Yield stress values of four thickened (high-density) mineral tailings at varying solids concentrations were determined using three different techniques. The first set of values was measured using the modified slump test approach with an open-ended cylinder having an aspect ratio of 1.2. A second set was derived from measurements obtained from a coaxial cylinder fitted to a Rheolab® MC1 rheometer. The results were graphically compared with those obtained using the miniature vane technique, a popular and well-accepted method of measuring yield stress. Empirical relations developed from the modified slump test appear to predict reasonably accurate yield stress values up to about 200 Pa when compared with the vane and rheometer results. It is concluded that, although the time-dependent nature of the tailings tested may induce errors as much as ±30% for some samples, the modified slump test provides a reliable and simple test for evaluating the yield stress of thickened tailings. The method can therefore be employed as a first approximation of the changing parameters of high-density thickened tailings in the field without necessarily resorting to sophisticated equipment.Key words: high-density thickened tailings, rheology, shear yield stress, slump, torque.

One method for the placement of thickened tailings is the central discharge technique. The beach slopes achieved with this method have generally been in the order of 2 to 3%. With such small beach slopes, it is important to be able to predict the slope angles accurately. This is often done using laboratory flume tests. However, these predictions have tended to overestimate slope angles. This paper presents a model that takes account of the wall friction in flume tests and illustrates the folly of using flume tests indiscriminately. The model does not account for issues such as deposition rate or initial velocity, but serves to quantify the potential errors of using flume data for direct extrapolation to field applications.

We are pleased to respond to the valuable comments provided by the discussers. They present results that show a poor correlation between the yield stress of nonplastic tailings determined by the vane test and those determined by the "modified slump test" which is contrary to our results and to those presented by other workers, such as Pashias et al. (1996). Before addressing this specific anomaly, we address other issues raised by the discussers.Contrary to the assertion that the correlation we discuss was not established below a yield stress of 100 Pa, a number of tests using one or all of the three techniques described (vane, slump cylinder, and rheometer) were carried out on material having a yield stress less than this value. This can be seen from Fig. D2 provided by the discussers.Our reference to the time-dependent nature of the tailings was simply that it is desirable to keep all factors such as time of mixing and time between mixing and testing as constant as possible to avoid any time-related changes to the tailings structure but that this is not completely controllable when using different types of tests that have different establishment times. The effect of time of mixing (as well as relaxation time) can be fairly significant, and it is commonly accepted that when trying to predict field behaviour from laboratory tests these effects should be simulated as closely as possible.The discussers query the type of rheometer used in our study. Although it is mentioned in the text, perhaps we did not make it explicit enough: a Paar Physica MC1 Rheolab instrument, capable of both stress-controlled and straincontrolled shearing, was used.We turn now to the crux of the discussion, namely the discrepancy reported by the discussers between yield stress values derived from slump and vane tests. They do not explain whether their slump data interpretations were made using the exact theoretical solution or the approximate result provided in the original work by Pashias et al. (1996). Nevertheless, it is worth considering their results. The slump tests provided in Fig. D1 of the discussion indicate virtually no change above 76% solids. This could be partly due to the difficulty of measuring very small slumps accurately. In addition, as shown by Clayton et al. (2003), when the slump amount is small (and the yield stress thus large), it is advisable to use larger cylinders for the slump test, maintaining a 1:1 height to diameter ratio. Although one would have expected the discussers to do this, the size of cylinders used is not given, and perhaps they only used a 100 mm high cylinder. The significant deviation of the slump results in Fig. D1 from the vane data would then be consistent with the findings of Clayton et al. (2003), who indicate that above a (vane determined) yield stress of 400 Pa, one should be using larger cylinders. Perhaps much of the difficulties reported by the discussers could be solved through this simple check.The reason that a linear correlation between the slope angle established in a laboratory flu...