The geographically isolated annual teosinte from the coastal plain and embayments/estuaries near the Gulf of Fonseca, Nicaragua, is differentiated from Zea luxurians from southeastern Guatemala by its much longer and more abundant tassel branches, larger number of spikelets per branch, and longer, more pronouncedly transversely rugose outer glumes, as well as by its low-elevation coastal habitat. Molecular (ribosomal ITS) evidence and other data, placing it basal to Z. luxurians, seem to support this taxonomic segregation. RESUMEN. El teosinte annual que crece en aislamiento geogrAfico en la planicie costera y los bajos inundados y esteros cerca del Golfo de Fonseca, Nicaragua, difiere de Z. luxurians del sureste de Guatemala por el nuimero mayor de ramas y espiguillas por rama de la inflorescencia masculina, por la rugosidad transversal pronunciada de las glumas inferiores, adem-s de altura baja de su h ibitat costero. Evidencia molecular ubica a esta especie como mas primitiva a la especie Z. luxurians y apoya a esta segregaci6n taxon6mica. "There are very recent but as yet unverified reports of teosinte [in Honduras] from the mountains of the Department of Francisco Mordizan, not far from Tegucigalpa, from the mountains of Copdn, and even Nicaragua" (Standley, 1950: 61). In November 1989, entomologist Allan H. Hruska, following up on a tip from Adrian Ramos, a fellow worker in the CARE Nicaragua maize improvement project, located a population of teosinte (a wild species of Zea) at Km 178 on the Chinandega-Somotillo highway in southwestern Nicaragua. Subsequently, ripe seeds were collected and, in 1990, sent to his mentor, Lowell ("Skip") R. Nault, eminent leafhopper taxonomist and maize virologist at the Ohio Agricultural Research and Development Center, Wooster, Ohio, who promptly sent two of these to Iltis at the University of Wisconsin-Madison. Iltis, in turn, after noting their resemblance to those of Guatemalan Z. luxurians (Durieu & Ascherson) Bird, forwarded the two seeds to John Doebley (then at the University of Minnesota) for preliminary analysis. This revealed great morphological similarity to, but also considerable differences in developmental behavior from, Guatemalan teosinte. The general area of Hruska's 1989 discovery in southwest Nicaragua was visited by the authors on 5 October 1991. No teosinte was encountered there; however, repeated inquiry about the location of additional stands of "teosinto," the local name, hinted at by Hruska in a letter, led us several days later to the discovery of a small population some two hours by horse from Rancho Apacunca, some 19 km to the west of Hruska's site. Now evidently extremely local and rare, the teosinte at this location is remarkable for its ability to grow in as much as 0.4 m of standing or slowly moving water. Zea nicaraguensis Iltis & Benz, sp. nov. TYPE: Nicaragua. Chinandega: Pacific coastal plain near Golfo de Fonseca, edge of now flooded [by mid-December, bone-dry] area between Estero Paimica and Estero Palo Blanco, at a place seasonally occupied by cattle herders known as El Rodeo (12'53'45"N, 86?59'W), ca. 4-5 km NNW of Apacunca (12'52'30"N, 86o57-58' W), 18-20 km SW of Villa Nueva, alt. 9 m, 8 Oct. 1991, H. H. Iltis, B. F Benz & A. Grijalva 30831 (holotype, HNMN; isotypes, K, MEXU, MO, NY, US, WIS-2). Topotypic paratype of same population, in ripe fruit, 18 Dec 1991, H. H. Iltis, M. Castrillo, C. Medina & C. Aker 30919 (HNMN, WIS). NovoN 10: 382-390. 2000. This content downloaded from 207.46.13.111 on Sun, 22 May 2016 05:32:52 UTC All use subject to http://about.jstor.org/terms Volume 10, Number 4 2000 Iltis & Benz Zea nicaraguensis (Poaceae) 383 Similis Zeae luxurianti, sed robustior, culmis ad bases 3-5 cm diam., 2-4(-5) m altis, inflorescentiis masculinis robustioribus et 23-32 cm longis, cum 27-38 ramis, differt. Maize-like erect annuals, unbranched to candelabra-branched, 2-4(-5) m tall (Fig. 1A), sometimes with 1 to 4 stout basal tillers (suckers), these to 2 m long in giant plants, with prominent prop roots (Fig. 1D). Male inflorescences (tassels) 23-32 cm long, with 27-38 slender branches (including infrequent secondary branches from the base of basal primary ones), the central one undifferentiated [all plants of the type not quite mature as of date of collection (8 Oct.), hence weak, with all the branches drooping downward together on one side] (Fig. 1B); branches each with ca. 36 to 64 sessilepedicellate spikelet pairs, the outer glumes with two prominent keels that merge at the apex, ca. 911 mm long, strongly transverse rugulose especially when young (Fig. 1C); culms (tassel peduncles) immature in the type specimens, only ca. 3-6 cm long, but dramatically elongating to 25 cm or more by maturity; tassel branching axes variable, 8-12 cm long. Female spikes in sheathed complex clusters, 6-18 cm long, on peduncles of variable length, enclosed by glabrous, linear-lanceolate spathes, these 6-10 cm long; fruitcases ca. 4 to 10 per spike, trapezoidal appressed-cylindric, very smooth and shiny, ? uniformly colored a light to dark (coffee or earthen) brown, 7-10 mm on the long (glume) side, 3-6 mm on the short side, 4-6 mm diam. (Fig. 1C). Flowering (tasseling) from mid-October to early November, with fruitcases mature in mid-December. Common name "Teosinto." The still somewhat immature tassel of one of the two WIS isotypes (Fig. 1B), symbolized by the second highest filled circle on Figure 2, was used in 1992 for the University of Wisconsin Herbarium's traditional New Year's card, photocopied directly and unreduced from the specimen onto herbariumsize sheets and sent to over 400 herbaria, botanists, and friends all over the world. Distribution, ecology, and habitat. (All locations on lower half of Nicaragua topographic sheet Villa Nueva 1:50,000, 1989): Zea nicaraguensis appears to be restricted to the Department of Chinandega, Nicaragua, in the lowlands of the Gulf of Fonseca that drain to the nearby Pacific Ocean, at elevations of 9 m (type collection) to 75 m (according to Hruska's seed collection, presumably from near Israel; see below). We were told by some local old campesinos that this very flat landscape was once covered mostly by savannas of Crescentia, with gallery forests only near the confluence and levees of seasonal streams. (1) The Apacunca (holotype) locality: The nearby savannas had in the past often been cleared for agriculture, once for rice, then for pasture, and are now heavily grazed by cattle. At the October date of collection, this vast treeless plain was under 0.21.5 m of very slow-moving water but deceptively covered with a dense, lush, and green, rooted floating mat of mostly introduced Old World and some native New World grasses, such as Brachiaria mutica (ForsskMl) Stapf, Echinochloa polystachya (HBK) Hitchcock, E. pyramidalis (Lamarck) Hitchcock & Chase, Oryza latifolia Desvaux, Paspalum virgatum L., and Hymenachne amplexicaulis (Rudge) Nees, as well as colonies of Neptunia sp., Thalia geniculata L., and Echinodorus paniculatus Micheli, the last two often locally dominant. The dense gallery forest, where the teosinte grew, was located on a slightly higher river levee, which formed somewhat of an island, where, inaccessible to cattle during flooding, the teosinte was thus protected especially in its early stages from grazing (Fig. 3). Teosinte, now rare, was reported by the old cowboys at Rancho Apacunca to have been "ubiquitous" here in the past. With floodplain savanna on one side, and a seasonal flooding river (estero) on the other, the strip of gallery forest of Ceiba pentandra (L.) Gaertner, Pithecellobium dulce (Roxburgh) Bentham, Enterolobium cyclocarpum (Jacquin) Bentham, and Bravaisia integerrima (Sprengel) Standley had an understory of Coccoloba (2 spp.), Bactris minor Jacquin, Cassia reticulata L., Caesalpinia spp., and Crescentia alata HBK, and here, also, towering thickets of gigantic plants of teosinte. Growing both in shade and on the forest's sunny margins, their stems and prop roots were covered by shallow standing water 10-40 cm deep, this by 8 October beginning to recede (Fig. ID). But only a very few of the several hundred teosinte plants were now starting to "shoot out" their male inflorescences, and, despite diligent searching, all we could find were nine somewhat immature tassels, which became the type collection (see Discussion). (2) The Cayanlipe locality (for paratype and second population) lies in the flat, mostly treeless Comarca Cayanlipe 4 km south and east of the village of Cayanlipe in the Llano de Toreras and 7 km east-southeast of the Apacunca station. Although this area gets flooded as well as during the rainy season, the dirt road nearby is not elevated like the one leading to Rancho Apacunca, a circumstance that prevented us from visiting this teosinte population in October. Zea plants here are ca. 3.0-3.5 This content downloaded from 207.46.13.111 on Sun, 22 May 2016 05:32:52 UTC All use subject to http://about.jstor.org/terms