viridis. Third, I examined whether changes in soil-quality alter coffee-plant quality (N
and caffeine), and how soil-quality and the presence of ant-mutualists regulate population
size of C. viridis, the abundance of ant partners, the abundance of natural enemies, and
the insect community as a whole. The laboratory experiment suggested N fertilization
resulted in greater leaf N and phloem caffeine exudation. However, root, stem, leaf, and
total caffeine did not differ with fertilization treatment. In the field, I found a positive
association between leaf N and the density of C. viridis. Stepwise multiple regression
revealed that A. instabilis ant activity and leaf N explained 45 % in the variation in C.
viridis abundance. Caffeine concentration in coffee leaves and phloem exudates had no
relationship with C. viridis density. Finally, in the field experiment, high-soil-quality
resulted in higher plant growth rate, greater leaf N, and greater phloem caffeine. Highsoil-quality
plants had 45% greater C. viridis population sizes than plants in the lowquality
treatment. Additionally ant-exclusion also increased scale population size by
34% relative to control plants. However, greater C. viridis density did not result in greater
recruitment of A. instabilis. Overall, there were 40% more arthropods on high-soil-quality
treatment plants relative to low and 20% more arthropods on ant-excluded plants than
control. However, this result was driven by changes in the abundance of ants and
hemipterans. Additional experimentation revealed Pheidole sp. ants increased recruitment
by more than three times to C. viridis on high-soil-quality plants relative to low. Thus,
soil-quality, plant growth rate, and percent N appear to be important in the distribution
and population size of C. viridis and for recruitment of Pheidole ant-mutualists.
However, other factors likely more strongly affect A. instabilis tending of C. viridis.
Further, caffeine may not be an important regulator of C. viridis in field settings. Overall,
v
soil-quality had important impacts at the community level through changes in abundance
of the dominant arthropod group (ant-hemipteran mutualism) on coffee seedlings. This
study contributes to the under-represented literature describing the effects of bottom-up
forces in tropical systems by 1) showing how host-plant quality can affect the density and
growth of a herbivore in coffee agroecosystems and 2) comparing the indirect effect of
soil-quality on the recruitment of multiple ant species to hemipteran mutualists.