An accurate measurement of the n-p differential cross section, in the angular range 116' -180', at 96 MeV is reported. Between 150' and 180' the angular distribution is steeper than earlier measurements and potential predictions. The sensitivity of the differential cross section to isospinsinglet, J&4 phase-shift parameters was studied, and it was found that the observed difference at the most backward angles is sensitive mainly to L)3 phase shifts.PACS number(s): 13.75.Cs, 21.30. y The upgraded Gustaf Werner cyclotron at the The Svedberg Laboratory, Uppsala, Sweden, has been equipped with a facility designed to produce wellcollimated and energetically well-defined neutron beams in the energy region from about 50 to 200 MeV. A program to study isovector excitations induced by the (n,p) reaction in various nuclei is at present in progress. The measured (n, p) cross sections are normalized to the n pscattering cross section and therefore precise knowledge of the latter at backward angles is of great importance. Accurate measurements of the n pobservables -are also required for a unique phase-shift description of the nucleonnucleon interaction. The relatively sparse and inaccurate previous measurements of the n-p difFerential cross section in the 100 MeV region motivated a new measurement in the angular range 116' -180' (c.m. ).Since the neutron facility has been extensively presented recently [1], only a brief description will be given here. The neutrons are produced by the Li(p, n) Be reaction, using 100 -200 mg/cm thick discs of lithium, enriched to 99.98% in 7Li. A system of collimators defines the shape of the neutron beam which at the (n, p) target position, 8 m from the neutron production target, has a diameter of 7 cm. A multitarget system, which consists of a stack of thin (n, p) target layers, interspaced by multiwire proportional chambers, is used. In this way it is possible to determine in which layer the reaction occurred, and corrections for the energy loss in the subsequent targets can be applied. The momentum of the charged particles emerging from the reaction target is determined in a spectrometer consisting of a dipole magnet and four drift chambers. The scattering angle is found from the trajectory through the first two drift chambers. The trigger signal is generated by a triple coincidence between two large plastic scintillators, located behind the last drift chamber, and a thin scintillator, positioned immediately after the multitarget box. The neutron timeof-fight (TOF) is measured by using signals from the thin scintillator and the radio frequency of the cyclotron as start and stop signals, respectively. This information is used to reject events from low-energy neutrons in the tail of the neutron spectrum. Together with information on the particle momentum, the pulse heights from the two large scintillators are used for particle identification, enabling separation of protons from other charged particles.Measurements of the iH(n, p) cross section were performed at an incident neutron energy ...