?

I¡¯m dangling somewhat inelegantly from the end of a rope, 200ft above the sea on Skomer Island off Wales¡¯ Pembrokeshire coast. Bracing my feet against the cliff face, I gingerly direct the tip of a long fibreglass fishing rod towards a bunch of guillemots. The colony smells something like a fishy pig farm and the noise is deafening. Beneath me the sea is pounding the rocky shoreline; behind me is my climbing buddy and research assistant; and in front of me a hundred adult guillemots are belching out deep guttural roars of parental agitation, and their fluffy offspring are squealing like demented songbirds.

The tip of the rod has a wire hook, which I gently manoeuvre around a chick¡¯s leg before pulling the bird towards me. The roar of the parents increases, but I now have the chick in my hand. Passing it back to my climbing buddy, he places it in a net bag. On and on it goes, chick after chick, hour after hour, and all the while my assistant keeps up a relentless banter - mainly about failing ropes or dirty chicks - so that we are continually laughing despite our precarious position. Once all the chicks are taken, we secure ourselves on tight ropes and begin the process of ringing them: a metal ring on one leg, a plastic one (made of a material called Darvic) on the other. The Darvic ring is a different colour each year, with a number engraved in large print on it so that if the chick returns to the colony in future, we can read the number through a telescope from adjacent cliffs and note its wearer¡¯s identity.

As soon as the chicks are ringed, I use the fishing rod to return each one to its rightful ledge and its anxious parents.

For four decades, the annual ringing expedition to Skomer has been the highlight of my year. It is both exhilarating and exhausting. Wielding that fishing rod at arm¡¯s length for days on end and knowing that one false move could pitch a chick to certain death makes it tense work. The specially designed metal guillemot rings require considerable pressure and dexterity to close, and snapping the ringing pliers shut on one¡¯s fingers is a distinctly unpleasant occupational hazard - especially when one is already splattered with copious amounts of rank guillemot shit. Working on cliffs is stressful, too. I used to use the Whillans sit harness (designed by climber Don Whillans), which allowed me to dangle securely from my rope with both hands free. But that security came at a price: the harness was a cojone-crusher - no joke for someone who studies sperm competition back in the laboratory.

É«ºÐÖ±²¥

My study started in 1972, the year I graduated and began my PhD. It turned out to be a long-term thing.

Like many doctoral students, after I gained the degree I felt that if I had my time again I could do it so much better. In most cases the opportunity does not arise, but in my case I acquired a lectureship within a year of my doctorate and was soon able to arrange for an extremely competent and highly motivated student to redo my PhD research - rectifying my failures and building on my successes.

É«ºÐÖ±²¥

The justification for a repeat performance was easy. The rationale for my doctorate was that guillemot numbers on Skomer had declined enormously, from about 100,000 birds to 2,000 in just 30 years. Numbers stabilised at about 2,000 during my PhD years, but four years later - to my amazement and delight - they started to increase. A comparison of a stable versus an increasing population was irresistible and my student, the aptly named Ben Hatchwell, did such an outstanding job that I realised that the study was worth continuing. By the time Hatchwell finished his doctorate, we both knew a great deal about the pros and cons of studying a bird that lives at incredible densities on sheer cliffs on remote islands. I like a challenge, and so does he.

What is the point of all this effort? My PhD research was - as is so often the case - overly optimistic. In retrospect, I really hadn¡¯t a hope in hell of finding out in just three years why guillemot numbers on Skomer had collapsed. The evidence for the decline was unequivocal, however: photos taken in the 1930s on a large-format camera showed birds in numbers almost unimaginable in the 1970s. Yet even in the latter decade, no one knew how to make a proper census of guillemots. You have only to look at a cliff face full of the birds to see how difficult they are to count; and even when you can count them, figuring out what it means in terms of breeding pairs is no trivial task.

I had a go at understanding why the numbers had declined, and to do that I tried to estimate three main parameters: breeding success (the proportion of pairs rearing a chick to fledging); survival of those chicks to breeding age; and annual adult survival (the proportion of breeding birds that survive from one year to the next).

Measuring breeding success was the easiest of these difficult tasks and merely required sitting in a hide for six to eight hours a day for three months recording when birds laid their single egg, and seeing whether the chick hatched and fledged. It sounds straightforward, but when birds are bunched together at densities of up to 70 pairs per square metre, it¡¯s tricky. That is what the ringing is all about. We found that young birds don¡¯t return to Skomer until they are two or three years old, and most don¡¯t breed until they are seven years old. A few duffers are still unpaired at 10! Today we know that about half of all chicks that fledge survive to become breeders. And the adult birds? Year-to-year survival is about 94 per cent, indicating a breeding life of about 25 years.

Those are the basics. We now know that the high survival rate of young birds is responsible for the continued increase in numbers on Skomer: 2,000 in the 1970s to more than 20,000 birds today - but there is still a long way to go to get back to 1930s levels.

Our results provide strong evidence for the effects of climate change: guillemots breed two weeks earlier now than they did in the 1970s. But we also know that adults survive less well in the warm, wet, windy winters that are becoming more frequent as sea temperatures rise. During the course of our study there have also been four major oil spills: although often occurring hundreds of miles from Skomer, such events double the adult mortality rate of our birds.

We have learned a lot about guillemot social behaviour, too: birds maintain long-term pair bonds but, like humans, aren¡¯t averse to a bit on the side. Some 7 per cent of chicks are fathered by the chap next door - this is low compared with some birds, but high for a seabird.

É«ºÐÖ±²¥

Forty years is a long time to keep a study going. How does one do it? Tenacity is the key, obviously. ¡°I¡¯ve started so I¡¯ll finish¡± runs through my head each time I visit Skomer. Understanding the population dynamics of a long-lived bird species requires long-term commitment: simply working out how old guillemots are when they first breed required a minimum of 10 years. On top of that, environmental conditions vary - the climate change effect attests to that. Also, one finds better ways of doing things. It sounds trivial, but the use of the fishing rod to catch birds was an innovation that increased our efficiency enormously and at the same time minimised disturbances.

É«ºÐÖ±²¥

Long-term studies require commitment in terms of keeping the data in order. Yes, we now have a fabulous customised database, but it wasn¡¯t that long ago that we were taking huge numbers of handwritten index cards to and from Skomer each field season.

Researchers do not survive by tenacity alone; money is essential. But finding it is difficult for short-term studies, let alone long-term ones. Finding funding for 40 consecutive years has been a major challenge. Cash is necessary to pay field assistants¡¯ salaries, for travel, accommodation, subsistence and safety equipment. Most research grants are for three years, very occasionally for five - hardly suitable for studying long-lived animals.

Even if one is lucky enough to secure consecutive research grants, there¡¯s always the risk - especially in today¡¯s lottery-like research funding system - that you might fail to get a grant and thus miss a year or more of data. Such breaks in long-term studies are disastrous, and force researchers to seek alternative funds; some have even used their own salaries to keep things going. Over the years we have been lucky enough to have had research council funding, but we have also had much-appreciated support from less conventional funding bodies - notably the Countryside Council for Wales. To obtain research council funding one has to demonstrate results, but with long-lived species they can be a long time coming. As a result of this uncertainty, researchers typically have shorter-term studies running alongside the main one. Indeed, that is one of the benefits of long-term studies - they provide numerous opportunities for additional work that capitalises on the accumulated data or knowledge from the main one.

Back in the 1970s, one of my original goals was to establish whether the Skomer population was self-sustaining. It had been suggested that the huge fluctuations in guillemot numbers at certain colonies might be the result of mass movements of birds between different parts of the North Atlantic. It has taken us 40 years to acquire sufficient data to convincingly show that the Skomer population is indeed self-sustaining. Along the way we have completed numerous other investigations, as well as training PhD students and field assistants.

It is deeply ironic that uninterrupted funding for long-term studies is so difficult to obtain. Many such studies are disproportionately successful in terms of their contribution to our understanding of the ecology, evolution and behaviour of particular species. Those based on just three field seasons - the duration of a typical grant or PhD project - may sometimes yield lucky results, but often with little understanding of the year-to-year variation in the species in question¡¯s biology.

By contrast, and for obvious reasons, long-term studies are rare. In total there aren¡¯t many more than a dozen or so. Two of the most famous are the research run by Tim Clutton-Brock, Prince Philip professor of ecology and evolutionary biology at the University of Cambridge, which is examining red deer on the Isle of Rum in Scotland (started in 1972); and the study of great tits in Wytham Woods currently overseen by Ben Sheldon, Luc Hoffmann professor of field ornithology at the University of Oxford, which has been running continuously since 1947.

Long-term studies of individually marked animals provide a deep understanding of a species¡¯ biology and hence much more accurate and meaningful measures of their social interactions, of their relationship with their environment, and of basic biological processes such as selection and senescence. For example, they tell us about the way a female¡¯s position in a social hierarchy affects her reproductive success and longevity as well as providing information about the sex ratio of the offspring she produces and their survival. Such studies allow researchers to measure lifetime reproductive success rather than merely taking a three-year snapshot and hoping that it is representative. It is precisely because of the insights that long-term studies provide that publications reporting those results typically appear in high-ranking journals and make an important contribution to research assessment exercises.

Although the original objectives of a study may remain unchanged, methodologies and technologies continue to improve. Working with Tim Guilford, tutorial fellow in zoology at Merton College, Oxford, we now use various tracking devices to see - literally - where Skomer birds go to find food for their chicks and where they spend the winter. Mercifully, the climbing equipment has improved, too. The Whillans sit harness (affectionately known as the ¡°shit harness¡±) has now been replaced by a less restrictive and much more comfortable arrangement.

É«ºÐÖ±²¥

Despite the fact that Skomer is officially uninhabited, I haven¡¯t been alone: Ben Hatchwell continues to collaborate on the project, and we have relied on a succession of field assistants and island wardens to provide companionship and wonderful logistical support. The study will continue as long as we continue to secure funding: whether I¡¯ll manage 40 more years is another question.