Fig. 1.

Parasitic wasp preference for volatiles from Fabaceae plants infested with aphids with or without the endosymbiont. Parasitic wasp (Aphidius ervi) response to volatiles emitted by plants infested with aphids (Acyrthosiphon pisum) with or without the endosymbiont Hamiltonella defensa. For each test, the bars show standard errors and the asterisks the significance of the deviation from no choice (t-test: **p < 0.01; ***p < 0.001). a Tests where Vicia faba plants had been induced by 20 wingless aphid females each (n = 10; column i—cured 101: t ₉ = −8.78, p < 0.0001; ii—cured 132: t ₉ = −5.82, p = 0.0001; iii—cured 302: t ₉ = −3.41, p = 0.0039, iv—microinjected 132: t ₉ = −4.27, p = 0.0010; v—microinjected 404: t ₉ = −3.11, p = 0.0062, vi—cured 132: t ₉ = −3.56, p = 0.0031). b Tests with V. faba plants induced with double the number of symbiont-bearing compared to symbiont-free aphids (n = 10; column vii—cured 101: t ₉ = −6.57, p < 0.0001; viii—cured 132: t ₉ = −8.73, p < 0.0001). c Tests with the original host plants Ononis spinosa (n = 10; column ix—cured 101, t ₉ = −6.58, p < 0.0001) and Lotus pedunculatus (column x—cured 132, t ₉ = −6.28, p < 0.0001). Microinjections were performed into an aphid clone naturally lacking any secondary symbiont, and which was collected on Lathyrus pratensis. Note that the same symbiont strain and aphid clone may have been used in different tests