A couple weeks ago right here on this very channel, we told you about how aluminum changes this molecule to this one and how that should change the color of hydrangea sepals from red to blue.

But when we tried it in our lab (aka our studio), uhh yeah it didn't go so well.

But that's ok.

In chemistry -- and science in general -- failure is at the heart of the process.

Science is a cycle - you try something, it fails, you change something and try again, it usually fails again, you notice something you hadn't before and then you make another change and maybe it fails again but in a different way this time so you-- You get where I'm going with this.

So in that spirit let's figure out where we might have gone wrong.

Remember from our last hydrangea episode that aluminum ions change the color of this pigment in hydrangea sepals from red to blue.

By incubating these cut hydrangeas in an aluminum solution, we hoped to change the color of their sepals.

And look, it sort of worked, but there are a lot of things we could have done better.

First of all, we ran this experiment... here - - in our windowless basement studio under hot lights.

Sean, can you cut out for a second?

Does this look like the best place for flowers?

No.

It does not.

Oh, and that blue backdrop when you're looking for a blue color change?

Not my best idea.

Secondly, we only used two of each type of hydrangea.

Ideally, you'd do a bunch of math to figure out how many flowers you would need to make sure the color change didn't just happen by random chance, and you'd use... that many flowers.

Then you'd know how confident you could be in your results.

We didn't do that, for two reasons: one, the flowers were pretty expensive and we're only a wee youtube channel, not Michael Bay.

Second, we were trying to see if this would work at all in a small scale pilot experiment, rather than trying to go big right out of the gate.

If failure is inevitable (which it is) you wanna fail small and learn things before you try big.

We also asked the scientist whose work we adapted for this video, [visual: Dr. Henry Schreiber, Professor Emeritus at the Virginia Military Institute] about the composition of our buffer.

We used 0.05 molar aluminum and 0.05 molar citric acid and citrate for our first attempt.

That doesn't sound like much, but Dr. Schreiber told us that even that amount of aluminum risks poisoning the plants.

So we tried again.

This time we set up the experiment in a place with real sunlight, tried lower aluminum and citrate concentrations, and used three flowers instead of two (I know, 30 would have been better, but we're still in the pilot experiment phase here).

And look, it actually worked!

Mmm, pretty.

Now I just want to be clear: this result is not statistically significant.

But it does give us some guidance on how we might set up a larger experiment with 50 or 100 flowers.

Why are we telling you all this?

You've already heard "if at first you don't succeed, try try again" a million times.

And it's true.

But you don't just try the exact same thing over and over.

You guess at what went wrong, and you make changes, and you try again, and you repeat that cycle until something starts to work.

In this case, we were trying to replicate a published experiment, which means we had a specific thing we were going for.

But that's not usually the case.

If you're testing a brand new hypothesis, you might notice something interesting and make adjustments in that direction.

And if an experiment fails, it might still tell you something.

Negative data isn't a crime.

Dr. Schreiber told us that, while his group has managed to turn red hydrangeas blue using this method, they have not yet managed the reverse.

So he encouraged us, and you, to experiment further.

His group also has all kinds of cool tricks to turn blossoms green, or yellow, or multiple colors at once.

So thanks for coming along this journey of failure and then success with us.