Stem morphology and anatomy in Amaranthus L. (Amaranthaceae)-taxonomic significance

Journal of the Torrey Botanical Society, Jul-Sep 2001 by Costea, Mihai, DeMason, Darleen A

COSTEA, MIHAI (Department of Botany, University of Agronomical Sciences, Bucharest, Romania) and DARLEEN DEMASON (Botany and Plant Sciences, University of California, Riverside, CA 92521). Stem morphology and anatomy in Amaranthus L. (Amaranthaceae) taxonomic-significance. J. Torrey Bot. Soc. 128: 254-281. 2001.-The range of variation within the genus Amaranthus L. (Amaranthaceae) is described for a number of stem characters including: morphology, epidermis, primary stem vasculature and mechanism of secondary growth. The results provide new characters (phyllotaxy, complexity of leaf vascular supply and relative amount of secondary growth) that support (1) a new infrageneric classification (subgenus Amaranthus vs subgenus Albersia (Kunth)Gren. & Dodr.), and (2) the separation within the "hybridus" complex of cultivated amaranths (A. caudatus L., A. cruentus L. and A. hypochondriacus L.) from their presumed wild ancestors (A. hybridus L. subsp. quitensis (Kunth) Costea & Carretero, A. hybridus L. subsp. hybridus and A. powellii S. Wats. subsp. powellii respectively).

Key words: Amaranthus, stem, morphology, anatomy, trichomes, primary vascular system, secondary growth, taxonomy.

In spite of the fact that the genus Amaranthus has been the subject of many taxonomic studies, it is still poorly understood and is widely considered to be a "difficult" genus. It consists of about 70 species, of which about 40 are native to the Americas and the rest to Australia, Africa, Asia and Europe. The infrageneric classification of the genus is still an unresolved problem. The classification most frequently used was suggested by Sauer (1955) in which he designates 2 subgenera: Acnida (L.) Aellen ex K. R. Robertson, which includes the dioecious species; and Amaranthus, which includes the monoecious species. Traditionally, the subgenus Amaranthus has been divided in two sections: Amaranthus (= Amaranthotypus); and Blitopsis Dumort sensu lato (Thellung 1914; Covas 1940; Morariu 1940; Aellen 1959; Brenan 1961, 1981; Gusev 1972; Frey 1974; Carretero 1979; Robertson 1981; Hugin 1986, 1987; etc.). Carretero (1985, 1991) divided the section Blitopsis sensu lato in two sections: Blitopsis, which includes those species having indehiscent fruits and x = 17; and Pyxidium, which includes those with dehiscent fruits and x = 16. Another section, Puncticulatae, was proposed by Kowal (1954) for A. viridis and A. acutilobus, but it was not validated by later studies (Klopper and Robel 1989a; Costea 1997c). Recently, Mosyakian and Robertson (1996) proposed the subgeneric rank [subgenus Albersia (Kunth) Gren. & Godr.] for the section Blitopsis sensu lato. This infrageneric classification with 3 subgenera (Acnida, Amaranthus and Albersia) is based on classical characters, such as those of inflorescence and floral characteristics, and it would be interesting to see if other characters support it too.

At the species level, most of the taxonomic problems involve the most studied group of species, the A. hybridus aggregate. The actual taxonomic treatments in this group of species, assuming that the nomenclatural problems are solved, range between two extremes. At one extreme is that of Sauer (1950, 1967) in which he recognizes the cultivated taxa (Amaranthus caudatus, A. cruentus and A. hypochondriacus) as species. And at the other is that of Greuter (1981, 1984) who lumps the cultivated species with their putative wild progenitors (A. quitensis, A. hybridus and A. powellii respectively). All possible intermediate possibilities between these two opposing treatments, many of them published since the beginning of the century by Thellung (1907, 1914, 1919), have also been used (Aellen 1959, 1964, 1972; Dostal 1950; Morariu 1952; Brenan 1961, 1981; Gusev 1972; Ehrendorfer 1973; Townsend 1974, 1985, 1988; Carretero 1979, 1985, 1990; Stace 1991, 1997; Lambinon 1992; Cserepanov 1995; etc.). Detailed studies of the relationship among the amaranth species using cytological or molecular methods are often contradictory. However, the combined results support separating the cultivated and wild taxa, over combining them (Pal and Khoshoo 1972, 1974; Hauptli and Jain 1984; Wilkin 1992; Sammour et al. 1993; Greizerstein and Poggio 1994, 1997; Transue et al. 1994; Lanoue et al. 1996; Chan and Sun, 1997).

Some common, unusual features characterize the stem anatomy of many families in the Caryophyllales (including Amaranthaceae). Among these are anomalous secondary thickening, occurrence of two or more rings of primary vascular bundles and complex organization of leaf traces associated with leaf gaps (Gibson 1994). Some of these features have been used to characterize relationships between families within this order (Eckardt 1976; Cronquist 1981; Thorne 1983; Gibson and Nobel 1986).

Although there are some detailed morphological studies of seeds (Kowal 1954, Klopper and Robel 1989a; Costea 1997b), fruits (Klopper and Robel 1989a; Costea 1997a), and pollen (Costea 1998), anatomical characteristics of vegetative and reproductive organs have not previously been seriously used for the resolving taxonomic relationships of the genus. Viana (1993) described some general aspects of anatomy in A. viridis. The influence of herbicides on the stem structure was studied in A. retroflexus and A. powellii (Rugina et al. 1984; Nita 1997), and ecological factors in A. blitoides (Toma et al. 1994). The primary vascular system has been analyzed only in A. caudatus (Gravis and Constantinesco 1907), A. hybridus and A. graecizans (? = A. albus) (Wilson 1924).