The validation of the “geophysical products” derived from observations of the satellite ocean colour sensors requires the collection of the same parameters from in situ instrumentation. In this report, we focus on the validation of the water-leaving reflectance, and we examine a protocol issue, which is linked to the different ways of deriving at sea a water-leaving reflectance value that is suitable for the validation of this product as it is derived in particular from the MERIS ocean colour sensor observations (see at http://www.envisat/esa.int). The two techniques that we have tested use two very different instrument types. The first one measures the upwelling irradiance in the water column in a free-fall profiling mode, using a multi-channel submersible radiometer (see section 3 for the detailed instrument description). During the profiling, the above-water downwelling irradiance is recorded on deck. Extrapolation of the vertical profile of irradiance allows the irradiance value just below the water-air interface to be determined. The abovewater reference is corrected for the transmission from air to just below the sea surface, which allow the ratio of upwelling to downwelling irradiances to be formed at null depth. This is the reflectance, noted R. Deriving the water-leaving reflectance in a given viewing direction then requires the knowledge of the Q factor (Morel and Gentili, 1993), i.e., the upwelling irradiance to upwelling radiance ratio. The second technique measures directly the water-leaving radiance exiting the ocean at a nadir angle of about 40° and an azimuth difference with respect to the sun position of about 135°, as well as the sky radiance reflected in the same viewing direction by the wind-roughened sea surface, plus a small, albeit unavoidable, contribution from combined reflection on the ship superstructure and the sea surface. The instrument is a portable hand-held radiometer that has been designed for that purpose (Fougnie et al., ???; again see section 3 for the detailed instrument description). It is called the SIMBADA, and is an evolution of the former SIMBAD radiometer (see at http://www-loa.univ-lille1.fr/recherche/ocean_color/src ). Inter-comparison between above-water and in-water techniques have been already performed several times elsewhere using other instrument types (Frouin et al., 2000; Toole et al., 2000; Hooker et al., 2002; Zibordi et al., 2002), and sometimes with a considerable amount of details and with extremely controlled procedures (Hooker and Morel, 2002). The aim of the latter work was precisely to examine the effects of various perturbations that interplay in forming the signal measured from above the sea surface, and to assess their importance in the final error budget of such measurements, as well as the ability we have (or we don’t have) to correct for these perturbing effects. It was not our goal here to repeat such studies, and rather the work reported here had two very specific goals, as follows : (1) The SIMBADA radiometer is a newly developed instrument, and the processing of its measurements as well uses quite new concepts. Both the instrument and the data processing code need some qualification with respect to other instruments whose history is longer and which are better understood thanks to intensive use in marine optics work in the past decades. We aim at providing some elements to help in this qualification process. This includes possible improvements in the data processing procedures. (2) The suitability of using the SIMBADA from the Téthys-II research vessel (see next sections) remains to be established, when the collected data are to be used for the validation of the water-leaving reflectance products derived from the MERIS observations. Indeed, the ship perturbations, which have been identified as being sometimes preventing the derivation of the water-leaving reflectance with the desired accuracy (Hooker and Morel, 2003), are by definition specific to the ship. A specific work was therefore needed in the context of the BOUSSOLE project.