“…This "hydrolytic weakening" has been observed in both diffusion and dislocation creep regimes (Mei and Kohlstedt, 2000a, Mei and Kohlstedt, 2000b, Karato et al, 1986, Demouchy et al, 2012, Girard et al, 2013a, Tielke et al, 2018, Jung and Karato, 2001, Hirth and Kohlstedt, 2003, Mackwell et al, 1985, as has an increase in the silicon diffusion rate (Costa andChakraborty, 2008, Fei et al, 2013) with these two mechanisms hypothesised as linked (Fei et al, 2016). The strain rate of forsterite has been found to increase in the presence of water from around half an order of magnitude (Mei and Kohlstedt, 2000a, Mei and Kohlstedt, 2000b, Fei et al, 2013, Girard et al, 2013b, Tielke et al, 2018, Demouchy et al, 2012, Umemoto et al, 2011 to a couple of orders of magnitude (Costa and Chakraborty, 2008, Karato et al, 1986, Jung and Karato, 2001) with these differences often explained by experimental differences in strain rates, compositions and water content with grain boundary water likely an important confounder of results. The two works at the highest pressures (4-8 GPa (Fei et al, 2013, Girard et al, 2013a) where [HSi] should be the largest found the lowest weakening which could be evidence that either [HSi] is not important in Si diffusion or that Si diffusion is not important in the rheological strength of forsterite (which would also explain the difference between direct strength measurements and those from Si diffusion measurements) but may be evidence that other studies have overestimated the weakening effects.…”