Cell-wall microfibrils formed on the surface of plant cells may be either deposited without obvious order, as in primary wall formation of higher plants, or in parallel orientations, as during secondary wall formation of higher plants and in the crossed microfibrillar layers in various algae. The mechanism of orientation is an actual cytological theme with controversial hypotheses trying to correlate components of the cytoskeleton with the microfibril orientation [1, 2]. Mainly based on inhibitor experiments, there is broad agreement that in most cases microtubules play a key role in the orientation mechanism. With the recent freeze-fracture findings of "roset te" particle complexes in the plasma membrane of a variety of cellulose forming algae, moss, fern and some higher plant systems (reviews in [3, 5]), it has become rather probable that these rosette structures, and perhaps associated "terminal globules" in the other half of the membrane, are involved in cellulose microfibril formation, or even are the synthases themselves ([3-8], for discussion see [3, 4]). We now have freeze-fractured the plasma membrane of higher plant cells involved in secondary wall formation to check for the possible existence of oriented rosette arrangements during the deposition of parallel wall microfibrils. We used mung beans (Vigna radiata (L.) Wilczek) cultivated for 6 days with twice daily watering. The hypocotyl was cut into thin slices, and immediately after sectioning was frozen between double replica holders in nitrogen slush. The further procedures were as described [3]. The plasmatic fracture face (PF) of the plasma membrane of (not further defined) cells with parallel microfibril arrangement (as judged from the microfibril imprints) was relatively rich in rosettes, and showed unequal particle density depending on the region of the cell surface. Some rosettes did not show a correlated position with respect to other rosettes, but quite frequently the rosettes occurred in tracks (Fig. 1 a). There was no regular distance between the individual rosettes in such a track, and the alignment was somewhat zigzag, with up to ca. 50 nm lateral deviation from the extrapolated center line of such a track following a microfibril imprint (arrows). This arrangement would allow cooperation of adjacent rosettes in formation of microfibril bands, if each rosette would form a 3.5 nm elementary microfibril (Fig. 1 b) as postulated earlier [3]. It furthermore would allow formation of many parallel microfibrils. The adjacent tracks of rosettes were often, but not always, in parallel, and not with regular spacing to the neighbour track. These findings of an oriented arrangement of rosettes in mung bean plasma membranes are the first such case for a higher plant system, and lend further support to the postulated role of rosettes in cellulose microfibril formation [3-8]. There are some indications that individual rosettes arranged in lines also occur in the basal parts of Funaria caulonema cells [8]. The arrangement of rosettes in tracks somewhat resembles the rows of rosettes involved in formation of the parallel microfibrils in Closterium, but there the groups of rosettes show equal center-to-center spacing [5]. From a variety of experimental data, mainly involving serial section analyses of cortical microtubule arrangement with respect to microfibril orientation after inhibition of the crystallization process, it has become rather