The detachment of the gelatinous layer (G-layer), often observed on microtome cross sections, has led some authors to believe that the G-layer cannot act as the driving force of longitudinal shrinkage in tension wood. The aim of this study was to observe the detachment of the G-layer along fibers. Green wood blocks were cut transversely into two samples. One sample was kept in water and the other was oven-dried. With one face being common to both samples, the detachment of the G-layer was studied on the same fibers. Observations were performed after blocking deformation by embedding. This revealed that the detachment of the G-layer is an effect produced by the act of cutting the transverse face of the wood block to be embedded. At distances greater than 100 mu m from this primary surface of the sample, no detachment was observed. Drying shrinkage shows little or no effect on this detachment. The result seems to explain well why the detachment of the G-layer occurs during sectioning using conventional sliding microtomy. These observations prove the adhesion of the G-layer in massive wood and confirm the active role of the G-layer in tension wood properties.