344 Diane W. Davidson and Steven C. Cook
Table 20.1 Comparison (on average) of leaf-foragers versus trophobiont-tenders with respect to traits hypothesized
to confer efficacy in defense against plant herbivores and pathogens.
Trait LF relative to TTPotential for pathogen transmission (through trophobionts or wounding) LF<TT+
Parasitism of plant resources LF<TT+
Alteration of plant metabolism and development LF<TT+
Percentage of workers searching independently LF>TT+
Amount of plant surface (leaves, stems) covered per foraging worker LF>TT+
Prevention of pathogen development in wounds LF>TT+
Potential for larvae to consume solid foods (prey)a LF>TT+
Length of daily activity period LF<TT−
Total numbers of workers on plant surfaces LF<TT−
Average N-deprivation of ant taxab LF=TT=Notes: Leaf-foragers (LF): Formicinae, Ponerinae, Ectatomminae, Paraponerini, Pseudomyrmecinae, Cephalotini and
Cataulacini, following Bolton (2003). Trophobiont-tenders (TT): mainly dolichoderines, a few formicines, and perhapsCephalotes
attratus; see Blüthgenetal.(2000) and Davidsonetal.(2004). Superscripts “+”, “−” and “=” highlight, respectively, cases where
LFs should provide greater protection (or do less harm) than do TTs, where the reverse likely holds, and where effects of LFs
and TTs are apparently equivalent.
aFrom Wheeler and Wheeler (1976).
bSee Davidson (2005).
including Old World Dolichoderus thoracicus
(Khoo and Ho 1992) andOecophylla smaragdina
(e.g., Offenberget al.2004), and New World
Azteca chartifex(De Medeiroset al.1999). Never-
theless, just a few existing studies document plant
resource losses to tended trophobionts and ants
(butseeKayetal.2004),andconvincinglydemon-
strate even intermittent and context-dependent
net positive effects of ants (Messina 1981, Horvitz
and Schemske 1984, Gaumeet al.1998, Oliveira
and Del-Claro 2005).
If leaf-foragers are generally more desirable
associates than are trophobiont-tenders, plants
should have been selected to favor the former
species over the latter. In this light, Davidsonet al.
(2004) have resurrected Becerra and Venable’s
(1989) hypothesis that EFNs evolved in part to
entice trophobiont-tenders to desert their asso-
ciates and feed directly from plants, effectively
short-circuiting sap-feeders from the interaction.
In its simplest form, this hypothesis is refuted by
evidence that ant colonies respond numerically
to tend both EFNs and trophobionts on the same
plants (Buckley 1983), and by refutation of the
theory’s correlate, that nutritive values of EFNs
should exceed those of honeydews (Fiala 1990,
Blüthgen and Fiedler 2004a). However, a more
complicated revision of the theory might suggest
that EFN plants have evolved disproportionately
CHO-rich nectars (Figure 20.2) to support the
wide-ranging and energy-demanding activities of
leaf-foragers. Members of that guild could both
convey greater plant protection (Table 20.1) and,
by virtue of better exploitative competitive abili-
ties, keep the resource too low to attract takeovers
by large-bodied trophobiont-tenders, especially
Dolichoderusspecies. Because total sugar concen-
tration is directly correlated with total amino acid
content in EFN (Blüthgen and Fiedler 2004a), and
since SUCmin and AAmin are also positively cor-
related (P=0.0008 in Spearman rank test), this
suggestion need not contradict Blüthgenet al.’s
(2004a) assertion that identities of ant associates
respond to the amino acid component of EFN. EFN
plants may attract aggressive, territorial species
with cheap and abundant sugars but increase
amino acid production only when damaged by
herbivores (Smithet al.1990).
Finally, because some small-bodied ants tend
to exhibit individually low resource requirements,
high N-limitation, and carnivory (see above;
Davidsonet al.2003, Davidson 2005), as well