Regardless of venue, the ritual is the same. The dive boat reaches the dive site and ties up to the mooring. The sound of the boat's engine brings in the sharks from their dispersed areas. The divemaster/shark feeder swims the fish-filled crate down to a clearing in the coral heads on the reef and waits. Divers enter the water and make their way to the bottom where they sit in a semi-circle, backed against the coral heads, about 6m from the bait box. A dozen or more Caribbean reef sharks mill around, swimming in from down-current, attracted to the scent of the fish carcasses from the crate. The panorama is awash with bubbles from the exhalations of excited tourist divers, which is somewhat off-putting for the sharks, but their instinct to feed overrides their natural wariness and they come close to the divers, often to less than a metre away.

When the mood strikes, the shark feeder, who wears a chainmail gauntlet from fingertips to shoulder on each arm, thrusts a metre-long metal pole into the box and spears out a fish carcass. He offers it to the most dominant shark, although the excitement causes several sharks to swarm energetically around the feeder. By keeping the bait enclosed and selectively offering food at delayed intervals, the shark feed can be drawn out for up to an hour. Even when the supply of food is gone, the sharks continue to mill around, although at a more cautious distance from the divers.

Not to be outdone, dive operators on Grand Bahama Island (notably unexso, the Underwater Explorers Society, and Xanadu) have added a touch of the medieval to their shark feeds. The divemaster dons a full-length suit of chainmail to do battle with the local sharks. Not a bad idea, considering the hands-on approach taken: the shark feeder holds out the bait and lets the sharks take the fish right from his or her hands. Occasionally, a shark will clamp down and shake the feeder like a rag doll, to the delight of the tourists. Often, the feeder will reach out and grab a passing Caribbean reef shark and stroke it, just so, until the shark stalls and seemingly falls into a trance. Because of the association the sharks make between humans and the doling out of food, non-chainmail-attired guests must kneel on the seabed, well away from the feeder, with their hands tucked up under their armpits. This is done to prevent the sharks mistaking a lily-white hand for a piece of bait and the obvious bad feelings it would engender should the shark sample the mistaken food item.

Of all the organised shark feeds in the Bahamas, the most spectacular, and the most unique, is conducted on an irregular basis at Walker's Cay, northernmost island of the Bahamas at the top of the Abacos chain. Walker's Cay is a private, intimate, self-contained island (100 acres in total) owned by an American industrialist, and was originally created solely as a sport-fishing resort. During the high season, with each boat-slip occupied, its horseshoe-shaped marina is filled with more than $50 million worth of fibreglass, aluminum, teak, marine electronics and fishing tackle. Diving was first added as an optional activity for guests who didn't fish or who sought variety, but all that changed in 1991, when Gary Adkison and Barry Albury began to experiment with shark feeding at a dive site where they'd often seen sharks. Seven years later, Walker's Cay's main drawing ratio has changed from fishing 90 per cent, diving 10 per cent of its bookings, to fishing 10 per cent, diving 90 per cent.

And what started as a 'typical' shark feed with 30 or more participants has since evolved into an event known as Shark Rodeo, which regularly features 100 or more sharks. The star of the rodeo - at least as far as the sharks are concerned - is Gary Adkison's own invention: the 'chumsicle.' This is a mass of fish carcasses, weighing approximately 75 kilos, packed around an iron frame, which is placed into a large garbage can and frozen solid. The chumsicle is taken to the site and lowered into the water. One line runs from the bottom to a light anchor and another from the top to a large float. Thus, the mass of chum is suspended in mid-water, 4m from the sandy bottom, the centre of feeding activity with 100 or so sharks circling, as if orbiting, the food source. Additional swirling 'satellites' that find the free meal irresistible are an abundance of yellowtail snappers, groupers, jacks and other reef fish.

The dive site is a natural amphitheatre upon a sandy plateau, 15m deep, surrounded by coral walls. An ideal setting, but before the chumsicle is put into the water, Gary Adkison briefs the divers on exactly what they are going to witness and how the operation will take shape.

'This dive is a myth-breaker,' Gary begins, before going on to explain how the divers are free to leave their position by the reef once the sharks start feeding. The activity is monitored by Gary and his divemasters, who signal to the divers if there is any sudden change in behaviour - most notably, if a chunk of fish breaks off from the chumsicle. When this occurs, the sharks compete for the food item and the result is Shark Rodeo.

'When a piece of food breaks free from the chumsicle and is free in the water column,' Gary explains to the divers, 'the sharks race each other to get to the fish. That's when we'll ask you to move back. It's not that they're going to bite you - they're pretty smart, they know you're not food. They think you're another predator, and you are - and we tell everyone this - you're much worse, you're the worst predator in the ocean. What we're concerned about is that they're going to bump into you accidentally. If they whack you with their tail, they can really, really hurt you.'

Upon reaching the site, boat captain 'Snoopy' Cooper revs the engines for five minutes. There are sharks already in attendance but more come quickly. The divers enter the water and swim to the bottom near the coral walls.

Sharks are present as far as the eye can see. They mill around, circling, at all levels, from the bottom to the surface. Two species are prevalent: Caribbean reef sharks and blacktip sharks. The actual number is said to be 80 to 100, but at certain times of the year the figure can be twice that amount. The sharks keep a distance of 4m or so from the divers, who by now are exhaling at a rate bordering on hyperventilation. When all the divers are settled in, a divemaster surfaces and signals the boat to lower the chumsicle.

Activity changes to an electric excitement with the addition of a food source. The most vicious predators are a swarm of yellowtail snappers - it is difficult to make out the chumsicle through their cloud. But the more dominant sharks make passes, parting the snappers, bumping the chumsicle and gnawing at it. The bait mass is too large and too solidly frozen for them to get a mouthful at first, but over the next 30 minutes to an hour, as the chumsicle thaws and is worked on by fish and shark, filleted carcasses separate from the frozen form and bursts of feeding activity are witnessed.

The sharks become oblivious to the divers and their noisy, distracting bubbles. They approach close enough to touch and, indeed, bravado reigns as diver confidence increases, prompting many divers to stroke a passing shark or grab hold of a tail. The sharks shrug off such contacts without any retaliation.

It becomes apparent that of the 100 or so sharks, only a small percentage actually tries to feed on the chumsicle. There is a social hierarchy at work here, known as 'bite or flight', and the dominant sharks keep subordinates in check.

And then, ultimately, one shark will dislodge the remaining mass from the frame. It is too large to swallow, but the shark will attempt to make off with his prize, carrying it like a football in his jaws as he swims away. And, just as in a football game, all the other 'players' will then suddenly rush after him, trying to take possession of the 'ball'. The spectacle of so many sharks swarming around the leader is the highlight of Shark Rodeo - and a signal that the encounter is winding to a close. When the chumsicle is no more, the sharks will remain in the area, but will now keep a distance from the divers. Once the stimulus to feed is removed, sharks revert to their normal state of wariness of human visitors.

Critics of the business of shark feeding for entertainment are many and their arguments seem reasonable at face value. At the core of all arguments is the underlying concern that a human being may be injured. This fear, which is not backed up by the weight of statistics, is usually annexed to the additional thought: 'and there could be a backlash against sharks'.

There have been accidents, although not well publicised and certainly no one in the Bahamas wants to go on record about them. Primarily, those bitten were the ones feeding the sharks. Most injuries were minor. And as far as a backlash against sharks is concerned, that argument is null and void. Any backlash in this day and age is more likely to be settled in a court of law and by insurance companies, than taken out on unidentified animals in the wild.

This unreasoning fear of sharks comes from a lack of field experience. Statistically, more people will be injured or will die as a result of putting on scuba gear (or driving a car to their diving destination) than by experiencing sharks, fed for entertainment or otherwise.

You can almost hear the plaintive wail of the armchair ocean specialist: 'But sharks are wild animals, sharks are unpredictable.' Unpredictable! Every time I read this word or hear it uttered by experts on television I feel a wave of nausea. It is the mantra of the 'shark expert', always said with a grave, knowing expression and sotto voce. Sharks are very predictable. Given a stimulus to feed, they feed. It is the product of the evolutionary process that placed them high in the food chain. Get in between their food and their teeth and you can predict some misfortune.

The concern for diver safety is paramount for all professional shark-feeding operations (with the exception of a few frauds running white shark operations in South Africa, see Dive International, July 1997). If that concern is due to fear of lawsuits rather than fear of sharks, so be it. The point is moot. Safety dictates the protocol of each operation.

In the Bahamas, one could argue the Shark Rodeo is the safest way to run a shark-feeding operation, because the feeding association with a human is removed, other than that there are humans present. Next would come the operations that feed sharks with the help of a pole. Because these keep the food away from human hands, the sharks do not associate the human hand with contact with food items. Finally, the chainmail suit could also be considered safe because the electrical field emanating from the suit tends to focus the sharks' interest on the feeder and away from the watching divers.

In Australia, the use of an anti-shark cage is ultimately safe because a truck would have trouble getting through the bars. In South Africa, a lesser-gauge mesh is used for cages, but then white sharks in South Africa are much less aggressive than in Australia.

There is an argument that feeding sharks on a regular basis and at a designated site alters their behaviour, conditions them to contact with humans (associating human activity with food) and affects the environment. It's all true, and as unimportant in the grand scheme of things as the environmental impact of spitting in the ocean.

Shark feeds can only take place where sharks occur naturally. If there are no sharks in the area then all the blood, fish parts and offal in the world will not draw them in. This is one reason why shark feeds are not advertised in Bonaire or in the Virgin Islands. There are sharks in all waters, but if they don't regularly visit a particular area then they're not going to start just because a dive operator has a bucket of fish guts. The silky sharks were resident under the autec buoy before anyone had the notion to feed them; the Caribbean reef and blacktip sharks regularly aggregated around Walker's Cay before the first chumsicle was lowered.