Formica ant and Cinara aphid mutualisms on limber pines (Pinus flexilis)

Charles Chen1, Lori Liu2, Skai Peterson3

1University of California, San Diego; 2University of California, Riverside; 3University of California, Santa Barbara

ABSTRACT

Ant-aphid mutualisms play important ecological roles in many ecosystems. While Formica ants are well studied model organisms, pine-specialist Cinara aphids remain understudied. Moreover, there is little research on ant-aphid mutualism in subalpine ecosystems. We investigated multiple species of Formica ants and their mutualistic interactions with Cinara aphids on the limber pine Pinus flexilis in a subalpine ecosystem. Cinara apini colonies were found tended to by silver, black, and red Formica ants, and they occupy older branches and trunks of the pines. A different, unknown Cinara species forms species-specific mutualisms with the brown Formica ant, occupy new growth twigs, and are more aggressively defended by brown ants. All ants forming mutualist colonies with aphids exhibited little modification in behavior and defense in the presence of sugar bait. The results of our study give insight into several Formica and Cinara life histories as they relate to ant-aphid mutualisms in a subalpine ecosystem.

Keywords: ant, aphid, mutualism, subalpine

INTRODUCTION 1993). After feeding, they secrete a sugary waste called honeydew. Ants will guard Mutualisms are a common form of species aphid patches to consume honeydew, and in interaction, in which both species gain a net the process, remove waste from aphid fitness benefit. Species involved in a aggregations and protect aphids from mutualism may gain resources that would predation (Mooney 2005). Honeydew can be otherwise be difficult to acquire, or receive a vital source of sugar for ants, especially protection from predators and parasites. when fruits and nectar are not abundant One such mutualism is the relationship (Mooney 2005). Aphids experience very high between ants (family Formicidae) and aphids predation pressure, and thus may benefit (superfamily Aphidoidea), in which ants greatly from ant protection, especially when defend aphids from predation in exchange their population numbers are low (Breton for honeydew. Aphids are gregarious 2005). In this way, ant-aphid mutualisms phloem-feeders which reproduce in vast provide considerable advantages to both numbers, thus acting as predators of many taxa. plant species, and prey of many animals (Ives

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 1/10

Mutualistic relations with aphids can allow Aphid colony defense by ants has been ants to thrive in food-limited environments, extensively studied - studies indicate that thereby expanding the range and impact of when ants are tending to a larger colony of these ants. Ant-aphid mutualism can also aphids, they play a less significant role in have varying consequences for host plant aphid defense (Breton 1995). Moreover, species. Host plants may benefit via ants may decrease tending behavior or cease presence of ant guards—in the presence of tending aphids all together when exposed to ant-aphid mutualists the majority of plants a more preferable source of sugar (Sakata benefit by experiencing reduced damage, 1995). However, this varies by sugar content increased growth, and increased and concentration, and previous studies reproduction (Styrsky 2007). Ant-aphid have indicated diminished interest of ants mutualism can also harm plants, as for sucrose bait during the summer months Kilpelanin et al found that Formica rufa and (Sudd 1985). Cinara aphids mutualism decreased 30-year- Ant-aphid mutualisms are geographically old Norway spruce (pine family) growth by widespread (Way 1963) and well-researched 7.3% but had no effect on other age amongst crops and temperate and tropical standings. Alternatively, ant-aphid ecosystems (Styrsky 2007), but relatively mutualisms on host plants can have understudied within subalpine ecosystems relatively negligible effects; pine growth will and amongst pine-feeding Cinara aphids. sometimes go unaffected by ant-aphid There has been no previous research interactions (Kilpeläinen 2009). These documentation of ant associations and previous studies exemplify the interspecific microhabitat preferences of any known variation that exists within ant-aphid aphid species found on the subalpine- mutualisms. dwelling limber pine Pinus flexilis. The three Among aphid species, there is documented Cinara aphid species on P. considerable variation in interspecific flexilis have been understudied in their associations and aphid microhabitat natural history and in their interactions with preferences. For example, pine-specialist ants. However, such mutualisms are Cinara are known to form associations with important to research because ants have Formica wood ants as well as Lasius ants large ecological impacts on subalpine (Akyürek 2016). Cinara exploit various habitats. For example, Formica rufa colony microhabitats among their pine hosts: C. mounds were linked to forest tree species brauni feeds on current year and one year- composition, slope aspect, and canopy old growth of Pinus nigra (Massuti 2005), C. closure (Risch 2005). piceae on older branches and trunks of In order to expand our understanding of spruce (Picea) trees (Blackman 2000), and C. Formica ants and Cinara aphids in a piceicola on spruce (Picea) needle bases and subalpine setting, we studied these ant- shoots (Binazzi 2009). Despite this research, aphid mutualists on limber pines (Pinus the interspecies associations and flexilus) in the White Mountains of microhabitat preferences of many species California. In our observational study, we remains unknown. noted species and microhabitat trends of ant-aphid associations. We hypothesized

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 2/10

that ant-aphid associations would not be patch, we recorded ant and aphid species species-specific, and that there will be identity, and ant and aphid numbers. We interspecific differences in aphid classified the microhabitat of the patch as microhabitat preferences. We hypothesized either the main trunk and branches, or the that aphid to ant ratios will be negatively needle bases. Tagged patches were revisited correlated with defensive behavior of ants. two more times at approximately 24 hour We also hypothesized that the presence of intervals to see if patches remained in their honey would have little effect on aphid to original location. Patches that were absent ant ratio and defensive behavior. We were from their original site and patches that motivated to expand the knowledge on ant- reappeared on following days were classified aphid mutualism, as well as the natural as mobile. To test defensive behavior, we history of Cinara aphids on P. flexilis. prodded aphids within a patch with a 0.5mm wide pencil lead and counted the number of METHODS prods required to elicit a defensive response, Aphids were prodded up to 10 times at one 2.1 Study Site second intervals. A defensive response was classified as at least one ant attempting to Our study was conducted near Crooked bite the pencil. Defensive behavior was Creek Field Station, at the University of classified into one of three categories: California White Mountain Research Center immediate response (1–2 prods), in the White Mountains at the Inyo National intermediate response (3–9 prods), and no Forest. We studied interactions amongst response (no aggression after 10 prods). four Formica ant species and two Cinara aphid species on Pinus flexilis trees from 2.3 Manipulative Study August 2–5, 2018. One aphid species, the “grey aphid,” was identified as Cinara apini. For the manipulative test, we selected 10 The other species could not be identified, sites for each of the four ant species, and will be referred to as the “red aphid” recording aphid and ant species composition based on coloration and morphology. The and aphid and ant numbers prior to Formica species could not be identified administering honey. We placed 5 mL of beyond genus, and will be referred to as honey 1 cm away from each site, and the “silver ant,” “brown ant,” “black ant,” and ants at the site were observed for 5 minutes. “red ant.” Samples were collected of each After 5 minutes, we counted the number of Formica and Cinara morphospecies. ants eating honey and the number of ants tending aphids. We also tested defensive 2.2 Observational Study behavior of the aphid-tending ants using the pencil prodding method. We examined P. flexilis trees within the Crooked Creek station for ant-aphid patches. 2.4 Data Analysis Patches initially featured at least one aphid being visibly attended by ants. Ant-aphid All data analysis was conducted using JMP mutualistic patches and their host trees 14 statistical software. We analyzed the were marked and numbered. For each

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 3/10

differences in ant and aphid taxonomy marginally higher cluster sizes than red associations with a chi-square test. We aphids (two-tailed t-test: N = 149, T = 1.81, P examined whether ant or aphid abundance = 0.07). With greater number of aphids per varied among species with ANOVAs. The cluster, number of ants increased (ANOVA: 2 effect of number of aphids on ants and ant R = 0.49, N = 153, F(1,143) = 146.17, P < 0.01, to aphid ratio was tested with an ANOVA. Figure 2). However, ant to aphid ratio The association between microhabitat type decreased as aphid cluster size got bigger 2 and aphid species was analyzed with a chi- (ANOVA: R = 0.10, N = 147, F(1,158) = 17.54, P square test. We used a chi-square test to < 0.01, Figure.3). Microhabitat types affect investigate whether ant species differed in the colonized aphid species, with red aphids their defensive behavior. The effect of ant typically on needle bases and C. apini on number, and aphid to ant ratio on defensive branches (chi-square test: N = 159, Χ2 = behavior was analyzed with logistic 109.93, P < 0.01, Figure 4). regressions. Differences in mobility of aphid clusters across time and species were analyzed with chi-square tests. To examine interspecific differences in response to honey, we compared the number of ants eating honey between species with an ANOVA. The effects of aphid to ant ratio on number of ants eating honey were also analyzed with an ANOVA. In order to examine the effect of honey treatment on Figure 1. Ant species associated with different aphid defensive behavior, we derived one sample species. Red bar regions represent colonies of red from observational study and one from our aphids and grey bar regions represent colonies of C. manipulative test, from both of which we apini (grey aphids). Brown ants largely associated excluded responses after 10 prods, and with red aphids and other ants associated with C. apini. compared the difference in defensive behavior with a two-way ANOVA.

RESULTS

3.1 Ant-Aphid Association

Silver ants, red ants, and black ants largely associated with species of C. apini, whereas brown ants typically associated with red aphids (chi-square test: N = 145, Χ2 = 97.24, P < 0.01, Figure 1). There was no dramatic difference in number of ants on a cluster among different ant species (ANOVA: N = Figure 2. Number of ants increased with aphid 161, F(3,157) = 1.90, P = 0.13). C. apini had colony size. Black dots represent each cluster of aphids with ants. Blue line is line of best fit.

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 4/10

However, aphid to ant ratio did not have an effect on defensive behavior (logistic 2 regression: N = 145, Χ (2) = 1.33, P = 0.51).

Figure 3. Ant aphid ratio decreases as aphid number increases. Black dots represent clusters of ants and aphids. Blue line is line of best fit. Figure 5. Ant species respond differently in defending aphid patches. Based on the number of prods required to trigger a bite, defensive response was categorized as immediate (1-2), intermediate (3- 9), or no response (>9). The quicker an ant species reacted to our attack, the more aggressive they were. Brown ants were the most aggressive in defending their aphids and black ants defended least actively.

Figure 4. Microhabitat affected the distribution of different aphid species. Red bar regions represent colonies of red aphids and grey bar regions represent colonies of C. apini (grey aphids). Red aphids tended to feed on limber pine tips whereas grey aphids tended to feed on branches.

3.2 Ant Defensive Behavior

Brown ants exhibited the most intense Figure 6. Defensive response became more rapid as defensive behavior, but black ants least number of attending ants per ant-aphid colony actively responded to our prods (chi-square grew. Defensive response is more quickly enacted by tending ant colonies of larger sizes. Error bars test: N = 145, F(6) = 24.89, P < 0.01, Figure 5). represent ±1 standard error. The more ants on one cluster the more aggressive they were (logistic regression: N 2 = 144, Χ (2) = 18.41, P < 0.01, Figure 6).

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 5/10

3.3 Aphid Mobility

From day 1 to day 2, the mobility of aphid clusters did not differ between two aphid species (chi-square test: N = 54, Χ2(3) = 0.74, P = 0.39), and it also did not differ among associated ant species (chi-square test: N = 54, X2(3) = 2.40, P = 0.49). From day 2 to day 3, however, the mobility of red aphid clusters was much lower than that of grey aphid clusters (chi-square test: N = 53, Χ2(3) = 6.35, P = 0.01, Figure 7). Clusters guarded by red and brown ants also moved more frequently (chi-square test: N = 51, Χ2(3) = Figure 8. Mobility of ant-aphid clusters differed by Formica ant species. Brown and red Formica ants 12.57, P < 0.01, Figure 8). largely remain with aphid colonies over the observed day 2 to day 3 period, whereas some mobility was observed in the aphid colonies guarded by black and silver Formica ants.

3.4 Behavior in Presence of Food

More silver ants ate honey than brown ants (ANOVA: N = 40, F(3,36) = 3.42, P = 0.03, Figure 9). Aphid to ant ratio did not affect the number of ants eating honey (ANOVA: N = 40, F(1,38) = 0.05, P = 0.83). After the honey treatment, interspecific differences in defensive behavior did not change (chi- square test: N = 26, Χ2(6) = 12.85, P = 0.05, Figure 10). However, number of ants that Figure 7. Aphid colonies in C. apini (grey) and red aphids persisted to differing extents. Red aphid remained on the clusters had no effect on colonies all persisted unmoved throughout the day 2 defensive behavior (logistic regression: N = to day 3 period, whereas more C. apini colonies 26, Χ2(6) = 0.91, P = 0.63), so did aphid-ant moved. ratio (logistic regression: N = 26, Χ2(6) = 12.85, P = 0.05).

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 6/10

DISCUSSION

4.1 Ant-Aphid Association

While silver ants, red ants, and black ants all associate with C. apini, brown ants predominantly associate with red aphids. Furthermore, the brown ant-red aphid patches were predominantly found on only one tree. Red aphids and C. apini differ in both microhabitat and associated ant species, which lends doubt to the notion that these two species are competing for space or ant care. Alternatively, the condition of the Figure 9. After honey treatment, number of ants tending honey varied by ant species. Silver ants took host plant may be discouraging C. apini and honey bait more often than other ant species. Error their associated ant species from settling. bars represent ±1 standard error. The tree in which we observed most of the brown ant-red aphid patches displayed many browning needles, and some species of aphids have been shown to flee and avoid host plants which are in poor health (Hough 1922). Conversely, the poor health of the tree may actively attract red aphids, as some species of aphids are known to prefer diseased host plants (Macias 1969). If sick trees are a limited resource for red aphids, this would explain why red aphids were Figure 10. Ant species differed in defensive behavior uncommon on other trees, and after honey treatment. Depending on the number of concentrated on one sick tree (Pyke 1977). prods required to trigger a bite, a defensive response While brown ants are most frequently was categorized as immediate (1-2), intermediate (3- observed with red aphids, they have 9), or no response (>9). Response speed measured ant aggression levels. Brown ants were the most occasionally been observed with C. apini. aggressive in defending their aphids and black ants Moreover, while silver ants, red ants, and defended least actively. black ants predominantly tend C. apini, they have also been observed with red aphids. There was no effect of honey treatment on This suggests that the ant-aphid associations defensive behavior (two-way ANOVA: N = we studied display trends in species and 81, F = 0.01, P = 0.983), ant species did not microhabitat, but can still occur outside of differ in their defensive behavior (two-way their preferred circumstances. ANOVA: N = 81, F = 2.39, P = 0.08) and ant species did not have an effect on treatment (two-way ANOVA: N = 81, F = 0.66, P = 0.58).

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 7/10

4.2 Defensive Behavior resource (food, mates, oviposition target) or more prone to deserting as an escape Among the ant species, brown ants were response (Hough 1922). Likewise, as most aggressive in defending their aphid individual colony movement was not clusters, while black ants were least tracked, it is possible that the measure of aggressive. Brown ants may simply have the mobility used may have been confounded by most aggressive archetype, which may also predation or involuntary weather-induced manifest in competitive exclusion of other displacement. Rain and hail were ant species (Czechowski 2013), as previously experienced between the first and second speculated. Ant to aphid ratio had no effect observational periods, and aphids have been on defensive behavior, which contradicted shown to disperse from host plants in the our initial hypothesis. A 1995 study by face of harsh weather (Mann 1995). Breton et al. had shown that ants play a less However, aphid patch mobility was higher proactive role in aphid tending when the ant between the second and third observational to aphid ratio is lowered, but this did not periods, when the weather was mostly clear. appear to diminish defensive behavior in the ant-aphid mutualisms we studied. 4.4 Behavior in Presence of Food

4.3 Mobility of Ant-Aphid Patches Most ant species readily took the honey bait, however the proportion that did varied Brown ant-red aphid patches were amongst different ant species. Brown ants observed being less mobile than other ant- were least inclined to accept bait, and largely aphid associations, perhaps because of continued tending to their aphid colonies. brown ant aggression. Red aphids under This behavior along with elevated defensive careful protection by brown ants may tendencies in brown ants suggests different experience reduced predation pressure, and life history strategies in regards to aphid this may reduce the need for patch attendance in brown ants relative to other movement. Alternatively, aphids may be aphid-tending Formica species. Brown ants under “stricter” control by brown ants via may be more reliant on their aphid aphid wing amputation, chemical secretions mutualists and consequently more inhibiting wing development (Kleinjan 1975) protective overall. or chemical “fencing” (Oliver 2007) . Similarly, Sudd et al. noted alternative Although C. apini is a “sedentary” aphid food sources such as sucrose-water mixtures type (both C. apini and red aphids were have limited acceptance by tending ant largely wingless nymphs and apterae), populations, especially from mid-spring mobility rates among C. apini patches were throughout the summer months (Sudd high. This may be due to high mobility in red, 1985). Previous studies also indicate that the black, and silver ant species (Dixon 1958), as trisaccharide composition of honeydew ants are known to herd their aphid towards produced by aphids is more desirable than better phloem sap extraction sites (Collins monosaccharides or disaccharides (Völkl 2002). Alternatively, grey aphids may be 1999)—primary components of the sucrose more mobile in search of a particular and honey baits used (White 1980). The lack

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 8/10

of change in ant defensive response in 10.21973/N3KM2J. We are also deeply presence of honey bait also supports a lack indebted to Robert Dransfield (author of of interest, indicating that honey is a lower- Influential Points) and Andy Jensen (author value food source than honeydew. of Aphid Trek), who assisted us with aphid identification and biology. Lastly, we are 4.5 Significance and Future Studies grateful to instructors Tim Miller and Krikor Andonian for their support and assistance Our study examined ant-aphid mutualisms with this project. among Formica ants and the relatively understudied Cinara aphids. Ant-aphid REFERENCES mutualisms are ubiquitous around the world, yet they remain understudied outside Akyürek, B., Zeybekoglu, Ü., Görör, G., & Karavin, M. of economically important plants and 2016. Reported aphid (Hemiptera: Aphidoidea) and ant (Hymenoptera: Formicidae) species temperate and tropical ecosystems. The associations from Samsun Province. Journal of the results of our study provide insight into the Entomological Research Society 18(3):97. dynamics of ant-aphid mutualisms in subalpine ecosystems where ants occupy Binazzi, A., & Scheurer, S. 2009. Atlas of the important ecological roles across several honeydew producing conifer aphids of Europe. Rome: Aracne. trophic levels. Ants act as predators, prey, and decomposers, which is why it is Blackman, R. L., & Eastop, V. F. 2000. Aphids on the important to expand our understanding of world's crops: an identification and information ant ecology and its relationship to other guide (No. Ed. 2). John Wiley & Sons Ltd. organisms. Brady, S. G., Larkin, L., & Danforth, B. N. 2009. Bees, There are several possible directions for ants, and stinging wasps (Aculeata). The timetree future research that could expand on the of life, pp. 264–269. findings of our study. Examining the chemical composition of aphid honeydew Breton, L. M., & Addicott, J. F. 1992. Density- could provide insight into ant preferences dependent mutualism in an aphid-ant interaction. Ecology 73(6):2175–2180. regarding aphids and other sugar sources. We could also study the association between Collins, C. M., & Leather, S. R. 2002. Ant-mediated red aphids and tree health to see if the red dispersal of the black willow aphid Pterocomma aphids have a preference for sick trees. salicis L.; does the ant Lasius niger L. judge aphid- Furthermore, a long-term monitoring study host quality? Ecological entomology 27(2):238–241. could help us understand how ant-aphid Czechowski, W., Marko, B., Radchenko, A., & ŚlipińskI, mutualisms respond to changes in weather P. (2013). Long-term partitioning of space between and seasonal conditions. two territorial species of ants (Hymenoptera: Formicidae) and their effect on subordinate ACKNOWLEDGMENTS species. European Journal of Entomology 110(2).

Research was conducted at the University Collins, C. M., & Leather, S. R. (2002). Ant-mediated dispersal of the black willow aphid Domisch, T., of California White Mountain Research Finér, L., & Jurgensen, M. F. 2005. Red wood ant Center, Crooked Creek Field Station, doi: mound densities in managed boreal forests. In

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 9/10

Annales Zoologici Fennici (pp. 277–282). Finnish Mooney, K. A., & Tillberg, C. V. 2005. Temporal and Zoological and Botanical Publishing Board. spatial variation to ant omnivory in pine forests. Ecology 86(5):1225–1235. Dixon, A. F. G. 1958. The escape responses shown by certain aphids to the presence of the coccinellid Oliver, T. H., Mashanova, A., Leather, S. R., Cook, J. Adalia decempunctata (L.). Transactions of the M., & Jansen, V. A. 2007. Ant semiochemicals limit Royal Entomological Society of London 110(11): apterous aphid dispersal. Proceedings of the Royal 319–334. Society of London B: Biological Sciences 274(1629):3127–3131. Binazzi, A., & Scheurer, S. 2009. Atlas of the honeydew producing conifer aphids of Europe. Rome: Aracne. Pyke, G. H., Pulliam, H. R., & Charnov, E. L. 1977. Optimal foraging: a selective review of theory and Folgarait, P. J. 1998. Ant biodiversity and its tests. The quarterly review of biology 52(2):137– relationship to ecosystem functioning: a review. 154. Biodiversity & Conservation 7(9):1221–1244. Risch, A. C., Jurgensen, M. F., Schütz, M., & Page- Hough, W. S. 1922. Observations on two mealy-bugs, Dumroese, D. S. 2005. The contribution of red Trionymus trifolii Forbes and Pseudococcus wood ants to soil C and N pools and CO2 emissions maritimus Ehrh. Entomol. News 33:171–176. in subalpine forests. Ecology 86(2):419–430.

Ives, A. R., Kareiva, P., & Perry, R. 1993. Response of Sakata, H. 1995. Density-dependent predation of the a predator to variation in prey density at three ant Lasius niger (Hymenoptera: Formicidae) on two hierarchical scales lady beetles feeding on aphids. attended aphids Lachnus tropicalis and Myzocallis Ecology 74(7):1929–1938. kuricola (Homoptera: Aphididae). Researches on Population Ecology 37(2):159–164. Kleinjan, J. E., & Mittler, T. E. 1975. A chemical influence of ants on wing development in aphids. Styrsky, J. D., & Eubanks, M. D. 2007. Ecological Entomologia experimentalis et applicata, consequences of interactions between ants and 18(3):384–388. honeydew-producing insects. Proceedings of the Royal Society of London B: Biological Sciences Kilpeläinen, J., et al. 2009. Does the mutualism 274(1607):151–164. between wood ants (Formica rufa group) and Cinara aphids affect Norway spruce growth?. Sudd, J. H., & Sudd, M. E. 1985. Seasonal changes in Forest ecology and management 257(1):238–243. the response of wood-ants (Formica lugubris) to sucrose baits. Ecological Entomology 10(1):89–97. Macias, W., & Mink, G. I. 1969. Preference of green peach aphids for virus-infected sugar beet leaves. Völkl, W., Woodring, J., Fischer, M., Lorenz, M. W., & Journal of Economic Entomology 62(1):28–29. Hoffmann, K. H. 1999. Ant-aphid mutualisms: the impact of honeydew production and honeydew Mann, J. A., Tatchell, G. M., Dupuch, M. J., Harrington, sugar composition on ant preferences. Oecologia R., Clark, S. J., & McCartney, H. A. 1995. Movement 118(4):483–491. of apterous Sitobion avenae (Homoptera: Aphididae) in response to leaf disturbances caused Way, M. J. 1963. Mutualism between ants and by wind and rain. Annals of Applied Biology honeydew-producing Homoptera. Annual Review 126(3):417–427. of Entomology 8(1):307–344.

Masutti, L., Battisti, A., & Faccoli, M. 2005. Insect White, J. W., & Doner, L. W. 1980. Honey composition fauna of the Pinus nigra group in Italy. and properties. Beekeeping in the United States Entomological research in Mediterranean forest Agriculture Handbook 335:82–9. ecosystems. Paris: INRA editions, pp. 79–87.

CEC Research | https://doi.org/10.21973/N3VQ1P Summer 2018 10/10