Section XVII
Stability, Buoyancy and Flooding

17-1. The stability, buoyancy and trim of a submarine present a unique problem, involving three main phases: (1) operation as a surface vessel, (2) the period of submergence or surfacing and (3) underwater operations. While operating on the surface, the stability and trim of a submarine and calculations thereof are the same as for an ordinary surface ship. The calculation of stability for the submerging period is based on the same principles as for the surfaced condition but is considerably more complicated since arbitrary assumptions must be made as to the rate of flooding of various tanks, the angle of trim while the boat is diving, etc. Calculation of stability in the fully submerged condition is a very simple matter since both the transverse and longitudinal metacenters move to the center of buoyancy. The metacentric height or GM, uncorrected for free surface, is then measured both transversely and longitudinally by the vertical distance between the center of gravity and center of buoyancy. Balance calculations for weight and buoyancy changes in the submerged condition involve only weight and moment computations and are familiar to all submarine officers.

17-2. War experience has repeatedly demonstrated that changes in stability and weight and buoyancy distribution on submarines will frequently be caused as a direct or indirect consequence of damage and that where such changes are relatively large, ship control submerged may be lost or excessive list and trim angles may result on surfacing. In several instances our submarines have been subjected to serious military hazard following damage sustained while submerged due to inability to retain depth control, forcing bottoming or surfacing, and in at least one instance an excessive list assumed by a U.S. submarine upon surfacing after damage would probably have caused her loss had the enemy been more alert (paragraph 17-4).

17-3. Dragonet, Salmon and Scamp are among those submarines which assumed large lists upon surfacing after receiving damage.1 In each of these instances, the lists were caused by unsymmetrical liquid loading in external tanks. Dragonet was reported to have assumed approximately 15 degrees port list due to inability to expel water more than a few feet below the tops of two ruptured port main ballast tanks, a fact unknown to her at the time the remainder of the ballast tanks were blown dry on surfacing. In Dragonet's case, the list caused by the heeling moment of the two partially flooded port tanks was also slightly aggravated by the free surface effect of the water in the flooded forward torpedo room. However, calculations indicate that the reduction in metacentric height resulting from the free surface and added weight of the water in the forward torpedo room, considering the water level to have been approximately two feet above the forward torpedo room flat, was only about 0.18 foot. This was a minor reduction compared with


Dragonet's total probable GM before damage of about 1.50 feet,2 and caused the list to increase by only about one degree. Calculations also indicate that the list was probably of the order of 10 degrees rather than the reported 15 degrees. Scamp assumed about 17 degrees port list upon surfacing, due to inability to expel water from two port main ballast tanks and the displacement of fuel oil in one port fuel ballast tank with heavier sea water through a rupture in the tank top plating. Salmon assumed about 15 degrees starboard list on first surfacing due to leaking starboard ballast tank master vent valves and the displacement of fuel oil in one starboard fuel ballast tank with sea water through a damaged riser. In both of these latter cases, the lists were no doubt increased to a minor extent by loss of GM due to the free surface effect of the sea water which accumulated in internal compartments from various sources of leakage.

17-4. It should generally be expected that a submarine which sustains damage to external tanks, risers, master vent valves or the low pressure blow system to individual tanks will almost invariably assume a large list upon first surfacing. This will occur primarily because the extent and nature of such external damage will not be known while still submerged and will manifest itself only when the ship actually does list upon breaking the surface or after following normal tank blowing procedure. Large lists assumed upon surfacing should not normally be dangerous since rapid compensation can of course by made either by venting ballast tanks on the high side or by selective blowing of ballast tanks on the low side. It has been demonstrated on several occasions that submarines can quickly get underway on their surface Diesel propulsion plants even with lists as high as 15 degrees. Scamp and Dragonet are two such examples. However, should a large list persist on a submarine after surfacing, and if the submarine were unable to resubmerge if attacked during this period, due to damage, low air reservoirs or depleted batteries, gun operations and ship handling would certainly be hampered, if not impossible. For example, Salmon would very probably have been destroyed after surfacing in her action of 30 October 1944 had the enemy escorts not delayed their attack until the ship's force had taken damage control measures and successfully righted the boat.

17-5. Calculations made for SS285 Class submarines show that a boat which is initially in normal diving trim will assume a 45 degree list if all main ballast and fuel ballast tanks on one side only are completely flooded to the tank tops while the corresponding tanks on the opposite side remain dry. This might be considered to represent the maximum possible listing moment for a surfaced submarine. Calculations have also been made to determine what angle of heel might result in the hypothetical case occurring if all main ballast tanks on one side only (not including fuel ballast tanks) were blown completely dry while submerged. The boat would then surface rapidly, of course, and while rising would steady at a list of about 70 degrees, assuming the boat rises through a sufficient distance for equilibrium conditions to occur before surfacing. On surfacing, the list would be sharply reduced to about 33 degrees since the introduction of water plane area greatly increases the


metacentric height (note that in this case only 33 degrees list results after surfacing rather than the 45 degrees calculated for the case above, since the fuel ballast tanks on both sides are considered to remain flooded).

17-6. The freeboard and reserve buoyancy of U.S. submarines now in service are sufficient even in emergency diving trim condition to permit several main ballast tanks to remain free flooding for list control, when necessary. However, if a large number of ballast tank master vent valves or risers are damaged, a submarine may then experience difficulty after surfacing in blowing up far enough to achieve satisfactory seaworthiness. For example, Salmon remained just barely awash for several minutes after surfacing due to loss of tank air through leaking master vent valves and was not able to blow up until the emergency vent valves were closed. This delay, in addition to her large list, hampered damage control efforts and gun operations.

17-7. While returning to base with her forward torpedo room partially flooded, Dragonet reported reaching a roll of 63 degrees from upright to one side as measured by the control room inclinometer.3 At the time this occurred, heavy quartering seas and winds of gale force were being encountered. There are many other similar cases on record where submarines have reported extreme rolls. For example, Snapper (SS185) on her eighth war patrol and Seadragon (SS194) on her shakedown cruise after commissioning reported rolls reaching 45 degrees or more at times. Also, while operating off Newfoundland on 17 November 1946, Dogfish (SS350) and Sablefish (SS303) reported maximum rolls from upright to one side of 52 degrees and 63 degrees, respectively, and average rolls from upright to one side of 35 degrees and 20 degrees, respectively. In all of these five cases the rolls were measured by inclinometer. It should be noted that this type of instrument is not accurate for recording large angles of roll and will in such cases indicate somewhat greater angles than actually occur. The circumstances attendant at the time the extreme rolls occurred were quite similar in each case: high wind velocities and heavy quartering seas. The periods of roll were long and the vessels seemed to "hang" at the end of the large angle rolls before returning to upright. In three of the five cases, water was shipped through the outboard induction valves and conning tower hatches. Sablefish reported that on her 63 degree roll it was necessary to blow ballast tanks on the low side with high pressure air in order to right the ship. Dragonet reported that on her 63 degree roll it was necessary to swing the ship's heading through 105 degrees before the vessel slowly commenced to right herself and then steadied at about 20 degrees list. With the exception of Dragonet, all of the above vessels are believed to have been in emergency diving trim condition.

17-8. As nearly as can be determined, Snapper and Seadragon each had a GM of approximately 1.00 foot and Dragonet, Dogfish and Sablefish each had a GM of 1.25 feet or larger. As mentioned previously, the loss in stability on Dragonet due to the flooded forward torpedo room was minor, having the effect of reducing her GM from about


1.50 feet to 1.32 feet. These values of GM are more than sufficient to assure a range of stability considerably beyond 90 degrees and it is difficult to explain the tendency of the submarines cited above to "hang" at extreme angles. This "hanging" behavior is similar to that which is often reported by destroyers and minesweepers, or occasionally even larger surface vessels, under identical conditions, ie., heavy quartering seas. The extreme rolls evidently occur as the vessels roll away from the crests of waves which are momentarily in synchronism. All vessels are subject to this action in a seaway under the right conditions. The extreme example is the broaching of landing craft in a surf. The smaller the vessel, the larger the part that must be played by seamanship to avert disaster. On a submarine, the tendency to "hang" under certain wave and wind conditions is probably aggravated by the fact that the tank tops and part of the superstructure deck will be submerged at large angles of roll, these surfaces tending to act as a plane to hold the submarine at the extreme angle of roll while it is displaced sideways due to the dynamic action of the waves.

17-9. Another factor which may contribute to the tendency to "hang" is that at large angles of roll it is possible for a submarine to spill air from ballast tanks on the low side and take on water through the flood openings, with resultant increase in heeling moment. A study made in 1942, taking into account the loss of air and consequently the partial flooding of low side main ballast tank, indicated that statically the effect on stability was maximum between 50 degrees and 60 degrees of heel but was only a minor consideration, the loss of righting arm being less than 0.10 feet. This study also indicated that the static stability improved beyond 60 or 70 degrees due to the addition of low weight by partial flooding of the low side tanks. The additional heeling effect of such weight would be large, of course, and tank blowing may be requited in order to again right the ship. Tank blowing will only partially remove the water in the low tanks at extreme angles but should have some beneficial effect and is probably the correct course of action to take when such a case is believed to exist. It should also be noted that when air spillage occurs from low side tanks, the submarine upon righting itself will not return to the vertical but will assume a list in the direction of the low side tanks, due to the added water in those tanks, and tank blowing will then become necessary to reach the fully upright condition. Whether "hanging" at an excessive angle is due only to the synchronism of dynamic wind and sea forces, or whether progressive loss of tank air with increasing roll angle contributes appreciably to the "hanging" tendency, is not clear at present. Rolling, heaving and pitching, accompanied by heavy seas, will result in variation in head over the tank flood openings and will tend to cause loss of tank air, allowing partial flooding which will in turn reduce stability due to increased free surface effects.

17-10. Although submarines have a very poor hull form for surface transverse dynamic stability, nearly circular in cross section with very little freeboard and extreme tumblehome, the possibility of capsizing


through loss of stability is remote. They have a positive righting moment over a wider range than most ships due to a lower center of gravity and have a smaller area exposed to the force of the wind. Submarines now in service have a range of stability beyond 90 degrees and, in addition, the hull is kept more completely closed in normal operation than in other types of vessels, the conning tower hatch and engine air induction outboard valve being the only openings. The criterion of surface stability in current use for modern fleet type submarines is that the GM in normal diving trim condition shall not be less than 1.25 feet. This figure wad decided upon as a result of Bureau studies and wind heeling moments combined with synchronous rolling, and the curves of righting moments of fleet type submarines. It results in righting arms throughout the range of stability which provide a considerable margin of safety over those of Snapper at the time of her reported heavy rolling.

17-11. When submerging or surfacing submarines pass through a period of low metacentric height. This condition occurs when the moment of inertia of the reduced waterplane reaches a low value concurrently with free surface effects due to partially filled ballast tanks. In modern double-hull submarines, this point of minimum GM occurs when the ballast tanks are about 70 to 80 per cent full. The detail specifications for these vessels require a minimum GM of 3 inches when submerging or surfacing. However, even with slightly negative GM when trimming down, the righting moment soon becomes positive as angle of heel increases and no real danger exists for the vessel. In addition, the condition of minimum stability is transitory, existing for probably less than a second during a normal dive.

17-12. Damaged stability and flooding studies have been made for submarines as for other combatant vessels. The amount of flooding a submarine can sustain and still operate submerged is limited at most to the relatively small capacity of the variable ballast, safety and negative tanks plus a small additional amount which can be compensated for by use of the diving planes and angle on the boat at high submerged speeds. Usually the amount of flooding which can occur while submerged without causing loss of depth control will be much less than the capacity of the above tanks since any longitudinal moments introduced by the flooding must also be balanced by the use of the same tanks, assuming that diving planes and angles on the boat are not used for moment compensation. For surface conditions, however, a submarine can absorb a considerable amount of flooding of internal compartments and tanks before upending or capsizing of the vessel will occur, particularly when the flooding occurs in amidships compartments and the submarine is in normal rather than emergency diving trim, i.e. when the fuel ballast tanks are empty.

17-13. The table below presents several hypothetical cases of flooding for SS285 Class submarines and the resulting effects


on transverse stability and longitudinal trim when surfaced. The calculations were made for both the normal and emergency diving trim conditions. The submarine is considered to upend when the longitudinal moment tending to trim the ship after flooding remains greater than the opposing longitudinal moment of buoyancy until a near vertical position is reached. The initial transverse GM in normal diving trim was taken as 1.63 feet and in emergency diving trim as 1.59 feet, both figures being considerably greater than corresponding values for SS285 Class submarines now in service. Therefore, the figures given for lists resulting from off-center flooding are somewhat lower than would actually be the case for SS285 Class submarines in service.

View of submarine compartments

Spaced Flooded
(to new waterline)
Normal Diving Trim Emergency2 Diving Trim
Trim on Boat
After Flooding Occurs
Loss in
Transv. GM
Trim on Boat
After Flooding Occurs
Loss in
Transv. GM
Forward Torpedo Room 4'4" By Bow 0.43' 6'11" By Bow 0.61'
Forward Torpedo Room
and MBT No. 1
7'1" By Bow 0.17' Ship upends ----
Forward Torpedo Room
and For'd Battery Comp't.
Ship upends ---- Ship upends ----
Forward Battery Compt.
Control Room MBT No. 2A
6'8" By Bow
(List 5.3° Stbd)
0.38' Ship upends ----
Control Room After
Battery Comp't FBT NO. 3A
2'7" By Bow
(List 7.7° Stbd)
0.59' Ship upends ----
After Battery Comp't.
For'd Engine Room
FBT No. 5A1
1'9" By Stern
(List 9.5° Stbd)
0.77' 1'7" By Stern
(List 3.6° Stbd)1
For'd Engine Room
After Engine Room
MBT No. 6C
5'8" By Stern
(List 12.6° Stbd)
1.04' Ship upends ----
After Engine Room
Maneuvering Room
7'1" By Stern 1.01' Ship upends ----
After Torpedo Room 4'3" By Stern 0.56' 4'9" By Stern 0.78'
After Torpedo Room
MBT No. 7
6'6" By Stern 0.77' 9'0" By Stern 0.77'


The above table emphasizes the marked increase in ability of a surfaced submarine to absorb flooding when in normal diving trim as compared with emergency diving trim, due to the additional freeboard and reserve buoyancy of the lighter load condition. The extreme susceptibility of a damaged submarine to upend through reduction of longitudinal stability should also be noted. A submarine will usually be lost by upending, rather than by capsizing or bodily sinkage. These studies indicate that recent types of submarines, while surfaced and in emergency diving trim, i.e., in usual wartime surface condition, can probably survive the flooding of one major compartment and adjacent tanks with the possible exception of flooding of the forward torpedo room. In the latter case, blowing the bow buoyancy tank, if such remains possible after damage, should assure that upending will not result.

17-14. Also to be considered as a possible damaged condition is the flooding of the conning tower while submerged combined with the damage control measure of blowing the safety tank to maintain neutral buoyancy submerged. For modern submarines, the loss of GM will be only about 0.13 feet when submerged, and 0.18 feet after surfacing. In either case, this slight decrease in GM will not have a serious effect on the ship. It should be noted that the capacity of the safety tank is purposely designed to be slightly greater than that of the conning tower so that it can provide a rapid means of compensating for a flooded conning tower.

17-15. The longitudinal GM of a completely submerged submarine is very small as compared to the surfaced condition, being essentially equal to, or, considering free surface effect, slightly less than the transverse GM, both being a function of the distance between the center of gravity and the center of buoyancy. Consequently, relatively small changes in longitudinal moment will have relatively large effect on trim and therefore on the ability to maintain satisfactory ship control when submerged. In addition, changes in total weight will place a submarine in a condition of either positive or negative buoyancy and will affect depth control. When considering the ability of submerged submarines to sustain war damage, it is apparent that this sensitivity to small changes in moment and weight constitutes one of their most vulnerable aspects.

17-16. The flooding of a submarine major compartment when submerged is a serious matter. If the damaged compartment is at either end of a submarine, and if flooding is rapid, blowing of all ballast tanks to surface would probably result in serious loss of air and in upending. Under such conditions the pressure in the tanks at the high end is less than in the low end and the blowing would therefore unwater the tankage at the high end first and increase the trim angle to such an extent that the air would be spilled out of the flood openings and blowing would become comparatively ineffective. It would be far better to first blow the tankage at the low end and decrease the trim angle before unwatering the tankage at the other end. In the case of Squalus, it is believed that had only the after tanks been blown at first, the ship might have been surfaced.


If a submarine settles on the bottom after flooding a major compartment, and assuming the bottom at a depth less than the designed collapse depth of the hull or of the watertight compartment bulkhead, various methods of self-salvage may be attempted. The first and most logical method is, of course, to dewater the flooded compartment by use of the internal salvage air system, as was successfully accomplished in Dragonet's case.4 If, for some reason, dewatering of the flooded compartment is not feasible, then it is still possible to surface the submarine, without upending, by calculating moments and blowing only those tanks which will provide both a slight margin of reserve buoyancy and an overall longitudinal moment balance such that reasonable trim control can be maintained while rising, either by use of propulsion and the diving planes or by such means as shifting men and pumping, blowing or flooding forward and after variable tanks. In the case of a flooded end compartment, or where more than one compartment is flooded, normal fuel oil tanks should be considered for blowing as well as main ballast, fuel ballast and variable tanks. Blowing should progress from the damaged end. After surfacing, certain additional tanks may then be safely blown to provide additional freeboard and reserve buoyancy since the longitudinal GM will increase tremendously when water plane area is once established.

17-17. A interesting example of flooding in amidships compartments is the experience of Crevalle (SS291) during her fifth war patrol. On 11 September 1944 Crevalle surfaced with her ballast tank vents unintentionally still open and the bridge was manned. Fifteen seconds later the ship took a large down angle and resubmerged with the upper and lower conning tower hatches still open. The ship was making standard speed on battery propulsion at this time and as she submerged the ahead motion accelerated her descent and the down angle increased rapidly. Water poured into the conning tower and thence to the control room and pump room. The forward and after bulkhead doors of the control room were quickly closed and prevented flooding forward and aft. At 150 feet the upper hatch seated but by this time the water in the conning tower had reached armpit depth, the pump room was completely flooded and the control room was flooded above the door to the forward battery compartment (at the large down angle). At about this time, full speed astern was ordered by a quick thinking MoMM 1/c who had tried in vain to contact the control room and conning tower and then telephoned the maneuvering room. With propulsion reversed, descent was checked at 190 feet with a down angle of about 42 degrees. The ship then surfaced, backing full all the way. Since all electrical equipment in the conning tower, control room and pump room was grounded out, Crevalle was forced to terminate her patrol at once.

17-18. In over twenty-five instances during World War II, U.S. submarines suffered marked loss of ability to retain depth control after sustaining damage from enemy action while operating submerged. In most of these cases, loss of control was only temporary and had no serious consequences. In several instances, however, loss of depth


control was prolonged or permanent and resulted in oscillations over a wide range of depths, broaching, bottoming, or forced early surfacing with consequent increase in hazard.

17-19. Derangements which prevented use of either or both the bow and stern planes for short periods of time were responsible for most of the cases of temporary loss of depth control, although added weight due to relatively minor flooding was sometimes an additional or the sole factor involved. The instances in which serious or prolonged loss of depth control was experienced were for the most part caused by the additive effect of several of the following factors: flooding of engine air induction and ship's supply ventilation systems; flooding of sanitary, trim and auxiliary tanks; relatively large intake of water into one or more compartments through various sources of leakage; loss of propulsion power; temporary or permanent derangement of bow and stern planes; and displacement of fuel oil with heavier sea water in damaged external fuel tanks. The suddenness with which damage occurred was often a contributory cause of loss of depth control. For example, rapid flooding of an induction system or sudden disabling of diving planes when set on large rise or dive angles frequently resulted in a boat assuming a considerable angle or large change in depth before corrective measures could be taken. In such cases restoration of control therefore became more difficult. The outflow of water ahead of the expanding gas bubbles of close underwater detonations in some cases also forced submarines to assume large angles of inclination or changes in depth (See paragraph 15-8).

17-20. The following instances of loss of depth control are particularly interesting and serve to illustrate typical situations which submarine officers may expect to occur following war damage: Kingfish (SS234), Section V; Tunny (SS282), Section VI; Salmon (SS182), Section VII; Scamp (SS277), Section IX; Sargo (SS188), 4 March 1942, Appendix I; Jack (SS259), 26 June 1943, Appendix I; Gurnard (SS254), 6 July 1943 Appendix I; (SS218), 10 November 1943, Appendix I; Tinosa (SS283), 22 November 1943, Appendix I; Bluegill (SS242), 22 May 1944, Appendix I; Sailfish (SS192), 4 November 1944, Appendix I; and Hawk bill (SS366), 18 July 1945, Appendix I. The methods which were employed to regain depth control varied with the circumstances of each case. Weight compensation by flooding, venting pumping or blowing variable, safety and negative tanks was resorted to whenever the tactical situation permitted. With extreme weight and trim changes, submarines were forced to blow ballast tanks or back with both propulsion shafts in order to check downward descent. After becoming heavy overall as a result of damage, several submarines managed to retain fairly satisfactory depth control by resorting to fast speeds ahead at large upangles. Two such examples are the experiences of Sculpin (SS191) on 28 September 1942 and Whale (SS239) on 30 October 1942, briefs of these actions being contained in Appendix I. It should be noted that once a submarine becomes heavy and has to use pumps, air or high speed to maintain depth control, the resulting noise increases the possibility of detection by enemy vessels using sonic listening gear.


17-21. The maximum extent of flooding which a submarine can sustain and still operate submerged depends on the ship's ballasting, condition of loading, location of flooding and ability to maintain depth control by use of planes, "bubbles" in tanks, and angles on the boat at high speeds. Flooding introduces added weight plus a trimming moment both of which must be counteracted if depth control is to be retained. Weights added in amidships areas are obviously easier to compensate with the means available than weights added either forward or aft since the safety tank, which is amidships, can be blown, the resultant of the lift forces obtained by putting an up angle on a submarine at ahead speeds is exerted amidships, and the trimming moment introduced by an amidships weight is much less than an equivalent weight added forward or aft. Calculations based on SS475 Class model tests indicate that an added weight amidships of at least 50,000 pounds can be compensated by proceeding at an ahead speed of six knots with an up angle of 20 degrees on the boat.5

17-22. Several interesting tests were made on Entemedor (SS340) during April 1946 to determine how quickly the descent or rise of a heavy or light submarine could be stopped when not under propulsion.6 The submarine was first balanced dead in the water. Then, with the boat about 5000 pounds light or heavy, and allowing sufficient time for the limiting velocity of rise or descent to be reached, it was found that in every case the vertical motion could be stopped in from 10 to 20 feet of travel by cancelling the 5000-pound force and adding 5000 pounds in the direction opposite to the motion. A test was also conducted during this same period to determine the hull angle required on Entemedor to counteract the effect of the bow planes at various angles so that depth control could be maintained with the stern planes alone At speed 6.4 knots and with neutral buoyancy, it was found that if the bow planes were locked on 25 degrees rise angle, a 5 degree down angle was required on the boat to maintain constant depth. With the bow planes locked at 25 degrees dive angle, a 2.5 degree up angle on the hull was required. Since there are many war damage cases where bow planes have been temporarily or permanently disabled, such as occurred on Tunny (Section VI), it is interesting to note that even with the bow planes locked on hard rise or dive angles the effect on depth control is not serious provided the stern planes and propulsion remain available.

17-23. As the safety tank is a pressure tank located near the center of balance and has a capacity of about 20 tons which can be quickly blown with high pressure air, it is ideally suited for rapid compensation of a submarine which suddenly becomes heavy overall due to damage. Upon recommendation by ComSubPac in May 1943, the installation of liquidometer gauges in the safety tank of all fleet type submarines was authorized to permit finer control when necessary to blow the safety tank partially and quickly, and to afford a means of avoiding the release of an air bubble through over-blowing if ever necessary to blow the safety dry during evasion maneuvers.7 The inboard vent lines from this tank were also cross-connected and led into the control room to permit operation at that location.8


Table of Contents
Previous Section (XVI) * Next Section (XVIII)


1. See Sections VII, IX and XIV for narratives of Salmon, Scamp and Dragonet.

2. Condition VII-B, Emergency Diving Trim, Revised Inclining Experiment Data for Dragonet.

3. See paragraph 14-18.

4. See Section XIV for Dragonet's narrative.

5. Experimental Towing Tank conf. report No. 309 of December 1946, prepared by the Stevens Institute of Technology (Hydrodynamic Coefficients of a Self-Propelled Model of the SS475 Class Submarine).

6. Preliminary Report entitled "Performance Characteristics of Submerged Submarines, Tests on U.S.S. Entemedor (SS340)", dated 23 April 1946, prepared by The Woods Hole Oceanographic Institution.

7. ShipAlt SS206 of 5 December 1943.

8. BuShips ltr. SS/S48-23(515) of 13 August 1943.

Transcribed and formatted for HTML by Patrick Clancey, HyperWar Foundation