J.S.F.A., 1, Nov.

, 1950 WHITTLES AND LITTLE-ANALYSIS OF SOIL EXTRACTS 323

A COLORIMETRIC METHOD FOR THE mum development of the weak and strong colours. The
colour obtained was also found to be unstable to light,
DETERMINATION OF POTASSIUM especially sunlight.
AND ITS APPLICATION TO THE (a) Reduction of nitrite in alkaline medium with ferrous
ANALYSIS OF SOIL EXTRACTS sulphate in presence of silver, and distillation of a m m ~ n i a . ~
It would appear that this method is unsuitable for routine
By C. L. WHITTLES and R. C. LITTLE work since it involves a distillation.
Parker15 has described an absorptiometric method for deter-
.I colorimetric procedure is described for the determination of mination of nitrite, based on the use of a-naphthylamine.
small amounts of potassium. The soil is extracted with Bray's This has been applied by Kenyon, Ovenston and Parker' to the
extracting reagent and the available potassium is precipitated as determination of potassium in the pug.-range. Their paper' gave
cobaltinitrite. The determination of potassium precipitated in-
volves the diazotization of sulphanilic acid, using the nitrite groups a good review of the literature of cobaltinitrite techniques for
of the cobaltinitrite precipitate, and subsequent coupling with determination of potassium, as did Tinsley's 1948 paper1' ; and
dimethylaniline to give methyl orange. The colour is measured hints regarding possible sources of interference can be obtained
with a Spekker absorptiometer. The method is applicable to the from both these critical publications. The method of Kenyon
routine determination of potassium in soil extracts.
et al. is sensitive when applied to relatively pure substances
or solutions, but it seems not to have been tested with soil
Introduction and literature review extracts. The information contained in the paper of Kenyon
et al. does not appear to conflict with our conclusions.
The need was felt for a rapid and reliable chemical method For the determination of ' available ' potassium in routine
for the determination of potassium in soil extracts in con- soil analysis, use of the nitrite groups in the potassium
nexion with the routine examination of large numbers of soil cobaltinitrite precipitate seemed to us to form the basis of
samples for advisory purposes. The method here described a quantitative method offering the greatest promise of success.
was intended for this purpose and has been found adequate I n a search for suitable reagents for diazotization and coup-
in practice. In the search for a method, the requirements ling reactions, we found that sulphanilic acid and dimethyl-
of speed, accuracy and applicability t o large numbers were aniline served our purpose. Use of these reagents results in
borne in mind. These considerations suggested that a colori- the formation of methyl orange, which is stable, and possesses
metric method would be most suitable. the desired properties for application to colorimetric work.
Of the several methods hitherto proposed for the chemical A method is described below whereby potassium is leached
determination of small amounts of potassium, most have been from the soil and precipitated as potassium cobaltinitrite.
based on precipitation with sodium cobaltinitrite ; a few are Sulphanilic acid is then diazotized by the nitrite, and coup-
based on precipitation with chloroplatinic acid or iodoplatinic ling with dimethylaniline follows. The colour intensity of the
acid. solution is measured, and the potassium determined (as K,O)
The amount of potassium cobaltinitrite precipitate can be from a light-transmittance curve prepared from standards.
measured by several methods. The turbidimetric method has
been proposed"? l9 but is open t o criticism. Variations in Method
results may be caused by slight variations in particle size Reagents
due to varying conditions of precipitation, e.g. temperature I. Bray's extracting reagent.2-Iooo g. A.R. sodium acetate
or concentration of reagent ; sedimentation may also cause is dissolved in 1600 ml. distilled water and filtered. To seven
difficulties. Colorimetric methods are based on determina- parts by volume of this solution are added three parts of a
tion of either the cobalt or the nitrite of the precipitate. mixture of equal volumes of water and concentrated nitric
(i) Cobalt.-Methods employing the cobalt of the precipitate acid.
include : 2. Sodium cobaltinitrite reagent.l3-(a) 226 g. finely ground
(a) a-Nitroso-/?-naphthol method'$ : this method is some- cobalt acetate is dissolved in 500 ml. distilled water with
what cumbersome for routine work. 200 ml. glacial acetic acid and made up to 1000 ml. with
(b) Modified thiocyanate method101 l*$20 : this method distilled water.
also has disadvantages for routine work on account of the (b) 440 g. sodium nitrite crystals is dissolved, and made
many manipulations involved. up to 1000 ml. with distilled water.
(c) Choline ferrocyanide' : this method has too low a sensi- For use, equal quantities of (a) and (b) are mixed in a
tivity to be generally applicable for estimating potassium in small suction flask and the nitrous oxide fumes removed by
soils. evacuation (preferably left standing under vacuum overnight).
(a) Nitroso-R-saltls : this method appears to have too The solution should be stored in the dark, and filtered as
limited a range, and inadequate sensitivity, for routine work. required for use.
(e) Terpyridyl has been suggested14 for cobalt estimation, 3. Formaldehyde, A.R.
but that reagent is not readily available. 4. isoProfly1 alcohol.
(ii) Nitrite.-Methods for determining the nitrite of the 5. Acetone (dilute 30%).-30o ml. A.R. acetone is mixed
precipitate include : with 700 ml. distilled water.
(a) The Griess reaction4 using a-naphthylamine and sul- 6. Acetone, A.R.
phanilic acid. 7. Sulflhanilic acid reagent ( N / ~ o ).-17.3 g. sulphanilic acid
(b) Herzner's modification5 of the Griess reaction4 whereby crystals is dissolved in 400 ml. distilled water, and added to
the red is converted into a yellow or orange by the addition 4 1. distilled water. 60 ml. concentrated hydrochloric acid
of strong alkali. is diluted to 600 ml. with distilled water and added to the
Both these ' nitrite ' methods have the grave disadvantage, solution, making 5 1. in all.
in quantitative work, that the colour continues to develop 8. N-Sodium carbonate.-z6.5 g. anhydrous sodium car-
indefinitely. bonate is dissolved and made up to 500 ml. with distilled
(c) The formation of the nitroso addition compound12 with water.
dimethylaniline. This method was unsatisfactory in our 9. Dimethylaniline solution.-6 ml. dimethylaniline is dis-
hands on account of the difference in time required for maxi- solved in 500 ml. 50% aqueous ethyl alcohol.
324 TVHITTLES AND LITTLE-ANALYSIS OF SOIL EXTRACTS J.S.F.A., 1, Nov., 1950
Procedure again mixed, the flasks are made up t o the mark with dis-
The following is a detailed description of the procedure tilled water, stoppered and mixed by inversion. After at
adopted for the routine determination of ' available,' i.e. least I hr. the colour intensity is measured on the Spekker
extractable, potassium (as K,O) in soils. absorptiometer, using filter KO.6 (blue) ; the readings are
5 g. air-dry soil, which has passed through a 2-mm. sieve, interpreted by means of a calibration curve.
is weighed out and then put into a 60-ml. beaker. 10 ml. The method has been employcd in the routine examination
of Bray's reagent is added and the mixture is stirred with of soil samples for advisory purposes a t the soil laboratory
a rubber ' policeman ' vigorously for 30 scc. The contents of the West of Scotland Agricultural College since September,
of the beaker are then filtered into a small tube through a 1946. I t has been found possiblc for one worker to maintain
7-cm. Whatman S o . 42 filter paper. I ml. of the filtered an output of 48 determinations per day ; the analysis of a
extract is pipetted into a centrifuge tube (21 in. x 4 in., fresh batch of 48 samples is begun each day and the deter-
tapered) containing I ml. formaldehyde and three drops of minations are completed on the following day.
isopropyl alcohol. The contents are mixed by shaking.
0.5 ml. sodium cobaltinitrite reagent is added (or more if Yre;baration of calibratioit curzie
necessary) and the tube is shaken again. The tube is then A series of standard solutions containing known amounts
stoppered and allowed t o stand overnight in the refrigerator of potassium was prepared. For this, known amounts of
at 5 ° C . potassium chloride were dissolved in Bray's reagent. The
Next day the tube is centrifuged for 15 min., at not less foregoing routine procediire was adhered t o strictly ; tlic:
than 3000 r.p.m. (The centrifuge used has a special ' solid ' potassium was precipitated and the colour was devchoped
head able t o take 24 of the small centrifuge tubes.) The from the precipitate. The results obtained were plotted
supernatant liquid in the tube is decanted down a glass rod (Fig. I) ; the Spekker values uscd were thc means of a series
(bent at a right angle) into a beaker, care being taken not to
disturb the precipitate. 2 ml. 30% acetone is added to the
tube, which is then centrifuged as before. The operations
of decanting and Centrifuging are repeated. After the third
centrifuging the liquid is decanted, the precipitate is worked
up in I ml. pure acetone, using a fine glass rod, and the glass
rod washed with a further I ml. acetone: the tube is then
centrifuged as before. The acetone-washing operation is
repeated after the fourth centrifuging.
The precipitate is worked up with I ml. om+NaOH and
the glass rod is washed with a further I ml. o m - N a O H . The
centrifuge tube and its contents are then heated in a water
. at least 10 min. and until there is no
bath at 5 0 - 1 o o ~ ~for
remaining trace of the yellow precipitate.
The tube is removed from the bath and allowed t o cool.
When cold, the contents of the centrifuge tube are first poured,
then washed with cold water through a long-stemmed funnel
into a Ioo-ml. graduated flask which has been freshly rinsed
with cold water, and which contains 5 ml. sulplianilic acid
reagent. The contents of the tube and washings amount to
about 90 ml. Cold distilled water below 1 5 " ~ is . used for
these rinsings and a glass rod fitted with a small rubber . SPEKKER REAOINGS (means)
' policeman ' is employed to ensure cornplcte transference. FIG.r.--.CuZihvnfioii c w u c
After an interval of 30 min., I ml. sodium carbonate reagent
is added, and the contents of the flask are mixed b y swirling. of determinations. The results, of which 'fable I presents
I ml. dimethylaniline solution is then added, the contents are specimens, wcrc esamined statistically.

Table I
Slntistical enaniirratioii of ralihmtioir dola :
hctual amount vf I i 2 0 in precipitate from I ml. standard
KCl solution. pg. .. .. ., .. .. .. 5 50
Mean Spekker reading .. .. .. .. .. .. 0485 "'453
Mean determined concentration . . . . .. .. .. 49'5
Standard tleviation .. .. .. .. .. .. + 4'97
1.00 11 .+.OR
:\mount o f I<,O dctcrminctl using grapli o f mean readings, pg. 2.75k7)
4'75
5'25
4'75
6.25
6.00
5'25
4'75
Significance of poorest values (4)
Student's f,,,* . . .. .. .. .. .. .. 2.08
P .. .. .. .. .. .. .. o.ro-o.05
This Table shows the sensitiveness o f the method, which, nevertheless, was not intended t o detect quantities of the order of 5 pg.
A low value o f available Ii,O in soil would be 4-5 mg. per 100 g. soil, equivalent t o 20-25 pg. K,O per inl. soil extract.
* Seven degrees of freedom.
J.S.F.A., 1, Nov., 1950 W H I T T L E S A N D LITTLE-ANALYSIS OF S O I L EXTRACTS 325
Discussion Efect of time allwed for diazotization.-In order to deter-
In actual routine work every tenth sample of soil is analysed mine the effect of time allowed for diazotization, tests were
in duplicate. From 290 successive pairs of duplicates obtained made on a series of six solutions covering the range of nitrite
in actual work the percentage of variation from the mean was concentration used in this work. The results are shown in
calculated for each pair. From the summation curve (Fig. 2 ) Table 111, from which it will be seen that all solutions were
fully diazotized within 30 min.
Table I l l
Effect of time between adding sulplia?rilic acid atid adding
sodium carbonate
' 8 0 -- O O r
Sodium Spekker scale readings
nitrite, when the time between addition of sulphanilic acid
N. and of sodium carbonate is
j min. 10 min. 20 min. 30 min. 60 min.
I
0 .. 2'00 2'00 2'00 2'00 2'00
I I I I 22
I10
..
..
1.86
1.56
I '84
1.51
1.79
1.45
1.78
1'43
1.78
"43
220 .. 1.36 1'33 1.26 1.24 1.23
1100 .. 1'01 0.930 0.900 0.87j 0.870
2200 .. 0.81j 0.800 0'730 0.720 0.710
4 400 .. 0.680 0'635 0.580 O'j45 0.550

Efect of time allowed for colour development after adding the

coupler.-h this investigation the intensity of the colour was
read after various periods of time. It was found that the
0 5 10 I5 20 25 30 3s maximum intensity was reached after about 60 min.
PER CENT DEVIATIONS FROM MEAN Efect of varying the amounts of reagents.-The effect on
FIG.-,.---Suin~mhm of disfribution of percentage deviations froni lhe results of varying the amounts of reagents used is shown in
mean in routine duplicate deferinination Table 1V. The length of time for diazotizing was held con-
stant at 30 min. and the time of coupling at 60 min.
it will be seen that 20% of the replicates gave agreement,
and that over 90% showed a variation of less than 10% of Table IV
the mean. The coefficient of correlation between replicates Effect of irarying the amounts of sulphanilic acid, sodium carbonate
was found to be +
0.989 f 0.001. The fact that fairly satis- and dimelhybniline
factory agreement obtains over the range was also shown Sodium Spekker scale readings with
when the first determination was plotted against the second nitrite, recommended twice the three times the
Leg. quantities recommended recommended
on a logarithmic scale : the resulting graph gave a very close quantities quantities
approximation to a straight line. 0 .. .. 2'00 2'00 2.00
From the results obtained by using standard solutions of 22 .. .. 1.80 1'79 1.80
sodium nitrite it seems that the main errors in the deter- II0 .. .. 1.46 1'44 1'45
mination of potassium are introduced before diazotization. 220 .. .. 1.26 1.26 1.26
From the curve in Fig. I it can be seen that readings were
I100 .. .. 0.890 0.875 0.880
2200 .. .. 0.760 0'730 0.720
readily obtained for amounts of K,O per ml. of extract,
ranging from 5 to 200 pg., using the yellow alkaline colours Constancy of reactions and products.-It was thought desir-
of methyl orange when the Spekker was set a t 1.00. For able to estimate the completeness of the diazotization and
higher values of 'potash ' it is optional to make suitable coupling reactions. This was done by preparing a series of
dilutions of the colour or t o set the Spekker at 2.00 ; if the standard solutions using A.R. methyl orange. The colour
latter is chosen, the maximum concentration which can be intensities of these solutions were measured and a graph was
dealt with is 1000 pg. K,O per ml. of extract. drawn (Fig. 3) plotting log(pg.-equivalents of methyl orange)
In order to estimate the effect of various conditions on
colour development a series of tests were made on standard
solutions of sodium nitrite.
Effect of temperature of diazotization.-Standard sodium
nitrite solution was used as a diazotization reagent under
three sets of temperature conditions ; subsequently the pro-
ducts were coupled and the colour intensity measured. The
results (Table 11) show that working at about 15" c. had no
detrimental effect on the reading.

Table I1
Effect
ternperalure on draaotizatran
o/
Concentration Spekker scale readings
of sodium Water at oo c. Water at 0 - j o c . Water at room
nitrite, pg. in flasks kept in flasks allowed temperature
at 0-5°C. to stand at in flasks allowed
room to stand at
temperature room
(about 12-1 j n c.) temperature
0 .. .. 1'000 1'000 1'000
IIO .. .. 0505 0.500 0'505
220 .. . . 0'34j 0'350 0'350 FIG. 3
*
326 C H I L D A N D NATHANAEL-SUGAR I N COCONUT T,VATER J.S.F.A., 1, Nov., 1950
against Spekker readings. Plotted on the same graph are lo I,ewis, A. 11. and Marmoy, F. n., J . So(. ~ h e ~I in.d . , 1933. 52,
Spekker readings v . calculated pug.-equivalents of methyl '77T
l1 Mellor, J. W., Trans. cerani. S O C . , I ( ) I ~ - z o . 19, 132
orange developed by the diazotizing and coupling procedure, l2 Miller, E. H., i l n d y s f , 1912. 37, 345
using known amounts of sodium nitrite. It will be seen from l3 Morris, R. L.,ibid., 1 ~ 1 2 3 .48, 256
the graph that the two curves coincide, showing that the 1 4 Moss,M. L. with .\lellon, 31. (;,, Indtcstr. Engng C l i e ~ (analyl.. Ed.),
reactions go t o completion. I t also indicated that one 1943. 15, 74
equivalent of nitrite reacts to give one equivalent of methyl Parker, C . X., Analy.sl, 1919, 74, I I L
1 6 Sideris, C. P., Indzistr. I:'ngnR. Chenz. (analyt. E d . ) , 1937, 9, 145
orange, as is to be expected. l7 Tinsley, J., A m l y s t , 1048, 73, 86 ; 1949, 74, 167
Log(pg.-equivalents of K,O) plotted against the Spekker Tomula, E. S.. Z. Analyt. Chem., 1931, 83, 6 (abstr. in Chem.
readings gave a straight line except in the lower values, i.e. , ~ ? J s . , T ~ 25 1 1 8 1 )
19 Wo$B., jndzrstr. I:'ngng Chenr. (anal.yf. E d . ) , 1043, 15, 248 ; 1944,
the curve was straight above 2 0 pg. per ml. Log(,ug.-equiva- 16, t z i
lents of sodium nitrite) were also plotted against the Spekker ZinzadzC, C., Ckinr. el Indrdsfr.. 1932, 27, Special So., 841 (abstr.
readings. The. two lines thus plotted were found to be almost in C h e w A ~ J s .1, ~ 3 3 2 ,26, 3751)
but not quite parallel. The disparity may be due to the
presctnce of cobalt (possible as hydroxide) or an effect of the
colour filters. The parallelism of the two curves indicates
that the composition of thc cobaltinitrite precipitate remains
practically constant throughout the range studied. CHANGES IN T H E SUGAR COMPOSI-
TION OF COCONUT WATER DURING
Practical notes MATURATION AND GERMINATION
The flexibility of the method makes it adaptable to the
quantitative estimation of ' available ' potassium or ' potash ' By R . CHILD and W. R . N. NATHANAEL
throughout the whole range of soils to be encountered. The
only difficulty arises from very peaty soils in which the Changes in the sugar content of the water of the coconut have
abundance of organic matter sometimes leads to the floccu- been followed during ripening and gcrmination. From the most
immature fruit in which the cavity has just formed to about the
lation of organic colloids on the addition of the sodium seven-month-old fruit when i t is full size but still unripe, the
cobaltinitrite reagent. This precipitate, though not itself sugars present, which are almost entirely reducing sugars, increase
interfering appreciably with the method, is thought t o retain t o a maximum of about 596 concentration. Thereafter non-
some of the sodium cobaltinitrite reagent ; hence it gives a reducing sugars appear, but the total concentration of sugars
falls t o about 2% in the fully ripe nut, 1 2 t o 13 months old.
more intense colour than that which corresponds to the The results were much the same in the three varieties of COCOS
potassium extracted from the soil. This difficulty is p ~ r t l y nttcijera L. studied-the ordinary tall palm, a dwarf variety, and
overconie by dilution. 'The amounts of potassium (as mg. a variety with orange-coloured fruit, known in Sinhalese as Rath-
K,O per IOO g. air-dry soil) commonly range from about Tliembili-except that in the water c:f dwarf nuts small amounts
of non-reducing sugars were present in the earlier stages.
3-20 in uncultivated and agricultural field soils, and up to During tlie early stages of germination the concentration of
300 or more in soils used for market-gardening or for crops reducing sugars continues t o fall, but that of total sugars remains
under glass : all such values are satisfactorily embraced by fairly constant until the fourth month of germination, when it
the method given. Necessarily the lowest values are not also falls.
strictly accurate ; but if a soil reveals only 3-4 mg. K,O
per roo g. its potassium status is likely to be much too low Analyses have been carried out on the water present in
for good agricultural production, and no purpose is served coconuts at various stages of development from the first for-
by striving after fractional accuracy difficult t o attain by mation of a liquid-containing cavity in the small immature
any method. fruit u p to complete maturity, and thereafter during the
When large numbers of soils h a w to be analysed, time and germination of the nut. The coconut palm produces new
trouble spent in searching for appropriate dilutions are kept flowering branches a t approximately monthly intervals. A
at a minimum if soil samples are grouped so that those pre- period of 12 of 13 months elapses between the emergence of
sumed ' high ' or ' low ' (respectively horticultural, and forest the inflorescence anti complete maturity of the fruit. At any
or agricultural soils) are processed together. one time, therefore, the palm is carrying twelve or more
bunches of fruit at successive stages of development, each
Acknowledgments representing a growth interval of about a month. Changes
The authors express thanks to Mr. E. K. Schofield-Palmer in composition of nut water during maturation were studied
ior statistical suggestions and to Dr. Hugh Sicol for help by stripping all the nuts from each palm under investigation ;
and valuablc criticism. the volume of water in each nut was measured and the desired
analyses carried out on each sample of nut water.
Soil Laboratory, \Vest of Scotland Agricultural College To study changes during germination, 300 ripe coconuts
Auchincruive were put down in a nursery and 2 0 removed for examination
A y r , Scotland at fortnightly intervals ; on these, determinations were made
Received Julie 1 2 , 1950
similar to those in the maturation study.
References
Methods
Xtack. 1:. \\-., ,I. S O L .( h p J J , , l m i . , 191j . 34, G41
Bray, I<. H . , , I . -4wer. Sor. .-lpon., 1932, 24, 312
2 Reducing sugars were determined by the method of Lane
C n t t e , J . ant! Ihicet, G., Anrr. .-lgron, 1946, 16, 2 2 5 (abstr. in
J
and Eynon' on suitably diluted samples, before and after
Soils 6. I;ertil., 1947, 10, 135) inversion. For inversion, IOO ml. of suitably diluted nut
4 Griess, I'., I l r r . dtsch. C h e w . & s . , rX7S, 11, 624
Herzner, I t . :\., Biocken?. %., 1931, 237, 1 2 9 water was allowed to stand overnight with 3 ml. conc. HCI ;
6 Hurwitz, C a n d T3atclielor. H . W.. Sod S r i , , 1947, 63, 3.j after neutralizing with sodium hydroxide to litmus i t was
'
Jacobs, I f , l i . I). and Hoffman, W. S., J . b i d . Chenz., 1931. 93, diluted t o a suitable standard volume. Sugars present are
68.5 reported as ' g. per IOO ml. of reducing sugars calculated as
Jones, E. G.,ilnulyst, 1918, 43, 317
9 Kenyon, C., Ovenston, T . C. J . and Parker, C. h.,ibitl., 1950, invert sugar,' and ' g . per 100 ml. of non-reducing sugars
75, 269 calculated as sucrose.'