Dispersal agents observation
Fresh and mature seeds were collected the in the field, and several of them are randomly put in study site as baits to attract the seed dispersal agents. Digital video recorder was set to record the whole dispersal process. Totally more than 500 seeds are used in 2009 & 2010. To our preliminary observation, ants are the dispersal agent of A. macranthum. Agent species, moving behavior and foraging model (grouped or solitary) were recorded to reveal the seed dispersal mechanism of A. macranthum.
The dispersal agent is defined which could transport seeds more than 5 cm. Animals directly gnawing or eating the seeds without transporting will not be recognized as dispersal agents. All agents are preserved in 70% ethanol in department of Life Science, National Taiwan Normal University, Taipei, Taiwan for further identification.
Measure dispersal distance
In order to realize the seed fate after dispersal and the distance that agents carry. I put mature seeds in 100 small depots. Each depot contained three seeds, and was settled randomly in the study site without overlapping continually. From the previously observation, I have already confirmed the major seed dispersal agent are ants. Therefore, when seed were removed, researchers followed the foraging route back to their nest, and measured the linear dispersal distance from source (seed depot) to the sink (ant nest). Any seeds dropped in midway will not be considered. All the data were imputed to the statistic software- JMP to map the frequency distribution of the dispersal distance and to analysis the influence that different ant species contributed to seed dispersal.
Function of elaiosome
Surveying the previously image that captured by the digital video recorder, ants
carry seeds by gnawing at the structure “elaiosome”. In order to figure out the function and importance of elaiosome in whole dispersal process, 504 seeds are divided into two parts. Half of those seeds are manipulated with their elaiosomes removed by forceps and washed for twice with sterile water. 14 seeds are used at a time; 14 seeds are put in one depot (7 with elaiosome attached, and 7 peeled off) and observed the ant‟s preferences on both control seeds and manipulated seeds. Each depot is recorded for two hours with digital video recorder, and finally 36 depots are presented in the field without overlapping. The number and ratio that seed are transported and agent‟s behavior were recorded, and use chi-square test and Tukey-Kramer test to check for difference between these two manipulations. In short, use this “loss of function” test to reveal the importance and function of the structure
“elaiosome”.
Spatial analyses
Video has shown that ants transported the seed of A. macranthum directly back to their nest. In order to realize the spatial relationship between ants and A.
macranthum, a 3 x 13 m2 area in the Mt. Ergo was set for spatial analyses. In order to understand the relationship between ant nests and Asarum, I recorded the localities of every A. macranthum (mature individual and juvenile are recorded separately) and every ant nest. Because ant nests just form a small entrance to the surface, it is hard to find it directly. Therefore, baits are used to trace ants back and help to document their nest. For spatial analysis, the locality data are presented with binary code in every 39 quadrants (if target exist in this quadrant, I mark “1” as a symbol; if target are not exist in this quadrant, I mark “0” as a symbol). Fisher‟s exact test was conducted to reveal the spatial relationship between the ant nests and A. macranthum plants.
Results
Dispersal agents observation
Nine ant species were presented as seed dispersal agents during two-year videotaping (2009-2010). Nine ant species were different in their body size (1.5mm~9mm), foraging behavior (group or solitary) and foraging efficiency. Smaller ants (fig 22-A~C) were with small mandible, usually grouped around seeds and transported seeds slowly. Larger ants (fig 22-H-I) can gnaw elaiosomes with their huge mandible, and transport seeds solitarily and efficiency. All of them can transport seeds more than 5 cm, even directly move to their nest. All the transporting behaviors were recorded in digital video recorder (figure 23).
Figure 22- The dispersal agent composition of Asarum macranthum
Nine species are proved to be the dispersal agents of Asarum macranthum. They are arranged by body size.
Figure 23- Behavior of myrmecochory Each ant species is presenting by three graphs.
Nine ant species are marked with scale, and arrange by body size.
The field behavior was captured from digital video recorder.
Figure 23- behavior of myrmecochory Each ant species is presenting by three graphs.
Nine ant species are marked with scale, and arrange by body size.
The field behavior was captured from digital video recorder.
Measure dispersal distance
The linear distance was measured from source (seed depot) to sink (ant nest), 211 observations were recorded and mapped into dispersal curve (fig24). The average dispersal distance is 77.5 cm. Most of the data are less than 1 m, showing the pattern of short distance dispersal. Although still some fortuitous long distance dispersal occurred, which were caused a long tail reach to a value of 4.2 meters. There are nine species of ants transport Asarum seeds in the study site, and each species shows different dispersal pattern and foraging behavior. Among these nine ant species, three of them were chose to do further analyses due to their frequent transporting. Therefore, the dispersal distance data of these three main kinds of ants were analyzed separately.
In figure 25, Pheidole rinae tipuna is a small-size ant in this study, showing an averagely 33.45 cm dispersal distance (N=38, std err=8.55). Pheidole formosensis is a middle-size ant, with an averagely 71.56 cm dispersal distance (N=114, std err=4.94).
Aphaenogaster gracilenta is a large-size ant, with an averagely 130.9 cm dispersal distance (N=50, std err=7.45). By Tukey-Kramer test, the dispersal distances of these three ant species are significantly different (p<0.0001). Besides, ant body size affects their mandible size, causing different dispersal efficiency, and different ant species also affect their nest distribution pattern, causing the variation of foraging range and behavior.
Figure 24- Ant dispersal curve of distance
Mean=77.50 cm; Median=62.50 cm; N=211; Std Dev=61.69
Function of elaiosome
The function of elaiosome is evaluated by a “loss of fuction” experiment. The difference between ant‟s preference and removal rate can be shown from the results of videotaping. The remaining number of seed were calculated after 2-hours observation and checked for the difference between two manipulations. 36 sets of data are analyzed with chi-square test (fig 26). In the manipulated group, the average removal rate is 51.76%, and in the control group, the removal rate is 87.88%. A significant difference between the removal rates of manipulated and control seeds (p<0.0001).
The results reflect that elaiosomes are really an important structure to facilitate the transport efficiency.
Figure 25- Dispersal distance of three main ant species A- N=38, mean=33.45cm, std err=8.55
B- N=114, mean=71.56cm, std err=4.94 D- N=50, mean=130.9cm, std err=7.45
Spatial analysis
The distribution data of mature individuals, seedlings, and ant nests in 39 quadrants are analyzed by Fisher‟s exact test. A significant correlation is presented between locations of ant nests and seedlings (p=0.0069) (fig 27-A). It means that seedling have a strong probability to exist around ant nests than any other region.
However, the distribution of mature individuals of A. macranthum is not strongly correlated with the locations of ant nests (p=0.0684) (fig27-B). That may be a result that the mature individuals were established long time ago, so it couldn‟t fit the distribution of ant nest in the present day. The ant nests location at that time may influence by several factors, including weather, food resource and migration. It is properly the reason why there is no significant relationship between mature individuals and present ant nest location. Besides, no correlation between the locations of mature individuals and juveniles (p=0.1823) (fig27-C). Mature individual do not influence on the distribution of juveniles. It means that most of the seed are transported by agents, and only few of them are established around parents.
Figure 26- Removal ratio of “loss of function experiment”
Mean removal rate of “elaiosome exist”=87.54%, std Dev=23.99 Mean removal rate of “elaiosome remove”=53.06%, std Dev=35.38 t-test, p-value<0.0001
dark dash line – average removal rate
The graph A shows the correlation between the distributions of ant nests and seedlings. When ant nests are absent in a 1x1 m2 quadrants, only 35.3% of seedlings exist in the same quadrants.
However, when ant nest exist in a 1x1 m2 quadrants, the probability that seedling exist in the same quadrants achieve to 81.8%. Significant correlation (p=0.0069) between the distributions of ant nest and seedlings.
The graph B shows the correlation between the distributions of ant nests and mature individuals.
When ant nests are absent in a 1x1 m2
quadrants, only 16.7% of mature individuals exist in the same quadrants. However, when ant nest exist in a 1x1 m2 quadrants, the probability that mature individuals exist in the same quadrants is 63.6%. Weak correlation (p=0.0684) are
presented between the distributions of ant nest and mature individuals.
The graph C shows the correlation between the distributions of seedlings and mature individuals.
When seedlings are absent in a 1x1 m2
quadrants, 73.3% of mature individuals exist in the same quadrants. However, when seedlings exist in a 1x1 m2 quadrants, the probability that mature individuals exist in the same quadrants is 91.7%. There are no significant correlation between the seedlings and mature individuals.
A
B
C
Figure 27- Spatial pattern analysis
A- Correlation analysis between the distributions of seedling and ant nest.
B- Correlation analysis between the distributions of mature individual and ant nest.
C- Correlation analysis between the distributions of mature individual and seedling.
Discussion
different foraging model and behavior. Therefore, different ant species compositions around the world present various dispersal models. Slight differences are caused on average dispersal distance. 77.5cm in this study is less than the global mean due to different ant composition or other dispersal steps. Seeds are moved directly to the nest, and elaiosomes are utilized as food resources to feed the larvae. After that, seeds are leaved abandoned. Ants gather up these abandoned seeds in their depositories (Headley, 1949). The depositories are just below the soil surface and suitable for seeds germination. Effective dispersal and suitable germination environment make this myrmecochorous behavior contribute to the dispersal success.Mutualism relationship
Plants and ants have evolved a stable relationship. Plants obtain benefits from ants in several ways. First, seeds are directly taken to the ant nests, which are full of nutrients and trace elements for seedling growing (Culver and Beattie, 1978, 1980).
Second, seeds are moved to the nest underground. Seeds predators are hard to discover the seeds (Smith, 1989). And third, it successfully avoids the parent-offspring competition (Gomez and Espadaler, 1998). On the other hand, ants also benefit from plants in several ways. In this case, the diaspores of genus Asarum are recognized as viola-odorata type (Sernander, 1906). This type is characterized by its huge and a well-separated elaiosome. Therefore, when ants take seed to their nest, they can easily gnaw the elaiosomes off and feed to the larvae. Moreover, the
elaiosomes contain several essential fatty acids that few insect can synthesize (Hughes et al, 1994). Ants and Asarum benefit from each other in this myrmecochorous relationship.
In the “loss of function” experiment, significant difference (p<0.0001) between two manipulations demonstrate the importance of elaiosome. But there are still 53%
“elaiosome-removed” seeds are transported by ants as well. The rewardless seeds must remain something attractant. In the study of animal behavior, when animals die for more than two days, their corpses appear to emit a chemical signal. This signal can urge ants to carry the corpses away to the refuse pile (Wilson et al, 1958). This chemical substance “oleic acid”, a normal constituent of animal fat (Agosta, 2001), is proved to be the main trigger of necrophoric behavior of ants (Wilson et al, 1958;
Gordon, 1983). Same fatty acid also exists in the elaiosome of myrmecochorus plants (Pfeiffer et al, 2010). Not only the oleic acid, but the whole fatty acid composition is very similar to insect hemolymph (Hughes et al, 1994) Therefore, chemical mimicry may contribute to this stable relationship between ants and plants. Elaiosome continually emits the corpse-like odor, which likes a “dead insect analogue” (Carroll
& Janzen, 1973). The odor strongly attracts ants to move the seed (the scent may still remain on the seed even the elaiosome is removed). Except the corpse-throwing signal that emitted by oleic acid, other fatty acids exist in the elaiosome can also work in coordination. These chemical cues including the feeding signal that emitted by linoleic acid, ants are attracted to take seed back to feed their larvae. The food signal that emitted by 1, 2-diolein attracts ants to carry it to nest as a food (Marshall et al, 1979; Beattie, 1985). The ordinary dispersal behavior may implicate to the delicate cooperation of combined signals and the mutualism of myrmeochory behavior (Agosta, 2001).
Evolution implication
The distance generated by ants compared with other animals (birds and mammals) is relatively small (Gomez and Espadaler, 1998). In this study, the average seed
dispersal distance of A. macranthum is 77.5 cm. The short dispersal distance
generated by ants strongly influences the population distribution. The seedling spatial distribution is strongly correlated to ant nest distribution according to the correlation test. In other words, the seedlings are restricted in the specific area with ant nests nearby. Therefore, by ant dispersal, the whole population seems undergo a slow and restricted expansion event. The short-distance dispersal pattern may restrict gene flow among populations. Therefore, this patern enhances the probability of speciation. That may be the reason why most Asarum species present a series of small and
well-isolated populations (Yinger, 1983).
Beside of dispersal by ants, some studies mention the possibility about abiotic dispersal forces. Because some Asarum plants grow on the steep along the crest line, seeds may fall by way of gravity or rainfall. But in our observations, majority of seeds are transported to ant nest, only small portion are transported by gravity or rainfall.
Short-distance dispersal restrain the population expansion and may cause prolong inbreeding, and may cause some drawbacks on evolution and population dynamic.
But in the case of A. macranthum, several advantages may rise on the foundation of short-distance dispersal. First, species of genus Asarum are all distributed in a consistent environment, which is understory, and high-moisture (Yinger, 1983).
Therefore, requirement of strict limitation on habitat can be satisfied by short-distance dispersal features, which dispersed seeds to a similar environment than to the high risk place by long-distance dispersal. Second, the pollinators of A. macranthum are
“fungus gnat” and “isopods”. These pollinators can just transfer the pollen grains within a short distance. Therefore, restrict and clumped populations are perfectly fit this pollination model and successfully enhanced the pollination rate. In the case of Asarum dispersal, long-distance dispersal may scatter the population and cause much
more cost and uncertainty (Horvitz et al, 2002).
Moreover, elaiosome on the seed is full of lipid-rich nutrient. High predatory risks are against in natural environment. Short-distance dispersal by ants makes seeds move to a safe place underground within a short time. These features successfully increased the survival rate of seeds and enhanced the fitness of Asarum plants.
In short, limitation to the habitat, pollinator efficiency and safe seed locations contribute to the advantages of short-distance dispersal, and in this case, by ants.
References
1. Agosta, W. C. 2001. Thieves, deceivers and killers: tales of chemistry in nature.
Princeton University Press. 30-33.
2. Beattie, A. J. 1985. The evolutionary ecology of ant-plant mutualism. Cambridge University Press, Cambridge.
3. Berg, R. Y. 1975. Myrmecochorous plants in Australia and their dispersal by ants.
Aust. J. Bot. 23: 475-508.
4. Carroll, C. R. and D. H. Janzen. 1973. The ecology of foraging by ants. Annu. Rev.
Ecol. Syst. 4: 231-258.
5. Culver, D. C. and A. J. Beattie. 1978. Myrmecochory in Viola: dynamics of seed-ant interactions in some West Virginia species. J. Ecology. 66: 53-72.
6. Culver, D. C. and A. J. Beattie. 1980. The fate of Viola seeds dispersed by ants.
Amer. J. Bot. 67: 710-714.
7. Giladi, I. 2006. Choosing benefits or partners: a review of the evidence for the evolution of myrmecochory. /Oikos. 112: 481-492.
8. Gomez, C. and X. Espadaler, 1998. Myrmecochorous dispersal distances: a world survey. /J. Biogeogr. 25: 573-580.
9. Gordon, D. M. 1983. Dependence of necrophoric response to oleic acid on social context in the ant, Pogonomyrmex badius. J. Chem. Ecol. 9: 105-111.
10. Handel, S.N. and A.J. Beattie. 1990. Seed dispersal by ants. Sci. Am. 8: 76-83.
11. Hanzawa, F. M., A. J. Beattie, and D. C. Culver, 1988. Directed dispersal:
demographic analysis of ant-seed mutualism. Am. Nat. 131: 1-13.
12. Headley, A. E. 1949. A population study of the ant Aphaenogster fulva ssp. aquia Buckley (Hymenoptera, Formicidae). Ann. Entomol. Soc. Am. 42: 265-272.
13. Horvitz, C. C., M. A. Pizo, B. B. y Bello, J. LeCorff and R. Dirzo. 2002. Are plant species that need gaps for recruitment more attractive to seed-dispersing birds and ants than other species? In Levey, D. J. et al. (eds.), Seed dispersal and frugivory:
ecology, evolution, and conservation. pp.145-157.
14. Hughes, L., M. Westoby, and E. Jurado. 1994. Convergence of elaiosomes and insect prey: evidence from ant foraging behaviour and fatty acid composition.
Funct ecol. 8: 358-365.
15. Liao J. P. and Qi-Gen Wu. 2000. A preliminary study of the seed anatomy of Zingiberaceae. Bot. J. Linn. Soc. 134:287-300.
16. Marshall, D. L., A. J. Beattie, and W. E. Bollenbacher. 1979. Evidence for diglycerides as attractants in an ant-seed interaction. J. Chem. Ecol. 5: 335-344.
17. Pfeiffer, M., H. Huttenlocher, and M. Ayasse. 2010. Myrmecochorous plants use chemical mimicry to cheat seed-dispersing ants. Funct ecol. 24: 545-555.
18. Sernander, R. 1906. Entwurf einer Monographie der europäischen
Myrmekochoren. Almquist & Wiksells Boktrycheri- A.-B., Uppsala, Sweden.
19. Smith, B. H., P. D. Forman, and A. E. Boyd. 1989. Spatial patterns of seed dispersal and predation of two myrmecochorous forest herbs. Ecology.70:
1649-1656.
20. Whitney, K. D. 2002. Dispersal for distance? Acacia ligulata seeds and meat ants Iridomyrmex viridiaeneus. Austral Ecol. 27: 589–595.
21. Wilson, E. O., Durlach, N.I. and Roth, L. M. 1958. Chemical releasers of necrophoric behavior in ants. Psyche. 65: 108-114.
22. Yinger, B. R. 1983. A horticultural monograph of the genus Asarum senus lato, in Japan. Master thesis, University of Delaware. Newark, DE.