Some tricky photography
One of the testing requirements we had to meet was to produce photographs of Armour Piercing Discarding Sabot (APDS) projectiles from a new anti-tank gun, recorded within a few feet of leaving the muzzle. This was necessary to check that the discarding sabots (a three segment outer casing designed to carry the projectile up the gun barrel) really were breaking away cleanly from the heavy projectile once their task was done.

We owned a super high speed "Framing" (movie) camera made by Vinten Ltd., of the U.K. It was a masterpiece of engineering and could feed a 400 ft reel of 35mm film through at seven hundred frames per second with the film stopping in the gate for each exposure. The only snag was that as Leslie Barnes wryly pointed out, it needed three Ph.D's kneeling over it in silent prayer to get it to work. The mechanism had such fine tolerances that the slightest hiccup in film feed would cause instant devastation and on one occasion the result was that the entire mechanism became welded solid with smouldering film sprouting out of it in all directions.

We therefore decided to try some other tack and the first attempt was made with a scheme which used flash photography; a "dark-box" made of welded armour plate was constructed in which a ruggedised camera was installed, firmly anchored to the bottom plate. The box had large holes at each end, covered with blackout material, for the passage of the projectile, and a heating element on the top side, illuminating an infra-red sensor on the bottom. The idea was that a micro-flash unit would be triggered by the passage of the projectile between the heating element and the infra-red sensor, which would trigger the flash and freeze an image of the projectile and its discarding sabots on the film. We had not reckoned with the incredibly destructive power of the blast from the projectiles, travelling at more than three thousand miles an hour. The vacuum following the blast from the very first shot caused the bottom plate of the box to "ping" upwards like a piece of tin, catapulting the camera up against the top plate and destroying it completely.

There was another type of moving film camera made by Kodak which they called the "Fastax". This simply allowed the film to whiz through without any attempt at stopping it for individual frames. A rotating prism was used to move the image to match the film speed for each frame. A modified version of one of these things was being used at CARDE at Valcartier on some experimental project or other and I and some others from Nicolet went to see a demonstration. The prism had been removed and replaced by a narrow slit, so that there were no individual frames.

The idea was that as the projectile hurtled past, the tiny image formed by the camera lens would move at the same speed as the film for the brief instant that the projectile was in the field of view. The image was literally smeared onto the film via the narrow slit, so that there was no fuzziness in the final result. The technique was not new and was called "Smear Photography". They loaded a 400 ft reel of 35mm film into the camera and let it rip. It took about two seconds to get up to maximum speed at which point the gun was fired. It then took about half an hour to process the film (which required an automated processing machine) and another half an hour to comb through all four hundred feet of it to find the single image which was the picture of the projectile. We were not impressed, it may have been alright for a single experiment, but it was clearly not on for a routine operation.

I realised that it would be a viable technique if only there was not so much film to deal with. It was then that I remembered the rotating drum camera on the old Southern Instruments recording system. It was designed to spin at speeds of 3000 rpm, or 50 revolutions per second. If the lens which it had could be replaced, for example by a good standard press-camera lens, having an accurate shutter setting to give a fiftieth of a second exposure, then the drum with the one foot or so of film which it took, could be brought up to speed and held there spinning in the dark until the gun was fired. If then the shutter could be synchronised to open as the gun was fired, the image of the projectile would be recorded somewhere on the one foot of film, which took just a fiftieth of a second to make one revolution past the lens.

I went back to Ottawa and started to get down to the nitty gritty details of how this scheme might be made to work. I knew that there were some physics and electronics problems to solve if it was ever going to be a practical proposition. The first was to make sure that the speed of the film on the rotating drum could be made to match the speed of the image of the projectile formed by the lens, and to come up with a design for the narrow slit (for example how narrow?). It was a while since I had done any calculations involving optics, but I revisited some of the University text books and lecture notes which I had brought with me from the U.K. and was able to make some reasonably accurate estimates. From these it looked as if the camera could indeed be placed at a distance from the trajectory of the projectile such that a good sharp image could be obtained, and that a good compromise could be worked out between film speed and width of the slit (The narrower the slit the better the image, but the shorter the effective exposure time).

The next step was to find a suitable lens, which was not too difficult. The tricky part was to come up with a method of opening the shutter at the right time. It turned out that Graflex (the lens manufacturer and a big name in press photography equipment at that time) offered the option of a solenoid to allow the shutter to be actuated by an electrical switch. I worked up a preliminary design for modifying the drum camera and had the necessary machining and so on done down at the Range to accommodate the lens. The thing that I needed to know then was the delay time between energising the solenoid and the shutter actually opening.

When I got word that the modified camera was ready I arranged to spend several days down at the Range to conduct some experiments with it. By that time I had enough experience to know that Murphy's law; "if something can go wrong - it will go wrong", applies particularly strongly when conducting any sort of experimental work. I also knew that it might need quite a few shots to fine-tune the thing and get all the parameters under control. Artillery rounds were very expensive and needed a full complement of proof officers and gun crew, plus clearance to fire across the lake and so on, so I decided to make the initial trials using a 7.62mm rifle, which I could fire all day long for a fraction of the price of one artillery shell and with virtually no disruption of other ongoing operations. There was a risk however in that the image on the film of a 7.62 mm bullet was going to be pretty small and hard to see under the best of circumstances.

I used an electronic counter, a light and a photo tube to measure the solenoid delay time. A switch started the counter and energised the solenoid simultaneously, when the shutter opened the photo tube was illuminated by the light and sent a signal to stop the counter. The time interval recorded by the counter was thus the time it took for the solenoid to open the shutter. Having got that number I calculated how far down the trajectory the camera would have to be so that the shutter would have time to open before the bullet was in the field of view. I used one of the photoelectric "sky screens" (normally used for muzzle velocity measurements) set down close to the muzzle of the rifle to detect the passage of the bullet and thereby provide the signal to energise the solenoid.

I had bought several rolls of standard 20 exposure black and white 35mm film with a 400 ASA film speed with me and finally after a couple of days of preparation it was time to try the first shot. I loaded the film onto the drum and let it get up to speed before giving the signal to fire. I retrieved the film with trembling hands, knowing that after all the work that had gone into this idea, I had a lot of personal credibility riding on it. We took the film into the dark room and developed it, then dried it off and stretched the roll out on a light table. The film was of course a negative version, so that what we were looking for was a white blip somewhere in a strip of black film. At first nobody saw anything and then I saw a white blip. Was that it, or was it wishful thinking or a flaw in the emulsion? The Range Photographer who was on hand, made a positive enlargement, there was no doubt about it then, the characteristic pear shape of the bullet was unmistakeable. Everyone was impressed, nobody more than me, I was just amazed to have got all the numbers right on the first shot.

More shots were taken to confirm the operation and to tweak up a couple of parameters, all of which were recorded satisfactorily. It was a considerable relief to me, because there had been some scepticism that such a relatively simple idea, using fairly simple gadgetry would ever work at all - let alone as well as the sophisticated and expensive Kodak Fastex camera. With that assurance it was then time to try the real thing, the anti-tank gun. A test firing was arranged for the following week. I had other more urgent fish to fry which meant that I would have to go back to Ottawa and hope for the best. I briefed the technicians and Proof Officers who would do the test on the things that were critical and left them to it, after making them swear to phone me with the results immediately afterwards on pain of death.

The test was indeed successful and the prints were sent up to me in Ottawa. The pictures showed exactly what was needed and plans were immediately put in hand to purchase and modify a second camera from Southern Instruments in the U.K. A minor snag was that the drum rotation was in a direction such that the camera had to be placed on the side of the trajectory facing the sun, which meant that the pictures were essentially silhouettes. The problem was solved by one of the technicians at the Range who found a way to make the drum spin in the other direction. This allowed the camera to be placed on the other side of the trajectory, thereby having the sun highlight the details of the projectiles and the discarding sabots.

This improvement produced startlingly detailed pictures of remarkable clarity. The people at the Range went one step further once the second camera was operational and set them up as a stereo pair to photograph a single projectile. The three dimensional images seen with the aid of a stereoscope revealed even more of what was happening to the projectiles and their sabots as they parted company. This home- made "smear" camera, was used extensively later on in another project with which we were to become heavily involved, the McGill University High Altitude Research Project "HARP", of which more later.

A better mouse trap
I did quite a bit of work on the pressure measurement problem which confirmed that the pressures in the FN-7.62mm rifle were indeed much higher than had been generally recognised. It also revealed that there was quite a large variation from one production lot of ammunition to another, even though the muzzle velocities were all well within specifications. There was sufficient concern at one point that the rifle was temporarily withdrawn from service to make some design modifications. This prompted a question in the House of Commons from a young MP from the riding of Hamilton and the Islands, demanding to know for how long the Canadian Army was going to be defenceless (or something like that). The MP in question was Judy LaMarsh, who was later to become health minister in the Liberal government of Lester Pearson, which ended the Diefenbaker years. She died of cancer at a tragically early age, but left her mark on politics with her book "Bird in a Gilded Cage", an account of the obstacles women faced in making a career in politics.

A 7.62 mm FN rifle with a solenoid operated trigger and a
microswitch (top), actuated by the firing pin mechansim,
which triggered an oscilloscope to record the pressure-time trace
The mechanics of the pressure measurement using the miniature piezo-electric pressure transducers were rather tedious. for each shot the pressure time curve had to be recorded by taking a polaroid print of the trace generated on the screen of an oscilloscope, using a special polaroid camera designed to fit onto the scope. The peak height then had to be measured against a calibration trace superimposed on it to arrive at the peak pressure. I realised that if the technique was ever going to be adopted for routine use, some sort of instant result would be needed. The transducer was also not very suitable because it had a tiny piston to transmit the gas pressure to the quartz crystal which generated the electrical signal. The piston would distort under the pressure and would jam unless it was polished and maintained in tip top condition. I had long discussions with the engineers at the Kistler Instrument Corporation and tried to convince them to produce a version with a sealed diaphragm rather than the piston arrangement.

I ended up going down to N.Tonawanda New York to see their plant and have some round table discussions with them. They were very friendly and even insisted that I borrow the company car to drive to my hotel which was some way from the plant. It was a Ford Thunderbird, very big and very powerful. I drove gingerly through the town and the first time I braked for a traffic light I nearly pitched myself through the windscreen, the thing had power brakes which I had never encountered before. The visit was worthwhile because I succeeded in persuading them that it would be worth their while at least to try the diaphragm approach. I also saw what they were up to in their research and development lab.

They were evaluating something called a field-effect transistor or "FET", under development by the Fairchild Camera and Instrument Corporation. This was the solid state analogue of a vacuum tube in that it did not draw any current from the signal that it was amplifying, which all other transistors did. These devices were incredibly fragile and the slightest electrostatic charge could destroy them (and frequently did). Today the FET which evolved from that early experimental version is a rugged almost indestructible component which will operate at hundreds of volts and pass tens of amperes of current. One of the main uses is to drive the loudspeakers of the too-powerful sound systems which regularly pollute the urban environment.

The Kistler people did eventually produce a sealed diaphragm gauge which was virtually trouble free, thereby bringing the technique one step closer to a routine operation. By that time I had bought some very compact transistorised electronic counters with "Nixie Tube" digital readouts to replace the ancient and outmoded ones that were still being used for projectile velocity measurement. The new ones were light years ahead in terms of ease of use and portability and I thought how neat it would be to have something like that to give an instant digital readout of the peak pressure in pounds per square inch (psi). Such a development would make the routine use of electronic pressure measurement a viable and realistic alternative to the copper crusher gauge, because it could then be used by people other than electronics technicians. As things stood, the only way to get the required peak pressure was to go through the business of photographing an oscilloscope screen and making measurements on the pressure-time trace that was produced.

As is so often the case in the research and development business, once having seen the possibilities of something like that, it was just impossible to let go of it. I had little or no idea how such a thing might be done, something called an Analogue-to-Digital Converter was needed, such things are now in every CD player, but at that time the development of digital technology was just beginning. I knew that those counters were the key to any practical realisation of this idea, they were there and waiting and ideal for the purpose. Furthermore, every other proving ground involved in similar work would be certain to have electronic counters of some sort.

I had a lot of sleepless nights trying to come up with some way of converting a voltage amplitude into a proportional time interval, which was essentially the problem involved in trying to use the electronic time interval counters for the readout of a peak pressure. Eventually I found a reasonable way to do it which was to allow a condenser to charge up to the peak amplitude of the signal generated by the pressure gauge via a diode, so that it would retain the charge after the pressure dropped. The condenser was then discharged at a constant current by a pentode valve, the time for the discharge was exactly proportional to the peak pressure.

In fact this technique of analogue-to-digital conversion was not original as I later discovered. It was the basis of digitising nuclear radiation spectra, something that was just beginning to be done in some nuclear research laboratories, not surprisingly I was not aware of developments in that field. Once I had solved the basic riddle of how to convert the peak pressure to a time interval, it was then a question of coming up with a practical design. There were the usual peripheral problems, one of which was how to ensure that the pressure gauge could be calibrated using the readout device. Calibration was done by installing the gauge in a hydraulic pressure tester, which generated precise pressures using weights on a piston. The piston applied pressure to oil in a chamber which was transmitted to everything connected to the chamber, including of course the gauge. The gauge generated a negative signal when the pressure was suddenly released through a quick-release valve. My device had to accommodate that and produce a reading on the counter that was numerically equal to the pressure applied by the tester in psi. I found a way to build in a control that could be adjusted by a knob so that the exact reading could be arrived at by a few iterations of trial and error.

Eventually I had a working version which was demonstrated to bigwigs who came to visit from other NATO countries and the reports which I wrote describing the technique evoked quite a lot of interest, particularly in France, the U.K. and Belgium, where similar work was underway to replace the copper crusher gauges with something more reliable. Meanwhile another project began to take a large part of my time.

The McGill University Harp Project (HARP)
During 1962 We were approached by a young scientist who had worked on the hyper-velocity gas gun project at CARDE, but who had left and gone to McGill University in Montreal, to pursue an ambitious and risky project that he had conceived. his name was Dr. Gerald Victor Bull, the man who many years later achieved international fame and notoriety as the brains behind the "Super- Gun" and also as an unscrupulous international arms dealer, meeting an untimely end from an assassin's bullet. A sad and sour end to the career of one of Canada's most brilliant and most ignored scientists.

Gerry Bull was basically an aerodynamicist who had done some brilliant work in his field, and was one of the youngest students ever to receive a Ph.D., from the University of Toronto, which he did in 1951. His idea was to perfect a cheap alternative for the ruinously expensive sounding rockets being used by the U.S. space agency NASA, for collecting meteorological and other data in aid of the Mercury and Gemini orbital manned spaced flights. These rockets required powerful thrusters to get the instrument packages up into the stratosphere and sophisticated guidance systems to keep them on course. Most of the expense was incurred in the first portion of the journey, because of the weight of the fuel that was needed, which dwarfed the weight of the instruments that were the raison d'etre of the mission. Bull's idea was to use a conventional artillery gun to boost the package for that first portion of the journey, literally firing it up into the stratosphere, where a much smaller and cheaper rocket motor would carry it for the remainder of the mission.

A smear camera photo of a 'mini-Martlet'
in flight showing sabots separating
By the time we became involved Bull and his team had already designed the projectile, which was to be fired from a sixteen-inch World-War-one naval gun. A suitable gun had already been found and the rifling had been bored out, bringing the diameter up to eighteen inches, making it a very big gun indeed. The projectile, which they dubbed the "Martlet", was about four feet long with fins at the rear and designed to have the lowest possible "drag" coefficient (i.e. it was as streamlined as possible). The same principle of discarding sabots was to be used that we were familiar with from the anti-tank gun tests. The sabots in this case were made of plywood and encased the Martlet from top to bottom in a cylindrical package that fitted tightly into the gun barrel. A "pusher plate", made from a sandwich of metal and nylon, was used to separate the package from the heat of the burning gun propellant. There was no question of conducting full- scale firings until the details had been verified using small scale models in a much smaller gun and that is where we came in. The Range with its specialised instrumentation and other facilities was an ideal place to conduct these preliminary tests. The capability for gun pressure-time recording and smear-camera photography of the projectiles in flight, were of particular interest to Bull and his team.

The HARP project had had a cool reception from Official Ottawa, which was absolutely outrageous because it was a highly innovative and original concept - and one hundred percent Canadian, an all too rare occurrence, then and now. It was a sad commentary on Canadian politics and politicians of the time and not much has changed in the thirty years since then. As in most western nations the government at any given time usually consists of a collection of mostly lawyers with a cross section of car dealers, fast-food franchisers, farmers and furniture salesmen thrown in for good measure. None of them have the slightest clue about science and engineering and consequently are quite incapable of making judgements about which projects are important and which are not. They are more aware now than they were then of the pivotal role that engineering, science and technology now plays in the fate of nations, but there are still no plans afoot to provide the necessary incentives that would make it attractive for more people to run for public office with the necessary background to make informed judgements on these matters.

Despite the lukewarm reception that Bull and his project got elsewhere in Ottawa, Leslie Barnes managed to enlist the official support of Inspection Services and from that point on we had a lot to do with the initial stages in the development of the project. The first order of business was to find a smaller gun, that was easily solved because we had an old six-inch naval gun mounted on a gun- carriage designed to run on standard railway tracks. It was basically a demonstration piece, when it was fired there was a window- shattering boom, accompanied by a brilliant flash from the naval "WM" cordite.

smear camera photo shows
buckling of a dummy martlet
The recoil then propelled the whole thing backwards along a short length of railway track which took it up a steep mound, where it came to rest and then trundled down again into a set of buffers which finally stopped it. It was almost tailor-made for the purpose, the only modification required was removal of the rifling in the barrel. when this was done it became an ideal test bed for firing "Mini-Martlets" designed to fit it, which were scaled down in every detail from the full sized version. A number of test firings were made with the smear cameras in place and the photographs showed some structural weaknesses that had not been anticipated. Some "stove-pipe" dummy projectiles were made up to pin down the strength that was needed to prevent a Martlet from crumpling, concertina-fashion, under the unimaginable acceleration that it would receive in the barrel of the gun. This was typically ten thousand 'g' loads, or a force equal to ten thousand times its own weight.

Eventually this and other problems got sorted out with the aid of the small scale test firings and it was time for the big event, the full scale test firings of real Martlets from the mammoth eighteen inch gun. These experiments were done in Barbados, which at that time was part of the British West Indies. Because I had been closely involved in the small scale tests, Gerry Bull asked if I could be seconded to the HARP project team in Barbados for a week or two to help them with various instrumental measurements that they wanted to do. One of the parameters which was needed was accurate velocity measurements, but the gun was going to be fired almost vertically, so that there was no question of laying out "skyscreens" along the trajectory.

We had a radar-type apparatus made by Marconi in the U.K. It worked on the same principle as the Police radar speed measurement devices which are commonplace today, but at that time microwave technology was implemented with Klystrons and vacuum tubes and the equipment occupied two large boxes. It was built to military specifications ("Mil-Spec") and that included "Tropicalisation", just as well, as Barbados is only ten degrees or so north of the equator. As a backup I took some stick-on strain gauges with me, which I got from some group in the National Research Council Laboratories. My idea was to stick these on at each end of the barrel and use an oscilloscope to record the signals as the pressure wave driving the projectile passed each set of strain gauges in turn, causing a measurable strain in the metal on the outside. The time interval between the signals from the two sets of strain gauges would then enable at least the velocity in the barrel to be computed.

The HARP project paid for all the expenses involved and I flew down there in June 1963. It was my first experience of a tropical climate and I found out that everything that I had ever heard about it was all true. I had brought my little clockwork Yashika 8mm movie camera to record anything and everything of interest. It worked on the old standard, using a regular 16mm roll of film which was exposed first on one side and then on the other. This meant having to open the camera in the dark at some point and turn it over to use the second side, which was absolutely infuriating when it needed to be done at some critical time when you were shooting something in broad daylight, miles from anywhere evenly remotely dark. It was a Kodak standard for "home movies" and was a truly dreadful compromise because less than half the available film area for each frame was used. Some years later the so-called "super 8mm" was introduced with a film made for the job. The big selling point was "revolutionary new film cassette allows daylight loading", which made it within shouting distance of being as good as the French 9.5mm Pathescope cine film camera which my Father had had since the mid 1920's. It had always had a daylight loading cassette and the images were razor-sharp, because about ninety five percent of the available film was used.

The huge WWI naval gun, bored out to 18"
It was the beginning of the rainy season which took the form of frequent torrential showers, they began with teacup sized rain drops plopping down here and there and within about fifteen seconds became a veritable celestial waterfall. I discovered land crabs that rustled through the hedgerows, and assorted other tropical creepy-crawlies. The project HQ building was a beautiful British colonial mansion complete with rose coloured exterior and spacious rooms with high ceilings and tall windows. It was called "Paragon House", not the most propitious choice of names as things turned out. There had been some test firings using less than a full charge of propellant and stove-pipe dummy projectiles, to try and get a feel for how everything would work. This was the first time however that a real Martlet was to be fired with a full charge of propellant.

Loading a Martlet with its wooden
sabots into the breech of the big gun
The preparations for this important shot took the better part of a week. One of the chores was the business of cleaning the inside of the huge gun barrel, the local Barbadians who were hired to do this were for ever cheerful and their melodious voices were amplified and distorted by the giant organ-pipe effect created by the barrel, resulting in some of the weirdest sounds that I have ever heard. As time went by I became a little uneasy at the rather ad hoc and relaxed approach to the management of what was after all a pretty complex undertaking, with a lot of credibility riding on the outcome. For example a contingent of people were sent out in small boats to the area where the Martlet was expected to splash down, after attaining what was hoped would be a record-breaking altitude of more than four hundred thousand feet.

The gun elevated, ready for the shot
The idea was that they would hear the impact and note the time on their synchronised watches. This would pin down the total time of flight, allowing an estimate of the maximum altitude to be calculated. I was a bit surprised that people would be sent out miles offshore in relatively small craft, in barracuda infested waters, to hang around in the target area and wait for the thing to fall somewhere in their midst. When I mentioned this to Gerry his response was that the chances of anyone actually being hit were so remote in statistical terms that for all practical purposes there was no measurable risk. I knew that mathematically speaking he was right, but it did not exactly inspire confidence in the project management philosophy just the same.

The "Martlet" projectile
I had set up my radar transmitter/receiver unit looking skyward to bounce the beam off the base of the receding projectile, with the associated oscilloscope and polaroid camera, (yes, another one, it was the only way in those days to capture many transient events of that sort), being set up in the block house in which the control room was located, about fifty feet from the gun. The Martlet in its cylindrica l sabot of segmented plywood was designed to be a very tight fit in the barrel and it had to be forced in via the breech using a special jack. The gun was then slowly elevated to 89 degrees, one degree short of vertical. Gerry Bull was becoming more agitated by the minute as the count down progressed. I suggested that the blast from a gun of that size with a full charge of propellant would be considerable and that it might be prudent to open the windows of the block house to prevent them from being shattered. He agreed rather offhandedly as he paced the floor and then turned around quite suddenly and asked if he could borrow my movie camera to get some footage of the actual firing.

The tension mounted as the time ticked down to zero and the adrenalin was running very high indeed. My finger was practically numb from maintaining just enough pressure on the lens shutter to be sure that it was ready to open at "Tee-minus- three" to record the circular sweep of the trace as the radar beam sent back its vital signal. The moment came, I opened the shutter and three seconds later there was a mammoth explosion and blast wave. I saw a blur on the oscilloscope instead of the usual clear trace and knew that I had not got a reliable velocity measurement, "..probably too much partly burnt propellant.." I muttered to myself, knowing that that would cause enough ionisation to upset the radar return signal. I went in to the control room to find a somewhat shaken Gerry Bull wearing the easel of a blackboard which had been blown over by the blast wave. My movie camera was on the floor with one of the three lenses in the rotating turret knocked out. He was most apologetic and immediately offered to get me a replacement camera. In fact the damage turned out to be minor and was easily dealt with.

Gerry Bull with his brainchild
The team assembled to await word from the beleaguered band of brave souls cruising around the target area in their cockle-shell craft, that the Martlet had re- entered. Nothing happened. Gerry Bull was obviously mystified, he and the other aerodynamicists on the team had gone over the calculations again and again, the drag coefficient of the Martlet was well established, the gun elevation had been checked several times, where the hell had the thing gone, not into orbit surely? Even the most optimistic estimate of the muzzle velocity could not possibly have taken it that far up. Eventually it became clear that it had not fallen in the intended target area, for whatever reason. It was about an hour later that there was a phone call, not to report a sighting from the observers stationed out in the ocean, but from much nearer home, the Barbados Seaview airport, to inform us that a long heavy object had fallen into a sugar cane field and would someone from the project go and take a look.

About half an hour later the remains of the Martlet were brought back to the HQ and laid out on a table like a cadaver. The facts were clear enough, it had been shot out of the gun into the deep blue yonder, but instead of landing miles out in the ocean, it had wound up in a sugar cane field half a mile behind the gun, a major embarrassment for a shot that was supposed to have set a world altitude record for a free flying ballistic projectile. Absolutely nothing had worked as it should have, the Martlet had a fuse in it which was supposed to ignite and activate something or other after it was clear of the gun and it too had not worked. Gerry Bull was furious that so much had gone wrong and I remember standing with one foot outside the door as he vented his frustration by attacking the fuse mechanism with a penknife of all things, in a vain attempt to find out why it hadn't gone off.

The reason for the disaster was symptomatic of what ailed the whole project - lack of professional management. A short investigation showed that the Martlet had been loaded without the pusher plate. the vital component that had been designed to insulate it from the heat of the propellant and form a rigid cylinder with the discarding sabots to cushion it from local shock waves. As a result the rear fins had been sheared off and the projectile had simply tumbled end over end like a piece of pipe. Elementary oversights like this were of course ruinously expensive and would never have happened if there had been more attention to detail in the way the operation was conducted. There was much soul-searching about what to do next and all further test were put on hold pending a much overdue review of checks and balances.

I flew back on a Bristol Britannia turbo-prop, island- hopping across the Caribbean. I remember coming in to land at Antigua rather vividly; it was bad weather with rain and high winds and I had a nasty feeling that we were coming in much higher than we should have been on the final approach. How right I was, we touched down much too far along the runway and the pilot slammed on the wheel brakes as soon as he could. We were thrown forward in our seats as the plane shuddered under the deceleration and finally came to rest with less than a hundred yards to the edge of the cliff overlooking the sea. Everyone looked pretty ashen-faced, but as far as I was concerned this was just one more cliff-hanger, (a literal one in this case) to top off my experiences over the previous two weeks.