Multi-element chemistry of some Amazonian waters and soils
K.O. Konhauser", W.S. Fyfe a and B.I. Kronberg b aDepartment of Geology, University of Western Ontario, London, Ont. N6A 5B7, Canada
bDepartment of Geology, Lakehead University, Thunder Bay, Ont. P7B 5El, Canada
(Received November 19, 1992; revised and accepted April 28, 1993)
ABSTRACT
The rivers flowing through the Amazon Basin are both physically and chemically heterogeneous. Through detailed geo- chemical analyses, variability is shown to be primarily controlled by both substrate lithology in the source region and soil geochemistry in the erosional regime. The solute-rich waters of the Rio Solimres reflect the drainage from the Andes Mountains and the fertility of the varzea. In contrast, the solute-deficient waters of the Rio Negro reflect the infertility of the lateritic and podsolitic terrains of the Central Amazon. Rivers flowing through the Precambrian Shield have an inter- mediate composition. Such observations suggest that it is possible to classify the chemical composition of Amazonian rivers according to the soil types in their catchment regions and vice versa. These findings have profound implications for using water chemistry as an indicator of the agricultural and mineral potential of a region. Results from this study also confirm that precipitation is a significant source of major cations and trace metals to solute-deficient river systems such as the Rio Negro.
1. Introduction
The Amazon Basin is composed of the world's largest tropical forest ecosystem, cov- ering an area of > 7.106 km 2. The basin ex- tends from 5 °N to 17 ° S and from 46 °W to 79°W (Sioli, 1984) - - almost the entire con- tinental width. The Basin is drained by the 6518 km long Amazon River, which is fed by more than 1000 tributaries (Salati and Vose, 1984). The Amazon's average annual dis- charge of 175,000 m 3 s-1 is 4.6 times that of the Congo River and 11 times that of the Mis- sissippi River (Oltman, 1965 ). In both drain- age area and volume of discharge the Amazon is the world's largest river (Gibbs, 1967 ).
Early studies in the Amazon Basin showed that the waters were chemically and physically very heterogeneous. The first classification of Amazonian rivers was based on their physical appearance (Sioli, 1950). Whitewater rivers are rich in dissolved solutes and are extremely
turbid owing to their high concentrations of suspended sediment. Blackwater rivers are rel- atively infertile rivers characterized by low suspended sediment yields and their "tea-col- ored" (Hedges et al., 1986, p.719) acidic waters that are rich in dissolved humic mate- rial (Ertel et al., 1986). Finally, clearwater riv- ers are relatively transparent, green-colored rivers that are neither turbid with detrital ma- terials or colored by humic compounds (Sioli, 1984).
The chemistry of Amazonian rivers have long been attributed both to the inhomogene- ity of the geological source and to the erosional regime through which the rivers flow. Early studies by Raimondi (1884) and Katzer (1897) observed that the low dissolved inor- ganic concentrations in some lowland rivers contrasted with those of rivers draining the Andes. It was later shown that the rivers flow- ing through the central lowlands were typically blackwater in composition, whitewater streams
were distinctive of the waters draining the An- des Mountains, while clearwater types were characteristic of rivers originating in the Pre- cambrian Shields (Sioli, 1968, 1975; Stallard and Edmond, 1983 ).
More recently, several papers have begun to deal specifically with the elemental composi- tion of Amazonian rivers (Gibbs, 1970, 1972, 1973, 1977; Boyle, 1978; Stallard, 1978; Furch et al., 1982; Stallard and Edmond, 1983; Furch, 1984). Studies such as Stallard and Edmond ( 1983) have indicated that the distribution of major cations and anions in the dissolved load were controlled by substrate lithology in the source regions. This relationship was later ex- tended to several trace elements (Furch, 1984).
On the basis of distinct differences in geol- ogy, soil types and vegetation, Fittkau et al. (1975) have divided the Amazon Basin into three major geochemical provinces (Fig. 1 ): the western peripheral region, where soils and waters are rich in nutrients; the Central Ama-
zon, an area where the soils and waters are nu- trient-deficient; and the northern and south- ern peripheral regions with intermediate compositions (Furch, 1984).
In the western peripheral region, the soils are directly influenced by erosion of the different lithologies (acid to intermediate volcanics, shales, limestones and arenitic sandstones) in the headwater regions of the Andes Mountains (StaUard and Edmond, 1983 ). In these geolog- ically active sites, the addition of new volcanic materials and substantial weathering along the steep slopes, result in enormous quantities of fresh detrital material being carried down- stream. The whitewater rivers are estimated to carry ~ 85% of the total suspended solids dis- charged by the Amazon River at its mouth (Gibbs, 1967). This sediment is systemati- cally deposited over the entire catchment area from the broad valleys at the foot of the Andes to the central Amazonian lowlands, and along the banks and floodplains of the river systems (Sioli, 1975). The resulting alluvial soils,
72 60 48 36
12
Fig. 1. The major geochemical provinces of the Amazon Basin, after Fittkau et al. ( 1975 ). Location of the study areas in Manaus and Caraj~is are shown.
MULTI-ELEMENT CHEMISTRY OF SOME AMAZONIAN WATERS AND SOILS 15 7
known as the varzea, represent the only natu- rally fertile terrains throughout the Amazon (Klinge et al., 1981 ). The constant fertility of these soils are sustained by the annual deposi- tion of new sediments during seasonal floods (Sioli, 1975).
In contrast, the Central Amazon is com- posed of highly weathered Tertiary and Pleis- tocene sediments of fluvial and lacustrine ori- gin (Fittkau et al., 1975). In these lowlands, lack of exposed rock and intense chemical weathering of the humid tropics over millions of years has resulted in the development of thick, lateritic soils (Kronberg et al., 1979) with low erosional rates (Stallard and Ed- mond, 1983). In the Rio Negro Basin, open caatinga forests and igapo forests (a forest type that is adapted to partial or complete flood- ing) allow the heavy rains to wash away the fine clay particles, leaving behind sandy, pod- solitic soils that are characterized by a bleached quartz A horizon and an underlying alumi- nous clay layer, cemented with humic mate- rials derived from decaying surface vegetation (Klinge, 1965; Leenheer, 1980). River origi- nating in these extremely infertile soils contain low levels of dissolved solids and suspended matter (Sioli, 1984), and are frequently col- ored black due to leaching of the humic mate- rials (Ertel et al., 1986).
The areas in the northern and southern pe- ripheral regions are dominated by low-lying crystalline platforms of the Precambrian Shield (Klammer 1984), composed of granites, gneisses, acid and basic volcanics, and meta- sediments (Putzer, 1984). In these stable cra- tonic regions the soils tend to be del~cient in nutrients (Fyfe et al., 1983), giving rise to the poor soils of the Amazonian high grounds m the so-called terra firme. The closed terra firme forest (which protects the soil against direct impact of rainfall, thereby reducing surface erosion) generally is characterized by clear- water rivers (Sioli, 1984). Due to local condi- tions and seasonal fluctuations these rivers may assume black or whitewater characteristics, lo-
cally or for short time periods (Sioli, 1967 ). Despite the great compilation of published
work on the chemistry of Amazonian waters, most studies have dealt specifically with a lim- ited number of major elements and trace met- als. Therefore, to supplement the information already published, this paper was designed to: (1) provide a more extensive geochemical analysis of Amazonian surface waters, involv- ing 50 elements; and (2) to relate the chemical composition of the various rivers in the Ama- zon Basin to the geochemistry of the soils in the three geochemical provinces discussed above.
2. Study area
Studies were conducted in both the Manaus area, in the state of Amazonia, Brazil (Fig. 2a), and the Caraj~ls mining area in the state of Pardi, Brazil (Fig. 2b). The sampling area in Manaus provided the opportunity to sample the water, soils and sediment of both the varzea (which serves as a projection of the western peripheral region) and the Central Amazon lowlands, while Caraj~ls allowed sampling to be under- taken in the southern peripheral area (Pre- cambrian Shield).
The city of Manaus is built at the confluence of the Rio Solim6es and the Rio Negro in cen- tral Amazonia, 1600 km from the Atlantic coast. The Rio Solim6es, also known as the Upper Amazon, is a typical whitewater river that drains the Andes Mountains and the west- ern peripheral region of the Amazon Basin. The Rio Negro, volumetrically the largest tributary of the Amazon, is a blackwater river that drains the Central Amazon. The frequently inun- dated area between the two Amazon tributar- ies is partly influenced by both water types. 11 consists of the varzea forest that borders the Rio Solim6es and a mixture of caatinga forest., and the less dense igapo forest, which is an. nually flooded by the Rio Negro (Hedges et al. 1986).
The Caraj~ts Range is located within the ten.
15 8 K.O. KONHAUSER ET AL.
~ " ) \ MA~ 12 /11£
N
. . . . . ,,o
%-N t~] 0 es ed " ~ I ~ deforested
~l ~ , o :8 ~ ~ - - ~ - ~ = - ~ t sample site
Fig. 2. Location of study areas and sample sites in: (a) Manaus; and (b) Carajts.
tral Brazilian Shield, ~ 550 km south of Bel6m in the state of Pardi. The area is almost com- pletely covered by an equatorial rainforest, broken only by clearings of sparse vegetation. Most of the Caraj~is area is served by the Ita- caiunas River which flows into the Tocantins River at Marab~i. The main tributary of the Itacaiunas River, the Paraopebas River, marks both the eastern boundary of the Caraj~ls area and an abrupt transition from a forested to de- forested landscape. Recently, large areas of
Pani have been deforested as raw material for charcoal to be used in smelting pig iron (Fearnside, 1989).
3. Sampling and analytical methodology
Water samples were collected upstream of Manaus on the Rio Negro (and some of its tri- butaries) and on the Rio SolimGes on two con- secutive sampling periods in mid-July 1990 and late May-early June 199 l, following the
MULTI-ELEMENT CHEMISTRY OF SOME AMAZONIAN WATERS AND SOILS 15 9
annual rainy season in the Amazon. Rainfall was collected on both May 30 and 3 l, 1991. Soil and sediment samples were also collected from several locations along both rivers. Soil samples M8, MI1, S4A, S4B and $5 were col- lected within 50 m of the river bank from the upper 10 cm of the soil profile, while samples M T, MI3A and N2B were collected randomly at greater depths along the soil profiles ex- posed on the embankments of the Rio Negro. All stream sediment samples (except sample MI4) were collected from the river bottom (near-shore) by diving to depths of ~ 5 m be- low the water surface. Sample MI4 was col- lected at a greater depth ( ~ 15 m) by a dredg- ing device operating during our sampling period.
In Caraj~is, sampling occurred during three consecutive summers in late July 1989, middle July 1990 and late June 199 I. Water and sedi- ment samples were taken from several small streams draining completely forested regions, while soil samples were collected throughout the area. Soil samples K3C and K7A were also collected from the upper l0 cm of the soil pro- file, in close proximity to the river embank- ments, while both samples C5D and P0 were taken several kilometers from any water systems.
Approximately 1-1 samples of water were collected for multi-element analysis. The water samples were placed in two 500-ml Nalgene ® bottles.and then acidified on site with 5 ml an- alytical grade HNO3. Rainfall was collected with 25-cm-diameter plastic funnels into 500- ml Nalgene ® bottles and treated as above. Prior to multi-element analysis, most particulate materials in suspension were removed in the laboratory by pressure filtration through 0.2- #m Nuclepore ® membranes (Costar Corp.). To determine the concentration of metals in the "dissolved" size fraction (the size fraction typically distinguished as being smaller than 0.45 #m, regardless of whether the metals are in a soluble form, a colloidal form, or in an or- gano-metallic complex; Salomons and FiSr-
stner, 1984), the water samples were then ana- lyzed by inductively coupled plasma-mass spectroscopy: an ICPMS Model PQI by Ele- mental Ltd. (U.K.) was used. As a check on accuracy, standards ERI CNI and ERI CN2 were run and found to be in agreement of +_ 20%, while a check on precision, through the use of duplicates, suggests good reproducibil- ity (typically within +_ 15%). In addition, Ele- mental Research Inc. (Vancouver, Canada) ran standard methods of analysis which in- cluded the use of blanks.
All bulk soil and sediment samples were analyzed for major and minor elements by X- ray fluorescence spectroscopy (XRF) on a Philips ® PW-1450 automatic sequential spec- trometer, in the Department of Geology, Uni- versity of Western Ontario. The percentage of major and minor elements were determined from the preparation of fusion disks, while the concentration of trace elements were deter- mined using pressed powder pellets (Wu, 1986). Similar to the water samples, common reference standards for major oxides (SY-2 and MRG-1 ) and trace elements (G-2 and BHVO- 1 ) were run, in addition to several duplicate samples. The results demonstrate that the XRF was both extremely accurate and precise, with errors within _ 15% and +_ 5%, respectively.
To analyze the different varzea soil size frac- tions, unmilled samples were disaggregated with an agate mortar and pestle, and dispersed in distilled water using an ultrasonic probe. The resulting supernatant was decanted into a set- tling column and the residual > 20-#m frac- tion was removed. The 2-20- and < 2-#m size fractions were removed by settling accordin~ to Stokes law, and subsequently dried in a pre. heated oven. All size fractions were analyzec by XRF.
Mineralogies for all soil and sediment sam. ples were determined by X-ray diffractiot (XRD) on a Rigaku ® rotating anode diffrao tometer (Co-K,). The scans were carried ou at 160 kV and 45 mA, from 2-82 ° (20) at rate of l0 ° rain -1. Sample preparation in
Crus ta l a b u n d a n c e s are f rom Fairbr idge ( 1972).
TABLE 3
Minera logy o f soils and sed imen t s in s tudy areas
Average soil Average sed imen t
Centra l Shield varzea Central Shield A m a z o n A m a z o n
varzea
Quar tz m m m m m m Kaol ini te m m m n m n m m n Gibbs i te m n m n Goeth i te m m n Hemat i t e m n m n m n t m n m n Anatase m n m n m n t t m n Plagioclase m m K-fe ldspar m n m n Micas t m n t m n Illite t m n t* t m n Chlor i te t t Smect i te t m n t* t m n Zircon t t t t Tourma l ine t Ruti le t
m = ma jo r const i tuent ; m n = m i n o r const i tuent ; t = trace const i tuent . " t*" indicates s ed imen t sample M 14 only.
volved packing the milled soils and sediments into AI sample holders, so as to retain a ran- dom orientation. From the < 2-/~m size frac- tions, ~ 2-ml subsamples were pipetted to a glass slide and analyzed by X R D under the fol-
lowing conditions: untreated, 54% relative hu- midity and glycol solvated.
In conjunction with the X R D work, thir sections of several soil samples were made anc studied with a transmitting light microscope
162 K.O. KONHAUSER ET AL.
This proved useful in determining the pres- ence of minerals found in only trace amounts.
4. Soil and sediment chemistry and mineralogy
Soils in the Amazon Basin are collectively some of the poorest soils in the world. This nu- trient deficiency is typified by the soils in the Central Amazon. These highly leached soils have had most of their base cations reduced to very low levels with some samples having lev- els below those detectable by XRF [Table 1, (a) ]. In these soils, > 95% of the macrochem- ical components are accounted for by a SiO2- A12Oa-Fe203 assemblage. Compared to crus- tal abundances (Table 2 ), these soils are more siliceous and aluminous, with considerably lower levels of major cations. At this advanced stage of weathering the soils have a rather sim- ple mineralogy dominated by quartz, Fe + Al- oxides and kaolinite, with a few refractory minerals such as anatase and zircon (Table 3 ).
Soils from the Shield areas are also highly leached, with the major-cation levels showing only slight enrichments over Central Amazon soils [Table 1, (b) ]. This enrichment is prob- ably due to the presence of trace amounts of illite and smectite, in addition to quartz, ka- olinite and Fe + A1 + Ti-oxides. These soils also show an extreme variability in their SiO2- A1203-Fe203 ratios, with some samples being very siliceous (KTA) and others very alumi- nous and ferruginous (CSD). Collectively soils from the Shield are both more Fe-rich and A1- rich than crustal abundance and other Ama- zonian soils.
The only nutrient-rich soils in the Amazon are those of the varzea. Compared to the Cen- tral Amazon and the Shield, these soils show the highest cation to silica ratios [Table 1, (c)] , with the concentration of Ti, Mn, Mg, Ca, K, P and Na most closely approaching crustal abundance. This in turn is reflected by a more diverse mineralogy. The most noticea- ble difference is the presence of primary min- erals such as feldspars and micas and a high
proportion of diverse clay minerals ( ~ 11% of sample) such as kaolinite, illite, chlorite and smectite.
The geochemistry of the near-shore river sediment samples from all three geochemical provinces closely mirror those of the soils (Ta- ble 1 ). In the Central Amazon, the sediment is extremely siliceous, dominated by quartz and kaolinite, with trace amounts of hematite and anatase. Sample M14, collected from the mid- dle reaches of an off-branch of the Rio Negro, contains a greater diversity of clays, including trace amounts of illite and smectite. This sam- ple further shows a high LOI (loss on ignition) value, expected due to the high concentration of particulate organic material in suspension (Hedges et al., 1986). Rivers draining the crystalline Shield, in the Caraj~is region, have sediment comprised of quartz, kaolinite, and trace micas and illite. These sediments are typ- ically more siliceous and less ferruginous than their soils. Finally, the presence of feldspars and cation-rich clays (smectite, illite and chlo- rite) in the Rio Solimfes sediment reflect drainage from the Andes Mountains and the varzea soils. The considerable load of enriched detritus in this whitewater river leads to the fertile sediments of the varzea, which contain the highest concentrations of major cations. A comparison of the soil and sediment samples shows similarities in both composition and mineralogy, indicating that the floodplains are formed almost exclusively by river sediment (Victoria et al., 1989 ).
The variability between the soils of all three geochemical provinces is also shown by their trace-element composition (Table 4). The in- fertile soils of the Central Amazon typically have the lowest abundances of trace elements with only Nb and Zr having greater concentra- tions than the other soils (Table 5 ). In addi- tion, only refractory metals such as Nb, Zr, Pb, Th and U (all commonly found in zircons) are enriched relative to crustal abundance (Table 5).
In soils of the Shield, several elements are
MULTI-ELEMENT CHEMISTRY OF SOME AMAZONIAN WATERS AND SOILS 163
enriched relative to both the other soils (e.g., Cu, Co, Cr, V, As, Ga, U and Th) and crustal abundances (e.g., Nb, Zr, Y, Pb, Cu, V, As, Ga, U and Th). These enrichments clearly reflect the mineral wealth (e.g., bauxite, copper, gold,
iron ore, manganese, nickel and tin deposits) discovered and exploited in the Caraj~is area (Shaw, 1990). The soils also show the highest levels of organic debris present [Table 1, (b) ], suggesting that some elements may be exhibit-
stantial enrichments (relative to the crust) of several elements, including Nb, Zr, Y, Rb, Pb, Zn, Ba, As, Ga, U and Th. The high trace-ele- ment concentrations suggest that elemental enrichments may be associated with the pres- ence of clay minerals (Kronberg et al., 1979; Larocque, 1989). Where primary minerals are present, the levels of most inorganic species in the soil solutions are sufficiently high for smectite [ e.g., montmorillonite ((Ca,Na) 0.7- (A1,Mg,Fe)4(Si,A1)8 O2o (OH)4"nH20) and illite ((K,H20)A1ESi3A1Olo(OH)2) to form (Fyfe et al., 1983 ) ]. In these complex clays al- most any transition metal can substitute for A1, Mg and Fe, while halogens can be fixed in the hydroxyl sites (Fyfe et al., 1983). Therefore, as long as these clays are present with abun- dant unweathered minerals, many trace metals will be concentrated in the weathering process (Kronberg et al., 1979). This is shown in the < 2-~tm size fraction of the varzea soils which were shown to concentrate significant quan- tities of trace metals, including Y, Rb, Pb, Zn, Cu, Ni, Co, Cr, Ba, V, As and Ga (Table 6 ).
Once the primary minerals are eroded the smectite clays will be degraded to the more
simple kaolin clays (A12Si205 (OH)4) with a corresponding decrease in major elements and trace metals (Fyfe et al., 1983). At this ad- vanced stage of weathering, the cation-ex- change capacity is minimal and enrichments in the soils will generally be limited to the ele- ments incorporated in refractory minerals such as anatase (Ti) and zircon (Zr). This situa- tion is also consistent with our findings in the Central Amazon where kaolinite dominates as the stable clay phase, and the trace-metal con- tents are typically depleted relative to the soils from the varzea and crustal abundances.
5. Surface water chemistry
The concentration of 50 elements in river water ("dissolved" size fraction) are given for several sample sites in the Rio Negro, the Rio Solim~es and Shield streams in Caraj~ls [Table 7, (a), (b) and (c), respectively]. The 1991 values for Cu, Ni, Zn and Pb were omitted due to contamination from the acid bottle during the sampling procedure.
These analyses indicate that there is sub- stantial variability among individual sampling sites in all three areas. This is shown in the large standard deviation values shown in Table 7. The greatest variability is shown in the waters from Caraj~ts, where the standard deviation often exceeds the average concentration (e.g., B, P, Cu, Br, Mo, Nd). These high standard deviations may partially reflect the differences in location and time of sampling, as well as the low concentrations (ppb) that we are dealing with. The least variability is exhibited by the Rio Negro samples.
If the chemical composition of Amazonian rivers are primarily determined by the geolog- ical source and erosional regime, then varia- tions should be observed in water chemistry. This shown in Fig. 3 which compares surface waters from all three regions. The largest vari- ance exists between the whitewaters of the Rio SolimSes and the blackwaters of the Rio Ne- gro. With the exception of Zn and Se, all other
166 K.O. KONHAUSER ET AL.
r-- "~
<d
7~
o
0 0 0 0 0 ~ 0 0 0 d d d ~ ~ d d d d ~ d d d ~ d ~ d d d d d d d d d d d ~ d d d d d d d d d d d d d d d d d
~ o ~ o ~ ~ ~ o ~ ~ d d d ~ 4 o g m - - o d ~ d ~ d d d 4 d N d 4 m d d d d d d o d e o o o o d d d d d d d d d d d d d v V ~ ~ e
~ V ~ m V V ~ V ~ V V V V V V V V V V V V V V V V V V V V V
~ 0 0 0 ~ 0 0 0 ~ 0 ~ 0 ~ 0 0 ~ 0 0 ~ d d 4 d d " 4 d d d d v v = ~ ~ V ~ V
d N 4 d d I I I d 4 d 4 m d d d d d o 4 d ~ d 4 d d d d d d d d o d d I d d d m V V V V V V V V V
~ 0 0 0 0 0 0 ~ 0 0 ~
d d d ~ d o 4 d d d d d d ~ d d i V V ~ ~ V m ~ m
o ~ ~ o o ~ ° ~ ° ~ ° m ~ ~ ~ ~ o ~ o ~ o o o o o o o o o ~ ~ o i ~ d d d 4 ~ o o d d d d d d ~ d 4 d o d d d d d d d d d ~ d o d
V V V V V V V V V V V V V V
d 4 ~ 4 d d d d d d d ~ d ~ d d d o d d d d d d o d d i d d d o o ~ o o ~ o o o o o ~ o i i i ~ o ~ V ~ ~ V m V V V V V V V V V V
~ 0 0 0 0 0 0 0 0 ~ ~ 0 ~
~ d d d d d ~ d d 4 d ~ d d d d d ~ d 4 d d # d d d d d d ° ° ~ ° ° ° ° d d d ° d d d d d ° d d d ~ o o m V ~ ~ V - - ~ V V V V V V V V V V V V V V V V V V V
o o o o o o ~ o o - ~ q ~ o o o o o o o o ~ o o ~ o o o o ~ o o ~ d d ~ d o d d d ~ d 4 d d ~ d d d d d d o o ~ d o o o d d d d d d d d d o d d d
V ~ _ ~ ~ V V ~ V V V V V V V V V V V V V V V V V V V V
0 0 0 0 0 0 0 0 ~ 0 0 ~ o ooo o ooo o ooooo_ ooooooooooo ooooo ~ o ~ _ ~ o ~ V ~ V V ~ V V V ~ V V V V V V V V V V V
ooooooo _oo .oo o_oo - o o q o o o o ~ o o q ~ ~ d d d d ~ d d 4 d d d ~ d d d d d " 4 d 4 N d d o d £ d £ d d d d d d d o d d d d d d d o d
~ - - ~ ~ V V V ~ V V V V V V V V V V V V V V V V V
~ - - o o o o o o o o ~ - - ~ o ~ o o - - o o ~ 0 0 0 0 0 0 0 ~ 0 ~
d d d d d d ~ d d ~ d d ~ ° d d ° m ~ d : ~ d d d d d d d d ~ d d d d ~ d d d d d d d d d d d ~ ~ o ~ V V ~ ~ V V ~ V V ~ V V V V V V V V V V V
MULTI-ELEMENT CHEMISTRY OF SOME AMAZONIAN WATERS AND SOILS 167
I,z,1 o .~h2
< Z
ii d d d d 4 d ~ 4 ~ d 4 4 d ~ d d d d d d ~ d 6 d d d d ~ d d d ~ d 4 d d d d g d d d d d d d d d d a a e n ~
. d d d 4 d ~ d N 4 d d d d o d d d d d o d d o V V V ~ V V v V V V V V V V V V V V V V V ~ o ~ ~
V
~ 0 0 0 0 0 0 0 ~ 0 0 0 0 0 0
d d ~ d d d ~ d d d 4 ~ 4 ~ d 4 - - ~ V V V V V V V
~ 0 0 0 0 0 0 0 ~ 0 0 0 0 0 0
d d ~ d d d 4 d d d ~ 4 ~ d ~ V V ~ 0 ~
~ o o ~ o o ~ o o ~ o d d d 4 d d d d N d d C d Z d N 4 d d d d o d d o d d d o o
V V V V
~ o ~ ~ N ° ° ° ° ° 0 0 0 0 0 ~ 0 0 0 d d ~ d d d ~ d d d 4 ~ 4 ~ d 4 V V ~ ~ o
d d ~ d d d ~ d q d ~ C d d V V ~ o ~ ~ o
o o n n ~ o o d d d d C d ~ 4 ~ 4 d d d d o d d o d d o o o - - ~ V V V V V V V
° ~ ° ~ ° ~ ° ~ o ~ 0 0 o ~ ~ o ~ o o o o ~ ~ ~
4 d ~ 4 d d d d d d C d ~ 4 N 4 d d d d o d d o d d d o d V ~ ~ V V V V V V V V V
o ~ o o o o o o o o o o o o o o ~ o ~ O O o o o o ~ o ~ o o ~
~ d 4 ~ d d d d d d 4 d ~ N ~ M d d d d d d d o d d d o o V V V V
d d ~ d d d ~ d d d Z ~ 6 d d o o o ~ o o ~ o d d d ~ o o d d d d d d ~ J 4 4 d d d d d d d o d d d o o V ~ ~ o V VV VV VV
I
I ~°°°°°°°°°°-°°~°°°~- ooooo~ ~ - ~ - i d d ~ d d d ~ d d ~ d d ~ o o C ~ d d d ~ ' " d d d ~ d 4 d d d d d o o o o ~ o - - o o o ~ o o V ~ V ~ o o o o d d d d d d d d o o
V ~ o ~ V V V V V V V V V V
n
~ o ~ q q o o o q ~ m ~ ~ o ~ o - o o ~ m ~ 0 0 0 0 O 0 d d o d d d d d d d d d d d o d ~ o - - d ~ d O ~ d o ~ o ~ o o d 4 d d d 4 Z d ~ d 4 ~ ° ° d ° d d d d ~ d d d o o o o ~
V ~ ~ o ~ V V ~ V V V V V V V V V V V V V V
I
168 K.O. KONHAUSER ET AL.
d
d d d d ~ ~ d ~ d d ~ d ~ d d ~ d ~ d d d d ~ d d d ~ d d d d d d d d d d d ~ d d d
~ o o o o ~ o ~ ~ o ~ ~ - o o - o o o o o ~ 4 d d ~ d ~ d ~ 6 ~ 4 4 6 "d--dN~ned456ddmdbddddddddddddd~dd6 V V ° ~ ~ m V ~ V ~ ~ V ~ V V V V V V V V V - - V ~ V V V V V V V V V V ~ V V V
~dddddddd4dd~d ~ d d ~ - - ~ d d d d d o - - d ~ d d d d d d d o d d d d d d ~ V ~ o ~ ° ~ ° o o V V ~ ~ - - V V V V V V V V V V V V V V ~ V
o ~ O O o • oooo
o ~ o o o - - ~ v ~ ~ V ~ V V V V -- V V V V V V V V V V ~ V
o ~ o o o ~ o ~ o ~ o o o ~ o - ~ "d~ddd~dd4~d~dd4~dd~4~d.dod~d~o~ddddddddodd~dd~ ~ o ~ o ~ v v-- ~ v~v v v v- v v v ~ v v v v v v v v v- v 0 ~ V
~ 0 0 ~ 0 0 0 0 0 ~ ~ 0 O 0 ~dddd "ddd~d~gd6 "4d~44~d46d~d~d~doddddddoddm o d o o ~ o o ~ V ~ V V V V ~ V V - - V V V V V V V ~ V V ~ q v ~ q v
4ddddd~dd~ddN4dd~466~ " d ~ d ~ d d d d ~ d b d d d d d d d d d d o d d ~ d d d o o o ~ v ~ V V ~ n v ~ V v ~ v - ~ ~ V ~ V V V V ~ V V V V V V - - V V V
m ~ o o o o ~ o o o o ~ O O O - - q o ~ o ~ o ~ ~ ° ° ~
V ~ o m o ~ ~ V ~ ° ° ~ V V ~ V V V V V V V ~ V V V V V V m V
0 ~ 0 0 0 0 0 0 0 0 0 ~ 0 0 0 ~ 0 0 0 0 0 0 0 ~ ~ 0 0 0 0 ~ 0 0 0 0 ~ 0 0 0 ~ 0 0 ~ ~ 0 0 0 0 0 ~ O 0 0 - - q N O ~ ~ ~ ° ~ - o o 0 ~ ~000~0~ d d d ~ d d d d ~ d d M M d d 4 ~ £ M 4 M ~ d ~ d d d d ~ ~ 4 ~ 6 d d d d d o d d d d d d
~ o ~ o o o ~ ~ ~ V ~ V V V V V V V V V V ~ V ~ ~ V ~ V V ~ V
0 0 0 0 0 0 0 ~ 0 0 ~ 0 ~ ~ ~ 0 0 0 0 o o d d d d d d d o d d o • " d d d d d ~ d ~ o o d d
~ ~ V V m V V V V V V V V V V V V V V V V
o
o
8 o
V
,-k
"~, 0
~.~
N N
r~ i( I'
MULTI-ELEMENT CHEMISTRY OF SOME AMAZONIAN WATERS AND SOILS 169
1 0 0 , 0 0 0 -
~ " 1 0 , 0 0 0 0. .
c 1 , 0 0 0 -
Z 0 1 0 0 - I - -
r r 10 I--- z t.u 1 (.3 z 0 0.1 ¢..)
0 .01 C a K M g
1 O, 0 0 0 ,
Sr B a R b Li Os B e Si AI P Br B I Se As Sb
10,000
1,000
100
10-
1
0 .1 -
0.01
0.001 Fe Mn Ti Zn V Cu Cr Sc Co
lO0
/ 0.01
0.001 Sn Ni Pb Mo Cd Th Zr U Ag
Fig. 3. Comparison of the "dissolved" elemental concentrations in the surface waters of the Rio Solim0es, the Rio Negro, and Shield streams in Caraj~is, with world river average. World river average values are taken from Bowen ( 1979 ).
elements are consistently found in greater con- centrations in the former. These differences are commonly greater than an order of magnitude, and include Mg, Ca, Ti, V, As, Br and Sr. This pattern vividly reflects the differences in the metal concentrations in the surface soils.
Unlike the clear differences in chemical composition between the rivers draining the Andes Mountains and those draining the Cen- tral Amazonian lowlands, rivers in the Shield areas of Caraj~is are much more variable. While these streams display both the highest concen- trations of some elements (e.g., Mg, Li, Be, B, I, Se, Zn, Cu, Cr, Sc, Ni, Pb, Mo and Ag) and lowest concentrations of others (e.g., A1, As, Fe, V, Sn, Th and Zr) the overall chemistry suggests an intermediate composition.
Aside from differences in concentrations, the heterogeneity of Amazonian waters is also
confirmed by the relative proportions of some of the major elements (Ca, K, Mg and Si). Calcium is the dominant element in the Rio SolimSes and represents 52% of the sum of the major cations, followed by silicon at 32%, po- tassium 8.4% and magnesium 7.6% [Table 7, (a) ]. In the Rio Negro, silicon dominates at 65%, potassium 16%, calcium 14% and mag- nesium 5% [Table 7, (b)]. Similarly, in the Shield streams, silicon dominates at 51%, magnesium 18.4%, calcium 18.6% and potas- sium 12% [Table 7, (c) ]. The clear domi- nance of calcium suggests that the solute-rich waters of the Rio SolimSes could be classified as "carbonate waters" (Furch, 1984, p.186), and are indicative of the carbonate source (mostly limestones) in the Andes Mountains (Stallard and Edmond, 1983 ). The solute-de- ficient waters of the Rio Negro are character-
170 K.O. KONHAUSER ET AL.
ized by a completely different proportion of major elements. These waters show an ex- tremely high Si to cation ratio (Stallard and Edmond, 1983; Berner and Berner, 1987) which is indicative of the podsols in the Cen- tral Amazon, and a greater proportion of K to those of the alkaline-earth metals Ca and Mg, suggesting an alkali dominance (Furch, 19 8 4). Lastly, the Shield streams are once again rep-
120
tl" 0 ~ 100 ~ O
g z ~ g £~ ENRICHED w~ > J
MI., n l l n n n n i . . . . . . . . . .
20 - - - - -
E ~ 0 D E P L E T ~ ' n=U
03~ O O< ~69 ~- O~ornN u. 69 < (o~- Jfl3 ~ ~<o3
ao q o~ - b R t u n -
, ~ E i . o o
, , , w I - > o ENRICHED
iit O~-i~ ~ ~ , r - D ~' -20 ~ DEPLETED
69 O) F ' - E : O ~ C [30 I - k N ( - ) Z ' : ~ ~ 60 < O < N 69 13. ~ O
o3~-~ O t) o3 fl~ u- < ~7 O<r- Z~{) 69 690 N~ N ~_<
Fig. 4. Chemical composition of: (a) the Rio Negro; (b) the Rio Solim6es; and (c) Shield streams in Caraj~s, nor- malized to world river average, indicating an enriched or depleted status for each element in the river systems.
Fig. 5. Comparison of the chemical composition of rain- fall collected over Manaus (May 30, 1991 ) normalized to: (a) the Rio Negro; and (b) the Rio Solim~es, indicat- ing an enriched or depleted status of each element in the rainfall, relative to each stream.
Fig. 6. Chemical composition of the rainfall normalized to world river average, indicating an enriched or depleted status for each element in the rainfall.
resented by an intermediate composition. To ascertain the individual status of each
element, the average chemical composition of the Rio Negro [Table 7, (a ) ] , the Rio Soli- mSes [Table 7, (b)] and the Shield streams [Table 7, (c) ] were compared to world river
MULTI-ELEMENT CHEMISTRY OF SOME AMAZONIAN WATERS AND SOILS
TABLE 8
Chemical analysis of rainfall ( in ~tg 1-1 )
171
Manaus rainfall
May 30, 1991 May 31, 1991
Manaus rainfall
May 30, 1991 May 31, 1991
Li <0.38 0.75 Be < 0.07 < 0.07 B 3.10 4.70 Mg 910.00 480.00 AI 2,300.00 160.00 Si 3,000.00 1,600.00 P 200.00 23.00 K 12,000.00 6,200.00 Ca 5,200.00 8,000.00 Sc 0.51 0.34 Ti 5.80 3.30 V 120.00 7.90 Cr 5.60 3.60 Mn 92.00 6.60 Fe 1,700.00 130.00 Co 1.20 0.67 As <0.02 0.13 Se <0.26 0.70 Br < 0.56 9.60 Rb 35.00 11.00 Sr 30.00 51.00 Zr 0.36 0.39 Mo < 0.09 2.60
Ag 0.05 0.02 Cd 1.70 0.24 Sn 13.00 8.00 Sb 0.39 0.33 I 0.56 2.00 Cs 0.44 0.14 Ba 50.00 16.00 La 11.00 0.32 Ce 9.30 0.26 Nd 5.80 0.23 Sm 0.55 0.10 Eu 0.04 < 0.02 Tb 0.10 < 0.02 Dy 0.33 < 0.03 Yb 0.16 <0.02 Hf < 0.02 0.03 Ta < 0.02 < 0.02 W 0.05 0.20 Au < 0.02 < 0.02 Hg <0.09 0.12 Bi 0.11 0.05 Th 0.20 0.15 U 0.13 0.08
average (Fig. 4a, b and c, respectively). In these graphs, world river average (taken from Bowen, 1979) has been normalized to 1, such that each element in the studied rivers was de- termined to be either enriched or depleted. From the 30 elements compared, the Rio Soli- mres is shown to have the greatest number of elements with a similar to enriched status (16). These include Sn, Th, Mn, V, Co, Cd, Se, Ba, Fe, A1, Rb, P, Cr, As, Ti and Ni, with Sn being the most enriched element at 82 times the world river average. Equally, the Shield streams (collectively) have 16 enriched elements, in- cluding Mo, Se, Cr, Sn, B, Pb, Zn, Mn, Ba, Ni, Rb, Co, I, Th, Cd and Cu. Lastly, the Rio Ne- gro is shown to have only 6 enriched elements (e.g., Sn, Se, Th, Rb, Cd and Mn). These re- suits confirm that the Rio Solim~es is the most solute-rich river studied, while the Rio Negro
is extremely solute-deficient when compared on a global basis.
6. Chemical analysis of rainfall
Chemical analyses of rainfall collected over Manaus (May 30 and 31, 1991 ) are given in Table 8. A comparison of the rainfall collected between the two consecutive days indicates an extreme difference in chemical composition. Rainfall from May 31, 1991 is consistently de- pleted in dissolved solutes, suggesting that: (a) many of the metals may have been washed out in the previous storm, or (b) the second day storm was greater in intensity, and therefore, a dilution effect may have taken place.
A comparison of the dissolved metal con- centration in the rainfall (May 30, 1991 ) with the concentration of dissolved metals in the Rio
172 K.O. KONHAUSER ET AL.
Negro (normalized to give individual ele- ments in the rainfall either an enriched or de- pleted status relative to the river) illustrates that all metals (with the exception of Br, I, Mo, As and Se) were found in higher concentra- tions in the precipitation than in the fiver: sev- eral metals were precipitated at > 10 times their concentration in the surface waters, in- cluding Sn, Rb, P, K, Cd, Ca and A1, while V was precipitated at > 100 times (Fig. 5a). Since the Rio SolimSes (also normalized) is inherently more solute-rich than the Rio Ne- gro, the same rainfall shows a less pronounced effect on the whitewater river, where only V, Sn and Rb were found in the rainfall at con- centrations 10 times greater than in the surface waters (Fig. 5b).
A further comparison between the chemical composition of the rainfall and world river av- erage (normalized to give the rainfall an en- riched or depleted status) also indicates that several metals (e.g., Sn, V, Rb, Cd, Mn, P, A1, Th, Co, Cr, K, Ba, Fe, Sb, Se and Ti) are en- riched in the rainfall (Fig. 6). Although only two rainfall samples were analyzed, precipita- tion is clearly shown to be an additional source of dissolved metals to Amazonian river sys- tems. This may, in part, account for the high concentration of Sn, Se, Th, Rb, Cd and Mn in the Rio Negro and Sn, Th, Mn, V, Co, Cd, Se, Ba, Fe, A1, Rb, P, Cr and Ti in the Rio Soli- mres, relative to world river average (Fig. 4).
7. Conclusions
Similarities between the major-cation and trace-metal levels in the soil samples, and the dissolved metal concentrations in surface waters flowing through the varzea, the Central Amazon and the Shield, suggest that the rivers are in chemical equilibrium with their drain- age basins. In geologically active areas, such as the Andes Mountains, crustal elements such as Si, A1, K, Na, Ca, Be, Li, Zn, Mo, F and B are mixed with newly extruded mantle elements such as Ca, Mg, Cu, Co, Ni, S, V, Cr and Se
(Leonardos et al., 1987). With high rates of erosion of the varied lithologies in the source regions, enormous quantities of unweathered minerals (feldspar and mica), metal-rich clays (smectite, chlorite and illite) and dissolved metals are transported downstream by white- water rivers, resulting in the fertile soils of the varzea which are almost exclusively derived from the sedimentation of the suspended load. The chemistry of the dissolved size fraction in the Rio Solimres, therefore, reflects the sub- strate lithology in the Andes, with the high concentration of calcium, for example, indic- ative of a limestone source (Stallard and Ed- mond, 1983 ), while continuous leaching of the varzea soils further supplies the solute-rich river with additional metals.
The high concentration of dissolved metals in the Rio SolimSes is contrasted by the solute- deficient waters of the Rio Negro which drains the highly weathered lateritic and podsolitic terrains of the Central Amazon. Due to a lack of exposed rock, and the low rates of weather- ing in conjunction with the development of thick, siliceous and aluminous soils, the sus- pended sediment are typically cation-de- pleted, consisting almost entirely of quartz and kaolinite (Stallard, 1988 ), while the dissolved load is dominated by silicon, with extremely low levels of major cations and trace metals.
Rivers that drain the Precambrian Shield typically carry a limited suspended and dis- solved load (Gibbs, 1967), reflecting the tec- tonic stability which leads to low erosional rates and high leaching (Leonardos et al., 1987). Furthermore, in these stable cratonic regions, which are favored by sedimentary rocks such as orthoquartzites, with no nutrients, or gran- itic and metamorphic rocks, with a limited nu- trient spectrum and low erosional rate (Stal- lard, 1988), the soils are typically depleted in nutrients (Fyfe et al., 1983). Although these conditions should give rise to extremely sol- ute-deficient rivers, the waters flowing through Carajds have an intermediate composition. This discrepancy may largely reflect the fact
MULTI-ELEMENT CHEMISTRY OF SOME AMAZONIAN WATERS AND SOILS 17 3
that Caraj~ls is a very metallogenic area, rich in mineral deposits, which upon weathering could supply dissolved metals to local river systems.
These observations, collectively, suggest that we should be able to classify the chemical com- position of rivers according to the geochemis- try and mineralogy of the soils through which they flow. While previous work has demon- strated a relationship between geology/soil geochemistry and the chemical composition of Amazonian rivers based on a few cations (Stallard and Edmond, 1983) and trace met- als (Furch, 1984), our research extends this relationship for 50 elements.
These findings have profound implications for using water chemistry as an indicator of the agricultural and mineral potential of a region. A comparison of three key macronutrients and the total dissolved inorganic solids (TDS) of several great rivers of the world speak elo- quently about the state of the soils (Table 9). The Danube, Mississippi, Yangtze and Nile river systems all have high dissolved metal concentrations, reflecting the young, fertile soils through which they flow. Since plants re- quire a wide range of macronutrients and mi- cronutrients, their solute-rich waters indicate a high capacity to support bioproductivity. In contrast, the extreme chemical poverty exhib- ited by the Rio Negro is indicative of the poor nutrient status of the Central Amazon. The in- ability of the lateritic soils to support agricul- tural activity (Leonardos et al., 1987 ) is prob- ably the reason why historically this region has never been overpopulated by humans (Fyfe,
1989). Although the fertility of a soil depends on more than just chemical criteria (i.e. poros- ity, permeability, etc. ), these other factors be- come of secondary importance when a soil is almost completely deficient in nutrients. Therefore, in regions such as the Central Am- azon, chemical analysis of the river systems should reflect the incapability of the soils to sustain long-term crop production. Clearly then, using river chemistry as an agricultural indicator would be a very efficient way of as- suring soil potential prior to massive develop- ment projects.
Knowing the chemical composition of a river system may also be of use in determining the mineral potential of a region. The high concen- tration of a number of metals (such as Zn, Cu, Cr, Ni, Pb, Mo, Co, Mn, Rb, Ba, B, I and Se) in the waters draining the Shield areas in Ca- raj~is (relative to both Amazonian rivers and world river average) confirms the presence of one of the greatest mining areas of the modem world. While the high aqueous concentration of some of these metals may be associated with local mining activity, many of these rivers drain completely forested areas with no an- thropogenic influence. Under these condi- tions, river chemistry may be useful in mineral exploration.
Results from this study also indicate that precipitation is a significant source of dis- solved metals (e.g., Sn, V and Rb) to solute- deficient river systems such as the Rio Negro, with a less pronounced effect on the other Amazonian rivers. This corresponds to earlier
TABLE9
Comparison of dissolved solutes in principal world rivers (in #g 1- ~ )
World river values from Berner and Berner ( 1987). TDS = total dissolved solids.
174 K.O. KONHAUSER ET AL.
work by Gibbs (1970) who observed that in many low-salinity rivers of the Amazon, the chemical composition was determined pri- marily by the dissolved salts supplied by pre- cipitation. The marine component of the pre- cipitation was shown to consist of Na, K, Mg and Ca in approximately sea-salt proportions (Gibbs, 1970; Stallard and Edmond, 1981), and this flux may partially account for the high concentrations of K, Mg and Ca found in the rainfalls over our sampling locations.
While the marine origin can clearly expli- cate the high concentration of a limited num- ber of dissolved metals in the rainfall, the question which needs to be addressed is, what is the source for the numerous other metals found in elevated concentrations in the rain- fall? In the Amazon Basin, it has been esti- mated that 52% of the rainfall in the region comes from water vapor from the Atlantic Ocean, while the remaining 48% is water va- por evapotranspired from the forest ecosystem (Marques et al., 1977). Although evapotran- spiration off the surface of the rivers or vege- tation would supply some metals to the atmo- sphere, this process would be limited in its efficiency in supplying particulate matter to the atmosphere. A more likely factor determining the nature of the rainfall is that as the water vapor moves across the continent, it accumu- lates aerosols (needed to serve as nuclei for cloud formation) from multiple sources, in- cluding soil dust, natural biological exudates and biomass burning (Stallard and Edmond, 1981 ): pollution in the Amazon Basin is neg- ligible due to the lack of industry. Initial stud- ies on the impact of biomass burning as a nu- trient source into Amazonian river systems, using strontium (87Sr/86Sr) and lead (2°6pb/ 2°4pb) isotopic ratios, suggests that these emis- sions may greatly enrich the rainfall in dis- solved metals (Konhauser, 1993). Unques- tionably more research into the significance of such anthropogenic activity in the Amazon Basin is required.
Acknowledgments
Supported by the Natural Sciences and En- gineering Research Council of Canada (NSERC). We thank Steve Hicock, Fred Longstaffe, Neil MacRae and Mike Powell of the Department of Geology, University of Western Ontario for reviewing the manuscript.
References
Berner, E.K. and Berner, R.A., 1987. The Global Water Cycle. Prentice-Hall, Englewood Cliffs, N.J., 397 pp.
Bowen, H.J.M., 1979. Environmental Chemistry of the Elements. Academic Press, London, 333 ppp. (see es- pecially pp. 237-272).
Boyle, E.A., 1978. Trace element geochemistry of the Amazon and its tributaries. Eos (Trans. Am. Geo- phys. Union), 59:276 (abstract).
Ertel, J.R., Hedges, J.I., Devol, A.H., Richey, J.E. and Ri- beiro, M.N.G., 1986. Dissolved humic substances of the Amazon River system. Limnol. Oceanogr., 31: 739- 754.
Fairbridge, R.W., 1972. The Encyclopedia of Geochem- istry and Environmental Sciences - - Encyclopedia of Earth Sciences Series, Vol. 1VA. Van Nostrand and Reinhold, New York, N.Y., 1321 pp.
Fearnside, P.M., 1989. Deforestation in Brazilian Ama- zonia: The rates and causes of forest destruction. Ecol- ogist, 19: 214-218.
Fittkau, E.J., Irmler, U., Junk, W.J., Reiss, F. and Schmidt, G.W., 1975. Productivity, biomass, and population dynamics in Amazonian water bodies. In: F.B. Golley and E. Medina (Editors), Tropical Ecological Systems - - Trends in Terrestrial and Aquatic Research. Sprin- ger, New York, N.Y., pp. 289-311.
Furch, K., 1984. Water chemistry of the Amazon Basin: The distribution of chemical elements among fresh- waters. In: H. Sioli (Editor), The Amazon, Limnology and Landscape Ecology of a Mighty Tropical River and Its Basin. Dr. W. Junk Publishers, Dordrecht, pp. 167- 200.
Furch, K., Junk, W.J. and Klinge, H., 1982. Unusual chemistry of natural waters from the Amazon region. Acta Cient. Venez., 33: 269-273.
Fyfe, W.S., 1989. Soil and global change. Episodes, 12: 249-254.
Fyfe, W.S., Kronberg, B.I., Leonardos, O.H. and Olorun- femi, N., 1983. Global tectonics and agriculture: A geochemical perspective. Agric., Ecosyst. Environ., 9: 383-399.
Gibbs, R.J., 1967. The geochemistry of the Amazon River System, Part 1. The factors that control the salinity and the composition and concentration of the suspended
MULTI-ELEMENT CHEMISTRY OF SOME AMAZONIAN WATERS AND SOILS 175
solids. Geol. Soc. Am. Bull., 78: 1203-1232. Gibbs, R.J., 1970. Mechanisms controlling world water
chemistry. Science, 170:1088- 1090. Gibbs, R.J., 1972. Water chemistry of the Amazon River.
Geochim. Cosmochim. Acta, 36: 1061-1066. Gibbs, R.J., 1973. Mechanisms of trace metal transport
in rivers. Science, 180:71-73. Gibbs, R.J., 1977. Transport phases of transition metals
in the Amazon and Yukon Rivers. Geol. Soc. Am. Bull., 88: 829-843.
Hedges, J.I., Clark, W.A., Quay, P.D., Richey, J.E., De- vol, A.H. and Santos, U.M., 1986. Compositions and fluxes of particulate organic material in the Amazon River. Limnol. Oceanogr., 3:717-738.
Katzer, F., 1897. Das Wasser des unteren Amazonas. Sit- zungsber. BiShm. Ges. Wiss.-Math. Naturw., 17: 1-38.
Klammer, G., 1984. The relief of the extra-Andean Ama- zon Basin. In: H. Sioli (Editor), The Amazon, Lim- nology and Landscape Ecology of a Mighty Tropical River and Its Basin. Dr. W. Junk Publishers, Dor- drecht, pp. 47-84.
Klinge, H., 1965. Podzol soils in the Amazon Basin. J. Soil Sci., 16: 95-103.
Klinge, H., Furch, K., Irmler, U. and Junk, W., 1981. Fundamental ecological parameters in Amazonia, in relation to the potential development of the region. In: R. Lal and E.W. Russell (Editors), Tropical Agricul- tural Hydrology. Wiley, New York, N.Y., pp. 19-36.
Konhauser, K.O., 1993. Multiple fluxes influencing Ama- zonian river chemistry. Ph.D. Thesis, University of Western Ontario, London, Ont., 150 pp.
Kronberg, B.I., Fyfe, W.S., Leonardos, O.H. and Santos, A.M., 1979. The chemistry of some Brazilian soils: ele- ment mobility during intense weathering. Chem. Geol., 24:211-229.
Larocque, A., 1989. Trace element behaviour during weathering of till from eastern Ontario, with implica- tions for geochemical exploration. M.Sc. Thesis, Uni- versity of Western Ontario, London, Ont., 201 pp.
Leenheer, J.A., 1980. Origin and nature of humic sub- stances in the waters of the Amazon River Basin. Acta Amazonica, 10: 513-526.
Leonardos, O.H., Fyfe, W.S. and Kronberg, B.I., 1987. The use of ground rocks in laterite systems: An improve- ment to the use of conventional soluble fertilizers? In: Y. Ogura (Guest-Editor), Proceedings of an Interna- tional Seminar on Laterite, October 14-17, 1985, To- kyo, Japan. Chem. Geol., 60:361-370 (special issue).
Marques, J., Santos, J.M., Villa Nova, N.A. and Salati, E., 1977. Precipitable water and water vapor flux between Bel6m and Manaus. Acta Amazonica, 7: 355-362.
Oltman, R.E., 1965. Some observations of Amazon River hydrology. S.C. Eng., 16: 6-12.
Putzer, H., 1984. The geological evolution of the Amazon Basin and its mineral resources. In: H. Sioli (Editor),
The Amazon, Limnology and Landscape Ecology of a Mighty Tropical River and Its Basin. Dr. W. Junk Publishers, Dordrecht, pp. 15-46.
Raimondi, A., 1884. Aguas Potables del Peril. F. Masias, Lima, 206 pp. (see especially pp. 127-134).
Salati, E. and Vose, P.B., 1984. Amazon Basin: A system in equilibrium. Science, 225:129-138.
Salomons, W. and Ftirstner, U., 1984. Metals in the Hy- drosphere. Springer, Berlin, 349 pp.
Shaw, A., 1990. CVRD's Caraj~is Project. Min. Mag., Au- gust iss., pp. 90-97.
Sioli, H., 1950. Das Wasser im Amazonasgebiet. Forsch. Fonschr., Amazonasgebiet, Arch. Hydrobiol., 43: 267- 283.
Sioli, H., 1967. Studies in Amazonian waters. Atos Simp. "Biota Amaz6nica", Rio de Janeiro (Limnologia), 3: 9-50.
Sioli, H., 1968. Hydrochemistry and geology in the Bra- zilian Amazon region. Amazoniana, 1: 267-277.
Sioli, H., 1975. Tropical rivers as expressions of their ter- restrial environments. In: F.B. Golley and E. Medina (Editors), Tropical Ecological Systems. Trends in Terrestrial and Aquatic Research. Springer, New York, N.Y., pp. 275-288.
Sioli, H., 1984. The Amazon and its main affluents: Hy- drography, morphology of the river courses, and river types. In: H. Sioli (Editor), The Amazon, Limnology and Landscape Ecology of a Mighty Tropical River and Its Basin. Dr. W. Junk Publishers, Dordrecht. pp. 127- 166.
Stallard, R.F., 1988. Weathering and erosion in the hu- mid tropics. In: A. Lerman and M. Meybeck (Edi- tors), Physical and Chemical Weathering in Geo- chemical Cycles. Kluwer, Dordrecht, pp. 225-246.
Stallard, R.F. and Edmond, J.M., 1981. Geochemistry of the Amazon, 1. Precipitation chemistry and the ma- rine contribution to the dissolved load at the time of peak discharge. J. Geophys. Res., 86: 9844-9858.
Stallard, R.F. and Edmond, J.M., 1983. Geochemistry of the Amazon, 2. The influence of geology and weather- ing environment on the dissolved load. J. Geophys. Res., 88: 9671-9688.
Taylor, S.R. and McLennan, S.M., 1985. The Continental Crust: Its Composition and Evolution. Blackwell, Ox- ford, 312 pp.
Victoria, R.L., MartineUi, L.A., Richey, J.E., Devol, A.H.: Forsberg, B.R. and Ribeiro, M.N.G., 1989. Spatial and temporal variations in soil chemistry on the Amazor floodplain. GeoJournal, 19: 45-52.
Wu, C.T., 1986. Operation manual for X-ray fluorescenc~ analysis. Department of Geology, University of West. ern Ontario, London, Ont., 8 pp.
