Loops and Whorls

Whether it's in an elementary statistics class or, apparently, graduate-level regression analysis, it seems to be a fairly common early point to go through different kinds of predictors and the most common examples of each one. The rundown in my last stats class went something like: 

Continuous variables (eg, height, weight)

Binary variables (eg. dead/alive, gender)

Categorical variables (eg. treatment group)

Ordered categories (eg. Likert scales)

and so on

Fairly standard list, basic examples you'd see, right? Nothing to raise an eyebrow at?

Obviously I spoiled the punchline in my title, but one of these things is a pretty terrible example for its category. Unlike lots of things that only have two meaningful states you might want to think about in a statistical analysis (dead or alive, heads or tails, success or failure in an experiment), to represent with a dummy variable that can equal 1 or 0, gender is not a binary.

(And you can't weasel around me for the rest of this post by arguing that you really mean sex and not gender, having read just enough to know there's a difference, at least if you're taking about people. Sex has some elements to it that are binary, but the whole picture doesn't break down so neatly. And even if you're not talking about humans, it's probably actually way more complicated than you think.)

Yes, some people identify as men and some identify as women, and nothing else. Those two aren't the only options, and lumping gender in with real binary variables excludes all those people. When stats professors and authors of textbooks ignore that reality, they reinforce discrimination and exclusion of non-binary people (including their non-binary students). 

If that wasn't reason enough, it's also not really accurate for how data is collected in many fields. Almost every survey or form I've filled out recently has given more than two options for gender or sex. Sometimes it's just three or four ("Male", "Female", "Other", and sometimes "Prefer not to disclose"). Facebook lets users chose from 56 possible options (which sounds like a goldmine for a statistician to me). 

Luckily, unlike many problems with inclusion and accessibility in academia, this seems like a pretty easy problem to fix. Just pick another example for your binary variable. (I like dead/alive, as it leaves room for many an ex-parrot joke or a reference to one of my favorite Wikipedia categories.) Introduce dummy variables with a treatment assignment where there's only two groups. 

Science!

This is what my lab bench looks like. The brownish petri dishes are leaves in sodium hydroxide, which dissolves out all the pigments. The pink petri dishes are leaves that have been stained with a chemical called Saffranin-O and are soaking in alcohol to remove the stain from places it's not supposed to be. I spent today mounting stained leaves between transparency film so I can photograph them under a microscope later, and then map out the patterns of the veins.

The leaves here are from Asclepia tuberosa (a prairie plant related to milkweed). It has pretty orange flowers and you probably see it a lot if you live in the midwest. 

The double pendulum is a neat application of chaos theory - processes that are so sensitive to initial starting values that they look random - but if you put in the exact same initial conditions, you'd get the same path every time. This can happen in biological populations - small changes in starting conditions can create wildly different results. 

Can you talk about plants that can move under their own power? Like venus flytraps and things like that? Basically, how unrealistic are triffids and ents? :)

Sorry this took so long! I was out of town with my lab all weekend, and this took a bit of research on my part. Plant movement is super cool and really complicated. 

There's a couple of ways plants can move - for me they fall into a couple of big categories. They can change how big or small their cells are - this is controlled by hormones, and is what makes a plant bend toward or away from the light - like how sunflowers rotate around to keep their developing flowers warm. They can change where they grow (like how climbing plants climb or how the walking palm moves its stilt roots. 

The group you're asking about - venus fly traps and other plants with faster movement - move by changing their turgor pressure - the amount of water stored in the cells that can keep them stiff or floppy. 

A venus fly trap has little hairs around its traps. When these hairs get touched (by a fly or a person messing around), it triggers a chemical signalling pathway that moves water around so that the trap closes. 

A similar process happens in sensitive plants (Mimosa pudica) - when they're touched, a chemical process triggers them to move water from the leaves into the stems, so the leaves drop down and look wilted. These kinds of movements can happen quite fast. 

Here's a recent technical paper (that should be open access) looking specifically at the biomechanics in venus fly traps. 

Where does that leave ents and triffids? You could think of ents as much, much faster versions of walking palms - setting down new roots and picking up old ones in order to move forward. (My wife thinks ents are more likely to be a fungus, like a giant intelligent slime mold, which can move around by reforming its stalk. I think walking palms still work!)

So many things! I think the coolest thing, though, is the incredible diversity. Like, you take the same basic bits (roots, stem, leaves that have photosynthesis, and maybe flowers) and you get huge, huge ranges of diversity. 

There are teeny, tiny plants like the duckweeds that pretty much only have a little root, two little leaves, and even little flowers, that float on air trapped in the hairs of the leaves: 

(Wolffia arrhiza via Wikipedia)

On the other end of the scale, there's the mountain ash, the tallest flowering plant:

(Eucalyptis regnans via Wikipedia)

I can think off the top of my head of three totally different kinds of carnivorous plants, that live in really nutrient poor places like bogs and use My favorite are the pitcher plants: 

(Sarracenia sp. via Wikipedia. There are two totally unrelated families of pitcher plants (North American and Australian) - these are the North American ones!)

There are plants that have almost entirely lost their photosynthesizing parts and take most of their nutrients from the roots of other plants: 

(Orobanche hederae via Wikipedia)

And that's not even scratching plants that vine, plants that grow in horribly in hospitable places, succulents full of water and special acids to keep from drying out, plants that grow on top of other plants, and the thousands and thousands of different kinds of flowers. 

I think the single most interesting thing to me is the diversity in patterns and shapes there are across leaves? The thing that made me want to study leaf venation were the melastomes: 

(Miconia calvescens via Stephan Buchan, Flickr)

It makes them really easy to identify (which was helpful for me at the time, because I knew very little tropical botany the summer I did this project, and identifying tropical plants can be really, really hard!), but it's also really unusual, because the veins are joined up at the bases and then at the tips, and the overall structure is unique to the family. There are similar looking, like cinnamon (which tripped me up in a greenhouse once - got all excited then looked at the tag), but none quite the same.

(Cinnamomum varium via Wikipedia)

So I started asking what might this pattern be for? Is it just an artifact of development, or does it give some benefit to the plant? Do other plant groups have venation as similar among close relatives, like melastomes, or does it vary even in closely related species? Can plants change their venation in different kinds of stressful environments? 

Those are the questions I'm hoping to answer in my PhD!