The Asian elephant (Elephas maximus) is one of the few remnant megaherbivores and the largest animal in Asian ecosystems. The extreme body size confers to Asian elephants unique functional properties and they are hence assumed to play an important role in ecological processes. In spite of this, little is actually known about this role. Asian elephants are endangered and have already disappeared from most of their historical range. Given this trend, it is urgent to understand their role in Asian ecosystems to evaluate the consequences of their local disappearance.

The aim of this PhD thesis is to contribute to the understanding of the ecological role of Asian elephants, particularly as agents of seed dispersal. Seed dispersal is a key process in the life cycle of plants with profound consequences for the structure and trajectories of plant populations and communities. Many plants rely on mutualistic interactions with animas to disperse their seeds, and Asian elephants have attributes to be efficient seed dispersers. Elephants eat large amounts of a broad diversity of plants, which can result in the dispersal of large quantities of seeds from a variety of plant species. Moreover, the combination of long gut passage times and wide-ranging behavior can result in a broad spatial scale of dispersal. The ecological flexibility of Asian elephants can result in strong context-dependency of their ecological role. Therefore, a variety of ecological scenarios must be considered to understand the role of Asian elephants in seed dispersal processes.

I present here four interconnected studies. First, I describe the food habits of Asian elephants in a mixed-deciduous forest of central Myanmar (Burma), with special focus on the plants potentially dispersed by them. Second, I describe the type of plants dispersed by Asian elephants in two dry monsoon environments of southeastern Sri Lanka, looking at the influence of landscape structure upon the type of seeds dispersed by elephants. Third, I describe the spatial scale at which elephants disperse seeds in the two previous environments and test hypotheses about the influence of behavioral and physiological factors on it. Finally, I study the role of Asian elephants on the dispersal of an invasive plant species in southeastern Sri Lanka comparing with that of other sympatric mammalian dispersers.

1. Food habits of Asian elephants in South East Asia

Little is known about the basic ecology, including food habits, of Asian elephants in South East Asia. This is partially due to the difficulty to study elephants living at low densities in habitats with low visibility. In countries like Myanmar there are thousands of captive elephants living in semi-natural conditions that might be used as a proxy to study elephant ecology. I visited an elephant camp in central Myanmar and interviewed mahouts and veterinarians to describe the diet of captive elephants feeding on natural vegetation of a mixed-deciduous forest. Elephants showed a broad dietary breadth (103 plant species from 42 families); consumed mostly browse plants (94% of species); and were very selective about plant parts (e.g. many trees were eaten exclusively for their bark (22%) or fruits (14%)). I recorded the fruits from 29 plant species as eaten by elephants. I found several of these as fruit remains, seeds, or seedlings in elephant dung, suggesting that elephants play a role in seed dispersal in this environment. Work elephants and their mahouts proved to be a rich source of information to understand wild elephant ecology.

2. Seeds dispersed in Sri Lanka and the effect of landscape structure

Although frequently ignored, the ecological role of animals in seed dispersal is highly context-dependent. Human alteration of landscapes (e.g. forest fragmentation) can thus alter this role. I analyzed the species and quantity of seeds in the dung of Asian elephants at two locations of southeastern Sri Lanka. These areas differ in their landscape structure. Kotiyagala is a highly fragmented area in which the original forest has been largely replaced by grasslands; Mattala, on the other hand, is a relatively continuous scrubland. I collected approximately 30 dung samples per month for a period of one year in Kotiyagala but in Mattala sampling collection was interrupted after five months. I found seeds in most (93.5%, N=479) of dung piles with an average of 1.6 (±1.0 SD, max6) species per sample. The seeds retrieved from elephant dung were categorized as 69 different plant species (only 14 present in both areas). I classified plants into four gross categories: (1) native woody and (2) herbaceous plants, (3) invasive plants, and (4) crops. In the period of five months sampled in both areas there was an interaction between sampling area and the presence (frequency and abundance) of seeds in elephant dung. In the scrubland elephants dispersed native woody plants at higher frequency and quantities while in the fragmented area invasive and crop plants were far more common. This study shows the importance of human alteration of landscapes over the ecological role of animals, and undercovers a new form of conservation conflict - the spread of invasive plants by an endangered wildlife species.

3. Spatial scale of dispersal and factors affecting it

The digestive physiology and movement patterns of animal dispersers determine deposition patterns for seeds dispersed by endozoochory. I combined data from feeding trials, germination tests, and GPS telemetry of Asian elephants to (1) describe the spatial scale at which Asian elephants disperse seeds; (2) assess whether seasonal differences in diet composition and ranging behavior translate into differences in seed shadows; and (3) evaluate whether scale and seasonal patterns vary between two ecologically distinct areas - Sri Lanka's dry monsoon forests and Myanmar's mixed-deciduous forests. The combination of seed retention times (mean 39.5 h, max 114 h) and elephant displacement rates (average 1988 m in 116 hours) resulted in 50 % of seeds dispersed over 1.2 km (mean 1222-2105 m, max 5772 m). Shifts in diet composition did not affect gut retention time and germination of ingested seeds. Elephant displacements were slightly longer, with stronger seasonal variation in Myanmar. As a consequence, seed dispersal curves varied seasonally with longer distances during the dry season in Myanmar but not in Sri Lanka. Seasonal and geographic variation in seed dispersal curves was the result of variation in elephant movement patterns, rather than the effect of diet changes on the fate of ingested seeds.

4. Role of Asian elephants and other mammals mediating a plant invasion

The mesquite (Prosopis juliflora) is a leguminous tree from Central and South America that has been planted in tropical arid areas across the world, where it has frequently become invasive. Mesquite was introduced in Hambantota (southeastern Sri Lanka) in the early 1950s and has subsequently expanded, invading neighborhood areas such as Bundala National Park. I investigated the presence of mesquite seeds in the feces of Asian elephants and other two functional groups of dispersers (cattle and primates) at sites located at different distances from the current distribution of mesquite trees. I also analyzed the effect of consumption by these animals upon germination of mesquite seeds. Elephants were the only animals to disperse seeds at all the spatial scales considered; cattle (cows Bos Taurus and buffaloes Bubalus bubalis) dispersed the largest number of seeds although mostly at short distances; primates (grey langur Semnopithecus entellus and tokey monkeys Macaca sinica) dispersed few seeds only at short distances. Seeds consumed by cattle showed improved germination rate (28.3%) compared with control (17.0%) and seeds consumed by elephants (9.2%) and monkeys (4.5%). Germination time was increased by animal consumption - while control seeds stopped germination after 17 days, seeds consumed by animals continue to germinate three months after being planted. Animals enhance mesquite's invasion expanding its range and helping to colonize new areas (elephants) and consolidating current populations (cattle). Human management of wild and domestic mammals should take this into consideration.

This thesis is the most comprehensible study on the role of Asian elephants in ecological processes as well as one of the most complete studies regarding seed dispersal by a megaherbivore. Overall, elephants proved to be an efficient seed dispersal agent that consumed and dispersed large amounts of a broad variety of plants, from different functional types, at long dispersal distances. The role of elephants as dispersers is context dependent, showing variations across seasons, geographical areas, and human alterations of landscape structure. Although elephants can play a positive role promoting forest maintenance and succession, their efficiency can also have a negative side e.g. promoting the spread of invasive plants.

アジアゾウ (Elephas maximus)は、アジアの生態系で最大の植食動物であり、独特な生態、行動、形態的特性をもっており、生態学的に重要な役割を担っていると考えられる。しかし、その実態はほとんどわかっていない。本種は本来の生息域のほとんどから姿を消しており、その局所絶滅の影響を評価するために、生態学的役割を早急に解明する必要がある。本論文のねらいは、アジアゾウの生態学的役割、特に種子散布者としての役割を明らかにすることである。

本論文は、以下の4部構成をとった。1. ミャンマー中心部の落葉混交林におけるゾウの食性の記述。2. スリランカ南東部にある2か所のモンスーン乾燥地域においてゾウにより散布される植物の記載と、景観構造がそれらの植物の種子形状タイプ等にあたえる影響評価。3. ミャンマーおよびスリランカにおいてゾウが種子を散布する空間スケール解明と、そのスケールに行動的・生理的要因があたえる影響評価。4. スリランカ南東部において、ゾウが侵入植物種の分布拡大に果たす役割解明と、同所的に生息する他のほ乳類種子散布者の役割の比較。これらの結果を統合し、アジアゾウの種子散布者としての役割を考察した。

1. 東南アジアにおけるアジアゾウの食性

ミャンマーでは、多くのゾウがほぼ自然環境下で飼育されており、野外研究の代用にできる。ゾウ飼育がおこなわれているキャンプを訪問し、湿潤落葉樹林の自然植生で飼育ゾウの食物メニューを記述した。ゾウは植物42科103種の植物を採食しており、ほとんどが双子葉植物だった(全食物メニューの94%)。また、そのほとんどは皮部分(22%)や果実(14%)のみを採食しており、採食部位に関して高い選択性を示した。果実は29種だった。このいくつかは、ゾウの糞中に果実、種子、実生が残っており、ゾウが種子散布の役割を担っていることが示唆された。

2. スリランカにおいてゾウが散布した種子と景観構造の影響

種子散布における動物の役割は、森林分断化などの人為改変によって役割が変化することがある。スリランカ南東部の景観構造が異なる2か所 (森林分断化が進んだ地域と連続した低木林) で調査したところ、ほとんどの糞 (93.5%, N = 479) に種子がはいっており、1糞中に平均1.6 (± 1.0 SD, 最大 6) 種が確認された。糞から回収された種子は69種にのぼったが、両調査地で確認された植物は14種のみであった。低木林のゾウは土着木本を高い頻度と量で散布していたが、分断化が進んだ地域では侵入種や穀物を散布していた。この結果は、動物の生態学的役割に対する景観の人為改変が及ぼす効果の重要性だけでなく、新しい形の保全上の矛盾、絶滅に瀕した野生動物による侵入植物の種子散布という問題を提示している。

3. 種子散布の空間スケールとそれにあたえる要因

種子散布動物の消化生理と移動様式は、被食散布型の種子の散布様式に影響する。そこで、給餌実験と発芽実験の結果と、GPS追跡の結果から、(1) ゾウが種子散布する空間スケール、(2)ゾウの食物メニューおよび遊動によるseed shadow (種子散布距離の確率分布)の変化、(3) スリランカの乾燥モンスーン林とミャンマーの湿潤落葉樹林における散布様式の違いを明らかにした。種子が体内を通過する時間とゾウの移動速度の組み合わせから、散布種子の50%は1.2 km以上遠くに運ばれることがわかった (平均 1222-2105 m, 最大 5772 m)。食物メニューは消化管通過時間や消化管を通過した種子の発芽率と関係なかった。ミャンマーにおいては種子散布曲線が季節変化し、乾期に長距離運ばれるというパターンがあったが、スリランカではそのような傾向はなかった。

4. アジアゾウおよび他のほ乳類が外来植物侵入を介在する効果

中南米原産のマメ科低木メスキート (Prospis juliflora) は、世界各地の熱帯乾燥地域に定着している。メスキートの分布域から異なる距離に位置する地点で、ゾウおよび散布機能が異なる他のほ乳類(牛類と霊長類)の糞中のメスキート種子を調べた。また、これらの動物による消化がメスキート種子の発芽にあたえる効果も調査した。調査したすべての空間スケールにわたって種子散布していたのはゾウのみであった。牛類は多くの種子を散布していたが、ほとんどは短距離移送であった。霊長類はほとんど種子を散布せず、移送も短距離であった。牛が消化した種子については、コントロール(17.0%)やゾウ(9.2%)、サル(4.5%)が消化したそれよりも発芽率が高かった(28.3%)。ゾウはメスキートの分布拡大に影響しており、牛や霊長類は現在の分布域内での密度増加に影響することで、メスキートの侵入を助けている。在来ほ乳類の管理をおこなう際には、この問題を考慮する必要がある。

本論文はアジアゾウの生態学的役割に関するもっとも包括的な研究であると同時に、大型草食動物による種子散布に関するもっとも完全な研究のひとつである。アジアゾウの生態学的な役割、特に種子散布者としての役割を解明し、関連の生態系保全のあり方についても考察した重要な研究と考えられる。したがって、本研究は基礎、応用両面から学術上貢献するところが大きく、審査委員一同は本論文が博士(農学)の学位論文として価値あるものと認めた。