I saw this funny article titled

Plants do sums to get through the night. It claims:

The John Innes Centre scientists show that to adjust their starch
consumption so precisely they must be performing a mathematical
calculation – arithmetic division.

I probably don't need to tell you that this claim is false. The author of the article has confused a mathematical

*model* for its

*subject*. The two are not the same!

After all, rocks don't compute parabole when tossed in the air. We use pen and paper to compute rock trajectories, and starch consumption in plants. But rocks and plants just do what they do, with no need for

**computation**.

That said, this claim

*could* have been true. A plant

**could** start with two molecular concentrations, representing the numerator and denominator, then end up with one concentration of molecules, equal to the quotient. Such a biological system would be, in effect, an analogue calculator; and that would be awesome!

Sadly, this is not the case. The paper cited by this article, titled

Arabidopsis plants perform arithmetic division to prevent starvation at night, appears to have been written by plants, for plants. It hardly deserves to be called a paper. It is disorganized, overcomplicated, and does not support its conclusions.

The paper was published in some open-access journal called

eLife, which has no established track record. Well, now it has a track record of publishing garbage...

In the paper, there are three models, which each use hypothetical molecules

**S** and

**T**. These represent the starch storage and estimated time until dawn, respectively. Basically,

**T** inhibits

**S**, while

**S** breaks down starch, so that the rate of starch consumption is equal to

**S** divided by

**T**.

Even if this is how plant metabolism worked, this wouldn't be called

*calculation*. It would be called

**behavior**. This model lacks a third molecule to encode the

*quotient*. Instead, the quotient is inherent in the rate of starch degradation. It's a controlled enzymatic reaction. This is how all inhibitors work, in all living things. It's not a calculation. It's not arithmetic. It's not even novel biochemistry.

Even worse, there is one jaw-dropping flaw in the paper. The reported evidence does not support any of the over-complicated models.

Essentially, the models attempt to explain why the starch concentration is always (so close to) zero at dawn. In other words, why is the right hand side of this graph a strait line? How does the plant know what slope to make that line?

You don't need the

**T** molecule to accomplish this. The smooth slope can already be accomplished by their definition of

**S**. Specifically, that

**S** is proportional to the quantity of starch stored during the proceeding period of light.

More intense light (or more daytime) leads to more starch, and therefore more

**S.** More

**S** leads to speedier consumption of starch, during the next dark period. In other words, the starch degradation rate is encoded in the quantity of

**S**, by the action of starch generation during daylight.

See? No complicated models (or division) necessary.

Sure, you can throw several variables into a pot, stir them up, and boil on best-fit for 30 pages.

You'll get a tiny-ass p-value.

But that's not science.