Previous Page Table of Contents Next Page

Amazon forests

Technical Assistance Officer, FAO1

1Mr. G. J. W. Pitt now teaches in the Department of Forestry, Oxford University.

Possible methods of regeneration and improvement

THE principal object of management in the Amazon forests may be regarded as being to replace a forest composed of trees of all ages and of a large number of species, many of which are of little or no known use and in which the more useful species are usually the rarer ones, with a forest composed of a large number of individuals of a few useful and very useful species and of about the same age.

The vast area of the Amazon jungle, some 3.5 million square kilometers, gives many people a false sense of security on account of the apparently limitless supplies of timber. But the distances over which first-class commercial timbers have to be transported from stump to sawmill are already rapidly increasing, and the need therefore arises for improvement of those forests which are closer at hand.

Two of the main obstacles to the development of forest management in the Amazon are the lack of silvicultural information about many of even the better known species, and about the properties and possible uses of the lesser known ones. It was to assist in the former that a silvicultural section was included in 1955 in the FAO Forestry Mission to the Amazon.2

2See Unasylva, Volume 12, Number 4.

To carry out the task of assisting in the introduction of appropriate silvicultural methods, various preliminary investigations have been necessary. For such a vast area as the Amazon it is not practicable to investigate all the possibilities. Work therefore has been confined to a few special areas, namely, on the Curuá Una river 105 kilometers from Santarém; near Santarém; and near Macapá.

In the discussions later on in this paper the various generalizations will probably apply to most of the region though their possible limitation must be kept in mind. A full report on the first five years' work done is to be issued shortly by FAO.


There is a definite dry season in the second half of the year varying from about three to six months according to the locality. The mean annual rainfall may vary from 8.50 millimeters (C.B.A., Santarém; to 2,750 millimeters (Belém). Mean maximum monthly temperatures vary from 24°C. to 35°C., and the minimum from 19°C. to 25°C. Nights are usually cool, especially in the forest.

The level of the rivers can vary considerably. At Belém the tidal range is about 2 meters; at Santarém; dry and wet season levels vary by about 3 to 4 meters, and at Manáus it once reached 16 meters. These changes would naturally affect work on the "varzeas" the periodic or seasonal swamp forests.

Principal forest types

1. Swamp

(a) Varzea - seasonal or tidal
(b) Igapó - permanent

2. Terra firma

(a) Sandy soils - generally on quaternary deposits - flanco
(b) Clay soils

(i) "Planalto" - tertiary
(ii) Undulating - may be a more denuded tertiary formation, or over old rocks, usually of the Brazilian shield.

Varzea forests, especially in the lower reaches, provide much of the timber today, though exploitation is confined almost entirely to floaters.

Igapó forests are poorer and provide little commercial timber.

Sandy soils carry forest generally low in volume and with few desirables; in places they carry only fairly low savanna.

Planalto clays can carry the highest volume (up to over 300 cubic meters per hectare of trees over 25 centimeters d.b.h.) and are rich in species; usually there are very few economic species, such as Cedrela, on a single hectare, though there may be several desirables.

It is only on the terra firma soils that exploitation by mechanical means is possible over most of the year, and the following notes refer only to these soils, and in particular to the planalto area by the Rio Curuá Una.

Possible degrees of exploitation

Creaming. The density or stocking of the "economics" is so low, e.g., 10 Cedrela per square kilometer, that the forest is virtually unaffected, except for occasional gaps and extraction tracks. Hand labor, with its limited range of about half a kilometer, is usual for moving the logs from stump to stream or river; lorries are sometimes used, and occasionally even tractors for the first stage.

Heavy selective. In the Curuá, where the FAO team is logging, the planalto forest has some 120 different species over 25 centimeters d.b.h. with about 35 to 45 species on any one hectare. Of these some 30 to 35 percent are desirables, and represent about 40 percent of the stems or 30 percent of the standing volume. Sound trees of all the desirables are being felled and taken to the sawmill training center at Santarém; or used locally for camp buildings.

Intense. This implies more or less clear felling for pulp, with the best logs going to sawmills and/or plywood or veneer mills, and suitable second quality logs going to sleepers. On the planalto area referred to above, "100 percent inventory over 13 hectares has shown" standing total volume of all trees over 25 centimeters d.b.h. of about 250 cubic meters per hectare (190-315). Recently a 100 percent inventory of trees over 45 centimeters d.b.h., but excluding the badly fluted trees, was carried out over a block of 100 hectares by Glerum and Smit (11).3 The result shows an irregular stocking in the compartment over rather short distances. In this block, six species (five economic and one potentially so) gave 60 percent of the total volume - i.e., some 90 cubic meters per hectare, a very high figure for a tropical forest.

3 The italicized figures in parentheses refer to literature cited at the end of the article.

FIGURE 3. - Parapara (Jacaranda copaia) 2 ¼ years old raised from small (20 cm.) striplings planted at 2.5 X 2.5 m. The canopy is already closed.


The value of pure silvicultural research is not to be disputed, particularly in such undeveloped forests as occur throughout the Amazon. However, in view of the terms of reference of the assignment, silvicultural work had to bear some relation to the possible degree of exploitation and utilization. Economic considerations therefore imposed considerable limitations. Valuable work, however, could be done now to improve the forests for exploitation at some future date - "refining" vide Dawkins (5) - but this is an aspect of management which is usually not considered, especially in tropical forests, owing to budget limitations. There is scope for considerable development in the intensity of the utilization of the planalto forests; until this takes place, the silviculturist can do little beyond showing how the forests can be improved both before, and after, exploitation.

Possible methods

Much has been written about tropical silviculture, and the methods which are most likely to be suitable to Amazon conditions are:

1. Natural regeneration

(a) inducing natural seedling regeneration of desirables prior to exploitation, or encouraging regeneration normally under some degree of shelterwood;

(b) natural regrowth of seedlings, coppice and perhaps suckers, following intensive exploitation.

2. Artificial regeneration

(a) enrichment planting;
(b) plantations.

Natural seedling regeneration. This would entail some degree of canopy opening by creeper cutting, killing some trees and cutting back the undergrowth of undesirable species, followed later by periodic killing of further undesirable trees and recutting the undergrowth.

Natural regrowth. Unless plantations are envisaged, natural regrowth would be encouraged to follow intensive exploitation. The initial seedling regrowth is likely to consist mainly of softer-wood, light-loving species and will result initially in a forest of a rather different composition. During the first few years it may be advisable to cut back the regrowth of the less useful species two or three times. Some management aim may perhaps dictate favoring only those soft-wooded species more useful for pulp, or encouraging useful harder timber species as well.

A heavy selective felling (eight trees per hectare) was carried out early in 1959. Some 18 months later, in most of the more open areas there was already enough regeneration of several desirables to make it possible to decide in the first cleaning whether to aim at a final crop composed mostly of the softer-wooded species, or of the harder timber species such as Goupia.

Enrichment planting. This could be on a relatively small area, say one to two percent of a creamed area. Only valuable species, tolerant to shade, should be used. Such planting could be adopted in those parts of areas where natural regeneration has not been as successful as required. Similarly it might be suitable for accessible areas near centers of population where the present stocking of desirables does not justify exploitation.

Plantations. For these it would be necessary to clear fell an area and then bum it at the end of the dry season. Parts of easily accessible forest already creamed or with only a few desirables could be replaced by this method, particularly if they are near large centers of population. If plantations were undertaken it would be most important to have a variety of species and to avoid large pure stands.

There is no question of "taungya" at present as there is no land shortage. If intensive exploitation were to start on a large scale, then there would be definite possibilities for providing food for the laborers and for growing trees at the same time.

Work in hand

All the above methods have been demonstrated by the writer, often on two or three different sites.

No intensive exploitation is being carried out at present; therefore small areas (two to six hectares) have been clear felled and burned, after removal of the larger economic species. These have been used for camp building or sent to the FAO sawmill training center nearby.

FIGURE 4. - Cupiuba (Goupia glabra) 15-months-old raised from 40 cm. striplings planted at 2.6 x 2.5 m. (1,600 p. ha.). Note absence of weeds following an intense burn of clear-felled forest. Compare with Figure 5.

Natural regeneration

Natural seedling regeneration. Plots have been laid down in three localities with two or three degrees of canopy opening; the size varies from one to four hectares, plus a surround. Weed trees have been poisoned with contact arboricides; various concentrations and mixtures have been tried; the most satisfactory seems to be a one to two mixture of 2,4-D to 2,4,5-T (both 80 percent acid equivalent) at three to five percent in diesel oil, depending on the species and size of the trees to be killed. A few resistant species are frilled before spraying. The degree of opening was controlled by the number and size classes of tree poisoned, and later checked by girth measurements for basal areas (only trees over 25 centimeters d.b.h. were girthed). Recently a prism has been used for measuring basal areas. The original basal areas varied from 18 to 22 square meters per hectare and were reduced to about 11 square meters for a heavy opening and to about 15 square meters for a light opening. In a heavy opening, besides some large undesirables, some understory trees were also poisoned and the undergrowth of all undesirables slashed back.

Good regeneration of several desirables has been obtained in some plots two to three years after treatment. In one plot it has been considered advisable to cut back the weed regrowth after two years, and to thin out the useful regeneration to intervals of about 1.5 meters. A recent "milliacre" survey showed 98 percent of the quadrats (2 meters square) stocked with generally more than one seedling of the three main species, or occasionally with one or two of 11 other desirables.

Plots have been laid down to encourage regeneration of Cedrela, probably the most important species of the Amazon. There are three degrees of canopy opening, each with three ground treatments. A good seed year is expected.

FIGURE 5. - Very successful natural regeneration of Cupiuba (Goupia glabra) 2½, years old, following a light burn of clear-felled forest. Alter two cleanings and one thinning the stocking is now 1,200 Cupiuba and 750 other desirables.

Gregarious patches of seedlings of Clarisia and Hymenolobium have been encountered. As these trees are rare in the forest, these patches are being observed carefully; over some the canopy is being opened and the seedlings are being thinned out.

Natural regrowth. One of the problems is to recognize the seedlings. Seeds of desirables are being collected as opportunity offers and plants raised in the nursery. Later, the plants are put out into arboreta on the main soil types to obtain more silvicultural information.

Natural regrowth is being observed in the surrounds of plots clear felled for trial plantations. In the which was first felled and was lightly burned in December 1957, good regeneration of Goupia, mainly from seed, has been obtained: 62 percent of the 2-meter quadrats were stocked by August 1958, and 49 were still stocked by June 1960 (average height of tallest in each plot was 2.5 meters). Weed growth had to be out back in April 1959 and a 1.5 meter stick thinning was carried out by uprooting the surplus seedlings; seedlings of several more desirables have also appeared; some are in quadrats already stocked with Goupia, and some in 30 percent, more of the quadrats, thus giving a total stocking of 79 percent with various desirables on about twice the minimum (40 percent) considered adequate. The plot can be regarded as fully stocked - there are very few gaps of over 10 or 15 meters - though more slashing back of weeds will be necessary for a year or two.

Where the logging unit has been operating, tight burns are now being carried out where there is a dense mass of branches. These should result in good regeneration of Goupia and a few other desirables. A fierce burn is to be avoided, as in such patches the soil will still be comparatively bare, even after two wet seasons. However, if the burn is fierce and no natural regeneration appears, the patches can be safely planted up the next wet season with quick-growing species. These patches should not need a cleaning until the natural regeneration in the surrounding areas is being cleaned; it will thus not be necessary to look separately for these planted patches.

Artificial regeneration

Enrichment planting. Various spacings and degrees of opening are being tried. Even with shade-tolerant species, growth is usually slower than when these species are planted in the open. The main scope for enrichment planting lies in improving the composition of young secondary growth. It is also indicated in areas of fairly intense exploitation where the natural regeneration of desirables has not been adequate. It is clear, however, that in virgin high forest, poor in desirables, little improvement can be made except by a heavy, and therefore expensive, opening of the canopy and by using that rather rare phenomenon, quick-growing shade-tolerant species.

Both contact arboricides and girdling are being used to kill weed trees along the enrichment lines. Direct sowing of Carapa. shows some promise and so do transplants of Carapa, Clarisia, Vochysia, Hymenolobium and Virola after one year. Spraying with 2,4-D is being tried to keep down the weed regrowth along some of the lines.

Plantations. Practically all aspects of this work are being demonstrated. In the nurseries, besides the local soil, seedbeds been have made of "terra preta" (black sandy more fertile soil of uncertain origin and a little less acid, pH 5.6 as against 4.4), sawdust and river sand. Transplants are being raised in beds, boxes, galvanized iron tubes and earth pots (torräo paulista); plants from the beds are put out as naked rooted, ball rooted, striplings and stumps. Only one or two methods are tried for any one species in one area. Everything is first tried out as naked rooted transplants. The only uniform procedure is the spacing, 2.5 X 2.5 meters, giving 1,600 plants per hectare. Work on various spacings can follow when the best technique has been established for each species.

Fertilizer trials with phosphorus are also in hand. On a poor degraded campos clay soil, nothing has succeeded without phosphorus, and even then the initial response with phosphorus was not maintained in the second and third years, except with some eucalypts. Cover crops (Crotalaria) are being sown to help check weed growth and improve soil fertility. Pruning and thinning will come later.

There has been considerable interest in the possibilities of intensive exploitation for pulp. The technical and economic aspects of this subject are outside the scope of this paper. However, if such exploitation starts and silvicultural methods can be introduced to give a forest with fewer species of more uniform properties, then a great step forward will have been made. To this end work is in hand on introducing pines, as these would also improve the furnish very considerably. The best pine is likely to be P. caribaea; difficulty was experienced in the first year in obtaining adequate mycorrhiza; and postal delays over seed cost us another year. Several trial plots with well inoculated plants were, however, put out early in 1959; results are promising.

P. khasya has not succeeded and postal delays twice resulted in P. merkusii losing its viability; we were more fortunate with the third lot and good plants were put out in 1960. Some 40 local species and 30 exotics including eucalypts are also on trial.

FIGURE 6. - Part of the Curuá Nursery. Terminalia superba on left, Gmelina arborea between men. In six months time the former will be put out as stumps or striplings depending on their size, and the latter will go out as stumps.

The main pest is saúva, the leaf-cutting or parasol ant; in general, indigenous species seem less delectable, but they too have been attacked. The ant seems more frequent on sandy soils and appears to be more abundant where there is poor or young secondary growth following clearings for cultivation.

Other operations

Pre-exploitation. These consist of creeper cutting, freeing young advance growth and freeing and thinning occasional patches of desirable gregarious seedlings. When more information is available as to the density of canopy required to encourage good regeneration of one or more desirables, the poisoning of weed trees will be undertaken.

Postexploitation. These consist of further freeing of advance growth, cutting back damaged young desirables and cutting young weed trees which, when larger, are resistant to arboricides. Exploitation, even when fairly intense, leaves some large patches often untouched. In these patches some weed trees are being killed to favor any young desirables. When these are absent, enrichment planting is indicated, provided suitable species have been found; in this case, poisoning of weed trees is necessary to open the canopy over the lines. Reference has already been made to the burning of branchwood in heavily felled patches.


When introducing silvicultural methods to a virgin forest area it is impossible to go straight ahead with only field-scale operations. A tremendous amount of information about the composition of the forest has to be found out, and about the silvics, and silvicultural requirements, not only of the desirables, but also of, the weed trees. This is quite apart from deciding which species are desirables and which are weeds.

A line usually has to be drawn between experimental work and research work. Where that line lies is a matter of opinion. Since there is so far no forestry section developed at the National Research Institute for Amazonia (INPA) at Manáus, plots of a more purely research nature have been established along with our demonstration and experimental plots.

Observations or measurements are being made on:

1. inducing natural regeneration by canopy opening and, in the case of cedar, by ground preparation;
2. resistance of all larger species to different concentrations of arboricides and to girdling;
3. amount of slashing back of weed growth needed to favor established or induced regeneration;
4. suitability of various species for enrichment planting; 5. various methods of raising transplants;
6. various methods of establishing plantations;
7. effect of using phosphorus fertilizers;
8. rates of growth of indigenous and exotic species in plantations on various soil types;
9. rates of diameter growth in virgin and in partly opened forest.

During 1960 some work was started by INPA near Manáus and an agronomist came three times to the Curuá for some training.

The pedologist with the FAO/UNESCO mission has visited all the silvicultural centers, studied soil profiles and had samples analyzed. The main features have been the very low fertility values as shown by the amount of elements present, and the high acidity (pH 3.9 to 4.9). This high acidity is probably a limiting factor in the introduction of some exotica and may explain why such species as teak have failed in most trials. Much work could be done in tropical forests on the pH requirements of the better known species and thus avoid many of the failures encountered when introducing exotica. The difficulties with mahogany in the nursery may also be due to acidity; where it was seen growing naturally by the writer on soils of much older geological origin, the pH was much higher - about 6.

Space does not permit to discuss in this article the various methods used in sampling the stocking and regeneration. These are being dealt with in the report which is being prepared for the Brazilian Government.

FIGURE 7. - Pinus caribaea. 15 months old on young (5 years) secondary forest soil, Macapá nursery. Note the 3 ¼ year old pine plants beyond and Parapara (Jacaranda copaia) on right, three years old.


In these early stages, management must be extensive rather than intensive; stress should be more on natural regeneration than artificial, at least until intensive exploitation for special purposes starts. The principle should be: "Follow Nature; perhaps guide her, but do NOT dictate to her." The introduction of exotica should be considered, especially if there is any question of future exploitation for pulp.

From the point of view of natural regeneration it is better to have an area with a relatively high volume of a second-class desirable, which it is fairly easy to regenerate and encourage, e.g., Goupia, than to aim at a first-class desirable which can be established only with difficulty, e.g., Cedrela.

It is far too soon to weigh the relative values of natural regeneration against artificial regeneration. Much more muddy water must flow down the Amazon before any reputable forester should give his considered opinions on this nice point. Fortunately, the Brazilian Government has asked for technical assistance to continue in the Amazon at least to December 1962.

Cost of operations

The figures given below are what it is thought the operations should cost on a field scale with experienced labor. In many cases the work has so far been on a small scale, often with untrained labor.


1. Felling the largest trees a year or so ahead, and the medium-sized trees six months or so ahead of the small trees and undergrowth may result in a better burn and so avoid the need for collecting and reburning large branches and stems.

2. Planting costs on a large scale will vary considerably according to the type of nursery plants; naked roots transplants, striplings or stumps will cost less to plant than transplants from containers such as pots or boxes.


Man days

per ha.

per km.

Linear regeneration surveys




Natural regeneration


Creeper cutting and freeing advance growth


Underbrushing (selective)




Poisoning (light opening)


Poisoning (heavy opening)


Postexploitation - after clear felling and burning

First cleaning (at 18 months)


Second cleaning and thinning (at 2-2½ years)


Postexploitation - after heavy selection felling

Freeing advance growth





Line cutting













Underbrushing (try to avoid)


Clear felling


Clear felling (staggered) (Note 1)


Felling desirables


Clear felling remainder


Collecting and reburning large branches and stems (Note 1)




Planting (Note 2)


Sowing cover crops





1. AUBRÉVILLE, A. 1957. Sylviculture dans les forêsts tropicales hétérogènes. In FAO. Tropical silviculture. V. 2:46-56.

2. BANERJI, J. 1957. Tropical rain forest. In FAO. Tropical silviculture. V. 2:36-45.

3. BARNARD, R. C. 1950. Linear regeneration sampling. Malayan Forester, 13:129-142.

4. CLARKE, E. C. 1956. The regeneration of worked-out Greenheart (Ocotea rodiei) forest in British Guiana. Emp. For. Rev., 35:173-183.

5. DAWKINS, H. C. 1957. Contact arboricides for rapid tree-weeding in tropical forest. In FAO. Tropical silviculture. V. 2-109-112.

6. DAWKINS, H. C. 1958. The management of natural tropical high forest with special reference to Uganda. Imperial Forestry Institute, Oxford. Institute Paper No. 34. 155 p.

7. DAWKINS, H. C. 1958. The measurement of basal area increment in tropical high-forest. Paper, Inter-African Forestry Conference, Pointe Noire.

8. DE FAYS, E. P. & HUYGEN, J. P. 1957. Enrichissement des forêts ombrophytes hétérogènes dans la Province Orientale du Congo belge. In FAO. Tropical silviculture. V. 2:129-138.

9. FINCH, H.D.S. 1957. Plotless enumeration with angle gauges. Forestry, 30:173-192.

10. FOURY, P. 1956. Comparaison des méthodes d'enrichissements utilisées en forêt dense équatoriale. Bois et Forêts des Trop., 47 (3): 1525.

11. GLERUM, B. B. & SMIT, G. 1960. Report to the Government of Brazil on a forest inventory in the Amazon valley. Pt. 6. 100 % survey Curuá Una. Rome. FAO/ETAP Report No. 1271. '

12. HAIG, I. T., HUBERMAN, M. A. & U AUNG DIN. 1958. Tropical silviculture. V. 1. Rome. Forestry and Forest Products Studies No. 13. 190 p.

13. HEINSDIJK, D. 1957-58. Report to the Government of Brazil on a forest inventory in the Amazon Valley. Pt. 1. Tapajós to Xingú. FAO/ETAP Report No. 601. Pt. 2. Rio Xingú to Rio Tocantins. FAO/ETAP Report No. 949. Pt. 3. Rio Tapajós to Rio Madeira. FAO/ETAP Report No. 969. Pt. 4. Rio Tocantins to Rios Guamé e Capím. FAO/ETAP Report No. 992.

14. HOLDRIDGE, L. R. 1957. The silviculture of natural mixed tropical hardwood stands in Costa Rica. In FAO. Tropical silviculture. V. 2:57-66.

15. KADAMBI, K. 1957. Methods of increasing growth and obtaining regeneration of tropical forests. In FAO. Tropical silviculture. V. 2:67-78.

16. LANDON, F. H. 1957. Malayan tropical rain forest. In FAO. Tropical silviculture. V. 2:1-12.

17. MAGINI, E. & TULSTRUP, N. P. 1955. Tree seed notes. Rome. FAO Forestry Development Paper No. 5. 354 p.

18. MARÉCHAL, P. 1957. Notes sur le traitement applicable aux peuplements forestiers du Lac Léopold II. In FAO. Tropical siviculture. V. 2:139-143.

19. MEIJER DREES, E. 1957-59. Report to the Government of Brazil on the organization of a forest research centre in the Amazon region. Pt. 1. FAO/ETAP Report No. 756. Pt. 2. Appendices. FAO/ETAP Report No. 1068.

20. PARRY, M. S. 1956. Tree planting practices in tropical Africa. Rome. FAO Forestry Development Paper No. 8. 302 p.

21. NAIR, K.N.R. 1957. Tropical rain forest. In FAO. Tropical silviculture. V. 2:103-108.

22. PICKLES, A. 1958. Methods employed in West Africa for the destruction of redundant forest trees. Emp. For. Rev., 37:421-428.

23. RENIER, H. J. 1957. L'aménagement des forêts, naturelles au Kasai et an Ruanda. In FAO. Tropical silviculture. V. 2:144-159.

24. ROSAYRO, R. A. DE. 1957. The silviculture and management of tropical rain forest with special reference to Ceylon. In FAO. Tropical silviculture. V. 2:79-101.

25. WADSWORTH, F. H. 1957. Tropical rain forest. In FAO. Tropical silviculture. V. 2:13-23.

26. WILTEN, W. 1957. Aspects de la sylviculture au Mayumbe. In FAO. Tropical silviculture. V. 2:160-167.

Previous Page Top of Page Next Page