mildei alone and A. pacifica alone treatments, respectively. Effect of treatment on numbers of leafhoppers remaining at the end of the experiment was assessed using anova and a Tukey HSD test for multiple comparisons. Of the C. mildei initially added to cages, 53·3% were recovered during the final collection in C. mildei-only cages. Both C. mildei and C. inclusum were present in cages where they were not introduced (Fig. 1a, b). Cheiracanthium mildei-addition treatments contained more C. mildei spiders than non-addition treatments (Mann–Whitney test, U = 13·65, P < 0·001; Fig. 1a). Numbers of C. mildei were not affected by the addition of other spider species to additive treatments Epigenetics Compound Library that received C. mildei (anova, A. pacifica: F1,16 = 0·05, P = 0·82; T. melanurum: F1,16 = 0·15, P = 0·71). Only 26·6% of the A. pacifica initially added to the cages were recovered in the final collection in the A. pacifica-only treatment. Anyphaena pacifica numbers were higher, however, in A. pacifica-addition treatments than non-addition treatments (Mann–Whitney test, U = 19·70, P < 0·001; Fig. 1c). The addition of other spiders to additive treatments that received A. pacifica did not impact A. pacifica (anova, C. mildei: F1,17 = 1·01, P = 0·33; T. melanurum: F1,17 = 0·00, P = 0·96), although regression analysis uncovered a negative relationship between A. pacifica and C. mildei numbers per cage (Fig. 2a). Numbers of T. melanurum recovered at the end of the experiment amounted to 53·3% of those initially added to the T. melanurum-only treatment. More T. melanurum were in T. melanurum-addition treatments (Mann–Whitney test, U = 31·96, P < 0·001; Fig. 1d). The addition of C. mildei negatively impacted T. melanurum (anova, F1,16 = 7·00, P = 0·02) while A. pacifica had no effect (F1,16 = 0·00, P = 0·96). A significant negative relationship emerged between numbers of T. melanurum per cage and C. mildei across additive treatments that received T. melanurum (Fig. 2b). Before treatments were imposed, leafhoppers abundance was not different between additive treatments (anova, F7,32 = 1·67, P = 0·15), although there was a marginally significant effect when the replacement-series treatment was included in analysis (F8,36 = 2·00, P = 0·08). Final densities of adult grape leafhoppers tended to be lower in treatments that received C. mildei (Fig. 3a). Grape leafhopper numbers were significantly affected by the blocking factor and the addition of both C. mildei and A. pacifica to treatments, while T. melanurum and spider interactions had no detectable effects (Table 1). Adult variegated leafhoppers were also different across treatments (Fig. 3b), and were impacted by the addition of C. mildei (Table 1). No other factors affected variegated leafhopper numbers. Numbers of both leafhopper species were also significantly affected by C. mildei in regression analysis. Quadratic functions provided the best fit (Fig.