We have electrophoretically analysed the variation in queen mating and worker paternity across and within seven populations of the ant Lasius niger in northwestern Europe. Populations were panmictic and not genetically differentiated (FST = 0.003 ± 0.004; range c. 1000 km). Queens (n = 535) were shown to mostly mate with a single male, but double mating occurred in all populations and triple mating was found in one case (the total number of worker offspring analysed for paternity was 4825). The genetically effective queen mating frequency was 1.16 on average across populations (range 1.04–1.42). Double sampling of six out of the seven populations showed that most of the variation in queen mating occurred among populations and not within populations among years. Also, paternity skew in colonies with double-mated queens was relatively constant per site but varied across populations. Paternity skew was high in populations with low frequencies of double queen-mating, low in populations with intermediate frequencies of double queen mating and ambiguous in a single population where more than half of the queens mated multiply. Double-mated queens were only collected halfway through a nuptial flight, suggesting that double mating is time consuming and that the mating swarm sex ratio may affect the likelihood of multiple mating towards the end of a flight. No difference in fresh weight between single- and double-mated queens from the same population was found.