HEATHERE.LAPINANDPAULA.MARINO PaloVerde
DEFENSEACTIVITYOFTHREEANTSPECIESONACACIACOLLINSIIINSUNANDSHADE
SARAM.HELLMUTH,PETERN.CHALMERS,MATTHEWT.KEMP,
Abstract:Localabundancesofthreeacaciaantspecies,Pseudomyrmexflavicornis,P.nigrocincta,andP.spinicola,are differentiallycorrelatedwiththelightenvironment(shadevs.sun)oftheirhosttrees.Wehypothesizedthatthis is due to physiological differences among ant species in their metabolic responses to sunlight. However, we foundnodifferencesamongspeciesinhowtheirdefensiveactivityvarieswithsunlight.Alternativemechanisms to explain the observed distributional pattern include (1) differences among ant species in their metabolic re sponsetoairtemperatureand(2)interspecificcompetitionamongants.
Keywords:lightgradient,myrmecophyte,Pseudomyrmex
INTRODUCTION Acaciacollinsii (Fabaceae) has a mutualis tic relationship with ants in the genus Pseudo myrmex.Theantsattackherbivoresthatdisturb the host plant and they trim away neighboring competitive vegetation. The ants benefit from their association with the acacia tree by taking refuge in its thorns and consuming the plants nectarandproteinaceousBeltianbodies(Janzen 1983). Three species, P. flavicornis, P. spinicola, and P.nigrocincta, have been shown to vary in distributionbetweensunandshadehabitats:P. flavicornisis more abundant in highlight envi ronments; P. nigrocincta is found primarily on shaded trees; and P.spinicola is equally abun dant in both light environments (Ermentrout et al. 2003). We hypothesized that this distribu tionalpatternistheresultofhabitatselectionto optimize metabolic activity. One indication of such optimized metabolic activity would be a heighteneddefensiveresponseunderlightcon ditions in which they occur more abundantly. Wepredictedthatdefenseactivity,measuredas theresponsetophysicaldisturbanceofthehost plant, would be higher for P.flavicornisin high light environments than in the shade, whereas theresponseofP.nigrocinctawouldbehigherin
1
the shade than in the sun. We also predicted that P.spinicola would show an equal intensity of defensive response between habitats of sun andshade. METHODS Thisstudywasconductedinadrytropi cal forest near the OTS Biological Station, Palo VerdeNationalPark,Guanacaste,CostaRicaon 10 January 2004. For each of three ant species (P.flavicornis,P.spinicola,andP.nigrocincta) we sampled four trees that had a majority of their branches in the sun and four trees that had a majority of their branches in the shade. All study trees selected were 1 3 cm in trunk di ameterand23mtall(anattempttocontrolfor antcolonysize). Oneachtree,weselectedthebranchclos est to 1.5 m from the ground and 1 m long to standardize the location of the disturbance among trees. Predisturbance ant activity was measured by counting the number of ants that crossed a line on the trunk level with this branch for 1 min. We then rapped sharply on the branch five times with a pen to simulate a disturbance,andcountedantscrossingthesame lineduringthenext1minperiod. Wequantifieddefensiveresponsetodis
DartmouthStudiesinTropicalEcology2004
turbance as DR = [postdist + 1]/[predist + 1]), where predist and postdist = ant counts before andafterdisturbance,respectively.Thismetric standardized ant activity based upon pre disturbance abundance of patrolling ants (adding 1 was to avoid division by zero). Dif ferencesinsunandshaderesponseforeachant species were analyzed with a ttest. We also evaluatedDRwithatwowayANOVAthatin cludedspecies,lightenvironment,andspeciesx light. RESULTS Both pre and postdisturbance counts varied widely among trees (predisturbance range 0119 ants/min; postdisturbance range 1403 ants/min) (Fig. 1). Predisturbance activity ant was correlated with postdisturbance ant activity, and this correlation was not af fected by light differences (Fig. 1). There were no differences in defensive response (DR) be tweensunandshadecoloniesofanyantspecies (t < 1.02, df = 6, P > 0.34; Fig. 2). ANOVA re vealed no significant differences in defensive response among the three ant species (F2,23= 1.04, P = 0.37) or between light environments (F1,23=1.23,P=0.28;F2,23=0.91,P=0.42forspe ciesxlightinteraction). DISCUSSION Our results refuted the hypothesis that sunlightwouldhavedifferenteffectsonthede fensive responses of three sympatric species of Pseudomyrmex. This suggests that the distribu tion patterns of ants, which are correlated with light gradients (Janzen 1983, Ermentrout et al. 2003), are not attributable to differences among species in their direct metabolic responses to sunlight. Interspecificcompetitionisanalternative hypothesistoexplainthedistributionalpatterns
2
of Pseudomyrmex at Palo Verde. Beadell et al. (1996) showed that tree health is correlated to light availability. If P.flavicorniswere competi tivelydominantovertheotherspeciesinsunny environments, P. flavicornis may exclude other Pseudomyrmex species from the healthier (and
300 250 200 150 100 50 0 0 20 40 60 80 100 Sun Shade P. spinicola
Post-disturbance ant activity
800 P. flavicornis 600 400 200 0
0 200
50 100 150 200 250 300 350 P. nigrocincta
160 120 80 40 0
0
20 40 60 80 100 120 140 160
Pre-disturbance ant activity
Figure 1. The relationship between pre-disturbance ant activity (number of ants) and post-disturbance ant activity (number of ants).
PaloVerde
therefore presumably nectarrich) sun trees. The other two species would then be forced into suboptimal (shadier) habitats. Under this hypothesis, P. flavicornis is predicted to win competitiveinteractionswiththeothertwospe cies. Alternatively,temperaturemightexplain the pattern in a way we have not examined. Ginsburg and Kaveeshwar (1995) found that P. flavicornis becomes more active and P.spinicola less active with increasing temperature. Be cause we took all of our measurements during one3hperiodinthemorning,itispossiblethat the temperature diffe...