UC Berkeley

Coevolution between phylogenetically distant, yet ecologically intimate taxa is widely invoked as a major process generating and organizing biodiversity on earth. Yet for many putatively coevolving clades we lack knowledge both of their evolutionary history of diversification, and the manner in which they organize themselves into patterns of interaction. This is especially true for mutualistic associations, despite the fact that mutualisms have served as models for much coevolutionary research. In this dissertation, I examine the codiversification of an obligate, reciprocally specialized pollination mutualism between leafflower moths (Lepidoptera: Gracillariidae: Epicephala) and leafflower trees (Phyllanthaceae: Phyllanthus sensu lato [Glochidion]) on the oceanic islands of southeastern Polynesia.

Leafflower moths are the sole known pollinators of five clades of leafflowers (in the genus Phyllanthus s. l., including the genera Glochidion and Breynia), and thus this interaction is considered to be obligate. Female moths actively transfer pollen from male flowers to female flowers, using a haired proboscis to transfer pollen into the recessed stigmatic surface at the end of the fused stylar column. The moths then oviposit into the flowers' ovaries, and the larva which hatches consumes a subset, but not all, of the developing fruit's seed set. This interaction is known as a pollinating seed predation mutualism, because the same insect responsible for pollinating the flowers also feeds on the seeds of the same plant as a larva. Furthermore, as best as is known, this mutualism is characterized by high reciprocal species-specificity, with each leafflower species associated with only one or two moth species, and vice versa. In these salient characteristics, the leafflower/leafflower moth mutualism is thus analogous to the classically known mutualisms between fig trees and fig wasps, and yucca plants and yucca moths.

In my dissertation I examined the best-studied leafflower clade, the genus Glochidion, with 300 described species distributed from Pakistan to Japan, south to Australia, and east to the Pitcairn Islands. Because of their high species diversity and presumably complex biogeographic history, I focused on the co-radiation of 24 Glochidion species and their Epicephala moths on the oceanic islands of southeastern Polynesia (Cook Islands, French Polynesia, and the Pitcairn Group). All these islands are formed by volcanoes as the Pacific plate moves over a series of stationary hotspots in the earth's mantle, and the majority are of Plio-Pleistocene age (>5 Ma). All the Glochidion species are endemic to this region, and nearly all are described as endemic to single archipelagos; they constitute one of the largest endemic plant radiations in this biodiversity hotspot. These species differ primarily in the morphology of female flowers. Most islands have 1--3 species, but large islands have 4--7, with multiple species occurring sympatrically. All these factors suggest this diversification has occurred over very short evolutionary timescales.

In Chapter 1, I demonstrate that the mutualism between Glochidion and Epicephala, previously described on continents, is present in southeastern Polynesia. Glochidion had previously been reported from nearly all high islands in this region, but Epicephala from only two (Nuku Hiva and Fatu Hiva, Marquesas). Because of the difference in generation time between Glochidion and Epicephala, it is unlikely that they could colonize a new island in a single dispersal event. With my co-authors, I observed and photographed Epicephala pollinating Glochidion flowers on two islands in the Society Archipelago (Tahiti and Raiatea), discovered pollen on museum specimens of Epicephala from several other Pacific islands, and reared Epicephala from the fruit of 19 species of Glochidion on 17 islands where it had not previously been reported. These results indicate that this mutualism has been able to repeatedly dissemble and reassemble on remote islands. This conclusion is of interest because it is in contrast to a long-standing island biogeography paradigm that organisms with specialized biotic interactions should be unable to colonize remote islands, and also because of concern over the resiliency of specialized mutualisms to global change.

In Chapter 2, I demonstrate as a result of a three-year rearing inventory on 21 islands that not only Epicephala, but several other lineages of insects specialized on the leafflower family (Phyllanthaceae) have colonized southeastern Polynesia. These include the leafmining moth Diphtheroptila (Gracillariidae) and seed-feeding moth Tritopterna (Tortricidae), and potentially also the leafrolling moths Caloptilia (Gracillariidae) and Dudua (Tortricidae). This indicates that niche conservatism in host plant use may play an important role in the assembly of phytophagous insect communities even on remote islands. I discuss this result in the context of the literature on adaptive radiation (onto new host plants) and niche conservatism (on the same host plant) in phytophagous insect faunas on oceanic islands generally.

In Chapter 3, I use molecular phylogenetic methods to reconstruct the evolutionary history of southeastern Polynesian Glochidion and their Epicephala within the geographic context of the entire Asia-Pacific region. Glochidion trees have colonized southeastern Polynesia twice, with one colonization spreading throughout the southern Cook, Society, Austral, Marquesas, and Tuamotu archipelagos, and a separate colonization of Mangareva (Gambier Islands). Epicephala moths have colonized southeastern Polynesia at least twice. An older colonization has spread throughout the southern Cook, Society, Marquesas, Tuamotu archipelagos and Rapa (Australs), and a younger colonization has spread through the southern Cooks, northern Australs, and Societies. Not only has this obligate mutualism diversified in a manner that has not resulted in phylogenetic congruence, but one pollinator lineage (the younger) has rapidly shifted onto a large number of novel host species in Glochidion over a wide geographic area. These results indicate a role both for geographic isolation and host-shifts in the diversification of this mutualism. More importantly, these results indicate that despite all the constraints that might be expected to prevent multiple colonizations by Epicephala (geographic isolation, reciprocal specialization between Glochidion and Epicephala species, and competitive exclusion between older and younger Epicephala clades), Epicephala have been able to establish multiple times on these remote islands and spread rapidly onto new hosts once they arrive. Consequently, patterns of association are likely to be similarly dynamic in these between Glochidion and Epicephala on continents. Implications of these findings for understanding how pollinating seed-predation mutualisms diversify are discussed.

In Chapter 4, I use molecular phylogenetic and network analysis methods to ask whether the patterns of interaction between Glochidion and Epicephala species are reciprocally specialized and modular to the same degree as in continental regions. I focus on the three youngest Society Islands with Glochidion present (Huahine, Moorea, and Tahiti), which collectively have 10 species of Glochidion. I find that unlike on continents, Glochidion-Epicephala networks on Tahiti and Huahine lack modularity. These findings may provide a view onto the early stages of diversification in pollinating seed-predation mutualisms, and suggest that coevolving clades may pass through different patterns of interaction as they diversify.

Taken together, these results indicate that these free-living mutualists have colonized remote oceanic islands repeatedly, but not congruently. These repeated colonizations have led to a dynamic coevolutionary history over only a few million years, with evidence for host-shifts, diversification on different archipelagos, and patterns of interaction which are less modular than are previously known from this and similar systems on continents.