Bamboos are perennial woody plants, but the stems, known as culms in all grasses, generally only live for 5 to 10 years, and need to be replaced. As culms lack bark and thus the secondary thickening that enables palms or trees to increase diameter and height progressively, bamboos grow to their full diameter and usually to full height within one growing season. To attain substantial height in limited time, bamboo culms cannot attempt purely apical growth from a terminal meristem, which would be too slow. Instead they use intercalary growth, the extension of each internode from a basal ring of meristem. Combined simultaneous intercalary growth occurs in the many preformed nodes and internodes condensed into an underground bud and young shoot. This produces the rapid shoot extension, in the order of 25 cm a day, for which large bamboos are famous.
Among all the grasses, the woody bamboos are the best group for illustrating the segmented nature of the various component parts of the grass plant. Each segment constitutes a basic building block, and is composed of a node and internode. The node supports a sheath that surrounds both the internode and the bud that develops at the internode base. The sheath is modified during its development, according to its location on the plant and the physiological gradients that apply during its development. The end product can be bract, culm sheath, foliage leaf, prophyll, lemma, palea, anther, lodicule, or ovary. In some locations it can be greatly reduced to a tiny bract, small tuft of hairs, or thickened ring around the node. Under tissue culture environments, this plasticity of developmental potential becomes very apparent.
Various combinations of terminal and lateral growth in rhizomes, branches and inflorescences leads to an array of different morphological growth forms. This, along with the segmented underlying structure and a high degree of plasticity in growth according to physiological influences, has led to a more critical analysis of growth and morphology in bamboos than in other grasses. This has led to different interpretations of structures such as inflorescences, which in most grasses are considered to be much more distinct from vegetative structures. Such contemplation of the origins of organs, and also more flexible use of terminology does not always please the traditional grass taxonomist, and several of the homologies suggested for particular components of bamboo morphology are speculative and by no means proven or universally accepted.
Several different architectural growth forms have evolved in particular habitats according to climate, soils, other vegetation, browsing of animals and even the occurrence of fire. These adaptations make particular bamboos very successful under the right conditions. It also makes many of them somewhat site-specific, and also unfortunately vulnerable to alterations in their environment.
Fundamentally, the components of the bamboo plant must allow the photosynthesizing leaves to grow into the light, while the roots seek out water and nutrients below ground, all the while competing with other plants for available resources, and recovering after browsing, drought, or fire damage.
To help woody bamboos achieve these objectives, two vegetative characteristics have evolved, which are unique in the gras family. Firstly the culms are seriously woody, which allows them to stretch out or upward further, so that the leaves are further from the rhizomes than in any other grass, and thus closer to sources of light. Secondly the leaf blades are narrowed where they join the sheath, allowing these blades to be presented perpendicular to the direction of incoming light, and rotated if that direction changes. This mechanism can also used to allow blades to hang vertically in winter to prevent accumulation of snow and ice. These two characteristics, along with the extreme infrequency of flowering, which is usually heavy and well synchronized over a large area, suggest that early bamboos evolved in a forest understorey environment of low light intensity, with light winds making cross-pollination difficult, but more successful if flowering is synchronized.
Examples of vegetative architectural strategies:
Extensive, thin, underground rhizome systems spread through soft soils, moist from early spring right through a long, warm summer, producing an open network. From this erect, tall, self-supporting culms with few light branches arise, rapidly establishing a delicate but deep and extensive canopy under bright light levels found under scattered trees or in open areas. The few, light branches are all that is required as they are not heavily predated, being beyond the range of most browsing animals, and can provide a deep many-layered canopy structure of bamboo leaves, oriented in different directions and kept through the winter.
Compact, thick rhizomes remain in a fixed locality, often on a rocky or steep site, growing only a short distance during a short cool growing season at high altitude. From their fixed origin, pendulous culms hang out in all directions, with cascading masses of dense foliage in heavy mats, the blades oriented towards light filtering through a gap in a heavy canopy of trees. The short culms produce many branches with dense leaves, weighing the culms down further to hang out into new light, to absorb as much of the low levels of light as they can, to overshadow competing plants, and the large number of branches means that those eaten by animals can be replaced from basal buds. To avoid snow-laden leaves breaking the heavily leaf-clad culms in winter, and to reduce the desiccation of those leaves once the ground is frozen and roots can no longer supply water, varying proportions of leaves fall each winter according to habitat, those from the highest altitudes dropping nearly all their leaves.
Thin underground rhizomes extend to form a dense mat, from which quite closely spaced culms develop, usually to only a relatively short height. The close culms produce large leaf blades, and the density of this foliage effectively shades out all competition on the forest floor, giving a uniform understorey or groundcover. With dense impenetrable culms and large leaves a single branch is sufficient, with a rather palmate leaf arrangement to minimise mutual shading but improve competition for light with other plants. In winter, when water is unavailable from frozen soils, leaf blade edges are sacrificed, rather than dropping a proportion of leaves, as these leaf blades with necrotic margins will continue to shade out other plants.
Compact, thick rhizomes infrequently produce very tall, thin culms that initially grow upwards but then hang down and continue growth laterally using the support of tree branches. From the nodes two categories of branch develop. Central branches that can remain dormant for many years, grow to be as large and as long as the original culm. These can scramble over tree branches in many directions. Lateral branches are much shorter and thinner, and form sprays of leaves that can continue terminal growth to produce long curtains of foliage hanging from the tree branches. The rhizome has no need to extend far as the branches can spread at a more effective height through the tree canopy, where support is assured.
Other strategies can also be found. In the truly climbing genus, Dinochloa, culms and branches twine around the trunks of tropical trees in SE Asia to attain great heights. In Vietnamosasa, slanting subterranean culm bases with many dormant buds develop from deep rhizomes, and these produce new culm-like branches after seasonal fires in fire climax savannah areas in Indochina. In S America there are bamboos in Neurolepis with no branches, little culm development and a palm-like basal rosette of tough broad leaves. In the Brazilian genus Glaziophyton erect naked rush-like culms have scarcely any leaves at all until they flower, and this illustrates another characteristic of woody bamboos, their exposed green and photosynthesizing culms, which can power the growth of new branch and leaf shoots after defoliation that would require re-growth from the ground in other grasses.
Flowering: adapted architecture for regeneration
Woody bamboos are characterised by heavy flowering, which is usually periodic after a long interval of purely vegetative growth. Flowering is often synchronized over large areas, and the clumps often die after flowering, especially if seed is produced in large quantities or environmental stress is encountered. Sporadic flowering is also seen between gregarious events, and may be the norm in some species.
Various explanations for this behaviour have been put forward, the most famous, by the ecologist Janzen in 1976, hypothesized predator satiation as the key, with prolific seed production being necessary to avoid consumption by animals such as rats. Tales of rat infestations after bamboo flowering may be over-stated however. Another hypothesized benefit is the reduction in competition between parents and offspring when the parents all die, but this is a risky strategy when young seedlings are so vulnerable. The most plausible explanation would be the simple need to cross-pollinate for continued genetic health, which is very difficult for wind-pollinated plants in a low-wind forest understorey environment. While collecting seed of Dendrocalamus hamiltonii across dense forest in East Nepal, I found that clumps near to ridges had better seed production than those in hollows. On another occasion, seed production in Dendrocalamus hamiltonii was very good where pollen was being actively collected by bees. To them the capitate inflorescences were vivid purple-red balls of anthers, and they would no doubt also be attracted to the very similar purple-red balls of stigmas that precede anthesis in this species. Various mechanisms may have evolved in woody bamboos with long, synchronized flowering cycles just one component of a strategy to maximize cross-pollination, while avoiding total self-incompatibility.
Nobody knows the mechanism by which flowering is initiated in woody bamboos after a period of up to 150 years of vegetative growth. An internal counter of some form has to be involved, but what would trigger each count remains a mystery. A limited ability to replicate DNA indefinitely, in some form of ageing process, seems a possible explanation. All that is clear is that once flowering stimuli affect meristematic tissue in a bamboo, then the characteristics of new axes and their units change, sometimes gradually and progressively in larger bamboos, usually more rapidly in smaller ones. As meristematic tissue is found in different locations, terminally, laterally, in separated buds or in adjacent buds in branch complements, the patterns of inflorescence produced can differ surprisingly, both in space and in time.
As the production of flowers takes over the entire plant, lasting sometimes for several years, vegetative sheaths and leaves usually all fall off. In a large segmented plant that has so many flowering axes and little else, it is hard to define what exactly represents an inflorescence. It would seem sensible to start to define it at the points where normal vegetative growth ceased and modified, floral growth starts. This makes each capitate cluster in tropical bamboos a separate inflorescence, rather than considering the branches on which the sessile clusters are borne to be part of the inflorescence. The prophyll basal to each cluster is modified, and can be helpful in identification.
Understanding the inflorescences properly is only possible if the vegetative architecture of branching patterns are understood first. Branching near to the culm in the ‘branch complement’ is not the most important aspect of architecture when flowering starts and spikelts, florets and glumes are produced instead of leaves. Genera vary substantially in whether leaves develop over the years from terminal growth of the branchlet, or from lateral buds. Oldeania produces scores of leaves on long pendulous branchlets, while Phyllostachys produces very few leaves on each axis, starting anew from lateral buds each year, so that a complex branching system develops at the branchlet level. When growth changes from vegetative to flowering, the inflorescences produced are substantially determined by a pre-existing vegetative architectural template.
The first inflorescences produced are often terminal to a leafy branchlet, with leaf blades progressively reduced in size, succumbing to a physiological gradient that can sequentially alter them from large sheaths with functional leaf blades to small, fertile lemmas on a single axis. Lateral axes developing from dormant buds can produce rather different styles of inflorescence later. They are more compact and ramified, and not associated with larger sheaths with leafy bracts. Cephalostachyum is one genus that is particularly prone to production of very different forms of inflorescence. Whether inflorescences are produced later from lateral buds depends on whether and where the genus actually has buds.
Flowering in bamboos involves a physiologically controlled transition from vegetative to analogous floral components, with the flowering structure predetermined by a template provided by its vegetative precursor. The often leafy sheath that precedes floral analogues on a newly flowering axis is arguably not part of the inflorescence, but such sheaths are often referred to as ‘spathes’, and are often said to ‘subtend’ the inflorescence. Like many sheaths in bamboos, whether it is noticeable, persistent or quickly deciduous varies, even within one plant, and may not be consistent. When flowering starts there may be a sudden transition from production of leafy sheaths to glumes, and if collected at that time inflorescences can have ‘spathes’ or be ‘subtended by leafy bracts’, while later inflorescences have none at all. The presence of bracts within the inflorescence proper is of course more consistent and often more critical, but is usually overlooked.
Taking a complete system of bamboo axes and units as the fundamental basis or template for the structure of any part of the bamboo plant helps in understanding, especially for inflorescences and how they differ. They can consist of a system of axes that has extended internodes, giving an open inflorescence, or they can have unextended internodes, giving a compressed inflorescence. They can have all the expected sheaths in place, in a bracteate inflorescence, or they can have many or all sheaths severely reduced to a miniature bract or tuft of small hairs, in a largely ebracteate inflorescence. Prophylls, the first sheaths on axes, like any other sheath may be present or absent in an inflorescence. The buds on each internode may be present or absent throughout the inflorescence branches and/or within the spikelet, and they may remain dormant, or may develop rapidly into further axes to give clusters (iterauctant). Branches may develop separately or they may be fasciculated by compression of their basal internodes. Branches may be erect, or deflexed by the growth of small swellings, pulvini, in their axils.
As the vegetative and floral architectures are thus intrinsically linked, and they both have adaptive value, one may dictate the other, which could be an interesting line of investigation.
The inflorescences of bamboos should thus be analysed carefully and described accurately according to their architectural structure as well as their components. Various terms such as determinate, indeterminate, semelauctant, iterauctant, spathed, bracteate and ebracteate, prophyllate or not, have been coined, but then they have been applied in very different ways by different people, and require a degree of caution, as their use may be the result of superficial appearance rather than a critical examination of real characteristics. It is always better to describe what is seen in simple language than to try to apply clever sounding terminology without understanding it properly or using it in a different way as others. That architecture is so important, and the terminology that describes it equally so.