1. Although plants and herbivores interact under varying soil resources and natural enemy effects, little is known about how these factors influence plant-herbivore interactions and shape the evolution of plant and herbivore traits. Here we ask whether soil fertility and parasitoids shape selection on fruit number imposed by a seed predator (SP) on the perennial herb Ruellia nudiflora. 2. We used a common garden where half the plants of 14 genetic families were fertilized, and recorded the abundance of cleistogamous (CL) fruits and seeds, SPs, and parasitoids. We calculated relative fitness per family based on CL seed number under the following three scenarios: Three trophic levels (accounting for SP and parasitoid effects), two trophic levels (accounting for SP but not parasitoid effects), and one trophic level (fitness in absence of SPs), and compared selection strength on fruit number between trophic scenarios and fertility environments. 3. In unfertilized conditions, SPs selected for increased CL fruit number, whereas parasitoids dampened (but did not eliminate) this selective impact. With fertilization, however, selection by SPs was reduced and unaffected by parasitoids. 4. Synthesis: Overall, we show that parasitoids can shape herbivore selection on plants, but that both herbivore and parasitoid selective impacts depend upon the abiotic environment. These findings underscore how linkages between abiotic factors and trophic complexity influence the ecological and evolutionary outcomes of species interactions.
Fruit, seed predator, and parasitoid abundance
Data were collected in the field from a common garden experiment conducted in 2011, near the town of Molas (Yucatan, México). These data include the number of fruits, seed predators, and parasitoids recorded per plant, used to analyze and describe insect functional responses. Abbreviated headlines are as follows (from left to right): plant maternal family ("genotype"), fertility treatment (NF = not fertilized, F = fertilized, "treatment"), individual plant sampled ("plant"), number of sampled fruits ("fruit#"), number of fruits with a seed predator or seed predator abundance ("seedpredator#"), and number of attacked fruits by the seed predator with a parasitoid or parasitoid abundance ("parasitoid#"). Data were collected over a two-month period (November-December) in 2011.
fruit_insect_abundance.csv
seedsrescued_fitness
Data were collected in the field from a common garden experiment conducted in 2011, near the town of Molas (Yucatan, México). These data include the number of fruits, seed predators, and seeds rescued by parasitoids from seed predator consumption, used to calculate relative fitness under scenarios of: one trophic level (potential seed number, i.e. fruit#*mean # of seeds per fruit), two trophic levels (potential seed # - seeds consumed by seed predators assuming no seed rescue [seed predator #*mean # of seeds per fruit]), and three trophic levels ([potential seed # - seeds consumed] + seeds rescued by parasitoids). Abbreviated headlines are as follows (from left to right): plant maternal family ("genotype"), fertility treatment (NF = not fertilized, F = fertilized, "treatment"), individual plant sampled ("plant"), number of recorded fruits ("fruit#"), number of fruits with a seed predator (i.e. seed predator abundance, "seedpredator#"), number of seeds consumed by seed predators assuming no seed rescue ("consumedseed#(no rescue)"), and the number of rescued seeds by parasitoids ("rescuedseed#"). Data were collected over a two-month period (November-December) in 2011.
Seeds per fruit per genotype
Data were collected in the field from a common garden experiment conducted in 2011, near the town of Molas (Yucatan, México). Data are the mean number of seeds per fruit for each of 14 maternal plant families, shown separately for each of two levels of soil fertilization. These means were used to calculate the potential number of seeds produced per plant (# of fruits recorded*mean # of seeds for the appropriate family-treatment combination). Abbreviated headlines are as follows (from left to right): plant maternal family ("genotype"), mean number of seeds per fruit in fertilized conditions ("#seedsperfruitF"), and mean number of seeds per fruit in unfertilized conditions ("seedsperfruitNF"). Data were collected over a two-month period (November-December) in 2011.
seedsperfruit.csv