This is a response to Prakash (2026). Discussion: Salinityand oven-drying effects on the plasticity of a marine soft clay. Géotechnique, Link to the cited article.The authors thank Dr Prakash (the discusser) for his thoughtful discussion of the paper. He rightly highlights the pioneering contributions of Indian researchers to the study of marine clays. The authors acknowledge the work of Jose et al. (1988), Rao et al. (1989, 1990), Chandrakaran (1990), Sridharan (1991, 2014) and Pandian et al. (1991). These studies presented experimental evidence providing an essential framework for interpreting marine clay behaviour. While some of those previous contributions are cited in the paper under discussion (Suwal et al., 2025), the authors welcome the opportunity to explicitly note the additional contributions and situate their work within this historical context.The discusser suggests the analysis of the experimental results in terms of the mineralogical composition of Ballina clay and its comparison against previous studies carried out on Cochin and Mangalore clays, two natural marine clays from India. Although the quantitative mineralogical composition and the pore-water chemistry analysis described in Table 1 underpinned the interpretation of the results presented in the paper, the authors agree that a comparison with other soil deposits may provide additional insight into the mechanisms controlling the plasticity of marine clays. Table 5 presents the mineralogical composition, pore fluid salinity and ion concentration for Ballina clay as well as values for Cochin and Mangalore marine clays reported by Rao & Sridharan (1992). It is important to note that fractions reported by Rao & Sridharan (1992) only consider clay minerals, whereas values reported in Table 5 include all fractions present in the natural soil (illite, kaolinite, interstratified illite/smectite, amorphous minerals, quartz, plagioclase, K-feldspar, calcite, pyrite and mica). Table 5 compares the estimated clay fractions for the three clays in which values with asterisks were calculated considering only the clay fractions from Table 1. Ballina clay shows a similar proportion of kaolinite to Cochin clay and a similar content of interstratified illite/smectite to the proportion of smectite in Mangalore clay. The illite fraction is larger in Ballina clay compared to both Cochin and Mangalore clays. The pore fluid salinity in Ballina clay is around 2·5 and 39 times greater than Cochin and Mangalore clays, respectively. Similarly, the total ion concentration in the pore fluid is 1·2 and 17·5 times higher for Ballina clay than Cochin and Mangalore clays, respectively. Data in Table 5 show that, despite some similarities, there are important differences in chemical and mineralogical composition between Ballina clay and both Cochin and Mangalore clays. Those differences play a role in the behaviours discussed below.The discusser rightly points out the well-documented differences between the behaviour of expansive clays, mainly controlled by the double layer repulsive forces (mechanism 1), and non-expansive soils (e.g. kaolinitic clays), whose behaviour is mainly controlled by attractive forces, which have an impact on the soil fabric arrangement (mechanism 2). Despite their complex mineralogical composition, previous studies carried out by the discusser and co-workers have indicated that mechanism 1 is dominant in Cochin clay whereas the behaviour of Mangalore clay is primarily controlled by mechanism 2.The results presented by Suwal et al. (2025) demonstrate that Ballina clay’s behaviour is predominantly fabric-controlled (mechanism 2), consistent with its kaolinite/illite-rich mineralogy, although minor contributions from diffuse double layer effects are evident. The following findings support this conclusion.Mineralogical analyses are not common in many engineering works and design. Nevertheless, referring to index properties seems to suggest that a misleading prediction of soil behaviour may be obtained when empirical correlations are used to estimate soil parameters without proper consideration of salinity and drying effects.Figure 11 shows four correlations commonly used in geotechnical practice, including: soil activity A = PI/% 3σ′yield) once the natural soil structure is erased by mechanical loading. Although Cc reduces to values around 0·75 upon leaching and then it increases up to 1·05 during re-salinisation, the value for natural soil is not recovered. The Cc for oven-dried soil mixed with deionised water (series 6) is 0·66 and hence the maximum error expected in practice would be 0·60 (half the value for natural soil). Other parameters being equal, this reduction in Cc would lead to a large underprediction of the total consolidation settlement (e.g. Terzaghi et al., 1996) in these soft soil deposits.In stability calculations, the shear strength obtained from laboratory or in situ tests requires correction to account for the time to failure. The time to failure correction factor for laboratory data μlab proposed by Terzaghi et al. (1996) is shown in Fig. 11(c). By using the PI for natural soil, a correction factor of 0·81 is obtained. The correction factor increases due to leaching and then reduces upon salinisation. A value of 0·88 is obtained for oven-dried soil mixed with deionised water, which represents 7% maximum expected error. The overestimation of μlab obtained by using a misleading PI will lead to unsafe stability calculations. A similar trend is observed in the reduction factor for undrained shear strength from field vane tests, μvane. Fig. 11(d) shows the correction chart proposed by Chandler (1988), where the curve for a time to failure of 104 is suggested to be used for embankments under normal rates of construction. Although the correction factor for natural soil (series 1) is 0·62, it increases up to 0·73 if the plasticity for oven-dried soil mixed with deionised water is used (11% over-prediction). As in the case of μlab, this would lead to unsafe stability calculations due to the over-prediction of the undrained shear strength.The effects of soil drying on the plasticity of Cochin clay were studied by Rao et al. (1989). They compared the variation in liquid limit for natural (moist), air-dried and oven-dried (at 110°C) clay. Changes in particle size distributions were also estimated. Results are summarised in Table 6. Three main findings are reported by Rao et al. (1989). First, there was a strong reduction in liquid limit for dried soil with minor variation in plastic limit, in line with the behaviour observed in Ballina clay. Second, there were minor differences in liquid limit and plastic limit between air-dried and oven-dried specimens (this aspect was not explored by Suwal et al. (2025)). Third, there was a strong reduction in the clay fraction in both air-dried and oven-dried specimens. It is worth noting that the reduction in clay fraction was ‘compensated’ by an increase in sand content of similar magnitude, that is without important changes in the silt fraction. This phenomenon, associated with soil aggregation, is clearly stronger in the case of Cochin clay compared to the aggregation reported in Fig. 8(b) for Ballina clay.Values in Table 6 are used in Fig. 12 to evaluate the variation in soil activity A, Cc, μlab and μvane for Cochin clay. Soil activity reduces from 1·55 (moist) to 1·21 and 1·25 for air-dried and oven-dried clay, respectively. Interestingly, the decrease in PI is somewhat compensated by the reduction in clay content, which explains the small reduction in soil activity. However, Cc reduces from 1·06 (moist) to 0·58 (air-dried) and 0·50 (oven-dried). This represents a reduction to half the value for natural soil, similar to the decrease estimated for Ballina clay. The time to failure correction factor for laboratory data μlab increases from 0·82 for natural soil to 0·89 (air-dried) and 0·90 (oven-dried), which represents 8% maximum expected error. Similarly, the correction factor for field vane tests μvane changes from 0·65 (moist) to 0·75 and 0·77 for air-dried and oven-dried soil, respectively. Similarly to the values reported in Fig. 11, this represents a 12% over-prediction of the undrained shear strength for Cochin clay.
Suwal et al. (Fri,) studied this question.