THE TAXONOMY OF "SEAGRASSES" SURVEYED FROM THE HIGHER TAXA DOWN THROUGH TO THE FAMILY LEVEL

Florida International University

December 1994

ABSTRACT

There is no one formally accepted system of taxonomy that fully describes the twelve recognized genera of "true seagrasses". This may be the result of two factors: a) "seagrasses" are inherently a polyphyletic group, and therefore the 12 genera do not represent a 'natural" taxon, b) there is wide spread disagreement among authors as to what families in which to place these 12 genera, as well as disagreement how these families themselves may be placed into higher taxonomic categories. This paper attempts to place the twelve genera of seagrasses within the context of Cronquist's system of taxonomy, a system that has widespread acceptance internationally.

INTRODUCTION

The term seagrass is used to collectively describe 12 genera of plants that are not in the grass (Poaceae) family. These twelve genera are: Zostera, Phyllospadix, Heterozostera, Posidonia, Halodule, Cymodocea, Syringodium, Thalassodendron, Amphibolis, Enhalus, Thalassia, and Halophila. They are not even closely related to each other as a whole. The term seagrass is therefore "artificial", a construct of human invention, as opposed to being "natural" in the sense that the plants have enough shared vegetative and generative characteristics to justify categorizing them into a single (i.e. "natural") group.

So what are the "artificial" requirements needed to name a plant a seagrass? They were first delineated in 1920 by Arber, and she cited four properties as being essential. The plant must grow in a saline (marine) environment (being an obligate halophyte), its growth habit must be completely submerged, it must be able to anchor itself well enough to withstand tidal action, and it must be able to complete its entire life cycle while fully submerged, meaning that it must be able to complete pollination while submerged as well, (hydrophilous pollination). The twelve genera named above satisfy all of these requirements. Five other genera of plants, Zannichellia, Lepilaena, Althenia, Ruppia, and some species of Potomogeton, apparently satisfy these requirements as well, but they grow primarily in brackish water (and tolerate a wide range of salinity), and are therefore not truly marine.

Now that we know why these twelve genera are placed in this "artificial" group of seagrass, is there any "natural" group or groupings that they may be placed in for context? I believe we can create a context using principles of higher taxonomy, but how "natural" the grouping is will depend upon what kind of taxonomic approach is taken and which system of taxonomy is used. Even taxonomy itself is a somewhat artificial construction, by which humans attempt to categorize living things into units that make sense in a human context. Since our own context is always changing, taxonomy is always changing. (I have begun to realize this rather frustrating fact during the course of my preparation to write this paper.) Needless to say it is always open to interpretation as well.

Taxonomy is a study that seeks to place "natural" groupings within a hierarchy that hopefully contextualizes various plants relationships to each other. Approaching taxonomy from a top down approach, describing the larger categories first and then fitting the increasingly smaller taxa within them is one way to accomplish this contextualization. Another is to start from the "particulars", the smaller taxa, and work up to the "general" or larger categories. Most authors working in seagrasses seem to take this second approach. What I had hoped to do was to relate the taxonomy of seagrasses by the first approach. This is difficult because much disagreement exits among botanical systematists especially at the level of class, subclass, superorder, order, and family, there is a difficulty with ranking plants to begin with. The characteristics that define a plant as being a seagrass, are as I mentioned earlier human constructs and actually say more about there ecology than their natural taxonomy, and while authors are very good at describing the 12 genera and elucidating the species contained within them, they are not really concerned much with putting them within a total taxonomic context. But I still believe it can be done, and I think it is useful to do so, because following these plants through the "big picture" may add another valuable perspective to understanding these unique organisms, and may help make them more accessible to other students of systematics and taxonomy.

For purposes of simplicity I will be taking my top down approach to seagrass taxonomy using the system of Arthur Cronquist, an American Botanist who wrote "An Integrated System of Classification of Flowering Plants" in 1981, the definitive edition of a taxonomy that he had developed around 1968. His system has proven to be very useful and is for example widely employed in Europe. It is an easy system to familiarize oneself with and makes a good starting point for discussions that concern differences of taxonomic opinion. The most important author concerning the seagrasses themselves is of course the Dutch Botanist C. den Hartog whose classic "Seagrasses Of The World" (1970) provides an excellent taxonomic account of seagrasses at the Genus and Species level, but which I find less helpful in getting a clear picture of the Family and Order level.

METHODS

The basic method involved in compiling this paper involved studying in detail Cronquist's taxonomical system outlined in his 1981 work "An integrated system of classification of flowering plants", and finding out where the 12 genera of plants commonly regarded as being seagrasses are classified under that system. The diagnostic features were noted and summarized starting with the higher taxa all the way down to the family level. Comparisons were then drawn to C den Hartog's 1970 work on seagrass taxonomy "Seagrasses of the world" which is basically a work concerned with the genus and species level, although using it some family comparisons can be made. The issues in taxonomy that I wished to address in this paper concerned the family level and higher. I also looked cursorily at a few other taxonomies to see where they worked out their hierarchies. These included the systems of Bentham and Hooker, Dahlgren and Takhtajan (1969).

RESULTS AND DISCUSSION

Seagrasses are all flowering plants, it is part of their "artificial"definition, and this does actually place them into one very broad "natural" category: the Division angiospermae, as it is traditionally known, also called anthophyta, and called magnoliophyta by Cronquist. The angiosperms may be further divided into two other apparently "natural" categories: the Class monocotyledonae and the class dicotyledonae, called liliopsida and magnoliopsida respectively, by Cronquist. It is generally accepted by most taxonomists that the monocots are derived from dicots and show many features of vegetative and generative reduction.

The seagrasses are all monocots. The monocots are in turn divided into various subclasses or orders depending on the system being used. So this is apparently where the "natural" categories seem to break down. Cronquist divides the monocots into 5 subclasses and 19 orders. Bentham & Hooker, a very influential system dating from 1883 and used for decades by Kew Gardens in Britain divides the monocots only into 7 Orders. The current system used by Kew Gardens most approximates the one set fort by R. M. T. Dahlgren in 1983. He divides the monocots into 10 subclasses and 23 orders. Takhtajan, a Russian Botanist who was a colleague of Cronquist divides the monocots into 11 subclasses, 7 superorders and 38 orders and 6 superfamilies (Takhtajan 1969). There are numerous cases where families are put into different orders in different systems.

Noting all this where does one take the next step down in the taxonomical assignment of the seagrass take us? Apparently most major systems of plant taxonomy in dividing the Monocots delineate a category (usually a Subclass) that is comprised of aquatics. This is unusual because most other categories are described based on floral characteristics. Usually it is the first one listed, meaning it is considered the most "primitive", or to put it perhaps more accurately, the most ancestral and least derived. Sometimes this grouping is known as the Helobiae (from the Greek helos - marsh, and the Latin bius - mode of life) or marsh plants. Cronquist, and a few others call it the Alismatidae, taking the root name from the genus Alisma, the water plantain. So in this respect it seems quite logical to place the 12 seagrass genera into the subclass Alismatidae.

It is interesting to note that this is the only subclass in Cronquist to be described by primarily by habit rather than floral characteristics (although floral characteristics and noted as well). Why is this so? Cronquist believes that monocots evolved from fresh water aquatic dicots, perhaps from a group resembling the modern dicot order Nympheales (the group that contains the water lily and the lotus). So the ancestral condition of Monocots is considered to be aquatic. Subsequent evolution of the terrestrial growth habit is considered to be a derived characteristic. Alismatidae contains all the families that Cronquist believes have retained their ancestral aquatic growth habit. The plants in this subclass are not considered to contain members that had evolved a terrestrial growth habit and then returned to an aquatic one. Cronquist does indeed place all the seagrass genera within the subclass Alismatidae. The idea that the marine aquatics evolved from fresh water aquatics is the prevailing one and was also supported by Arber (1920). C. den Hartog (1970) takes exception and believes that the seagrasses evolved from salt tolerant terrestrial ancestors (with the kind of habit that plants associated with mangrove ecosystems have) that evolved greater and greater tolerance for immersion in salt water. He notes that two genera Amphibolis and Thalassodendron, show vivipary, a trait quite often found in mangrove plants. He also believes there is fossil evidence to support this, but there is debate as to whether some of these fossils can be reliably identified. Still he maintains that "In my opinion the Angiosperms originally invaded the sea and fresh water quite independently along different lines."

Den Hartog (1970) does not describe what higher taxa he believes the seagrass genera should be placed, so it is difficult to deduce which other plant families he feels they are allied to, or possibly derived from. The fact that he believes in a terrestrial origin for them would certainly mean they are extremely derived and he would have to place them in another group besides the Alismatidae. Of the 4 other subclasses of monocots within Cronquist's system (Arecidae containing the palms and aroids, Commelinidae - containing the spiderworts, grasses, sedges, and rushes, Zingiberidae - containing the gingers bromeliads and bananas, and Liliadae - containing the lilies, aloes, yams and orchids), perhaps he would assign them to the Commelinidae, since this subclass contains the orders that are superficially at least most vegetatively similar, although quite different florally.

Cronquist's subclass Alismatidae, besides showing an aquatic habit, has other characteristics unique to it. The leaves are simple and parallel veined, usually sheathing at the base. The stem usually has small scales at the nodes. The vascular system is quite reduced, and vessels are confined to the roots or absent altogether. In the flowers the perianth is generally reduced or lacking, the carpels are generally free, the pollen is almost always trinucleate and monosulcate (possessing a single furrow in its outer coating) and lacking an aperture. The fruits are mostly follicles (dry non fleshy fruits developed from a single carpel) and the seeds lack endosperm.

Within Alismatidae there are 4 Orders: Alismatales, Hydrocharitales, Najadales and Triuridales. Each of these orders are sorted out from each other primarily with floral characteristics. It is at this level that the 12 seagrass genera are first partitioned up in this system. Of the twelve genera, 3 are apparently more closely related to each other than they are with the remaining three, and form a more cohesive group. Enhalus, Halophila and Thalassia are all placed by Cronquist in the Order Hydrocharitales (from the Latin hydro -water, and the Greek charis - grace or beauty) within a single family. Zostera, Phyllospadix, Heterozostera, Posidonia, Halodule, Cymodocea, Syringodium,

Thalassodendron and Amphibolis are all placed in the order Najadales (from the Latin naiad - water nymph) within 3 different families.

The order Hydrocharitales, and its single family, the Hydrocharitaceae, has diagnostic features as follows. They may be annual or perennial aquatic herbs and they may have creeping monopodial rhizomes. Tanniferous cells are present. Leaves generally have a stipular expansion, and always subtend an axillary scale at their base, and their epidermal cells contain chlorophyll. Flowers are most often unisexual, contain a perianth comprised of 3 distinct colored petals and three distinct green sepals and are in addition usually subtended by connate (joined) bracts. The inflorescence is usually compact and in the case of pistillate flowers usually consists of a single flower. There may be one or more carpels and they may be weakly connate. There are as many styles as there are carpels, and the stigmas are dry. The ovaries are inferior, and unilocular, with laminar or parietal placentation. The ovules are generally anatropous (with micropyle pointing down towards funicle). The anthers are tetrasporangiate opening by longitudinal slits. The pollen grains are globose and most often lacking an aperture, and in marine genera are united into thread like tetrads. Fruits are submerged and open irregularly. The seeds lack endo sperm and the embryo is straight with a pronounced radicle. Germination is epigeal with the cotyledon emerging.

Besides the three marine genera contained in the family Hydrocharitaceae, also known as the Frog bit family, there are 12 other fresh water genera of plants, including the familiar Elodea, and Vallisneria. In general Hydrocharitales are considered to be a rather primitive group, sharing many similarities with the most primitive order in the Alismatidae, the Alismatales, and in particular the family Butomaceae.

The order Najadales, with has 10 families, is described as having the following features. The leaves often have a ligule at the juncture of the sheathing base and the free portion, but without a stipule. Flowers are most often unisexual with a perianth (when present) not differentiated into sepals and petals, and not subtended by bracts. Ovaries are superior. Carpels are often united to form a compound ovary, with ovules solitary in each locule and variously anatropous, campylotropous, or orthotropous. Anthers are mostly tetrasporangiate, but sometimes unisporangiate, and open longitudinally. Pollen grains often lack an exine. Fruit is often single seeded. Endosperm is absent in the seeds.

The Najadales on a whole are considered more derived, or advanced than the Hydrocharitaceae (even despite the superior ovary). They show many signs of floral reduction, which is considered a consequence of abandoning insect pollination, and seems to follow a general trend of more specialized habitats, in going from fresh, to brackish and finally marine environments. The ten families within Najadales are defined mainly by generative characteristics, however three families contain exclusively marine genera. These three families are Posidoniaceae (from the Greek god Poseidon), Cymodoceaceae (from the Greek Cymodoce a water nymph or nereid, an attendant to Poseidon) also known as the Manatee grass family, and Zosteraceae (from the Greek zoster - meaning girdle, belt or zone) also known as the eel grass family. Three other families in Najadales contain the five brackish water genera mentioned at the beginning of this paper that are sometime allied with but not considered to be seagrasses. Ruppiaceae, the Ditch weed family, contains Ruppia. Potomogonaceae, the Pondweed family, contains Potomogeton, and Zannichelliaceae, the Horned Pondweed family, contains Zannichellia, Lepilaena and Althenia.

The family Posidoniaceae contains only the genus Posidonia. The family Cymodoceaceae contains the genera Cymodocea, Syringodium, Amphibolis, Halodule and Thalassodendron. The family Zostera contains the genera Zostera, Heterozostera and Phyllospadix. The families Posidoniaceae and Zosteraceae are considered to be more closely related to each other than they are to Cymodoceaceae. Cymodoceaceae is in fact considered to be most closely related to Zannichelliaceae.

The characteristics that define Posidoniaceae include the following. They are perennial, and have laterally flattened monopodially creeping rhizomes. Tanniferous cells present. Vascular bundles are closed and they lack vessels completely. Stems are erect. Leaves are alternate, distichously arranged, and ligulate at the juncture of sheath and blade, with the sheath being persistent after the blade has been shed. Flowers are perfect, sessile, proterandrous, hydrophilously pollinated and born in terminal spikes. The spikes themselves are subtended by bracts, but not each individual flower. The perianth is scale like or absent. There is only one carpel with a sessile stigma. Anthers are also sessile. The pollen grains are filamentous, and lack an exine. Pollination is hydrophilous. The single ovary is superior. The ovule is solitary and orthotropous. The fruit is buoyant in water. Seed lacks endosperm and the embryo has a lateral cotyledon and a terminal plumule.

The characteristics that define Cymodoceaceae include the following. The rhizome may be herbaceous and monopodial or woody and sympodial. Tanniferous cells present. Vessels are lacking in all tissues. Leaves are alternate or opposite and distichous. The basal sheath is open and there is a ligule at the juncture of blade and sheath. Scales are present at the nodes. Flowers are unisexual and the plants are dioecious. They lack a perianth completely and are either born in small cymes or in pairs in the leaf axils. There are two distinct carpels each with a separate style. The ovaries are superior and the each orthotropous ovule is solitary within its carpel. The anthers are paired and either sessile or nearly sessile. The pollen grains lack an exine, are filamentous, and up to l mm long. Pollination is hydrophilous. The fruit is single seeded and indihiscent.

The characteristics that define Zosteraceae include the following. Plants are perennial and the rhizomes may be tuberously thickened. Rhizome creeps sympodially in Heterozostera and monopodially in Zostera and Phyllospadix. These plants lack tanniferous cells. Leaves are alternate and commonly distichous. The leaves are ligulate, and the sheath may be open or closed, and may be flanged. Flowers are unisexual and arranged in two rows on one side of a flattened spadix and pistillate and staminate flowers are inclosed within the same spadix except in Phyllospadix which is dioecious. Within the spadix accessory to the flowers are small bractlike structures of unknown morphological origin called retinacula. The flowers are otherwise lacking a perianth and subtending bracts. Pistillate flowers contain two carpels joined to form a compound unilocular ovary, and contain two basally connate styles. The ovules are solitary and orthotropous. Staminate flowers consist of a single sessile stamen with two anthers. Pollen grains are filamentous and pollination is hydrophilous. Fruit is usually a drupe. The seed lacks endosperm and the embryo has a basally closed and sheathing cotyledon lying in a groove of enlarged hypocotyl.

At this point I think it would be useful to describe briefly what den Hartog (1970) has to say about seagrass taxonomy. He separates the 12 genera along the same lines more or less as Cronquist does. He places Enhalus, Halophila and Thalassia in the family Hydrocharitaceae. Not much difference there. But he goes on to place each genus in a separate subfamily. The three subfamilies he recognizes are: Marine Vallisnerioideae (containing Enhalus), Thalassioidae (containing Thalassia), and Halophiloidae (in which he divides the genus Halophila up further into 4 sections). It is not clear how many other genera he would consign to Hydrocharitaceae, because he never mentions any other than the seagrass genera he is concerned with. This illustrates how difficult it is to work with taxonomies that do not extend up beyond the family level, they do not help delineate relationships among larger groups of plants.

The remaining genera he consigns to the family Potomogetonaceae, naming three subfamilies: Posidonioideae, Cymodoceoidae, and Zosteroidae. He divides the genus Zostera into to 2 subgenera. This is also not such a drastic departure from Cronquist, the basic difference being that Cronquist considers the characteristics of these 3 subfamilies to be strong enough to split them off from Potomogetonaceae and give them full family status. Again there is a problem with not knowing how large den Hartog (1970) considers the entire Potomogetonaceae family to be, since only the seagrass genera are mentioned, so there is no way to relate how he considers them related to other groups of plants.

Den Hartog (1970) recognizes 49 species of seagrass, 14 in Hydrocharitaceae and 35 in Potomogetonaceae. He distinguishes the genera apart from each other largely based on floral characteristics. However he does note that it is often very difficult to obtain flowering specimens for comparison and that many herbaria specimens are sterile so he does include some vegetative features in his key to distinguish the genera as well. Here follows a summary of some of the diagnostic features for each genus.

Family Hydrocharitaceae:

1) Enhalus: (from the Greek en + halus - in salt)

Floral - Numerous flowers in male inflorescence, flower buds detaching before flowering and floating to surface, pollen grains free

Vegetative - leaves distichous, lacking ligula, 1 1/4 to 1 3/4 cm wide, not differentiated into petiole and blade; rhizome: I cm thick, covered with persistent black strands, roots cord like.

2) Thalassia: (from the Greek Thalass - sea)

Floral - Single flower in male inflorescence, flower shed after flowering, pollen grains shed in moniliform chains, 6-8 styles divided into 2 stigmata in female flower

Vegetative - leaves: distichous, lacking ligula, not differentiated into petiole and blade, 1/2 1 cm wide; rhizome: 1/2 cm thick, internodes longitudinally grooved, scale at each node.

3) Halophila: (from the Greek halus + phila - salt loving)

Floral - single flower in male inflorescence, flower shed after flowering, pollen in moniliform chains, 3-5 styles in female flower not divided

Vegetative leaves in pairs or psuedowhorls, lacking ligula, differentiated into petiole and blade

Family Potamogetonaceae

subfamily Zosteroidae (family Zosteraceae in Cronquist)

4) Zostera: (from the Greek zoster - girdle)

Floral - monoecious, sessile spadix always covered by spathal sheath, retinacula often absent

Vegetative - leaves: ligulate, lacking tannin cells; rhizome: monopodial, internodes elongate with long roots at nodes

5) Phyllospadix: (from the Latin phyl + spadix - leaf shaped like a spadix)

Floral - dioecious, spadix projecting out of spathal sheath

Vegetative - leaves: ligulate, lacking tannin cells; rhizome: monopodial, internodes short, and short roots at internodes

6) Heterozostera: (from the Greek hetero - different, + Latin zoster - girdle)

Floral - monoecious, spathal sheath inflated, spadix stalked, retinacula always present

Vegetative - leaves: ligulate, lacking tannin cells; rhizome: sympodial

subfamily Posidonidonioidae (family Posidonaceae in Cronquist)

7) Posidonia: (from the Greek sea god Poseidon)

Floral - bisexual flowers

Vegetative - leaves: ligulate tannin cells present; rhizome: monopodial, thick, with persistent leaf bases

subfamily Cymodoceoidae (family Cymodoceaceae in Cronquist)

8) Halodule: (from the Greek halus + doulos - salt slave)

Floral - dioecious, flowers solitary, ovary with 1 undivided style

Vegetative - leaves: flat, ligulate, tannin cells present, with 3 nerves; rhizome: monopodial, thin, leaf sheathes persisting, short shoot at each node

9) Cymodocea: (from the Greek nereid Cymodoce an attendant to Poseidon)

Floral - dioecious, flowers solitary, style divided into 2 stigmata

Vegetative - leaves: flat, ligulate, tannin cells present, with 7-17 nerves; rhizome: monopodial, thin, leaf sheathes persisting, short shoot at each node

10) Syringodium: (from the Greek syring - tube)

Floral - dioecious, flowers in cymose inflorescence

Vegetative - leaves: round, ligulate, tannin cells present; rhizome: monopodial, leaf sheathes persisting, short shoots at each node

11) Thalassodendron: (from the Greek thalass + dendron - sea tree)

Floral dioecious, flowers subtended by bracts, anthers entirely connate, style with 2 stigmata

Vegetative - leaves: ligulate, tannin cells present, parallel nerved, leaf sheath shed with blade; rhizome: sympodial, sparsely branched shoots at each 4th node, roots at internode preceding stem bearing nodes

12) Amphibolis: (from the Greek amphibol - ambiguous)

Floral - dioecious, flowers enclosed by normal leaves, anthers ventrally connate, style with 3 stigmata

Vegetative - leaves: ligulate, tannin cells present, psuedoparallel nerved, leaf sheath shed with blade; rhizome: sympodial, profusely branched shoots every 4-8 internodes, roots at each internode

CONCLUSIONS

The seagrasses although they are a multiphyletic group can be seen to have some taxonomic cohesion. They are all angiosperms, monocots, and belong to the same aquatic subclass. They are placed in 2 separate orders, Hydrocharitales if they have complete flowers, inferior ovaries and stipulate leaves, and Najadales if they have reduced floral parts and ligulate leaves. Hydrocharitales is monofamilial with three genera, Enhalus, Halophila, and Thalassia. Najadales contains many families but three are diagnostic as having marine genera. Differentiation among the marine families is based on floral characteristics. Posidoniaceae if the flowers are bisexual (Posidonia). Cymodoceaceae if the inflorescences are small to singular and the carpels are free (Halodule, Cymodocea, Syringodium, Thalassodendron, Amphibolis). Zosteraceae if the inflorescences are arranged in a spadix enclosed in a spathe and the carpels are fused (Zostera, Phyllospadix, Heterozostera). Within each genera it is possible to use both vegetative and floral characteristics for field identification. Important features include if the leaves are ligulate, if the rhizomes are monopodial or sympodial, thick or thin, and the positioning of the stems and roots with respect to the internodes.

ACKNOWLEDGMENTS

I would like to thank Dr. J. Fourqurean, at Florida International University, for enabling me to get access to a copy of C. den Hartog's "Seagrasses of the world", which is currently out of print.

REFERENCES

Arber, A., 1920. Water plants, a study of aquatic angiosperms. Cambridge University Press, 436 pp.

Allaby, M., 1992. The concise oxford dictionary of botany. Oxford UniversityPress, New York.

Brummit, R. K., 1992. Vascular plants families and genera. Royal BotanicGardens, Kew, pp. 1-6, 753-804.

Cronquist, A., 1981. An integrated system of classification of flowering plants. Columbia University Press, New York pp. 1031-1041, 1052-1072.

Den Hartog, C., 1970. Seagrasses of the world. North Holland Publishing Company, Amsterdam.

Heywood, V. H., 1993. Flowering plants of the world. Oxford University Press ,New York, pp. 7-26, 271-277.

Mabberly, D. J., 1987. The plant book: a portable dictionary of the higher plants. Cambridge University Press, Cambridge.

Takhtajan, A. 1969. Flowering plants, origin and dispersal. Smithsonian Institution Press, Washington. 310 pp.

Return to BOT 5647 Student Projects