THE NORTH AMERICAN SPECIES OF PASPALUM By Agnes Chase INTRODUCTION BASIS AND METHODS OF WORK The revision of Paspalum here offered is based primarily upon the collections in the United States National Herbarium. The collections of PaspaHum in several of the large herbaria of this country and in Europe have also been examined. As in other groups, nearly all the early descriptions of American species of Paspalum were published by European authors. It has been necessary, therefore, to seek the basis of their work in European herbaria. The writer studied the collec- tions in several of the herbaria listed in the revision of Panicum,1 and visited besides Freiburg, where, at the Botanical Institute, is pre- served DoelPs own herbarium; Leiden, where types of Persoon and Steudel are preserved in the Rijks Herbarium; and Pisa, where Raddi's types are preserved at the University. The Hackel Herbarium, listed in the revision of Panicum under Attersee, the home of Prof. Eduard Hackel, is now in the Natural History Museum in Vienna. Dr. J. Daveau kindly lent for study the collections of Paspalum in the Institut Botanique, Montpellier, and Dr. Ren6 Viguier those in the Institut Botanique at Caen. The plants studied by Fliigge, author of Graminum Monographiae, Part 1, Paspalus (Reimaria), published in 1810, have not been located. Johann Fliigge was a physician of Hamburg. He appears to have used the Willdenow Herbarium, preserved in the herbarium of the Botanical Museum in Dahlem, Berlin, at least for the Humboldt and Bonpland collections. In the herbarium of the British Museum are a number of grasses from the herbarium of Ernst Ferdinand Nolte, with labels bearing names and data agreeing with those in Flugge s work, the plants themselves agreeing with the descriptions. According to Murray and Britten2 the British Museum purchased a selection of plants from the Nolte Herbarium in 1875, the year of * Contr. U. S. Nat. Herb. 15: 2-4. 1910. 1 Hist, Coll. Nat. Hist. Brit. Mus. 1:171. 1904. 1 2 CONTRIBUTIONS PROM THE NATIONAL HERBARIUM Nolte's death. Nolte seems to have purchased a number of classic herbaria, the collections of Forsk&l, Cavanilles, Delile, and Allioni, being among those in his herbarium. It may be that he purchased Fliigge's herbarium. Fliigge's name nowhere appears on the speci- mens examined in the British Museum. TYPE SPECIMENS Like the revision of Panicum the work on Paspalum has been done on a type basis. For a discussion of this method of work the reader is referred to the revision of Panicum and especially to Hitchcock's revision of Agrostis* The basis of each name, valid or synonymous, is stated in each case, and also the herbarium in which the type specimen is preserved, SPECIES, SUBSPECIES, AND FORMS The attitude expressed in the revision of Panicum4 has been strengthened during subsequent years. Judgment concerning the taxonomic rank of groups of Paspalum has been based on a great amount of herbarium material and on somewhat extended field work. In some species, such as Paspalum coryphaeum, field work has shown that what in the herbarium appear to be fairly distinct species are not even taxonomic forms, but phases, from young simple plants to old branching ones, or habitat forms. Others, like P. laeve and P. pli- catvlum, might be divided into a great number of varieties and forms, but having so divided them it would still be difficult to assign many plants to any particular form, the characters failing to make constant combinations. The fact seems to be that the species actually are variable in length of leaf, pubescence, and, within narrow limits, size of spikelet. Some, such as the allies of Paspalum setaceum, form a network of closely related species. The entire Setacea group might be reduced to one or two species with subspecies, varieties, subvarie- ties, and forms, and the Laevia to another, but I do not see that greater definiteness would be gained thereby. A name is the expres- sion of a taxonomic idea.5 If it were obvious which of the species involved was the parent and which the descendants, and the order of their descent, it would express a taxonomic idea to make the parent form a species with the others ranged under it, according to their descent. But the study of a vast amount of material in the field and in the herbarium does not indicate any definite descent. I surmise the parent form is long since extinct, that P. cUiatijolium, P. pubescens, and P. stramineum are brothers, that the others are cousins in various 1U. S. Dept. Agr. Bur. PI. Ind. Bull. 08: pp. 12-14. 1905. * Hitchc. & Chase, Contr. U. S. Nat, Herb. 16:7. 1910, * Greene, Landmarks of Botanical History, 122. (Smithsonian Misc. Coll. 54: no. 1870.) 1909. CHASE—THE NORTH AMERICAN SPECIES OF P ASP ALUM 3 degrees removed. Darwin's diagram6 best expresses the case. To place P, pubescens as a subspecies of P. ciliatijoliurn, with P. strami- neurn and P. mpinum under that as varieties,. and P. setaceum as another subspecies, with P. debUe and P. longepedunculatum as varieties under it, would express an exactness of relationship that the facts do not at all warrant. All we can truthfully say is that the species are closely related and appear to intergrade, whether by hybridization or by reverting to ancestral forms we do not know. It may be, as some geneticists hold, that there are far more real species than we recognize, but if these " elementary " species can not be recognized, it seems needless to cumber bibliography with them. The descriptions are based on the great majority of plants studied, and specimens that disagree slightly with these descriptions are cited as exceptional or intermediate. GEOGRAPHICAL DISTRIBUTION The species of PaspcUum, numbering probably nearly 400, are plants of the Tropics and warm temperate regions, especially of the Western Hemisphere, and are particularly abundant in Brazil. The genus extends to the northern United States, but no species are known from Canada or from the Rocky Mountains north of Colorado. Two littoral species and two tropical weedy ones are of world-wide distri- bution but are undoubtedly of American origin. The relatively few species not native to the Americas are mostly allies of PcLspalum scrobicvlaium L. and related to our dark-fruited species of the Plica- tula group. The species of Pa&palum are found in various habitats from sandy coasts, marshes, savannas, and prairies, to forested or brushy slopes, where some assume a clambering habit. In much of the high campos of Brazil Paspalum is the dominant genus of grasses. TEXT FIGURES Each species is illustrated by a text figure showing a portion of the inflorescence, only enough to suggest its general character, natural size, two views of the spikelet and one view of the fruit magnified 10 diameters. The four full page habit drawings were made by Mary Wright Gill, the other figures by the author. ECONOMIC IMPORTANCE OF FASPALTTM Paspalum is one of the most economically important genera of Paniceae, the tribe to which it belongs, only Panicum, containing the common millet, Ohaetochloa (Setaria), containing the foxtail millets, and Pennisetum, containing pearl millet and elephant or Napier grass, exceeding it in value. ' Origin of Species, chap. 4, opposite p. 90, Amer. ed. 1887. 4 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM A large number of native perennial species of Paspalum, especially P. circvlare, P. laeve, and related species, and P. pubescens and its allies furnish excellent grazing from the Middle Western States south- ward. They are mostly late summer and fall grasses, furnishing fresh forage when earlier pasture grasses are dry or have been con- sumed. Paspalum pubiflorum and more particularly P. pubiflorum glabrum are excellent drought-resistant pasture grasses in the Gulf States and northward to Kentucky and Oklahoma. They, as well as P. laeve and P. circulate, are cut for hay to a limited extent. The chief value of the native perennials, however, is for grazing. Paspalum boscianum, commonly called bull-grass or bull-paspalum, a succulent annual, furnishes good hay in the Southeastern States, where it makes abundant volunteer growth in fields after cultivation of the crops has ceased. It reseeds freely and is not planted. It is especially valuable for dairy cows, but is hard to cure unless dried on frames. Paspalum dUaiatum, widely known as paspalum-grass, water-pas- palum, water-grass, or more commonly simply paspalum, and in recent years as Dallis grass, was introduced into the southern United States from Uruguay or Argentina about the middle of the last century and is now common throughout the Gulf States where it is considered a valuable pasture grass especially for dairy cattle, furnishing excellent late summer and autumn feed, withstanding close grazing, and not being injured by moderate frosts. It is most important in northern and central Georgia, Alabama, and Mississippi, where it is cut for hay as well as grazed. The names water-paspalum and water-grass are misleading because, though this species is naturally a plant of moist lowlands, records show that it stands extreme drought where there is a good rainfall at some time of the year. In the Hawaiian Islands, particularly on the island of Hawaii, and in Guam Paspalum dUatatum, introduced from the Southern States, is building up a livestock indus- try. In the Philippine Islands it is widely planted by dairymen and ranchers for pasture, but is not cut for hay. Paspalum urviUei, Vasey grass, is related to P. dilatatum and was introduced somewhat later from Uruguay or southern Brazil. It is coarser and taller than P. dUatatum, and though readily grazed while young, becomes woody and unpalatable with age. In some sections of the South, particularly Arkansas and Louisiana, it is cut for hay. It was introduced into Australia and South Africa, where it is now regarded as a useful fodder grass. The early introductions of this grass into the British dominions were misidentified as Paspalum vir- gatum and that name still persists, an unfortunate error, true P. mrgatum being one of the unpalatable group called in Spanish Amer- ica "cortaderia" because of the razor-edged blades. The name P. CHASE—THE NORTH AMERICAN SPECIES OP PASPALUM 5 virgatum for P. urvUlei has recently found its way into Hawaiian agricultural literature. Paspalum notatum has been introduced as a pasture grass in the Gulf States under the name Bahia grass. It is proving of value, especially in Florida, thriving on ty>th clayey and sandy soil and being readily grazed. It forms the main constituent of native pasture in Cuba and Porto Rico and in parts of Costa Rica, southern Brazil, Uruguay, and Argentina. It has recently been introduced into East Africa sjad is said to furnish excellent lowland pasture in Uganda. Paspalum notatum is coming into use for golf courses in the West Indies and in Panama, because it forms a tough heavy turf even on rather light sandy soil. Less important are the native Paspalum lividum and P. hartweg- ianum in Texas and Mexico. The latter has recently been grown in Mississippi as a forage grass. It is sensitive to frost but is promis- ing for pasture along the Gulf Coast. Paspalum vaginatum is an important soil binder on low sandy coasts in the Tropics and sub-Tropics, and P. distichum, known as knotgrass, jointgrass, and Fort Thompson grass, is a valuable soil binder along streams subject to erosion in the Tropics and sub-Tropics. It fur- nishes excellent grazing in flat regions near the coast. In the West it is spreading along irrigation ditches and sometimes invades rice fields in California. In Australia it is known as water-couch and siltgrass. It is there regarded as a valuable grass on alluvial flats. A wide-spread aquatic species, Paspalum repens, floating-paspalum or water-paspalum, is sometimes a water weed, making dense growth in drainage canals in the Southern States. It has recently been reported as troublesome in the Panama Canal Zone where it is chok- ing the outlet of Pedro Miguel River and affording a breeding place for mosquitoes. This species is, however, greedily eaten by cattle wherever they have access to it, the animals wading far into the water to get it. It would seem that cattle might be utilized to hold this grass in check or to exterminate it if desirable. Paspalum conjugaium a common widely creeping species of the Tropics and sub-Tropics, not grazed by cattle where other forage is available, makes a fairly good grass for golf grounds in southern Louisiana. According to W. L. McAtee, of the Biological Survey, the seeds of paspalum are eaten by a large number of birds, wild ducks and bob- whites especially feeding on them extensively. HISTORY AND LIMITATION OF THE GENUS. The genus Paspalum was first known from American plants. There are but two known pre-Linnaean figures of Paspalum, both cited by Linnaeus under Panicum diseectum, the first species of 6 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM Paspalum described. Neither belongs to this species. The first, Plu- kenet's "Gramen paniceum, spica simplici, ad caulem intervallata, binis granorum ordinibus uno verftu constante Americanum," plate 350, figure 2,7 is unidentifiable, but may be Paspalum boscianum. The figure seems to be meant to illqptrate the American species cited by Linnaeus and also "Gramen paniceum distachyophoron s. spica gamella, binis granorum ordinibus uno versu constante Ind Orient. Ponnevaragupille Malabarorum,'' coupled with it by Plukenet. The latter phrase name probably refers to Paspalum scrobiculatum and the figure may have been intended for that, though Linnaeus did not cite it when later he described P. scrobiculatum. The second pre-Linnaean figure, Sloane's "Gramen dactylon majus, panicula longa, spicis plu- rimisnudis eras sis/' plate 69, figure 2,8 is Paspalum virgatum, and is later cited by Linnaeus under that species. In the Spfccies Plantarum (1753) Linnaeus included the one species of Paspalum known to him in Panicum, describing the genus Pas- palum in 1759, with four species, all still retained in the genus. The type spd6ies,* the earliest described, belongs to a relatively small group, in which the rachis is foliaceous-winged. Subsequent authors expanded the genus to include species of other groups. Lamarck10 included species of Syniherisma (Digitaria), Axonopus, EriocMoa, and Cynodon (Capriola), and was followed by Poire t.11 Fliigge12 includes all these except Cynodon. Kunth, Trinius, and Nees, in their various works, and most subsequent authors retain species of Axonopus in Paspalum, but place Syniherisma in Panicum, mostly as section Digitaria, or recognize it as a genus under the name Digitaria Scop, or Syntherisma Walt. Hooker13 and a few other modern authors retain Syniherisma in Paspalum. EriocMoa H. B. K. and Cynodon Pers. (or Capriola Adans.) have been generally recognized as distinct genera. Bentham and Hooker" and Hackel15 include Cabrera and Anasirophus as sections of Paspalum. Chase18 differen- tiates Axonopus Beau v., with A. compressus (Swartz) Beau v. as the type, on the reversed position of the subaessile, solitary spikelet, with Ana&trophus Schlecht. as a synonym and Cabrera Lag. (including A, aureus) as a section. 7 Plukenet, Mant. 94. pi. 350. /. 2. 1700. 8 Sloane, Oat. Fl. Jam. 34. 1696; Voy. Jam. I: 112. pi. 69. f. £. 1707. • See page 7. » Tabl. Encycl. 1: 175-176. 1791. " In Lam. Encycl. 4: 28-35. 1804. u Monogr. Pasp. 1810. »F1. Brit. Ind. 7: 10-20. 1896. w Gen. PI. 3: 1098. 1883. 16 In Engl.