|Vascular Architecture in Plants
Above: primary plant stem vascular tissue in cross-section. Primary stems are generally green and photosynthetic and
have no or little secondary growth and so are non-woody and occur in young shoots and in herbaceous plants. This
diagram illustrates the structure of the vascular cylinder or stele (surrounding stem tissues are not shown). White
indicates phloem, xylem is shown in black and central parenchymatous pith is dotted. Protosteles occur in many
fossil plants, in psilopsids and club mosses and in some roots. In these steles there is a central solid cylinder of xylem
ensheathed by phloem. For example, the psilopsid Psilotum nudum has an actinostele (in which the xylem cylinder is
ridged) and the club moss Lycopodium clavatum has a plectostele, in which plates of phloem infiltrate the xylem, which
is nevertheless still continuous.
Xylem has to be rigid so that it can draw up sap under suction pressure without collapsing. This makes xylem good for
plant support and in woody plants it is the main supporting or skeletal tissue. To provide greater support without
manufacturing more expensive xylem, the plant stem can contain a hollow cylinder of xylem. Just as in the hollow long
bones of mammals, it is possible to have a bone which contains less bony material but which is nevertheless stronger,
as in having greater flexural stiffness - a wide cylinder is much harder to bend than a narrow cylinder, though we have
to maintain a minimum thickness of the walls of the hollow cylinder to prevent buckling. This gives rise to a
siphonostele, in which the vascular tissue forms a hollow cylinder or siphon, although the central cavity may be filled
with parenchymatous pith. In the ectophloic siphonostele in which the phloem remains on the outside surface of the
xylem or an amphiphloic siphonostele (also called a solenostele) in which phloem coats both the inner and outer
surface of the xylem cylinder. Siphonosteles occur in ferns, and in some gymnosperms and flowering plants.
The vascular cylinder is not generally a complete cylinder, because at the leaf-bearing nodes of the stem xylem and
phloem vessels must arc away from the stele to enter the leaf as a leaf trace. If the leaf is a small scale-like leaf or
microphyll, the vascular cylinder may remain more-or-less intact, otherwise, if a larger trace enters a large leaf or
macrophyll a gap may result in the xylem cylinder, called a leaf gap. This gap only extends so far up the cylinder before
the cylinder closses over at the internode above to become a complete circle again in cross-scetion.
Prerequisite: Water Transport in plants.
Vascularisation is essential for any large land plant. Water, carrying mineral nutrients from the soil need to travel from
the roots to other parts of the plant in the xylem vessels, whilst sugary sap must, for example, translocate from
photosynthesising leaves and storage organs to growing parts and ripening fruit in phloem vessels. Plants are often
divided into vascular and non-vascular, which roughly parallels evolution with large terrestrial forms appearing later in
evolution as vessels evolved sufficiently to supply their needs. However, this division is a simplification. Seaweeds
might not require xylem, since they can absorb water across their entire body surface when submerged, but they do
have a well-developed network of phloem vessels, called trumpet hyphae, interconnecting the various parts of the
seaweed body. Mosses are tiny and although terrestrial, many require damp conditions and so it is not surprising that
they do not require advanced vascular systems. However, even small mosses may have specialised conducting tissue,
functioning as a prototype vascular system.
In the dictyostelium we essentially have a modified siphonostele in which the stem has very short internodes, so that
leaf gaps of the nodes above and below overlap to completely and in cross-section the stele appears to be broken up
into a series of strands. This type of stele occurs in some ferns.
In the eustele (literally 'true stele') the vascular tissue is broken up into discrete longitudinal strands or bundles. This
type of stele occurs in the internodes of the horsetail Equisetum and in some gymnosperms and flowering plants. It is
the 'typical dicotyledonous' primary stem illustrated in text books, e.g. Helianthus.
Finally, in the atactostele we have the discrete bundles of vascular tissue scattered throughout the stem. This type
of stem is characteristic of monocotyledonous flowering plants.
After attaining their initial width and length, the parts of many plants undergo secondary growth, increasing in
thickness and strength. New secondary xylem and phloem will be produced to supply the growing plant. The covering
of the epidermis may become replaced with periderm bearing 'breathing pores' or lenticels. In a eustele, a layer of
undifferentiated cells between the phloem and xylem of each vascular bundle (fascicle) my form a growth zone or
meristem where new cells are produced. This is called the fascicular cambium and is continuous with a layer of
similar cells inbetween the vascular bundles, the interfascicular cambium. Thus there is a complete cylinder of
meristematic tissue only one or a few cell layers thick.As these cells divide by mitosis, they both replace themselves to
maintain the meristem and produce cells which differentiate into new xylem vessels on the inside of the stele and new
phloem on the outside.
This type of secondary growth is extensive in woody plants and indefinite in many trees, with the secondary xylem
forming the wood and the secondary phloem the innermost layer of the bark (or just beneath the bark proper).
Herbaceous plants may also undergo some secondary growth, depending on species. In Pelargonium (Geraniacea)
the vascular bundles are so close together that secondary growth readily produces an entire fused cylinder of
vascular tissue. In Helianthus (sunflower, Asteraceae) a similar continuous cylinder may form at the base of the stem,
but higher up there is no secondary growth, but the interfascicular parenchyma (parenchyma inbetween the vascular
bundles) forms sclerenchyma to toughen the stem and give it extra strength. A periderm may not form in Helianthus,
but the epidermis continues to produce new cells and expand.
As has been discussed elsewhere (see wood) trees often grow by adding annual rings of wood such that the stem
increases in girth. If we picture a tree stem as consisting of a cone of wood then essentially a new cone is added over
the top of this each year or growing season.
Review the detailed structure and function of wood and xylem vessels.
In this article we will look at some additional features of xylem architecture in trees. A single xylem vessel does not
generally extend the whole height of the tree but may be a meter or so in length and communicates with neighbouring
vessels so that the sap can flow the whole height of the tree from vessel to vessel. The xylem sap is drawn up the tree
from the roots by a suction pressure generated by water loss, chiefly through stomata in the leaves of the tree
canopy. This loss of water to the atmosphere is transpiration and the stream of water flowing through the xylem,
carrying valuable mineral salts from the soil, is called the transpiration stream. This does mean, however, that the
lower branches are nearer the source of the flowing sap and may tap more than their fair share, leaving insufficient
sap for the upper canopy which potentially needs it more. To combat this trees have a special architectural feature to
slow the movement of xylem sap into the lower branches - concentric circular vessels.
Above: left, the complete vascular cylinder (siphonostele) of an internode. Middle, a small
vascular trace arcs away from the stele to enter a microphyll, leaving a notch in the vascular
cylinder which extends part-way up. Right, a leaf trace entering a macrophyll results in a gap,
or elongated slit, in the vascular cylinder, called a leaf gap.
Above: the grain in a tree with its bark removed. Note the concentric circular or elliptical
vessels at the bases of the lower branches.
An important part of a tree's insurance policy is the production of dormant buds. Most of these buds will never open
or develop, but should the canopy become damaged some of them may become active and replace the damaged
canopy. Each dormant bud has its own vascular supply which must elongate if the bud is to remain at the surface of
the stem as an epicormic bud. As new layers of wood are added to the trunk, the vascular traces to the buds
elongate to maintain the buds at the surface (some buds may fail to keep up and become buried in the wood).
Above: a longitudinal section of a tree trunk showing the vascular supply to the dormant
epicormic buds. (Based on Busgen and Munch, 1929, in Thomas,2000; trees: their natural
history, Cambridge University Press).
Epicormic buds originally form as normal buds in the axils of leaves on young sheets but which remain dormant. In
some trees the majority of such buds remain dormant and some may abort. Trees can also form new buds de nova
when they are damaged, from any parenchyma tissue (adventitious buds). Branches also have vascular traces
which can be traced to the centre of the trunk as a narrowing cone (a 'spike knot'). As new wood is added the
growth of the branch keeps pace and the knot consists entirely of wood firmly anchored to the surrounding wood of
the trunk. However, if the branch dies then wood added to the trunk will simply grow over it, encasing the dead
branch complete with its bark. The bark around the dead wood does not integrate well and such encased knots
easily fall out of a plank of cut wood.
The physics of water movement in plants and how plants control this is a very complicated subject! However, an
understanding of this topic sheds considerable light on plant anatomy and physiology, including the reasons why
conifers and hard woods have different types of wood, for example. Here we will discuss some of the principles and
results of published studies. This section will probably be expanded periodically to include additional material.
The mechanics of water transport in plants can be modeled using Ohm's law for electrical circuits, namely that the
rate of flow (current) is equal to the 'driving force' (voltage) divided by the resistance. (Note we are using the word
'force' in a colloquial rather than a physical sense). Increasing the resistance reduces the rate of flow, whilst
increasing the 'driving force' increases flow rate. In the case of the plant, we are of course concerned with the flow
of water rather than the flow of electricity. The 'driving force' of water flow is pressure (rather than voltage) or the
pressure gradient (pressure per metre).
Note that the units for pressure are Pa which is equivalent to force per unit area, and that of the pressure gradient
are Pa/m and so this is not a force in the strict physical sense. Conductance is used more often than resistance in
plant physiology. Conductance is simply the reciprocal of resistance (conductance: k = 1/resistance).
Flow rate can be measured in terms of the mass of water passing a given point each second, i.e. the mass flow rate
in kg/s; or in terms of the volume flow rate in metres cubed / s. The flow rate is given by:
Flow rate = Pressure x Conductance
kg/s = Pa x kg/Pa/s
Flow rate = Pressure gradient x conductivity
kg/s = Pa/m x (kg/Pa/s)m
Where conductivity is conductance x length of the pipe (similarly resistivity is resistance x length).
Another way to look at this is to say that flow rate is proportional to pressure, with conductance as the constant of
proportionality, or that flow rate is proportional to pressure gradient with conductivity the constant of proportionality.
Thus, for example we have the following definition of conductivity:
If Q, the flow rate of water, is given in kg/s, for example, and pressure in Pa (pascals) then the units of k will be:
(kg/s) / (Pa/m). Often pressure is inputted into this equation in MPa, but it is generally better practice to stick to SI
We may be considering flow along an individual xylem vessel or along an entire branch. There is one further
modification, we often talk in terms of leaf-specific (hydraulic) conductance (LSC) when considering flow along
a stem segment. This is the conductivity of that stem segment divided by the total surface area of leaves irrigated
by the same stem segment. This gives us a measure of how readily water can be supplied to the leaves. For
example, the LSC of a stem will incorporate all the leaves supplied by that stem (i.e. the leaves downstream of the
stem), but for a branch it will include only those leaves supplied by that branch. This gives us a measure of how
well the axis (stem or branch) can supply its leaves with water.
If we assume that flow rate is constant throughout the system (i.e. no fluid enters or leaves the system and the flow
is steady) then this equates to requiring the sum of the radii cubed of branches of any order being equal to the
same constant. That is, if a parent vessel branches into several daughter branches, then the sum of the cubed
radii of the daughter branches is equal to the radius cubed of the parent vessel.
This proportionality can be derived from the (Hagen-)Poiseuille law for laminar (i.e. non-turbulent) flow in a
cylindrical pipe and the assumption that power consumed (i.e. the rate of energy consumption) is minimised. This
power is equal to the power used to drive the flow and the power used to maintain the circulatory system. It is
further assumed that, for animals, the main maintenance cost is in production and maintenance of the blood itself,
rather than the vessel walls (which are thin in all but the larger arteries). This means that the power for
maintenance is proportional to the volume of the vessel. Vessels are assumed to be cylindrical. The derivation is
given below for those who are curious:
Application of Murray's Law to Plants
The circulatory systems of plants are fundamentally very different to those of vertebrates. These differences can
be summarised in three key principles with specific reference to xylem:
Redundancy. In plants, individual conducting channels or pipes (tracheids or xylem vessels) are in
parallel bundles with the side-walls connected by pits (and pit fields). This means that flow can divert
sideways from one pipe to another (even though the main flow is still along the pipe). This means that
there is a high level of redundancy in the system: blockage of one pipe can be circumvented. This is
functionally important, since in plants the conducting pipes of the xylem are under negative pressure and
will sometimes cavitate (become blocked by air bubbles) or branches may become damaged by the
elements or by grazing herbivores.
Integration. The xylem system of plants is highly integrated: each root is connected to and supplies all
branches. This is partly a result of redundancy.
Compartmentation. The xylem system rarely consists of freely open pipes along the whole length of the
plant. More typically vessels may extend a few cm or m before connecting to another vessel, perhaps
sideways. Tracheids are typically only 1-2 (occasionally up to 10) mm long. This means that if air enters
one compartment, then it can be contained and stopped from spreading to other compartments.
We also need to consider how the power required for maintenance differs for plants. The xylem sap is
largely water and dissolved minerals drawn up from the roots and is assumed to have a low cost of
manufacture (however roots do actively expend energy in transporting minerals into the xylem sap). The
greater investment comes from the relatively thick walls of xylem conduits, which must be rigid to resist
collapse since the xylem operates under negative pressure (suction) rather than positive pressure as in
vertebrates (the latter requires more flexible vessel walls). However, xylem pipes are also approximately
cylindrical and this cost is still approximately proportional to the volume of the pipe.
With these ideals borne in mind, it can be concluded that Murray's law is also applicable to the xylem
system of plants, although many plants do not reach this ideal of minimal energy expenditure due to
physical restrictions, as we shall see.
Blood vessels in animals typically bifurcate or branch into smaller vessels, but may also give rise to many
branches. We can see that the furcation (forking) ratio is greater than or equal to 2. Things are not so
simple in plants. Since flow can also travel sideways between vessels in the xylem systems of plants,
more arrangements are possible, as shown in the diagram below:
In this diagram, a typical bifurcating blood vessel of an animal is shown on the left. Three different arrangements in
plant xylem are illustrated. This is not exhaustive, all we can see about the xylem system is that F is greater than
zero (F > 0).
Let us put all this together on a couple of graphs and see what we get!
The first of these graphs shows leaf specific conductance (LSC) as a function of 'conduit taper'. (Based on data in:
McCulloh and sperry, 2005). To remind the reader this is the hydraulic conductivity of a stem or branch divided
by the total surface area of leaves supplied by that stem or branch. In this case, conductivity has been defined as
flux (the flow rate of water in kg/s) divided by the pressure gradient driving that flow (in Pa/m). The conduit
taper is a measure of how successful branches of the vascular system conduits narrow: it is the ratio of diameters
of the daughter branch to the parent branch.
Efficiency of water transport
Why do the leaves need so much water? If it was the case that they only needed a set amount then plant
hydraulics would be tuned to supply that amount. However, the water rising in the xylem brings mineral nutrients
absorbed by the roots to the cells of the leaf. Leaves also need to obtain carbon dioxide from the air through pores
in the leaves called stomata. This necessarily exposes the interior of the leaf to the outside air causing water loss
by evaporation (evapotranspiration). This water loss creates the suction (negative pressure gradient) which is
mainly responsible for driving water up the xylem in the transpiration stream in the first place. Transpiration is
necessary to replace the water lost so that the cells can maintain access to carbon dioxide (if the leaf wilts the
stomata close and photosynthesis shuts down). For these reasons, conductivity correlates well to plant growth.
Opposing the need for higher conductance are several costs. One of these is the cost of manufacturing the
vascular system. Murray analysed the blood circulatory systems of vertebrates and devised Murray's Law which
states that blood vessels taper when they branch so as to maintain a flow rate, Q, proportional to the sum of the
cubes of the radii of all the branches at every branch level in order to maximise conductance for a given
investment in vascular tissue. That is:
Somebody probably ought to invent a name for the awkward units of hydraulic conductance.
Why trees and vines have different woods
The above graph shows the relationship for LSC as a function of conduit taper when the ideal of Murray's law holds
(solid line). It also shows the relationship for three different furcation ratios (F). Compound leaves of trees and
shrubs (with high F) follow the Murray's law ideal. Vines can also approximate this ideal. Trees, however, can only
approach this ideal with F = 1 and a conduit taper close to 1.0. Note that to achieve this ideal, trees have had to
settle for a lower LSC (conductivity/leaf area) which is a result of having a low F. The apparent reason for this is
that the xylem conduits of tree stems and woody branches also function to provide support and mechanical
strength. This forces trees to adopt a hydraulic architecture which optimises, or at least compromises, the need for
support. To remain stable the total cross-sectional area (proportional to the sum of r-squared) of a tree at each
branch level must either reduce or remain constant, otherwise the tree would be top-heavy like an inverted cone.
This line of mechanical stability is indicated by the dashed line. Trees must stay below this line and so can only
approach Murray's law by having a low F and a low LSC.
Vines rely on other plants for support and so their xylem conduits are less important for support and so can move
above the line of stability to approach Murray's law. They do this by having very large xylem vessel diameters in the
main stem, which can conduct rapidly and result in a higher F. Larger vessels, however, give xylem tissue a lower
density of wall material and it is this wall material that provides the bulk of the mechanical support. This means that
trees need to have narrow xylem conduits. Compound leaves have different mechanical requirments (taken over by
parenchyma, collenchyma and sclerenchyma) and so their xylem vessels are freed of mechanical constraints and
can be high conducting.
Why conifers and hardwoods have different woods
Conifers, such as pine trees (Pinus) have xylem conduits composed of narrow tracheids. This gives good
mechanical support at the expense of high hydraulic conductivity. This would lower maximum growth rates, but
these trees compensate in part by being evergreen to lengthen the growing season. Narrower conduits are also
less prone to cavitating (becoming blocked by air) under cold conditions so conifers compete well in cold climes.
Caveat: available data is still scanty and more studies need to be done to compare hydraulic conductivity in
different woody plants growing in different conditions. However, what data is available supports the following
Hardwood trees have larger xylem vessels which can conduct sap at a higher rate, especially those with ring-
porous woods (such as Ash). These trees also have a higher furcation ratio (F) in their woody parts (though still
lower than in vines and compound leaves) since the vessels in the main stem are especially wide. (The reason for
this is that conductivity remains high throughout the main stem so that topmost branches are not deprived by lower
These trees can afford to do this because they have more sclerenchyma fibres in their woods. In conifers about
90% of xylem tissue consists of conduits, in ring-porous hardwoods this is as low as 10% (and diffuse-porous
woods are intermediate at about 25%). This means that the xylem vessels can sacrifice mechanical strength and
have larger diameters and higher conductivity since other tissues now take on the main supportive role (division of
It is remarkable that detailed studies of hydraulic architecture have shed so much light on the perplexing problem of
the functional benefits of different wood types. I have previously pondered this problem to no avail.
The graph below shows a different way of presenting this kind of data. This shows the deviation from Murray's law
as a function of F, by plotting the ratio of total radii cubed for each order of branching relative to the parent branch
on the vertical axis. For example, we could consider the radii cubed of the main boughs divided by the radius-
cubed of the trunk. This graph is based on data in: McCulloh et al. (2004).
Cruiziat, P., H. Cochard and T. Ameglio, 2002. Hydraulic architecture of trees: main concepts and results. Annals of
Forest Science 59: 723-752.
McCulloh, K.A. and J.S. perry, 2005. patterns in hydraulic architecture and their implications for transport efficiency.
Tree Physiology 25: 257-267.
McCulloh, K.A., J.S. Sperry and F.R. Adler, 2004. Murray's law and the hydraulic vs mechanical functioning of wood.
Functional Ecology 18: 931-938.