Something's bugging me about today's reporting of the imminent extinction of insects.
Don't think I'm not worried. I really am. I love insects. I spent six years of my life studying them. Like many others, I am deeply concerned about the global declines in some of our best-loved and most ecologically important bugs.
But making claims about the diversity or abundance of insects which aren't necessarily supported by the facts risks undermining the power of the case desperately being made by scientists that more must be done to protect them.
The research doesn't present any new data. It's something called a meta-analysis that pulls together other studies. The researchers searched only for papers that reported a decline in insects. In so doing they excluded studies that found insect numbers stable or growing.
In science, that's called bias. It's totally acceptable, when, as in this paper, the researchers acknowledge it. But it's the kind of detail that's often omitted by journalists.
One of the most alarming studies included in the paper is one published in 2018 that illustrates the decline of insects in an unspoiled forest environment in Puerto Rico.
Comparing data gathered in 1976 with that from 2012, the study found declines of between 10% and 60% in insect biomass - that's the sheer weight of insects. It's a mind boggling decline and shows something deeply concerning is going on in this ecosystem.
But it's far from perfect. It's a study based on just two data points.
And one thing ecologists already know is that insect populations show huge amounts of natural variability year-to-year.
Take insects in the UK. This recent study, published in the Journal of Insect Conservation, showed so much variation year-to-year in butterfly and moth numbers that it was hard to know which were really threatened with extinction.
And other studies have shown that depending on where you look in the UK, declines in one area may not be found in others.
One 2009 study looked at total abundance of insects in four locations in southern Britain comparing 1973 with 2002.
It found a marked decline in a sampling site in Hereford, but none in three others. It emerged that the decline in Hereford was probably explained by the loss of one species of fly, its decline most likely explained by a change in farming practices. A really important detail for conservationists, which could be very relevant in that locale but not in others.
Then look at numbers of moths. I am a particular fan of moths in the UK. The latest data shows significant declines in species like the garden tiger, once common and one of my favourite moths as a child.
But increasing in abundance is its warm-weather cousin, the Jersey tiger, which I now see regularly even in my gloomy north London garden.
Undoubtedly the Jersey tiger has been driven here by climate change, which is pushing the garden tiger north, potentially - depressingly - to oblivion.
Overall, the picture from the time series found here is really mixed. Some moth species are in decline, others are stable, others increasing.
For example, the dingy footman is doing great since huge improvements in UK air quality have allowed the lichen on which it feeds to thrive.
In summary, headlines along the lines of "insects could be extinct by the end of the century" are plain wrong.
And overblowing the headlines risks undermining the very urgent need to change the way we grow food and protect landscapes.
Already, the Global Warming Policy Foundation, a think-tank devoted to undermining the science of global warming, has seized on the Puerto Rico study and criticised its claim that increasing temperatures are the driver behind insect losses there.
There is undoubtedly a global decline in insect numbers on the scale that justifies calling it a sixth extinction, but we urgently need more data on the status of insect populations around the globe.
When you're as small as an insect, details are important.