NUMERICAL ANALYSIS OF SOILS IN SIMPLE SHEAR DEVICES
MuNIRAM BunHu* and ARUL BRITTo**
ABSTRACT
The results of finite element analyses of soils in simple shear devices, assuming that these soils can be modelled either as an elastic or an elastoplastic material, are presented. The results indicate that an elastic analysis produces larger levels of stress concentrations than an analysis using the modified Cam-clay model. The predicted stress-strain behavior of a very loose sand and speswhite kaolin using the modified Cam-clay model agrees very well with simple shear test results deduced from measurements made at the sample core of the top boundary of the samples for constant load tests. A satisfactory match of experimental and modified Cam-clay stress-strain results was not obtained for the constant height test. The simple shear devices (Norwegian Geotechnical Institute and Cambridge types) can be expected to give good quality results for monotonic loading from carefully prepared samples if measurements of stress and pore water pressures are made at the sample core on either the top or bottom or both horizontal boundaries of the sample.
Key words : clay, deformation, direct shear test, sand, stress distribution, stress-strain curve, (non-uniformities) (IGC : D 6)
INTRODUCTION
Laboratory apparatus which can impose a state of plane strain and allow the rotation of the principal axes of stress are desirable since this condition appears to simulate the stress state to which soils are subjected in many practical situations. For example, soils adjacent to a friction pile or beneath the foundation of an offshore platform can be expected to deform in a manner similar to simple shear strain. Simple shear strain is
a form of plane strain in which an initially cuboidal element is transformed into a parallelepiped without change in volume as shown in Fig. 1. For a uniform stress state to occur under simple shear strain, equilibrium demands that complementary shear stresses be developed on the vertical sides of the element normal to the plane of deformation.
Two apparatus with different purposes have been developed to supposedly impose simple shear strain to soil samples. One, originally proposed by Roscoe (1953) for research and
* Assistant Professor, Dept. of Civil Engineering, State University of New York at Buffalo, Buffalo, NY 14260, USA.
** Senior Research Associate, University of Cambridge, Cambridge, UK, formerly Visiting Assistant Professor, SUNYAB, NY 14260, USA.
Manuscript was received for review on May 27, 1986.
Written discussions on this paper should be submitted before January 1, 1988, to the Japanese Society of Soil Mechanics and Foundation Engineering, Sugayama Bldg. 4 F, Kanda A waji-cho 2-23, Chiyoda-ku, Tokyo 101, Japan. Upon request the closing date may be extended one month.
2. Cambridge simple shear apparatus
TOP PLATEN
REINFORCED RUBBER MEMBRANE
Fig. 3. Norwegian Geotechnical Institute apparatus
continuously upgraded at Cambridge University, tests a cuboidal sample (100 mm x 100 mm x 20 mm high) between rigid boundaries (Fig. 2). Simple shear strain is applied through two hinged end flaps which rotate when the bottom boundary of the device is horizontally displaced. The inner walls of these end flaps are smooth so that significant shear stresses cannot be generated there. The other, originally proposed by Kjellman (1951) and modified by Bjerrum and Landva (1966) at the Norwegian Geotechnical Insti· tute (NGI) for practical use, tests a cylindrical sample (80 mm diameter x 20 mm high) laterally confined by a wire reinforced mem· brane between rigid top and bottom platens (Fig. 3). Simple shear strain is presumed
to be imposed by displacing the top boundary.
Neither of these apparatus allows the development of complementary shear stresses on the vertical sides normal to the plane of deformation. As a result, the shear and normal stresses must be non-uniformly distributed to satisfy equilibrium (Wood, Drescher and Budhu, 1979 ; Budhu, 1984 ; Airey, Budhu and Wood, 1985).
In order to evaluate the practical effects and actual distribution of normal and shear stresses, studies have proceeded along four different directions- analytical/ numerical (Roscoe, 1953 ; Duncan and Dunlop, 1969 ; Lucks et al., 1972 ; Prevost and Hoeg, 1976 ; Hara and Kyota, 1977 ; Shen et al., 1978) ; experimental (Cole, 1967 ; Stroud, 1971 ; Budhu, 1979 ; Airey, 1984) ; comparisons of simple shear test results with results from either the triaxial or hollow cylinder appara· tus or both (Saada et al., 1983) ; and discussion (V ucetic and Lacasse, 1982).
In the analytical/numerical studies, elastic material behavior was assumed and it was shown that for the Cambridge simple shear apparatus (SSA) at least the middle onethird of the sample and for the NGI SSA, seventy percent of the sample could be expected to deform uniformly. Detailed experimental work in these devices is tedious and require sophisticated instrumentation and as such cannnot be conducted on a routine basis. However, such tests in sands by Stroud (1971) and Budhu (1979, 1984)did show that the middle one-third of the samples in monotonic tests deform uniformly in these devices but that non-uniformities predominant at the ends spread rapidly during cyclic loading. Airey (1984), using a modified version of a specially instrumented NGI type device originally developed for sand by Budhu (1979), showed that the normal stresses are much more uniform for clays than for sands.
Comparisons of simple shear test results with results from other devices-the triaxial and hollow cylinder apparatus-show that the simple shear devices produce lower soil strengths and stiffnesses and as a result sim-
BUDHU AND BRITTO
Fig. 1. Simple shear strain
LOAD CELL
Fig.
ple shear devices have been dismissed (Saada and Townsend, 1981). There is no obvious reason why simple shear test results should agree with results from triaxial or other devices unless the stress paths followed are identical and the stresses imposed produce similar changes in soil fabric.
In comparing results from various testing devices, it is necessary to select appropriate stress and strain parameters. In simple shear tests, the intermediate principal stress (a 2) is neither independent of nor is it equal to either the major or minor principal stresses. Consequently, the effect of the intermediate stresses should not be ignored as a rule.
Recent tests on clays (Airey, 1984) in an elaborately instrumented NGI type simple shear device show that as the plasticity of the material increases so does the uniformity of boundary stresses. It appears that elastic analyses produce results which perhaps show up the non-uniformities in simple shear devices too pessimistically. Of course, soil is neither isotropic nor elastic. It is the intention of this paper to examine the distribution of stresses and the mechanical behavior of a sand and a clay in the simple shear devices assuming these soils behave as either elastic or elastoplastic materials.
ELASTOPLASTIC SOIL MODEL
The model used is the modified Cam-clay model which has been shown to predict the mechanical behavior of normally and lightly overconsolida ted soils (overconsolidation ratio less than 2) reasonably well (see for example, Wroth, 1977). The details of this model have been elaborated on by others (for example, Roscoe and Burland, 1968 ; Wroth and Houlsby, 1985) and need not be attempted here. Essentially, the model consists of an ellipsoidal yield surface in p, q space (Fig. 4(a)) where:
and P=(1/3) OiJaii
q= (3/2 SiJSiJ) 112
p is the mean effective stress q is the deviatoric stress
l p
Fig. 4. Modified cam-clay yield surface and e-p relationship
OiJ is Kronecker delta au is the effective stress tensor siJ is the stress deviator
During loading the yield surface expands and this expansion is linked to the isotropic normal consolidation line (Fig. 4 (b)). The soil parameters required for this model are:
...{ -the slope of the consolidation line tc -the slope of the swelling line <P or M-the effective angle of friction or the frictional constant at failure in (q, p) space respectively
G or p-the shear modulus or Poisson's ratio
e;. -void ratio at unit effective mean stress on the critical state line.
For simple shear strain tests, a knowledge of either the magnitude of the intermediate principal stress or its relationship to the other two principal stresses is needed to determine M. Unfortunately, the intermediate principal stresses in simple shear devices are not routinely measured or cannot be measured as is the case for the NGI SSA.
Drained tests in the Cambridge SSA conducted by Stroud (1971) and Budhu (1979) on 14/25 Leighton Buzzard (LB) sand (average grain size 1 mm) showed that
O'z=m(al+aa) (3)
where m is a constant, a 1 and a 3 are the major and minor effective principal stresses. For 14/25 Leighton Buzzard sand m=O. 37 (Stroud, 1971). The authors are not aware of any such measurements in clays. Assuming values of m=O. 4 to 0. 5, then for simple shear conditions :
M=.J3sin¢
where sin¢=(a 1 -a 3 ) r1t+r1a f
and the subscript f denotes failure.
Numerical analyses were also conducted assuming that the test samples behave as isotropic elastic materials. In this case the elastic modulus (E) was calculated from the expression :
(5)
K where e is void ratio.
SIMPLE SHEAR TESTS
In conventional simple shear tests, the soil sample is consolidated under K 0 condition by applying a vertical stress and then shear displacements are subsequently imposed. Two types of tests are often carried out. One, called the constant load test, permits the passage of water from the sample through porous stones placed on either the top or bottom boundaries or both, and the top boundary is allowed to move vertically to follow the compression or the dilation of the test specimens. Constant load tests on sands are generally carried out on dry material. The other, called the constant height test, is similar to the constant load test but the height between the top and bottom platens is kept constant. The reduction or increase in vertical stress required to maintain constant sample height is then related to the pore water pressures that might be developed in an undrained test under constant vertical
FINITE ELEMENT ANALYSIS
CRISP (Gunn and Britto, 1984) is a finite element program developed at Cambridge University. The program allows prediction of plane strain, axisymmetry and three-dimensional boundary value problems. Constitutive models included in the program are the critical state models (Cam-clay, modified Cam-clay) elastic and elastic perfectly plastic models. The modified Cam-clay and the elastic models are used in the present work.
A typical mesh used is shown in Fig. 5( a). The mesh includes the top and bottom metallic platens. These platens are assumed to behave elastically. Each element is an eight noded linear strain quadrilateral and the displacements are assumed to be quadratic functions of position coordinates. A 3 x 3 integration scheme was employed.
The constant load test is simulated by firstly consolidating the sample and then prescribing horizontal displacements on the top platen with the bottom platen fixed. The top platen is permitted to move vertically. The vertical sides are assumed to be
Fig. 5 (a). Finite Element Mesh used
Fig. 5 (b). Displacements imposed on vertical sides
on rollers. The nodes on the vertical sides are displaced following a uniform triangular distribution as shown in Fig. 5( b).
The constant height test is simulated in the same manner as the constant load test except that the top platen is prevented from moving vertically.
The displacements are applied in small crements and the nodal coordinates are dated after each increment. Equilibrium is, therefore, satisfied in the final (deformed) configuration. The updated coordinate scheme adopted here is not a rigorous ment of finite deformation analyses. But, since significant displacements are imposed on the test samples (final shear strains are about 10 percent), this scheme is preferred to the alternative of satisfying equilibrium with respect to the original configuration.
The effect of the stiffness of the reinforced membrane of the NGI SSA is not included in this study. Shen et al (1978) using isotropic elastic material behavior showed that the uniformity of stresses and strains improves by increasing the stiffness of the reinforced membrane. Budhu (1979) showed perimentally, for LB sand tested in an NGI type SSA, that stiffer reinforcing wires proves the uniformity of stresses. V ucetic and Lacasse (1982) reported that the use of thicker than standard membranes (thickness varies from 0. 56 mm to 0. 67 mm) resulted in lower shear strengths and moduli.
The simulation used here is intended to represent conditions on the principal third of the test samples in both the Cambridge and NGI type SSA (see Fig. 6). For the Cambridge SSA, the size of the test samples analyzed is 100 mm x 100 mm x 20 mm high and for the NGI type SSA, the sample size 1s 110 mm diameter x 20 mm high. The mesh dimensions are chosen to make direct comparison with results from specially instrumented tests conducted by Budhu (1979) and Airey (1984). Even though it may seem inappropriate to compare the test results from the NGI type SSA apparatus with a plane strain analysis, it is reasonable to make such a comparison for the core
CORE
Fig. 6. Principal third and sample core part of the sample.
MATERIAL PROPERTIES
It is assumed that the platens are made out of aluminium and the test samples are either 14/25 Leighton Buzzard sand or white kaolin. The following properties were selected.
Platens : E=65 x 10 6 kN/m 2 , x 10 6 kN/ m 2 , p =0. 33 (Roark and Young, 1975) Leighton Buzzard Sand : A=0. 025 (Airey et al., 1985), tC=O. 005, ¢ =35° (Stroud, 1971), e 2 =0. 927 (Airey et al., 1985), 11=0. 37 (Stroud, 1971), K 0 =0. 425 (Budhu, 1979)
Speswhite Kaolin (Pl=35%) : A=O. 25 (Lawrence, 1980), tC =0. 05, ¢ =21 o (Nadarajah, 1973), e1 =2. 5, t-t=O. 33, K 0 =0. 685 (Airey, 1984)
COMPARISON OF NUMERICAL AND EXPERIMENTAL RESULTS
Distribution of Vertical Stresses
The vertical stress distribution on the top and bottom boundaries for very loose tive density :::::::7 percent) LB sand and Speswhite kaolin obtained from the numerical analyses for constant load tests are compared with those from the experimental observations in Figs. 7 and 8. All the results reported in this paper for LB sand and white kaolin were obtained from tests carried out in the Cambridge SSA (Budhu, 1979) and a specially instrumented NGI type SSA (Airey, 1984) respectively. The experimental vertical stress distributions were deduced from load cells placed on the boundaries of these devices (Budhu, 1984). It is assumed here that the centroidal vertical stresses in
Experiments (Budhu, 1979)
Modified cam-clay Elastic (a) .., o.6%
(b> Y=l4%
experiments modified cam clay elastic
Fig. 7. Vertical stress distribution on loose sand (constant load test)
the elements in the first and last rows CAB and CD, 1 mm thick, Fig. 5) of the sample obtained from the numerical analysis are appropriate for comparisons with the experimental results. The vertical stresses (ay)
Experiments (After Airey, 1984)
Modified Cam-clay Elastic '( = 10%
Fig. 8. Vertical stress distribution on kaolin (constant load test) \ are all normalized with respect to the applied vertical stress aa· By default all stresses are effective stresses. For LB sand the latter value is 100 kN/m 2 and for kaolin it is 200 kN/m 2 •
At low shear strain levels where elastic behavior is expected to be predominant the modified Cam-clay and elastic analyses give sensibly the same results for LB sand (Fig. 7 ( a)). However, there are significant differences between the numerical and experimental results at this strain level. The LB sand used in the experiments was in a very loose state and during the consolidation phase very small lateral movements of the hinged side walls of the Cambridge SSA was sufficient to cause a collapse of the loose sand structure in the vicinity of these side walls. Thus, the differences between the numerical and experimental results may at least be partially attributed to the non-uniformities inherited during the consolidation phase. At high strain levels (Figs. 7( b) and 8) where plastic behavior is predominant, the elements
KAOLIN
Fig. 9. Comparison of quintic, cubic and linear vertical stress
Exp:riments (After Airey, 1984)
Modified Cam-clay
y = 10%
Fig. 10. Vertical stress distribution on kaolin (constant height)
close to the ends A and D (Fig. 5( a)) which are subjected to very high vertical stresses yield quickly reaching critical state, and transfer stresses to adjacent elements. The elastic analyses, as expected, continuously show very high stresses at these ends.
The experimental vertical stress distribu-
Experiments (Budhu, 1979)
Fig. 11. Comparison of stress strain behavior of LB sand (constant vertical load)
tions were obtained by fitting the information genera ted by load cells placed on the top and bottom boundaries by a quintic polynomial (Budhu, 1984). Other forms of distributions are possible. For example, a cubic polynomial can be fitted to the data using the least square method or a simple linear distribution can be used as done by Stroud (1971). A typical comparison of these distributions is shown in Fig. 9 where it is seen that the cubic polynomial produces less wiggles than the quintic polynomial (as expected). The linear stress distribution shows sudden changes in stresses which cannot, in practice, occur. The quintic polynomial distribution also shows the development of tensile stresses which are unsustainable by soils.
Fig. 10 shows the vertical stress distribution for a constant height test on kaolin at r 10%. The agreement between the experimental and analytical results at this rather large strain level is encouraging. Again, the results from the elastic analyses show much larger vertical stresses at ends A and D than those obtained from the experiments and the modified Cam-clay model.
Stress-strain Behavior
Fig. 11 shows the comparison of the nor-
BUDHU AND BRITTO
experiments (After Airey, 1984)
modified cam cloy
Fig. 12. Comparison of stress strain behavior of kaolin (constant load test)
malized shear stress ratio-shear strain response of LB sand if measurements were made at the sample core (Fig. 6) and over the top boundary for a constant load test. The latter measurements are those normally made in routine tests. The experimental and modified Cam -clay results show good agreement. Both results show a difference of shear stress ratio about 4 percent between the sample core and the average on the top boundary.
A similar comparison for kaolin, except now the shear stress on the horizontal boundary of the sample is reported (Airey, 1984) rather than the normalized shear stress ratio, is shown in Fig. 12. The numerical results give no significant difference in shear stress between the sample core and the average over the top boundary up to about 10 percent shear strain whereas the experimental results show a difference of about 16 percent. The numerical analysis did not account for slippage between the platens and soils. Although techniques have been developed to minimize slippage (Stroud, 1971; Budhu, 1984 ; Airey, 1984) some slippage at large strains could be expected during experiments using the simple shear devices. To account for slippage at platens-soil interface requires the use of slip elements in the numerical analysis. This would be the
Exp€riments
(After Airey, 1984)
Modified earn-clay y (%)
Fig. 13. Comparison of stress strain behavior of kaolin (constant height test)
subject of separate communication by the authors.
Results from constant height tests for kaolin are shown in Fig. 13. The numerical results give a maxim urn shear stress about 40 percent greater than that observed in the experiments. It appears that the results from the plane strain version of the modified Cam-clay model cannot satisfactorily match the experimental shear stress data for constant height tests.
Constant Height/Undrained Test Relationship
Bjerrum and Landva (1966) proposed that the constant height test in the NGI device is similar to an undrained test and that the changes in vertical stresses required to maintain constant height are equivalent to the changes in pore water pressures generated in a constant load test if drainage of water from the samples were prevented. There is controversy on this proposal. Saada et al. (1983) reported results from simple shear (NGI type device) and hollow cylinder tests on Edgar Plastic Kaolin and Reid Bedford sand which show that the changes in vertical stresses in constant height tests are not equiv-
Fig. 14. Comparison of changes in vertical stresses with changes in pore water pressures
alent to the pore water pressures in constant load tests. They found that none of their normalization procedures was able to bring the two results to equivalency. However, V ucetic and Lacasse (1984) pres en ted results from anisotropically consolidated triaxial tests on Drammen clay which are contrary to the results reported by Saada et al. (1983).
Saada et al. (1983) were only able to measure the average vertical stresses on presumably the top boundary of their samples and it is not clear where pore water pressure measurements were made. There is no obvious reason to suppose that the triaxial test results presented by V ucetic and Lacasse (1984) would be true for soils tested in the simple shear devices. It seems intuitively plausible, however, that if a soil sample in the simple shear devices were to deform as a single element, then there should be some relationship between the changes in vertical stresses in a constant height test and the changes in pore water pressure which would occur in a constant load test under undrained conditions.
Fig. 14 shows the finite element results of the changes in vertical stresses ..day and pore water pressures (Liu) from a constant height test and a constant load test (un· drained conditions) respectively on the top
1 mm layer at the sample core (Figs. 5( a) and 6). These results represent predictions using the modified Cam-clay model and the properties of speswhite kaolin presented earlier in this paper.
The changes in vertical stresses and changes in pore water pressures in the two types of test show a maxim urn difference of about 10 percent at a shear strain of 10 percent. Of course, these predictions will vary depending on the choice of soil model. But, it is clear from Fig. 14 that if the sample were to behave as a single element (as demonstrated for the sample core by analysis and experiments) then one could expect (for all practical purposes) that the changes in vertical stresses in a constant height test would be directly related to the changes in pore water pressures in a constant load test conducted under undrained conditions. However, good quality simple shear tests in which the pore water pressures and stresses are measured on the sample core are required to further validate Bjerrum and Landva's supposition.
CONCLUSION
It is indisputable that non-uniformities of stress and strain develop in samples of soil tested in the two available simple shear apparatus. It is, however, not necessary to dismiss these apparatus since meaningful results, regardless of the plasticity of the soil, on the mechanical behavior of soils under monotonic loading can be obtained by measuring the stresses and pore water pressures on the sample core.
It was shown in this paper that elastic analyses tended to produce larger stress concentrations at the ends of the sample in the plane of shear deformation than analyses based on the assumption that the behavior of the soil specimens can be modelled by the modified Cam-clay model. The experimental evidence suggests that the stressstrain behavior of very loose 14/25 Leighton Buzzard sand and speswhite kaolin from constant load simple shear tests is satisfactorily
40
BUDHU AND BRITTO
predicted by the modified Cam-clay model. This is, however, not the case for constant height tests. Even though the vertical effective stress is reasonably well predicted by the finite element program for kaolin, the shear stresses are overpredicted by about 50 percent.
Experimental data from the sample core of good quality tests are needed to validate the supposition that the changes in vertical stresses in constant height tests are equivalent to the changes in pore water pressures in constant load undrained tests in spite of the good agreement obtained from the numerical analysis.
ACKNOWLEDGEMENT
The authors are grateful to Dr. D. M. Wood and Dr. D. W. Airey for comments on the draft version of this paper. Permission to use the test results on NGI type SSA apparatus by Dr. D. W. Airey is appreciated.
NOTATION
P=mean effective normal stress
q=deviatoric stress
u=pore water pressure
E=elastic modulus
e,<=void ratio at unit mean effective stress on critical state line
G=shear modulus
M=slope of critical state line
·=<:trP<:<: deviator tensor
r=engineering shear strain
Oij=Kronecker delta
K=slope of swelling line
A=slope of consolidation line
.u=Poisson's ratio
O'ij=effective stress tensor
O"t> 0' 2, 0' 3 =effective major, intermediate and minor principal stresses
O'a=effective applied vertical stress
O"x,O'y, 0' 2 =effective normal stresses in x, y and z directions respectively
'Z"yx=shear stress
¢=effective angle of internal friction
¢'cv=effective angle of internal friction at the critical state
REFERENCES
1) Airey, D. (1984) : "Clays in circular simple shear apparatus," Ph. D. Thesis, University of Cambridge, Cambridge, United Kingdom.
2) Airey, D., Budhu, M. and Wood, D. M. (1985): "Some aspects of the behaviour of soils in simple shear," Developments in Soil Mechanics and Foundation Engineering-2, Editors, P. K. Banerjee and R. Butterfield, Elsevier Applied Science Publishers, Ltd., London, pp. 185-213.
3) Bjerrum, L. and Landva, A. (1966): "Direct simple shear tests on a Norwegian quick clay," Geotechnique, Vol. 16, No.1, pp. 1-20.
4) Budhu, M. (1979) : "Simple shear deformation of sands," Ph. D. Thesis, University of Cambridge, Cambridge, United Kingdom.
5) Budhu, M. (1984) : "Non-uniformities imposed by simple shear apparatus," Canadian Geotechnical Journal, Vol. 21, No.2, pp.125-137.
6) Cole, E. R. L. (1967) : "The behaviour of soils in the simple shear apparatus," Ph. D. Thesis, University of Cambridge, Cambridge, United Kingdom.
7) Duncan, J. M. and Dunlop, P. (1969) : "Behavior of soils in simple shear tests," Proceedings 7 th International Conference on Soil Mechanics and Foundation Engineering, 1, pp.101-109.
8) Gunn, M. J. and Britto, A.M. (1984) : "CRISPUser's and programmer's guide," University Engineering Department, Cambridge U ni versity, Cambridge, U.K.
9) Hara, A. and Kyota, Y. (1977) : "Dynamic shear tests of soils for seismic analyses," Proceedings of the 9th ICSMFE, Tokyo, 2, pp. 24-250.
10) Kjellman, W. (1951) : "Testing the shear strength of clay in Sweden," Geotechnique, Vol. 2, No.3, pp. 225-232.
11) Lucks, A. S., Christian, J. T., Brandow, G. E. and Hoeg, K. (1972) : "Stress conditions in NGI simple shear test," Journal of the Soil Mechanics and Foundations Division, ASCE, Vol. 98, SM 1, pp. 155-160 .
12) Lawrence, D. (1980) : "Some properties associated with kaolin soils," M. Phil. Thesis, Cambridge University, Cambridge, U.K.
13) Nadarajah, V. (1973) : "Stress strain properties of lightly over-consolidated clays," Ph. D. Thesis, University of Cambridge, Cambridge, U.K.
14) Prevost, J. H. and Hoeg, K. (1976) : "Reanalysis of simple shear soil testing," Canadian Geotechnical Journal, Vol. 13, No.4, pp. 418-429.
15) Roark, R. J, and Young, W. C. (1975) : Formulas
for Stress and Strain, McGraw Hill.
16) Roscoe, K. H. (1953) : ''An apparatus for the ap· plication of simple shear to soil samples," Proceedings 3 rd International Conference on Soil Mechanics and Foundation Engineering, Zurich, Switzerland, 1, pp. 186-191.
17) Roscoe, K. H. and Burland, J. B. (1968) : "On the generalized stress-strain behavior of 'wet' clay," in Engineering Plasticity (J. Heyman and F. A. Leckie, eds.), Cambridge University Press, Cambridge, United Kingdom, pp. 535609.
18) Saada, A. S. and Townsend, F. C. (1981) : "State of the art : Laboratory strength testing of soils," Laboratory Shear Strength of Soils, American Society for Testing and Materials, STP 740, (Eds., R.N. Yong and F. C. Town· send), pp. 7-77.
19) Saada, A. S., Fries, G. and Ker, C. C. (1983) : "An evaluation of laboratory testing techniques in soil mechanics," Soils and Foundations, Vol. 23, No.2, pp. 98-112.
20) Shen, C. K., Sadigh, K. and Herrmann, L. R. (1978) : "An of the NGI simple shear apparatus for cyclic soil testing," ASTM Dynamic Geotechnical Testing, STP 654, pp. 148-162.
21) Stroud, M.A. (1971) : "The behavior of sand at
low stress levels in the simple shear apparatus," Ph. D. Thesis, University of Cambridge, Cambridge, United Kingdom.
22) Vucetic, M. and Lacasse, S. (1982) : "Specimen size effect in simple shear test," Journal of the Geotechnical Engineering Division, ASCE, Vol.108, GT 12, pp. 1567-1585.
23) Vucetic, M. and Lacasse, S. (1984) : Closure "Specimen size effect in simple. shear test," Journal of the Geotechnical engineering Division, ASCE, Vol.llO, No.3, pp. 439-453.
24) Wood, D. M., Drescher, A. and Budhu, M. (1979) : "On the determination of stress state in the simple shear apparatus," Geotechnical Testing Journal, GTJODJ, Vol. 2, No.4, pp. 211-221.
25) Wroth, C. P. and Houlsby, G. (1985) : "Soil Mechanics-property characterization and analysis procedures," Proceedings of the Eleventh International Conference on Soil Mechanics and Foundation Engineering, San Francisco, Vol. 1, pp. 1-55.
26) Wroth, C. P. (1977) : "The predicted performance of soft clay under a trial embankment loading based on the Cam-clay model," Finite Elements in Geomechanics, Editor, G. Gudehus, John Wiley and Sons Ltd., U.K., pp. 191-208.
THE TREATMENT OF THE MISDEMEANANT
[This excellent summary of the present correctional treatment of the misdemeanant was read by Amos W Butler, chairman of the Committee on Corrections, at the National Conference of Charities and Corrections in Memphis, in May, 1914.]
Who knows how many persons are confined in local jails in the United States? Who comprehends the magnitude of the problem involved? How many appreciate its relation to the individual, to his family, to society?
The Bureau of the Census tells us that 452,055 persons were committed to county and municipal prisons in 1910, under sentence or for non-payment of fine. We call them misdemeanants. It is the name ordinarily given one whose offense the law does not deem sufficiently serious to warrant a State prison sentence. If the ratio of commitments to the whole number received is the same throughout the United States as in Indiana, it is probable that one and one-half million people annually come under the influence of these local prisons.
We know that jails are the spoil of partisan politics. They are maintained largely on the fee basis. Most of them were built without any proper idea of the purpose they were intended to serve. As a rule they are insanitary, they lack proper provision for separating the sexes and there is no means of employment. Often they are crowded beyond their capacity. There is little attempt to classify the prisoners. They congregate in the corridors and the older and more experienced in criminal ways instruct the others in vice, immorality and crime. In how many such institutions are women not only waited upon but searched by men?
The late Samuel J. Barrows once remarked: “Back in 1876 a committee of the New York Legislature said: ‘There is no one source of crime more operative in the multiplication of thieves and burglars than the common jail’ and that statement still remains true of a large number of jails throughout the country.”
Many of you have heard Dr. F. H. Wines’ scathing denunciation of these institutions. Delegates to the International Prison Congress who visited this country in 1910 declared our local jail system as bad as it was centuries ago in Europe. “Every jail I saw ought to be wiped off the face of the earth,” said Thomas Holmes, secretary of the Howard Association of London, and this was the general verdict of these distinguished prison officials and penologists. It was the idleness of the prisoners, the lack of fresh air, the indiscriminate mingling, the long-delayed trials that impressed them so unfavorably. “I asked two colored men how long they would be in and they said they did not know; that they had waited eleven days for a trial,” said Dr. Eugene Borel, professor of law in the University of Geneva, Switzerland. “That is a shocking travesty of justice. In Europe a prisoner gets a hearing within twenty-four hours.”
Yet under such conditions as these we detain hundreds of thousands of persons:—The vagrant, the drunkard, the witness, the runaway boy, the first offender, the hardened criminal, the man awaiting trial, the convicted law-breaker. What can we expect but that they will degenerate in body, mind and morals? Even where work is provided, as is done in some larger jails and workhouses, it is under the old contract system, which we should like to see abolished. In some States misdemeanants are employed on the public highways. However successful this may be in some parts of the country, it would probably not be in conformity with the public sense in the more thickly settled communities, or practicable to any great extent in the more northern latitudes.
The whole question of the apprehension, treatment and release of the misdemeanant is of tremendous importance. While prison reforms are coming with surprising rapidity, they have been confined largely to the felon. The misdemeanant has been neglected.
In the first place, what are the qualifications of the average policeman? Ordinarily he is without training or experience. His politics have usually had more to do with his appointment than any other consideration. What part can such a policeman play in an enlightened system of penology?
In a number of States the constitution proclaims that the penal code shall be founded on principles of reformation and not of vindictive justice. How far has that been interpreted in the statute laws? The provision of most of our State constitutions that justice shall be administered “speedily and without delay” is wholly forgotten. We generally think that the day of imprisonment for debt is past. Yet many jail prisoners are held for debt—the fine assessed against them. The man of means pays his fine and goes free; the man without money suffers imprisonment under conditions which menace health and morals. It frequently becomes necessary for his family to ask for help; sometimes he loses his job. If he becomes embittered, or vindictive, need we wonder at it?
The picture is not all dark. Here and there light is breaking through. We are coming to understand that the policeman can be a social agent, a next friend, an instructor in obedience to the law. We are beginning to regard as the best officer the one who makes the fewest, not the most arrests. In some cities women are being added to the police force, and there are police matrons, and jail matrons, and even women judges.
“Humanizing the courts” is an expression coming more and more into use. Instead of sending to jail men who are unable to pay their fines, judges are releasing them conditionally and giving them a chance to earn the money. The plan works admirably. Judges are finding that their confidence is seldom misplaced. This principle has been enacted into law in Massachusetts and other States. Elsewhere it has been practiced without special authority of law. In New York State the probation law provides for the collection of fines on probation and also restitution on probation.
We are further coming to believe that too many persons are sent to prison. Some States have adopted a system of probation, under
which many lawbreakers are reclaimed to society without the stigma of a prison sentence. Probation has been successfully tried in Massachusetts, New York and other States.
We have learned, too, the value of the farm colony for the open-air employment of almost all classes of public wards. This has been applied to the insane in Wisconsin, Massachusetts and Indiana; to epileptics in New York, New Jersey and Indiana; to feeble-minded in Massachusetts, New Jersey and Indiana; to both dependent and delinquent children in many States, and more recently to certain classes of prisoners, including all kinds of misdemeanants of both sexes. What is being done at Cleveland, at Occoquan and at Guelph is well known. The most recent development of this movement is the New York State Farm for Women Misdemeanants. The simple, inexpensive, yet substantial form of buildings, the freer life and the opportunity to contribute in part at least to their own support, make it far better for the inmates and cheaper to the taxpayer.
In Indiana the first step in this direction came about through the establishment in 1907 of a State workhouse for women misdemeanants as a branch of the Women’s Prison at Indianapolis. The institution is entirely in control of women. Then the Board of State Charities began a vigorous campaign for a state farm for male misdemeanants. Conditions in the county jails were shown forth in the following paragraph:
HOW PRISONERS LIVE AND LEARN IN INDIANA COUNTY JAILS.
They live in idleness at the expense of the taxpayer.
They learn vice, immorality and crime.
They become educated in criminal ways.
They degenerate both physically and morally.
In 1913 an appropriation was secured, the land has now been purchased, and work on the buildings will soon begin. The law
contemplates that the construction work shall be done largely by State Prison and Reformatory men. The new institution is for men who have a jail sentence of sixty days or more, and prisoners may be transferred from the State institutions whenever room for them exists at the farm. Eventually there will probably be several such farms in the State, and this movement, with proper amendments to existing laws, should in time do away with the use of the county jails for the confinement of convicted offenders, and leave them only as places of detention.
We now look forward to the time when we shall have a form of indeterminate sentence for misdemeanants. The success of this form of sentence for felons in Indiana as in many other States justifies our belief that it will prove valuable in the treatment of misdemeanants. Certainly some improvement can be made over the present illogical short sentence, which benefits neither the individual nor the public, in whose name he is held.
New York has already taken an important step in this direction. Misdemeanants between the ages of sixteen and twenty-one are to be committed under an indeterminate sentence to the reformatory for this class of offenders, authorized by the legislature of 1912. The site for this new institution has not yet been located. It is the purpose of those interested to place it on a farm and to make it one of the most complete and modern of reformatories.
Another needed reform, State control of county jails, is receiving some attention. J. S. Gibbons, Chairman of the Prison Board of Ireland, said: “I tell you what I think you lose sight of in this country: That all these splendid reformatories deal with merely a drop in the ocean compared with the county and city jails to which your thousands of prisoners go, and where many are manufactured. We were in exactly the same condition up to 1877 when we brought county and city jails out from under local authorities in the United Kingdom. We found the antecedent to all reform was State centralization. In 1877 every prison and jail was put under central administrative authority, and the expenses paid out of the imperial funds. Three acts were passed simultaneously for the three
kingdoms. We then began at the bottom, closing all the superfluous ones, and in that way we were able to close about half.”
I have the following statement from Sir Evelyn Ruggles-Brise, of England: “The Prison Act, 1877, transferred the Local Prisons of this Country (i.e. prisons for the confinement of all classes of prisoners other than those sentenced to penal servitude) from the control of local ‘Visiting Magistrates’ to that of the State. The Act came into effect on the 1st of April, 1878, 113 Local Prisons being so transferred. Since that date, their number has been reduced to 56. At the time of their transfer, the Local Prison population stood at 21,030 —the highest known. From that date a continuous fall was recorded until 1885, when the numbers reached slightly over 15,000. After a series of fluctuations below and above this number, the population stands at 15,000 at the present time. Relatively to the total population of the country, the figure for 1878 represented 686 committals per 100,000, while that for the year ended 31st March 1912 was the lowest on record, viz., 439 per 100,000.”
Massachusetts, perhaps, has led the agitation in this country for State control of county jails In other States there has been some publicity in favor of such action. An offender against the federal law becomes a prisoner of the United States and is under the direction of the federal judge. Why should one who violates a State law not be a prisoner of the State? That is the theory that underlies the new law for jail supervision in Indiana. Offenders against the State law have been placed under the oversight and authority of the judge of the circuit or criminal court, who is a State official. This judge may say where and how the prisoner shall be detained, and if the jail is unsatisfactory he may condemn it. He is authorized to prescribe rules formulated by the Board of State Charities, which has supervision of all jails and other public charitable and correctional institutions. A violation of these rules, once entered in his order book, is in effect a contempt of court.
While these advance steps have been taken, the reform is by no means general. Most of the States continue to use, unchanged, the system long since discarded in Europe, whence it came. The results are not reformatory. On the contrary, they are destructive alike to the
individual and those with whom he later comes in contact. Local jails are recruiting stations for our larger State correctional institutions. We should make greater progress in reformation if we did not first pollute the stream we are going to treat.
The outlook is not bright, but it is by no means hopeless. The evils which exist are the natural result of the system we adopted. Let us change the system. Let us begin at the bottom and study all the steps in the treatment of the offender—his apprehension, detention, trial, conviction, probation, confinement, treatment, employment, conditional release, final discharge. Let us set as our goal:
1. A system of police recognizing character, merit, and efficiency in the personnel and a proper social view for its operations.
2. A prompt hearing for every person arrested.
3. The establishment of juvenile courts for all children’s cases.
4. Provision for the care and detention of delinquent children outside the jail.
5. A probation system for adults similar to that of juvenile courts.
6. Separate trials for women offenders.
7. A modification of the present system of fines in order not to discriminate against the poor.
8. Classification of prisoners, confinement of individuals apart from each other and absolute sex separation in county jails.
9. The prohibition of the use of the jail for any other purpose than that of temporary detention.
10. The abolition of the fee system.
11. State control of all minor prisons.
12. The establishment of industrial farms for the convicted misdemeanants.
13. A form of indeterminate sentence for misdemeanants.
14. Their release on parole under supervision.
15. The abolition of contract labor.
The following members of the Committee on Corrections approve and sign the report:
Major R. W. McClaughry
Julian W. Mack
Arthur W. Towne
Frank E. Wade
Joseph P. Byers
John J. Sonsteby
Quincy A. Myers
W. H. Moyer
A. J. G. Wells
J. A. McCullough
John H. Dewitt
Archdeacon B. M. Spurr
Mr. E. Stagg Whitin disagrees with what is said about the employment of convict labor on public highways in the more thicklysettled northern States.
EYTINGE
By Charles Wheatley.
[In the State Prison of Arizona is a life prisoner who has made a remarkable record. The Delinquent has followed with great interest Mr Eytinge’s business career while inside the prison, and now, through an article from the Ohio State Journal, presents Mr. Eytinge’s story in part.]
Down in Florence, Arizona, in the State Prison they have penned up a Daytonian for life on a charge of murder. This man went into prison an outcast ready to die—or rather about to be killed by tuberculosis. He had been a “con man,” forger and all-around crook, and today, seven years after his conviction, he occupies a prominent place in the business world.
The man in question is Louis Victor Eytinge, born in Dayton in 1878. At the age of three his parents separated. His mother was a talented musician, while his father had been actor, broker and gambler. Sometime after the separation young Eytinge went to live with an uncle. He was educated in the First District school, Central high school, spent two years at St. Mary’s Institute and one year at the University of Notre Dame.
At the age of 16 Eytinge started on his life of crime by forging a check. He went from bad to worse, and was soon known from east to west and north to south by the list of bad paper he had left behind. Many times his relatives kept him out of jail by straightening out difficulties into which the young man drifted, but finally they became tired of this and refused to have anything further to do with him.
Cut off from home’s ties, Eytinge became even more desperate and finally ended up in the Ohio State Penitentiary following his arrest and a daring attempt at a big jail delivery. While in the penitentiary Eytinge was known as a “bad one” and when released it was found
that he was suffering from the advanced stages of tuberculosis. This was at the age of twenty-seven. When relatives learned of his physical condition they weakened in their determination to do no more for him. They sent him to Arizona with an allowance of one hundred dollars per month. He was to remain away from Dayton and live within the law.
Eytinge had not been in Arizona very long, however, until he was under suspicion of murder. A man with whom he had become quite friendly and whom he had taken care of in a financial way was found dead in the desert one day after Eytinge had been seen to leave the city with him for a buggy ride. When the body was found Eytinge had disappeared. He was found later in California and in his pockets were some of the belongings of the dead man.
Eytinge was brought back to Arizona and placed on trial for his life. He was convicted on circumstantial evidence after William A. Pinkerton, one of the greatest detectives of the time, had declared that it was improbable that Eytinge had killed the man “as his criminal bend did not gallop in that direction.” Judge A. C. Wright, of Phoenix, who defended Eytinge, was so sure of the prisoner’s innocence that he went into his own pocket to pay for the appeal of the case.
Eytinge was convicted and sentenced to life in prison. How long can he live in prison? That was the question asked by all who had seen him on the witness stand. Eytinge was near death’s door at the time. His allowance had been cut off and he did not have the funds with which to purchase fresh milk and eggs, two delicacies almost absolutely necessary for the tubercular patient. The State prison was then at Yuma, Arizona. It was a hell hole with its mosquitoes which swarmed into the cells that looked out on the Colorado river. The quarters were damp, dark and dirty, and afforded excellent means of killing a man who at the time was ready to die.
But right there is the story Eytinge was not ready to die. He was not 30 years old. He clung to the hope that something would bring about his release. He wanted to live, wanted to breathe again the fresh air
of freedom and hobnob again with the pals, association with whom had placed him behind steel bars. Eytinge was still a crook.
Enough of the crook! Let us get to Eytinge the business builder and man. It is a great, broad jump we admit, but the telling of how Eytinge accomplished it is a thousand times easier to do than the battle which Eytinge himself had to go through in bringing about that end. Seven years ago Eytinge was a crook of the first water, seven years, mind you, and—today Eytinge is a man. He is looked up to by professional advertising writers the country over. His business letters are the best that are being written today. Eytinge is the star of them all.
The strange part of this great accomplishment is the manner in which it was brought about. Eytinge made the road light for himself. He started out with weak hands to carry the torch that showed him the way. Today Eytinge is his own light; he has thrown away the torch and walks alone. Honesty is his greatest asset, and it was through the writing of business letters that Eytinge found his way.
When Eytinge entered the State prison he realized that he must find some means of making money. One could not forge checks in prison. Eytinge was placed in the chronic ward. Around him were Indians, Mexicans and others who in their spare moments made hat bands, watch fobs and other articles out of horse hair. They decorated these with rosettes hammered out of Mexican dollars. These articles were sold whenever possible to chance visitors at the prison, but it was a slow way of making money.
And then one day the big idea came to Eytinge! Opportunity knocked on the steel bars and Eytinge was there to talk to Her! The crude handiwork of the prisoners was the vehicle by which Eytinge began to make his way. He began by writing letters to retail dealers on the outside asking them to take the agency for the hat bands, belts, watch fobs and other articles. It was not long until the money began to pour in. There was a big demand for the articles, but it was the letters that Eytinge had written that opened up the field. In these letters Eytinge had given the dealers a straightforward account of what the articles really were. He told the truth absolutely and
unvarnished, and one day he discovered that he had been doing this.
Does honesty pay? This was the question that Eytinge asked himself. He did not have to ask others—he had the proof before him. He had discovered that honesty did pay and that it paid well. But just at that time this new business received a severe setback. Prison officials became suspicious that many of the letters going out from the prison were not what they purported to be. Some of the salesmen were suspected of filling their letters with fake statements which were little short of appeals for help.
Word went out from the warden’s office that each prisoner could write but two letters a week. Eytinge had been dealing with forty retail merchants. Many a man would have thrown up the sponge and quit. To have thirty-eight customers taken away in that manner would have proved a shock that few concerns would have been able to withstand. But Eytinge did not admit defeat. He soon recovered from the shock and saw that something must be done. Two letters must be made to do what forty had been doing before.
The wholesale field was the only thing left. Eytinge began again through that channel. And the letters he wrote! They were full of business from the “Dear Sirs” to the “Yours Very Truly.” These letters did not whine. Eytinge did not have the time nor space for that. He had to talk business, and he talked it so well that it was not long until the two letters were doing all that the forty had ever done.
In the meantime a new administration had taken hold of affairs in Arizona. The State prison was removed from the banks of the Colorado up the Gila Valley to Florence, and with this change came a change for the better in Eytinge’s health. When he entered the penitentiary at Yuma he weighed 119 pounds and now he weighs 190 pounds, is the picture of health and physicians declare that he is cured of tuberculosis.
Finding himself had given Eytinge two things. It gave him a new self, one of which he was as ignorant of as you or I, and it also gave him health. The necessity for money with which to fight off death had done this and it had been his desire for money that had placed him
in a position where he had to get it by some other means than by placing some one else’s name on a check.
By this time Eytinge’s letters had begun to attract the attention of people other than those with whom he was dealing. He had become a student of criminology also, and when Governor George P. Hunt, the first Statehood governor of Arizona, began his fight for better penal laws after his inauguration in February, 1912, Eytinge had published a booklet of which he was the author, dealing with this question. A noted sociologist has pronounced this to be the cleverest thing ever published on the question.
Magazines soon began to find Eytinge a man worth telling their readers about. Advertising booklets such as “Letters,” published in Chicago; “Printers Ink,” and others sent men to Arizona to “write Eytinge up.” In this way he became well known to men who held high positions in the advertising field. His advice was sought by business houses whose sales forces needed bolstering up, and he has also been offered numerous positions as business manager and business counsel.
Eytinge has made thousands of dollars out of his business since in prison, but in an article a fellow prisoner has written about him it is said that Eytinge is in debt. Why? Well, principally because he is a man, and goes out of his way to help others. These laws of ours give a discharged prisoner a five-dollar bill and a cast-off suit of clothing when he starts out to face the world. Eytinge says this is not enough, and thousands agree with him, so when he sees a man leaving prison who is deserving and in need, he is the first to come forward with financial aid. He has helped hundreds of men fight through the courts for their freedom, and the families of hundreds of prisoners can thank Eytinge for money that has tided them over the prison term of the head of their house.
Don’t forget that Eytinge now before you is the man who seven years ago entered the penitentiary many thought to die. Yes, the old Eytinge is dead—Eytinge, the crook, has passed away, but it has left an Eytinge whom the world will be glad to greet. Peter Clark Macfarlane, a Collier’s writer, says this of Eytinge: “They had penned
