CHEMICAL COMPOSITION OF RIVERS AND LAKES 



G7 



Errors of this sort will be minimal in considering 

 the chemical discharge of large streams into the ocean, 

 for such streams do not display as marked fluctuations 

 as their small tributaries. These errors may be more 

 important in computing the salt discharge into basins 

 of internal drainage. 



A more serious error, and one that affects all the data 

 that are available for stream transport of mineral 

 substances, stems from general carelessness in dis- 

 criminating between dissolved and colloidal or sus- 

 pended material and a cavalier disregard for all the 

 mineral matter that does not meet the arbitrary 

 criteria of solution set up by a particular investigator. 



For purposes of practical industrial water chem- 

 istry the errors introduced by these habits of thought 

 and analytical procedure may not be important. 

 Obviously the bedload of a stream will have little 

 effect upon its suitability as a source of boiler water, 

 but in calculations of the role of streams in geochemical 

 cycles these errors are more serious. The few careful 

 investigations, such as Strakhov's (1948) work with 

 iron in natural waters, deal with a very small number 

 of elements only. 



The implicit assumption behind this practice of 

 ignoring all mineral matter that is not dissolved in 

 stream water seems to be that mineral matter is pres- 

 ent in only two states, as true solutions and as sus- 

 pensions, and that the suspended material consists 

 of unmodified rock, so that transport of the solid 

 material effects no chemical fractionation of the 

 earth's mantle. The transport of clay, silt, sand, and 

 gravel has been regarded as the sphere of the geo- 

 morphologist rather than of the geochemist. 



For material of large-grade size this may be almost 

 true. Gravel carried by streams is probably not 

 changed very much chemically from the parent rock 

 from which it came, although one would like to have 

 more definite evidence of this. One should know to 

 what extent it is unmodified primary rock, and to 

 what extent it is the residue from which the elements 

 being carried in true solution by rivers have been 

 removed. 



It is in considering the finer grades, however, that the 

 seriousness of ignoring suspended material becomes 

 apparent. Even such readily ionized and extremely 

 soluble elements as sodium and chlorine can be bound 

 in considerable quantity to fine mineral particles by 

 various sorption processes, of which the most important 

 is probably ion exchange. Data concerning the sorp- 

 tive capacities of suspended river solids are extremely 

 scarce. The study of Carritt and Goodgal (1954) on 

 Chesapeake Bay silts shows that the sorptive capacity 

 may be considerable, and further that it may be in- 

 fluenced by surrounding conditions, such as pH, in 



such a way that ions strongly sorbed to the river silt 

 when it is in fresh water may be released as soon as 

 the river water mixes with the sea. The opposite 

 transfer is also possible. A river silt may enter the 

 sea with its sorptive capacity at a very low level of 

 saturation, and may, on entry into the ocean, immedi- 

 ately pick up a large quantity of ionized material 

 from the sea water and precipitate it on the ocean 

 floor. This is a separate process from the chemical 

 precipitation of certain dissolved components which 

 has been recognized ever since it became apparent that 

 there were discrepancies between the chemistry of 

 the sea and that of the rivers that nourished it. 



The pernicious variability of filtration procedures al- 

 ready referred to must be considered. If a sample is 

 carefully filtered, the analysis should give a figure that 

 is representative of the dissolved material, using the 

 term "dissolved" to mean "consisting of aggregates 

 small enough to pass the particular filter used." If the 

 sample is not filtered prior to analysis, then not only 

 dissolved material, but also any sorbed material re- 

 moved by the method of analysis, will be included in 

 the result. 



In the absence of exhaustive data on the sorptive 

 capacity and saturation of river silts, it is impossible to 

 evaluate the magnitude of this error exactly. It will 

 vary from element to element and from river to river. 

 For the principal components in most rivers it will not 

 be of very great importance. The quantity of sus- 

 pended material usually carried by streams is hardly an 

 order of magnitude greater than the quantity of dis- 

 solved material, according to usual methods of discrim- 

 inating between them. As only a small part of the 

 suspended material consists of sorbed components, esti- 

 mates of the total amount of mineral matter carried by 

 rivers or of the principal components of the mineral 

 matter are unlikely to be seriously in error. 



Any complete consideration of the geochemical role 

 of rivers, however, cannot be restricted to the major 

 elements that are strongly ionized, but must include 

 the trace elements, plus more abundant elements that 

 vary widely in solubility under the conditions that 

 prevail in the hydrosphere. 



Elements so scarce that they tend to limit the growth 

 of aquatic organisms provide another example of the 

 danger of trying to carry out geochemical calculations 

 with only the dissolved component of river solids. 

 Phosphorus, for example, may be reduced to so low a 

 level in waters that is is not detectable in inorganic so- 

 lution. Under such circumstances an analysis for phos- 

 phate ion will be completely misleading. There may 

 be appreciable phosphorus in the water in the bodies of 

 plants and animals or in the form of dissolved organic 

 materials. The same is true for nitrogen and silicon, 



