Sieve-element Plastids Research Articles

Haustorial structure of the holoparasitic angiosperm Cynomorium coccineum L. invading host roots

Absract The perennial angiosperm holoparasite Cynomorium coccineum L. (Cynomoriaceae) was found infecting the roots of halophytes in a Mediterranean coastal salt marsh in Egypt. The parasite consists of a subterranean rhizome, which carries haustoria and sends aboveground annual inflorescences. Although the published research on the botany and taxonomy of C. coccineum dates back to approximately 150 years ago (Weddell, 1860; Engler, 1889; Baccarini and Cannarella, 1899), it lacked sufficient information on the anatomy of the haustorium parasitizing the host roots. Here, we described for the first time, the anatomy and ultrastructure of the haustoria of C. coccineum infecting host roots. The hosts recorded in this study were Limonium narbonense Mill. (Plumbaginaceae) and Atriplex portulacoides L. (Chenopodiaceae). The anatomy of haustoria was investigated by light microscopy, by histochemical techniques to stain lignin and callose and by scanning and transmission electron microscopy. The invasion of the host root by the haustoria of C. coccineum is largely accomplished via mechanical pressure. The presence of sieve tube elements in the haustorium at the interface between the phloem of the host roots and the parasite, was detected by light microscopy. The sieve element plastids of the host H. portulacoides are of the P-type, which is characteristic of Chenopodiaceae. The sieve elements in the haustorium are either differentiating or mature; the formers have thick undulating walls, disintegrated protoplast, and callose occluded sieve plate pores. The mature ones have fibrillar nacreous walls, clear lumen and fine sieve plate pores occluded with non-callose material. Sieve element plastids and P-protein bodies were not observed in the phloem of the haustorium. Plasmodesmatal connections are lacking at the contacts between the host and the parasite. The contacts between the tracheary elements of H. portulacoides and the parasite occur by the abutments of their walls, pits, or parenchyma cells.

Development of phloem connection between the parasitic plant Orobanche cumana and its host sunflower

Orobanche cumana is a root parasitic plant causing considerable yield losses in sunflower cultivation. The holoparasite fulfills its entire demand for water, minerals, and organic nutrients from the host's vascular system. In this study, the ultrastructure of the phloem connection between the haustorium of young O. cumana tubercles and the sunflower root has been examined for the first time. Parasite and host tissues were intermingled at the contact site and difficult to distinguish, but sieve-tube elements of O. cumana and sunflower could be differentiated according to their plastid ultrastructure. While O. cumana sieve-element plastids were larger, often irregular in shape and contained few, small starch inclusions, sieve-element plastids of the host were significantly smaller, always roundish with more and larger starch inclusions. This made it possible to trace the exact contact site of host and parasite sieve elements to show a direct symplastic phloem connection between the two species. The interspecific sieve plate showed more callose on the host side. This allowed detection of newly formed plasmodesmata between host sieve-tube elements and parenchymatic parasite cells, thus showing that undifferentiated cells of the parasite could connect to fully differentiated sieve elements of sunflower. Furthermore, the arrangement of phloem within the O. cumana tubercle as well as differences in sieve-element plastid ultrastructure during shoot development in O. cumana were investigated and are discussed in this paper.

Starch-accumulating (S-type) sieve-element plastids in Hydatellaceae: implications for plastid evolution in flowering plants

TEM investigation of sieve-element plastids in three species of Trithuria, the sole genus of the small aquatic family Hydatellaceae, show that P-type plastids are absent from this genus and only starch-accumulating (S-type) sieve-element plastids are present. This discovery is consistent with the recent transfer of Hydatellaceae from the highly derived monocot order Poales (grasses and their allies) to the early-divergent angiosperm order Nymphaeales (waterlilies) based on molecular phylogenetic data. Species of Poales consistently possess P2-subtype plastids, in common with other monocots, but only S-type plastids are present in Nymphaeales. The results confirm that Hydatellaceae do not belong in monocots. Optimisation of the two major types of sieve-element plastid onto a recent phylogeny of early-divergent angiosperms confirms that S-type is the primitive form and indicates that P-type sieve-element plastids have evolved more than once in angiosperms.

Can sieve-element plastids in Panicum maximum (Poaceae) leaves act in the blockage of injured sieve-tube elements?

The ultrastructural features and the plastid changes caused by sample preparation were studied in sieve elements of Panicum maximum leaves. Samples of expanded leaves, taken near the ligule region, were fixed and processed by common light and transmission electron microscopy methods. In mature sieve-tube elements, the protoplast is electron-translucent and plastids are the most frequent organelles. Mitochondria and smooth endoplasmic reticulum segments are also visible and occupy a parietal position within the cell. The plastids are globular and show electron-dense proteinaceous inclusions in the stroma. The protein crystals are predominantly cuneate, but thin crystalloids and amorphous and/or filamentous proteins also occur. The presence of intact plastids plus others in different phases of plastid envelope rupture were interpreted as evidence that this rupture is a normal event in response to injury. This plastid envelope rupture is possibly activated by the release of pressure in the sieve-tube element. After plastid membrane vesiculation, the stroma and the protein crystals are dispersed within the sieve-element ground cytoplasm. The vesicles originating from the plastid envelope move to one cell pole, while protein crystalloids move to the opposite pole and agglomerate in the sieve-plate region. Our findings indicate that these protein crystalloids, which deposit in the sieve plate, may act in sieve-plate pores occlusion, preventing the release of phloem sap, similar to the role of P-protein in dicotyledons.

Sieve-Element Plastids and Evolution of Monocotyledons, with Emphasis on Melanthiaceae sensu lato and Aristolochiaceae-Asaroideae, a Putative Dicotyledon Sister Group

Abstract Monocotyledons are distinguishable from dicotyledons by their subtype P2 sieve-element plastids containing cuneate protein crystals, a synapomorphic character uniformly present from basal groups through Lilioids to Commelinoids. The dicotyledon genera Asarum and Saruma (Aristolochiaceae-Asaroideae) are the only other taxa with cuneate crystals, but their sieve-element plastids include an additional large polygonal crystal, as is typical of many eumagnoliids. New investigations in Melanthiaceae s.l. revealed the same pattern (polygonal plus cuneate crystals) in the sieve-elementplastids of Japonolirion osense (Japonoliriaceae/Petrosaviaceae), of Harperocalfsflava, Pleea tenuifolia, and Tofieldia (all: Tofieldiaceae). In Narthecium ossifragum a large crystal, present in addition to cuneate ones, usually breaks up into several small crystals, whereas in Aletris glabra and Lophiola americana (Nartheciaceae) and in all of the 15 species studied and belonging to Melanthiaceae s.str. only cuneate crysta...

Systematics and evolution of Velloziaceae, with special reference to sieve-element plastids and rbc L sequence data

Sieve-element plastids and rbc L nucleotide sequences were analysed for Velloziaceae and all families sometimes discussed as close allies (Acanthochlamydaceae, Amaryllidaceae, Bromeliaceae, Cyclanthaceae, Haemodoraceae, Hypoxidaceae, Pandanaceae, Pentastemonaceae, Stemonaceae). Velloziaceae (37 taxa investigated) contain sieve-element plastids defined here as subforms P2c a–c, P2c al, P2c ap, P2c alf, and P2c apf and distinct from those of any other monocotyledon by (1) angular crystals replacing to a large extent the monocot-specific cuneate crystals, (2) additional loosely-packed crystals, and (3) their very small average diameters. This sieve-element plastid syndrome is a synapomorphy of Velloziaceae and indicates the monophyly and early isolation of the family. Sieve-element plastids of the monotypic Acanthochlamydaceae share their small sizes with Velloziaceae, but all their crystals are cuneate, i.e. loosely-packed crystals were not found. Among the other families studied the sieve-element plastids of (a) some Bromeliaceae ( Ayensua in particular) contain loosely-packed crystals, (b) several Amaryllidaceae incorporate additional orthogonal crystals, and (c) Pandanaceae belong to form-Pcf, but cuneate crystals are present throughout and the plastids are significantly larger than in Velloziaceae. The new information on the rbc L nucleotide sequences of 24 taxa of the Velloziaceae and of 20 additional taxa from putatively related families strongly support the monophyly of Velloziaceae and define Acanthochlamydaceae as its sister family. At the infra-familial level four evolutionary lines are discovered which correspond to major generic groups and their geographical distribution. The divergence time for the split between New Word and Old World taxa has been calculated as 40–70 million years, that between African and Madagascan species maximally 25 million years. The cladistic analysis requires that the monotypic genera Nanuza and Talbotia be included within Vellozia and Xerophyta, respectively. At the ordinal level the analysis of the rbc L-data strengthens the Pandanales as a monophyletic clade.

Systematics and evolution of Velloziaceae, with special reference to sieve-element plastids and rbcL sequence data

Sieve-element plastids and rbcL nucleotide sequences were analysed for Velloziaceae and all families sometimes discussed as close allies (Acanthochlamydaceae, Amaryllidaceae, Bromeliaceae, Cyclanthaceae, Haemodoraceae, Hypoxidaceae, Pandanaceae, Pentastem

P-type sieve-element plastids present in members of the tribesTriplareae andCoccolobeae (Polygonaceae) renew the links between thePolygonales and theCaryophyllales

Form-Pfs sieve-element plastids were found inTriplaris, Ruprechtia, andCoccoloba (Polygonaceae) while other genera of the family and those studied from the often associatedPlumbaginaceae contain S-type sieve-element plastids. The rareness of form-Pfs plastids among the angiosperms, their similarity to the peculiar form-P3fs plastids of theChenopodiineae, and the comparatively small plastid diameters measured for all forms present in theCaryophyllales, Polygonales, andPlumbaginales suggest close relationships between these taxa. The restriction inPolygonaceae of form-Pfs plastids to the closely allied tribesTriplareae andCoccolobeae is discussed with regard to both the intrafamilial and ordinal phylogeny, and also considering possible connections to the only magnoliidaean Pfs-taxonCanella.

Systematics of Acorus: Ovule and Anther

This article identifies and discusses the critical characters in the systematics of Acorus, a significant taxon among "basal" angiosperms. New data are presented on tapetum development and ovule structure, and old characters are revisited to establish their homologies. Many of the character states found in Acorus (such as trimerous flowers, successive microsporogenesis, orthotropous ovules, secretory anther tapetum, ethereal oil cells, and lack of raphides) occur both in other monocotyledons and "primitive" dicotyledons. Several characters indicate close monocot affinities but provide little further evidence of relationships: P2-subtype sieve-element plastids (Behnke 1995), seedling structure (Tillich 1995), and ensiform leaves. Two characters that potentially link Acorus with palaeoherb outgroups, anther wall formation and cellular endosperm, both require further review. Apart from the strong signal from molecular data, there is still virtually no nonmolecular evidence to confirm a taxonomic placement for Acorus as the sister taxon to all other monocotyledons.

Developmental features of protophloem sieve elements in roots of wheat (Triticum aestivum L.)

Protophloem sieve elements (PSEs) in roots of wheat (Triticum aestivum L.) are arranged in single vertical files. The number of PSEs within the files increases by symmetrical divisions, which take place after the completion of asymmetrical (formative) divisions and before the initiation of differentiation. The divisions are preceded by well defined pre-prophase bands (PPB) of microtubules, which surround the nucleus in an equatorial position. In the cytoplasmic region between the nuclear surface and the PPB, perinuclear and endoplasmic microtubules were observed. The perinuclear microtubules are considered as part of the developing spindle, while the endoplasmic ones interlink the perinuclear microtubules with the PPB. Dividing cells do not show any signs of incipient differentiation. The first and most reliable indication of a commencing differentiation is provided by the sieve-element plastids that begin to accumulate dense crystalloid inclusions in the very young PSEs. In mature PSEs plastids contain two kinds of crystalloid inclusions, dense and thin, in a translucent stroma. Depending on the plastid-inclusions criterion it was shown that: (a) the PSEs of a given root do not initiate differentiation at exactly the same stage, (b) the developmental sequence extends to a span of 7–9 actively differentiating PSEs arranged in a single vertical file, and (c) each PSE needs about 16–21 h to pass through the whole developmental sequence. In the last two differentiating PSEs of a file, mitochondria were found to be enveloped by single cisternae of ER. The association is temporary as it is lost in the first PSEs with an autolysed lumen. During differentiation, Golgi bodies were abundant and active in producing vesicles involved in cell wall development. Golgi vesicles were also found among the microtubules of the PPB, but no local thickening was observed. Golgi bodies disorganize in the last stages of autolysis and disappear in mature sieve elements.

P-type sieve-element plastids and the systematics of theArales (sensuCronquist 1988)?with S-type plastids inPistia

The sieve-element plastids of 126 species of theArales were investigated by transmission electron microscopy. With the exception ofPistia (with S-type plastids) all contained the monocotyledon specific subtype-P2 plastids characterized by cuneate protein crystals. While the species studied from bothAcoraceae andLemnaceae have form-P2c plastids (i.e., with cuneate crystals only), those of theAraceae belong to either form P2c (14 species), P2cs (the great majority) or P2cfs (Monstera deliciosa, only, with form-P2cs plastids in the otherMonstera species studied). The form-P2cs plastids of theAraceae are grouped into different categories according to the quantity and quality of their protein and starch contents. The subfamilyLasioideae is redefined to comprise all aroid P2c-taxa and those P2cs-genera that contain only one or very few starch grains. Only little starch is also recorded in the sieve-element plastids ofGymnostachys (Gymnostachydoideae), with the other plastid data denying a close relationship toAcorus. While equal amounts of starch and protein are generally present in sieve-element plastids of the subfamiliesPothoideae, Monsteroideae, Colocasioideae, Philodendroideae, andAroideae, maximum starch content and only very few protein crystals are found in form-P2cs plastids ofCalla (Calloideae),Ariopsis (Aroideae), andRemusatia (Colocasioideae?). In the latter, both morphology and size of sieve-element plastids are close to those ofPistia.—In theAraceae the diameters of the sieve-element plastids exhibit a great size range, but are consistent within a species and within a defined part of the plant body. Comparative data are mainly available for stem and petiole sieve-element plastids.—The accumulated data are used to suggest an affiliation of the species to subfamilies and to discuss the phylogeny of theArales. Forms and sizes of their plastids support a separation of bothAcoraceae andLemnaceae from theAraceae. The presence of S-type plastids inPistia does not favour direct and close relationships to the form-P2c genusLemna.—The prevailing form-P2cs plastids might support proposals that place theArales (together with also form-P2cs plastid containingDioscoreales) in the neighbourhood of basal dicotyledons. BesidesAsarum andSaruma (Aristolochiaceae), with monocotyledonous form-P2c plastids,Pistia (with dicotyledonous S-type plastids) gives another example for a link between the two angiosperm classes.

Further studies of the sieve-element plastids of theCaryophyllales includingBarbeuia, Corrigiola, Lyallia, Microtea, Sarcobatus, andTelephium

The sieve-element plastids of 69 species of theCaryophyllales were investigated by transmission electron microscopy. All contained the specific subtype-P3 plastids characterized by a peripheral ring of protein filaments. The presence or absence of an additional central protein crystal and their shape being either polygonal or globular as well as the average sizes of the sieve-element plastids are useful features in the characterization of some families.—Barbeuia contains sieve-element plastids that confirm its placement within thePhytolaccaceae. Lyallia differs fromHectorella by including small starch grains in their sieve-element plastids, which otherwise by their globular crystals negate a closer connection to theCaryophyllaceae. The lack of a central protein crystal in its form-P3fs plastids placesMicrotea best within theChenopodiaceae. Sarcobatus, a so far uncontested member of theChenopodiaceae, contains form-P3cf plastids, i.e., including a central crystal not found elsewhere in this family.Telephium andCorrigiola, shifted back and forth betweenMolluginaceae andCaryophyllaceae, have form-P3cf(s) plastids with a polygonal crystal which favor their placement within theCaryophyllaceae.

Distribution and Evolution of Forms and Types of Sieve-element Plastids in the Dicotyledons

Sieve-element plastids may contain any combination of protein crystals (c), protein filaments (f), and starch grains (s), or none of these. All structurally distinct combinations (=forms) possible (s, cs, cfs, c, cf, fs, f, o) are recorded among the 2100 dicotyledons investigated so far with the transmission electron microscope (representing 381 of more than 460 families described). The six forms that include c and/or f define the P-type and are found in some 620 species, mainly confined to Magnolianae, Caryophyllanae, Vitanae, and Rutanae (Fabales, Linales, Rhizophorales). A detailed family-by-family analysis and quantitative form-by-form comparison suggests that form-Ss plastids give rise to all other forms. Based on a logic model connecting all forms via one-step-alterations (loss or gain of starch, protein crystals, or filaments), the distribution of P-type plastids within dicotyledons, their interconnection, and their relationships to the S-type taxa are discussed. Due to the diversity of sieve-element plastids found within one taxon, the Magnolianae (P- and S-type) are held to be a center of P-type evolution and more primitive than Caryophyllanae (P-type, only) or other superorders. The significance of sieve-element plastid data to the shaping of some of the higher dicotyledon taxa is discussed.

An Additional Note on the Floral Morphology and Affinities of Medusagyne Oppositifolia (Medusagynaceae)

Further observations on the floral morphology and anatomy of the monotypic genusMedusagyne from the Seychelles are presented. The unique synovarious gynoecium is composed of numerous carpels that are attached to the side of a broad, elongated torus. The biovulate carpels are differentiated into a narrow ovary, a stout, erect, three-trace style, and a lobed, discoid stigma. The carpel wall is initially composed of thin-walled parenchyma, but at later stages of development the peripheral region becomes subdivided into many lignified sterile extensions. Carples show pronounced ontogenetic adaxial deformation, during which the upper ventral surfaces become expanded, greatly extending the ventral margins and forming a domelike cushion of carpel-derived tissue over the summit of the torus. The distal cushion remains separated from the flattened torus by a compressed cavity and eventually undergoes lignification, perhaps protecting the underlying gynoecium from chewing insects and exposure. Comparisons are made between the mature gynoecia ofMedusagyne and those of the genusDillenia. The occurrence of S-type sieve-element plastids inMedusagyne is reported for the first time, a feature in agreement with suggested thealean affinities of the family.

A phloem-specific, lectin-like protein is located in pine sieve-element plastids by immunocytochemistry

Two co-purifying phloem polypeptides of 24 and 25 kilodaltons (kDa) were isolated from homogenates of Pinus sabiniana Dougl. phloem by differential centrifugation, selective solubilization and electrophoresis, and rabbit antibodies raised against them. The antisera were found to be specific for doublet bands between 23 and 25 kDa in Western blots of whole phloem extracts of Pinus species; no xylem polypeptides were labelled, nor did labelling occur in blots of phloem extracts from other genera in the Pinaceae. Solubilized phloem polypeptides bind strongly to chitin (oligomeric N-acetylglucosamine) columns and are sensitive to thiol reagents, both characteristics which relate them to phloemspecific lectins isolated from angiosperm species (C. Allen, 1979, Biochem. J. 183, 133-137; A.K. Gietl et al., 1979, Planta 144, 367-371). Fluorescence microscopy and immuno-gold electron microscopic cytochemistry demonstrated antigenic sites specifically associated with protein crystals peculiar to the sieve-element plastids of the Pinaceae.

Contributions to the knowledge of sieve-element plastids inGunneraceae and allied families

P-type sieve-element plastids were found in theGunneraceae, while S-type plastids are present in theHaloragaceae andHippuridaceae. The specific characters of the sieve-element plastids (e.g., their size and the morphology of their contents) are discussed in relation to other taxa of theRosidae containing P-type plastids and to the systematic position of theGunneraceae.

Sieve element characters and the systematic position ofAustrobaileya, Austrobaileyaceae?with comments to the distinction and definition of sieve cells and sieve-tube members

The ultrastructure of the sieve elements ofAustrobaileya is compared with that of angiosperm sieve tubes and gymnosperm sieve cells (mostly fromCycadales). Except for the size of the sieve poresAustrobaileya shares all ultrastructural characters (e.g., chromatolytic nuclear degeneration, presence of p-protein, formation of sieve pores from unbranched plasmodesmata) and other features (e.g., companion cells) with angiosperm sieve tubes. Gymnosperm sieve cells on the contrary are characterized by pycnotic nuclear degeneration, absence of p-protein, formation of sieve areas from branched plasmodesmata with median cavities. — The exact ordinal assignment ofAustrobaileya within the subclassMagnoliidae is still disputed, a placement close to eitherMyristicaceae andWinteraceae orMonimiaceae being possible as judged from both S-type sieve-element plastids and p-protein bodies. — On the basis of the ultrastructural results fromAustrobaileya it is proposed to reconsider concepts and terminology of sieve elements. i.e., to include features from sieve pore development, nuclei degeneration and presence of specific proteins into the definitions and to restrict the term sieve cell to gymnospermous sieve elements which differ much from those of other vascular plants.

CONTRIBUTIONS TO THE KNOWLEGE OF P‐TYPE SIEVE‐ELEMENT PLASTIDS IN DICOTYLEDONS II. EUCRYPHIACEAE

SummarySieve‐element plastids of Eucryphia are of the P‐type, the characters of which are discussed in relation to its suggested position between the Connaraceae and Cunoniaceae. Paracryphia contains S‐type plastids extremely similar in size and ultrastructure to those in the Dilleniaceae‐Theaceae‐Actinidiaceae.

Ultrastructure of Sieve-Element Plastids of Myrtales and Allied Groups

The core families of Myrtales (69 species investigated) contain S-type sieve-element plastids. In these, the presence of several medium-sized spherular starch grains is probably a common trend in the order. Among those families usually closely associated or more or less distantly related, the great majority also have S-type plastids, some with the same characteristics. However, Rhizophoraceae, in contrast to the close allies and Connaraceae, Gunneraceae, and Rhabdodendraceae of the distantly related families, have developed P-type plastids. That subtype PV-plastids, which exclude all starch, are found in all tribes of Rhizophoraceae except Anisophylleae supports family recognition of the latter. Sieve-element characters investigated with the transmission electron microscope have repeatedly contributed to the circumscription and classification of higher taxa, e.g., Leguminosae (Behnke & Pop, 1981), Monocotyledoneae (Behnke, 198 lb). Of these, the most intensively studied feature is the ultrastructure of sieve-element plastids. The great number of investigated angiosperms (some 2,000 species from nearly 400 families-as of 1983-12-31) allow a subdivision of the plastids into Pand S-type, a number of subtypes, and a great many of characteristic forms by using both chemical and morphological composition at the ultrastructural level (see Behnke, 198 la, for more detail). The successful delimitation of the order Caryophyllales to the PIll plastid families (Behnke, 1976; Mabry, 1977) and the use of sieve-element plastid data in separating Vitidales from Rhamnales (Behnke, 1974; Dahlgren, 1980) and Buxaceae from Simmondsiaceae and Euphorbiaceae (Behnke, 1982a) stimulated, among others, the screening of all the core families of Myrtales and of those closely or more distantly related to the order, with the results to be included in this symposium report. MATERIALS AND METHODS Preferentially young stem pieces of the plant species listed in Table 1 were cut into longitudinal sections fixed in a formaldehyde-glutaraldehyde mixture followed by 1% buffered OsO4 and dehydrated in acetone. Small pieces containing phloem tissue were embedded and polymerized in epoxy resins, cut with an ultramicrotome, and the final ultrathin sections viewed and photographed with a transmission electron microscope (for exact procedures see Behnke, 1 982a). RESULTS AND DISCUSSION Sieve-elements of many of the myrtalean families are extremely difficult to fixate in a closeto-natural condition. There is less plasmatic content which is more labile and, during preparation, sieve-element plastids burst more often than in the average dicotyledon family. Nevertheless, all of the 69 investigated species of the core families of Myrtales (cf. Dahlgren & Thorne, 1984) could be shown to contain S-type plastids (cf. Table 1). Number, size, and shape of the starch grains in these plastids certainly vary (Figs. 1-14), but some common trends may be recognized. There are always several starch grains in one plastid-often six or more (Figs. 1-4, 6, 81 1), their size being variable (sometimes all of the same size) but never very large and their shape almost exclusively is a spherule. Considerably smaller starch grains are present in adI For permitting collection of living material used in this study the author wishes to thank the directors and curators of the Botanical Gardens in Berlin-Dahlem, Bogor, Bonn, Copenhagen, Edinburgh, Heidelberg, Kew, Leiden, Mainz, and Utrecht. I am particularly indebted to A. Assi (Abidjan), P. Berry (Caracas), D. Clark (Kirstenbosch), I. Dorr (Kiel), D. M. C. Fourie (Pretoria), I. S. Gottsberger (Botucatu), A. Irvine (Atherton, Qld), A. Juncosa (Petersham), K. Kubitzki (Hamburg), Lee Ying Fah (Sandakan, Sabah), G. Merz (Heidelberg), H. B. Nicholson (St. Michaels-on-Sea, RSA), H. T&h6 (ORSTROM), R. Tracey (Indooropilly) for collecting and shipping living plant material, A. P. Bennell (Edinburgh) and L. B. Jorgensen (Copenhagen) for helping with fixations, and A. R. A. Noel (Pietermaritzburg) and P. H. Raven (St. Louis) for making contact to collectors. I gratefully acknowledge the expert technical help of Mrs. L. Pop and Mrs. D. Laupp (both Heidelberg). This work was supported by grants from the Deutsche Forschungsgemeinschaft. 2 Zellenlehre, Universitat Heidelberg, Im Neuenheimer Feld 230, D-6900 Heidelberg, Federal Republic of Germany. ANN. MISSOURI BOT. GARD. 71: 824-831, 1984. This content downloaded from 157.55.39.35 on Thu, 01 Sep 2016 05:46:04 UTC All use subject to http://about.jstor.org/terms 1984] BEHNKE-SIEVE-ELEMENT PLASTIDS 825

Leaf Histology and Its Contribution to Relationships in the Myrtales

An examination of the leaf histology of a wide array of families thought at one time to be included in the Myrtales has led to a restricted circumscription of the order as follows: Lythraceae, Rhynchocalycaceae, Trapaceae, Oliniaceae, Combretaceae, Alzateaceae, Penaeaceae, Melastomataceae, Crypteroniaceae, Psiloxylaceae, Myrtaceae, and Onagraceae. Data used in testing this concept were obtained from lamina transsections cut to include the midrib. Observations on midrib shape, venation configuration, shape and extent of extraxylary fibers, sclereids, dermal and mesophyll features were most discriminating when examining taxa with the most elaborate and/or generalized anatomy. While leaf histological trends of specialization are not independently producible from this data, correlation with information from leaf architecture should eventually make this possible. Seldom recorded features which have proven useful in this systematic context include length/width ratios of palisade cells and the shape of the boundary between the mesophyll and the midrib ground tissue. The sharpness of tissue differentiation also appears often to be a family characteristic. In most respects, plant families comprising the core of the Myrtales form a coherent group and the recognition of the order has not been particularly controversial (Dahlgren & Thorne, 1984). At the same time, many of the included taxa pose problems as to their level of recognition as well as to which higher taxon they should be assigned within the order. In addition, more than 30 other families have been assigned to the Myrtales at some time in the recent past (Dahlgren & Thorne, 1984) demonstrating insufficient knowledge of the boundaries of the order as well as its evolutionary background. A review of the general literature on the circumscription of the order and the evidence used in its definition has been ably dealt with in the other papers from this symposium from the standpoint of general systematic review (Dahlgren & Thorne, 1984; Johnson & Briggs, 1984), wood anatomy (van Vliet & Baas, 1984), leaf architecture (Hickey, 1981), pollen (Patel et al., 1984), and sieve element plastids (Behnke, 1984). I will restrict my comments to the contribution to be made by leaf histology, recognizing that a clear understanding will depend upon a synthesis of all of the above plus data yet to be obtained. Some formidable problems must be faced when one attempts to use features from young vegetative anatomy in a systematic investigation. First is the insufficient number of studies of this sort to provide any reliable trends of specialization. I am convinced, however, that leaf histology contains not only valuable diagnostic characters, but that these characters will eventually demonstrate great utility in evolutionary studies. Much of the value will be based on both an increasing number of systematic studies of leaf histology, as well as character correlation with studies of leaf architecture (cf. Hickey, 1981). Leaf structure studies of both types should continue to accumulate rapidly. Leaves are obviously the most easily obtained of all plant materials both from the field and from herbarium collections. While leaves respond readily to evolutionary pressures toward xeromorphy or other modifications, their endomorphic characters commonly retain the ground plan of the phylad to which they belong (Keating, 1984; Dickison, 1970). In this study a particular problem concerns the interpretation of the data from my sample. The sample appears comprehensive in that 176 species were examined representing 55 genera from 19 families. Actually, the specimens at hand amount to a very small sample of many of the families. Even though attempts were made to choose specimens representing geographic and taxonomic diversity within each family, obviously the family samples most likely do not include the total spectrum of characters to be found in them. The absence of a character from one group does not I This work was partially supported by National Science Foundation grants DEB 77-15571 to the author and by DEB 78-23400 to Peter H. Raven. I am grateful to curators from many institutions for liquid preserved specimens and to P. H. Raven for his assistance in obtaining many specimens. 2 Department of Biological Sciences, Southern Illinois University, Edwardsville, Illinois 62026. ANN. MISSOURI BOT. GARD. 71: 801-823. 1984. This content downloaded from 157.55.39.178 on Fri, 05 Aug 2016 05:09:46 UTC All use subject to http://about.jstor.org/terms TABLE 1. Selected histological features of leaf transections of families of Myrtales. O