Pollen related with functional optimization of aroid traps?

The Araceae, subfamily Aroideae, provide examples of floral traps with functional perfection: insects are (usually deceptively) attracted by scent, glide down into a pollination chamber, deposit pollen possibly brought along, are captured until pollen is released from the anthers, are loaded with pollen, and finally are released. Starting with the classical studies of Knoll in 1926 on Arum nigrum, a surprising diversity in morphological and functional patterns of such floral traps has been discovered in the family. Recent palynological and molecular data now cast new light on the evolution and the functional differentiation of floral traps and their preceding forms (starting with "non- traps", i.e., spadices with spreading spathes, or with only loosely enclosing spathes, followed by "semi-traps", ending with "perfect kettle-traps"). Study of the stratification of the pollen wall revealed marked differences as to the six subfamilies currently recognized. The most surprising pattern was found in the Aroideae, namely a complete lack of sporopollenin in the outer pollen wall layer. This pattern is not only characteristic of the most advanced subfamily of Araceae, but unique within angiosperms. It definitely requires functional explanation. The hypothesis was put forward (by the applicant M. Hesse) that the lack of sporopollenin is correlated with the presence and timing of elaborate trap mechanisms, e.g., by the short and precisely defined time of stigma receptivity and pollen release the plants were able to drop the costly production of sporopollenin and an elaborate exine. However, knowledge is still poor, especially regarding the temporal aspects of aroid traps. It is also poor regarding the evolutionary aspects, the diversification in space and time and the functional optimization of non-, semi- and perfect traps. While it is easy to construct a rough morphological series leading to perfect traps, e.g., the Arum type (and misleading to believe in an orthogenetic elaboration), the evolutionary development was without doubt a complex and branched process, with many sideways, one-ways and dead ends. Molecular data (and also pollen data!) have brought new insights regarding the relationsships of the taxa, but correlation to functional evolution is still in its infancy. Goals: The present research project intends a detailed investigation of the inflorescences of Araceae, focussing on functional and temporal aspects of anthesis, combined with further palynological studies (e.g., role of pollen surface coatings), and, as far as required, molecular analyses of taxa of special interest. It will test the above hypothesis that the pollen characters are correlated with the development, the evolution and the function of inflorescences and the timing of anthesis lapse. It will aim at a better understanding of the evolutionary steps and processes leading to perfect floral traps ("kettle traps") in Araceae. It will combine taxa evolution with character evolution and functional elaboration in order to reach a better understanding of some of the most complex and intriguing plant-animal-interactions found on earth. As Araceae are a predominantly tropical family it will provide better understandings of the biotic ecology of tropical ecosystems.

One aim of our study was to investigate the evolution of trap blossoms (kettle-traps) in the Araceae in a phylogenetic context. Besides the morphological structures we also studied the ecological context (i.e. pollinators) in which trap blossoms evolved. We found that trap blossoms have evolved at least 10 times independently in the family of Araceae. One reason for the high abundance of traps is that the precondition for the evolution of kettle-traps a floral chamber formed by the spathe - already evolved in the early diverging lineages of the family. Later-diverging lineages show a higher degree of synorganization in the inflorescence and thus traps have a more complex bauplan in this clades. The multiple evolution of trap pollination in different clades is correlated with pollination by flies, not beetles. Already the common ancestors of lineages with traps were pollinated by flies. We suppose that they had a mutualistic relationship with drosophilid flies which used the inflorescences as a brood site. In rewarding taxa of Colocasia preadaptations for trapping insects such as spathe movements and papillate epidermal cells have been found. Field observations and investigations of the cell ultrastructure indicate that these papillate epidermal cells are not slippery and serve for odor production only. We postulate that such surfaces represent an ancestral condition from which slippery surfaces evolved. A detailed study of slippery surfaces in the genus Arum showed that the general bauplan of the trap was very uniform in the different species, indicating a stable evolutionary condition of such trap mechanisms. However, the size and shape of papillate cells differed significantly in the various species, indicating an adaptation to the physical properties of the particular pollinating fauna. Pollen of the common ancestor of subfamily Aroideae most probably lacked an elaborated sporopolleninous exine. There is no correlated evolution between traps and the loss of sporopollenins, as the Aroideae-traps evolved later, irrespectively of an elaborated sporopolleninous exine. We found that both pollen characters, the type of ornamentation and bi- or tricelled pollen, are correlated with the systematic position within the subfamily Aroideae and the pollinators (flies, beetles). Tricelled echinate pollen is typical for the higher, fly pollinated Aroideae clades, whereas bicelled psilate pollen is typical for the lower, beetle pollinated Aroideae clades. Therefore most probably the ornamentation type and the number of cells are determinated genetically as well as ecologically (beetle pollination is dominant in the lower clades, fly pollination is dominant in the higher clades).