Previous Page Table of Contents Next Page

Rainforest treatment


GEORGE N. BAUR is a research forester, Forestry Commission of New South Wales, Australia; and was an FAO André Mayer Fellow, 1961-62. The FAO-André Mayer fellowships are awarded for the promotion of research of interest to member countries and having a relationship to the Organization's program of work. They are intended either for experienced research workers who need to complete more or less independently a particular line of research, or for younger individuals who have demonstrated an aptitude for research work and who will benefit from further training in methods at recognized institutes and in field studies.

Since these fellowships were started in 1966, a total of 75 swards has been made, of which nine have been in the field of forestry and forest products.

Some observations of an FAO André Mayer Fellow

RAINFOREST can be defined as a tall, dense plant community, composed essentially though not invariably of moisture-loving, evergreen, broad-leaved trees which tend to be arranged into a number of more or less distinct layers or stories. The trees may show curious stem forms such as buttressing, stilt-roots, and stem fluting; shrubs are often plentiful in the under-story, while climbers and epiphytes are usually frequent.

Within quite recent historical times vegetation of this type has covered nearly 10 percent of the earth's land surface and, although much has been destroyed during the past century or so, rainforest still occupies very large areas in the humid tropics, notably in tropical America, in west and central Africa, and in the southeast Asia-Pacific region. Many countries within these regions rely upon the rainforests to provide timber and a host of other products not only for their local requirements, but also as valuable items of external trade. It is therefore not surprising that in many different areas attempts have been made to treat the rainforests silviculturally, with the view to bringing them to a state where sustained yield management can be practiced (Figure 1).

In addition approaches have also been made to bring the rainforest to a more readily managed condition by more overtly artificial means, such as direct clearing and planting, the taungya system of agrisilviculture, and enrichment planting: these approaches are not considered here, though their significance is well appreciated.

Attempts at managing the natural stands have been made along different lines in different countries. Many of these attempts appear to have met with appreciable success, and it is instructive to compare these approaches. However, before an appraisal of these methods can be made it is necessary to have some understanding of the growth processes of natural undisturbed rainforest: the maxim that "silviculture is applied ecology" is a basic truth of forest science.

Growth processes

Flowering and fruiting

A mature patch of undisturbed rainforest typically contains several recognizable layers of plants, and for simplicity these can be termed the overstory of very tall, large trees; smaller trees forming the understory which may in fact consist of several independent layers of trees; a low shrub layer; and finally a ground layer of herbaceous plants and seedlings. In such a patch the climbers usually contribute to the canopy formed by the overstory trees.

The flowering habits of plants growing under such diverse microclimatic conditions can be expected to be most varied. Some species flower almost continuously: these are mostly understory and shrub species and species characteristic of recently disturbed sites, the so-called "secondary species" such as the west African Trema guineensis (Taylor, 1960). Few overstory trees, in which foresters are usually most interested, are of this type. These larger trees at best tend to be seasonal flowerers and frequently they flower at very irregular intervals. Many Malayan dipterocarps, for example, flower only at intervals of from two to seven years, while the valuable African Triplochiton had only five good flowering years in the period between 1933 and 1958 (Mackenzie, 1961).

After flowering, the fruits of most rainforest trees ripen rapidly, usually within six months, and in markedly seasonal rainforest climates there is a tendency for the seed to be shed towards the start of the wet season. Seed dissemination occurs in various ways, though distribution by birds and other animals is common in all stories. Wind dissemination is largely confined to the overstory species, including many vines.

Most species retain their seed viability for very short periods only. Many dipterocarps, for instance, lose all viability within three weeks of collection (Barnard, 1954). However, some large fruited species such as the Brazilian Bertholettia and the Australian Gmelina spp. may germinate over a period of years following seed fall, while there is evidence that some "secondary species," such as Musanga cecropioides, build up in the soil a store of seeds which only germinate on exposure to warmth and light (Keay, 1960). Conversely, some species are inhibited from germinating in strong light, the seedlings only appearing in shaded conditions: Guarea cedrata behaves in this way (Gilbert, quoted by Jones, 1955-56).

Seedling development

The seedlings of rainforest trees can be grouped into three main classes on the basis of their subsequent behavior: these can be termed the "secondary species," the truly tolerant species, and the gap opportunists. These classes are not entirely clear-cut, but they are sufficiently distinct to provide a useful key to the understanding of rainforest regeneration.

The "secondary species" require almost complete light for survival and growth (and, in most cases, for germination also). They are usually extremely fast growing, at least in their early stages, and are frequently protected by stinging hairs (e.g., Laportea spp.), thorns (e.g., Fagara spp.), or myrmecophily (e.g., Cecropia spp., Barteria spp., Macaranga spp.). Most are short-lived and only grow to small trees but some, with slower later growth, can become very large trees (e.g., Goupia glabra, Laportea gigas). For the most part they are of very limited economic significance though there are some notable exceptions such as Ochroma lagopus, yielding the balsawood of commerce; Didymopanax morotonei, used for match veneer in Trinidad; and Cecropia spp., being used for paper pulp in Peru.

The truly tolerant species are typically those of the understory, requiring considerable shade for their early development. They include, however, some trees which ultimately reach the overstory, such as many Lauraceae including the highly valued Greenheart, Ocotea rodiaei, and Queensland walnut, Endiandra palmerstonii. Some of the slower growing dipterocarps such as Balonocarpus helmii; and the Balau group of Malayan Shorea spp. also appear to belong in this class. Growth rates of these species are usually slow (Figure 2).

The gap-opportunist species include many of the more valuable trees that ultimately make up much of the rainforest overstory, and for this reason this group tends to be of the greatest interest to foresters. Among the species belonging to this class are many Meliaceae, Dipterocarpaceae, and Flindersia spp. Like the tolerant species, the seeds of this class can germinate on the undisturbed forest floor, but unlike the previous class the seedlings make little further development unless they subsequently receive increased light: instead they go into a state approaching dormancy, but retain the ability to commence active, and frequently rapid, growth if light conditions are suddenly improved. With reliable annual seeders (e.g., Terminalia superba and T. ivorensis) this dormancy may only last a year by which time most seedlings have died, only to be replaced by a new crop from the next seed fall. With many irregular seeders, such as the Meranti group of Malayan Shorea spp., the seedlings survive for longer periods, sometimes as much as 10 years. The result of this behavior is to establish on the forest floor a reservoir of regeneration which is available at all times to take advantage of any opening in the forest canopy. The quantity of this regeneration varies from time to time and in different areas, but an idea of its extent is given by some studies in virgin dipterocarp rainforest in North Borneo where over a period of years there was an average stocking of 10,000 desirable seedlings per acre (25,000 per hectare), with a minimum during this period of 4,000 (10,000) seedlings (Nicholson, 1968). Seedlings of these species, when released from dormancy, generally grow at a somewhat slower rate than the "secondary species" and thus, during the short life-span of the latter, tend to be retarded, suppressed and sometimes deformed by these. However, as the "secondary species" die it is usually found that a reasonable stocking of the gap-opportunists survives to take over dominance of the patch of regrowth. A few gap-opportunists, such as Shorea leprosula, are capable of growing as fast as most of their "secondary" associates.

FIGURE 1.- View of part of African Timber and Plywood (Nigeria) Ltd. mill at Sapele, Nigeria, showing sawn timber and plywood being loaded for export to Europe.

FIGURE 2 - Clump of young Meranti tembaga (Shorea leprosula) some 18 months After release from dormant condition by logging and poison-girdling. Note that growth of the shorea is equal to that of the associated "secondary species". Pasoh forest reserve, Malaya.

This ability to survive for lengthy periods while making negligible growth, and then to respond rapidly to improved growing conditions, is a major silvical characteristic of many of the most desirable rainforest trees. It is a widespread characteristic, and is usually retained throughout much of the species' life, enabling them to withstand periods of suppression and then to resume active growth when conditions improve. It is, however, not confined to desirable species; but is found also in many vines such as the serious Nigerian weeds Acacia ataxacantha and Calamus spp. and related climbing palms.


The basal area of mature stands of virgin rainforest ranges in different types of rainforest from under 100 to over 300 square feet per acre (23 to 69 square meters per hectare), with values of 160 to 200 square feet per acre (36 to 46 square meters per hectare) common in many better quality rainforest areas. Growth in such saturated stands is, as aptly described by G.G.K. Setten, 1 a study of dynamic stagnation: any increment is balanced over a period of years by mortality in the stand. Under these conditions the growth rates of individual stems vary tremendously: over periods of 15 or more years one tree may average 0.6 inch (1.3 centimeters) of diameter growth or more a year, while a nearby stem of the same species and of similar initial size may show virtually no diameter increment. These great differences are due to various causes, notably the health and size of the trees' crowns and the extent to which the crowns are free from interference (R.W.J. Keay, pers. comm.; Dawkins, 1956, 1958).

1 Formerly Chief Research Officer, Forest Research Institute, Kepong, Malaysia.

Despite these variations there can be recognized a general relationship between the diameter and the increment rate for various species (Jones, 1955-56; Keay, 1961; Schultz, 1860). Many of the gap-opportunist species tend to reach their maximum rate of diameter growth in the intermediate sizes, from 10 to 26 inches d.b.h. (26 to 63 centimeters), the growth curve rising in the smaller size classes and falling off in the larger sizes. This differential growth rate means, first, that in conjunction with the variation between stems of similar size, size classes are most unreliable indicators of age classes in rainforest. Second, it means that trees pass through the intermediate sizes faster than through the smaller and larger sizes. Thus, even if the same number of trees are being recruited annually, there will be an apparent deficiency of stems in the intermediate sizes. This deficiency has been widely recognized, especially for certain important west African species (Khaya ivorensis, Entandrophragma cylindrica, Terminalia superba), but the growth rates indicate that in many cases it is a deficiency more apparent than real.

Growth pattern in virgin rainforest

With this background the general pattern of growth in virgin rainforest can be understood. The uppermost layer is composed of large trees, often bound together with climbers: these species are typically light-demanders, and tend mostly to be gap-opportunists in their regeneration habits. Below this are several further layers of trees and shrubs, mostly tolerant species but including immature and suppressed stems of light-demanding species. On the ground is usually a reservoir of dormant regeneration, while within the soil may be a further store of viable seed unable to germinate until light (and probably temperature) conditions improve.

The overstory the most critical of these, since it dominates all others. Three alternatives can ultimately occur in this story. First, one of the trees may die and break up where it stands. As the branches collapse and fall a small gap is created below. This is too small to permit "secondary species" to enter, but it is usually sufficient to release from dormancy the gap-opportunist regeneration beneath, whilst smaller trees adjacent to the gap obtain improved growing conditions and are able to speed up their rate of growth, often assuming dominance of the gap.

Second, and probably by far most common, one of the large trees blows over during a storm. Often large crowned, and laced to its neighbors by vines, this stem tears and crashes to the ground creating a larger gap which is frequently up to an acre in area. In this gap the dormant regeneration is released and secondary species enter, producing a small patch of even-aged regeneration.

Finally some catastrophic occurrence such as a severe cyclonic storm may Batten an extensive area which is then occupied by secondary species and by the release of dormant seedlings. Disturbances of this type have been widely recorded, e.g., the Malayan "storm forest" in Kelantan (Browne, 1949), and are paralleled by heavy, uncontrolled logging, such as occurred in parts of Malaya during the Japanese occupation (1942-45).

Each of these three occurrences results in the creation of a gap, and although the type of regeneration appearing tends to be different in each case, all produce a patch of relatively even-aged regeneration to replace the original tree or stand. Thus virgin rainforest can be regarded in most areas as a patchwork, or mosaic, of smallish, even-aged stands. Each patch is of different size and composition and, because of the differential growth rates, the boundaries between patches soon become indistinct a favored, fast-growing stem in a new patch soon becomes larger than slower stems in an older patch. The units of the mosaic are not permanent, new gaps very rarely coinciding with their predecessors, while the species occurring in the gap and ultimately replacing the old tree appear to be determined largely by chance factors and may or may not be those found in other gaps in the area.

Treatment of rainforest

Silvicultural treatments applied to natural rainforest are of two basic types. The first type is improvement, and the second can be termed regeneration establishment, though it usually covers greater scope than just establishing regeneration.

Improvement treatments are a passing phase in the conversion of probably all previously unmanaged stands to managed forest, e.g., the timber stand improvement (T.S.I.) treatments applied to pine-hardwood forests in the United States and to eucalypt forests in Australia. Because rainforest management is of relatively recent introduction, improvement operations are an intrinsic part of treatment even where the main emphasis is on regeneration establishment. The operations are concerned primarily with removing stems which, because of faultiness, poor form or species, are locally unmerchantable. This destruction of useless basal area allows the remaining desirable stems and regeneration to grow at a faster rate than is otherwise possible and, while it is usually carried out in conjunction with regeneration establishment, it can be applied to areas unlikely to be logged and brought under regeneration treatment for a long period. This permits the potentially useful stems to attain a greater volume when logging is finally carried out, and it leaves the forest in a better condition for subsequent regeneration treatments. Uniformisation par le haut in the Congo (Donis and Maudoux, 1951) was a treatment of this type, and a similar treatment has recently been recommended for certain forests in Malaya (Wyatt-Smith, 1962).




1. Forest not currently accessible for exploitation


1x. Forest available for exploitation now or in near future


2. No finance available for treatment

Reserve and protect

2x. Finance available for treatment

Improvement treatment (e.g., Congo: uniformisation par le haut)

3. Management for indirect benefits (e.g., watershed protection, recreation) paramount

Selection system (e.g., Puerto Rico)

3x. Management for timber production paramount


4. Intermediate size classes plentiful; royalty rates make retention of these desirable


4x. Intermediate size classes relatively scarce


5. Severe opening of stand deleterious

Selection system (e.g., New South Wales)

5x. Severe opening not deleterious

Group selection (e g., North Queensland)

6. Desirable regeneration adequate in virgin forest, or occurs readily with exploitation


6x. Regeneration not naturally adequate


7. Regeneration capable of responding to sudden and complete increase in light and exposure

Clear-cutting (e.g., uniform system of Malaya, North Borneo)

7x. Regeneration needing partial shelter for some years

Post-exploitation shelterwood (e.g., T.S.S. of Trinidad)

8. Regeneration induced by canopy opening and cleaning


8x. Regeneration not readily induced naturally

Artificial regeneration, possibly combined with some other type of treatment (e.g., Reunion; North Queensland in part)

9. Regeneration, once induced, responding to complete light and exposure

Pre-exploitation shelterwood (e.g., T.S.S. of Nigeria)

9x. Regeneration after inducement still requiring shelter for some years

Extended shelterwood (e.g., Andamans)

FIGURE 3. - Destroying unwanted rainforest stem by frill-girdling and applying to the frill a solution containing 2 lb (1 kg) sodium arsenite to 1 gallon (4.5 liters) of water. Kemasul forest reserve, Malaya.

Establishment treatments, along with the subsequent care of the developing crop, are usually regarded as forming the main part of the various silvicultural systems which have been developed in rainforest areas. As noted above, these currently have a large content of improvement operations in the scheduled treatments, though in subsequent rotations this content should greatly diminish. The systems in use can be divided into those producing an even-aged stand, and those maintaining an uneven-aged, irregular type of forest.

The type of treatment applicable to any particular rainforest area is determined partly by local economic and policy considerations, and partly by the silvicultural characteristics of the main species present in the stand, as discussed earlier. In Table 1 an attempt has been made to set out the various economic and silvicultural features that may occur, and to indicate the type of treatment which appears most suited to each combination of features. In the table, items 1 to 4 are primarily economic and items 5 to 9 silvicultural: the high place given to economic considerations is not coincidental, but reflects the pre-eminence of economic factors in determining soundly based forestry activities.

Treatment operations

Any treatment applied to rainforest is made up of various silvicultural operations. Some of these may be applied together or they may all be given independently, but in all cases a relatively small number of distinct operations are found to be involved.

Soil treatments to make the ground more receptive for germination are rarely given in rainforests, though the disturbance caused by tractor logging often amounts to such a treatment. In North Queensland, local patches deficient in desirable regeneration may be raked clear of litter around seed trees in order to induce regeneration (Anon., 1954).

Canopy opening is a very important operation in most rainforest areas. Four phases of this can be recognized:

1. Climber cutting (CC)². This is required in most areas, and serves several purposes including facilitating access through the stand, creating an appreciable lightening of the canopy, and destroying a serious weed group.

2. Exploitation (F). This is normally the operation resulting in the greatest opening of the canopy. In addition to the removal of the felled stems, further opening usually results from the damage caused to adjacent trees by the falling stem.

² The symbols CC, F. Ru, Ro, etc., are used in Table 2 to indicate these venous operations.

3. Understory removal (Ru). The tolerant understory trees frequently possess very dense crowns, and removal of these is necessary both to induce and to release regeneration. Removal is most commonly accomplished by frill-girdling and poisoning with sodium arsenite solution, though hormone-type poisons (e.g., 2, 4, 5-T), ring-barking without poisoning, and even felling the smaller stems are also used.

4. Overstory removal (Ro). This is an essential operation to remove the large useless stems which otherwise occupy considerable growing space. Frill-girdling and poisoning is a method commonly used (Figure 3).

Artificial regeneration (P), by way of enrichment planting or more rarely sowing, e.g., in Reunion (Miguet, 1955), may be used in treatments that are nonetheless primarily applied to natural stands (Figure 4). The aims are twofold, first to regenerate local patches where it has not been possible to induce natural regeneration, e.g., North Queensland (Anon., 1954), and second to innoculate the area with valuable species which are rare or do not occur naturally in the stand, but which it is hoped will regenerate naturally in subsequent rotation, e.g., Trinidad (Moore, 1957).

Undergrowth cleaning (C) is widely used to remove shrubs, stemless palms, tall herbs, and other unwanted undergrowth likely to interfere with the establishment or release of regeneration. However, because studies have shown that the operation frequently results in the destruction, by accident, of much desirable advance growth, there is a present tendency for the operation to be confined to those areas where the density of undergrowth is most likely to preclude the establishment of natural regeneration, e.g., Malayan ridges with a dense ground cover of the stemless palm, Eugeissona triste.

Liberation treatments are intended to maintain the growth of established regeneration at a rate approaching the maximum by removing useless or less valuable stems interfering with the growth of more desirable stems. Two types of operation can be distinguished:

1. Removal of impeders (L). This is usually a selective operation applied to young stands, in which unwanted stems, e.g., many of the fast-growing "secondary species," are removed where they overtop or otherwise compete with a desirable sapling.

2. Thinning (T). This is applied to older regrowth stands and aims at favoring the best stems in the stand by removing less satisfactory stems, of the same or other species, where these are in competition.

Diagnostic sampling (S), the final operation, is not truly silvicultural. However, under the influence of Malayan experience (Barnard, 1950), it is now widely used at various stages of treatment to indicate the need for various alternative sequences of operations. It-is usually accomplished by establishing transects of small plots (milliacre, 1/160 acre, 1/40 acre) through the area under treatment and recording the composition and condition of the regeneration on each plot. The results can then be used to diagnose the type of treatment required in the area (Barnard, 1950; Dawkins, 1958; Wyatt-Smith, 1960).



Clear-Cutting (Malaya)

Pre-exploitation shelterwood (Nigeria)

Post-exploitation shelterwood (Trinidad)

Extended shelterwood (Andamans)


70 yrs

100 yrs

60 yrs


n - 5

(Ru) (1)

S. CC, Ru

C, Ru* (9)

n - 3

(C) (2)

C, Ru* (9)

n - 2

(CC) (3)

C, Ru* (9)

n - 1




S, F. Ro, Ru (4)



F. Ru*

n + 1

Ru*, Ro* (P) (8)

L, C, CC

n + 2

(L) (5)

CC, L, Ro


L, C, CC

n + 3


L, C, CC, Ro

n + 4


n + 5

S. (CC, L, T. Ru) (6)

CC, L, Ro

n + 6



n + 9

S (7)

n + 10

S (CC, L, T. Ru) (6)

+ Based on moat characteristic sequences in use: operation symbols as indicated in section dealing with "treatment operations" plus: * Operation of selective nature only (i.e., partial, not complete) ( ) Operation given in selected sites only, or if shown (usually by S) to be necessary

NOTE: (1) Given only in selected areas (viz., somewhat seasonal rainforests of northwest) where regeneration is difficult to establish. (2) Given in forest infested with dense stemless palms or bamboo. (3) Given in forests where vines are particularly thick (usually old shifting cultivation areas). (4) First 8, then F, followed by removal of unwanted stems by poisoning. (5) Given only where aim is to produce the slower-growing "heavy hardwoods" (e.g., Balanocapus -, certain Shorea spp.) which require more assistance: rarely used in practice (6) S at 6 yrs (¼ ch ² plots) and 10 yrs (¼ ch ² plots) is followed by such operations as are indicated to release the more desirable stems. (7) 8 at n + 9 is proposed, and will probably serve similar purpose to Malayan S at n + 10 (see (6) above). (8) Canopy Opening is Partial to produce a light shelterwood, enrichment planting may be employed to inoculate area with stems of valuable exotic species (e. g. Chlorophora, Terminalia spp.) primarily as seed bearers for future rotations. (9) Initial operations are timed to occur with good seed years, which occur at about 4-year intervals: hence 4-6 annual crops receive initial treatment (C, Ru*) at the same time, with subsequent operations occurring in the prescribed working plan years n - 1, n etc., for the compartments concerned.

FIGURE 4. - Queensland maple (Flindersia breyleyana), six years old, planted in rainforest under treatment in a gap which lacked natural regeneration. Reserve 99, Western Queensland.

FIGURE 5. - Unloading logs extracted from rainforest at riverside for transport by water to mill at Sapele. Okomu forest reserve, Nigeria.

FIGURE 6. - Regeneration of Kapur (Dryobalanops aromatica) making rapid growth after release from dormancy by recent logging and poison-girdling. Kluang forest reserve, Malaya.

FIGURE 7. - General view of area logged and poisoned about 18 months earlier under the Malayan uniform system. Note the dead, poisoned stems and the retained stems of immature trees, with dense regrowth of "secondary species" (including much Thema op.) and desirable regeneration. Pasoh forest reserve, Malaya.

Regeneration of even-aged stands

In most areas where a sincere effort is being made to manage the natural rainforest the emphasis is on producing even-aged stands. The reasons for this emphasis are varied and include:

1. Lack of success under old selective logging methods. Most early rainforest logging was of an uncontrolled, highly selective nature. This was stand degradation, not silviculture, but the fact that it failed to produce worthwhile regeneration has created a bias against selection systems of management in many areas.

2. The development of mechanical logging methods, which for economic working require a high yield from each hectare logged and which, in the normally low-yielding rainforest, tend to result in as heavy a logging operation as the stand will permit (Figure 5).

3. The greater ease of management in even-aged stands.

4. Damage to other stems during logging, making it desirable to remove all stems rather than to leave damaged stems standing.

5. Rising standards of utilization, so that the forester is less frequently faced with the need to destroy large stems of good form but of currently unmerchantable species.

6. Realization that most desirable species are light-demanders which are well suited to regeneration in even-aged stands (Figure 6).

Four basic systems of regenerating rainforest in even-aged stands can be recognized (see Table 1, items 7, 7x, 9, and 9x). In three of these the creation of a shelterwood is regarded as essential either to induce the establishment of regeneration or to foster it over the first few years, while in one no shelterwood is required. Details of typical treatment sequences currently applied in these four systems are given in Table 2, though in few cases are these sequences rigid, the treatments varying with local conditions and with the results of diagnostic sampling. It should be noted that, in all cases, immature stems of desirable species are retained so that the treated forest never has the appearance of a completely clear-felled, even-aged stand (Figure 7).

It should also be noted that in those Malayan areas where understory removal is carried out ahead of exploitation ³ the system is no longer clearcutting, but rather a pre-exploitation shelterwood. In Nigeria the shelterwood is formed 5 years ahead of exploitation and removed shortly after exploitation; in Trinidad it is formed at exploitation and removed some 5 years later; and in the Andamans it is formed at the time of seedfall some 2 to 5 years ahead of scheduled exploitation, and retained for about 3 years after exploitation. Modifications of these systems are used in other areas besides those quoted in Table 2.

³ See note (1), Table 2.

Regeneration of uneven-aped stands

Although even-aged systems are undoubtedly more simple to apply and have been shown over a wide range of conditions to be well suited to rainforest management, uneven-aged systems are employed in a number of areas for very sound reasons. These examples are in addition to the still considerable areas where uncontrolled selective logging occurs with no real attempt to manage the rainforest.

Reasons for favoring uneven-aged stands include:

1. The ability to retain small, healthy stems of low current value. Such stems are in fact retained in most even-aged systems, but particularly where they are numerous and where the royalty rates provide a higher payment for a given volume of larger logs, there is a strong incentive to retain these immature stems in such quantity that even-aged management is not practicable.

2. The ability to retain the existing size-class distribution of the forest.

3. The protection afforded the soil from exposure and possible erosion. Against this, even in the drastic clear-cutting system of Malaya, the soil is rarely bared for any length of time.

4. Avoiding the risk of destruction from cyclones, even-aged stands tending to be more vulnerable than uneven-aged stands (Wadsworth and Englerth, 1959).

5. Better protection against other climatic hazards. This is probably of greater moment in temperate rainforest than in tropical rainforest (Baur, 1962).

6. The maintenance of a more attractive looking forest, a feature of some importance where the forest has high recreation value.

7. The opportunity for retaining species currently unmerchantable, but for which a market may develop within the felling cycle.

The main disadvantages of uneven-aged systems, as already mentioned, are: (a) that most important rain-forest species tend to be light-demanders; (b) the desirability of having few concentrated logging operations rather than frequent light operations; (c) the damage occurring during logging to stems intended for retention (Dawkins, 1958).

Two examples of uneven-aged systems are probably sufficient to illustrate the main features. The most important point in both examples is the great emphasis given to improvement operations by the removal of unwanted stems, to the extent that the first example should perhaps be regarded as an improvement rather than an establishment technique.

Puerto Rico. Treatment in the Luquillo mountains is experimental rather than routine, though appreciable areas have been treated. The aim is to create a stand with a basal area of about 80 square feet per acre (18 square meters per hectare), balanced as regards size-class distribution and occurrence of species groups, and maintained by a cutting cycle of 5 or 10 years. The stands have mostly been degraded by past logging and, being in an important catchment and recreation area, care has to be taken to avoid excessive opening of the existing stand. Treatment involves the removal of trees where crowns are in contact, the avoidance of gaps larger than 25 feet (7.62 meters) in diameter, and allowing remaining stems about 6 feet (2 meters) of crown freedom on all sides. The emphasis in carrying out treatment is on the removal of the larger stems and of stems of the less desirable species or of poor form. Ultimately the stand would be managed as a true selection forest in which the more tolerant species could be expected to predominate (Wadsworth, 1947; 1952; 1957).

FIGURE 8. - Dense regeneration of Queensland maple (Flindersia brayleyana) resulting from routine treatment of rainforest 8 years earlier. Reserve 99, Western Queensland.

North Queensland. The rainforests of the Atherton tableland have much in common with those of the Luquillo mountains, both being classed as submontane rainforest and occurring in a cyclone belt. There is a fairly strong demand for timber in the area, while the stumpage rates give a strong price differential both between the most desirable species groups (e. g., Flindersia brayleana Figure 8) and less-sought-after species, and between large and small stems of a single species. The price differentials produce an incentive to retain smaller (though merchantable) stems which will become increasingly valuable per unit volume as they grow. This has led to the introduction of uneven-aged management which, after the initial stage with its emphasis on improvement is complete, will resemble the group selection silvicultural system widely used in Australian eucalypt forests (Jacobs, 1955). A cutting cycle of 15 to 20 years is envisaged, with the main sequence of operations being:

1. Climber cutting and removal of useless undergrowth.

2. Eradication of any seedlings or shrubs of Laportea moroides (a virulent stinging-tree).

3. Tree marking to stated minimum girth limits, but aiming:

(a) to retain any oversize stems of the more desirable species for seed where the existing regeneration stocking is low;

(b) to remove any undersize stems of poor form or vigor;

(c) to thin any dense patches of undersized stems.

4. Exploitation of the marked stems, followed by further eradication of L. moroides.

5. Removal of useless stems and, where necessary, to favor potentially more valuable stems, of stems of the less desirable species.

6. Soil treatment in areas where seed trees are present but regeneration is scarce, and enrichment planting where seed trees are also scarce.

7. Liberation after 3 to 4 years.

The girth limits, below which stems are normally retained, range from over 8 feet (2.43 meters), 30 inches (76 centimeters) d.b.h., for the most desirable groups down to 6 feet (1.82 meters), 23 inches (58 centimeters) d.b.h., for the least desirable.

Expected yields

Although these approaches to rainforest treatment differ considerably in outline, it is perhaps relevant to mention that all are based on adaptations of the way virgin rainforest grows and regenerates naturally: the clearcutting system in Malaya is no more "unnatural" than the selection system in Puerto Rico (Figure 9).



Rotation/cutting cycle

Yield from initial exploitation

New crop annual increment (1)

Yield (1)











860 (2)


30-60 (3,4)


2 100-4 200





400 (5)


60 (5)


6 000 (6)





1 300 (7)


89 (8)


5 000


Puerto Rico




60 (9)




North Queensland



c. 400 (10)


40 (11)




NOTE (1) Increment is M.A.I. for even-aged stands, P.A.I. for uneven-aged stands yield is obtained by multiplying the increment by the rotation or cutting cycle. (2) Vincent, 1960. (3) Cousens, 1967 - from comprehensive sampling of forest regenerated before 1940 in Perak, under the former "regeneration improvement felling" system (a form of pre-exploitation shelterwood system, which is a direct ancestor of the current Nigerian system) (4) The higher value was obtained from plots located in well-stocked stands of old regenerated forest (33-50 yrs of age). (5) Rosever, 1952. (6) The final yield includes an estimated 2 000 cu ft/acre (140 m³ /ha) obtained from Intermediate fellings. (7) Data supplied by D. Moore (pers. comm.); sawlog yield of 125 cu ft/acre and 16 cords (say 1200 cu ft/acre) (84 m³ per ha) of fuelwood. (8) Plot mentioned by Moore, 1967. (9) Sawlog increment only; Wadsworth, 1967. (10) Partial exploitation only. (11) Plots mentioned by Volck, 1960.

The likely yields of timber resulting from these treatments are hard to estimate: the oldest rainforest stand under management known to the writer is the magnificant Kapur (Dryobalanops aromatica) stand on Kanching forest reserve in Malaya, which was regenerated some 50 years ago, but in most areas serious management dates only from after the second world war. In Table 3 an attempt has been made to estimate likely yields in different areas.

While Table 3 shows annual increments up to 90 cubic feet per acre (6 cubic meters per hectare), it appears that over any extensive area treated rainforest will rarely produce more than 50 cubic feet per acre (3.5 cubic meters per hectare). This can be compared with increments of from 200 to 600 cubic feet per acre (14 to 42 cubic meters per hectare) commonly being obtained on rainforest sites with certain Pinus, Araucaria, and Eucalyptus spp. in plantations. Several reasons for these immense differences can be suggested:

1. Inherent characters which enable a small group of species (including many conifers and eucalypts) to produce much greater volumes than most other species. As an example, plantations of Flindersia brayleana in North Queensland (Figure 10) and of Tectona in India and Trinidad rarely produce a mean annual increment exceeding 150 cubic feet per acre (10.5 cubic meters per hectare).

2. Different standards for assessing volume growth: plantation stems are often assessed on the basis of yielding pulpwood, poles, etc., while natural stands are assessed on the production of relatively large sawlogs.

3. Different intensities of management: in plantations virtually all growth is concentrated on desirable stems, while even in managed rainforest much growth goes on useless species and undergrowth.

It seems that the absolute maximum production from most rainforest species is about 150 cubic feet per acre (10 cubic meters per hectare) per year, and on a routine scale in managed forest about a third of this is all that can be hoped for. The effects of compound interest soon indicate that, if the cost of rainforest treatment is to be recovered, this limited growth must be obtained as quickly as possible by concentrating the maximum rate of increment on what will prove to be the final crop trees: in even-aged stands these are commonly estimated to number about 30 to 40 trees per acre (2 to 2.8 cubic meters per hectare). The conclusion surely is that liberations are the most important phase of rainforest treatment.

FIGURE 9. - Stand of Meranti tembaga (Shorea leprosula), 34 years of age and with a volume mean annual increment of 60 cubic feet per acre (4.2 cubic meters per hectare) per year. This area was regenerated by the regeneration improvement felling system which was used in Malaya before the second world war and on which the present pre-exploitation shelterwood system of Nigeria is based. Pasoh forest reserve, Malaya.

FIGURE 10. - Plantation of Queensland maple (Flindersia brayleyana), 31 years old, established on rainforest site and showing volume mean annual increment of about 160 cubic feet per acre (11.2 cubic meters per hectare) per year. Observe the poor form of the trees grown in this way. Reserve 310, Gadgarra, Queensland.

While there is no doubt that in many parts of the world rainforests are being managed successfully and with profit, and while there is an undoubted case for the continued management of increasing areas for generations, yet the considerations outlined above strongly confirm the conclusion of Wadsworth (1960): " population and demands for land and forest products grow and as other types of agriculture progress, intensive plantation culture may prove to be the only economical source of timber in the tropics."


ANONYMOUS. 1954. Rules for tree marking, for logging and for natural regeneration treatment, North Queensland rain forests. Queensland Forest Service Circular.

BARNARD, R. C. 1950. Linear regeneration sampling. Mall For., 13: 129-142.

BARNARD, R. C. 1964. Manual of Malayan silviculture for inland lowland forests. Malayan Forest Research Institute. Res. Pamph. 14.

BAUR, G. N. 1962. Silvicultural practices in rainforests of northern New South Wales. Forest Commission N.S.W., Div. For. Mgt., Res., Note No. 8.

BROWNE, F. G. 1949. Storm forest in Kelantan. Mall For., 12: 28-33.

COUSENS, J. E. 1967. The sampling of regenerated forest in Malaya. Seventh Brit. Comm. For. Conf. paper.

DAWKINS, H. C. 1956. Rapid detection of aberrant girth increment in rainforest trees. Emp. For. Rev., 36 (4): 448-454.

DAWKINS, H. C. 1958. The management of natural tropical high forest with special reference to Uganda. Imp. For. Inst. Paper No. 34.

DONIS, C. & MAUDOUX, E. 1961. Sur l'uniformisation par le haut. Publ. INEAC, Ser. scient. No. 61.

JACOBS, M. R. 1966. Growth habits of the eucalypts. Canberra, Forestry and Timber Bureau.

JONES, E. W. 1955-56. Ecological studies on the rain forest of Southern Nigeria. IV. The plateau forest of the Okomu forest reserve. J. Ecol., 43 (2): 664-694; 44 (1): 83-117.

KEAY, R.W.J. 1960. Seeds in forest soils. Nig. For. Inf. Bull. (n.s.), 4.

KEAY, R.W.J. 1961. Increment in the Okomu forest reserve. Nig. For. Inf. Bull. (n.s.), 11.

MACKENZIE, J. A. 1961. Phenology of Triplochiton. Dept. For. Res. (Nig.) Tech. Note 1.

MIGUET, J. M. 1966 .Boisement et régénération de forêts reliques en zone tropicale humide. Les forêts de St-Philippe et de la Réunion. Rev. for. franc., 7: 187-200.

MOORE, D. 1967. Effects of an expanding economy on the tropical shelterwood system in Trinidad. Seventh Brit. Comm. For. Conf. paper.

NICHOLSON, D. L. 1968. Natural regeneration of logged tropical rainforest, North Borneo. Mall For., 21: 66-71.

ROSEVEAR, D. R. 1962. Yield of the high forest, Nig. For. Inf. Bull. 3.

SCULTZ, J. P. 1960. Ecological studies on rain forest in northern Surinam. (Vol. II, Vegetation of Surinam.) Amsterdam, Van Eedenponds.

TAYLOR, C. J. 1960. Synecology and silviculture in Ghana. Edinburgh, Nelson.

VINCENT, A. J. 1960. A study of the merchantable yield per acre from final felling in inland forest: Malaya. Malayan Forest Research Institute. Res. Pamph. 30.

VOLCK, E. 1960. Silvicultural work in the tropical rainforests of North Queensland. Austr. Timber Ind. Stabil. Conf. paper.

WADSWORTH, F. H. 1947. An approach to silviculture in tropical America and its application in Puerto Rico. Car. For., 8 (4): 246-266.

WADSWORTH, F. H. 1962. Forest management in the Luquillo mountains. III. Selection of products and silvicultural policies. Car. For., 13 (3): 93-119.

WADSWORTH, F. H. 1967. Tropical rainforest. The Dacryodes Sloanea Association of the West Indies. In Tropical silviculture, vol. II, p. 13-23. Rome, FAO.

WADSWORTH, F. H. 1960. Regeneration of tropical forests by planting. Fifth World For. Conf. paper.

WADSWORTH, F. H. & ENGLERTH, G. H. 1959. Effects of the 1956 hurricane of forests in Puerto Rico. Car. For., 20: 38-61.

WYATT-SMITH, J. 1960 .Diagnostic linear sampling of regeneration. Mall For., 23 (3): 191-208.

WYATT-SMITH, J. 1962. A review of Malayan silviculture to-day. Eighth Brit. Comm. For. Conf. paper.

Previous Page Top of Page Next Page