Echinochloa crus-galli var. oryzoides (Ard.) Lindm., Svensk Fanerogamflora 69. 1918.

Hoste, Ivan & Verloove, Filip, 2022, Taxonomy of the weed species of the genus Echinochloa (Poaceae, Paniceae) in Southwestern Europe: Exploring the confused current state of affairs, PhytoKeys 197, pp. 1-31 : 1

Echinochloa crus-galli var. oryzoides (Ard.) Lindm., Svensk Fanerogamflora 69. 1918.

Basionym.

Panicum oryzoides Ard., Animadv. Bot. Spec. Alt. 2: 16, pl. 5. 1764.

Type.

LINN 80.68. Image available at https://linnean-online.org/1302/.

Note.

According to Carretero (1981), LINN 80.68 is a plant sent by Arduino to Linnaeus, possibly collected in Italy. It was designated as the lectotype for that name by Crespo et al. (2020a).

Remarks on E. crus-galli .

Echinochloa crus-galli s.l. is taxonomically the most complex Echinochloa occurring as a weed in Southwestern Europe. As we understand, this species occurs in a number of varieties, but E. oryzicola is not one of them and is accepted as a separate species (see below). As a result of a long and complex evolutionary history, including significant modifications in the recent past (after the introduction of agriculture), the differences among the varieties are often slight. Furthermore, introductions of several taxa as weeds in a range of crops far outside their natural range have contributed to obscuring their original geographical distribution. Rather than aiming at precisely describing the limits and defining features of varieties of E. crus-galli occurring in Southwestern Europe, we restrict ourselves primarily to indicating where unsolved problems remain.

Being extremely polymorphic, numerous varieties of E. crus-galli have been described, many of them based on the presence or absence of awns. As the development of awns is influenced by environmental conditions ( Michael 1983), the value of varieties or forms based on such characteristics as the presence or absence or the length of awns is quite limited. Inflorescences that develop later in the season frequently differ from the terminal inflorescence. Other characteristics on which the description of varieties has been based include the coloration of the plant (inflorescence, leaves, stem nodes, etc.), structure and position of the inflorescence (erect, bent or nodding; primary branches more or less patent or not, alternately positioned on the main axis or whorled), the arrangement of the spikelets on the branches and the dimensions of the spikelet. The importance of the length of the spikelets is emphasised by Michael (1983), who in his identification key for the annual Echinochloa separates the species characterised by spikelets measuring 3-5 mm from those with either shorter or longer spikelets. Applied to specimens collected in Southwestern Europe, the criterion of spikelet length works well to separate only the two rice mimics, E. crus-galli var. oryzoides (Ard.) Lindm. (syn.: E. oryzoides [Ard.] Fritsch) and E. oryzicola , from the remaining taxa of E. crus-galli s.l. with smaller spikelets.

There is a broad consensus that E. crus-galli var. crus-galli occurs in large parts of Europe and Asia, but authors differ on how to appropriately define it. Ibaragi (2020) stated that Asian var. Echinochloa crus-galli crus-galli slightly differs from plants in Europe, "but the differences are difficult to formally distinguish." Thus, the need for additional research on the morphological and genetic variation of the type variety throughout its range is evident.

According to Michael (2019), Echinochloa crus-galli var. hispidula (Retz.) Honda is the appropriate name for E. crus-galli with non-pyramidal panicles and usually prominently awned spikelets that are widespread in sub-tropical areas of Japan and Southern China. With slightly larger spikelets than var. Echinochloa crus-galli crus-galli , this taxon is often treated as a separate species, Echinochloa hispidula (Retz.) Nees ex Royle; however, Ibaragi (2020) completely ignored it, and Shouliang and Phillips (2006) interpreted it as synonym of var. Echinochloa crus-galli crus-galli . Its extreme variability ( Carretero 1981, as E. hispidula ) makes it hard to distinguish var. Echinochloa hispidula hispidula from var. Echinochloa hispidula crus-galli , which is characterised by a usually more or less procumbent habit (the lower nodes often rooting), floppy leaves, the whole plant or parts of it more often than not purple-tinged, erect to strongly bent pyramidal inflorescences with the branches alternately placed or sometimes whorled, with at least the lower branches usually more or less patent, and spikelets with or without awns, the length of the awns and the percentage of awned spikelets within a single inflorescence exhibiting considerable variation (Fig. 12 ; description based on material from maize fields in Belgium, where var. Echinochloa hispidula hispidula , a taxon of sub-tropical climates, is considered not established as a persistent weed). If accepted as a separate taxon, the strongly bent inflorescence with appressed branches, the green colour of the plants and the stiffer leaves (the latter two features shared with var. Echinochloa hispidula oryzoides and E. oryzicola ) might help to separate E. hispidula from E. crus-galli var. crus-galli ( Jauzein 1993). To this could be added the less bristly spikelets of var. Echinochloa crus-galli hispidula ( Martínez-Azorín and Crespo 2021). Whether the branches of the inflorescence are whorled or not ( Michael 1983) seems to be a less reliable trait to use. The synonymizing of E. erecta (Pollacci) Pignatti, characterised by an erect inflorescence, with E. hispidula (see, e.g., Ardenghi et al. 2015) underscores the wide morphological variation of Echinochloa hispidula and the difficulty to define it as a unit clearly different from the equally variable var. Echinochloa hispidula crus-galli . Interestingly, the illustration of E. hispidula given in Shouliang (1990) shows a spikelet with both the upper glume and lower (sterile) lemma with a short awn, a feature not mentioned in the recent literature; however, apparently, it corresponds with the specification "calycibus hispidis biaristatis" in the original description of the species ( Retzius 1789).

Within E. crus-galli as interpreted here, var. Echinochloa crus-galli oryzoides is the most easily identified variety, clearly distinguished by the large size of its spikelets. Although the descriptions given in triplet 4 in the key above may suggest otherwise, it is not always easy to distinguish between var. Echinochloa crus-galli oryzoides and E. oryzicola ; see the discussion about the latter species below. At one time, the name Echinochloa hostii (M. Bieb.) Link was used by Italian botanists ( Pignatti 1982). Previously, Pirola (1965) merely cited this name as a synonym of E. crus-galli subsp. oryzoides , but Pignatti (1982) accepted the name at species rank for the taxon that is here named E. crus-galli var. oryzoides . However, from Pignatti’s identification key, it is clear that the name E. hostii was used for the species that today can only be identified as E. oryzicola , based on the quite diagnostic glume characteristics. Unfortunately, we were not able to trace type material of Panicum hostii M. Bieb. which according to Tsvelev (1984) is preserved in LE. Thus, we do not know whether P. hostii is indeed identical with E. crus-galli var. oryzoides as stated by nearly all contemporary authors. Nonetheless, we are certain that the binomial E. hostii was wrongly applied by Italian authors (particularly Pignatti 1982) for E. oryzicola .

Recently, Martínez-Azorín and Crespo (2021) accepted E. crus-galli var. oryzoides as a species, just like the similar-looking E. oryzicola . The strong similarities shared by these two taxa are explained by a shared ancestor - tetraploid E. oryzicola being one of the parent species of hexaploid E. crus-galli - and recent convergent evolution as rice mimics derived from E. oryzicola and E. crus-galli (Fig. 13 ). This evolutionary trajectory provides an argument for assigning the rank of variety to E. oryzoides . Further, it would seem logical to reduce the rice mimic E. oryzicola to the rank of variety (provisionally 'var. Echinochloa oryzicola infestans ' in Fig. 13 ) as well, but since it is not known whether or how the ‘original’ E. oryzicola of pre-agricultural times differed from today’s E. oryzicola - because it is now extinct or goes undetected - this is not an option.

Among the forms with small spikelets, subsp. Echinochloa oryzicola spiralis (Vasinger) Tzvelev (no combination available as a variety) and var. Echinochloa oryzicola praticola Ohwi have been mentioned as occurring in Europe. Apparently solely based on the small spikelets, both names were synonymised by Scholz (2002), who noted that subsp. Echinochloa oryzicola spiralis - a taxon with a huge distribution area and possibly indigenous to Europe - and subsp. Echinochloa oryzicola crus-galli are polymorphic and that no sharp distinction between the two is possible solely based on the spikelet length. Martínez-Azorín and Crespo (2021) interpreted var. Echinochloa oryzicola praticola as probably no more than an impoverished form of E. crus-galli , and Tison and de Foucault (2014) seriously doubted the taxonomic value of subsp. Echinochloa crus-galli spiralis and var. Echinochloa crus-galli praticola . In Belgium also, E. crus-galli with small spikelets has been recorded, but nowhere have such plants been known to establish as noxious weeds.

A rather distinct form of E. crus-galli with spikelets ca. 3 mm long or a little longer (somewhat smaller than average var. Echinochloa crus-galli crus-galli ) has occasionally been observed in Belgium, including in the border of maize fields where, however, it seems not to establish easily and disappears after only a few years. These plants usually have an erect habit and rather stiff leaves. The inflorescence is erect, with patent branches. The purple-tinged spikelets are usually unawned (but a few spikelets may have a long awn), and some have a glabrous, convex and shiny sterile lemma. The lower leave sheaths vary from glabrous to densely covered with short retrorse hairs. Scholz (2002) included specimens with small spikelets with a convex, shiny sterile lemma in subsp. Echinochloa crus-galli spiralis , mentioning that the spikelet morphology resembles Echinochloa glabrescens Munro ex Hook.f. Another name for E. glabrescens is E. crus-galli var. formosensis Ohwi ( Yabuno 1983; Ibaragi 2020); this name was used by Japanese authors for a weed of wetland rice fields. The habitats in which the plants were found in Belgium stand in contrast to those preferred by var. Echinochloa crus-galli formosensis in Japan. Adding to the confusion, the name E. glabrescens has also been applied to plants with spikelets 3.5-5 mm long by Bor (1960), Shouliang and Phillips (2006; as E. glabrescens Kossenko) and Xia et al. (2011). Echinochloa with small spikelets and a shiny lower lemma collected in Europe requires more study in order to reveal its true identity and its relation with similar taxa having small spikelets in Asia and E. oryzicola ; see, e.g., Yasuda and Nakayama (2019). Although a quite distinctive feature, the convex and shiny sterile lemma might prove to be of little value taxonomically. Bor (1960) wondered whether the "most peculiar" feature of the indurated sterile lemma in E. glabrescens was sufficient to make it a good species, and Yabuno (1966) indicated that in E. oryzicola , the convex lemma is a simple dominant characteristic (Fig. 4B, C ).

Considering the preceding discussion, we accept, for the present, only few varieties of E. crus-galli as occurring in Southwestern Europe. Indigenous and quite variable var. Echinochloa crus-galli crus-galli , usually with a less erect habit and more floppy leaves, is by far the most widespread variety, especially towards the north. Part of the variation observed in Europe is perhaps due to the involuntary introduction and establishment of populations of var. Echinochloa crus-galli crus-galli , with slightly different morphological features, from Asia. Echinochloa crus-galli var. oryzoides , characterised by large spikelets, a more erect habit and stiffer leaves, is a rice mimic in rice fields of Southern Europe. Echinochloa crus-galli var. hispidula , in some respects resembling var. Echinochloa crus-galli crus-galli and in others var. Echinochloa crus-galli oryzoides , appears to us not to deserve a separate status and is, therefore, included in var. Echinochloa crus-galli crus-galli .

Plants with small spikelets are the most difficult to interpret. Probably representing more than one taxon - quite possibly including taxonomically irrelevant forms of var. Echinochloa crus-galli crus-galli - they require additional study, which will need to include material of Asian origin.

Finally, it can be argued that Echinochloa esculenta (A. Braun) H. Scholz, a cultivated taxon derived from E. crus-galli , should be included in E. crus-galli ( Banfi and Galasso 2021). Here, in line with most recent floras, it is pragmatically accepted as a well-defined separate species. However, species rank is justifiable based on morphological features, as the two taxa clearly differ from each other and identification of E. esculenta is usually not much of a problem.

Figures

Figure 4. Spikelets of Echinochloa oryzicola A fertile lemma with the tip differentiated from the coriaceous body of the lemma (upper glume removed) B, C two spikelets with convex shiny sterile lemma D spikelet with long lower glume and non-shiny sterile lemma. (Photograph: Andre De Kesel, Meise Botanic Garden).

Figure 12. Habit of Echinochloa crus-galli var. crus-galli growing as a roadside weed, Belgium. Although extremely variable, the usually more or less procumbent habit (the lower nodes often rooting) and the floppy leaves are among the features that distinguish var. Echinochloa crus-galli crus-galli from the obligate rice weeds E. crus-galli var. oryzoides and E. oryzicola which are characterised by a more erect and stiffer habit. (Photograph: Luc Audenaerde).

Figure 13. Schematic reconstruction of the evolutionary history of Echinochloa oryzicola, E. crus-galli and two rice mimics derived thereof. It is hypothesised that the taxon that today is called E. oryzicola has only recently evolved from a wild taxon that seems no longer to exist or has not yet been identified. In the absence of information on this original species, it is impossible to distinguish between a long-existing taxon (' var. Echinochloa oryzicola oryzicola ') and a recently evolved rice mimic (' var. Echinochloa oryzicola infestans ').