UNIVERSITY OF CALIFORNIA PUBLICATIONS 
 
 IN 
 
 AGRICULTURAL SCIENCES 
 
 Vol. 1, No. 10, pp. 291-339 April 29, 1916 
 
 FUNDAMENTAL INTERRELATIONSHIPS 
 
 BETWEEN CERTAIN SOLUBLE 
 
 SALTS AND SOIL COLLOIDS 
 
 BY 
 
 L. T. SHARP 
 
 INTRODUCTION 
 
 "While engaged in an extensive investigation of the physio- 
 logical effects of NaCl, Na 2 S0 4 and Na 2 C0 3 on crop plants as 
 grown in the Davis clay loam, in cylinders, under field condi- 
 tions, the writer observed that the soil to which the salts had 
 been previously applied became so impervious during the course 
 of the experiment as to retard markedly the rate of percolation. 
 So pronounced was this effect that during the winter and early 
 spring months all of the salt-treated soils were continuously 
 covered with standing water. The appearance of this striking 
 modification in the permeability of the soil to water in the salt- 
 treated soils, together with the inferior cultivating qualities 
 exhibited by them, impressed us as evidence of the fact that the 
 salt treatments under the field conditions of the experiment had 
 effected a fundamental change in the physical constitution of the 
 soil. The occurrence and nature of this change and its relation 
 to soil colloids, interior surface, and other properties of soils, 
 form the considerations with which this paper is chiefly con- 
 cerned. 
 
 Just such an effect on the physical condition of the soil as 
 described above had been anticipated as the normal result of 
 adding Na 2 C0 3 to the soil, for this salt has generally been con- 
 ceded, by soil experts, to be an active deflocculating agent. But 
 
292 University of California Publications in Agricultural Sciences [Vol. 1 
 
 to find, on the other hand, that a similar state of diffusion existed 
 in the soils to which NaCl or Na 2 S0 4 had been added appeared 
 wholly inexplicable in the light of the prevailing conception of 
 these salts as agents capable of producing flocculation in the 
 soil colloids. However, this apparent contradiction of well- 
 established facts brought out by the field observations becomes 
 more intelligible as the accumulating data define more clearly the 
 conditions necessary to produce the remarkable effects observed. 
 Thus it has been demonstrated, not only in the field but also in 
 the laboratory, that the removal from the soil by water of NaCl 
 or Na 2 S0 4 , together with the water-soluble products of their 
 chemical reaction with the soil constituents, either wholly or in 
 part, is the initial step in creating a condition favorable for the 
 diffusion of the soil colloids and possibly for the formation of 
 new colloidal matter. Therefore the net result of salt application 
 to and subsequent washing of a soil is to render the soil com- 
 paratively impervious and to injure seriously its physical con- 
 dition. The leaching out of added Na 2 C0 3 from the soil also 
 presents some interesting phenomena, which are discussed below. 
 
 Although the alteration in the physical condition of the soil 
 was first observed by the writer as purely incidental to an investi- 
 gation primarily designed to ascertain the toxicity limits of the 
 common alkali salts for crop plants, yet it has proved, at least 
 in the case at hand, a most perplexing factor in the production 
 of crops. Our experience would lead us to believe that these 
 after-effects of salt treatments, which appear during the course 
 of leaching the salts from the soil, would have some application 
 to the management of alkali lands, and perhaps some significance 
 with respect to fertilizer treatments. However, the literature 
 on these subjects, with a few exceptions, seems quite devoid of 
 any pertinent reference to the possible importance in these 
 problems of the factor discussed above. 
 
 A survey of the literature concerning soils reveals the chief 
 exceptions just mentioned in the following important contribu- 
 tions to Hi is subject. Thus, a brief but significant chapter. 
 ' ' Veranderung der Durchlassigkeit (lurch Auswaschen der 
 Sake/' by Adolph Mayer, 1 records a somewhat sudden reduction 
 
 • Porschungen auf dem Gebiete der Agrikultur-Physik, vol. 2, 1879, p. 251. 
 
1916] Sharp : Soluble Salts and Soil Colloids 293 
 
 in the permeability of a soil when NaOH is washed from it. A 
 similar experiment performed by Mayer with sodium phosphate 
 was not accompanied with the sudden reduction in permeability 
 of the soil as was noted when NaOH was washed from the soil. 
 Likewise the washing out of lime-water produced no marked 
 effect on the rate of percolation, but the leaching out of added 
 NaCl, either with pure water or with lime-water, reduced percola- 
 tion to a minimum. Mayer ascribes the poor physical condition 
 of drained sea-shore lands to the washing out of the salt, and 
 significantly remarks that this effect, which frequently appears 
 in the second year, is probably more injurious to crop plants 
 than the toxicity of the salt itself. 
 
 Van Bemmelen 2 in his classical researches on colloids has also 
 observed a similar decrease in the rate of percolation when loosely 
 bound salts are washed from clays or from the hydrated oxides 
 of tin, silica, and manganese. Moreover, he noted that these 
 colloids, when subjected to salt treatments followed by leaching 
 with water, invariably exhibited a high degree of diffusion upon 
 suspension in water for a second time. He further asserts that 
 this process can be indefinitely repeated by alternately adding 
 to and washing salt from the colloids. The colloidal particles, 
 as remarked by Van Bemmelen, become so exceedingly small, 
 during the process of washing salts from them, as to pass through 
 the filter paper. 
 
 Warington 3 also refers in a general way to the appearance 
 of somewhat similar phenomena when soils, previously treated 
 with acids, are washed with water. 
 
 It would appear, with the above exceptions, that those who 
 have studied absorption, adsorption, or other physico-chemical 
 effects of salts on soils, have failed to recognize the existence of 
 any relation between the washing out of salts and the subsequent 
 condition of the soil. In fact, Hall and Morison 4 assert that the 
 flocculating effect of salt solutions on kaolin are reversible, that 
 is to say, upon the removal of the added salt the kaolin resumes 
 the original condition of diffusion. 
 
 2 Journ. prakt. Chem., 2nd ser., vol. 23, p. 388, 1881. 
 
 3 Physical Properties of Soil (Cambridge Univ. Press, 1900), p. 30. 
 
 4 Journ. Agri. Sci., vol. 2, p. 244, 1907. 
 
294 University of California Publications in Agricultural Sciences [Vol. 1 
 
 That these earlier findings have a greater significance in 
 agricultural practice than was formerly attributed to them be- 
 comes very evident when the practical phase is attempted of 
 cropping the diffused, salt-treated, water-washed soils in our 
 cylinders. Heretofore, the residual effects left upon clays and 
 clay soils after salts have been washed from them have been 
 chiefly considered in connection with their application to the 
 ceramic industry, and have been largely omitted from considera- 
 tion by those dealing with soils. 
 
 Since the striking soil behavior under discussion was first 
 observed by us, as above explained, in crop cylinder experiments, 
 it appears desirable to introduce here a brief description of these 
 experiments. 
 
 The Cylinder Experiment on the Physiological Effects of 
 NaCl, Na 2 S0 4 and Na 2 C0 3 on Crop Plants 
 
 The cylinders used in this experiment are of galvanized iron, 
 open at both ends, coated inside and out with asphaltum, and 
 are fourteen inches in diameter and five feet long. They were 
 placed, during September, 1913, in a clay loam soil at Davis, 
 California, by digging holes to the depth of five feet. During 
 the progress of digging the holes the soil from each foot was 
 carefully removed and set aside separately, thus permitting the 
 cylinders to be refilled with the soil layers in the same order as 
 they exist naturally. 
 
 The soils within the cylinders were then supplied with vary- 
 ing percentages of the three sodium salts, NaCl, Na 2 SO t , and 
 Na 2 CO ;r It seemed a difficult task to secure a satisfactory mixture 
 of the salts with the soil already in the cylinders, and it was 
 therefore deemed wise to apply the salts to the surface of the 
 soils by means of a solution containing the weighed quantity 
 of salt in 8800 cc. of distilled water. Accordingly, each cylinder 
 received the same quantity of water, but of a different salt con- 
 centration. In addition to the salts some of the soils also re- 
 ceived other treatments which, together with the salt treatments, 
 are detailed in the following chart. The quantities of salts were 
 
1916] Sharp: Soluble Salts and Soil Colloids 295 
 
 calculated and applied on the basis of percentage of weight of 
 the five-foot column of soil in every cylinder. 
 
 As mentioned above, all of the salt-treated soils developed 
 every indication of a thoroughly diffused condition. As a result, 
 percolation through the treated cylinders practically ceased, so 
 that the rain-water collected in the two-inch cylinder rim and 
 remained so persistently as to prove seriously detrimental to the 
 
 V_y v*co) v*op) v$oc) \t40p) v$og) \f$o$) \f3o? yeog) ysg) \ggj) ^ooo) V_y 
 
 ©©©©©©©©©©©©©^ 
 
 ©©©©©©©©©©©©© UQZ °° 5 
 ©©©©©©©©©©©©© N026 ° 4 
 
 ©©©©©©©©©©©©© Naa 
 ©©©©©©©©©©©©© N ^ 
 
 ©©©©©©©©@@@© N ° 25 ° 4 
 
 (5j\ /j$\ K&\ (£q\ (&\ rS\ (30\ /3*\ (5o\ tt\ /^\ (&\ UaCI 
 
 ©©©©(^)©©©@©0© N ° 2C ° 3 
 
 (5o\ fo\ (5$\ (&\ (J$\ (ji\ fo\ (J$\ /tt\ (Zo\ fiS\ (zo\ Nfl 2 G0 3 
 \_J ViaoJ Y&/ Vf&J V**?J [ <K*J V*o) \Z9oo) \qooJ \f9ooj Kqo) \$300j 
 
 /9\ (Zs\ /2o\ (&\ (5o\ (o\ (j&j (*o\ /3£\ tt\ ft&\ (Zo\ y\ C\ 
 
 ©©©©©©©©©Q©© U0ZC ° 6 
 
 Fig. 1. The Salt and Manure Applications to the Soil in the 
 
 Cylinders at Davis. 
 
 The upper figures refer to the percentages of salts in the 5-foot 
 
 column of soil. The lower figures refer to the grams of barnyard manure 
 
 added. In row 1, no. 1 contains Na 2 S0 4 instead of NaCl. In row 10, no. 
 
 2 likewise contains Na 2 S0 4 instead of Na 2 C0 3 . 
 
 growth of crops. As a final measure, holes were opened down 
 through the soil columns to provide drainage. In sharp contrast 
 with the treated soils, even immediately after heavy rain-storms, 
 no standing water was ever observed on the control soils, and 
 the splendid growth of the crops planted on them showed a con- 
 genial condition for crop development. Both the winters of 
 
296 University of California Publications in Agricultural Sciences [Vol. 1 
 
 1913-14 and 1914-15 were accompanied with similar manifesta- 
 tions of the extreme retentiveness and imperviousness of the salt- 
 treated soils. 
 
 As previously remarked, all of the check soils responded 
 readily to cultivation and excellent seed beds were easily pre- 
 pared, while the soils to which NaCl and Na 2 S0 4 had been added 
 were in an unworkable condition. A crust, an inch or so in 
 thickness, formed as soon as the surface dried, and was so hard 
 that even heavy tools made but little impression on it. Just 
 below the crust the soil was far too wet for plant roots and re- 
 sembled putty in consistency. Contrary to expectations, the 
 soils treated with Na 2 C0 3 exhibited better cultivating qualities 
 than the soils receiving the other salts, but inferior to those 
 of the control soils. 
 
 With a view to obtaining further information concerning the 
 great dissimilarity between the control soils and the salt-treated 
 soils, it seemed wise to examine some of the surface water stand- 
 ing in the cylinders as well as the soil in contact therewith. These 
 proved to contain but mere traces of any salts. Of next im- 
 portance a study of the vertical distribution of the salts in the 
 cylinders disclosed the fact that the first few inches of soil 
 contained relatively small quantities of the salts, the first foot 
 more, the second still more, and the third foot most salt. From 
 these results it was concluded that the addition of NaCl or Na 2 S0 4 
 to a soil and subsequently washing the added salt, or at least 
 part of it, from the soil with water, produces a high degree of 
 diffusion in the soil colloids. Moreover, this alteration in diffusion 
 was accompanied with proportional changes in the other physical 
 characteristics of the soil. These conclusions based on our field 
 experience were soon verified by laboratory experiments. 
 
 Laboratory Experiments Reproducing Field Observations 
 
 It has been found that the field conditions can be readily 
 reproduced in the laboratory. Thus a simple procedure, illustrat- 
 ing qualitatively the lessened percolation, is to place 10 grams or 
 more of soil in a filter paper in a funnel, adding thereto either 
 NaCl or Na 2 S0 4 as a solid or in a water solution, and then wash- 
 
1916] 
 
 Sharp: Soluble Salts and Soil Colloids 
 
 297 
 
 ing the soil practically free of salt, meanwhile washing with 
 pure water a similar portion of soil from v/hich the salt has 
 been omitted to serve as a control. The difference in the rate of 
 percolation of the two soils has proved to be sufficiently great to 
 be easily discernible even in the case of sandy soils. Frequently 
 
 108144 180 216 252 286 324 360 396 432 468 504 540 576 
 
 ""• • » 1 • 1 r~ -t— 1 1 1 1 1 1 
 
 Fig. 2. Curves Representing the Downward Movement of Water Through 
 the Davis Soil to which Salts Have Been Added. 
 The penetration of the water downward in the soil column is expressed 
 in inches by the ordinates, while the abscissas represent the time of 
 observations in hours. 
 
 the fine material mentioned by Van Bemmelen and Warington 
 appears in the percolate. 
 
 It has also been found possible to simulate fairly the condi- 
 tions existing in the field cylinders by means of glass tubes care- 
 
298 University of California Publications in Agricultural Sciences [Vol. 1 
 
 fully filled with the Davis clay loam to which chemically equiva- 
 lent salt solutions were added, much in the same manner as the 
 salt applications were made to the cylinder soils. A constant 
 head of water was then carefully maintained in the tubes. The 
 downward movement of the water was observed from time to 
 time and is graphically portrayed in Figure 2. 
 
 From the curves in it it is very apparent that the influence of 
 these salt additions, accompanied by subsequent surface applica- 
 tions of water to the soil, has been to retard, markedly, the down- 
 ward movement of the water through the soil. In less than 33 
 hours the water had penetrated the entire 14^2 inches of soil in 
 the control tube, while at the end of 552 hours, under a constant 
 head of three inches, the maximum distance reached in any of 
 the salt-treated soils was 13% inches in the case of the soil 
 receiving NaCl, followed by 13*4 inches in the Na 2 C0 3 soil, and 
 but 11 inches in the soil to which Na 2 S0 4 had been added. At 
 the expiration of 56 days no percolation had occurred, though 
 the entire soil columns were moist. The total absence of percola- 
 tion in the salt-treated soils was attributed to the growth of algae 
 in the tubes. 
 
 To test further the points under consideration, a second series 
 was prepared to study the rates of percolation through Davis 
 clay loam to which salts had been added in various ways. The 
 results of this experiment are embodied in Table I. 
 
 In the above experiment the chemically equivalent quantities 
 of salts were mixed uniformly with the dry soil before placing 
 the mixture in the paramne-coated brass percolation tubes, except 
 in the case of Nos. 12, 13, and 14, through which salt solutions 
 were passed of a strength (calculated on the basis of a previously 
 determined water-holding power of the soil) to give compara- 
 tively equal quantities of salts with those soils receiving the dry 
 salts. A head of one to two inches was carefully maintained on 
 the soil in all the tubes throughout the experiment. 
 
 The results recorded in the foregoing table fully corroborate 
 the field experience. Thus, the average rate of percolation of 
 the check soil was 1.59 cc. per hour, while that of the soil to 
 which NaCl has been added was reduced to .19 cc. per hour. The 
 percolation through the other two soils which received solid 
 
1916] 
 
 Sharp: Soluble Salts and Soil Colloids 
 
 299 
 
 Ps 
 
 
 JO 
 
 
 P 
 
 
 
 
 CD 
 
 > 
 
 
 rt 
 
 OS 
 
 
 OS 
 
 P - 
 
 HO'OOOSOlOl^W 
 
 to co co co 
 
 i — ' i — >■ ' — ' i — ' i — 1 i — >• r>~ m 
 
 co co co to to to c rS,W 
 
 h hj 
 p g. 
 
 ol 
 
 b\ 
 
 bi 
 
 bi 
 
 bn 
 
 bi 
 
 bi 
 
 bi 
 
 bi 
 
 bi 
 
 Ol 
 
 Ol 
 
 Ol 
 
 Ol 
 
 
 
 
 
 
 
 cs 
 
 CS 
 
 CS 
 
 CS 
 
 C5 
 
 rs 
 
 C5 
 
 C5 
 
 CS 
 
 CS 
 
 CS 
 
 CS 
 
 "5 
 
 rs 
 
 
 
 
 
 
 = 
 
 3 
 
 3 
 
 B 
 
 B 
 
 3 
 
 B 
 
 3 
 
 3 
 
 3 
 
 3 
 
 5 
 
 3 
 
 5 
 
 3 »-» 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1— ' jo 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 tO p 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o M . 
 
 B 3 
 
 
 
 hd 
 
 hj 
 
 ^ 
 
 !zj !z 
 
 03 H 
 
 
 
 
 
 
 
 
 CD 
 CS 
 
 CD 
 CS 
 
 CD 
 hi 
 
 C5 
 
 p 
 
 1= 
 
 p 
 
 p 
 
 p 
 
 p 
 
 p 
 
 M 
 
 99 
 
 £. M 
 
 
 
 
 
 
 O 3 
 
 
 
 Q 
 
 QC 
 
 rs 
 
 Q 
 
 X 
 
 q ■ 
 
 b x 
 
 2 
 
 
 
 
 
 
 
 S- ^ 
 
 
 
 o 
 
 o 
 
 O 
 
 O 
 
 o 
 
 
 O 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 p 1 
 
 p' 
 
 p" 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 <T+- 
 
 rt- 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 CO p 
 
 
 
 CD 
 
 CD 
 
 CD 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ° St 
 
 
 
 Dj 
 
 P-, 
 
 Oj 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 »' s 
 
 
 
 CO 
 
 o 
 
 co 
 
 O 
 
 CO 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 w 
 
 CD 
 
 CO c+ 
 
 oo 3 
 
 ££ 
 
 o p 
 
 
 
 5" 
 
 
 
 5" 
 
 J_ i 
 
 io 
 
 to 
 
 '— ' 
 
 io 
 
 to 
 
 — • 
 
 
 
 
 
 
 
 
 
 
 ► 
 
 
 
 00 
 
 4- 
 
 o 
 
 X 
 
 4- 
 
 o 
 
 00 
 
 4- 
 
 o 
 
 
 
 
 
 g 
 
 
 
 
 
 CO 
 
 OS 
 
 o 
 
 00 
 
 or. 
 
 o 
 
 00 
 
 - - 
 
 O 
 
 c+ J/J (D 
 
 
 
 
 
 H 
 
 
 o 
 
 O 
 P 
 O 
 Hs 
 
 o 
 o 
 
 Hs 
 
 o" 
 o 
 
 Hs 
 
 
 
 
 o 
 
 O 
 
 a 
 
 
 
 
 " 3 
 
 
 
 CS 
 
 5' 
 
 CD 
 
 5" 
 
 O 
 
 - 
 
 3 
 
 p 
 
 X 
 
 fe! 
 
 3 
 
 tzj 
 
 G 
 
 Q 
 
 Q 
 
 p 
 O 
 
 p 
 
 p 
 X 
 
 p 
 
 X 
 
 
 
 
 02 to 
 
 
 
 o 
 
 p- 
 
 
 * s, 
 
 p 
 
 p 
 
 b 
 
 p 
 
 P 
 
 c 
 
 p 
 o 
 
 p 
 Q 
 
 p 
 Q 
 
 O 
 
 O 
 
 p 
 
 
 
 
 P t3 
 
 
 
 CD 
 
 
 § 5 
 
 h- 2 
 o © 
 
 p p 
 
 c 
 
 to 
 
 O 
 
 O 
 
 O 
 
 to 
 
 to 
 
 to 
 
 
 
 
 
 
 
 d 
 
 CD 
 
 o 
 
 
 w 
 
 
 
 
 
 WWW 
 
 
 
 
 
 
 
 O 
 
 o, 
 S" 
 
 
 
 b 
 
 o 
 o 
 o 
 
 OS 
 
 bo 
 
 o 
 © 
 o 
 
 00 
 
 bi 
 
 o 
 o 
 
 o 
 
 OS 
 
 to 
 
 OS 
 
 
 
 
 b 
 
 b 
 
 O 
 
 
 
 
 
 
 
 p' 
 o' 
 
 o 
 
 CO 
 
 M 
 
 CS 
 
 30 
 H 
 
 co 
 P 
 
 cs 
 
 H 
 CO 
 
 Si 
 
 C5 
 
 H 
 
 CO 
 
 p 
 
 to 
 
 io 
 
 to 
 
 p_, 
 
 p_j 
 
 j_, 
 
 
 
 
 
 
 3 
 
 
 H] 
 
 ^ o 
 % * 
 
 2. p 
 
 o 
 
 O 
 
 CS 
 
 w 
 b 
 
 P 
 
 H 
 b 
 
 CS 
 
 w 
 b 
 
 o 
 
 O 
 
 o 
 
 o 
 
 o 
 
 o 
 
 
 
 
 rt-3 CD 
 
 H] 
 
 
 5' 
 
 o 
 
 3 
 
 3 
 P- 
 
 d 
 
 Q 
 
 H 
 
 X 
 
 o 
 
 crq o 
 
 ^ ST 
 h- J 
 
 5' " 
 
 4- 
 Ol 
 
 4- 
 bc 
 
 
 o 
 
 to 
 
 4i. 
 
 bi 
 
 
 -1 
 
 io 
 
 b 
 
 be 
 
 '-) 
 
 ^ 
 
 95 w.O 
 
 ST 1 
 
 5" 
 
 o 
 
 o 
 
 
 en 
 
 00 
 
 Oct. 15 
 
 Jj 
 
 to 
 
 o 
 
 Ol 
 
 CO 
 
 H^ 
 
 CO 
 
 ^1 
 
 OS 
 
 o 
 
 4^ 
 
 to 
 
 O 
 
 -J 
 
 H 1 Qpq fa 2 
 
 c 
 
 
 
 
 5T © 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 09 
 
 5 
 
 CD 
 
 
 
 72 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 o 
 
 -*> 
 
 
 
 » 
 
 CD 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 E=l 
 
 e-t- 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 « 
 
 
 
 O 
 
 3 o 
 
 tc 
 
 co 
 
 M 
 
 co 
 
 to 
 
 
 
 
 
 
 
 
 
 
 to ' 
 
 
 
 2 
 
 
 
 M 
 
 CD 
 
 P £ 
 
 co 3> 
 
 10 
 
 b 
 
 f_i 
 
 b 
 
 4^ 
 
 OS 
 
 -i 
 
 bi 
 
 b 
 
 OS 
 
 4^ 
 
 4- 
 
 CO 
 
 to 
 
 b 
 
 
 
 o 
 
 Ol 
 
 CO 
 
 Oct. 21 
 
 < 
 
 X 
 
 Ol 
 
 X 
 
 Ol 
 
 Ol 
 
 i- 1 
 
 0J 
 
 4- 
 
 o 
 
 4- 
 
 4^ 
 
 ^1 
 
 Ol 
 
 oo 
 
 o 
 
 5 
 
 
 p* 
 
 
 X 
 
 P CD 
 
 H 
 
 CD CO 
 
 
 to 
 b 
 
 1— » 
 
 b 
 
 to 
 
 to 
 
 '•<! 
 
 bi 
 
 bi 
 
 4- 
 
 -t 
 
 bi 
 
 bo 
 
 4- 
 
 DC 
 
 to 
 
 4- 
 
 
 
 c' 
 3 
 
 CO 
 00 
 
 Oct. 26 
 
 53 
 
 2 a 
 
 w 
 
 4- 
 
 o 
 
 CD 
 
 CO 
 
 OS 
 
 OS 
 
 O 
 
 4- 
 
 4- 
 
 GO 
 
 O 
 
 o 
 
 1-1 
 
 oo 
 
 5 
 
 
 
 
 
 > 
 
 ?3 
 
 P c+- 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 CD 
 
 
 l-t 
 
 
 
 
 e-»- M- 
 
 to 
 
 JO 
 
 j- 
 
 Hi 
 
 Ka 
 
 
 
 
 
 
 
 
 
 
 M 
 
 CD 
 
 
 o 
 
 (0 
 
 
 3 
 
 02 
 
 O P 
 
 to 
 
 bi 
 
 b 
 
 b 
 
 bi 
 
 OS 
 
 -1 
 
 4- 
 
 bo 
 
 OS 
 
 4^ 
 
 4- 
 
 oo 
 
 io 
 
 X 
 
 
 
 o 
 
 Oct. 30 
 
 Hs e£ 
 
 CO 
 
 o 
 
 Ol 
 
 4^ 
 
 OS 
 
 03 
 
 M 
 
 ^3 
 
 o 
 
 to 
 
 4^ 
 
 4- 
 
 *o 
 
 4- 
 
 t— ' 
 
 
 
 
 
 
 e*- ©' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 „„. 
 
 
 
 X 
 
 &- P 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 d 
 
 
 
 > 
 
 CD 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 d 
 
 cs' 
 
 
 
 43 
 
 T3 Hs 
 CD 
 
 4- 
 
 <i 
 
 bo 
 
 b 
 
 
 4^ 
 
 4- 
 
 to 
 
 4- 
 
 4- 
 
 io 
 
 io 
 
 to 
 
 'r- 1 
 
 -i 
 
 
 
 10 
 
 Nov. 4 
 
 H3 
 
 CS M- 
 
 m 
 
 r^ 
 
 co 
 
 co 
 
 o 
 
 3 
 
 rf^ 
 
 y 
 
 ffl 
 
 O 
 
 -<l 
 
 ~i 
 
 to 
 
 4^ 
 
 ^1 
 
 w 
 
 
 rt 
 
 
 
 w 
 
 O co 
 
 p" £ 
 
 CD CD 
 
 X 
 
 CO 
 
 ^ 
 
 CO 
 
 bi 
 
 bt 
 
 bi 
 
 CO 
 
 bi 
 
 bi 
 
 03 
 
 CO 
 
 to 
 
 
 bi 
 
 CD 
 O 
 
 o 
 
 3 
 
 
 CD 
 
 i" 
 
 45> 
 
 Nov. 9 
 
 
 * P^ 
 
 4- 
 
 Cl 
 
 CO 
 
 4^ 
 
 ^ 
 
 o 
 
 -i 
 
 4- 
 
 -i 
 
 H* 
 
 to 
 
 to 
 
 X 
 
 X 
 
 co 
 
 CD 
 
 
 CD 
 
 
 
 2! 
 
 3 
 P 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 H 
 
 
 
 M 
 
 50 
 
 to 
 
 4- 
 
 b 
 
 to 
 
 'co 
 
 bi 
 
 OS 
 
 'c: 
 
 bi 
 
 bi 
 
 bo 
 
 CO 
 
 CO 
 
 ^j 
 
 OS 
 
 CD 
 
 
 
 Ol 
 
 to 
 
 Nov. 11 
 
 
 CD 
 
 tc 
 
 - 
 
 1-1 
 
 M 
 
 Ol 
 
 CO 
 
 c 
 
 OS 
 
 X 
 
 4- 
 
 Ol 
 
 Ol 
 
 H 1 
 
 •O 
 
 1—1 V 
 
 
 
 
 
 
 
 Ha 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 ► 
 
 
 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 rt-C 
 
 
 <1 
 
 
 
 
 H 
 
 
 to 
 
 
 H> 
 
 M 
 
 
 
 
 
 
 
 
 
 
 - 1 o"5g 2. 
 
 - 
 
 a. a 
 p 
 
 
 
 
 CD 
 
 X 
 
 bo 
 
 b 
 
 rf^ 
 
 "co 
 
 0( 
 
 OS 
 
 w 
 
 OS 
 
 bi 
 
 bo 
 
 w 
 
 co 
 
 — i 
 
 w S 2 x 
 
 v. 
 
 
 
 
 ~q 
 
 h-i 
 
 o 
 
 ^1 
 
 o 
 
 h^ 
 
 o 
 
 sc 
 
 00 
 
 4^ 
 
 OS 
 
 o 
 
 
 O 
 
 -D M 5l 
 
 
 CD 
 
 
 
 
 CO 
 P 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Op 
 
 ^^ 
 
 J> 
 
 
 
 
 e+ 
 
 
 
 M 
 
 to 
 
 H> 
 
 
 
 
 
 
 
 
 
 
 5'm P- 0- 
 
 3 » 
 
 
 
 
 co 
 
 
 
 b 
 
 4^ 
 
 b 
 
 
 
 
 
 
 
 
 
 
 
 
 o> Hi 
 
 
 
 
 O 
 
 
 
 ^i 
 
 to 
 
 4- 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 P 
 (T? 
 CD 
 
 
 
 
 et- 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 P 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Op 
 
 Ol 
 
 <5 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 oS e. 
 
 3 Cfi 
 
 
 
 
 h3 
 
 to- 
 
 
 
 *-i 
 
 bn 
 
 OS 
 
 
 
 
 
 
 
 
 
 
 d ™ J 
 
 ? 
 
 ^89 
 CD 
 
 
 
 
 
 
 4 
 
 i. 
 
 t 
 
 3 
 
 
 a 
 
 
 
 
 
 
 
 
 
 
 a 
 
 < r+ 
 
 p pr 
 
 =- 
 
 
 
 
300 University of California Publications in Agricultural Sciences [Vol. 1 
 
 Na 2 S0 4 and Na 2 C0 3 was also decidedly reduced, as shown by 
 the respective percolation rates of .31 cc. and .37 cc. per hour. In 
 significant contrast thereto, the maintenance of a comparatively 
 uniform salt content in the soil column, produced by substituting 
 solutions of the salts for the distilled water as in the case of 
 Nos. 12, 13, and 14, creates a favorable condition for percolation, 
 except in No. 14, which received Na 2 C0 3 . In this case percola- 
 tion averaged a rate of 1.07 cc. per hour, which was less than the 
 control, but three times that of the soil receiving Na 2 C0 3 and 
 distilled water. It is of interest to note the immediate depression 
 in the rate of percolation when distilled water is used instead of 
 the salt solutions of Nos. 12, 13, and 14. Eventually the rate of 
 percolation from these soils under applications of distilled water 
 approached that of the soils originally treated with salts and 
 which in addition had received only distilled water throughout 
 the experiment. 
 
 Another percolation experiment, arranged somewhat similarly 
 to the preceding, demonstrated that the relative position of the 
 salt in the soil column, or the manner of adding it, had little or 
 no influence on the depressing effect noted on percolation. It was 
 further observed that a soil, the percolation rate of which had 
 been diminished through the agency of NaCl and H 2 applica- 
 tions, failed to recover its original rate of percolation even when 
 a solution of that salt was applied for the second time. 
 
 The general trend of the results thus far secured is in accord 
 with Beeson's observations, 5 in which he recorded a delayed 
 absorption of water by soils containing sodium salts as well as 
 a pronounced retarding of percolation through soils to which 
 various salts had been added. Like many other investigators 
 who have observed similar effects resulting from salt treatments, 
 Beeson failed to recognize any connection between the physical 
 condition of the soil and the removal of the salt, but attributed 
 the peculiar changes observed in the soil to surface-tension 
 phenomena, or other alterations in the physical properties of the 
 liquid phase, or, as some investigators are inclined to believe, 
 to a shifting of the soil particles to new positions through the 
 influence of the added salt. Undoubtedly a movement of soil 
 
 BJourn. Am. Chem. Soc, vol. 29, p. 620, 1897. 
 
1916] Sharp : Soluble Salts and Soil Colloids 301 
 
 particles does occur when salts are added to soils, but under the 
 conditions obtaining in many of the experiments described in the 
 literature it would seem highly improbable that enough of the 
 added salt remains in contact with the soil to bring about such 
 a movement of the particles. In our experience, the removal of 
 the salt creates among the soil particles a new adjustment which 
 seems of greater importance than that effected by the addition 
 of the salt. Furthermore, if our laboratory experience is properly 
 applicable to field conditions, it would seem of greater advantage 
 to apply CaC0 3 rather than CaS0 4 when draining soils contain- 
 ing NaCl and Na 2 S0 4 , the reverse being apparently true in case 
 of soils impregnated with Na 2 C0 3 . 
 
 Laboratory Investigations of the Causes of Salt 
 Effects on Soils 
 
 The laboratory studies herein reported have been purposely 
 designed to throw some light on the possible causes contributing 
 to the above-noted effects of salts on soils, or at least to ascertain 
 if a relationship exists between the formation of these peculiar 
 physical conditions and the simultaneous occurrence of certain 
 other events. Certain well-known theoretical considerations, re- 
 inforced by concrete laboratory experience, have directed the 
 attempts to locate these fundamental causes into three well- 
 defined channels. The first of these is based on an assumption 
 that the salt-and-w T ater treatments have actually increased the 
 quantity of the colloidal matter of the soil. It involves neces- 
 sarily a study of the soil itself and of the amount and degree of 
 diffusion of the colloidal matter therein contained. The second 
 line of reasoning connects the increase of calcium and magnesium 
 in the percolate from the salt-treated soils, and the absorption of 
 sodium, with the appearance of conditions suitable for the 
 formation of new colloidal matter, as well as with favoring the 
 extreme diffusion of that already present. The third considera- 
 tion ascribes the diffused condition of the soil colloids to the 
 presence of a small quantity of OH-ions in the soil solution. It 
 requires the theoretical assumption that these ions are associated 
 in some manner with the absorption of sodium. Although these 
 
302 University of California Publications in Agricultural Sciences [Vol. 1 
 
 three lines of argument, as outlined, may be closely allied or even 
 parallel, yet in view of our work they seem sufficiently well de- 
 fined to warrant a separate discussion of each. 
 
 At the outset it may be said that it has been extremely diffi- 
 cult to find appropriate and reliable methods of attacking this 
 problem. Certainly opportunity is not lacking for future in- 
 vestigations to perfect accurate methods for definitely measuring 
 certain soil properties or soil constants, which are discussed 
 below. The Davis clay-loam soil was used throughout the fol- 
 lowing experiments, except as otherwise noted. The designation 
 NaCl + H 2 used before the word soil signifies that the soil 
 was treated with from 0.2 to 0.5 per cent of NaCl, most or all 
 of which, together with the soluble salts formed, has been subse- 
 quently leached from the soil with water. It must be added that 
 a quantity of Davis clay-loam soil was given the same amount 
 of washing with water to serve as a control soil, and hereafter 
 will be referred to as H 2 soil. A comparison of the H 2 soil 
 with unwashed Davis soil disclosed no important differences. 
 
 Some Physico-Chemical Observations on the Salt-Treated, 
 Water-Washed Soil 
 
 It seemed reasonable to expect that a study of the NaCl + H 2 
 soil itself would reflect, in some degree, the causes contributing 
 to the pronounced imperviousness of such soils. Therefore, as 
 an introduction to this subject, determinations were made of the 
 suspended matter derived from soils which have been subjected 
 to various treatments, and they appear in Table II. The soils 
 and suspending media were placed together in tall hydrometer 
 jars, and were thoroughly shaken for 45 minutes. After standing 
 undisturbed for eighteen hours, aliquot portions of the sus- 
 pensions were removed by means of a pipette, evaporated to 
 dryness in platinum dishes, dried at 110° C, and weighed. 
 
 The data reported in Table II confirm by an entirely different 
 procedure the original belief with respect to the intensity of the 
 diffusion of the colloids of the salt-treated, water-washed soils. 
 Under the conditions of the experiment 40 grams of normal soil 
 yielded 0.310 grams of suspended matter, while 40 grams of 
 
1916] Sharp: Soluble Salts and Soil Colloids 303 
 
 TABLE II 
 
 Suspended Matter in Soils Which Have Received Various Treatments 
 
 Per cent of 
 
 Weight of Suspended 
 
 Previous Suspending Suspended Matter 
 
 No. Davis Soil Soil Treatment Medium Matter-grs. in soil 
 
 1 40 grs. Washed with dis- 
 
 tilled water H 2 .3100 .77 
 
 2 40 grs. Boiled in H 2 H 2 2.5075 6.27 
 
 3 40 grs. .290 grs. of NaCl 
 
 washed out H 2 2.9555 7.39 
 
 4 40 grs N/50Na 2 CO 3 .2665 .66 
 
 5 40 grs N/50NaOH 2.7975 6.99 
 
 NaCl -j- H 2 soil yielded 2.9555 grams, or nearly ten times the 
 amount of suspended matter found in the untreated soil. It is 
 an interesting coincidence that the rate of percolation previously 
 shown for the untreated soil is almost ten times that for soil 
 to which NaCl has been added and subsequently leached out. It 
 may be properly inferred from this that percolation varies in- 
 versely as the degree of diffusion, though our present knowledge 
 does not indicate a relation capable of expression in simple 
 mathematical terms. Furthermore, the inadequacies of the 
 method employed to secure the data in Table II make it impos- 
 sible to express a positive view with reference to the quantity of 
 colloids in the soils tested, but it is evident that the colloidal 
 matter present is in a much higher state of diffusion in certain 
 of the soils than in the control H 2 soil. The three treatments, 
 boiling the soil in water, suspending it in NaOH of certain con- 
 centrations, and leaching added NaCl from it, produce approxi- 
 mately the same degree of diffusion in the soil colloids, as indi- 
 cated by the similarity in the results of the quantitative esti- 
 mations of the suspended matter derived from soils so treated. 
 This agreement in the behavior of the soils receiving the different 
 treatments suggests a similarity or possible relationship between 
 the processes by which these treatments affect the soil or soil 
 colloids. Boiling the soil in water has been assumed b} r soil 
 physicists to disintegrate the colloidal aggregates. If this be the 
 case and no new colloidal substances are formed by this pro- 
 cedure, then the similarity in colloidal content of the boiled soil 
 and the salt-treated, water-washed soil militates against the sup- 
 
 
304 University of California Publications in Agricultural Sciences [Vol. 1 
 
 position that the latter treatment has increased the colloidal 
 matter of the soil. On the other hand, boiling the soil in water 
 may bring about a more profound change in the physical condi- 
 tion of the soil than was formerly attributed to it by soil 
 physicists. Some data withheld from publication at this time 
 indicate that the effect of the boiling treatment is of a different 
 nature from that of the salt-and-water treatment, though the 
 soils receiving the two different treatments yield about the same 
 quantity of colloidal matter. On the other hand, the similarity 
 of the colloidal contents of the boiled soil and the soil suspended 
 in NaOH admits of a more plausible explanation on the basis 
 of NaOH as a deflocculating agent. 
 
 It was thought that the quantitative data reported above 
 might show sufficient dissimilarity to indicate an actual increase 
 in the soil colloids, but a second series of determinations, reported 
 in a later paragraph, are somewhat contradictory to the above, in 
 that the boiled soil yields a suspension slightly richer in colloidal 
 matter than that derived from a NaCl + H 2 soil. This point, 
 however, deserves more investigation before a final conclusion is 
 reached. 
 
 In the light of certain theories more properly discussed in 
 connection with the third hypothesis, it is of great interest to 
 note the general similarity between the NaCl + H 2 soil when 
 suspended in H 2 and the normal soil when suspended in NaOH. 
 One might infer that this peculiar agreement in the behavior of 
 the soils in response to two widely different treatments is not 
 accidental. It is also to be observed further that the data under 
 consideration indicates that NaOH and Na 2 C0 3 are not pro- 
 ductive of like results on soil suspensions. 
 
 The failure of the suspension method to secure trustworthy 
 results on the quantity of colloidal matter present prompted the 
 adoption of other means for this purpose. But thus far the 
 determination of the hygroscopic coefficient and the dye- 
 adsorption capacity have given negative results, in that they have 
 not indicated any increase in the total interior surface of the 
 soils which have been subjected to the salt treatments. From a 
 theoretical consideration, a soil rich in colloids, or containing 
 colloids in a high state of diffusion, should expose more interior 
 
1916] Sharp: Soluble Salts and Soil Colloids 305 
 
 surface for the deposition of hygroscopic moisture and for dye 
 adsorption, and since these phenomena are presumably direct 
 functions of surface the soils under observation should therefore 
 register increased hygroscopicity and dye adsorption. The find- 
 ings to the contrary cast some doubt on the validity of these 
 measurements as a reflection of the quantity of colloidal matter 
 present or its degree of diffusion. A third method suggests 
 itself, that developed by Mitscherlich, 6 to study the interior sur- 
 face through energy exchanges when the soil is moistened with 
 water, but this has not as yet been tried. 
 
 The use of the centrifugal machine as employed by Briggs 
 and McLane 7 to ascertain the moisture equivalent of soils in the 
 study of the salt-treated, water-washed soils, has yielded some 
 highly satisfactory results which will be reported in a future 
 paper. The method proposed by Lynde and Dupre 8 for esti- 
 mating the capillary powers of soils has not proven entirely 
 satisfactory in our hands, when employed for investigating the 
 properties of the salt-treated soils. 
 
 It was further questioned whether the physical condition of 
 the soil had been permanently changed or whether the injured soil 
 would completely recover its original condition in response to a 
 second addition of NaCl. Such a supposition naturally implies 
 that some of the reactions involved in producing the increased 
 diffusion partake of the nature of reversible reactions. In order 
 to test this point from a chemical standpoint, it would be neces- 
 sary to treat the injured soil with its own percolate. Work of 
 this character is reported under the second series of experiments. 
 In this connection, however, the effects of the added salts were 
 considered chiefly in their physical aspects, and accordingly the 
 following experiment was performed : Three grams of NaCl -4- 
 H 2 soil were suspended in 10 cc. of NaCl solutions of various 
 concentrations in test-tubes. The time required to clear the 
 supernatant liquid denoted the effect of the NaCl. A similar 
 comparative series with H 2 soil was also prepared. The results 
 are given in Table III. 
 
 G Bodenkuncle fiir Land- und Forstwirte, p. 51. 
 
 7 U. S. Dept. Agr., Bur. of Soils, Bull. 45. 
 
 s Journ. Amer. Soc. Agron., vol. 5, no. 2, p. 107, 1913. 
 
306 University of California Publications in Agricultural Sciences [Vol. 1 
 
 TABLE III 
 
 Effect of Second Addition of NaCl on Time Eequired to Clear 
 Suspensions of NaCl 4- H 2 Soil and H 2 Soil 
 
 No. 
 
 1 
 2 
 3 
 4 
 5 
 6 
 
 9 
 10 
 11 
 12 
 13 
 
 Soil 
 3grs. 
 3grs. 
 3 grs. 
 3grs. 
 3 grs. 
 3 grs. 
 3 grs. 
 3 grs. 
 3 grs. 
 3 grs. 
 3 grs. 
 3 grs. 
 3 grs. 
 
 Previous 
 Treatment 
 
 NaCl washed out 
 
 NaCl washed out 
 
 NaCl washed out 
 
 NaCl washed out 
 
 NaCl washed out 
 
 NaCl washed out 
 
 H 2 washed 
 
 H,0 washed 
 
 H 2 washed 
 
 H 2 washed 
 
 H 2 washed 
 
 HoO washed 
 
 Concentration 
 
 of NaCl as 
 
 Suspending 
 
 Medium 
 
 N/1000 
 
 N/500 
 N/250 
 N/100 
 
 N/50 
 
 N/1000 
 N/500 
 
 N/250 
 N/100 
 
 N/50 
 
 Time Required to 
 Clear Suspension 
 
 Almost clear after 600 hrs. 
 
 Almost clear after 600 hrs. 
 
 Almost clear after 600 hrs. 
 
 Clear after 600 hrs. 
 
 Clear after 360 hrs. 
 
 Clear after 20 hrs. 
 
 Clear after 164 hrs. 
 
 Clear after 140 hrs. 
 
 Clear after 117 hrs 
 
 Clear after 53 hrs. 
 
 Clear after 1 hr. 
 
 Clear after y 3 hr. 
 
 Clear after 140 hrs. 
 
 The results of this experiment seemed of sufficient interest 
 to warrant a repetition of the work on a larger scale. Accordingly 
 the following experiment was prepared, much in the same manner 
 as the above, but using 25 grams of soil, 250 cc. of, solution, and 
 250 cc. graduates, instead of test tubes. In this case the weight 
 of the suspended matter was also determined after the mixtures 
 had stood undisturbed for three days. A photograph of this 
 series taken two days after the final shaking is shown (Figure 3). 
 
 ;j 
 
 Pig. 3. A Photograph of the Suspensions Described in Table IV 
 The graduates containing the suspensions as shown in Figure 3 cor- 
 respond, from left to right, to the Nos. 1 to 12, inclusive, of Table IV. 
 
1916] 
 
 Sharp: Soluble Salts and Soil Colloids 
 
 307 
 
 
 
 
 TABLE IV 
 
 
 
 
 Effect of 
 
 Second Addition of 
 
 NaCl ON THE 
 
 Amounts of Suspended Matter 
 
 
 
 From NaCl + H 2 Soil and H 2 
 
 Soil 
 
 
 No. 
 
 Soil 
 
 Previous 
 Treatment 
 
 Concentration 
 
 of NaCl Cc. of 
 
 Suspending Suspending 
 
 Solution Medium 
 
 Weight of 
 ; Suspended 
 Matter 
 
 Per cent 
 
 of 
 
 Suspended 
 
 Matter 
 
 on Dry 
 
 Soil 
 
 1 
 
 25 grs. 
 25 grs. 
 
 NaCl washed out 
 
 
 250 
 
 .8345 grs. 
 .6375 
 
 3.33 
 
 2 
 
 NaCl washed out 
 
 N/1000 
 
 250 
 
 2.55 
 
 3 
 
 25 grs. 
 
 NaCl washed out 
 
 N/500 
 
 250 
 
 .5615 
 
 2.36 
 
 4 
 
 25 grs. 
 
 NaCl washed out 
 
 N/250 
 
 250 
 
 .4570 
 
 1.82 
 
 5 
 
 25 grs. 
 
 NaCl washed out 
 
 N/100 
 
 250 
 
 .2305 
 
 .92 
 
 6 
 
 25 grs. 
 
 NaCl washed out 
 
 N/50 
 
 250 
 
 .0170 
 
 .07 
 
 7 
 
 25 grs. 
 25 grs. 
 
 H 2 washed 
 H 2 washed 
 
 
 250 
 
 .0695 
 
 .27 
 
 8 
 
 N/1000 
 
 250 
 
 .0740 
 
 .29 
 
 9 
 
 25 grs. 
 
 H 2 washed 
 
 N/500 
 
 250 
 
 .0660 
 
 .26 
 
 10 
 
 25 grs. 
 
 H 2 washed 
 
 N/250 
 
 250 
 
 .0615 
 
 .24 
 
 11 
 
 25 grs. 
 
 H 2 washed 
 
 N/100 
 
 250 
 
 Lost, but al- 
 most clear 
 
 
 12 
 
 25 grs. 
 
 H 2 washed 
 
 N/50 
 
 250 
 
 
 
 13 
 
 25 errs. 
 
 Boiled 
 
 
 250 
 
 .9060 
 
 3.62 
 
 The evidence presented in the last two tables supports the 
 view that the NaCl -f- H 2 soil has suffered some physical altera- 
 tion which is not readily reversed by the second addition of 
 NaCl. The NaCl added to the soil already diffused by previous 
 treatment with that salt and water is here considered as a 
 physical agent, possessing the power to flocculate clay colloids. 
 To test the reversibility of the chemical reactions occurring when 
 salt solutions are allowed to act upon soils requires a soil treat- 
 ment involving the application of solutions containing, in appro- 
 priate form and quantity, the elements removed by the salt ap- 
 plications. Thus, to restore normal conditions in a soil which has 
 been diffused by a salt-and-water treatment would require the 
 replacement of the absorbed sodium by such metals as were 
 originally present in the soil. The larger amounts of NaCl re- 
 quired to flocculate the colloids of the diffused soil may be due 
 to the increased quantity of colloids present, or to the degree of 
 
308 University of California Publications in Agricultural Sciences [Vol. I 
 
 diffusion of the colloids, or to a change in the nature of the 
 colloid. However, the fact is patent that more NaCl, or a 
 greater length of time for equal quantities of NaCl, is required 
 to produce effects on the NaCl + H 2 soil commensurate with 
 those on the check H 2 soil. Table III shows clearly the relative 
 effectiveness of varying concentrations of NaCl in flocculating 
 the colloidal matter as measured by the time required for clearing 
 the suspension. Thus N/100 NaCl flocculates the colloidal matter 
 of the washed soil in one hour, while 360 hours were required to 
 accomplish the same result with the NaCl -f- H 2 soil. Never- 
 theless, the colloidal matter of the NaCl -f- H 2 soil seems to be 
 more readily flocculated per unit of NaCl than does that of the 
 H 2 soil, as is shown in Table IV. Thus the NaCl + H 2 soil, 
 when suspended in distilled water, yields a suspension containing 
 0.8345 grs. of solid matter, while the same soil suspended in N/250 
 NaCl yields but 0.4570 grs. of solid matter. The deposition of 
 0.3775 grs. of solid matter, in this case, was brought about by 
 0.058 grs. of NaCl, or at the rate of 6.5 mgs. of solid matter to 
 1 mg. of NaCl. Similar calculations for the same suspensions 
 of H 2 soil showed that but 0.14 mgs. of suspended matter was 
 flocculated per mg. of NaCl. Although the NaCl is relatively 
 more effective on the suspended matter of the NaCl + H 2 soil 
 than on that of the control soil, yet, in but one instance, that of 
 the comparatively strong solution of N/50 NaCl, is the influence 
 of the added salt sufficient to flocculate completely the colloidal 
 matter of the NaCl -f- H 2 soil. 
 
 Furthermore, it seemed possible that the transformations 
 manifestly occurring in the physical condition of the soil might 
 also be reflected in some measure in the chemical composition of 
 the variously treated soils and of their colloidal substances. Ac- 
 cordingly these materials were subjected to analysis by the strong 
 hydrochloric-acid digestion method, as recommended by Hilgard 
 for chemical soil studies. However, the results secured up to 
 the present time have not confirmed the above presumption. But 
 to what extent future analytical work will enable us to decipher 
 the relationship of the various factors involved in producing the 
 condition under consideration, is still an open question. 
 
1916] 
 
 Sharp: Soluble Salts and Soil Colloids 
 
 309 
 
 cd cd 
 hrt p p 
 
 p-j * 
 
 p 
 B 
 
 pj 
 ft 
 
 -3 O 
 
 to o 
 
 en 
 
 o o 
 © o 
 o o 
 
 © © 
 co to 
 to I— ' 
 
 © © 
 
 i- 1 © 
 
 OX © 
 
 *> CO 
 
 £1 bO tO 
 
 o o 
 
 p p 
 
 w 
 
 cd 
 
 1-1 
 
 Cb 
 
 CD 
 
 -"■1 © 
 to © 
 
 en 
 
 •o © 
 © © 
 © © 
 
 © © 
 
 h-> © 
 
 © rf^ 
 
 *«. © 
 
 © © 
 
 © © 
 
 to © 
 
 -i co 
 
 > Q < 
 
 CD © p-j M ^ CD 
 
 S» B J P 
 
 Q O 
 
 > > 
 
 b b 
 p p 
 
 en en ►> 
 
 ° °^og 
 crq o S,o 
 
 3 - § 
 
 -i © 
 
 to © 
 
 en 
 
 crq 
 
 
 O © S5 p 85 
 
 ^ © £.Ho 
 oo oo ■ • 
 
 crq 
 
 2-3 
 
 gi&ss 
 
 oo 
 OTQ 
 
 CO 
 
 ■££.3" 
 
 CO g 
 
 O" 
 
 |> o 
 
 
 CD 
 
 CD 
 
 
 CD 
 
 
 CD 
 
 CD 
 
 CD 
 
 p 
 
 P 
 
 ^ 
 
 P 
 
 S^ 
 
 P 
 
 P 
 
 P 
 
 1-1 
 
 i-l 
 
 i-l 
 
 t-i 
 
 i-l 
 
 H 
 
 
 
 C 
 
 
 cr 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 P^ 
 
 
 P-. 
 
 
 
 
 "^ 
 
310 University of California Publications in Agricultural Sciences [Vol. 1 
 
 Chemical Studies of the Calcium and Magnesium Content of the 
 Percolate and the Absorption of Sodium 
 
 Our attention was first directed to the possibility of the ex- 
 istence of a relationship between the two factors mentioned in 
 the heading above, by the following experiment, the outline and 
 results of which appear in Table V, page 309. 
 
 A review of these data discloses the fact that the application 
 of NaCl to a soil increases the calcium and magnesium found in 
 the percolate as compared with the quantities found in the 
 percolate of the normal soil when leached with distilled water. 
 This is in accord with the results of Kullenberg, 9 Van Bem- 
 melen 10 and the work of others, a good list of which is given by 
 Sullivan. 11 To exemplify the above remarks concerning calcium, 
 let us examine the table with respect to that element. The addi- 
 tion of NaCl to the Anaheim sandy loam and Oakley sandy soil 
 has practically tripled and doubled, respectively, the amount of 
 calcium found in the percolates over that of the percolate from 
 the salt-free soil. In the Berkeley adobe and Davis clay-loam soils 
 the NaCl brings about a much more marked increase of calcium 
 in the leachings. It may be remarked also that the results secured 
 with magnesium are quite parallel to those concerning calcium. 
 
 In addition to altering the calcium and magnesium contents 
 of the percolate, the salt treatments materially affected the physi- 
 cal condition of the soil, as shown by a marked retardation of the 
 rate of percolation. All of the soils responded alike to the NaCl 
 treatments, in that they became more impervious and their col- 
 loidal matter exhibited a higher degree of diffusion when sus- 
 pended in water, as shown in the second part of Table V. The 
 degree of imperviousness and diffusion varied, however, with 
 the different soils, and also appeared to be roughly proportional 
 to the increase of calcium and magnesium in the leachings from 
 the soils receiving NaCl. Thus the salt-treated Berkeley and 
 Davis soils, showing notably greater increases in calcium and 
 magnesium in the percolates, were also more highly diffused than 
 
 9 Jahrb. Fortsch. Agri. Chem., vol. 8, p. 15, 1865. 
 loLandw. Vers. Stat., vol. 21, p. 135, 1878. 
 " U. S. Geol. Surv., Bull. 312, 1907. 
 
1916] Sharp : Soluble Salts and Soil Colloids 311 
 
 the other two soils, which were much less affected in both these 
 particulars. Evidently the extent of decomposition of the salt 
 by the soil is a factor in determining the final physical condition 
 thereof, and would certainly seem to indicate that the changes of 
 the physical condition of the soil resulting from salt treatments 
 are considerably more than a mere shifting of the soil particles 
 to new positions, as some investigators would lead us to believe. 
 Moreover, it appears proper to infer that these after-effects of salt 
 treatments can be more properly referred to the salt as a chemical 
 agent than as a physical agent. 
 
 If it be true that the chemical reaction between soils and salt 
 solutions results in a chemically equivalent exchange of bases, as 
 Sullivan 12 states, then it seems proper to assume that the calcium 
 and magnesium in the leachings from soils treated as described 
 above represent approximately the amount of absorbed sodium. 
 At least it would appear that the calcium and magnesium in the 
 solutions, less the quantity normally present in distilled-water 
 extractions, is an index of the absorbed sodium. This statement 
 has received further justification in some solubility studies, the 
 data of which have not been published. Undoubtedly consider- 
 able quantities of sodium have been removed from the salt solu- 
 tions by the soil, which in return has given up calcium and mag- 
 nesium. The absorbed sodium has become so firmly fixed in the 
 soil that no amount of washing can dissolve it; otherwise the 
 wash water after passing through the soil would be slightly alka- 
 line. On the contrary, the first portions of the wash water com- 
 ing through appear to be slightly acid, which is in accord with 
 the work of Sullivan, 13 Parker, 14 and others, but if the washing 
 is continued the leachings eventually become neutral or just per- 
 ceptibly alkaline, as is the case when distilled water is in contact 
 with the normal Davis soil. These remarks, supported with 
 stronger evidence from the rapidly accumulating literature on 
 chemical exchanges between salt solutions and soils or silicates, 
 point to the formation, under the conditions herein reported, of 
 a sodium alumino-silicate compound, or possibly a series of such 
 
 12 Loc. cit., p. 27. 
 
 is Loc. cit., p. 8. 
 
 1 4 Journ. Agric. Eesearch, vol. 1, no. 3, p. 1, 1913. 
 
312 University of California Publications in Agricultural Sciences [Vol. 1 
 
 compounds, which resemble most natural silicates in possessing a 
 comparatively inert chemical nature, and which suffer but slight 
 decomposition when in contact with water or are scarcely more 
 than appreciably soluble in water. This hypothesis receives 
 further confirmation in the contention of R. Gans 15 that artificial 
 aluminum silicates behave like natural zeolitic silicates. More- 
 over, Way 16 clearly recognized the formation of such compounds, 
 and Van Bemmelen 17 is inclined to consider the absorbed salts as 
 fixed in loosely bound chemical combination. 
 
 It was originally considered that the substitution of sodium for 
 calcium and magnesium in the soil was a potent factor in bring- 
 ing about the diffusion of the salt-treated, water-washed soils. 
 However, this first conception attributed the possible effect of 
 the chemical exchange to the double decomposition of the organic 
 salts of calcium and magnesium by the NaCl, which would result 
 in the formation of CaCl 2 and MgCl 2 and organic compounds of 
 sodium. Since these end-products are all soluble in water, a 
 continuous leaching of the soil with that solvent would be likely 
 to deprive the soil of a portion of its organic matter and of a 
 certain amount of calcium and magnesium. If the organic matter 
 herein referred to can be properly catalogued under those elastic 
 and indefinite terms, humus or humates, then it appears reason- 
 able in the light of Schloesing's work 18 to expect that the physical 
 condition of the soil, depleted of such organic matter, would be 
 seriously affected in the direction already suggested. The loss 
 of such well-known flocculating agents as calcium and magnesium 
 might also be reflected in the changed physical aspect of the 
 soil. 
 
 But the cogency of the argument just presented is consider- 
 ably lessened by the results of the experiment to be described 
 next. Twenty-five gram portions of the Davis soil were subjected 
 to the treatments outlined in Table VI. The soils were then 
 allowed to dry, after which 10 grams of each soil were sus- 
 pended in 100 cc. of distilled water for 48 hours before the sus- 
 
 15 Cited from Exp. Sta. Kec, vol. 31, no. 1, p. 22, 1914. 
 ia Journ. R. Agric. Soc, vol. 13, p. 123, 1852. 
 i7 Loc. cit.; also in Landw. Vers. Stat., vol. 35, p. 121, li 
 is Cited from Ililgard, Soils (1911), p. 111. 
 
1916] 
 
 Sharp: Soluble Salts and Soil Colloids 
 
 313 
 
 TABLE VI 
 
 The Effect of Eemoving Humus from the Soil in its Relation 
 Diffusion Appearing in the Soil After Salt and 
 Water Treatments 
 
 to the 
 
 No. Soil 
 
 1 25 grs. (Davis soil) 
 
 2 25 grs. (Davis soil) 
 
 4 25 grs. (Davis soil) 
 
 5 25 grs. (Davis soil) 
 
 6 25 grs. (Davis soil) 
 
 7 25 grs. (Davis soil) 
 
 Treatment 
 Washed with H,,0 
 
 50 cc. 4N HC1 then washed with H 2 
 HC1,* H 2 0, NH 4 OH, and H 2 
 HC1, H 2 0, NaCl, and H 2 
 HC1, H 2 0, NH 4 OH, H 2 0, NaCl, and H 2 
 NILOH 
 
 Weight of 
 
 Suspended 
 
 Matter 
 
 from 
 
 10 grams 
 
 in 100 cc. 
 
 of water 
 
 .0184 
 
 .0118 
 
 .0250 
 
 .1216 
 
 .1842 
 
 .0176 
 
 * The symbols used herein designate the soil treatments on the filter paper 
 with the various substances; as indicated, H 2 applications were generally made 
 between applications of the other materials. Solutions approximating normal 
 strength were generally employed. 
 
 pension above the deposited material was drawn off. An aliquot 
 of this suspension was evaporated to dryness, gently ignited, and 
 weighed to obtain the data of column 4, in Table VI. 
 
 The weight of the suspended matter secured from suspensions 
 of Davis soil which had been subjected to the various treatments 
 gives an index of the extent of the physical effects of such treat- 
 ments. The treatment with II CI, followed by washing until the 
 nitrate was practically free from chlorides, seemed to reduce the 
 quantity of material capable of being held in suspension, yet the 
 soil had no doubt lost a considerable quantity of its original 
 calcium and magnesium content. This fact certainly indicates 
 that the loss of calcium and magnesium from the soil bears but 
 little relation to the diffusion of the soils as a general proposition. 
 
 The soil treatment in the case of No. 4, which simulates the 
 procedure generally employed in humus determinations, with 
 the exception that water is used as a final application, gives a 
 suspension slightly richer in solids than the soil treated with 
 water alone. If, in addition to the treatment just mentioned, 
 NaCl is then added to the humus free soil, 19 the suspension be- 
 comes a very turbid liquid rich in solids in a highly diffused state, 
 
 is The term humus here signifies that portion of the soil's organic 
 matter removed by the treatment in vogue for that purpose. 
 
314 University of California Publications in Agricultural Sciences [Vol. 1 
 
 as shown in the case of No. 6. Evidently the loss of humus and 
 its organic colloids increases, rather than decreases, the amount 
 of material finally found in suspension, for No. 5, receiving HC1 
 and NaCl treatment, yields a suspension containing less colloidal 
 material than does No. 6. In all probability the increase of 
 colloidal material in No. 6 over No. 5 is in no way connected with 
 the loss of organic matter of No. 6, but is more likely to be due 
 to the conditions and treatments involved in extracting the 
 humus. Furthermore, the organic colloids of soils seem to have 
 but little direct relation to the phenomena which appear when 
 certain salts are washed from soils. 
 
 That the deflocculated condition of the salt-treated soils might 
 be due in some measure to the loss of calcium and magnesium 
 from the soil, in a manner somewhat analogous to that outlined 
 by Foerster, 20 appeared as a secondary consideration. The actual 
 loss of calcium and magnesium, however, cannot be considered 
 the sole factor in producing the striking conditions observed. 
 For, if that were the case, the application of acids followed with 
 washing should bring about similar results. On this point War- 
 ington has already noted that normal diffusion of the colloids 
 will reappear in a soil which has once been flocculated by acids 
 when the acids and such salts as they have formed are removed 
 by washing. Our experience with soils treated in the manner 
 just described fails to show any marked increase in colloids, as 
 is the case when the soil is treated with NaCl. The treatment 
 with HC1 followed by leaching with water seems to retard to a 
 notable extent the rate of percolation through the Davis soil. 
 However, if any new colloids are formed by this treatment, or if 
 the original colloidal content of the soil is thoroughly diffused 
 thereby, as the case may be, then these effects are entirely oblit- 
 erated by the subsequent drying of the soil, for a suspension of 
 the dry soil so treated yields no additional colloidal matter. This 
 fact tends to confirm the view that the loss of calcium and mag- 
 nesium is not alone responsible for the diffusion of the soils. 
 
 Furthermore, the writer has shown clearly by a rather simple 
 procedure that a considerable exchange or a direct addition of 
 bases is essentia] for the production of the diffused condition. 
 20 Chem. Ind., vol. 28, no. 24, p. 733, 1905. 
 
1916] Sharp: Soluble Salts and Soil Colloids 315 
 
 A solution of N/10 NaCl was allowed to act upon successive 
 portions of Davis soil until practically no more calcium or mag- 
 nesium was taken up by the solution, so that a solution of 
 chlorides was formed which was quite inert toward this soil. A 
 solution secured in this manner appears to approach the satura- 
 tion point with respect to calcium and magnesium compounds 
 of the soil. This solution remained practically N/10 with respect 
 to chlorine throughout its successive periods of contact with the 
 soil, and at no time was the solution entirely depleted of its 
 sodium content by the interchange with calcium and magnesium. 
 The application to the Davis soil of a solution secured in the 
 fashion described above and followed by washing with water did 
 not result in such pronounced diffusion as was observed upon 
 the application of N/10 NaCl solution followed bj' washing with 
 water. It may therefore be inferred that the chemical exchange 
 of bases plays a significant role in the mechanism by which dif- 
 fusion is produced in soils which have been washed after an 
 addition of NaCl. 
 
 By the various lines of reasoning outlined above, the majority 
 of the more obvious possible causes for the extreme diffusion of 
 the soils under discussion have been completely eliminated or at 
 least reduced to factors of little significance. For this reason, 
 more mature thought on the subject has directed our attention 
 to the absorption of sodium, which has been previously men- 
 tioned, as the keynote to the deteriorated physical condition of 
 the salt-treated, water-washed soils. The substitution of sodium 
 for the calcium, magnesium, or other bases in the silicate complex, 
 or the direct addition of sodium to such a complex, results, ac- 
 cording to our proposed hypothesis, in the formation of new 
 jelly-like colloids, capable of becoming highly diffused when in 
 contact with water. The fact that leaching Na 2 C0 3 , NaOH, 
 NaHC0 3 , NaCl, Na 2 S0 4 , and NaN0 3 from the soil brings about 
 the same appearance of the soil and the same physical manifesta- 
 tions of deflocculation, becomes more comprehensible if the view 
 is accepted that the absorption of sodium and the new compounds 
 formed thereby are, in the main, the factors responsible for the 
 deflocculated condition of the soils so treated. Moreover, these 
 various treatments are accompanied by the appearance of widely 
 
316 University of California Publications in Agricultural Sciences [Vol. 1 
 
 different quantities of calcium and magnesium in the percolate, 
 thus affording additional evidence that the increased solubility 
 and the consequent loss of these elements from the treated soils 
 is not the prime factor in causing the ultimate diffusion of the 
 soil colloids. The facts herein presented concerning the absorp- 
 tion of sodium and the formation of new colloidal substances cor- 
 responds in the main with the idea of unsaturated silicates and 
 absorption phenomena, as developed by Cameron, 21 Harris, 22 and 
 Loew, 23 and others, in their work dealing with acid soils. 
 
 The conception of the causes of the apparent deflocculation, 
 as discussed above, receives tangible substantiation in the results 
 of the following experiment. Twenty grams of Davis soil were 
 subjected to the treatments outlined in Table VII. After wash- 
 ing with water, the soils were dried at room temperature and 
 then 10 grams of each were suspended in 100 cc. of water to 
 secure the figures in the last column. Meanwhile the calcium 
 and magnesium in the total percolates were determined, in the 
 usual fashion. 
 
 TABLE VII 
 
 Effects of Washing Various Sodium Salts from the Davis Soil 
 
 Soil 
 No. (Davis) Treatments 
 
 1 20 grs. H,0 only 
 
 2 20 grs. 100 cc N/10 NaOH fol- 
 
 lowed with water 
 
 3 20 grs. 100 cc N/10 Na 2 CO, fol- 
 
 lowed with water 
 
 4 20 grs. 100 cc. N/10 NaCl fol- 
 
 lowed with water 
 
 5 20 grs. 100 cc. (approx.) N/10 
 
 NaHC0 3 followed with 
 water 
 10 20 grs. 100 cc. N/10 NaNO y fol- 
 lowed with water 
 
 Calcium 
 in Total 
 Percolate 
 
 Magnesium 
 in Total 
 Percolate 
 
 Weight of 
 Suspended 
 Matter 
 from 10 
 grams in 
 100 cc. of 
 
 grs. 
 .0017 
 
 grs. 
 .0014 
 
 water 
 .0348 
 
 .0015 
 
 .0005 
 
 .4860 
 
 .0020 
 
 .0037 
 
 .6500 
 
 .0108 
 
 .0109 
 
 .4720 
 
 .0035 
 
 .0086 
 
 .4880 
 
 .0088 
 
 .0103 
 
 .4340 
 
 It is clear that all of the sodium salts used in the above ex- 
 periment produce approximately the same effects on the colloidal 
 matter of the soil when washed from it, as indicated by the 
 
 ^i The Soil Solution (Chemical Publishing Co.), 1911. 
 22 Journ. Phys. Chem., vol. 28, no. 4, p. 355 ; and Michigan Agric. 
 Exper. Sta. Tech. Bull. 19. 
 
 28 Porto Rico Agric. Exper. Sta., Bull. 13, 1913. 
 
1916] Sharp: Soluble Salts and Soil Colloids 317 
 
 amount of material finally found in suspension. All of the 
 salt-and-water treatments yielded suspensions containing more 
 than ten times the suspended matter found in the water-washed 
 soil. From the evidence reported in the above table it seems 
 proper to infer that neither the nature of the added sodium salt, 
 nor the loss of calcium and magnesium from the soil, have much 
 part in the production of the deflocculated condition noted in the 
 salt-treated, water-washed soils. The data of Table VII are also 
 discussed later in their relation to the possible effects of NaOH 
 and the OH-ion on the physical condition of soils. 
 
 The interchange of ions between the soil silicates and neutral 
 salt solutions like NaCl or Na 2 S0 4 , which seems to result in the 
 formation of new colloidal substances, also necessitates the simul- 
 taneous presence of free acid or of new salts of the free acid as 
 the calcium or magnesium salts, in the solution bathing the soil 
 particles. As previously discussed, both of these conditions have 
 been encountered in the supernatant liquid of a neutral salt 
 solution in contact with soil and also in the first portion of the 
 percolate coming through a soil to which a neutral salt has been 
 added. The constant presence of the free acid and its soluble 
 salts, together with more or less of the salt originally added, oc- 
 casions the maintenance of a flocculated condition of the soil 
 colloids, so that any additional colloidal matter which may have 
 been produced is not sufficiently effective on the physical char- 
 acter of the soil to be easily recognized. 
 
 If it be assumed that new and additional colloidal material 
 is formed in the soil by virtue of the salt-and-water treatments, 
 then it is most likely in existence prior to the washing process, but 
 its effects on the physical condition of the soil are not manifested 
 until the surrounding medium has become sufficiently dilute, with 
 respect to salts, to allow of a more or less complete diffusion of the 
 soil colloids. The addition of the neutral salts either produces 
 new colloidal matter simultaneously with the chemical inter- 
 change of bases, which occurs independently of the w r ashing with 
 water, or in some manner disintegrates the existing colloidal 
 aggregates. The former seems the more plausible, especially in 
 view of certain phases of the work herein presented. In addition 
 we also have evidence that the washing with water alone does not 
 
318 University of California Publications in Agricultural Sciences [Vol. 1 
 
 measureably increase the colloidal content of the soil, nor does 
 it materially affect its physical condition. 
 
 At least one interpretation of these facts seems plausible, 
 namely, that the neutral salts or their ions function as the creative 
 agent whereby diffusible colloidal matter is formed, while the 
 washing with water serves in the entirely separate capacity of 
 removing from the sphere of activity any flocculating agents in 
 the shape of soluble salts which may have been present. 
 
 Certain modifications of the hypothesis just presented must be 
 considered in order to account for the action on soils of the salts 
 which give an alkaline reaction. First of all, the chemical pro- 
 ducts that can possibly be formed when the latter class of salts 
 is allowed to act upon soils are on the whole comparatively in- 
 soluble and hence possess relatively small flocculating powers. 
 Owing to this fact the washing process, which seems to be essen- 
 tial for the appearance of the diffusion in soils treated with 
 neutral salts, is not such an important factor in case of soils 
 treated with salts giving rise to an alkaline reaction. 
 
 Instead of attributing the deflocculation of soils which have 
 received NaOH or alkaline carbonates, wholly to the OH-ion 
 content of the solution or to the alkaline reaction so produced, 
 our present theory, supported by the facts already presented, 
 proposes to account for the diffusion of the soils so treated by the 
 formation of colloidal sodium alumino-silicate complexes under 
 conditions which permit of an immediate deflocculation. Therein 
 lies the difference in the behavior of neutral salts on soils, as 
 compared with that of salts of an alkaline reaction. In all prob- 
 ability somewhat similar compounds are formed in the two cases, 
 but in the first case, with the exception of certain circumstances, 
 the conditions are such as to prevent deflocculation, while in the 
 second case deflocculation is at least permitted and perhaps 
 accentuated. 
 
 A Possible Relation Between the Composition of the Soil 
 
 Solution and the Diffusion Phenomena in Certain 
 
 Salt-Treated, Water-Washed Soils 
 
 The first two hypotheses formulated to explain the diffused 
 
 condition of the salt-treated soils dealt largely with the soil itself 
 
 and with the absorption of sodium by the soil. A third hypothesis 
 
1916] Sharp: Soluble Salts and Soil Colloids 319 
 
 proposes to attribute the diffusion of the soil colloids to changes 
 in the composition of the medium, namely the soil solution. Such 
 a diffused condition of the soil colloids might be brought about 
 by an increase in the OH-ion content of the solution bathing the 
 soil particles. This concept is based on the fact that clay is a 
 negatively charged colloid, and according to the views now held 
 the further addition of negative ions to such a colloidal suspension 
 causes these particles to assume greater charges of like sign, so 
 that they repel each other and thus remain distributed through- 
 out the medium in a stabilized condition. On the other hand, it 
 is held that the introduction of ions bearing an opposite charge 
 to that of the colloidal particles neutralizes the charge associated 
 with the particles, so that they no longer repel each other but 
 gather together in aggregates or noccules. For a more complete 
 discussion and bibliography dealing with these phenomena, the 
 reader is referred to the work of Whitney and Ober. 24 In a more 
 recent review Tolman 25 has advanced a clear conception of colloids 
 and their behavior, which affords us a satisfactory working basis 
 for studies on these substances. According to this author, the 
 surface tension existing between the particles and the surround- 
 ing liquid is the factor which determines the degree of dispersion 
 of the particles in the liquid. Thus systems of zero surface tension 
 are at equilibrium. Those possessing a negative surface tension 
 increase, automatically, their degree of dispersion until the zero 
 value is reached, while those of positive surface tension tend to 
 become less dispersed. Since the surface tension referred to is 
 the resultant of many forces, it may be readily affected in numer- 
 ous w T ays, as by the mechanical process of grinding, by heating, by 
 the addition of electrolytes, or by the passage of an electric cur- 
 rent. Through the application of these considerations we may be 
 able to decipher more clearly and definitely the problems involved 
 in the effects of salts on the physical condition of soils. 
 
 It is possible that deflocculants other than the OH-ion may 
 have been introduced into the soil solution by means of the salt- 
 and- water treatments, but the latter factor obviously appears as 
 the most significant deflocculating agent likely to be present 
 
 24 Journ. Amer. Chem. Soc, vol. 23, p. 842, 1901. 
 
 25 Journ. Amer. Chem. Soc, vol. 35, no. 4, 1913. 
 
320 University of California Publications in Agricultural Sciences [Vol. 1 
 
 under the conditions of the experiments and hence merits first 
 attention. Probably the most striking evidence of the possibility 
 that the OH-ions in the soil solution may be justly deemed the 
 cause of the diffusion of the soils with which this paper is chiefly 
 concerned, lies in the well-known deflocculating effect of dilute 
 solutions of NaOH on the soil colloids. Thus the poor tilth and 
 cultivating qualities of lands impregnated with black alkali 
 (Na,C0 3 ) has been rather vaguely attributed to the OH-ions 
 derived from the hydrolysis of the Na 2 C0 3 . This expression fails 
 to offer any explanation of the mechanism whereby the OIT-ion 
 induces the observed effects, neither does it allow of any possible 
 effect of the Na-ion on the physical properties of the soil. How- 
 ever, if our interpretations be not too far amiss, there are several 
 reasons, not based on theoretical considerations alone, but sub- 
 stantiated by facts, for suspecting that the OH-ion is of much 
 smaller significance than the accompanying Na or other basic ion 
 in the final effect of the chemical compound on the physical condi- 
 tion of the soil. 
 
 The common conception that substances which yield an alka- 
 line reaction on hydrolysis occasion the deflocculation of the soil 
 colloids is frequently accepted without qualification, even by 
 those working with alkali soils, notwithstanding some facts now 
 extant which deny its validity. Thus, as early as 1874 Durham 20 
 pointed out that clay suspensions cleared more rapidly in strong 
 Na 2 C0 3 solutions than in distilled water. Whitney and Straw 27 
 have also shown that NaOH in dilute solutions tends to stabilize 
 suspensions of colloidal silver, china clay, and lampblack, and 
 that emulsions of turpentine, carvene, and carvol are also acted 
 upon in a similar manner. The behavior of these substances also 
 gave evidence that the maximum stability occurred at certain 
 concentrations of NaOH, above which flocculation was produced 
 and below which the effect of the NaOH was not so pronounced. 
 The investigations of Hall and Morison already referred to sub- 
 stantiate, in the main, the previous citations on the point under 
 discussion. Quite recently Maschhaupt 28 has found that even 
 
 26 Chern. News, vol. 30, no. 676, p. 57, 1874. 
 
 27 Journ. Amer. Chem. Soc, vol. 29, p. 325, 1907. 
 
 28 Landw. Vers. Stat., vol. 83, p. 467, 1914. 
 
1916] Sharp : Soluble Salts and Soil Colloids 321 
 
 0.015 N NaOH has a flocculating effect on the colloidal matter 
 of a sandy loam soil, while more dilute solutions of NaOII stabil- 
 ized the diffusible colloidal matter. He further asserts that alkali 
 carbonates act much in the same manner. Despite the preceding 
 evidence, Rohland 29 still contends that the flocculation of clay 
 soils by lime is due to the OH-ion. 
 
 While working along similar lines, the author has noted that 
 a suspension of the Davis clay loam settles out more rapidly in 
 a 0.05 N solution of NaOII than in distilled water, but that solu- 
 tions of greater dilution were stabilizing in their effects. On the 
 other hand, no solution of Na 2 C0 3 30 proved effective as a stabiliz- 
 ing agent as compared with distilled water, while solutions of 
 Na 2 CO a stronger than 0.022 N had a decided flocculating effect. 
 This difference in the behavior of NaOH and Na 2 C0 3 has not come 
 to the writer's attention before in the literature of the subject, 
 and fails to support the widespread teaching that Na 2 C0 3 and 
 salts which hydrolize similarly deflocculate the soil colloids 
 through the agency of the OH-ion. But the most striking fea- 
 ture of the action of NaOH on the soil suspension was the marked 
 resemblance of the soils suspended in that medium to the 
 NaCl + H 2 soil suspended in distilled water. The yield of 
 suspended matter from these two suspensions proved to be of 
 about the same magnitude, as shown in Table II. A cursory con- 
 sideration of this fact obviously supports the contention that the 
 OH-ion content, or the alkalinity of the suspending solution, may 
 be responsible for the diffusion of the soil colloids in the cases 
 under consideration. However, an interesting point to the con- 
 trary lies in the fact that suspensions of the Davis soil in very 
 dilute solutions of NaOH can not be distinguished from sus- 
 pensions of the same soil in distilled water. That is to say, NaOH 
 solutions of a concentration of N/1500 to N/2000 exercised no 
 recognizable effect on the soil colloids. The concentrations here 
 referred to approach the same order of alkalinity as that found 
 in the solution bathing the particles of the highly diffused 
 NaCl -f- H 2 soil. Evidently some other factor than the OH-ion 
 
 29 Landw. Jahrb., vol. 44, no. 3, p. 437, 1913; and Landw. Vers. Stat., 
 vol. 85, nos. 1-2, p. 123, 1914. 
 
 so Baker's analyzed Na 2 C0 3 was used in these experiments except where 
 otherwise noted. 
 
322 University of California Publications in Agricultural Sciences [Vol. 1 
 
 content of the solution is either responsible for the diffused con- 
 dition of the soil colloids, or at least is of material assistance in 
 producing this effect. 
 
 Moreover, it has been demonstrated that the washing out of 
 the excess NaOH does not materially benefit the injured physical 
 condition of the soil, although the alkalinity of the soil solution 
 was thereby reduced to a concentration comparable with that of 
 the dilute solutions to which reference has just been made. 
 Actual determinations of the alkalinity of the final portions of 
 the percolate from soils to which NaOH has been added show 
 only such alkalinity as the percolate from the normal soil. If 
 we accept the view brought forth by Cameron, 31 which seems 
 justifiable, then the percolates from these soils approach, as a 
 limit, the chemical composition of the soil solution and hence 
 we may conclude that the alkalinity of the soil solution is of the 
 same order of magnitude as that of the percolate. The fact that 
 washing the NaOH from the soil is not accompanied by an im- 
 provement in the physical condition of the soil may be fairly 
 interpreted as substantially affirming the view that the OH-ion 
 is of little moment in the diffusion of the NaCl -|- H 2 soil, and 
 possibly has but little connection with the deflocculation of soils 
 to which NaOH has been added. Furthermore, the facts just 
 discussed, taken in conjunction with those considered under the 
 heading of sodium absorption, lend an appearance of reality to 
 the assumption that the sodium, even in the case of direct addi- 
 tion of NaOH to soils, is the principal agent in creating the 
 diffused condition in soils so treated. 
 
 Moreover, the writer has found that NaHC0 3 , Na 2 C0 3 , and 
 NaOH have distinctive effects on the soil colloids. In higher 
 concentrations all three salts prove to be flocculants ; in very weak 
 concentrations, as of the order of N/2000 or less, they seem to 
 resemble distilled water in their behavior toward the soil col- 
 loids. In medium concentrations, that is, those less than 0.05 
 N, NaOH is a deflocculating agent, Na 2 C0 3 acts much like dis- 
 tilled water, while NaHC0 3 seems to be a pronounced flocculant. 
 At least the NaOH and Na 2 C0 3 yield a certain amount of 
 OH-ions, and in the wide range of concentrations employed some 
 
 3i Eighth Intern. Cong. Appl. Chem., vol. 15-16, p. 49, 1912. 
 
1916] Sharp : Soluble Salts and Soil Colloids 323 
 
 point must have been encountered at which the Oil-ion concen- 
 tration of these two was the same or approximately so. Yet the 
 two salts exercised a different influence on the soil colloids at all 
 the concentrations used, with the exception of the highest and 
 lowest concentrations. This fact indicates that some other factor 
 than the OH-ion is effective in determining the degree of dis- 
 persion of the soil colloids. What application this may have to 
 black alkali lands is a question, for under natural conditions it is 
 most likely that all three compounds, NaOH, NaHC0 3 , and 
 Na 2 C0 3 , and their respective ions occur. 
 
 But the most striking feature of this phase of the problem 
 lies in the fact that the washing out of the soluble matter from 
 separate portions of the Davis soil receiving NaOH, Na 2 C0 3 , and 
 NaHC0 2 results in the same way. The soil becomes very im- 
 pervious and diffuses when shaken with distilled water. Thus 
 the cylinder soils receiving Na 2 C0 3 have been exposed to condi- 
 tions permitting the leaching out of the soluble salts, at least 
 from the surface soil, so that they now exhibit the same peculiari- 
 ties as the soils treated analogously in the laboratory. 
 
 The systems so far considered have been largely made up of 
 soils to which various quantities of NaOH, Na 2 C0 3 or NaHC0 3 
 have been added. To ascertain to what extent the facts so gained 
 are applicable to the conditions existing in the laboratory sam- 
 ples of NaCl -f- H 2 soil or to those of the field cylinder soils 
 receiving NaCl and Na 2 S0 4 , necessitates the measurement of the 
 alkalinity obtaining in these soils. We have attempted to secure 
 some information with regard to the quantity of OH-ions or the 
 alkalinity in the solution containing a suspension of the 
 NaCl -f- H 2 soil, but the persistent color of this solution has 
 made it impracticable to employ it directly with the various indi- 
 cators. It would seem, however, from actual titrations made in 
 the usual manner using methyl orange as the indicator, that the 
 alkalinity of the soils in the cylinders receiving NaCl and Na 2 S0 4 
 had been somewhat increased over that of the control soils. But 
 solutions secured from soils treated in the laboratory in a manner 
 similar to the treatment of the field soils failed to verify con- 
 sistently the above observations. At the present time we are en- 
 gaged in a study of the reaction of the soil suspensions by the use 
 
324 University of California Publications in Agricultural Sciences [Vol. 1 
 
 of the hydrogen electrode. The results thus far secured indicate 
 that there is not a sufficient quantity of OH-ions in the NaCl + 
 H 2 soil to produce the degree of deflocculation observed therein. 
 
 If the OH-ion content of the medium is assumed to be the 
 causal agent for the diffusion of the soil colloids, it would appear 
 an equally plausible assumption that the transference of the sus- 
 pending medium from a diffused soil to the original untreated 
 soil would carry with it the deflocculating agent, so that the 
 second soil would register to some degree the physical manifesta- 
 tions of deflocculation. To test the correctness of this hypothesis 
 the following experiment was undertaken. Twenty-five grams of 
 NaCl + H 2 soil were suspended in 250 cc. of water. The col- 
 loidal matter of such a suspension normally remains in a stabil- 
 ized condition for at least three or more weeks, when exposed to 
 laboratory conditions. This suspension was cleared, however, at 
 the expiration of 48 hours by passing it through the Pasteur 
 Chamberland filtering candle under pressure. This was accom- 
 plished with considerable care so as to conform to the conditions 
 found by Briggs 32 to yield the most satisfactory results. The 
 solution secured in the manner described from the NaCl -f- H 2 
 soil was then used as a suspending medium for 10 grams of un- 
 treated Davis soil. A comparison of the suspension thus formed 
 with a suspension of the same soil in distilled water disclosed no 
 recognizable difference. Thus it would appear that a transference 
 of the suspending medium of a diffused NaCl -f- H 2 soil to a 
 second untreated soil does not carry with it the agent causing the 
 deflocculation of the first soil. Hence, any hypothesis ascribing 
 the altered physical condition of the NaCl + H 2 soil to changes 
 in the soil solution seems untenable. 
 
 The value of the results secured by the procedure outlined 
 above may be open to question. To avoid such criticism, the 
 writer made the following experiment, which distinctly shows 
 that the transference of a soluble soil deflocculant from one soil 
 to another by the method employed is susceptible of proof. A 
 N/50 NaOH solution in contact with the Davis soil tends to 
 maintain the soil colloids in a stabilized condition. Such a dif- 
 fused suspension was subjected to the filtering treatment out- 
 
 82 U. S. Dept. Agr., Bur. of Soils, Bull. 19, p. 31, 1902. 
 
1916] Sharp: Soluble Salts and Soil Colloids 325 
 
 lined above and on passing through the Pasteur filter yielded a 
 clarified solution which had not lost its power to deflocculate the 
 Davis soil. The writer is fully aware that conditions might arise 
 wherein such treatments would considerably modify the defloc- 
 culating power of the solute — for example, in case of very dilute 
 solutions where absorption by the soil and filter would be rela- 
 tively large and possibly of sufficient magnitude to markedly 
 diminish the quantity of deflocculant in the filtrate, which would 
 result undoubtedly in a decrease in the deflocculating power of 
 the solution. A similar reduction in deflocculating power of a 
 solution would also be likely to appear if a solution were sub- 
 jected to repetitions of the procedure described above. 
 
 Despite the evidence above some doubt may still be enter- 
 tained as to the absence of significant quantities of OH-ions in 
 the films of water on the immediate surfaces of the colloidal 
 particles, especially in view of the probability that the colloidal 
 matter in the diffused soils consists largely of compounds of the 
 chemical nature of sodium silicate, which hydrolizes to some 
 extent in water and eventually gives rise to OH-ions, thereby 
 lowering the surface tension as described by Tolman 33 and like- 
 wise producing the conditions obtaining in the "natant" col- 
 loids of Hall and Morison. 34 Moreover, the results secured by 
 Briggs 35 on the absorption of alkali hydrates by silica tends to 
 confirm the proposed conception of "natant" colloids. If this 
 be the case, the rate of diffusion of these ions from the films into 
 the more dilute medium would be the factor determining whether 
 it would be possible to transfer a sufficient quantity of these 
 ions, by means of the solution, to be effective on the physical 
 condition of a second, otherwise untreated, soil. Under the condi- 
 tions of the experiment just cited, a contact period of two days 
 was allowed for such diffusion, which would appear to allow 
 ample time therefor. 
 
 Furthermore, as observed by Whitney and Straw 36 and others, 
 the fact that NaOH in certain concentrations tends to stabilize 
 
 33 Loc. cit. 
 
 3-t Loc. cit. 
 
 35 Joum. Phys. Chem., vol. 9, p. 617, 1905. 
 
 se Loc. cit. 
 
326 University of California Publications in Agricultural Sciences [Vol. 1 
 
 colloidal matter of a comparatively inert chemical nature forms 
 the basis of an argument which opposes the conceptions herein 
 presented, and which favors the view that the OH-ion may, after 
 all, play an important role in the diffusion of colloids. Thus, 
 systems composed of colloidal silver, lampblack, or relatively 
 pure organic substances in contact with NaOH probably do not 
 offer opportunity for the direct addition of the sodium or the 
 exchange of ions whereby sodium is taken up, a factor which 
 forms an essential link in the scheme proposed by the writer to 
 explain the effects of the alkali salts on soils. Also the experi- 
 ments of Bliss, cited by Whitney and Straw, 37 give an indica- 
 tion that NaOH may be an effective agent, through the medium 
 of the OH-ion. No attempt is made, however, to deny the com- 
 plete ineffectiveness of the OH-ion, for it undoubtedly has an 
 important influence on the physical condition of colloids, as is 
 the case with many ions. But with regard to the physical condi- 
 tion of the salt -treated Davis soil, it is very evident that the 
 OH-ion is a factor of much less significance than the other ions 
 associated with it. 
 
 The Precipitating Effect of Various Acids and Salts 
 on the Soil Colloids 
 
 There seemed some possibility that the presence of some ion 
 or ions, other than the OH-ion, in the soil solution might be held 
 accountable for the extreme diffusion of the NaCl -(- H 2 soil 
 and soils similarly affected by other salt treatments. Conse- 
 quently an extensive series of test-tube experiments dealing with 
 the effects of various acids and salts on the soil suspension were 
 undertaken. The results of these tests are briefly referred to at 
 this time in their relation to the condition existing in the salt- 
 treated, water-washed soils. 
 
 It was found that N/2000 HC1 and H 2 S0 4 perceptibly floccu- 
 lated the soil colloids, as compared with distilled water sus- 
 pensions of the same soil. To attribute this action wholly to the 
 II-ion of the acid would, in the writer's opinion, be erroneous, 
 for no doubt, salts are immediately formed when the acid comes 
 
 ■■< hoc. cit. 
 
1916] Sharp : Soluble Salts and Soil Colloids 327 
 
 into contact with the soil. Nevertheless, it is significant that 
 acids of the low concentration here employed are more efficient 
 flocculants than their salt solutions of similar concentrations. 
 Hence the nnneutralized acid in very dilute concentrations un- 
 doubtedly exercises a pronounced flocculating power, so that this 
 factor is eliminated as the possible deflocculant in the case of 
 NaCl + H 2 soil. 
 
 Solutions of the chlorides, sulphates, and bicarbonates of 
 calcium, sodium, ammonium, and potassium, were also studied 
 with regard to their effects on the colloidal matter of soils. These 
 salts in solutions ranging in concentration from N/1500 to N/500 
 possessed distinct flocculating powers. Solutions of higher con- 
 centrations were likewise flocculating in their effect, while more 
 dilute solutions behaved similarly to distilled water. Hence the 
 diffused condition of the NaCl + H 2 soil cannot be attributed 
 to the mere dilution of the neutral sodium salts or of the simple 
 salts formed by reactions between the added salts and the soil 
 silicates. As previously indicated, this statement must not be 
 construed to include the systems soil + NaOH, or soil + Na 2 C0 3 , 
 but must be considerably modified to express the facts existing 
 in those cases. 
 
 The Effects of Washing Various Salts from the 
 Soil with Water 
 
 The washing out of KC1 and NH 4 C1 from the Davis soil with 
 water gave essentially the same results as accompanied the re- 
 moval of the sodium salts by the same means. On the other 
 hand, the leaching out of calcium salts seemed to leave the soil 
 colloids in a more flocculated condition than in the normal, un- 
 treated soil, though this effect was by no means so pronounced 
 as when the calcium salts were allowed to remain in contact with 
 the soil particles. According to the ideas herein presented, the 
 effect of a salt solution, either while in contact with the solid 
 particles or after its removal from the soil by washing with dis- 
 tilled water, upon the physical condition of soils or complex 
 silicates of such a character as to admit of chemical exchange of 
 ions, is, in a measure, dependent upon the nature of the chemical 
 
328 University of California Publications in Agricultural Sciences [Vol. 1 
 
 bodies formed by virtue of this interchange of ions. The soluble 
 products, including the salts of calcium, magnesium or other 
 bases, which are formed when salt solutions are allowed to act 
 upon soils, must be considered, if a true conception of the physi- 
 cal effects of the added salts is to be secured. In addition, the 
 individual properties of the new silicate complex, which is 
 formed simultaneously with the exchange of ions, must be recog- 
 nized as bearing significantly on the final effects of the added 
 salt on the physical condition of the soil. Undoubtedly, these 
 factors have a bearing on the effects of salts on the capillary 
 movement of water in soils, and make impossible the direct appli- 
 cation of physical laws concerning surface tension and densities 
 to the phenomena of capillary rise, as Briggs and Lapham 38 have 
 attempted. Instead, the factors discussed above may help to 
 harmonize the observations made by these investigators with the 
 more recent findings of Kossovich. 39 
 
 The addition to soils of KC1, NH 4 C1, NaN0 3 , Na 2 S0 4 or NaCl, 
 and in fact most neutral salts, has been shown by Way, 40 Eich- 
 horn, 41 Henneberg and Stohman, 42 Peters, 43 Van Bemmelen, 44 
 Bradley, 45 Curry and Smith, 46 and others, to result in increasing 
 the quantity of calcium, magnesium and other metallic ions in the 
 solution secured from soils so treated, as compared with the same 
 soils receiving distilled water. Furthermore it has been shown 
 by Hall and Morison, 47 Davis, 48 Patten and Gallagher, 49 Masoni, 50 
 and others that calcium and magnesium salts possess more pro- 
 nounced flocculating powers than the corresponding salts of the 
 alkali metals. This fact may account for the different rates at 
 
 1912 
 
 ss U. S. Dept. Agric, Bur. of Soils, Bull. 19, 1902. 
 so Cited from Exp. Sta. Eec, vol. 25, no. 9, p. 824. 
 
 40 Loc. cit. 
 
 41 Cited from Sullivan, loc. cit., p. 12. 
 
 42 Cited from Sullivan, loc. cit., p. 13. 
 
 « Landw. Vers. Stat., vol. 2, p. 113, 1860. 
 
 44 Loc. cit. 
 
 45 Oregon Agrie. Exper. Sta., Bull. 112, 1912. 
 
 4« New Hampshire A.gric. Exper. Sta., Bull. 170, 1914. 
 
 47 Loc. cit. ■ 
 
 4* U. 8. Dept. Agric, Bur. of Soils, Bull. 82. 
 
 4» U. S. Dept. Agric., Bur. of Soils, Bull. 52, 1908. 
 
 • r >o Abstract in Journ. Chem. Soc. London, vol. 102, no. 597, p. 677, 
 
1916] Sharp : Soluble Salts and Soil Colloids 329 
 
 which the same strength of different salt solutions operates to 
 clarify suspensions from different soils. Moreover, the addition of 
 these salts seems to bring about a decided increase in the colloidal 
 content of the soil, which is obviously manifested when the 
 soluble salts are removed by distilled water. On the other hand, 
 the introduction of calcium salts to a soil does not apparently 
 bring about the production of a colloidal complex, although some 
 interchange of ions occurs. In fact, the addition of CaCl 2 fol- 
 lowed with washing seems to have the opposite effect on the Davis 
 clay-loam soil. For these reasons it seems evident that the in- 
 soluble as well as the soluble products formed by the interaction 
 of salts on soils must be considered as important factors in such 
 practices as prescribe the use of soluble salts on soils. 
 
 It is apparent that at least the Na, K and NH 4 ions form a 
 colloidal substance upon their introduction into the silicate com- 
 plexes of the Davis soil. The Ca-ion, and probably others, form, 
 on the contrary, non-colloidal substances when introduced into 
 the complex silicates of the Davis soil. It is also made clear 
 that the acid ion is of little importance in the phenomena follow- 
 ing the washing out of salts from the soil, by the fact that NaN0 3 , 
 NaCl and Na 2 S0 4 all operate to produce approximately the same 
 results. 
 
 Some General Observations on Other Peculiar Appearances 
 
 Accompanying the Washing Out of Certain 
 
 Salts from Soils 
 
 Stewart 51 in a recent paper attributes the color of the nitre 
 spots to the action of sodium or potassium nitrates on the organic 
 matter of the soil. Bearing somewhat upon this subject, we have 
 noted that the addition of NaCl or Na 2 S0 4 to a soil materially 
 increases the organic matter in the supernatant solution if the 
 depth of color of such a solution can be relied upon as an index 
 of the quantity of organic matter contained therein. But the 
 most striking results in this connection appear when these neutral 
 salts are removed from the soil by washing. The soil itself as- 
 sumes the typical gray color of alkali soils in the field, and the 
 
 si Journ. Amer. Soc. Agron., vol. 6, no. 6, p. 247, 1914. 
 
330 University of California Publications in Agricultural Sciences [Vol. 1 
 
 filtrate takes on a rich, dark brown color, not unlike the usual 
 NH 4 OH humus extract. Frequently the soil when dry is covered 
 with a thin layer of hard, black, organic matter. Indeed the re- 
 semblance between the NaCl + H 2 soil, the water in contact 
 with it, or its percolate, and the natural alkali soils and the water 
 bathing their particles, is so striking as to deceive even those 
 experienced in the handling of ' ' black alkali ' ' soils. 
 
 The increased solubility of the organic matter in the salt- 
 treated, water-Avashed soils, as indicated by the increased depth 
 of the brown color, cannot be due to the presence of the NaCl 
 as a chemical entity, for the color does not appear until after 
 some of the salt has been removed from the soil by washing. There 
 can be no doubt that the NaN0 3 by means of an interchange of 
 ions with the organic salts of calcium or magnesium does affect a 
 greater solubility of organic matter, but this solubility may be 
 greatly increased if the added salt is washed from the soil. This 
 fact may be of assistance in accounting for the color of the nitre 
 spots. 
 
 In agreement with Van Bemmelen 52 and Warington, 53 we have 
 also noted the appearance of fluffy, flocculent colloids passing 
 through the filter paper, when the added salt is practically 
 washed from the soil. These colloids coagulate when in contact 
 with the filtrate containing the soluble salts from the soil. 
 
 One other point which has not received prior attention, but 
 which seems of sufficient importance to be mentioned at this 
 time, is the fact that all of the salt need not be washed from the 
 soil to produce the effects noted. In another series of experiments, 
 which will be reported in detail later, 3200 cc. of a 1.5 per cent 
 solution of NaCl (which was intended to represent the strength 
 of the soil solution when 0.3 per cent of NaCl is added to the soil 
 under optimum water conditions) was passed through 1600 grams 
 of Davis soil. The soil was affected to a greater extent than a 
 soil receiving a similar quantity of NaCl, namely 4.8 grams, or 
 than a third soil receiving a similar quantity of NaCl, which in 
 addition was leached with 3200 cc. of water. The injury to the 
 first soil was greater than to the other soils despite the fact that 
 
 52 hoc. cit. 
 
 53 Loc. cit. 
 
1916] Sharp: Soluble Salts and Soil Colloids 331 
 
 a small quantity of salt remained in contact with the soil. These 
 observations lend further confirmation to the fact previously ob- 
 served, that a second addition of NaCl to a NaCl + H 2 soil 
 does not cause the reversion of the soil to its original condition. 
 
 General Discussion op the Effects Noted and Their 
 
 Application to Soil Studies and the 
 
 Management of Soils 
 
 In the discussion on the possible relation of the calcium and 
 magnesium content of the percolate to the diffused condition of 
 salt-treated, water-washed soils, attention has already been 
 called to the fact that the phenomena, first observed in case of 
 the cylinders containing Davis soil, also appeared in three other 
 soils of widely different types. Evidently the effects observed 
 when added salts are washed from soils by distilled water, are 
 not peculiar to any particular type of soil, but are more or less 
 general in their application. Some soils, as those which do not 
 offer replaceable calcium or magnesium to exchange for sodium 
 or other ions of salt solutions, would probably react to a much 
 less extent than those soils which do react with salt solutions. 
 The degree to which the soil is affected by the salt-and-water 
 treatments is apparently dependent upon the amount of exchange 
 of bases, or the direct addition of bases which occurs. This, of 
 course, varies markedly with different soil types and is probably 
 closely associated with complex silicates of which mention has 
 already been made. This conception conforms, in the main, with 
 the ideas brought out by Knop, 54 Van Bemmelen, 55 and AVaring- 
 ton, 56 with respect to the absorption of salts by soils. Moreover, 
 Kossovich 57 has noted that NaCl exercises a more pronounced 
 effect on capillary rise in clay soils than in sandy soils. 
 
 There can be but little doubt that the effects of salts on the 
 physical and chemical properties of soils has a wide range of 
 application to alkali soils and their management. Moreover, the 
 
 54 Cited from U. S. Dept. Agric, Bur. of Soils, Bull. 52, p. 19. 
 ss Loc. cit. 
 
 56 Loc. cit. 
 
 57 Loc. cit. 
 
332 University of California Publications in Agricultural Sciences [Vol. 1 
 
 peculiar phase of these effects under consideration would seem to 
 have particular significance in that connection. First of all, 
 these soils contain, as a general rule, sufficiently large quantities 
 of salts to be commensurate with those used in the experiments 
 reported above. Secondly, the natural rainfall, irrigation prac- 
 tices, and drainage operations often wash the salts out of the 
 top layers of soil, so that this soil would be exposed to the con- 
 ditions likely to produce the inferior physical qualities attendant 
 upon deflocculation. 
 
 During the late winter and early spring months the writer 
 had occasion to observe the standing water in the depressions of 
 the alkali lands near Fresno, California. To all appearances the 
 underdrainage of these low areas would be practically impossible, 
 owing to the imperviousness of the soil, which upon examination 
 reveals all the characteristics of the diffused salt-treated, water- 
 washed soils. Besides, the gray appearance of the soil when dry, 
 and the striking resemblance of the supernatant liquid to that 
 above the soils in the cylinders, would certainly lead one to be- 
 lieve that even under natural conditions the washing out from 
 and the dilution of the salts in alkali soils has much in common 
 with the artificial production of similar conditions. Heretofore, 
 the depressions and the remarkable imperviousness of their soils 
 have been attributed to the presence of Na 2 C0 3 . 
 
 It would seem that the deflocculation effects exhibited by soils 
 when added salts are washed from them would have particular 
 bearing upon the reclamation of alkali soils by underdrainage. 
 Yet the possible application of this feature to alkali reclamation 
 seems to have been omitted from the literature dealing with this 
 subject. 58 Cameron and Patten 59 observed somewhat similar 
 lowering of the rate of percolation, followed, however, by an 
 increase later as the washing out of the salts progressed. Hare, 60 
 in a recent paper describing tank experiments dealing with the 
 effects of alkali salts on soils and crops, does not note any pos- 
 sible effect of the added salts on percolation through the soils 
 receiving them, but does comment on the observation that Na 2 S0 4 
 
 58 Hilgard, Soils, 1911, chap. 22, and U. S. Dept. Agric, Bur. of Soils, 
 Bull. 35, 1906. 
 
 69 Journ. Amer. Chem. Soc, vol. 28, p. 1639, 1906. 
 oo New Mexico Agric. Exper. Sta., Bull. 88, 1913. 
 
1916] Sharp : Soluble Salts and Soil Colloids 333 
 
 was more difficult to leach from the soil than Na 2 C0 3 or NaCl. 
 Specific information on the physical condition of the drained 
 alkali soils of the various drainage experiments is lacking. Even 
 the recorded observations of a general nature are indefinite and 
 discordant. Thus at Salt Lake, 61 Tempe, 62 and North Yakima, 63 
 the effect of drainage has not appeared to have benefitted the 
 physical condition of the soil, while at Fresno, 64 , Billings, 05 and 
 elsewhere, the drained soil appeared to be superior with respect 
 to physical condition than the undrained soil. Owing to the in- 
 adequacy of these reports in that no definite measurements of 
 alteration in the physical condition of the soil are given, they 
 cannot be accepted as a final expression of the effects of drainage 
 of alkali soils. 
 
 The behavior of the Davis soil in the cylinders and the general 
 trend of the laboratory experiments seem to the writer convincing 
 arguments that the drainage of these cylinder soils in their 
 present condition would be practically, perhaps absolutely, im- 
 possible. Underdrainage, supplemented with powerful flocculat- 
 ing agents, might serve as a feasible plan whereby the excess salt 
 could be removed from such soils. 
 
 That natural alkali soils are not strictly comparable with the 
 cylinder soils to which single salts have been added is patent, 
 and this may account, in a measure, for the general discrepancy 
 between the behavior of certain natural alkali soils and that of 
 alkali soils made artificially by the addition of salts. The wash- 
 ing out of salt mixtures, as in the case of natural alkali soils, may 
 have an entirely different effect than the washing out of a single 
 salt. It has, however, been shown in this laboratory that the 
 washing out of certain mixtures of salts from the Davis soil has 
 resulted in the same way as washing out single sodium salts. On 
 the other hand, the time for the reaction of the salt in the soil 
 
 6iDorsey, C. W., U. S. Dept. Agric. Bur. of Soils, Bull. 43, p. 16, 
 1907, and Bull. 35, p. 181, 1907. 
 
 esDorsey, C. W., U. S. Dept. Agric, Bur. of Soils, Bull. 35, p. 190, 
 1907. 
 
 «s Loc. cit., p. 189. 
 
 64Mackie, W. W., U. S. Dept. Agric, Bur. of Soils, Bull. 42, p. 43, 
 1907. 
 
 65Dorsey, C. W., U. S. Dept. Agric, Bur. of Soils, Bull. 44, p. 18, 
 1902. 
 
334 University of California Publications in Agricultural Sciences [Vol. 1 
 
 being comparatively unlimited in case of the natural soils, it may 
 considerably alter the end-points and end-products and thus 
 introduce a new factor. The constant up-and-down movement 
 of the soluble salts in the soil, the alternate drying and wetting 
 of the soil, and the effects of thermal changes on the soil, may 
 partially obliterate the true effects of washing certain salts from 
 soils. In addition, the relation of calcium, magnesium and other 
 bases to the sodium may be modified by the conditions just men- 
 tioned, so that different equilibria are established. Furthermore, 
 the nature of the soil, especially with respect to the original con- 
 tent of colloidal substances and hydrated silicate complexes, must 
 be taken into account. Thus the effect of draining salts from 
 sandy soils free of colloids may be so small as to pass unrecog- 
 nized, while similar drainage from heavier types of soils would 
 produce marked effects. The colloidal content, too, and the state 
 of diffusion thereof, in our drained alkali lands, may be a direct 
 result of the first natural introduction of the salts into and their 
 subsequent removal from the original soil, the exact physical 
 condition of which may be only surmised. Hence we have no 
 means for securing evidence as to the exact salt effects under 
 natural conditions. 
 
 As suggested by Mayer, 66 the physical effects resulting from 
 washing sea salts from the soil may be as injurious to the growth 
 of plants as the direct toxicity of the salt itself. Our experience 
 with the cylinders serves to confirm this idea, for olespite the 
 fact that the top foot of soil is almost free from salts, it seems 
 impossible to secure a satisfactory growth of plants. A pot 
 experiment, to be reported upon later, also gives substantial 
 proof that the deflocculated soil, comparatively free from salt, 
 offers a far less congenial home for barley plants, at least during 
 the early stages of growth, than the same soil containing as high 
 as 0.3 per cent of NaCl. 
 
 The inability of the plants to make a satisfactory growth in 
 the highly deflocculated soils appears, from a cursory considera- 
 tion of the conditions, to be due to the lack of available moisture. 
 Although the deflocculated soils may seem to be moist, yet the 
 quantity of water which the plants have at their disposal is 
 
 66 hoc. Cit. 
 
1916] Sharp: Soluble Salts and Soil Colloids 335 
 
 probably insufficient to promote normal growth. At least this 
 condition is suggested by the moisture equivalents of the defloc- 
 culated NaCl + H 2 soil as determined with the centrifugal 
 machine developed by Briggs. In addition to the low availability 
 of the moisture in such soils as an inhibiting factor, the rate of 
 movement of water through such soils must also be considered 
 in its relation to the growth of plants. From all appearances, it 
 seems justifiable to predict that the movement of water through 
 the highly diffused soils is so slow as to fail to resupply the area 
 occupied by the roots from the more moist soil layers lying 
 adjacent to the root area. 
 
 The observed effects of washing salts from soils may also have 
 some bearing upon fertilizer practices and the results on crop 
 plants secured therefrom. The salts used as fertilizers undoubt- 
 edly effect certain modifications of the physical condition of 
 soils. It is also true that the washing out from the soil of at 
 least some of these salts brings a far greater change in the 
 physical properties of the soil than that accompanying the ap- 
 plication of the salts. Probably no instance w r ith respect to this 
 feature, as a phase of soil management, is so patent as that con- 
 cerning the use of NaN0 3 . Hall 67 comments upon the inadvis- 
 ability of applying this material under certain conditions be- 
 cause of the deflocculating effect of the alkalinity due to residual 
 sodium carbonate. Warington 08 also notes the deflocculating 
 effect of NaN0 3 and refers to it as being particularly noticeable 
 after heavy rains. More recently McGeorge 69 has observed a 
 marked retardation of percolation through Hawaiian soils receiv- 
 ing applications of NaN0 3 , but attributes this effect to some 
 reaction between the added salt and the organic matter of the 
 soil. Undoubtedly the application of NaN0 3 to soils under cer- 
 tain circumstances has resulted in a deterioration of the physical 
 condition of the soil, although the NaN0 3 , as a salt, is in itself 
 a flocculating agent. 
 
 The writer has produced marked deflocculation in the Davis 
 soil by applying NaN0 3 and subsequently washing the soluble 
 
 67 The Soil, p. 252, 1910. 
 
 es Loc. cit. 
 
 eo Hawaiian Agric. Exper. Sta, Bull. 35, 1914. 
 
336 University of California Publications in Agricultural Sciences [Vol. 1 
 
 salts from the soil with water. The use of the N0 3 radical by 
 plants or bacteria, thereby leaving the sodium, could not have 
 occurred to any extent in this case, for the whole process re- 
 quired no longer than two hours, and was carried out in a filter. 
 The filtrate at no period in the washing process showed sufficient 
 alkalinity to account for the deflocculation. Hence the former 
 conception that residual Na 2 C0 3 causes the deflocculation seems 
 untenable, and some such hypothesis as has been advanced in 
 this paper must be adopted to explain the unfavorable physical 
 condition frequently existing in soils receiving applications of 
 NaN0 3 . 
 
 With respect to fertilizer salts in general, Hall 70 calls attention 
 to the deflocculation following the use of neutral salts on soils, 
 but believes the effect is due to alkalinity arising from the ab- 
 sorbed base. Hessler 71 has noted the increase in coherence of the 
 soil particles when NaCl, NaN0 3 and kainite have been applied 
 to soils. Hoffman, 72 however, could not detect any difference in 
 the interior surface of soils due to fertilizer applications of the 
 usual magnitude. It is of interest to note that no attempt is 
 made to correlate the inferior physical condition of salt-treated 
 soils with the process of leaching the salts from them. 
 
 However, potassium and ammonium salts are effective in the 
 same direction as sodium salts when washed from the soil, al- 
 though they are generally considered flocculating agents when 
 in contact with certain colloidal particles. Hence, to be fully 
 comprehensive and expressive of the whole truth, studies on fer- 
 tilizer effects should involve not only the conditions under which 
 the salt or salts are present, but also those conditions which not 
 infrequently arise in nature, whereby the soluble salts are re- 
 moved from the soil by processes analogous to washing the soil 
 with water. 
 
 The nature of the clay colloids and the cause of the relatively 
 high degree of deflocculation which they assume when suspended 
 in water have long been subjects of much conjecture and much 
 
 70 7,0c. cit. 
 
 7i Cited from Exper. Sta. Rec, vol. 31, no. 2, p. 123, 1914. 
 
 72 Landw. Vers. Stat., vol. 85, nos. 1-2, p. 123, 1914. 
 
1916] Sharp : Soluble Salts and Soil Colloids 337 
 
 debate. Hilgard 73 has consistently maintained that the power 
 of such colloids to remain in a stabilized condition is not to be 
 attributed solely to the fineness of division of such particles, but 
 that other factors may be responsible for that condition. In 
 agreement therewith, our experience, which has been presented 
 in this communication, has led us to believe that the chemical 
 nature of the body itself and that of the medium, determine in a 
 large measure the condition in which the colloid may exist. 
 
 The extent to which the factor considered in this paper may 
 be applicable to agricultural practice can be better surmised than 
 asserted at the present writing, but it seems highly possible that 
 the modifications of the physical condition of the soil, due to 
 washing out the soluble salts, under circumstances involving 
 fertilizer applications or natural alkali soils, will be reflected in 
 the inferior tilling qualities of the soil, in the increased resistance 
 offered by such soils to root penetration, in the lack of air space 
 and air movements, in the deflocculated soil, and in the moisture 
 and temperature relations of such diffused soils. The movement 
 of moisture in soils by surface tension and osmotic pressure or 
 under gravitational attraction, appears to be particularly de- 
 pendent upon the degree of deflocculation of the soil colloids. 
 
 Besides affecting the physical condition of the soil, the leach- 
 ing out of soluble neutral salts from soils is, as has been previously 
 shown, frequently accompanied by notable quantities of calcium 
 and magnesium in the percolate. Thereby the soil sustains a con- 
 siderable loss of calcium and magnesium, which may in the course 
 of time be of sufficient magnitude actually to deplete the available 
 supply of these plant-food elements in the soil. The bacterial 
 flora and bacterial activity of soils subjected to the treatments 
 outlined above are liable to be considerably modified and to all 
 appearances in a harmful direction. Thus all the factors of soil 
 fertility are likely to be measureably affected through the pro- 
 cess of washing salts from soils ; consequently the crop-producing 
 power of the salt-treated, water-washed soil is apt to be con- 
 siderablv modified. 
 
 Soils, chap. VI. 
 
338 University of California Publications in Agricultural Sciences [Vol. 1 
 
 Summary 
 
 1. The Davis claj^-loam soil to which surface applications of 
 solutions of NaCI, Na 2 S0 4 and Na 2 C0 3 had been made became 
 very impervious to water, difficult to cultivate, and manifested 
 the characteristics of a high degree of diffusion, although these 
 salts have been shown to exercise flocculating powers on sus- 
 pensions of this soil. 
 
 2. The salt-treated soils referred to were in cylinders exposed 
 to natural conditions. 
 
 3. Examination of the soils showed that the salts had moved 
 downward into the lower layers of soil and that only the surface 
 soil had been affected in the direction described. 
 
 4. The deflocculated condition resulting from adding certain 
 salts to and subsequently washing them from soils can be repro- 
 duced in the laboratory. 
 
 5. The deflocculation of soils treated in the manner described 
 above is intimately associated with the leaching of the NaCI and 
 Na 2 S0 4 down into the lower layers of soil by water. In the case 
 of Na 2 C0 3 the leaching process is not so essential for the diffusion 
 of the soil colloids. 
 
 6. The addition of NaCI, Na 2 S0 4 and Na 2 C0 3 to the Davis 
 soil when followed with applications of water was particularly 
 effective in diminishing the rate of percolation through the soil 
 so treated. 
 
 7. NaCI and Na 2 S0 4 in constant contact with the Davis soil 
 increased the rate of percolation, except when a comparatively 
 dilute solution of NaCI was slowly passed through the soil for a 
 considerable period of time. 
 
 8. The Davis soil treated with NaCI, NaOH, Na 2 C0 3 and other 
 salts, followed by leaching with water, yields a suspension in 
 water containing approximately ten times as much solid matter 
 as the same soil washed with water only. A real diffusion in such 
 salt-treated soils seems evident. 
 
 9. The soil once diffused by washing out added NaCI requires 
 considerably more salt to completely flocculate it, than does the 
 water- washed soil. Likewise the injured physical condition of 
 such soils is not readily repaired by a second addition of NaCI. 
 
1916] Sharp: Soluble Salts and Soil Colloids :;:;!» 
 
 10. The portion of the organic matter of the soil known as 
 humus has little or no connection with the appearance of dif- 
 fusion in salt-treated, water-washed soils. 
 
 11. The diffusion in soils treated as described above seems to 
 be closely associated with the direct addition of sodium to, or 
 with the absorption of sodium by the soil, thereby producing a 
 new silicate complex of a colloidal character in the soil. 
 
 12. This silicate complex is formed simultaneously with the 
 interchange of ions occurring between the salt and the soil. 
 
 13. The washing process serves, in the case of neutral salts. 
 to remove flocculating agents. 
 
 14. The loss of calcium and magnesium from the soil bears 
 little or no relation to the flocculation appearing in salt-treated, 
 water-washed soils except in so far as it may be a measure of the 
 absorbed sodium. 
 
 15. The presence of the OH-ion does not seem to be an essen- 
 tial factor in the diffusion of salt-treated, water-washed soils. 
 
 16. Na 2 C0 3 and NaOH produce markedly different effects on 
 suspensions of the Davis soil. 
 
 17. The acid ion of the salt is not an important factor in the 
 deflocculation phenomena following the washing out of salts 
 from soils. 
 
 18. Sodium, potassium, and ammonium seem to produce the 
 colloidal silicate complex when salts of these metals are applied to 
 soils, while calcium does not. 
 
 19. Dilute solutions of acids and salts possess flocculating 
 powers on suspensions of the Davis soil. 
 
 20. It is not essential in every case to wash all of the salt 
 out in order to bring about diffusion. 
 
 21. The facts presented are discussed in their relation to the 
 reclamation and management of alkali lands, and in their appli- 
 cation to the use of soluble salts as fertilizers. 
 
 The writer desires to express his thanks to Professor Charles 
 B. Lipman for many valuable suggestions and for his critical 
 reading of the manuscript, and to L. E. Bailey for his timely 
 assistance in the analytical work. 
 
 Transmitted November 16, 1915.