A Buzz in the Meadow

by Dave Goulson

  One question that is often raised is what will farmers use to control the pests of crops such as oilseed rape, now that neonics are going to be withdrawn (at least temporarily)? It seems to me that we should return to IPM, minimising the use of pesticides by monitoring pest problems and boosting the numbers of natural enemies, using chemical controls only when they have to be used. Prophylactic use of persistent pesticides is not a sustainable approach because it leads to pesticides accumulating, damages the populations of bees and the natural enemies of pests, and is highly likely to lead to the evolution of resistance in the pests, as has already happened in the USA, where some Colorado-beetle populations are nearly immune to neonics. Exactly the same argument explains why doctors are reluctant to prescribe antibiotics, and why they would never prescribe them prophylactically to avoid illness – if they did, bacteria would swiftly evolve resistance and the drugs would no longer work.

  I would also question whether farmers always need to replace neonics with something else. The evidence that crop yields directly benefit from using neonics is surprisingly hard to pin down. Oilseed-rape yields are no higher now than they were in 1991, when no neonics were available. Recent studies of soya-bean farming in Brazil suggest that farmers would obtain the same or greater yields for less expenditure on pesticides if they switched from using neonics to using IPM. Studies from the United States have shown that soya-bean yields are not improved one jot by using neonicotoids. Yet in both Brazil and the USA nearly all soya-bean farmers use neonics.

  Why would farmers use chemicals if they don’t need them? Much of the advice given to farmers on what chemicals to use comes from agronomists, most of whom work for large companies involved in supplying pesticides. This is hardly independent advice, and is surely prone to overselling. If a farmer is advised that product X will provide his crop with the best protection, he is likely to buy it; it would be a brave person who ignored the advice of his agronomist. Yet he has no way of knowing that the agronomist isn’t recommending product X because it gives him the best profit. Farmers growing oilseed rape are generally recommended to use a neonic-dressed seed, but they are also told to spray several times in the autumn, spring and summer with other pesticides such as pyrethroids, so what function is the seed dressing serving?

  There are some clear lessons to be learned from all of this. Safety tests for new generations of agrochemicals need to be realistic, taking account of subtle sub-lethal effects that cannot be revealed by lab tests. As the French Science paper shows, even putting a hive immediately next to a treated flowering crop will not reveal much in the way of impacts if the compounds affect the bees’ ability to navigate over long distances. And safety tests are currently carried out by the agrochemical companies themselves, or by private companies whose main income derives from them. Does anyone really think that a company that has invested millions in developing a new chemical will be entirely unbiased when presenting its safety-testing data to the authorities? As someone who carries out experiments and analyses data all the time, I am aware that it is all too easy to influence the outcome of experiments, either accidentally or deliberately, and that data can always be analysed in a number of different ways, and these may not all give the same answer. And of course if the ‘wrong’ result emerges, it is easy to argue that the experiment was flawed in some way and that it should be repeated. I have no evidence that such things go on, but it would be surprising if they didn’t, with so much money at stake. All such tests and analyses should surely be carried out by strictly independent bodies.

  It is just over fifty years since Rachel Carson wrote Silent Spring, in which she highlighted how agrochemicals were devastating wildlife. At the time DDT and its organochlorine relatives, dieldrin and aldrin, were the main culprits. These chemicals have long since been banned in most developing countries in favour of safer chemicals. Are neonics so different? The influence of big industry seems to have pushed farming far away from the sustainable approaches that I was taught about in the 1980s.

  At this point in time it is hard to predict what will happen in the EU after the proposed two-year ban expires. Neonics will continue to be used extensively for non-flowering crops such as winter wheat, and with oilseed rape generally being grown in rotation with wheat, there are still likely to be neonic residues in the nectar and pollen. Even if we completely stopped using neonics, they would be in our soils for years to come. So any benefits from the partial moratorium will not be apparent within two years. In any case there seems to be no plan to monitor the benefits, so if they did occur (which is unlikely) we wouldn’t even know. All in all, it seems an ineffectual half-measure, a political manoeuvre perhaps intended to keep as many people happy as possible, but not one that has resolved the situation satisfactorily or permanently.

  I think environmentalists will triumph on this issue in the end. But new generations of neonics are rumoured to be in development, to which the current moratorium does not apply, and no doubt other compounds too. Environmentalists may win this battle, but they are losing the war, and wildlife is steadily paying the price. Until we fundamentally change the system for testing and approving new chemicals, and for advising farmers, how long will it be before something worse comes along? In the past we were assured that organochlorides were safe, until it transpired that they weren’t. Then we were told that organophosphates were all safe, until many of them were banned. Now Defra repeats the mantra that all pesticides are subject to rigorous safety testing, and that all of the ones that are currently licensed are perfectly safe. History suggests it is wrong. It is remarkable and sad that, half a century on from Silent Spring, we don’t seem to have learned any lessons at all.

  I must add one final note of caution. I am not claiming for one second that neonics are the only problem that bees or other wildlife face in the modern world. Bee declines are undoubtedly due to a mixture of factors, probably including diseases, Varroa mites (in the case of honeybees), lack of flowers, a monotonous diet and exposure to multiple pesticides, all providing a potent cocktail of stressors. It is very likely that these factors interact; bees that are mildly poisoned will be more susceptible to disease, less able to cope with starvation, and so on. We can’t easily solve all of the problems affecting bees, but we can stop poisoning them. Isn’t it time we did so?

  CHAPTER FOURTEEN

  The Inbred Isles

  8 July 2013. Run: 40 mins 48 secs. Slow again today, perhaps a little too much Côtes du Rhône last night. People: none. Dogs: 7. Butterfly species: 8. Damp morning; there was heavy rain in the night and ominous, bruised clouds to the west suggest more to come. I nearly trod on a young fire salamander that was taking advantage of the humidity and trundling slowly across the track near the Épenède water tower, a wonderful creature with a glossy black skin marked with golden-yellow streaks and spots. I placed it in the long grass for safety – perhaps I should knock ten seconds off my time to allow for salamander rescue?

  In the 1960s, some of the oddest ecological experiments ever to be performed were carried out by Daniel Simberloff in the Florida Keys. Dan was a grad student under the legendary E.O. Wilson, one of the great champions of the importance of insects in natural ecosystems and one of my personal heroes. Dan’s doctoral studies focused on testing some of the predictions of ‘Island Biogeography Theory’ (known as IBT to its close friends). IBT was an idea hatched by E.O. Wilson and his collaborator Robert MacArthur to try to explain the relatively small numbers of species found on oceanic islands compared to the mainland. MacArthur and Wilson’s theory is one of those things that now seem fairly obvious, but at the time it was proposed it was seen as ground-breaking and revolutionary.1 It had long been known that small, remote islands tend to have few species, whereas big islands and those very close to the mainland tend to have more. MacArthur and Wilson argued that the number of species on any island was a balance between immigration and extinction. Animals and plants occasionally get blown or washed off the mainland, and are more likely to arrive in o
ne piece at islands close to where they came from, or on big islands (simply because they are more likely to bump into them). Thus one might imagine that a big island just offshore from a major continent might be exposed to a constant stream of potential colonists washing up (immigration is high). In contrast, a tiny island in the middle of the Pacific is likely to receive few colonists (immigration is low). In addition, big islands are likely to have a greater diversity of habitats, such as mountains, forests and lakes, while small islands will have few. Thus colonisers arriving on a big island are more likely to find somewhere suitable to live.

  MacArthur and Wilson also proposed that the arrival of new species on islands will be offset by extinctions of existing species, so that once an island has been in existence for a while, we would expect there to be an approximately stable total number of species, with occasional extinctions balancing occasional new arrivals. Extinctions of island populations naturally occur every now and again, due to a multitude of causes – disease, storms, and just plain bad luck. Extinction is more likely, the smaller the population; in a big population there is more chance that a few lucky individuals will survive whatever catastrophe has struck. This pattern is likely to be exacerbated by inbreeding; in small populations it takes just a few generations for everybody to become related to everybody else. For reasons to which I will return, it is not a good idea to marry your brother or cousin, but you may have no choice if you live on a small island. Inbreeding is likely to weaken a population, again making small islands prone to high extinction rates.

  You may wonder why I am telling you all this, but the relevance will eventually emerge, so bear with me. To recap, few new species ever arrive at small, remote islands, and when they do they are quite likely to go extinct there before too long, and this is essentially why these small, remote islands tend to have fewer species than large islands near a mainland source of colonists. Now this all seems like an eminently logical theory, but it is hard to test. You would need to manipulate the size or remoteness of an island, which of course isn’t possible. Unless you are Daniel Simberloff. Dan hit upon the extreme idea of manipulating the size of mangrove islands off the coast of Florida. Mangroves are salt-tolerant trees that grow in the shallow, muddy waters off tropical coasts, and they support a wealth of insect, crustacean and bird life, as well as providing a spawning ground for numerous tropical fish. Clumps of trees grow wherever the sea is shallow and sufficiently sheltered, so the Florida coast is dotted with mangrove islands of various sizes and shapes, and of varying distance from the shore. Dan couldn’t easily make these islands bigger or move them further offshore, but with the enthusiastic deployment of a chainsaw he could certainly make them smaller, and this he proceeded to do. He found, as you will by now expect, that if he shrank the size of an island, he increased the rate of extinction, so that the number of species quickly dropped and then reached an equilibrium. This supported part of MacArthur and Wilson’s theory – the bit to do with island size and extinction rates – but left untested the parts relating to colonisation rate. How could one study the rate at which new species colonised an island?

  Dan devised another boldly destructive approach to tackle this question. He enlisted the help of a local professional exterminator named Steve Tendrich, who one might imagine must have thought this a highly eccentric project. Together they constructed scaffold frames over entire mangrove islands (admittedly rather small ones – some were just a few metres across). Plastic sheeting not being available at the time, they hung rubberised netting over the scaffold to encase the islands and then pumped in highly toxic methyl-bromide gas, fumigating the little islands and killing everything but the mangroves themselves. They managed to repeat this on six small islands at varying distances from the mainland (not a great sample size, but quite impressive, given the effort involved). After denuding the islands of their fauna, Dan monitored how quickly they were recolonised over the next two years. Lo and behold, islands near the mainland were quickly repopulated with a range of species, while those further offshore were colonised more slowly and never reached the level of species richness that was rapidly recovered on the inshore islands.

  If MacArthur and Wilson’s theory only applied to oceanic islands, it would never have created the stir that it did, and I would not be writing about it now. However, it quickly became apparent that the theory could be applied to habitat ‘islands’ separated by a ‘sea’ of unsuitable habitat. Obvious natural examples might include ponds or mountaintops, each of which supports specialist creatures that cannot survive for long in the intervening habitat. More importantly, man’s activities have created habitat patches – fragments of natural or semi-natural habitat surrounded by a sea of heavily modified land, such as farmland or urban areas. The large tracts of woodland that once covered much of Europe in a near-continuous blanket now exist as fragments, separated by fields of crops that are inhospitable to most woodland inhabitants. Flower-rich grasslands such as the meadow at Chez Nauche were once scattered thickly across the countryside so that no patch was far from any other patch, but they are now few and far between.

  Of course habitat islands are not quite the same as oceanic islands; for most animals, crossing an arable field is much easier and less daunting than crossing an expanse of sea, so – all else being equal – we might expect habitat islands to have more colonisers than oceanic islands. Nonetheless the broad principles remain and have profound consequences for conservation, for they make predictions as to how many species a nature reserve might be likely to support, and how we might prioritise conservation efforts and expenditure.

  Conservation efforts are limited by funds, and are constrained by other pressures on the land. In our modern, crowded world we can only have so much in the way of woodlands and meadows, for we need to grow food and build houses, out-of-town supermarkets, industrial estates and roads.2 Supposing we are faced with a situation in which we can only save, say, 1,000 hectares of woodland, would we be best to save one big wood or dozens of little ones? In 1975 Jared Diamond used Island Biogeography Theory to argue that the former would be the best strategy, for the larger patch would support more species. I’m sure he didn’t realise it at the time, but in doing so Diamond helped to spark one of the longest-running debates in ecological history, one that has rumbled on in various forms till today. It became known as the SLOSS debate: ‘Single Large Or Several Small’.

  Perhaps surprisingly, given that Diamond was essentially building on Dan Simberloff’s work, Simberloff himself was one of the first to disagree with Diamond, pointing out that the logic only worked if small habitat patches simply contained subsets of the species found in a larger habitat patch. In reality, this is rarely likely to be the case. Suppose, for example, that we had the job of prioritising woodland conservation in the UK – all but 1,000 hectares was to be swept away to provide extra parking. Which bit(s) should we save? One huge chunk of, say, the New Forest, or lots of little patches, from the twisted, lichen-encrusted oaks of Wistman’s Wood on Dartmoor, to a bit of ancient Caledonian pine forest at Abernethy in the Cairngorms and a patch of Breckland forest on the sandy soils of Norfolk? In this example, one would certainly save many more species by having lots of little patches, for these woods all have different characters and each supports a different range of species.

  Maybe this example is a little unfair, and it is certainly not a very plausible scenario. So let’s suppose instead that we had a single chunk of reasonably homogeneous habitat – let us say one large flower-rich meadow – and that much of it was to be lost to make way for a housing development. Would we be best to save one large piece, or lots of little ones scattered among the houses? This seems like the sort of scenario that planners might regularly face in the real world. David Quammen writes beautifully about the SLOSS debate in the The Song of the Dodo, in which he coins the Persian Rug analogy. Imagine we take a ten-foot-by-ten-foot Persian rug and cut it into 100 pieces. Do we get 100 perfect replicas of the original? Of course we do not; inste
ad, we get 100 worthless, frayed scraps of carpet. The same argument could be applied to the meadow: the little patches would be trampled by the residents, invaded by garden weeds and would be unlikely to support much more than a handful of the original diversity of life in the meadow.

  These two arguments lead to opposite conclusions: the first that we should save lots of small fragments; the second that we should save large, intact tracts. Resolving this difference occupied quite a number of the world’s ecologists for several decades in the later twentieth century, and there is still no clear agreement, although interest in the debate has finally waned.

  Of course to some extent all of this misses the point. In the real world we save what we can where we can, and generally make do with what is possible, rather than worrying about what would be ideal. The housing planners are likely to be much more driven by the practicalities of road access, drainage and so on, when deciding where to put the houses, than they are with worrying about biodiversity. Where a habitat has been largely lost we might try to re-create patches of it, but the size and location of such patches are rarely determined by any ecological theory, but rather by what is available. When I bought Chez Nauche I did not have unlimited choice as to where it would be, or how large the meadow. As I had very limited funds, its location was largely determined by the cheapness of property and land in the Charente compared to other parts of France, and by what was for sale. Nonetheless, many of the ideas that emerged from the arguments over Island Biogeography Theory are useful. For example, the distance from a source of colonists plays a role in deciding how to manage my meadow. If there were a lovely flower-rich grassland in one of the neighbouring fields, there would be a ready source of wild-flower seeds and the recovery of my meadow to a flower-rich state would have been rapid, just as Simberloff’s fumigated islands were colonised quickly when they were close to shore. Unfortunately there was not, so I have had to collect seeds from some flowers – particularly those with heavy seeds and no long-distance mechanism for dispersal – and sprinkle them in the meadow (avoiding, of course, my experimental plots, which would mess up my experiments).