and unknown missiles as judged by the time lost from flyingstatus.

There is a striking difference in the severity of wounds dueto Plexiglas fragments as compared with wounds due to flak and other missiles.Thus,

593

FIGURE 283.-Locationof 85 wounds due to Plexiglas fragments in 75 WIA battle casualties.

TABLE 217.-Distributionof 104 aircrew battle casualties due to Plexiglas fragments and othersecondary or unknown missiles, by disposition

Disposition	Plexiglas fragments		Other secondary or unknown missiles
Disposition	Number	Percent	Number	Percent
Returned to flying status after loss of-
0-1 day	24	32	5	18
1-7 days	31	41	5	18
7-30 days	14	18	4	14
30-90 days	1	1	2	7
Total	70	92	16	57
Permanently grounded	6	8	8	29
Killed in action	---	---	4	14
Total	6	8	12	43
Grand total	76	100	28	100

594

none were killed and only 8 percent of men wounded by fragments of Plexiglaswere permanently grounded, whereas 43 percent of men wounded by flak or othermissiles were permanently grounded or killed.

CASUALTIES DUE TO MISSILES FROM ENEMY FIGHTER AIRCRAFT

Causes of Casualties

In the present survey, 50 battle casualties (4.5 percent) were known to bedue to missiles fired from enemy aircraft. Their distribution according tomissile (figs. 284, 285, and 286) and type of casualty is shown in table 218.Cannon shells (20 mm.) accounted for 88 percent of the casualties.

FIGURE 284.-Primarymissiles (7.92 mm.). The top specimen is the steel core of an armor-piercing7.92 mm. Mauser bullet. The fragments in the bottom group are from the jacket ofthe same type of bullet.

TABLE 218.-Distributionof 50 aircrew battle casualties, by category and by missile fired fromenemy aircraft

Missile	Category		Total casualties
Missile	WIA	KIA	Number	Percent
	Number	Number
20 mm.	37	7	44	88
13 mm.	3	1	4	8
7.92 mm.	2	---	2	4
Total	42	8	50	100

595

FIGURE 285.-Primarymissiles (13 mm.). Each fragment or group of fragments was found in fatal woundsof aircrew personnel. The two on the left are from armor-piercing and the two onthe right are from high explosive shells. The 13 mm. cannon shell was thesmallest caliber missile in which a high explosive charge was used.

Distribution of Casualties According to Combat Position

Table 219 shows the distribution of the 50 casualtiesaccording to combat position. Although the number of casualties is quite small,it may be noted that, as in the case of casualties due to flak, the waist gunnerappears to be the man most vulnerable to fighter attack. This is at leastpartially accounted for by the fact that two waist gunners were frequentlycarried in heavy bombers. The tail gunner is probably most vulnerable to fighterattack. This is in agreement with the findings of the Eighth Air ForceOperational Research Section that the greatest directional density of hits onheavy bombers by enemy cannon and machineguns is from dead astern.

TABLE 219.-Distributionof 50 aircrew casualties due to missiles fired from fighter aircraft, by categoryand combat position

Combat position	Category
Combat position	WIA	KIA
	Number	Number
Pilot	2	1
Copilot	4	---
Navigator	2	---
Bombardier	---	---
Radio operator	7	2
Waist gunner	11	2
Ball turret gunner	5	---
Top turret gunner	3	1
Tail gunner	8	2
Total	42	8

596

FIGURE 286.-Primarymissiles (20 mm.). Each fragment or group of fragments caused, and was found in,a fatal wound of aircrew personnel. All except the bottom group were from highexplosive shells. The bottom specimen has been reconstructed from the retainedfragments found in a through-and-through fatal wound produced by a 20 mm.armor-piercing incendiary cannon shell. The shell broke up in the wound as theresult of its having perforated a flak suit worn by the casualty. Plates andfragments of the man's body armor were also found in the wound.

It is apparent from the distribution of casualties in table 219that bombardiers and navigators were least likely to be casualties from enemyfighter attack. This might be expected on the basis that enemy fighters areleast likely to attack the nose of an aircraft from the front. The low incidenceof casualties due to missiles from fighter aircraft as compared with the highincidence due to flak for the bombardier and navigator positions substantiatesthe finding that these positions are susceptible to an increased density of flakhits because of their leading and exposed positions with respect to the rest ofthe aircraft.

Regional Distribution of Wounds Due to Missiles From EnemyFighter Aircraft

Tab1e 220 shows the distribution of 83 wounds in 50casualties struck by missiles fired from enemy fighter aircraft. Thewounddistribution in this

597

group of casualties differs from that in flak and othermissile casualties in that there is an increase in the occurrence of wounds ofthe vital regions of the body. Less than 25 percent of flak wounds occurred inthe head and trunk regions as compared with an incidence of 35 percent for headand trunk wounds due to missiles from enemy fighter aircraft.

TABLE 220.- Distributionof 83 wounds in 50 aircrew battle casualties due to missiles fired from fighteraircraft, by category of casualty and anatomic location (regional frequency) ofwounds

Anatomic location	WIA casualties		KIA casualties		Total casualties
Anatomic location	Number of wounds	Percent	Number of wounds	Percent	Number of wounds	Percent
Head	13	19	6	38	19	23
Chest	3	5	3	19	6	7
Abdomen	2	3	2	12	4	5
Extremities:
Upper	20	30	3	19	23	28
Lower	29	43	2	12	31	37
Total	67	100	16	100	83	100

Single and Multiple Wounds Due to Missiles From Enemy FighterAircraft

Table 221 shows the frequency with which one and more than oneregion of the body was wounded by missiles fired from enemy fighter aircraft inthe 50 casualties. The increased multiplicity and severity of wounds in thisgroup of casualties may be compared with those sustained by casualties due toall missiles (table 186) and flak casualties (table 201). In both of the latter,only 15

TABLE 221.-Distributionof 50 aircrew battle casualties due to missiles fired from fighter aircraft, bycategory and number of regions wounded

Regions wounded	WIA casualties		KIA casualties		Total casualties
Regions wounded	Number of wounds	Percent	Number of wounds	Percent	Number of wounds	Percent
Number:
1	25	59.5	4	50.0	29	58.0
2	10	23.8	1	12.5	11	22.0
3	6	14.3	2	25.0	8	16.0
4	1	2.4	1	12.5	2	4.0
Total	42	100.0	8	100.0	50	100.0

598

percent were wounded in more than one region as compared with 42percent of casualties wounded in more than one region due to missiles from enemyfighter aircraft.

Altitude at Which Casualties Sustained Wounds

The altitude at which 27 of the 50 (54 percent) casualtiessustained wounds was known. Table 222 shows distribution of the casualties andthe type of aircraft in which they were wounded.

There are no significant variations in the distribution ofB-17 casualties according to altitude. Of 5 B-24 casualties, 4 were wounded orkilled below 22,000 feet. This is in agreement with the observations madepertaining to the altitude at which casualties due to flak were sustained;namely, that B-24 aircraft usually operated at a lower altitude than B-17aircraft.

The difference in the occurrence of casualties in the twotypes of aircraft is marked (76 percent in B-17's and 24 percent in B-24's).However, the frequency with which B-17's and B-24's were attacked by enemyfighter aircraft is not known. Thus, the relationship between fighter damage toaircraft and the occurrence of casualties was not determined, and an evaluationof the significance of the difference in the occurrence of casualties in the twotypes of aircraft could not be made.

TABLE 222.-Distributionof 50 aircrew battle casualties (38 in B-17's; 12 in B-24's) dueto missiles fired from enemy fighter aircraft, by altitude

Altitude in feet	Casualties in B-17's			Casualties in B-24's
Altitude in feet	WIA	KIA	Total	WIA	KIA	Total
Above 26,000 feet	1	---	1	---	---	---
24,001 to 26,000 feet	7	1	8	---	---	---
22,001 to 24,000 feet	1	2	3	1	---	1
20,001 to 22,000 feet	2	---	2	---	1	1
18,001 to 20,000 feet	7	---	7	1	1	2
Below 18,000 feet	1	---	1	1	---	1
Unknown	14	2	16	6	1	7
Total	33	5	38	9	3	12

KIA CASUALTIES-JUNE THROUGH NOVEMBER 1944

During the 6 months from June through November 1944, the bodies of 164 KIAbattle casualties from the Eighth Air Force and Ninth Air Force and TroopCarrier Command were examined in the laboratory of the Medical ORS.

599

During the last 2? months of the survey period, the Office of the Surgeon,USSTAF, provided additional facilities and personnel to aid the ORS in theexaminations.⁸

Causes of the Fatalities

The missiles causing fatal wounds in 164 casualties areshown in table 223. The proportion of fatalities due to flak (87.8 percent) isapproximately the same as the incidence of all aircrew battle casualties due toflak (86.2 percent).

TABLE 223.-Distributionof 164 KIA aircrew casualties, by missile

Missile	Casualties
Missile	Number	Percent
Flak	144	87.8
From fighter aircraft	¹10	6.1
Secondary	4	2.4
Unknown	6	3.7
Total	164	100.0

¹Of these, 8 casualties were due to20 mm. missiles and 1 each to 13 mm. and 7.92 mm. missiles.

Distribution of KIA Casualties According to Combat Position

Table 224 shows the distribution of 164 KIA casualtiesaccording to combat position. The positions of KIA casualties of the Ninth AirForce and Troop Carrier Command not accounted for in heavy bombers are includedunder "Position unknown." The occurrence of fatal casualties and thusthe case fatality rates for casualties in any combat position do not appear todiffer significantly except as previously noted in table 184.

Regional Distribution of Wounds

Table 225 shows the regional distribution of all entrancewounds due to all missiles in this larger sample of 164 KIA battle casualties.The distribution of wounds in the different body regions is not appreciablydifferent from that of the smaller sample of 110 KIA casualties describedpreviously in table 185.

Regional Frequency of Wounds

Table 226 shows the regional frequency of wounds in thelarger sample of KIA casualties.

⁸From the examination of the casualties referred to in this section, complete data pertaining to the morbid anatomy, histopathology, arteriography, and X-ray appearance of fatal wounds were compiled. Lack of space, however, precludes inclusion of this information on the 164 casualties examined. Another report on these casualties is also contained in AAF Memorandum Report, TSEAL-3697-7B, 1945, by J. B. Hickam, Aero Medical Laboratory, Wright Field, Ohio.

600

TABLE 224.-Distributionof 164 KIA aircrew casualties, by combat position

Position	Casualties
Position	Number	Percent
Pilot	9	5.5
Copilot	10	6.1
Navigator	20	12.2
Bombardier	22	13.4
Radio operator	17	10.4
Waist gunner	32	19.5
Ball turret gunner	7	4.2
Top turret gunner	10	6.1
Tail gunner	28	17.1
Unknown	9	5.5
Total	164	100.0

TABLE 225.-Distributionof 451 wounds in 164 KIA aircrew casualties due to all missiles byanatomic location

Anatomic location	Wounds
Anatomic location	Number	Percent
Head	105	23.3
Thorax	58	12.9
Abdomen	32	7.1
Extremities:
Upper	145	32.1
Lower	111	24.6
Total	451	100.0

TABLE 226.-Distributionof 164 KIA casualties due to all missiles, by anatomic location (regionalfrequency) of wounds

Anatomic location	Casualties
Anatomic location	Number	Percent
Head	50	30.5
Chest	16	9.8
Abdomen	3	1.8
Extremities:
Upper	---	0
Lower	11	6.7
Multiple regions	84	51.2
Total	164	100.0

601

Comparison of tables 225 and 226 reveals again that the differences in theregional incidence of wounds by the two methods of tabulation, that is, woundsversus casualties, are quite marked. These differences are characteristic of allsamples of KIA casualties and as stated before are due primarily to the highincidence of multiple wounds in the dead.

Single and Multiple Wounds

Table 227 shows the incidence of single and multiple woundsin KIA casualties. The KIA casualties were struck in more than one region fourtimes as often as the WIA casualties. In the 3 months' survey of allcasualties as discussed earlier in this chapter, multiple wounds were more thanthree times as frequent in KIA as in WIA casualties (39.0 percent in the KIA ascompared with 12.2 percent in the WIA). The incidence of multiple regionswounded in this larger sample of KIA casualties (includes 89 of the 110 KIA fromthe smaller sample) was 50.0 percent. The increased number of multiple wounds inthe larger sample was most marked in about half of the sample; that is, thosecasualties sustained during September, October, and November 1944. Such asignificant increase, which would be even more marked if the incidence ofmultiple wounds for the first 3 months were to be compared separately with thatfor the second 3 months, may be regarded as being due to the increased use ofhigher burst velocity shells by the enemy.

TABLE 227.-Distributionof 164 KIA aircrew battle casualties, by single and multiple regions wounded

Number of regions wounded	Number of casualties	Percent of casualties
1	82	50.0
2	49	29.9
3	22	13.4
4	7	4.3
5	4	2.4
Total	164	100.0

Incidence of Fractures in KIA Casualties

Table 228 shows the distribution of 265 fractures according to body regions.Of the fractures in KIA casualties, 72.9 percent were associated with wounds ofthe vital areas of the head and trunk regions as compared with 14.3 percent inWIA casualties (table 189).

The 91 percent incidence of fractures in KIA casualties (in 149 of the 164)reported here as compared with 85.3 percent reported for the smaller sample (87of the 102) may be explained by the increased multiplicity of wounds in thelarger sample.

602

TABLE 228.-Distributionof 265 fractures in 149 KIA aircrew battle casualties due to all missiles,by anatomic location

Anatomic location	Number of fractures	Percent of fractures
Head	102	38.5
Chest	86	32.5
Abdomen	5	1.9
Extremities:
Upper	34	12.8
Lower	38	14.3
Total	265	100.0

Causes of Death in KIA Battle Casualties Due to AllMissiles

Table 229 shows the regional distribution of wounds which werethe causes of death in 164 KIA casualties. There were 11 casualties in whicheither of 2 hits could have been fatal and 2 casualties in which any 1 of 3 hitscould have been fatal. However, for this tabulation, the following criteria werefollowed in order to determine the primary fatal wound:

1. Only the severest one of multiple fatal wounds wasregarded as the cause of death in any one casualty.

2. When the severity of a head and a chest or abdominal woundappeared to be the same, the cause of death was arbitrarily attributed to thehead wound.

3. When the severity of a chest and an abdominal woundappeared to be the same, the cause of death was attributed to the chest wound.

4. Decapitations were regarded as causes of death due towounds in the head-and-neck region in cases where the head was missing as wellas in other cases where a head wound was very extensive and associated withcomplete evulsion of the brain.

TABLE 229.-Distributionof 164 aircrew battle casualties due to all missiles, by anatomic locationin which the primary fatal wound occurred

Anatomic location	Number of casualties	Percent of casualties
Head	74	45.1
Chest	63	38.4
Abdomen	12	7.4
Extremities:
Upper	---	0
Lower	15	9.1
Total	164	100.0

603

TABLE 230.-Distributionof 164 KIA aircrew battle casualties due to all missiles, by anatomiclocation and type of fatal wound

Anatomic location and type of wound	Number of casualties	Percent of casualties
Head:
Penetration	65	39.6
Decapitation	7	4.3
Fracture, without penetration	2	1.2
Total	74	45.1
Thorax:
Penetration	60	36.6
Mutilation	3	1.8
Total	63	38.4
Abdomen, penetration	12	7.4
Lower extremity:
Perforation	11	6.7
Traumatic amputation	4	2.4
Total	27	16.5
Grand total	164	100.0

5. In the case of extensive mutilating wounds, the cause ofdeath was attributed to a wound of the region of the body nearest the center ofthe area of mutilation.

Table 230 gives the breakdown of cause of death data. Figures287, 288, 289, 290, 291, and 292 depict typical examples of fatal wounds inaircrew casualties.

SUMMARY AND CONCLUSIONS

The survey of aircrew casualties presented here covers aperiod of 3 months of operational missions carried out by heavy bombers of theEighth Air Force (D-5 to D+86). A survey of KIA casualties was continued for afurther 3 months and was extended to include casualties from the Ninth Air Forceand Troop Carrier Command. A total of 69,682 heavy bomber sorties (39,724 byB-17's and 29,958 by B-24's) was credited during the 3 months' period.This represents 657,096 man-combat missions completed of which 357,516 were inB-17's and 299,580 were in B-24's. During the period, 693 heavy bombers (390B-17's and 303 B-24's) were reported "Missing-in-Action." Thus,casualty data pertaining to 6,540 (1.00 percent)

604

FIGURE 287.-Radiooperator in B-17 aircraft. Typical example of fatal wound in the neck regioncaused by large low-velocity flak fragment. A. Wound of entrance 1.5 x

4.2 cm. B. Bruised areasoverlying flak fragment. C. Flak fragment in larynx. D. Flak fragment 47.30grams.

605

FIGURE 288.-Waistgunner in B-17 aircraft. Example of through-and-through fatal wound produced byhigh-velocity flak fragment. A. Wound of entrance, right (4 x 4 cm.), and woundof exit, left (4 x 7 cm.). B. Missile track laid open showing extensivemutilation of thoracic cage. C. Thoracic viscera showing widespread damage tolungs and posterior mediastinum.

606

FIGURE 289.-Radiooperator in B-17 aircraft. Example of a fatal wound of the unprotected flank andextending into the chest, produced by a large low-velocity flak fragment. A.Entrance wound 5.5 x 14.7 cm. B. Damage to abdominal organs. C. Extensivelaceration of heart with missile in situ. D. Flak fragment 83.66 grams.

officers and enlisted men were not available. Casualty data wereavailable and studied on the 99 percent of aircraft and personnel thatsuccessfully completed and returned from 68,989 sorties or 650,556man-combat-missions. There were 1,117 known battle casualties sustained by theEighth Air Force during the 3 months of the survey of whom 110 had been killedand 1,007 wounded as a result of enemy gunfire. The 1,117 casualties representan overall casualty rate of 0.172 percent (1.72 percent per 1,000 man-combatmissions completed). When distributed according to types of aircraft, thecasualty rates in B-17's and B-24's were 2.10 and 1.26 per 1,000 man-combatmissions, respectively. The case fatality rate was 9.8 percent and did netdiffer significantly for casualties in the two types of aircraft.

The ratio of MIA personnel to known casualties wasapproximately 6 to 1. The data pertaining to casualties among MIA personnel,could they have been included in the study, might have materially influenced theobservations that have been made. Of aircrew personnel, 1 percent (10.1 per1,000 mancombat missions) were known to be missing in action. The incidence ofMIA

607

FIGURE 290.-Navigatorin B-17 aircraft. Example of fatal cranial wound due to unexploded 88 mm. higherexplosive antiaircraft shell. A. Entrance hole in the nose of a B-17 aircraft ofan unexploded 88 mm. high explosive antiaircraft shell. B. Fatal wound of thehead produced by unexploded 88 mm. shell as it passed through the nose of a B-17aircraft.

aircraft and personnel for the two types of bombers did notdiffer significantly.

Flak fragments caused 86.2 percent of the casualties. Since7.8 percent of the casualties were due to secondary missiles, that is, those setin motion usually by flak, 94 percent of all the casualties studied may beregarded as being due to flak. Of the total casualties, 4.5 percent were causedby missiles from enemy fighter aircraft and the remaining 1.5 percent werecaused by unidentified missiles.

The incidence of multiple wounds in KIA casualties during thefirst 3 months of the survey was 39.1 percent. This incidence increased to 50.0percent when the KIA casualties examined during the second 3 months wereincluded in an analysis of all KIA casualties. The increase in the multiplicity

608

FIGURE 291.-Radiooperator in B-17 aircraft. Typical example of fatal wound in axillary regioncaused by flak fragment. A. Wound of entrance 3 x 3.6 cm. B. Wound of exit 1.5 x2 centimeters.

of wounds may be regarded as evidence of an increase in the useof higher burst velocity shells by the enemy. Further evidence of this is givenwhen a comparison is made of the incidence of fractures in the two samples ofKIA casualties. During the first 3 months, the incidence of fractures was 85.3percent, whereas for the 6 months the incidence of fractures increased to 91percent.

The severity of wounds sustained by aircrew battle casualtieswas evaluated on the basis of time lost from flying status. The period ofobservation after injury was limited to 90 days. Of the total number ofcasualties (including the KIA), 33.8 percent were permanently lost from flyingstatus. Of the WIA casualties, 9.8 percent lost a day or less from flyingstatus, 25.4 percent

609

FIGURE 292.-Pilotof B-24 aircraft. A. Wound of entrance 8.4 x 13.2 cm. B. Multiple wounds ofexit. C. Partial reconstruction of a 20 mm. armor-piercing incendiary cannonshell from retained fragments, 75.46 gm. D. Pieces of body armor and otherpersonal equipment in fatal wound.

lost a week or less, 55.9 percent lost a month or less, and 17.5percent lost from 1 to 3 months.

Approximately 70 percent of the casualties in B-17'soccurred at an altitude of 24,000 feet or above, whereas 92 percent in B-24'soccurred at 23,000 feet or below.

Of all WIA aircrew battle casualties, 90 percent receivedadequate surgical treatment in hospitals within 4 hours after they were wounded.

Since by far the majority of casualties was caused by flak,an independent analysis of all flak casualties was made. Their distributionaccording to combat position is shown in the order of frequency (table 231).Heavy bomber aircraft formerly carried two waist gunners, which probablyaccounts for the highest incidence of casualties in that combat position. Thelowest incidence of casualties in the ball turret gunner's position is atleast partially due to the fact that only B-17 aircraft carry a ball turretgunner.

610

No fragments smaller than 1 gm. were recovered from fatalwounds due to flak in KIA casualties; 92.6 percent of those recovered weighed 5gm. or more. In WIA casualties, 39.1 percent of the fragments weighed less than1 gram.

Plexiglas fragments set in motion by other missiles produced88 wounds; of these 85 were on the exposed regions of the face and neck and 3were on the forearms. There were no Plexiglas wound fatalities. Of those woundedby Plexiglas fragments, 92 percent were returned to flying duty within 90 days.It would appear that protection of the eyes and circumorbital regions with anyrelatively thin, shatterproof, transparent material would probably haveeliminated most casualties due to Plexiglas fragments.

TABLE 231.-Distributionof flak casualties sustained according to combat position, in order of frequency

Order of frequency	Position of-
Order of frequency	WIA casualty	KIA casualty
1	Waist gunner	Waist gunner
2	Bombardier	Tail gunner
3	Navigator	Bombardier
4	Tail gunner	Navigator
5	Radio operator	Top turret gunner
6	Top turret gunner	Radio operator
7	Pilot	Pilot
8	Copilot	Ball turret gunner
9	Ball turret gunner	Copilot

Of the casualties due to missiles from fighter aircraft, 88percent were produced by 20 mm. shells. The tail gunner was found to be the mostvulnerable combat position, while bombardier and navigator were the leastvulnerable to enemy fighter aircraft. This is the reverse of the relativevulnerability of the same combat positions to flak and is in accordance with thefindings of the Operational Research Section, Eighth Air Force, that enemyfighter aircraft usually attack heavy bombers from the rear.

A comparison is made of the regional distribution of woundsin flak casualties with and without protective body armor. From table 232, itmay be seen that the incidence of flak wounds of the trunk has fallen from 13.3percent in unarmored casualties to 8.2 percent (38 percent decrease) incasualties wearing body armor. It is apparent that the thoracic and abdominalregions have been protected by the wearing of body armor. It has been observedthat the neck and axillary regions are the most highly vulnerable to penetrationby enemy missiles on men wearing body armor.

611

The separation of casualties into "unarmored" and"armored" in table 232 does not hold so far as the occurrence ofwounds of the head is concerned. Data pertaining to the protective value of headarmor (steel helmet) were not sufficient to evaluate statistically.

TABLE 232.-Meanprojected body areas and regional distribution of flak wounds in unarmored andarmored battle casualties

Region	Mean projected area of region	Wounds in unarmored casualties	Wounds in armored casualties
Region	Percent	Percent	Percent
Head	12	16.1	16.1
Chest	16	9.8	5.5
Abdomen	11	3.5	2.7
Extremities:
Upper	22	26.4	32.5
Lower	39	44.2	43.2
Total	100	100.0	100.0

RETURN TO TABLE OF CONTENTS