While intercropping is known to have positive effects on crop productivity, it is unclear whether the effects of mixing species start at the very beginning of the plants’ life, i.e. during germination. We tested whether germination is affected by mixing cover crop species, using two legume species with a contrasting response to water availability and temperature, alsike clover (AC) and black medic (BM). We set up four experiments in each of which we compared a 1:1 mixture against the two monocultures, and combined this with various other experimental factors. These additional factors were (i) varied seed densities (50%, 100% and 150% of a reference density) in two field trials in 2016 and 2017, (ii) varied seed densities (high and low) and water availability (6 levels, between 25% and 100% of water holding capacity, WHC) in a greenhouse pot trial, (iii) varied seed spacing in a climate chamber, and (iv) varied temperatures (12°C, 20°C, 28°C) and water availability (4 levels between 25% and 100% of WHC) in a climate chamber. Across all experiments, the absolute mixture effects (AME) on germination ranged between -9% and +11%, with a median of +1.3%. Within experiments, significant mixture effects were observed, but the direction of these effects was inconsistent. In the field, AME on germination was significantly negative at some of the tested seed densities. A positive AME was observed in the climate chamber at 12°C, and the mean AME decreased with increasing temperature. Higher density was associated with decreased germination in the field, indicating negative interaction through competition or allelopathy, among seedlings. Our findings indicate that seed interaction in mixtures may be ongoing as early as during germination, but that the direction of the mixture effect is affected by complex interactions with abiotic and biotic factors.