History of sequential herbivore attack affects resistance against larvae of P. xylostella
The sequence of herbivore attack affected the capability of B. nigra plants to defend against larvae of P. xylostella (F1,12=3.42, P<0.001, Figure 2). A sequence of attack by four episodes of leaf-chewing caterpillars of P. xylostella induced resistance to larvae of P. xylostella that were feeding on these plants as fifth herbivore (t1,2=-2.17, P=0.032, Figure 2). No resistance was induced by a sequence of four episodes of attack by the other leaf chewer A. rosae. Performance of P. xylostella caterpillars feeding on plants induced by a history of four rounds of A. rosae did not differ from P. xylostella caterpillars feeding on plants that did not receive herbivory. Sequential attack by four rounds of aphid attack by either M. persicae or B. brassicae did not affect the performance of P. xylostella caterpillars as compared to performance on undamaged plants. However, specific sequences of herbivore attack affected the performance of P. xylostella caterpillars compared to performance on undamaged plants. Plants became more resistant to P. xylostella larvae after plants had been exposed to sequential attack by M. persicae – M. persicae – A. rosae – A. rosae (t1,2=-4.42, P<0.001, Figure 2) and by M. persicae – A. rosae – M. persicae – A. rosae (t1,2=-2.34, P=0.020, Figure 2).
Importance of specific events in sequential attack for resistance to P. xylostella caterpillars
The differential effect of specific orders of sequential herbivore attack on performance of P. xylostella was determined by the number of switches between herbivore guilds, the identity of the first and last herbivore as well as the specific order of herbivore attack.
First, the number of times a plant was exposed to the same feeding guild of attacker as part of sequential herbivore attack, 0-, 2- or 4-times phloem feeders or leaf chewers affected plant resistance to larvae of P. xylostella (F1,3=3, P=0.04). Plants that were attacked four times by chewers tended to be more resistant to larvae of P. xylostella than plants exposed to four times attack by phloem feeders (t1,2=-1.67, P=0.097). Second, the feeding guild of the first attacker of the sequence did not affect plant resistance to larvae of P. xylostella (Figure 3a). However, the specific species identity of the first attacker of the sequence affected performance of P. xylostella (MLM: F1,4=2.89, P=0.023, Figure 3b). Plants that had been attacked first by M. persicae, were more resistant to larvae of P. xylostella independent of the order and identity of the second, third and fourth herbivore in the sequence (t1,2=-2.33, P=0.02, Figure 3b). Third, the feeding guild as well as the species identity of the last and thus fourth attacker of the sequence, affected plant resistance to P. xylostella larvae (MLM, Guild: F1,3=-8.12, P<0.001; Identity: F1,4=5.11, P<0.0001, Figure 3c, d). When the last attacker had been a chewer, plants were more resistant P. xylostella larvae (MLM, chewer t1,2=-2.33, P=0.021) and these effects were particularly apparent for the leaf chewer A. rosae (MLM, t1,4=-2.78, P=0.006, Figure 3c, d).
In summary, the history of sequential herbivore attack affected B. nigra resistance to larvae of P. xylostella. The number of times there was a chewer in the sequence, the species identity of the first attacker in the sequence and the guild and the identity of the last (i.e. fourth) attacker in the sequence affected the capability of plants to mount resistance to P. xylostella after previous exposure to multiple incidences of herbivore attack.
To further separate effects caused by number of switches between feeding guilds from effects by feeding guild, species identity and order of arrival, we analysed these effects within subsets of treatments that were equal in the number of herbivore switches. Brassica nigra plants attacked by sequences containing two times chewers, and two times phloem feeders, were more vulnerable to P. xylostella larvae when the first attacker was a chewer, compared with plants that were first attacked by a phloem feeder (F1,2= 10.85, P=0.001). After first attack by a phloem feeder, the order of subsequent attackers influenced plant resistance to P. xylostella, as larvae grew more on plants attacked by the sequence aphid – chewer – aphid – chewer (B. brassicae – P. xylostella – B. brassicae – P. xylostella as well as M. persicae – A. rosae - M. persicae – A. rosae) than on plants attacked by the sequence aphid – aphid – chewer – chewer (B. brassicae – B. brassicae – P. xylostella – P. xylostella as well as M. persicae – M. persicae – A. rosae – A. rosae) (MLM, F1,2=5.06, P=0.03; Figure 4a). When the first attacker was the phloem feeder B. brassicae, plants were equally vulnerable to larvae of P. xylostella, irrespectively of the order of the subsequent attackers (F1,2=1.34 P= 0.25, Figure 4c). In contrast, when the first attacker was M. persicae the order of the attackers mattered and larvae of P. xylostella grew more on plants where the identity of the herbivore switched every time (larvae grew more on plants that had been exposed to M. persicae – A. rosae – M. persicae – A. rosae than on plants that had been exposed to M. persicae – M. persicae – A. rosae – A. rosae) (F1,2=4.08 P=0.04, Figure 4d). In contrast to the effect of herbivore order when the first attacker was an aphid, the order of herbivore arrival after the first herbivore was a leaf chewer did not affect plant resistance to P. xylostella larvae (F1,2=0.75 P=0.39, Figure 4b). These effects were also similar for the two leaf chewers P. xylostella or A. rosae (P. xylostella first: F1,2=0.05 P=0.83; A. rosae first: F1,2=2.08 P=0.15, Figure 4e, f).
In summary, B. nigra plants attacked by sequences containing two chewers, and two phloem feeders, were more vulnerable to P. xylostella larvae when the first attacker was a chewer, compared with plants that were first attacked by a phloem feeder. Additionally, when plants were first attacked by an aphid, the order of subsequent attackers in the sequence influenced plant resistance. Plants that were attacked by the sequence aphid – chewer – aphid – chewer were more vulnerable to P. xylostella larvae than plants attacked by the sequence aphid – aphid – chewer – chewer.