The Cocos plate offshore Costa Rica comprises igneous oceanic crust covered by 450 to 500 m of pelagic deposits. Along the Costa Rica convergent margin, this plate is being subducted beneath a prism wedge attached to the Caribbean plate (Morris, Villinger, Klaus, et al., 2003, and references therein) (Figs. F1, F2). During Ocean Drilling Program (ODP) Legs 170 and 205, the sedimentary succession of the subducting plate was drilled at reference sites oceanward of the deformation front (Sites 1039 and 1253) and was also recovered at a prism site (Site 1040) below the décollement (Fig. F2).
At reference Site 1039, three sedimentary units and one intrusive unit were recovered (Kimura, Silver, Blum, et al., 1997) (Fig. F3). Unit U1 consists of dark olive-green diatomaceous ooze intercalated with ash layers. Below a sharp contact, Unit U2 is distinguished by a rapid decrease in biogenic sediment and consists of dark olive-green silty clay interbedded with light olive-green calcareous clay and ash layers. Unit U3 exhibits a dramatic increase in biogenic sedimentation, changing sharply from the nearly barren clays of Unit U2 to ivory to light green and mottled siliceous calcareous oozes interbedded with calcareous clay and ash. The basal oozes of Subunit U3C are metaliferous. The igneous Unit U4 consists of fine-grained gabbros that are interpreted as intrusive sill (Kimura, Silver, Blum, et al., 1997). Below this ~35-m-thick igneous unit, another ~20 m of pelagic sediment (Subunit U3C) was recovered during Leg 205 at Site 1253 (Morris, Villinger, Klaus, et al., 2003).
Drilling at Sites 1040 and 1043 and subsequent studies (Morris, Villinger, Klaus, et al., 2003, and references therein) show that the sediment section beneath the décollement repeats the complete lithology and sequence of the subducting plate cored at Site 1039 (Fig. F3). Also, the seismic observations indicate complete sediment subduction past the prism front (Fig. F2). The thinning of the underthrust section seen between Sites 1039 and 1040 (Figs. F2, F3) must then reflect compaction and dewatering processes. This leads to a change of the corresponding lithologies from silty clay and calcareous oozes at Sites 1039 and 1253 to claystones and calcareous chalk in the underthrust section at Site 1040.
The progressive subduction of the pelagic sediment section along the Costa Rica convergent margin offers the possibility to study in detail compaction and dewatering processes during early subduction. The signature of such processes should be seen in chemical proxies due to diagenesis and alteration of the pelagic sediments. By comparing the chemical signature in the sediments at the reference sites with those of the underthrust section, effects of compaction and dewatering processes on diagenesis could be estimated. In this data report, we present results on stable isotope compositions that show significant differences in the 
18O values between the unaffected carbonates in sedimentary Unit U3 at reference Sites 1039 and 1253 and the equivalent underthrust section at Site 1040. These results could be used in further studies to test models of sediment compaction and dewatering processes during early subduction.
In addition, the stable isotope analyses of the pelagic carbonates in the reference site oceanward of the deformation front (Site 1039) are used as a chemostratigraphic tool to complement the bio- and magnetostratigraphy of the carbonate section (Kimura, Silver, Blum, et al., 1997; Muza, 2000). The occurrence of positive shifts in carbonate 13C values of up to 1.5
in the middle Miocene (known as the Monterrey excursion, between 13 and ~17.5 Ma) has been described by Vincent and Berger (1985) in a data set from Deep Sea Drilling Project (DSDP) Site 216 in the tropical Indian Ocean and correlated to several stratigraphic sections all over the world (Woodruff and Savin, 1991; Jacobs et al., 1996; John et al., 2003). The Monterey excursion comprises several higher frequency (~400 k.y. cycles) peaks (carbon isotope maxima [CM]) (Woodruff and Savin, 1991), presumably caused by orbital climate forcing corresponding to long eccentricity cycles (Cramer et al., 2003). Reproducing these isotopic peaks in the analyzed samples from ODP Leg 170 leads to an isotope stratigraphy that is supportive of the biostratigraphy established by Muza (2000), especially in the range between 270 and 370 meters below seafloor (mbsf) (Site 1039), where biostratigraphic datums are lacking and magnetostratigraphic datums reveal controversial ages.
