The use of registered names, trademarks, etc. in this pUblication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.

The influence of grain size and water content on the high-temperature plasticity of olivine aggregates was studied, using a gas-medium high-pressure deformation apparatus. The specimens used were hot-pressed, dense olivine aggregates with controlled grain size ranging from a few to 70 /•m, with or without added water. Mechanical tests were made at 1573 K and 300 MPa confining pressure and at strain rates of 10 -3 to 10 -• s -• The results reveal two distinct mechanisms of deformation, depending on stress level and grain size. At relatively high stress and large grain size, the strain rate is proportional to about the cube power of the stress and is nearly independent of grain size. In this regime, microstructural observations gave evidence of intragranular deformation involving dislocation motion. At low stress and small grain size, the strain rate depends almost linearly on stress and decreases markedly with increase in grain size. In the latter regime, little evidence was found for intragranular deformation.These observations suggest that the deformation mechanism in the grain size insensitive regime is dislocation creep, while that in the grain size sensitive regime is diffusion creep. In both regimes, water was found to enhance the creep rate.The absence of grain size sensitivity in the dislocation creep regime and comparison with single-crystal data indicate that the water weakening effect is mainly an intragranular process. However, the existence also of a water weakening effect in the diffusion creep regime indicates that water also enhances diffusion.The extrapolation of the present results to coarser grain sizes indicates that the transition from dislocation to diffusion creep occurs at 0.1-1 MPa for 10-mm grain size. Therefore it is suggested that this transition may occur in the upper mantle and that, in both regimes, the presence of trace amounts of water will result in significantly lower creep strength than under strictly "dry" conditions. motion, although the details of the controlling the effect of water. It is well appreciated that the processes are still controversial [e.g., Hobbs, 1983; earth's mantle contains some fluids including water Kohlstedt and Ricoult, 1984].However, the and carbon dioxide [e.g., Ringwood, 1975]. Earlier studies have suggested a significant weakening effect of water [Carter and Av•

L'analyse des calibrations existantes pour différentes substances montre que l'intensité de l'absorption des I.R. de 3 µm par les hydroxyles dépend de la fréquence. Quand on prend en compte le facteur d'anisotropie γ, une raie unique de calibration pour le coefficient d'absorption molaire intégral I peut, en première approximation, être utilisée pour une variété de substances telles que les verres de silicate ou autre, le quartz et les diverses formes de l'eau. La relation utilisée est I/γ = 150(3 780 — v) où v est le nombre d'onde en cm⁻¹ et I est donné en cm⁻² par mole H/1. Avec cette relation, on peut, en première approximation, estimer le contenu d'hydroxyle associé à l'absorption en bande large ou de type gel dans le quartz et les substances semblables. Ceci implique que la distribution des forces de liaison des hydroxyles est beaucoup plus affectée aux grands nombres d'onde que ne le suggère a priori la forme du spectre.

Specimens of two natural dunites (Anita Bay, of 100 •m grain size, and •heim, of 900 •m grain size) have been deformed in these dunites are of the order of 0.01 wt %, 1000 ø to 1300øC and constant strain rates between amounts much less than most estimates of the water 10 -3 and 10 -6 s -1 at a confining pressure of 300 content of the earth's undepleted •upper mantle. MPa [Chopra and Paterson, 1981•. We report here Such water may have an important role in determining upper mantle flow behavior. in which this water resides in the rocks of the upper mantle is unclear. It ha s been proposed on the one hand that such small amounts of water could be incorporated into nominally anhydrous silicates such as olivine and pyroxene by atomic substitution [e.g., Martin and Donnay, 1972]. On the other hand, Wyllie [1971] suggests the Experimental Procedures New rheological experiments were performed with the gas apparatus and procedures described previously •Paterson, 1970; Chopra and Paterson, 1981]. The apparatus uses argon as the confining medium and incorporates an internal furnace and internal load cell. These features permit the establishment of very low thermal gradients within the specimen region and a high sensitivity in the measurement of applied load. With only a few exceptions, as indicated in the text, the 7861 7862 Chopra and Paterson: Role of Water in Deformation of Dunite experiments reported herein used long sealing jackets of iron to prevent ingress of argon to the specimens [Paterson et al., 1982]. Solid alumina spacers and thin platinum or iron foils were interposed between the specimens and the pistons. The specimens of 10-or 7-mm diameter and 20-mm length were diamond core drilled from blocks of Anita Bay dunite and •heim dunite perpendicular to their respective foliation planes. The cores were subsequently dehydrated by heating them to 1200øC for >60 hours in a controlled oxygen fugacity furnace by using a mixture of C02 and CO gases producing an fo 2 of 10-5 Pa. Following this treatment, the cores and specimen assembly ß ß ß ß ß ß ß ß ß ß Am. Mineral., 59, 81 1-819, 1974. Wyllie, P. J., The D•namic Earth: Textbook in Geosciences,

No abstract