“WE WERE SHOCKED,” says Brad Lister, of Rensselaer Polytechnic Institute in Troy, New York. “We couldn’t believe the first results. I remember [in the 1970s] butterflies everywhere after rain. On the first day back [in 2012], I saw hardly any.” Dr Lister is describing the Luquillo forest of Puerto Rico, where he recently carried out a census of insect life and found it had been almost wiped out in 40 years. But he could be talking about many other places. Over the past few years, scores of scientific studies have found declines in different measures of insect life and health, all of the order of 50-80%, in areas as far apart as Germany, California and Borneo.
The findings have triggered alarm, almost panic. Animals, mostly insects, pollinate 87% of flowering plants, according to a recent study by the UN’s Food and Agriculture Organisation (FAO). Without insects, most plants could not reproduce. They break down and recycle the nutrients that plants need for photosynthesis. They decompose organic waste and feed a large proportion of all birds and bats. E.O. Wilson, an American biologist, calls insects “the heart of life on Earth.”
The studies suggest that such life is in peril. One talks of “the dreadful state of insect biodiversity”. Its authors give warning of “the extinction of 40% of the world’s insect species over the next few decades”. If insects really do face extinction, it would be an immense environmental crisis. But how real is that possibility? What do the data actually tell us?
In terms of the number of species, insects are by far the most abundant of life forms. Scientists have identified and described over 1m species of insect, compared with only about 6,000 mammals and 18,000 birds. Insects are so numerous that they contain three times as much mass as humans and 30 times that of all wild mammals. “To judge by his creation,” a geneticist, J.B.S. Haldane, once quipped “God must have an inordinate fondness for stars and beetles.”
Little about this astonishing bounty is known. Using computer models of ecosystems, Nigel Stork of Griffith University, Queensland, estimates there are 5.5m species of insect and 6.8m of terrestrial arthropods (the wider category that includes spiders and crustaceans). That implies over 80% of insects remain undiscovered. Arthur Shapiro of the University of California, Davis, tells the story of travelling by bus across Patagonia, when he broke down in the middle of nowhere. In the two hours it took to mend the engine, and while he stayed within sight of the vehicle the whole time, he found three species of butterfly new to science. “That’s how little we know.”
Even when individual species are described, the process yields only partial information. Scientists have little hard data on what the vast majority of insects eat, how mobile they are or what determines their reproductive success. There has been almost no long-term monitoring of their numbers. “Around and beneath our feet,” writes Dr Wilson, “lies the least explored part of the planet’s surface.”
For many years, it did not seem possible or necessary to study insect populations. In their murmuring trillions, insects seemed safe from the pressures that were driving predatory mammals towards extinction. It was unimaginable that so many could disappear. As a haiku by Basho, a Japanese monk, puts it, “Nothing in the cry of cicadas suggests they are about to die.”
In the 1970s a few disturbing signs of decline began to appear. Long-distance drivers across America and Europe reported that their windscreens were no longer splattered with bugs. Pilots in the Arctic Circle now describe the same thing. Urban streetlights are no longer enveloped by clouds of photophilous moths. Insect-eating birds began to disappear. But these signs could be explained away by, say, more aerodynamic car designs, or changes to farming. The evidence was anecdotal.
There are, though, a few exceptions to the rule that no long-term databases of insect populations exist. The biggest is kept by Butterfly Conservation, a charity based in Dorset in southern England. It has records from 1690 but its most important data begin in 1976 when, concerned by the anecdotes of decline, two government scientists designed a simple monitoring system. Every week in summer, volunteer butterfly-spotters walk slowly along fixed paths, or “transects”, and log every butterfly and moth they see within 2.5 metres of their path. In Hethfelton Wood, near Wool, in Dorset, the path loops through a forest that, now partially felled, is reverting to heathland. Volunteers have logged 35 species of butterfly there since 2000, ranging from Graylings and White Admirals to thriving species such as the Silver-Washed Fritillary, which is rare elsewhere.
The project, called the United Kingdom Butterfly Monitoring Scheme (UKBMS), proved such a hit with Britain’s amateur naturalists that the scientists were overwhelmed with data. Now, 2,000 volunteers monitor more than 2,500 sites and produce 3m records a year. It is by far the most detailed insect data set in the world and, unlike most, measures both whether a species occurs in an area and how abundant it is. It shows that between 1976 and 2014, 32 of Britain’s 56 native butterflies declined in numbers, 21 by more than 40%. The biggest falls were among species classed as habitat specialists, with limited ranges or which feed on a small number of plants.
A second long-term set of data also monitors butterflies, is older and, heroically, is conducted by just one person. Every two weeks since 1972, Professor Shapiro of UC Davis has trekked along ten transects in central California, noting the butterflies he sees (159 species and subspecies). Since 1972, the number of species has fallen in half the transects and risen in one; 2017-18, he says, was “a terrible, perhaps even catastrophic butterfly year”.
The third data set is kept by the Krefeld Entomological Association, a group of professional and amateur naturalists in a town near Dusseldorf in western Germany. Their headquarters is lined with wooden cases full of insect specimens found in the surrounding grasslands, dunes and woods, meticulously labelled. In 1989 the society began setting up so-called Malaise traps, large tent-like structures that trap flying insects, in local sites in spring and summer. There are now 63 sites. This collection method records the total biomass of insects in the trap, a good measure of the amount of food available to birds and other predators but which tells you nothing about which species are being caught or how many of each there are. More than half the traps have only been checked once since 1989 and even those that are checked more than once are not monitored in consecutive years. The results are not those of a classic longitudinal study.
But when Caspar Hallmann of Radboud University in the Netherlands combed through the data in 2017, all doubts about their significance were silenced. Between 1989 and 2016 he found the biomass of flying insects in this corner of western Germany fell 77%, or over 5% a year. Making the results more remarkable is that the traps were set up in nature reserves which, though hardly pristine, are better protected from clouds of insecticides than most land in western Europe. “We were amazed,” says Dave Goulson of the University of Sussex, one of the co-authors. The study was the third most frequently cited scientific study (of all kinds) in the media in 2017 and pushed the governments of Germany and the Netherlands into setting up programmes to protect insect diversity.
Since then, more surveys have confirmed the results. Early this year, Francisco Sánchez-Bayo of the University of Sydney and Kris Wyckhuys of the University of Queensland, reviewed all the studies they could find mentioning insect decline. Theirs was the first study of studies. The authors found that 53% of lepidoptera (butterflies and moths) were in decline; 49% of coleoptera (beetles) and 46% of hymenoptera (bees, wasps, ants and sawflies). Over a third of insect species, they claimed, are threatened with extinction.
That species are failing in some places is not in dispute. What is less clear is whether the decline is global. Drs Sánchez-Bayo and Wyckhuys found a mere 73 papers. That is not enough, argues Alex Wild of the University of Texas, Austin, to say much about anything globally.
There have been no surveys of wild insect numbers in India, China, Siberia, the Middle East or Australia and only a single study each in South America, Sub-Saharan Africa and South-East Asia. These areas include almost all the tropics where the majority of insect species are thought to live.
Most of the biggest declines have been measured in Europe and the United States, where the human footprint lies heaviest on the landscape and where modern agricultural methods are almost universal. Given the paucity of evidence, it is impossible to say whether insect numbers really have declined the most in these two areas or whether they have fallen everywhere but these are the places that have been studied.
It is true that all 73 studies show declines. But that is because the authors went looking for that result. They typed the search terms [insect*] and [declin*] and [survey] into a database. “Estimates based on this ‘unidirectional’ methodology,” argues Chris Thomas of the University of York in Britain, “are not credible.” Nor do all studies show a decline (though they were not captured by the search). A recent study found pollinators are increasing in undisturbed habitats in south-eastern Spain. “This provides evidence,” that report says, “that pollinator declines are not universal beyond anthropogenic ecosystems.”
No less important, the relationship between declining insect numbers and damage to ecosystems is not a simple linear one. Both Drs Sánchez-Bayo and Wyckhuys and the UKBMS found that generalist pollinators (such as Brown Argus and Gatekeeper butterflies) are doing less badly than specialists, such as the White Admirals, which are now extinct in some regions. Generalists are presumably moving into niches vacated by specialists, a process that is not without risks. At low levels of diversity, ecosystems become vulnerable to diseases that can sweep through a single species. Nevertheless, this process is a reminder that insect ecosystems are more resilient than they sometimes appear.
That is also true of the floral ecosystems that depend on insect pollinators. Jane Memmott of the University of Bristol investigated what happens when insect species die out by gradually removing one pollinator after another in a controlled environment and keeping track of what happened to the plants. She found that, depending on which species was removed first, most plants managed to hang on even after more than two-thirds of insects had been removed, illustrating the resilience of ecosystems. Only when more than 90% of insect species were removed, did floral diversity collapse.
All of this provides reasons to be cautious about extrapolating too far from the data. But there are still three big reasons to worry about what is known.
First, the scale. Declines of more than 50% in most measures of insect health seem more severe than the diminution in other taxonomic ranks. A British study from 2004 found that insect species are declining faster than birds or plants. Almost all species and subspecies are affected, regardless of the altitude at which they live. Declines have been drastic even in protected reserves.
Second, it is worrying that so many influences are contributing to the decline. It is as if, in the insects’ world, everything is going wrong at once. The main causes seem to be, in order of importance: habitat loss (97% of wildflower meadows have been grubbed up in Britain since the 1930s); intensive farming, which leaves fewer unproductive parcels of land for wildlife; pesticide use; and the spread of diseases and parasites such as the varroa mite, once confined to East Asia where local bees had a measure of resistance, but which is now killing honeybees worldwide. David Wagner of the University of Connecticut calls this “death by a thousand cuts”.
Third, insects pose a less familiar sort of environmental problem, that of dwindling abundance. Biologists often think of biodiversity loss in terms of extinctions, especially of top-of-the-pile predators. But when a species is abundant, ecosystems come to depend on profusion, and a decline short of extinction can disrupt their workings profoundly. As Dr Memmott’s experiment showed, the impact may be delayed but it will occur eventually.
One of the unnerving possibilities of insect decline is that it may have been going on decades before long-term monitoring started. After all, modern agricultural practices were well under way in the 1920s. This in turn could mean that the decline documented in Europe and America is even greater than it seems. That does not mean a global insect collapse is imminent, but the data do suggest there are good reasons for concern. “In the past three months,” says Dr Lister, “my fear level has gone up. I worry that we might be reaching a point where insect decline becomes irreversible.”