The spatfall of mussels on artificial collectors sustains the production of similar to 350,000 t of mussels per year in Chile, which is known as "capture based aquaculture" (CBA), although little is known about the processes that determine a successful or unsuccessful spatfall. We studied the vertical variability of the larval supply and spat of Mytilus chilensis, Choromytilus chorus and Aulacomya atra in a marine farm located in one of the most important fjords used for spatfall in Chile. Larval supply was sampled five times per day (between 8 am and 6 pm) on three dates, in four depth strata between 0 and 13 m. Environmental variables from the water column were also recorded during each sampling. The spatfall was sampled monthly from December 2014 to June 2015 in the same four depth strata by installing 13 m long collectors in the study area. Larval supply of the three Mytilidae species showed few clear patterns, although the water column showed wide variability in temperature, dissolved oxygen, salinity, and sigma-t. The spatfall of M. chilensis and A. atra showed almost homogeneous vertical distribution at first. However, after six months, a stratified distributional pattern of spat emerged, with M. chilensis dominating between 0 and 7 m depth, and A. atra dominating between 7 and 13 m depth. In both species there was similar to 90% spat loss in six months from their maximum abundance, which needs to be studied. Our results suggest low production efficiency of spatfall in the study area and we concluded that improving the efficiency of spat collection and monitoring mussel spatfall is a priority, including their potential relationship with natural beds in the vicinity.