Architectural Models of Plants

Most animals, including humans, have a unitary type of organization. A unitary organism (from Latin unitas — unity) develops from a fertilized egg according to a strictly defined plan. By the time of birth or hatching, such an organism already possesses all its major organs, and their number is generally fixed: one head, four limbs, one heart, etc. The growth of a unitary organism primarily involves an increase in the size of already existing structures, not the creation of fundamentally new organs (Bell, 1991). Of course, some animals are capable of regeneration, but their overall body plan remains rigidly determined.

Plants — and this is a crucial difference from animals — are modular organisms. Their body is built from repeating, similarly constructed structural units called modules (Halle, Oldeman & Tomlinson, 1978; Savinykh, 2006).

Imagine that a plant is not like a monolithic building but like a long train to which new carriages are continuously added throughout its life. Each such “carriage” is a shoot with leaves and buds. As it develops, a plant does not simply increase in size; it continually increases the number of its modules, and some of these modules may be modified (e.g., turn into flowers or thorns). The number of modules in a single plant can reach tens or hundreds of thousands, and it is never strictly fixed.

The concept of modular organization is key in modern biomorphology. It was fully recognized and developed in the Russian botanical school, primarily by Ivan G. Serebryakov and his followers (Serebryakov, 1962; Serebryakova, 1977; Savinykh & Cheremushkina, 2015). According to this concept, the life cycle and form of a plant are the history of the emergence, functioning, and death of successive modules.

In plant structure, several levels of modular organization are distinguished. We will be primarily concerned with those related to the construction of the shoot system.

Metamers are formed sequentially from the apex (apical meristem) of a shoot. The time interval between the initiation of two successive metamers is called the plastochron (Halle et al., 1978).

Thus, modularity is not just a theoretical abstraction but a key principle of plant organization. It is precisely because of the ability to repeatedly repeat and modify modules that a plant can possess enormous plasticity, adapting to changing environmental conditions, recovering from damage, and attaining huge sizes.

From this fundamental feature — modularity — all other properties of plant architecture follow. Knowing that a plant consists of repeating units, we can move to the next question: by what rules (algorithms) does their creation and spatial arrangement occur? These rules are what the architectural model describes.

In the next section, we will examine the hierarchy of shoot structural units in more detail and discuss what an “elementary shoot” and “metamer” are, as well as growth rhythm — one of the key traits for classifying architectural models.

Metamer (synonym: phytomer) is the smallest, indivisible repeating unit of a shoot, formed by the apex over one plastochron. The classical definition of a metamer includes a node with a leaf (or leaves) arising from it, an axillary bud, and the internode below (Bell, 1991; Barykina & Gulenkova, 1983). Some authors may define metamer more broadly, but for architectural analysis it is convenient to adhere to this canonical view.

The stem, therefore, represents a sequential series of metamers. It is in their arrangement and properties that characteristics such as the following are encoded:

The metamer is the “atom” of the shoot system. However, to describe plant development over longer time periods, this single unit is insufficient.

The next level of hierarchy is the elementary shoot (annual shoot, growth unit). This is a shoot that develops from one bud over one growth cycle. In temperate plants with rhythmic growth, the elementary shoot typically corresponds to the annual shoot — the current year’s increment (Halle et al., 1978; Barthelemy & Caraglio, 2007).

Boundaries between successive elementary shoots on the stem are often clearly visible: they are the so-called bud scale scars — remnants of bud scales that fall off after the bud breaks. In plants with continuous growth (e.g., many tropical trees or palms), external boundaries between elementary shoots may be smoothed, but they can still be identified by changes in leaf angle or microstructural markers (Halle & Martin, 1968).

The elementary shoot itself consists of a sequential series of metamers formed during one growth cycle.

The term “module” in biomorphology can be used in different meanings, sometimes leading to confusion. We will follow the approach proposed by N. P. Savinykh (2006, 2015), who distinguishes three categories of modules, differing in complexity and lifespan.

In a broad sense, the “main module” is a large shoot system that determines the type of life form and architectural model. In herbaceous plants, the main module is often a partial shrub — a set of shoots formed as a result of branching and functioning as a relatively independent unit (Serebryakova, 1977). In trees, the analogue of the main module can be a skeletal branch or even the whole trunk, if it is formed by sympodial growth.

To clarify terminology, N. P. Savinykh (2006) proposed distinguishing:

Such a three-level hierarchy allows a rigorous and unambiguous description of plant structure at different stages of its ontogeny (see Table 2 in Savinykh & Cheremushkina, 2015). In the context of architectural models, the properties of universal and main modules — how they grow, branch, and complete their development — are of greatest importance.

The idea that plants are built from repeating units (metamers) proved extraordinarily fruitful. It formed the basis of formal languages for describing development — L-systems, which allow modeling plant growth as a process of sequential replacement of modules according to given rules (Lindenmayer, 1968; Prusinkiewicz & Lindenmayer, 1990; Prusinkiewicz & Remphrey, 2000).

A key point is that each metamer carries information about its type (e.g., a metamer with a short internode and a scale leaf, or one with a long internode and a large assimilating leaf). The plant’s developmental program specifies which type of metamer follows which. Determinate growth is the property of a shoot whose apical meristem is programmed to produce a limited number of metamers, after which it turns into a flower or dies (Bell, 1991). An example is the flowering shoot of a tulip. Indeterminate growth means that the apical meristem is potentially capable of forming metamers indefinitely. This property is characteristic of vegetative shoots, such as the main trunk of a pine.

Thus, the basic units of architecture — from metamer to main modules — form a scalable system of description. Knowing the rules for the formation and replacement of these units, one can decipher the architectural model of any plant. In the following chapters, we will move to the practical application of this knowledge and consider the specific types of models identified by Hallé and his coauthors.

**Architectural model is a genetically determined program of growth and branching that defines the sequence and pattern of formation of all axes of a plant throughout its life, from seed germination to natural death.

This definition contains three key points.

First, the model is genetically determined. This means it is encoded in the plant’s genome and inherited. Different species (and even different cultivars) may follow different models, which is a hereditary characteristic (Halle et al., 1978).

Second, the model is a program, i.e., a process unfolded in time. It describes not a static picture of the “adult tree” but the entire sequence of changes that occur in the plant from seed to old age. This is a dynamic, not static, concept (Barthelemy & Caraglio, 2007).

Third, the model defines the growth and branching rules for all axes. It does not simply say that “this tree has a main trunk,” but explains how this trunk is formed (monopodially or sympodially), how branches are laid down on it, how they are oriented in space, and when flowers appear on them.

It is very important not to confuse the concept of architectural model with the more traditional concept of life form (biomorph). Their relationship can be understood through an analogy with building architecture.

Life form is, rather, the “functional purpose of a building.” For example, “residential house,” “school,” “factory.” This category answers the question: what is the object for? In botany, the life form (tree, shrub, herb, liana) reflects the general adaptation of a plant to climatic and soil conditions (Raunkiaer, 1934; Serebryakov, 1962). Two different architectural solutions (a wooden log house and a concrete multi-story building) can belong to the same life form (“residential house”).

Architectural model is the “construction principle” or “type of structure.” It answers the question: how is the internal space organized? For example, “tower type” with a single load-bearing column, or “post-and-beam type” with a system of supports. In botany, the architectural model does not directly depend on the size of the plant. The same model can be manifested in a giant tropical palm and in a modest annual herb (Halle & Oldeman, 1970).

Thus, the life form is more of an ecological category, while the architectural model is a morphogenetic one, describing the endogenous structural plan. Within a single life form (“trees”), many architectural models can be realized (e.g., Rauh’s, Leeuwenberg’s, Troll’s models, etc.) (Halle et al., 1978; Savinykh & Cheremushkina, 2015).

How did scientists manage to identify these programs by observing giant tropical trees? Hallé and his coauthors proposed working backwards: analyze not the final (often environmentally distorted) form, but the sequence of axis initiation and development in plants growing in optimal or controlled conditions (Halle et al., 1978). A key observation was that when comparing thousands of species, all the diversity of their architecture reduces to a small number of combinations of a limited set of traits.

These traits (criteria) were briefly listed in the introduction, and now we consider them as the basis of classification:

Combining these traits, one can build an identification key that allows any plant to be assigned to a particular model. Importantly, the models are not discrete, isolated units. Halle and Oldeman (1970) introduced the concept of the “architectural continuum” — a continuous space of forms where models are merely “attractors,” and transitions between them are possible. In addition, in response to injury or changes in light, a tree may initiate reiteration, where dormant buds begin to “replay” the architectural program, creating complex, repeatedly repeated structures (Oldeman, 1974).

The concept of architectural models revolutionized plant morphology by providing a powerful tool for analyzing development. Knowing the model allows one to:

In the next chapter, we will finally apply this theoretical foundation to the analysis of specific architectural models. However, before moving to their review, we need to examine in detail the four key traits by which models differ. This is the subject of the next section.

How does a plant grow in length — evenly throughout the year or in separate “bursts”? This trait describes growth rhythm.

Rhythmic growth. Characterized by alternating phases of active shoot elongation and phases of rest. During the rest phase, the apical meristem is in a state of relative dormancy, often protected by bud scales — modified low leaves (Halle et al., 1978). This type of growth is typical of most temperate trees (oak, birch, apple), where annual growth is clearly limited by the winter period. However, rhythmic growth also occurs in the tropics, where it may be linked not to temperature but to alternating wet and dry seasons, or even be endogenous (internal) in nature, as in the rubber tree Hevea brasiliensis (Halle & Martin, 1968).

Continuous growth. In this case, the rest phase is absent, and the shoot grows (at a variable rate) continuously. There are no bud scales, and leaf primordia develop sequentially without long pauses. Continuous growth is typical of many herbaceous plants, palms, and some trees of humid tropical rainforests (e.g., mangroves of the genus Rhizophora) (Tomlinson & Gill, 1973).

The difference between rhythmic and continuous growth affects architecture: in plants with rhythmic growth, bud scale scars are often visible on the stem — traces of scales delimiting annual increments. These scars allow precise determination of branch age.

This is perhaps the most important and visible trait. It answers the question: does the main axis retain its dominant role throughout the plant’s life, or is it replaced by lateral axes?

Monopodial branching. The main axis (trunk) grows indefinitely due to a continuously functioning apical meristem. Lateral shoots (branches) are always subordinate to the main trunk and generally lag behind it in growth (Bell, 1991). Such a system resembles a slender pyramid or column. Classic examples are most conifers (spruce, pine), and among broadleaves, the Lombardy poplar. In flowering plants, monopodial branching is often combined with lateral flower placement to avoid interrupting the growth of the main axis (Halle et al., 1978).

Sympodial branching. In this case, the apical meristem of the main axis sooner or later ceases growth (turns into a flower, dies, or simply stops). Its “leader” function is taken over by one or more lateral buds located just below. This creates a so-called false (sympodial) axis, composed of several successive lateral shoots (Barthelemy & Caraglio, 2007). Sympodial branching is the dominant type in most angiosperm trees (linden, birch, willow). It gives the crown a more spreading, “branchy” character.

Important note. Sympodial growth may not be obvious externally. In some trees, the replacement of axes occurs so quickly and inconspicuously that an illusion of monopodium is created. This phenomenon is called pseudomonopodium (Troll, 1937). To recognize it, one must study the scars of dead apices on the trunk.

This trait describes not so much the branching pattern as the geometry of shoot arrangement.

Orthotropic axes. These are shoots that grow vertically upward (or, rarely, strictly downward). They possess radial symmetry: leaves and lateral branches are arranged evenly around the circumference. Orthotropic shoots serve as “scaffolding” — they form the trunk and main skeletal branches that lift the crown toward the light (Halle et al., 1978).

Plagiotropic axes. These are shoots that grow horizontally or at an angle to the vertical. They possess dorsiventral symmetry — they have a clearly defined upper (adaxial) and lower (abaxial) side. Leaves on plagiotropic shoots are often arranged in one plane (distichously), forming a kind of “fan.” The main function of plagiotropic branches is to efficiently exploit horizontal space for photosynthesis (Halle et al., 1978; Barthelemy & Caraglio, 2007).

It is important to understand that the same species may have axes of different orientations. For example, in spruce, the main trunk is orthotropic, while the lateral branches are plagiotropic. Moreover, some models are characterized by “mixed” axes, which grow orthotropically in the basal part and switch to plagiotropic growth distally (Mangenot’s model, Troll’s model).

This trait is closely related to the previous ones and answers the question: where and how are flowers and inflorescences formed?

Terminal position (terminal flowering). The flower or inflorescence is formed at the very tip of the shoot. Such flowering is a determining factor: once the shoot has flowered, its elongation ceases (Halle et al., 1978). If a terminal inflorescence forms on the main trunk, this inevitably leads to sympodial branching (as in linden or horse chestnut). Examples: tulip, snowdrop, many Apiaceae.

Lateral position (lateral flowering). Flowers or inflorescences are formed in leaf axils (laterally), while the apical meristem continues its vegetative growth. In this case, axis growth is not stopped by flowering. Lateral flowering is characteristic of species with monopodial branching (e.g., plantain, many palms, oak) (Bell, 1991).

Each of the four described traits individually does not define a model. It is their combination that matters. For example, the combination “monopodial branching + lateral flowering + orthotropic trunk” yields one model (Rauh’s model, typical of spruce and many tropical trees). The same combination but with plagiotropic branches leads to a different model (Massart’s model).

Hallé and his coauthors (Halle et al., 1978) identified 23 such stable combinations, which they called architectural models (often named after the botanists who first described them). Some of these models we will examine in detail in the next chapter. However, it should be remembered that these 23 models are not rigid boxes but rather reference points in a continuous spectrum of possible architectural solutions (the architectural continuum). Knowledge of the four key traits gives us a tool for independent analysis and understanding of the logic of morphogenesis of any plant.

In the next section, we will proceed to an overview and comparison of the main architectural models, to see how these traits work in “living” examples.

The ability of a plant to change leaf shape and shoot growth pattern as it ages is called heteroblasty (Bell, 1991; Jones, 1999). Classic examples:

In the Russian biomorphological school, a distinction was made between the external appearance of a plant at a specific moment in its life and the final, “adult” life form.

Ontobiomorph — the habitus of a plant at a certain stage of ontogeny (seedling, juvenile, generative, etc.) (Khokhryakov, 1981; Savinykh & Cheremushkina, 2015). For example, a small oak seedling with a few leaves is one ontobiomorph, while a hundred-year-old mighty oak is a completely different one.

Principal life form — the habitus of a plant in its mature generative state; this is usually what is meant when referring to the life form of a species (Serebryakov, 1962).

Plant ontogeny can be seen as a regular succession of ontobiomorphs. Studying this process is a crucial task of biomorphology, as it allows us to understand how and when key architectural rearrangements occur (Savinykh & Cheremushkina, 2015).

Architectural dynamics can be not only gradual but also discrete, with the plant sequentially realizing different models. Hallé and his coauthors (Halle et al., 1978) introduced the concept of the architectural continuum, where models are merely “attractors.” In ontogeny, a tree may transition from one model to another.

Transition related to flowering. In palms, for example, which realize Corner’s model with long-term monopodial growth, flowering may trigger Holttum’s model with a terminal inflorescence and subsequent death of the trunk (as in the coconut palm?). In other cases (e.g., agaves), a monocarpic shoot is formed only after a long vegetative period. This demonstrates that a change in functional state (vegetation → flowering) can radically alter the architectural program.

Transition related to change in axis orientation. Many tree species begin life as upright shrubs (with orthotropic shoots), but later, under the weight of the crown, their trunks and branches become plagiotropic, and the tree begins to grow in width. An example is Troll’s model, where initially plagiotropic axes secondarily become orthotropic in the basal part (Halle et al., 1978). In some species, this occurs only with age.

A special case of ontogenetic dynamics is reiteration — the “repetition” of the architectural model at a higher level (Oldeman, 1974). Dormant buds on an old trunk or branches may awaken and give rise to shoots that literally copy the development of a young tree (with all its phases, from juvenile to generative). This leads to the appearance of complex “trees-within-trees”: the old trunk becomes overgrown with young crowns, each built according to the same architectural model as the mother tree.

Reiteration can be adaptive (in response to damage or improved light) or automatic (endogenously programmed in some species) (Halle et al., 1978; Barthelemy & Caraglio, 2007). It is thanks to reiteration that old trees acquire their characteristic, often whimsical shape. Reiteration essentially allows a tree to “rejuvenate” and repeatedly use the same architectural program to exploit new space.

Ignoring ontogenetic dynamics can lead to erroneous identification of the architectural model. A juvenile plant (e.g., ivy) would be assigned to one model, and its adult form to a completely different one. Only by tracing the entire development cycle from seed to seed can one confidently say that we are dealing with a single model in which age-related plasticity is an inherent property.

Thus, the architectural model is not a static blueprint but a developmental algorithm that includes rules for ontogenetic transformations. Understanding these rules is necessary for correctly interpreting plant form in nature and for their skillful cultivation (e.g., for pruning, crown training, predicting habitus).

In the next section, before concluding the article, we will summarize the practical significance of the architectural model concept for agronomy, forestry, and landscape design.

The classic task of a gardener is to shape the crown of a fruit tree so that it is convenient for care and harvesting, well-lit, and yields high-quality fruits. Without knowledge of the architectural model, pruning is often “blind” and can harm the plant.

Model with monopodial branching and lateral fruiting (e.g., Rauh’s model, characteristic of many pome crops — apple, pear). Such trees have a clearly expressed leader (central axis), and fruit structures are formed on lateral branches. The agronomist’s task is to maintain leader dominance by shortening or removing competing shoots, and to stimulate fruit wood formation on the periphery. Attempting to cut off the central leader in such a tree (unless a special training system calls for it) leads to loss of leadership and proliferation of “water sprouts” — vigorous vertical shoots that take a long time to bear fruit (Halle et al., 1978; Savinykh, 2006).

Model with sympodial branching (e.g., Leeuwenberg’s model, typical of many stone fruits — cherry, sweet cherry, and some tropical fruits like mango). In such plants, the main trunk ceases growth after flowering (or by itself), and is replaced by a lateral branch. The crown forms as if from “tiers” and tends to become dense. Here, the agronomist should instead thin the crown, removing old fruited branches and stimulating the growth of replacement shoots from dormant buds. Knowledge of the sympodial pattern allows one to choose the correct cutting point and direction for the future branch (Barthelemy & Caraglio, 2007).

Champagnat’s model — characteristic of many shrubs and subshrubs (e.g., raspberry, blackberry, some rose species). Their shoots live for two years: growing the first year, flowering and fruiting the second, then dying. Knowledge of this model underlies all care of raspberry plantings: removal of fruited canes, thinning of young suckers. Without it, the plantation quickly becomes an impenetrable thicket, and yield drops sharply.

Desired crown architecture is an important breeding goal. Knowing how architectural traits (branching type, branch orientation) are inherited allows the development of cultivars with compact crowns suitable for mechanical harvesting.

Columnar apple trees — a classic example of a mutation in architectural model. They have strongly suppressed lateral branching, and fruit spurs are formed directly on the main trunk. This mutation affects traits defining Rauh’s model, making the plant dwarf and high-yielding per unit area. Breeders have fixed this trait and created entire series of columnar apple cultivars (Prusinkiewicz & Remphrey, 2000).

Dwarf rootstocks — another example of using architectural features. By grafting a commercial cultivar (with its own model) onto a rootstock with a genetically determined weak growth vigor (a different model or its modification), compact, early-fruiting trees are obtained, convenient for intensive orchards.

Breeding for “pyramidal” or “spreading” form. Different climatic zones and planting systems require different crown types. Knowledge of the genetic basis of architecture speeds up the breeding process.

In forestry and agroforestry, understanding architectural models is necessary for predicting tree growth and their interactions in a stand.

Stand structure. Species with a monopodial leader (e.g., pine, spruce) form straight, well-pruned trunks, which are highly valued in forestry. Species with sympodial branching (linden, maple) produce more knotty wood, but are often valued for their ornamental and protective properties. When establishing forest plantations, the architectural features of each species are taken into account to ensure optimal planting density and reduce the risk of stem leaning or curvature due to competition for light (Sachs & Novoplansky, 1995).

Forest growth modeling. Modern computer models of forest ecosystem dynamics increasingly incorporate modules describing architectural tree growth (so-called functional-structural models). This allows prediction not only of stemwood volume but also of crown development, which is important for assessing light regimes, fire hazard, and wildlife habitat (Godin & Sinoquet, 2005; Palubicki et al., 2009).

In intensive horticulture and greenhouse production, knowledge of the architectural model is used to manage the production process, i.e., the distribution of assimilates between vegetative growth and fruiting.

Planting systems. For crops with plagiotropic branches (e.g., some grape varieties, lingonberry), planting schemes are designed to maximize filling of horizontal space. For orthotropic crops (e.g., maize or sunflower), optimal stand density is designed so that each stem receives enough light.

Flowering management. In species with terminal flowering (Holttum’s model, Champagnat’s model), flowering inevitably leads to cessation of shoot growth. This is used to program yield and growing season (e.g., in strawberry or cereal crops). In species with lateral flowering (Rauh’s model), part of the generative buds can be removed to improve the quality of the remaining fruits or to stimulate vegetative growth.

Diagnosis of stress. Changes in architecture (appearance of apical dominance in sympodial species, increased branching in monopodial species, change in shoot orientation) can serve as early indicators of stress (drought, salinity, nutrient deficiency, disease) (Barthelemy & Caraglio, 2007). By knowing how to read architectural signs, the agronomist can adjust care in time.

Control of root-sprouting weeds. Knowledge of the modular organization and architecture of underground shoots (rhizomes, stolons) is the basis for developing control strategies against perennial weeds (couch grass, sow thistle, Canada thistle). Mechanical destruction of rhizomes (e.g., rotary tilling) provokes the awakening of a huge number of dormant buds and massive weed proliferation (Serebryakov, 1962; Gatsuk, 2008). Understanding the architecture of their shoot system allows one to choose the most vulnerable phase and control methods (e.g., depletion in shade or herbicide treatment at the time of active assimilate translocation to rhizomes).

Thus, the concept of architectural models serves as a bridge between fundamental botany and practical crop production. It turns the art of pruning and training into a science-based technology, helps predict plant growth and development, and enables the development of environmentally friendly weed management methods. For an agronomist, knowledge of a crop’s architectural model is as necessary a tool as knowledge of its temperature or water requirements.

In the concluding section, we will summarize and formulate the main conclusions about the significance of the architectural approach for modern biology and agronomy.