Awn length variation and its effect on dispersal unit burial of Trachypogon spicatus (Poaceae).

Trachypogon spicatus, formerly known as Trachypogon plumosus, is a dominant grass in some savannas of Northern South America. Its dispersal unit, like many other species of the Andropogoneae tribe, bears a hygroscopic awn which facilitates its establishment in favorable microsites. Some authors have previously proposed that there is a positive correlation between awn length and dispersal unit burial, and that this relationship increases the probability of seed survival in the event of a fire, since soil acts as insulator. In this study we experimentally tested this relationship for T. spicatus. A total of 192 diaspores were placed in randomized blocks, in aluminum trays filled with soil under greenhouse conditions. Diaspores were sprayed with water daily for a month to guarantee awn movement; on the last day of the experiment, they were sprayed with red aerosol paint to determine burial depth. The effects of awn length, presence of caryopses, and presence of a pivot for the passive segment of the awn on diaspore burial were evaluated. Germination viability was tested using a tetrazolium salt test for 35 caryopses. No significant differences in diaspore burial were observed between diaspores with and without caryopses (F(2,126) = 0.034, p=0.853). A positive correlation between awn length and diaspore burial was observed only if the passive awn lacked a pivot (r(66)=0.394, p<0.05). Diaspores whose awns had a pivot point achieved significantly deeper burial distances than their counterparts (F(2,126)=7.063, p=0.005). Viability test found that 0% of caryopses tested were able to germinate; this is possibly due to the time difference between sampling and testing. We considered the presence or absence of caryopsis as an important factor, since previous studies have not yet considered it and the high production of sterile diaspores in grasses. These results suggest that the physical mechanism behind T. spicatus diaspore burial is awn torque. This would explain why our results partially support the hypothesis previously proposed by other authors regarding the effect of awn length upon dispersal unit burial, but would also account for the differences observed in burial depth between diaspores with an available pivot point for the passive awn and diaspores lacking thereof.