This question is based on a scenario on DeviantArt. The presented image is here:
The description is as follows:
By the start of the Cryocene, bamboo has become one of Serina's most successful plant groups and makes up entire forest communities in some subtropical equatorial habitats. Fast-growing, hardy and adaptable already in its ancestral form, it has begun to evolve in several new directions by fifty million years PE.
Illustrated above are examples of the growth habits of four groups of bamboo present on Serina by this time, labeled A, B, C, and D. A is the ancestral plant from which B, C, and D descend and grows individual bamboo canes directly from a large underground rhizome, which can spread almost indefinitely in any direction via underground stems. This ability to quickly spread across Serina in its early history initially gave bamboo an advantage over most other plant groups, though this rate of spread was only possible in very fertile and wet climates. Bamboo forests of this growth habit typically grew and spread for a lengthy period of time, anywhere from a few decades to over one hundred years, before flowering, dying, and slowly regenerating over the next ten to twenty years from seedlings - a habit retained in many Serinan species today, including the desert bamboo, which takes this behavior to its extreme. The individual canes of these colonies are short-lived, lasting only a few years, and are typically determinate in growth; they can only grow to a certain height before they stop developing. To keep the plant growing, new canes are constantly produced from the root system to replace those which die. The roots are shallow, without deep anchoring tap roots.
Over time, however, mutant populations of bamboo would gradually arise with less aggressive rates of spreading, but which concentrated more of their energy on producing fewer but larger, taller, and longer-lived individual stems (B). Their growth rate was slower, but it was indeterminate - given time, they could grow larger than their competitors, with canes that last much longer. This was the first step to reaching a more tree-like shape and growth pattern and was particularly beneficial in drier ecosystems or those where the soil was fairly poor in nutrients. By concentrating on just a few trunks and increasing its size more gradually, a plant could better grow in less ideal environments, including in the shadow of more vigorous type-A bamboo colonies. Biding their time on what little nutrition they can get that the more aggressive bamboo doesn't suck up around them, they can eventually grow beyond the canopy and slowly shade out their competition. B-type bamboo still produce canes from the underground root but far fewer than the ancestral bamboo and much closer to the main trunk, resulting in a small clump of trees together rather than a sprawling colony. In the natural pattern of forest succession, B-type bamboo forest tend to gradually replace A-type forests over the course of about a century, for though type-A bamboo is quick to colonize open land, by the time it flowers, the ground will be too shaded by maturing type-B bamboo for most of their seedlings to survive.
Next is the type-C growth pattern. An extension of type-B, this bamboo produces only one trunk and few or no secondary canes. Its rhizome is reduced, and it instead produces a single deep taproot to probe the soil for water and deeply buried minerals. Its growth is perhaps the slowest of all, but it is the only type which can reliably take root and survive on dry exposed hillsides or areas with little soil. Its singular trunk can become several meters wide, though remains hollow and never becomes woody like a proper tree; the only living tissue within it occurs in a ring-shaped cross section. This makes a type-C bamboo "tree" lighter than a woody plant like a sunflower tree, which makes it potentially more vulnerable to high wind or weather when young but considerably less likely to topple over under their own weight at very large sizes. The first representative of a "type-C" bamboo was the monovitus tree of the middle Tempuscenic. Like this ancestor, many modern examples still flower only once, often at a very advanced age, in such an extravagant show that they use up all their energy and die back to give their offspring a chance to grow in their place. Others however have become truly indeterminate growers, growing throughout their life and flowering less abundantly but much more frequently, even every year, producing fewer seeds at a time but surely many more over the course of a longer lifetime. In the case of trees that don't die back after flowering, methods to ensure their seeds get far enough away that they don't compete with their parent have had to evolve, and the descendants of the monovitus tree use their symbiotic ants. Now not only do these ants live within their tissues, prune away competitors, and clean them of insect pests, but they disperse their seeds for them across wide reaches of land, each new queen taking a seed with it when it flies away to start its own new colony. It finds a suitable sunny patch of soil and begins to dig a burrow for its new colony, taking the seed underground. When it sprouts they instinctively guard it and keep its patch of soil weeded, and when it is big enough, it begins to grow swollen bolls at its nodes which the ants then move into. A single colony will stay with a tree for up to twenty years - the maximum lifespan of certain queen ants, but eventually a queen will die and the colony will collapse. Fortunately for the tree, however, it takes very little time for another young queen to move in and start the cycle again - indeed, so valuable is a mature tree as a nest site that dozens of queens will fight for the right to use it. Several may even start nesting at different heights within its trunk, but eventually the colonies will meet and battle. That with the largest army always wins. Because Type-C bamboo doesn't reproduce asexually at all and only spreads via seeds, it has a higher rate of mutation than other bamboos, giving it improved resistance to pathogens and a better ability to adapt to changing environmental conditions than bamboos that only rarely reproduce sexually.
Type C bamboo has taken the ant symbiosis to its greatest pinnacle so far by the start of the Cryocene, being entirely dependent on certain ant species to reproduce, but representatives from all growth types provide ants with shelter and sometimes food in exchange for keeping them free of pests and competition. Type-A bamboo is the least dependent on them, however, for it grows the most aggressively already that it can compete against most other plants well enough by itself that its symbiotes are simply extra help. The ancestral type-A bamboo's rapid growth may in fact have been a contributing factor in later slower-growing species adopting the help of ants in the first place, for they could counteract their less vigorous biology if their helpers pruned away other plants from around them before they were overtaken. Eventually, however, the more vigorous bamboo would begin to attract ants of their own specifically for the purpose of putting the odds back in their favor. Today hundreds of bamboo species exist with many more symbiotic ant species, all fighting a largely hidden battle for turf against each other, trimming away the branches of their neighbors if any two plants get too close to each other. The Serinan bamboo forest exhibits universal crown shyness to a particularly obvious degree, because anywhere that the branches of two different trees with different ant colonies touch is soon pruned away by the ants living on one or the other.
~~~
Type-D bamboo is the result of an entirely different evolutionary path. A mutation has occurred in these plants that results in two very different methods of growth in a single plant. Canes rise up from the rhizome and expand their mass from the top, adding whorls of leaves and gaining in height by building upwards like almost every other plant, but this type of bamboo also grows from the bottom up, random sections of the rhizomes gradually lengthening underground until they rise out of the soil, in some places as high as ten meters, raising the bases of the plant's canes up with it, from under the ground to quite a ways above the soil. In this way, Type-D bamboo - which is otherwise very similar in its manner of growth to type-A bamboo - is able to reach much greater heights, as it becomes supported by multiple woody trunks in its old age and towers above its competition that doesn't exhibit the mutation. Of all the bamboos so far to evolve, only this one could be called a tree in the strictest of sense, for with its elevated rhizome, it could be said to have a proper solid trunk. Canes still die back frequently and re-sprout from the trunk, but with the rhizome now high above the forest floor, they don't need to struggle up through the shaded understory and can grow in the sun from the start.
Nonetheless, though type-D bamboo is more competitive in dense forest environments than type-A bamboo, it's much less suited to colonizing open expanses of land, for when its rhizome rises out of the soil, the canes that grow from it lose their rooting and become wobbly and vulnerable to being broken off. In the forest they can lean on their neighbors for support but in exposed areas they typically blow over in the wind once the rhizome breaks out of the soil and its stabilizing presence on the shoots. The two bamboo types have thus begun to niche partition, Type-D thriving in forests and type-A in open environments. Because Type-D cannot survive except in a forest environment of mostly type-A bamboo to support it, however, its long-term survival is linked with the other growth type, meaning that in order to continue to survive, it must remain competitive, but not to the point of totally outdoing its competitors, for in an ironic twist, it needs them to survive itself.
Type-D, like type-A, eventually flowers and dies back to the ground as well, and is succeeded just like it by types B and C in a healthy forest community.
This is a situation set in the future, but could this have happened in the past, too? Say, in an extinction event five million years ago that killed off most if not all of the eudicots (angiosperms that have two cotyledons in the seed, leaves with a network of veins radiating from a central main vein, flower parts in multiples of four or five, and a ring of vascular cambium in the stem)?
