Humans are still way ahead of sharks in the competition to be top predator. Last year, while thousands of sharks were eaten by people, only 60 people were bitten by sharks, 20% fewer than the previous year. The international file on shark attacks, compiled in Florida, reported only three deaths worldwide (www.flmnh.ufl.edu/fish/Sharks/Statistics/2002attacksummary.htm). Sharks have always preferred surfers to other water users (56% of attacks), but the odds against injury are still stacked in the surfers' favour.

If you are still worried about sharks, why not try snowboarding instead? No shark attacks were reported among snowboarders last year, though there are plenty of other ways to hurt yourself. This month's British Journal of Sports Medicine reports two cases of chronic subdural haematoma, presenting several weeks after unremarkable bumps on the head (2003;37:82-3). Both patients were fine after burr hole surgery.

Nova Scotia has no hospices, so patients with terminal cancer must die in hospital or at home. Ten years ago, four fifths of patients with cancer died in hospital. Most still do, but there is a measurable trend towards death at home—a 52% increase between 1992 and 1997 (Canadian Medical Association Journal 2003;168:265-70). The researchers who reported it suspect the trend has something to do with closing hospital beds. They also say, however, that most patients would choose to die at home if they could.

Sir Isaac Newton, one of Britain's greatest scientists, was no stranger to spin. An essay in Science (2003;299:831-2) says he manipulated his public image by releasing different portraits to different audiences. To some he appeared as a country gentleman, to others a visionary Roman swathed in fine robes (his favourite). The solitary troubled genius painted by Godfrey Kneller, now his most famous image, wasn't seen in public until after his death.

William Harvey, another great scientist of the 17th century, was probably the first to record the link between mortality and air pollution. In his account of an autopsy, transcribed in Occupational and Environmental Medicine (2003;60:147), Harvey clearly blames London's filthy air for killing Thomas Parr in 1635. Parr was said to be 152 when he died.

Recent studies show that ibuprofen interferes with the effect of aspirin on platelet aggregation. Ibuprofen may limit related benefits in people, according to a Scottish cohort study (Lancet 2003;361:573-4). People with cardiovascular disease who were discharged from hospital with prescriptions for aspirin and ibuprofen died sooner (on average) than patients taking aspirin alone. The effect seems confined to ibuprofen: other non-steroidal anti-inflammatory drugs had no impact on mortality in this cohort.

Sober, sensible readers who get to bed by half past ten may have no fun, but they may be less likely to get heart disease than the rest of us. Stay up late partying every night (or even working) and you could increase your risk of heart disease by 39%, according to data from a long-running cohort of American nurses (Archives of Internal Medicine 2003;163:205-9). Between seven and nine hours of sleep a night seems about right.

It's hard to measure how much fun children are having, so paediatric researchers looking for a health dividend had to measure what parents spent on fun instead (American Journal of Public Health 2003;93:288-9). They found a direct association between children's health and the amount of money their parents spent on trips to the zoo, skating rinks, and swimming pools. Perhaps having fun is good for children, or perhaps it's just harder for them to have fun when they are sick.

Randomised trials are a superior kind of evidence: a gold standard, a holy grail. It's a pity, then, they are so hard to do. One team of frustrated surgeons describe their efforts to recruit patients for a trial of pain relief after cholecystectomy, a common operation (Archives of Surgery 2003;138:59-62). After five months they had only seven participants, having excluded another 59. The commonest reason for exclusion? Straightforward refusal by the patient.

There were no such problems for surgeons from one Singapore hospital who managed to persuade nearly 200 patients to participate in their randomised study comparing two operative techniques for haemorrhoidectomy (British Journal of Surgery 2003;90:222-6). The surgeons don't say how long recruitment took but the end results show that surgical technique (staples or diathermy) makes no difference to patients' risk of perioperative bacteraemia.

British and American scientists have developed a topical human monoclonal antibody that seems to protect monkeys from sexually transmitted simian HIV (Nature Medicine advance online publication, www.nature.com/nm/10.0138/nm833). The new agent, applied to the vagina in a gel or solution, prevented infection in nine out of 12 macaque monkeys who were challenged with vaginal inoculation of simian HIV. Twelve of 13 control monkeys became infected after the same inoculation.

Minerva usually skips papers reporting new genes for old diseases, but she couldn't ignore a new gene for sleepwalking reported in Molecular Psychiatry (2003;8:114-7). It's been clear for some time that sleepwalking runs in families. The inheritable component of this common illness is probably due to a family of genes called DQB1. A preliminary study in 60 sleepwalkers showed the genes were considerably more common (35%) among sleepwalkers than among controls.

Old myths about colds simply refuse to die, at least in Boston, USA, where a recent survey shows that 60% of parents believe colds are caused by changes in the weather, and 40% of parents think you can catch a cold by going out with wet hair (Pediatrics 2003;111:231-6). Two thirds of parents think bacteria can cause colds, and over half believe antibiotics are a useful treatment. Parents who believe in antibiotics are more likely than other parents to take their children to a doctor.

Figure.

A 35 year old laser engineer with an amblyopic left eye (permanent poor left vision since childhood) momentarily removed his safety goggles while repairing a YAG laser used for removing skin tattoos. Unexpectedly the laser fired, and a flash reflected from a dark surface into the engineer's good right eye, causing a foveal burn with retinal haemorrhage, macula hole, and consequent severe loss of acuity. These untreatable injuries which result in permanent and sometimes devastating loss of vision are being seen more frequently yet can be avoided by adhering to safety protocols.

M Ramchandani, specialist registrar, A Calcagni, senior house officer, Birmingham and Midland Eye Centre, City Hospital NHS Trust, Birmingham B18 7QH