Experimental Setup
All experiments were done on bean leaves. Females and males were isolated from the base population on detached leaves at the quiescent stage, immediately before completing the last moult. This way, all individuals used in the experiment shared the same age-at-maturity, and virginity was ensured in both sexes before they were allocated to different treatments.
Previous work shows that to choose between virgin and mated females male spider mites use volatiles emitted by females and substrate cues left by females on patches . Based on this information, we created with our setup the conditions for mate discrimination to take place in an environment in which there was matching information (e.g., virgin females on patches previously impregnated with cues released by virgin females) or in which there was a mismatch between the information released by the females present on the patch and the cues that were left on the substrate by previous females (e.g., virgin females on patches previously impregnated with cues released by mated females). To create these different environments for males, multiple groups of 10 virgin females were randomly assigned to patches (leaf discs of 2.55 cm2) with 3 virgin males. Behaviour was observed for 1 hour and, when matings occurred, females were transferred in groups of 10 to a new empty patch of the same size. Simultaneously, groups of 10 virgin females were directly transferred to similar empty patches without ever being in contact with males. Both types of females were left on those patches during 24h such that they could release cues that remained on the substrate. Those females were then removed, and 5 new females (either mated or virgin) were placed on those patches. Subsequently, one focal male was added to all patches. Henceforth, for simplicity, we refer to the cues left on the patch by virgins or by mated females that were removed prior to the beginning of the mating sessions as “substrate cues” and the information emitted by females present on the patch (including their own behaviour) as “female mating status”. Note that, although we are aware that chemical information is sufficient for mate choice (Rodrigues et al. 2017), it is possible that other cues or signals (e.g., tactile, visual) are used by males as well (Royalty et al. 1993). In any case, we assume that the information provided by the presence of a female is composed of multiple cues that are concordant, and thus, for simplicity, consider “female mating status” a unit of information. Likewise, substrate cues include web, faeces, eggs and any chemical information left behind by females; once more we consider those a unit of information. We cannot rule out the possibility that new substrate cues are being deposited by the females present in the patch, but given the short duration of trials, these are not expected to play a significant role in the behaviours described.
Male and female behaviour, i.e., the number of male mating attempts, the frequency of female acceptance, the number of mating events and copulation duration were observed for 1 hour. A mating attempt was registered whenever a male touched the female with the two front legs and started bending its opisthosoma . Whenever a mating attempt resulted in the insertion of the male aedeagus into the female abdomen for more than one minute, the observer registered it as the occurrence of a mating event . It has been shown that females can reject male mating attempts by moving away from them (Rodrigues et al. 2020). As such, we also studied the frequency of female acceptance, calculated as the number of mating events over the number of mating attempts. Copulation duration was registered as the time in seconds a male spent with his aedeagus inside a female. Note that females were not removed from the patch during a mating session, so mating events could have occurred with mated females in patches with virgin females at later stages of the mating session.
Subsequently, males were transferred individually to a new patch (2.55 cm2), made from uninfested bean plants, and their survival was followed daily, to measure whether different mating histories would translate into a longevity cost. Death was classified as natural (i.e., the corpse was found on the patch) or censored (i.e., males died by drowning or by being accidentally stuck in the leaf or squeezed).
This experiment was carried out in 21 mating sessions divided in 8 days, and in total, 84 males and 420 females, corresponding to 21 males and 105 females per treatment (i.e., combination of type of substrate cues and female mating status) were observed.