THE above is an extract from the text of an article on the supply of arms and ammunition to the Mujahideen written by a journalist who had spent some weeks in Pakistan trying to unravel the complexities of a system that stretched half way round the world, involved six countries other than Pakistan and Afghanistan, and was costing, by 1987, over a million dollars a day. Perhaps, not surprisingly, Mr Jack Anderson's comments were largely guesswork, but in this case not too far from the mark. As I have explained in the previous chapter the CIA has much to answer for with its wasteful purchasing system. Nevertheless, it was not the CIA's pipeline that provided weapons to the Mujahideen. As soon as the arms arrived in Pakistan the CIA's responsibility ended From then on it was our pipeline, our organization, that moved, allocated and distributed every bullet that the CIA procured. But even the ISI did not actually give the guns and ammunition to the Mujahideen who were to use them in battle. The last stretch of the line into and across Afghanistan was in the hands of the seven Parties and the Commanders in the field. To understand how the arms reached the battlefield from places as remote from Afghanistan as the USA or Britain it is necessary to know that the pipeline was divided into three distinct parts. The first part belonged to the CIA, who bought the weapons and paid for their delivery to Pakistan; the second stretch was the ISI's responsibility, getting everything carried across Pakistan. allocated to, and handed over to the Parties at their headquarter offices near Peshawar and Quetta; the third and final leg of the journey belonged to them. The Parties allocated the weapons to their Commanders, and distributed them inside Afghanistan.

When a Mujahid dropped a bomb down the barrel of his mortar it was the end of a journey that had involved being loaded, or off-loaded. at least fifteen times; being moved many thousands of kilometres by truck, ship, train, truck again, and pack animal, before being carried to the firing point by the mortar men themselves. Seldom, even in guerrilla operations, can a line of communication have been so tenuous and fraught with frustrations long before any item reached hostile territory. A British general once said something to the effect that, for every thought that a commander bestows upon his enemy, he probably directs a hundred anxious glances to his own supply line in his rear. I would endorse that comment.

As far as I was concerned my major headache was logistics getting sufficient supplies forward, in time, to the right people, and at the right place. Everything else was of secondary importance. My difficulties were compounded by the fact that I only had direct control over the centre section of the pipeline, both ends being in the hands of others (see Map 8). With the CIA and the Parties I could only plead, explain, cajole or persuade. I could not intervene directly when things went wrong and I could not use my own resources to put matters right. The task of my logistics colonel was certainly the most unenviable within my bureau, if not within the entire ISI organization. His was the daily grind of keeping supplies moving, of worrying about ship or aircraft arrivals, lack of manpower, late supply of railway wagons, insufficient vehicles, mechanical breakdowns, and above all security - preventing any leaks as to what we were doing getting to the public, or over-inquisitive foreign journalists and enemy agents. What he achieved was a minor miracle, as the system was never exposed, never disrupted by sabotage inside Pakistan, during the period 1984-87. In 1983 some 10,000 tons of arms and ammunition went through the pipeline. By 1987 this amount had risen to 65,000 tons, all of it handled by 200 men from the Ministry of Defense Constabulary (MODC) with four fork-lift trucks, working seven days a week, month after month.

At the CIA end of the pipe it was not just that we often received inappropriate and outdated weapons systems that we did not want, but their scheduling of shipments frequently took no account of our capacity to handle the huge quantities involved. It was usually a feast or a famine situation. I repeatedly stressed that we needed a smooth, regular flow of arrivals at Karachi port, of about one or two ships a month. This we could manage. Such a steady build-up would prevent bottlenecks when our stores were overflowing, or periods when they were virtually empty. Perhaps I was asking the impossible, because sometimes three, or even four, ships would arrive within a month, or there would be a long period when nothing came.

A small proportion of arms arrived by air at Rawalpindi (Chaklala Air Base). Until 1986 many of these deliveries were the cause of increasing friction between ISI and the Pakistan Air Force (PAF). The trouble seemed to stem from Saudi Arabia, where the CIA had arranged to dump supplies at Dhahran Air Base either for onward delivery by Saudi aircraft or collection by the PAF. For some reason unknown to me, these flights always seemed to go wrong. The CIA provided the liaison, but even when they had an agent at the Saudi airfield invariably our planes were not allowed to land on schedule, or were even turned away. When Saudi aircraft flew to Pakistan they did not arrive on time, or came completely unannounced, causing the PAF to go on unnecessary alerts. After nearly two years of this I managed to get the system stopped and USAF planes used, but our relationship with the PAF was badly soured.

Once the arms arrived on Pakistani soil we took over. I should explain that the supply system had been established before I arrived at ISI and that overall security of our methods was greatly enhanced by our operating under martial law. The military was in complete control. They were both the makers and executors of the law. For example, the normal bureaucratic stream of paperwork was suspended as far as we were concerned. Until later, when the size of the supply system expanded so enormously, nothing was committed to writing. Officials in government agencies or departments that had to be involved in keeping things moving were briefed verbally on what they had to do. If they became too curious they were told that the consignment or work was related to the widely publicized, but 'secret', project of making a nuclear bomb. This was normally enough to get cooperation.

At Karachi the port authorities were paid their dues in cash, the ship's manifest merely stated 'defense stores' and the customs department was not involved. From the ship the crates were loaded on to between ten and twenty freight wagons for the journey by rail to my warehouse at Ojhri camp or, for a small proportion, direct to Quetta. The wagons were accompanied by armed MODC escorts. This rail movement was a daily event. Ten wagons would take about 200 tons, although I could cope with up to 400 as a maximum. If several ships arrived in quick succession the system broke down, with a build-up of stores at the docks as my men toiled to shift them, while my colonel contended with the railway functionaries for more rolling stock.

At Rawalpindi we had a fleet of 200 vehicles, mostly five- or ten-ton trucks with false and frequently changed number plates, with which to move the weapons further down the pipeline. All the boxes had to be brought to the camp from the railway station to be separated, checked and stored in the warehouse. Everything had to be taken on charge and stock lists updated daily. I insisted on having this information on my desk early every morning.

Next came the business of breaking bulk, and dividing up the weapons and ammunition into consignments for the Parties at Peshawar. This was done in accordance with the allocation priorities that I will describe later. Arms and ammunition is useless sitting at a depot, it needs to be in the hands of the user, so I made it a point of honour to keep the flow moving towards Afghanistan I preferred an almost empty warehouse to a full one - it was also a much less tempting target should a saboteur obtain knowledge of its location, or in the event of an accidental fire. We were, after all, sitting on a potentially very large bang, close to a built-up area. At least 80 per cent of all arms and ammunition used in Afghanistan passed through my warehouse at Ojhri, yet I believe our secret was kept. Despite the daily volume of traffic and activity to and from the camp, we never experienced any incident that suggested security had been breached during those four years.

Every morning, at times varying from 5.00 am to noon, a convoy of trucks with MODC drivers in civilian clothes would leave for Peshawar. It was a 150-kilometre drive, which had to be completed by evening so nothing could leave Rawalpindi later than noon. Every afternoon empty vehicles from the previous day's delivery arrived back. Our workshop staff were kept frantically busy on maintenance.

To call these vehicles convoys is perhaps misleading, in that the 50-60 trucks did not follow one another in a long column - far from it. We sent them off in small packets of two or three lorries at intervals of five to ten minutes. They merged with the civilian traffic, with the man riding shotgun in the cab having his weapon concealed on the floor. Once, when traveling from Peshawar with the local CIA chief, I challenged him to try to spot any of our vehicles. He failed to do so.

Our main worry was the possibility of road accidents. An officer always traveled with the leading truck and another at the end of the convoy, and we included one or two empty vehicles in case of breakdowns. General Akhtar refused to countenance the possibility of accidents, although he knew our difficulties and that the law of statistics made a few inevitable. He remained adamant that they were not to happen, so I was forced to increase the number of officers on convoy duties to the detriment of training and operational requirements.

In 1986 I calculated that my trucks travelled well over a million kilometres. With such distance accidents happened. The most unfortunate was when a truck hit a car head on. By the time the officer at the rear reached the scene the casualties had been evacuated, and despite enquiries at the nearby hospital he failed to locate the car passengers. It turned out that the car occupants were Army officers and that two had died at the military hospital. There was a lot of unpleasantness as the Army blamed the ISI, despite witnesses testifying that the car was at fault.

At Peshawar the Parties took over. Their consignments were off-loaded at their warehouses while the drivers overnighted before the return journey next day. This was the system I used for the great bulk of arms and ammunition. There were a few exceptions, apart from those going direct to Quetta from Karachi. These involved rocket launchers and SAMs, weapons that were scarce and were needed in particularly critical operational areas such as around Kabul, or airbases, or along the Salang Highway. Everybody wanted to shoot down helicopters or launch rocket attacks from 10 kilometres from the target as this added to a Commander's prestige, but I insisted they must be deployed in accordance with the overall guerrilla strategy. With these weapons I insisted I deliver them direct to the Commanders from those key areas, but in consultation with the Parties. Similarly, if there was a special operation being mounted, such as the sinking of barges on the Amu that required limpet mines, or a coordinated large-scale attack on a large enemy garrison, then I would also arrange a direct issue to the Commanders involved, again with the Parties' concurrence.

Despite the allegations of corruption leveled at those involved with the arms pipeline, I remain totally convinced that as far as my organization was concerned nothing much went astray. The middle section of the pipe was virtually corruption-free. The beginning section under the CIA was riddled with opportunities for fraud and, as I have shown, there was ineptitude and probably dishonesty as well.

I should explain that there were also charges that the ISI diverted arms to the Pakistan Army. These were correct to the extent that some 200 14.5mm machine guns, RPG-7s and SA-7s were given to the Army to be deployed in emergencies on the western border when the Soviet/Afghan forces stepped up their air and artillery violations into Pakistan. I can say with absolute authority that no other weapon was so diverted. It was foolish of us to do it without taking the CIA into our confidence, as I am sure they would not have objected. As it was, they found out, so there was a flurry of accusations and denials which damaged our relationship unnecessarily. Perhaps more detrimental, and increasingly so as time went on, to our association with the CIA was the endless bickering over their persistent demands for control over the allocation of all arms and ammunition entering Pakistan.

During General Akhtar's eight years as Director-General of ISI it was the policy, on which he rightly remained unmoveable, that ISI decided who got the weapons, how many, and what types. By this I mean that, after the formation of the Alliance, the detailed allocation to each Party was our responsibility. It bears repeating. No one outside the ISI, including President Zia, had any say or control over the allocation of arms, ammunition and allied logistic stores from our warehouses at Rawalpindi and Quetta. It was not only the CIA who criticized us over this matter. The US Ambassador was often outspokenly disapproving, as were the US Congress, foreign journalists, senior generals in the Pakistan Army and the Parties themselves. They all thought they knew best. They all had their own political and personal motives for knocking the system, so they took every opportunity of pressuring ISI to alter allocations. Parties and Commanders routinely clamoured for more, for bigger and better weapons, while the Americans insisted we favoured the Fundamentalist Parties, particularly Hekmatyar. It was a never-ending source of friction, in-fighting and frustration.

The US felt that as they were paying for at least half of all the arms they should have a say in who used them. As the war progressed, and especially when the Soviets started talking about withdrawing, US officials started becoming more and more concerned that the next government in Kabul might be an Islamic Fundamentalist one, possibly with Hekmatyar becoming another Khomeini. This fear was eventually to lead to a deliberate US policy of withholding support to prevent a Mujahideen victory, but during my tenure it manifested itself in mounting allegations of partiality over arms allocations.

My job was to apply military pressure inside Afghanistan to get the Soviets out. I was a professional soldier, with a soldier's ambition to win on the battlefield. With these as my motives I decided who got the means to win -- the weapons and ammunition -- on the basis of maximum combat effectiveness. I had to implement a campaign strategy to influence operations without the ability to issue orders to subordinates, without any military infrastructure to sustain or implement decisions. I had to coordinate attacks on strategic targets and maintain the initiative over an area of 260,000 square miles by exhortation, supported by animal pack trains and, for most of the time, a system of messengers that had not much changed since Alexander's days. Concentration and cooperation are two immutable principles of war. Success in battle is often dependent on both being applied simultaneously at the right time and place. The only way I could influence the Parties and Commanders, get them moving in the right direction, was through the allocation or withholding of supplies and training.

As I have emphasized before, weapons have always played an important part in an Afghan's life. The more modern the rifle that a man owns the higher his standing. For Mujahideen the possession of heavy weapons and plentiful ammunition was a common goal, for which they were willing to show some flexibility, some inclination to listen, or to follow instructions. My giving assurances that a certain operation would be backed up with extra weapons or more missiles, and that success would lead to further supplies, was sometimes the only way I could obtain cooperation. I had a carrot to offer. My stick was to withhold the weapons. Had the ISI not retained this prerogative my task would have been hopeless.

Eighty per cent of all arms and ammunition was allocated to the Parties for onward distribution. Commanders had to belong to a Party in order to get weapons, the only exception being when they came for training for special operations, but, even though they were then given the weapons direct, they came from their Parties' allocation. Our American allies favoured giving arms direct to Commanders. This had been the system before I took over, when the supply was a trickle, before the Quetta incident described previously, and prior to the formation of the Alliance. By the mid-eighties such a policy was unworkable. It was daunting enough trying to get results dealing with seven Parties; to attempt to do so by direct contact with hundreds of rival Commanders, each anxious to enhance his own reputation, was to invite chaos.

Every three months an operational conference would be held between General Akhtar, myself, and my officers of lieutenant-colonel or above. A key matter for discussion and decision was always the arms share and any modifications needed to existing arrangements. Because it was such a critical and controversial matter, I spent many hours before the conference going over the problems with my staff. I needed their opinions before making firm recommendations to the General. Frequently this subject would generate long debate at the conference and, although the final decision was Akhtar's, he seldom overruled our recommendations. For planning purposes we worked on a rough percentage basis for each Party. These were not permanently fixed; they varied slightly for operational reasons, and sometimes they were deliberately reduced if a Party was seen not to be pulling its weight in the field. Such reductions were normally gradual and followed a verbal warning to the Leader.

The criteria we used in drawing up these rule-of-thumb percentages were all related to battlefield competence. The numerical following of a Party as such was not a factor. For example, Khalis' Party was comparatively small but its combat effectiveness was greater than a large Party like Mujaddadi. The location of Commanders of each Party in Afghanistan was an important consideration. The majority would not fight outside their own area, even their own valley, so it was pointless pouring arms down the pipeline to a Commander far removed from strategic targets. Any Party strong around Kabul could rely on a higher percentage, likewise those operating against sensitive spots such as airfields, or main lines of communication. By using the word 'strong' I do not mean large numbers of Mujahideen at a given place, but the frequency of successful attacks in the area. To assess this I was indebted to the radio interception service which often provided me with confirmation of activities claimed by Commanders and Parties. Similarly, we used the CIA's satellite photographs to establish the validity of damage claims. I, and my officers, well understood the Mujahideen's inclination to exaggerate. Debriefing of individuals, the CIA and MI-6 weekly intelligence reports and the careful sifting of all information from various sources were important ways of verifying who was actually fighting and who was not.

Then we looked at the Parties' control over such dubious activities as the illegal sale of arms. I had a Major working fulltime on gathering this information. If a Party could not control its Commanders in this respect then their share would be cut. Nevertheless, I should qualify this condemnation of the sale of arms by the Mujahideen by saying that it is my belief there is probably no Commander in Afghanistan who has not, at some time, sold or bartered weapons. So long as it was done in Afghanistan between Mujahideen, for the Jehad, we never penalized them. Sometimes, in an emergency, it was the only way to obtain food, evacuate a casualty, or secure urgently needed ammunition. If the sale took place in Pakistan for the Commanders' personal enrichment or comfort, then we treated the offence as serious. Several Leaders were lax with their Commanders on this, and these tended to be the Moderates, part of the reason being that they were always short of funds. These Parties employed permanent staff, often Western-educated men who were not satisfied with the meager $100 a month salary paid by the Fundamentalists. They demanded, and got, three times this amount, plus free housing. There was an ever-present temptation to sell weapons they had been given at 100 per cent profit to make up cash shortfalls.

The final factor we considered was the general efficiency of the Party and their own logistic system, which I shall describe shortly. A sure way of judging a Party's competence was to visit their warehouse regularly. If my officers reported a warehouse was always full, sometimes for months, it meant that the Party was less than enthusiastic at prosecuting the war, and as such never qualified for an increased share of arms. Nabi's Party was a prime culprit in this respect. Despite having great potential, with some fine Commanders in the field, plus a numerous following, together with a former Afghan general as his military representative, Nabi and his officials never seemed able to improve their efficiency. In marked contrast was Sayaf, whose warehouses invariably held the minimum of stocks, although I must admit he had the singular advantage of receiving generous extra financial aid direct from rich Arab supporters.

In 1987 the broad percentages allocated to the Parties were Hekmatyar 18-20 per cent, Rabbani 18-19 per cent, Sayaf 17-18 per cent, Khalis 13-15 per cent, Nabi 13-15 per cent, Gailani 10-11 per cent, and Mujaddadi trailing with 3-5 per cent. Certainly the Fundamentalists came out on top with 67-73 per cent, much to the CIA's chagrin, but using strictly military criteria it could never be otherwise. My critics were taking into account political considerations and biases which, as a soldier, I was fortunately able to ignore.

I wish I had calculated the total cost of getting a weapon or bullet from the seller to the firer; it would have been a staggering statistic; shipment costs, rail and truck movement to Peshawar, followed by carriage over the border deep into Afghanistan, multiplied the purchase price a hundredfold. Probably the most expensive leg of the journey was the last sector of the pipeline from the Parties to the Mujahideen who would use the weapons. In some cases, where the supplies were going to Kabul or the eastern provinces, this was the shortest part of the journey, in which case the costs were more manageable, but charges to get arms to the crucial northern provinces were constantly rising, and by 1986 were little short of extortionate. By this time the going rate was $15-20 per kilogram. This meant the cost of moving a mortar from the Pakistan border to the Mazar-i-Sharif area was approximately $1100, while just one bomb cost around $65. Little wonder that the monthly expenditure by the Parties on transport and allied expenses was $1.5 million.

The CIA placed funds each month in the ISI-controlled bank account. This money had to pay for Party offices, construction and maintenance of warehouses, purchase of software (rations, clothes), subsistence allowance for Leaders, salaries for Party officials/employees, and transport. This latter included buying vehicles, and paying contractors to carry all supplies forward into Afghanistan, but not the purchase of mules from China (or later of horses from Argentina) which the CIA did themselves. Normally every Party had exhausted this source of money within 10-12 days. Without cash, supplies got stuck in the pipe, which meant in Party warehouses at Peshawar or Quetta. I recall how horrified I was when I first visited their warehouses in Peshawar, which at that stage were merely houses within the city. There were no proper storage or security arrangements as they were run in the most casual and unmilitary fashion. In one warehouse the 'storeman' was sitting on an upturned anti-tank mine cooking his meal over an open fire. Things did improve marginally and I managed to get funds to move all seven warehouses several kilometres outside the city, but there was little I could do to make up cash shortfalls.

Parties and Commanders did have other sources of finance. Until late 1984 local taxes were levied by Commanders in their valleys in Afghanistan, but as the Soviets progressively pounded the villages, smashed the irrigation systems, burnt crops and drove survivors into refugee camps, these taxes became impossible to collect. Captured weapons were used, sold or bartered. According to Islamic law war booty must be divided so that a fifth goes to the state (Party). I know Mujahideen sometimes found it cheaper to buy weapons or ammunition from Soviet or Afghan posts. I can vouch for this happening on a small scale on numerous occasions.

It was largely Arab money that saved the system. By this I mean cash from rich individuals or private organizations in the Arab world, not Saudi government funds. Without these extra millions the flow of arms actually getting to the Mujahideen would have been cut to a trickle. The problem was it all went to the four Fundamentalist Parties, not the Moderates. Sayaf, in particular, had many personal religious or academic contacts in Saudi Arabia, so his coffers were usually kept well filled. This mean. the Moderates became proportionately less efficient, lack of Arab money being one of the causes of their inability to match the Fundamentalists in operational effectiveness. Their income was less, their administrative and bureaucratic expenditure greater, thus making it harder for them to come up to our allocation criteria.

When my vehicle dumped the arms and ammunition at the Party warehouses responsibility for its distribution to the Mujahideen passed to the Parties (except for certain special types, or that earmarked for special operations). If some Commanders failed to receive their supply, or they felt their share was insufficient, there was little I could do about it. Each Party had its own method of deciding allocations to its Commanders. Sometimes it was on a fixed percentage basis - a hopelessly ineffective system which allowed Mujahideen in quiet areas to receive the same as those where fighting was frequent. At times supplies were sent to a single Provincial Commander for further distribution, on other occasions it was several Commanders who shared out between their sub-Commanders. Now and again all Commanders in a province would come to collect direct from the Party bases at the border.

How did the Parties move their supplies? It was one of the most complicated, chaotic and time-consuming operations of the war. Trucks and tractors, carts and camels, mules and horses all played their part, as did the backs of the Mujahideen themselves.

The larger Parties owned up to 300 vehicles of all types. These were civilian-pattern trucks which blended with the normal cross-border traffic. A number were Afghan vehicles purchased in Kabul, which were used for the longer journeys by road. They were more numerous than ISI's transport as often these vehicles undertook journeys of several days or more, with no possibility of returning empty on the second day. A truck could be absent from the Party pool for weeks. It was sometimes possible to drive all the way to the northern or western provinces, journeys of over 1000 kilometres, while on other occasions only pack animals could be used. Inside Afghanistan deals were often struck with local Afghan commanders for the use of Afghan Army transport. One of the peculiarities of the war was that on occasion the Mujahideen could have their arms delivered to them in their enemy's trucks. This occurred more often with sabotage operations in Kabul or other important cities, and included KHAD vehicles as well as military ones. At times, such activities were provided free, but normally money would have to change hands.

It was by lorry that the Parties moved their freight forward along the next stage of the pipe to the frontier. Here, some fifty-five border bases were located just inside Pakistan, mostly clustered around the main entry points near Parachinar and Chaman, NW of Quetta. To reach them the vehicles had to travel through the restricted areas of NWFP, Baluchistan, and the Tribal Areas (see Map 2). Throughout these regions the Pakistan Army, Border Scouts and Police were always on an active-service footing. Passage was controlled and subject to permits, check points, or vehicle search. To facilitate progress ISI issued all trucks with a 'let go vehicle' permit which gave all details of the lorry, except its cargo. C heck points en route were given lists of the trucks expected to pass through in advance. These vehicles were immune from search on the outward journey only, as a precaution against smuggling drugs or arms into Pakistan. Most of the time the system worked, but it was far from perfect. At times police check points would exact a 'fee' to avoid delays. Pay a bribe and the barrier was raised at once; refuse, and all sorts of excuses and telephone calls to non-existent or absent superiors could halt vehicles for hours.

I had an amusing personal experience of exactly this sort of difficulty when I took Congressman Wilson to the border prior to his secret visit to Afghanistan. I had sent an officer ahead to tell the police posts that our car was not to be delayed. At the first checkpoint a civilian official refused to let us proceed without 'clearance'. I showed him my Army identity card, but he insisted he would have to check back. His telephone call conveniently found his superiors out of office - more waiting. After 15 minutes I exploded. I told the official that if he did not raise the barrier I would get my three escorts, who were armed with AK-47 rifles, to empty their magazines into him. They cocked their weapons and the barrier was forcibly raised. Later, I told the officer who had gone ahead what had happened, and that I was not impressed with his 'smoothing' of our route. On the return trip the same officer was travelling in a second vehicle. At the checkpoint this officer ordered his men into the post, and had the unfortunate individual dragged out at gunpoint. Shrieking protests, he was bundled into the vehicle to be driven off. I am sure he thought his end was near. After driving about 15 kilometres, during which time the wretched man was weeping and apologizing pathetically, he was dumped at the roadside to fend for himself. He had picked the wrong person from which to try to extract a 'sweetener'.

Close to the border, especially around Parachinar, Miram Shah and Chaman, everybody was involved in the war in some way or other. There were tens of thousands of refugees in their camps, the bases teemed with Mujahideen, hundreds of transport contractors milled around with their animals, and scores of trucks were being loaded for their final journey to the end of the supply pipeline. Every day of every month, winter permitting, arms and ammunition were on the move. These areas contained the main jump-off points from the Mujahideen's base of supply. The Durand Line was to the Mujahideen what the Amu River was to the Soviets. Here Commanders came to collect their supplies, here the trucks from Peshawar and Quetta were off-loaded, and here the pack trains of animals assembled and loaded up.

In the early days a Commander would arrive in Pakistan with his own horses, perhaps a hundred, to collect his weapons, but as the quantity of stores multiplied and horses were lost, this system became totally inadequate. Thousands of animals were needed, and they became casualties like the men, so a reliable replacement organization was necessary. The answer was the contractor, although costs were high and climbed steadily every year. The contractor was a businessman, he owned the animals, he accompanied them into Afghanistan and he fed them. As it was his livelihood, he looked after his beasts, taking away this responsibility from the Mujahideen. Our CIA comrades did not like the system. They advocated specially formed animal transport companies operated by the Mujahideen. I did not agree, as the Mujahideen would not own the animals, would not have a personal financial interest in their care and half the animals would be needed to carry fodder. From experience I knew such companies would be as expensive, but less efficient, than the contractors.

The animals used were camels, horses and mules. The camels were usually employed on the long routes to the southern provinces where the land was arid. Horses were by far the most numerous pack animals. The Afghan pony was ideally suited to the task, having been reared in the country over the centuries for precisely this type of work. The horse tended to be the long-distance, strategic carrier of supplies from the border to the operational bases in the provinces. As more and more died or were killed, we resorted, in the later stages, to importing Argentinean horses. There were far less mules than horses. The mule was not bred in Afghanistan. There were some in Pakistan, and China had mule farms which provided an additional source of supply. These animals were usually to be seen as operational, or tactical, carriers. Normally it would be mules that packed the mortar, the heavy machine gun or the SBRL and its ammunition to the actual firing point, or very close to it. The mules, together with some horses, were given to Commanders to keep at their operational bases as what the military would term 'F' Echelon transport - transport that carried weapons on to or near the battlefield. The CIA would buy the animals, then give them to the Parties to issue to their Commanders. They were quite separate from the contractors' animals.

Apart from the branch from Karachi to Quetta there was really one main pipeline via Rawalpindi and Peshawar to the border. From then on numerous branch lines spread out into Afghanistan. I would liken our system to a tree. The roots represented the ships and aircraft bringing supplies from various countries to Pakistan. The trunk lay from Karachi almost to the border, at which point the main branches lay across the frontier. These branches divided into hundreds of smaller ones inside Afghanistan, taking the sap (arms and ammunition) to the leaves (the Mujahideen). Lop off a small branch, even a large one, and the tree survives, and in time others grow. Only severing the roots or trunk kills the tree. In our case only the branches were subject to attack. Unlike the Soviets, whose lines of communication were confined to main roads, ours made use of scores of tracks and trails through the mountains and valleys. If a road was blocked a route round could always be found.

There were six main routes leading into Afghanistan (see Map 9). Starting in the north, from Chitral a high route led to the Panjsher valley, Faizabad and the northern provinces. This was the shortest, cheapest and safest passage to these regions, but it was closed by the snow for up to eight months every year. We could only use it from June to October. Next came the busiest route. From Parachinar (the Parrot's Beak) via Ali Khel into Logar Province was the gateway to the Jehad, through which some 40 per cent of our supplies passed. This was the shortest route to Kabul, only a week's journey away. We also used it for journeys north over the mountains to the plains around Mazar-i-Sharif, although this could take a month or more. The disadvantage lay in the strong enemy opposition that tried to bar the way. When the Soviets wanted to decrease pressure on Kabul it was in the eastern provinces that they launched their largest search and destroy missions.

A little further south, the third route started around Miram Shah via Zhawar, again into Logar Province. Supply trains could either swing south near Gardez or Ghazni, or north to join the second route over the mountains. This was another busy route, but enemy interference was relatively light.

The fourth route started in Quetta, crossed the frontier at Chaman, before leading towards Kandahar and nearby provinces. There was much open country which meant vehicles were required to shift the bulk of the supplies quickly. We aimed to get trucks to their destination in one day's or night's fast driving. Suspicious vehicles were subjected to enemy ground or air attacks.

Over 400 kilometres further west, on the southern border of Helmund Province, was the smaller and unpopular base at Girzi-Jungle. It was used to replenish Helmund, Nimroz, Farah, and Herat Provinces. It was unpopular as vehicles were so vulnerable to attack. Seldom did we send in a convoy without incident. It was an arid, open area, sparsely populated, with little possibility of early warning of attack. Trucks travelling north were easily spotted from the air and were often shot up by gunship or ambushed by heliborne troops pre-positioned ahead of them. To reach Herat by vehicle took a week.

Finally, the sixth route was via Iran. A glance at Map 9 will show that to get supplies quickly and safely to Farah and Herat Provinces a long drive west along the Baluchistan border to Iran, then another 600 kilometres north from Zahedan in Iran to the Iran-Afghanistan frontier opposite Herat, a three-day journey, was the answer - in theory. In practice it was very different. Although we did use this route it took up to six months for the Iranians to grant a special permit, then only small arms could be carried, while every convoy was checked, inspected and escorted by Revolutionary Guards. It was the same when our empty vehicles re-entered Iran.

Such was our pipeline. For all its complexity, cost and length, somehow it worked. Of course there was much bellyaching from aggrieved Commanders, who protested bitterly that they were starved of supplies. In some cases they did go short, but I know of no battle that the Mujahideen lost for lack of ammunition, certainly not during the years 1983-87. Most often it was Commanders whose Parties were inefficient, or who operated in areas remote from strategic targets, or who lacked vigour in the fighting, who had cause for complaint.

My problem was in getting the right type of weapon and sufficient ammunition to the right Commander, at the right place, at the right time. If I achieved this it was usually the prelude to operational success. It involved thinking months ahead. Up to nine months were needed to organize operations in the north. It was this inescapable time lag between the conception of a plan and its execution that outsiders, such as the CIA, so often failed to comprehend.