Answers are grounded in this paper and cite their sources.
Many species of aphid have intimate associations with species of ant. These relationships, while variable in the level of their inter-dependence and conditionality, are often considered mutualistic, and the benefits to both parties are widely appreciated (Jones, 1929; Way, 1963; Breton & Addicott, 1992a; Dixon, 1998; Stadler & Dixon, 1999; Yao et al., 2000). The benefits to ants stem largely from the ready supply of carbohydrate available in aphid honeydew and, in certain cases, from the protein of the aphids themselves (Sakata, 1995), while aphids may benefit in terms of protection from natural enemies, improved colony hygiene, increased feeding rate, transport and dispersal (Way, 1963; Buckley, 1987; and references therein). Given the importance of the ant partner to the success of tended aphid colonies, the presence, or potential for recruitment, of ants is likely to influence host selection by these aphids, as in myrmycophilous lycaenid butterflies (Pierce & Elgar, 1985) and in the membracid homopteran Publilia concava (M. Morales, pers. comm.). A more direct contribution of the ant partner, made by dispersing the tended aphids themselves, has been observed in subterranean aphid species, where ant-aided dispersal may be obligate (Way, 1963), and in species in which eggs are sheltered in ant nests over winter (Pontin, 1960); however the suggestion that the black garden ant Lasius niger L. and other ant species may transport aphids above ground between host plants (Nixon, 1951) has not found support in later studies (El-Ziady & Kennedy, 1956; Banks, 1958; Pontin, 1960). Although Lasius fuliginosus L. transports young fundatrices of Stomaphis quercus L. from their overwintering egg site on the base of the tree to suitable feeding sites in the upper canopy of oak trees (Goidanich, 1959), this subject has received little attention in recent years. The investigation reported here resulted from the chance observation of a L. niger, carefully carrying a third-instar black willow aphid Pterocomma salicis L. in its jaws away from an established aphid colony. The ant deposited the aphid on another host plant, previously uncolonised by the aphid, and remained with it for some time until settling behaviour was observed. Pterocomma salicis does not have an obligate dependency on attendant ants and the first experiment was designed to assess rapidly whether attendance by L. niger influences P. salicis in terms of population growth and adult mass. The second experiment was designed to investigate whether ant attendance leads to increased dispersal of aphids from an established colony and whether there is an interaction between the presence of ants and the choice of new host tree. The aphid clone used in these experiments was collected originally from Roves Farm, Sevenhampton, U.K. (51°37′Ν, 1°38′W) and has been maintained in parthenogenetic culture for >2 years. Aphids were kept on potted, growing willow trees Salix viminalis‘Bowles Hybrid’ at 18 ± 1 °C, LD 16:8 h photoperiod. Plant material was also sourced from Roves Farm. The stem-wood cuttings used were taken from a single coppice stool to minimise genetic variability. Twenty-four similar-sized cuttings of the hybrid willow Salix viminalis‘Jorr’ were planted in 4 litres of multi- purpose compost (J. Arthur Bowers, William Sinclair Horticulture Ltd, Lincoln, U.K.) and left to establish in an open greenhouse for 20 days. Each young tree was allocated randomly to one of two treatments, accessible or inaccessible to ants, and inoculated with a single pre-reproductive adult P. salicis. Ants were present and active in substantial, but uncounted, numbers in the greenhouse and were excluded by application of a 3-cm wide strip of Oeco Tak A5® (Kimpton, U.K.) to the outside of watering saucers. After 28 days of aphid infestation, the number of P. salicis present on each tree was counted, the dry weight of five adult aphids from each tree was determined, and the mean adult dry weight for each tree was calculated. Drying time was 72 h at 40 °C. Data preparation was in Microsoft Excel (Microsoft, Seattle, Washington). t-tests of dry weight and square-root-transformed count data were performed in S-Plus (MathSoft Inc., Seattle, Washington). Twenty source colonies of P. salicis were established for 2 weeks on 3-month-old willow trees S. viminalis‘Jorr’ growing in 4 litres of multi-purpose compost (as above) in an open greenhouse. At the start of the experiment, source colonies consisted of 50–60 aphids, of which the majority was in the second to fourth instar and thus pre-reproductive. This population distribution and the short duration of the experiment avoided confounding effects of host quality on the performance of the aphids (Zhou & Carter, 1991; Duffield et al., 1997) with its effects on dispersal. Target trees were generated by planting 40 similar-sized cuttings of the hybrid willow S. viminalis‘Jorr’ in 1-litre pots containing a 1:1:1 mixture of sand : grit : peat. In order to produce trees of different host quality to aphids, half were selected randomly and watered twice weekly with 50 ml of a 0.125% solution of ‘Miracle Gro’ (N:P:K 15:30:15, I.C.I. Garden Products, Haselmere, U.K.); the remainder received 50 ml of tap water. After 50 days of establishment and growth, these saplings were size-standardised by pruning all trees. Poor quality trees were slightly paler in colour but had similarly sized leaves to the good quality trees. Each replicate consisted of a 30 × 45 cm tray containing three trees: a source tree, a high quality target tree, and a low quality target tree. The rims of the pots were in contact with one another but there was no contact between the foliage. Ten of the 20 trays had a 3-cm wide strip of Oeco Tak A5® applied to all sides to exclude ants. The experiment ran for 7 days and the number of aphids established on each target tree was counted daily. Time was then collapsed into a factor with two levels: early (days 1–4) and late (days 5–7). Data preparation was in Microsoft Excel. Analysis using S-Plus was by 2 × 2 × 2 contingency table testing the null hypothesis that ant presence, host quality, and time are independent in their effects on the numbers of aphids present on newly colonised trees. The presence of ants had no effect on the dry weight of adult P. salicis (t = 1.34, d.f. = 22, P = NS), however there were significantly more aphids in ant-tended colonies after 28 days (t = 2.19, d.f. = 22, P < 0.05) (Fig. 1). Attendance by ants clearly benefited the aphids. Although the adult aphids were no larger when ants were present, the number of aphids present in a colony 28 days after founding was almost doubled. Whether this is a result of increased protection from predators, corralling, or more rapid growth and/or reproduction is not known; it can only be concluded that 4 weeks of attendance by L. niger has a net positive effect on population numbers. (a) The mean (± SE) mass of adult apterous Pterocomma salicis in single fundatrix, 28-day-old colonies as a function of attendance by the ant Lasius niger. There was no significant difference between treatments (t = 1.34, d.f. = 22, P = NS). (b) The mean (± SE) number of the aphid Pterocomma salicis in single fundatrix, 28-day-old colonies as a function of attendance by the ant Lasius niger. The difference between treatments was significant (t = 2.19, d.f. = 22, P < 0.05). The interaction between ant presence and host quality was dependent on time (χ2 = 7.9, d.f. = 1, P < 0.01), i.e. more aphids dispersed in the treatment with ants than in the treatment where ants were excluded, and when ants were present many more aphids established on the high quality host trees (Fig. 2), but this pattern was not recognisable in the early phase (days 1–4) of the experiment. As this experiment only covered 7 days and the source aphid populations were small compared with the carrying capacity of the saplings (C. M. Collins, pers. obs.), it is unlikely that the observed dispersal occurs because of an aphid-induced decline in the host quality of source trees in turn promoting an aphid exodus. On several further occasions during the experiment, ants were observed bearing third- and fourth-instar aphids between source and target trees. The direct ant-borne dispersal is, however, unlikely to be the only dispersal route for tended aphids as adult aphids in attended colonies can become noticeably more restless and may also disperse locally by wandering (El-Ziady, 1960). Local wandering dispersal is likely to be a costly strategy as it exposes aphids to ground-dwelling predators (Sunderland et al., 1986) and longer distance aerial dispersal has a very low success rate (Ward et al., 1998). Dispersal vectored by ants could provide a safer and targeted means of dispersal, which may well lead to an increase in aphid populations within the foraging zone of ant nests. The cumulative mean (± SE) number of Pterocomma salicis per new host tree as a function of tree quality and ant attendance by Lasius niger. The interaction term between ant attendance and new host quality was significant (χ2 = 7.9, d.f. = 1, P < 0.01) in the late stage of the experiment (days 5–7). The fact that the ant : host quality interaction was not apparent until the latter part of the experiment (days 5–7) is not surprising and may well be a result of low replication or the relatively low numbers of dispersing aphids. It may also result from the ants not being able to identify immediately the quality of new aphid hosts per se, and using an aphid-mediated cue such as the composition of the aphid honeydew, which is known to vary both with host quality and ant attendance (Llewellyn et al., 1974; Fischer & Shingleton, 2001). Other aphid behavioural cues, such as their readiness to settle and feed, that relate to the aphids' perception of host quality (Dixon, 1998) may also be used by the ants to assess the quality of host trees. While this work examined only the influence of ants on aphid dispersal rather than classical tritrophic interactions, the results are in contrast with those of Breton and Addicott (1992b), who did not demonstrate a significant effect of host plant quality on aphid populations in the presence of ants. In their experiment, more ants tended the aphid populations of high quality plants but this did not result in more aphids. Here it was shown that host plant quality can influence the spatial distribution of an aphid–ant mutualism. It is increasingly apparent that interactions between farming ants and stock aphids are more subtle and complex than thought previously (Cushman & Addicott, 1989; Bristow, 1991; Shingleton & Foster, 2000; Yao et al., 2000; Fischer & Shingleton, 2001). The fact that ants appear to be able not only to recognise differences in the host quality of tended aphids but also to invest in inter-host plant transport opens a new area of research in the field of ant–aphid mutualism and of positive species interactions. We wish to thank Rui G. Rosado for his contribution to the maintenance of these experiments, Rufus Sage and The Game Conservancy Trust for their support of this work, Mike Way, John Addicott, Bernhard Stadler and Jim Hardie for their comments on the manuscript.
No takes yet. Share an insight, caveat, or question.
Collins et al. (2002) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: