For masses larger than 2 % + MT, narrow peaks in two proton invariant (missing) masses spectra : M (M ) are now well established. The existence of such PP x narrow peaks is not so unquestionable below 2 $ + Mr. Tensor analyzing power T20 + of the invariant mass M has been studied using the p(d,pp)~ reaction. Depending PP on the assumption done for background substraction -a hble or an oscillation mayappear at M = 1941 MeV. Additional data are under analysis and will improve the statistics. PP It is well known that narrow structures have been obsenred in two nucleon invariant masses, or missing mass experiments/l/. These structures have been tentatively attributed to isovector dibaryons although no proof has been advanced to associate them with quark degrees of freedom. This assumption relies only on the fact that such structures are not present in classical nuclear physics using meson exchanges and nucleon or delta degrees of freedom. The experimental certainty depends mainly on the quantum numbers and the mass of studied dibaryons. Up to now, none has been observed in T = 0/2/ or T = 2/31 isospin states. In case of T = 1 isospin, for the range of masses larger than 2 $ + M* (M > 2.016 GeV) there is a good agreement for narrow peaks between results obtained at Saturnel41 : measurement of d2a/dQdMx for 38e(p,d)~ reaction, and Lanpf/5/ where Ay was measured. These data correspond to the second generation of experiments, since they have used electronical detectors. ('1 Present address : Ben Gourion University of the Negev, P.O. Box 653 84105Beer Sheva, Israel. Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1990631 C6-372 COLLOQUE DE PHYSIQUE Below 2.016 GeV, many structures have been presented, corresponding usually -but not alwaysto the first generation of experiments studying bubble chamber pictures. Here the masses of experimentally observed structures do not concentrate around a few values, but are more or less distributed from 2 % to 2 % + M?/1/. The necessity of a precise second generation experiment close to the two nucleon mass appears then naturally. 3 Such an experiment has been carried out at SATURNE with the p(d,pp)X reaction. Tensor polarized deuteron beam of 1.722 GeV has been used and the two protons detected at 0 using the SPES3 facility. The properties of the spectrometer : pmax/pmin z 2.5, and As2 = sr allow the study of the following two proton invariant mass range : 2 % < M < 1.963 GeV, PP the counting rate decreasing regularly from small to larger masses M . This result is mainPP ly due to pp FSI giving rise to a narrow maximum a few MeV above 2 M . Due to the granulaP rity of the detection, when two trajectories intercept a same scintillator of one among the two major trigger detection plans or give rise to two mixing tracks on the focal pla~e, the event is lost. This produces a loss of efficiency in the first three MeV of the spectra.This explains that our FSI is located around 1880.7 MeV, that is ?. 3 MeV higher than the peak maximum from low energy experiments. The detection has been described elsewhere/3/. The empty target accounts for about 5 % of the counting rate. The FWHM of the experimental wcdth for M is 1.2 MeV. The data have been therefore binned into 2 MeV intervals. The M hisPP PP togram presented in Fig. 1 is quite continuous, without any discontinuity having a number of standard deviations (S .D .) larger than 2.2. id = 1.722 GeV epp =0° 0.2 188