Northeastern Naturalist Vol. 22, No. 3 C.T. Philbrick, P.K.B. Philbrick, and B.M. Lester 2015 643 2015 NORTHEASTERN NATURALIST 22(3):643–647 Root Fragments as Dispersal Propagules in the Aquatic Angiosperm Podostemum ceratophyllum Michx. (Hornleaf Riverweed, Podostemaceae) C. Thomas Philbrick1,*, Paula K.B. Philbrick2, and Brandon M. Lester1 Abstract – Most aquatic flowering plants spread via specialized vegetative propagules. Such propagules do not occur in Riverweeds (Podostemaceae), which constitute the largest family of strictly aquatic flowering plants. This study was undertaken to test whether root fragments of Podostemum ceratophyllum Michx. (Hornleaf Riverweed) can reattach and thereby serve a dispersal role. In field experiments, root fragments re-attached with tenacity sufficient to challenge removal. We conclude that fragments can provide a vegetative means to disperse plants in rivers. While the plant tested is the only species of the family in North America, the results have broader implications for the lar gely tropical Podostemaceae. Introduction Aquatic flowering plants comprise a diverse assemblage of evolutionary lineages. Even so, a number of shared traits are common and have evolved independently many times in the various groups. One such trait is reduction in sexual reproduction in favor of vegetative means of propagation (e.g., Cronk and Fennessy 2001, Hutchinson 1975, Philbrick and Les 1996, Sculthorpe 1967). Individual plants spread clonally, and specialized vegetative structures, including modified roots, stems, or buds, are important in dispersal to new locations (Philbrick and Les 1996 and references therein). It is notable that the largest family of aquatic flowering plants, Podostemaceae, is an apparent exception. Species within that family are highly sexual but lack specialized propagules for dispersal. The species can produce millions of seed per square meter (e.g., Philbrick and Novelo 1997). Notwithstanding, clonal growth is common in the family and accounts for local expansion of plants, but dispersal through vegetative propagules is unknown. However, herein we report that root fragments have that potential. We use Podostemum ceratophyllum Michx. (Hornleaf Riverweed, hereafter Riverweed) as a model organism. This species is the only temperate representative of the Podostemaceae in North America and is common regionally. Riverweeds are native to the regions where they occur and play important ecological roles (e.g., Hutchens et al. 2004). They are unusual angiosperms in that they grow on rocks in the turbulent current of river rapids and waterfalls where firm at- 1Department of Biological and Environmental Sciences, Western Connecticut State University, Danbury CT 06810. 2Department of Ecology and Evolutionary Biology, University of Connecticut-Waterbury Campus, Waterbury, CT 06702. *Corresponding author - philbrickt@wcsu.edu. Manuscript Editor: Glenn Motzkin Northeastern Naturalist 644 C.T. Philbrick, P.K.B. Philbrick, and B.M. Lester 2015 Vol. 22, No. 3 tachment is crucial (e.g., Philbrick and Novelo 2004). Roots grow prostrate across the rock surface, with stems arising from them. These plants attach tenaciously— holdfasts on the underside of roots and the base of upright stems achieve such firm attachment that plants are best removed by scraping between the rock and holdfast with a knife. Riverweed’s biology is closely tied to seasonality in the water level. Plant growth occurs when the water level is high and plants are inundated (Philbrick and Novelo 2004) and flowering occurs as water levels drop, and plants are exposed. Seed production is high for most species, but little is known about the effective seed dispersal range and recruitment rates (Philbrick and Novelo 200 4). The current study is an elaboration of research conducted by Hammond (1936), who reported the ability of roots, stems, and leaves of Rivereweed to regenerate after they had been removed from rocks and placed in petri dishes or on wire screens sprayed with water. Hammond’s studies, however, did not address whether vegetative structures could reattach to the substratum during the process of regeneration. Our focus was on roots because they play a central role in plant attachment. The purpose of this study was to assess whether detached root fragments can reattach in the natural habitat. If they could, then fragments have the potential to serve as propagules of dispersal and gene flow. Field Site Description We conducted our field studies in the Pootatuck River, about 200 m downstream of the Church Hill Road bridge, Newtown, CT (41.421225°N, 73.283358°W, 80 m amsl). At this location, the river bottom is covered with granite stones and outcrops, and Riverweed is abundant. Methods On 15 June 2013, we gathered 5 rocks with extensive growth of Riverweed from the river. We collected and cut 15 linear roots into 4-cm lengths and removed all shoots and leaves. The roots were out of the water for 30–45 minutes. We built 5 brick-tile assemblies to secure the root fragments. A red fired brick (19 cm x 9.5 cm x 4.5 cm) served as a platform for each. We drilled 2 pairs of 7-mm-diameter holes on each brick and used hydraulic cement (UGL Drylok Fast Plug©, United Gilsonite Laboratories, Scranton, PA) to secure a threaded 8.9-cm stainless steel bolt (6.25 mm diameter) in each hole. We employed 15 cm x 2.5 cm x 1.5 cm granite or porcelain tiles as substrata for root attachment. Next, we drilled two 7-mm holes in Plexiglas® strips (12 cm x 1.5 cm x 0.6 cm) so that they could be placed on the tiles with the bolts projecting through the holes. We transferred the root fragments to the brick–tile assemblies as follows. First, we placed a granite tile on the brick parallel to the long axis and equidistant between the pair of bolts at each end and laid down fifteen 4-cm root fragments along the tile perpendicular to its long axis. We put a porcelain tile on top of the root fragments, aligned 1 Plexiglass® strip onto each pair of bolts, and affixed it in place with hexagonal nuts. We tightened the nuts against the Plexiglass strip to hold the root Northeastern Naturalist Vol. 22, No. 3 C.T. Philbrick, P.K.B. Philbrick, and B.M. Lester 2015 645 fragments in place. We returned the brick–tile–root assemblies to the river a few meters downstream from where the original rocks with plants had been removed and lodged them between rocks with about 3 dm of water flowing o ver them. We recovered the brick–tile–root assemblies from the river on 28 July 2013, removed the tiles, and collected data. For each root fragment that remained, we noted: the substratum type to which the fragment attached (granite, porcelain, red brick); total root-fragment length; and number of holdfasts per cm of root length. For comparison, we selected 5 rocks upon which plants were naturally growing and assayed fifteen 3–6-cm-long roots from each for the same 3 variables measured on the experimental root fragments. Results Of the 75 root fragments initially placed in the brick–tile assemblies, 41 reattached. The remaining 34 root fragments washed away. The number of root fragments that attached to granite or porcelain tiles on each brick ranged from 0–5 and 0–14, respectively. Some roots also attached to the brick and/or a stainless steel bolt. There was considerable variation among the number of roots that attached to the substrate types, but at least some fragments attached to each. The original design was confounded by the varying number of roots that had washed away, which discounted any meaningful statistical comparison. Therefore, we present only combined results. The root fragments that attached to our experimental substrates all grew over the 6-week period. New growth ranged from 1–7 cm (mean = 2.9, SD =13.2), and most bore shoots and leaves. The number of holdfasts per cm of root length ranged from 0.4–16 (mean = 3.3, SD = 2.8) for roots attached to tiles. The number of holdfasts per cm of root on naturally growing plants ranged from 1.4–16 (mean = 5.4, SD = 3.6). Roots growing on naturally occurring rocks had more holdfasts per cm of root length than those growing on tiles (2-sample t-test; P > 0.05). Discussion Riverweed root fragments can reattach to substrates. Over half of the root fragments used in the study did so during the 6-week experiment. Notably, naturally growing plants attached to rock with more holdfasts per cm of root than our experimental fragments. This result is not surprising because they were physically held in place and had more time to grow. That our experimental fragments attached to tile, in addition to brick and adjacent rock, requiring that they be scrapped off with a knife for assay, demonstrated that they had securely attached in swift current. Hammond (1936) was the first to show that fragments of Riverweed can regenerate; we demonstrated that root fragments can become re-established on several substrates. Our results lend support to the hypothesis that root fragments serve as dispersal propagules in Riverweed. Did the conditions of the study reflect what happens in nature? We propose that the conditions under which our study was done represent an applicable model of Northeastern Naturalist 646 C.T. Philbrick, P.K.B. Philbrick, and B.M. Lester 2015 Vol. 22, No. 3 natural phenomena. Hand removal of root fragments from rocks and lodging them between solid surfaces mimics what likely happens naturally in river rapids. Riverrapids are characterized by haphazard arrangement of rocks of varying sizes and shapes, resulting in numerous spaces between rocks where debris becomes trapped. In addition, water currents are strong and water levels periodically rise and fall. When the latter happens, plants die back and fragments are washed away. Such evidence, albeit circumstantial, lends support to the idea that vegetative fragments of Riverweed are released into the water current. Empirical support for this idea is not yet available. There would certainly be logistic challenges associated with documenting such stochastic events as the dispersal of plant fragments, especially in river-rapid habitats. Results from our simple experiment have implications for understanding the biology of tropical riverweeds. Most aquatic angiosperms produce vegetative structures (e.g., buds, stems, roots) specialized for dispersal and establishment of new individuals (Philbrick and Les 1996 and references therein). For members of the Podostemaceae, the lack of such specialized structures has supported the contention that vegetative propagules play no role in dispersal. In fact, our results support the hypothesis that root fragments are specialized enough to serve a dispersal role. The strategic elements of such a scenario are present. Whether this mechanism happens in nature, and to what degree, remains to be assessed. The only riverweed in temperate regions of the Americas is Hornleaf Riverweed; the other ~135 species are tropical (Philbrick et al. 2010). The temperate species serves as an appropriate model organism to investigate a range of biological issues that are equally applicable to its tropical relatives. Herein we provide the first evidence that root fragments have the potential to serve as propagules of dispersal and re-establishment. Given that the closest relatives of Horned Riverweed occur in the tropics (Philbrick and Novelo 2004), this phenomenon is likely applicable to tropical species as well. The role of fragments in plant dispersal within tropical rivers has implications for understanding the spread of species. Acknowledgments This study was supported by National Science Foundation Grant (DEB-0444589) and Connecticut State University-AAUP research grants to C.T.P. Literature Cited Cronk, J.K., and M.S. Fennessy. 2001. Wetland Plants: Biology and Ecology. Lewis Publishers, Boca Raton, LA. 482 pp. Hammond, B.L. 1936. Regeneration of Podostemum ceratophyllum. Botanical Gazette 97:834–845. Hutchens, J.J., J. Bruce, and E.D. Ramaniszyn. 2004. Role of Podostemum ceratophyllum Michx. in structuring benthic macroinvertebrate assemblages in a southern Appalachian river. Journal of the North American Benthological Society 23:713–727. Hutchinson, G.E. 1975. A Treatise on Limnology. Vol. 3: Limnological botany. John Wiley and Sons, New York, NY. 660 pp. Northeastern Naturalist Vol. 22, No. 3 C.T. Philbrick, P.K.B. Philbrick, and B.M. Lester 2015 647 Philbrick, C.T., and D.H. Les. 1996. Evolution of aquatic angiosperm reproductive systems. Bioscience 46:813–826. Philbrick, C.T., and A. Novelo R. 1997. Ovule number, seed number, and seed size in Mexican and North American species of Podostemaceae. Aquatic Botany 57:183–200. Philbrick, C.T., and A. Novelo R. 2004. Monograph of Podostemum (Podostemaceae). Systematic Botany Monographs 70:1–106. Philbrick, C.T., C.P. Bove, and H.I. Stevens. 2010. Endemism in Neotropical Podostemaceae. Annals of the Missouri Botanical Garden 97:425–456. Sculthorpe, C.D. 1967. The Biology of Aquatic Vascular Plants. St. Martin’s Press, New York, NY. 610 pp.