C.A.P. 127

MINISTRY OF TRANSPORT AND CIVIL AVIATION

CIVIL AIRCRAFT ACCIDENT Report of the Court of Inquiry into the Accidents to Comet G-ALYP on 10th January, 1954 and Comet G-ALYY on 8th April, 1954

LONDON: HER MAJESTY'S STATIONERY oFFICE EIGHT SHILLINGS NET MINISTRY OF TRANSPORT AND CIVIL AVIATION

CIVIL AIRCRAFT ACCIDENT Report of the Court of Inquiry into the Accidents to Comet G-ALYP on 10th January, 1954 and Comet G-ALYY on 8th April, 1954

LOND ON: HER MAJESTY'S ST ATIONER Y OFFICE 1955

House ofLords. 1st February, 1955. SIR, Your predecessor directed that Public Inquiries should be held into the causes and circumstances of two accidents which occurred in the Mediterranean to Comet aircraft, the first on the 10th January, 1954, to G-ALyP, the second on the 8th April, 1954, to G-ALYY. In pursuance of that direction I was, on the 8th September, 1954, appointed by the Lord Chancellor to be the Commissioner to hold the Inquiry and Sir William Scott Farren, C.B., M.B.E., F.RS., Professor William Jolly Duncan, C.B.E., D.Sc., F.RS., and Air Cottlmodore AlIen Henry Wheeler, O.B.E., were appointed Assessors. I now have the honour to present my Reports, each of which has been signed by all three Assessors to signify their agreement with it. With the agreement of the parties the two Inquiries were conducted at the same time in order to avoid unnecessary duplication. I have attached as appendices to the Report on the accident to G-ALYP: I. lists of the witnesses who gave evidence at the Inquiry, IT. lists of the parties represented thereat and of those representing them, lIT. dates and places of the public hearings. As will be seen from Appendix IT there were present accredited representatives of the Government of , Colonel R Miniero .and Signor R. Roveri and accredited representatives of the Government of the Union of South Africa, -Colonel L. E. Lang and Major J. J. Granzier. A copy of the transcript of the oral evidence and addresses of Counsel will be available to you and I have handed to the Branch of the Chief Inspector of Accidents all the documents and other exhibits that were put in evidence except that I have returned to the. Director of the Royal Aircraft Establishment, Farnborough, the metallic exhibits which emanated from that establishment. The case was most carefully prepared and very clearly presented on behalf of the Attorney-General. It is right that I should place on record that every assistance within their power was rendered to the Court by the makers of the aircraft, Aircraft Company Limited, the operators, British Overseas Airways Corporation and the Air Registration Board. The contentions of all the parties represented were put forward with cogency and force and at the same time with fairness and moderation. All parties displayed a genuine desire to arrive at the true causes of the accidents and to make constructive suggestions which would enable a re-designed Comet aircraft to take the air once more with safety to the public and benefit to the nation. In exercise of the powers conferred on me by Section 9 (9) of the Civil Aviation (Investigation of Accidents) Regulations, 1951, and with the consent of all parties represented I allowed affidavits from certain witnesses resident in Italy to be used as evidence and these affidavits are included in the documents handed to the Branch of the Chief Inspector of Accidents. All possible assistance was rendered to the tribunal by the Chief Inspector of Accidents and his staff. In particular he arranged at an early stage in the hearings for the Court and the parties represented to visit the Royal Aircraft Establishment at Farnborough. We saw for ourselves the reconstructed wreckage of aircraft G-ALyP and heard an account of the reconstruction and of the tank experiment, with the tank and the aircraft G-ALYU before our eyes. The Assessors and I are most grateful to the Director for the arrangements he made. Without this experience it would have been impossible for us adequately to appreciate the evidence which was afterwards given. In Inquiries of this kind a legal Commissioner must always be very dependent on his expert advisers. I think this was particularly so in the present investigation having regard to the suspected cause of the accidents. Be that as it may, I cannot too strongly express my deep sense of gratitude for the help I have derived from the expert knowledge and practical experience of my Assessors. I am glad to find that there is no material point on which they have failed to agree with one another and I have felt no difficulty in accepting the advice they tendered to me. . I cannot conclude this letter without expressing my very sincere thanks to Mr. J. N. B. Penny, barrister­ at-law, who was specially appointed secretary to the tribunal and has been of great assistance to me throughout the proceedings. I have the honour to be, Sir, Your obedient Servant, (Signed) COHEN

The Right Honourable J. A. Boyd-Carpenter, M.P., Minister of Transport and Civil Aviation 3 94722 A2

CONTENTS REPORT ON THE ACCIDENT TO COMET G-ALYP Paragraphs I. Introductory ... 1-10 IT. History of the Comet Project 11-25 lIT. The Accident 26-28 IV. The Aircraft ... 29-31 V. The Crew 32-39 VI. Passengers and Cargo 40 VII. Pre-flight Incidents ...... 41-43 VIII. Weather Conditions at the time of the Accident 44 IX. Action taken after the accident and prior to the accident to Comet G-ALYY (a) Local salvage and medical investigation 45-46 (b) Action taken by the Ministry of Transport and Civil Aviation 47-48 (c) Naval Search for Wreckage 49-51 (d) The Abell Committee 52-53 (e) Resumption of Comet Services .. . 54-57 X. The accident to G-ALYY (Yoke Yoke) .. . 58 XI. Investigation of the accidents to G-ALyP and G-ALYY (a) Investigation by RAE 59-79 (b) Investigation by the de Havilland Engine Company Limited 80-91 XII. The RAE Report ... . 92-94 XIII. The Court's Conclusions as to the cause of the accident (a) The main finding of the RAE Report ...... 95 (b) The alternative suggestion made by Mr. B. Jablonsky 96-104 (c) Mr. Tye's evidence 105-106 (d) The possibility of over-pressurisation ... 107 (e) Certain,defects referred to in the RAE Report 108-111 (f) The possibility of damage by jet efflux ... 112-113 XIV. Responsibility (a) Introductory 115 (b) Criticism of de Havillands' design work 116-117 (c) Memorandum by Assessors 118-129 (d) Criticism of de Havillands' repeated loading tests in 1953 130 (e) de Havillands' method of dealing with cracks ... 131-139 XV. Future (a) Statements on behalf of the Attorney-General and de Havillands 140 (b) Further suggestions directed to guarding against 141-145 (c) Use of available Government facilities .. . 146 (d) Avoidance of flight by unlicensed crew ...... 147 (e) Suggested Scientific and technical investigations 148-155 (f) Observations on certain suggestions made in the course of the Inquiry 156-163 Pages XVI. Questions and Answers 32-33

REPORT ON THE ACCIDENT TO COMET G-ALYY I. Introductory Paragraphs (a) Matters in common with the Report on G-ALYP 1-2 (b) Arrangements with South Mrican Airways . 3 IT. The Accident 4-5 lIT. The Aircraft ... 6-11 IV. The Crew 12-18 V. The Passengers and Cargo ... 19 VI. Pre-flight Incidents ... 20-24 VIT. Weather Conditions at the time of the Accident 25 VIII. Action taken after the Accident ... 26-28 IX. The Court's Conclusion as to the cause of the Accident 29 X. Responsibility 30-31 XI. Future 32 Pages XII. Questions and Answers 47-48 Appendices I-VIII ... 35-42 Figures 1-19 ... Facing page 24 5

~

THE CIVIL AVIATION ACT, 1949 THE CIVIL AVIATION (INVESTIGATION OF ACCIDENTS) REGULATIONS 1951

Report of the Public Inquiry into the causes and circumstaRces of the accident which occurred on the 10th January, 1954, to the Comet aircraft G-ALYP AIRCRAFT: Comet G-ALYP ENGINES: Four de Havilland Ghost 50 REGISTERED OWNERS AND OPERATORS: British Overseas Airways Corporation CREW: Captain A. Gibson-Killed First Officer W. J. Bury-Killed Engineer Officer F. C. Macdonald-Killed Radio Officer L. P. McMahon-Killed Steward F. L. Saunders-Killed Stewardess J. E. Clarke-Killed PASSENGERS: 29-All Killed PLACE OF ACCIDENT: Over the Mediterranean off Elba. DATE AND TIME: 10th January, 1954, at about 1000 G.M.T.

All times in this Report are G.M.T.

PART I consist of two members appointed by the Minis­ INTRODUCTORY ter and sixteen other members. Of these sixteen four must represent operators of aircraft. four (a) Definitions must represent constructors of aircraft, four must 1. In this Report the following expressions bear represent insurers engaged in aircraft insurance the following· meanings:- business and the remaining four are co-opted. It is " ARB." means the Air Registration Board provided that of the two members to be nomi­ incorporated as a company limited by guaran".. nated by the Minister one is to be an independent tee under the Companies Act. 1929, on the person and the other a person who has had not 26th February. 1937. less than five years' professional experience as " AS.B." means the Air Safety Board a pilot of civil aircraft. It is further provided that appointed by the Minister of Transport and the co-opted members are to be persons represen­ Civil Aviation. tative of some interest connected with civil aviation. "de Havillands" means the de Havilland Aircraft Company Limited. 3. By section 7 of the Civil Aviation Act, 1949. which reproduces section 2 of the Air Navigation " R.AE." means the Royal Aircraft Estab­ Act. 1936. it is provided that the Minister may lishment controlled by the Minister of Supply. by order provide for delegating to a body appear­ " B.O.AC." means British Overseas Airways ing to him to be constituted as is ARB. under Corporation. its Memorandum and Articles. such of the admin­ istrative functions of the Minister with respect (b) The Air Registration Board to the matters mentioned in the subsection as 2. The primary object of A.R.B. is to carry out may be specified in the order and for entrusting such administrative and advisory functions with to that body such advisory functions in connec­ regard to the design. construction and mainten­ tion with any of such matters as may be ance of aircraft and matters connected therewith specified. as may from time to time be delegated to A.RB. 4. In pursuance of this section the Minister by by the Minister of Transport and Civil Aviation. the 'Civil Aviation (Air Registration Board) Order Under its Articles of Association A:R.B. is to of 1951 (which replaces Orders wade under the 1936 Act) delegated a number of his adminis­ Council are identical with the members of the trative functions to A.R.B. and entrusted to it Board. The Council are advised by a technical certain advisory functions. Under section 1 of staff of about 125 of whom about 84 are em­ the Order the Minister delegated to A.R.B. the ployed on inspectional duties. The Chief Execu­ following functions (inter alia):- tive Officer is Mr. R. E. Hardingham and the "(a) the formulation and publication of tech­ Chief Technical Officer of the Board is Mr. W. nical'requirements as regards the design, Tye. construction and maintenance of aircraft 6. To enable A.R.B. to discharge its functions and engines, components, accessories, it prepares and from time to time publishes instruments, equipment and apparatus of detailed requirements which' inform manufac­ aircraft; turers of the minimum conditions with which, (b) the investigation of aircraft (including prima facie, they have to conform if they are to their engines, components, accessories, in­ obtain a Certificate df Airworthiness. To assist struments, equipment and apparatus (ex­ A.R.B. in the preparation of these requirements cluding radio apparatus) and the manner they have appointed an " Airworthiness Require­ of the installation of the same) for the pur­ ments Co-ordinating Committee" which includes representatives of-the Ministry of Supply, R.A.E., poses of the issue and renewal of certifi­ manufacturers of aircraft, operators of aircraft cates of airworthiness or of validations and A.R.B. itself. of such certificates and for the purposes of the variation of particulars and conditions 7. Requirements are not, however, treated by specified in such certificates or any flight A.R.B. as being as immutable as the laws of the manual or performance schedule issued Medes and Persians. On the one hand, during therewith ; 'the development of a new type, requirements more exacting than those prescribed in the pub­ (c) the making of recommendations to the lished regulations are often imposed or adopted Minister as to the issue of certificates of by the manufacturer concerned. On the other airworthiness and of validations of such hand, on occasions certain deviations from the certificates and as to the variation of par­ prescribed conditions are accepted by A.R.B. pro­ ticulars and conditions specified in such vided that they are satisfied that the safety of the ,certificates or any flight manual or per­ aircraft is not thereby jeopardised. formance schedule issued therewith; (d) the renewal of certificates of airworthiness (c) The Air Safety Board and of validations of such certificates and 8. A.S.B. is a purely advisory body and has no to such extent as may be determined by statutory authority behind it. It was appointed in the Minister in writing the variation of November, 1946, with the following terms of particulars and conditions specified in reference: "To keep under continuous review such certificates 0r any flight manual or the needs of safety ,in British civil aviation and performance schedule issued therewith ; to recommend measures calculated to promote (e) the making of any investigation required safety in respect of both (a) the operation of in connection with an application for a British civil aircraft throughout the world, and special permission for an aircraft to fly (b) the efficiency of the system of ground facili­ without a certificate of airworthiness being ties provided for civil aircraft of all nations in force in respect thereof and the making operating over the United Kingdom." Its mem­ of recommendations to the Minister as to bers are appointed by the Minister and at the the giving of such a special permission; material date consisted of Air Chief Marshal (f) the approval of engines for aircraft; Sir Frederick Bowhill, Lord Brabazon, Sir (g) the making of inspections of organisations Leonard Bairstow, Air Commodore Banks and of persons or firms desiring to furnish re­ Mr. (now Sir) Amold Hall. ports or certificates as to compliance by aircraft and engines, components, acces­ (d) The Royal Aircraft Establishment sories, instruments, equipment and 9. R.A.E. is controlled by the Minister of apparatus of aircraft with airworthiness Supply. The main establishment is at Farn­ requirements, the approval of any such borough but there are branch establishments in firm or persons as qualified to furnish other parts of the country. In this Report I am such reports of certificates, and the mainly concerned with the work done at Farn­ acceptance of such reports or certificates;". borough. The Director of R.AE. is Sir Amold 5. The Chairman of A.R.B. is the Rt. Hon. Hall. The Head of the Structures Department is Lord Brabazon of Tara. The members of the Dr. P. B. Walker. The only other member of \s the staff who need be mentioned by name is 12. At some date in 1951 it was arranged that Mr. E. L. Ripley who was responsible for the the first two prototypes should be delivered to work in connection with the reconstruction and the Ministry of Supply but that the remaining investigation of the wreckage recovered after the aircraft to be supplied under the contract should accident. I should, however, add that R.AE. be delivered to B.O.AC. and that the approval has its own flight testing facilities which were of the Ministry of Supply to them should no fully used in the investigations which took place longer be required. after the accident. 13. AR.B.issued a number of special category certificates of airworthiness to enable the requisite (e) The de Havilland Aircraft Company Limited tests, both in. this country and overseas, to be 10. de Havillands were the manufacturers of carried out, but it was not until early in 1952 that the Comet aircraft and the engines were made a full Certificate of Airworthiness was issued. by a subsidiary company, the de Havilland This enabled the passenger service to be started Engine Company Limited. Mr. R. E. Bishop is and it was actually commenced on the 2nd May, the Chief Designer of de Havillands and his Chief 1952. The personnel for the service had received Assistant is Mr. C. T. Wilkins. Mr. R. H. T. intensive training. B.O.A.C. had established a Harper is the Chief Structural Engineer and Mr. school for the training of pilots and crews and H. Povey is the Director in charge of Production. made full use of a special school which had been de Havillands have an Inspection Department established by de Havillands for the training not entirely separate from their Production Depart­ only of pilots and crews, but also of station ment and the independence of the Inspection engineers. By the 8th April, 1954, when the Comet Department is secured by the provision that it fleet of B.O.AC. was grounded after the disaster reports direct to the M(;lnaging Director and is near Naples, Comet aircraft had flown almost not in any way under the control of the Produc­ 25,000 hours, representing, on the basis of 400 tion Department. de Havillands have been miles per hour, a mileage of 10,000,000 miles. approved under paragraph 1(g) of the Civil A via­ 14. Dealing more specifically with the technical tion (Air Registration Board) Order of 1951 as aspect of the development of the project between qualified to furnish reports and certificates as to September, 1946, and the 2nd May, 1952, de compliance with airworthiness requirements. Havillands' outlook and practice underwent virtually no change. In order to provide an PART II economically satisfactory payload and range, at the high cruising speed which the turbo-jet engines mSTORY OF THE COMET PROJECT offered, it was essential that the cruising height 11. Mr. Bishop stated that at the end of the should be upwards of 35,000 ft.-double that of war de Havillimds were faced with the problem the then current air-liners-and that the weight of recommencing the manufacture of civil air­ of the structure and equipment should be as low craft. During· the war they had been building as possible. o:t;tly military aircraft. They decided that it would 15. Throughout the design they relied upon be inadvisable merely to build another version well established methods, essentially the same as of the conventional aircraft; they had had some those in general use by aircraft designers. But years' experience with jet fighters and concluded they were going outside the range of previous that with the help of their engine company they experience and they decided to make thorough should be able to produce a useful civil aircraft tests of every part of the cabin structure. They which would be a step ahead of the current type. had not only to prove to their own satisfaction With this end in view' they commenced design that their design was basically sound, but also to by the end of September, 1946. Some idea, how­ investigate the effect, on the large variety of ever, of the amount of work involved is indicated materials involved, of the extreme conditions by the fact that it was not until the 27th July, which would be met. They gave special attention 1949, that the first prototype Comet made its first to the structural integrity of the pressure cabin flight. de Havillands were, however, fortunate that The difference* between the internal and external B.O.A.C. and the Minister of Supply were willing pressure (8! lb.! sq. in) was about 50 per cent. to enter into a contract for the purchase of greater than that in general use and there was in Comet aircraft without waiting for the prototype addition a larger difference between the internal to be available. This enabled de Havillands at and external temperatures. once to do preliminary work in the Production Department. The contract was entered into on 16. Their policy of testing in the laboratory the 21st January, 1947 and under it B.O.AC. was not a novel one, nor indeed were they alone started th~ir proving flights in April, 1951. * This difference is sometimes referred to hereafter as ' P '.

94722 A3 in their belief in it. They recognised. however. 2 P some 30 times to the test section of the front that testing alone is not sufficient. Every test is part of the cabin together with a series of 2,000 to some extent a compromise. since the conditions pressurisations to rather over P. These tests were to be met in service can seldom be represented not intended as a test of the fatigue resisting completely in the laboratory and in many cases properties of the structure, but rather as providing are not accurately known. The result must. there­ an assurance that the cabin would be satisfactory fore. be reviewed in the light of calculations based as a pressure vessel. But they undoubtedly contri­ on fundamental knowledge. and on general buted to de Havillands' confidence in the sound­ experience and practice. ness of the cabin. 17. For the design of the basic structure of the 20. Simultaneously with the design and testing cabin they adopted a multiple of the working of the pressure cabin. all other parts of the struc­ pressure difference. p. in excess of current require­ ture were receiving treatment based on the same ments in any country. The British Civil Air­ outlook--design to at least the current require­ worthiness Requirements (B.C~A.R.) called for a ments, coupled with exhaustive tests. The wing "proof" pressure of It P (under which the cabin is of special interest, since it is here that require­ must show no signs of permanent deformation). ments specifically directed to resistance to fatigue together with a "design" pressure of 2 P (at first became important. During the period 1949 which the material may reach its ultimate to 1951 there had been growing among all air­ strength). These requirements were the same as craft designers and users a realisation that the those of the International Civil Aviation life of the essential structure of an aircraft is not Organisation (LC.A.O.) and also those of this unlimited. The effects of atmospheric turbulence country for military transport aircraft. de had produced unexpected and relatively early Havillands used a design pressure of 2! P and failure of the wings of certain transport aircraft. tested the cabin to 2 P. Two test sections of the Gusts are most severe near the ground and in the cabin were built. The front part. 26 ft. in length, tropics. Methods had beer4 devised, and have extended from the nose nearly to the front spar since been improved and extended, for determin­ of the wing, and included typical windows, ing their frequency and intensity. In the light of hatches and door. The centre part, 24 ft. in this knowledge, repeated loading tests* of the length, extended from a few feet in front of the wings of transport aircraft became accepted as front spar to a few feet aft of the rear spar. necessary. Tests of the Comet's wing were made covering the. large cut-out containing the wing in close co-operation with RAE. structure. 21. Until about the middle of 1952 the likeli­ 18. Their reasons for adopting these sub­ hood that the fatigue resistance properties of a stantially higher figures were two. They believed. pressure cabin demanded further precautions, and this belief was shared by ARB. and other either in design or by test, than were provided expert opinion, that a cabin which would survive by the current static strength requirements had undamaged a test to doubl~ its working pressure, not been realised. The matter first came to de 2 P, would not fail in service under the action of Havillands' notice through Mr. Harper's associa­ fatigue* due to the pressurisation to working tion with the problem on Service (R.AF.) trans­ pressure, P, on each flight, and to other port aircraft, as a member of the Joint Airwortlfi­ fluctuating loads to which it is subjected in ness Committee (J.AC.) of the Ministry of operations. Supply. Draft Requirements (Paper 579, Oct., Secondly, they considered that it would ensure 1952) called for a static test to 2 P, a proof test to a larger margin of safety against the possible It p, together with repeated loading tests of It failure of windows, doors, and hatches. These P applied 10,000 times. are contingencies which had been shown by 22. At about the same time ARB. were re­ experience to be a serious risk. for even if viewing the civil position. In due course they nothing worse happens, the resulting loss of pres­ issued proposals in Paper No. 230 (19th June, sure may be rapid. 1953) which called for the same static test to 2 P 19. So much importance did they attach to this and proof test to It P but raised the number of latter consideration that· they made many tests applications of It P to 15,000. At the same time of window panes to very high pressures. In addi­ the paper suggested that certain structural parts, tion, they applied pressures of between P and such as riveted joints, door and window fr~es, etc., might have to be designed to 3 P (on the * There is attached hereto as Appendix IV a note on the subject of fatigue in metals and its bearing on the * In which the appropriate load is applied and removed design of engineering structures which has been prepared many times, simulating the effects of gusts, or any other for my assistance by my Assessors. cause of variation of load 10 ultimate strength of the material), in order to at 0931 hours on the 10th January, 1954, on a meet these requirements. It also stated that the :flight to London. After taking off the aircraft figure of 15,000 was intended to cover the num­ was in touch with Ciampino control tower by ber of applications of P during the life of an air­ radio telephone and from time to time reported craft. and that the test pressure of 1;1- P was in­ its position. These reports indicated that the tended to cover the phenomenon of "scatter"* :flight was proceeding according to the B.O.A.C. in the fatigue strength of different cabins built :flight plan and the last of them, which was re­ to the same design. ceived at 0950 hours, said that the aircraft was 23. The result of these developments was that over the Orbetello Beacon. The Captain of in July, 1953 de Havillands reconsidered the posi­ another B.O.A.C. aircraft, Argonaut G-ALHJ, tion of the Comet's cabin. Up to that time no gave evidence of communicatiolls which passed Comet had exceeded 2,500 hours :flying-say 800 between him and Yoke Peter. The last such pressurised :flights. In order to satisfy themselves message received by the Argonaut began of its safety, and also to discover its probable safe " George How Jig from George Yoke Peter did working life, they carried out repeated loading you get my " and then broke off. The Captain of tests of the test section of the fore part of the the Argonaut gave it as his opinion that the cabin, applying the working pressure P about message was not merely interrupted by another 16,000 times. By September, 1953, this specimen aircraft but that transmission ceased after the had withstood 18,000 applications of P in addi­ word "my" and he estimated that the message tion to some 30 earlier applications of pressures was received by him at approximately 0951 between P and 2 P. hours. Shortly after 1000 hours the Ciampino Traffic Control Clerk heard a sound which he 24. These tests were ended by a failure of the suggested might have been an unmodulated skin in fatigue at the corner of a window, origi­ transmission from Yoke Peter. nating at a small defect in the skin. But the num­ ber of pressurisations sustained was so large that, 27. The evidence of four witnesses from Elba in conjunction with the numerous other tests, it as to things seen and heard by them on the 10th was regarded as establishing the safety of the January suggests that Yoke Peter must have Comet's cabin with an ample margin. crashed into the sea at about 1000 hours and it 25. Meanwhile, on the 2nd May, 1953, Comet therefore appears that something happened to G-ALYV had crashed in a tropical storm of the aircraft with catastrophic suddenness which exceptional severity near Calcutta. An inquiry may have accounted for the interruption of the was directed by the Central Government of India transmission of the last message to the Argonaut. and was held under Rule 75 of the Indian Air­ It is also clear from the evidence of the Elba wit­ craft Rules 1937. The Court reported on the nesses that part of Yoke Peter fell into the sea 26th May, 1953, that the accident was caused by in flames. structural failure of the airframe during :flight 28. The chart, which is Figure 1 of this through a thundersquall. In the opinion of the Report, was prepared from all the information Court the structural failure was due to over­ available and produced by a Navigating Officer stressing which resulted from either:- from B.O.A.C. The estimated :flight track of the (i) Severe rgusts encountered in the aircraft and the position in which bodies and thundersquall, or wreckage were found can be seen on the chart (ii) Overcontrolling or loss of control by and the witness gave it as his opinion that at the pilot when :flying through the 0951 hours the aircraft was probably approach­ thunderstorm. ing a height of 27,000 feet. Fatigue failure of the cabin was not then suspected as a cause and in my opinion the evi­ PART IV dence adduced in the course of the present Inquiry affords no sufficient reason for doubting THE AIRCRAFT' the conclusion of the Indian Court. 29. Yoke Peter was designed and constructed by de Havillands and was of the type properly PART III described as DH106 series 1, commonly known THE ACCIDENT as the Comet I. It was designed for high speed long distance, passenger and freight transport at 26. Comet G-ALYP (sometimes hereinafter high altitude and was propelled by four de called Yoke Peter) left Ciampino Airport, . Havilland Ghost 50 turbo-jet engines- mounted * See Appendix N. within the wings, each engine developing a static u 94722 A4 thrust of 5,000 lb. The crew and passenger com­ PART V partments were pressurised, so that when flying at 40,000 ft. a cabin pressure equivalent to THE CREW atmospheric pressure at an altitude of 8,000 ft. 32. Captain Alan Gibson, D.F.C., who was in was maintained. The cabin pressure was regu­ command of Y okePeter at the time of the acci­ lated to a maximum pressure difference between dent was aged 31 years and 3 months. He held cabin and outside atmosphere of 8·25 Ib/sq. in. Airline Transport Pilot's Licence No. 22713, and a safety valve was set to open at a pressure valid until the 24th February, 1954, which difference of 8·5 Ib/sq. in. The dual flying con­ entitled him to fly in command of Comet air­ trols were power operated by hydraulic servo craft and he had a valid Instrument Rating. control units. The fuel for the engines was kero­ Captain Gibson also held Flight Navigator's sene carried in a centre section tank made up of Licence No. 1442 which was valid until the 19th four inter:-connected bag tanks and in four February, 1954. He entered the employment of integral wing tanks. The authorised maximum B.O.A.C. under contract in 1946 having previ­ all-up weight was 107,000 lb. Yoke Peter first ously been employed by B.O.A.C. on secondment flew on the 9th January, 1951, and was granted from the . While in the Royal a Certificate of Registration No. R.3162/1 on Air Force Captain Gibson had a total flying the 18th September, 1951, in the name of experience of 1,348 hours of which 1,175 were B.O.A.C. as owner. A Certificate of Airworthi­ flown in command. He had flown a total of ness No. A.3162, valid until the 12th March, 4,062 hours by day and 1,156 hours by night with 1953, was granted on the 22nd March, 1952. B.O.A.c. and most of these were flown as first The aircraft was delivered to B.O.A.C. on the pilot. He had flown Comets· for 84 hours by day 13th March, 1952, and from that date was and 48 hours by night as second pilot and for operated by B.O.A.C. On the 2nd May, 1952, 79 hours by day and 80 hours by night as first having by then flown a total of 339 flying hours pilot. During the six months preceding the acci­ in experimental, test and training flights on dent he had flown 79 hours by day and 80 hours behalf of de Havillands and B.O.A.C. it entered by night as first pilot of Comets and 47 hours by scheduled passenger service and was the first day and 31 hours by night under supervision. jet-propelled passenger carrying aircraft in the 33. While with B.O.A.C. Captain Gibson was world to do so. concerned in an accident involving the forced 30. On the 11th March, 1953, the Certificate landing of a Hermes aircraft in 1951 and was of Airworthiness was renewed for one year and complimented by the Operations Manager for his was,· therefore, valid at the time of the accident. conduct on that occasion. He was successful in On the 11th November, 1953, after the aircraft both his flying checks during the period when had flown 3,207 hours and following a repair to he was flying Comets and I am satisfied that he the passenger entrance door the fuselage was was fully equipped to carry out his normal duties subjected to a proving test to 11 Ib / sq. in. The as a pilot and as a captain and to deal with airframe and engine log books show that the emergencies. airframe and engines had been regularly inspec­ 34. The second pilot of Yoke Peter was First ted and maintained in accordance with the Officer William John Bury whose age was 33 Approved Maintenance Schedules and that the years and 10 months. He held Airline Transport number of flying hours of each engine since its Pilot's Licence No. 27251 valid until the 8th April. last complete overhaul was well within the 1954, and a valid Instrument Rating. In addition approved life. he held Flight Navigator's Licence No. 2583 31. In accordance with. the Approved Main­ valid until the 9th October, 1954. He had flown tenance Schedules a Check I inspection a total of 1,917 hours in the Royal Air Force of was completed on the 6th January, 1954, at which 1,735 were as first pilot, all in· piston London Airport and a Certificate of Mainten­ engined aircraft. With B.O.A.C. he had flown ance, signed by properly licensed airframe and 2,355 hours by day and. 643 by night as second engine maintenance engineers and valid for 75 pilot and 11 hours by day and 1 hour by night as flying hours, was issued on the 7th January, first pilot and altogether had flown 153 hours by 1954. At the time of the accident the aircraft had day and 109 by night in Comets, all as second flown only 40 hours since the issue of the Certifi­ pilot. I am satisfied that First Officer Bury was cate of Maintenance and its total flying life was fully equipped to carry out his normal duties and 3,681 hours. An Aircraft Radio Station Certifi­ to support his captain in emergencies. cate of Serviceability was issued in respect of 35. The Engineer Officer was Mr. Francis Yoke Peter on the 7th January, 1954, with the Charles Macdonald who was aged 27 years and remark" no items unserviceable." 11 months. Since joining B.OA.C. on the 21st January, 1952, he had 439 hours flying as 2A. Moreover, no load was shown on the Load Engineer Officer in Hermes aircraft and 281 hours Distribution and Trim Sheet for hold 3, whereas in Comets of which 225 hours were flown during there was evidence that 15 kilograms of baggage the six months preceding the accident. Mr. were placed in that hold. I am satisfied, however, Macdonald's Flight Engineer's Licence was No. by the evidence of Mr. B. J. Folliard that these 428 and had expired on the 11th December, 1953. errors in the Load Distribution and Trim Sheet During its validity this licence included Comet would have left the loading and trim of the air­ aircraft. Had he applied to renew his licence he craft well within the prescribed safe limits. would have been required to give Log Book evidence of six hours flying as engineer-in-charge including six flights during the 12 months preced~ PART VII ing the date of application and would have been PRE-FLIGHT INCIDENTS required to pass a medical examination. 36. On joining B.O.A.C. Comet Fleet 41. The last three flights made by Yoke Peter Mr. Macdona1d obtained an endorsement to his prior to that which ended in disaster were from licence which made it valid in respect of Comet Karachi to Bahrein, Bahrein to Beirut and Beirut aircraft and he completed a form giving details to Rome. During refuelling at Karachi a defect of his licence. In completing this form he stated, developed in the port wing refuelling-valve wrongly, though no doubt in good faith, that his actuator and in order to complete the refuelling licence was valid until the 24th April, 1954. He of the port wing tanks the Engineer Officer of himself made no application to renew the licence Yoke Peter adopted a procedure known as before its expiry nor was he given any reminder " off-load" refuelling which is authorised for use to do so by B.O.A.C. This matter is further in such an emergency. It involves holding the referred to in paragraph 147 of this Report. refuelling switch in the" off-load" position* and 37. I am satisfied that Mr. Macdonald's flying releasing it when refuelling is complete. In fact experience was sufficient to support an applica­ the Engineer Officer did not release the switch in tion for renewal of his licence but I have no time and about five gallons of fuel escaped from evidence as to his medical fitness. However, I the airvent on the under surface of the mainplane. have no reason to suppose that he was in fact There was no repetition of this incident at Bahrein unfit at the time of the accident. but at Beirut, after the Engineer Officer had explained to the ground engineer, who was 38. The Radio Officer was Mr. Luke Patrick assisting him with the refuelling, what had hap­ McMahon who was aged 32 years and 2 months. pened at Karachi, a further. incident occurred. He held a First Class Flight Radio Telegraphy When the Engineer Officer returned to the port Operator's Licence No. 1235 which was valid wing after inspecting the starboard tanks he until the 16th October, 1954, and had done 2,946 noticed fuel emerging from the port air vent. The flying hours with B.O.A.C. in various aircraft refuelling switch was in the neutral position from before the 3rd October, 1952, and 629 hours in which fact, and from the fact that fuel was Comets thereafter. During the six months pre­ obviously entering the tank, he deduced that ceding the accident he had flown 207 hours in somebody, intending to put the switch to the Comets. I am satisfied that he was a capable "off-load" position, from which it should auto­ officer. matically have returned to neutral when released, 39. The other members of the crew were must have failed to do so and that the switch, Steward Frank Leonard Saunders and Stewardess instead of returning to neutral, had remained half Jean Evelyn Clarke, both of whose services had open. He attempted to close the switch by moving at all times been entirely satisfactory. it to the full "off-load" position and releasing it but this had no effect and the flow of fuel was eventually stopped by shutting down the bowser. PART VI 42. As a result of this incident the actuator THE PASSENGERS AND CARGO was removed and as no replacement was available 40. Yoke Peter carried a total of 29 passengers, it was tested, found satisfactory and refitted. all of whom were killed in the accident. The These incidents were reported by the Engineer cargo carried did not include any items which Officer to Mr. Macdonald when the aircraft was could have been relevant to the cause of the handed over at Rome. The practice of" off-load" accident. The comparison between the amount refuelling is further referred to in paragraph 111 of cargo known to have been carried and that of this Report. shown in the Load Distribution and Trim Sheet -*-Th-e-n-o-rm---'-al-p-u-rp-o-s-e-o-r -thi-·s-po-s-iti-· o-n-i-s -to-e-n-a-bl-e-th-e showed a discrepancy of 27 kilograms in hold tanks to be emptied. 13 43. Two other items were also unserviceable board and he himself put to sea after he had during the flights from Karachi to Rome. These made all the necessary arrangements. In these sal­ were the No. 1 engine hydraulic flow warning vage operations 15 bodies. various mail bags and light and the automatic temperature control some aircraft wreckage and personal effects were selector. The former device is designed to draw recovered. The ships had been assisted in their the attention of the pilot to a possible failure of search by the collaboration of aircraft. On the two the engine-operated hydraulic pump. On this following days the search was continued. No more occasion. when the flow warning light appeared bodies were found but various pieces of wreckage faulty. the operation of the pump was tested by and articles were recovered. other means and found satisfactory. The auto­ 46. Under Lieutenant-Colonel Lombardi's matic temperature control selector is intended to directions the bodies were taken to the local control automatically the temperature of the crew cemetery at Porto Azzurro and devoutly placed and passenger compartments. When it was found in the chapel there. At the request of the exam­ to be faulty the temperature was controlled ining magistrate at Portoferraio an examination manUally. I am satisfied that neither of these of the bodies recovered was carried out by Pro­ faults. both of which were drawn to the attention fessor Antonio Fomari who was acting under the of Mr. Macdonald. can have endangered the direction of Dr. Folco Domenici. Director of the aircraft in any way. Institute of Forensic Medicine in the University of Pisa. Professor Fomari gave evidence before PART VIII me and he put in a report which' had been pre­ pared by him and Dr. Domenici. The substance WEATHER CONDITIONS AT THE TIME of their report is to be found in the conclusions OF THE ACCIDENT at p .. 60 of the translation of the report and may 44. From take-off at Rome at 0931 hours on be summarised as follows:- the 10th January. 1954. to'the time of the acci­ (1) Death was caused by impact against parts dent at approximately 27.000 ft. near Elba Comet of the aircraft. G-ALyP experienced essentially good weather (2) There were serious lesions resulting from conditions. The climb was made through only explosive decompression and deceleration. thin and broken layers of cloud with no rain (3) The probable point of impact between the and with negligible icing conditions. At the time bodies and the structure of the aircraft was the and position of the accident it is probable that forepart of the fuselage. perhaps in the vicinity some turbulence in clear air may have existed of that part of the fuselage which lies above due to the proximity of a narrow high velocity the engines. wind current called a " jet stream ". Such turbu­ lence. if encountered. would be less than aircraft (4) There were burns on the bodies of all frequently experience in turbulent cloud condi­ the victims but they presented post-mortem tions. It can. therefore. be assumed that the state characteristics from which the inference was of the weather was not a contributory cause of that the burns took place after death. the accident. (b) Action taken by the Ministry of Transport and Civil Aviation PART IX 47. News of the accident was received by the ACTION TAKEN AFTER THE ACCIDENT Accidents Investigation Branch of the Ministry AND PRIOR TO THE ACCIDENT TO of Transport and Civil Aviation at 1200 hours COMET G-ALYY on the 10th January. 1954. and both the Senior (a) Local salvage and medical investigation Inspector of Accidents. Mr. Nelson. and the 45. At 1150 hours on the 10th January. 1954. Senior Investigating Officer. Mr. Morris. left for the Harbour Authority at Portoferraio in the Isle Italy that evening. of Elba was informed of the occurrence of the 48. On arrival Mr. Nelson got into touch with accident. being told that an aircraft had exploded the Commission which had been convened by the in the air and crashed in flames into the sea Italian aviation authorities and went with the south of Cape Calamita roughly in the direction Commission to Elba. Some days later it was of the island of Monte Cristo. With commendable agreed that the responsibility for the investigation promptness Lieutenant-Colonel Lombardi. the of the accident should be handed over to the Acci- Officer Commanding the Harbour Authority of dents Investigation Branch of the British Ministry Portoferraio. despatched all available craft to the of Transport and Civil Aviation but Colonel scene of the accident with a doctor and nurse on Miniero and Signor Roveri. who have attended 14 this Inquiry. were appointed accredited represen­ relate only to the external structure. Figure 4 is tatives to the British investigators and gave them a photograph showing the reconstruction of the every possible assistance. The Minister of Tra~s­ fuselage and tail unit from the wreckage and port and Civil Aviation was also in touch WIth Figure 5 is a photograph showing the reconstruc­ the Admiralty and it was arranged that the Com­ tion of the front fuselage. mander-in-Chief Mediterranean. Earl 51. The amount of wreckage recovered was Mountbatten. would cause an intensive search to greatly in excess of the expectations entertained be made. for the wreckage. The Chief Inspector in March. 1954. when the decision to allow the of Accidents. in accordance with normal practice. Comets to fly again was taken. A remarkable arranged for the wreckage recovered to be sent to fact was the small amount of damage which had and examined at RA.E. Mr. Nelson and Mr. been caused to the structure either by immersion Morris remained in Elba. examined the wreckage in sea water or in the process of salvage. recovered and arranged for its transport back from Elba to the mainland and thence to Rome. (d) The Abell Committee whence it was flown direct to the United King­ 52. Immediately on receiving news of the acci­ dom. but certain very large pieces had to be sent dent B.O.A.C. had decided to suspend their by sea. normal Comet passenger services. for the pur­ (c) Naval search for wreckage pose of carrying out a detailed e~amination ?f the aircraft of the Comet operational fleet ID 49. Forsberg was placed in charge collaboration with ARB. and de Havillands and of the operations. Special vessels. H.M.S. Barhill to this end the Chairman of B.O.A.C. had called and H.M.S. Sea Salvor, were fitted up to carry a meeting at London Airport for the 11th J anu­ 200 tons of heavy moving gear. An observation ary. 1954. which was attended by representatives chamber. television gear. an 8 toothed grab and of B.O.AC.• the Accidents Branch of the Minis­ other equipment were obtained from England and try of Transport and Civil Aviation. de the necessary modifications to the vessels were Havillands. the de Havilland Engine Company made in the dockyard at Malta. This was all done Limited and ARB. As a result of that meeting in under a fortnight and the two vessels and a committee under the chairmanship of Mr. C. H.M.S. Wakeful, in which the television equip­ Abell. the Deputy Operations Director

THE RA.E.· REPORT PART XIII 92. The Report (which was part of the evi­ THE COURT'S CONCLUSIONS AS TO dence before the Court) is divided into 12 parts. THE CAUSE OF THE ACCIDENT The first part contains an outline of the investi­ gation and states the opinion RA.E. formed as (a) The main finding in the R.A.E. Report to the cause of the accident. I have included the 95. The opinions expressed in the Report first part, which is intelligible without reference were supported by the evidence of Sir Arnold to the other parts, as an appendix to this Report Hall, Dr. Walker and Mr. Ripley. Their con­ (Appendix VD. Para. 4 thereof which states clusions were accepted by de Havillands and the opinion of RA.E. is in the following terms: - B.O.A.C. All parties appearing at the Inquiry "We have formed the opinion that the acci­ paid a warm, and in my opinion well-deserved, dent at Elba was caused by structural failure of tribute to the Report and to all who had co­ the pressure cabin, brought about by fatigue. We operated in the work done at RA.E. As I have reach this opinion for the following reasons:- already indicated and for the reasons I have (i) The low fatigue resistance of the cabin given I have accepted the main conclusion of has been demonstrated by the test the Report that the cause of the accident to Yoke described in Part 3, and the test result is Peter was the structural failure of the pressure interpretable as meaning that there was, cabin brought about by fatigue. at the age of the Elba aeroplane, a definite (b) The alternative suggestion made by risk of fatigue failure occurring (Part 3). Mr. B. lablonsky (ii) The cabin was the first part of the aero­ plane to fail in the Elba accident (Part 2). 96. The only rival suggestion was made by (iii) The wreckage indicates that the failure in Mr. Jablonsky. His experience of structural the cabin was of the same basic type as problems in aeronautics has been concerned that produced in the fatigue test (Parts 2 mainly with propellers having blades of highly and 3). compressed wood. He is, therefore, familiar with adhesives, and with the problems which have to (iv) This explanation seems to us to be con­ be overcome in using them to make components. sistent with all the circumstantial evidence. (v) The only other defects found in the aero­ 97. In the construction of the Comet wide use plane (listed in Section 3) were not con­ is made of a metal-to-metal adhesive known as cerned at Elba, as demonstrated by the , mainly for the purpose of attaching wreckage. members, generally known as" stringers ", to the Gkin both of the wing and of the cabin. In the Owing to the absence of wreckage, we are un­ cabin there are about forty stringers more or able to form a definite opinion on the cause of less evenly spaced around the circumference and the accident near Naples, but we draw attention running longitudinally. They are not structurally to the fact. that the explanation offered above for continuous from end to end, the largest uninter­ the accident at Elba appears to be applicable to rupted length being about 25 ft. de Havillands that at Naples." were pioneers in using Redux for such purpos~s It should be added that the medical evidence in aircraft structures, and have had long expen­ KS to the state of the bodies recovered was con­ ence of it. It is in effect an alternative to the sistent with the conclusion thus reached. conventional riveting. 22 98. Mr. Jablonsky's argument proceeded on 100. I deal below with these points the following lines:- separately: - (a) The skin of the cabin is exposed under (a) During the experiments made in flight on service conditions to a large variation in Comet G-ANAV at R.A.E .• measurements temperature. He suggested a range of were made of the difference in temperature 80°C on the ground in the tropics to between the skin and the stringers in typical -$5°C at about 40.000 ft. The rate of climb positions in steady flight at cruising altitude. of the Comet is fairly high and the tem­ They led to the conclusion that the maxi­ perature of the skin might change over this mum probable steady difference in tempera­ range in about 30 minutes. The stringers. ture is about 10°C. I am advised that the however. although inside the skin. are out­ shear stress in the Redux caused by the side the insulating lining of the cabin and relative contraction between the skin and therefore not exposed to the full tempera­ the stringers due to a temperature difference ture. of the warm cabin air. His argument of this order would be well within its contemplated a difference in temperature capacity. between skin and stringer of as much as 60° or 70°C. This would have the result Mr. Jablonsky did not agree that any reliable that the skin would contract relative to the inference about the conditions on an operational stringer. in the direction of the cabin's climb could be drawn from these experiments. I length. The adhesive would. therefore. be recognise that this comment has some force but subjected to a shear stress which might be I base my conclusions on this aspect of his sufficient to cause it to fail. criti~ism on the more general considerations set out in paragraphs 101. 102 and 103 below. (b) Even if this did not cause the adhesive to fail statically (that is on the first occasion (b) No evidence was submitted of. the effect, on when such a difference of temperature be­ the fatigue strength of a Redux joint. of the tween the skin and the stringers occurred) level of temperature of the adhesive. But frequent repetition of the shear stress might I am advised that the wide experience of its produce fatigue in the adhesive. and cause use by de Havillands in the structures of it to fail. other aircraft. where alternations of load on (c) Mr. Jablonsky recognised that the depend­ the glue line have certainly existed in num­ ence on temperature level of the properties bers far in excess of any likely to have been of Redux is well known. He suggested. how­ experienced in the cabin structure of the ever. that frequent and rapid variations of Comet. and over a wide range of tempera­ temperature would reduce its strength sub- ture of the Redux itself. is satisfactory stantially. . evidence that this is not a probable cause of failure of the Redux joints in the Comet's (d) It is generally recognised that the satisfac­ tory use in engineering structures of any cabin. form of adhesive (or. indeed. of processes (c) de Havillands made special tests to investi­ essentially similar. such as the welding or gate the effect on typical joints of repeated soldering of metals) can be ensured only by alternation of temperature between 60°C the development and maintenance of higher and -50°C. I am advised that these show standards of workmanship and process in­ that alternations of temperature within this spection than are necessary in the use of range have no appreciable effect on the riveting. While Mr. Jablonsky recognised strength of a Redux' joint. that 'de Havillands' production technique (d) At my request. de Havillands submitted a for Redux had been developed after many statement which summarised the history years' study of it$ properties, and that their and present state of their production experience of its use in other aircraft had methods in the use of Redux. with particu­ been highly satisfactory. he suggested that it lar reference to its application to the con­ was not a process sufficiently reliable for struction of the Comet. Mr. Povey. the use in the primary structure of a pressure Director responsible for production, gave cabin. evidence on the point. I am advised that this 99. Mr. J ablonsky said in evidence that in his statement and evidence show that de Havil­ inspection of the wreckage at R.A.E. he had seen lands fully appreciated the importance of examples of failure of the " glue line" which had this aspect of the use of an adhesive in satisfied him that weakness in it was primarily essential structural components. and that responsible for the failure of the structure of the the methods they have devised. including cabin. process control and inspection, tests of s~ples of every joint, and periodic strip­ proceeded on the basis that the 9,000 hours pmg of complete stringers from the skin, (3.000 flights) at which Yoke Uncle burst could provide all the assurance that could reason­ be regarded as a fair average life for the fuselage ably be required. and to have been impressed by the improbability, 101. However, the final test of a process of this on this basis, of both Yoke Peter and Yoke Yoke type i~ recognised to be experience in service. failing from fatigue after only about 3,000 hours No eVIdence was produced of any failure of de 0,000 flights). He was unable, however, to sug­ Havillands' methods of dealing with the same gest any other cause. He admitted that he could problem in aircraft such as the Hornet and the find no evidence either (a) of excessive internal Dove, in both of which Redux is widely used. pressure in the cabin or (b) of excessive stresses Moreover, inspection of Yoke Uncle at RAE., in the cabin structure due to external action such both. before and after it was tested under repeated as gusts or failure of the control system. lIe loa~g, showed no signs of any defiCiency in the agreed also that he could not name any alterna­ glue lIne. It must be remembered that before it tive cause of the failure which RAE. had failed was delivered to RAE. for tests, this aircraft to consider. had done 3,521 hours of flying on B.O.AC. ser­ 106. Bearing in mind that Mr. Tye is the vices, experiencing the conditions of temperature, Chief Technical Officer of ARB. and as such and of temperature variation between the skin will be responsible for advising ARB. when an and the stringers, contemplated by Mr. lablonsky. application is made for a new Certificate of Air­ 102. Finally, examination of the wreckage led worthiness for Comet aircraft, his caution is Mr. Ripley to conclusions, contrary to those in­ understandable, but I have the duty of express­ ferred by Mr. Jablonsky, for reasons which he ing my conclusion on the evidence. I rely in this explained in detail. connection on an answer given by Mr. Tye to 103. It has been established to my satisfaction Sir Lionel Heald which seems to me to represent that the rear part of the fuselage, substantially the proper approach for me to adopt in the cir­ intact, hit the surface of the sea' at high speed, cumstances of the case. Mr. Tye said "I think open end downwards. This caused the equiva­ in concluding on the likelihood of the cause one lent of an explosion in it, whose effects were has to take the thing as a whole; one has to take n':lturally most acute near the open end (see the tank test evidence and say that that shows FIgures 3 and 4). I am advised that the failure, that fatigue is possible, although on my argument under these circumstances, of the adhesion not necessarily probable, that is the tank test by between the skin and the stringers cannot be itself; one then has to look at the other half of regarded as evidence of the failure of the the matter, namely, all the other possible causes, adhesive to meet the requirements of the normal and if in the process of eliminating possible use of the aircraft. There was in this neighbour­ causes you become completely confident that you hood abundant evidence of the failure of all the have eliminated every other possible cause, then methods of attaching the various structural com­ you are driven to say that the possible fatigue ponents to one another. Moreover, the numerous rises to the most probable cause." Applying these places where the skin had parted from the observations to what was done in the course of stringers exposed the glue line to examination the investigations by RAE. and by the de and Mr. Ripley said that he had been unable to Havilland Engine Company Limited and to the find any sign of any unsatisfactory features in evidence given in the Inquiry before· this Court, the process employed by de Havillands, or of I unhesitatingly come to the conclusion that any weakness in the adhesive. RAE. were right in their conclusion that the accident at Elba was caused by structural failure 104. In the light of these considerations I of the pressure cabin in the region of the AD.F. have no hesitation in rejecting Mr. Jablonsky's window, brought about by fatigue. In reaching suggested alternative cause of the failure of the this conclusion I am fortified by the advice I cabin. have received from my Assessors. (c) Mr. Tye's evidence (d) The possibility of over-pressurisation 105. The only other witness who did not com­ pletely accept the suggestion advanced in the 107. I considered nevertheless that although Repor~ was Mr. Tye. He did not dispute that the RAE. Report contained a full investigation the pnmary cause of the accident was the burst­ of the equipment used for controlling the pressure ingof the cabin structure, but he expressed him­ in the cabin, including both an examination of self as not entirely satisfied that fatigue was the the possible causes of mal-functioning and of the cause of that disruption. He appears to have condition of the equipment recovered from the 24 CHART OF AI RCRAFT TRACKS G-ALYP AND G-ALYY Rc I,

~3'N f£ 10£. JJ£. 12£ 13£. IlE. 15E. '(;4: I7E .• .. . ,:J I. I I 145 N.

.~ " 2} G .AL'fP. WRECK~E~TC: """" ""''B' '0

o.'ll I . I I 1 - -'- 1 - m_ ~-h . \ _ J.c ~42·N .

COAST LINES A.RE APPROlCIMATE. f NL\'

4t"1l1 \ ~ <::::: t.,.. 14(N. G-lIlYY I ,. I '!i~E" NLI6IIT ... lloooh ~MM1V ~~. 21000h

C-AL"/y. R~PoI(rSb 7i..,~ 1851 e:.n .... rEb 7iN£ /95. ..!!e.1HT ;. 2'QOoO ;;r:

40,..1 l £·... 17 · ",",NT, ... d"""f'l =::." / , I,v ... , ~ , ~~.

39·.... 1 I \ "),. 1~9·N. \: "" o o (3 Cl a.

3'4, I I. I V ~ I'=', ct"'=:t-=c::T ! Cl 1 ,~ 138'N E. loE:. 1.1£. ,~'£.. /SE. Ji'£ ' 15·£.. ,,'t I7"€. . FIG. 2. DIAGRAM SHOWING AMOUNT OF WRECKAGE 'RECOVERED_G_AL YP. FIG. 3. DEVELOPMENT DRAWING OF FUSELAGE SHOWING AMOUNT OF" WRECKAGE RECOVERED-G-AL YP. FIG. 4. RECONSTRUCTION OF FUSELAGE AND TAil UNIT WRECKAGE-G-ALYP. THIS PANEL (INCLUDING CREW ENTRY DOOR) RECOVERED . SEPARATELY AND SHOWS WATER IMPACr DAMAGE ON OUTER FACE.

STARBOARD WINDSCREEN MISSING.

FORWARD EQUIPMENT BAY REAR BULKHEAD TORN FROM FRONT EXTEt\SIVE CRUSHING FUSELAGE AND OF WHOLE UNDERSIDE REMAINED ATTACHED OF FRONT FUSELAGE TO WING. BY WATER.

NOSE DOWN BENDING FAILURE OF PICK UP POINTS ON WING WITH NOSE WHEEL FINALLY SOME BENDING RETRACTED WHEN TO STARBOARD. SALVAGED.

FIG. 5. FRONT FUSELAGE-VIEW ON STARBOARD SIDE-G-ALYP. FIG. 6. AERIAL VIEW OF COMET G-ALYU IN TESTING TANK. FIG. 7. VIEW FROM INSIDE OF FAILURE AT THE FORWARD ESCAPE HATCH ON THE PORT SIDE-COMET G-ALYU D="R .... ME. 13 '4 '"1 18\ '9 171 '''j : RIVtT5 P'lJL.LEO OUT ,MM THE. I I' , TO"!> OF THESE SUa~I'RAA1E5 , , ! , I , : ITENSION I I ~: I/~ I f -I---~- - --V-- - < -CQ-NC'VC"~Y1' ____-+___ -+--4--__ . ___ L",r:-.~eu~,,:.:.:...--- r. ~ -l T_~ -auCKLE.-- .+ -_.

AIN jauCKLESLREL£ASE CAUSEDOF PReSSURE or ""'LuRE AT FAAME 19 I_I 11 I 1 IrTHROUGH LINE OF RIVETS iI' ON AFT EDGE OF FRAME lB.

FRAME:; 18 BY TEi\F\ING OUT 4 1 OF RIVETS ON FORWARO ... ~ SECONDARY FAIL.URE AT ]------___ ---j:~~~~------i--_~':------I EDGE 0" "R"ME. ~ .... ~ _____1

~ ALONG. THIS ~ ~:1Te:ARING L.'NE'l WINOOW REMAINEO IN PLJI\CE eu r--- ORICIN OF' F'A.ll..URE. ~ I( OF RIVETS UNQER FRAME SPRUNC OUTWAROS AT ( TENSION I RIVETS TORN TO! BOTTOM CORNER.S. [ EDGe: OF PLATE I ~ ~~ <---- .\ ----.J+l[) ,I bJ- ' , , ~ r~~----+-~i~~:=~==~:E~~~==~:===~.nr I ' ." ~j--®- --e I I I ------, ,SMALL FATIGUE CRACKS IF"ATIGUE AND OISCOLOUF\.ATIONt I I AT EOGe: OF THIS I AT THIS RIVET HOLE. I I , ..FI:!..S£.LAGE: DATUM , --- _ IRIVET HOLE. J _ • --..:.r: .------, ,, J LAP ..JOINT. , I I I I : r FRAME NO. KEY. I 14 1 16 17 18 I QIRECTION OF" SKIN SE:PA.RATION SHOWN By A.RROWS SKIN MOVING INWARD SHOWN ®­ SKIN MOVING OUTWARD SHOWN 0---- DIRECTION O~ TENSION L..OAOS TO CAUSE: F' ..... ILURE4'o-- __» o o o

FIG. 8. THE PROGRESS OF THE FAILURE OF THE FORWARD ESCAPE HATCH ON THE PORT SIDE-COMET G-ALYU. a. GENERAL VIEW LOOKING FORWARD.

b. CLOSE-UP OF SKIN AT BOTTOM REAR CORNER OF ESCAPE HATCH.

FIG. 9. FAILURE OF FRONT FUSELAGE AT IOAlbjin2 (3057 TOTAL FLlGHTS)-G-ALYU. a. GENERAL VIEW LOOKING AFT.

b. CLOSE-UP OF SKIN AT BOTTOM FRONT CORNER OF ESCAPE HATCH.

FIG. 10. FAILURE OF FRONT FUSELAGE AT 10.4 Ib/in 2 (3057 TOTAL FLlGHTS)-G-AL YU. CENTRE FuSELACllE. SPL.IT AL.ONCll TOP CENTRE L.INE THROU",H A.D.F. AERIAL WINDOWS AND OPENED OUTWARDS

REAR FU5ELAI::jE />ND TAIL UNIT SEPARATE.D AT. REAR SPAR STARBOARD OUTER WINC'j ----_~_ .L------P>.TTACHMENTS IN A DOWNWARD DIRECTION

SEPARATION Of CENTRE SECTION _____~;;c..".. AND OUTER WINCi WRECKACiE. MAIN FAILURE OCCURRED BETWEEN RIBS 12 AND 13 IN A DOWNWARD DIRECTION

WINCi CENTRE SECTION COMPLETE WITH ENC;INES ANO UNDERCARRIACiE

PORT oUTER WIN<7j

SEPARATION OF CE.NTRE SECTION AND OUTER WINCi WRECKA~E. MAIN FAlLUI\E OF PORT WING NOT FRONT FUSELACllE SEPARATED AT FRONT SPAR ATTACHMENTS f'.ECOVEf'.ED. F"ILURE PROBABLY IN IN A DOWNWARD DIRECTION DOWNWAI\O DIRECTION AND ROUGHLY SYMMETRICAL WITH STAf'.BOAI\D WING FAILUf'.e::. FIG. 11. LOCATION AND DIRECTION OF MAIN FAllURES-G-ALYP. DIRECTION OF PROPAGATION OF MAIN FAiLURES

FIG. 12. PHOTOGRAPH OF WRECKAGE AROUND ADF AERIAL WINDOWS-G-ALYP. (FOR LOCATION ON AILERON SEE FIG. 16.)

FIG. 13. PAINT TRACES OVER PORT AILERON UPPER SURFACE-G-ALYP. PARTICLES OF FIBREGLASS AND WOOD JAMMED INTO SKIN AND UNDER LAP JOINT PROBABLY FROM FUSELAGE CABIN FLOORING

GENERAL PATTERN OF METALLIC SCORES AND PAINT TRACES OVER INBOARD FACE PROGRESSING FROM WING SKIN TO UPPER EDGE OF FENCE:

NO COMPARABLE DAMAGE ON OUTBOARD FACE OF FENCE

fiG. 14. PORT BOUNDARY LAYER FENCE. IMPACT DAMAGE TO INBOARD FACE-G-ALYP. PAINT TRACES MADE BY PORTION OF CABIN WALL (FOR LOCATION SEE FIG. 16).

HEAVY METALLIC SCORING WITH BLUE PAINT TRACES CONTINUOUS OVER FRACTURE EDGE (FOR LOCATION SEE FIG. 16).

FIG. 15. PAINT TRACES ON PORT WING UPPER SURFACE-G-ALYP. STARBOARD BOUNDARY LAYER FENCE

METALLIC SCORES AND BLACK IMPACT NARKS SPANNING MAIN SEPARATION OF STARBOARD WiNG

IOlOUR SCHEIoIE: PAINT TRACE ON UPPER SURFACE BLUE OF PORT AILERON. (FIG.I)' YEllOW WHITE POLISHED UNPAIUTED IAETAL

METAlliC SCORES MW BLUE PAINT TRACES ON UPPER SURFACE OF PORT UPPER AIR 6RAKE.

IMPACT MARK MADE BY FRACTURED EDGE OF BLUE BAND ON CABIN WALL. PHOTOGRAPH Of IMPACT MARK - FIG.15 PHOTOGRAPH OF CABIN WAlL - FIG.12 LARGE PATTERN OF PAINT TRACES AND METALLIC SCORES OH UPPER SURFACE OF PORT wING, FORMED BY WRECKAGE MOVING OUTWARDS O'lER WING. PAINT TRACES "'AINLY BLUE, BUT PORT BOUNDARY LAYER FENCE OAI4AGEO. SOME YELLOW AND WHITE PRESENT. AlSO BLACK IMPACT ON INBOARD FACE BY SPANWISE IMPACT MARKS, POSS laLY 14AOE aY RUBBER. DIRECTEO FROW I"BOAr.D, FENCE SURFACE CONTAINS NETALLIC SCORES. PAINT TRACES AND PARTICLES OF FUSELAGE FLOORING .tAMMEO IN SKIN JOINTs AND CREVICES. (FIG.HI FIG. 16. IMPACT DAMAGE TO UPPER SURFACES 0'= WINGS-G-ALYP. '1'31\ FRAME 16

PEE.L.IN4 OFF FAIl.. URE. I

SECONDARY FAICURE BY SENOINq OF CENTRE poRTION OVER oUTER PORTION

FRAMES

DIRECTION OF PROPAC;ATION POSSIBLE E~TENT OF ORIC;INAC OF MAIN FAICURES FAILURE. WRECKAC;E NOT ___RECOVERED L ______---1

TOP t OF AIRCRAFT. ____~--~------_7.~~--~~------f._+~~~~__+ ~~~~~~~.------.------~---=-L------i AERIAL SI(iN5 OF FATIC;UE FOUND WINDOW IN SKIN AT THIS CORNER I I

SS.CONOAR'w' I FAILUR'E I FORWARD I I f­ a: o I a. I I I I , ___ ' ______-----J SKIN PULLE.O OVE.R (------) RIVETS ON WINDOW FRAME I I I PEEL..IN(i OFF FAICU RE I I I I I I \----_ ... ' \ _____ ,....1

oeVECOPMENT DRAWIN" SHOWIN" I.OCATION OF TliESE PANEi-S IN RE.LATION TO SURROUHDINq WRECKA"e. IS "IV EN ON FI".IB. PHOTO':;RAPH OF PANECS IS "IVEN ON fie;. 12 . FIG. 17. SALIENT FEATURES OF DISRUPTION OF PRESSURE CABIN-G-ALYp. r------, KEY------,, : ~ REMOVED BY : I DISRUPTION OF I I PRESSURE CABIN I : ~ REMOVED WITH I I ~ FRONT FUSELAGE i " ~ REMAINED ATTACHED " 10 WING CENTRE I SECTION I

REMOVED WITH ,Ill", REAR FUSELAGE " L ______.J

FORWARD

------~

----j

FIG. 18. DEVELOPMENT DRAWING OF WRECKAGE AROUND ADF AERIAL WIN DOWS-G-ALYP. r (a) PLAN VIEW OF FAILURE IN SKIN.

(b) VIEW OF EDGE A B.

(c) VIEW OF EDGE C D. FIG. 19. SIGNS OF FATIGUE IN SKIN AT STARBOARD REAR CORNER OF REAR ADF AERIAL WINDOW-G-ALYP. wreckage. de Havillands should be asked to had sustained damage of the type indicated in produce further evidence directed towards estab­ Part 6 of the RA.E. Report (which deals with lishing that the precautions taken in the Comet this subject). such damage was not the cause installation. to ensure that the pressure could not of the accident to Yoke Peter. There had. how­ rise appreciably above the normal working pres­ ever. been a recorded instance of trouble due sure. were reliable. Mr. Wilkins. an Assistant to this cause and it is to be observed that Chief Designer of de Havillands. who was de Havillands have indicated their intention of responsible for this aspect of the design. gave devising a method of removing the possibility evidence on the matter. and a statement was pro­ of damage of this kind (see Appendix VIII). duced by de Havillands summarising the method (f) The possibility of damage by jet efflux of operation of the essential controlling ~nd safety valves. Messrs. N ormalair Limited. the 112. During the operation of B.O.A.C. ser­ firm responsible for the pressurisation control vices. there had been some experience of small equipment. also produced full information about damage to the cabin skin. due to the buffeting the essential parts. Taken together with the by the efflux from the jet engines. This damage RA.E. Report. this additional evidence satisfies was partly in front of and partly behind the me that the possibility of the development of pressure dome of the cabin. As soon as it was excessive internal pressure in the cabin. of an observed. a systematic inspection was made of amount sufficient to endanger its structure. was all Comets. and where any signs of cracking were detected a repair was made according to a :;'0 remote that it can be excluded as a probable cause of the bursting of the cabin. scheme specially devised by de Havillands. In­ ternal inspection showed that the buffeting was also causing slight loosening of the joint between (e) Certain defects referred to in the R.A.E. the stringers and the skin in this region. and Report rivets were therefore inserted in order to ensure 108. I turn now to the other defects discovered that this would not give rise to danger. by RA.E.and already referred to in paragraph 113. This point of possible weakness was 93 of this Report. I see no reason to differ from under continuous observation. The steps taken the conclusion reached by RA.E. that none of to deal with it may be considered to be satis­ these defects was in any way the cause of the factory. particularly since. where the repair had accident. been carried out. no further trouble occurred. , 109. It is clear that the separation of both 114. It is. however. recognised by de port and starboard outer wings from the centre Havillands that a situation in which it is known section (see Figure 11) was not the primary cause that such cracks are likely to occur is unsatis­ of the accident. for there is ample evidence from factory. and among the improvements they the distribution of paint marks and scratches intend to make on future Comets is one which on both wings that they were made by parts they believe will reduce the cause of this damage. of the cabin structure. and form a pattern (see namely. a slight change in the direction of the Figure 16) which is consistent only with the jet pipes at their exits. with the object of whole wing having been intact when they were diverting the jets away from the sides of the made. For the same reason. the known point of cabin. fatigue weakness in the wing skin near the edge of the wheel-wells is not suspect. Moreover the PART XIV fracture of the wings occurred some distance RESPONSIBILITY' outside this region. (a) Introductory 110. As regards escape of fuel from the fuel tank venting system. examination of the 115. No suggestion was made that any party wreckage disclosed that fire did not start until wilfully disregarded any point which ought to after the disruption of the cabin. It is clear. have been considered or wilfully took unneces- therefore. that escape of fuel from the tank vents sary risks. But. in the course of the evidence. ouring take-off or climb had nothing to do with questions were put which make it necessary for the accident. me to consider a number of points in the light of the conclusion I have already expressed as to 111. Turning to refuelling. the danger appre- the cause of the accident. hended could only occur by a concatenation of five events. The risk was. therefore. said to be (b) Criticism of de Havillands' design work a remote one and in any event in the present 116. Dealing first with the period prior to the case RA.E. stat~ tha~ examinatioI?- of th~ Elba commencement of the scheduled passenger ser- wreckage made It plam that even if the arrcraft vice on the 2nd May. 1952. the calculations made 25 f.J by de Havillands were criticised and it was 65,000 lb/sq. in.) they'were confident that they suggested that the tests they carried out were could demonstrate by static test that there would inadequate to guard against the risk of fatigue be no failure at twice the working pressure, and in the cabin structure. In support of this con­ that there would be a considerable reserve in tention particular reference was made to certain hand. Their tests of panels about 3 ft. square, calculations included in paragraph 4 of Part 3 including a Window, substantiated this view. of the R.A.E. Report and to other calculations produced by Sir Arnold Hall in the course of 120. We note, however, that in these tests the his evidence. It is, however, to be observed that panel was supported on the face of a stiff steel the primary object of de Havillands was to lay "pressure box", and not in conditions truly the foundation for extensive tests which they representative of those which existed near the regarded as the soundest basis for the develop­ window in the pressure cabin itself. It is not ment of a project rather than to arrive at a possible to say what the effect of this would be. precise assessment of the stress distribution at de Havillands were reassured by the results of the corners of the cabin windows. the tests, in which the specimen withstood nearly 20 lb./ sq. in. without failure. 117. I do not think that they can justly be criticised for this approach to the problem. In 121. de Havillands used the same approach to arriving at this conclusion I have been assisted the design of the whole pressure cabin. The by a Memorandum which has been prepared static tests which they made on the two parts for me by my Assessors and which confirms the of the pressure cabin, respectively 26 and 24 ft. impression I formed from the evidence of the long, gave them confidence in the integrity of witnesses that de Havillands were proceeding in the whole cabin. Since they believed, with accordance with what was then regarded as good general support from then current practice and engineering practice. I am also satisfied that in opinion, including that of A.R.B., that this basis the then state of knowledge de Havillands can­ of design and static tests would give ample assur­ not be blamed for not making greater use of ance against risk of failure under repeated appli­ strain gauges than they actually did or for be­ cations of the working pressure, and other lieving that the static test that they proposed known causes of fatigue, they felt that the cabin to apply would, if successful, give the necessary was good for the life of the aircraft (say 10,000 assurance against the risk of fatigue during the pressurised flights, or 10 years). working life of the aircraft. The Memorandum 122. Here again, however, we note that the to which I have referred is included as para­ test sections of the cabin differed from the cabin graphs 118 to 129 of this Report. as fitted to the aircraft in several respects. In the first place, each was incomplete, and incapable (c) Memorandum by Assessors of sustaining pressure if it had not been fitted 118. During the design of the Comet de with a stiff bulkhead at the open end or ends. Havillands did not make use of calculations in It is not possible to say whether the constraint an attempt to arrive at a close estimate of the which these bulkheads imposed on the structure stress distribution near the corners of the cabin would make it stronger or weaker than when it windows. We have examined such of their calcu­ formed part of a complete cabin. But it must lations as had a bearing on this question; these be recognised that the stresses in the structure led to the stress of 28,000 lb./sq. in. mentioned near the bulkheads would be appreciably affected by Mr. Harper. It is clear that this stress refers by the constraint, and the reliability of deduc­ to an area of the skin in the neighbourhood of tions about the strength of the cabin would the corners, and may fairly be said to be an thereby be reduced. Secondly, neither section average value over a width of 2 or 3 inches. was fitted with the complete number of windows. de· Havillands believed that their method was etc. Moreover, the windows of special interest in satisfactory for the purpose they had in mind, this Inquiry, which were in the front test section, namely, the design of a test specimen. They did were rather near the bulkhead mentioned, so that not consider that a closer estimate of the highest the stresses in the skin round them might have value of the stress could be made by any method been appreciably different from those in similar which they would regard as reliable. They pre­ places in, the complete cabin. ferred to rely on tests of specimens designed on 123. The increasing attention which de the basis of their calculations. Havillands gave, during the period mid 1952 to 119. Since their estimate of the general level end 1953, to the fatigue life of pressure cabins of stress in the region investigated was less than has been mentioned in paragraphs 21 to 24. In half the ultimate strength of the material (about their repeated loading tests the front test section 26 of the cabin survived 16,000 applications of just figure "70 per cent. of the ultimate strength" over the working pressure. They felt confident was obtained by adding to the 43,000 lb/sq. in. that the Comet's cabin would have a safe life (mentioned above) due to the working· pressure, well beyond their target of 10 years in service. another 2,700 lb/sq. in. due to other known 124. The repeated loading test on Yoke Uncle loads, leading to a tota~ of 45,700 lb/sq. in. This at R.A.E. led to an unexpected failure after was contrasted with de Havillands' own estimate of 28,000 lb / sq. in. It has already been pointed , some 3,000 applications of load. Though this out that de Havillands' figure relates to an was about three times the life of Yoke Peter average over a considerable distance near the at-Elba or Yoke Yoke at Naples it was surpris­ corner of the window, and due only to the work­ ingly short, and led directly to the inference ~hat ingpressure, whereas the estimate made by there were high local stresses. Steps were, there­ RA.E. relates to a particular point where the fore, taken at RA.E. to measure the stresses near stress would be expected, on general grounds, the corner of the window, using strain gauges to reach a maximum. A direct comparison placed as near as possible to the edge of the between them is therefore misleading. Having skin where the failure started. These measure­ regard to the different approach the two figures ments led to an estimated stress of 43,000 lb / sq. cannot be said to be inconsistent. in. at the edge at the normal pressure difference of 8i lb./sq. in. 129. It is natural that de Havillands and RA.E. should have approached the problem of 125. This estimate of the stress was regarded the "safe life" of the pressure cabin of the by de Havillands as unreliable, partly because Comet from different points of view. de Havil­ the process of deriving it from the experimental lands were the designers and looked at the prob­ measurements involved some extrapolation, but lem as designers would, having confidence in also because it would imply that in their own test their methods based on their experience. RA.E. to twice the working pressure, there was a local had had virtually no previous knowledge of the stress of double this amount, namely 86,000 design background of the Comet, since it is a tb/sq. in., which is some 30 per cent. above the civil aircraft and their connection with it before ultimate strength. of the material. This apparent the 8th April, 1954, was primarily advisory in paradox can be explained by recognising that character and was wholly concerned with fatigue it neglects to take account of the effect of the of the wings. In the early stages of the Inquiry ductility of the material in relieving "stress there was, therefore, a sharp disagreement be­ concentrations" (see on this subject paras. 148 tween them on the interpretation of their calcu­ to 153 below). lations and tests. These differences of opinion 126. Calculations were made by Sir Arnold diminished in the course of the Inquiry as Hall to explore the problem in the light of such greater mutual understanding developed. While theoretical solutions as were known of the there are still minor points on which they do problem of stress distribution, round a cut-out not quite see eye to eye, a situation which is by of the shape of the cabin windows, in a cylin­ no means unusual in technical problems of such drical shell of metal under pressure. These cal­ difficulty, there is now no longer any substantial culations were not put forward as exact, but, with disagreement between them. Our own interpre­ due allowance for the fact that the window frame, tation of the situation, so far as it can be deter­ and the cabin stringers and hoop frames, would mined by existing evidence,. is set out above, and influence the result, they supported the reason­ we believe that it would be accepted by de ableness of the estimate made from measure­ Havillands and RA.E. ments on Yoke Uncle. 127. It is our view that the two results taken (d) Criticism of de Havillands' repeated loading together point strongly to the conclusion that the tests in 1953 stress in the skin at the edge of the window near 130. Another criticism of de Havillands was the corner was far higher than had been sus~ connected with the r~peated loading tests carried pected by de Havillands, and was probably over out by them in 1953. When the RA.E. test 40,000 lb/sq. in. under the normal pressure revealed the short life of the cabin structure of difference. Yoke' Uncle the question. arose as to how to 128. In the course of the Inquiry much atten­ reconcile the result of that test with the result tion was paid to an estimate, given in Part 3, of these earlier repeated loading tests. Sir Arnold para. 6 of the R.A.E. Report on the tests on Hall suggested that the explanation might well Yoke Uncle, of the stress which might be pre­ be that the 1953 tests were carried out on a nose dicted on the basis of their measurements by section which had previously been subjected to strain gauges, as probably existing in flight. The static tests up to a differential pressure of 16t 2'1l lb / sq. in. and that the effect of such a test might tors in collaboration with the designers. However, be to prolong the life of the specimen subjected if after such repair the crack continues to spread, to it. Mr. Harper said that he was aware of this it is considered as a matter of major concern, possibility but he considered that if there was possibly requiring a radical modification to the any increase in life of the nose section attribut­ design to reduce the stress which gave rise to [to able to pre-Ioading the tests so amply covered 135. For small cracks in regions not highly the life of the aircraft both at the time of the stressed the method of location is generally found tests and for the immediate future that de Havil­ to prevent further spread, provided that care is lands could safely accept the test as satisfactory. taken to ensure the inclusion of the end of the In the then state of knowledge I think this con­ crack in the hole drilled. All witnesses who dealt clusion was reasonable. with this matter in the Inquiry were agreed that location was a reasonable method of dealing with (e) de Havillands' method of dealing with cracks such cracks. 131. There is one other question bearing on 136. I am also advised that most aircraft ex­ responsibility to which I must refer. This con­ perience cracks due to one or more of the causes cerns certain cracks, revealed by the examination mentioned above and that it would, indeed, be of the wreckage (see para. 78), which had hardly practicable to insist on a standard of occurred in the process of manufacture and had design and construction which would preclude been dealt with by location, Sir Arnold Hall said completely the possibility of any crack in the that such manufacturing cracks might form foci skin. for fatigue and thus shorten the life of the struc­ 137. The methods employed by de Havillands ture. It was suggested in cross-examination that in dealing with manufacturing cracks were in no the fatigue which led to the disintegration of way different from those used to deal with other Yoke Peter had originated in these cracks, that deviations from the strict requirements of the they ought not to have been dealt with as they drawings to which the aircraft was being built. were and that accordingly some responsibility Defects whether discovered by the workman or ought to attach to de Havillands for/allowing the the inspector would be dealt with by the aircraft which contained them to be put into procedure known as " Concession" pro­ service. • cedure which varied according to whether 132. It will be convenient to deal with the the defect was classed a~ major or minor. Mr. subject of cracks generally before giving my Povey said that manufacturing cracks were re­ opinion on the specific question of responsibility quired to be dealt with as· major defects with the mentioned above. This course may also enable result that." Concession Notes" containing the the whole matter to be viewed in proper perspec­ proposals for dealing with them would have to go tive. Public concern may have been aroused by forward to the Chief Inspector and, if approved what was said during the Inquiry and it is by him, would have to be submitted to the Design important that groundless fears should be allayed. Department for final approval. In the case of 133. I am advised that it has been the general Yoke Peter three cracks were discovered in the experience that certain parts of the structure of reinforcing plates of the A.D.F. windows. The aircraft develop cracks as the result of fluctua­ action taken, which was approved by the Chief tion of load, vibration or casual damage and that Inspector and the Design. Department, was the external skin, whether in the wings, tailor " splits have been located with a 1 !16th dia. drill fuselage is particularly vulnerable. Cracks which hole". According to the then current engineering: occur during manufacture do not differ practice this action would have been appropriate materially, in their significance, from those which had the stresses been as low as de Havillands may develop subsequently save, of course, that believed them to be, but was, in fact, inappro­ their presence may indicate an unsatisfactory priate as the region concerned was one in which manufacturing process. there were high stresses. However, as I have 134. It is the ordinary practice to make careful already stated in paragraphs 116 and 117 my inspection of the structure, both during manufac­ opinion that de Havillands cannot be blamed ture· and subsequently, particularly in regions for their ignorance of the true state of affairs, it known to be specially susceptible and, if cracks follows that no responsibility attaches to them. are found, to deal with each case on its merits 138. The evidence disclosed other cracks in in the light of a now very wide experience of the Comet aircraft. Thus in the wreckage of Yoke problem. Where frequent inspection shows that Peter there was a crack in the skin at the star­ a particular crack is likely to spread, it is dealt board front corner of the rear A.D.F. window. . with by a carefully considered repair scheme, This had been located at both ends. No Conces­ either prepared by the designers or by the opera- sion Note was available in relation to this crack .28 and it would appear that there had been a defect 142. In Appendix IV para. 4 (ill). reference in the operation of the Concession procedure. is made to the problem which ·arises owing to Although this crack had spread during the life the variation among the lives. under a given load­ of the aircraft beyond one of the points at which ing cycle. of nominally identical parts. known it had been located. th~ actual fracture did not as "scatter". In the pressure cabins of aircraft take place there nor was there any sign of fatigue. there are probably a number of causes of scatter. Other cracks were referred to in Yoke Uncle and Tests of a large nUIIlber of specimens are how­ Yoke Yoke but in no case was there any evidence ever virtually impracticable and. in order to that the crack had contributed to the failure of ensure a safe life well above the minimum that the aircraft. is economically acceptable to an operator. 139. I need not pursue further the question of methods must be devised of ensuring that design. manufacturing cracks of this type since the combined with a reasonable programme of tests, statement put in on behalf of de Havillands (see can guarantee that the pressure cabins of Appendix VIII) records that if in future a crack transport aircraft will be entirely safe. does occur at any time either in manufacture or 143. The policy which de Havillands propose subsequently during the life of an aircraft, no to adopt for the Comet is directed to achieving repair scheme for such· a crack will be sanc­ this end. primarily by reducing both the general tioned unless it ensures that. after it has been level of stress and the local excesses. due to all carried out. the part of the aircraft concerned known causes. above the general level of stress. will be as strong and will have as long a life The knowledge which has been acquired as a as it would have had. had there been no crack. ;result of the investigation of the accident to , Yoke Peter. and the tests made on Yoke Uncle PART XV at R.A.E., strongly suggests that steps should FUTURE be taken to determine by calculation. by tests of typical parts of the cabin. and by tests on one (a) Statements on behalf of the Attorney-General or more complete cabins, both the distribution and de Havillands of stress throughout the structure in considerable 140. By s. 9 (12) of the Civil Aviation (In­ detail, the influences which determine both vestigation of Accidents) Regulations 1951 the the highest static load which it will sustain, and duty is imposed on me of making such recom­ its life to failure under repeated loading. In the mendations as I think fit with a view to the present state of knowledge. it is likely that two preservation of life and the avoidance of similar complete cabins will have to be tested, one under accidents in future. I have been greatly assisted static loads and one under cycles of repeated in that part of my task (a) by the statement as loads. to future policy made by Sir Lionel Heald on 144. From the evidence of Sir Arnold Hall the 12th November. 1954 on behalf of the and from advice I have received from my Attorney-General after consultation with the Assessors it became clear that there exist methods Ministry of Transport and Civil Aviation and of calculating the stress distribution in the struc­ A.R.B. : (b) by the statement put in by Sir ture of a pressure ca6in which could with Hartley Shawcross on the 23rd November. 1954 advantage be employed more widely. Moreover recording the action which de Havillands now the result of R.A.E.'s investigation satisfied me propose to take to deal with the problem of that in tests of pressure cabins or parts of them, fatigue and with the other defects referred to in the stress distribution should be determined by the Report of R.A.E. These statements are of wide use of strain gauges. This procedure will such importance that I have attached them to enable the calculations used in the design to be this Report as Appendices VII and VIII. I verified or amended, and will lead to a fuller respectfully agree with the course therein pro­ understanding of the problem. posed to be adopted. 145. When these measures have been applied (b) Further suggestions directed to guarding and the tests completed, de Havillands will no against fatigue doubt ask A.R.B. to recommend the grant of a 141. The problem of securing an economically Certificate of Airworthiness to the re-designed satisfactory safe life of the pressure cabin of an Comet aircraft. It would not be desirable for me aircraft needs more study. both in design and to say anything which might in any way limit by experiment. if. the lightest possible safe struc­ the discretion of A.R.B. but I may perhaps ture is to be achieved. This is recognised by appropriately express the hope that this pro­ de Havillands in their policy in regard to the cedure will reassure the public as to· the integrity future of the Comet (Appendix VIID. of pressure cabins and will justify Sir Arnold 29 Hall's confidence that the Comet aircraft will internal and external pressure on the cabin. It fly again. is perhaps simplest to look at this problem in the light of the situation which develops as the pres­ (c) . Use of available Government facilities sure in the cabin is increased from the working pressure P up to the value somewhat below that 146. In the course of the evidence there was at which it fails under a static test. some suggestion that prior to 1954 inadequate use was made in the development of the Comet 149. In the first place it is essential to appre­ of the unrivalled facilities available at R.A.E. ciate that. although it would from many points to the civil aircraft industry. This may have been of view be desirable that the stress in the skin exaggerated. Be that as it may. in view of the should be the same everywhere. in practice con­ importance of that industry to the national siderable variations are unavoidable. There will. economy it is essential that in the future manu­ therefore. be points. generally near to the cut­ facturers should be aware of. and should make outs. where the stress is appreciably higher than full use of. such facilities as the research estab­ the average. and it is on these points that the lishments of the Ministry of Supply can offer. designer's attention is naturally focussed when The Court was informed that in practice there considering the strength of the structure. had been close personal association between 150. As the pressure difference in the cabin members of the staffs of A.RB. and RA.E. and rises from P to. say. I! P the stresses everywhere that RA.E. was represented on the Airworthi­ will rise in the same proportion. But as the pres­ ness Requirements Co-ordinating Committee of sure difference approaches. say. 2 P the stress in A.RB. It is desirable. nonetheless. to strengthen the more highly stressed regions will reach that the liaison between A.RB. and all the research at which the material is no longer elastic. Its establishments of the Ministry of Supply and it extension will then be of a plastic nature. that might be worth considering whether. when the is to say. one which, does not disappear when Council of A.RB. is being strengthened in the stress which caused it is removed. Over most accordance with the statement made by Sir of the skin the stress will remain within the range Lionel Heald (see Appendix VII). it should not in which the material is still elastic and the also receive such additional reinforcement as removal of the pressure will restore this part of will encourage the full use by manufacturers. the skin to its original dimensions. But in areas operators and A.R.B. of all available facilities. where the stress was high. there will remain a permanent stretch. The pre-Ioaded cabin is (d) A voidance of flight by unlicensed crew therefore physically different from a new one. 147. Reference has been made in paragraphs if the pre-Ioad has exceeded a certain level. 35 and 36 to the fact that the Engineer Officer 151. Although the permanent extension of the of Yoke Peter was not in possession of a valid material in the areas where it has stretched plas­ licence at the time of the accident. I was informed tically. but without fracture. is small and unde­ by Counsel for B.O.A.C. that their system for tectable by visual inspection. it may have a ensuring the prompt renewal of licences had profound effect on the distribution of stress in been overhauled and that adequate steps have the material when the working pressure is been taken to prevent a recurrence of this lapse. applied a second time. Without going into It is clearly of the first importance to ensure that details. the general nature of this will be to no aircraft flies save with a crew not only fully reduce markedly the stress in the areas where it qualified in knowledge and experience but also was previously greatest. The stress concentration properly licensed. in such areas is therefore relieved. 152. This is a process whose general nature (e) Suggested scientific' and technical is understood. and there are examples where it investigations has been deliberately used in order to improve 148. There are certain scientific and technical resistance to fatigue. It has indeed been sug- matters on which. acting on the advice of my gested that it might be used in such structures Assessors. I recommend that research can use- as a pressure cabin. But there are obvious diffi- fully be undertaken. in the interest of increasing culties. not to say dangers. in applying it. Never- knowledge of the problems of the design of theless. the subject should undoubtedly receive pressure cabins. The first arises from the influ- more study. if only to ensure that tests during ence of the ductility of the from design are not rendered unreliable by failure to which the skin of the cabin is made. on the appreciate its significance. manner in which the stress distribution in the 153. Though there can be no direct proof. there skin is related to the difference between the is no doubt that the phenomenon described above ,30 provides at any rate a partial explanation of the inspection of aircraft parts is delegated by AR.B. apparent anomaly presented by the failure of the to manufacturers. By this system, the operation pressure cabin of Yoke Uncle at R.A.E. after of which is set out in an AR.B. pamphlet on 3,000 cycles, in spite of the survival of the test " The Approval of Inspection Organisations and specimen of the forepart of the cabin to over the Maintenance of Airworthiness", manu­ 16,000 cycles when tested by de Havillands. The facturers' own inspectors have the duty of super­ maximum pressure difference which had ever vising all the work done in building civil aircraft. been applied to Yoke Uncle was It P, whereas This inspection organisation is supervised by the test' ,specimen had been subjected to two A.R.B. through their own inspectors to ensure applications of 2 P in addition to nearly twenty that it is adequate. AR.B. inspectors do only of between P and 2 P. such detailed inspection of work as is needed to assure themselves that the system is working satis­ 154. The second question which needs study factorily. Evidence was given by Mr. Povey may be put shortly as follows: what is the true illustrating how this system worked at static strength of the complete Comet cabin? de Havillands. Reasons have been given in paras~ 120 and 122 why the tests made on sections of the cabin may 158. The suggestion was made that the system have been somewhat misleading. A test con­ for inspection would be more satisfactory if all ducted in the tank at R.AE., with the most com­ the inspectors were responsible direct to A.R.B. prehensive exploration of the stress distribution, and not to manufacturers, or alternatively that would be invaluable. Not only would it clear there should be a duplicate system of inspection up such uncertainties as remain from our Inquiry, whereby both manufacturers and AR.B. would but, in conjunction with the repeated loading have inspectors. Reference was also made to the tests already made on Yoke Uncle, would pro­ method of inspection of shipping by Lloyd's as vide an invaluable body of information for the an example of how such a system might work basis of design of future pressure cabins. but no evidence was produced as to this method. I cannot, therefore, form any conclusion on the 155. The remaining question which requires suggested analogy. study relates to the system used to operate the aircraft controls. Most of the evidence on this 159. It is plain that there would be inherent subject was concerned with the alleged excessive dangers in duplication. Responsibility for the "break-out" force and indicated a difference of quality of his product must rest with the prodt;lcer. opinion, among pilots, as to whether the existing It is, therefore, essential for the producer to have system was satisfactory in this respect, though his own system of inspection. Any additional none suggested that the alleged defect had in any system would add to expense, but not, it was way contributed to the accident. A different argued, to safety. criticism was made by one of the Assessors to 160. I have come to the conclusion that the the Indian Court of Inquiry into the accident to . present system of inspection by manufacturers G-ALYV and apparently prompted that Court's approved and supervised by AR.B., is essentially second recommendation, which was as follows: satisfactory. It is, of course, subject to human " That consideration should be. given to the de­ errors, but it has the beneficial effect of creating sirability of modifying the flying control system a sense of responsibility in manufacturers without of the Comet aircraft in order to give the pilot which aircraft could not be designed and built a positive' feel ' of airloads exerted on the control to the requisite standard of reliability and safety. surfaces." Only a passing reference was made 161. The second suggestion arose out of some to this before me. As advised by my Assessors, I criticism which was levelled at A.R.B. on the am satisfied that the characteristics of the control ground that their flight testing organisation is system of the Comet should be reconsidered by relatively small compared with similar flight de Havillands and by AR.B. in the light of both. test teams at aircraft firms and at the Ministry the criticisms which have been made. of Supply Experimental Establishments. A sug­ gestion was, therefore, put forward that AR.B. (f) Observations on certain suggestions made in flight testing and aircraft approval would be the course of the Inquiry made more effective if an active pilot were 156. I cannot conclude this part of my Report appointed to their Council and if civil aircraft without mentioning two suggestions made during were sent to a Ministry of Supply test establish- the Inquiry which, after full consideration, I feel ment where a much wider and more experienced unable to recommend. opinion on flying qualities could be obtained from a larger organisation, instead of the some- 157. The first of these arose out of some what restricted assessment at present available criticism which was made of the system whereby to A.R.B. 31 162. Although I am satisfied that there is no Question 3. . Was the accident due to the act reason to criticise the flight testing of the Comet I or default or negligence of any as carried out by de Havillands and A.R.B.• party or of any person in the em­ I think serious consideration should be given ployment of that party? to the poss~bility of obtaining the best available opinion on the flight characteristics of future Answer. The accident was not due to particularly when they incorporate the wrongful act or default or to novel features in design which effect those the negligence of any party or of characteristics. As I have mentioned in para. any person in the employment of 146 of this Report. such facilities are available any party. in Ministry of Supply Establishments. and the importance of the civil aircraft industry to the Question 4. At the time of the accident: economy of this country seems to warrant Question 4 (a). Had the aircraft been properly making the fullest use of those facilities. maintained in accordance with the 163. With reference to the suggested appoint­ current approved maintenance ment of an active pilot to the Council of A.R.B.• schedUles? If not. did any defect there are clearly difficulties in such an arrange­ in maintenance affect the safety ment since the pilot woUld be unable to do of the aircraft or contribute to the his job as an airline pilot and at the same time accident? be available to give his advice to the Council. Answer. Yes. The second part of the I have no reason to believe that the present question does not arise. representation on the Council has been in any way lacking in the past .and I hesitate to recom­ Question 4 (b). Was the aircraft airworthy so mend any change. If an active pilot were to far as coUld reasonably have been be appointed the post would have to be made then ascertained? a whole time paid employment and it woUld not be long before he ceased to possess the Answer. Yes. qualifications upon which those who advocated the appointment laid stress. On the whole I Question 4 (c). Was there a valid Certificate of think it is better to rely on the Minister to Airworthiness in respect of the secure that the person he nominates to the aircraft? Council as possessing professional experience . as a pilot of civil aircraft is always someone Answer. Yes. who is reasonably up-to-date. Question 4 (d). Was there a valid Certificate of Maintenance in respect of the air­ craft? Answer. Yes.

PART XVI Question 4 (e). Was the radio station of the aircraft serviceable and was there QUESTIONS AND ANSWERS a valid Certificate of Serviceability My answers to the questions submitted on in respect thereof? behalf of the Attorney-General are as follows:- Answer. Yes. Question 1. What was the cause of the accident? Question 4 W. Was the aircraft properly loaded and trimmed within the limits Answer. The cause of the accident was specified in the Flight Manual? the structural failure of the pressure cabin brought about by Answer~ Yes. fatigue. See para. 95. Question 4 (g). Were all members of the crew Question 2. If several factors caused the properly licensed and adequately accident what were such factors experienced to make the flight? and to what extent was each con­ If not. did any defect in the tributory? licence of any member of the crew affect the safety of the aircraft or Answer. This does not arise. contribute to the accident? 32 Answer. All members of the crew were Question 5. Upon consideration of all facts adequately experienced to make disclosed by this Inquiry what the flight but the flight engineer, steps should be taken to increase Engineer Officer F .. C. Macdonald the safety of civil aircraft? was not properly licensed to make the flight (see paragraph 35). This Answer. See Paragraphs 140-155 of this defect did not affect the safety of Report. the aircraft or contribute to the accident. (Signed) COHEN. W. S. FARREN. W. J. DUNCAN. A. H. WHEELER. 1st February, 1955.

33

ApPENDIX I LIST OF WITNESSES ERIC NEWTON, Chief Investigating Officer; Accidents PATRICK GRAEME TWEEDIE, Chief Inspector of Branch, Ministry of Transport and Civil Accidents, Ministry of Transport and Civil Aviation. Aviation. CECIL MONTIE MACK, Inspector, Grade I, employed ERIC STANLEY MOULT, Doctor of Philosophy by B.O.AC. (Engineering), Chief Engineer, de Havilland MAURICE RussELL OVENDEN, Inspector, Grade I, Engine Company Limited. employed by B.O.AC. DOUGLAS ERIC HOLLAND-MARTIN, Captain, Royal ARCHIBALD AMos ELLIOTT, Higher Executive Officer, Navy. Ministry of Transport and Civil Aviation. BERNARDO Russo, Captain, Italian Air Force. ERNEST EDWARD RODLEY, Flight Captain, B.O.AC. RICHARD HERBERT WILLIAM CLARKSON, Station Comet 11 Fleet. Operations Officer employed by B.O.AC. at WALTER LANSDOWNE BENNETT, Fleet Engineer Ciampino. Officer, B.O.AC. Comet Fleet. SIR FREDERICK WILLIAM BOWHILL, Air Chief ALBERT MEAGHER, Captain employed by B.O.AC. Marshal, Royal Air Force (Retired), Chief Aeronautical Adviser to the Ministry of Transport GEORGE GORDON STUART, Flight Engineer employed and Civil Aviation and Chairman of AS.B. by B.p.AC. LEONARD EAST, Inspector, Grade I, employed by WILLIAM GEORGE THOMAS LATIMER, Station Officer B.O.AC. employed by Corpora­ tion (hereinafter called B.E.A) at Ciampino LEWIS PmLIP BROMLEY BOWLES, Inspector, Grade I, Airport, Rome. employed by B.O.AC. GERALD ARTHUR BULL, Licensed Aircraft Main­ DONALD WILLIAM FOOTE, Aircraft Maintenance tenance Engineer employed by B.O.AC. at Engineer employed by B.O.AC. at Ciampino. Cianipino. J OSEPH EDWARD COOK, Radio Maintenance Engineer BENJAMIN JESSE FOLLIARD, Assistant Inspector of employed by B.O.AC. at Ciampino. Accidents employed by B.O.AC. PETER Ross WARDEN, Aircraft Maintenance PETER CLAUDE PINFIELD. Station Superintendent Engineer employed by B.O.AC. at Ciampino. employed by B.E.A at Ciampino. FILIPPO CUFFARO, Ground Engineer employed by LUCIANO LIPPERA, Operations Assistant employed by B.O.AC. at Ciampino. B.O.A.C. at Ciampino. CHARLES ABELL, Deputy Operations Director JOHN RICHARD JOHNSON, Captain employed by (Engineering), B.O.AC. B.O.AC. ERIC LEWIS RIPLEY, Head of the Accidents CHARLES EvANS, Member of the staff of the Opera­ Investigation Section, R.AE. tions Controller, B.O.A.C. and formerly Fleet THE RIGHT HONOURABLE LORDBRABAZON OF TARA, Navigation Officer, B.O.AC. Comet Fleet. Chairman of AR.B. PATRICK JOHN MEADE, Senior Meteorological Officer, SIR VICTOR HUBERT TAIT, Air Marshal, Royal Air London Airport. . Force (Retired), Operations Director, B.O.AC. J AN MARTHINUS BOTHA BOTES, Comet Line Captain, WALTER TYE, Chief Technical Officer of AR.B. South Mrican Airways. BRUNO JABLONSKY, Aeronautical Engineer and CHARLES GERALD FORSBERG, Commander, Royal Director of J ablo Plastics Industries Limited. Navy. ALFRED JOHN MURPHY, Director of the Department ANTONIO FORNARI, Professor and Assistant to the of Metallurgy and Professor of Industrial Director of the Institute of Forensic Medicine, Metallurgy, University of Birmingham. University of Pisa. ROBERT HENRY TRAVELL HARPER, Chief Structural ROBERT DONALD TEARE,· Doctor of Medicine, Engineer, de Havillands. Assistant Pathologist and Lecturer in Forensic RONALD ERIC BISHOP, Director and Chief Designer, Medicine, St. George's Hospital, and Lecturer in de Havillands. Forensic Medicine, St. Bartholomew's Hospital. HARRY POVEY, Director in charge of production, de SIR ARNOLD ALEXANDER HALL, Fellow of the Royal Havillands. Society, Director of R.AE. CHARLES TIMOTHY WILKINS, Chief Assistant PERCY BROOKSBANK W ALJ;S:ER, Doctor of Philosophy, Designer, de Havillands. Head of the Aircraft Structures Department, The evidence of the following was given by R.AE. affidavit :- GEOFFREY CORFIELD, Fleet Inspector, B.O.AC. RONALD FREDERICK YOUNG, Station Officer (Opera­ Comet Fleet. tions) employed by B.O.AC. at Ciampino. 35 LIST OF WITNESSES-cont. DONALD EDW ARD HILL, Traffic / Operations Officer BRUNO BORGOGNlNI, Traffic Clerk employed by employed by B.O.A.C. at Ciampino. B.E.A. at Ciampino. ANTONlO DE TOMMASO, Clerk employed by B.E.A. BENJAMIN PERCIVAL LOUIS WELLMAN, Duty Officer in the Load Control Section at Ciampino. employed by B.E.A. at Ciampino. ERNESTO ASTORINo, Station Assistant employed by ORESTE MARlNI, Freight Clerk employed by B.E.A. B.E.A. at Ciampino. at Ciampino. NlNUCClO GERI, Sailor, of Capoliveri, Elba. MARlO RUIA, Control Clerk employed by B.E.A. at VASCO NOMELLlNI, Farmer, of Portoferraio, Elba. Ciampino. LEOPOLDO LORENZlNI, Driver, of Castagneto FEDERICO CONTE, Apron Controller employed by Carducci, Leghorn.· B.E.A. at Ciampino. UMBERTO DAMIANI, Resident of Portoferraio, Elba. GlOVANNI MIRAGOLl, Control Clerk employed by B.E.A. at Ciampino. GIUSEPPE LOMBARDI, Lieutenant-Colonel and Officer Commanding the Harbour Authority, Portoferraio. FERDINANDO GIANNONTONI, Apron Control Super­ Elba. visor employed by B.E.A. at Ciampino. LUIGI P API, Fisherman. of Porto Azzurro. Elba. GlOSUE CARDUCCI, Traffic Clerk employed by B.E.A. at Ciampino. ANmONY ROBERT PIRIE, Translator. COLlN REGINALD DRURY, Duty Officer employed by GERT CORNELlS DRY, employed by the Government B.E.A. at Ciampino. of the Union of South Mrica in the Division of HENRY WYNDHAM FOLKARD, Operations Officer Civil Aviation of the Department of Transport. employed by B.O.A.C. at Ciampino. IRENO CAFIERO, Load Control Supervisor employed CLlFFORD WILLIAM ROGERS, Licensed Aircraft by B.E.A. at Ciampino. Engineer employed by B.O.A.C. at Ciampino.

APPENDIX 11 LIST OF REPRESENTATIONS THE RIGHT HON. SIR LlONEL HEALD, Q.C., M.P., personal representatives of the late First Officer MR. J. P. GRAHAM, Q.C., and MR. P. J. STUART Bury. BEVAN (instructed by The Treasury Solicitor) MR. J. WHITFORD (instructed by Messrs. Waterhouse appeared on behalf of H.M. Attorney-General. & Co.) appeared on behalf of Messrs. Aero­ THE RIGHT HON. SIR HARTLEY SHAWCROSS, Q.C., Research Ltd. M.P., MR. RODGER WlNN and MR. H. A. P. FISHER MR. G. D. EVERINGTON (instructed by Messrs. (instructed by Messrs. Linklaters & Paines) Barradale & Co.) appeared on behalf of Messrs. appeared on behalf of the de Havilland Aircraft Normalair Ltd. Company Limited. MR. J. M. SHAW, M.C. (instructed by Messrs. Janson. MR. E. J. RIMMER, Q.C.• and MR. PATRICK BROWNE Cobb, Pearson & Co.) appeared on behalf of the (instructed by Mr. K. H. Staple, O.B.E., and Mr. personal representatives of the late Captain R. M. Forrest) appeared on behalf of British J. A. Collings, M.B.E.; (instructed by Messrs. Overseas Airways Corporation: and (instructed Cardew-Smith & Ross) appeared on behalf of the by Mr. A. Conradie) appeared on behalf of South personal representatives of the late Mr. and Mrs. Mrican Airways. Brooks; (instructed by Messrs. Ward, Bowie & MR. L. G. SCARMAN, O.B.E .• and MR. ANTHONY Co.) appeared on behalf of the personal repre­ ALLEN (instructed by Messrs. Stanley & Co.) sentatives of the late Miss D. M. Eady ; (instructed appeared on behalf of the Air Registration Board. by Mr. F. R. Howell) appeared on behalf of the personal representatives of the late Miss N. Young; MR. M. A. L. CRIPPS (instructed by The Treasury (instructed by Messrs. Stikeman & Co.) appeared Solicitor) appeared on behalf of the Ministry of .on behalf of the personal representatives of the Supply. late Mr. J. Rosenberg; (instructed by Messrs. MR. D. A. GRANT, D.S.O. (instructed by The Kenneth Brown, Baker, Baker) appeared on behalf Treasury Solicitor) appeared on behalf of the of the personal representatives of the late Mr. Ministry of Transport and Civil Aviation. F. H. Greenhaugh; and (instructed by Messrs. MR. R. A. MAcCRINDLE (instructed by Messrs. Evan Charles Robinson & Son) appeared on behalf of Davies & Co.) appeared on behalf of the British the personal representatives of the late Mrs. Air Line Pilots Association: (instructed by Messrs. E. S. D. MacLachlan. Heppenstall. Rustom & Rowbotham) appeared on MR. T. HUMPHREY TILLING (instructed by Messrs. behalf of the personal representatives of the late Ingledew, Brown, Bennison & Garrett) appeared Captain A. Gibson: and (instructed by Messrs. on behalf of the Navigators and Engineer Officers Kingswell & Berney) appeared on behalf of the Union. 36 LIST OF REPRESENTATIONS-cont. COLONEL R. MINIERO and SIGNOR R. ROVERI MR. R. G. LLOYD, C.B.E. (instructed by Mr. Stuart appeared as the Accredited Representatives of the H. Lewis) appeared as an observer on behalf of Government of Italy. Messrs. British Thomson-Houston Company Limited. LIEUTENANT-COLONEL L. E. LANG and MAJoR MAITRE LEOPOLD DOR (instructed by Messrs. J. J. GRANZIER appeared as the Accredited Repre­ Holman, Fenwick & Willan) appeared as an sentatives of the Government of the Union of observer on behalf of Union Aeromaritime de South Africa. Transport.

ApPENDIX III The Court sat at Church House, Westminster, as follows:- 19th October, 1954 20th 21st " " 22nd " " 25th " " " " 28th 29th " " 1st November," "1954 2nd 4th " " 5th " 8th " 9th " " 10th " " 11th " 12th " " The Court sat at the Institution of Civil Engineers, Great George Street, Westminster, as follows:- 17th November, 1954 18th 19th " " 22nd " " " 23rd " 24th " " " " On the 26th October, 1954, the Court visited the Royal Aircraft Establishment at Farnborough.

ApPENDIX IV NOTE ON METAL FATIGUE BY THE ASSESSORS (1) The fatigue of metals has been studied for applied once, the lower is the number of applications upwards of 100 years, and there is a world wide of that load under which failure by fatigue will literature about it. The name correctly suggests that occur. And for some materials (in particular steel, metals (and indeed other materials) suffer from a type though not in general alloys of aluminium) there is of weakness not unknown to human beings. They an appreciable load below which fatigue is borne will break under a load which is repeatedly applied indefinitely ~ and then removed, though they can support a much (3) Enough is now known about the fundamental larger steady load without distress. physics of fatigue for engineers to be aware that there (2) There is one generalisation which applies to is still much to be learnt. Research is continuous. all failures due to fatigue. The higher the intensity of But this relates more to discovering how to improve the internal stress caused by the external load, rela­ the ,resistance of materials to fatigue, than to any tive to that which would just cause failure when doubts about the principles which should govern the 37 NOTE ON METAL FATIGUE BY THE ASSESSORS-cont. design of engineering structures which must have accordingly short-a few thousand cycles of a long life under loadings which vary. This is true loading. both of mechanisms. such as engines. where the (b) Failure under repeated loading when the level number of cycles of loading in an economic life of stress is low, and the life accordingly long­ runs into millions. or hundreds of millions. and of . many thousands, or even millions. of cycles. relatively static structures. such as bridges. or the wings or pressure cabins of aircraft. (6) The latter is the type of failure due to fatigue with which engineers are mostly concerned. and with (4) These principles have been arrived at by long whose visible symptoms they are most familiar. This experience. and are widely known and soundly is because most components whose design is governed founded. They are also simple. The highest stress to largely by considerations of repeated loadings must which the material will be subjected during a cycle have very long lives and ample margins of safety, must be limited by three considerations:- since failure would be costly. if not calamitous. The (i) For a given "life". expressed as the number very existence of multitudes of high speed mechan­ of cycles of loading. there is for a given isms. engines. turbines. etc.. testifies to the under­ material an upper limit to the allowable standing of it which now exists. The appearance of stress. This can be determined only by the corresponding type of failure is characteristic of experiment. a process which has endured for some time. (ii) In the material of an engineering structure. (7) Failure due to fatigue at a relatively high level or components of a mechanism. there exist of stress is comparatively rare. Virtually no engineer­ inequalities of stress. some known and some ing components are designed to have a short life unknown. For the former. which should be under alternating loads. Nor. owing to the influences avoided as far as possible by design. allow­ mentioned above in para. 4. would it be easy so to ance can be made. The latter arise for design them satisfactorily. Not only are the symptoms various reasons. such as unavoidable im­ of failure when it occurs after relatively few cycles perfections in workmanship. or the use of therefore less familiar. but they are also less specific. riveted joints. For these provision can be The process has not endured long. and most of the made only by adopting a conservative out­ symptoms of a disease which spreads gradually are look based on laboratory experiments and on absent. experience in .service. and by meticulous inspection during manufacture. (8) In the evidence given by Professor Murphy. Professor of Industrial Metallurgy at the University (iii) When due allowance has been made for the of Birmingham. are two remarks which summarise points mentioned above. there remains a the situation as seen by an experienced metallurgist. phenomenon known as " scatter •• -the He said: -" I am strongly impressed by the high appreciable. in some cases large. variation general level of stress round these spots;* and I do among the lives. under a given loading cycle, not feel it necessary, or even relevant. to try to of nominally identical parts. Tests of a pursue the last final speck. as it were. which set number of specimens are necessary in order off the fatigue failure. I think in various places. to assess it. and the "safe life" under there was a readiness for fatigue failure. and it re­ service conditions must be determined by quired only a mild stress raisert to initiate failure." reference to them. "The prime consideration is, I think. to get the (5) In certain evidence and exhibits in the Inquiry general stress level down so that you are not in the terms " high level ., and " low level ., fatigue were jeopardy in this way. If you can get the general used in a way which may have suggested that two stress level down. you can tolerate the variations distinct kinds of fatigue are known to exist. The which are very liable to happen in practical manu­ distinction was one of convenience only. and arose facture. If you have a high stress level you are setting because of the contrast between the circumstances yourself the task of guarding against most minute of two classes of failure due to fatigue. namely :- variations from the ideal." (a) Failure under repeated loading when the level of stress is high in relation to the ultimate * The corners of the cabin windows tensile stress of the material, and the life is t Such as a rivet or bolt hole

38 APPENDIX V DIARY-MAIN ITEMS-COMET G-ALYP Date Date Item recovered received R.A.E. Floating wreckage, Private Trawling (including one part of centre fuselage) ... 26.1.54 onwards Pressure dome, parts of rear fuselage 22.2.54 18.3.54 Main items, wing centre section (engines about same time) 15.3.54 5.4.54 Front fuselage ... 3.4.54 15.4.54 Starboard outer wing (17-20) (other parts followed later) 7.5.54 22.5.54 Tail Unit 26.5.54 28.5.54 Centre fuselage-port side piece 16.6.54 21.6.54 Fin 9.6.54 30.6.54 Search re-orientated-6.7.54 Centre fuselage-piece containing A.D.F. aerial windows 12.8.54 31.8.54 Outer portion, port tailplane 21.8.54 31.8.54

ApPENDIX VI Accident Note No. 270 September, 1954

ROYAL AIRCRAFT ESTABLISHMENT, FARNBOROUGH Report on Comet Accident Investigation PART 1

OUTLINE OF THE INVESTIGATION

1. Introduction Details of investigations of the structural strength Thls report gives an account of the investigation of the aircraft. including its resistance to metal of the Comet 1 aircraft undertaken by the Royal fatigue. will be found in Parts 3. 4 and 5. Aircraft Establishment after accidents to Comet Arguments relating to the possibility of excessive G-ALYP near Elba on 10th January. 1954. and to pressures being built up in the fuel tanks and the Comet G-ALYY near Naples on 8th April. 1954. cabin. covering explosion and pressure refuelling in­ When this investigation was started. little of the vestigations. are outlined in Parts 6 and 7. wreckage of the aircraft which met disaster near The possibility of loss of control due either to Elba had been recovered. and it was not known aerodynamic characteristics or to malfunctioning of how much of the wreckage would prove to be re­ the automatic pilot and power control systems. is coverable. A series of technical investigations of the examined in Part 8. properties of the aircraft were therefore initiated and The results of tests on 1/ 36th scale dynamic proceeded simultaneously with the examination of models of the Comet are given in Part 9. the wreckage as it was received. Inevitably examina­ tions were made of items which were subsequently The results of flight trials of Comet aircraft shown. from the wreckage. to have played no part G-ANAV are given in Part 10. in the disaster at Elba although some may be rele­ Relatively small miscellaneous investigations are vant to the Naples accident. from which very little contained in Part 11. wreckage has been recovered. In Parts 2 to 11 of A note on the me!lical evidence. contributed by this report win be found accounts of all these investi­ the Institute of Aviation Medicine. is contained in gations. but detailed reference will be made in this Part 12. summary only to those shown to have possible rele­ vance to the accidents near Elba and Naples. 2. Discussion of the accident at Elba The circUInstances of the accidents and the results The examination of wreckage show_s that struc­ of the detailed -examination of the wreckage of tural failure of the aircraft occurred in the following G-ALyP are set out in full in Part 2. pattern. First. there was a violent disruption of the 39 central part of the pressure cabin. Then the fuselage We notice in the wreckage (part 2) that cracks had aft of the rear spar, the nose of the fuselage, and formed near a cut-out during manufacture of the the outer portions of the wing fell away, all under Elba aeroplane, and their ends had been drilled to the action of "downward" forces. Then the main stop propagation. Their presence suggests that part of the wing, now a separate entity, caught fire. incipient cracks, undetected at the time, may have The fuselage aft of the rear spar, complete with been induced in the material by the processes used tail unit. fell into the sea in a single piece, falling in in the manufacture of the aeroplane, and this may an " open end first, tail plane aft" attitude. The main have brought about some acceleration of the onset part of the wing hit the sea in an inverted position. of fatigue. The full evidence on which the above outline is based The possibility that the failure of the cabin might is to be found in Part 2. have been brought about by the development of The flight plan which the aircraft was following abnormally high internal cabin pressure, due either would have brought it to an altitude of about 30,000 to mis-operation of the ventilation and pressurisation feet at the time of the accident. Supporting evidence system, or to explosion in the main cabin or in the showing that the break-up occurred at this height is equipment or luggage bays is dealt with in Part 7. as follows. Metallurgical examination of the burnt No evidence can be found to suggest that such an centre section shows that the fire was burning for event occurred, or is likely to occur in a form about three minutes (part 2) and calculations and leading to disruption of the cabin, unless the cabin model tests (part· 9) confirm that the time of descent was already weakened, or unless an explosive were of the centre section from 30,000 feet altitude would deliberately placed; no evidence has been found to be of this order. Tests on models (Part 9) confirm suggest that the latter occurred. . that a "break up" of the kind outlined, at this Although the development of abnormally high -tail altitude, would produce on the ground a pattern of loads might lead to the disruption of the cabin, it wreckage similar to that found on the sea-bed near is unlikely that they would cause an initial break at Elba, and that the motions of the larger parts would the place at which it is thought (part 2) the cabin be of the. type which would lead to impact with the first failed. Apart from this, however, the following sea in the directions which are thought to have evidence appears to be relevant. Strength tests on the occurred in the accident (part 2). tail-plane (Parts 4 and 5) showed it to be structurally The sequence of events outlined above appears to sound. High tail loads might have developed if the be consistent with the evidence obtained from automatic pilot or the power-operated-control system examination of the bodies of the victims of the mis-operated in a serious way. No evidence has been accident (part 12). found from the wreckage to suggest that this To produce a violent disruption of the pressure occurred, and examination of the system suggests cabin, there must either have been a structural failure that such mis-operation is unlikely (part 8). High under normal flight loads (including the internal tail loads could also occur were the pilot to "over­ pressure load), or the internal pressure must have control " the aircraft, perhaps subsequent to a violent been raised to an abnormally high value, or a disturbance of his flight path by a gust. A consider­ failure must have been induced due to the develop­ able sample of gust records made in previous flights ment of abnormally high tail loads. of Comet aircraft have not suggested that gusts of such In Part 3 is an account of an investigation of the violence are likely to be encountered at the altitude strength of the cabin in its ability to resist " fatigue". of the disaster (Part 10). Opinions of test pilots who This work demonstrates beyond doubt that the have handled the aircraft in the course of these fatigue life of the cabin is relatively short* . The investigations are that it is unlikely that they would failures produced on test originate at points of con­ be led into difficulties by the control system (part 10), centration of stress near the corners of "cut outs" although they would prefer that the "break out" in the structure of the cabin which accommodate force associated with the elevator control were windows and hatches. The general character of the smaller. Experiments using a simulator which repro­ concentration of stress is similar at each cut-out. duced, in the laboratory, the characteristics of control Evidence from the wreckage suggests strongly that and response of the Comet aircraft, and which are the initial failure in the accident occurred at one of reported in Part 8, did not reveal any likelihood of these stress concentrations (part 2) and was of the such loss of control. same nature as the fatigue failures produced on test 3. Other technical matters at other similar points of concentration (Part 3). It is shown in Part 3 that, quite apart from the direct In the course of these investigations, it was shown evidence from the wreckage, the test results can be (Part 3) that the wing has a relatively low resistance interpreted as meaning that fatigue failure of the to fatigue. This did not cause the Elba accident, for cabin, at the age of the aircraft involved in the Elba the wing did not fail in the "fatigue prone" region, and Naples accidents, is possible. A failure of the but it may have a bearing on the Naples accident, type envisaged would cause the violent disruption of from which there is no wreckage to examine. There the centre fuselage which occurred in the accident. are some reasons (Part 3) for supposing this to be unlikely in that particular case. * A structure with an ample reserve of strength in its early It has been shown (Part 10) that fuel from the fuel­ life may in the course of time fail" in fatigue" through repeated applications of load. The term " fatigue life" means tank venting system can enter the trailing edge area the period required to produce failure under these conditions. of the wing, near the jet pipe shrouds, and there is 40 therefore some possibility of fire occurring in this (ii) The cabin was the first part of the aeroplane region of the aircraft, in which there are no fire to fail in the Elba acident (part 2). detectors or fire fighting units. Such a fire was not the (ill) The wreckage indicates that the failure in the cause of the Elba accident, as is shown by the cabin was of the same basic type as ,that wreckage. produced in the fatigue test (Parts 2 and 3). Investigations into refuelling, which are discussed (iv) This explanation seems to us to be con­ fully in Part 6, indicate that the outer wing tanks are sistent with all the circumstantial evidence. liable to sustain internal damage during refuelling under conditions which, though abnormal, may some­ (v) The only other defects found in the aeroplane times have occurred in practice. An accident is (listed in Section 3) were not concerned at unlikely to be caused by such damage, and, in par­ Elba, as demonstrated by the wreckage. ticular, it did not cause the Elba accident, as is shown Owing to the absence of wreckage, we are ,unable by the wreckage. t.o form a definite opinion on the cause of the accident 4. Opinion near Naples, but we draw attention to the fact that We have formed the opinion that the accident at the explanation offered above for the accident at Elba Elba was caused by structural failure of the pressure appears to be applicable to that at Naples. cabin, ,brought about by fatigue. We reach this opinion for the following reasons:- 5. Acknowledgement (i) The low fatigue resistance of the cabin has We ,are grateful for the help which we have been demonstrated by the test described in received m. conducting this investigation from all Part 3., and the test result is interpretable as authorities and organisations concerned in tbecon­ meaning that there was, at the age of the struction and operation of Comet type aircraft. and Elba aeroplane, a definite risk of fatigue from those responsible for the recovery of the failure occurring (Part 3). wreckage.

APPENDIX VII STATEMENT MADE TO THE COURT BY SIR LioNEL HEALD, Q.C., M.P., ON THE 12TH NOVEMBER, 1'954 When I addressed the Court at the opening of this Government, I think that you will feel that this is Inquiry I expressed the confident belief that everyone ,an important matter. concerned would give willing co-operation, not only In the first place the Board has already indicated in ascertaining the cause of the Comet disasters, but its intention that complete cabins of pressurised air­ also in. the taking of constructive measures to improve craft shall be submitted to tank tests similar to those the safety and efficiency of British aviation. I think used at Farnborough, at any rate until knowledge the Court will agree that that is the spirit in which of the fatigue problem has become much more exact. this Inquiry has been conducted. It certainly has Before the certification of certain prototypes can be ,enormously facilitated my task and I believe it has contemplated by the Board, therefore, provision will also lightened the labours of the Court. now be required for fatigue testing of entire com­ I have been hopeful frQm the very beginning of ponents, and in the Board's view at least two this Inquiry that it might be possible for me, as airframes of each prototype will have to be made representing the Crown in the interests of the public, available, one for static testing and the other for to put before the Court some suggestions, however fatigue testing. broad in outline, as to how the machinery for Secondly, as you have heard in evidence, the whole ensuring the safety of civil flying might be of the country's technical resources can already be strengthened and improved, but it obviously would made available to the Board, but in view of the have been of no value to attempt to do this-indeed difficult problems facing aircraft designers and the it might have seemed to be mere impertinence on Board at the present time, careful consideration has my part-unless I could first be assured that any been given to the advantages which can be derived suggestions that I might make would have the from the inclusion in the Council of the Board of support of all those most closely concerned with the one or more additional members, specially selected matter. I have 'therefore, consulted the Ministry of for their eminence in, the scientific field. It is pro­ Civil Aviation and the Air Registration Board, and posed to take action on these lines. I am glad to be able to inform the Court of certain practical steps which the Air Registration Board Various other measures are under consideration propose to take, with the Minister's approval. as a but I do not think it would be of assistance to the result of the consideration that they have both been Court to go into these in detail. I may, however, giving to this matter for several months past. perhaps usefully mention two points. First, a purely practical matter, the value of test flights W. the shape Since the Air Registration Board, as the Court is of extensive flying on proposed routes with fully well aware, now represents all the main interests in , instrumented aircraft, especially in relation to aircraft aviation, and since the Minister represents the with engines of which there is no background of 41 experience. for example. in the Royal Air Force. high grade light alloys has been sponsored for a The second point is concerned with research. As long time by the Ministry of Supply and work on you have heard in evidence. the Board has been the programme has been intensified during the past concerned for some time about the structural two or three years. The subject is continually before integrity of modern high performance pressurised the appropriate Committees of the Interservice transport aircraft. The importance of research on Metallurgical Research Council. Although the subject fatigue in aircraft has also become mcreasingly is a difficult one and rapid results cannot be expected. apparent in recent years and much valuable work it is fully recognised that the best brains and has been done on this by the Air Registration Board. resources ought to be directed to it. and the Ministry in the Industry and in and the Ministry of Supply; of Supply has agreed to give the fullest co-operation In particular. research on the problem of fatigw: in in this work.

ApPENDIX Vill STATEMENT FOR SUBMISSION TO THE COURT ON PROPOSED FUTURE ACTION BY DE HAVILLANDS 1. Now that the danger of high level fatigue in and climb. These consist of Modification 755 and pressure cabins has been. generally appreciated. de Modification 1404 (the siphon break). In addition Havillands will take adequate measures to c;lea1 with the vent exit will be taken into the jettison pipe so this problem. Naturally these measures will be taken as to ensure that any fuel which may be vented in full consultation with A.R.B. To this end we pro­ will be discharged clear of the aircraft. pose to use thicker gauge materials in the pressure cabin area and to strengthen and re-design windows 5. Refuelling and cut-outs. and so lower the general stress to a de Havillands recognise the desirability of re­ level at which local stress concentrations either at moving the possibility of damage from this source rivets and bolt holes. or such as may occur by reason and are devising a method of doing this. As Mr. of cracks caused .accidentally during manufacture Wilkins stated it is not possible to particularise at or subsequently. will not constitute a danger. In present what the modification will be but de addition de Havillands are already engaged on an Havillands have no doubt that the risk of damage extensive programme of detailed testing in order to . can be removed. In addition de Havillands will establish a design technique which will minjmise advise operators to incorporate a flow meter into the effect of such local stress raisers as are necessarily the refuelling unit. inherent in the design. Further. de Havillands will carry out repeated loading tests on the lines indicated 6. Handling Characteristics by Mr. Bishop.* de Havillands are considering a modification of 2. Every possible precaution will be taken by the the break out force to suit the convenience and com­ use of appropriate design and manufacturing tech­ fort of the pilot. They are investigating the possibility niques and by stringent inspection procedure further of considerably reducing the break out force below to mjnimjse the possibility of cracks occurring in the present figure of 18-20 lbs. A reduced break out . force has already been incorporated in the design manufacture. If a crack does occur at any time either of the Comet ill. in manufacture Ol" subsequently during the life of an aircraft. no repair scheme for such a crack will 7. Hydraulic Fluid be sanctioned by de Havillands unless it ensures de Havillands have carried out exhaustive investi­ that after it has been carried out the part of the gations into the possibility of using non-inflammable aircraft concerned will be as strong and will have as, hydraulic fluid. We shall continue to carry out fur­ long a fatigue life as it would have done had there ther investigations and we hope that high priority been no crack. will be given to the development of these fluids to make them suitable for use in modern commercial 3. Wing Fatigue aircraft. de Havillands are re-designing those parts of the wing structure which have been shown to be prone 8. Jet Buffet to fatigue in order to reduce the stress level. The de Havillands have reduced det buffet on the measures to be taken will probably include an in­ underside of the fuselage by slightly altering the crease in the thickness of certain parts of the bottom angle of' the jet engines in Comets II and ill. The skin and reinforcing the area aft of the wheel well. increase in the thickness of the fuselage skin (see Para. 1) will also serve further to mjnjmjse the risk 4. Fuel Venting of damage from this source. Modifications have been devised which it is be­ lieved will prevent the venting of fuel during take-off 9. Accidental damage to doors and hatches Attention will be given to reinforcement of areas * Mr. Bishop indicated that de Havillands would test specimen sections before testing the complete pressurised which are susceptible to damage from loading of fuselage. passengers and freight. 42 THE CIVIL AVIATION ACT, 1949 THE CIVIL AVIATION (INVESTIGATION OF ACCIDENTS) REGULATIONS 1951

Report of the Public Inquiry into the causes and circumstances of the accident which occurred on the 8th April, 1954, to Comet aircraft G-ALYY AIRCRAFT: Comet G-ALYY ENGINES: Four de Havilland Ghost 50 REGISTERED OWNERS: British Overseas Airways Corporation OPERATORS: South Mrican Airways (under charter) CREW: Senior Captain W. K. Mostert-Killed First Officer B. J. Grove-Killed Navigation Officer A. E. Sissing-Killed Flight Engineer Officer A. R. Lagesen-Killed Radio Officer B. E. Webbstock-Killed Steward J. B. Kok-Killed Air Hostess P. Reitz-Killed PASSENGERS: 14-All Killed PLACE OF ACCIDENT: Over the Mediterranean, S.E. of Naples. TIME OF ACCIDENT: 8th April, 1954, at about 1910 G.M.T.

All times in this Report are G .M. T.

PART I arrangements made between British Overseas Airways Corporation (hereinafter called INTRODUCTORY "B.O.AC.") and South African Airways. I (a) Matters in common with the Report on need not go in full into the history of the G-ALYP arrangements between the two operators.· Suffice it to say that the arrangements were revised on 1. In my Report of today on the accident to the 3rd October. 1953 and it was agreed. Comet aircraft G-ALYP (sometimes called Yoke amongst other things. that South African Air­ Peter) I gave a short explanation of the constitu­ ways should participate with B.O.AC. in the tion and functions of the Air Registration Board operation of the. standard class services between (AR.B.) and of the Air Safety Board (AS.B.) England and the Union of South Africa by which I need not repeat here. It is also operating Comet aircraft chartered from unnecessary for me to repeat the account I gave B.O.AC. The Corporation trained the necessary in that Report of the origin and history of the South African Airways crews to carry out this Comet aircraft. arrangement. Amongst the aircraft so chartered 2. As the two Inquiries were conducted to South African Airways was Comet G-ALYY together. the evidence in the Inquiry into the (sometimes hereinafter called Yoke Yoke). loss of Yoke Peter is the evidence in the present Inquiry. I need not. therefore. append any lists of the witnesses or parties represented at the PART 11 hearings or the dates of such hearings. THE ACCIDENT (b) Arrangements with South African Airways 4. Yoke Yoke left Ciampino Airport. Rome. 3. South African Airways are the national at 1832 hours on the 8th April. 1954 on a flight operators· of the Government of the Union of to Cairo. After taking off the aircraft from time South Africa. Air communication between to .time gave its position by radio telephone to London and South Africa was carried on under Rome Air Control at Ciampino and on the last 43 such occasion at about 1857 hours reported that engineers and expressed to be valid for 75 flying it was abeam Naples and climbing to 35,000 ft. hours, was issued. Further reference to this Cer­ This position and those given earlier indicated tificate is made in paragraphs 21. and 22 ?f this that. the flight was proceeding according to the Report. On the 7th April, 1954, an Aircraft B.O.AC. flight plan. At 1905 hours Cairo Radio Station Certificate of Serviceability was received a signal from the aircraft reporting its issued and showed no items unservic.eable. departure from Rome and giving its estimated 10. At the time of the accident Yoke Yoke time of arrival at Cairo. Thereafter no message had had a total flying life of about 2,704 hours, was received from Yoke Yoke and all attempts including 841 since the renew:al of its Certificate to make contact failed. of Airworthiness and including less than 75 hours 5. A chart, which is Figure 1 of my Report since the issue of .the Certificate .of Maintelilance on Yoke Peter, was prepared by a Navigating on the 2nd April, 1954. Officer of B.O.AC. from all the information 11. From examination of the airframe and available, and shows the probable flight track engine log bo()ks and maintenance records it of the aircraft. It also indicates the position in appeared that all routine inspections of airframe which bodies and wreckage were found on the and engines had been regularly carried out day following the accident. It is evident from withiN. the limits of time specified by the the chart that something catastrophic happened Approved Maintenance Schedules and that the to the aircraft at about 1910 hours when it must flying life of each of the engines since its last have been at or near the end of its climb to complete overhaul was within, and in two cases 35.000 ft. very well within, ·the approved life between com­ plete overhauls. Save as mentioned ·in paragraphs PART III 21 and 22 of this Report the evidence disclosed THE AIRCRAFT no c irregularity in connection with any 'sl!lch inspection. 6. Yoke Yoke was the same in all relevant respects as Yoke Peter. Details of Yoke Peter PART IV are given in my Report thereon and I need not repeat them here. THE CREW 7. Yoke Yoke was granted a Certificate of 12. Senior Captain Willem Karel Mostert. Registration No. R.3221/l on the 18th who was in command of Yoke Yoke was born September, 1951 in the name of B.O.AC. as on the 27th April, 1916. Before joining South owners and ,first flew on the 10th September, African Airways he had flown 2,812 hours in 1952. On the 23rd September, 1952 it was certi- the South African Air Force and had served as fied and approved by AR.B. for the issue of its a flying instructor. He joined South African Air- Certificate of Airworthiness and this Certificate. ways on the 10th June, 1946, was promoted No. A.3221, was issued by the Ministry of Civil Captain on the 1st November, 1946 and on the Aviation on the 30th September. 1952. After 15th June, 1949 became a Flying Instructor. On approval by AR.B. on the 21st September, 1953 the 15th May, 1953, he became Senior Hying the Certificate of Airworthiness was renewed on Instructor and ,on the same day was promoted the 23rd September. 1953 and was valid at the to .the rank of Senior Captain. In June, 1953, time of the accident. Captain Moster! was transferred to the ,Comet 8. After the accident to Yoke Peter on the Line -of South African Airways and became the 10th January, 1954, special checks, in addition Comet Line Instructor. In South African Air- to the routine Check 4 in accordance with the ways. captams who are appointed Line Instruc- Approved Maintenance Schedules, were· carried tors have to spend two-thirds of their time on out on Yoke Yoke and a number of modifica- route flying and one-third on instruction within tions were made affecting the airframe, the con- the line. During his service with South African troIs and the fire detection and protection at the Airways Captain Mostert flew a total of 8,159 engines. On the 15th February, 1954, the fuselage hours of which about 51 hours by day and 35 was subjected to a proving test to 11 Ib I sq. in. hours by night were flown in Comets within the The aircraft was returned available for service six months -preceding the accident. on the 24th February. 1954. 13. Captain Mostert's last "six monthly 9. On the 2nd April. 1954, following a Check check" prior to the accident was carried out on 1 inspection in accordance with the Approved the 19th December, 1953 and his report was: Maintenance Schedules, carried out at London " Proficient. (Very well executed flight)". He Airport. a Certincate of Maintenance signed by 1lad not been involved in any previous ac~dent. d;uIy licensed aidram.e and engine maintenance Captaj;]a Moster! was -the holder of a Umon of 44 ~'llt1lh\ Abica .Arir Line Transport Pilot's Lieence Licence No. 348 valid until the 30th April, 1954 N@'. 65A valid unti11ihe 11th June, 1954. A rating and I am satisfied that he was a capable officer. fur Comet aircraft had been added to this licence . 17. Flight Engineer Officer August Ranwald by the British Ministry of Transport and Civil Lagesen was born on the 22nd May, 1920. He Aviation. I am satisfied that Captain Mostert had wide experience of several types of aircraft was fully equipped to carry out his normal duties both during the war and after rejoining South as a pilot and as a captain and to deal with African Airways on the 16th February, 1945. emergencies. There was no positive evidence relating to his 14. The seeond pilot was First Officer Barent flying hours prior to the 11th May, 1950 but such Jacobus Grove who was born on the 15th July, records as were available suggested that up to 1922. After service in the South African Air that date he had flown a total of about 4,300 Force, in which he had flown a total of 1,640 hours. After the 11th May, 1950 he had a total houts. he joined South African Airways on the flying time of 2,290 hours 35 minutes. He had 29th January, 1953, as a First Officer and was flown about 203 hours in Comets including about posted to the Comet Line on the 26th February, 141 hours during the 90 days preceding the acci­ 1953. While with South African Airways First dent and had completed a Comet Conversion Officer Grove flew for a total of 54 hours, Course on the 2nd September, 1953, a Comet including about 47 hours in Comets during the Refresher Course on the 19th December, 1953 90 days preceding the accident. There was no and a further refresher course and flight training evidence of First Officer Grove having been programme on the 21st March, 1954. He was involved in any previous accidents Save as a examined on the 19th December, 1953 and found result of enemy action. His last check took place proficient. Flight Engineer Officer Lagesen was on the 20th February, 1954, when he obtained a the holder of a Union of South Africa Aircraft satisfactory pass. First Officer Grove was the Maintenance Engineer's Licence No; 387, valid holder of a Union: of South Africa Senior Com­ until the 26th February, 1955, and Flight nietcial Pilot's Licence No. 48 (S), valid until the Engineer's Licence No. 10 valid until the 22nd 11th June~ 1954, to which a Comet rating had February, 1955. I am satisfied that he was a been added on the 2nd March, 1954. I am satis­ capable officer. fied that he was fully equipped to carry out his 18. Air Hostess Pamela Reitz, who was born normal duties and to support his captain in . on the 16th February, 1932 and Steward Jacobus emergencies. Bruwer Kok, who was born on the 18th Decem­ 15. Navigation Officer Albert Escourt Sissing ber, 1918 had both flown extensively with South was born on the 1st January, 1917. After training African Airways. in the South African Air Force he joined South African Airways on the 16th October, 1946 and PART V from then until his death had 4,840 hours flying THE PASSENGERS AND CARGO experience including about 155 hours in Comets . . in 1953 and about 51 hours in Comets during 19. Yoke Yoke carried 14 passengers all of 1954, all of the latter during the 90 days pre­ whom were killed in the accident. There was ceding the· accident. At his last six monthly nothing in the cargo which could have been rele­ check, in March, 1954, he passed in Comet vant to the cause of the accident and I am satis­ Refresher Flight Planning and Plotting. Naviga­ fied that, despite the off-loading of a small bag tion. Officer Sissing wM the holder of a Union of aircraft spares at London after the Load Sheet of SOuth Africa Navigator's Licence No. 17(N) had been completed, the aircraft was loaded and valid until 1st December, 1954 and i am satis­ trimmed within the prescribed limits. fied that he was a capable officer. 16. Radio Officer Bertram Ernest Webbstock PART VI was born on the 17th June. 1917. He joined South PRE-FLIGHT INCIDENTS African Airways on the 23rd April, 1946 and after spending some time on the London service 20. Yoke Y oke~ in common with the rest of passed a Comet course on the 20th June, 1953 the Comet fleet of B.O.A.C., had been grounded and thereafter flew only in Comets. His total by B.O.A.C. after .the 3;ccident to Yo~e Peter. flying hours were 4,373 of which about 98 houts The circumstances m which Comet servIces were were during the 90 days preceding the accident. resumed are fully stated in paragraphs 54 to 57 He was passed as proficient in his Comet check of my Report on the accident to Yoke Peter and on. the 5th October, 1953. Radio Officer I need not repeat them here. Webbstock was the holder of a Union of South 21. Yoke Yoke arrived at Ciampmo on the AIrica First Class Flight :Radio Operator's 7th April from London and was due to depart 45 from Ciampino the same evening. However, on was encountered either during the climb through completion of refuelling it was discovered that the cloud layers or in the clear air above. It can. the centre tank contents gauge showed no read­ therefore, be assumed that the state of the ing although the tank was full. The fault was weather was not a contributory cause of the eventually traced to a co-axial cable for which accident. a replacement had to be flown from England and the departure of the aircraft was consequently PART VIII delayed for about 24 hours. While the fault was ACTION TAKEN AFTER THE ACCIDENT being traced a number of bolts were found lying 26. As in the ca,se of the accident to Yoke about in the port wing of the aircraft and further Peter the assistance of the Royal Navy was in­ inspection revealed that an equal number of bolts were missing from the inspection panel voked and on the 9th April, 1954, H.M.S. providing access between the rear spar and the Eagle and H.M.S. Daring proceeded to search wheel-well wall and that the remainder of the for Yoke Yoke. Avenger aircraft of H.M.S. bolts securing the panel, though in position, were Eagle were used to assist in the search as also not properly tightened. The missing bolts were were certain United States aircraft. A replaced and all were properly tightened. The number of dead bodies as well as some maintenance engineer who supervised this work aircraft seats and other wreckage were was satisfied from visual examination and from identified in the water and jn due course the readiness with which the missing bolts were recovered .. The depth of water where the bodies refitted that no distortion of the panel or adjacent and wreckage were found varied between structure had occurred during the absence of the approximately 520 fathoms and 580 fathoms bolts. and the evidence established that at that depth the prospect of further recovery was hopeless. 22. As has been stated in paragraph 9 a Check 1 inspection was carried out on Yoke 27. The six bodies recovered were not Yoke before the issue of the Certificate of Main­ examined by Professor Fornari, who had tenance on the 2nd April. It is quite clear that examined the bodies recovered at Elba, but four it must have been during that inspection that of them were examined at Uxbridge on the 12th the panel was removed and incorrectly refitted April, 1954 by Dr. Teare, one was not subjected and I was informed that disciplinary action had to autopsy and the other was examined by the been taken against the inspectors concerned. Italian authorities. 23. The arrangements for safeguarding the air­ These examinations did not disclose anything craft during its stay at Ciampino were the subject inconsistent with the view that the accident to of a great deal of evidence. For the greater part Y oke Yoke was attributable to the same cause of this period Yoke Yoke was under observa­ as the accident to Yoke Peter. tion by B.O.A.C. officials whose duties, however, 28. As a result of the accident to Y oke Yoke were not primarily concerned with. security. For the Royal Aircraft Establishment (hereinafter the rest of the time it was guarded by an Italian referred to as RA.E.) were directed to conduct Finance Guard whose main duty was to prevent a full investigation into it and the accident to smuggling. In all the circumstances I consider Yoke Peter. In the absence of any wreckage from it unlikely that any unauthorised person gained Yoke Yoke RA.E. could only proceed with their access to the aircraft. investigations in the light of a priori reasoning 24. Apart from the above-mentioned defects, and experiments and of conclusions to be drawn the Refuel and Departure checks disclosed from the wreckage of Yoke Peter. I have dealt nothing unusual. at length ·with the RA.E. investigations and Report in my Report on the accident to Yoke Peter. PART VII WEATHER CONDITIONS AT THE TIME PART IX OF THE ACCIDENT THE COURT'S CONCLUSION AS TO 25. From the take-off at Rome at 1832 hours CAUSE OF ACCIDENT on the 8th April, 1954 until the time of the 29. RA.E's conclusion as regards the cause accident, which was approximately 1910 hours, of the accident to Yoke Yoke is expressed in the Y oke Yoke climbed through three moderately following paragraph: _co Owing to the absence thick layers of cloud. In the top layer there may of wreckage, we are unable to form a definite have been slight to moderate icing conditions opinion on the cause of the accident near Naples, but these would have been insufficient to cause but we draw attention to the fact that the anxiety. It is unlikely that any severe turbulence explanation offered for the accident at Elba 46 appears to be applicable to that at Naples". I PART XII agree with this conclusion and have only to add that it is impossible in the case of the Naples QUESTIONS AND ANSWERS accident to be dogmatic that defects of the kind My answers to the questions submitted on considered in paras. 108-144 of my Report on behalf of the Attorney-General are as follows:- Yoke Peter were not contributory causes to the Naples accident. I am therefore glad to note Question 1. What was the cause of the acci­ that the programme of future action putlined by dent? the de Havilland Aircraft Company Limited and set forth in Appendix VIII to my Report on Answer. Owing to the impossibility of Yoke Peter includes measures to deal with those salvaging any appreciable part of defects. the wreckage of the aircraft no positive answer can be given to this question but the fact that this accident occurred in similar PART X weather conditions, at approxi­ mately the same height and after RESPONSIBILITY approximately the same lapse of 30. I have dealt at length with this question time after take-off from Rome as in my Report on the accident to Yoke Peter. that to G-ALYP makes it at least There is, however, one matter on which criticism possible that the cause was the was made which is applicable only to Yoke Yoke same as in that case. The state of and that is the decision, after the accident to the bodies recovered was, as in Yoke Peter, to allow the Comet passenger ser­ the case of G-ALyp, consistent vices to be resumed on the 23rd March, 1954. with the accident being due to I have set out in paras. 52 and 53 of my Report failure of the cabin structure owing on the accident to Yoke Peter the nature of to metal fatigue. the full investigation carried out by the Com­ mittee under the chairmanship of Mr. Abell, Question 2. If several factors caused the the Deputy Operations Director (Erigineering) accident what were such factors of B.O.A.C. and the modifications made on the and to what extent was each con­ recommendation of that Committee. tributory? Answer. I cannot usefully add anything 31. Before deciding to authorise the resump­ to my answer to Question 1. tion of the Comet passenger services the Minister of Transport and Civil Aviation con­ Question 3. Was the accident due to the act sulted A.R.B. and A.S.B. Both of these bodies or default or negligence of any recommended that consent should be given. party or of any person in the em­ When they did so, there had been only one ployment of that party? accident to a. Comet aircraft for which no explanation had been furnished. According to Answer. There was no evidence on which the evidence it was certainly not the practice I could ·attribute the accident to either in the United Kingdom or elsewhere to the wrongful' act or default or ground all aircraft of a type because of an un­ negligence of any party or of any explained accident to one aircraft of that type. person in the employment of any The evidence indicated that steps had been taken party. to deal with what the experts then considered to be all potentially dangerous features. In these Question 4. At the tinle of the accident: circumstances I am of the opinion that no blame can be attached to anyone for permitting the Question 4 (a). Had the aircraft been properly resumption of the services. . maintained in accordance with the current approved maintenance schedules? If not did any defect in maintenance affect the safety' of PART XI the aircraft or contribute to the FUTURE accident? 32. I cannot usefully add anything to what Answer. ~he .aircraft had been properly I have said on this branch of the Inquiry in mamtamed save that on arrival my Report on the accident to Yoke Peter. at Rome a' number of bolts were 47 found lying in the port wing of Question 4 (e). Was the radio station of the the aircraft and further inspection aircraft serviceable and was there revealed that an equal number of a. valid Certificate of Serviceability bolts were missing from the in­ in respect thereof? " spection panel providing access be­ tween the rear spar and the wheel­ Answer. Yes. well wall and that the remainder of the bolts securing the panel though in position were not Question 4 (t>. Was the aircraft properly loaded properly tightened; The missing and trimmed within the limits bolts were replaced and all were specified in the Fight Manual? properly tightened and I am satis­ fied that this defect in mainten­ Answer. Yes. ance did not affect the safety of the aircraft or contribute to the accident. Question 4 (g). Were all members of the crew properly licensed and adequately Question 4 (b). Was the aircraft airworthy so experienced to make the flight? If far as could reasonably have been not did any defect in the licence then ascertained? of any member of the crew affect the safety of the aircraft or con­ Answer. Yes. tribute to the accident? Question 4 (c). Was there a valid Certificate of Answer. Yes. The second part of the Airworthiness in respect of the air­ question does not arise. craft? Answer. Semble yes. I do not find it necessary to deal with the legal Question 5. Was the Minister of Transport question wpether the default in re­ and Civil Aviation properly ad­ assembly referred to in paras. 21 vised in March. "1954 that Comet and 22 of this Report had any services should be resumed? effect on the validity of the Cer­ tificate of Airworthiness since I Answer. Yes. See paragraph 31 of this am satisfied that this default did Report. not· contribute to the accident.

Question 4 (d). Was there a valid Certificate of Question 6. Upon consideration of all facts Maintenance in respect of the air­ disclosed by this Inquiry what craft? steps should be taken to increase the safety of civil aircraft? Answer. Semble yes. See my answer to Question 4 (c) on Certificate of Answer. See paragraphs 140-155 of my Airworthiness. Report on Yoke Peter.

(Signod) COIff!N. W. S. FARREN. W. J; DUNCAN." A. H. WHEELER. 1st February, 1955.

48 (81687) Wt. 4517- K12 2/55 D.L. Crown copyright reserved

Printed and published by HER MAJESTY'S STATIONERY OFFICE To be purchased from York House, Kingsway, London w.c.2 423 Oxford Street, London w.l P.O. Box 569, London s.E.l 13A Castle Street, Edinburgh 2 109 St. Mary Street, Cardiff 39 King Street, Manchester 2 Tower Lane, Bristol 1 2 Edmund Street, Birmingham 3 80 Chichester Street, Belfast or through any bookseller Price Ss. Od. net

Printed in Great Britain

s.o. Code No. 56-14*