Each one is a single insect, alone on white paper – and the Latin name tells you which.

These are my watercolour specimens: original-watercolour insect studies in the old language of natural history, where the subject is set down on bare paper with nothing around it and nowhere to hide. Some are anatomically faithful. Some I invented outright. The binomial underneath is an honest contract – it tells you whether I kept the creature real, named it for a feeling, or made it up.

Apis fratris is the kept-real one, and the apex – a male Buckfast bee, anatomically exact, painted from life over four days among my own flowers, in honour of the bee and Brother Adam, the monk who bred it. Neuroptera spectralis is real too, but named for how it struck me – a lacewing so faint it seems barely there, laid down in the thinnest washes I could manage. Phyllium imaginarium is the invented one, and the name admits it – a real leaf insect given a wing the living ones don't have. A beautiful forgery, signed as one.

All three came out of the same garden as Green Lush, where they lived among the leaves before I ever lifted them onto the bare paper. A photograph could catch any of them in a moment. These took the opposite – a hand choosing every hair, every wing-vein, every grain of pollen by eye, the original of each no bigger than twelve or thirteen centimetres. That is what you are really looking at. Not the insect, but the attention.

Real, named for a feeling, or invented – each one tells you the truth about itself.