---

Underwater Trajectories:

It is clear from this information that Hood may indeed have been destroyed due to the action of a projectile which penetrated her belt.  But a shell need not necessarily have penetrated the belt to destroy her.  It might have gone over her armor belts.  Or it might have gone under.

Both British Boards of Inquiry considered the possibility that Hood was destroyed by a projectile which fell slightly short of the target, and, after travelling some distance underwater, penetrated the hull and then detonated well below the waterline.  This hypothesis is often associated with a similar underwater hit suffered by Prince of Wales in the same action.  This damage was not discovered until the ship was drydocked some time after the action, and therefore the exact time and range at which the hit was scored is uncertain, although because the projectile apparently entered the hull from about 45° relative, it seems most probable that the hit occurred shortly after Hood was destroyed.  Prior to that time, Prince of Wales was not under 380mm fire, and shortly after that time she turned away and presented her stern to the enemy.  The hit, caused by a 380mm projectile from Bismarck, first contacted Prince of Wales about 8.5 meters below the waterline.  After piercing the skin, it penetrated four additional internal light bulkheads before ending up nose forward just outboard of the ship's main torpedo bulkhead alongside the main engineering spaces.  This hit certainly had the potential to destroy the ship had it exploded, and had it struck adjacent to the magazines.

Especially because it appears that the Germans made no specific effort to enhance the underwater capabilities of their projectiles, this represents a truly remarkable performance.  Although the detailed underwater trajectory of any individual projectile remains difficult to predict with any precision, the general principles governing the underwater trajectories of most standard projectile types are relatively well understood.  The usual tendency of an ogival headed projectile impacting the water at an angle of fall of about fourteen degrees would be to travel about eighty calibers submerged in an upwardly curving path, and to re-emerge once again with its velocity greatly diminished.  The projectile would normally penetrate to a depth of about six calibers during its underwater run, corresponding to a depth of about 2.28 meters for Bismarck's 38cm shells.  Provided the windscreen remained attached, which would be atypical, a mathematical treatment⁵⁹ yields a trajectory radius of about 110 meters for Bismarck's shells, with a maximum depth of the trajectory of approximately 2.3 meters (c. 6.1 calibers) and a corresponding run to emergence of approximately 44 meters (c. 115 calibers).  For a projectile of this type, therefore, penetration to a depth of 8.5 meters, or more than 22 calibers, is remarkable.  Only the Japanese "suichu dan" projectiles, which were designed specifically to maximize their underwater performance, would be expected to approach this capability - in technical terms their practical lift coefficient seems to have been only about 0.11, corresponding to a trajectory radius of about 325 calibers.  Such a projectile striking at an angle of fall of 12° would typically be expected to penetrate to a depth of about 18-20 calibers (6.85-7.6 meters) after an underwater travel of about 110 calibers (41.8 meters).  At that point it would be expected to have retained about half of its original striking velocity.

In order to achieve this performance, the suichu dan projectile was designed with a specially weakened windshield and armor-piercing cap which broke away upon impact with the water and which presented the flat nose which is typical of projectiles designed for effective underwater penetration.  Although the German projectiles used by Bismarck did not have the special breakaway windscreen and cap characteristic of the Japanese diving projectiles, they were equipped with a rather brittle aluminum windscreen and a fairly flat faced 'knob and ring" armor-piercing cap design.  It is therefore quite possible that if the windscreen were removed by water impact, the comparatively flat armor-piercing cap may have accidentally approximated the effects of the suichu dan design.⁶⁰  In fact a broken windscreen surrounding a knob and ring cap might by chance have formed what a hydrodynamicist commonly calls a "stagnation cup" at the nose, a shape well known for its underwater stability.⁶¹

Even had this taken place however, there remains some doubt concerning the action of the projectile fuze.  The fuze delay of the standard German projectiles of the time was approximately 0.035 seconds, which at the range at which Hood was struck corresponds to a distance of only about 19 meters even if the projectile were traveling in air.  Because of the additional retarding force, the equivalent distance under water would be much less - certainly less than 17 meters, and probably closer to 13 or 14.  At the angles of fall which concern us here, this corresponds to a depth at detonation of something less than 3.5 meters.  Assuming a fuze delay of 0.070 seconds, or twice the nominal,⁶² results in an underwater travel of from about 19.5 to 31.1 meters depending upon one's (rather arbitrary) assumptions about the shell's underwater behavior.  If the trajectory were relatively straight, this corresponds to a depth of from about 4 to 7 meters.  Insofar as Hood's belt projected only about 960mm below the water line, even a projectile with a "normal" underwater trajectory could have easily struck beneath it, and penetrated well into the ship before exploding.  The second board of inquiry obviously considered a penetrating underwater hit a distinct and dangerous possibility, and prepared a special drawing of Hood showing the wave profile along the side at a speed of 28 knots.  At least in the stern of the vessel, this drawing shows that the effective "draft" of the ship at any point along the side might range between 10.5 meters just forward of the mainmast to only 9.5 meters just forward of 'X' turret.  The drawing also shows that the main armor belt, with its lower edge located about 7.75 meters above the keel, would normally be covered by only from 1.75 to 2.75 meters of water.  This was obviously a point of great vulnerability.

The investigators who examined the fuze of the dud projectile which had struck Prince of Wales found that it had been actuated, probably upon water impact, but that shortly thereafter the powder train ". . .  went out." This is not surprising when one considers that shells not specially designed for stable water entry quite commonly experience transient yaws exceeding 90° within the first twenty or so calibers of underwater travel.  It is in fact possible (though improbable) that the projectile which struck Prince of Wales penetrated the water normally, rotated nearly 180° shortly after impact, and stabilized base first for the remainder of its journey.  Such violent rotation would almost certainly have rendered the fuze inoperative, though after it was completed the flat base might have acted like a suichi dan nose.  The fact that the projectile was recovered nose forward in the bilges tends to support this somewhat unusual hypothesis.

Provided the fuze operated correctly, however, an underwater hit represents one of the most plausible explanations for the loss of the ship.  Projected upon the ship's cross sections, a projectile with an angle of fall of 10.6°-13.9° striking about 6 meters short could have penetrated the hull just below the 308mm belt and penetrated almost unimpeded directly to the area of the after magazines.  On the not unreasonable assumption that unarmored ship's structure along the path would have approximately the same retarding characteristics as water, the nominal fuze delay range of from 0.035 to 0.07 seconds places the likely point of detonation squarely in the after magazines.  The views below show the areas of vulnerability on Hood at the beginning and end of the turn.
 
 

A computer-generated plot of the internal arrangements of Hood corresponding to an angle of fall of 12° and a target angle of 53°.  This shows a "shell's-eye" view of Hood as she probably appeared at the beginning of her final turn.  The 12° angle of fall corresponds to an average value that would occur over the range of probable conditions. Click on this image for a larger view.

A computer-generated plot of the internal arrangements of Hood corresponding to an angle of fall of 12° and a target angle of 73°.  This shows a "shell's-eye" view of Hood as she probably appeared at the end of her final turn, if it was completed.  The 12° angle of fall corresponds to an average value that would occur over the range of probable conditions. Click on this image for a larger view.

Two cross sections of Hood taken from the ship's plans, showing possible paths to the after magazines.  The explosion points roughly correspond to fuze delays of 0.035 and 0.070 seconds respectively.  The upper trace shows how the shells were unlikely to have penetrated on the slope of the main deck, as the board had hypothesized.  The assume angle of fall is 10.5 degrees, corresponding to a range of c. 15,100 meters.  A passage through the after engine room would have been even less impeded. Click on this image for a larger view.

The boards of inquiry located, and most accounts since then have also mentioned, a band of potential weakness about 450mm wide located slightly above the probable waterline, and stretching from frame 259 to frame 280, i.e., abreast the after engine room.  In this area, they considered it possible that a projectile could reach the area of the magazines by passing just over the top of the 305mm belt, penetrating the 178mm belt, and sneaking in just under the flat protective main deck by striking the portion of the deck that sloped down to meet the side.  Assuming that the angle of fall of Bismarck's shells was somewhere between 10.6° and 13.9° and that the target angle was 53°, when superimposed on a cross section of the Hood, this corresponds to a resolved angle of between 13.2° and 17.2°.  Searching the cross sections from the ship's plans at this angle of fall (instead of the 20° resolved angle of fall assumed in the Board's investigations) indicates that such an event, though not impossible, would have been rather improbable.

As noted above, it is also possible that the fatal projectile may have reached the magazines by passing over the belt, rather than under it.  An attractive byproduct of this hypothesis is that it also lends itself to an easy explanation of the observation of witnesses that the fatal explosion seemed to originate in the vicinity of the mainmast rather than amongst the after magazines.  As the orthographic diagrams above show, one plausible route for a projectile heading toward the after magazines and passing over the belt takes it in the immediate vicinity of the starboard torpedo tube nest.  Because of the mantlet armor, a projectile traveling this path would not have been capable of detonating a torpedo, but it still might have done collateral damage along the way.  There was an obvious fire on the boat deck, and perhaps one under the boat deck as well, in the vicinity of the torpedo tubes.  If the shell or collateral fragments from the hull had perforated a charged torpedo air flask, for example, the result - especially in an area where an extensive fire was in progress - could have been a spectacular burst of flame around the mainmast, followed a split second later by the detonation of the after magazines.  This is, in fact, exactly what many witnesses observed.⁶³

Unfortunately it does not appear likely that a projectile from Bismarck could have penetrated to the magazines under these conditions.  The angle of fall of the incoming shells was only from 10.6° to 13.9°, which is a very high obliquity for deck penetration indeed.  Graphs of German armor penetration for the 380mm shell (see previous page) stop at an obliquity of 20° and must be extrapolated to obtain any figure at all for an obliquity greater than this.  Further, the best estimates obtained by such extrapolation seem to yield a total penetration capability of only about 65mm.  The total thickness of deck armor protecting the magazines in this area, admittedly distributed amongst several layers, amount to approximately 120mm.  On average, penetration via this route, especially by a damaged projectile, is therefore considered unlikely.

It has however been suggested that a projectile striking over the engineering spaces could reach the forward bulkhead of the magazine group by penetrating both the 50.8mm forecastle and the 19mm main decks and then detonating between the main and lower decks just forward of station 280.  The orthographic diagrams above reveal this to be an attractive alternative hypothesis.
 

Magazine Detonation:

A comprehensive British survey undertaken by the boards of inquiry covering twenty-two cases of war damage involving magazines, found only one case [4.5%] where the magazines had been known to explode, i.e., in Hood herself, six others [27.3%] where magazines were suspected to have exploded and fifteen cases [68.2%] where magazines had not exploded in spite of severe damage.⁶⁷  Coupled with American experience, which seems to have indicated that catastrophic propellant explosions were rather rare even in cases where shells or bombs had detonated directly inside magazine spaces, this would seem to indicate that catastrophic explosions were rather rare.

It must be noted however, that British double-based propellants, which contained a substantial amount of nitroglycerine in their makeup, were significantly more susceptible to ignition than their single-base American counterparts.  In 1945 the U.S. Navy Bureau of Ordnance conducted systematic tests to determine the susceptibility of various propellant formulations to accidental ignition.⁶⁸  Using a nozzle mechanism capable of generating a reproducible flash, they found that while British Cordite type propellants would ignite while still some 530mm from the vent, standard American single base propellants would not ignite until the distance was reduced to 120mm, and the relatively new U.S. "SPCG" flashless powder, incorporating nitroguanadine, would have to be within 25mm of the nozzle before ignition took place.  These are very substantial differences.  Assuming the flame of the explosion to expand in a spherical front, the same explosion which would ignite one cubic unit of standard American powder, would be capable of igniting almost seventy-five times as much cordite.  In the confined space of a magazine, the relative amounts of gas evolved, and the ensuing internal pressures could easily spell the difference between disturbance and disaster.  Had Hood carried single base propellant instead of cordite, there is in fact a good possibility that the fatal explosion might never have occurred.⁶⁹
 

Detonation of Torpedoes:

"Evidence of eye witnesses, REPULSE, and an Officer who had recently served in HOOD leaves little room for doubt that the mantlet doors were closed.  A warhead could still, however, have been detonated or exploded by a direct hit from BISMARCK'S shell.  There is no direct evidence that such a hit occurred, but it may have done so on either side of the ship.  If a single warhead had gone off one other, but probably not more than one, the other warhead would also have gone off.  . . . Expert opinion suggested that the explosion of two warheads would produce an all round almost instantaneous flash.  It would not have produced the very high column of flame of appreciable duration, which was seen by so many witnesses.  Nor was the noise, reported as being heard, compatible with that of a T.N.T. detonation or explosion.  The consensus of expert opinion was definitely against the characteristics of the explosion as given in evidence by eyewitnesses being that of T.N.T."⁷⁰

Sir Stanley Goodall, then Director of Naval Construction, held to a dissenting opinion, however:

"If one or more shells from the 5th salvo burst in this devastated area [where fire was already burning], where there are eight torpedo heads, four each side, each containing about 500 lbs of TNT at the base of the mainmast, and if one or more of these warheads detonated, the result would be an explosion where it was actually observed.  Such an explosion could break the ship's back already weakened in this neighbourhood by the earlier damage.  With the force on the after bulkhead of the engine room due to the ship's speed of 28 knots and the low reserve of buoyancy of the after part of the ship, this portion would rapidly sink.  The foregoing is an alternative explanation of the occurrence which is as likely as the explanation in the finding of the court."⁷¹

Sir Stanley was probably wrong.  Although no formal calculations were ever done, the court had looked long and hard into the issue, and received the testimony of a number of expert witnesses.  Typical of these was Capt. John Carslake, R.N., an explosives expert who testified on Friday, 29 August.  After confirming that although he was not an expert on cordite explosions per se, but that he ". . . knew a fair amount" about high explosive detonations, the court subjected him to a detailed examination.

"Would a 15-in or 8-in shell striking Hood's side abreast the torpedo tubes detonate a warhead with a pistol in it?" they asked.  His answer, based on trials, was "No, not unless it penetrated and detonated inside the mantlet.  If it detonated outside it would not detonate the warhead."  "Would a direct hit, either from a shell or a splinter detonate a warhead without a pistol in it?" the court inquired.  "If the shell detonated on impact with the warhead, it would detonate the warhead.  If the shell hit the warhead but did not detonate, it would not detonate the warhead . . ."  When asked what would be the probable effects if the warheads detonated, he replied:

"I would expect the mantlet, the ship's side, and the forecastle deck to be nearly demolished, but that the major venting would have been as I suggested horizontal.  Immediately after the explosion it is anticipated that the observer would have seen a gap in the ship's side, probably some 15 or 20 ft. radius down to the top of the 12-in belt.  . . . I would expect the boat deck above the tubes also to be blown away.⁷²

B.A. Fraser, Controller, summarized his particular objections to the torpedo scenario in a memo dated 7th July:⁷³

"D.N.C. has raised the question of whether the above water torpedoes in 'HOOD' were responsible for the destruction of the ship.  I disagree with his view, and accept the report of the Board of Inquiry for the following reasons:

 a) from trials, a 15-in shell burst outside a torpedo tube protected by a mantlet will not detonate the torpedo.

 b) Although a direct hit may detonate one torpedo, it is extremely unlikely that others will be countermined.  In 'KHARTOM' a torpedo was fixed⁷⁴ into the after galley by an air vessel bursting.  The head did not detonate but it burnt to detonation in the fire after a considerable period, about 20 minutes, and in 'HOOD' the interval between the first hit and the destruction of the ship seems to have been under 3 minutes .

These descriptions are consistent with the results of other experience.  To take one example, on 26 December, 1943 the destroyer U.S.S. Brownson (DD-518) was struck by a Japanese bomb which apparently detonated one or more of the 374kg torpex loaded torpedo warheads in her after quintuple torpedo mount.⁷⁵  The resulting explosion removed or completely demolished all ship structure over a radius of approximately 10 meters.  Fifteen minutes later, Brownson sank.  Although the detonation of a torpedo tube nest could result in the sinking of a destroyer in fifteen minutes, it is clearly unlikely to have caused the loss of a battlecruiser in three.  The chances of a torpedo explosion near the mainmast detonating even the closest of the ship's magazines twenty-five meters and two decks away are inconsequentially small.
 

The Boat Deck Fire:

"The 4-in. ammunition supply doors in the ship were closed until the Captain passed the order 'supply 4-in.  ammunition.' This organization resulted from an incident when the ship was bombed at the end of September or beginning of October 1939, as it was then found that 4-in. ammunition was being replenished before any order had been given.  I can visually confirm that this procedure was rigidly adhered to as I was on the boat deck during the subsequent bombings of the ship and during the action off Oran.  . . . The water-tight door organization was extremely thorough and efficient and was practiced every day at sea.  All doors were marked in plain language."

With regards to the gasoline stowage, he testified:

 When the ship proceeded to sea, all petrol, except approximately 2 gals, was left in the drifter.  This was kept behind to enable the power boats to be started up on return to harbour.  There was a special organization for landing a number of 50 gallon drums which were kept on the boat deck adjacent to the sea' lifebuoy abreast the mainmast.  The quick release drums originally supplied to the ship were released into the sea and not recovered or replaced, on the first occasion of the ship being attacked by aircraft, about 3 weeks after the outbreak of war.  The drums were stowed on trays and the stoker who looked after this stowage was so proud of his care that he told me he could not fill a petrol lighter."⁷⁷

It would thus appear that the fire on the boat deck, though spectacular, probably played no direct part in the loss of the ship.
 

A Hit from Prinz Eugen:

Speculations concerning Prinz Eugen's shells usually revolve around the idea that although her projectiles would have had little or no chance of penetrating Hood's belt armor at the specified range and obliquity, due to their "plunging" trajectory, they might have set off the after magazines after penetrating Hood's relatively thinly armored decks.  In support of this, one article purporting to reproduce a diagram of the relative trajectories of Bismarck's and Prinz Eugen's guns has been drawn so as to give an angle of fall exceeding 35 degrees.⁷⁸  A quick examination of the range tables, however, shows that such an angle is highly exaggerated, to say the least.  In reality, due largely to the higher initial velocity of Prinz Eugen's guns, at the range at which Hood was engaged the angles of fall of both Bismarck's and Prinz Eugen's guns were remarkably similar.  Over the ranges of interest, the angle of fall of Prinz Eugen's projectiles was only about 13°-19°, an angle which cannot in any meaningful sense be construed as "plunging fire."  Further, the striking velocity of Prinz Eugen's shells could not have exceeded about 460 meters per second.  Even assuming that the shells could have found a spot to hit the deck directly, at the calculated angle of fall the official German armor penetration curves for this gun, though not reproduced here, allow them a penetration of only about 40-60mm of homogeneous armor at best.  In fact, as was the case of the 380mm gun, the curves for these conditions are effectively "off the graph," strongly implying almost no possibility of intact penetration at all.  Even at the closest possible range, belt armor penetration at the calculated obliquity of 47° would have only been about 100mm for an intact projectile and 103mm for a broken one, both of which are well under the thickness of even Hood's uppermost and thinnest belt.

Although a hit from Prinz Eugen could possibly have caused the fire in Hood's after superstructure, it would have been almost impossible for such a hit to have penetrated to the after magazines.⁷⁹  Prinz Eugen may have been able to hurt the Hood, but she would almost certainly be unable to kill her.
 
 

Footnotes for Part 3:

59 All normal underwater trajectories are found to develop an upward curvature as the projectile slows.  A rough approximation of this curvature is given by the equation:

R = 2m/p A Cl y where:

R = Radius of trajectory
m = Mass of projectile
p = Density of fluid
A = Area of projectile
Cl = Lift coefficient
y = yaw angle

60 It is known for certain that German windscreens could be detached on water impact, as one fetched up on destroyer Zulu while engaging Bismarck a few days later.  This is not surprising when one considers that the instantaneous load on a windscreen upon typical water impact could well exceed fifty tons.  In fact, during World War II, the U.S. Navy had difficulty with armor piercing projectile caps and windscreens coming loose through normal handling.  A U.S. 16-inch projectile recovered from Casablanca Harbor in 1960 was found to be missing not only its windscreen but its armor-piercing cap as well.

61 Other similar nose forms which serve the same purpose are known colloquially as "spades," "pickle barrels," and perhaps somewhat more officially as "kopfrings."

62 Evidence from U.S. sources indicates that such an increase in delay would not be unusual.  For example, U.S. Navy specifications for the Base Detonating Fuze Mark 21 considered the fuze action satisfactory if the detonation occurred between 0.030 and 0.070 seconds after impact when set for a nominal 0.033 second delay.  British experience at the River Plate and elsewhere lead them to estimate the effective German fuze delay as being about 0.05 seconds.

63 Captain William Weilcose Davies, who had left Hood less than a year before her loss, testified that the torpedoes would have been stowed with pistols installed if shipped in the tubes, and that spare torpedoes without pistols but with warheads in place would have been slung above the ready use torpedoes, with the nose section run into the protective mantlets.

64 "Enclosure 'B' to D.N.O.'s Minute of .6.36 - Trials to obtain the effect of shell fire on 4.7" Q.F. fixed ammunition R.U. magazines" extracted in ADM 116/4352 pp. 386.

65 "Shellite" was the British term for a mixture of picric acid and dinitrophenol, with approximately the same explosive power as TNT.  See C. B01831 "Memorandum on Armour, Shells, Fuzes and Aerial Bombs, 1928," ADM 186/174 for more details.

66 CB 1594, "PROGRESS IN GUNNERY MATERIAL, 1921," ADM 1861251 x/L00326.  See also C.B. 01831, "MEMORANDUM ON ARMOUR, SHELLS, FUZED AND AERIAL BOMBS, 1928" ADM 186/174 ERD17626, describing complete trials with essentially the same results.

67 The full list is given and discussed in ADM 116/4351 pp. 116-119.  Evidence to support the second category was characterized as typically "scanty" in nature.

68 Bulletin of Ordnance Information, No.245, pp. 54-60.

69 I am indebted to John Howard Oxley of Halifax, Nova Scotia for bringing to mind the role that powder characteristics might have played in the disaster.

70 ADM 116/4351, pp. 105-106.

71 Raven & Roberts pp. 350-351.  A variant of this memo, reaching essentially the same conclusion and dated 2 July 1941, appears in ADM 116/4351 p. 7.

72 ADM 116/4351 pp. 385.

73 ADM 116/4351 pp. 8.

74 This is obviously a misprint for "in KHARTOUM a torpedo was fired" and has been so corrected by an unknown hand.

75 BuShips War Damage Report #51, "Destroyer Report - Gunfire, Bomb, and Kamikaze Damage.  17 October 1941 to 15 August 1945," pp. 80 et. seq.

76 ADM 116/4351 pp. 368.

77 ADM 116/4351 pp. 368 et. seq.  Astute readers will recall that AB Tilburn had earlier testified that he did not know whether the petrol drums on the boat deck had been dumped or not, but this testimony suggests that even if they had been aboard, they would have likely been empty.

78 In fact, even projectiles from Bismarck have been characterized in the popular literature in similar terms, e.g., "The fatal salvo, fired at a 19-km/12mls range [the longest range reported yet!], screamed down almost perpendicularly and went straight through Hood's unstrengthened deck into the magazine deep below the waterline." John Howard Oxley sent me this quote from John Macdonald's Great Battles of World War II.

79 Paul Schmalenbach, who should have known, himself discounted the possibility of Prinz Eugen's shells causing the fatal explosion.  He saw his shells hit Hood's boatdeck and start the fire, but stated flatly of theories that Prinz Eugen sank the Hood, "Diese Behauptung its eine reine Erfindung" [such assertions have no basis in evidence].  Vide Kreuzer Prinz Eugen, pp. 139.
 
 

Back to the INRO Index Page

This article transcribed by Tony DiGiulian.  Send  with questions or comments.
Copyright © 1999 / 2000 / 2001 / 2002 / 2003 / 2004 / 2005 / 2006 / 2007 / 2012 International Naval Research Organization and NavWeaps.com.
Last modified:  20 August 2012