4/29

T H E O R Y O F P IL E S IN S W E L L IN G A N D S H R IN K IN G SOILS

T H E O R IE DES P IE U X DANS DES SOLS S U J E T S A U G O N F L E M E N T OU AU R E T R A IT

T E O P M H C B A fï B H A B yX A lO lU H X 11 Y C A flO IH b lX T P y H T A X

H .G. P O U L O S & E .H . D A V IS , C iv il E n g in e e rin g , U n iv e r s ity o f Sydney ( A u s tra lia )

SYNOPSIS. Piles are frequently used for foundations in swelling and shrinking soils to sup­
press structural movements. Current methods of analysing such foundations usually only con­
sider the ultimate uplift or downward load capacity of the pile and little or no consideration
is given to prediction of the pile movements. In this paper, a method of analysing the m ove­
ment of a pile in swelling soil is presented. The analysis is based on the simplifying
assumption that the soil can be treated as an elastic material but modifications to the analysis
are described which allow consideration of such factors as no n-h omogeneity, slip at the pile-
soil interface and crushing or tensile failure of the pile.

In order to illustrate the major factors affecting pile movements, a number of idealized
theoretical solutions are presented. Comparisons between the predicted characteristics of b e ­
haviour of piles in swelling soils show encouraging agreement with the behaviour observed in a
number of full-scale field tests.

INTRODUCTION c) the movement of the piles due to the net

5 effects of uplift forces and the structural
»■Foundations in expansive clays are frequently load must be less than the prescribed
subjected to severe movements arising from limit.
moisture changes within the clay with conse­
quent cracking and damage due to distortion. Attention is focussed on the latter two as­
Piles have been used extensively for founda­ pects since few problems arise in obtaining
tions in swelling soils in order to anchor adequate load-carrying capacity in expansive
the structure down at a depth where changes in soils. An analysis based on elastic theory
moisture content are negligible, so that mo v e ­ is described, and typical results are given.
ments of the structure are minimized. However, Some comparisons between theoretical and ob­
considerable uplift forces are then induced in served pile behaviour are also presented.
such piles due to the action of the swelling
soil. An analogous problem arises with piles Existing methods of analysis of piles in
in soils undergoing shrinking or consolidation, swelling and shrinking soils are generally
When downdrag forces are induced in the piles confined to the estimation of the forces in­
due to negative friction. In this paper, the duced in a pile by soil movement. Typical of
effect of soil movements on the behaviour of such approaches are those described by Collins
floating piles will be considered. Because of (1953), Mariotti and Khalid (1969) and
the similarity in approach between a pile in Bozozuk (1972), which in general assume that
a swelling soil and a pile in a consolidating full slip occurs between pile and soil along
or shrinking soil, attention will be concen­ the shaft. Sahzin (1968) has obtained ex­
trated on the case of a swelling soil. pressions for the movement of a pile in swell­
ing soil by considering the work done by
In designing piles in swelling soils, there friction forces in the upper portion of the
are three requirements - pile tending to lift the pile and the applied
load and the friction forces in the lower half
a) the pile must be able to carry the struc­
of the pile tending to resist uplift. A more
tural load safely i.e. there must be ade­
satisfactory analysis can be carried out by
quate ultimate load capacity.
employing elastic theory in a similar fashion
b) the pile must have sufficient tensile to that described by Poulos and Davis (1968)
strength to withstand the tension develop­ for pile settlements and Poulos and Mattes
ed in the pile due to uplift forces. (1969) for negative friction on end-bearing
piles.
169
4/29

THEORETICAL ANALYSIS
P„ + I Pi"d L/n + pb TTdb 2/4 = 0 (3)
Basic Analysis
The problem is illustrated in Fig.l,A cir- where P Q = applied downward load on pile top.
cular pile,length L, diameter d a nd base di­
ameter dfe, is situated in a soil mass in which, Equations (2) and (3) may be solved to obtain
away from the pile, a general specified dis­ the displacement p and the distribution of
tribution of movement, S (either swelling or shear stress p from which the load in the pile
shrinking) with depth occurs. The pile is P at any depth can be calculated.
divided into n cylindrical elements each with
a uniform shear stress pj acting on the peri­ The above basic analysis fails to take account
phery. In the basic analysis the soil is of several factors which are likely to be im­
assumed to be homogeneous and linearly elas­ portant in real situations. Modifications can
tic and it is assumed that no slip occurs at be made to allow for these factors.
the pile-soil interface.
Local Failure Between Pile and Soil

The effects of pile-soil slip along the shaft

JL* can be allowsd for by specifying a limiting
value of shear stress, xa , at each element
along the pile. T a will usually be expressed
in terms of effective stress by the Coulomb
expression. In addition, a maximum value of
pase stress pj, can be specified, corresponding
to the ultimate bearing capacity cf the base
Tm TtTT if px is negative (see sign definition in
u -H-st Fig.l) or to the ultimate uplift capacity of
w Pb
the base (usually taken as zero unless tne
(a) Pii« (b) Stressas on (c) Stressas on (d) Spacifiad Soil
Pila Soil Movamant base is enlarged) if p>, is positive. The solu­
tion is re-cycled until tixe shear stresses and
FIGURE 1 PILE IN SWELLING OR CONSOLIDATING SOIL. base stress do not exceed these limiting values
(Poulos and Davis,1958).It should be noted
Defining downward soil displacements as Dosi- that if full slip occurs along a pile, the
tive,and shear stresses positive as s h o w n i n load distribution in the pile will remain con­
Fig.1, the soil displacements along the pile stant and the movement of the pile subsequent
can be expressed as to full slip will be the same as that of the
soil adjacent to the pile at the point of shear
(sr/=-l, r i j {»)*(«) <-> reversal.
where /Ig/ = the (n+1) by (n+1) matrix of dis­ Compression Failure of Pile
placement influence factors.
Eg = Young s modulus of soil. When the soil is shrinking,allowance can be
{S? = the (n+1 ) vector of soil move­ made for compression failure due to downdrag
ments, negative for swelling. on the pile by including in the computation
{s p} = the (n+1 ) vector of soil dis­ the requirement that,when the load in any
placements adjacent to pile. pile e l e m m t reaches the compression strength
{p } = the (n+1 ) vector of shear stresses of the pile, it remains constant at this value. 5
at the soil-pile interface and Since the load on such a crushed element must S
the base pressure, positive be the same at the top as at the bottom of the
directions as in Fig.l. element, the shear stress between this element
and the soil drops to zero. A redistribution
The elements of [ls ] are obtained by double of the shear stresses elsewhere at the soil-
integration of the Mindlin equations as des­ pile interface therefore occurs and there is
cribed by Poulos and Davis (1968). Poisson's an increase in pile displacement. The pro­
ratio of the soil v s is not a very important cedure is explained more fully in Poulos and
parameter of the values of Is . Davis (1972).

Without slip the pile displacements, pP = sp. Tension Failure of Pile

For the general case of a compressible pile,
the pile displacements must be compatible When the tensile load reaches the tensile load
with the elastic properties of the pile, and capacity of the pile, the pile is assumed to
the analysis could proceed along the lines fracture and in effect becomes two piles. Two
given in Mattes and Poulos (1969). Here only equilibrium equations now apply:
the simple case of an incompressible pile is
considered and therefore pp = g p = p = con­ a) for the upper fractured portion
stant. It follows that
(4)
d ip} = E g [Is]_ 1 {p-S} (2)
\lso, from equilibrium of the pile, where m = number of elements in fractured por­
tion.
170
 4/29

b) for the lower portion from 0.36ES for L = 5d to 0.64ES at I = 20d;

these values correspond approximately to a
irdb
(5) •soil having <(>1 = 30°.

A new variable, the displacement of the frac­

tured portion of the pile, is now introduced
so that the n+3 equations may now be solved
for the n unknown shear stresses, the base
pressure, and the displacements of the upper
and lower portions of the pile.

Non-Uniform Soil H- oieJ—

PH« Soil Pile-Sofl
Movement Strength
Approximate allowance for the effect of vari­
ation of the modulus of elasticity of the soil
with depth can be made by substituting the
matrix [Is/E s ] for [I? ]/Es in E g . (1). By this
procedure, the soil displacement at a particu­
lar depth will be calculated as if the modulus
at that particular depth was also the modulus
at all other depths. This approximation will
clearly be unsatisfactory when the variation
of modulus from top to bottom is very large
or there are sudden major changes.

Variation wittt Time

For cases in which the soil movements are
time-dependent and the variation of pile loads
and displacements with time is required a con­
solidation analysis may be combined with the -Uniform diameter pile d ^ =1
above analysis (Poulos and Davis, 1972). -Enlarged base d^j=2
Alternatively, appropriate values of soil dis­
placement S may be input at each time consid­ FIGURE 2 EFFECT OF PILE LENGTH AND BASE
ered, and the solution carried out as before. DIAMETER.

THEORETICAL TRENDS IN PILE BEHAVIOUR Fig.2 shows that as would be expected, the
pile movement decreases as the pile length in­
To illustrate the influence of various factors creases. iVhen the pile is entirely situated
on pile behaviour,a number of solutions far in the swelling zone (L=5d),movement of the
relatively idealized cases are examined. In pile continues after full slip has occurred
all cases, the pile is assumed to be incom­ aloncf the shaft. For piles founded below the
pressible, ten elements have been used to div­ zone of swelling a limiting pile movement is
ide the pile, and the soil modulus along the reached after a certain soil movement occurs.
pile is uniform, \is being 0.3. Attention is The advantage of founding a pile below the
concentrated on piles in swelling soils. As swelling zone is obvious. The maximum tensile
long as elastic conditions are preserved the load in the pile- generally increases markedly
solutions also holds for shrinking soils ex­ as the length increases; relatively small
cept for a change in sign but, if failure loads are developed when the pile is entirely
occurs between soil and pile or within the within the swelling zone.
pile, the solution for shrinking soil may be
somewhat or even significantly different. In The presence of an enlarged base leads to a
all the cases examined slip between soil and decrease in pile movement although the effect
pile starts to occur at the top of the pile is relatively small, especially for L = 5d
and requires a very small swelling movement and L = 20d i.e. when the pile is entirely
to initiate it. in the swelling zone, or anchored well below
the swelling zone. In the latter case, the
The Effect of Pile length and Base Diameter enlarged base has virtually no effect. The
corresponding maximum loads are considerably
For a given soil profile with a linear dis­ greater for the enlarged base piles except for
tribution of soil swelling from Sp at the sur­ the L = 20d pile. It is therefore apparent
face to zero at a depth of lOd, the variation that the enlarged base has the greatest in­
of pile movement and maximum pile force with fluence when the pile is situated at or near
increasing soil movement is shown in Fig.2 the bottom of the swelling zone, and that the
for three different pile lengths. For each most efficient means of reducing pile movements
length, both a uniform diameter pile and a is either to use a uniform diameter pile
pile with a base diameter twice the shaft di­ founded well below the swelling zone (of
ameter is considered. The pile-soil interface length about twice the depth of this zone) or
shear strength xa varies linearly from zero at to use an underreamed pile founded at or just
the surface to 0.01ES at a depth of 20d, and below the bottom of the swelling zone.
the base bearing capacity is assumed to vary

171
4/29

For the uniform diameter piles considered in

Fig.2, the load distributions are shown in
Fig.3 for various values of dimensionless soil
movement SD/d. As the pile length increases,
the load in the pile increases and the d i s ­
tribution of load also changes; the relative
position of the maximum load moves towards the
top of the pile. For the L=5d and lOd piles,
slip occurs along the whole length of the pile
at relatively small soil movements. For the
L=20d pile, no change in load occurs after
SQ/d reaches about 0.12 but slip only occurs
along the upper half of the pile.

is°

—[OlEsI—
Pile Soil Pile-Soil
Movement Strength

0 0 S. Values of
\ Values of z\ Values of z
L L L V 0 ' 0 8
0-5 ^V-^003' 0-5 0-5
^ ^ £ > •0 6 12
°'01) 002 )

10
lr ,
. .1 0
-0 005 -0010 0
P P
Esd2 E,d* (b) Maximum Pile Load
&
L =5d (b) L =i0d (c) L =20d FIG U R E 4 INFLUENCE O F PILE DIAMETER
FIGURE 3 TYPICAL LOAD DISTRIBUTIONS

The Effect of Pile Shaft Diameter

For a given pile length and soil swelling pro­

file, the effect of pile diameter is shown in
Fig.4. As the diameter increases, the pile
movement decreases but the maximum load in­
creases. However the rate of decrease of pile
movement is almost negligible for diameters
greater than about 0.0 3L, and even a slender
pile (d=0.01L) moves only about 20% more than
a relatively large-diameter pile (d=0.2L). Strength S w elling w ith Depth
The theory therefore suggests that small-
diameter piles founded well below the swelling
zone, can satisfactorily suppress upward mov e­
ments in swelling soils. Donaldson (1967a),
has described the successful use of small-
diameter piles to support conventional brick
buildings on expansive soils in South Africa.

The Effect of Distribution of Soil Movement

The effect of the shape of the soil swelling

profile, for a given depth of movement, is
shown in Fig.5. Both the pile movement and
the maximum pile load are greatest when the
soil movements decrease slowly with depth near
the top of the pile (case iii) and least when
the soil movements decrease rapidly (case ii).

FIGURE 5 INFLUENCE O F SOIL S W E L L I N G

PROFILE.

172
 4/29

The Effect of Pile-Soil Strength Distribution

In Fig.6 the effect of a uniform distribution

of soil-pile strength on the behaviour of a
pile with a length to diameter ratio of 20 is
compared with that of a strength distribution
which increases linearly with depth but has
the same average value. For a given soil
movement, the pile movement and maximum pile
load are considerably greater for the uniform
case.

FIGURE 7 EFFECT O F AXIAL LOAC ON PILE

MOVEMENT
T ' 1
-d 10d| I/
L=20d:
\ upward more rapidly than before failure, while
I \
.O O lE s.1
the lower part, according to the analysis,
suffers a small downward movement at failure
0 0 0 5 £ jH -
(i) (ii) and than an upward movement as the soil move­
Pile Soil Pile-Soil Strength ment increases. As the tensile strength in­
Movement creases, the soil movement required to cause
tensile failure increase also. In this example,
F IG U R E 6 EFFECT OF PIL E - S OIL STRENGTH
the maximum tensile load developed in the pile
DISTRIBUTION
is 0.393 E sd 2 so that tensile failure will not
occur if the tensile strength exceeds this
The Effect of Axial Load value.

The effect of axial load on the pile movement

is shown in Fig.7. In the case considered,
the downward movement increases almost linear­
ly with increasing axial load. Comparison of
Figs.7 & 2 shows that the axial load required
to prevent upward movement is only about half
of the maximum tensile load developed in the
uploaded pile. Also shown in Fig.7 are the
pile movements calculated on the assumption
that the effects of axial load (Poulos and
Davis, 1968) and soil swelling can be super­
posed. Although superposition is not strictly
valid in this case because slip occurs along <°
> Ve,a* ■
02 <w 3
part of the pile shaft, it nevertheless pro­
vides an approximate estimate of pile movement. FIGURE 8 EFFECT OF TENSILE FAILURE OF PILE,
Taking account of pile-soil slip in the solu­
tion for axial load would lead to increased
movements and closer correlation with the It should be noted that the accuracy of solu­
complete solution. tions in which tensile failure occurs depends
to a large extent on the previous history of
The Effect of Tensile Failure of the Pile the pile, since, once failure has occurred
and the pile has separated, it remains separ­
An example of the effect of tensile failure ated. If too large an increment in soil move­
on pile movement is shown in Fig.8, where ment is taken, the point of separation may be
pile movement is plotted against soil surface inaccurately computed; thus, small increments
movement for two tensile strengths. When are desirable for accuracy.
tensile failure occurs, the upper portion moves

173
4/29

APPLICATION OF THEORETICAL ANALYSIS TO in K s of 1.0 to 1.5 in such situations. The

Pr a c t i c a l p r o b l e m s cohesion term is also not necessarily neglig­
ible. For soils swelling from an initially
In applying the theoretical analysis described unsaturated condition the effective stress
in the previous sections to practical problems, concept is of doubtful validity and, as a
estimates are required of the following quant­ practical approach it may be better to use the
ities : total instead of the effective overburden
6tress in e q . (6) and appropriately determined
(i) the soil movement profile
values of adhesion and pile-soil friction
(ii) the pile-soil interface strength
angle. If the soil is severely cracked before
(iii) the soil modulus.
swelling starts, the high values of Ks given
above may not be attained in the earlier
Prediction of Soil Movement Profile
stages of swelling, but, for design purposes,
it would be unwise to rely on the reduction in
Two cases may be considered:
upward force on the pile this implies.
(a) movement in saturated soils
(b) movements in unsaturated soils. Soil Modulus

The first case applies to many problems invol­ The most satisfactory means of estimation is
ving consolidating soils and soil movements to backfigure E s from a pile load test in-situ,
may be predicted at various depths by conven­ using the theoretical solutions for an axially-
tional methods of settlement analysis. Methods loaded pile. If such a load test is not pos­
for predicting movements in unsaturated soils sible, a rough estimate may be made by using
are not as well established, and a variety of the correlations between E s and the undrained
approaches has been suggested e.g. Salas and cohesion c„ of the clay projbosed by Poulos
Serratosa (1957), Blight (1965), De Bruijn (1971). The value of cu should be that appro­
(1961). Van der Merwe (1964). Approximate priate to the final moisture content of the
methods for estimating variations of pore soil.
suction, and hence soil movements, due to
moisture changes have been described by COMPARISONS BETWEEN OBSERVED AND THEORETICAL
Richards (1965) and Blight (1965). BEHAVIOUR

Pile-Soil Interface Strength An extensive series of observations on the

movements of thirty-seven houses in South
It is commonly assumed that the shear -strength Africa founded on expansive soils was present­
between soil and pile increases with depth, ed by Collins (1958), twenty-four of these
approximately in proportion to the overburden beinc founded on underreajned piles. The
pressure. In situations where the soil is houses were divided into two groups, on e in
normally consolidated and is going to reiftain which the average load per pile was 5 tons and
saturated (zero air voids but not necessarily the othar in. which the average load was 10 tons,
positive pore water pressure), the limited and each house was founded on piles taken on a
experience from measurements of negative fric­ specific depth. It is interesting to note that
tion on piles in soft normally consolidated the bouses on piles, founded at 30 ft. depth
clays (for example Johannessen and Bjerrum, showed a greater movement than those at 20 or
1965) may be taken to be reasonably applicable 25 ft. It was subsequently found that tne 30
to the case when the pile is affected by ft. piles had failed in tension. Considerably
shrinking or consolidation of the soil. This smaller movements were experienced with the
experience suggests that, in the equation piles carrying 16 ton loads. Theoretical pre­
dictions of the behaviour of the piles at the
Leeuhof site were made using the measured soil
movement profile, the drained strengths to ^
y/here ca ' = adhesion estimate the distribution of pile-soil strength
4>a ' = effective pile-soil friction angle xa and the undrained strength to estimate the
av ' = effective overburden stress distribution of soil modulus E s . A summary of
Ks = a coefficient of horizontal earth the measured and predicted pile movements is
pr ess ure , shown in Fig.9, together with the chosen para­
meters for the analysis. The agreement is good
the combined term K s tan <t>a ' usually lies with­ for the piles with 5 ton loads but for the 10
in the range 0.2 to 0.3 and that ca ' can be ton loads, the theory somewhat over-estimates
neglected. For <J)_1 “ 20° this corresponds to the heave. It is significant to note that the
a range in K s of 0.5 to 0.8, i.e. somewhat theory predicts that tensile failure occurs
greater than the coefficient of earth pres­ for the 30 ft. piles, resulting in larger pile
sure at rest KQ . movements. This prediction is substantiated
by the observations.
For swelling situations, especially when swell­
ing occurs from relatively arid conditions, the Donaldson (1967b) has described an instrumented
soil must be in an overconsolidated state and, test pile which was also installed at Leeuhof
corresponding to the higher values of K0 for South Africa and in which the development of
overconsolidated soils, a higher value of K s tensile force with time was measured. The
may be expected. South African experience pile was a 9 in. diameter concrete pile, 34 ft.
(Collins, 1953; Baikoff and Burke, 1965; long, with a central steel pipe core containing
Donaldson, 1967b) appears to suggest a range strain gauges and acting also as tensile re-
174
 4/29

Tbpsoil and
laterttic soil. Tensile Force (Kips)
0 3 10 15 20
Depth Clay with 0
ft. sllckensides.

5
Yellow
bedded Es lb/sq. In xQ ib/sq.ln
clay.
Vfarved (a) Assumed Parameters in Analysis. 10
B cloy.
Organic cloy.
Shale
Soil Profile Depth
ft. 15

20
Pile Pile
Length Length
ft. ft.

25

(b) Piles with 5 ton (c) Piles with 10 ton

Loads. Loads. 30
-----Theoretical
o Measured (Coltins,1958)
FIGURE 9 COM PA RISO N S BETWEEN MEASURED AND
PREDICTED PILE MOVEMENTS
FIGURE 10 C O M P A R I S O N B E T W E E N
M E A S U R E D A N D PREDICTED
F O R C E IN PILE.
in f or c e m e n t .Heave of the Bail around the
pile was a ccelerated by f il l i n g , w i t h water, pile founded just below the swelling
fou r holes of 4 in.diameter,drilled t o zone.
depths of 1 5 ft. at 3 ft. radius f r o m the
pile, a n d hence f l oo d in g the site. The ten­ (iii) small-diameter piles founded well below
s i l e force was f o u n d bo increase significan the swelling zone are as effective in
tly as the heave incre a se d .A theoretical cal­ suppressing upward movements as larger
culation w a s carried out f o r the latest load diameter piles.
distribution above the test pile, using soil
JJ data similar to that used for Collins' tests, (iv) an applied axial load of about half the
fc The soil surface movement remote from the pile maximum developed tensile load is re­
teas assumed to be 9 mm. A comparison between quired to prevent upward movement of the
the theoretical and measured load distributions pile.
Is shown in Fig.10. Within the obvious limi­ (v) tensile failure of a pile leads to an
tations imposed by the uncertain soil data, increased rate of upward movement if the
the agreement is satisfactory. soil movement continues to increase.

CONCLUSIONS Some suggestions have been made regarding the

estimation of the input data required for the
An analysis has been presented for the move­ analysis of practical cases. Reasonable agree­
ment of, and load distribution in, a pile in ment between theoretical and observed behaviour
a swelling or shrinking soil. The analysis has been found in some reported field cases.
is based on elastic theory, but can be mod i ­
fied to take account of such factors as pile- ACKNOWLEDGMENT
soil slip and tensile or compressive failure
of the pile. From typical solutions obtained The work described in this paper forms part ot
for idealized cases, the following main con­ a general programme of research into the
clusions have been drawn: settlement of all types of foundation and is
supported by the Australian Research Grants
(i) in many cases, there is little differ­
Committee.
ence between the behaviour of a pile in
a swelling or a shrinking soil, apart
from a change of sign in stresses and
displacements.
(ii) the most efficient means of suppressing
pile movements is either to use a uni­
form diameter pile founded well below
the swelling zone or an underreamed

175
4/29

REFERENCES POULOS, H.G. & MATTES, N.S. (1969). "The

Analysis of Downdrag in End-Bearing Piles-'.
BAIKOFF, E.M.A. & BURKE, T.J. (1965), "Deter­ Proc.7th Int.C o n f .Soil Mechs, Vo l . 2, p . 203.
mination of Foundations for Heaving Soils".
Trans.Sth.African Instn.Civ.Engrs, Vol. 7, POULOS, H.G. (1971). Discussion to "Load-
No. 8. Deformation Mechanism for Bored Pil e s ”, by
R.D. Ellison, E. D'Appolonia & G.R. Thiers.
BLIGHT, G.E. (1965). "The Time-Rate of Heave Jnl. Soil Mechs. & Fndns.Divn, ASCE, Vol.97,
of Structures on Expansive Clays". Moisture N o .S M 1 2 , p . 1716.
Equilibria and Moisture Changes in Soils Be­
neath Covered Areas, p. 78. Butterworth, RICHARDS, B.G. (1965). "An Analysis of Sub­
Australia. grade Conditions at the Horsham Experimental
Road Site using the Two-Dimensional Diffusion
BOZOZUK, M. (1972).-"Downdrag Measurements on Equation on a High-Speed Digital Computer".
a 160 ft. Floating Pile Test Pile in Marine Moisture Equilibria and Moisture Changes in
Clay". To be published. Soils Beneath Covered Areas. Butterworths,
Sydney, p . 24 3.
COLLINS, L.E. (1953). "A Preliminary Theory
for the Design of Underreamed Piles". Trans. SAHZIN, V.S. (1968). "Method of Determining
Sth African Instn.Civ.Engrs, V o l . 3, N o . 11. the Magnitude of Rise of Piles Following
Flooding of Swelling Soils". Soil Mechs &
COLLINS, L.E. (19581 . "Some Observations on Fndn.Eng, No.l, p . 49.
the Movement of Buildings on Expansive Soil in
Vereeniging and O dendaalsrus". Trans.Sth. SALAS, J.A.J. fit SERRATOSA, J.M. (1957).
African Instn.Civ.Engrs, Vol.7, No.9. Foundations on Swelling Clays". Proc. 4th
Int.Conf.Soil Mechs. Fndn.Eng, Vol.l, p.424. 5
DE BRUIJN, C.M.A. (1961). "Swelling Character­ V
istics of a Transported Soil Profile at Leeuhof, VAN DER M E R W E , D.H. (1964). "The Prediction
Vereeniging (Transvaal)". Proc. 5th Int.Conf. of Heave from the Plasticity Index and the
Soil Mechs.F ndn .En g., Vol.l, p.43. Percentage Clay Fraction". The Civil Engr.
in Sth.Africa. Vol6, N o . 6, p . 103.
DONALDSON, G.W. (1967a). "The Use of Small-
Diameter Piles in Expansive Soil". Proc. 4th
Reg.Conf. for Africa on Soil Mechs. & Fndn.
Eng., Vol.l, p.249.

DONALDSON, G.W. (1967b). "The Measurement of

Stresses in Anchor Piles". Proc. 4th Reg.Conf
for Africa on Soil Mechs & F nd n.E ng. , Vol.l,
p. 253.

JOHANNESSEN, I.J. & BJERRUM, L. (1965).

"Measurements of the Compression of a Steel
Pile to Rock Due to Settlement of the Surround­
ing Clay". Proc. 6th Int.Conf.Soil Mechs,
V o l .2, p . 261.

MARIOTTI, M. & KHALID, R. (1969). "Comporte-

ment de Pieux dans des Sols Surconsolides
Expansifs". Proc. 7th Int.Conf.Soil Mechs,
Vol.2, p. 165.

MATTES, N.S. & POULOS, H.G. (1969). "Settle­

ment of a Single Compressible Pile". Jnl.Soil
Mechs. & Fndns.Divn, ASCE, Vol.95, No.SMI,
p . 189.

POULOS, H.G. & DAVIS, E.H. (1968). "The

Settlement Behaviour of Single Axially-Loaded
Incompressible Piles and Piers". Geotechnique,
Vo l . 18, p . 351.

POULOS, H.G. & DAVIS, E.H. (1972). "The

Development of Negative Friction with Time in
End-Bearing Piles". To be published in Aust.
Geomechanics Jnl, Vol . 2.

176