Published January, 1959
Application of Phase Rule to the Chemical Behavior of Monocalcium Phosphate
Monohydrate in Soils1
WALTER E. BROWN AND JAMES R. LEHRS
ABSTRACT
. When a granule of monocalcium phosphate monohy-
drate, Ca(H2PO4)2 • H2O, dissolves in a soil system, the
fraction, f, of the phosphate left at the site of the granule
as a dicalcium phosphate, CaHPO4 or CaHPO4 • 2H2O, is
shown by the relation
f = (1 - R)/(2 - R)
where R, the mole ratio of CaO/P2O5 in the composite so-
lution removed from the granule, can be determined ex-
perimentally or estimated from phase-rule data for the
system CaO-P2O5-H2O. The composition of the solution
leaving the granule approximates the invariant-point solu-
tion in equilibrium with Ca(H2PO4)2 • H2O and either
CaHPO4 (the stable invariant point) or CaHPO4 • 2H2O (a
metastable invariant point).
Equilibrium data from the literature suggest that a
residue of CaHPO4 will contain some 28% of the P. Ex-
perimental data are presented from which a residue of
CaHPO 4 -2H 2 O is predicted to contain about 21% of
the P.
Reaction of. some of the solute phosphate with cations
such as iron and aluminum can cause precipitation of addi-
tional calcium phosphate.
P LACEMENT OF A GRANULE of monocalcium phosphate
monohydrate in soil initiates physicochemical processes
that are the subject of a companion article (17). A consid-
eration of their quantitative aspects aids greatly in relating
these processes to the uptake of phosphorus by plants from
fertilizers containing monocalcium phosphate monohydrate.
Shown here is the use of phase-rule data to predict the
amount of dicalcium phosphate residue, the composition
of the solution moving from granule to soil, the amount of
additional dicalcium phosphate that may form incidental
to solution-soil reactions, and effects of salts on the amount
of residue and on the composition of the solution.
Of many reports on the system CaO-P2Os-H2O, several
are pertinent here (2 to 9, 11, 19). Investigators of the
25° C. isotherm agreed fairly well except as to the degree
of hydration of the dicalcium phosphate. Cameron and
co-workers (4 to 7) reported that their data applied to the
dihydrate, whereas Bassett (2) reported the anhydrous
form over all the range of stability of the dicalcium salt
except the dilute end, where it was mixed with the dihy-
drate. Bassett concluded that the isotherms for the two
forms were nearly coincident.
The more recent work cited above shows that the solu-
tion compositions reported by Cameron and by Bassett
were essentially correct. A preponderance of evidence indi-
cates, however, that the saturating dicalcium phosphate
phase in their studies, and in all others at 25° C., was the
anhydrous salt, the only exception being at low concen-
trations, where conversion of the dihydrate to the anhy-
drous form is quite slow and where the two isotherms
probably are near coincidence.
A considerable difference between the isotherms for the
two salts in the concentrated range is indicated by meas-
urements at 40° C. (8) and by the results reported here
for 25° C. Despite its instability with respect to the anhy-
drous salt at normal soil temperatures and humidities, the
dihydrate is the usual initial product of hydrolysis of mono-
calcium phosphate monohydrate in the soil (16)—a point
that focused attention on the need for information on the
metastable isotherm at soil temperatures and led to the
experiments reported here.
The results of various studies of equilibrium in the sys-
tem CaO-P2O5-H2O are diagramed in figure 1. Lines OS
and SC represent the compositions of solutions in equilib-
rium with CaHPO4 and Ca(H2PO4)2-H2O, respectively.
The point of intersection of these isotherms, S, is the stable
invariant point. Metastable solubility isotherm OM for
CaHPO4-2H2O and metastable invariant point M are from
data presented here.
Metastable and stable isotherms are equally valid ther-
modynamically; all are lines of constant chemical potential
as defined by the relations
For OS:
^CaO+1/2'p205+1/2*H20=<<0CaHP04 f1]
For OM:
K C aO + CaHP0 4 -2H 2 O [2]
1/2 + 2 = f j tp
^P 2 0 5 ^ H , 20O ~
For CSM:
"CaO+"pl0i + 3
"H20="0Ca<H2P04>2-H20 ^
 SOIL SCIENCE SOCIETY PROCEEDINGS 1959
where the terms on the left are the chemical potentials
of the components in the solutions and those on the right
are the standard free energies of formation of the solids.
Measurements in the metastable region may be compli-
cated by supersaturation phenomena that cause no diffi-
culty in the stable region.
When monocalcium phosphate monohydrate is hydro-
lyzed in pure water, the fraction, f, of the phosphorus left
in the residue of dicalcium phosphate depends solely on
the mole ratio CaO/P2O5 in the solution formed in the
hydrolysis. The hydrolysis reaction may be written
1/2 Ca(H2 PO 4 ) 2 • H2 O + water = f CaHPO4 4-
(1/2 - f) CaO (in sol.) + (1 - f )/2 P2 Oj (in sol.)
[4]
With the mole ratio CaO/P2O6 in the solution defined as
R, equation [4] yields
R = 2(1/2 - f ) / ( l - f)
from which
f = (1 - R)/(2 - R) (1>R> 0) [5]
In soil systems, R may be evaluated from phase-rule
data, provided the mechanism of dissolution is such that
the solution leaving the granule to enter the soil is con-
tinually saturated with respect to monocalcium phosphate
monohydrate and a dicalcium phosphate. This assumption
is supported by observations described in the companion
paper (17) and by several other points.
1. The solution within the granule rapidly reaches an
equilibrium composition. In an exploratory equilibration
experiment, the solution reached a constant composition
in less than 6 minutes. In contrast, several days were re-
quired for dissolution of the monocalcium phosphate mono-
hydrate from tablets in soil (17).
2. The actively dissolving part of the granule is fairly
well isolated from the soil, because (a) the outer shell of
the granule soon becomes a coarsely porous structure of
drained dicalcium phosphate, and (b) granule-soil contact
usually is limited to a relatively small area.
3. The outward capillary flow of solution sweeps back
materials that otherwise would diffuse into the granule
through liquid bridges.
4. Water entering the granule through vapor transport
does not bring in impurities.
5. Concentrations in the solution within the granule
are high in comparison with those in the soil solution.
These factors combine to limit entry of ions from the
soil which might alter the conditions of hydrolysis of the
monocalcium phosphate monohydrate. The effects of some
of these factors are nullified, however, in water-saturated
soils where blotting is impeded and diffusion of ions into
and out of the granule is promoted.
The hydrolysis process may be visualized as one in
which a small amount of water enters the granule and
quickly establishes an invariant-point composition, the
solution is blotted from the granule, another increment of
water enters the granule, and so on until all the mono-
calcium phosphate monohydrate is dissolved. Dissolution
of the granule then almost ceases because of the low solu-
bility of the residue of dicalcium phosphate. The value
of R to be used in equation [5] under these circumstances
can be calculated from the invariant-point composition for
the dicalcium phosphate that has formed in the residual
phase.
EXPERIMENTAL METHODS
Since compositions along line OM in figure 1 are metastable
with respect to anhydrous dicalcium phosphate, conditions were
selected that would minimize the effects of its precipitation.
High solid/liquid ratios (thick slurries) speeded attainment
of the metastable equilibrium, and a large excess of dicalcium
phosphate dihydrate rapidly replaced the ions removed by any
precipitation of anhydrous dicalcium phosphate.
Equilibration mixtures were prepared with recrystallized
monocalcium phosphate monohydrate and dicalcium phosphate
dihydrate (both acetone-washed) that contained traces of
anhydrous dicalcium phosphate as the only contaminative phase
detectable by petrographic examination.
For the 25° C. isotherm, bottles containing the mixtures
were almost completely submerged in a water bath at 24.98
± 0.01° C. The charges were agitated by stirring or tapping
every few hours and prior to each sampling the first day—twice
a day thereafter. Continuous stirring was avoided lest it en-
hance formation of anhydrous dicalcium phosphate seeds.
Upon equilibration of a mixture, about 5 g. of the liquid
phase was drawn through a fritted glass funnel (at room tem-
erature) into a weighing bottle. Suction of air through the
E Iter was avoided. Solid phases were identified and their rel-
ative amounts estimated petrographically. The liquid phase
was analyzed for calcium gravimetrically (22) and for P by
the AOAC alkalimetric titration method (1).
In locating the metastable invariant point at 0° C., con-
version of dicalcuim phosphate dihydrate to the anhydrous
salt was so much slower than at 25° C. that the procedure was
simplified through use of a lower solid/liquid ratio and
through preparation of only 1 mixture from which several
samples were removed. The bottle of mixture was rotated con-
tinuously in a water-ice mixture through the first 3 samplings,
then was rotated for 30 minutes before each later sampling.
The fritted funnel was packed in ice.
Initial compositions of the mixtures are shown in table 1.
Series A and B were designed to establish the composition
at the metastable invariant point through respective approaches
to equilibrium from supersaturation and undersaturation with
respect to dicalcium phosphate dihydrate. The composition of
the solution used in preparing series B was too close to that at
the metastable invariant point, however, to provide a good
test. Series C and D were respective approaches to points on
the metastable isotherm from supersaturation and undersatura-
tion with respect to dicalcium phosphate dihydrate. Series E
was a determination of the metastable invariant-point com-
position at 0° C.
RESULTS OF EQUILIBRATION EXPERIMENTS
The results are summarized in table 2 and plotted in
figure 1. The average of series A and B compositions is
plotted as point M in figure 1, and the average composi-
tions of the other series are plotted as points C, D, and E,
respectively.
An examination of table 2 shows that the compositions
of nearly all the solutions were constant by the time of the
first sampling. Exceptions are the 144-hour sample of
series A and the 460-hour samples of series C and D. The
composition of the 144-hour sample of series A falls on
stable invariant point S, figure 1, in accord with the petro-
graphic observation that the preponderant dicalcium phos-
phate was the anhydrous salt. A linear extrapolation of the
stable portion of the solubility isotherm for monocalcium
phosphate monohydrate passes through point M. This fact,
along with the reproducibility and constancy of the com-
positions in series A and B equilibrations, leads to the
conclusion that point M is the metastable invariant point.
The average compositions in series C and D equili-
brations (excluding the 460-hour data) fall exactly on a
smooth curve drawn between the origin and point M with
about the same curvature as the stable isotherm—a signifi-
cant fact in that opposite approaches to equilibrium are
represented.
Table 1—Initial composition of equilibration mixtures.
Series Salt Liquid phase
Ca(H2 PO4 )2 • H2 O CaHPO4 • 2HZ O Cao Pj O5 Volume
% % ml.
g. g.
0 15
A 30 10 0
B 10 30 6.25 22.15 15
C 4.052 35 0 0 15
D 0 35 0 8.47 15
E 100 10 0 0 70
 BROWN AND LEHR: APPLICATION OF PHASE RULE TO BEHAVIOR OF MCP IN SOILS

Table 2—Metastable equilibrium in the dicalcium phos-

phate region of the system CaO-P2O5-H2O at 25° C.

Series* Equilibration Liquid phase CaHPOj in

time CaO solid phaset
P!°S
hr. % % %
A 0.5 6.30 21.9 No increase
2 6.25 21.8 No increase
6 6.31 22.0 0
24 6.27 21.9 1
48 6.29 22.1 5
79 6.29 22.0 15
144 5.77 24.4 50
Average^ 6.28 21.9
A' 0.5 6.34 21.8 _
3 6.32 21.8 No increase
3 6.31 22.0 No increase
24 6.29 22.0 0.1
30 6.30 21.7 Tr.
52 6.41 22.3 1 to5
53 6.31 21.9 0.1 to 1
100 6.30 22.0 10
144 6.30 22.0 20
145 6.2S 21.9 25
Average 6.32 21.9
B 3 6.24 21.8 No increase
3 6.31 21.9 No increase
24 6.32 22.0 —
26 6.28 21.9 Tr.
52 6.21 22.1 _
73 6.32 22.0 1
170 8.29 22.0 25
Averagefi 6.28 22.0
C 1 3.30 10.5 No increase
3 3.29 10.4 No Increase CaO,%
27 3.29 10.4 Tr.
28 3.23 10.3 ——
51 3.28 10.4 Tr. Figure 1—Solubility isotherms in the system CaO-P2O5-
51 3.27 10.4 No increase H2O.
123 3.28 10.4 No increase
460 3.39 10.8 1
Averages 3.28 10.4
D 0.5 3.80 12.4 No increase value reported by Flatt and co-workers (13) (19.43% P2OB,
3 3.80 12.3 No increase 6.01% CaO).
30 3.80 12.4 Tr.
53 3.78 12.3 0.1 When the polytherm connecting the stable invariant
460 3.92 12.8 5
Average 3.80 12.4
points at 40°, 25°, 5° C. (19) is extrapolated to 0° C., it
E 0.5 6.12 19.9
misses Flatt's result unless a sharp curvature is introduced
1 6.06 19.9 below 5° C. This suggests that Flatt and co-workers were
3 6.01 19.9
24 6.07 20.1 dealing with the metastable instead of the stable invariant
90 6.11 20.0 point. The difference in P2O5 contents of the two 0° C.
Average 6.07 20.0
invariant points seems too great, however, to be attributed
*8erles correspond to those in table 1.
tEstlmated from petrographlc examination.
to experimental error alone.
{Excluding data for 144 hr. The polytherm for the stable invariant points shows an
SExcludlng data for 460 hr.
almost constant CaO content above 25° C. The CaO con-
The compositions of the solutions in the 460-hour equili- tent increases as the temperature is lowered below 25° C.,
brations fall nearly on the metastable isotherm but at whereas that at the metastable invariant point decreases.
higher concentrations. The change in concentration pre- The results of the present work indicate that the poly-
sumably resulted from evaporation of water around the therms for the stable and metastable invariant points inter-
screw tops of the equilibration bottles. Other processes, sect slightly below 0° C. At this temperature and compo-
quite slow but essential for true metastable equilibrium, sition, monocalcium phosphate monohydrate, dicalcium
may have occurred in the interval between 120 and 460 phosphate dihydrate, and anhydrous dicalcium phosphate
hours—hydration or dehydration of ions (21), for exam- would be in equilibrium with the same solution. Bassett (2)
ple. The shift is not particularly important here; it was concluded that this equilibrium would be found at 21° C.
relatively small and was largely along the isotherm. Points —a conclusion in conflict with (I) the invariant-point com-
C and D, together with metastable invariant point M and positions at 5° and 0° C., which show the invariant point
the origin, adequately delineate the metastable isotherm. for the dihydrate to be metastable relative to the anhydrous
salt down to 5° C. (probably below 0° C.), and (2) the
These experiments provide information on the relative invariant-point compositions at 25° C. when it is consid-
rates of processes occurring in the equilibration mixtures. ered that the dihydrate has a negative temperature coeffi-
When the proportion of anhydrous dicalcium phosphate cient of solubility.
in the solids was estimated at 25%, its surface area, be-
cause of its small particle size, was at least 10 times that RESIDUES—PREDICTED AND OBSERVED
of the dihydrate. Yet the solution had the same composi-
tion as when the anhydrous salt was absent. This means To facilitate comparisons of residues predicted from
that the specific rate of dissolution of the dihydrate invariant-point compositions and equation [5] with resi-
was much higher than the specific rate of precipitation of dues observed in the companion study (17), the known
the anhydrous salt. A corollary conclusion is that the rate- invariant-point compositions are summarized in table 3.
controlling process was the precipitation of anhydrous di- Temperature differences in the range usual in soil have
calcium phosphate, rather than the diffusion of ions. little effect on the amount of residue. A much greater
Conversion of dicalcium phosphate dihydrate to the effect is associated with the extent of hydration of the
anhydrous salt was not detected in series E equilibrations. residual dicalcium phosphate.
The composition at the 0° C. invariant point, E, (20.0% The amounts of residue left at the tablet sites from pure
PoO5, 6.07% CaO) was 0.6% higher in P2O5 than the monocalcium phosphate monohydrate in 5 soils at 2 mois-
10 SOIL SCIENCE SOCIETY PROCEEDINGS 1959

ture levels (table 2 of ref. 17) are compared in table 4 Table 5—Residues from pure monocalcium phosphate
with the amounts predicted. Predictions were based upon monohydrate after 7 months.*
the major residual phase (table 3 of ref. 17), although the
presence of a considerable amount of the other dicalcium Temperature Moisture Fraction of P left at tablet site, %
phosphate made arbitrary some of the choices between •C. equivalent Observed Predicted
redicted values. Observations and predictions agree excel- value t
E ;ntly at 0.5 ME, whereas the observations at 1.0 ME
exceed the predictions by some 9% of the P.
11 1.0
1.8
Humus (-)
25
21
Humus (+ )
27
31
19.9
19.9
30 1.0 29 31 28.9
The proportion of the P left in the residue from tablets 1.8 19 33 21.7
containing gypsum in addition to monocalcium phosphate *From table 6, companion paper (17).
monohydrate in soils at 0.5 ME (exp. 2 of ref. 17) was tFrom table 5t companion paper (17). and table 3, this paper.
Hartsells 22.8, Ruston 23.1, Carrington 25.9, Webster 23.9,
and Dewey 25.2%. The average of these values, 24.2%, Table 6—Compositions of solutions saturated with respect
is 3.7% lower than the prediction of 27.9%, as derived to monocalcium and dicalcium phosphates
from a report (4) that a solution in equilibrium at 25° C. for several systems.
with monocalcium phosphate monohydrate, dicalcium
phosphate (probably anhydrous), and gypsum contains Composition of solution, g. -moles
77 g. CaO, 317 g. P2O5, and 0.6 g. SO3 per liter. (This com- NHt P04'- NO,-
position, with allowance for the gypsum-derived calcium, 8.57 4.04 __ 33.3 __ 298 29 (12)
yields a value of 0.612 for R.) The agreement between 8.95 2.70 -—— 33.3 —— 342 29 (12)
8.05 ——• 6.6 ——
observation and prediction is fairly good; the spread 47.5 —— —
33.3
2.1 93.7
260
317
28
30
(10)
(14)
includes experimental error as well as possible soil 18.3 —— — 26.3 21.0 446 29 (15)
10.1 — 33.3 344 28
interactions. —— —— (H)
The results of the factorial experiment (expt. 4 of ref. 17)
are rearranged in table 5 and compared with predictions soil-granule interaction and the reasons for residues in ex-
for the amounts of residue as listed in table 3. Despite cess of the predictions are not known. Diffusion of calcium
the high levels of moisture, observations and predictions and phosphate ions from the granule at rates somewhat
agree quite well for experiments at the low (—) humus disproportionate to their concentrations would explain the
level, the average difference being 2.1%. Only one of the discrepancies. Such differences would be related to the
predictions at the high (-J-) humus level is in fair agree- specific mobilities of the ions or to distortions of concen-
ment with observation. tration gradients caused by reactions of the ions with con-
To summarize the comparisons, predictions and obser- stituents of the soil.
vations agreed fairly well for experiments at 0.5 ME and A high level of moisture favors entry of foreign ions into
for about one-third of the experiments at 1.0 or higher the granule, with possible shift of the equilibrium compo-
ME. In two-thirds of the experiments at the high moisture sition to a lower value of R, but the low ionic concentra-
levels, particularly when the soils contained about 10% tions in the soil solution make unlikely an effect as large
leaf mold, the values observed were considerably higher as the one observed. In some of the comparisons where
than those predicted. dicalcium phosphate dihydrate was the residual phase,
The highly significant effects of humus and of the agreement between observation and prediction would have
second-order interaction humus-moisture on the amount been better had the anhydrous salt been the residual phase
of residue (17) constitute, in themselves, sufficient evidence that controlled the composition of the solution. It might
that the granule system cannot be considered as isolated be argued that the original residue was the anhydrous salt
from the soil under all conditions. The exact nature of the —that it hydrated before the petrographic examination
was made. Even when the humidity is near 100%, how-
Table 3—Invariant-point compositions at ever, as in these experiments, the anhydrous salt was more
various temperatures. stable than the dihydrate, possible exceptions being the
experiments at 11° C., where the relative stabilities are
Temperature Stable Metastable not known.
CaO P!OS f CaO P,0S f
•c. % %
SALT EFFECTS
% % % %
0 __ __ .__ 6.07
——
20.0
——
18.7
——
Phase rule can be applied in an analysis of the effects
5 5.86 21.54 23.7
11 —— —— 26.3* —— —— 19.9* of other salts on the uptake of phosphorus by plants from
25 5.79
——
24.3
——
28.4 6.30
——
22.0
——
21.5
21.7*
superphosphate to the extent that the effects are related to
30 28.9*
40 5.79 26.47 30.9 —— —— —— the amount of residue and the composition of the solution
'Estimated from values at other temperatures.
entering the soil. Ideally, the dissolution process for the
mixture in question is examined in detail to determine
Table 4—Residues from pure monocalcium phosphate which phase-rule data are needed, then the composition
monohydrate. of the solution in equilibrium with monocalcium phosphate
monohydrate and the appropriate dicalcium phosphate is
Fraction of P left at tablet site, % established through a phase-rule study of the system. The
Moisture Moisture examinations reported here for tablets containing gypsum
equivalent, 0,5 equivalent, 1.0
Observed*
appear to be the only such studies that have been made
Hartsells 28 30
of salt effects. Speculation about certain salt effects is
Ruston
Carrington
27 28 warranted, however, in the light of chemical considerations
28 31
Webster 40J 34 and some phase-rule data that are available.
Dewey 30 30
Average 28 30 The CaO/P2OB ratio of the solution in equilibrium with
Predictedt monocalcium phosphate monohydrate and a dicalcium
28.4 21.5 phosphate generally should be affected little by a salt,
*Room temperature. such as KC1, that neither hydrolyzes strongly nor introduces
T25' C. a common ion. A salt such as Ca(NO3)2, NH4H2PO4, or
{High result caused by dicalcium phosphate in soil fragments adhering to tablet
residue; this value was omitted in the comparison. KH2PO4 may introduce a considerable common-ion ef-
 BROWN AND LEHR: APPLICATION OF PHASE RULE TO BEHAVIOR OF MCP IN SOILS 11

feet, whereas materials such as calcite will dominate the downward shift of the isotherms (figure 1) will cause some
situation. of the dicalcium phosphate to dissolve.
Some representative compositions are listed in table 6 More information is needed on the rates at which the
for solutions reported (10 to 12, 14, 15) to be saturated invariant-point solutions react with constituents of soils
with respect to monocalcium phosphate monohydrate and and on the kinds of crystalline materials that may be
a dicalcium phosphate (presumably CaHPO4) at 25° G. formed. The work of Lindsay and Stephenson (18) pro-
in 4-component systems comprising CaO, PaOs, and H2O, vides valuable information of this type.
together with ammonium, potassium, or nitrate. The calcu-
lated values of f for the 4-component systems are surpris-
ingly close to that for the ternary system CaO-P2Og-H2O
(at bottom of table). Here, the effects of the fourth-
component ions are so small that an examination of the
details of the dissolution process may not be important.
The values of f in table 6 were calculated on the
assumptions that the nitrate, potassium, and ammonium
were derived from Ca(NO3)2, KH2PO4, and NH4H2PO4,
respectively. The secondary salts of potassium and ammo-
nium would yield slightly higher values of f—for example,
31% instead of 29% for the first entry.
POSSIBLE PRECIPITATION REACTIONS IN SOIL
The solution leaving a source of monocalcium phosphate
monohydrate may encounter additional Ca++ ions in the
soil, with resultant precipitation of dicalcium phosphate,
but an external source of Ca is not essential to the precipi-
tation. At least three other processes may contribute.
1. As it enters the soil, solution of composition M, fig-
ure 1, is supersaturated with respect to anhydrous dical-
cium phosphate. Aside from reactions with the soil, which
perhaps will occur faster, anhydrous dicalcium phosphate
can precipitate from the solution until the composition
intersects the isotherm for the anhydrous salt. About 8%
of the P originally in the granule can be precipitated
in this way.
2. A partial precipitation through direct reaction with
soil-derived Ca entails the precipitation of additional di-
calcium phosphate. Assume that soil-derived Ca shifts the
composition from Q! to Q2 in figure 1. Dicalcium phos-
phate dihydrate will precipitate until composition Q3 is
reached. Since the Ca content at Q3 is less than at Qi, the
precipitation of Ca exceeds the addition from the soil.
3. Removal of P from the solution through reaction
with ions foreign to the system CaO-P2O5-H2O to change
the composition, say, from Pj to P2 in figure 1 will lead to
precipitation of dicalcium phosphate until the composition
reaches P3. Cations such as Fe+ + + and A1+ + + might
lead to this situation. Their principal sources (oxides, hy-
droxides, carbonates, silicates, and exchangeable combina-
tions) introduce impurities also, but the amounts probably
are small compared with the concentrations of Ca and P.
We have observed a precipitation of anhydrous dicalcium
phosphate in a few soils high in iron oxides and purported
to have high P-fixing capacity.
Since metastable equilibrium along the dicalcium phos-
phate dihydrate isotherm is reached quickly, the composi-
tion of the solution may follow the isotherm as the solution
reacts with the soil. The fraction of the P that would be
precipitated as dicalcium phosphate in process 3 is esti-
mated, through graphical integration, at 21% along the
stable isotherm or 25% along the metastable isotherm.
These proportions, together with the dicalcium phosphate
left at the site of the granule, can account for nearly half
the P applied as monocalcium phosphate monohydrate.
Inorganic and organic materials which do not dissolve
can provide exchange sites capable of removing Ca+ +
ions and displacing the composition to the left of the iso-
therm, thus decreasing the amount of dicalcium phosphate
that may precipitate. The amounts of dicalcium phosphates
that may precipitate are influenced also by shifts in the
positions of the isotherms caused by materials dissolved
from the soil. A material, for example, which causes a
12 SOIL SCIENCE SOCIETY PROCEEDINGS 1959