Amaranthus tunetanus (Amaranthaceae), a new species from Tunisia and a diagnostic key to the North African taxa in subgen. Albersia South African Botany

(Amaranthaceae), A . tunetanussp.nov. fromMonastirGovernorate(centralTunisia) is described and illustrated. The new species is morphologically similar to A. crassipes , A. crispus , A. graecizans subsp. aschersonianus , and A. scleropoides from which it differs mainly by characters of syn ﬂ orescences and ﬂ owers. Distribution in Tunisia, notes of its preferred habitat, phenology, and the IUCN status of conservation are also provided. A diagnostic key to the 14 taxa of subgen. Albersia occurring in North Africa is also presented. 2017 by Elsevier B.V. reserved.


Introduction
Amaranthus L. (Amaranthaceae Juss.) is a genus of about 70 mostly annual monoecious and dioecious species with a worldwide distribution. Approximately 40 species are native to the Americas, the remaining ones to all the other continents except Arctic and Antarctic regions (see e.g., Mosyakin and Robertson, 2003;Iamonico, 2015a). Several American species are used as ornamentals, and some of them are able to escape from cultivation, negatively impacting the agricultural systems through a reduction in both productivity and crop quality (see Iamonico, 2010).
The genus Amaranthus is critical from a taxonomic point of view due to its high phenotypic variability which has resulted in the current nomenclatural disorder and misapplication of several names (see e.g., Costea et al., 2001;Bayón, 2015;Iamonico, 2016aIamonico, , 2016bIamonico, , 2016c. According to Le Floc'h et al. (2010) and Iamonico (2015b), 10 species of Amaranthus currently occur in Tunisia.
As part of an ongoing investigation on the Tunisian Amaranthaceae s. lat. (Sukhorukov et al., 2016;El Mokni, 2016, 2017), we found some populations whose morphological characteristics do not match those of any other known Amaranthus taxon. We here propose A. tunetanus Iamonico & El Mokni sp. nov. as a new species to science.

Material and methods
The work is based on field surveys, analysis of relevant literature and examination of specimens preserved at CAT, G, FI, HFLA, K, P, RO, NY, and US (acronyms according to Thiers, 2017 +) and in the personal collection of one of the authors (R. El Mokni) deposited in the herbaria of the Faculty of Pharmacy of Monastir and the Faculty of Sciences of Bizerta (not listed in Index Herbariorum).

Etymology
The species epithet is dedicated to the country Tunisia.

Phenology
Two periods of flowering and fruiting were observed in the populations found. The first period ranges from November to February (flowering time November-January; fruiting time December-February), the second period ranges from April to July (flowering time April-June; fruiting time May-July).

Distribution and habitat
Amaranthus tunetanus is restricted to two localities of the Monastir Governorate: Monastir (two populations), and Jemmel (two populations). The species grows on clay and sandy substrates in meadows, margins of cultivated fields and railroads, uncultivated lands, and flower beds, at 32 m a.s.l. (type recorded at 3 m a.s.l.). Pyšek et al., 2004) The question regarding the status as native or not to the region is a crucial matter (see e.g., Pyšek et al., 2004). Webb (1985) and Preston (1986) provided the following useful criteria for presuming the origin status of a taxon: paleobotanical data, historical records, habitat, geographical distribution, genetic diversity, reproductive biology, and related phytophagous. According to Pyšek et al. (2004), most of these criteria are still relevant. Lacking desirable fossil and/or DNA evidence for Amaranthus tunetanus, we must consider the other criteria for evidence of its native status.

Origin status (sensu
No historical records were confirmed for Amaranthus tunetanus as we have not traced specimens that could be matched to the new species, including any of the European and American herbaria consulted (see e.g. the list of 45 herbaria in Iamonico, 2015a). The distribution area of the new Tunisian amarant appears to be restricted to the centraleastern part of Tunisia. According to Webb (1985: 233-234), the restricted distribution of A. tunetanus indicates that it is a native of Tunisia. In addition, another criterion to which we can refer is the reproductive biology (Webb, 1985: 235). Since Amaranthus tunetanus is observed to reproduce only by seeds, we may assume that A. tunatanus is native. In contrast, its preference for man-made habitats suggests that A. tunetanus is an alien species (Webb, 1985: 234). Pyšek et al. (2002Pyšek et al. ( , 2004 specified that plants introduced to a region after ca. 1500 should be named 'Archeophytes'. Pyšek et al. (2004: 138) also pointed out that "Since many archaeophytes now occur only in human-made habitats, we can ask, on the basis of the knowledge of their ecology, whether we can identify their potentially native habitat in the landscape before it was affected by people". Hence, A. tunetanus could be considered as archeophyte, but potentially native.
In addition to the criteria proposed by Webb (1985) and Preston (1986), and the discussion made by Pyšek et al. (2002Pyšek et al. ( , 2004, we can consider the following taxonomic and nomenclatural question: are the plants found in central-eastern Tunisia already been named elsewhere? No plant showing morphological characteristics of A. tunetanus appears to be described up till now. This is based on our I) studies at national or continental levels (e.g., Iamonico, 2015aIamonico, , 2015bIamonico and Das, 2014;Iamonico and El Mokni, 2017), II) analysis of recent literature (e.g., Bayón, 2015;Das, 2015;Iamonico, 2015b;Sánchez-Del Pino et al., 2017see also literature therein) and III) discussions with many other Amarants' experts (e.g., N. Bayón from S-America, I. Sánchez-Del Pino from C-America, S. Das from India, J. Palmer from Australia). Furthermore, any of the hundred names which were studies from the nomenclatural point of view (analysis of protologues, original material and types) by one of us (DI, see see Iamonico, 2014aIamonico, , 2014bIamonico, , 2016aIamonico, , 2016bIamonico and Clementi, 2016) matched the Tunisian material of A. tunetanus.
All things stated, we here consider Amaranthus tunetanus as a native species of Tunisia.

Conservation status
The distance between the two sites where A. tunetanus occurs is about 30 km. The Monastir populations are about 4 to 5 km apart while the Jemmel's populations are 3-4 km apart. The known AOO is about 27.87 km 2 , while the EOO is 110.94 km 2 . No protection plans exist for the areas where the new species occurs. Threats are represented by: residential and commercial development (urban areas), transportation and service corridors (roads), human intrusions and disturbance (recreational activities), and pollution (agricultural effluents). On the basis of the application of the IUCN criteria B1 and B2a, biii (2014), A. tunetanus can be categorized as Endangered (EN).

Taxonomic notes
Amaranthus tunetanus is morphologically similar to species belonging to the subgenus Albersia  (Table 1). All these species share a common character, viz. the shape of the tepals which are spathulate and expanded distally. However, they differ in a number of ways: A. scleropoides has circumscissile dehiscent fruits, whereas the fruits of A. tunetanus as the other species (i.e. A. crassipes, A. crispus, A. standleyanus) are indehiscent. The new Tunisian species differs from A. scleropoides and A. crassipes in having the synflorescence not thickened at any stage of the life cycle, while in the latter two species the synflorescence is thickened, becoming indurate (= hardened) at maturity, a unique feature of Amaranthus. A. standleyanus differs from A. tunetanus by the following characters: shape and length of bracts [ovate-lanceolate, 0.5-2.0 vs. narrow ovate, (2.3 −)2.5-3.0(− 3.2) mm], structure of the perianth (tepals free vs. connate in the proximal 1/4-1/3), expansion of tepals of pistillate flowers (uniform vs. hyaline in the distal part), length of tepals of pistillate flowers [1.5-2.8 vs. (2.7-)3.0-5.5(−5.7) mm]. A. crispus can be easily distinguished from A. tunetanus by its habit (prostrate to ascendent vs. erect), leaf blade margin (highly undulate vs. entire or slightly undulate), expansion of tepals of pistillate flowers (uniform vs. hyaline Smooth at the centre, slightly striate on the margin in the distal part), length of tepals of pistillate flowers [1.3-2.0 vs.
(2.7-)3.0-5.5(−5.7) mm], seed surface (dotted on the margin vs. slightly striate on the margin). A further taxon which could be confused with A. tunetanus, due to their leaf shape and the terminal spike-like synflorescence, is A. graecizans L. subsp. aschersonianus Thell. The main difference between these two taxa is that the tepals are spathulate with an expanded distal part in A. tunetanus, as opposed to ovate-lanceolate without an expanded distal part in A. graecizans.

Taxonomic notes on Amaranthus subgen. Albersia in North Africa
According to SANBI (2012 and literature therein), Iamonico (2015b) and Sukhorukov et al. (2016), the number of taxa in subgen. Albersia sensu Mosyakin and Robertson (1996) currently occurring in North Africa is 14 (including three subspecies and two varieties). We here present a diagnostic key to these taxa (the sexual characters reported in the key refer to the pistillate flowerssee e.g., Bayón, 2015;Iamonico, 2015a).
Key to the N-African species of Amaranthus belonging to subgen. Albersia sensu Mosyakin and Robertson, 1996: 1a