The World’s Largest Open Access Agricultural & Applied Economics Digital Library

This document is discoverable and free to researchers across the

globe due to the work of AgEcon Search.

Help ensure our sustainability.

Give to AgEcon Search

AgEcon Search
http://ageconsearch.umn.edu
aesearch@umn.edu

Papers downloaded from AgEcon Search may be used for non-commercial purposes and personal study only.
No other use, including posting to another Internet site, is permitted without permission from the copyright
owner (not AgEcon Search), or as allowed under the provisions of Fair Use, U.S. Copyright Act, Title 17 U.S.C.
 ~ 12.8 IIE~ 1.0 E:.i il~ 2.8 IIIII~
1.0 Ii.::
~W ­
U
~W
~
2.2
II.i lid
~ W ~W
II.i
II.i
:r W :r W
1.1 ......
.. M
1.1 ...
... ....
M

II I
111111.25 111111.4 111111.6 111111.25 111111.4 111111.6

MICROCOPY RESOLUTION TEST CHART MICROCOPY RESOLUTioN TEST CHART

t<ATIONAL BUREAU or 51 ANDARDS·J963-!< NATIONAL BUREAU OF STANDARDS-J963-A
 . ~~~.~..==~~==
TECHwcA L BUL~TtNNQ. 170 ~ JANUARY , 1930

UNITED STATFS DEPARTMENT OF AGRlCULJ:URE

WASffiNGTON, D. C.

A PIPETTE MET HOD OF MECHANICAL

. ANALYSIS OF SOILS BASED iON IM­
PRO VED DISP ERS ION PROCEDURE
By L. B. OLMSTE AD, Phys£cist, LYLE T. ALEXAN DER, J1mior
Physicist, an? H. E..
MIDDLE TQN, Associate PhyS'icist, Dilliiion (Jf Boil
Chemistry and PhY81C8, Botl
.bivestigat£ons, Bureau of Chemistry and Soils 1

CONTEN TS

Introduct ion ___ .. ____________________ ---___._

Dispersioll_....._•• • ____...... ...__..... ..._,
Hydrogen peroxide trel1tment ____.______
Pllg~
3
I
Outline oC metflOd__________ •_______________ •
General statement _______. _.______ ••___ ._
Pa~:
14
3 Preparatio n and dispersion oC sample___
Acid treatment.. ___ •__ •__ ., ••___________
1Vashing• • __ ._.. _______..____.. _._. ___ ._
5 Separation into size classes._______________ 16
15
Dispersing agents_______ •_____•__. _., .••• \I Calculation lIC results____________________ .18
9 Apparatu s._.__ ... __ •______• _____________.___ III
~~e:~[g~!~~:.~~s_:~::::::::::::::::::::: :::
Separntio a into size classes_ ••_. __••••_.__ ._.. U ~W~::;~ro-citeiC:::::::::::::::::::::::::::: i'l
12
Size clusses employed .__ •••_••__.. ..... __ \2
Method oC 5'lparation•••_•••••_._. __ •• ___ 13
Sediment ation velocity and particle size_ \3

INTRODUCTION
In all metho ds of mecha nical soil analysis some separa
based on the well-Imown .fact that the smaller particles tions in
are
water
suspensions fail more slowly than larger ones_ In elutria tion metho
the smaller particles ar-e borne upwar d by a ~ising water colunm ds
while larger parti~les fall. In the early work of the Bm:eau .of S9.i!~
the Osborne (20)~ beaker metho d was used to make .silt and clar'
separa tions. Whitn ey (5) appare ntly was the first to use the centn­
fuge to increas e the settlin g velocities of soil particles. The centri­
fuge metho d was then worked out indepe ndentl y by Hopki ns (12)
and Snyde r (27). In 1904 the Bureau of Soils published its cen.tri­
fuge metho d (5), which with modifications (9) has been used until
recent ly., .
The centrifuge metho d was much more rapid than the beaker
metho d and gave better dispersion than elutria tion me.tho In
this metho d the silt and clay were separa ted from the sands byds.dec!1!l­
.
tations from a shalring bottle, and the silt was separa ted from the
I The authors desire to express their appreciati
due especially to G. D, DoUmnn oC the Universiton to those who have assisted in thls work. Thanks nre
y oC Oalifornla for sofl samples Cor uualysls; to W. H. Fry
Cor microscopical examinati on oC soli separates; to.Hubert Lakin Mrs. Dorotha M. Darnell, and Wllllrun
Swvens oC the soli physics luborntory for assistance in the analytical
)Jail of soils investigations' instrumen t shop for ,lid in the design andwork; and to C. J. Crawley and .J. N.
1 Reference is made hy italic numbers In parenthese
for the construction.oC ap~tus.
s to Liternture cited, p. 21.
82395°- 29-1

GIFT

JAN S 0 1930
2 T]lCHNICALBUL."LETIN 170, U. S. DEPT. OF AGRICULTURE

clay by centrifuging. With trained men checking their silt and clay
separations with a microscope micrometer, the analyses generally ,
were reproducible. Dispersion was obtained by shaking a 5-gm. 1
soil sample with 150 c. c. of distilled water containing 1 c. c. of con~ ;
c6D.trated amm@nia. As many as 40 centrifugings were necessary ill
sO,me cases to give a complete separation. In such cases rubbing the
wet sample ,,,ith a l'ubber policeman hastened the operation. Later
(6) rubber balls with lead centers were put into the shaking bottles
.to assist in cleaning colloid coatings from sand grains and in br(~al;ing
colloidnl aggregates.
I About 1919, when attempts were made to determine guantiLatively
the amount of colloid in soils, it was ioundthat dispersion was not
always so complete as had previously been believed. Davis and
:Middleton (8) then studied a number of methods of dispersing soils \
preparatory to their sepa~ation ~y the centrif~lge ..lnethod. They
concluded that some chenucal aSSIstance was necessary for complete
dispersion and recommended treatment with 0.1 N HOI for one hour
followed by wl1.fY>.mg in the centrifuge to remeJve excess acid and
exchanged ions. 'fhe sample was .then sha:ken seven hours in 0.05 N
NaOH. .
The introduction of the pipette method developed independently
by Robinson.in 'Wales (23), Krauss in Germa.ny (16), and Jennings, .
'fhomas, and Gardner in the United States (13), necessitated a more;
stable and more complete dispersion than was reguired in the centri- ;
fuge method. A study of dispersion was undertaken by the first
commission (soil physics) of the International Society of Soil Science, j
that resulted in the adoption at the First International Soil Science
Oongress held in Washington in 1927 of two methods for mechanical \
soil analysis (19). One method, called thepmotical method, secured i
dispersion by shaking wi$ ammonia, boiling, and rubbin~. The
other method.. designed to secure complete dispersion, prescnbed (1)
treatment with hydrogen peroxide to remove organic matter, (2)
treatment with 0.2 N HOI followed by washing to remove alkaline,
earth carbonates and adsorbed bases, and (3) treatment with ammonia 1
or sodium carbonate to secure dispersion. The latter method will·be '
referred to hereafter as the international method. No attempt was
made to fL~the 8ize classes, but an arbitrary relilihlonship between
sedimentation velocity and particle size was established.
These two methods, one intended toO give partial and the other
complete dispersion, represent two diverse purposes of mechanical
analysis. In so-called practical methods such as the first the endeavor
is mad£'. to avoid removing the coating from soil particles or breaking
apart slightly cemented sand ~ains or aggregate.s of finer material
wbich would not be resolved ill tillage operations (;1' under severe
weather conditions in the field. Hilgard (11, p. 88) objects to the,
use of acids to disintegrate compound particles, particularly.in cal- ~
careous sands. In prnctical methods it is difficult to detennine just!
when a proper degree of dispersion has been reached. For example,. I
Bouyoucos (4), in an effort' to secure an analysis of th~ ultimate
natural soil structure, does not give the sample as rigorous treatment
as the international practical method proposes.
A mechanical analysis bllsed on complete disporsion does not give i
so true a picture of field Rtructure anel Itssociated physical properties'
as does the practical method/ but it does give information more vaI­
 A PIPETT E METHO D OF ANALY SIS OF BOILS
3
uable. for mos.t pm'poses, It gives the amoun t of coarser materi
strippe d of adhering material., the Itmount of colloid, and usually al
approx imate amolm t of organic matter . Since the colloid the seat the
of most soil chemical reiwtions and governs or strong ly isinfluen ces
i~.
most of the physical properties, the quanti tative determ ination of the
£nest soil fraction is the most impor tant single determ ination in
mechanical analysis. In the case of concretions of irreversiblea
colloid almate :dal having adsorptive prope:rties less than that of dis­
persed colloid but lllore than that of solid mine:ral grains, there is
somE' question as to how far dispersion should be ctUTied.
This bulleti n reports an investi gation of metho ds of obtain
persion of 'soil material, with special reference t.othe effect ing of
dis­
acid
treatm ent and the use of allmline dispersing agents. A
the ]~ellloval of organic matter from manga nese soils ismetho include
d for
11 procedure for mechanical soil analysis, based on the most satis­ d.
factory of severnl method s of soil dispersion investi gated, is here
outlined.. This is the metho d now used in the Bureau of Ohemistry
and Soils. This method. is designed to secure complete dispersion
of soil mate:rial below 2 nun. in diamet er after organic matter and
water-soluble Ill/twi nl have been removed. Sieves are used to make
all separations except those at 5 II- and 2 11-, where the pipette is used.
DISPERSION
Good dispe:rsion is the key to accura te mechanical analysis. By
comple te dispersion is meant the removal of colloid coatings on sand
and silt grains and the separa tion of aggregates into single grains
groups smalle:r than 2 II- equiva lent diameter. The proble m of dis­or
persion is, (1) to obtain, and (2,) to mainta in suspensions free from
aggregation. The chemical and mechanical r~ids to dispersion con­
sidered are treat.ment with (1) hydrogen pe:rliixide, (2) hydroc
acid, (3) water, !ilid (4) dispersing agents 3; electrodialysis;hloric
and
shaking.
HYDROG EN PEROXI DE TREATM ENT
In the analysis of soils high in organic matter G. W. Robinson (22)
iotmd that dispers ion was aided by the remov al of organic matter
with hydrog en pe:roxide. He ascribes this effect to the remov al of
materi al which cements soil particles into aggregates strong ly enough
to resist ordina ry dispersion method s. Althou gh it is not
sary to remov e ol'ganic matter in orde:r to secure dispersion neces­of the
mine:ral portion of the soil (14), its presence interferes with th'3
separa tion of the materi al into size classes. For this reaaon all sam­
ples for routine analysis in this labora tory are treated with 6 per cent
H20~1 in accord ance with the intel'llational metho d.
When manga nese dioxide is presen t in sufficient quanti ty
possible to remov e the organic matter by the usual treatm ent.it isSince
not .
manganese dioxide is the only form of manganese which catalyt ically
decomposes hydrog en peroxide so rapidly that the organic matter
very slowly attacke d, the obvious l'emedy is to conver t it into some is
other man~anese compound. W. O. Robinson (25) found
small fractlo n of 1 per cent of manga nese dioxide was presen that if a
be. de.stl·oyed by evapor ating the sample with a very s~lall texcess
it .could
of
, AmmQDiUlU hydroxide , sodiuUl hydroxide, sodium carbonate , and
BodillDl oxalate.
4 TECHNICAL BULLETIN 170, U.S.DEPT. OF AGRlCUI.TURE

oxalic acid. In this investigation, if manganese dioxide was known

to be present or its presence was indicated by the violence of the
reaction, acetic acid was added, followed by hydrogen peroxide, the
purpose being to convert the manganese dimdde into the manganous
form which does not interfere with the decomposition of organic
matter by hydrogen peroxide.
The effect.of ace.tic acid trea.tment was determined on duplicate
samples of Wabash silt loam iTOm Nebraska. The samples contained
about :3 per cent organic matter and 0.11 per cent manganese, calcu­
lated as manganous oxide (MnO). The rate of decomposition of
hydrogen peroxide indicated that the manganese is not in the form of
manganese dioxide. To 10-gm. samples of this soil were added various
amOtmts of finely precipitated manganese dioxide and acetic acid.
The amount of carbon dioxide evolved upon the addition of 20 c. c.
or more of 6 per cent hydrogen peroxide is shown in Table 1. The
amount of carbon dioxide evolved decreased rapidly with the addition
of ma! ganese dioxide. Th.e -presence of 0.1 per cent (10 mg.) of
manganese dioxide did not ~,'ter~ere with the decomposition oi organic
matter whe!l 10 mg. of acetil~ aCld was used. Blanks rtm by treating
acetic acid with hydrogen 1>ero:\.-ide in the presence of manganese
dioxide showed practically no evolution of 00 2 due to o:\.-idation of
the acid.
TABLE I.-Effect of manganese dioxide and acetic acid on the liberation of carbon
dioxide jr!Jm soil treated with hydrogen peroxide

Carbon
Soil dioxide'
evolved

};Ig. c. c. illIg.
Wabnshsiltloatn (lOg11l.) .............................. o o 20 lOS.S
Do................................................. 10 o 20 55.3
Do................................................. 50 o 20 17.5
Do............. '., ................................ 100 o 20 4.0
Do................................................. o 10 20 112.2
Do•.•••... , ••.•••.•.•.......•.••••..•••.••••••••••• 10 10 20 112.7
Do................................................. 50 20 20 10.5
Do................................................. 50 20 40 74.1
Do................................................. 50 20 60 174.2

TABLE 2.-0rganic multer removed by the Robinson method from manganese soil
treated with acetic acid

·Mangllnese. . 'd Hydrogen Organic

Soil dioxide ",CO.t-IC IlCI peroxide 11Iatter
lidded added (6 ~dc;ft) removed

.Mg. .Mg. C.C . Per cellt

Wllb(\.~h silt 108111 (10 gm) ____ •••••••••••••••••••••••.•• 100 o 40 0.81
Do••••••••••••••••••••.••••••••.•••..•••••••••••••• 100 100 40 1.38
Do••.•••••••••••••••••••••••••••••••••••••••••••••• 100 o 60 1.111
Do................................................. 100 100 60 1.73
Do.•.•••.••.••.••••••••••••••••••••••••••••••••.••• 1003 o 80 1.48
Do••••••.••.••••••••••.•••••.••••••.••••••••••••••• 100 100 SO 1.85
Do................................................. o o ('l 3.10

(Il(I<'l
Do................................................ 10 110 2•. 95
Do................................................. 100 1100 !l.85
Do................................................. 200 '200 2.03

IA slight e:<cess over the CCHliv81ent Ilmount of uc.Jtic Ilcid was used.
IAn excess of hyd~ogcJl peroxide.
 ~ PJ;PETl,'EMljl'l'HODOF ANALYSIS bF SOILS 5
"'ben soil if, treated with hydrogen peroxide the organic matter
ilither is destroyed with the evolution of carbon dioxide or is brought
into solution. Table 1 shows only the amount of organic carbon
oxidized to 002• Table 2 gives the total organic matter removed by
both mellns when the same soil, containing amounts of manganese
dio~ide, is treated with different amounts of hydrogen peroxide.
The quantity of acetic acid added was .It slight excess over the
molecuhtr eq1:rivalent of the manganese dioxide present. The
addition of acetic .acid is qmte effective ;when the soil contains 1 per i
cent of manganese dioxide, 40 c. c. of hydrogen peroxide removing 4
nearly as much organic matter when acid was present as was removed
by 80 c. c. of peroxide without acid. In the case of the last four
,samples hydl''Ogen peroxid(~ was added until decomposition of the
Ol'~anic matter was complete. The sample was then washed and
dned and the nmount of organic matter determined by loss in weight,
according to the method of W. O. Robinson (25), who removed 3.03
per cent of org~Lnic mattE\!' from this soil. The results show that
organic matter can be completely removed in the presence of man­
ganese dioxide up to 2 per cent. It is possible that soils containing
both manganese dioxide and lime carbonates would not respond so
readily to acetic acid and hydrogen peroxide treatment. The amount
of peroxide required increases with the amOlmt of mnnganese dioxide
present. The removal of organic. mat.ter from the sample containing
2 per cent manganese dioxide reqmred the use of excessive quantities
of h:rdrogen peroxide. The removal of organic matteI' is not essential
to th6 method described in this bulletin and may well be omitted
in the case of soils lrigh in manganese dioxide and low in organic
matter.
In some cases hydrogen peroxidp, decomposes the organic matter
slowly and incompletely. In one case, l'ecently not.ed, hydrogen
peroxide attacked the organic matter only slightly. After thp -p.ddi­
tioli of 1 or 2 mg. of manganese dioxide theol'ganic matter ,was
quickly attacked and apparently completely decomposed.
ACID TREATMENT

In the pipette method of mechanical analysis it is necessary to

prepare soil suspensions in such mannel' that they may be kept in a
sedinlentation chamber for several days without coagulation into
flocs larger than 2p. equivalent diameter. For this purpose treat­
ment of the soil sample with hydrochloric acid has been found helpful
because it dissolves alkaline earth carbona-iies and l'emoves "exchange
bases II which later might cause flocculation of the suspension. There
are, however, some objections to the use of acid treatment.
In the international method the soil, after decomposition of organic
matter, is treated with enough hydrochlol'ic abid to decomfose the
carbonates and still leave 250 c. c. of solution of 0.2 N HO. Such
rigorous treatment dissolves sesqmoxides of iron and aluminum and
some silica, plthou~h it may not dissolve all the carbonates. An
exarnple of this is gIven by Bodman (3). When It sample of Monte­
zuma clay adobe subsoil from Oaijfornia was boiled with 0,1 N HOI,
the amotmt of acid consumed was equivalent to 9.08 per cent calcium
carbonate; but if treated with 0.2 N HOI at l'oom temperature in
accordance with the international method, acid eqmvalent to only­
1.27 per cent of calcium carbonate- was consumed, and part of this
 consumption must be ascribed to the solution oJO.81percenti<)f
.sesquiopdes and silica. The amount of carbonates in this soil, de­
, ternrined by the official method (1 p. 22-24), was 8.28 per cent. The
mechanical analysis oj this material is given later in Table 5.
Even though carbonates might be removed completely -and ,easilYt
such removal is Wldesirable in sO.me soils, such as glacialso!!s contain",
ing large amounts of .limestoneor dolomite, because it \vould be
likely to change soil teA-tural classifications. Other soils -may have
relatively permanent grains or crystals of precipitated carbonate
whose presence should be expressed in the results of mechawcal analy­
sis. :An example of this is the Goliad :fine sandy loam subsoil from
!,exas. Ten-grRID: samples. of this material '.I{er~treated W,ith vary..
mg amounts of a~ld, followmg the reguli1r peroXide treatment. The.
washed residues were shaken with 5 c. c.of normal sodium carbonate
and. analyzed by the pipette method described later. The results
are shown in Table 3. An examination of the materialbyWillia.m
H. Fry, soil petrographer of this division, showed.tliatthe carbonate
material, amounting to about 38 percent, was present largely in
the form of small calcite crystals.
TABLE a.-Effect of acid treatment on the res.ul/8 of mechanicalanalY8t8 1 of Goliail
fine 8andy·loam (35/0 48 inches) ,

Silt 0.05 Olay Colloid .Hydrogen

'. Treatment Sands 2.0 to 0.005 <0.005 <0.002 peroxide
to 0.05 mm. mm. mm. and solu·
mm. tion.1oss

Per cent Pti cent Per cent Per cent Per cent
No acid••••••••••.•..•.•••••••••••••••••••• 36.0 38. 6 24.2 18. 3 0.2
2 c. c. N HOslI.O•• " ••••••" •••••••••___•• 36.6 37.9 24.1 18.1 1.4­
10 C.c. N HOl____• __ •.• __ ._••.••.••••.__ ._ 36.6 3~. 7 .23.. 3 5.3
Oarbonates remQved+lO c. c. N HOl_••• _ 31.4 9.6 19.9 19.4\
18.0 &9.0
,No acid, dispersed with sodium oxalate._._ 37.0 34.3 28.4 21.0 .2'

I JtesuIts in all mccba;tlical8llll1ysis tables are on oven-dry basis (105° 0.), and silt is reported by diller- .!
ence. '

For some special purposes, such as dam construction or other

engineering uses, an analysis based on vigorous acid treatment might
be preferred for a soil such as the Goliad, although an analysis with­
out acid treatment togeth!3l" with a determination of total carbonates
would be preferable. Analyses by both me.thods are required to
furnish a complete description of the soil. .
These analyses (Table 3) are presented primarily to showt1;tat
acid treatment may alter the amount of material in the different
size classes enough to change markedly the~xtural classific8.tion.
Incidentally they indicate that acid treatmen;t is not essential to
good dispersion, a matter which will now be examined more 'fully. '
Since ltappears that treatment with 0.2 N HOI is notalwayseffec­
tive in removing the carbonates from soil and in some cases pre­
sents a distorted picture of the soil, it seemed desirable to ascertain
whether complete dispersion can be obtained without its use.
Table 4 gives the results of mechanical analysis of a number of i
soils with and without acid treatment. .All samples (10 gm.) were.
:first treated with hydrogen peroxide. Those given acid treatment
were allowed to stand overnight in HOI solution of ·sufficient strength
to leave 150. c. c. of 0..2 N HOl after the carbonate content is removed.
 )j,',
\ The'd!iference in'dry wei~htof the sru;nple bef?re and after pretrea~
moot 'lS reported 'as solution loss. This comprISes the 10ssoi 'orgaruc
matter, sesquioxides (R20s),silica,.1ime carbop.a~es,or:other dissolved
,material. All sample!) were peptized by shaking wIth 150 c. ,c. of
w!lte.,1," conta~~ 10 c. c.of 0.5 N N a2C20 4 •• These~a.mples emhr8;cea
wIde range of soil types and are representatI~e of soil samples receIved
:for analysis. .
TABLE 4.:-EJlect of miid pre/reatmenton 'Tl4echanicdl.analysisof soils

Sam­ Treat-
Sand Snt Co.l Sclu I
CaCO. 2.0 to. O.o.'Hoqlay JOid- tlcn- Ih03
ple SI}U t:fll6 and source Depth
No.. ment 0.05 0.1\05 <51' <2 I' loss loss
mm. nun.
1---------1·--- ---.----- --' ----:--.-.
Per Per Per Per Per Per
Incht3 .\cld____
Perct1lt cent ctnt cent cent cent ce'rlt
425250 Buston fine SIludy ,loam po to. 70__ 0 56.2 13.9 29.7 25. 6 0.1 0.2
(MissWslppl) •
_____dD___.-________________ ___de ____
A25250 No. ncld_ 0 56.~ 14:.1 29.3 27.0 .1 .00:
3(}tO!l ____.dc_____________________ 1 to 2____
Isabcllalolltn (Michigan) _ ___dc____ Acld____ 0 58.0 30.5 6.8 4.2 4.5 .5
3(}t041 Ncacid_
Acld ____ 0 57.0 29.0 8.7 3.8 4:.3 .J.
2437200 Co.w!lle clay loam (Scuth 5 to. 15___ 0 17.7 .26.2 55.1 45. 5 1.2 .6
Carclina).
_____dc_____________________ ___dc____
2437200 Ncacid_ 0 17.5 25.2 56.3 48.8 1.0 .05
376337 :PlUten very fine sandy 24 to. 44_ Acld ____ 5.6 52.1 27.0 11.4 0.4 8.6 1.3
IcaIn (Nebraska).
_____dc___________________ __ dc ____
376337 No. acld_ 5.6 55. 1 24:.5 29..0 16.1 .'4 •. 01
----,-- Hcustcn black c.1ay Otel2__ ...\ cld ___ 1,11.7 '4.1 25.0 56.3 50.7 14. 7 1.7
___(Texas).
dc _____________________ __ dc_____
No. acid_ 111.7 5.9 28.5 63.8 53.8 1.0 .1
-_
,,- .. ..... ___
llndcb (Sudan·Qoziroy__ • __
dc ____________________
dc ____
___dc____
.\cld____
Ncncid_
5.0
5.9
6.3
8.4
.18.0
14.7
69.7
75. 9
62.4
64.6
6.0 2.2
1•. 0 .94
.Rendzina (Cejc) __________ ___dc_____
--_ ....- .. _____de--___________________ ___dc ____ Acid ____ 5.1 26.4 30.2 25.7 11.6 ,2.'5
:Pcdsol (Zdnr) _____________ ___de ____ No. ncid_ 5.1 31.8\
33.0 29.0 34. a 28..8 , 3.6 .,6
-..---- .. _____dc__________________ Acld____ . 0 25. 7 35.4 37.1 39.0 1.6 1.3
___dc____ ~cncid_ 0 25.6 36.4 37.8 30.2 .2 .05

I Dcterminaticn by Rcblnsen and Holmes (£ff). I>

The last three sets of analyses are from samples stibmittedfor

analysis to some 20 laboratories by a committee of the iirst commis­
sion of the Intemational Society of Soil Science.. The unpubliShed
results collected by ,this committee and furnished. us by R. O..E.
Davis give the amo.lllit of organic matter in the badob, l'endzina,and
podsol as. 1 per ~!.lent! 4.6 per cen(),and .0.4 per ~ent, re~pecthely.
The average (unpublished) result collected by thIS COmmlttee from
the l'eporting laboratories for the percentage .2p.clay (colloid), when
hydrogen peroxide and hydrochloric-iacid trea.tment was used, was
60.5 for the badob, 26.7 for the renclzina, and .28.6 for ;thepodsol.
Joseph and Snow (14-), using their method which consists of sho.lcing
with sodium carbonate, puddling, and decanting by the beaker
method .without the use of peroxide or acid, obtained 63.3 per cent,
29 per cent, and 27 per cent for the aIhountof 2,p. clay in the badob,
rendzina, and podsol, respectively. W.O. Robinson (25) reports '3.6
per cent organic matter in the Ho.ustonblackclay, determined by
the hydrogen peroxide method, and 4.2 per cent determined by com­
bustion. It is appareht from the results in Table 4 that the organic
matter was not all removed 'by peroxide treatment from either the
Houston ,black clay or the rendzina.
;Recently Bodman (3) published analyses of five well-selected Cali­
fornia soils, using two methods. At request of the writers Doctor
Bodman furnished this laboratory with samples of 'these soils. The
results of the analyses, together \with those of Bodman, are shown in.
•
8 TECHN,IC;AL BULLETIN 170, U. S. DEPT. <OF AGRICULTUEE

Table 5. The Montezuma clay adobe soil has a "high 'content ,of
organic ma.tter, and its subsoil has a high content of lime. The
Fresno sandy lo~m has a ¥gh content of sodi~ carbonate; the ~ken
clay loam subsoil has a high percentage of Iron; and the last ,IS ,the
difficultly dispersible hardpan horizon of the San ,Joaquin sandy loam.
TABLE5.-il,fechanical analyses of soils by c01!J.parative methods

Sand Silt 0.05 Clay

Soli typo Depth Method I 2 to 0.05 to 0.005 <0.005 Colloid'
.mm. mm. .mm.
- - - --- ---, ---
Inchu Per cent Per cent ;Per Ct1lt Per cent ;
Montezuma
Uu____
ell.), adooo ___ ._•.•.__ ••••__ 0 to 12.__ ._
__ •_____ • _____ • _______ • ______ do_____ _ A 25.0 32. 5 36.0 28.1
.~_. ;
Do____ . ________________________________do_____ 13 24.0 30.3 3l.I 15.0
Do___..., __ •• ___•.__________________ 32 to 40__._
~

C
A
23.8
30.8
32.8
34.1
40.2
32.3
au
26.2 1
B~=::::::=::::::::~:::::::=::::::::=
5nu\ly IOIUl._________ • __ .________ =J?r~:::::
B 2i.5 43.0 26.7 14.4
Q 27.3 34.6 37.9 31.7
.Fresno 0 to"12____ :A- 61.S 20.3 7.5 3.7
Do.___ . _. _____________________•________do __ • ___ B, 7.,8
Do____ . _____•_______________ •__________do _____ _ 64.2 26.4 3.7
Q 62. 3 28.8 8.2 5.5
Aiken clay loam ________________________ 12 to 14___ _ :A- 6.5 34.8 58.3 45.7
Do_____________ •• _________•_. ____ .. ____do _____ _
Do____,. ________________________________do _____ _ 13 7.0 35.2 57.3 43.4
San Joa!(ui,il sandy loam ________________ 25 to 30__.. .c 8.2 3.1.3 58.1 48.6
Do__ •_________________.---_-- ___ ._.____do- ___ ._ A 70.7 18.0 10.7 4.9
B 58.4 25,5 15.,5 6.4
DO __ t, ___ .. ___ .. _______ .. _,._ .... _.. ~ __ ".,",_" ... __ do_ .. ", .. ~ .. C 66.1 15.7 17.9 11.8

1 Method A, hydrogen poroxide·hydrochloric acid pretreatment, shaken with ammonia. Method B,

rubbed with am.monla, shakon with ammonia; separates treated with hydrogen peroxide. Metbod Q,
hydrogen-poro:dde pretreatment, shaken with sodium oml~te.
2 Methods A and ,B, colloid pipetted ut 11'. Method Q, colloid pipetted at 2 1'.
~ ~.

The results given under methods A and B are those ,published by

Bodman, and lmder method 0 those obtained in this laboratory.
Method A ,is the international method designed to give complete
dispersion. ':' In methods B and 0 hydrochloric-acid treatment is
omitted, hydrogen peroxide being used ,as the last step in the analysis
of the former and the first step in that of the latter. In method B
the sample, wet with a dilute ammoniacal solution, is rubbed with It
rubber pestle for 10 minutes before being shaken. In methods A
and B 5-gm. samples dispersed with ammonia and in method 0 10-gm.
samples dispersed with SOdilUll oxalate are used.
That the larger amount of colloid obtained by method 0 is not
entirely due to the separation being made at 2 p. instead of 1 J.I. is
shown by the fact that the amount of clay is also greater by this
method.
The results'in Tables 4 and 5, typical of hundreds of analyses of
soils developed from a wide range of materials and under a wide range
·of climatic conditions, are presented ,to show that in no case is there
any indication that acid treatment is necessary to dispersion. Joseph
and Snow (14) reach the same conclusion in the case oithe heavy
alkaline soils of Sudan. In their beal\:er (or centrifuge) method in
which ~e repeated pud~g and dec~tation of suspensions produces
a washing effect, flocculatJ.on merely mcreasesthe number of decant a­
tions required for complete separation into size classes. In the pipette
method, however, it is necessary ior the material to remain in the
sec.linlentationcylinder wijihout coagulation for at least :t~o or ·three
days. Duplicate pipettiDgs of th~ same suspensio~ taken a week
apart show no evidence of flocculation, even though acid pretreatment
was omitted. Since acid treatment appears to serve no essential
 .A: PU'l!iTTEME'l'HODOF ANAL'YSISOFSOILS
<l
purpose, and may be objectionable, it is omitted in the method iii
r01ltineanalysis.
WASWNG

• After the soil sample has been pretreated it is washed to remove

soluble organic matter, dissolved lime, l'eplaced ions (especially
bivalent and trivalent cations), hydrogen peroxide, and excess acid if
hydroehIoric 01' acetic acid is used. Experiments by Wiegner (29)
indicate that the l'emoval of electrolytes by thorough washing is
necessary in order to obtain a high degree of dispersion by any subse­
quent treatment. Soils from humid regions are low in soluble salts
and usually do not require much washing to prevent flocculation.in.
the disperElion cylinder. For soils subject to low rainfall, however,
thorough washing is essentinl to good dispersion.
As washing proceeds the soil] flocculated by acid treatme:i;""Lt, gl:adu­
ally cleflocculates again, slmving the filtering rate. Occasionally, how­
ever, a soil such as the Cecil of Georgia does not deflocculate 011
washing. For this reason it was chosen for the followin~ test. Eight
duplicate IO-gill. samples were given hydrogen peroxid.e and hydro­
chloric-acid treatment. The samples were then washed 1,2, 4, 5,6,
7, 9, and 15 times, respectively, by the method described later, and
the tllll0unt of clay was determined by the pipette method. There
was no correlation between the munber of washings and the percentage
of day, the l'esults agreeing within the limits of experimental error.
The Cecil, It soil developed in It humid clin!ate, is low in soluble salts
and does not require much washing. The l'outine treatment is 6
,Yashings of about 125 c. c. each, removing the solution completely
each time. The entire operation l'e<Quires about one day. When acid
treatment is omitted only a :fraction of 1 pe'!: cent of soluble matter is I.
removed by \'"fishing, and no chemical analysis of it is made.
DISPERSING AGENl'S

The dispersing agents most used .in mechanical analysis are

ammonia, sodium hyro'oxide, and sodium carbonate. In the centri­
fuge method, in w.1J.ich the soil after dispersion is l'epeatedly :washed
and the pm'centl1geof clay determined by difference, any of the dis­
persing agents may be used. When ammonia is used in the pipette
method, the amotmt of ammonia held by the dried colloid is somewhat
uncertain. . The pipetted solutions are alkaline and consequently
absorb carbon dioxide from the air, giving an uncertain increase of
weight of the aliquot. If soditim oxalate is used as a dispersing agent
this difficulty disappears.
Ammonium and sodium hydroxides give good dispersion when
calcitun and magnesium carbonates have been removed from the soil.
T.his removall'equires a thorough acid pretreatment and washing of
the sample. If calcium carbonate remains in the soil after pretreat­
ment, sodium carbonate is a better dispersing agent than either sodium
or ammonium hyro'oxidebecause the carbonate decreases the solu­
bility of the calcium carbonate while the hydroxides may produce large
amounts of the :flocculating calcium ions. Sodium oxalate is even
better as a dispersing agent ,than sodium carbonate. Unlike the
hydroxides, sodium oxalate frees no all\:aline earth hyro'oxides asa
result of chemical equilibriums established. Also, the solubility of the
oxalates of calcium and magnesium is ,l;lluch Iower than that of the
82395"-29----"2
 .'.l1ECml'IC.AlTj) SU,LLE'l'l,W 170,:'U•.S. :J>EJ>'l'. ',QFAGRICULTURE

corresponding .carbonates. 'The ·o:x:ala.tedoesnot i:eact with ,carhon .

·.dioxide 1'rom the :air. ,
Surface layers ,of ,thre,e soil types were treated with hydrogen
"peroxide but no 'acid, :andshaken with ilie four dispers41g age'nts,
.equimdlecular concentratioIis(rOC. c. ~) being used.rhisisasmall~r
am.ount oJ ,ammonia than is generally used but·is·an adeguatequan.,
tity of the.other,deflocculents. Table 6 ,gives the·quautities of 5 p. ,and
2 ,i,J,clay obtained 'by the pipette method. 'Dhe total lime computed
:ss ,GaOOa was leSS than 2 jl,er cent for the Garrington and Wabash
soils and 13.9 ,per cent for the HOllston. The i:eSults indicate that
am,moniv,m .andsodiumhYdroxides arenotsuitabledeflocculei1ts.!for
soils which have not had acid preti'eatment,and, that, ·of ,the other two, .
sodium oxalate gives consistently ,higher results. The Mtivity,ofthe
oxalate ion may cause this increased dispersion. :Oxal\C acid is ·often
llsed to clean minerals for petrographic,examinatio;n.Tests .are soon
to be made upon the efficiency of lithium,oxaiateasadeflocculent.
T~LE 6.-Yield of clay obtained with variou8 dispersing agents·1

,Percentage of cl1\Y obtained from, treatment with­

• Soil type and source ,~onia

"
I Sodhll!l hy-
dro:ude
Sodium cnr-
bpnnte
Sodium oxn,­
late'

51' 21' 51' I 21' 51' 21' 51' ~I'

;
Wabasb slit loam (Nehraska)._______
Per cent Per
30.7 24.1
cent~II'per
33.2
ctnt Per cent Per cent P~r cent Per cent
28.51 33.1 28.3 34.5 20.6
Houston black clay (Texas) ____• __.__ 38.4 26.4 44. 5 30.6 61.3 53. 5 63.8 53.8
CnrrIngtonloam (IoWa) __ ' ___ "_"_'1 7.8 6,4 24.9 21.8 23.41 19.9 I 24.7 22.9

l'\'vernge of duplicnte d~terminntioD.~.

After it was determined that sodium oxa1ategave the/best ,disper­

sion oi soils containing. . calcium carbonateJ tests were made on soils
containing gypsum, which is much more soluble in water. Table 7
gives the results ·of comparative analyses ,of sO.me special samples of
gypsum soils unclassified as to soil type. Nos. 28U2and 28743 ,are
from South Dakota, Nos ..28655 and 28668 from Montana, No. 29735
from Hungary,and No. 5218 from KRI;lsas. Analyses care made iby
the international method, dispersing with sodium carbonate and ,by
our methodJ described later, in which acid treatrofill.t is omitted and
sodium oxala.te is used~s the disp'ersing agent~ . With :theexception
of sample N 0.5218J which contams only 1 Jler cent oigypsum, the
.routine washing was insufficient toremdve ·all .thegypsum ,from the
soils not given acid treatment. The l'outin,eamount,of sodium oxalate
(10 c.c. N /2) was sufficient to remove from solution the calcium ions
formed during the three days in which the .suspension stood in the
sedimentation chamber. There was D.oindic.ation of flocculation.
The .amount of clay and colloid obtained was, in each case, greater
by method 2 thRJ,l by method 1. It is likely that in soils containing
larger quantities of a soluble calcium salt, such as gypsum, mOre
thor~ugh lYashing, .acid treatmen~, .orelectr~dialysis may ,be required
iordisperslOn unless largerquantitlesoisodiumoxalate are used.
 A PIPETTE METaODOF W,ALYSIS0F .SOILS H
TABLE 7.-Mechanical analysis of soils containing gy.psun~

Sample Sand. 2 SUt.0.05 Clay' 1 Colloid Solution

No~
Gypsuml Method to 0.05 to 0.005 loss
mm. mm. <51' <21'
------ '~--. - - -- - - - - - - - -
Rer cellI Percent Per cent Per cent Per cellt j>er.ce.nl
28742 2. 52 '1 37.9 14. 5 42.9 40.1 4.7
28742 2.02 '2 37.3 16.8 43.8 40.8 2.0
28743 2.28 1 55.1 10.6 26.3 24. 9 8.2
28743 2.28 2 56.2 10.9 31. 7 26.3 1.3
28655 1.61 1 14.8 21.1 51.2 38.9 12.9
.28655' 1.61 ~ 15.5 23.5 60.0 44.3 1.2
.28668 5.29 1 31.6 31.0 31.0 27.1 0.5
28668 5.29 2 31.7 30.7 32.8 28.6 4.8
29735 2.82 2 3.3 46.1 48.2 38.4 2.4
5218 1.04 2 3.0 34.0 60.9 52. 7 1.6
I
I Computed lIS CnSO•.2H,0 from analysis of total S03.
'Hydrogen peroxide and .hydrochloric acid treatment; dispe~d with sodium car.bonate.
3 Hydrogen peroxide troutment only; dispersed with soditim Ol<alat~.

Occasionally a sample flocqulates .in the sedimentation chambers.

This may occur with any of the dispersing agents and with or without
a mild acid treatment. The difficulty may -be due to the presence of
some soluble soil material or of .an excess or deficiency of dispersing
aO'ent. It is impossible to determine accurately from inspection
wblch is the cause. No investigation of the electrokinetic potentials
of these suspensions has oeen made, but a simple !remedy has been
found which USu.ally restores dispersion. A Pasteur-Chamberland
suction filter is placed in the suspension, and the liquid volume is
reduced fro~ 1.,000c. c. t? some smaller definite volume, ~ay -200o:r.
250 c. c. DIstilled water IS then added to make up the ongmal vol­
ume. The suspension is thoroughly s.tirredand allowed ,to stand.. If
flocculation again takes place measured amounts of defioccultmt. ~are
added until the suspension is stable. This procedure leaves some
uncertainty as to the weight ·of dispersing agent to be deducted from
the pipetted aliquots. It is assumed that the filtrate l'emovedhas
the same concen.tration as the material left in the cylinder.
,ELECTRODIALYSIS

Another means of effecting dispersion is electrodialysis. For sam­

ples dispersed in this way the Mattson cell (17) was used. By this '
method carbonates and "exchange bases "as well:asthe chlorides,
sulphates, and phosphates can be removed; sothatJ in addition to
giving an excellent pretreatment for dispersion., electrodialysis en­
ables one to detennine the amount of "exchange -bases "in the soil.
The advantage of electrodialysis over hydrochloric acid treatment is
that it does not introduce .intothe soil a foreign ion which must be
washed out later. No washing is required after electrodialysis. The
method requires co.nsiderableextra equipment, extra time, .additional
transfers of the sample with attendant possibilities of loss, and, where
determination of exchangeable bases ~s not desired, yields no results
that can not more easily be obtained by acid treatment.
SHAKIN'G

After the hydrogen peroxide treatment, the sample is wash~d,

dried at 105 0 C., and weighed. It is this weight which is later maae
the basis. for the determination of the percentage of material in each
size class when the results are used for textural classification. The
.;d,iied sample ls !transfetred tci a.250c. c.. nm:sing hottle, thedispers:o
ing ,agent is ,a,dded1 :a;rrd ,the ma,terialis shake.n ovetn,igh,t.The.
shaker has a hOI~ontal moyement ,of 10 .cm.. ,.and ,a ,l'ateof 120 'com'­
;pIe.te; osciJlat,ions _;p~r -n$lute: .. ,Josepn;,and Snow{1~)find ·~h(\"Ul.'S.' ....
shaking to he suffiCIent. In :theexpernnents of DaYls .and MIddleton, .
(8) -the mat.eri!l1 was shaken '1 hours." Since thes.amples.,used in the
laboratory were fitstfuiedat 105 0 ,0. it is possible th,at10;ngershaking
is re.qJ,ti!ed. Although 16110urs 'is long~' shaking ,than ,necessar,v, ~tis
co;rrvellient ,to le.t.the shakerrlJ.ll ,0'Vel'D,lght.The ;amo.\mto£gn;n.ding
,of soil grams is negligible.. ..' .
:SEPARATION INTO ,SIZE ,CLASSES
SIZE -CLASSES 'EM.PLOYED

Early in the ')York of the'fonner Bureau of Soils, th~l's0ils~ple ;(or

mechaniCal analysis 'WaS sElparated into :the ifollowi,ng seven size
class.es: '
;Diameter•.zrun. iConv.entlonal.name
2;0 ,to 1.0__ .,._..:, __ ----.-~_.;-.-----.,)------'-:..-----~ Fine gra~el.
l;O,to 0.5 ___ .. __________________________ ------ ,CQarse sand;
0.5 to 0.25 ______ :.____________________________ MediUm sand.
0.2,5 to '0.10____....: _____.: __.--._______ .- ________ :.__ 'Fine sand.,
'0.;10 to 0.05 ________,_______ '.._______ .._._________ Yery:finesand;
0;05 ,to,O;OQ5__ .,. __ - ______ ----.-~ ____ .;__ :.~. ______ :Silt.
<O;005 _________ -' _________:__ ...;-~-...;.,.-.,.--_;.._,...;- __ ,Clay.
~alyses hased ,o~' the 'percent~e iby;eight 'J>! ;these siz,e ,dasse~;~re .
.used ,bY.the lbureau for ;the ,textUT.itl ,clasSificl),tion .ofso,ils Cr).: '. '';Dhe'
bureau'.s r,ecords contain sQm,e50;000,an,alyses ~b.~ed ,.on.,these siz~
classes, and fox ,this reasOn no change !has :beenma,de in fJreseelasses
,except tOl'eport .alsothe amount'of material \below'2 'p. e:ffectiv,e
diameter. This ~s ,called the i";c.olloid"£r:action.Thf)5 'fL,:clayand
iihe2 ,p. "colloid'''' both contain ,all,particleslbelow- ,diameters ,pi 5 !II.
and '~ p;, respectiv:ely. Usually less tthan 10 ;perc,en.t Qf the day is .'
between '5p.and 2,p.. in :e:ffectivedj.ame~r, '. Most ,of the ma;teria;l .
below:2 p. has·.adiameter ,of .only a'fewtenths of ,a ipicron ,or less, so', .'.
Jihis fraction may :be. regarde.das ,essentially.c,01l6idal inchar.Rcter..~.'
Xnadditio.:n .to ,theeig.ht~ize. "cIa,sses; this '.la.bo.;ratory ~e'p.,on.s .the.llo~s
by :pretreatment of orga;ruc. matter and soluble mateoal. ·w;l:l,en :a~}ld,
treatment :is omitted the .solution loss xepres,Elnts roughly ,the ,amount .
of organic matter. (See'irable4.) Neither it:4esolution !loss p.or ;the
amount of 2p. colloid is 'uf:jed inte';Ktur:aiclas~ifica:tion. 'They'l~re J'e,.
ported fOl: informationoniy..· . ;'. · . i

. The number .ofsand £rac,tions may he .g:re.ater ,than;neces!jary, ,but

,these separations .are.quickly :and,easily mad~, ''iJ;'hedecimtilsystem< .'
.of .grading, with separations madeah :2;0, ,0.2,0:.0'2, ,and 'Ol002 ,rm;n:j .
wouldno.t furnish an .adequa.te numhero~size:clllcsses' lor .theex:ten­
sive syste.mof texturaJ. Cla~sification of soils uS.ed l,h;the VnitedSta:tes,
.Although the .coarseclay is .siltyin ,character landiihe :larger silt might'
,be classed as very fine sand it now ,seems ·desirable itO ,main.tain ,the
preseJ;lt size classification. The simplest changes ·wo.1,1ld :be to. omit
the 5 p. separation, place the material between 5 .•JL .and2p. in the silt
class, and also omit ,the l. inm.s.and separation, ,throwing ,the mate­
rial :betwe.en .2 mIil. and ,0:.5 mm..intp ;one class.. . '
(I A PIPE'ri':E) METRO]) OF ANALTSIS OF SOILS

,1\lETIlQDOFSEPARAl',lON

,After ,the. sample has ·he.en shaken withdispersingl),g.ent it is wet

sievedthrQugh a Tyler 300-meshphQf:lphor brQ~etwilled ;w;ire-clbth .
screen, ;which passes material finer ,than 50 JJ. equiyalentdiamet,Eil'..
AU 'the clay and t~e diner PQ)'tiQn Qfthe .silt is ;\v:l,tshe.dthrQl!ghthe .
sien into a 1-liter sedimc;mtation ohamber. The' siltremain,ing: with
the sands is dry .sifted thrQughanQther300-mesh sieve added to the
nest Qf .sand sieves. .. .
Tnef:ledimentatiQnchamber cQntainsQnly fine material w:i,th lQW
settling velQ~ities.where the ;accurac;v {)f pipettes8;IDr>1i?g is !gre,atest.
L The suspenSlOn 'ill the sedimentatiQn chamher 18 stIrred, .and.an
aliquQtis quickly taken wltha 25 c. c.LQwyautQmaticpipette.. After'·
the suspensiQn has stoQd 77 minutes (at 20° c.), ,tindagain .-after
st&nding 8 hQUl"S,aliquQts are taken at a depth ·Qf 10 cm .., the mate,rial
in the chamber being stirredthQrQughlyeach time at ,the beginning
Qf the sedimentation periQd. . • .
SEDIMENTA'IlON VELO.CI:!,Y ANDPAR'IlCL}) SJZJ~

All decantatiQn 0,1' pipette methQds Qfdeterlnining~e .classes

depend Qn the relatiQn Qf settling ll'elQcity to. p8J:tkle size---:andthis
,relatiQnship is nQt accurately knmv:n.The rela#Qnsbip mQst uS,eg is
Stokes's, formula (28) fQr the fallQ!.a SQlid sphere in ,a visc.ousflUid,
2 (d-iJ,')
V=-9 9 7J. r
:vhere V i~ the settling velQcityQf I!- sph;ere .0,£ radJ.us r ;and Aen~~tyd
ill a medium Qf denSlty d' and VlSCQSlty 1/, ,under a grawtational
acCeleratiQn Qr centrifugal fQree,g. ....
. The internatiQnal m.ethQda.rbitrarily nxed 8 hQmsas ,the ,time
\', fQr a particle 2 JJ. in diameter to £all 10 em. in waterat2000. .Thls
J is the va,lue determ:iriedexperhnentallyby Atterherg (2), $:lthoughne
,.dQesnQt state the teniperature. If StQkes'.slaw'h,blds, asitis:assumed
in determining settling velocities in other sizes, ;then;fuingthe ;re­
·latiQn between settling velQcity and. particle size isequiv:alent' to ,
fixing the density Qf tIie SQil particles. Assuming d' = 1, takin~.the
values Qf the v:l8cQsitYQf water given in the internationalcntical
tables,and the .abQve values fQr V and T,StQkes's formula ~ves2,61
for the density of the soil particle:2 .JJ. in diameter.Th~tthis set,t-ling
velocity may nQtbe very a,ceutate is indicated by the iactthat'
densities Qf larger particles, cQmputed :from Atterberg'ssettling
velocities and StQkes's fQ~mqla, range II'om 1.64 ,to. 2:03.
Joseph and SnQw (14-) use a settling velQcityof IOcm. in 7.5 hours.,
which cQrresPQnds to' an average' particle .densitYQf 2:71.~ Most
specific-grav:ity ~easurements Qf clay give values :abQve '2.7,. Rcihin­
SQn and Hobnes (26) have. shQwn -that the average chemicalcQm­
PQsitiQn Qf SQil CQllQid is quite ~di:fferent frQmthecbmpositiQn ,Qf ,the
sand fractiQn. The higher sp.ecific gravitY-Qf the collQid may !be ,due,
then, either to difference incQmpQsitlQn 0,1' to the change in density
Qf water in the micrQPoresand.in the surface films .Qf high curyature
.and large tQtal area, surrQunding the ;partieles. FQr purposes of
:mechanical .analysis it is impQrtant ,to. knQW nQtQnly the :aver.age
density of the soil particle Qf diameter 2 JJ., .bl,ltalso ·the ,chaJ;"acter ,Qf
the particle, whethei' it is .asQlidmineral grain ,Qr ,an aggregate,Q.r
 14 !l'ECHNlCALBULLETIN 170, U.:S.DEPT. ·OF AGRICUL'1'U:RE

whether it is agel or hass. gel coating. The particles are too small
to permit the application of the usual petrographic .methods. .
- There is also someuncettainty as to the validity ofStokes'slaw Ifor
such a large ,range of particle shapes and .sizes.Millikan (18) found
it necessary to add a correction to Stokes's 'formula for spheres falling
in a gas when ,the mean free path of the gas molecules exceeded.a
small fraction of the diameter of the sphere. In applying :Stok~s's
Jaw to the settling of slud~es, C. :So Robinson (21) use.dthe viscosity
and density of the suspenSlOn instead of the solution and a coefficie'rit
other than 2/9 for the particle shape factor. It is proposed to study
the relationship of settling velocity and particle siz2in the near
future. In the meantime it seems bestro conform to the .relationship
used in the international method. .'
The average density of soil particles at any given size is chiefly a
matter of academic .interest if size fractions are expressed .in terms
of set~ling v£\locities, as recommended by G. W. Robinson (24) rather
than in terms of equivalent diameters. But in a ,method in whi.ch
five separations are madu by sieving .and only two by settlin~ veloci­
ties, and the results a£e for the use of field investigators, It seems
preferable to hold to the concept of particle size.
Joseph and Snow (14) have shown that, for five soils, considerable
variation in temperature, time, or depth of pipetting produces.a
negligible error in the determination .of 2p. clay but that a hi~her
relative accuracy is required for larger particles. It has been notIced
in this laboratory that the higher the .sedimentationvelocity 'the mote
difficult it is .to secure good checks from pipette sampling. One reason
is that, after the suspension has beenstirr.p~, several .minutesare
required for the turbulent motion to die down and permit settling to
proceed undisturbed.
Formerly when all material smaller than 100.p, was sieved into the i

sedim~ntation chamber, a pip~tte determin!Ltion was made at 50 p,.

The pIpette was then fitted WIth a metal tIP, closed at the bottom,
and provided with si.'( small equally spaced holes drilled horizontally
through the wall of the tip near its lower .end. This tip . assisted
greatly in securing good duplicates, but it was found preferable to
make this separation with a finer sieve. The tip Was then used in
the determinations of clay and colloid in .an effort tO'secure a 'cross
section of the liquid at the deSired depth. Observations on particles
of lampblack indicated that a pear-shaped volume, with the tip
near the bottom of the pear, was removed by 'this means. The results
thus obtained were no more consistent than those obtained with
the pl~ straight glass tip so that the removable tip was aban~one~.
According to Kohn (15), who has worked out the hydro,dynamlc
principles involved, a pipette with any kind of tip, atraight or bent,
shol1ld remov.e a sphere of liquid whose center is the opening of the
pipette tip. It is necessary that the rate of filling the pipette be high
in comparison with the settling veloQity of the largest particles.
OUTLINE OF METHOD
GENERAL STATEMENT

In view of the foregoing facts the writers desire to present a detailed

method of procedure for carrying out the mechanical analysis of soils.
This procedure is the one now in use in the Bureau of Chemistry ,and
 A PlPETTE M.ETHODOF ANALY~IS OFSQ!LS 15
Soils. It differs from the in~:mational method in certain particulars,
viz, the omission of hydrochloric acid treatmeIlt,the method of filter­
mg and washing, the drying and weighing of the sample af~r pre­
t.reatmeIlt,the use of sodium oxala~ as disperSing agent, the use ofa .
finer sieve for we.t sieving, and the fact that there is only a single
transfer of the sample before separation into size classes. This pro­
cedure .is more rapid and, it is believed, more accurate than the
international method. .
.PREPARATION AND DISPERSION OJ'SAMPLE

The sample of soil for analysis is mi...:ed with alarge spatula, and is
then quartered. The quarter reserved for analysis is ,rolled with a
wooden rolling-pin to break up clods and then is passed through .0.
sieve with 2-mm. round holes. Oare is taken in all mixing and
quartering to see that the fine material on the paper is properly
distributed. Samples are run in sets of eight.
Two samples of each soil are weighed out at the same time, a 10-~m .
. sample for analysis and a 5-gm. sample for moisture determinatIOn.
The latter is dried at 105 0 C. for 16 hours, cooled.in a desiccator, and
weighed. The 10-gm. sample is putiuto 0.250 c.c. PyJ:ex electrolytic
beaker. ~
To the 10-gm. sample is added 40 c. c. of 6 per cent hydrogell per­
o~;de, and the beal\:er is covered with a watch glass. The bealc~:ris
frequently shaken in order to bring all the organic matter in contact
with the peroxide. When the reaction has quieted down, m!)re
peroxide may be added where experience indicates that it is needed,
and the beaker is placed on a slow steam bath overnight. Thenif all
organic matter does not appear to have been removed, more hydrogen
peroxide of full strength (about 30 per cent) is added. If manganese
dioxide is known to be present or its presence is indicated by decom­
position of the hY!ITogen peroxide without removal of organic matter;
then a small quantity of acetic acid is added. About 50 mg. of glacial
acetic acid is usually sufficient for a soil containing 0...5 per cent Mn:02'
Repeated additions of small amounts of hydrogen peroxide appear to
give best result.s. After decomposition of the organic mat.ter the
excess peroxide is boiled off.
If acid treatment is desired the sample is cooled an.dthen treated
with 10 c. c. of N HCI, made too. volume of 100 c. c. and allowed to
stand overnight. If it is desired to remove all the carbonates, enough
acid is added so that the concentration of acid in the beaker, after the
carbonates are dissolved, will be 0.1 N. This really.involves a prior
~etermination of carbonates. In this laboratory, method B of the
Association of Official Agricultural Chemists (1) IS used for carbonate
determination.
The sample with or without acid treatment is now ready for wash­
ing. This is done by removing as completely as possible the solution
from the sample in the beaker with a short Pasteur-Chamberland
suction filter. The lower 12cm. of the filter is sawed off, Jittedwith
a removable stopper, and used for this purpose. The liquid is then
removed from the hollow filter core and replaced with distilled water.
By means of a 'rubber bulb back pressure is applied to the filter in
order to remove adhering soil ma.terial. The sample is well stirred
with 125 c. c. of distilled water and the solution again removed by
 :filtration, Si... sueh washings !lJ'e usUally sufficient. If it alumin :is desired
um
to determ ine the quanti ty of dissolv ed lo~des of iron and
y
the washings from each sample are saved sepru:at.ely, made slightl on a
alkaline with ammonia, and boiled. The precip itate is washed
.
filter paper, dried, ignited , and w(}ighed.
.,#ter cleaning and temov ingth,e suction filter, the soil an sample ,is
i~va.porate.d to drynes s on It steam hath and then dried
in elecbic
(overn ight). Upon coo.!.io gin a desicca tor
oven at 1Qi5° C. for 16 hours mo.tic
the sample, still in the beaker , is quicld yweigh ed on a Ohaino up
balance. If this weighing is done rapidly the sample will not takefew
more than 1 mg. of moistu re. The sample.-~s then soaked for 0. an,
minute s with about 25 c. c, of water, stirred with a .rubber policem again
and tmnsfe rred to a .250 e.. c. nm'Sing hottle. The beaker is of the
dt.ied and weighed, the differe nce in wei~ht being the portion
sample left after Qrgallic matter and solutio n loss. Tills is the portion when
of the sample upon which the percen tage of size classes is based, g bottle
the results are used for textura l classifi cation. To the shakin
up
is added 10. c. c. of 0.5 ij:sodi um oxalate, and the volume is made shaker
to 150 c. c. It is .stlaken overni ght on the recipro cating
desc.ribed in Bureau of Soils Bullet in 84. (9).
SEPAltA TlON .INTO SIZE CLASSES

11 sieye fitted with Tyler )~OO-mesh wire sc.reen cloth is clampe d

above a I-liter gradua ted gliiss sedime ntation cylinde r on foot, and
h
the clay and liner silt are decant ed from the nursin g bottle, throug gs the
the sieve, into the sedime ntation cylinde r. After a few pourin
clay is completely remov ed from the shakin g bottle. The entire of
conten ts are then trans~erred to the sieve by means of a stream distilled
distilled water. The cylinde r is :filled to the I-liter mark with is
water and set aside for sedime ntation . ECQnO.lnical use of water
necess ary in decant ation and washin g on the sieve., becauae the total
volume of the silto.n d clay suspension isfunit ed to 1 .liter. platinum.
The sands and remain ing siltma .y then be washed into a sieve in a
dish and dried, but the preferr ed proced ure is to set the wet r
shallow alumim un dish, place in an oven to dry, and .then transie are
the dry sand and silt into the platinu m dish. While the sands
for
drying, aT;lother sieve is used on the next sample. Afte'" drying weighe d,
two homs at 105° C.uh~ sands are cooled in a desicca tor,
separa ted by a nest of graded sieves into size classes , and the weight
the
of each class determined. The summa tion metho d of weighing is
silt and sands is used. The first sand is weighed, the second sand the
added, the total weight determ ined, and so on. If the sum of d thn.t
weigh.t s of the fractio ns is equal to the total weight it"is assume
.no error in weighing or recording weight.s has been made.
The materi al in the sedime ntation cylinde r is stirred with a motor­is
driven propeller, shown in Figure 1, and immed iately the pipette pipette
filled, before the coarser silt htis had time to settle out. Theand the
is emptie d into a low-fo rm glass-c overed weighi ng dish,
gs from the pipette added. The time at ,,~hich this stirrin g
washin diame ter
was made is noted. The time require d for a particl e 5 IJ. aindensity of
to settle 10 em. is compu ted from Stokes 's formul a using
pure water for variou s temper atures (77
2.61 and the viscosity of tes
minutes. at 20° 0.). A curve is plotted with temper atures as ordina
 A PIPETTE METHOD OF ANALYSIS OF SOILS 17
and time as abscissae. At the time shown on the graph for the
tempemture, as read from a thermometer hanging in air near the
cylinder, the suspension is again pipetted for clay. This time the
pIpette, clamped vertically to a SUppOl't (fi~. 1), and with stopcock
closed, is carefully lowered by a rn.ek and pInion until the tip is just
10 cm. below the liquid surface, as shown by a pointer moving over a
:£i.\:ed scale. In case pipetting can not be done at the correct time the
additional depth required for any delny can easily be computecJ., time

FIGCIlI, J.-.I.t Jell, motor lind propeller used to stir suspensions; at right, pipette and stand

and depth being directly proportional. This time the pipette is filled
with an eyen suction in about 40 seconds, and the entire contents run
into a weighing dish. The pipette is washed two or three times and
tho washings added to the dish.
Pipetting of the colloid is similar to that of clay except in this case
settling time is fixed at six and a half hours, and a graph (fig. 2), com­
puted from Stokes's law, giving the relation of depth to temperature
is used.
18 ffECHNICAL Rw,wETlN 170, u.s. DEPT. OF JH~RIC.'Ql.i'1'URE c;:?'
\!
.In Ol'dinary xo.utine.suspensio.ns fro.m eight soil,samples are placed
in ,a set o.fsedimentatio.n cylinders, which areco.v~red ,to. ,prevent
evapo.ration, and are allo.wed to. stand until aCQnvenient time fQr
pipettin~. Each suspensiQn is stirred and the 'first aliqUQt taken at
once. Ii Qr ,the secQnd aliqUQt the suspensiQns are stirred at6-minute
intervals and samples taken at the end Qfthe required time interval,
as xead frQm the time-temperature chart. (Fig. 2.) FQr the third
aliquQt the suspensions are stirxed early .in the day and the samples
taken after sh: and Qne-half hQurs at the CQrrect depth as read '.frQm
the depth-temperature chart. (Fig. 2.) If desired,J:>Qth charts may
be prepared fQrpipetting at CQnstant time intervals or at CQnstant
depth. .
It is essential that the pipette be dried e!l.(}htime befo.re :filling and
that the entire CQntents be emptied into. the wElighing dish. The
pipette is calibrated fQr total v91ume and nO,t fQr delivery. The ali­
qUQts are evapQrated to. dryness Qn a steam bath, dried at 105°0. for
16 ho.urs (Qvernight), cOQled in a desiccatQr, and weighed. Since any

./
V
110 liD
~.,
2~

i"-.. ,,. . ~
!"'-..
~.
1'. .,,::;'"
.......
: ./
...... / ' 1/
1'....
.,./I"- 1-.
/ ...:;~
./ ~ "'"
~
.,.V ~~ aD
,
./ V '''1- I--.
./
50 7/J
~ 10 .JQ .IS
Z5
Tevn,o<!"r'If?rur<!" '"
("<0/
FlOUR& 2.-Sedimentntlon velocity graphs used. in the pipctting of clay and colloid

lQss' of material Qr errQr in weighing is magnified nearly fQrtyfQld ,in

the final result, it is necessary t.Q use care in the determmatiQnQf the
pipetted fractiQns. Weighings are carried to. 0.1 mg. and the weight
of the dishes, is determined each time they are used.
If the rQom temperature changes rapidly the cylinders are cQvered
with bakelite jackets, but in general rQQm-temperaturechan,ges prQ­
duce Qnly negligible variations. If there is a temperature difference
be\tween QPPQsite sides Qf the cylinder there will be a circulatQry mQve­
ment of the suspensio.n, a co.nditiQntQ be aVQided.
OccasiQnally sets of separates are submitted to. William H. Fry Qf
this labo.ratQry fQr micrQscopic examinatio.n by his methQds (10). He
usually finds the sands practically free fro.m adhering co.llQid and the
silt relatively free frQm cQllQidal aggregates.
CALCULATIo.N OF RESULTS

FrQm the mQisture determinatiQn Qf the 5-gm. sample, the dry

weight Qf the sample fQr analysis is cQmputed. After perQxide treat­
ment, and acid treatment if used, fQllQwed by wRshing,the sample is
 APIPETrE METaOD OF AN.ALYSIS OF SOILS 19
dried and weighed. The difference between the ,two weights is the
hydrogen peroxide-solution loss. The first dry weight, which includes
organic matter and solution loss, is taken as the base for calculation
of results except when the analysis is to be used for textural classific8,­
tion only. In the latter case the d..-y w6~ht, aiter washing, is taken
as the base. .
From the weight of the first pipetted aliquot, after deducting the
weight of the dispersing agent, the dry weight of silt and clay in the
sedimentation chamber can be determined directly. It can also be
determined by taking the difference between the total weight of the
sample before dispersion and the wei~ht of the dry sifted silt and
sands. These weights should check, tmt in every case the weight
directly determined by pipe.tting is slightly greater, varying from a
small fraction to as much as 1 per cent or more. The reaso:Q. for this
discrepancy has not yet been found. In l'outine analyses the dry
weight of material in the sedimentation chamber is obtained by dif­
ference, but the amolmt is also determined by pipetting, and if the
two results fail to check within 2 per cent the analysis isrej:>eated. .
From the weight of the clay and colloid aliquots is deducted the
weight of added sodium oxalate and the total amount of material
smaller than 5p. and 2p. is computed. No size class between these
two limits is established. If the weight of clay is deducted from the
total dry weight in the sedimentation cylinder the difference is silt.
This is added to the amount of silt sifted from the sands to give total
silt. . '
Below are given the data obtained and the results computed from
the analysis of a sample of soil:
Sample No. 561860.
Weight of sample, air-dry, 10 gm.
Moisture (determined on duplicate sample) =4.89 per cent.
Dry weight of sample at 105 0 C.=10.000-0,489=9.511 gm.
Dry weight of sample (after pretreatment and washing) 9.366 gm.
Solution loss=9.511~9.366=0.145 gm.=1.6 per cent.
Weight of sand <+ some silt)=2.U6 gm.
Silt=0.046 Size Weight Per
mm. gm. cent
Sand 2.070 gm. 2.0 -1.0 O. 249 2..7
1.0 - . 5 .394 4..2
.5 - ..25 .312 3.3
.25- .1 .567 6.1
.1-.05 .557 5.9
W~ht of silt+clay aliquot '0.1905 gm.
Weight of clay aliquot 0.1127 gm.
Volume of pipette=25.12 c. c.=1/39.8 liter.
Weight of N8.:rC20~ in 25.12 c. c.=0.0065 gm.
Weight of clay= (0.1127-0.0065) 39.8_____________._________ 4.227 45.1
Weight of silt.=9.356-(2.070+4.227) _____________________ 3.059 32.7
Total except solution loss _____________________________ 9.365 100.0
Weight of colloid aliquot=0.0897
Weight of colloid=(0.0897-0.0065) 39.8____________________ 3.311 35.4,

APPARATUS
A few special pieces of apparatus used in mechanical analysis that
have not previously been described are the wei~hing dishes, pipette,
stirrer, sieves, and shaker. The pipette (fig. 1) IS a Lowy automatic,
j A determination of silt may be made os foUows:
Weight of silt aliquot=O.1005-0.1127=O.07i8 gm.
Weight o(silt=O.07iSX39.8+O.!J.I6=3.H2 gm.=33.6 per cent.
 ji'"""
20 TECHNICAIi'BULLET:h'l" t;7(), U. S. DEPT. OF AGRICULTURE
equipped with a stopcock having an air vent. When the pipette is
full ,a 90 0 rotation of the stopper will close the stopcock, and an addi­
tional 90 0 rotation will permit d,divery. The house vacuum is con­
n~cUJd through a :£!ne capillarYl1':..id a"s~aJl sulphmic Bcid. trap to ~c
pIpette. The capillary pernu~';,the deSIred slow and· uniform filling
of the pipette, and the sulphuric acid trap keeps mois,ture from int,er­
fering with, the air flow through the capillary . The pipette has ,a
delivery vo:b,une of 25 c. c. but is calibrated with mercury for total
'\~olume.
The stirrer shown in Figure 1 is a motor-driven 4-blade propeller
ona sliding shaft which can be clamped at any height to the driving
pul~ey. The variable speed motor is run as fast as possible without
whlWping air into the suspension. Mter a suspension has stood for
tW\j or three days it is sometimes difficult, with hand stirring to
break up the flakes of silt and clay aggregates which have settle~ on
the bottom of the chamber.
The weighing dishes are of low, flat form in two sizes, one 65 mm.
in diameter and 30 mm. high and the other 50 mm. by 30 mm., with
correspondingly numbered glass lids ground to fit over the outside of
the dish. The small dish is used for the clay and colloid pipel,ting,
and the larger dish is used for the :first or silt pipetting, because of the
extra washing required to remove the coarser particles from the
pipette.
The sieve :frames, mnde from 2~-inch extra heavy brass pipe, are
fitted with removable rings to permit the easy change of screens.
The two larger sieves have round holes of 1 and 0.5 mm., respectively.
The next two screens conform to the Bureau of Standards specifica­
tions for 60-mesh and 140-mesh sieves. They have square openings
of 0.25 and 0.105 mm., respectively. The finest sieve is 300 mesh
and holds particles larO'er than silt size. The screen wire cloth on
the 300-mesh sieve is frequently renewed, as the wire is so 'fine that
it is easily stretched, worn, or broken. ,
SUMMARY
The method of mechanical soil analysis developed for use in the
soil physics laboratory of the Bureau of Ohemistry and Soils is
described. In pretreatment the organic matter is removed .with
hydrogen peroxide, but the hydrochloric acid treatment, used in the
international method, ordinarily is omitted. A method of removing
organic matter in the presence of manganese dioxide is described.
Soluble matter is removed by washing and filtering with Pasteur­
Ohamberland suction filters. The sample is dried and weighed, and
this weight is the basis of calculation of percentages of material in
each size class, when the results are for use in textural classification.
In all operations up to dispersion the sample remains,in an extra tall
form beaker. The sample is deflocculated by shaking in a dilute
sodium oxalate solution. The colloid, clay, and 'fine silt are separated
from the sands by means of a 300-mesh sieve. The clay and colloid
are determined by sedimenta,tion, the pipette method being used.
The procedure is designed for accurate and rapid analysis.
Investigation of dispersion aids, incidental to this method, discloses
the fact that acid treatment introduces undesirable solution losses
and is not necessary for dispersion, even in calcareous soils, particu­
larly if sodium oxalate is used as the dispersing agent.
 ".

A .l'lP,E'l"l'E ME'l'HOl)OF .ANALX~IS ,OF SOILS .'21

:LI'l.'ERATURE CIT..ED
(1)' ASSOCIA'l'ION .OF OFFICIAL A,G.RICUL'l'.URAL 'CHEBUS'l'S. '
1925. OFFICIAL AND TENTA'l'IVE ME'l'HODS OF ANALYSIS. COMPlLED JlY
'l'HECOMMI'l"l'EE ON EDI'l'ING ME'l'H.ODS OF ,ANALYE!IS. .REVISED
TO JULY i, IOU. Ed•.2, 535 p., ,ilius. Washington, D.C.
(2) A'l"l'ERBERG, A.
~908. S'l'UDIEN AUF DEM GEBlE'l'E DER JlODENKUNDE. Landw. Verso
Sta. 69: [93]-143.
(3) BODMAN, G. B.
1928. 'l'HE HYDROGEN PEROXIDE-HYDROCHLORIC ACID 'l'REATMEN'l' ,01"
SOILS .ASA ME'l'HOD OF .DlS~ERSION .IN MECHANICAL ANALYSIS.
Soil Soi. 26: 459-4,70, illus. .
(4) BoUYoucos, G. B.
1929. !I'HE ULTIM.A!C1'.l NATURAL S'l'RUCTURE OF THE SOIL. Soil Soi. 28:
127-37.
(5) BRIGGS, ,L.J., MAR'l'LN, F:O., and PEARCE, J. R.
1904. T,lIE CEN'l'RIFUGAL METHOD OF MECHANICAL SOIL ANALYSIS. U. S.
Dept. Agr., Bur. Soils Bul. 24, 38 p., illus.
(6) DAVIS, R. O. E.
1925. COLLOIDAL DETERMINATION IN MECHANICAL ANALYSIS. Jour.
Amer. Soo. Agron.17: 2.75-279.
(7) - - - and BENNETT, H. H.
1927. GROUPING OF SOILS ON THE BASIS OF MECHANICAL ANALYSIS. U. S.
Dept. Agr. Ciro. 419, 14 p., illus.
(8) - - - and MIDDLETON, H. E.
1927. DISPERSION OF SOIU3 FOR MECHANICAL ANALYSIS. First Internatl.
Congo Soil Soi. 1: 366-3.77.
(9) FLETCHER, C. C., andBRYAN, H.
1912. MQDIFICATION OF THE METHOD OF MECHANICAL SOIL ANALYSIS.
U. S. Dept. Agr., Bur. Soils Bu!. 84, 16 p., illus.
(10) FRY, W. H.
1923. !I',HE MICROSCOPIC ESTIMATE OF COLLOIDS IN SOIL SEPARATES.
• Jour. Agr. Research .24: 879-883.
(11) HILGARD, E. W.
1906. SOILS; THEIR FORMATION, PROPERTIES, COMPOSITION, AND RE­
LATIONS !I'O CLIMATE AND PLAN!I' GROWTH IN THE HUMID AND·
ARID REGIONS. 593 p., illus. New York and London.
(12) HOPKINS, C. G.
1899. A RAPID METHOD OF MECHANICAL SOIL ANALYSIS, L.'lCLUDING THE
USE OF CENTRIFUGAL .FOUCE. U. S. Dept. Agr., Div. Chern.
Bul. 56: 67-68.
(13) JENNINGS, D. S., THOMAS, M. D., and GARDNER, M.
1922. A NEW METHOD OF MECHANICAL ANALYSIS OF SOILS. Soil Soi. 14~
485-499, illus.
(14) JOSEPH, A. F., and SNOW, O. W.
19.29. THE DISPERSION AND MECHANICAL ANALYSIS OF HEAVY ALKALINE
SOILS. Jour. Agr.Soi. 29: [106]-120, illus.
(15) KOHN, M.
1928. BEI'l'RAGE ZUU !I'HEORIE UND PRAXIS DEU MECHANISCHEN BODEN-
ANALYSE. Landw. Jahrb. 67: 486-546, illus. '
(16) KUAUS, G.
1923. ERGXNZENDER JlERICH!I' UBEU EINE DEM PBAGER BODENKUND­
LICHEN JWNGRESS VORGETRAGENE NEUE METHODE DEU MECH­
ANISCHEN BODENANALYSE,SOWIE .EIN EINFACHESGRAPHJSCHES
VERFAHREN ZUR BESTIMMUNG DER KORNOBERFLXCHE, .FERNER
PRAKTISCHES GEMT ZU"R PROBEENTNAllME ZWECKS ERMI'l"l'ELUNG
DER LAGERUNGSWISE. Iuternatl. Mitt. Bodenk. 13: 147-160,
illus.
(17) MA'l"l'SON, S.
1926. ELECTRODIALYSIS OF THE COLLOIDAL SOIL MATERIAL AN.D TIlE
,EXCHANGEABLE BASES. Jour. Agr. Researoh 33: 553-567, illus.
(18) MILLIKAN, R. A.
[1924]. THE ELECTRON, I!I'S ISOLATION AND MEASUREMENT AND !I'HE DETER­
MINATION OF SOME OF ITS PUOPERTIES. Ed. 2, 293 p., illUB.
Chioago.
 TECHN.ICAL BU.I,4.E,!l'lN 170,U. S. DEPT. OF AGRICULTURE
C
(19) NovlK, V.
1927. CONCLUSIONS OF ',l'HE FlRS',l' COMMISSION MEE',l'ING AT .JlOT.HAM­
STED, llARPEND.ENl 926. :Internatl. SOC. Soil Sci. ;16 p. :Brno,
Czechoslovakia. [In English, p. 5-'-8; German,p.9-12; ;French,
. ~p. 13-16:1
(20) OSBO~NE, T. B.
18.86. THE METHODS OF MECHANICAL SOIL .ANALYSIS. Conn. Agr. Expt.
Sta. Ann. Rpt. - - : 141-158.
(21) ROBINSON,C. S.
1926. SOME FACTORS INFLUENCING SEDIMENTATION. Jour. lndus. and
Eng!!l. Chern. 18: 869-8.71, iUus.
(22) ROBINSON, G. W.
1922. NOTE ON THE MECHANICAL ANALYSIS OF .HUMUS SOILS 1922. Jour.
Agr. Sci. 12: [287]-291. .
(23)
1~.22~ A NEW ME~HOD .FOR T.RE MECHANICAL aNALYSIS OF SOlLS :AND
OTHER DISpERSIONS. JoUl:. Agr:. Sci. 12:,,[306]-321, lilus.
(24) ~l--
1927. THE GROU.Pl'NG OF FRACTIONS ,IN MECHANICAL ANALYSIS. First
Internatl/ Congo Soil Sci. 1 :359-365.
(25) ROBINSON, W.O. .
192.7. THE DETE,R!,luNATION OF ORGANIC MATTER IN SOILS BY MEANS OF
HYDROGEN .PEROXIDE. Jour. Agr. Research 34: 339-356.
(26) - - - and HOLMES, R. S.
1924. THE CHEMICAL COMPOSITION OF SOIL COLLOIDS. U. S. Dept. Agr.
'Bul. 1311, 42 p.
(27) SNXDER,H.
1896. THE CHEMICAL AND .MECHANICAL ANALYSIS OF SOILS. Minn. Agr.
Expt. Sta. Ann. Rpt. 1895: [32]-63, lilus.
(28) .STOKES, C. G. '
1856. ON TBE .EFFEC',l' OF THE INTERNAL FRICTION OF FLUIDS ON tl'HE
MOTION OF .PEND.ULUMS. Cambridge Phil. Soc. Trans. (1851­
56) 9 (Pt. ll): [8]-[106].
(29) WlEGNER,G.
192.7. METHOD OFPREPARAtl'ION OF SOIL SUSPENSIONS AND DEGREE OF
DISPERSION AS .MEASURED BY THE WlEGNER-GESSNER .APPARATUS.
Soil Sci. 23; 377-390,
 ,QRGANIZATION 'OFl'HE
,Ul\1ITED ,STATES DEPART.MENT 'OFA~RIC.u'Ll'l1Rl)l
,Dec:eml>ek-l2. 1929

Becretaryoj Agri~lture ___________________ MTHURI\d. H:Y~E.

ABs:istant Secretary _______________________ i.R. 'W. DUNLAP.
Director>ojBcientific >Wor.~_---------~----- A. iF. WOODS.
iDireclorof Regulat:>ry 1firirk_ ___ ___ _____ __ _ 'WALTER 'G. {:lAMPBELL.
Di'reciorof Extension ,Work ________ -- _____ Q. W. WARBUllTON.
J):ireclor of PerB01I,nel .and B1!8ineliB Admin-W. W. ·STOCE:BERGER.
istration.
i(#rectoroj Information ________________ --- M. 'S. :EII'IEN.HOW,ER
Weather Bur~u_::. __________ ----- ________ 'CHARIiES;F. MAR:VIN, 'Chief.
$ureau of Animal indu8try _______________ JOHN R. MOHIiER, ·Chief.
BU'rllau of Dairy :Industry _________________ O. E. ,RE.ED, 'Chief.
BUTeau of Plant Industry________ ._________ WILLIAM A. TAl'LOR, Chief.
Forest Service ______ :. ______ ~---- _________ R. Y.Sl1UART, .Chie/.
Bureau of Chemi8try and BoiI8 _____________ H.G. .KNIGHT,Chief.
;Bureau of Entomology____________________ C. :L. MARLATT, {Chief.
;Bureau of Biological Survey._______________ ,PAt'L G.REDINGr,J:Ol{, Ohiej.
Burllau of Public Roads __________________ THOMAS H. MACDONALD, >Chief.
Bureau of Agricultural !8c9nomic8 __________ NILS A. OLSEN,Chief. >

;BurealL of Home Economics _______________ LOmSESTAJ."'lLEY, ·Chief.

Plant Quarantine and Control Administra- LEE A.STRQNG,Chief.
tion.
-Grain F'Uture8 AdminiBtration _____________ J. W. T. D.u:vEL, 'Chief.
Food, Drug, and Jnsecticide Admini8tration __ .WALTERG.CAMPBELL, .Directorof
Regulatory ,Work, in 'Charge.
'Office of Experiment Station8_._._:..__________ , Chief.
·Office of Cooperative Extension Work ________ 'C. B.SMITH,'Chief.
:Libra'1I_________________ - ______________ >CLAID.BEL ,R. 'B~NETT, Librarian.

This bulletin is 8. cOlltribution from

;Bureau of Chemi8try and Soil8 _____________ H. G. KNIGllT,Chief.
'Soil Investigations ___________________ A.G. MCCALL,·Chief.
Division oj Soil Chemistry and H. ,G. BYERs, Principal ·Chemist,
Physics. in Charge.
23

U.s, GOVERNMENT PRINTING qFF1C~' .IZ.

For anle by tbeSqperlntelldellt of Documents, Wnshington, .0. C. - .- _ _ - ,PrIce 5 cents
 .'

". r