Hydrogen isotope compositions have been measured on pore waters from sediments of Leg 129 sites in the Pigafetta and East Mariana basins (central western Pacific). Total water (pore + sorbed waters) contents and their dD have been analyzed for three samples that contain smectite but no zeolite so that sorbed water can be attributed to interlayer water. The H budget for pore and total waters implies that interlayer water is 20 per mil to 30 per mil depleted in D compared to pore water. Because the interlayer/total water molar ratio (0.25 to 0.5) in smectitic sediments is very high, interlayer water represents an important reservoir of D-depleted water in sediments.dD depth profiles for pore water at Sites 800 and 801 show breaks related to chert and radiolarite layers and are relatively vertical below. Above these chert units, pore waters are similar to modern seawater but below, they are between -10 per mil and -5.5 per mil. These values could represent little modified pre-Miocene seawater values, which were D-depleted because of the absence of polar caps, and were preserved from diffusive exchange with modern seawater by the relatively impermeable overlying chert layers. At Site 802, dD values of the pore waters show a decrease in the Miocene tuffs from 0 per mil values at the top to -8 per mil at 250 mbsf. Below, dD values are relatively uniform at about -8ë. Miocene tuffs are undergoing low water/rock alteration. A positive covariation of dD and Cl content of pore water in the tuffs suggests that the increase of dD values could result from secondary smectite formation. Low diffusive exchange coupled with D enrichment due to alteration of preglacial waters could explain the observed profile.

Supplement to: France-Lanord, Christian; Sheppard, M F (1992): Hydrogen isotope composition of pore waters and interlayer water in sediments from the central western Pacific, Leg 129. In: Larson, RL; Lancelot, Y; et al. (eds.), Proceedings of the Ocean Drilling Program, Scientific Results, College Station, TX (Ocean Drilling Program), 129, 295-302