Colour me beautiful

Racehorses, green fluorescent mice and red canaries do not seem to have much in common, but, in their own way, they are all the product of genetic engineering. It is often assumed that genetically modified organisms are a result of the molecular revolution in the 1990s and they are epitomised for many by Dolly the sheep. But Dolly was more a product of reproductive technology than genetic tinkering.

Racehorses, green fluorescent mice and red canaries do not seem to have much in common, but, in their own way, they are all the product of genetic engineering. It is often assumed that genetically modified organisms are a result of the molecular revolution in the 1990s and they are epitomised for many by Dolly the sheep. But Dolly was more a product of reproductive technology than genetic tinkering.

The genetic manipulation of animals and plants is not new. In the late 17th century, English racehorse owners were importing Arab stallions but not Arab mares, and were forced to breed their stunning studs with less-than-glamorous home-grown mares. Using the pragmatic trick of mating the female offspring of these mixed parents back to other Arab stallions over successive generations - a technique they called "grading up" - racehorse breeders with no knowledge of genetics created an almost pure Arab bloodstock.

Producing Arab racehorses in this way was ingenious, but not particularly difficult. Arab stallions and English mares are members of the same species after all and perfectly happy to breed together. The only thing it required was patience since it needed several generations to create a high-quality racehorse. It would have been much more difficult to put the Arab genes into a cow to produce a sleek milking machine. Indeed, it would not have worked. The two species are too dissimilar to produce viable hybrid offspring.

Today's technology, however, permits researchers to place the genes of one organism into the genome of a very different one, and such transgenic organisms are now produced routinely. One of the most useful of these is the green fluorescent mouse. First created in 1998, it contains a jellyfish gene. Many marine animals, including some jellyfish, glow in the dark - possibly so they can find each other in the murky depths - and in 1992 the jellyfish's green fluorescent protein gene (GFP for short) was "cut and pasted" into a bacterium. From there it was eventually introduced into a number of other organisms, including mice and plants.

GFP is now an invaluable tool for molecular biologists. It can light up those cells currently expressing particular genes, allowing researchers to see an otherwise invisible process. Transgenic animals also have more specific medical benefits. Chickens with human genes produce eggs containing human proteins in their albumen which - it is hoped - will eventually be used to treat cancer. Getting human genes into the chicken genome was just the first step. Getting these genes to do what they are "supposed to do" is the second and more difficult stage. Increasingly clever tricks are now used to get foreign genes into the chicken genome. These include infecting the chick embryo with a viral carrier bearing the human genes, or adding foreign genes to rooster sperm and relying on a more conventional mode of delivery.

Back in 1920, when the first transgenic animal was made, this would all have sounded like science fiction. The first truly transgenic animal was the red canary. The original, wild canary - which is still abundant on Tenerife and Lanzarote - is a grey-green bird. Years of selective breeding between 1400 and the mid 1600s by enthusiasts transformed it into the familiar all-yellow bird. By the early 1900s there were also deep-yellow canaries that bordered on orange and some fanciers fantasised about going that final step and creating a red one.

The impetus for doing this was the red siskin - a tiny, South American finch - that bird dealers started to import into Europe in the early 1900s. Centuries before, the Spanish had successfully hybridised red siskins with ordinary canaries. Knowing this was enough to convince Hans Duncker, a German school teacher and amateur bird enthusiast, that he could create a red canary.

Duncker planned to use exactly the same technique that the English had employed so successfully with racehorses. His impeccable logic was based on several bits of existing knowledge. First, he knew from his previous studies that the domesticated canary's yellow plumage is recessive to wild type (green) plumage. He was also aware from the reports of bird keepers that the canary's yellow colouring was recessive to the plumage of any other finch it hybridised with.

Since the 1600s, fanciers had been obsessed with breeding canary hybrids. Male finches kept with female canaries were like men in singles bars. Initially extremely choosy about who they would mate with, but as closing time (or the end of the breeding season) approached, they'd take whatever they could. The resulting offspring invariably resembled the finch parent. Duncker relied on the fact that his siskin-canary hybrids would be red. His idea was then to mate these hybrids back to yellow canaries, reject the poorly coloured ones and retain only the reddest individuals. By repeating this over several successive generations, the end point would be - he predicted - a canary the colour of blood.

Duncker partly succeeded and partly failed. He was successful because he did get red siskin genes into the canary's genome, and against all the odds some of his siskin-canary hybrids were fertile. It is commonly assumed that hybrids, like the offspring of the horse and donkey, are sterile. Whether they are, though, depends on their sex, and, as the famous evolutionary biologist JBS Haldane pointed out in 1922, on whether they are the sex-determining sex. In humans and other mammals, males produce X or Y sperm in equal numbers. If an egg is fertilised by an X sperm, the resulting offspring is female (XX); but if it is fertilised by a Y sperm, the offspring is male (XY). So, in mammals, males determine the sex of their offspring. In birds and butterflies, it is the other way around and females dictate the sex of their offspring.

What Haldane noticed with hybrids was that the sex-determining sex (technically referred to as the heterogametic sex) is either dead or sterile. In birds this is the female, and bird breeders had known for years that their hybrid offspring are much more likely to be male than female. No one has any real explanation for Haldane's rule other than the fact that the sex chromosomes cause problems when the genomes of the two species attempt to fuse at fertilisation. Duncker's siskin-canary hybrids were mainly male and a few of them proved fertile when mated back to canaries.

Duncker did not quite succeed in creating a red canary however, because "grading up" across the species boundary was far more difficult than he had ever imagined. The birds he produced were a coppery colour, but not red. There were several reasons why Duncker struggled to produce a red canary. One was that genes in the wrong genomes do not always behave in the way they "should". Current technology may make it easier to produce transgenic animals today than in Duncker's day, but many of the same problems still exist and the failure rate is high. Not until the 1960s were the first truly red canaries produced when British and American fanciers realised that as well as "red genes" the birds also needed an environmental input in the form of dietary carotenoids - red pigments. Both nature and nurture were needed to create a red canary.

'The Red Canary' by Tim Birkhead is published by Weidenfeld & Nicholson, £16.99