Thank you Professor Pantazidou for your comments and questions. I'll respond to them in the order you listed them
Slide 22. Validity of empirical correlations.
That is a very good point that you make - that even within sedimentary soils many correlations are valid only for the particular sedimentary soil from which they were derived. Regarding residual soils, a historical example is the design and construction of the Sasamua Dam in Kenya in the 1950s. The soil involved was tropical red clay derived from andesitic ash and rich in the clay mineral halloysite. The designers were puzzled by the very good geotechnical properties of a very fine grained soil. They got hold of Terzaghi to advise them and he agreed that the properties were very good despite the very small particle size, and offered the explanation that the particles existed in 'clusters' or 'aggregations' so behaved as a coarser material. Terzaghi's explanation was wrong as the good properties came from the intrinsic nature of the halloysite particles and not because they were in clusters.
Another example is the even more extraordinary properties of a related clay mineral, known as allophane. Clays rich in this mineral have very high natural water contents and Atterberg limits, especially the liquid limit. These can be as high as about 180%, but their engineering properties are remarkably good. They are of low compressibility, high ϕ' values (close to 40o) and do not shrink or swell.
I think it is mostly residual soils of volcanic (andesitic) origin that do not conform to many soil mechanics "norms". The allophane soils all plot well below the A-line but they behave as clays - NOT as slits.
Slide 23. A definition of "cohesion".
I am with Bishop on this issue. The word cohesion should be used with its everyday meaning, namely the property by which particles are attracted to one another or "cohere" together. On this basis soils are commonly divided into cohesive and non-cohesive soils. If used for the "cohesive" component of shear strength then Bishop always used the term "cohesive intercept". I realise, however, that "cohesion" is widely and loosely used with the latter meaning
Slide 37. Basing bearing capacity for foundation design on the undrained shear strength. The point about possible dilatant behaviour is certainly valid, but I doubt that it would be of sufficient magnitude to negate the effect of consolidation over most of the loaded area. Also, it is normal (and sensible) practice to use large safety factors in designing foundations on clay, so I think it unlikely that there would be large enough deformations to cause significant dilatant behaviour.
Slides 41 - 42. Total vs effective stress analysis
Slides 41 and 42 are concerned simply with a particular soil element, or a small zone of uniform properties. I should probably have made this clearer in my presentation. With regard to Slide 42, it will be apparent that if the soil is on the point of failure, then the strength will be the same whether expressed as the undrained strength or the strength in terms of effective stress. This is because no additional shear movement is involved and no change in the pore pressure. This is the same point made by Bishop and Bjerrum in their 1960 paper, namely that for a vertical clay bank on the point of failure the undrained shear strength will be the same as the effective stress strength.
I think I agree with the remainder of Marina's comments. The undrained strength is very likely to vary with depth or water content, and even if the clay is completely homogeneous, the undrained strength is still dependent on the method used for measuring it. The strength measured in an undrained compression triaxial test will be significantly different to that in an undrained extension triaxial test. The latter may be little more than half the former.
With regard to the last issue about the undrained analysis of a constructed slope. Such an analysis simply represents the "end of construction" case, and says nothing about the long term situation. I was simply using this example as a practical illustration of why the two methods of analysis give different answers.
I''m not sure about the statement "We construct the embankment and then it immediately becomes fully soaked". It may do so if it is a water retaining structure, but even then the soil above the phreatic surface will be a zone of negative pore pressure and not liable to become "fully soaked" Also the embankment may be for a highway or railroad, and the pore pressure throughout the embankment may be permanently negative. This will be the case if the embankment is built of the same material as the foundation soil it is built on and the water table in the foundation soil is deep. In this case the compacted embankment soil is highly likely to be of significantly lower permeability than the undisturbed in situ soil, and thus acts as a barrier between surface rainfall and the deep water table. The pore pressure at the base of the embankment will then be permanently negative and will ensure that the pore pressure throughout the embankment also remains negative.
Once again I am grateful to Prof Pantazidou for her interest and comments. With regard to my slides, as far as I am concerned there is no restriction on these. I will send the two complete power points to anyone who wants them. My email address is:
Bishop, A.W. and L. Bjerrum (1960). The relevance of the triaxial test to the solution of stability problems. Proc. ASCE Research Conference on Shear Strength of Cohesive Soils, 437-501, Boulder, Colorado, USA.
Terzaghi,K .(1958) ''Design and Performance of the Sasumua Dam, Proceedings of the Institution of Civil Engineers,Vol. 9, 369-394