Most textbooks tell you that primates have a lousy sense of smell, but watch a monkey investigate a piece of food, and the first thing it does is have a good sniff - which doesn't seem like the reaction of an animal with a poor sense of smell. Mattias Laska was intrigued by this apparent conundrum,so he set out to test just how sensitive these animals are to different smells. After testing monkeys with a battery of scents, he's added a selection of alcohols to the primates' repertoire of detected smells, proving that monkeys are better `smellers' than anyone had supposed(p. 1633).

Laska has spent the last 15 years training animals to detect specific smells, and most of the time, this isn't easy. Even though dogs have some of the sharpest noses on the planet, very few people have successfully trained dogs in order to test their sensitivity to individual scents. Man is probably the easiest animal to train, so more is known about the human sense of smell. Even then, only 300 monomolecular scents have been tested on us. Having successfully trained a variety of creatures from honey bees up to mammals,Laska was confident that he could successfully train primates to tell him when they could detect smells.

He decided to test two primate species' responses to a class of chemicals they encounter in the wild. Ripe fruits emit alcohols at different stages of ripeness, so he tested the animals' sensitivity to eleven alcohol molecules that they might experience in their natural environment.

Laska knew that the best way to get the monkey's cooperation was if he designed a test that was more like a game. First he trained squirrel monkeys to recognise that small cups, labelled with a scent, carried a reward. Then he introduced the element of fun. Laska offered each monkey a choice of 18 small sealed containers, half of which had scent labels, and suspended them half way up a climbing frame so they had to work a bit to earn their reward. Laska tested which of the eleven alcohols they could detect, and the minimum concentration the monkeys needed to pick up the scent.

But when Laska tried the same test on larger macaques, he ran into problems. Instead of making a choice, the macaques stuffed all of the cups into their cheek pouches. Undeterred, Laska completely redesigned the experiment, made the containers larger, added a clasp to the lids and cut the choice down from 18 to two. This time the macaques cooperated, and Laska began collecting more data.

The results were incontrovertible, not only could both species sense all eleven alcohols, but they were most sensitive to the longest alcohol, octanol,even at a dilution of 1 part per billion! Laska also tested whether the primates were sensitive to different alcohol isomers, but neither species responded to the subtle chemical difference.

Laska attributes his success to asking the monkeys to play, rather than setting them a dull chore. Ultimately he's hoping to correlate the monkey's scent-world with the animal's behaviour, so that he can classify how certain smells affect their behaviour. Not bad for an animal with a poor sense of smell!