Time dependency of hydrogen Ion concentration and electric conductivity measurements in river water samples of central Amazonia

The hypothesis was tested, whether hydrogen — ion concentration and electric conductivity in river water samples of the Tertiary region of central Ama­ zonia were subject to changes in time and due to nature of storag. Over a period of about two months only changes in pH and conductivity of Rio Negro water samples were found to be significant. The total deviation of both parameters in rain forest stream samples v/as considered negligible for at least a period of 30 days, independant of the nature of storage (laboratory or icebox). The latter results may be generalized more or less unconditionally for the entire Tertiary formations of Amazonia.


INTRODUCTION
The hydrogen-ion concentration (pH) in natural waters is controlled by chemical re actions and equilibria among the ions in so lution.As far as the water samples studied were concerned, the pH was primarily deter mined by the chemical nature of the water sheds (substrates, soils, vegetation cover, etc), the chemical reactions under way in the river basins (decomposition -and remineralization processes, release of various acids (sulfuric, organic and carbonic), biological processes (photosynthesis, respiration, etc.) and the current (mixing capacity).
Although the hydrogen -ion concentra tion in river waters is not a main controlling factor in fresh water ecology (Reid, 1961), it should be recognized, that several water or ganisms have their specific pH tolerance ranges (Welch, 1952).
The electric conductivity is a measure to evaluate the total amount of ionized mate rial in water solutions.The total amount and the composition of dissolved solids in water control up to a certain extent, the variety and abundance of aquatic organisms.Although no exact relationship exists between electric conductivity and total dissolved solids in natural waters', the parameter (electricconductivity) is often used as a first approxi mation to determine the influence of total dissolved solids on the aquatic environment.
Hydrogen -ion concentration and electric conductivity in water samples are liable to changes during storage between sampling and laboratory measurements, owing to reactions to suspensions, hydrolyses, oxidation-reduc tion processes, loss of dissolved gases, etc.If samples are collected and stored with care, the early laboratory measurements are sup posed to approximate field conditions.II samples reach the laboratory after a two or three months, the reliability of the data is assumed to be dependant on the delay.This hypothesis was tested for several river water samples of central Amazonia.

MATERIAL AND METHODS
At June 6 th, 1971 four water samples were collected at Paricatuba, the Rio Negro nai rows, at about 15 Km upstream of the city of Manaus.Water was sampled in different depths and positions of the Rio Negro cross -section, as.; midstream (surface), mid stream (40m depth), 300m off right bank (20 m depth) and 300 m off left bank (20 m depth).
At August 8 th, 1971, water samples from four rain forest streams were collected along the BR -174, the Manaus -Caracarai Road at kilometers 61 35, 32 and 27.These streams drain watersheds of different sizes, but are undisturbed by human activity (slash and burn, cattle breeding, etc).The areas are covered with a terra firme rain forest, a climax forest type is closely related to heavy or extremely heavy yellow latosols.Along the river banks hydromorphic soil is common.Several small patches of tropical podzol soils occur scattered on the hill slopes.These soils are covered with "carrasco forests" or campinas.
The Rio Negro samples were stored in the laboratory.The BR -174 samples were divi ded.One set of samples remained under laboratory conditions, the other set of four samples was stored in an icebox.
The hydrogen -ion concentration was measured with a pH -meter, type 390 WTW,.Germany.The electric conductivity was read by means of a conductivity -meter, type LBR -WTW, Germany.Hydrogen -ion con centration and electric conductivity were repeatedly measured (average of 5 subsamples) for all three sets of samples concurrently up to September 9 th, 1971.
Changes in pH and conductivity during the time between sampling and first labora tory measurements occured, but these inter ferences were disregarded, because field ins truments showed an instrumental error of about 5-10 percent in dupplacated samples.
The Rio Negro samples (midstreamsurface) and all four BR -174 samples were analyzed for various constituents of ecological interest (see Table 1).
Sulphur compunds, organics and bacte rial and fungal activity were not evaluated, although they are of considerable importance for changes in pH and electric conductivity of the water samples.

I. The Rio Negro samples.
Hydrogen -ion concentration and elec tric conductivity data obtained, confinned the significant fluctuations of both parameters, depending on the sampling position in the river crosssection (see Table 2; Table 3).Con cerning chemical interactions in the solution, pH and conductivity data proved the lack of homogenity of the Rio Negro water body.Compared with other river systems in the world, the data are very low with respect to to both parameters.The deviation from the first laboratory reading over the period of measurements was signi ficant for pH (range : pH 0.3 -pH 0.7) and electric conductivity (range 0.7 uS2 0 -3.8 uS 2o ) respectively.While the pH data displayed no variation from the fust laboratory readings, the electric conductivity data, however, increased (samples 1,3 and 4) and decreased (sample 2) with time.

II. The BR -174 samples.
For all samples, hydrogen-ion concen tration and electric conductivity values were considerably low.The conducti vity readings of samples BR -174, Km 61, Km 32 and Km 27 were extremely low (see Table 2; Table 3); The devia tion in pH from the first laboratory measurements (range : pH 0.1 -pli 0.3) was insignificant for all samples and independent of storage (laboratory -icebox).Electric conductivity data showed a slight deviation from the first laboratory readings (range : 0.4 US20

11S20).
In sample BR-174 (Km 35) the con ductivity decreased with time.Some differences in total deviation from the first laboratory measurements were reported for BR-174 samples (Km 35; Km 32) as far as storage (laboratory -icebox) was concerned (see Table 3).

CONCLUSIONS
Hydrogen -ion concentration and elec tric conductivity data of all samples analyzed were extremely low, if compared with equiva lent river systems in the world.
The bicabornate -carbon dioxide system is effective in the control of the pH of natural waters in about the pH 4.5 to pH 8.0 range (Reid, 1961).Calcium and magnesium ion analyzed were below detection in all water samples (see Table 1), i.e. the solutions were poorly buffered (Brihkmann and Santos, 1971).
Hidrolyses, the most important type oi reaction controlling pH in natural waters must play a part in fixing the low pH as salt-: of iron and ammonium were present in the samples analyzed (see Table 1).These salts give an acid reaction upon the hydrogenion concentration.
As far as Rio Negro samples were con cerned, the socalled humic complex of these waters may have a control function on pH, but the reactions are no understood yet (see Table 1) .
According to Schmidt (1968) to tat carbon dioxide in Rio Negro waters was in about the 5 mg/1 range.If compared with free carbon dioxide data obtained for the BR -174 samples (see Table 1), the gas must have a profound effect on the pH of the rain forest stream waters.The high free carbon dioxide content of natural waters of the Tertiary regions of Amazon were also confirmed by other author?(Fittkau, 1964;Sioli, 1957).
As shown in Table 2, the fluctuations of hydrogen -ion concentration in Rio Negrc samples with time were significant, while the BR -174 samples were mainly stable, i.e. at least the storage of samples over a period of 30 days was feasible as far as pH was con cerned .
The electric conductivity data of the Ric Negro samples displayed significant variations with regard to sampling position in the river cross -section as well as in total deviation from the first laboratory measurements (sec Table 3).It must be understood, that sur face samples are not representative for the entire body of river water.As all water samples analyzed (Rio Negro, BR -174 sam ples) were extremely low in conductivity (ordinary single-distilled water used in the laboratory has a conductivity of 1 to 4 US20), only the Rio Negro samples (midstream sur face, 300 m off right and left bank, 20 in depth) and the BR-174 sample (Km 35) showed a significant change with time (10 -3U percent).As conductivity data were extre mely low, the storage of water samples is practical at least for a period of 30 days.No significant differences were found between samples stored either in the laboratory or the icebox.As similar conductivity data were reported for other Tertiary regions of Amazo nia (Fittkau, 1964;Sioli, 1957;Gessner, 1960), the results obtained may be generalized for the entire Tertiary formations of Amazonia.

1)
Granted by Banco Nacional de Desenvolvimento Econômico under contract INPA/FUNTEC 89/70.2) INPA is sponsored by the National Research Council of Brazil.