On Biomass and Trophic Structure of the Central Amazonian Rain Forest Ecosystem Author(s): E. J. Fittkau and H. Klinge Source: Biotropica, Vol. 5, No. 1 (Apr., 1973), pp. 2-14 Published by: The Association for Tropical Biology and Conservation Stable URL: http://www.jstor.org/stable/2989676 Accessed: 30-06-2015 18:11 UTC Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. The Association for Tropical Biology and Conservation is collaborating with JSTOR to digitize, preserve and extend access to Biotropica. http://www.jstor.org This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions On Biomass and Trophic Structure of the Central Amazonian' Rain Forest Ecosystem E. J. Fittkau and H. Klinge Max-Planck-Instituteo f Limnology, Department of Tropical Ecology, D232 Plon/Holstein, Federal Republic of Germany ABSTRACT The importance of litter in the total energy flow dynamics of a central Amazonian rain forest near Manaus, Brazil, is discussed. The study area is located in the hinterland of Manaus between the Rio Negro and the Amazon. Its sub- strate is Tertiary sediment. The area receives 1771 mm rainfall per year, and the soil is classified as yellow latosol. The forest comprises 93,780 dicotyledonous trees and palms per hectare reaching 38.10 meters in height. Over 500 species of palms and dicotyledonous trees above 1.5 m. in height are identified for a 2000 sq. m. plot. The estimate for fresh living dicotyledonous tree and palm biomass is 939.5 metric tons per hectare consisting of 1.9% leaves, 49.7% stems, 21.3% branches and twigs, and 27.1% roots. Lianas, vascular epiphytes, and parasites are estimated to comprise 46.2 mt/hectare in the fresh state. At the soil surface there are 59 mt/hectare of fresh litter. Living animal biomass is about 200 kg/hectare of which half is soil fauna. The high proportion of soil fauna, the type of humus, the decomposition of litter, the apparent dependence of soil fauna on fungi, and the low nutrient content of litter are all factors which strongly support a consumer food chain based almost entirely on dead organic matter. The fungi play a decisive role in concentrating the otherwise limited nutrient resources. ECOLOGICALS TUDIES carried out in the Amazon re- for 1963 and 1964. The average litter-fall for this gion (Fittkau et al. 1969) raised further questions period shows that 7.3 metric tons (mt) of dry mat- concerning the richness of the Amazonian ecosys- ter per hectare (h) per year are returnedt o the soil. tems, the distribution of their biomass, and impor- Dry matter per hectare is made up of 5.6 mt of tance of overall as well as trophic-level structure. leaves, or 76.6 percent of the total dry matter; the Additional experience obtained through later field- remainder is composed of flowers, small fruits, and work in Amazonia and the results of our studies in twigs. It is interesting to note that this amount of hydrobiology, ecology, landscape ecology, pedology, litter is smaller than the litter-fall reportedf or tropi- and terrestrialp roduction indicated the importance cal rain forests in Africa and Asia (Bray and Gor- of finding out what the relationship of all these ham 1964). Klinge and Rodrigues (1968) showed factors is in the entire central Amazonian' rain forest that Amazonian litter is poorer in nutrients when where the predominant soil type is latosol. Also compared with litter from other tropical forests. consideredm ust be the geochemicals tructureo f that Results from our chemical analyses indicated that region, the division of the area as demonstratedb y the following raw elements occur in the litter re- Fittkau (1969, 1970a, 1970b, 1971a, 1971b; fig. 1), turned to the soil in central Amazonia (kg per hec- and its influence on the nutrient supply available tare per year): 2.2 P, 12.7 K, 5.0 Na, 18.4 Ca, 12.6 to the biomass in terrestrial and aquatic environ- Mg, and 105.6 N. ments. In 1970, estimates were made of the amount of A discussiono f the bioenergeticso f a tropicalr ain woody material involved in litter-fall of the central forest is difficult because few studies have been Amazonian rain forest (Klinge, unpublished). The conducted either on the basis of the specific trophic results of the analyses show that one mt of stems levels involved or on the basis of the whole biomass. (stem-part of plant between soil surface and first As a consequence,a discussion of available nutrients ramification), two mt of branches (ramifications for specific species at various trophic levels is even without leaves), and 1.35 mt of twigs (ramifica- more difficult. However, comparisons of litter-fall tions bearing leaves), bark, etc. are involved. with the soil complex have been published, and the Because the 1963-1964 litter-fall experiment results are interesting enough to investigate the was not suitable for the measuremento f the total problem further. In a study by Klinge and Rod- fruit-fall, we calculatedr oughly the amount of fruits rigues (1968), the litter-fall of a central Amazon- involved in annual litter-fall by kind, weight, and ian tropical lowland rain forest was determined number of fruits in a 2000 m2 forest plot. Thus, we determined amounts of 0.35 mt of small fruits 1 In this paper, 'central Amazonia' and 'central Amazonian' (up to 5 g in weight) and 0.5-1.0 mt of larger refer only to the ecological unit of Amazonia defined by fruits (over approximately 5 g in weight). We Fittkau (1963) and illustrated in figure 1. 2 BIOTROPICA5 (1): 2-14 1973 This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions 80 70 60 50 40 ~~~~~~~~~~~~~~~~~~~0~~~~~~~~~~~~~~~~~1 Biomass estimation plot in the Manaus area .. .. .... ... 0 2~~~~~~~~~~ ~ ~ ~ ~ ~ ~ ~ ...... iii*k 7 20 -- 20 FIGURE 1. Subdivision of Amazonia according to Fittkau (1969). 1-Limit of Amazonian rain forest area. 2-Cen- tral Amazonia. 3-Northern and southern peripheral Amazonia. 4-Western peripheral Amazonia. guess that these amounts are underestimated if we rainy season of 1964 litter-fall was 400 kg per hec- compare our determinations with data for the humid tare per month; it was 700 kg per hectarep er month tropical forest on Barro Colorado Island, Panama, in the dry season; the maximum for the dry season, as given by Smythe (1970). In a systematic fruit measuredi n September 1964, was 1000 kg per hec- collection experiment, Smythe found about two mt tare per month. The seasonalityo f litter-fall is ex- of fruits per hectare per year. plained by seasonalityo f rainfall. About 90 percent Adding all Amazonian litter fractions, a total of of the total annual rainfall of 1771 mm per year at about 11 mt per hectare per year is obtained. This Manaus (Walter and Lieth 1960-1967) occurs dur- value resembles the litter amount given for a mon- ing the rainy season which lasts generally from tane rain forest in Puerto Rico which comprises May-June to October-November.H owever, in 1964 5.52 mt of small litter-fall, 1.9 mt of log-fall, and the dry season had begun alreadym uch earlier than 4 mt of brush-fall (Odum 1970). As wood is gen- usual. erally poor in nutrients, the incorporation of so In the period from June to November 1970, much woody matter in the litter-fall of the central which was unusuallyr ainy, litter-fall was again mea- Amazonian rain forest suggests the validity of the sured as was the total litter at the soil surface. The above-stated low nutrient content of Amazonian lit- results indicate that litter-fall and litter decomposi- ter. tion were rather well balanced because of the high We also found litter-fall to be seasonal. In the activity of litter decomposers. According to Stark Biomass of Amazonian Rain Forest 3 This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions TABLE 1. Height classes (meters) and fresh biomass (kilograms) of dicotyledonous trees and palms in central Amazonian rain forest per hectare. Subtotal Height class <0.2 >0.2 - 0.5 >0.5 - 1.0 >1.0 - 1.5 <0.2 - 1.5 >1.5 - 5.0 NUMBERS OF INDIVIDUALS Trees 6,535 40,725 21,400 12,575 8,950 83,650 Palms J 915 BIOMASS Leaves 91 128 294 513 3,208 } Stem 5 55 72 200 609 936 2,993 Twigs and branches 1,132 Total aboveground 55 163 328 903 1,449 7,333 Large roots 25 62 105 247 439 1,542 Fine roots - - - - - Total underground 25 62 105 247 439 1,542 Total above and underground 80 225 433 1,150 1,888 8,875 RATIOS Aboveground:underground 1:0.5 1:0.4 1:0.3 1:0.3 1:0.3 1:0.2 Leaves:twigs and branches:stem - - - - - 1:0.4:0.9 Leaves:wood - 1:0.8 1:1.6 1:2.1 1:1.8 1:1.3 (1971), decomposition is 5.4 g per m2 per day. Prior to mapping, all smaller plants were harvested. Klinge (1972a) found a decompositionr ate of 0.56 Total height, stem length, diameter at breast height, percent per day for leaves, and of 1.5 and 2.3 percent and crown diameter of plants above 1.5 m height per day for woody litter and fruits, respectively, in were measured using a steel tape. Taller plants were the biomass estimationp lot. measured after felling; smaller plants were measured There is only scattered information regarding with a wooden ruler before cutting. In some cases, the structure of forest which produces the above crown diameters were derived from projections made amounts of litter (Hueck 1966, Lechthaler 1956, from the ground. Takeuchi 1961). Rodrigues (1967) surveyed a for- All leaves, twigs, branches, and stems were sepa- est and included all trees over 25 cm diameter at rated; leaves by hand, twigs by machete, and branch- breast height (dbh), covering 137,000 hectares in es and stems using a portable chain saw. The frac- the area of the Manaus-Itacoatiarrao ad. This survey tions were weighed in the field, branches and stems includes the Walter Egler Forest Reserve where the after they were sawed into manageable pieces. 1963-1964 litter-fall measurementsw ere made. The Portable balances, weighing up to 100 kg, were used. area studied by Rodriguesa lso includes a plot which The root mass of 381 dicotyledonous trees and 51 was studied by Klinge and Rodrigues (1971, Klinge palms above 1.5 m height was estimated after ex- 1972a,b,c), for forest biomass estimation. Some re- traction of the plants by hand or by use of a simple sults of this recent study are describedb elow. jack. Plants below 1.5 m height were extracted by On a level site at km 64 of the Manaus-Ita- hand, separated into four height classes and counted, coatiara road, a rectangularp lot of 2000 m2, im- but they were not sampled for taxonomic determina- mediately adjacent the Walter Egler Forest Reserve, tions as in the case of the taller plants. Leaves, was marked off using a nylon line. The plot was shoots, and roots of small plants were also sepa- subdividedi nto 40 equal subplots. The area is terra rated and weighed. All fractions of individual dico- firma, i.e., terrain which is never reached by annual tyledonous trees and palms of the four height classes river floods. The soil is a yellow latosol of heavy under 1.5 m were sampled and weighed, air-dried texture (Anonymous 1969). The subsoil is Ter- in the field, and oven-dried in the laboratory at tiary sediment (Barreiras Series). The plot was Manaus. All plant parts were finally shipped to mapped to show the position of all palms and dico- Europe for subsequent determinations of nutrients tyledonous trees above 1.5 m height and of lianas and dry matter. These analyses are still progressing. of more than approximately 5 cm in diameter. The field work extended from mid-June to the 4 Fittkau and Klinge This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions Subtotal >5.0 - 10.0 >10.0 - 20.0 >20.0 - 25.0 >25.0 - 30.0 >30.0 - 35.0 >35.0 - 38.1 1.5 - 38.1 Total 1,480 725 175 160 55 25 9,155 93,780 45 15 0 0 0 0 975 1,161 3,253 2,696 3,887 2,277 1,150 17,632 18,145 8,864 37,832 103,241 114,444 126,317 73,410 467,101 467,101 3,193 14,994 53,417 58,419 46,953 21,089 199,197 199,197 13,218 56,079 159,354 176,750 171,547 95,649 683,930 685,379 1,951 6,002 11,952 11,489 10,270 5,356 48,561 49,000 - - - - - - - 206,040 1,951 6,002 11,952 11,489 10,270 5,356 48,561 255,040 15,169 62,081 171,306 188,239 181,817 101,005 732,491 940,419 1:0.2 1:0.1 1:0.08 1:0.07 1:0.06 1:0.06 1:0.07 1:0.4 1:2.8:7.6 1:4.6:11.6 1:19.8:38.3 1:15:29.4 1:20.6:55.5 1:18.3:63.8 1:11.3:26.5 1:11.0:25.7 1:10.4 1:16.2 1:58.1 1:44.5 1:76.1 1:82.7 1:37.9 1:36.8 end of November 1970. During this time many 1). The number of trees clearly diminishes with collections were made of large and conspicuous ar- height. This relationship is valid particularly for thropods,a mphibians,a nd reptiles. The vertebrates palms which are abundantb elow 20 m height. The collected during this time were submitted to Dr. P. ratios of abovegroundb iomass,u ndergroundb iomass, Muller, SaarbriickenG, ermany,a nd the invertebrates or leaves/branches + twigs/stems, or leaves/wood went to H. Schubart and E. J. Fittkau, Plon, Ger- vary rather consistently from short to tall height many for identification. When evaluating this ani- classes. mal sampling and our observations on animals in No data are given in table 1 for lianas, epiphytes, central Amazonia made over the last 10 years, we and parasites which are represented by the follow- also made use of any information gathered from ing amounts of fresh biomass: Vascular epiphytes Indians, settlers, hunters, and professional biologists. (mainly Araceae, Bromeliaceae, and Orchidaceae), In order to obtain weight data of fauna we listed 0.1 mt per hectare; parasites (Loranthaceae), 0.13 first those animal taxa (orders, families, etc.) which mt per hectare; lianas (various families), 46.0 mt were observed to have some bearing upon biomass per hectare; total 46.23 mt per hectare. Thus we because of their individual weight and/or frequency. find that lianas are a striking feature of the central Then we estimated density taking into account all Amazonian rain forest, whereas angiosperm epi- observations gathered by ourselves and the above- phytes and parasitesa re of nearly negligible biomass. named sources. Finally, density was converted into Total weight of the living abovegroundb iomass weight of the respectiveg roup by multiplying weight of plant origin in the forest is 730.7 mt per hectare. by density, using average weight of individuals con- Weight of undergroundb iomass (roots and under- tained in our collections or otherwise determined ground trunks of certain palms) is 255 mt per weight. None of the values thus obtained was ad- hectare or 25.9 percent of the total living plant justed with respect to data found in literature for biomass (Klinge, unpublished). Both weights are similar groups in tropical regions elsewhere. in agreement with data presented by Rodin and Bazilevic (1964, 1968) and Bazilevic and Rodin (1966). Dead plant biomass at the soil surface (ex- RESULTSA ND DISCUSSION cluding roots) amounts to about 44 mt per hectare Trees and palms below 1.5 m height were in much of wood and 15 mt per hectareo f fine forest detritus, greater abundance than taller plants and made up both weights in fresh state. There is, in terms of 86 percent of the total plant cover considered (table weight, twice as much stem wood as branch wood. Biomass of Amazonian Rain Forest 5 This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions Total plant biomass of the forest under study is thus about 1100 mt of fresh matter per hectare. TABLE 3. Soil fauna of central Amazonian terra firma Our herbarium material is actually determined rain foorest on latosol. Expressed as 106 indi- viduals per hectare; from Beck 1970, 1971. only to the family level. The plant families of which the forest is composed are listed in table 2. TABLE 2. Plant fa'milies with representatives abo-ve 1.5 m > height in the central Amazonian rain forest growing on a 2000 m2 plot of terra firma _ 0 latosol. oC D at* ' 0 Numbers of Numbers of Percentage of Plant family species individuals total individuals Acarina 612 115 727 Collembola 103 16.8 0.34 120 Leguminosae 62 171 8.6 Isopoda 0.7 0.05 0.8 Sapotaceae 43 139 7.0 Pseudoscorpiones 2.1 0.09 2.2 Lauraceae 40 88 4.4 Other Arachnida 0.7 0.2 0.9 Chrysobalanaceaea 38 96 4.8 Diplopoda 2.8 0.1 2.9 Rubiaceae 32 137 6.9 Other Myriapoda 3.5 0.03 3.53 Buirseraceae 27 230 11.9 Protura 2.8 2.8 Annonaceae 21 87 4.4 Diplura 1.4 0.01 1.41 Lecythidaceae 17 132 6.6 Larvae of Coleoptera Moraceae 17 69 3.5 and Diptera 4.0 0.04 4.04 Palmae 11 196 9.9 Coleoptera 1.4 0.07 1.5 Violaceae 10 223 11.2 Isoptera 0.9 0.4 1.3 44 other families 177 409 20.6 Formicidae 7.2 1.4 8.6 Indeterminata 7 9 0.5 Aphidina/Coccina 39.4 39.4 Total 502 1986 100.3 Opiliones 0.02 0.02 Blattaria 0.02 0.02 Refers to the tribe Chrysobalanoideae of Rosaceae but in Gryllodea 0.05 0.05 the familial sense. Total individuals 782 144 2.8 929 If there are in the literature very little data on Biomass (kg) 67.7 12 4.4 84 total plant biomasso f humid tropical forests (Green- land and Kowal 1960, Ogawa et al. 1961, 1965, Art ptera, Formicidae,C oleoptera and their larvae, and and Marks 1971, Rodin and Bazilevic 1964, 1968, larval Diptera. Moreover,t here are regularlyp resent: Bazilevic and Rodin 1966), there is even less in- Scorpiones, Palpigradi, Ricinulei, Uropygi (Schizo- formation regarding total animal biomass of these peltidia), Symphyla, Pauropoda,D iplopoda (Psela- forests (Goodnight and Goodnight 1956, Harrison phognata), Protura, Diplura (Campodeidae and 1962, Hopkins 1967). A rare exception to this Japygydae), Embiidae, Orthoptera,P hasmida, Forfi- lack of information is the study on structure and culidae, Mantodea, Blattaria, Psocoptera, Thysano- metabolismo f a red mangrove forest in Puerto Rico ptera, Heteroptera, Cicadina, Turbellaria, Hirudi- (Golley et ail. 1962). There are, however, good de- neae, Onychophora,S erpentes,I guanidae, and Dasy- scriptions of the fauna living in humid tropical for- podida. ests, including Amazonian forests (Bates 1965, The lack of any precise information on the pop- Dorst 1967, Mann 1968, Mertens 1948), but infor- ulation density and biomass of termites and ants, mation on total animal biomass of Amazonian for- which are so numerous in the environment of cen- ests is completely lacking. Meggers (1971) states tral Amazonia, is to be regretted. We may refer that the majorityo f wild animals there are small and only to Beck (1971), who assumed that three quar- solitary. Three studies on the central Amazonian ters of the soil fauna, in terms of biomass, are ants soil fauna (Beck 1967, 1970, 1971) do exist, how- and termites in the central Amazonian rain forest. ever. Beck, working chiefly in the hinterland of It is, therefore, not possible to evaluate the role of Manaus, studied mainly meso- and macro-soil fauna these soil animal groups in the bioenergy flow with- on sites comparablet o our plant biomass study plot. in the central Amazonian rain forest. It is impos- He presentedd ensity data for a series of soil animal sible to evaluate their role from the literature con- groups of which only Isoptera and Formicidae are cerning other tropical regions because of contra- under-representedo wing to the technique of collec- dicting statementsr egardingt he alimentationo f spe- tion which he used (table 3). cific species and population density of different The overwhelmingp redominanceo f Acarina and species in different environments. Regarding term- Collembola, over all other groups, is easily recog- ites, see Lee and Wood (1971), and Krishna and nized from table 3. Other important groups of Weesner (1970). The data given by Wiegert fauna are: Isopoda, Pseudoscorpiones, Araneae, (1970) for Nasutitermes costalis (Holmgren) in a Opiliones, Diplopoda (Chilognatha), Chilopoda,I so- Puerto Rican montane rain forest are inappropriate 6 Fittkau and Klinge This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions because of the different type of forest. further that competition for food between fungi The biomass of faunal elements of the central and faunal primaryd ecomposersi s much stronger in Amazonian rain forest which are important biologi- the tropics than elsewhere. As a consequence of cally, with respect to production, or because of this competition, soil animals are more or less ex- abundanceo r size, is given in figure 2. The values cluded by fungi from the decomposition of organic have been derived from our general observations detritus, and must feed on other material: thus they over the last 1Q years and from observationso f oth- feed on the fungi themselves. Animal primary de- ers, but not from actual counting and weighing, as composers decrease in number, therefore, and sec- stated above. Excepted are: soil fauna, total fresh ondary decomposers increase accordingly. For sec- biomass (84 kg per hectare; Beck 1970, 1971), and ondary decomposers,h owever, there is not enough certain arthropods and vertebrates which we col- organic matter pretreated by primary decomposers. lected in the biomass estimationp lot. Secondaryd ecomposersa lso change their food base Regarding the zoocoenosis of our forest we and change over to feed on fungi. The final result found that the following animal groups have a high is that fungi become the predominant primary de- density: soil fauna (mainly mites and Collembola), composers. Orthoptera (Locustidae,P hasmida), Mantodea,B lat- One of the proofs which Beck gives to support taria, Isoptera, Hemiptera (Heteroptera, Cicadina), his opinion concerns diplopods, which in temperate and Hymenoptera (Formicidae, Vespoidea). Soil- regions are primary decomposers,w hile in the cen- inhabiting invertebrates are the most conspicuous tral Amazonian rain forest they feed exclusively on group comprising the faunal biomass. Below is a fungal mycelia and spores. Regardingo ther tropical detailed discussion of the environment and impor- regions, some other authors have reported on fungi- tance of these soil animals. vorous soil animals. Strickland (1945) referred to The soil contains roots which penetrate more or fungivorous springtails, beetles, and larval Diptera Lesso ne meter into the ground. The main root- in Trinidad. Meyer and Maldague (1957) men- bearing zone, however, is only about 30 cm deep. tioned fungivorous soil animals for the Congo re- rhis upper layer contains about half of all fine roots gion, Bullock and Khoo (1969) for Malaysia, and (Klinge 1972c). Because the forest invertebrates Healey (1970) and Coleman (1970) for soil fauna listed in figure 2 belong to the soil fauna, and be- in general. cause the soil fauna prefers the uppermost soil layer The surface organic matter or humus of the soil supplied with organic debris and humified matter, is present as "moder" in the sense of Kubiena about half the total animal biomass inhabits a zone (1953), or a "transition between raw humus and only 10 cm deep. Thus, in the 40 m vertical range mull" (Jacks et al. 1966). It is characterizedb y an of the ecosystem under discussion, the most im- absence of mixing of organic and mineral matter, portant fraction of the zoocoenosis occurs in less that is, the organic debris layer of moder is separated than 0.5 percent of the total volume. This fraction from the mineral soil below it. The debris layer is of the fauna, however, does not just dwell in this well provided with roots and can, therefore, be re- superficial soil layer, but it also feeds on the mostly moved from the mineral soil like a carpet. Moder organic nutrient matter in it. lacks digging animals or other soil animals which Regardinga limentationo f soil fauna in the trop- could mix the organic and inorganic components of ics, Beck (1970, 1971), in his studies in central the soil. There are few Oligochaeta in the central Amazonia, and Schaller (1960, 1961), referring to Amazonian rain forest soil and other soil-digging South America and to the tropics in general, stated animals are absent (Beck 1971). There are large that tropical soil animals (excluding termites) do earthworms (for example, the Glossoscolecidae not feed so much on organic detritus, as do their Rhinodrilusp riollii Righi), but they do not feed on counterpartsi n temperate regions, but also on the organic detritus (Beck 1971). These earthworms fungi which decompose this forest detritus. In ex- do not cause active leaf burial, but only incorporate plaining this strikingly different feeding behavior casts among the components of the litter layer. The between temperate and tropical soil fauna, Beck casts are easily destroyedb y rain and are not humic- argues that fungi have optimal growth conditions in stained. Madge (1965) observed an earthworm the humid tropics where there are constantly high showing similar behavior in a Nigerian rain forest; humidity, high temperature,a nd acidic soil condi- during the wet season the earthworm,H yperiodrilus tions. This argument has also been stated by Eid- africanus (Beddard), produced 36.4 mt of casts per mann (1942, 1943), Maldague (1958), Maldague hectarep er year. There were 16.4 kg per hectare of and Hilger (1963), and others. Because of these earthwormsi n the soil. In the generally scarce lit- optimal growing conditions for fungi, Beck asserts erature on earthworms in tropical rain forest soils Biomass of Amazonian Rain Forest 7 This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions il faunal witou eptilia lttr A Compositiono f the totol onimol biomoss B Food-base of the total animal biomass Hemiptera Cebidae FormicidMlnae Bradipoidae ~~~~~~~te ruis Orthopteraeraand Formm iaida ereoihneacisa Tayassuidae C Composition of the InCseocmt afaousintiao n D of the mammalianf auna FIGURE 2. Animal biomass, composition and alimentation. (Bullock and Khoo 1969, Madge 1966, Moore and leaves, palm frond bases, etc. are rarely inhabited by Burns 1970, Schulze 1967), there is also informa- animals. Collecting Amblypygi, even by experienced tion on earthwormsw hich feed on leaves of the for- people, is mostly ineffective (even when a high est floor (Lyford 1969). Together with Rhinodrilus price is offered for one collected specimen). Col- priollii, a large terrestrial leech, Liostomus sp., lecting insects using light traps is also ineffective. Herpobdellidae, which probably lives on Glos- The rarity of rodents can be established from the soscolecidae,i s found frequently in the Amazonian fact that collections for blood parasite studies by the forest which we examined. zoology staff of Instituto Nacional de Pesquisas da The amount of animal biomass in the central Amaz?nia had to be suspended because of the dif- Amazonianr ain forest is extremely small when com- ficulty of securing the animals. Hunting for human pared with the plant biomass. The poverty in ani- food is never effective (even by experienced peo- mals is also indicated by the following observations ple). Food wastes at campsites regularly attract of this forest: Temporary water accumulations in only some Scarabaeidaea nd Diptera. On faeces, 8 Fittkau and Klinge This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions only some Diptera may be found. In the biomass which occur here and there in the forest. estimation plot, one uninhabited wasp nest and one Damage attributable to insect plagues was never small bee nest were found. Only one Aviculariidae observed nor reported (Voute 1945-1946, Schneider was collected there, and butterflies were seldom ob- 1939). The opinion that a great proportion of the served. Only termites and ants are frequent, judg- leaves of tropical rain forest plants is consumed by ing by the number of their nests in the earth, on the animals before they fall on the soil appears to us soil surface, and attached to tree trunks and branch- to be rather exaggerated and seems not to be gen- es, and by the trouble they cause to humans working erally valid for all tropical forests. Inspection of liv- in the forest. Curculionidaea nd Cerambycidaea re ing leaves in the course of leaf harvest as part of also relatively frequent. the biomass estimation, as well as of leaves lying The small percentage of animal biomass com- on the soil, never showed any signs of heavy attack pared with the total biomass of the centralA mazoni- by animals. Our observations agree with the state- an rain forest becomes much more evident if we ment of Madge (1969) who, in a study of litter compare similar figures for African steppes and decomposition in Nigeria, expressed the view that savannas, or for other tropical forests. In a mon- primary food consumption of leaf litter appears to tane tropical rain forest in Puerto Rico, the animal be less in the tropics than in temperate regions. biomass comprises 0.1 percent of the total biomass His opinion is also supported by Wanner (1970) (Odum 1970, Odum and Pigeon 1970, Odum et al. who wrote: ". . . direct grazing seems to be a 1970), while in central Amazonia it comprises only minor pathway of energy flow in rain forest as 0.02 percent of the total biomass. In Tanganyika elsewhere." and Uganda grasslands there are 100-300 kg per There is some controversy in the literature re- hectare of large herbivorousa nimals (Wiegert and garding leaves consumed by tropical rain forest Evans 1967). In the East African savannaa re 235.6 animals. The most extreme position is held by Eid- kg per hectareo f these animals,a nd in Ghana forests mann (1942, 1943) and Biichler (according to Mann there are 0.72 kg per hectare of ungulates and 1968) who reported that 25 percent of the leaf primates (Bourliere 1963). In an East African thorn- matter is consumed by insects in West African and bush savanna,h aving a dry matter productiono f 1-7 South American forests. Hopkins (1967) estimated mt per hectare per year, on each hectare there are that 67 percent of the primary production in a 50 kg of ungulates,4 kg of plant-eating small mam- Nigerian rain forest travels along the grazing path- mals, and 250 kg of plant-consuming soil animals way. Bray (1964), however, estimated the loss of which feed on the dry matter produced each year. leaf matter due to herbivorous grazing to be 1.5-2.5 Also present on each hectare are 0.3 kg carnivores percent of the annual leaf production. Odum and (Hendrichs 1970). Ruiz-Reyes (1970) estimated that 7 percent of the It is remarkablet hat the main part of the animal leaf area is consumed by animals of the Puerto Rican biomass of the central Amazonian rain forest is Tabonuco rain forest. Cruz Acosta (1964) stated present in that part of the ecosystem which is well that less than 6 percent of leaf matter was consum- supplied with forest debris and in which the debris ed by insects in a Costa Rican rain forest. Chry- is processed. Correspondinglyt,h e proportion of the somelids and butterfly larvae are mainly responsible animal biomass feeding on living plant matter is for the consumption of living leaves in a West Afri- rather small. The enormous proportion of wood can rain forest (Eidmann 1942). Butterfly larvae in the living plant biomass contributes little to the were only occasionally observed, and in small num- food resourceso f the fauna. Beetles and their larvae, bers, in the central Amazonian rain forest near and termites, are the most important faunal ele- Manaus. ments feeding on wood. Termites are supposed Cruz Acosta (1964) agrees with Bray (1964) to be the most effective wood-eaters,b ut, as stated that leaf utilization by animals is correlated with above, no precise information is available regarding the nutrient content of the leaves. We assume that energy and matter flow through this animal group our finding of a relatively low nutrient content in central Amazonia. in central Amazonian leaf litter supports this view The leaves, which comprise about 2 percent of (Klinge and Rodrigues 1968). Also it has been the living plant biomass of our forest, are utilized frequently observed in Amazonia that where fresh by a relatively small number of taxa and individuals mineral soil has been recently dug and exposed to (Orthoptera, Atta, Coleoptera, Hemiptera, Lepidop- the air, and especially if it is polluted by urine, but- tera, sloths, and parrots;p rimatesa lso consumey oung terflies, bees, and other insects can be found in large leaves, leaf buds, and flowers). Leaf-cuttinga nts are numbers visiting this matter, presumably in search of mostly responsible for the defoliation of small trees nutrients. A heavily perspiring man is strongly at- Biomass of Amazonian Rain Forest 9 This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions tacked by Meliponinae which suck up the sweat; this isms in their digestive tract, as in ruminants, the circumstance is especially true in the highly oligo- tropical termites, and other insects living in sym- trophic Rio Negro area. Both observations seem biosis with microorganisms. These animals can to us to be an indication of the shortage of nutrients digest large quantities of food, such as wood, to in Amazonia. The physico-mechanical properties obtain a minimal nutrient level for their existence. and organic composition of leaves will also play a Our contention that the main energy flow is role in making the leaves either attractive or un- undoubtedly through the detritus food chain (fig. 3) attractive to animals. is strongly supported by Cruz Acosta (1964) and by At any given time, fruits and flowers are only Went and Stark (1968). When studying the bio- present to a small degree (Bourliere 1972) because logical role of soil fungi in the Amazon rain forest, of the low species density of plants (table 2). These Went and Stark proposed first a very intensive fungal plant organs, therefore, represent only a small food development, some fraction of which is in the form base and do not serve as a continuous source of of mycorrhiza. Then they argue that fungi are the nutrients for a very important part of the fauna. primary decomposers of the forest litter and thus Meggers (1971) also adheres to our opinion re- agree with the conclusions of Beck (1970, 1971). garding Amazonian terra firma forests and points out Went and Stark further assert that the nutrients that in this region the distribution of individuals liberated during the decomposition of forest litter of the same plant species is scattered so that ripe by fungi are not released into the soil but are trans- fruits or seeds are not available in concentrated ferred to the tree roots through mycorrhizal fungi numbers. She included aboriginal man in her thesis which are restricted to the surface organic layers and stated that he, subsisting on wild animals and of the soil. They believe that the same fungal my- plants, affected the ecosystem in a way similar to celium acts as primary decomposer and as my- other kinds of large animals. corrhizal mycelium. In the past, reference has been made to a "fauna The dependence of the lower members of the of tree crowns" in tropical rain forests, thus indi- consumer chain on organic detritus and on micro- cating a specific environment in that part of the organisms feeding on it in the terrestrial environ- ecosystem. Our observations agree with the find- ment has its parallel in the aquatic environment of ings of Harrison (1965) regarding vertebrates in central Amazonia in which no primary production Borneo and contradict the assertion that life in the has been observed as yet. There is a striking num- crowns of trees is highly significant. We have ob- ber of such groups of aquatic animals which feed served certain kinds of adaptations of fauna to a on organic micro-drift and on "detritus" (Fittkau life in the crown area and on bark; for example, 1967, 1970b, Sattler 1963, 1967). protective or cryptic coloration of locusts, mantids. In conclusion, we believe that in the tropical cicadas, beetles, and reptiles. On the other hand, rain forest of central Amazonia dead organic mat- our opinion is that the fungi which inhabit the soil ter is processed by microorganisms which channel are the main primary decomposers and are highly organic matter and nutrients through their own significant in processing the forest litter and in con- matter into the consumer chain. This type of forest centrating the low levels of nutrients stored in the occupies the vast level plain built up by Tertiary dead organic matter. It is nevertheless true that sediments (Barreiras Series) of the terra firma which large amounts of plant matter are produced in the is never reached by the annual floods of the drain- tree crown zone of the forest. But the conversion ageways. The hinterland of Manaus where we con- of the bulk of plant matter to living animal matter ducted our studies forms part of central Amazonia must pass through the bottleneck of the dead plant and lies between the Rio Negro and the Rio Soli- matter which accumulates in the litter layer of the moes. It is clearly to be distinguished from the forest soil. This zone is inhabited by soil fungi sandy Rio Negro basin, despite the fact that small which can incorporate nutrients of the forest detritus sand patches are spread throughout central Amazonia. into their mycelia, thus concentrating the limited The rain forest of the central Amazonian terra firma nutrients and making them available for organisms is characterized by a very high number of dicotyle- at the next higher level of the energy pathway. donous trees and by a great proportion of palms The critical point in the energy and nutrient and lianas. Small, slender trees predominate. Total transformation is the inability of soil invertebrates plant biomass of the forest is about 1100 mt of to process large quantities of dead plant matter into fresh matter per hectare while its animal biomass living matter in strong competition with fungi as is only 0.2 mt per hectare. Invertebrates predomin- described above. Exceptions to this pattern are ate. Insects are well represented. Soil fauna, com- found in taxa having representatives of microorgan- prising approximately 50-75 percent of the animal 10 Fittkau and Klinge This content downloaded from 132.236.27.111 on Tue, 30 Jun 2015 18:11:33 UTC All use subject to JSTOR Terms and Conditions