Highlights: • We designed a trapping technique to capture earthworms moving through the soil. • Tree rows did not act as a source of earthworms to the nearby crop alleys in spring. • Earthworm fluxes probably depend on species and seasons. Abstract: Alley cropping is the combination of tree rows and crop alleys. The tree row is covered by an understory vegetation strip (UVS), providing a beneficial habitat for many soil fauna, which could disperse through spillover to the crop alleys. However, such movements have never been directly studied. Our experiment investigated earthworm fluxes in the tree row vicinity using a trap technique, in a Mediterranean agroforestry alley cropping field cultivated with peas and planted with walnut trees. We assessed earthworm density at different distances from the UVS (0 m, 0.3 m, 1 m and 6 m) by hand sorting soil monoliths (25*25*30 cm) in spring 2019, at the start and the end of a two-month experiment. During this period, we detected earthworm fluxes by placing directional traps at 30 cm from the UVS border. Traps consisted of three glued plastic walls placed vertically in the soil. They delimited a soil block of 25*25 cm by 20 cm depth and were open on one side. More epigeic earthworms were found in the UVS and up to 30 cm from the UVS border than in the middle of the crop alley. By contrast, the earthworm Allolobophora chlorotica presented a homogeneous distribution in the plot. Trapped earthworms were mostly of the All. chlorotica species, and 1.6 times more earthworms were found in traps open towards the crop alley than in traps open towards the UVS. These results suggest that in spring, earthworms are moving more from the crop alley towards the UVS than in the other direction, probably using the tree row and its vicinity as a refuge against adverse summer conditions in the crop alley. Graphical abstract: https://ars.els-cdn.com/content/image/1-s2.0-S1164556323001085-ga1.jpg