Treatment effects on food web variables.

Results of linear mixed-effects modeling of food web variables as a function of experimental treatment. Effect sizes are given for the three planned contrasts: high vs. low SL variance (HV–LV); fish present vs. absent (Fish – NF); enclosed vs. outside enclosures (In – Out), while the random block effect was included in the model but is not shown. Significant or marginally significant effects are indicated in boldface () or italics (). Variables marked with an asterisk (*) were log-transformed to improve normality.