A Study Of Adaptive Significance Of Long Corolla Tubes In Pedicularis Species

The Pedicularis species display striking variation in corolla morphology especially the corolla tube length which varies from less than 10 mm to greater than 120 mm. To explore the effects of the physical enviromnental factors on the evolution of corolla morphology especially the corolla tube length, the seedlings of long-tubed Pedicularis siphonantha and P. tricolor were transplanted in experimental populations in Shangri-La Alpine Botanical Garden. Our results showed that the light intensity, soil water content and their interaction significantly affected the four vegetative traits and six floral traits in the two species. Furthermore, phenotypic variation (lower mean CV values) in most floral traits was less than in vegetative traits in both two Pedicularis species. However, variation in the corolla tube length was greater than that in other floral traits and similar with variation in vegetative traits. Four vegetative traits and six floral traits were observed to be correlated respectively, meanwhile floral traits and vegetative traits also correlated both in two long-tubed Pedicularis species. These results suggested that the floral traits have less phenotypic variation than that of vegetative traits in response to abiotic environmental variation. The corolla tube was greatly elongated under the condition of low light intensity and high soil water content. Based on the striking difference of corolla tube length among the four treatment groups after transplanting the seedlings of P. siphonantha, we have examined several hypotheses related to the adaptive significance of long corolla tubes in Pedicularis.Firstly, the evolution of long corolla tubes has been hypothesized to be driven by long-tongued pollinators. Corolla tubes in Pedicularis species can be longer than 120 mm which may function as flower stalks to increase visual attractiveness to pollinators because these species provide no nectar and are pollinated by bumblebees. The corolla tube length was manipulated (longer) in Pedicularis siphonantha and (shorter) in P. tricolor in the field to examine whether longer tubes are more attractive to pollinators and produce more seeds than short tubes. Our results did not support the pollinator attraction hypothesis of the evolution of long corolla tube in Pedicularis.Secondly we test the hypothesis that the long pistil protected by a long corolla tube provided an opportunity for male competition when outcross conspeeific pollen and heterospecific pollen mixed deposited on the stigma simultaneously. P. siphonantha, P. cephalantha and P. densispica coexist and co-flower along much of their natural distribution range in the field. Interspecific pollen transfer between the three species has also been observed in the field. To test the pollen competition hypothesis of long corolla tube evolution, three hand-pollination experiments were conducted in the field, including self pollination (with pollen grains from flowers on the same plant), outcross pollination (with pollen grains from flowers of other individuals apart from at least 100 m) and mixed pollination (with pollen grains from outcross conspecific flowers mixed interference pollen grains from P. cephalantha and P. densispica for two consecutive years respectively). The seed set of long-tube group was significantly higher than that of short-tube group under mixed pollination but not under self and outcross pollination in P. siphonantha. Additionally, the pollen tube number under mixed pollination in the styles of long-tube group was significantly more than pollen tube number of short-tube group for 24-48 h pollen tube growth. In P. siphonantha, the pistil (style) could be served as an arena of race for conspecific outcross pollen outperforming the congeneric heterospecific pollen grains deposited on the stigmas simultaneously. Our results support the pollen competition hypothesis of the evolution of long corolla tube in Pedicularis.Thirdly, to examine the property of corolla tube length within and among Pedicularis species, we investigate how changes of internal cells accompany with changes of external organ shape under the influence of environments. We experimentally measured cell length of epidermis to investigate the roles of cell divisions and cell expansion controlled in determining the corolla tube morphology in 8 Pedicularis species. Our results indicated that the cell number and cell size were two crucial factors finally determining the organ size of plant. The corolla tube length was positively correlated with cell number and cell length among Pedicularis species. Contrary to a previous hypothesis of a discrete meristematic zone as the driver of organ growth, we found that the corolla tube of Pedicularis developed via anisotropic cell expansion which corolla tube length was significantly positively correlated with cell length intraspecific. The corolla tube length of P. siphonantha was greatly elongated by increasing soil water content and reducing light intensity. Furthermore, the longer corolla tube was also mainly caused by the anisotropic elongation of cell but not by the increase of cell number. Changes in cell anisotropy accounted for most of the corolla tube length variation in the genus, indicating that the real evolutionary innovation underlying the rapid radiation of Pedicularis corolla tube was due to a change of cell shape primarily.Finally, Darwin proposed that pollen size should be positively correlated with stigma depth rather than style length among species given that pollen tubes first enter the stigma autotrophically, then grow through the style heterotrophically. However, studies often show a positive relationship between pollen size and style length. Five floral traits were observed to be correlated among 42 bumblebee-pollinated Pedicularis species which stigmas are distinct from styles. Our phylogenetic independent contrast analysis revealed that pollen grain volume was more strongly correlated with stigma depth than with style length, consistent with Darwin's functional hypothesis between pollen size and stigma depth.