Monograph of Diplachne (Poaceae, Chloridoideae, Cynodonteae)

Abstract Diplachne P. Beauv. comprises two species with C4 (NAD-ME) photosynthesis. Diplachne fusca has a nearly pantropical-pantemperate distribution with four subspecies: D. fusca subsp. fusca is Paleotropical with native distributions in Africa, southern Asia and Australia; the widespread Australian endemic D. f. subsp. muelleri; and D. f. subsp. fascicularis and D. f. subsp. uninervia occurring in the New World. Diplachne gigantea is known from a few widely scattered, older collections in east-central and southern Africa, and although Data Deficient clearly is of conservation concern. A discussion of previous taxonomic treatments is provided, including molecular data supporting Diplachne in its newer, restricted sense. Many populations of Diplachne fusca are highly tolerant of saline substrates and most prefer seasonally moist to saturated soils, often in disturbed areas. Some populations of Diplachne fusca in southern Asia combine nitrogen-fixation, high salinity tolerance and palatibilty to livestock, which should be pursued with further research for purposes of soil reclamation. Diplachne fusca subsp. uninervia is the most invasive of the subspecies and is becoming weedy in some non-native areas, including in the Old World. This monograph provides detailed descriptions of all taxa, a key to the species and subspecies, geographic distributions and information on the anatomy of leaves, stems, lemmatal micromorphology and discussions of the chromosome numbers. Lectotypes are designated for: Atropis carinata Grisb.; Diplachne acuminata Nash; Diplachne capensis (Nees) Nees var. concinna Nees; Diplachne capensis (Nees) Nees var. obscura Nees, Diplachne capensis (Nees) Nees var. prolifera subvar. minor Nees, Diplachne halei Nash, Diplachne maritima E.P. Bicknel, Diplachne muelleri Benth., Diplachne reverchonii Vasey, Diplachne tectoneticola Backer, Leptochloa imbricata Thurb., Leptochloa neuroglossa Peter, Leptochloa uninervia var. typica fo. abbreviata Parodi, Triodia ambigua R. Br. and Triodia parviflora R. Br.

located in the field during this project (see comments under that species). Leaves were sectioned by a rotary microtome at Rancho Santa Ana Botanic Garden in the laboratory of Dr. J. Travis Columbus or by hand by the first author at the Missouri Botanical Garden. Data from lemmatal micromorphology (Snow 1996) and caryopsis morphology (Snow 1998b) are included from those publications.
Descriptive terminology for leaf anatomy largely follows Ellis (1976) and includes information from Snow (1997a). This includes keel to refer to the proliferation of parenchyma (abaxially, but usually more prominent adaxially) that surrounds the centralmost (median) vascular bundle, which typically extends laterally to at least the first vascular bundle on either side of the median bundle. Lacunae in the keel refer to areas  Figure 1. Maximum likelihood tree inferred from nuclear analysis of combined plastid (rpL32-trnL, ndhA intron, rps16 intron, and rps16-trnK) and ITS sequences. Numbers above branches are bootstrap values; numbers below branches are posterior probabilities from Bayesian analysis; color indicates native distribution (see legend) of Diplachne species. Scale bar: substitutions per site. Table 1.
Taxon voucher (collector, number, and where the specimen is housed), country of origin, and GenBank accession for DNA sequences of rps16-trnK, rps16 intron, rpl32-trnL, ndhA intron, and ITS regions (bold indicates new accession); a dash (-) indicates missing data; an asterisk (*) indicates sequences not generated in our lab.  The location of some accessions (Peterson 20786, Columbus 3111) of the amphitropical  D. fusca subsp. uninervia near the base of the fusca clade suggests the possibility of a South American origin of this subspecies, but that interpretation needs further testing and does not explain the presence of one accession from Estado de México, México (Peterson 21305) in a large polytomy elsewhere. Of all subspecies, D. f. subsp. uninervia has the greatest weedy tendencies and is now distributed across many non-native areas (e.g. Snow and Simon 1999;Pérez et al. 2010). The significant separation of the Mexican accession from the South American accessions also suggests that subsp. uninervia may not have had a singular origin.
The three accessions of the exclusively Australian D. f. subsp. muelleri likewise are not monophyletic, given that one (Badman 1282 from South Australia) forms a clade containing two accessions of D. f. subsp. fusca from Korea and another clade includes an accession (Jensen 2527) with two accessions of subsp. fusca (Snow 7249;Walsh 6558). The Korean sequences were obtained from GenBank, but vouchers have not been seen.
The molecular data may accurately reflect the true evolutionary relationships of the specimens sampled and thus support the lack of monophyly among the subspecies. Given the widespread geographical distribution of D. fusca and of subsp. fusca in particular, the lack of monophyly would not be surprising. If accurately reflecting the actual history, then the molecular data suggest that the morphological characters used to separate the subspecies may be inadequate. The frequent inability of morphological characters, by themselves, to accurately recover genera within Poaceae, is well known (Kellogg 2015). In contrast, if the subspecies are monophyletic as recognised herein (largely) by morphological data, then any of several factors may have contributed to the incongruity of geographical region of origin and clade placement among subspecies of D. fusca, including: multiple mutational hits among base pairs, lineage sorting, hybridisation and long-distance dispersal. Among long-distance dispersal events, some of these might include human-mediated dispersal events within the last few millennia. In the authors' view, a more in-depth sampling of populations is necessary to test further hypotheses of monophyly among the subspecies. Another possibility, of course, is that the taxonomy at the subspecific level needs adjusting, but that would require extensive re-analysis of the thousands of herbarium specimens that underlie the morphologically-based taxa. A high sampling priority in the authors' view is to obtain DNA samples of D. fusca subsp. fusca from South America, to see where they fit into the phylogenetic framework.
Elsewhere in Figure 1, two accessions of Dinebra divaricatissima formed a clade that rendered D. decipiens paraphyletic. If further sampling upholds this arrangement, then D. divaricatissima might be better treated as a subspecies of D. decipiens. An earlier placement  of Dinebra divaricatissima within Diplachne was based on a misidentified individual of Diplachne fusca.
Morphology. The elongate, broadly to narrowly acute membranous ligule of Diplachne and numerous, dorsally rounded or flattened florets (before seed development), most consistently diagnose the genus from species formerly placed in Leptochloa sensu lato (Snow 1997a. As reported previously , banding patterns from chloroplast DNA restriction site analysis of populations of D. fusca subspp. fascicularis and uninervia were virtually identical, but consistently different from other North American members of Leptochloa s.l. (Snow unpublished, 1991). More recently, Peterson et al. (2014) demonstrated that DNA barcoding generally can separate members of Leptochloa s.l. with relatively high levels of confidence.
Geographically localised morphological variation among populations of Diplachne fusca is widespread, which has led to numerous heterotypic synonyms. After observing this variation globally, particularly with regards to taxa here included in Diplachne fusca subsp. fusca, Snow (1997a) concluded that only four infraspecific taxa could be consistently diagnosed morphologically following a phylogenetic species concept or (later) a general lineage concept using a diagnosability criterion (Snow 1997b;. Although some authors might choose to re-elevate the subspecies of D. fusca to the specific level, it should be stressed that the nominative taxon, which is the most widespread and exhibits the highest levels of morphological variation, intergrades into each of the other three to some degree with virtually all morphological characters. Whether these taxa hybridise has not been tested. A collection from Orange County, California (Beetle 13066 [OKLA]) may be a hybid between L. f. subspp. fascicularis and uninervia. Diplachne gigantea is a rare, evidently emergent aquatic species endemic to a few narrow regions of southern Africa.
Leaf anatomy. Detailed anatomical descriptions follow each taxon. In general, leaves show C 4 (NAD-ME) anatomy; primary vascular bundles project considerably more than secondary bundles and a prominent parenchymatous keel is present in the midnerve region, which typically develops a single (often prominent) lacuna. Sutton (1973a) commented briefly on leaf anatomy of D. fusca subsp. fasicularis (as Leptochloa fascicularis [sic]) but not in a comparative context with other taxa. Earlier workers also made limited observations of leaf anatomy. Hattersley and Watson (1976) used the presence of an "intervening cell" to predict a XyMS+ anatomy for D. fusca (as Diplachne parviflora (R. Br.) Benth.). Clifford and Watson (1977) illustrated the transverse section of Diplachne parviflora. Ellis (1977) reported NAD anatomy in Diplachne fusca. However, Valls (1978) carried out the most extenstive investigation of species formerly placed in Leptochloa s.l., which now inludes that genus, Disakisperma Steud., Dinebra, Trigonochloa P.M. Peterson & N. Snow and Diplachne (Peterson et al. 2012;2014. Watson and Dallwitz (1992: 324) briefly mentioned leaf anatomy in Diplachne. Valls (1978) studied cross sections of all of subspecies of D. fusca (under various names). Based on the prominent adaxial projection of the primary vascular bundles of D. fusca, Valls (1978) reported them as being "nodular" in transverse section, meaning that the primary (and to some degree secondary) vascular bundles project from the adjacent tissues. He also noted the broad and prominent parenchyma midnerve region of the adaxial surface among subspecies of D. fusca (Valls 1978), which is simply referred to as the keel. Most specimens of D. fusca have prominent keels on the abaxial surface, but these become increasingly narrow distally.
Whereas Snow (1997a) did not examine cellular detail of leaf surfaces, Valls (1978) observed saddle-shaped silica bodies for Diplachne fusca subspp. fascicularis, fusca and uninervia. All three species of Disakisperma (Snow and Peterson 2013) also had saddleshaped silica bodies on the laminar surface (Valls 1978), as did Leptochloa malayana (C.E. Hubb.) Jansen ex Veldkamp and L. longa Griseb., species not included in molecular studies ) and which presently are incertae sedis at the generic level.
Stem anatomy. Comparative studies of stem anatomy in Poaceae have been limited (but see de Wet 1960;Ebinger and Carlen 1975;Sánchez 1979Sánchez , 1981aSánchez et al. 1989;Clark and Fisher 1987). The internodes of D. fusca are hollow, whereas stellate aerenchyma comprises the pith of D. gigantea. Stellate aerenchyma is known from some grasses (Pohl and Lersten 1975) and is not uncommon among plants that remain submerged to somewheat emergent in water, which appears to be the preferred, if not exclusive, habitat of D. gigantea.
Lemmatal micromophology. All examined specimens of Diplachne share a combination of micromorphological characters including the presence of long and short cells, silica cells, cork cells and bicellular microhairs (Snow 1996). All members of Diplachne fusca have macrohairs with obtuse or rounded apices, but these were absent or scant for D. gigantea and, if present, then restricted to the base of the lemma (Snow 1996(Snow , 1997a. Diplachne fusca and D. gigantea have papillate short cells (Snow 1996).
Reproductive biology. Hackel (1906) believed that some specimens of Diplachne fusca subsp. fascicularis approach a cleistogamous condition, which he defined strictly to be plants that showed intrastigmatic anthers crushed at the top of a mature fruit. In contrast, for D. fusca subsp. uninervia Valls (1978: 96) observed flowering in the first two panicles in a greenhouse-grown specimen (Valls 3514 [TAES]) that produced numerous seeds. Anthesis was not observed despite a daily check, suggesting complete cleistogamy during this interval. However, the next two inflorescences to emerge were observed to undergo anthesis (by virtue of exserted stamens) and production of seed, suggesting the possibility of seasonable shifts in chasmogamy and cleistogamy. Valls (1978) also reported observing crushed anthers in D. fusca subsp. muelleri, which also suggests the possibility of cleistogamy. In contrast, he reported no evidence of cleistogamy for D. fusca subsp. fusca. Given that anthers tend to be significantly shorter in the predominantly Neotropical subspecies of D. fusca (fascicularis and uninervia) compared to the native Paleotropical subspecies (fusca and muelleri), Valls (1978) suggested that predominately chasmogamous and predominantly cleistogamous lineages appear to co-exist in D. fusca, which he used also as evidence that the three subspecies he studied (all excluding D. f. subsp. fusca) could be treated conspecifically, as in the present paper. Further, he posited that cleistogamous lines were developed from polymorphic chasmogamous stock in D. fusca subsp. fusca (Valls 1978: 97).
Caryopsis morphology. The caryopsis of Diplachne is flattened dorsally, with the pericarps of D. fusca and D. gigantea easily separable when soaked in water at room temperature (Snow 1998b). Dorsal flattening and readily separable pericarps also occur in species now treated in Disakisperma and among other genera of Chloridoideae (Snow and Peterson 2013), but most caryopses of Disakisperma also are broadly concave at maturity. A slightly concave surface has been noted previously (Valls 1978) for some specimens of D. fusca, although the caryopses of Diplachne are generally elliptic in transverse section and, if concave, then notably less so than the normal condition in Disakisperma (Snow and Peterson 2013).
Embryo formula. Reeder (1957Reeder ( , 1961 took the results of previous studies of grass embryos and those of his own to create an "embryo formula", which typified variation in a few observable characters of grass embryos. Reeder used the letters "F" and "P" to represent the condition typically found in subfamilies Festucoideae and Panicoideae in their former, much broader circumscriptions (e.g. Hitchock 1951). In addition, Reeder (op. cit.) indicated presence (+) or absence (-) of an epiblast. The typical formula for genera in subfamily Chloridoideae was P+PF, indicating (in order): an elongation of the vascular system between the point of divergence of the scutellum and coleoptile (P); epiblast present (+); presence of a cleft between the lower part of the scutellum and the coleorhiza (second P); and embryonic leaf in cross section with non-overlapping margins and relatively few vascular bundles (F). According to Valls (1978: 109), Reeder based his summary of Leptochloa (i.e. including the species treated here as Diplachne) on four species, but did not indicate specifically which species were examined. Valls (1978: 110 [voucher: Valls 3468 {TAES}) confirmed the formula for D. fusca subsp. fusca as P+PF.
Pollen. The pollen of Diplachne has been studied minimally, but Liu et al. (2004) report "type 4" pollen (for Chloridoid grasses) for D. fusca, which has an annulate aperture, an insular exine and a "sculptural density" of 0.5-2 µm 2 . These authors also showed that Dinebra (=Leptochloa) panicea (Retz.) P.M. Peterson & N. Snow has two aperatures, in contrast with the single aperture of D. fusca, the latter of which has also a minute annulus surrounding the aperture. Further study may reveal characters of pollen that reliably diagnose Diplachne from related genera.
Etymology. From the Greek diploos (=double) and achne (=awn), referring to the 2-toothed condition of some specimens of the lemma.

Key to the species of Diplachne
Stem anatomy. Stems are hollow in Diplachne fusca (Canfield 1934;Ebinger and Carlen 1975;Brown 1975;Auquier and Sommers 1967;Valls 1978). When branching occurs in D. fusca it tends to be concentrated in the upper nodes (Valls 1978). Valls (1978: 33) observed that branching in D. f. subsp. uninervia tends to occur after the terminal inflorescence is fully developed, suggesting that branching is facultative and dependent on favourable growing conditions, presumably adequate soil moisture.
Phenology. Flowering throughout the year in tropical latitudes; usually commencing early to mid-summer in temperate areas.
Distribtution. Native: Widespread and common to abundant in warm-temperate and tropical areas, between approximately 49°N and 40°S in the New World and 40°N and 42°S in Old World; mostly below 2000 m. Non-native: See under subspecies. Vernacular names. Malabar sprangletop; Chinese: shuang fu cao (双稃草) (and see others under subspecies).
Comments. Localised populations of the Diplachne fusca complex can be somewhat distinct morphologically from conspecifics occurring elsewhere, which is reflected in the many names that have been created to reflect such variation. However, all characters intergrade when considered globally (Snow 1997a), suggesting that the localised morphological variants do not merit recognition at the specific level. Field observations, herbarium work and multivariate statistical studies (Snow unpubl.) based on eleven population samples (n=20) from North America, Africa and Australia, which included over 80 morphometric traits, supported the recognition of four subspecies, which generally can be differentiated with little difficulty. These include: D. fusca subsp. fusca, a polymorphic Paleotropical taxon adventive in a few areas in the New World (Nicora 1995;; D. f. subsp. muelleri, known from much of the interior portions of Australia, particularly the Northern Territory; D. f. subsp. uninervia, native to the Neotropics but adventive elsewhere (e.g. Snow and Simon 1999) and D. f. subsp. fascicularis, native to the temperate and tropical regions of the New World.
Differentiating between subspecies can be particularly difficult in parts of California and Argentina, where D. f. subsp. fusca is adventive and sympatric with subspecies fascicularis and uninervia and in the Middle East and Australia (Western Australia, Queensland), where D. f. subsp. uninervia has become established (Snow and Simon 1999). It seems unlikely that D. f. subsp. fusca has persisted in California.

Key to subspecies of Diplachne fusca
Leaf anatomy. Midrib present (or rarely absent) in mature leaves; central lacuna present. Primary bundles: protruding adaxially and abaxially and to a greater degree than secondary bundles; outer bundle sheaths continuous adaxially, interrupted abaxially; extension cells present adaxially; adaxial cells enlarged, abaxially cells not enlarged; sclerenchymatous girders present adaxially and abaxially, or abaxially only as strands. Colourless cells present between primary and secondary bundles; chlorenchyma continuous or discontinuous between adjacent bundles. Secondary bundles: protruding adaxially or not, flush abaxially; outer bundle sheath continuous or interrupted abaxially; sclrenchymatous girders present abaxially. ( Ellis 1908, 3410, 3653, 4779.) Stem anatomy. Stems can root at the nodes and functionally act as stolons (e.g. Hitchock 24410 [US], Drège 3900 [P]; Valls 1978). When branching occurs, it is mostly towards the base and at lower nodes. Elliptic air canals subjacent to the epidermis of culms typically are present in populations emergent in water.
Phenology. Flowering throughout the year but seasonally locally. Distribtution. Native: Much of paleotemperate and paleotropical areas; mostly open mesic areas, often in saline conditions. Elevation sea level to ca. 2000 m. (TDWG:  ANG, BOT, BZE-PB, BUR, CHA, ETH, JAW, IND, IRQ, ITA-IT, KEN, MDG,  MLW, MLY-SI, MOZ, MYA, NAM, NGA, NGR, NSW-NS, NTA, OFS, PAK, PHI,  QLD-QU, SEN, SOA, SRL, TAI, TAN, THA, (Beetle s.n., 1941 [NA]). It has been collected outside the plots in California, but evidently has not become widely established in that state. Conservation status. Widespread; not of concern. Etymology. The Latin fuscus refers to dark, dusky, or swarthy and possibly alludes to the dark green or plumbeous spikelets of some specimens. Vernacular names. Malabar sprangletop; Beetlegrass sprangletop. Japan: Hamagaya (Koyama 1987 Comments. Diplachne fusca subsp. fusca is the most widely distributed taxon in the complex and occurs throughout the range of the species in the Paleotropics, with some introductions in the Neotropics and North America (California, South Carolina). It is sometimes considered a weed of rice paddies (McIntyre and Barrett 1985, McIntyre et al. 1989. Distinguishing D. f. subsp. fusca from subsp. fascicularis where the two are sympatric can be difficult, particularly in Argentina , where the populations of subsp. fusca may be non-native. The authors' interpretation of taxa in the complex differs from Nicora (1995), who synonymised Diplachne uninervia (J. Presl) Parodi var. procumbens Parodi under D. fascicularis. In contrast, the authors synonymise the former under D. fucsa subsp. fusca. The subsp. fusca is differentiated from subsp. fascicularis primarily by the latter having panicle branches inserted at the base and having anthers shorter than 0.5 mm, whereas the former more typically has fully exserted panicles and anthers exceeding 0.5 mm (sometimes significantly so).
The floret morphology of some specimens from western Africa ( The length of the awns varies throughout the range of D. fusca subsp. fusca and is too inconsistent to be of taxonomic value. The sexuality of the florets varies widely. A population from Namibia (Snow & Burgoyne 7196 [with numerous duplicates distributed]) had male, female and hermaphroditic florets within a single spikelet. The number of stamens can likewise vary within a spikelet from one to three (e.g. Snow and Burgoyne 7196), which has been noted also by Nicora (1995). The anthers also vary considerably in length between specimens, but typically are consistent within a specimen and within populations.
Thrips (Order Thysanoptera) of an indetermined genus were found in virtually every floret of fresh material of a specimen of D. fusca subsp. fusca from Queensland, Australia . In a study of germination, populations in Australia were affected more by nighttime than daytime temperatures (Myers and Morgan 1989).
Some populations of this subspecies have high to very high levels of tolerance for saline or sodic soils (Beadle 1948a-b;McVaugh 1983;Hackett and Wickens 1984;Myers and Morgan 1989;Kernick 1990;Chapman 1996;Ahmad 2009). This tendency is most developed in Australia, Asia and Africa and is least well documented for subsp. uninervia. For example, the label of Trapnell 1112 in Zambia states the species grows amidst ca. 5 cm of solid salts. Another specimen from Zambia (Greenway 6228) states that the species becomes a dominant pioneer in areas trampled out in brackish pans in grassland dominated by the grass genus Hyparrhenia E. Fourn. Others also have noted the high levels of salinity tolerance (Reinhold et al. 1986;Hurek et al. 1987;Reinhold-Hurek et al. 1993). Joshi et al. (1983) reported salt excretion on the leaves via small apical pores in papillate structures that they designated as salt glands. Chapman (1996) also reported that some populations of D. f. subsp. fusca excrete salt from glands on the leaves and that this taxon has been advocated for its use for lowering salt levels in soils. The ability of many populations of this subspecies to tolerate and even thrive in saline soils was one of many factors contributing to the decision to recognise them at the subspecific, rather than specific rank (Snow 1997a). In southern Africa, D. fusca subsp. fusca can be common to abundant in pans (low-lying areas that accumulate seasonal moisture but which dessicate in the dry season), where it sometimes is decumbent to geniculate and stoloniferous (Breen et al. 1993; e.g. see Hitchcock 24410 [US]; Snow 6829 [MO]).
Herbarium labels from African collections sometimes indicate grazing, presumably by the native fauna, and others indicate good grazing for livestock. The subspecies may occur in pure or nearly pure stands in Africa. In Australia, Lazarides (1970) reported the taxon as having high palatability and being preferentially grazed, but that it typically occurs in relatively small populations. Denny (1993aDenny ( , 1993b and Ahmad (2009) indicated that among grasses generally, Diplachne fusca appears to maximally combine adaptations to wetness and salinity, combined with palatability, of any species of grass. Kernick (1990: 171) linked the salinity tolerance and palatability of D. fusca and suggested that additional screening for salinity tolerant races would be worthwhile. The salinity tolerance of D. fusca subsp. fusca, high viability of its seeds (Chapman 1996; Snow, pers. obs. in greenhouse), and high palatability of the subspecies among ruminants (Chapman 1996; Snow pers. obs.) collectively make D. fusca an excellent candidate for reclamation of saline soils (Haq and Khan 1971;Sandhu et al. 1981;Chapman 1996;Ahmad 2009).
A remarkable biological attribute of Diplachne fusca subsp. fusca is the presence in some Pakistani populations (where it is called Kallar grass) of a recently described nitrogen-fixing genus of bacteria, Azoarcus (Reinhold et al. 1986Hurek et al. 1987Hurek et al. , 1993Hurek et al. , 1994Reinhold-Hurek and Hurek et al. 1997). Subsequent studies documented Azoarcus to be a nitrogen-fixing endosymbiont of many grasses, including rice Krause et al. 2006). Given increasing problems of salinisation in some areas and the biologically and economically important properties of the Pakistani populations (nitrogen-fixing, salinity tolerance, perennial growth form and high palatability), these populations should be investigated further for their reclamation potential of saline soils, given that they also could provide useful fodder. More recent research has examined the rhizosphere for bacterial species of D. f. growing in industrial sites (Abou-Shanab et al. 2005).
Diplachne fusca subsp. muelleri most closely resembles D. f. subsp. fascicularis by virtue of the lower inflorescence branches that typically remain inserted at the bases, especially from lateral inflorescences. However, the upper glumes are generally much broader than the primarily North American subspecies.
Leaf anatomy. Midrib present; central lucunae present. Primary bundles: protruding adaxially or not; protruding abaxially; outer bundle sheath interrupted adaxially and abaxially; extension cells present adaxially; adaxial sclerenchyma present as girders or strands; abaxial sclerenchyma present as girders; adaxial cells of primary bundle sheath cells enlarged; abaxial cells of primary bundle sheath cells not enlarged. Colourless cells present between primary and secondary bundles; chlorenchyma continuous or discontinuous between adjacent bundles. Secondary bundles: protruding adaxially or not; flush abaxially; outer bundle sheath continuous or interrupted abaxially; adaxial sclerenchyma present as girders or strands; abaxial sclerenchyma present as girders. Internodes hollow. Inner sclerenchymatous ring present. Peripheral sclerenchymatous ring present. Peripheral sclerenchymatous girders connected to the outermost vascular bundles absent. Intervascular peripheral sclerenchymatous pillars not associated with outermost vascular bundles absent. Inner sclerenchymatous ring canal tissue present. Kranz sheath cells absent. Kranz sheath cell canal tissue absent. Vascular bundles nested in outer portion of Kranz sheath cell canals absent. Sclerenchymatous rings surrounding vascular bundles located inside inner sclerenchymatous ring absent. Sclerenchymatous rings (5-10 cells thick) surrounding outermost primary vascular bundles absent. Phloem not divided.
Phenology. Flowering mostly May to October in North America and throughout the year in South America during and after warm and wet seasons.
Distribution. Southern Canada (infrequently) through USA (where most common) south sporadically to Paraguay. The transition between the native and non-native distribution of L. fusca subsp. fascicularis is somewhat uncertain, particularly in its sourtherly range. Elevation from sea level to 2300 metres. Etymology. Possibly in reference to the fasciculate (bundled or clustered) arrangement of the lower panicle branches before they are exserted from the sheath. ). On rare occasions, a specimen growing in a saturated soil may branch and root so vigorously at nodes that it appears to be stoloniferous (e.g. Fleetwood 12215 [SMU]). Older specimens from Reno, Nevada, identified as Diplachne tracyi Vasey, have narrow, relatively long and somewhat cylindrical spikelets (e.g. the type specimen), but since this spikelet morphology recurs in the species complex from parts of Africa and Australia and is narrowly distributed, this variant is unworthy of taxonomic recognition. Those specimens may be waifs of D. f. subsp. fusca from the Paleotropics, but genetic analyses would be necessary to test this hypothesis. Some specimens from California (Twissleman 6491 [MO]) also closely resemble some forms of D. fusca subsp. fusca.
A morphotype in the United States described as Diplachne acuminata Nash in Britton has been recognised based on the relative elongation of glumes and lemmas and awned lemmas (Gleason, 1968: 188;Gleason and Cronquist, 1991: 785). Diplachne maritima E.P. Bicknell was described based on its prostrate habit, more pronounced lemmatal awns and ecological occurrence in brackish coastal areas from Massachussetts to Florida (e.g. Specimens have not been confirmed for this subspecies listed on websites from Sweden (e.g. Lackalänga, year 1949 [LD, S]) and Poland (Vogel, years 1916 and1918 [OHN]).
Diplachne fusca subsp. facicularis is more widespread in North America than subsp. uninervia, although the latter is more invasive outside of its range (e.g. Snow & Simon 1999). Specimens morphologically intermediate beween D. f. subspp. fascicularis and uninervia occasionally are seen (Hitchcock 93 [TEX]), but the taxa maintain their distinct morphology in sympatry (e.g. Snow 5899 and 5900 at Falcon Lake, Zapata County, Texas). Sutton (1973a) briefly described the anatomy of D. fascicularis (as "fascicularis") and placed Leptochloa into the tribe Chlorideae (Sutton 1973b), but these papers extrapolated the limited analysis beyond their usefulness. Clifford and Watson (1977) illustrated D. fusca subsp. fusca (as Diplachne parviflora). Renvoize (1984) illustrated the midrib of D. fusca and discussed its prominent air canal, but cited no voucher.
Leaf anatomy.  renchymatous pillars not associated with outermost vascular bundles absent. Inner sclerenchymatous ring canal tissue absent. Kranz sheath cells present. Kranz sheath cell canal tissue absent or present. Vascular bundles nested in outer portion of Kranz sheath cell canals not applicable. Sclerenchymatous rings surrounding vascular bundles located inside inner sclerenchymatous ring absent. Sclerenchymatous rings (5-10 cells thick) surrounding outermost primary vascular bundles absent.
Chromosome number. n=10 (Gould 1958); 2n=20 (Gould 1958(Gould , 1968. Phenology. Flowering throughout the year in both Northern and Southern hemispheres as limited by cold or moisture availability. Distribution. The native and non-native distributions of Diplachne fusca subsp. uninervia are uncertain, particularly regarding the southwestern USA west of Texas and east of the Mississippi River (but excluding Florida), where it is considered to be non-native based on relatively few older collections (early 20 th Century or earlier). In contrast, the occurrences north of Virginia are certainly all non-native. Given its greater abundance in parts of the USA, México and South America (e.g. Argentina, Paraguay), it is hypothesised that this taxon is an amphitropical disjunct, although it was not reported as such previously (Peterson et al. 2007). Native: In the New World mostly south of Latitude 37°N, south to Argentina, occasionally adventive in Old World; open mesic areas, agricultural lands, saline flats, mangrove swamps. Elevation Conservation status. Least Concern (IUCN 2012) given its wide distribution. Etymology. Uninervia, meaning one-nerved, may be a reference to the prominent nerve on the glumes. Vernacular names. Mexican sprangletop. Argentina: hierba paymilla, (Aguilar 1 [NY]). México: zacate salado mexicano; zacate gigante peruano. Venezuela: paja de ratón (Campos 519 [US]).
Comments. Diplachne fusca subsp. uninervia has a greater tendency to become invasive and weedy compared to other taxa in the genus (Snow and Simon 1999;Perez et al. 2010 MO]). The often truncate (or sub-truncate) and frequently mucronate apex of the lemma and smoky white glumes can make this subspecies appear similar to the southern African form some have called D. cuspidata (e.g. Geiss & Van der Walt 12632 [M, MO]; see Gibbs Russell et al. 1991), which the authors include as a synonym of D. f. subsp. fusca.
The report of D. fusca from the Canary Islands (Scholz and Böcker 1996) applies to D. f. subsp. uninervia.
Leaf anatomy. Not studied. Stem anatomy. The stellate aerenchyma of D. gigantea probably reflects adapations for its emergent growth habit in aquatic situations. A series of relatively large, elliptic air canals are subjacent to the epidermal layers (Fig. 10). Chlorenchymatous tissue is also present near the surface. Peripheral sclerenchymatous ring present and linked to outermost vascular bundles via girders. Inner sclerenchymatous ring canal tissue present. Sclerenchymatous rings surrounding vascular bundles located inside inner sclerenchymatous ring present. Kranz sheath cells present, with vascular bundles present in outer portions of sheaths; Kranz sheath cell canal tissue present. Vouchers: Simon & Williamson 2025 (BM); Smith 4126 (MO).
Chromosome number. Unknown. Phenology. Flowering December through May. Distribtution. Native: A few widely scattered collections from Botswana, Tanzania and Zambia. Growing in edges of swamps and margins of slow moving rivers. Elevation 1250-1710 metres, although upper and lower ranges uncertain. (TDW: BOT, ZAM, TAN). Non-native: See below (cultivated) in Zimbabwe.
Conservation status. Data Deficient (IUCN 2012). However, detailed localities are provided below given the evident rarity of Diplachne gigantea, which apparently has not been collected since 1983. Focused attempts by the first author in January of 1996 to re-collect at the two localities of Pete Smith in the Okavango region of Botswana were unsuccessful, even after speaking with Smith regarding his knowledge of the species. To the best of the authors' knowledge, Smith is the last person to have collected the species. He knew of no vernacular name for the species in Botswana which, if it existed, might help in relocating efforts and indicated that each of the two populations  from which he had collected were small. He confirmed, as some herbarium specimens labels and the stellate culm anatomy suggest, that the species is an emergent aquatic that can continue to grow on sandbars.
More specifically, on 28 and 30 January 1996 the first author hired a boat and driver at Drotskys Cabins in Botswana, located approximately 20 km south of the Namibian boarder (Caprivi Strip) and approximately 7 km southeast of the small town of Shakawe (Highway A 35). One entire day each was devoted to searching for D. gigantea moving southeast downstream on the Okavango River and returning by the same route on the opposite bank of the river and upstream and back from the same point of departure. No individuals were seen. Although only speculative, two hypotheses concerning the evidently limited distribution in relation to its habitat merit consideration. First, the stellate aerenchyma of the culm likely contributes to a relatively soft texture, which may render it particularly palatable for large grazers such as hippopotamus (Hippopotamus amphibius), which are common in the Okavango Delta and which consume large quantities of food. Second, given the extensive potential habitat for D. gigantea in the Okavango Delta, D. gigantea may occur more frequently than herbarium records suggest. However, given the paucity of collections to date, it likely is uncommon anywhere in its range.
Diplachne gigantea was cultivated in the early 1980s at the Lakeside National Botanic gardens in Zimbabwe (Browning 8; A, P). Gibbs Russell et al. (1991: 118) indicate that D. gigantea has been collected at the western edge of the Caprivi Strip in Namibia but the authors have not seen a voucher.
In summary, the data suggest that Diplachne gigantea is rare and additional knowledge of its distribution, ecology and relative abundance are critical for making a recommendation following IUCN standards.
Etymology. The word gigantea refers to the large stature of the species. Vernacular name. Giant diplachne. Comments. Diplachne gigantea is easily diagnosed from D. fusca by its much taller stature at maturity, its evidently obligate and usually emergent growth in hydric habitats, the stellate aerenchyma of the culms, the absence or near-absence of lemmatal macrohairs and the relatively lax panicle branches. The collection from cultivation at the Lakeside Botanic Gardens (see below) is indicated as growing from a "dense mat at the base", presumably meaning the culms arose from short rhizomes.
The fourth author personally recalled the species as being fairly commonly locally at the time of his collection in 1983 (Simon & Williamson 2025).