The thecamoebae ( Protozoa , Rhizopoda ) of small Amazonian forest streams and their possible use as indicator organisms for waterquality ( 1 )

A fauna das tecamebas em detritos orgânicos foi analisada em 7 igarapes da regiao de Manaus, entre eles o da "Bacia Modelo" (Reserva Florestal "Km 60" do INPA). A metodologia de analise foi publicada anteriormente (Walker & Lages 1980). Uma lista com figuras das 129 morfo-especies caracterizadas encontra-se no anexo deste trabalho, 39 especies foram observadas em condicao viva, e em 20 delas foram determinadas particulas de alimentos (75 tecamebas foram observadas). O alimento principal sao fungos que decampoem a materia orgânica na agua (folhico submerso, detritos; 67 individuos de 18 morfotipos foram observados inferingo fungo). A analise quantitativa mostra que a maioria das especies de cada igarape foi encontrada pela amostragem efetuada (64 - 75/igarape) mas no caso de coletas em demais igarapes espera-se um numero consideravel de novas especies. A frequencia e distribuicao das morfo-especies, assim como a densidade total, e especifica para cada igarape, e parece estar ligada a qualidade quimica da agua. A densidade total e de 360-530 tecamebas/1ml detritos em suspensao em aguas pretas, e de 45-330/1 ml em aguas cristalinas. A distância geografica, separando os igarapes, tambem contribui a uma reducao na similaridade em morfo-especies entre varios igarapes. indices de diversidade (Williams' α) e de similaridade (% morfotipos em comum) encontram-se na Tabela 2 e na Fig. 4. A distribuicao das especies frequentes (encontradas em > 10% das amostras) em 5 igarapes da Bacia do rio Taruma (Fig. 1) e apresentada na Tabela 3. Todos estes resultados sao suporte a hipotese de que populacoes integradas de tecamebas, tambem morfotipos especificos, podem servir como Indicadores para a qualidade das aguas de igarapes florestais em condicao natural, e sob perturbacao. Assim sendo, serao no futuro comparadas as populacoes de tecamebas nos igarapes da "Bacia I" (Bacia Modelo) e da "Bacia III" (Bacia do Manejo Florestal) na Reserva Florestal "Km 60" do do INPA. Nota-se que a riqueza de um total de 129 morfo-especies em 7 igarapes amazonicos (drenando menos de 50 km2) corresponde aproximadamente a fauna integral de tecamebas de toda Holanda (Cailleux, 1978).


INTRODUCTION
This study is not an attempt at thecamoebae systematics -this should be emphasized right at the start -rather, it is the adventure an unsuspecting biologist blunders into when taking a first look at the fauna of Amazonian "Terra Firme" streams.The density and variety of thecamoebae in the substratum of organic detritus is so staggering that one wants to know more about it, and one begins to wonder whether this phenomenon might not serve a useful purpose.What are these cells living off in these almost distilled waters?What is their role in the food web of the general Ilse Walker (2) stream fauna?Are species and/or their frequencies characteristic for particular streams and for the quality of their waters?Could the distance of geographic separation between river systems with their thecamoebae spectra tell us something about the process of evolution and speciation?
The analysis presented in this study concentrates on the question whether the spectrum of species distribution and frequency could serve as indicator for water quality, specifically also with the aim to compare, in future, the streams of INPA's experimental basins, the "Bacia Modelo" (Bacia I) in its primary, undisturbed condition, and the "Bacia do Manejo" (Bacia Ill) before and during management.Observations on food ingestion are reported and throw some light on the possible ecological role of these thecamebians in their habitat.Evolutionary aspects are barely hinted at, discussion of these problems is only possible after further collection and analysis.
For the purpose in mind it was necessary to recognize and to characterize distinct morphotypes.This has been done as far as was possible by a non-specialist.A list of these types with accompanying figures is attached at the end of this paper.The attempt to assign them to systematic groups should be regarded as what it is: the struggle of the unspecialized observer to orientate himself in an overwhelming complexity of forms; doubts are expressed by a profusion of questionmarks accompar.yingthe suggested systematic groups.No attempt has been made to name species, and whenever the term M species" is [,

~'
1 I used, its should be understood in the sense of "morphotype".The numbers identifying the various morphotypes are a compromise between the process of finding them during analysis of the coliected material and the first impression of similarity with already found types.After much debate I decided to stick to this number system despite its arbitrariness.This was, because the thousands of camera lucida drawings of the primary protocol and of the concocted tables are identified by these numbers, so that re-numbering would certainly have resulted in confusion, besides, renumbering according to supposedly better systematic interpretation would still reflect the degree of ignorance of the observer.
If, in the face of all these shortcomings a dozen pages of this Journal are filled with a list of morphotypes, it is for the following reasons: firstly, to convince the reader that the profusion of forms is not an emanation of my imagination; secondly, to preserve the information -the pristine condition of thase streams may have vanished by the time a dedicated systematist finds time and opportunity to study their thecamoebae fauna; and thirdly, to incite the interest of protozoologists in this unique fauna and to provide to future investigations all possible assistence, that may result from this study.

THE THECAMOEBAE (PROTOZOA, RHIZOPODA}
The thecamocbae are prevalently fresh water organisms; they also occur in moist, terrestrial habitats (soils, mosses, epiphytes ect); they are less common in marine environments.Unlike their marine relatives, the Foraminifera, which secrete tests of calcium carbonate, the thecamoebae construct their tests, or thecae, of particles selected in their environment (mostly sand), which they glue together with chitinous excretions.The relative ::imo1:1nt of foreign particles and secretion varies greatly; the thecae of the enormous genus Diffugia, for instance, consist of sand except for the minimum of excretion necessary 80 -to glue the particles together, whereas Nebeliidae use virtually no foreign particles, indeed, this family and the Euglyphidae secrete as accomphished, geometric, proteinous thecae as do the calcareous Foraminifera.The construction of thecae is specific to genera and species, as may be verified in any introductory text to Protozoa (for instance Kudo, 1966).Description of species relies often on the characteristics of the test alone, as the living cell may never have been observed.
According to Cailleux (1978) and Decloitre ( 1964Decloitre ( -1976) ) the total number of thecamoebae species and varieties described, worldwide, until 1975, is about 1600; only 78 were, until then, reported from Brasil, some of which are reported in a recent atlas of thecamoebae (Ogden & Hedley, 1980).Later, Bonnet (1979) described some soil amoebas from Brasll.I have no knowledge of a published report on Amazonian thecamoebae.
The material presented in this study was collected in the fine, organic detritus that accumulates on the sandy bottom of streams in places of low water flow.This detritus consists largely of faeces of aquatic ins~ct larvae which live in submerged leaf litter.As was verified by spot checks, thecamoebae occur also on the surfaces of dead, submerged leaves, on mosses and roots along river banks at high water level etc.However, efficient, quantitative sampling was possible only in the detritus habitat.
Examination of the quantitative samples led to the recongnition of 129 mo;•photypes which are presented in a list with accompanying figures.This "List of morphotypes" with 7 Plates is attached to the end of this paper in order to avoid interruption of the text in this chapter.It is an appendix with regard to position only -it is essential for the understanding of the results and the conclusions

THE STREAMS
The map in Fig. 1 shows the forest region drained by the three river systems Taruma, Cuieiros and Tarumazinho (= Taruma-Mi-Walker rim in Table 2).This map is re-drawn after the map "Projeto RADAMBRASIL", 1 : 250.000, which, in turn, is based on satellite photographs.Hence, the courses of very small forested streams are not clearly recognizable, the positions of the collection sites should therefore ~e regarded as approximations.With the exception of some spot checks in the rivers Ypiranga and Tarumazinho, all collections were made in such very small streams, 2°30 1 Fig. 1 -Area drained by the three river systems Taruma, Tarumazinho ( = Taruma-Mirim =: little Taruma) and Cuieiras.AA = Arquipelago das Anavilhanas.B = hydrographic basin with the stream "lgarape Bacia" and its two tributaries issueing near Km 15 and Km 17 of the road ZF 2. CA = lgarape da Cachceira.Cm = forest reserve "Campina" with black water stream "lgarape da Campina".D = forest reserve "Reserva Ducke" (Km 26 on road 010 to ltacoatiara) with the clear water stream "Barro Branco" draining into Taruma.Res = southern boundary of the forest reserva "Reserva Km 60" (according to new numeration situatad at Km 45 of highway BR 174 to Boa Vista).Yp = lgarape Ypiranga.The black dots indicate sample sites of thecamoebae and of other stream fauna not dealt with in this paper (Kensley & Walker, 1982).Problems with exact position of streams and sampling sites see above.(map after "Projeto RADAMBRASIL" 1: 250,000).
within the first 1-7 km of their source.The major collections were carried out between 1978 and 1980 in five streams of the Tarumasystem; during the high-water period, in June 1978 and 1979, the small tributaries of the "lgarape ( = small river or stream) da Cachoeira" were accessible by boat, and a limited number of samples were taken on these occasions.

Sampling.
A detailed description and critical analysis of the sampling method was published previously.(Walker & Lages, 1980).Samples of approximately 8 ml of detritus and sand were collected at distances of 6 to 12 meters in up-stream direction in series of usually 12 such samples.In the laboratory the material was placed in petri dishes under a low-power microscope, and I allowed myself 10 min to search for and to remove, unbroken thecae from this substrate.In the text, this 10-min-sample is referred to as "the sample".The number of thecae removed is proportional to the density of thecae per volume of substrate, (i.e. per volume of settled detritus in aqueous suspension, the dry weight of which is less than 100 µg/ml (including sand), thus, these samples reveal the relative density of thecae between streams.Series of 10-minremovals from a single sample allow for the approximate calculation of the absolute density of the thecae.After the 10-min-removal the thecae were examined under the hingh power microscope, assigned to various morphotypes and, whenever it seemed necessary or desirable, traced via camera lucida.Special attention was given to amoebae that were alive and active under the microscope.Documentation: Camera lucida drawings were made invariably at the microscopic magnification of 160 x and under fixed camera lucida setting with the exception of some very small forms which were drawn at higher magnifications.However, this necessitates observation under oil immersion, and this was not always possible during quantitative analysis.For this reason, some of the smallest forms are not adequately documented by drawings.Series of camera lucida drawings were transferred to tracing paper with a japanese felt brush.This "catalogue of morphotypes" displays the range of variation of the frequent types and a complete record of the forms that appeared only a few times.

Walker
--Details shown in these drawings inciude largest and smallest identifyable paritcle size with their typical distribution in tests constructed with sand, and in the Euglyphidae, the patterns of the hyaline test.(Copies of this catalogue will be deposited in the Invertebrate Collection, Divisao de Entomologia, INPA, and in the "Archives du service de documentation" of the CNRS, 15 Quai Anatole France, Paris).
From this catalogue the here published documentation of a total of 129 morphotypes was extracted, namely a "list of morphotypes" with accompanying figures on seven plates , both attached at the end of this paper with the necessary explanations.The information given by text and figures is complementary, what is evident from the figures is not repeated in the text unless for special reason.The arrangement of the morphotypes on the plates is a compromize between numbered sequence of types, systematic position and graphical necessity.In the list of morphotypes the plate number is indicated only for those which are difficult to localize.(See also pp. 90, 91-103).
Presentation of results: Tables and figures with accompanying explanations should need no further comment on methology with the exception of the concept of frequency used to express the abundance of morphotypes.
The larger amoebae and ciliates from amazonian waters I have so far observed, and occasionally bred over long periods (Walker, 1978).all show a similar population behaviour, in that they move around randomly without reproduction for several days in absence of food, or if food is scarce; yet when meeting with a food source they stay in this place and reproduce intensely, presumably as long as the food is adequate (also several days).A single detritus sample may thus ceintain several individuals of a rare species and still be deficient in frequent species For this reason, the frequency of morphotypes is not given in percent of individuals recovered, but in percent of samples in which the morphotypes appeared.

Abundance and distribution of thecamoebae
Sequence of sampling and appearance of new morphotypes.Fig. 2 relates the finding of new morphotypes to the sequence of sampling and to the magnitude of samples.It is evident tha! a serieo of 12 samples contains the majority of species present in a given area (100 -200 m along stream) at a particular time.However, if such series are repeated at intervals of several months, a considerable number of new types appears; only after the third series does the cumulative curve of the lg.da Bacia level off.Thus, we would still expect some new species in each stream with further sample series.The parallel condition of the two curves (Bacia, Barro Branco) is reassuring: it means firstly, that two streams with approximately similar numbers of samples can be compared; and secondly, that the rise in the number of species is not simply reflecting the learning process of the observer.
because the second point of the Barro Branco was determined almost a year before the first point of the "Bacia".The strong rise before Aug. -Sept.1979 is also characteristic of the Campina stream (Table 2) and is due to an increase of the number of small euglyphid types after an exceptionally long, hot and dry season.
The cumulative curve for individual streams as samples shows that any collector of thecamoebae can expect a fair number of additional species with each stream sampled, and this even in the proximity of Manaus alone.
The reasonableness to expect new species with repeated sample series is also shown by the values of the index of diversity (Williams' a.., Table 2): the index of the totals is larger than that of the individual sample series (Southwood, 1979;Williams, 1947).1).It is not surprising, then, that 18 of a total of 119 types in the 5 Taruma-streams, that is 15. 1 % of the types described for this river system, appear only in a single sample, and that the fraction of rare species (0,5 -5% of samples) is 79/119 = 66.3%.These results mean that some caution must be exercised when the quantitative spectra of species are compared between streams.With the present sampling conditions, even two streams with ideally identical species inventories could not show this by an identical list of morphotypes.

Frequency of morphotypes (Table
The highest abundances of types in terms of individuals per population sampled in a given stream, were r~ached by type 1. 2 in the Campina stream with 12.6% of all thecamoebae typed and counted, and by type 7. 5 in the Bacia stream with 13.6%. Pattern of distribution in space (Fig. 3).Not everything is pure probabilistics, though.

If we regard the relation between means ( x) and variance (s 2
). bearing in mind that in case of random distribution we expect x = s 2 , we find that density of individuals is clumped (s 2 > x), whereas number of morphotypes per sample has a tendency to regularity s 2 < x.
Both deviations from randomness are highly The thecamoebae ... significant P < 0. 01, sequential analysis, (Documenta Geigy, 1960).This pattern of distribution may be explained in biological terms.
Microorganisms tend to multiply opportunistically and rapidly on an accidental food source.The higher the overall population density, the more individuals and the more morphotypes wi 11 be found near such a food source.The experimental, and purely local, rise of population density in this area will reduce the chance of a later arrival to build up a population boom of its own type.In case of low population density, exponential build-up of local population density of accordingly less morphotypes will still prevent late, new arrivals to proliferate on the spot.These conditions result in the pattern observed: increased tendency to clumped distribution of individuals with increased density ( x), but consistent tendency towards regularity in number of morphotypes, regardless of their mean number.
These conditions may slightly distort comparative data between streams of high and low density of thecamoebae, specifically in case of too few samples (which does not apply in our case, though).For instance the mean number of morphotypes per-sample in the Barro Branco series ranges from 6. 5 -13, and in the Campina stream from 10 -16 only, despite the fact that density of individuals is 3 -5 times higher in the Campina, and that the total number of morphotypes is almost equal in the two streams (Table 2).

Comparison between streams
Density and diversity (Table 2).All precautions taken, we may proceed to compare the thecamoebae faunas of the various streams.
Neither the data in Table 2 nor the foregoing considerations (Fig. 2 etc) allow for the assumption that the total number of morphotypes recovered per stream varies between streams (P>0.15,sign test).To the contrary, the values are remarkably similar, particularly so if we consider that the highest value Ug. da Bacia) is based on a higher number of samples (41 as opposed to 36 in the other streams).The striking difference between streams lies in the density of their populations.Only the difference between the "Bacia" and its two tributaries "Km 15" and "\(m 17" is insignificant (P > 0. 05). the probabilities for similarity of density is below 0. 05 for any other two streams, and for the "Campina" and Barro Branco it is less than 1 % (Sign tests).These differences seem to be correlated with the "blackness" of the water.Black water is several times richer in thecamoebae than crystalline water, whereas the mixed water of the lg.da Bacia shows an intermediate value.
The few data of the Cuieiras system are in this repect in complete agreement with those of the Taruma-system.It is possible, that 86 •-higher conductivity, that is, higher mineral content, also favours population density to some extent, this may account for the higher value in "km 17" than in "km 15", but the difference is not significant, and single determinations of conductivity do not allow tor conclusions to be drawn (see p. 82).
The combination of similar number of species and massive difference in density results in marked variation of the index of diversity (Williams' a) between streams.
Similarity (Fig. 4, Table 3).Considering all morphotypes in Table 1, the rare and frequent ones, comparison of similarity between streams (Fig. 4) would agree with a compromise between geographic distance (Fig. 1) and water quality.The closest simiiarities are found within the hydrographic basin, between the three streams "Bacia", "km 15" and "km 17".Let us remember that the high frequency of rare species and the expectation of new species still to be found, do not allow for 100% common species, even in case of theoretical identity of two streams.The lg. da Bacia is equally simiiar -or dissimilar-to the ~arro Branco and the lg.da Campina, although, geographically, it is much nearer to the latter.The Campina stream is as different from "km 15" as from the much remoter Barro Branco; but then the waters of Barro Branco and "km 15" are quasi identical as far as tested (p.82).In Table 3 the rare species are excluded, only those morphotypes are considered that occur with a frequency of about 10% (of samples!see p. 83) in at least one stream.Even then, the Campina stream and Barro Branco have the lowest percentage of morphotypes in common (51 %).followed by "Campina" and "Bacia" (54%), whereas Barro Branco and "Bacia" are much more simiiar (79%).Apparently, some rather frequent species are absent from original, non-mixed, black water as that from "Campina ".This is conspicuously true for centropyxid forms (types 9. 1 -9.6).for type 5. 1 and for the two granulo-reticulose types 4. 1 and 7. 2. The streams of the hydrographic basin are almost identical (92% for "Bacia" and "km 17" for example).Table 3 is far too complex to be trans• lated into words, but a few conspicuous differences may be pointed out.Type 1 .6 is exclusive to the Campina stream; the frequent type 4. 2 is virtually absent from the "Campina ";types 3.4 and 3.4.1 were found in black water only.Black water is particularly rich in Euglyphidae; type 9 .4was not found in the Cuieiras system, nor were types 4. 1 and 7. 12 (Granuloreticulosa (?)).A few observations point to interesting questions with regard to speciation, in that closely related forms are separated by river basins.Thus types 2. 1 and 2. 2 for the hydrographic basin and the "Campina "; types 3. 4 and 3. 4. 1 are almost separated between these two basins; type 7. 21. 1 is specific to the Cuieiras system, but 7 .21 is found in both river systems, Cuieiras and Taruma.Some observations not shown in this table may be mentioned in this context: the dozens of drawings made via camera lucida indicate that the proportions and/or sizes vary between the Taruma and Cuieiras systems for the forms 4.2, 6.1 and 2.2; and for type 10.1 between the hydrographic basin ("Bacia", "km 15", "km 17") and the "Campina".

BIOLOGICAL OBSERVATIONS
Accumulation of empty thecae or biological pattern?Only a fraction of the collected thecae contained a living cell, although many more than noted in the list of morphotypes, where living organisms are mentioned only in cases of extended pseudopods or of food •.tacuoles with recognizable content.It may be Walker argued, therefore, that the material sampled represents an accumulation of dead, empty thecamoebae without any relation to the actual living population.The following observations speak against this interpretation: 1 .Organic detritus is light and is washed away with heavy rains; 51 morphotypes were found in 60 samples collected in free water after a heavy rain in the lg.da Cachoeira; this shows that the thecamoebae get washed away together with the detritus.2. Empty thecae do not accumulate on the sr.r:dy substratum; samples poor in detritus and rich in sand are usually also poor in thecamoebae.3. empty thecae do not accumulate in the zone of accumulation of light sand particles and of detritus !n the meanders in the larger streams (for instance 5 samples of the Rio Tarumazinho showed a mean of 16±5.3 thecae, included all broken ones).4. The pattern of distribution in Fig. 3. <md the pertinent comments on p. 83. 5.The correlation between water quality and density of thecamoebae.
All these observations do not excluds the possibility that some specimens might get washed into the streams from overhanging vegetation and from surrounding soils.A few such samples in the vicinity of the Campina stream turned up one specimen each of types 7.21, 8.5.1, 10.12 and 9.3.2(also recovered alive from stream detritus!). and 3 specimens of type 8. 7.
Still, all observations taken together point to the likely conclusion that the thecae recovered from stream detritus are characteristic for a population living in this substrate, and that empty thecae decay rapidly, probably as a result of decomposition of their proteinous fraction.
Feeding behaviour and trophic position.Al I living amoebae observed, with the exception of types 3. 3, 4. 1 and 7. 12 (Granuloreticulosa (?)), carried food vaculoes with clumped, brown material.The handy conclusion that offers itself is, that they are "detritus feeders".Observations of several dozen active cells under the microscope, though, show a very The thecamoebae .... different picture.If an extended pseudopod touches a clump of detritus, its plasma flux accelerates in that direction, and the entire amoebae may glide into this substrate.Its pseudopods then probe through this material, a multitude of particles gliding along their plasma margin towards the pseudostome.However, only the meticulously se!ected particles mentioned in the list of morphotypes get ingested, the bulk of the detritus is rejected before it enters the pseudostome.The overwhelmingly favoured particles were bits ct fungus hyphae and spores, and to a lesser extent living diatoms and small euglyphids (seep.104).
The fungus material in detritus is, according to numerous observations, undoubtedly the same as that found within dead, submerged leaves.These fungi are aquatic leaf decomposers, and the 18 morphotypes (and probably many more as judged from the general similar appearance of the contents of food vacuoles) of thecamoebae, that were found ingesting them, must be regarded as primary consumers.Their prodigious density, notably in black waters, suggests that thecamoebae lend no support to Janzen's (1974)   theory of the toxicity of amazonian black water.

GENERAL CONCLUSIONS
Sampling for the "normal" thecamoebae fauna of a stream: A series of 1 O -12 "tenminutes-samples" (Walker & Lages, 1980) is sufficient to characterize the thecamoebae fauna of a particular stream over an extension of 100 -200 m at a particular time, and six such series at 2-monthly intervals should reveal the "normal" spectrum of diversity and abundance of morphotypes of a particular stream.With this sampling proceedure, detection of 70 -80 morphotypes would indicate the normal abundance of species in a small, undisturbed, central amazonian forest stream (area drained ;S; 8 km 2 ).
Thecamoebae as indicator organisms for water quality: All results of the qualitative and quantitative analysis indicate that spectrum and frequency of morphotypes, as well as general density, are dependent on water quality, and that, therefore, populations of thecamoebae and/or single species r.1ay be suitable indicators for water quality.

Ecological position:
The thecamoebae with identifiable food particles were found to be primarily consumers of fungi which decompose submerged leaf litter and organic detritus of animals which, in turn, live off submerged litter.This points to the important fact that the food web of amazonian forest streams is essentially dependent on decomposition of dead, organic material provided by the forest, rather than on primary production within the streams (algae, water plants).For these reasons deforestation of these drainage basins would have altogether devastating consequences for the fauna of these streams.
Diversity of thecamoebae word/wide and in the Amazon forest: Cailleux (1978) published a graph in which the number of thecamoebae species known is related to the area studied.This graph includes the roughly 1600 species known up to 1976 and all countries which provided data.This analysis allows for the construction of a fairly reasonable curve, on which the number of known species increases with the size of the area studied.On this curve, 129 morphotypes, the total of our list of morphotypes, corresponds to an area of 30.000 -40. 000 km 2 • The real points nearest to 130 species are the total number of species known from Holland ( ~ 140), Yugoslavia ( ~ 120) or from Scotland ( ~ 130) for example.A generous estimate of the area drained by the five Taruma streams sampled, with their 119 species, is about 30 km 2 • The low number of species in the countries cited above does not necessarily reflect ignorance about this fauna; a recent revision of the Genus Trinema alone has a bibliography of some 130 publications (Decloitre, 1981).Thus, the thecamoebae fauna presented in this paper is just one more example for the amazing diversity of organisms in the amazonian forest.

APPENDIX LIST OF MORPHOTYPES; EXPLANATIONS
The morphotypes are arranged in numerical order, that is, in the historical/semi--systematical fashion explained in the introduction.In the cases of gross, systematic misplacement, the type numbers of other representatives of the respective Family are given in brackets.The attempts to classify the material are based on two documents mainly: an unphished list of the Thecamoebae of the world, by L. Decloitre, on 35 mm film at the "Archives originales du service de documentation du GNAS", 15 Ouai Anatole France, Paris.This list contains the known forms up to 1974, it is filmed in form of separate parts numbered 401 -404 (with sub-sections), and the numbers of the film spools are 38.850, 388.876.Each morphotype is represented by a drawing.The parts with forms similar to the amazonian morphotypes in our list are: 404 Faune Thecamoebienne du monde; 403 Faune Thecamoebienne d'Europe par pays et regions; 404 (6) Faune thecamoebienne du monde, especes nouvelles.The second document consulted for classification was "Treatise on invertebrate paleontology", Vol. 1, the Chapter "Thecamoebians" by A. R. Loeblich & H. Tappan (1978).
In the list of morphotypes, the source led to identification is indicated in parenthesis: (Moore 1978), (Decloitre 404); doubts are expressed by a question mark.
All authentic observations on live material are reported, and the number of amoebae to which the observations refer are given in parenthesis.With the exception of the very small forms all morphotypes are represented by figures drawn to scale by means of the camera lucida (Plates 1 -7).Approximate measures given in the list are therefore added merely to facilitate comparison with descriptions from the literature.Measurements and abreviations: Given are the two largest diameters (or their range) at right angles, and CS (for Cross-Section) indicates whether the section through the horizontal plane along the
7. 21. 1.As type 7. 21, but distinctly larger species: (180 -190) x (130 -140) µm.Heleopera sp., (Moore, 1978)..1.1, 1.1.2, 1.3.1, 5.2, 5.3, 5.4, with a total of 58 amoebae observed in living condition.Secondly, types that build their tests with rather coarse sand of any shape, with sometimes intermittent zones of tightly fitted, smallest particles.They appear to be opportunistic in the choice of their building materials, this, at least, is the impression of the observer.Their pseudopodia do not contain larger crystals.This sub-group includes types 1. 1, 1 .2, 1. 3, (all three perhaps belonging to the same species), 2.1, 2.2, 3.1 and 5.1, with a total of 60 amoebae observed in living condition.In the course of analysis it was the distinction between pseudopodia with, and without, crystals that led to the recognition that the architecture of the tests, too, was different, and that, therefore, types 1 . 1 .1, 1 .2. 1 and 1 .3. 1 are probably different species from types 1 .1, 1 .2 and 1 .3. For these reasons I suspect that morphotype 1 .5 may contain a mixture of species, as pseudopods with, and without, crystals were observed.Morphotypes 3. 4 and 4. 2 are alltogether different species with a very distinct architecture.

-
Of the 75 individuals, belonging to 20 morphotypes, which contained identifyable food particles or were actually observed to ingest food, 67 individuals belonging to 18 morphotypes ingested fungus hyphae and fungus spores.Other food particles include smallest thecamoebae (Arcellidae, Cryptodiffugiaceae and small Euglyphidae).diatoms and starch grain~ of an unidentified fruit; only a single observation refers to a possible uptake of bacteria (type 4 .1).

Fig. 3 -
Fig. 3 -Pattern of distribution.s2 = variance and x =mean of 12 samples each, open circles showing number of thecamoebae with tendency to clumping (concentration above 45°-line),and black circles showing number of morphotypes with tendency towards overdispersal (concentration below 45°-line).
Fig.4-Morphotypes common to two streams in % of total number occuring in the two streams represented by the branches flanking the numbers.* : % morphotypes common to all 5 streams of the Taruma system.

ACKNOWLEDGMENTS1
am thanking Dr. A. Berkaloff of the CNRS of France for the permission to use the facilites of their "Archives du service de documentation" in Paris; Dr. J. Cheveaugeon (ORSTOM) for the gift of a copy of Decloitre's catalogue of the the-104 -camoebians of ~he world (35 mm-film, will be deposited in the Collection of invertebrates, Divisao de Entomologia, INPA); Prof. Dr. P. Vanzolini, Dr. L. Decloitre and Mr. 1.Leipnitz for important information on the subject of thecamoebae fauna; Dr. P. Henderson and Dr. H. Schubart for critically reading the manuscript; Miss C. Botinelly for the difficult task of typing the English script, and Mr. J. B. da Silva and Mr. J.J. Palheta for their patient and concienscous help in the field.

TABLE 3 -
Frequency of morphotypes in amazonian streams in % of samples (n = 100%) that contain morphotype.only those morphotypes are considered that appeared at least in one stream with a frequency of ,.., 10%.Thick•set, spaced numbers: frequencies of > 45%.: morphotype absent.One, 2 or 3 dots: type present in 1,2 os 3 samples (adding up to less than 10%).