Lianas are large woody vines that climb other plants toward the forest canopy. Although lianas play significant roles in various aspects of forest ecosystems, studies on liana ecology have been relatively few and therefore we have limited information about their life history and diversity therein. In this thesis, I elucidated diversities in several life history traits among five sympatric liana species (Actinidia arguta, Celastrus orbiculatus, Akebia trifoliata, Schisandra repanda, and Schizophragma hydrangeoides) in a cool temperate forest in Nikko, Japan. In the first study (Chapter 1), I evaluated the relationship between preferences in regeneration habitat and shoot production patterns in the five species. The habitat preferences were assessed by comparing frequency of plants in an old forest, a young developing forest, and forest edges. In A. arguta and C. orbiculatus, small and middle-sized plants (< 8 m in height) were frequently found in the forest edges but were scarce in the forest understory, while in S. repanda and S. hydrangeoides, those plants were more frequent in forest understory. The result suggested that the former and the latter species were more successfully regenerated in the forest edges and understory, respectively. In A. trifoliata, plants were frequent both in the understory of the young forest and forest edges, but was absent in the old forest. I next examined biomass allocation between "searcher shoots," which have the ability to attach external structures with twining stems or adventitious roots and show extension-oriented morphology, and "ordinary shoots," which are self-standing short shoots without any specialization for support acquisition and show leaf display-oriented morphology, in young plants (3-8 m in height) that were climbing toward the forest canopy. The proportion of searcher shoots in total current-year shoot mass was varied among species from 6% in S. hydrangeoides to 60% in A. arguta. The species that favored the forest edges (A. arguta and C. orbiculatus) showed significantly higher proportions of searcher shoots than those favored the forest understory (S. repanda and S. hydrangeoides). Because proportion of leaf in shoot mass (leaf mass ratio: LMR) was much lower in searcher shoots than in ordinary shoots, LMR of whole current-year shoots were lower in the forest edge species. Leaf area per leaf mass (specific leaf area: SLA) was greater for the forest understory species and, consequently, leaf area per shoot mass (leaf area ratio: LAR) for the species were about twice as high as those for the forest edge species. A model simulation indicated that a plant with increased allocation to searcher shoots attained more rapid extension early in the growth period, while late in the growth period, a plant with increased allocation to ordinary shoots showed a greater total extension length because of the increased assimilation product due to the increased leaves. Accordingly, the forest edge species were suggested to aim the rapid improvement of light environment by short-term, large extension in the places with large variation in light availability, while forest understory species were suggested to aim steady growth by increasing assimilation in the environment with scarce variation in light availability.

In the nest study (Chapter 2), canopy dynamics of lianas in the forest canopy (about 15 m in height) and their impacts on growths and survivals of host trees were examined in four species (the species in Chapter 1 other than A. trifoliata) to elucidate how lianas cope with a dilemma associated with competitions with their host trees: interception of more light on the host canopies will lead not only to a larger benefit in photosynthesis, but also to a higher risk of fall from the forest canopy due to increased mortality of host trees. Main locations for leaf display greatly varied among the liana species from well-lit upper surface to under the shade of host canopies. A large part of leaves received more than 80% PPFD relative to that above the host canopy in A. arguta, 40~80% in C. orbiculatus, less than 40% in S. repanda, and less than 20% in S. hydrangeoides. In A. arguta, canopy mass, leaf mass, and axis (stems that formed the framework of a canopy) length of sample plants increased linearly with canopy age (years since the liana reached the forest canopy, which was estimated from number of annual rings of the basal stem at 8 m in height), while in the other species, leaf mass and axis length did not increase with the age, suggesting that these species spent for decades in the forest canopy without marked changes in canopy scales. The species that spread in upper positions in host canopies acquired larger number of host trees, and only A. arguta had a tendency to spread to more host trees as it grew in the forest canopy. The two species that distributed leaves in better-lit positions near the upper surface of host canopies, A. arguta and C. orbiculatus, generally decreased annual-ring widths of the host trees, while the two species that distributed leaves in more shady positions within host canopies, S. repanda and S. hydrangeoides, did not show significant effects on the ring widths of hosts. To evaluate potential impact of lianas on survivals of host trees, the length of the basal stem from rooting point to the attachment point to the current host tree (which was expected to depend on number and sizes of former host trees, on which the liana once attached before reaching the current host tree, and which have died and disappeared in the past) was examined for lianas in the forest canopy. The length was larger in the species that spread in upper positions of the host canopies, suggesting that host trees of the lianas that spread in upper positions may have higher mortality than those of the lianas that were located in more shady positions. The results suggested a strategic differentiation among the liana species in behaviors in the forest canopy. Actinidia arguta could be regarded as parasitic in that it intercepted a large amount of light on host canopies. In this species, however, the resultant risk of fall could be decreased by keeping on spreading into many host crowns. On the contrary, S. repanda and S. hydrangeoides were regarded to adopt a strategy close to a commensalism, in which they do not take much of the benefit in light acquisition and instead reduce the risk of fall caused by deaths of host trees.

本論文の内容は2章よりなり、日本の冷温帯林に同所的に成育するつる植物(第1章では、サルナシ、ツルウメモドキ、ミツバアケビ、マツブサ、イワガラミの5種、第2章ではその5種からミツバアケビを除いた4種)を対象に、その生活史戦略の多様性を評価した。これまでつる植物の生態的多様性の評価は皆無であり、森林の純生産や更新などに大きな影響をもつつる植物の研究において、本論文は大きな貢献をしたものと考えられる。

第1章では、高さ成長期の若い個体を対象に、つる植物の更新ハビタットと成長(シュート生産)パターンの関連を評価した。対象種は全て、性質の異なる2種類のシュートを個体内で作り分ける。一つは「探索枝」、これは支持物獲得の機能(巻きつく、付着根を出す)を持ったシュートで、葉を発達させずに長く伸びる(伸長成長指向のシュート)。もう一つは「普通枝」、これは支持物獲得の特殊な性質を示さない直立シュートで、比較的短く、大きな葉を密に展開する(展葉指向のシュート)。両シュートへの相対的な投資量を中心に各種の成長特性を評価した。林縁種のもつ長い探索枝は伸長の効率は低下させるが、新たな支持物の探索(高さ成長の継続、明るい場所への定着)には有利であることが示された。これらの種の成長特性は、光条件の空間的な変化が大きい環境において、近くの明るい空間と新たな支持物を速やかに獲得することに適すると考えられた。林内種のもつ普通枝へ大きな投資という成長特性は、同化物生産を優先しつつ、光資源が乏しい環境で効率良く成長することに適することが示された。

第2章では林冠に到達した成熟個体を対象に、樹冠動態を評価し、その戦略的意義の考察を行った。つる植物はホスト樹木の樹冠に対し、高い位置に多くの葉を展開するほど多くの光を獲得できるため、同化物生産に有利である。一方、光を奪われた結果としてホスト樹木が枯死した場合、上に載っているつる植物も共倒れになる可能性がある。このジレンマにどのように対処しているのかを明らかにするため、つる植物4種に対し、樹冠部の動態とホストの成長、生存に与える影響の両面から評価を行った。ホスト樹冠の上に出る前に成長を止める種は、光獲得の利益は小さいが、その代わりホストに与える負の影響も小さくし、共に長く生きることを目指す片利共生的な戦略であることが示された。一方ホスト樹冠の上に出て多くの光を奪う種はより寄生的であるが、多くのホスト樹冠に広がることによって個々のホストへの悪影響を分散し、落下のリスクを低下させるものと考えられた。

本論文第1章は舘野正樹、長嶋寿江との共同研究であるが、論文提出者が主体となって研究を行ったものである。

したがって、博士(理学)の学位を授与できると認める。