P1. Syst. Evol. 180:65-75 (1992) --Plant

Systemat lcs and

Evolution © Springer-Verlag 1992 Printed in Austria

Hydrophilous pollination and reproductive morphology in the seagrass Phyllospadix scouleri ( Zosteraceae)

P. A. Cox, P. B. TOMLINSON, and K. NIEZNANSKI

Received June 14, 1990; in revised version October 24, 1991

Key words: Marine angiosperms, Zosteraceae, Phyllospadix scouleri.- Abiotic pollination, dioecism, hydrophilous pollination, submarine pollination, surface pollination, seagrasses.

Abstract: In the seagrass Phyllospadix scouIeri (Zosteraceae) floral aggregates (spadices) occur on short lateral axes produced by subordinate vegetative shoots. The filamentous pollen is both dispersed on the surface of the sea as well as below the surface. In surface pollination, snowflake-like search vehicles (pollen rafts) float and collide with any rigid female inflorescences that emerge from the water surface. In submarine pollination, collinear bundles of pollen grains are dispersed. Analysis of seed set indicates ovule position within the inflorescence to affect likelihood of fertilization in submarine, but not surface-pollinated inflorescences. Agamospermy appears to be unlikely, but the strongly female-biased shoot sex ratio remains to be explained.

As defined by DELP1NO & AS CHERSON (1871), KNUTH (1898), ERNST-SCHWARZEN- BACH (1944), and Cox (1988) hydrophilous pollination involves pollen which is transported using water as a direct vector or as a vector for the transportation of pollen conveyances, such as the flowers of Vallisneria. Ecologically, hydrophilous pollination systems can be divided into three ecological categories (ERNST-SCHWAR-

ZENBACH 1944, COX 1988): category 1, in which pollen is transported above the water surface (example: Vallisneria spiralis); category 2, in which pollen is trans- ported on the water surface (example: Thalassodendron ciliatum, Cox 1990), and category 3, in which pollen is transported beneath the water surface (example: Thalassia lestudinum, Cox & TOMLINSON 1988). However, not all species fit within a single category. For example, the pollen of Enhalus acoroides (Hydrocharitaceae) is transported directly by the water surface (Category 1) as well as by floating male flowers (Category 2) (TROLL 1931, Cox 1988).

The seagrass genus Phyllospadix includes 5 species found along the coasts of Japan and the western coast of North America (DEN HARTOG 1970). The pollination mechanism of Phyllospadix torreyi was investigated by DUDLEY (1893, 1894) who reported that at low tide the anthers dehisce and the filamentous pollen grains spring "immediately to the surface of the water, while the filaments repel one another sufficiently to form at once a silvery arachnoid film, perhaps a centimeter

66 P.A. Cox & al.:

across" (DUDLEY 1893: 412). DUDLEY later described the pollen of P. torreyi as "floating on the surface of the sea when first escaping" (DUDLEY 1894: 384) but he also believed submarine pollination to be possible.

Unlike the other two genera of the Zosteraceae, Zostera, and Heterozostera, Phyllospadix is dioecious. DUDLEY (1893) reported female plants to greatly out- number male plants in P. torreyi and believed that paucity of male plants may result in a lower rate of pollination of female plants on exposed shores. Female plants also predominate in the flowering collections of P. scouleri, P. serrulatus, and P. torreyi as reported by PmLLIPS (1979), although the paucity of flowering male herbarium material may partially be an artifact of the relatively ephemeral male inflorescences. The mechanism of sexual determination is also unknown; multiple sex chromosomes were reported for two species of Phyllospadix by HARADA (1944) although STEWART & RUDENBER~ (1980) could not find sex chromosomes in P. torreyi.

Given the unique occurrence of P. scouleri on surf-exposed rocks, and the importance of this species in intertidal communi ty structure (PmLLn'S 1979, TURNER & LUCAS 1985), the relationship of the vegetative and reproductive morphology of P. seouleri to its pollination ecology is of interest. We therefore designed a combined morphological and ecological study o fP . scouleri in order to more clearly understand the relationship of the pollination ecology of this species to its repro- ductive morphology.

Material and methods

A population of Phyllospadix scouleri HooK. was studied at the coast of Pacific Grove, California, U.S.A. immediately south of Monterey. The study period, May 4-11, 1989 was chosen to correspond with the period of spring tides, since flowering times of many seagrasses is correlated with spring tides (Cox 1988).

An attempt was made to observe all stages of pollination under field conditions (Cox & KNox 1988), including release of the pollen onto the vector, transportation of the pollen by the vector, and deposition of the pollen by the vector onto the stigma. All pollination events were confirmed by three independent observers. Twenty-five pollen rafts, hereafter termed as "search vehicles" in accordance with the nomenclature of search theory (Cox 1983), were collected from the water surface on glass microscope slides, dried, and their maximum and minimum diameters measured. Seed set of 14 inflorescences collected 1 m below the lowest lower low water (LLW) (DoTY & ARCHER 1950) and 30 inflorescences to collected above the low tide level [i.e. those that had been exposed to air during lower low water (LLW)] was recorded on a per inflorescence basis. Fertilization data based on ovule position within each spadix was recorded for all 14 submarine inflorescences and for 8 of the inflorescences collected above LLW.

Shoot sex ratio was estimated in two ways. First, the intertidal population was sampled with rectangular quadrats of 23 x 20 cm in dimension placed at 1 m intervals in a 10 m x 5 m grid; the short axis of the grid was oriented to north with the grid positioned 5 m seaward from the low tide mark on shore. In each quadrat the number of male and female shoots were counted. Second, since Phyllospadix scouleri plants mainly occurred in clumps on the tops of rocks, 73 rocks were recorded. The number of male, female, or both male and female inflorescences was recorded for each rock.

Male and female inflorescences in various stages of development were collected and preserved in FAA. The specimens were then critical-point dried prior to coating with gold amalgam and subsequently studied with scanning electron microscopy. Herbarium vouchers

Hydrophilous pollination in Phyllospadix 67

of flowering specimens of both sexes (male-Cox 1418; female-Cox 1417) were made and deposited in the herbaria of Brigham Young University (BRY) and duplicates in the Gray Herbarium of Harvard University (GH).

Results

Reproductive morphology. Phyllospadix scouleri is an unusual seagrass because it grows on rocks, rather than on sandy or silty substrates (Fig. 1 a), and this is

Fig. 1. Phyllospadix scouleri, a Plants exposed at low tide, Pacific Grove, California. b Exposed female spadices, c Pollen rafts or search vehicles on water surface, d Pollen search vehicle adhering to exposed stigmas. Bar: 2 mm

68 P.A. Cox & al.:

reflected in the vegetative morphology. Axes are rhizomatous and closely adherent to the smooth substrate by means of short adventitious roots. These produce very fine, short root-hairs that penetrate minute irregularities of the substrate surface. Rhizomes have short internodes and are asymmetrical in section, being more or less elliptical but with the lower surface flattened to conform to the rock texture. Dorsiventrality is more pronounced than in most seagrasses; leaves are distichous, with the plane of distichy parallel to the substrate surface, and the branches also extend horizontally. In vigorous shoots (diameter > 3 ram) each node supports a lateral branch. The branches are extra-axillary since they occur below the node beyond the axillary leaf. This intercalary displacement from a strict axillary position seems comparable to that found in other Zosteraceae (cf. TOMLINSON 1982: pl. 6). Adventitious adherent roots arise as a root complex on the upper part of the internode opposite the vegetative branch, with roots in two ranks up to a maximum of 5 pairs. Root number is correlated with internode diameter, and the narrowest internodes may bear no roots.

Rhizomes may be described as monomorphic since they uniformly support foliage leaves and are monopodial. There is, however, a clear differentiation between major and minor axes, since lateral shoots remain suppressed and produce leaves at a slower rate than leading axes. This results in an apparent dimorphism between "long-shoots" (major axes) and "short-shoots" (minor, lateral axes). Major axes have internodes in the range 3-8 mm long and of comparable diameter; minor axes have short, narrow internodes (up to 1 mm long, 1-33 mm wide) often smaller leaves, fewer or no roots, and are more cylindrical (ToMLINSON 1982: fig. 13.4). Lateral axes remain suppressed relative to parent axes, but can substitute oppor- tunistically for damaged or broken leaders. TURNER (1985) described growth of Phyllospadix scouleri as "slow" with a maximum rate of increase in rhizome length of 6 cm/yr. She also observed colonization from seedlings, but as a rare event. Since there is frequent overgrowth of adjacent rhizomes, demographic analyses are dif- ficult.

Nevertheless, reproductive axes seem to be produced seasonally, as reported by DUDLEY (1894). In the population we studied, inflorescences occupy the position of vegetative shoots but are frequently restricted to the base of those minor first- order vegetative branches which function as "short-shoots" in the monopodial system. In the species, as reported by DUDLEY (1893, 1894) the inflorescence axes are short compared with P. torreyi. They include a broad, sheathing prophyll, two or more short alternate or subopposite foliage leaves and 1-3 flower-bearing units (spadices). The axis ends in such a unit, but one or two spadices may be produced on second-order axes subtended by the reduced foliage leaves. The large aggregate of spadices (rhipidia), found in the more elaborately branched elongate inflores- cences of P. torreyi, do not occur in P. scouleri. Individual spadices (Figs. 2, 3) have a flattened, truncate axis enclosed by the sheath of the foliage leaf immediately below (often described as a spathe). Flowers always project adaxially with respect to the spathe (Figs. 2 b, 3 c). Each margin of the spadix supports an alternating series of truncate scales (retinacules) that initially enclose the flowers (Figs. 2 a, 3 a). Male spadices are represented by about 17-20 pairs of bilocular anthers, with those of opposite sides interdigitating (Fig. 2 c). Female spadices support an equivalent number of solitary carpels that form two rows, with carpels of each row interdi-

Hydrophilous pollination in Phyllospadix 69

Fig. 2. Phyllospadix scouleri, details of male spadix, a Young spadix (spathe removed), retinacules enclosing anthers, bar: 1 mm. b Retinacules erected to form a marginal palisade, bar: 5 mm. c Retinacules diverged, bar: 5 mm. d Anther dehiscence and pollen release commencing at base of spadix, bar: 1 mm

gitating (Fig. 3 a). Each flower is represented by a single uniovulate carpel, basally hastate, and with one style that bifurcates distally to form two tapering stigmas

70 P .A. Cox & al.:

Fig. 3. Phyllospadix scouleri, details of female spadix, a Spadix at time of pollination with carpels enclosed by retinacules but stigmas protruding, bar: 1 ram. b Curved spadix sub- sequent to time of pollination, bar: 1 cm. c Pollinated stigmas beginning to shrivel, reti- nacules removed, bar: 1 mm. d Maturing spadix, the retinacules forced apart by expanding fruits, stigmas abscised, bar: 3 mm

Hydrophilous pollination in Phyllospadix 71

Fig. 4. Phyllospadix scouleri, details of pollen and stigmas, a Flattened tips of individual pollen grains, bar: 10 gin. b Pollen grains, bar: 100 gin. c Pollen grains adherent to submerged stigma, bar: 100 gin. d Carpel, showing styles and stigma pair

72 P.A. Cox & al.:

(Fig. 4 a). Female spadices include a minute vestigial pair of staminodia between each pair of carpels that seem less conspicuous than those described by DUDLEY (1894) for Phyllospadix torreyi since he reported the existence of occasional pollen within them. Male spadices lack pistillodia. We are not concerned here with the comparative morphology of the reproductive structures in terms of conventional flowers; the various interpretations are summarized in TOMLINSON (1982).

NO phenological observations were made but periodicity of flowering is sug- gested by the development of inflorescences in series on laterals at the mid-portion of vigorous rhizomes which lack inflorescences on younger parts.

Pollination ecology. We found Phyllospadix scouleri to have a mixed mode of pollination, i.e. with pollen dispersed both on and beneath the water surface. In the intertidal population, which is exposed at low tide (Fig. 1 a), pollination on the surface appears to predominate. Male inflorescences appear to move through 4 distinct phases of maturity:

1) Prior to anthesis, the inflorescences of the male plants are held rigid at right angles to the surface of the water. The retinacules are folded over the inflorescence, protecting the anthers from exposure (Figs. 2 a, 5 a).

Fig. 5. Phyllospadix scouleri, reconstruction of pollination events. A Immature male spadix with retinacules appressed. B Mature male spadix with retinacules divergent, anthers de- hisced, and pollen released. C Surface pollen aggregated in rafts or search vehicles. D Pollen carried on the water surface to female spadix. E Submerged pollen carried to submerged spadices. F Pollen capture by submerged spadices. G Pollen capture by emerged stigmas, the pollen raft "painted" up and down the spadix by wave action

Hydrophilous pollination in Phyllospadix 73

2) The bract subtending the inflorescence recurves, and the retinacules begin to open in acropetal order until they stand erect at the margins of the inflorescence collectively forming a palisade around the anthers (Fig. 2 b).

3) The retinacules continue to open until they lie within the plane of inflorescence exposing intertidal inflorescences to the full force of the waves (Fig. 2 c).

4) The anthers dehisce (Fig. 2 d), releasing the noodle-like pollen (Figs. 4 a, 4 b) into the water as the waves move up and down the rigid inflorescence.

Female inflorescences appear to move through 3 phases of maturity: 1) The filiform stigmas (Fig. 4 d) penetrate the folds (Fig. 3 a) of the two bracts

enveloping the stigma. 2) The bract subtending the inflorescence recurves (Fig. 3 b). 3) After pollination the stigmas wither (Fig. 3 c), and fruit development proceeds

(Fig. 3 d). In intertidal populations, most of the pollen is released on the surface (Fig. 2 c)

where it floats, forming search vehicles (Figs. 1 c, 2 d, 5 c, 5 d) which resemble snowflakes. Twenty-five floating search vehicles were collected on microscope slides from the water surface and found to have a mean large diameter of 4.5 mm (s.d. = 2.15 ram). Some pollen is released, however, underwater. Although the pol- len is neutrally buoyant, it will form a floating search vehicle if it encounters the water surface (Fig. 5 b). Otherwise it is dispersed beneath the water surface (Fig. 5 e) in clusters with the pollen "noodles" parallel to each other. Pollination on the surface occurs by collision of a floating search vehicle with the stigmas (Fig. 1 d); since the female inflorescence is held rigid and extends above the plane of the water surface, a single search vehicle can fertilize all of the stigmas as shallow waves "paint" the pollen up and down the inflorescence. We did not witness submarine pollination, but found circumstantial evidence for the submarine transfer of pollen. This indication is the dioecious condition of P. scouleri coupled with SEM analysis which reveals pollen on stigmas collected from beneath the level of the lowest tide (Fig. 4 c).

Seed/ovule ratio of the 14 inflorescences collected 1 m beneath maximum low tide had a mean of 0.66 (s.d. = 0.28) while those collected above the mean tide level had a mean seed/ovule ratio of 0.62 (s.d. = 0.17). These mean seed/ovule ratios are not significantly different as shown by a two tailed t-test for unequal variances for the data which were transformed with an arc sine transformation (t = 1.519, p < 0.145). However, 50% of the inflorescences collected below the LLLW showed significant position effects at p < 0.05 for seed development as shown by Wald-Wolfowitz runs tests for sequence randomness, while only 12.5% of the inflorescences collected above the maximum low tide level showed such a difference. In the subtidal inflorescences, ovules located at either end of the inflo- rescence were less likely to be fertilized than ovules in the center of the inflorescence.

Given the non-random pattern of seed set, ranging from 0 to 80% within a single inflorescence, we can tentatively exclude obligate agamospermy in P. scouleri, but facultative agamospermy remains a possible, but unlikely alternative to sexual reproduction in the species. Given the preponderance of female plants, further studies of seed sex ratio, recruitment, and survivorship of the two sexes are needed.

Chi-square tests showed shoot sex ratios departed significantly from unity. The shoot sex ratios, as measured both in the grid (male/female = 0.07) and on rocks

74 P.A. Cox & al.:

(male/female = 0.36) were strongly female-biased. In the grid, female shoot density was 0.21/m 2 while male shoot density was 0.012/m 2.

Discussion

Although Phyllospadix scouleri plays an important role in the intertidal communities of Western North America (TURNER 1985, TURNER & LUCAS 1985), surprisingly little investigation has been made of its reproductive biology since the pioneering studies of DUDLEY (1893, 1894) nearly a century ago. Our observations indicate that Phyllospadix scouleri has a mixed mode of pollination, with pollen dispersed both on and below the water surface. In this regard, its pollination biology is very similar to Zostera marina which also produces floating search vehicles as predicted by search theory (Cox 1983), and yet is also capable of submarine pollination (DE Cock 1980). Despite the supposed inefficiencies of abiotic pollination systems (FAEGRI ~; VAN DER PIJL 1973), and the extremely low percentage of male plants in the population, Phyllospadix scouleri has an high seed set. The combined mean seed/ovule ratio of 0.63 for P. scouleri is far greater than the mean seed/ovule ratio of terrestrial perennial outcrossing species, 0.49 (WlENS 1984). In general P. scouleri does not appear to be pollen limited, although there is high variance of seed set in subtidal, but not intertidal, female inflorescences. However, mean reproductive success appears to be nearly equal in both intertidal and sub-tidal individuals. Postion of an ovule within the spadix appears to be far more likely to affect probability of fertilization in subtidal than in intertidal female plants. This is probably due to the "painting" effect we observed in the intertidal population in which all of the stigmas of a female plant can be pollinated by a single search vehicle.

For surface pollination to occur, phenology of P. scouleri must be related to low spring tides. PHILLIPS (1979) found flowering of P. scouleri and P. torreyi to occur during April in populations ranging from S. Oregon to Sitka, Alaska. Phyl- lospadix torreyi has a similar flowering pattern in S. California (STEWART & RU- DENBERG 1980). Isozyme patterns from P. scouleri plants on Monterey Peninsula indicate a possibility of some hybridization of P. scouleri with the sympatric species P. torreyi (McMILLAN • PHILLIPS 1981). P. scouleri occurs in shallower water than P. torreyi (TURNER & LUCAS 1985). This habitat difference combined with the different morphologies of the two species may result in slightly different mechanisms of pollination: "painting" an upright, rigid female inflorescence which extends above the water surface with a surface-borne pollen raft (as we observed in P. scouleri) is scarcely possible for the lax inflorescence of P. torreyi. Surface pollination in P. torreyi, which was reported by DUDLEY (1894), may occur in a manner similar to that of Zostera marina and thus the pollination mechanisms of putative hybrids would also be of interest.

This study was supported by National Science Foundation Grant BSR-8452090. We thank STEVE ELIASON and REBECCA SPERRY for assistance in data collection.

References

Cox, P. A., 1983: Search theory, random motion, and the convergent evolution of pollen and spore morphologies in aquatic plants. - Amer. Naturalist 121: 9-31.

Hydrophilous pollination in PhylIospadix 75

- 1988: Hydrophilous pollination. - Ann. Rev. Ecol. Syst. 19: 261-280. - 1990: Hydrophilous pollination of a dioecious seagrass, Thalassodendron ciliatum (Cy-

modoceaceae), in Kenya. - Biotropica 22: 259-265. - KNox, R. B., 1988: Pollination postulates and two-dimensional pollination in hydro-

philous monocotyledons. - Ann. Missouri Bot. Gard. 75: 811-818. 1989: Two-dimensional pollination in hydrophilous plants: convergent evolution in

the genera Halodule (Cymodoceaceae), Halophila (Hydrocharitaceae), Ruppia (Ruppi- aceae), and Lepilaena (Zannichelliaceae). - Amer. J. Bot. 76: 164-175.

-- TOMLINSON, P. B., 1988: The pollination ecology of a Caribbean seagrass, Thalassia testudinum (Hydrocharitaceae). - Amer. J. Bot. 75: 958-965.

DE CocK, A. W. A. M., 1980: Flowering, pollination, and fruiting in Zostera marina L. - Aquatic Bot. 9: 202-220.

DELPINO, F., ASCH~RSON, P., 1871: Federico Delpino's Eintheilung der Pflanzen nach dem Mechanismus der dichogamischen Befruchtung und Bemerkungen fiber die Befruch- tungsvorgfinge bei Wasserpflanzen. Mitgetheilt und mit einigen Zus/itzen versehen von P. Ascherson. - Bot. Zeitung (Berlin) 29: 443-445, 447-459, 463-467.

DEN HARTOG, C., 1970: The seagrasses of the world. - Verhandl. Kon. Nederl. Akad. Wetensch. Nat. 59 (1). - Amsterdam: North Holland.

DorY, M. S., ARCHER, J. G., 1950: An experimental test of the tide factor hypothesis. - Amer. J. Bot. 37: 458--464.

DUDLEY, W. R., 1893: The genus Phyllospadix. - In The Wilder Quarter-Century Book, pp. 403-418. - Ithaca, New York: Comstock.

- 1894: Phyllospadix, its systematic characters and distribution. - Zoe 4: 381-385. ERNST-SCHWARZENBACH, M., 1944: Zur Blfitenbiologie einiger Hydrocharitaceen. - Ber.

Schweiz. Bot. Ges. 54: 33-69. FAEGRI, K., VAN DER PIJL, L., 1979: The principles of pollination ecology. 3rd edn. -

Oxford: Pergamon Press. HARADA, I., 1944: Sex chromosome of Phyllospadix. - Japan J. Genet. 20: 127-128. KNUTH, P., 1898: Handbuch der Blfitenbiologie. Band I, II. - Leipzig: W. Engelmann. MCMILLAN, C., PHILLIPS, R. C., 1981: Morphological variation and isozymes of North

American Phyllospadix (Potamogetonaceae). - Canad. J. Bot. 59: 1494-1500. PHILLIPS, R. C., 1979: Ecological notes on Phyllospadix (Potamogetonaceae) in the Northeast

Pacific. - Aquatic Bot. 6: 15%170. STEWART, J. G., RUDENBEP, C, L., 1980: Microsporocyte growth and meiosis in Phyllospadix

torreyi, a marine monocotyledon. - Amer. J. Bot. 67: 949-954. TOMLINSON, P. B., 1982: Helobiae (Alismatidae) VII. - In METCALFE, C. R., (Ed.): Anatomy

of the Monocotyledons. - Oxford: Clarendon Press. TROLL, W., 1931: Botanische Mitteilungen aus den Tropen II. Zur Morphologie und

Biologie yon Enhalus acoroides (LINN. f.) RICH. -- Flora (Jena) 125: 427-456. TURNER, T., 1985: Stability of rocky intertidal surfgrass beds: persistence, preemption, and

recovery. - Ecology 66: 83-92. - LUCAS, J., 1985: Differences and similarities in the community roles of three rocky

intertidal surfgrasses. - J. Exper. Mar. Biol. Ecol. 89: 175-189. WIENS, D., 1984: Ovule survivorship, brood size, life history, breeding systems, and re-

productive success in plants. - Oecologia 64:47 53.

Addresses of the authors: PAUL ALAN COX, KEVIN NIEZNANSKI, Department of Botany and Range Science, Brigham Young University, Provo, Utah 84602 U.S.A. - P. B. TOM- LINSON, Harvard Forest, Harvard University, Petersham, Massachusetts 01366 U.S.A.

Accepted November 25, 1991 by V. HEYWOOD