In a recent paper (1) it has been discussed the set of measurements performed at 300 ~ and 77 ~ in order to understand the temperature effect on the 2-u positron annihilation in crystals, according to which, by lowering temperature, the area under the angular-distribution curve decreases in single crystals but not in polycrystals. From the experimental data it arose a possible explanation (1) of the effect, as due to positrons annihilating in flight when the temperature decreases in single crystals. In these conditions, positrons would save a fraction of their initial energy because of channeling between lattice planes. This brings to a noncomplete thermalization of the positrons, which consequently annihilate at wide angles spread quite far from the usual range of the angular distribution. In fact, this measurement performed (1) for a Pb single crystal up to 30 mrad did not show any increase, within the experimental errors, on the tails of the curve detected at low temperature. Thus the h lost in the central part of the curve should have a distribution so large in angle that the corresponding mean energy saved has to be detectable on the V spectrum. For this reason we performed high-resolution energy spectra at 300 ~ 77 ~ and 4.2 ~ for the y-radiation of positrons annihilating in a Pb single crystal which was one of the samples with a very large temperatnre effect (1). The experimental apparatus reported in Fig. 1 consisted in a eryostat supplied by the Oxford Instruments, which allowed, by the cold-finger technique, the same experimental conditions at any temperature, and a Ge-Li detector whose resohltion on the 511 keV annihilation gamma line was of 3 kV FWHM, while the electronic chain was stabilized within 100 eV/h by a computer PDP 8/1 working on two leference peaks which went through it, together with the 511 keV annihilation line. For each temperature several spectra have been performed which have been compared with each other, both for the same and for different temperatures.