- introduction to favourable environments
What is a growing system?
The way in which planted trees are arranged in relation to existing trees, crops, etc. (D 10).
What is meant by a favourable environment?
One in which young trees have a good chance of surviving, becoming established and growing
But isn't it the trees themselves that improve the site?
Yes, indeed, that is one of the most important benefits of tree planting.
But success is much more likely if growing conditions for the young trees are favourable,
especially during the first few weeks after planting.
Yet they've got to be able to stand up to harsh conditions!
Yes, they often have to. But, however hardy they may become, most kinds of trees are
vulnerable to drying out when first planted, because their root systems:
- have been restricted, if growing in a container;
- are bound to have been somewhat damaged when the plants were taken from the
nursery, transported and planted;
- will not take up water very efficiently until they have had time to grow and branch
into fresh soil; and
- usually have to compete with already-established plants.
How can I give my trees a good start?
In several different ways, including:
- keeping sufficient trees when opening the canopy (D 50–51, D 60), and if possible
leaving litter, branches and logs on the ground (D 63);
- handling them and planting them carefully;
- planting at the right time;
- using branches or palm fronds as simple, temporary shading;
- mulching; and
- if possible, watering.
Points 2–6 are covered in Manual 5.
But what if the site is bare and dry?
It will certainly be harder to create favourable environments (D 50), but you could:
- use tree and shrub species that are colonisers or light-demanders (D 14), at least for the
- if the site is degraded, follow the suggestions for reclaiming it in sheets D 22 and D 32;
- use healthy plants with a good root system and not too large a shoot (Manual 3);
- consider direct sowing (Manual 2).
Can I also improve conditions for later on?
Yes, you can. It is very important to think ahead, for instance about:
- how many existing trees to keep, of what sizes (D 2, D 51–52);
- the species diversity of your planting (D 30, D 53), which is likely to affect the later
growth of the trees, besides increasing your options in the future;
- the pattern of planting (D 54), which greatly influences the survival and growth of the
trees, as well as determining how the trees fit in with other activities.
This all sounds too difficult!
It may seem a bit confusing, but this is because:
- the Manuals cover many different planting sites (D 20–29);
- there is a large choice of tree species, grown for various reasons (D 30–42);
- local needs for tree products may not be the same (D 33–40);
- trees can be planted in a variety of growing systems (D 51–55);
- available information and experience varies (D 4), and knowledge on growing some
tree species is still scanty (D 1, D 31).
However, the basic idea is straightforward - working out which species and growing system
will give the trees the best chance of success in the particular conditions.
What are the main things to decide about?
- The best way of managing the stand sustainably (D 5, D 24, D 26).
- How much (or how little) to open up the canopy of existing trees (D 51);
- Whether to grow trees in one storey, or more than one (D 52);
- How many different tree species to include (D 30, D 53);
- Whether to plant the trees in sizeable blocks, small patches, lines or as an intimate
mixture (D 54).
Are there some guide-lines to reduce the risks of failure?
- Go and see any field trials that compare different growing systems (D 29), and consider
doing some formal or informal experiments yourself (D 6);
- Try and combine past experience with fresh ideas, testing out new or modified systems
on a small scale first;
- Wherever possible, aim towards retaining or re-creating some of the basic features of
the natural vegetation for the particular site (D 10, D 14), keeping some existing trees
wherever this makes sense (D 2);
- If in doubt, provide plenty of shade at the time of planting, and reduce it gradually as
- Make a plan of work (D 4) and keep records (Manual 5).
- controlled opening of the canopy
Won't my trees need plenty of light?
It depends on the circumstances:
Too little light means that there may be:
- not enough sugars produced (D 10–11, D 14) to keep the trees growing well; and
- a lot of root competition from other trees.
But too much light (D 11) can lead to:
- direct damage to the shoots of the young trees, especially if they had been previously
growing under heavy shade;
- rapid drying out of the young tree, before its roots have had time to make new growth
that can take up sufficient water;
- soil conditions being too hot for important organisms (D 13, D 32);
- favouring of grasses or weeds over trees (D 14, D 25); and
- bark scorch in larger trees with thin bark (D 11).
Well how do I decide how much to open up?
Here are some suggestions to help:
- Think about how tropical ecosystems work (D 10–16);
- Then apply this to your planting site (D 20–29), taking into account how many (or few)
trees are there at the moment;
- Next consider the various benefits and useful products that are priorities, and choose
which sets of species to grow (D 30–42);
- Finally weigh up which growing system (D 50, D 52–55) is likely to give the best results
with the chosen species, remembering the risks of rapid soil erosion (D 23) and damage
to trees from wind and fire (D 66).
What if there is a complete canopy of trees already?
Unless the area is being kept as an ecological reserve (D 28), a moderate amount of opening-up
will be needed for the establishment of new trees. For example, you might:
- remove a few selected mature trees for timber (D 36–37);
- encourage immature trees, saplings and seedlings to survive and grow well (D 24,
- convert unthrifty pure plantations into mixed, 2-storey stands (D 2, D 52–53);
- allow farming to take place amongst the trees (D 3, D 21); and/or
- plant additional species and genetically-improved selections (Manuals 1 and 2).
When will more opening-up be needed?
- When planting light-demanding tree species (D 14).
- If growing some crops in agroforestry conditions (D 3, D 21).
- To allow grasses to flourish under the trees for grazing or fodder (D 15, D 34).
- For parks and planting near roads and buildings (D 28, D 41).
And when should there be less?
- When the area is well-stocked with a range of useful tree species;
- If you are planting shade-bearing trees;
- When many of the existing trees cast a light shade and/or lose their leaves for a month
- On sloping ground (D 23); or
- When one aim is to conserve the diversity of existing plants or animals.
Wouldn't leaving more trees mean too much root competition?
Yes, this might be a problem, although there can also be strong competition from weeds in
more open sites. You could perhaps cut the roots back when digging the planting holes (Manual
If the existing species include dominant trees or weeds with an ‘aggressive’ rooting habit, or
those which produce chemicals that inhibit growth of other plants (D 14), you could try localised
burning or cultivation (D 63–64). Otherwise, arrange your planting to avoid these areas, or use
a fast-growing tree species that can compete (D 32).
What if the area has already been partly opened up?
What you do depends on whether the opening was insufficient, about right or too much.
Wherever possible, it is best to plan the opening up yourself, to suit the species (D 53) and the
pattern of planting you have chosen (D 54).
In any case, it is often necessary to do some cutting of undergrowth (D 63), and perhaps also
of damaged trees, in order to be ready to mark out and dig the planting holes (Manual 5).
Because of the risk of degrading the soil, you could keep plenty of trees, allowing you the option
of taking out some more later on.
Supposing there are only a few trees left on my site?
It will usually be best to keep any tree that is not actually liable to cause problems, especially
if the site has been degraded (D 22).
Even if these do not yield many products, they will provide some much-needed protection for
the soil (D 60).
What if it's savanna land?
In general, the canopy would be more open than in forest areas, though:
- trees are not evenly scattered in natural savannas (D 25);
- fewer trees might suit certain crops (D 21); but
- more trees may be needed on slopes than on flat land (D 23).
Should a forest canopy be evenly opened?
You could aim to make it more even if:
- it is very uneven at the moment;
- you want to plant new trees all over the area;
- regular shading is needed for animals, an understorey crop or an experiment (D 67).
But it would be better to have the opening more irregular when:
- some of the ground is steep or rocky;
- you want to plant in groups (D 54);
- there are existing patches of seedlings or sapling trees that need more light;
- the growth of larger but immature trees can be increased by liberation thinning.
What is liberation thinning?
Cutting perhaps 15–20% of the trees of poorer form that are restricting the growth of more
valuable trees (D 24).
Why is it done?
In order to increase the survival, growth rate and straightness of immature trees, and so obtain
greater yields in the future.
Unlike extractive logging, it is a move towards a sustainable system of growing trees (D 5,
D 62), where as much emphasis is put on continuing and improving the stand for harvesting
in perpetuity as on present yields.
When should opening up be done?
- If the area is not being burnt: it should be done only a little while before planting, in
order to minimise the time the soil is exposed (D 60);
- If it is to be burnt: do it in time for the wood to dry out (D 63);
Avoid cutting too long beforehand, because this can lead to:
- rapid loss of soil fertility; and
- a greater risk of fires getting out of control and killing people, or damaging
neighbouring stands of trees and crops (D 66).
Should I remove trees with large, spreading crowns?
Yes, if they:
- cast a heavy shade and/or have agressive root systems;
- could become dangerous (D 66);
- are species with few uses.
No, when they:
- cast a moderate or light shade;
- are soil-improvers (D 32);
- are rare species (D 16).
What if they are too big and hard to cut?
- try and fit them into your planting pattern (D 54); or
- ring-bark and/or poison them.
What are storeys?
Tree crowns occurring in more or less distinct layers. Three types of leafy canopy are:
- single-storey stands, with most foliage occurring at a similar height above ground;
- two- or three-storey stands, with the foliage concentrated at two or three different
- multi-storey stands, where the leaves occur at many different heights in the canopy.
Is that because of tree age?
Storeys are often related to the ages of the trees, and the three types are sometimes called:
- An even-aged stand;
- An uneven-aged stand; and
- A selection forest, (after the ‘Selection System’, in which all ages are mixed together,
and a few trees are removed at intervals to maintain the same age distribution).
However, smaller trees are not necessarily younger, and some of them never grow tall. In
addition, many tropical species lack clear, annual growth-rings, so it is often easier to use size
rather than age.
Isn't it easier to grow trees that are about the same size?
Sometimes yes, it might be the best option available; but
Generally no, because it involves getting rid of all the existing trees to plant new ones, which
is usually undesirable.
What's wrong with clearing the existing trees away?
- It generally exposes the soil to rapid degradation (D 22–23) while the new trees are
small (D 12–13, D 20, D 60);
- It may not provide sufficient biodiversity for them to continue to thrive, especially if
they are all of one species (D 53);
- It wastes the potential yields represented in the existing immature trees (D 24, D 51);
- It sharply reduces the number of useful products available (D 33–40).
So the risks are much greater with complete clearing. In addition:
- there will be fewer options open in the future; and
- in some cases, the land use may prove unsustainable (D 5).
But suppose I want to grow a lot of trees for one purpose!
For many years, it has been assumed that large, single species, single-storey plantations are the
best way of doing this.
Now, however, there are considerable doubts about whether such policies are necessarily wise
(D 2), except in particular circumstances (D 53).
But what else could I do?
- Instead of one large stand, consider a number of smaller plantings (D 54);
- Instead of single species stands, think about the range of species that can fulfil a given
purpose, and the advantages of mixtures (D 30, D 53);
- Instead of single-storey stands, weight up the advantages of retaining existing larger
trees (D 20, D 50, D 54);
- Instead of separating forestry and agriculture, remember the advantages of
agroforestry (D 3, D 21).
What is the best way of getting more than one storey of trees?
By combining planting with natural regeneration (D 2).
How would I do that?
- Do an inventory (sample survey) of the site (D 24), to find out what kinds and sizes of
trees there are in it;
- Identify the individual immature trees and groups of saplings that could grow into
- If needed, open up the canopy a little to favour them (D 51);
- Plant trees in any gaps where there are few natural replacements.
Surely they won't grow well unless I remove the taller trees!
- Shade-bearing trees may grow better under moderate shade than in full light (D 14),
as for example dipterocarps under the shade of opened forest or Albizia falcataria.
- Light-demanding species may not grow as fast under a canopy, but before you remove
or poison all the overwood, think about:
- what the overall growth rates of the stand are;
- whether the older trees are ready to harvest yet;
- whether selective thinning of the overwood might not make sense;
- what effects you may expect on the soil (D 12–13, D 23);
- which conditions will be best for the young trees (D 50).
If you plant light-demanding trees in a sizeable gap (D 54), they could continue to receive plenty
of overhead light.
What are two storeys particularly good for?
- Most sites that have some or many existing trees on them (D 2);
- Underplanting with shade-bearing trees to convert a plantation that is not thriving
(D 20, D 53);
- Sloping ground (D 23);
- Many agroforestry situations (D 3, D 21, D 33).
Isn't it more dangerous to work in?
Not necessarily. Branches are more likely to fall (D 66) when all the overwood trees have been
How can I plant and get two storeys?
- By using mixtures of a faster growing light-demander and a slower growing shadebearer;
- By underplanting other species after soil-improvers (D 32) have reclaimed degraded
land (D 22);
- By adding trees to farms under shifting cultivation before they return to tree fallow
(D 3) or secondary forest;
- By planting appropriate trees on permanent farms (D 21).
Some examples are pollarded Erythrina or Leucaena over coffee, Gliricidia over tea, Inga or
Terminalia over cocoa.
How about more than two storeys?
This is especially suited to:
- some kinds of ‘uneven-aged’ forestry, including enrichment planting;
- certain types of agroforestry, including ‘Home gardens’ and ‘Orchard gardens’;
- parkland (D 28, D 41).
And multi-storey stands?
Having trees of all sizes is particularly appropriate for:
- obtaining natural regeneration (D 2) of tree species that are suited to being managed
on a Selection System;
- woodland used for hunting animals and collecting wild fruits, etc (D 33);
- Wildlife Reserves (D 28).
Is it a good idea to use several different growing systems?
Yes it is, because that is more likely to:
- meet multiple needs;
- be better protected against risks;
- allow more choices in the future;
- fit the different approaches of the various people involved (D 5).
Could single-storey plantations be one of them?
Yes, indeed. As long as these are not too large, they may be well suited for:
- the first steps in restoring degraded land (D 22);
- a set of fuelwood plots (D 35) that are coppiced in sequence;
- trees for forage for domesticated animals (D 34);
- trees producing wood pulp or latex (D 37), and some kinds of fruits (D 33).
Have all these systems been well tried?
Some have been practised for centuries, while others are more recent.
Some have been carefully studied, while others have not.
This is why it is important:
- not just to use one system, but think about alternatives;
- to retain older methods while developing new approaches; and
- to do both formal and informal experimental trials (D 6, D 29).
Are there any guidelines?
Here are a few:
- If you are a forester, avoid turning good quality forest into struggling plantations.
- If you are a sawmiller, remember the advantage of a sustained supply of wood.
- If you are a farmer, think about farming in the forest, rather than replacing the forest
in order to farm.
- If you are a planner, work out ways to encourage:
- keeping existing trees in many different situations; and
- planting of plenty of trees to replace them as they mature.
- mixtures or pure stands?
Aren't forest trees usually planted in pure plantations?
They often have been, throughout the world. But large areas of a single species in the tropics
have come to be questioned (D 2, D 30), because they may be:
- more risky, in relation to severe damage from climatic extremes (D 11), fire (D 66),
and pests or diseases (Manual 5);
- less flexible, as regards the diversity of goods produced (D 33–40);
- less effective, in providing cover for the soil and protection against high rates of erosion
(D 23, D 30); and
- perhaps unsustainable, in terms of a managed ecosystem that recycles nutrients and
continues producing in perpetuity (D 10, D 13).
But aren't tree crops like rubber and oil-palm grown pure?
Yes, they often have been, though they are also sometimes grown in smaller plots (D 54), or
the products are collected from the wild.
Establishing them in large pure plantations may:
- uproot a lot of people from land that they were living on and using in other ways;
- require big inputs in terms of clearing and levelling the ground;
- involve cover crops and repeated fertiliser applications; and
- be disastrous if the price falls or companies pull out.
Can mixtures avoid these problems?
They can often help considerably (D 2, D 14), especially if combined with:
- avoiding large clearings (D 51, D 54) and soil compaction from heavy vehicles (D 12,
D 60, D 62);
- involving local people in decisions (D 5), and with trials (D 6); and
- appreciating that it is better to have several yields rather than relying on a single
See D 30 for the advantages and disadvantages of mixtures.
What kind of mixtures could I use?
There are an almost unlimited number of different kinds of mixture, but all of them can be
grouped into three types. Either one tree species is grown:
- at a different storey from other species (D 52);
- in small groups or lines amongst other species (D 54); or
- intimately mixed with other species (D 30).
Various combinations of (A), (B) and (C) can also be used.
Isn't this all too complicated?
Not really. For instance a farmer or a family with a home garden may:
- have some tall fruit trees, palms and a few timber trees;
- grow small plots of cereals and perhaps a cash crop of palms or shrubs;
- plant young trees for soil improvement, to replace old trees or for multipurpose uses
Are there some useful guidelines?
- Find out how many species grow together in any local natural or semi-natural forest
or savanna woodland.
- Try and mimic these conditions when growing trees which produce important items
that are traditionally collected from the wild.
- Think about the mixtures that seem to have worked well in your area.
- Read about newer combinations including ‘domesticated’ trees that have been tested
- Use the most reliable mixtures for most of your planting, but try out new ones on a
small scale as well (D 6, D 30).
What mixtures should I avoid?
Those for instance in which one species:
- casts too heavy a shade for the others to thrive;
- has a very aggressive root system; or
- produces chemicals that inhibit the growth of other species (D 14).
Aren't there some places where pure stands are best?
Yes, there are. Such situations include:
- sites where a single tree species dominates the natural vegetation, for example in rocky
sites and coastal mangroves (D 26), and in some freshwater swamps and savannas
- the early stages of reclaiming degraded land (D 22), or when making wind-breaks
- to establish fuelwood plots (D 35), or for industrial plantings to supply paper mills, etc
- in small stands within agroforestry systems (D 3) or to establish an overstorey (D 52) of
shade trees (D 41);
- many kinds of experimental plots (D 6, D 29, D 55); and
- some ‘tree banks’, in which the genetic diversity within a species is being retained for
the future (Manual 2).
Could mixtures yield enough to keep a factory going?
Managers might prefer the idea of large plantations producing a single, uniform product.
However, they very often have to adapt to dealing with different species, variable quality or
In any case, successful mixing of tree species together can still produce large quantities of a
product, for instance from many small or moderate-sized pure plots within a mixed planting
pattern (D 54).
- choosing the planting pattern
What is meant by ‘the planting pattern’?
There are three interlinked aspects of planting patterns:
- The scale at which the tree planting is to be done;
- The arrangement of the trees to be planted; and
- How far apart the individual trees are planted (Manual 5).
How about the scale of the planting?
This depends a lot upon the circumstances. Planting areas can be prepared on a:
- small scale, of the order of 10–100 m2 (up to 10-2 ha);
- medium scale, of a hectare or so;
- large scale, involving hundreds of hectares; or a
- very large scale, covering many square kilometres.
Who might use a particular scale?
- Individuals often want to plant on a small scale;
- Communities may decide to establish medium scale plots;
- Local or international institutions could undertake large scale planting;
- Private companies or government departments might operate on a very large scale.
Which is best?
Larger scale planting can offer some advantages, including:
- having better access to professional advice, information, seeds, etc;
- being able to employ a team of staff with different skills;
- being capable of tackling a large scale problem promptly, such as the reclaiming of
large areas of degraded land, or the planting of extensive shelter-belts.
But smaller scale tree planting can also have major advantages, such as:
- being better adapted to tropical ecosystems;
- fitting the planting into existing vegetation and land-use;
- accommodating local traditions and needs;
- planting all the trees at the most favourable season;
- allowing more care to be taken over the planting;
- being able to look after the trees regularly;
- involving a manageable amount of work.
And what are their disadvantages?
Larger scale schemes can suffer from:
- integration with other land-uses being more difficult;
- poor communication (D 4), and antagonism from local people (D 5);
- the risk of policy being directed from outside the region;
- sudden interruption of financial support.
Smaller scale projects may have problems with:
- lack of a key person to correlate the work (D 4–5);
- getting hold of relevant information, appropriate seed, good plants, etc. (but see D 70–72);
- training the people involved, so that they gain new skills;
- insufficient resources to buy even small items.
But won't larger scale schemes achieve more planting?
As well as the scale of planting, the total number of trees successfully established in a region
also depends critically on:
- how many people are planting trees;
- whether the trees are of appropriate species (D 30–42) and genetic origin (Manuals 1
and 2); and
- what proportion of them establish, grow well and reach maturity.
So, many small scale schemes could perhaps achieve more than one very large project.
In what different ways can trees be arranged?
They can be put:
- in one block;
- as small groups scattered within other vegetation or farmland;
- as lines or strips throughout other vegetation, along rivers, on contoured terrace edges
(D 65), as shelterbelts (D 41) or forming boundaries (D 21); or
- as single trees, in a variety of situations.
Does it make a lot of difference?
Yes, it does. Choosing the most suitable pattern can allow the tree planting to be fitted into:
- the existing vegetation (D 2);
- what the land is being used for (D 20–28);
- the requirement for keeping the soil protected (D1, D 50, D 60) from soil erosion and
degradation (D 22–23);
- the need for increased instead of declining yields (D 3);
- the ecological factors (D 10–16) that permit yields to be sustainable (D 5);
- continuing, local supplies of many different products (D 33–40);
- the growth responses of the particular tree species (Manual 3) and
- the felling cycles and rotations for the different trees (D 4, D 24, D26).
Which way fits in best with how trees grow naturally?
A few pioneer tree species regenerate naturally in the larger clearings that occur when there
is a fire (D 16, D 63), destructive storm (D 11), etc. The growing system closest to this would
be planting a single-storey stand (D 51) in one moderate-sized block.
Many tree species regenerate naturally in gaps where a big tree has fallen. The growing system
closest to this is successive planting in small groups within a stand of larger trees.
How big should groups be?
The smallest effective group is about 25 m2, allowing 9 trees to be planted in a gap that one of
them may in time fill. The largest is around 2500 m2, beyond which a gap is really becoming
a small clearing.
Isn't it troublesome organising many small groups?
Not necessarily, especially if individuals living nearby take responsibility for each set of small
groups of young trees.
What about line planting?
In the past this was used for turning forest into plantations, keeping the shelter of some of the
bigger trees and the coppiced stems of others to protect the soil, but only while the young trees
were growing up.
Line enrichment planting of an existing stand with desirable tree species is more suitable,
because many of the larger trees are retained instead of all of them being cut or poisoned.
How wide should strips be?
They can vary from a narrow band, two or three trees and a few metres wide to a broader strip
30–50 metres wide. When larger than this, the strip begins to take on the features of a block.
Are there some useful guidelines for choosing patterns?
Yes, here are a few:
- Go and look into (and if possible get above) any natural or semi-natural local forest or
savanna woodland, and see what patterns have stood the test of time (D 10, D 14).
- Take into account the topography and local soil variations (D 12, D 60) when choosing
where to open up the canopy (D 51).
- Leave some big trees of good form around any cleared areas, in order to:
- provide male and female parent trees to produce pollen and fertile seed (Manual 2);
- maintain nesting and feeding sites for useful wild animals (D 15);
- protect the cleared area from wind and fire (D 66).
- A year or two before a farm under shifting cultivation returns to tree fallow, enrich it
by planting a variety of trees (D 3).
Is there anything special about growing systems for research?
Yes there is. The choice of the growing system depends a lot on:
- what is being studied (D 6, D 29);
- the mixture of species being grown (D 30, D 53);
- to what extent trees are already present (D 2); and
- the layout of the experiment, especially how to replicate treatments in diverse terrain,
soils and vegetation.
What is a layout?
A plan showing which trees, or plots of trees, in the experiment are allocated to particular
treatments and Blocks.
What is a treatment?
In ordinary language, it is something done to a stand or to individual trees, such as weeding or
mulching, for instance.
In an experiment, a treatment is something specific that is done to a group of trees, but not done
to the controls (D 6).
In field experiments, treatments are often applied to a specific plot of land, or to particular
groups of trees. Sometimes, however, trees scattered throughout an experimental area can be
allotted to each treatment.
How big should the plots be?
- For investigating genetic differences, types of planting stock and modifications to
site (D 29): from about 100 m2 to 2500 m2.
- For studying methods of site preparation: about 2500 m2 to 1 ha.
- When comparing contrasting sites: here the main aim is to find areas that are as
comparable as possible, except for the aspect(s) to be studied.
How many plots do I need?
Enough for adequate replication of each treatment, and of the control. If these only occur once
in the experiment, it may not be possible to separate the effects of treatment from chance
variation between plots.
So all the plots are generally replicated in Blocks, giving a total number of 20–50 plots, with
six being an absolute minimum.
What is a block?
In ordinary language, trees are sometimes planted in a sizeable block, instead of in small groups,
lines or strips (D 54).
In a field experiment, a Block consists of a complete set of plots (control plus one of each
treatment). Having several Blocks means that controls and all treatments can be properly
replicated and randomised.
How many Blocks shall I need?
Between four and six are often used, with a minimum of three.
Do I need space for anything else?
Yes, you may need:
- Extra trees surrounding the plots and Blocks, to prevent them influencing each other,
and to reduce edge effects; and
- Spare trees to try out the treatments on.
How do I decide where to put the Blocks?
- Walk the ground again, to see what it is like in detail;
- Make a rough sketch map;
- Find the most similar part of the site for each Block, keeping its plots as close to each
other as possible. On a slope, for example, Block I might be on the highest ground and
Block V at the bottom.
But I thought everything had to be randomised?
Which plot receives which treatment within a Block is allocated at random;
Where the Blocks are within the experimental area is deliberately chosen.
Why is this?
Randomising the plots avoids bias over which treatment goes where (D 6).
Choosing where the Blocks go improves the precision of the experimental design by reducing
the amount of unassigned variation (Manual 5).
If the positions for all the plots were completely randomised, one might end up with most
replicates of a particular treatment clumped together in one part of the experimental area.
Suppose I want to compare mulch with fertilisers?
Here you have two different factors, and a good way of starting is to do a simple 2 × 2
(0) control (no application);
(1) mulch only;
(2) fertiliser only;
(1+2) mulch plus fertiliser.
This allows you to:
- keep the experiment reasonably small;
- compare the effects of each factor alone; and
- discover any interactions (D 6): for example, were the effects of fertiliser dependent on
whether the plants were mulched or not?
Interactions will be discussed further in Manual 5.
What about preparing the site?
This is covered in sheet D 67.
Does all this apply to informal trials?
In practice, informal trials are often:
- more short-term;
- intended to answer an immediate practical problem.
However, all the statistical principles described here are still relevant to setting up an informal
trial in such a way that it:
- is not a disappointment through failure;
- doesn't give misleading indications; and
- won't turn out to be a waste of your time.
How would I set up an informal trial of tree species?
Here are some hints:
- You could choose a local species as a ‘control’ to compare the others against. If possible,
utilise a frequently used, local seed-lot or known clones for this.
- Make sure that each of the species to be compared with the control is representative,
rather than being taken from a single source;
- Don't try and include too many species, so that the trial spreads over too much land,
or becomes too big to manage.
What would be a simple layout?
Choosing a ‘control’ plus 3 other species means that you could design a useful trial that had:
- 3 square Blocks, perhaps as small as 15 × 15 m or 25 × 25 m;
- Each Block containing 4 plots (species), so that there were only 12 plots in the trial;
- Each plot consisting of 9 plants, so that there were as few as 108 trees altogether. Later
on, one tree from each plot could be allowed to grow on to a larger size.
How about marking out the experiment?
This is best done first with temporary poles or pegs that can be moved around until their
positions are finalised.
When treatments are to be applied, coloured tags are helpful to reduce the risk of mistakes.
Soon afterwards, permanent marker posts and labels will be needed (Manual 5).
Do pot experiments need a layout?
Although they can be moved about, it is important to randomise where each pot stands, to
avoid bias. In particular, try to:
- keep shading as uniform as possible;
- have all treatments equally represented amongst edge plants;
- avoid larger trees shading smaller plants;
- be careful about watering (A 51 in Manual 1), and don't let the trees root through into
the ground (A 54).
Pot experiments are to be further discussed in Manual 3.
What about layouts for demonstration planting?
In field experiments, the randomised layout often makes it difficult to look at treated and control
trees at the same time. For demonstration plantings, you are not doing an experiment, but
making an exhibit of the most important differences that have been found earlier on. So you
- choose an easily accessible site, and plant close to a road or track;
- use short lines or small plots that can easily be seen and photographed;
- arrange the layout so that the most important comparisons are demonstrated - and
don't forget the controls!
- don't concentrate only on the best - also include some examples of poor performance
to show the contrast (for instance A and D in the diagram).