Matthew found this question on the BBC’s Discover Wildlife site; click on the screenshot below to the “answer”, which I reproduce below. The picture is of the green hairstreak (Callophrys rubi), Britain’s only green butterfly.

First, though, this species is not completely green—the color is only on the bottom of the wings, which it folds as shown, perhaps for camouflage. The top of the wings look like this:

Now, onto the question and “answer,” which I find lacking. Here’s Jones’s answer (indented, plain text), with my responses in bold print and flush left:

Most ‘natural’ green insect pigments (in grasshoppers and plant bugs, for example) tend to fade, since they are chemically altered by light, and there is evidence that some are derived from chlorophyll eaten by the insects.

The green of the hairstreak, though, is not a pigment, but a metallic refraction effect caused by submicroscopic parallel grooves on the wing scales, which reflect only green light. Metallic green beetles use a similar mechanism.

In contrast, melanin (the default pigment across most animals) is highly stable, as are yellow and red pigments, which occur widely. There may be an evolutionary mechanism at work here.

Or maybe not! Jones’s answer is that green pigments tend to fade compared to others, so butterflies use them less. And that might be true, but, as Jones also notes, you can achieve a green color by altering the microstructure of the wings, which gives a stable green color. Further, it’s true, as Jones says, that orthopterans like grasshoppers and katydids are quite often green. Why orthopterans and not lepidopterans? Saying that “green pigments tend to fade” is not a good answer unless you show that that fading has particularly bad consequences for butterflies compared to other insects: 

Here, for instance, are the first four rows of images given by Google for “katydid”:

And for “grasshopper”:

If Jones is going to maintain that green pigments fade is part of the answer, he has to explain this difference between insect orders. I suppose he tries to do that in the last bit of his short answer:

Sedentary butterfly (and moth) larvae tend to eat green plants, and being all the same colour – as the caterpillars of many groups are – offers them camouflage.

Many butterflies tend to sit on green plants, too, and would undoubtedly benefit from camouflage. So that consideration doesn’t explain why there are so few green butterflies. 

But the day-flying adults need to combine bright colours (for mate recognition) with muted cryptic undersides (to hide or roost), so in this case green just may not be necessary.

I’m astounded at this bit. Green coloration is a a cryptic color, and so resting with green wings folded up seems a good way to achieve camouflage. As for needing bright colors for mate recognition, well, look at the green hairstreak above. It’s fricking BROWN on top! Further, we’re not sure if butterflies evolved to be brightly colored so they can recognize mates more easily. Some, for instance, are “aposematically” colored to warn predators of their toxicity, while others are mimics of the aposematically colored ones.  And if butterflies are brightly colored to recognize mates, why aren’t katydids or grasshoppers? 

The correct answer about why so few butterflies are green is this: “We don’t know.” Jones has simply offered a speculative answer, which has problems, and it’s an answer that a lay reader will take away as the truth. Jones may be correct in part, or there may be other factors at play: for example, it may be harder for butterflies to manufacture green pigments than for other insects. But finding a good answer requires either comparative studies (Do butterflies tend to sit more often on brown trees than on green plants?) or experimental studies (Do bright colors really help butterflies recognize mates? Do orthopterans have a pigment synthesis pathway not present in lepidopterans? If so, why are caterpillars—the juvenile stages of butterflies, able to make green?)

What we have here, in fact. is a just-so story. Not only does it not make a lot of sense to me, but appears to give a definitive answer when it doesn’t. It would be much better if Jones admitted up front that the answer isn’t known, and then speculated about some possible answers, suggesting how to test them. That would show people not only that scientists can admit their ignorance, and at the same time suggesting ways to remedy it, displaying the scientific mindset. 

Now what’s your (speculative) answer?