The Ants and Elephants of Ecosystem Restoration

From inside the dung of a macaque a seed sprouts. It will be carried in the swell of this wild Kalu River till it takes hold along the riverbank and bury its roots deep into the earth, making its mark in rewilding the Earth.

A zoochorous seed’s journey

Towards the beginning of winter when wild grass is heavy with seed, clumps of discarded seedcoats can be seen at terminals of tiny whitish lines drawn across the forest floor. These small lines belong to a specific ant – the Harvester Ant. In India, ants in the genus Monomorium, Pheidole, and Meranoplus discard parts of seedcoats in heaps – called ant middens – around burrows that lead to their underground nests.

Midden of a Harvester Ant (Monomorium sp.) in a meadow in northern Western Ghats.

Grass seeds are dispersed by wind, eaten up by passerines, and collected by mice. Ants are the lesser-known hoarders – more than handfuls of seedcoats of assorted grasses can be collected from the middens in the peak of the harvest season, just as we start harvesting rice in the tropics. And just as we spill some during the harvest, so do ants. Harvester Ants are seed predators, they store the grain as food – and in the process of transport they also disperse seeds. A study of the middens of eight species of Harvester Ants in Australia recorded 33 species of plants (Andersen et al., 2000). The unconsumed seeds have a chance to grow in areas not normally accessible to the plants.

Elaiosomes of Borago officinalis, Chelidonium majus, Corydalis cava, Corydalis sempervirens, Cytisus scoparius, Dicentra torulosa, Euphorbia lathyris, Euphorbia cyparissias, Jeffersonia diphylla, Pentaglottis sempervirens, Symphytum officinale, Viola elatior (most of these native to the European region). © Hans Stuessi, CC BY 4.0 <>, via Wikimedia Commons. Retrieved from:

Some plants have upped the game by growing a small edible ‘fleshy’ structure attached to the seed called elaiosome (literally oily-body in Greek). This is a plant’s way to get its seed picked up and dispersed in nooks and corners of the world only an ant can reach. The seeds of these plants may not be edible, an ant interests only in the nutritious elaisome. After consuming the elaiosome, the seeds are discarded, giving them a chance to grow, sometimes from right inside an ant’s colony. This animal-mediated seed dispersal – zoochory – is called myrmecochory which primarily engages ants. As many as 10,000 to 20,000 species of plants grow elaisosomes.

Also close to the ground, the next industrious seed harvester is the mouse – or rodents, in general. While the large rodents such as porcupines are solely seed predators, some rats and mice have a habit of scatter hoarding seeds for future consumption. These seed caches are either hidden over a wider area to protect them from pilferers or are stored inside chambers in burrows or tree hollows. These sites, too, act as nurseries for seeds to sprout since not all are recovered by their owner or sniffed-out by other pilferers.

A Coppersmith Barbet checks unripe figs of Ficus virens in Central India. The ripe ones it eats contain hundreds of tiny seeds – some of which pass through its gut undigested. Such seeds are dispersed through bird droppings as they fly about different areas.

Higher up in the canopy, hornbills, barbets, and frugivorous bats are crucial aerial seed dispersers. A study on dispersal mode and spatial distribution of trees in a north-east Indian rainforest found that 78% of the 128 tree species the researchers studied were dispersed by animals, 54 of which were primarily bird-dispersed. The study showed that the density of trees primarily dispersed by birds was higher than trees that relied on other animals for dispersal (Datta & Rawat, 2008).

A mother Asian Elephant and her calf  (both camp elephants) relish on the aquatic vegetation in Central India. Elephants eat a lot in a day, excreting nutrition-rich dung containing seeds some of which remain undigested – these sometimes sprout right through the dung many mountains thence as they traverse.

The other higher vertebrates, the primates and elephants, also disperse seeds far and wide. Spider monkeys and howler monkeys in a Peruvian rainforest eat fruits of 71 and 14 species, respectively, the seeds dispersed through their droppings. In north-east India, wild elephants showed dispersal over 9 to 11 times farther than by domestic bovids (Sekar et al., 2015); some studies put this distance from 20 to as much as 50 km in a wide range of habitats (Dudley, 2000; Sekar et al., 2015). The secondary dispersers such as dung beetles roll the dung of herbivores and bury it underground, where some of the seeds sprout – this is called diplochory. This intricate web is an important part of ecosystem functioning.

When a tree falls, who leaves first?

From miniscule grain of wild grass to fleshy fruits of emergents, from tiny ants to lumbering elephants, all that feed on seeds, also aid in dispersal – enabling plants to travel distances; from beyond the round in the woods with the help of an ant, beyond the mountain with the help of a hornbill, to beyond the river with the help of an elephant. An ecosystem is whole when its green vegetation interacts with its faunal diversity. To know that a particular ecosystem is whole, even without understanding the invisible interactions, makes understanding presence of some key species important. Elephants, frugivore bats, hornbills, even ants, are some of the keystone species of certain ecosystems they help shape. Just as threshold densities of certain apex carnivores indicate a balanced food-web; several animals that are specific to certain environments and sensitive to disturbances indicate the health of an ecosystem.

Grasslands are perhaps the most neglected ecosystems, perhaps for the lack of charismatic tree species. A healthy grassland, however, is as unique as a healthy forest – with birds, bees, deer, elephants, even tigers as seen here in a Central Indian grassland meadow.

We rely more on animals for the tell-tale signs of change because of their animate nature: some mammals and birds respond quicker to disturbances than do plants which may take several generations to show the effect. In other words, in a disturbed ecosystem, animals leave first; which animals leave when indicates the level of disturbance.

In sub-Saharan African rainforest, swarm-raiding army ants form massive columns that move about the forest floor. A community of birds called ‘ant-following birds’ stalk these swarms to feed on arthropods flushed by the marching ants. Researchers found the species richness of birds and sizes of bird flocks decrease with decreasing size of forest fragments, which are much lower in degraded forests than undegraded – as was the case with ants. Composition of flocks, too, was variable in smaller fragments and degraded forests than in undegraded and large fragments (Peters et al., 2008). In landscapes with highly fragmented habitats, the size of the fragment and the distance of these fragments shows greater loss of bird species diversity and population declines than is expected from habitat loss alone (Andrén, 1994).

These right-for-the-picking Ficus nervosa figs of the southern Western Ghats are eaten by a number of animals. Once the tiny fig wasps pollinate them, the ripe fruits are consumed by a variety of birds, rodents, bats, bears, and primates – enabling dispersal of this canopy tree of India’s rainforests.

How is this relevant for ecosystem restoration? Fast-forward to a restored ecosystem, the faunal diversity of this state should resemble the diversity of the original or the remnant ecosystem. Quite naturally, some animals are first to leave and last to return, hence, ecosystem restoration does not account for the period in-between their departure and homecoming, that is, during the process of restoration, the focus is on primary producers. That both – flora and fauna – are intricately connected is often not considered – perhaps for the lack of better understanding – affects selection of appropriate tree species and abundances for restoration. In tropical forests of north-east India, researchers found that in less disturbed forests, hornbill food trees were three times higher in abundance, corresponding with a 22-times higher abundance of hornbills, while logging reduced abundance of hornbill food trees just as hunting diminished hornbill – the primary disperser – abundances (Naniwadekar et al., 2015).

In grasslands, the intricacies of animal-plant interactions are just as entangled. A study in agriculture-wilderness mosaics in Europe showed that in semi-natural habitats, bee diversity declined for the lack of floral resources, especially when insect-pollinated plants were low in numbers. Researchers found both, plant-mediated effects on bee functional diversity and also of bee-mediated effects on flowering plant richness; interestingly, in presence of patchy habitat diversity, bee species richness benefitted plant richness. In other words, degraded patches can benefit from fragmented but closely located undisturbed patches so long as nesting and food resources are available (Papanikolaou et al., 2017).

Hand-picking seeds from neighbouring remnant forests is the best way to culture and plant them in restoration sites, however, care must be taken to consider abundances of dietary floral seeds which may be underrepresented in restoration sites. Studies on floral and faunal interactions, therefore, should be an important primer to restoration activities.

There are many reasons why an ecosystem cannot be fully restored. It could be that the soil is changed in its composition. It could be that the mycorrhiza in the soil is lost; the groundwater exhausted. In addition, the intricacies with which an ecosystem as a community of constantly interacting organisms – the flora and fauna – works is still not fully understood. For instance, what do – or did – the animals eat, what seeds did they excrete? Quite often, afforestation efforts underestimate the dietary trees of animal dispersers and disproportionately prioritize large amounts of non-dietary seeds – a nuance that is further ignored in projects solely looking at carbon sequestration which disregards other carbon-based lifeforms: animals. For restoration, while soil can be treated and water restored, animals for the large part are left out – they come into picture after restoration process is well underway. Are we missing something?

The missing mediators of ecosystem restoration

Ecosystem restoration follows succession – at an accelerated pace. In areas where surface soil is stripped of its nutrients, acceleration needs to be amended with soil conditioners. In threatened prairie grasslands, a study found that a combination of compost and biochar, and arbuscular mycorrhizal fungi with compost, resulted in an accelerated plant growth (Ohsowski et al., 2017). Other studies also found arbuscular mycorrhiza to accelerate succession (Koziol & Bever, 2016). Following this, a restored habitat substantially improves pollinator community diversity – more than double compared to degraded sites according to one study (Sexton & Emery, 2020). Pollinator diversity significantly improves in older restoration while younger restoration sites show only minute improvements, this means that pollinator communities may need more time or more of key plant species to achieve restoration goals (ibid).

A single grass frond alone can host a number of animals, seen here in a Prairie grassland is an aphid, the larva of a ladybird beetle which preys on aphids, and a Crabronid aphid wasp that parasitizes aphids.

Restoration goals should be about bringing back lost relationships: not only native herbs but bees, grass and ants, emergent and hornbills, large-fruited trees and primates and elephants, and combinations of all such interactions. We’ve already seen ‘why’ we need animals in the picture, ‘how’ do we do it? In other words, can restoration process be in-tandem for certain faunal species?

Studies show that degraded habitats lack the corresponding animal diversities, but studies on how much it takes – in terms of time and resources – to accelerate restoration mediated by animals are few. One study from north-east India looked at this scenario: if megaherbivores such as elephants were to be replaced by domestic ruminants, what effect will it have on seed dispersal – and by extension the forest composition? Seeds from three species of trees consumed by elephants were 2.5 to 26.5 times more likely to pass undigested from elephant dung compared to domestic bovids. Furthermore, wild elephants dispersed undigested seeds farther, with at least 20% of the seeds being dispersed farther than by bovids (Sekar et al., 2015). It isn’t practical bringing elephants back into the picture early on, but we can start small.

Soil moisture and biomass are critical to not only encourage growth of beneficial fungi and bacteria, but also detritivorous arthropods, which aid in succession of animals just as moist, humus-rich soils accelerate growth of herbs and shrubs. Because acceleration of ground biomass may promote invasive species to grow, weeding is important.

Like bees, some species of ants are used as indicators of ecosystem health and can indicate a restored ecosystem. In Australian tropics, a restored mining site showed as many as 43 species of ants at a site from seven species at an unvegetated site prior to restoration. The study indicates that the soil microbial biomass – the mulch and plant matter as it is decomposed by microorganisms – more than plant species richness, as an important precursor for the ant homecoming. Just as mycorrhiza is added to promote plant growth, can mulch be added to the soil to promote quicker microbial biomass?

The leaf-cutter bee is a short-range solitary pollinator bee making a nest in tiny pre-existing holes in various substrates – from wood to concrete. Artificial holes in a natural substrate such as wood are often adopted by these bees to nest. Here, Megachile lanata it is seen feeding on Crotalaria spectabilis, a native gregarious nectaring plant of Central India that can be used in biodiversity hubs.

Since restoration work is spatially large, the mediators can be introduced in ‘biodiversity hubs’: these slightly more labour-intensive hubs focus on accelerating animal activities in select spots within a restoration site: in the river ecotones, in marshy areas, in areas where animals – especially invertebrates, can be reintroduced from neighbouring remnants at small scales after a thorough study of the faunal community of the ecosystem in question. For example, species-specific artificial nests for native solitary bees such as the leaf-cutter bees and mason bees which are important short-range pollinators, can be introduced in these ‘hubs’ alongside well-watered growth of native herbs and shrubs to sustain their populations. Natural mulch can be added at these hubs to promote flies such as native soldier flies to breed, which are excellent decomposers as maggots and pollinators as adults.

Leschenault’s Rosuette (Rousettus leschenaultii) feeding on  flowers of a horticulture variety of banana, bats help pollinate as well as disperse wild bananas, however, horticulture bananas lack seeds. It is important to consider wild or near-wild varieties of native plants in hubs and islands to enable dispersal of seeds. Horticulture varieties, although fast growing and attractive to feeders, fail at promoting regeneration of a forest.

Caution is extremely important. Planting non-native plants in restoration sites to attract animals is counterproductive. There are two ‘tread-carefully’ instances here: First, in well monitored and well-funded restoration sites, islands of Lantana camara – a highly invasive but excellent nectar-bearing shrub, may be retained in the initial days of succession to allow pollinators that subsist on these shrubs to survive. Second, selecting horticulture varieties of plants in biodiversity hubs or islands is not viable, one such example is of commercial bananas: easy to grow, easy to flower and fruit, these attract bats from all corners, except, the seeds do not exist or are not viable, making dispersal a failed task. Bat Conservation International (2021) matter-of-factly state why we need bats and wild bananas: “The plants that produce all that tasty fruit are so genetically similar that a single disease could devastate the global crop. In fact, some scientists warn that commercial bananas may already be at great risk from a recently reported fungus.”  Selecting wild or native varieties of commercial crops – wild bananas in this instance – is a better substitute for bats, even elephants. We must remind ourselves that ecosystem restoration is not just about green and red (indicating animals) cover – it is about restoring native and lost genetic diversity, too.

A study found that increasing plant richness with legumes and forbs – largely insect-pollinated – to restore grasslands increased pollinator functional diversity and abundance, it also improved visitations to nearby crops pollinated by insects (Orford et al., 2016). If introduction of such animals is a resource-intensive process, ‘biodiversity islands’ can be created with fast growing native shrubs and trees that naturally attract native wildlife and aid in the restoration process of native fauna. As for elephants, restoration of degraded elephant passageways with plantation of native trees dispersed by elephants in greater proportion, in addition to creating biodiversity islands of fast-growing elephant food species as an additional measure to ensure they do not visit farmlands, as opposed to solely planting commercial fruiting trees in forests to take their attention away from farmlands, needs to be explored.

A signage in Central India to make people aware - and inform the authorities - on elephant movement across the road; to the right is a fenced-off coupe and to the left a degraded forest leading straight to farmlands. Animal passes with the right kind of vegetation are more productive than disturbed habitats which exacerbate human-wildlife interactions.

The missing mediator of an ecosystem you plan to restore might be anything, from an ant to an elephant. The question is, how do we incorporate them in the restoration process? Integrative studies that look not only at flora but also fauna and their interactions are important for any restoration undertaking; understanding this interaction prior to restoration is an important first step; it helps us not only in selecting the right floral species but also in right abundances relative to other species. The point is not about picking animals from one site and relocating them to the restoration site – though this works in some ecosystems where the animal itself engineers the habitat, the investments required are ginormous – it is about making the restoration process itself conducive to their natural – and early, if not quick – arrival to aid in restoration or to start reclaiming the restored site early on.

As the world comes to understand the delicate intricacies of ecosystem restoration, processes that accelerate the return of native fauna during restoration such as biodiversity hubs and islands, need to be tested carefully if restoration processes of succession are to be more holistic, where animals are not end-users but mediators.


Further reading in order of citation

Andersen, A. N., Azcarate, F. M., Cowie, I. D. (2000). Seed selection by an exceptionally rich community of harvester ants in the Australian seasonal tropics. Journal of Animal Ecology. 69. Ppp. 975-984. Retrieved from:

Datta, A. & Rawat, G. S. (2008). Dispersal Modes and Spatial Patterns of Tree Species in a Tropical Forest in Arunachal Pradesh, Northeast India. Tropical Conservation Science.

Sekar, N., Lee, C., & Sukumar, R. (2015). In the elephant’s seed shadow: the prospects of domestic bovids as replacement dispersers of three tropical Asian trees. Ecology. 96(8). pp. 2093-2105. Retrieved from:

Dudley, J. P. (2000). Seed dispersal by elephants in semiarid woodland habitats of Hwange National Park, Zimbabwe. Biotropica. 32(2). pp. 556-562. Retrieved from:

Peters, M. K, Likare, S., & Kraemer, M. (2008). Effects of habitat fragmentation and degradation on flocks of African ant-following birds. Ecological Applications. 18(4). pp. 847-858.

Andrén, H. (1994). Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: a review. OIKOS. 71(3). pp. 355-366.

Naniwadekar, R., Shukla, U., Isvaran, K., & Datta, A. (2015). Reduced hornbill abundance associated with low seed arrival and altered recruitment in a hunted and logged tropical forest. PLOS ONE.

Papanikolaou, A, D., Kuhn, I., Frenzel, M., Kuhlmann, M., Poschlod, P., Potts, S. G., Roberts, S. P. M., & Schweiger, O. (2017). Wild bee and floral diversity co-vary in response to the direct and indirect impacts of land use. Ecosphere. 8(11).

Ohsowski, B. M., Dunfield, K., Klironomos, J. N., & Hart, M. M. (2017). Plant response to biochar, compost, and mycorrhizal fungal amendments in post-mine sandpits. Restoration Ecology. 26(1). pp. 63-72.

Koziol, L., & Bever, J. D. (2016). The missing link in grassland restoration: arbuscular mycorrhizal fungi inoculation increases plant diversity and accelerates succession. Journal of Applied Ecology. 54(5). pp. 1301-1309.

Sexton, A. N., & Emery, S.M. (2020). Grassland restorations improve pollinator communities: a meta-analysis. Journal of Insect Conservation. 24. pp. 719-726.

Bat Conservation International. (2021). Bats and disappearing wild bananas. Retrieved from:

Orford, K.A., Murray, P. J., Vaughan, I. P., & Memmott, J. (2016). Modest enhancements to conventional grassland diversity improve the provision of pollination services. Journal of Applied Ecology. 53. pp. 906-915.