Search

PREMISE OF THE STUDY Hybridization is an important mechanism of speciation in plants and often results in complexes that comprise multiple sexual diploids and their hybrid offspring. However, the intricacy of these systems has prevented a thorough understanding of many groups. The Crepidomanes minutum species complex (Hymenophyllaceae) is a widely distributed, morphologically variable fern species complex. Although prior reports of apogamy and polyploidy suggest hybridization, it has never been the focus of a phylogenetic study. METHODS Morphology, nuclear (gapCp), and chloroplast (rbcL) DNA sequences, cytology, field observation, and spore counts were used to infer phylogeny and trace hybrid origins. KEY RESULTS The C. minutum species complex is composed of at least three major clades: the African clade, clade 1 (East Asia and the Pacific), and clade 2 (Southeast Asia and the South Pacific). Clades 1 and 2 differ strikingly in morphological variation (uniform in clade 1 vs. highly variable in clade 2) and occurrence of hybrids (rare in clade 1 vs. frequent in clade 2). Apogamy and polyploidy were confirmed as likely mechanisms of hybrid stabilization in clade 2. Despite the large genetic distance between clades 1 and 2, several specimens were observed with gapCp sequences from both; diploid genome size and sexual reproduction indicate maintenance of genetic diversity via introgression or incomplete lineage sorting, rather than ongoing hybridization, in these specimens. CONCLUSIONS The C. minutum species complex is a reticulate network including multiple diploid lineages and their stabilized hybrid crosses. Additional sampling focused on reproductive mode and ploidy level is needed to delimit diploid species and hybrids.

Two taxa of cupulate magnoliid fossil flowers, Cronquistiflora and Detrusandra, are described from the Late Cretaceous (Turonian, ;90 million years before present [MYBP]) Raritan (or lower Magothy) Formation of New Jersey. The fossil taxa are represented by flowers at various stages of development, associated fragments of cup-shaped floral receptacles with attached anthers, and isolated anthers. Both taxa have laminar stamens with adaxial thecae and valvate dehiscence. Pollen is boat-shaped and foveolate in anthers associated with Cronquistiflora and spherical with reticulate ornamentation in Detrusandra. Cup-shaped receptacles are externally bracteose in both taxa. The receptacle of Cronquistiflora is broader than the campanulate one of Detrusandra. Cronquistiflora also has more carpels ( ;50 in a spiral vs. ;5 in a whorl or tight spiral). In Detrusandra the carpels are surrounded by dorsiventrally flattened structures (pistillodes?) that are remote from the attachment of the stamens near the distal rim of the receptacular cupule. Detrusandra stigmas are rounded and bilobed, while those of Cronquistiflora, although bilateral in symmetry, are somewhat peltate. The fossil taxa share prominent characters with extant cupulate magnoliids (e.g., Eupomatia, Calycanthus), but also share characters with other magnoliids including Winteraceae. These fossils represent taxa that are character mosaics relative to currently recognized families. Inclusion of these fossils in existing data matrices and ensuing phylogenetic analyses effect changes in tree topologies consistent with their mosaicism relative to modern taxa. But such analyses do not definitively demonstrate the affinities of the fossils other than illustrating that these fossils are generalized magnoliids. Additional analysis of modern and fossil magnoliids is necessary to fully appreciate the phylogenetic significance and positions of these fossil taxa. However, the results of the phylogenetic analyses do introduce the possibility that extinct taxa of Magnoliales with cupulate floral receptacles were transitional between basal angiosperms and those with tricolpate pollen. The fossils provide insights into the timing of evolution of character complexes now associated with coleopteran pollination.

The Turonian flora from Sayreville New Jersey includes one of the world's most diverse assemblages of Cretaceous angiosperm flowers. This flora is made even more interesting by its association with a large insect fauna that is preserved by charcoalification as well as in amber. Floral diversity includes numerous representatives of Magnoliidae, Hamamelididae, Rosidae, Dilleniidae, and Asteridae (Ericales sensu lato). Included are hypogynous, five-merous flowers with uniseriate hairs on the pedicels and stamens in bundles most frequently borne opposite the petals. There is considerable variation in filament length, and some filaments are branched. On some anthers, strands of residue, suggesting the former presence of a liquid of unknown nature, partially occlude the apparent zone of dehiscence. In other cases, open anthers are fully occluded by an amorphous substance. Pollen is rarely found associated with anthers, but is common on stigmatic surfaces. Pollen is prolate and tricolporate with reticulate micromorphology. The superior syncarpous ovary is five-carpellate with axile/intruded parietal placentation and numerous anatropous ovules/carpel. Ovary partitions have closely spaced, parallel ascending channels (secretory canals?), and there are apparent secretory canals/cavities in receptacles, sepals, and petals. Individual stigmas are cuneiform with a central groove and eccentrically peltate. Styles are short and fused. In aggregate, the stigmas form a secondarily peltate stigma. Seeds have a reticulate sculpture pattern, a pronounced raphe, and funicular arils with sculpture similar to the seeds. Phylogenetic analyses of several data matrices of extant taxa place this fossil in a monophyletic group with the modern genera Garcinia and Clusia within the Clusiaceae. As such, these fossils represent the earliest fossil evidence of the family Clusiaceae. Some modern Clusiaceae are notable, in particular, for their close relationship with meliponine and other highly derived bee pollinators; the fossil flowers share several characters that suggest a similar mode of pollination. This possibility is consistent with other floral and insect data from the same locality.

In a broad survey of anther glands in the mimosoid legume tribes Mimoseae and Parkieae, representatives from 30 genera with anther glands were studied using scanning electron microscopy. Four kinds of anther glands could be distinguished. The Piptadenia-type gland, found in all but four of the genera surveyed, is usually spherical to ellipsoid in shape and often borne on a stipe. The cells making up the gland vary in size among species and are often sculptured. Six different kinds of sculpturing can be recognized: smooth, reticulate, striate, rugulate, scabrate, and papillate. The Gagnebina-type gland is the least specialized, consisting of a simple extension of the connective with irregularly projecting cells. The Prosopis africana-type gland is borne ventrally between the thecae, the connective extending hump-like over the apex of the anther. The Pentaclethra-type gland, found only in the genus Pentaclethra, is distinguished by a median dorsal furrow and a ventral conical structure similar to a food body or osmophore. Anatomical sections revealed two different subtypes within the Piptadenia-type gland. Some glands are composed of a homogeneous matrix of isodiametric cells, whereas others have two layers: a peripheral layer of large radially elongated cells, and a central sphere of smaller isodiametric cells and large air spaces. Some characters of anther glands have clear taxonomic significance, and more detailed surveys within genera will undoubtedly provide additional taxonomic markers.

Th e caption Liliacidites peroreticulatus to the pollen grain illustrated in Fig. 7I is incorrect. Th e morphotype illustrated is in fact undescribed but is provisionally assigned to Arecipites because of the very fi ne nature of the reticulate ornamentation (e.g. compare descriptions of Arecipites , Card 166, and Liliacidites , Cards 1489, 4280 and 4462, in Jansonius and Hills, 1976 ). Th is morphotype also occurs in a number of Maastrichtian deposits in southern Australia. We thank Professor Emeritus James A. Doyle (Earth and Planetary Sciences and College of Biological Sciences, University of California, Davis) for alerting us to the error. Th e online article at http://www. amjbot.org/content/102/12/2092.full.pdf+html has been corrected.

Xylem of the orchids studied provided unusually favorable material to demonstrate how conductive tissue evolves in monocotyledons. In the end walls of tracheary elements of many Orchidaceae, remnants of pit membranes were observed with scanning electron microscopy and minimally destructive methods. The full range from tracheids to vessel elements, featuring many intermediate stages, was illustrated with SEM in hand sections of fixed roots, stems, and inflorescence axes of 13 species from four subfamilies. Pit membranes in end walls of tracheary elements are porose to reticulate in roots of all species, but nonporose in stems of Cypripedioideae and Vanilloideae and porose to reticulate in stems of Orchidoideae and Epidendroideae. The distribution pattern of pit membranes and pit membrane remnants in end walls of tracheary elements of orchids parallels the findings of others. The position of Cypripedioideae and Vanilloideae as outgroups to Orchidoideae and Epidendroideae, claimed by earlier authors, is supported by clades based on molecular studies and by our studies. Little hydrolysis of pit membranes in tracheary element end walls was observed in pseudobulbs or inflorescence axes of epidendroids. The pervasiveness of network-like pit membranes of various extents and patterns in end walls of tracheary elements in Orchidaceae calls into question the traditional definitions of tracheids and vessel elements, not merely in orchids, but in angiosperms at large. These two concepts, based on light microscope studies, are blurred in light of ultrastructural studies. More importantly, the intermediate expressions of pit membranes in tracheary element end walls of Orchidaceae and some other families of angiosperms are important as indicators of steps in evolution of conduction with respect to organs (more rapid flow in roots than in succulent storage structures) and habitat (less obstruction to flow correlated with a shift from terrestrial to epiphytic).

The origin and processes creating the high diversity of plant species in neotropical rain forests and their floristic composition and multistratitified forest structure are still uncertain. Here, we studied one of the most common leaf morphotypes of the Cerrejon flora (middle-late Paleocene, ca. 60-58 Ma), Guajira, Colombia, that contains one of the oldest records of neotropical rain forest floras. Fifty-seven leaf specimens were carefully examined with a focus on general morphology, venation patterns, and cuticular characteristics. The analysis allowed us to recognize four new species that were assigned to the fossil-leaf genus Menispermites on the basis of an ovate leaf shape with cordate to truncate bases, actinodromous primary venation, brochidodromous secondary venation, percurrent tertiary venation, regular polygonal reticulate fourth and fifth venation, well-developed polygonal areoles, entire margin, and the presence of a fimbrial vein. This set of characters suggests a possible affinity with the pantropical angiosperm family Menispermaceae. The predominantly climbing habit of this family suggests that the Cerrejon Paleocene tropical rain forest was already multistratified. These findings represent the earliest record for the family in northern South America.

Although reticulate evolution has been a frequent occurrence during the history of plants, determining how it has contributed to plant evolution will require analyzing many nuclear loci and developing effective analytical methods. The objective of this study was to make progress toward meeting these requirements in the evolutionarily complex subtribe Machaerantherinae. The 5S ribosomal RNA (rRNA) spacer was investigated to characterize its structure and variation. Analysis of the spacer supported relationships that were mostly the same as those supported by ITS and ETS sequence data. Two methods were used to estimate support for 11 occurrences of incongruence between 5S/ITS/ETS and cpDNA data sets. The fi ve best-supported incongruences were proposed to have had reticulate evolutionary histories. For Arida blepharophylla , Xanthisma rhizomatum , and Pyrrocoma , 5S and ITS/ETS evidence supported the same or similar relationships, indicating that these two regions of the nuclear genome were descended from the same ancestor or from two closely related ancestors, with cpDNA coming from a more distantly related ancestor. The 5S and ITS/ETS evidence disagreed on the relationships of Arida riparia , suggesting that its ITS/ETS region came from one ancestor and its 5S region and cpDNA from a different ancestor.

Fossil angiospermous stamens with in situ pollen from the Turonian (ca. 90 million years before present, Late Cretaceous) of New Jersey are described and assigned to the Chloranthaceae. The fossil stamens, which are three-parted and bear two bisporangiate thecae on the central lobe and one bisporangiate theca on each lateral lobe, are indistinguishable from stamens of several extant species of Chloranthus. The pollen is spheroidal, 13-18 gtm in diameter, with a reticulate exine and apparently elongate/elliptical apertures. The pollen is similar to that in extant Chloranthus in grain size, shape, exine sculpture, and aperture structure. Like pollen of some extant species of Chloranthus, aperture number in the fossil pollen appears to be variable. Because fossil pistillate chloranthoid reproductive structures have not been found at this locality it is unknown whether the fossil stamens described here were bome on the side of the ovary, as in extant Chloranthus, or in another arrangement. The three-parted stamen of Chloranthus is unique in angiosperms and there has been considerable debate conceming the origin and evolutionary significance of the structure. Uncertainty as to whether the three-parted stamen represents a synapomorphy for the genus or a retained plesiomorphy in angiosperms is the primary reason why these fossil stamens are not assigned to the extant genus Chloranthus. The Chloranthaceae is one of several families that have been of special interest in studies of the origin and early diversification of angiosperms. Evidence from the fossil record, comparative morphology and anatomy of extant plants, and phylogenetic analyses suggests that taxa similar to the Chloranthaceae were present and may have

Nordenskioldia borealis Heer is described and its systematic position is reassessed based on examination of the type material and specimens recently collected from three Paleocene localities in North America (Almont, North Dakota; Melville, Montana; Monarch, Wyoming). The morphology of Nordenskioldia infructescences and fruits is clarified, and in particular, silicified specimens from Almont provide new details of fruit and seed anatomy. Fruits are schizocarpic, and individual fruitlets also dehisce to release flat reticulate seeds. These seeds occur in many Paleocene floras but have not been linked previously to Nordenskioldia. Anatomical details of infructescence axes are identical to those of distinctive long and short shoot systems that cooccur with Nordenskioldia, and neither the infructescence axes nor shoots have vessels in the secondary xylem. Comparison of the floras at Almont, Melville, and Monarch with those at other Paleogene localities in Asia, Europe, and North America provides association evidence supporting earlier conclusions that the Nordenskioldia plant bore simple, entire- to crenatemargined leaves with actinodromous venation. Such leaves have been previously assigned to extant genera such as Cercidiphyllum, Cocculus, and Populus but are treated here as Zizyphoides flabella (Newberry) comb. nov. Based on the combined morphological and anatomical details now available, the Nordenskioldia plant is assigned to the Trochodendrales as an extinct genus most closely related to extant Trochodendron.

Phylogenetic relationships among the 21 taxa comprising Helianthus sect. Helianthus and three outgroup species were assessed by restriction site mapping ofthe 18S-25S nuclear ribosomal RNA gene family. Wagner parsimony analysis of the 41 restriction site or length mutations observed produced a single 59-step most parsimonious tree. This tree was then compared to a cytoplasmic-based plastid phylogeny for this group. Several major discrepancies were observed between the two phylogenies suggesting both recent and ancient introgression. Furthermore, three cases of diploid hybrid speciation are unambiguously documented and a fourth case is suggested. These data are interpreted to suggest that evolution in Heliathus is reticulate rather than exclusively dichotomous and branching.

Transmission and scanning electron microscopy of exine development in Bougainvillea spectabilis (Nyctaginaceae) confirmed that the exine pattern is initiated by invagination of the microspore plasma membrane at the early tetrad stage. Invaginated plasma membranes take the form of a reticulate pattern that corresponds to the mature exine tectum. Protectum is the first exine layer to be deposited on the reticulate-patterned plasma membrane. Subsequently, probacules elongate basally on protruding sites of the plasma membrane under the protectum and in the lumina. These sites retreat as the probacules elongate. After the dissolution of the callose wall, a foot layer forms through the accumulation of lamellated structures. Clearly, the plasma membrane serves a determinative role in the initial pattern formation of exine. Deposition of primexine at the microspore tetrad stage has been traditionally regarded as initiating exine formation (Heslop-Harrison, 1963). The primexine, a microfibrillar cellulosic layer beneath the callose wall, functions primarily as matrix into which probacules are formed in the shape of radially directed rods. This distribution pattern determines the principal structural features of the mature exine. However, pollen developmental processes in some taxa are different from the primexine scheme. Dickinson (1970) proposed in Lilium that individual probacular initials were first formed in a primary fibrillar layer, and a rapid synthesis of lamellate material during the withdrawal of the plasma membrane caused a proliferation of the "tops" of these structures so that they met with their neighbors to form muri. Later, Dickinson and Sheldon (1986) modified

The transition from helical to reticulate wall patterns in the conducting elements of the xylem was located in etiolated soybean hypocotyls in conjunction with the distribution of elongation in the axis. The most acropetal reticulate elements are interpreted as the most acropetal metaxylem. Evidence is given that these elements experience strain as they mature. Predicted strain correlated significantly with maximum relative elemental rate of elongation (RELEL), which ranged from 0.08 to 0.12 hr-1. RELEL varied from 0.03 to 0.06 hr-1 at the point of transition from helical to reticulate xylem element wall patterns. Estimates of strain to be experienced in the reticulate elements varied from 0.1 1 to 0.13 in untransplanted plants and from 0.04 to 0.1 1 in plants transplanted to the same, fully watered medium. Transplanting suppressed elongation in the region of maximum RELEL for ca. 12 hr and caused an upward migration of the region of depressed rates on the outside of the axis that accounts for hook-opening. Attention is called to the fact that this region of depressed growth corresponds with the onset of transverse enlargement of the hypocotyl axis.

A B S T R A C T A comparative study of the leaves of 31 species of Lycopodiirn was made. The pattern of tracheids varies with the species. Usually they exhibit annular, helical, reticulate, modified or transition forms; none of the species have advanced phylogenetically beyond the scalariform tracheid stage. Mucilage canals have been found in four species. Tracheids are enclosed by a few to several layers of oblique ended, elongated parenchyma cells in most species, but an endodermis is absent. Vascular bundles are encircled by fiber cells in three species. The mesophyll of most species consists of identically shaped cells, although palisade-like tissue has been observed in a few species. The epidermal cells vary from elongated to isodiametric and have either undulated or smooth anticlinal walls, which are deeply pitted in some species. The outer epidermal walls are usually thick and heavily cutinized. Stomates are distributed on both surfaces in 18 species, on the abaxial surface in 11 species, and on the adaxial side in only two species. Most stomates are practically isodiametric in surface view, broader than epidermal cells, usually parallel to the vein, and at the same level as the adjacent epidermal cell. A typical guard cell has a prominent outer ledge and a less developed inner ledge of cutin in most species. Six groupings are suggested, based on similarity of leaf structure and the known chromosome numbers.

Pollen development in Hibiscus syriacus L. (Malvaceae) was studied with light (LM), scanning (SEM) and transmission (TEM) electron microscopes, with special attention to the formation of extremely long spines of the pollen grains. At the early tetrad stage, probacules are initiated directly on the plasma membrane and grow in coincidence with the height of primexine matrix within a callosic wall. Subsequently, a pretectum appears at the top of the probacules and then a foot layer is formed by accumulation of white line centered lamellations. Before dissolution of the callosic wall, a reticulate patterned pretectum is established around the microspores. There is not, however, any morphological indication on the initiation of the spines during the tetrad period within a callosic wall. It is after dissolution of the callosic wall that the spines of exine begin to form by the apposition of lamellated sheets. The lamellated sheets show a concentric configuration around the developing supratectal spines. The mature pollen grain is spheroidal, polycolporate, 160-170 ,tm in diameter, with supratectal spines 20-25 ,4m long. The supratectal spines of Hibiscus pollen are not homologous with the other exinous protrusions which are determined within the callosic wall during tetrad stage.

Pollen of the scrophulariaceous genera Amphianthus, Artanema, Curanga, Glossostigma, and Peplidium was examined by light and scanning electron microscopy. All have single, isopolar, radially symmetrical pollen grains with equally spaced meridional colpi and columellate exines. Intergeneric differences are established mainly on the differential correlation of colpus number, colpus morphology, and sculpturing of the mesocolpium. The colpi vary in length and range in number from 3 to 6. The colpus membrane can be variously granular, verrucate, rugulate, or spinulose. The exine varies from semitectate and microreticulate or reticulate to tectate and sparsely perforate. It may be psilate or ornamented with supratectal processes ranging from small granules to relatively long spinules or bacules. All the genera examined, with the possible exception of Artanema vs. Amphianthus, are distinguished from one another by pollen morphology, and the grains of both Curanga and Glossostigma have morphotypes not previously reported in tribe Gratioleae. Final taxonomic assessment of the pollen data in the genera examined must await further comparative studies in the Gratioleae and other tribes, but the morphological diversity observed is clearly of potential significance in evaluating specialization levels and clarifying current conflicting interpretations of relationships in these taxa. THE PRESENT REPORT examines the pollen morphology of Artanema D. Don, Amphianthus Torr., Curanga Juss., Glossostigma Wight & Am. ex Am., and Peplidium Delile, five infrequently studied, small scrophulariaceous genera variously considered to be closely related to one another or to have their affinities in divergent areas. Artanema, with four or five

Reticulate thickenings in Juglans nigra, J. major, J. microcarpa, and J. californica (including J. hindsii) appear, under the transmitted light microscope, to be a cluster of very large or "gash-like" vessel-ray pits. The use of scanning electron microscopy shows that these so-called pits are thickenings on the inside wall of the vessel. RETICULATE THICKENINGS ("gash-like" vessel pits) have been reported in some species of Juglans, but these structures have never been adequately described or photographed. In the late 1940's Dr. B. F. Kukachka of this Laboratory noticed the presence of what appeared to be very large or "gash-like" pits in some vessels of J. nigra and found that this character could be used for separating J. nigra from J. regia because this feature was lacking in the latter. Carpenter and Leney (1952) illustrated latewood vessel elements of J. nigra that had "reticulate intervessel pitting." In their "Identification of Hardwoods" Brazier and Franklin (1961) stated that J. nigra has very large "gash-like" pits in the vessels. Also, Stark (1953) reported "gash-like" intervessel pits in J. nigra, J. major, and J. microcarpa. No mention of "gash-like" pits were made in the descriptions of Juglans by Solereder (1908), Kribs (1927), Heimsch and Wetmore (1939), Record and Hess (1943), Metcalfe and Chalk (1950), Jane (1956), and Panshin and de Zeeuw (1970). However, Kribs (1927), Heimsch and Wetmore (1939), and Record and Hess (1943) described the vessel-ray pits in Alfaroa and Engelhardia as large and "gash-like." Their use of "gash-like" refers only to the typical coarse vessel-ray pits characteristic of these two genera. A recent investigation at the Forest Products Laboratory on the wood anatomy of Juglans confirmed the diagnostic value of reticulate thickenings. Juglans nigra, J. major, J. microcarpa, and J. californica (including J. hindsii) were the only species in which reticulate thickenings were found. Observations indicated that these supposedly very large vessel-ray pits occasionally were not in contact with rays. Thus, in an effort to determine which cells subtended these so-called "gash-like" I Received for publication 23 May 1975. The author gratefully acknowledges the technical assistance of Richard Kinney with the SEM microscopy and specimen preparation. The Forest Products Laboratory is maintained at Madison, Wis., in cooperation with the University of Wisconsin. pits the SEM (scanning electron microscope) was employed. METHODS AND MATERIALS-A wood sample of J. nigra (MADw 8986, collected in the Missouri Botanical Garden Arboretum, Grays Summit, Mo., by Beilmann, BWCw 8274) and known to have an abundance of reticulate thickenings was examined with the SEM. Blocks were split or smoothed with a microtome knife to expose either the radial or transverse face. Thin sections (51O 1/m) were also examined with the SEM. In all cases the specimens were coated with gold. Latewood material was macerated in a solution of equal parts of nitric and chromic acid (Jeffery's solution), and then washed thoroughly in water. A small portion was then placed on a slide and gently teased to separate the vessels from the fibrous elements and parenchyma. With no cover slip applied, the macerated material was examined under a transmitted light microscope to find a vessel member with reticulate thickenings. After such a vessel member was found, it was maneuvered so that no other cells were near and the reticulate thickenings were facing upward or on the top surface. The specimen was allowed to dry and then coated with gold and examined with the SEM. OBSERVATIONS AND DISCUSSION-Reticulate thickenings generally occur only on the radial walls of the smaller latewood vessel members, and resemble vessel pits under the transmitted light microscope. It was originally thought that reticulate thickenings were some type of very large or "gash-like" vessel-ray pits. Figure 1 shows reticulate thickenings under the transmitted light microscope. The space between the thickenings appears to be an opening, thus the concept of vessel-ray pits. The absence of rays near the reticulate thickenings led to the examination of these structures with the SEM. In Fig. 2, reticulate thickenings can be seen on a vessel member which was separated from all other cells by maceration. Fig. 3 shows the same

The phloem of Etapteris leclercqii and Botryopteris tridentata petioles is described from Lower Pennsylvanian coal balls. Petioles of B. tridentata are characterized in transverse section by an omega-shaped xylem trace, a phloem zone which extends from 2-10 cells in width, and 2-parted cortex. Etapteris leclercqii petioles exhibit a 4-9 cell-wide phloem zone surrounding the central clepsydroid xylem mass, and a 3-parted cortex. In both taxa a 1-2 cell layer parenchyma sheath separates the xylem from the extra-xylary tissues. The phloem of both species consists of sieve elements that average about 20 ,um in diam by 200 ,m in length in Botryopteris, and 100 Am in length in Etapteris, with horizontal-slightly oblique end walls. In transmitted light, the radial walls of the sieve elements form an irregular reticulate pattern enclosing elliptical lighter areas. With the scanning electron microscope, these areas appear as horizontal-slightly oblique furrows on the cell wall, with many small indentations lining the furrows. These indentations, because of their regular occurrence and size (from a few fractions of a micron up to 1.0 Atm in diam), are interpreted as sieve pores, and the elliptical areas that enclose them as sieve areas. The phloem of E. leclercqii and B. tridentata is compared with that described for other fossil genera and with that of extant ferns.

Pollen of the scrophulariaceous genera Melosperma and Monttea was studied with light and scanning electron microscopy. Both have single, radiosymmetric, isopolar, trizoniaperturate pollen grains with usually ruptured colpus membranes, tectate-perforate to semitectate exines with a columellate interstitium and pitted or perforate nexines. Intergeneric differences in pollen morphology are established based on the differential correlation of sculpturing characters. Comparisons of the pollen of Melosperma and Monttea with that of the tribe Minuleae sensu stricto reveals no difference in the case of Melosperma and the existence of intermediate morphologies and "reticulate" similarities in the case of Monttea; the taxonomic significance of this morphological intergradation is considered. The transfer of Oxycladus aphyllus to Monttea is consistent with the pollen evidence, but proposed affinities between Monttea and tribe Antonieae of the Loganiaceae are not supported.

A B S T RA CT The earliest attempts at host penetration by the primary haustorium of Phthirusa pyrifolia (H.B.K.) Eichl. were studied utilizing the techniques of light and electron microscopy. Cells located in the central portion of the haustorial core form a lens-shaped structure, the gland. Gland cells are of two kinds, possessing either a dense-staining cytoplasm or a light-staining one. The ultrastructural features of these unique cells are described. A cavity forms at the base of the gland and fills with a fibrillar substance, which appears morphologically similar to that characterizing the core parenchyma, and the contents of some disintegrated gland cells. The gland cavity extends through the contact zone, where its contents exude onto the host's surface. The primordial endophyte is initiated above the gland and grows through it to the host's surface. If not successful in penetrating the host, this organ may spread laterally between the host and the old contact layer. A new contact layer is subsequently formed. Several successive penetration attempts may be made until one is successful. Strands of vessel members may be followed from the hypocotyl to the gland. Along these strands the pitting pattern changes from scalariform to reticulate, respectively. THIS REPORT iS the second in a series on investiga

IN THE COURSE of a palynological study of the Araliaceae, a unique pollen type was found in Boerlagiodendron, a genus of about 40 species, distributed from Taiwan, the Philippines, Marianas, Palau Islands, and Malaysia through New Guinea and the Solomons, to the New Hebrides (Smith and Stone, 1968). It is one of the most distinct genera in the family because of its sympetalous corolla, pseudofruits, and fringed petiole (Philipson, 1970). The only palynological description of Boerlagiodendron in the literature is by Huang (1967) for the Taiwanese species B. pectinatum: ". . . colpi nearly as long as the length of P-axis; ora obscure; exine psilate to scabrate, with warts only at intermediate lobes . . . sexine . .. reticulate . . ." The present study, however, has brought to light certain differences with Huang's observations with regard to the sexine patterns and the apertural structure. Critical studies of the pollen of this genus, including that of B. pectinatum supplied by Huang, have revealed that the unexpanded pollen grains and the inconspicuous apertures may have caused his misinterpretation of the pollen structure of the species he studied. The present paper provides an account of the pollen morphology of eight species of Boerlagiodendron, with special reference to that of B. tetrandrum. An attempt has been made to compare several pollen features with other floral characters as well as with the geographical location of each species in order to reveal any significant correlations existing among them.

Forty-nine taxa of Sansevieria were examined to provide the first detailed description of leaf anatomy in the genus, as well as to determine the cellular organization of the water-storage tissue present in the central mesophyll of all leaves. Leaf form ranges from flat through cylindrical, with varying degrees of xeromorphic characteristics, including cuticle thickness, stomatal depth, fiber content, and fiber cell development. Mesophyll in all species is divided into an outer region of chlorenchyma and a central region of colorless water-storage tissue. The water-storage tissue comprises a highly branched 3-dimensional network of living cells, among which are many dead, thin-walled water-storage cells. Species with more extreme xeromorphic characteristics tend to have a greater percentage of their water-storage tissue composed of water-storage cells. In 28 taxa, water-storage cells have spiral or reticulate wall bands. These taxa tend to be more xeromorphic. The wall bands may serve an important structural role in drought tolerance.

Inflorescences from the Claiborne Formation of western Tennessee are remarkably similar to those of the tribe Hippomaneae, subfamily Euphorbioideae, of the Euphorbiaceae. The fossil inflorescences are spikes of bract-subtended cymules of at least three florets each. Florets are composed of at least three stamens. Palynological features of the fossils are also shared by the Hippomaneae. Fossil pollen is tricolporate, prolate (26.9 x 20.6 ,um; P/E = 1.3), with lalongate pores. Exine structure is tectate columellate with a perforate tectum. The exine is reticulate and the muri are conspicuously striate. These specimens represent the first fossil floral evidence of the Euphorbiaceae. It is surprising that inflorescences of the Hippomaneae so modern in aspect existed in the Middle Eocene, since the tribe is universally considered to be one of the most advanced in the family.

The pollen of three closely related genera, Aetoxylon, Amyxa, and Gonystylus is compared in SEM and TEM with that of Thymelaeaceae, s. s. The Thymelaeaceae have spherical, pantoporate grains with a crotonoid tectum in which the basic subunit is tnangular in shape and forms a continuous triangular array. Thin section (TEM) and fractures (SEM) revealed that these subunits are attached to a ringlike network of horizontal rods. Within the Thymelaeaceae, the triangular subunits vary in the number of subdivisions and degree of fusion and form a morphological continuum. Aetoxylon, Amyxa, and Gonystylus also have spherical, pantoporate pollen but with a tectum in which almost all of the distinction of the subunits appears to have been lost. The structure of the exine in Aetoxylon, Amyxa, and Gonystylus, however, is unique thus far within the angiosperms. Thin section revealed a thick tectum with a layer of short or even granular columellae, then a thin, discontinuous layer from which larger columellae appear to hang. There is no evidence of an endexine even in the region of the apertures. The distinctive exine structure would support the treatment of Aetoxylon, Amyxa, and Gonystylus as a separate family, Gonystylaceae, allied to the Thymelaeaceae. AETOXYLON (Airy Shaw) Airy Shaw, and Amyxa Van Tiegh., each with a single species in Borneo, and Gonystylus with 25 species distributed through most of Malaysia, the Solomon Islands, and Fiji, have been treated as a separate family, Gonystylaceae, by Hutchinson (1959), or included in the Thymelaeaceae by Domke (1934), Wagenitz (1964), Airy Shaw (Willis, 1966), and Cronquist (1981), all of whom acknowledge the distinction of these three genera by segregating them as a subfamily, Gonystyloideae. Aetoxylon, Amyxa, and Gonystylus can be distinguished from the remaining Thymelaeaceae by a combination of characters: capsule fruit, numerous petals, elongate style, and pellucid-punctate leaves. These genera also lack the internal phloem that characterizes the vast t Received for publication 6 August 1984; revision accepted 5 February 1985. We thank J. L. Bittner, S. G. Braden, and W. Chissoe for their fine technical assistance. We are also grateful for the pollen samples sent by the late Dr. Jan Muller. This study was supported in part by a grant from Fluid Research from S. Dillon Ripley, and National Science Foundation Grants BSR-8002585 and BSR-8216349. majority ofthe Thymelaeaceae, as well as most of the order Myrtales. Archangelsky (1971 ), who described the pollen morphology of the Thymelaeaceae and allied families (1971), referred to their predominant crotonoid pattern as stellate. He considered the majority of the Thymelaeaceae to have typical stellate sculpturing, but described the Gonystylaceae as having cryptostellate sculpturing, and cited the "absolutely special type of their exines as a further criterion" for their treatment as an independent family. Of further interest, Archangelsky suggested that stellate sculpturing was one of the complex variables of simple reticulate sculpturing. As part of an extensive study of pollen morphology in the orders Myrtales (Patel, Skvarla and Raven, 1983, 1984) and Euphorbiales (sensu Cronquist, 1981; Nowicke, 1984; Nowicke and Skvarla, 1984), the pollen of more than 70 species representing 39 genera in the Thymelaeaceae have been examined. Aetoxylon sympetalum, Amyxa pluriconis, and four species of Gonystylus were included (see Table 1).

The structure of tracheids in Lycopodium lucidulum, L. clavatum, and L. tristachyum was studied with the light microscope. Protoxylem development is at least sometimes and possibly always mesarch in indeterminate axes of all three species. Centrifugally formed protoxylem elements are reticulate and discontinuities in the secondary walls of these elements are sometimes conspicuously bordered. Wall thickenings of first formed protoxylem elements consist mainly of indirectly connected rings. Late centripetally formed protoxylem elements and transitional elements have a reticulate secondary wall pattern. The narrowest metaxylem elements have circular bordered pits while in wider metaxylem elements pits are bordered and may vary from circular to scalariform. Pitting is uniseriate to triseriate in tracheids of all three species, and intermittent tetraseriate pitting was occasionallv observed in L. lucidulum. Crassulae occur in tracheids of the three species, and in L. clavatum an additional framework, probably representing thickened compound middle lamella, is also present. Pits often appear helically arranged, and in all three species pits are connected by thin areas in the secondary wall. Macrofibrils approximately 0.5 ,uwide were observed in tracheids of the three species. In L. clavatum the arrangement of macrofibrils was predominantly bidirectional. FEW EARLIER workers expressed particular interest in tracheids of Lycopodium, and their descriptions of these usually comprised only small portions of broader studies (Hegelmaier, 1872; Wigglesworth, 1907). The most thorough study was made by Bierhorst (1960) and was based on observations of eight species (including L. clavatum and L. lucidulum, two of the three species included in the present work). But Bierhorst's observations of tracheids in the genus were limited since they constituted only a small fraction of his extensive studies of tracheary elements. Of the eight species he studied, Bierhorst illustrated tracheids of two, L. selago and L. obscurum. This report verifies most of Bierhorst's generalizations regarding tracheids of Lycopodium. However the tracheids of three species of this genus are described in greater detail and some different conclusions are presented. The work also parallels a study by Zamora (1960) of protoxylem in Selaginella. The genus Lycopodium undoubtedly has had a long history and similar extinct genera in the Lycopsida are known from Devonian times. Banks (1968) recently proposed, in fact, that the Lycopsida and other vascular plants may have 1 Received for publication 13 April 1970. This paper represents part of a masters' thesis submitted in partial fulfillment of the requirement for a M.S. degree in Biology at the State University of New York, Binghamton. The author expresses deep appreciation to Dr. James D. Grierson who advised him during the course of this study. He also thanks Dr. David W. Bierhorst for reviewing the manuscript and Miss Laura Gabel for her technical assistance. evolved separately from two basal stocks respectively, the zosterophyllophytes and rhyniophytes. It is clear, however, that in assessing the origins of the Lycopsida and their relationships to other vascular plants, we would benefit from a more detailed knowledge of both extinct and extant lycopods. It is hoped that the information and documentation presented here will contribute towards that end. MATERIALS AND METHODS-Material for the study of each of the three species was collected from at least two locations in central New York State. All of the species are common in this region although L. tristachyum is least abundant. Late in this investigation it appeared that plants from one locality,iwhich I orginally considered to be L. tristachyum, deviate slightly from typical L. tristachyum. Dr. R. T. Clausen kindly examined these plants in the field and concluded that they were a hybrid of L. tristachyum and L. complanatum var. flabelliforme that he described earlier (Clausen, 1945). Dr. J. Wilce, who recently monographed Section Complanata of Lycopodium (Wilce, 1965) also examined pickled specimens of these plants. She concluded, however, that the plants were most similar to L. tristachyum whether or not they represented a hybrid. In light of the above suggestions, a synthesis of characteristics of true L. tristachyum and a closely related plant, perhaps a hybrid, is presented. Voucher specimens are on deposit in the herbarium, University of M\lassachusetts, Amherst.

FOSTER, ADRIANCE S. (U. California, Berkeley), and HOWARD J. ARNOTT. Morphology and dichotomous vasculature of the leaf of Kingdonia uniflora. Amer. Jour. Bot. 47(8): 684-698. Illus. 1960.-An intensive study of the nodal anatomy, petiolar vasculature and open dichotomous venation of the leaf of Kingdonia has revealed a type of foliar vascular system of unusual morphological and phylogenetic interest. The vascular supply at the nodal level consists of 4 collateral traces which diverge from a single gap into the sheathing leaf base. This type of nodal anatomy is perhaps primitive, and comparisons are made with the unilacunar nodes and the 2and 4-parted leaf trace systems characteristic of many angiospermous cotyledons and the foliage leaves of certain woody ranalian genera. The petiole of Kingdonia is vascularized by 2 pairs of bundles which represent the upward continuation of the 4 leaf traces. A transition from an even (4) to an odd (3) number of strands occurs near the point of attachment of the 5, lobed, cuneiform lamina segments to the petiole. Each of the 2 abaxial bundles dichotomizes and the central derivative branches fuse to form a double bundle which enters the base of the median lamina segment. The 2 adaxial petiolar bundles diverge right and left into the bases of the paired lateral segments of the lamina. An analogous type of transition from an even to an odd number of veins occurs in many angiospermous cotyledons which develop a definable mid-vein. But, in Kingdonia, the bundles which enter the bases of the lamina segments give rise to systems of dichotomizing veinlets devoid of "mid-veins." Although the majority of the terminal veinlets enter the marginal teeth of the lamina segments, "blind" endings, unrelated to the dentations, occur in all the leaves studied. Typically, all of the vein endings in a given lobule of a lamina segment are derived from the same dichotomous vein system. However, in some leaves, a veinlet dichotomizes directly below a sinus and the branches diverge into the marginal regions of 2 separate lobules. The phylogenetic significance of the occurrence of open dichotomous venation in such an herbaceous angiosperm as Kingdonia is briefly discussed. From a purely morphological viewpoint, the Kingdonia type of venation invites direct comparison with the venation of Sphenophyllum, certain ferns or Ginkgo rather than with any of the known reticulate venation patterns of modern angiosperms. Although the foliar venation of Kingdonia may represent the result of evolutionary reversion, the very rare anastomoses which occur seem primitive in type rather than "vestiges" of a former system of

A B S T R A C T A single-layered disc of peripheral pronged cells and central prongless cells impart the typical gear shape to colonies of Pediastrum, while the walls of each cell have a characteristic reticulate triangular pattern. The two-layered wall forms in the cells during colony formation following zoospore aggregation and adhesion. The uniformly thin outer layer reflects contours resulting from differential thickening in the reticulate pattern of the inner, thicker, more fibrillar and granular wall layer. The reticulate pattern thus imparted to the outer wall layer persists in empty zoosporangia following the release of zoospores. Columns of electron-dense material extend through the outer wall layer except at the ridges and centers of the reticulum. Following mitosis and cleavage, the resulting zoospores are extruded within a vesicle membrane consisting of the inner wall layer. Separation of this membrane from the parent cell occurs in material of the inner layer adjacent to the outer wall. Vesicles containing swarming zoospores also contain a granular material which appears to become associated with the aggregating and adhering cells of

Fouquieria splendens, ocotillo or candle bush, with its switch-like thorny branches, two to five meters in height, its recurrent short-lived foliage and flaming inflorescence, is a familiar and striking form of the south-western desert areas. From the 1norphological and taxonomic standpoints, it has been fully described (I, 2, 3), and certain aspects of its physiology have also been discussed (4, 5); but with regard to the detailed anatomy of the genus, I have been unable to find anything but a few scattered references (6). In the present paper, the structure of the plant as a whole is briefly reviewed, and certain characteristic features, such as the development of the corky layers and the distribution and structure of the peculiar water storage cells present in the cortex, are described in greater detail. The striking appearance of the stem of the ocotillo (text fig. i) is due to the reticulate pattern formed by the corky layers. The smooth hard spines which originate from the petioles (2, 3) are spirally arranged, and pass

AFTER A fascicled root system had been observed by Bryan (1936) in greenhouse plants of Cycas revoluta Thunb., an anatomical study of the fleshy adventitious roots was undertaken.2 Of principal interest is a centrifugally arranged primary xylem in addition to the regular centripetal metaxylem; this root, therefore, is mesarch. Among the few descriptions of Cycad roots, that of de Bary (1884) is the only one to mention reticulate tracheids, appearing as if secondarily in a centrifugal direction from the primary xylem plate. Material was obtained from plants grown in the botanical greenhouses of the University of Wisconsin. Each root after being washed was cut into segments approximately 1 cm. long; these segments were numbered consecutively, fixed with formol-alcohol-acetic or chrom-acetic solution in separate vials, and imbedded in paraffin. Several cross sections were cut from the proximal end of each segment; longitudinal sections were cut from the remainder of each segment that showed a distinct variance from those segments above or below. Staining was with safranin and fast green. One of the roots (A) most thoroughly studied was 17 cm. long; its segments were numbered A-i, A-2, .etc., progressing consecutively upward to A-19 at the stem end. Series B contained 2 much older and larger roots. Series D contained the 7 continuous segments of a 15 cm. root. Series E and F were from roots 4 and 6 cm. long, respectively. OBSERVATIONS. -The gross feature of the adventitious roots that is most conspicuous is their fleshy character. Root A has a maximum diameter of 15 mm. at a level about 1 cm. from the proximal end. In cross sections of thisroot at any level,, the cortex occupies almost 3/4 the total diameter of the root. Even the growing tip is blunt and massive. All the cortical cells arise in the meristematic region and become differentiated at once into parenchyma. It is chiefly the enlargement of each cell of the cortex to several times its original size that results in the growth in diameter of the whole root. The number of parenchymatous cells, counted in a radial direction, is almost exactly the same in a cross section near the tip and in one near the proximal end. At the proximal end of each root examined there are either 7 or 8 protoxylem points. By a merging from time to time of neighboring points, this number regularly decreases as the root is traced downward. In root A, the decrease segment by segment is as follows: A-19, septarch; A-18, hexarch; A-17, pentarch; A-15, tetrarch; A-13, triarch; A-8, diarch.

Three new species of Protostelium (Order Protostelida of the Mycetozoa) are described: P. irregularis, P. zonatum, and P. pyriformis, all with rather wide distribution and occurring on dead attached plant parts, less often in soil and humus. The latter two species differ from the first and from the other two known species, P. mycophaga Olive and Stoianovitch and P. arachisporum Olive, in the endogenous origin of the stalk. P. mycophaga var. crassipes, with vesicular stalk bases, is the third variety of that species to be described. A key to the genus is included. PROTOSTELIUM Olive and Stoianovitch (1960) is the type genus of the family Protostelidae, order Protostelida of the Mycetozoa (Olive and Stoianovitch, 1966a). Up to the present time it contains only two valid species, P. mycophaga Olive and Stoianovitch (1960) and P. arachisporum Olive (1962). In the present paper three new species and one new variety are described and additional developmental details on members of the genus are presented. Protostelium is characterized chiefly by its predominantly uninucleate, non-flagellate, phagotrophic, amoeboid cells and slender-stalked sporocarps that bear the deciduous spores singly. N\Jembers of the genus occur primarily on dead plant parts (pods, capsules, berries, flowers, etc.) that have remained attached to the plants, though some species are occasionally found in soil, humus, or rotting wood. The trophic stage ingests such foods as yeasts, fungal spores and bacteria. Methods for isolating and culturing protostelids in the laboratory have been described in an earlier publication (Olive, 1967), as have the photographic techniques used in the present research. While uninucleate amoebae and spores are characteristic of the group, plurinucleate cells or even aplanate multinucleate protoplasts appear in cultures of some species. The multinucleate ones, however, do not develop into the extensive reticulate plasmodia that characterize such genera as Schizoplasmodium and Schizoplasmodiopsis (Olive, 1967). Cell division is by binary fission, following which the nucleolar material reappears in each daughter nucleus as several separate bodies that eventually fuse into a single larger nucleolus more or less centrally positioned in the nucleus. It is common to find considerable range in cell and nuclear size within an isolate, which may indicate a range in ploidy, but there are no cytological data to substantiate the latter. I Received for publication 17 March 1969. This research was supported by grants GB-5508, GB7392, and GB-501 from the National Science Foundation. No evidence of sexuality has been observed. Typical sporulating cultures are regenerated from single spores. Plurinucleate protoplasts appear to arise by cell anastomosis as well as by nuclear division without cytokinesis. Such protoplasts may divide into smaller ones by plasmotomy. The amoebae commonly produce filose pseudopodia, especially when placed in water. However, under certain conditions, such as prolonged flattening between a cover slip and a thin agar layer ("sandwich" preparation), they frequently produce lobose pseudopodia instead. Ontogeny of the protostelid sporocarp has been described elsewhere (Olive and Stoianovitch, 1966a) and, with the exception of the two new species with bell-shaped or pyriform spores, is applicable to the genus Protostelium. The stalk-generating part of the culminating protoplast is typically hemispherical, with a narrow extension passing into the upper part of the eveloping stalk, but in the two exceptional species the steliogen is cylindric and the stalk of endogenous origin. The mature sporocarp typically consists of a narrow stalk tube terminated by a single spore, the whole being enveloped by a thin, closely applied sheath. At least one species does not have a demonstrable stalk tube. All species are characterized by having deciduous spores, though in some the spores fall off more readily than in others. The manner in which the spore separates from the stalk tip is not well understood. A key to the species and varieties of the genus is presented below. KEY TO THE GENUS PROTOSTELIUM Spores globose to subglobose Protoplasts orange-pigmented; spores averaging less than 15,g diam; microcysts fairly regular in outline .... . mycophaga Stalk with conspicuous vesicular enlargement at base .... var. crassipes Stalk without conspicuous basal enlargement

PRAY, THOMAS R. (U. South. California, Los Angeles.) Ontogeny of the open dichotomous venation in the pinna of the fern Nephrolepis. Amer. Jour. Bot. 47(5): 319-328. Illus. 1960.-The venation of the pinna of Nephrolepis consists of a midvein and 2 lateral series of dichotomizing veins all of which terminate freely near the margins. The development of the pinna is analyzed with particular attention to the nature of the marginal meristem and the organization of the embryonic pinna as it appears in paradermal section. The arrangement of cells in pinna wings during the period of marginal growth displays a pattern which foreshadows the pattern of the mature venation. In contrast with the development of the leaves of angiosperms, marginal growth continues into a relatively late phase of pinna ontogeny and apparently is active throughout the phases of ontogeny concerned with blocking-out the pathways of procambial differentiation. Thus the pattern of venation appears to be correlated with the manner of activity of the marginal meristem and subsequent orientation of its derivatives. The theoretical aspects of the result of this investigation are discussed in relation to other studies of foliar venation ontogeny. THIS paper is concerned with the ontogenetic processes which result in the open dichotomous venation pattern of the pinnae of Nephrolepis exaltata. The prevalent pattern of dichotomous venation in the leaves of ferns offers an intriguing subject for investigation for several reasons. In its simplest form, a strictly open system of diverging veins, the pattern is in marked contrast to the various types of closed reticula which are nearly universal in angiosperm leaves. Also, within the ferns as a group, dichotomous venation has been considered primitive and various patterns of a reticulate nature as derived (Bower, 1923). It has been suggested recently, however, that fern leaves which display a dichotomous organization are derived from pinnate forms, at least so far as extant species are concerned (Wagner, 1952). In spite of the interest in fern foliar venation, the literature on leaf development in this group of plants is rather limited. The only information on venation ontogeny at the histogenetic level is contained in a sketchy account by Goebel (1922) of pinna development in a species of Adiantum. MATERIALS AND METHODS.-The leaf of Nephrolepis exaltata was selected for investigation because the young pinnae of the developing leaf of this commonly cultivated species are nearly flat during the earlier phases of ontogeny. Thus it was possible to obtain well-oriented paradermal sections of the pinnae which are necessary for a detailed study at the cellular level. Growth of the circinnate leaf tip is prolonged, frequently producing 50 or more lateral pairs of pinnae. Sectioning of a number of such tips at 8 u yielded good examples of the critical phases of pinna ontogeny. Some transverse sections of developing pinnae also were made. 1 Received for publication November 22, 1959. Cleared and stained specimens of mature and developing pinnae were prepared, but the latter proved unsuitable as a source of the precise information desired. The materials were fixed in formalinpropionic acid-ethyl alcohol and dehydrated and embedded by the tertiary butyl alcohol method (Johansen, 1940). The clearing procedures and the staining methods used for both cleared and sectioned materials were the same as those previously employed by the writer (Pray, 1955a). I am indebted to Dr. Sherwin Carlquist for his expert assistance in the preparation of the photographs used in this paper. RESULTS.-Venation pattern. The pinnae of the Nephrolepis form used in this study are 21/2-3 in. long with serrate margins and a basal lobe. The venation of a typical sterile pinnae (fig. 1) consists of a sympodial system with midvein and a series of alternating dichotomous lateral veins. In the distal portion, the laterals fork once; in the median and basal portions, the proximal or both branches dichotomize once more. The 2 basal laterals are somewhat more elaborate, particularly the one in the basal lobe. A secondary sympodial system of this latter is comparable to that of the pinna as a whole. Without exception, all branches of the venation system terminate freely and submarginally. All vein tips are markedly dilated into a hydathodic mass of short, broad tracheary elements. The venation of fertile pinnae is essentially the same with the sori located at the shortened tips of the distal member of each lateral branch. Pinna initiation.-The young leaf tip is occupied by a single apical initial with extended activity. This is true of all members of the Polypodiaceae (sensu lato) thus far investigated (Bower, 1923). The young rachis distal to emergent pinnae (fig. 2) is nearly cylindrical and continues to grow by [The Journal for April (47: 239-318) was issued April 20, 19601 AMERICAN JOURNAL OF BOTANY, Vol. 47, No. 5, May, 1960

INTEREST IN both the structure and functions of phloem has heightened in the past decade, and consequently it may be timely to offer some supplementary information concerning sieve tubes in the Monocotyledoneae. Esau (1939) has recently reviewed the literature on the general subject of the structure of phloem elements in vascular plants and it is unnecessary to duplicate her work in the present paper. Some of the contributions, such as those of Hemenway (1911, 1913), MacDaniels (1918) and Crafts (1939), somewhat resemble the nature of the present work, which is in the form of a survey of the phloem in a large number of representative plants. Other articles, such as those of Schulze (1893), Ross (1883), Lange (1910), Bouvier (1915) and many more, only incidentally mention the structure of the phloem in the Monocotyledoneae. A source of general but brief information on the sieve tubes in this group of vascular plants is Solereder and Meyer (1928, 1929, 1930, 1933). Esau (1939) pointed out the confusion of terminology in the literature concerning phloem and suggested certain changes, especially with regard to sieve cells and sieve tubes. It will be profitable, therefore, to define the terms2 used in this paper before the objectives of the present work are presented. 1. A sieve area is an area in the walls of parenchyma cells, sieve cells, or sieve-tube members in which pores with connecting strands are clustered. In sieve cells and sieve-tube elements, the sieve areas are usually specialized through a more or less pronounced enlargement of connecting strands and a deposition of callose around each strand. 2. A sieve plate is a wall or portion of a wall of a sieve-tube member bearing one or more highly specialized sieve areas. There are two generally distinct types of sieve plates. One is the simple type in which the pores are uniformly distributed (fig. 6), and the other is the compound.type in which the pores are clustered in scalariform, reticulate, or any other complex manner (fig. 4 and 5). 3. A sieve cell is a cell, enucleate at maturity, in which the sieve areas of all walls are of the same degree of specialization. 4. A sieve-tube element or member is a cell, enucleate at maturity, in which certain sieve areas are

EGGERT, DONALD A. (Yale U., New Haven, Conn.) Studies of Paleozoic ferns: Tubicaulis stewartii sp. nov. and evolutionary trends in the genus. Amer. Jour. Bot. 46(8): 594-602. Illus. 1959.-Tubicaulis stewartii, a new species of the order Coenopteridales is described. The specimen was derived from the Upper Pennsylvanian of Berryville, Illinois, and is characterized by having a lacunar middle cortex, a well-developed integumentary system bearing uniseriate hairs, and xylem parenchyma organized into vertically anastomosing strands. In addition, multiseriate (somewhat transitional to reticulate) bordered pitting is present in the petiolar metaxylem elements, while those of the stem stele are multiseriate scalariform. The habit is intermediate between that of a form such as Osmunda and a tree fern, having an upright tapering stem which gives off prominently decurrent petioles in a 2/5 divergence. A reinvestigation of the type specimen of the most closely allied species, T. multiscalariformis, of Upper-Middle Pennsylvanian age, has shown that it has similar features in the cortex, metaxylem, and integumentary layers. Tubicaulis multiscalariformis and T. stewartii form a distinct group in the 6 species now known, whose evolution has most likely involved the retention of a more primitive form of pitting (multiseriate scalariform) with parenchymatization of the xylem. The remaining species of the genus have not developed xylem parenchyma but have developed circular bordered pitting. The relationships of the genus to other genera in the Coenopteridales remain obscure. THE ORDER Coenopteridales is composed of Paleozoic fern-like plants not referable to modern families. These poorly understood plants make up a highly artificial group showing definitely fernlike features, but generally differing from extant ferns in frond symmetry, type of stele, petiole trace symmetry, and in sporangial morphology and position. Supposed relationships in the order are based almost entirely upon anatomical characteristics of stem and petiole parts, for in most instances neither the fructification nor the complete frond is known. Relatively small anatomical differences have been emphasized in the separation of the families and genera. This is in contrast to the classification of living ferns where sporangial morphology is the most important single taxonomic feature. The Coenopteridales possess many recognizably primitive features, such as protostelic vascular systems and sporangia borne on stems, as well as more advanced characters, such as medullated steles and synangiate fructifications. Classification of the group is largely a matter of convenience and expediency, in most instances, although there are some genera, such as Botryopteris, that appear to be roughly equivalent to genera as they are defined in living forms. One of the most recent taxonomic treatments is the admittedly artificial one of Morgan and Delevoryas (1954), in 'Received for publication February 20, 1959. The author wishes to thank Dr. T. Delevoryas, Dept. of Botany, Yale University, for his advice, constructive criticism, and encouragement during the course of the present investigation. His aid in preparing the photographs also is gratefully acknowledged. The species was named in honor of the collector of the type specimen, Dr. W. N. Stewart, Dept. of Botany, University of Illinois, who donated the specimen to our laboratory for study. which 4 families are delimited: the Stauropteridaceae, Botryopteridaceae, Zygopteridaceae, and the Anachoropteridaceae. The present paper deals with the genus Tubicaulis, placed in the family Anachoropteridaceae by the above authors. Discovery of the new species described in this paper has led to a reinvestigation of the type material of the allied species, T. multiscalariformis (Delevoryas and Morgan, 1952). Features have been found in that species which closely ally it to the newly described one, although the 2 are distinct. DESCRIPTION OF THE NEW SPECIMEN.-The specimen is Upper Pennsylvanian in age and was found in a coal ball from Lawrence County, near Berryville, Illinois. It consisted of 2 pieces of stem, each having attached petioles. One piece includes apical portions of the plant while the other includes more basal parts. The 2 pieces in all likelihood are from the same individual. A general scheme of the orientation, and relative size of the 2 specimens is shown in fig. 1. The preservation of the new specimen is the most nearly perfect yet reported for the genus, so that one can determine details of structure seemingly lost in previously described forms. General features.-The species apparently had an upright habit with elongate tapering stems and closely appressed, relatively large petioles. The habit may be considered as intermediate between the condition in Osmunda, in which closely spaced nodes are present, and that in tree ferns, where variously spaced leaves arise from an upright stem. Stem.-Because of the attached petiole bases, the stem has a lobed appearance (fig. 2) and is variable in diameter. Apically, the stem appears somewhat elliptical in outline when viewed in cross section (fig. 5), but sections made at lower levels are more

This paper includes a survey of the nature of the primary vascular system in a large number of extinct gymnosperns and progymnosperms. The vascular system of a majority of these plants resembles closely that of living conifers, being characterized, except in the most primitive forms which are protostelic, by a eustele consisting of axial sympodial bundles from which leaf traces diverge. The vascular supply to a leaf originates as a single trace with very few exceptions. It is proposed that the eustele in the gyn nosperms has evolved directly from the protostele by gradual medullation and concurrent separation of the peripheral conducting tissue into longitudinal sympodial bundles from which traces diverge radially. A subsequent modification results in divergence of traces in a tangential plane. The closed vascular system of conifers with opposite and whorled phyllotaxis, in which the vascular supply to a leaf originates as two traces which subsequently fuse, is considered to be derived from the open sympodial system characteristic of most gymnosperms. This hypothesis of stelar evolution is at variance with that of Jeffreywhich suggests that the eustele of seed plants is derived by the lengthening and overlapping of leaf gaps in a siphonostele followed by further reduction in the resultant vascular bundles. This study suggests strongly that the "leaf gap" of conifers and other extant gymnosperms is not homologous with that of siphonostelic ferns and strengthens the validity of the view that Pteropsida is an unnatural group. It supports the position that gymnosperms have evolved from progymnosperms rather than from ferns, and the inclusion of ferns and gymnosperms in separate subdivisioslS or divisions. IT HAS BEEN clearly established that the primary vascular system of a large number of living conifers is composed essentially of longitudinally oriented sympodia, axial bundles from which leaf traces diverge (Namboodiri and Beck, 1968a,b). While these features are found in all the specimens studied, variations are observed in the number and mode of formation of traces supplying a leaf. Thus in the forms with helical phyllotaxis the trace supplying a leaf is single at the point of origin (Namboodiri and Beck, 1968a), whereas in species with opposite leaves two traces originating from two different bundles are found to fuse to form the single strand supplying a leaf, thus giving rise to a reticulate vascular architecture. Some arguments have been presented in part II of the study (Namboodiri and Beck, 1968b) suggesting that the open system of the genera with helical leaf arrangement is primitive and the closed, reticulate system of conifers with opposite phyllotaxis is derived. This implies that the single-trace condition characteristic of the conifers with helical phylloI Received for publication 17 July 1967. This paper is based on part of a Ph.D. thesis by the first author. This study has been supported in part by National Science Foundation grant GB-3038 to C. B. Beck. 2 Formerly known as G. Kadambari Kumari. Present address: Genetics Program, University of North Carolina, Chapel Hill, North Carolina 27514. taxis is the primitive condition, a view that supports the observations of Pant and 'MIehra (1964). It also suggests the possibility that the leaf gap of conifers is not homologous with that of the siphonostelic ferns. These conclusions would be of greater phvlogenetic significance if they could be applied to the gymnosperms as a whole, especially the extinct forms. A knowledge of the primrary vasculature of fossil gymnosperms is necessary in order to draw any valid conclusions about long-range stelar evolution and homology of structures. It is with these goals in mind that the survey reported herein has been made. Such a survey presents several problems. Shoot apices of woody plants, in which the primary vasculature is clearly discernible, are seldom fossilized. Therefore, the main source of information about the primary vascular system is the stem with secondary growth. But when the maturation of the xylem is endarch it is difficult to discern accurately the details of the primary vascular system. Thus genera characterized by mesarch or exarch development of primary xylem provide most of the data for this study. Among living conifers the number of traces that supply a single leaf may not be the same at nodal levels as at subnodal levels. In the Abietineae, for example, the trace is single at the subnodal levels, but double at the nodal level

A B S T R A C T A new genus from a Middle Devonian locality near Cairo, N. Y., is described. Actinoxylon gen. nov. is based upon pyritic petrifactions. Three orders of branching are present: penultimate branch, ultimate branch, and leaf. The penultimate branch bears spirally arranged ultimate branches and leaves, the leaves apparently replacing the branches in the spiral. The ultimate branches bear opposite to subopposite and decussate leaves. The leaves are non-planated, unwebbed structures which show at least three dichotomies. Each segment of the lefaf is terete as are all other axes. Internally the penultimate branch has a six-lobed actinostele with mesarch protoxylem areas, one or two per lobe. Secondary xylem is visible in the oldest parts of several specimens. The xylem has helical-reticulate, reticulate, scalariform and circular-pitted elements. The presumptive areas of phloem are occupied by cells with dark contents. The cortex is composed of a parenchymatous inner region and a selerenchymatous outer region. The ultimate branch traces are at first three-lobed protosteles, later becoming four-lobed. Several ultimate branch traces also possess secondary xylem while within the cortex of the penultimate branch. The leaf traces are terete strands. Below each forking of a leaf segment there is a corresponding forking of the vascular strand. Actinoxylon is compared with the progymnosperms Actinopodium, Svalbardia, Archaeopteris, Siderella, and Tetraxylopteris. The anatomy of the penultimate branch of Actinoxylon is similar to that of Actinopodium, Archaeopteris macilenta, and Siderella. The ultimate branch traces of Archaeopteris and Actinoxy7on are similar. The ultimate branch stele and pattern of trace formation in Actinoxylon is similar to the stelar configuration and trace formation in the r + 2 axes of Tetraxylopteris schmidtii. The uinwebbed leaves are similar to those of Archaeopteris fissilis, Svalbardia, and the terminal units of the Aneurophytales.

IN PREVIOUS papers on the vessels of the Monocotyledoneae, the present author discussed the occurrence of vessels (1942), their origin and certain general trends of specialization (1943a), and their specialization in the late metaxylem of various organs (1943b). The primary object of the present report is to determine the course of specialization of vessels throughout the metaxylem within each organ of the plant. Insofar as possible, the protoxylem will receive similar consideration. As a secondary object, specialization of the early metaxylem from organ to organ will be compared with similar changes in the late metaxylem from organ to organ. The previous paper (1943b) which concerned such changes considered only the larger vessels in the late metaxylem. This secondary object should indicate whether or not the rate of specialization throughout the metaxylem has been the same in all organs. As pointed out in the previous papers of this series, there are few references in the literature which are pertinent in this discussion of vessels in the Monocotyledoneae. MATERIALS AND METHODS.-Materials from all available organs of 306 species in 204 genera among 34 families, as defined by Hutchinson (1934), were used in this study.2 The number of genera and species in each family is about the same as given in the latest paper in this series (1943b), and merely the families are presented here to indicate the scope of the investigation: Agavaceae, Alismataceae, Alstroenmeriaceae, Amaryllidaceae, Araceae, Bromeliaceae, Butomaceae, Cannaceae, Commelinaceae, Cyperaceae, Dioscoreaceae, Eriocaulaceae, Gramineae, Haemodoraceae, Hypoxidaceae, Iridaceae, Juncaceae, Liliaceae, Marantaceae, Musaceae, Orchidaceae, Palmae, Pandanaceae, Pontederiaceae, Potamogetonaceae, Ruscaceae, Scheuchzeriaceae, Smilacaceae, Sparganiaceae, Strelitziaceae, Trilliaceae, Typhaceae, Xyridaceae, Zingiberaceae. It should be pointed out that the number of genera and species in the highly specialized grasses is relatively high. Consequently, this family will be treated both separately and together with the remainder of the families. Using methods previously described (Cheadle, 1942), all materials were macerated and most of them sectioned as well. The macerations were often of greater use than the sections, although the latter were valuable in confirming positional relationships of the various elements of the meta1 Received for publication October 7, 1943. With the support of a grant from the American Philosophical Society, and two grants from the American Academy of Arts and Sciences. 2The author is grateful to all those who have had a share in the accumulation of data for this paper. He also expresses appreciation to the Biological Laboratories of Harvard University for the use of photomicrographic equipment. xylem. Incomplete macerations were also of similar value. OBSERVATIONS AND DISCUSSION.-Because only primary xylem is treated in this paper, it will be necessary to use both the terms protoxylem and metaxylem. These terms, as pointed out by Esau (1943), have been variously defined, most recently in a fashion that stresses features revealed in on5ogenetic development. In a report which is phylogenetic in nature, it is perhaps necessary merely to make clear what structures are meant by the terms which are being used in presenting the data and conclusions. It is unnecessary in this particular paper, for example, to report what exact type of element first appears ontogenetically in each organ. There seems to be no doubt that when a wide variety of conducting elements occurs in the primary xylem, those elements having annular wall thickenings mature first, those having spiral thickenings mature next, and those having typical bordered pits mature last. However, because of various intergradations there is no absolutely clear boundary between any two of the above elements. Furthermore, the conducting elements in the primary xylem of some fleshy representatives are not often easily identifiable from the standpoint of the presence or absence of pitting, especially in the shoot system. With regard to papers of this particular nature, it is perhaps fortunate that few of these fleshy plants have vessels. Only a small number of such plants are, therefore, involved in the present discussion. The use of the term protoxylem will be such as to involve those elements with either annular or spiral thickenings or both. All other elements will be considered as metaxylem. The earliest metaxylem will then normally consist of the intermediate forms such as those with reticulate sculpturing on the side walls. In some cases, the late metaxylem may consist of these so-called intermediate forms. When the actual meanings of the terms late and early metaxylem are given later, however, it will be seen that these unusual cases will not confuse the issues in this paper. The terms vessel, vessel member, scalariform and porous (simple) perforation plates, and tracheids will be used as defined by the Committee on Nomenclature of the International Association of Wood Anatomists (1933). The more convenient term vessel will be used instead of vessel member wherever possible and where its use will not be confusing. Little difficulty was encountered in choosing a method for arbitrarily dividing the metaxylem of most roots for the purposes of comparison. This was especially true in such large roots as those of the palms (fig. 1) and aroids, where the xylem strands extend for a considerable radial length. Certain of the smaller roots (fig. 3) were more troublesome in this regard, because of the smaller number of elements

A B S T R A C T The structure and ontogeny of the calyx and corolla of Downingia bacigalupii Weiler (Campanulaceae; Lobelioideae) were investigated for the purpose of comparing perianth development with previous observations on the floral bract, as well as elucidating the mechanism of development of the zygomorphic, sympetalous corolla. Sepals are uni-traced with a palmate, reticulate venation. They have basal and apical hydathodes, as well as storage tracheids. Sepals show a reduction in size, venation and hydathode number when compared to the bract. The pentamerous, zygomorphic corolla is bilabiate, consisting of a three-lobed adaxial lip and a two-lobed abaxial lip connected by a short tubular region. The constituent petal lobes are also uni-traced and have a reticulate venation, resembling that of the sepal and bract, but lack storage tracheids and hydathodes. Sepals arise in an adaxial to abaxial succession and are initiated in the outer corpus layer of the floral apex. Expansion of the floral apex follows and is accompanied by the establishment of a second tunica layer. Sepals undergo apical, marginal, and intercalary growth accompanied by acropetal differentiation of procambium. The petals arise simultaneously and are initiated in the second tuniica layer and the outer corpus cells. After initiation, the petals exhibit a period of apical and marginal growth followed by intercalary growth. Apical growth in petals is less protracted than in sepals, but plate meristem activity is more extensive. The free petal lobes become temporarily fused by an interlocking of marginal epidermal layers, but they separate at anthesis. Zonal growth beneath the originally free lobes forms the tube and lip regions of the syinpetalous corolla. Zygomorphy is evident from the time of initiation of petals and is accentuated by later differential growth. Comparative observations of corolla ontogeny in autogamous species of Downingia indicate that the reduced corollas in these taxa are derived by a simple process of neoteny.

A B S T R A C T A systematic study of leaf morphology and anatomy of native, aquatic Polygonum species is presented, treating 60 populations propagated in a common greenhouse. On the basis of this study it is now possible to distinguish American Polygonum species by leaf characters alone. Taxa in section Persicaria are contrasted and also compared with representatives of sections Bistorta and Echinocaulon. A complete description, utilizing 40 leaf characters, is given for each taxon. Complex glands, designated valvate chambers, are shown to be a consistent feature of the P. punctatum species-complex, whereas epidermal storage-idioblasts are present only in certain populations of this group. All other species lack valvate chambers but possess idioblasts, which may reach a frequiency of over 400/mm2 in P. hirsutum. The plate-like multicellular gland of P. opelousanum is found to occur sporadically in P. hydropiperoides, thereby losing its specific diagnostic value. Illucidation of the characteristic structure of this gland, however, lends evidence to discourage the theory that it is an abortive structure produced by a hybrid between glandular and non-glandular species. Reticulate venation and unicellular trichomes separate members of section Bistorta from the smartweeds, but P. meisnerianum (section Echinocauilon) is shown to simulate leaf struicture of the persicarias both internally and externally.

Mature thalli of Boldia erythrosiphon gen. et sp. nov. from Big Walker Creek, Giles Co., Va., are hollow, simple, or, rarely, branched tubular cylinders. The plants generally are from 1 to 20 cm in length and 0.1-1.5 cm in diameter, brownish-red or olivaceous in color, and grow singly or in clusters from a multicellular holdfast. The vegetative portion of the thallus is essentially monostromatic, comprised of isodiametric, or slightly elongate, uninucleate cells with numerous parietal, simple or lobed, ribbon-like chromatophore segments. Monospores are produced from narrow, branching filaments (derived from vegetative cells) which form a reticulate pattern in the intercellular matrix external to and at the juncture between groups of (usually 2, 4 or 8) vegetative cells. The spores are liberated singly and germinate into short, branched filaments. New plants arise from small, monostromatic discs of branching filaments united in a common matrix which become cushion-like and ultimately produce saccate thalli. While Boldia shows affinities with certain members of the Bangiales, it cannot be assigned to any of the presently recognized families. Accordingly, it is placed in a new and distinct family, Boldiaceae, distinguished by the unique manner in which asexual spores are produced from filaments derived from vegetative cells, the filaments developing within the intercellular matrix of the thallus, as well as by the parietal chromatophores of the vegetative cells. A MOST unusual freshwater red alga has been found growing in tributaries of the New River in Virginia. The first specimen of this plant was discovered by members of the animal ecology class of the Mountain Lake Biological Station of the University of Virginia, taught by Dr. William S. Woolcott in the summer of 1958. This class, engaged in an icthyological study of local streams, was asked to collect Lemanea for the plant morphology course and, in response, Dr. Woolcott brought to the Laboratory a snail with an attached saccate thallus approximately 3.5 cm in length. While the first inclination was to dismiss it, since its body form was so different from that of any freshwater alga known, after microscopic examination it became clear that it was a red alga with structure so unique that it could not be accommodated in any of the presently recognized familial taxa of this group. Return to the site of the original collection revealed the plant in abundance. The following account of its general morphology is based on the material collected then and in subsequent summers. DIAGNosIs-Boldiaceae famn. nov.-Thallus adultus vegetativus parenchymaticus, cellulis chromatophora parietalia exhibentibus. Monosporae ex filamentis formatae inter cellular vegetativas. Plantae praeterea notas Rhodo1 Received for publication May 25, 1964. The writer is indebted to Mr. Tyge Christensen, Copenhagen University for preparation of the Latin diagnosis and to Professors W. R. Taylor, R. C. Starr and G. F. Papenfuss for their helpful comments. Contribution from the Botanical Laboratory, The University of Tennessee, n. ser 238. phycearum Bangialium praebentes. Genus typificum Boldia infra descriptum. Boldia erythrosiphon gen. et sp. nov.-Thallus saccato-tubularis, simplex vel raro ramificatus, fusco-carneus vel olivaceus, 1-20 cm altus, ad 1.5 cm diam., unicus disco basali affixus vel complures e pulvinari communi hemisphaerico multicellulari emissi, in statu vegetativo vix ulla exceptione monostromaticus. Cellulae isodiametricae vel subelongatae, plerumque 10-20,u diam., quaeque unico nucleo instructa, chromatophoro interdum perjuvenis vel propagationis causa formata unico centrali, praeterea compluribus parietalibus, taeniiformibus, simplicibus vel irregulariter lobatis et ramificatis. Propagatio non-sexualis per sporas peracta e filamentis angustis copiose provenientes sub superficie exteriore thalli penetratibus, parietes cellulas vegetativas contiguas discernentes reticulatim secutis. Sporae in filamentorum cellulis singulae, per poros parietales liberatae, in filamenta brevia ramificata germinantes. Thalli juvenes initio disciformes, monostromatici, e filamentis radiantibus contiguis formati, postea pulvinati, demum in partes saccatas elati. Propagatio sexualis ignota. Typus die 5, Jul. anni 1958 in amne Big Walker Creek regionis Virginianae Giles County (folii quartae Pearisburg) Americae Borealis ab auctore lectus, in honorem Harold C. Bold ob phycologiam ut plantarum morphologiam ingenio excitante communicatam nominatus, in herbario cul nomen U. S. National Herbarium, urbis Washington depositus. [The JOURNAL for May-June (51: 465-573) was issued June 11, 19641 AMERICAN JOURNAL OF BOTANY, Vol. 51: No. 6, 1964

Euadiantum; species pulchra rachibus paleis subulato-capillaceis divaricatis persistentibus numerosissimis ubique praeditis insignis, habitu A. pulverulento non dissimilis sed ut videtur A. glaucescenti affinis, quod differt: rachibus gracillimis lucidissimis paleis carentibus, segmentis tenuiter membranaceis mediocriter lobatis, soris vero lunatis valide discretis partem mediam late retusam loborum insidentibus. Rhizome short-creeping, 2-5 cm. long, 4-7 mm. thick, coarsely and unevenly nodose from the stout pseudopodia of old fronds, closely paleaceous, especially toward apex; scales tufted, light to dark brown, 4-5 mm. long, reticulate, lance-subulate, deciduously ciliate toward the base, remotely denticulate toward the long-attenuate tip. Fronds 2 or 3, approximate, laxly ascending, 40-90 cm. long; stipes somewhat longer than the blades, stout, atropurpureous, obtusely trigonous, deeply sulcate ventrally, lustrous, conspicuously shaggy with very numerous, wide-spreading, subulatefiliform scales (4-5 mm. long, mostly with appressed fimbriate-stellate base), these extending freely to the rachises throughout; blades impari-bipinnate (rarely 3-pinnate at base), 25-50 cm. long, 20-40 cm. broad; pinnae 7-i6 (22), 2-4 cm. broad, linear-oblong to elliptic-oblanceolate, alternate, spreading, simply pinnate or (exceptionally) the 2 or 3 lowermost pairs fully bipinnate at base; segments I5-30 pairs, distinctly petiolulate, non-articulate, oblique, the median ones rhombic-oblong from a subrectangular or broadly cuneate inequilateral base, straight or lightly falcate, rounded-obtuse or (if fertile) often muticous at apex, I.5-2.5 cm. long, 5-8 mm. broad, the basal ones reduced, short-rhombic to orbicular-flabelliform, the terminal ones lanceolate to narrowly and unequally lobate-hastate; sterile pinnae entire at lower margin, unequally serrulate at upper margin and apex; veins prominulous, rather close, mostly 2 or 3 times forked; leaf tissue membrano-herbaceous, dark green, lustrous, and striolate above, beneath glaucous, often conspicuously so; sori borne mostly on segments of the lower half of the pinnae, oblong or sometimes sublunate, 2-3 mm. long, occupying the upper margin and apex, adjacent, the reflexed marginal lobe very narrow, closely revolute, glabrous, the true indusium ample (o.s-o.8 mm. broad), brownish, membranous, glabrous, erose to coarsely dentate-fimbriate; spores triplanate, obtusely 3-lobed, about 35 ,u in diameter, yellowish, granular. 17 I65

ILLUSTRATIONS OF many species of Spirogyra not previously illustrated are in preparation by the author. In the collections made toward the accomplishment of this larger work four new species have been found. These new species are described and illustrated in this present paper. SPIROGYRA echinospora sp. nov.-Vegetative cells 66-72 ,u X 50-125 ,u, with plane end walls, 6-7 (-8) chromatophores making 1/2-11/2 turns in the cell; conjugation scalariform, tube wide, formed by both cells, narrowest at the point of fusion and not sharply set off from either cell; fertile cells inflated on the conjugating side only; zygospores ovoidellipsoid, 67-82 u X 98-150 /; exospore very thin, smooth; mesospore thick, brown-black, echinate (fig. 1, 2). Cellulis vegetativis 66-72 I X 50-125 ,u, dissepimentis planis, chromatophoris 6-7 (-8), anfractibus /2-11/2 conjugatione scalariformi, tubo lato, ex utraque cellula conjugante emisso, maxime angusto in loco conjunctionis neque ab alterutra cellula discriminato; cellula fructifera solum in latere conjugationis inflata; zygosporis ovato-ellipsoideis, 67-82 ,u X 98-150 /i; exosporio tenuissimo, levi; mesosporio crasso, atrobrunneo, echinato. New Hampshire. Collected by Dr. Hannah Croasdale in Frog Pond, Hanover, September 8, 1942. This species is much smaller than S. echinata Tiffany which it resembles in many respects. SPIROGYRA sulcata sp. nov.-Vegetative cells 37-46 ,u X 50-160 t,u with plane end walls, 1 (-2) chromatophores which have a constant median furrow; chromatophores making 2-5 turns; conjugation scalariform, tube formed by both cells, and set off sharply from each cell; fertile cells inflated slightly on the outer side; zygospores ovoid, 43-46 Iu X 52-62 I,u exospore smooth, mesospore brown, thicker, reticulate (fig. 9-11). Cellulis vegetativis 37-46 I X 50-160 I, dissepimentis planis, cliromatophoris 1 (-2), sulco medio constante praeditis, anfractibus 2-5; conjugatione scalariformi, tubo ex utraque cellula conjugante emisso, ab utraque cellula abrupte discriminato; cellula fructifera parte in exteriore paululum inflata; zygosporis ovatis, 43-46 ,u X 52-62 ,u; exosporio levi; mesosporio brunneo, crassiore, resticulato. Wisconsin. Collected by the writer in springs, University of Wisconsin Arboretum, Madison, August 15, 1942. SPIROGYRA torta sp. nov.-Vegetative cells 4956 I, X 270-600 I,u with plane end walls, 3-5 chromatophores making 1-4 turns in the cell; conjugation scalariform, tube formed mostly by the male cell, widest at the point of fusion, mostly twisted and distorted during the maturation of the zygospore; tube sharply set off from the male cell, not from the female cell; fertile cell not inflated; zygospores