It is known that migration of the axons of neuronal cells is driven by guidance cues sensed by receptors located on the growth cone.1 Filopodia and lamellipodia, the highly motile structures extruding from the tip of the growth cone, explore the environment. Their motion has been analysed, but little is known about the force these neuronal structures exert on the structures they might find during their navigation. In fact, the analysis of this force has been limited to theoretical considerations and experimental analysis have been restricted to samples of isolated filaments2 or to migrating cells.3 In this study, we used optical tweezers4 to measure the forces exerted by filopodia and lamellipodia with a millisecond temporal resolution. We found that a single filopodium exerts a force not exceeding 3 pN. In contrast, the force exerted by lamellipodia ranged up to 20 pN or more with a duration varying from less than 1 second to more than 30 seconds. These measurements suggest that in the absence of actin polymerisation no force can be produced and that microtubule polymerisation is required in order to develop forces larger than 3 pN. These results show that neurons not only process information but also they act on their environment exerting forces varying by 1 to 2 orders of magnitude. Silica beads of 1 µm in diameter, functionalised with amino groups to reduce sticking, were trapped with an 1064 nm infrared (IR, mW power at the sample) optical tweezers close to the growth cone of a migrating axon (Fig.1a).5 The growth cone displaced the bead both laterally and axially from its equilibrium position by even 2 or 3 microns (Fig.1b). At the end the bead did not remain attached to the growth cone and could return to its original position in the trap (Fig.1c). We measured the lateral force exerted by the growth cone Fneu = (Fx, Fy) by following the bead position both by using back focal interferometry with quadrant photo diode (QPD)4 and by video tracking. When the bead was far from the growth cone the QPD recordings of Fx and Fy were quiet with a standard deviation σ of approximately 0.18 pN (Fig. 1d upper trace), but collisions producing a force larger than 5σ were observed when the bead was moved near the growth cone (Fig.1d lower trace). In several occasions, Fx and Fy increased within 1-10 seconds, reaching values of the order of 20 pN (Fig.1e) and, when the growth cone stopped pushing, the bead returned to its equilibrium position, often in less than 1 ms. As the presence of floating debris and wandering filopodia near the bead could affect the light pattern impinging on the QPD, a collision was considered reliable when the bead displacement obtained with the QPD and videotracking were in agreement (black and yellow traces respectively, in Fig. 1f) and the presence of a colliding filopodium or lamellipodium was verified by visual inspection of the movie. We have analysed collisions between growth cones and trapped beads in more than 200 experiments. Each experiment lasted for 3 minutes and in many experiments several collisions significant for statistical analysis were observed. These collisions produced maximal forces ranging from less than 1 pN to at least 20 pN with a maximal rate of increase of 10 pN/second. These collisions lasted from less than 1 sec to about 40-60 seconds. Larger forces were usually observed during long lasting collisions. As these forces extend over a wide range of intensities and durations, we measured separately the forces developed by filopodia and lamellipodia for hundreds experiments in order to have a good statistics.