Roses are red, bananas are blue

It’s easy to tell when a banana is ready to eat, because its ripening is handily color-coded. An unripe banana is green, its peel tinted by a surface layer of chlorophyll. As it ripens, this chlorophyll breaks down and allows the underlying yellow to become visible. According to a team of German chemists, chlorophyll breakdown also results in chemical compounds that fluoresce in UV light. The practical upshot of this is that if you stick a ripe banana under a black light, it glows blue.

Apparently it’s not unusual for chlorophyll breakdown products to be fluorescent. In most plants, though, these products are fleeting: their structure is unstable and almost immediately breaks down even further into non-fluorescent compounds. Bananas are different. Due to a slight chemical modification, the normally short-lived chlorophyll intermediate is stabilized and takes a lot long longer to disappear. So enough of it hangs around for it to be detectable in a ripe banana.

I read about this a few years ago, and I wanted to see it for myself. So I bought a set of bananas describing the best ripeness spectrum available in the supermarket, and borrowed a handheld UV light.

The bananas, ready for their big moment.

Banana rave.

This was disappointing. There was no visible difference in glow along the banana spectrum, and without any differences I couldn’t really tell if all of the bananas were glowing or if none were and they were just being illuminated by the UV light. I tried several bananas, both purchased and borrowed*, and no dice. The California bananas were not going to behave like the German bananas.

German bananas. They understand their role in this experiment.

In science, when something doesn’t work they way we expect it to, we feel compelled to figure out why by coming up with hypotheses and continuing our experiments. A hypothesis is pretty much just an educated guess that can be tested. Often, there’s more than one reasonable possible explanation, in which case we propose alternative hypotheses and tests. This testing part is critical. If it cannot be potentially falsified, it’s not a hypothesis. In the case of the bananas, my scientific thought process went like this:

Aw, shit. I really wanted blue bananas.

H1: The UV light I used was the wrong wavelength to excite the blue chlorophyll by-products.

Test: Try multiple UV wavelengths.

Result: Using the three wavelengths I could scrounge, the bananas remained stubbornly identical.

It’s always possible that there was some other issue with my UV light, since to be really systematic about it I would have tested every possible UV wavelength. But I went back to the paper, and these compounds, Mc-FCC-56 and related chemicals, glow more-or-less intensely under a range of wavelengths. And, in fact, one of the three UV lights I was able to borrow gave off 366 nm UV light—the precise wavelength used by the Germans to take the very awesome blue banana picture above.

Returning to the paper suggested an alternative hypothesis, which is that my bananas were inadequate. Apparently Mc-FCC-56 and its co-conspirators are present in a nice bell-shaped curve around optimal banana ripeness, like so:

Greatest bar graph ever produced in Excel. This is directly out of the paper with absolutely no image manipulation.

So, there should be some blue glow in all but the greenest and blackest bananas. Which leads me to an alternative hypothesis:

H2: My banana spectrum was inadequate. All of my bananas were glowing, since none were sufficiently under- or overripe to produce negligable quantities of luminescent chlorophyll breakdown products.

Test: Try again with a more extreme banana spectrum.

Testing of H2 was on hold for a few months, until I ran across a sufficiently wide banana spectrum. An anonymous patron of science donated two very black and squishy bananas to the cause (I found them on my desk one day), so I stuck them in the freezer until they were needed to hold up their end of the spectrum. And last week, Kevin presented me with some exceptionally green exotic bananas from the Indian market, along with their riper counterparts. I picked up a few garden-variety bananas from the local supermarket and it was time for an experiment.

The banana equivalent of lining up by height in kindergarten.

The banana equivalent of lining up by the keg in college.

This looked really promising, but I wasn’t quite convinced. Maybe this result was just an artifact of how I set up the experiment. I needed to test one more hypothesis before I could be comfortable declaring the Germans right.

H3: The UV light was most intense directly under the center of the light, causing the bananas in the middle to glow fiercely regardless of ripeness.

Test: Rearrange the bananas.

By variety this time: two plantains, two Saba bananas, and five Cavendish bananas**.

Case closed. Time for banana bread.

BOOYA! Blue bananas. The German chemists and their spectacular Excel graph are vindicated.

*For reference, “Is that your banana on the conference room table? Could I borrow it for a few minutes? I’ll bring it right back.” is surprisingly successful.

**Cavendish are ordinary supermarket bananas. Like most commercial bananas, they have three sets of chromosomes and are incapable of having sex. (Yes, most plants have sex lives.) I’ll probably come back to this in a later post.

Thanks to Chris Patton for the use of UV lights and help with banana photography.

Reference:

Moser S, Müller T, Ebert M-O, Jockusch S, Turro N, Kräutler B. (2008) Blue luminescence of ripening bananas. Angewandte Chemie International Edition 47: 8954-8957

Figure credits:

German banana photo and graph from Moser et al. 2008 (citation above)

All other banana photos by me, licensed under Creative Commons. Feel free to use them non-commercially as long as you attribute them to me and link back to ScienceFare.org.