(See Part I here, if you haven't already read it. Also while you're there, check out the comments. Some interesting stuff on Sansevieria's namesake.)
Okay, when we left off before, I was dangling in front of you the promise of understanding why variegated specimens of Sansevieria trifasciata revert to the plain varieties when you try to grow them from leaf sections, but not when grown from division. We'll get there, but you should bring some water with you, and maybe go to the bathroom – it's going to be a long trip, and we won't have time for a lot of stops.
First important concept: many variegated plants are chimeras (ki-MARE-uh). In the broadest use of the word, this means that not all cells in the plant contain identical genetic information. This is not all that uncommon, actually, in people, animals, or plants: if you are a woman who has had a baby, there are probably some of your child's cells running around in your body right now, however long ago it's been since you gave birth. Such cells have been implicated as maybe being involved in disorders like scleroderma, but at the same time may also protect against some kinds of cancer (ref 1, ref 2). The same also applies if you've ever had a blood transfusion: you're not guaranteed to have some of your donor's cells still roaming your body, but (especially with large or repeated transfusions) you may still have a small percentage of white blood cells which are not, genetically, your own. A few people out there initially develop in the womb as a fraternal twin, and "absorbed" their twin early in development, with the result that they contain two types of cells: those from one twin and those from the other. This isn't thought to be all that common, but of course how would you know? Previous cases have come to light as the result of paternity tests (in one notable case, a paternity test revealed that the "mother" of a child was not the mother of the child, which you can bet caused a lot of anguish for all involved) or in tests for potential tissue donors or recipients. I predict that we'll find out this is a lot more common than people think, whenever genetic tests are cheap enough to be routine.
Oh, and – obviously, if you absorb your identical twin at some point, no one will ever be able to tell by genetic testing, because, duh, identical. So there's no telling how often that happens. (We know it happens at least once in a while, because sometimes, er, the twin is only incompletely absorbed, and people discover tumors later in life that have teeth and fingernails in them, or weird shit like that. I'm not kidding. It happens.)
It gets waaaaaay weirder than that (Anglerfish! Mosaics! Hermaphroditism!), but I'm just trying to establish the concept here, so let's move on.
Botanical chimeras are easier to believe. Generally, what happens is that a cell in the developing tip of a plant mutates in such a way that it can no longer produce chlorophyll or some other plant pigment, or it still produces it but at a much slower rate, and then as the cells continue to divide and reproduce, this leads to some large sections of the plant which are a different color than the rest of the plant.1 There are three basic layers of cells in the developing growing tip, which are whimsically called, from outside to inside: I, II, and III (those wacky biologists!). Each of these give rise to different parts of the developing leaf. Mutations in I affect the very outermost edge of the leaf, like, say, the parts of 'Black Gold' which are yellow in this picture:
Mutations in III affect the center of the leaf, and mutations in II affect a section in between the two. In the three simplistic drawings of leaves below, which are loosely modeled after Chlorophytum comosum 'Vittatum,' Dracaena deremensis 'Warneckei,' Sansevieria trifasciata 'Laurentii,' and Dracaena deremensis 'Lemon-Lime,' we can see examples where each layer of cells, in turn, lacks chlorophyll-producing cells. The Chlorophytum-modeled leaf lacks chlorophyll in layer III, the Dracaena 'Warneckei'-modeled leaf lacks it in layer II, and the Sansevieria-modeled leaf lacks it in layer I (though it still expresses a yellow pigment of some kind). In the Dracaena 'Lemon-Lime' leaf, we have a double mutation: layer II expresses no pigment, and layer I expresses only a little chlorophyll, along with a yellow pigment of some kind.
But wait! It gets so much better, because this explains all kinds of crazy stuff. For example, we can now gain a little bit deeper insight into what was going on with the Dracaena deremensis 'Warneckei' I reported about a long time ago: obviously, when the stem was cut and the plant had to resprout, a mutated cell in layer III that couldn't make chlorophyll was allowed to spread and be expressed in that particular growing tip, along with the original 'Warneckei' mutation that produced a chlorophyll-less layer II. The result: bigger white patch in the middle.
You can play this game at home: which layers have to be mutated to form the reverse-variegated Chlorophytum comosum? What about turning Dracaena fragrans into D. fragrans 'Massangeana?' Does Tradescantia zebrina make more sense now, in some odd way?
Now for the disclaimers.
We need to note here that expression or lack of expression is not always as clean and precise as I've made it out to be. Dracaena sanderiana, for example, tends to have a single band of pure white or cream color along the outside edge of the leaf (the layer I cells), but then lighter and darker stripes through the rest of the leaf. My guess is that the lighter stripes are spots where cells descended from both layers I and II are laying on top of one another, so the resulting color is lighter than II but darker than I, and then the layer-III cells in the middle of the leaves are normal. But I could be wrong.
Also, all of the plants that I have discussed so far are monocots. Monocots typically have one initial seed leaf (which is where the name comes from: the opposing group, the dicots, have two: mono = "one" and di = "two"), flower parts occur in multiples of three (lilies and tulips being particularly vivid six-petaled examples of this), and leaf veins are typically parallel from base to tip, instead of branching out from a midvein like in dicots. The above discussion of variegation mostly applies to both monocots and dicots, though with dicots, layer I usually only forms the colorless epidermis of the leaf, so it would make no visible difference if those cells could produce chlorophyll or not, because you don't normally see any color in them anyway. Also it's easier to tell what's going on when all of the color variation happens along parallel straight lines, as it does with monocots. Dicots tend to have uneven boundaries between layers II and III, and so you get variegation that's either dark with light edges, or light with dark edges,
depending on which layer was affected by the original mutation.2
Wheeeee! Are we having fun yet?
So let's get to the original question, then: why, with all this crazy layering going on, can you keep the variegation in a Sansevieria trifasciata 'Laurentii' when you're propagating by division of a rhizome, but not when you're propagating from a leaf section? After all, there are yellow and green cells to begin with in both cases: why can't they just stay together?
The reason is,
A plantlet from division retains the same overall organization as the original growing tip: all three layers of its cells will, at some point, resume growing and dividing as though nothing had happened, and the descendants of green cells will be green, and the descendants of yellow cells will be yellow. There's nothing, of course, to completely rule out a new mutation springing up at some point down the road, but mostly it's going to keep going like it's been going already.
A plantlet from a leaf cutting has all kinds of obstacles to get over at first, though. Like, for one thing, it has to organize a new growing tip somewhere. First step in that process? Turning a cell into the new growing tip.
Now, I don't know how the cell decides which cell to start with, but there are really only three things that could happen from this point.
1) A layer I cell, in the yellow leaf edge, is selected as the new start point. All three layers in the new plantlet are then going to be cloned from this starting cell, and, consequently, all three layers in the new plantlet are going to share the mutation which is preventing the cells from producing chlorophyll. The new plant will have trouble making its own energy, and will be limited by the amount that the original leaf cutting can produce. When the original leaf cutting dies of old age, or when the original cutting's ability to produce food is outpaced by the demands from the new plantlet, which has got to happen sooner or later, the whole shebang dies, and it all ends with no new plantlets.
2) A layer II or III cell, which does not have the mutation preventing chlorophyll production, is selected as the new start point. All subsequent cells, in all layers, are descended from this one and are therefore capable of producing chlorophyll. This means that all three layers of cells will be green, and consequently the whole plant is green.
3) The new shoot begins with any cell, and at just the right moment, a mutation happens. A yellow cell gets the right reverse mutation to enable chlorophyll production, and we have a new cultivar, or a plant that would have been all-green gets chlorophyll production knocked out in one of the founding cells and some part of the new plant gets to be yellow, possibly resulting in a new cultivar. Either situation is a huge long shot: if you did nothing but plant Sansevieria leaf sections for the rest of your life, it's possible that you might see such a mutation occur . . . aaaaaand it's possible that you wouldn't. Or, I guess, technically one of your assistants would be the one to see it, not you, because you're not doing anything but planting leaf sections for the rest of your life. Like I said.
So that's why the variegation doesn't continue from a leaf section: one way the whole thing dies, one way it's all-green, and once every few million times, there's a new mutation that saves the whole thing but that's so rare that you'd be foolish to count on it. A similar mechanism is at work, I believe, when Hoya carnosa 'Krimson Queen' produces an a vine consisting of nothing but white leaves, though I claim no particular understanding of how Hoya plants decide to branch.
And so. Gosh, this has been a fun profile to write. I went off-track a number of times, I piled on the random details, the word count is outrageous compared to previous profiles, but I loved that moment where I understood how this type of variegation works, where it clicked, and I started putting together, oh, this is what was going on with the 'Warneckei' sport, and this is why you don't see leaves with this kind of coloration, and oh my god, why have I never noticed how similar these all are to one another before?
The structure beneath the beauty, the mechanics of it, has its own sort of beauty. I've really never understood the people who complain about how, you know, when you find out that a rainbow is just a bunch of refracted light it ruins the magic of the experience. It takes the mystery out of it, or whatever. I've always thought it made it better to have that kind of insight into the experience, to see the pieces fit together, to see the connections. Not that there's not also something to be said for just letting the pure sensory experience of a new thing (seeing Saturn in a telescope, meeting a new houseplant for the first time, hearing a piece of music that immediately becomes your favorite song ever, learning a new word) happen on its own, but if you never connect it to anything else you know, if it never does anything to flesh out your understanding of the world in general, then you've missed an opportunity to be impressed a second time by the same thing.
This is the sort of thing I love about science: it's not just that it has practical, technological applications, though those are often nice. What I like is that if you dig down deep enough, and see how it all works, there are almost always connections to be made to something you'd never have thought was related: if you're doing it right, you can almost always make familiar things strange, and vice-versa. It's the closest I ever get to feeling like a kid: that whoa moment when it clicks for me that the Earth's axis is tilted, or those dinosaur bones used to be an animal, or matter is made of atoms, or whatever.
This post is not the be-all and end-all of variegation. It works well for certain kinds of variegated plants, but you may have noticed that I didn't even try to explain where the horizontal bands of darker green come from, on Sansevieria trifasciata, so it's obviously incomplete.
It also doesn't work as an explanation for a lot of variegated plants. We need something else for Dieffenbachia, for example. And this isn't helpful with Aspidistra lurida 'Milky Way,' either, or many, many other variegates. But, you know, baby steps.
Photo credits: all are my own except for the Dracaena sanderiana picture, which was anonymously donated.
References: A lot of this post came from this site, which is actually sort of obnoxiously overcomplicated and jargony for my money, but on the plus side, it's overcomplicated and jargony enough that it sounds authoritative to me. A related post is here. The site which got me started on this whole thing, by throwing the previously-unheard of term "periclinal chimera" at me, leading me to google the term and then try to understand it, is here.
1 There are other types of botanical chimerism. Grafted plants, for example, are chimeric, and these are even more extreme than the ones we're talking about: they don't even necessarily have to be from the same species. We kind of covered this in Hylocereus undatus.
2 The uneven-edges thing is, as best as I can figure out, because some dicots have leaves which are arranged like a stack of pancakes. Each pancake in the stack has the same I-II-III organization we're talking about, but the pancakes are not identical. So, any given spot on the leaf may have no layers of pigment, one layer, two layers, etc., which results in blotches of color which get generally darker or lighter as you move from the center of the leaf to the outside, but don't do so in any kind of smooth way. It may help to look back at some of the pictures in this post, particularly Ficus benjamina 'Spearmint,' which appears to have this kind of structure. Citrus limon 'Pink Lemonade,' Codiaeum variegatum 'Andrew,' and Peperomia clusiifolia all also have leaves that look like this. They actually all look more related to one another than they do to their actual relatives.