Multilevel selection on offspring size and the maintenance of variation

Data used to model multilevel selection on egg size under sperm saturating (H) and sperm limiting (L) conditions for the broadcast spawning polycheate, Galeolaria caespitosa. We used a split-clutch design to conduct in vitro fertilisations, whereby each female's clutch was split across 2 replicate fertilisations (Rep) per sperm treatment. We then measured the size-specific fertilisation success of all eggs within these clutch subsamples as a binary response variable (0 = unsuccessful eggs; 1 = successfully fertilised eggs). Note that values of clutch-mean in the dataset represents the mean-egg size of each female's clutch measured from photographs taken prior to fertilisation. Egg size represents the inital size of each indiviual egg prior to fertilisation, and has been back-calculated from measurements taken after fertilisation using equations that describe these relationships.