THE JOURNAL OF THE LINNEAN SOCIETY. BOTANY. — VOL. XIM. W LONDON: SOLD AT THE SOCIETY’S APARTMENTS, BURLINGTON HOUSE; AND BY LONGMANS, GREEN, READER, AND DYER, AND WILLIAMS AND NORGATE. 1873. PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET. LIST OF PAPERS. Page BAKER, J. Q., Esq., F.L.S. Revision of the Genera and Species of Scilleæ and Chlorogaleæ 209 BENNETT, ALFRED W., M.A., B.Sc., F.L.S, On the Floral Structure of Impatiens fulva, Nuttall, with especial reference to the Imperfect Self-fertilized Flowers .......... 147 BENTHAM, GEORGE, F.R.S., P.L.S. Notes on the Styles of Australian Proteace@ ................ 58 Notes on the Classification, History, and Geographical Distribu- tion of Compositæ ... <... hs cose «ebro URINE 335 BERKELEY, Rev. M. J., M.A., F.L.S. Australian Fungi, received principally from Baron F. von MUELLER and Dr. R. SCHOMBURGK ................ «5 155 Curney, F., M.A., F.R.S., Sec. L.S. On a new Genus in the Order Mucedines .............. 333, 578 DALZELL, N. A., Esq., A.M. Note on Capparis galeata (Fresen. and C. Murrayi, J. o e Mr c 72 Remarks on Dolichos uniflorus, Lamarek.........oooomoo.o... 145 New Leguminose from Western India .................... 185 Dicxre, G., M.D., F.L.S, On the Marine Algte of the Island of St. Helena ............ 178 Dyer, W. T. TuiseLTON, B.A., B.Sc., F.L.S, On the Determination of three imperfectly known Species of Indiau Lernstræomiatem o sack vee kr beouevo eee ues 328 Hansury, DANIEL, Esq., F.R.S., F.L.S. Historical Notes on the Radix Galange of Pharmacy ........ 20 Note on Amomum angustifolium, Sonnerat ......ooomomoooo.. 154 iv. Page Haxcz, HENRY FLETCHER, Ph.D. «e. On the Source of the Radix Galange minoris of Pharmacolo- KEEN ou did a ee d 1 Supplementary Note on Chinese Silkworm-Oaks ............ 7 Notes on some Plants from Northern China ................ 74 Flore Hongkongensis Supplementum. A compendious Sup- plement to Mr. BENTHAw's. Description of the Plants of the Iand OF Hongkong -i ce ay Bee ree ee ae 95 LrE1IGHTON, Rev. W. A., B.A., F.L.S., F.B.S.Ed. On two new Species of the Genus Mycoporum, Flotow. ...... 326 LiwbBERG, S. O., M.D. Bryological Notes 0 rc cna a E EC Ke eva we ss 66 Remarks on Mesotus, Mitten... Te- ree RS 182 On Zopas, H. f Oe E CH 188 Masters, MaAxwELL T., M.D., F.R.S., F.L.S. &c. Note on the Genus Byrsanthus (Guill.) and its Floral Conforma- OE uper a ee iU cu O 15 On the Development of the Andrcecium in Cochliostema, Lem. . 204 MATEER, Rey. S., F.L.S. Remarks on the Tamil Popular Names of Plants ............ 25 Mitten, W., A.L.S, New Species of Musci collected in Ceylon by Dr. Thwaites.... 293 Munro, Major-General, C.B. Extract of a letter to Mr. BENTHAM, dated May 29, 1872 .... 331 REEKS, H., Esq., F.L.S. On the varieties of Aspidium angulare and aculeatum ........ 65 TuzasxE, L. R. and C. New Notes upon the Tremellineous Fungi and their Analogues, 31 WEALE, J. P. MansEL, B.A. Oxon. Notes on a Species of Disperis found on the Kagaberg, South eh a quos os se ee Re 42 Some Observations on the Fertilization of Disa macrantha .... 45 Notes on some Species of Habenaria found in South Africa. . 47 Observations on the Mode in which certain Species of A podes ars fertilized oo ec a 48 mr ET E CR OR: WM A 9e QE OE WOO aa Cree A a E | RP OR € I" 9b Ce «Wh THE JOURNAL OF THE LINNEAN SOCIETY. On the Source of the Radix Galange minoris of Pharmacologists. By Henry FrETOHER Hance, Ph.D. &e. {Read December 1, 1870.] Wuttst it is, I believe, fully established that the “ Greater Ga- langal ”” is produced by Alpinia Galanga, L., the plant which yields the lesser kind has hitherto remained altogether doubtful, though some writers have hazarded the opinion that it is the rhizome of 4. chinensis, Rosc. It is now more than twelve years since my attention was first drawn to the subject by my esteemed correspondent Mr. Daniel Hanbury, who begged me, if possible, to set the question at rest. I have never lost sight of Mr. Hanbury’s wishes; but al- though the drug forms a considerable article of export from Southern China*, my want of success will not seem surprising * Galangal is not used in British medical practice; and, even on the Conti- nent, Endlicher speaks of it as “exoleti fere usus.” The following statement of the export of this drug during the last three years is compiled from the offi- cial returns published by the Foreign Inspectorate of Maritime Customs, the quantities and value being, however, for greater convenience, reduced to Bri- tish weight and currency. From Canton. From Shanghae. Total. Years. Quantity. | Value. | Quantity. Value. Quantity. Value. Ite. £ & 4| te |£-«a d| e | 8 1807 ... 32,800 |123 10 10 | 79,200 | 354 9 9/112,000| 478 0 7 1868 ...| 15,233 | 57 10 0 | 162,308 [1149 3 5,177,041 [1206 13 5 1869 ..| None. | ie. | 370,800 ee 16 9) 370,800 |3046 16 9 | LINN. JOURN.—BOTANY, VOL. XIII, B 2 DR. H. F. HANCE ON THE SOURCE OF THE when it is borne in mind that many vegetable products shipped from Canton come from distant parts of the empire, and pass through a number of hands before they reach those of the native merchants, and that these latter are quite incapable of compre- hending the interest attaching to the solution of a doubtful scien- tific point, or of troubling themselves about what seem to them matters of aimless and puerile curiosity. Those who have tried know well how difficult it is to get reliable information from the natives, who will frequently invent answers, rather than seem ignorant, and are especially prone to reply in the affirmative to direct or leading questions, as if they supposed the object of an inquirer was rather to obtain the confirmation of his own views than to elicit the truth. In November 1867 I had the opportunity of making a visit— at the invitation of, and in company with, the Commissioner of Maritime Customs at Canton—to the Island of Haenan. During this excursion, and while at anchor off Pak-sha, a fishing-village on the south coast of Kwangtung, about seventeen miles from, and rather to the east of Hoi-haú, on the north coast of Haenan, we landed, and some of the party went about six miles inland to a ruinous walled city named Hoi-on; but, being slightly indis- posed, I preferred botanizing over the low hills near the coast. On their return, Mr. Sampson, who was one of the party, in- formed me that they had seen a large quantity of what he took for ginger (but which he described as bearing the inflorescence on the leafy stems) under cultivation; and another gentleman produced—asking if I knew what it was—some pieces of rhizome, of which quantities had been passed, exposed to the sun in shal- low bamboo baskets to dry. This I immediately identified as Galangal; and as some inquiries made of a linguist who had accompanied them left no doubt that the rhizome belonged to the plant seen growing, I had the mortification of knowing that the true Galangal plant had been met with, and no specimens obtained, whilst our arrangements did not admit of further delay. Fortunately, however, at the close of the year, another expe- From this table it would appear that the demand for Galangal is increasing ; but I cannot explain why the export of a product of the extreme south of China should be transferred from Canton (the nearest port) to Shanghae, si- tuated 8° further north. RADIX GALANGZE MINORIS OF PHARMACOLOGISTS. 3 dition to Haenan was planned; and on this occasion Mr. E. C. Taintor,an American gentleman in the service of the Imperial Customs, to whom I was indebted for the specimens of the Oaks on whieh the North Chinese wild silkworm is fed, respecting which I have already communicated a paper to the Society, ac- companied it. Mr. Sampson took great pains to indicate to Mr. Taintor the locality where the plant had been seen; and I am happy to say that Mr. Taintor's researches were crowned with complete success, he having brought back fine living plants with the rhizomes attached, an examination of which, and com- parison with authentic specimens of the drug from Mr. Hanbury and others, procured here, leave no doubt whatever of the species being the true officinal one. The following account from Mr. Taintor’s notes will explain how he obtained the plant. “The locality is about one mile north of the small village of Tung-sai, situated upon the Bay of Pak-shá, at the southern extremity of the peninsula of Lui-chau- fi, or Lei-chau-fü, and directly opposite Hoi-haú, the port of Kiung-chau-fü in Haenan. The plant was growing at an ele- vation of about 100 feet above the level of the sea, in a very dry hard red soil, evidently composed of disintegrated volcanic rock. The plant grew in masses, which had been originally planted and cultivated, but were now apparently neglected and running to waste. The roots were in dense masses of sometimes more than 1 foot diameter, and with as many as twenty-five or thirty stalks springing from each. Rarely more than one or two of these stalks, however, bore flowers at the date of collection, January 5th. My plan, to insure that I was getting the real plant, was to write the two characters Liang-kiang, R (mild or gentle ginger, the Chinese name), and tell an intelligent-look- ing villager that I wanted to see the flower. He led me, with- out the least hesitation, directly to the spot where I obtained the plants." F I must add that Mr. Swinhoe has since found the plant grow- ing wild in dense jungle on the south coast of Haenan, one of his specimens being now before me, and that he has informed Mr. Hanbury, as I quite recently learnt trom that gentleman, that there is good reason for believing that its fruit is the Bitter- seeded Cardamom, figured in Mr. Hanbury’s valuable paper * “ On some rare kinds of Cardamom.”’ * Pharm. Journ. xiv. 418, fig. 8. B2 4 DR. H. F. HANCE ON THE SOURCE OF THE In endeavouring to determine the specimens collected by Mr. Taintor, I found in my herbarium, for the purpose of compari- son, only the Hongkong species of Alpinia, and a few Moluccan ones, received from M. Teijsmann, of the Buitenzorg Garden; whilst, as regards books, I was restricted to Roxburgh's ‘ Flora Indiea, the writings of Wight and Miquel, and the very useful * Prodromus Monographie Scitaminearum’ of Prof. Horaninow, published at St. Petersburg in 1862. With these somewhat slender adminicula, Y was soon satisfied that the Galangal was either referable or else very closely allied to 4. calcarata, Rose. (which Roxburgh states to have been introduced from China into the Calcutta garden); and though I found some discrepancies be- tween the Kwangtung specimens and the description of A. cal- carata drawn up from the living plant by Roxburgh *, whose accuracy is so well known, yet these were apparently so few and unimportant that my chief ground of hesitation as to their identity was the extreme improbability that the rbizome of a plant widely cultivated within the tropics, and growing and flowering luxuriantly in the Calcutta and also, according to Thwaites t, in the Peradenia garden, should have remained for so long a period unrecognized, if really the same as the Lesser Galangal of commerce. It being evident that this question, of so much interest in itself, could not be solved with the means at hand, whilst an ap- proximate judgment would be valueless, I determined to let the matter lie over until I had access to more complete materials. Since then I have received, through the kindness of Mr. Han- bury, a sketch, with a single flower coloured, of the plate of A. calcarata given in Roscoe’s 'Scitaminem, and a full-coloured copy of that in the second volume of the ‘ Botanical Register ;' whilst my ever liberal friend Dr. Thwaites has sent me living rhizomes of the same species, whence have been reared fine healthy plants, though they have not as yet flowered, and, be- sides, copious specimens both of the flowering plant for the her- barium, and of the dried mature rhizomes. Mr. Taintor's Ga- langal plants have also again blossomed under culture, but set no fruit f; so that fresh flowering specimens of A. calcarata, * Flora Indica, ed. Carey, vol. i. p. 69. + Enum. Pl. Zeyl. p. 320. i Zingiberaceous plants when under cultivation, even in localities where they are native, are far less disposed to fruit than the same species in a wild state, the flowers usually dropping off as soon as they fade. a RADIX GALANGZE MINORIS OF PHARMACOLOGISTS. and fruit of both species being alone wanting, I may claim to have had at my disposal as good materials for comparison as ordinarily fall to the lot of a descriptive botanist. I have, to the best of my ability, made a careful and exact comparative exami- nation of living flowerless plants of each kind (including the rhi- zome), and of the mature rhizome of each; whilst I have com- pared the fresh and also the dried flowering plant of the Ga- langal with separate dried flowers, as well as herbarium specimens of the entire inflorescence, of A. calcarata. The result is, that I am now entirely satisfied that the plant which furnishes the Lesser Galangal root is, though very closely allied to Alpinia calcarata, Roscoe, a perfectly distinct and well-defined species, the two differing in several particulars of structure, as well as in sensible qualities, as the following brief comparative notes will show :— Alpinia calcarata. Galangal. Dried mature rhizomes chestnut- brown*, conspicuously furrowed lon- gitudinally; when cut across, with a stronger odour than Galangal, the cut surface remaining of a fus- cous hue ; of a bitter aromatic taste, much like cardamoms, with a di- stinct flavour of rhubarb superadded, but destitute of heat. Sheaths and bases of the young living stems or shoots more or less tinged with pink; tasting somewhat like rhu- barb, but without any hot flavour. Leaves of a full deep green; aro- matic, but not hot in taste. Li- gule 3-6 lines long, rounded or Dried mature rhizomes externally rufous-brown, only very finely stri- ated longitudinally; when cut across, surface becoming rufous ; aromatic and very warm in taste, as if made up of ginger and pepper, with a recognizable camphoraceous flavour, leaving a powerful sensation of heat in the mouth when chewed T. Bases of young shoots white; tasting very warm. Leaves of a rather lighter green; hot in taste. Ligule 9-15 lines long, aeutish. Racemes quite simple. Flowers without a bract- let. Labellum without the slightest trace of yellow, its veins very fine. * Described by Roxburgh as somewhat woolly and pale-coloured. Dr. Thwaites and myself find them perfectly smooth, both when young and at full growth. The young fresh rhizomes of both plants are quite white and succu- lent; but these can scarcely be alluded to: again, some dried rhizomes kindly supplied from the Calcutta garden are cinnamon-coloured ; but these are of small diameter, and evidently immature. The full-grown ones from Ceylon are, as described, of a chestnut hue externally, t Cesalpinus characterizes the rhizome very accurately, though briefly, as “ subrufa intus et extra, sapore Piperis, modice odorata” (De Plant. lib. iv. c. 62). 6 DR. H. F. HANCE ON THE RADIX GALANGZE MINORIS. truncate, and frequently bifid at apex. Racemes compound *. Flow- ers with an oblong concave bractlet at their base. Labellum “yel- lowish, minutely punctated with dull red, and with veins of a deep dull red colour? (Thw.)l, its veins thickish. The fruits of both species, when known, may afford other marks of distinction. A description of the Lesser Galangal plant, for which I pro- pose the name of Alpinia officinarum, drawn up very carefully from living specimens, may fitly bring these notes to a close. ALPINIA OFFICINARUM, n. sp. Rhizomatibus longe repentibus atque intertextis cylindraceis 6-9 lineas circiter diametro rufo-brunneis gla- berrimis squamis magnis pallidioribus fibrosis demum secedentibus annulosque irregulares sinuosos albidos relinquentibus copiose in- struetis, caulibus 23-33-pedalibus, foliis bifariis longe vaginantibus coriaceis glaberrimis nitidis anguste lanceolatis basi angustatis sed non petiolatis exquisite attenuatis 9-14 poll. longis medio 10-12 lin. latis ligula magna (9-15 lin. longa) oblonga scariosa erecta basi de- currente vaginas marginante apice acutiuscula auctis, racemo termi- nali simplici erecto densifloro brevi (plerumque haud 4-pollicari) foliis superato, rachi tenuiter tomentella, bracteis $ spathaceis invo- * So described by Roxburgh, and so I find them in all Dr. Thwaites's spe- cimens; but represented as simple in Wight's plate (Ic. Pl. Ind. Or. vi. 2028), and also apparently by Roscoe, and in the * Botanical Register.’ t Deseribed by Roxburgh as * solitary, boat-shaped, white, 1-flowered," and shown in the Bot. Reg. plate, and also (so far as I can make out from the sketch) in that in Roscoe, but omitted in Wight’s figure. Quite conspicuous in all Dr. Thwaites's specimens. t Roxburgh describes the labellum as “deeply coloured with dark purple veins on a yellow ground." The Bot. Reg. plate represents it as crimson in the centre, with a broad yellow border, into which veins from the centre run, though not very conspicuously ; whilst my copy of Roscoe’s figure gives an oblong yel- low centre dotted with crimson, and a broader margin striated with red and yellow, the latter colour slightly predominating. Considering the variation in colour of the flowers of Canna, and the differences of shade and marking in the labella of many cultivated epiphytes of the allied order Orchidacee, it is perhaps unsafe to attach any considerable weight to a character of this kind. § Though these exist equally in A. calcarata, it is curious that Roxburgh makes no allusion to them ; he would have called the two an involucre. There is likewise no indication of them in the figures of the * Botanical Register, Roscoe, or Wight. DR. H. F. HANCE ON CHINESE SILKWORM-OAKS. 7 lucrantibus binis exteriore viridi nunc folio abbreviato coronata in- teriore alba ambabus demum extus stramineo-arefactis nitidis intra margineque scariosis cucullatis flore pluries longioribus vel simul apicibus invicem convolutis basique solutis calyptratim secedentibus vel interiore paulo serius decidua, floribus ebracteolatis arcte sub- sessilibus 15 lin. longis, perigonio exteriore albo tubuloso tomentello apice breviter 2—3-lobo lobis scariosis rotundatis ciliatis, perigonii interioris albi tubo extus intusque tomentello lobis oblongis obtusis cucullatis 8-11 lin. longis 2-23 lin. latis tertio paululum majore et latiore, labello albo medio striis vinoso-rubris juxta apicem in macu- lam distinctam flabellatim dilatatis percurso aliisque pallidioribus a lineis medianis interioribus marginem versus pinnatim radiantibus elegantissime picto sessili ovato integro apice acutiusculo vel bilobo crispulo-eroso 10 lin. longo 8-9 lin. lato basi corniculis binis rigi- dulo-carnosis subulatis subreflexis 1-13-linealibus pilis capitatis con- sitis basique glanduloso-incrassatis conniventibus tubum occludenti- bus aucto, stamine labello dimidio breviore, ovario densissime albo- tomentoso, stylo apice sensim dilatato paulo ultra antheram producto, stigmate concavo margine ciliato, glandulis epigynis j-linealibus luteolis oblongis apice truncatis integris vel lobulatis. Habitat in interioribus insulze Haenan ; vix dubie etiam in silvis austra- liorum imperii Sinensis provinciarum, ubi commercii ergo large co- litur (Exsicc. n. 16866). British Vice-Consulate, Whampoa, September 1870. Supplementary Note on sie Silkworm-Oaks. By Henry F. Hance, Ph.D., &c. Kai onptxod vnparos o$aíivew Aóyovs rj QiMa Cidwory y KAwWOw róxn. Puike, In Verm. Seric. 1. [Read December 1, 1870.] In a paper read before the Linnean Society in May 1868 *, I showed that the tree on which feeds a larva spinning cocoons from which large quantities of silk are manufactured in the north of China is Quercus mongolica, Fisch. ; and I gave reasons for the opinion that Q. dentata, Thunb., is also used to feed the same worm. When writing this paper, I had accidentally over- looked a valuable article by Dr. D. B. M‘Cartee * On some Wild Silkworms of China," published in the ‘Journal of the North * Journ. Linn. Soc., Bot. x. 482. 8 DR. H. F. HANCE ON CHINESE SILKWORM-OAKS. China Branch of the Royal Asiatic Society’ *, from which the following passage is an extract:— During my late sojourn at the Port of Chefoo, I learned from conversations with Chinese there, that three species of silk are manufactured in Shantung from the cocoons of wild silkworms; and the account given by them of the worms, &c., corresponded with the accounts given by the Jesuit Missionaries of similar operations in the province of Szchuen. The cocoons from which one of the kinds of silk is derived are called Ta-kien (large cocoons), or Tso-kien (oak cocoons). A species of oak was pointed out to me at Chefoo as the oak upon which the silkworm was reared, and is called by the natives Tsoshu, or sometimes Po-lo-shu t; but my informant was unable io give me the Chinese characters for the latter designation. The tree corresponded with the description given by M. d'In- carville, Quercus orientalis Castanec folio, glande recondita in capsula crassa et squamosa i. It is said that only young trees, not more than three or four years old, can be used." The only oaks in Northern China agreeing at all with Tourne- fort's phrase, quoted by D’Incarville, are Q. serrata, Thunb., and Q. chinensis, Bge. ; and my friend Mr. Mayers, now H.M. Acting Consul at Chefoo, who has been so kind as to make inquiries for me on the spot, says he can only find that one species is met with in that neighbourhood; and the specimens of this he has communicated belong to the first-named tree. I have lately had an opportunity of seeing a Chinese work, entitled * Chih Wuh Ming Shih T'u K'ao," or * Nomenclature and Description of Plants, illustrated with Plates, the figures to which are very superior.to those in the majority of Chinese * New Ser. iii. 75 (1866). Dr. M‘Cartee identifies the moth as Saturnia Mylatta, t Regarding this name, Mr. Mayer writes as follows in a letter to me, ac- companying specimens of Q. serrata, Thunb. “The Chinese call it Po-/o, the same as at Newchwang; and, after some hesitation, characters have been given me for the sound which are apparently mere phonetic devices to represent a foreign or at least unwritten name. I should not wonder if the word Po-lo were Corean or Manchu, as the Chinese admit that it is not the correct name for the oak, which should be called Siang." 1 Père d'Incarville quotes this as the designation given by “nos botanistes,” referring, in fact, to Tournefort, and supposing that the Chinese oak is identical with some Asia-Minor species, most likely Q. va//onea, Kty. The fruit of this is exceedingly like those of Q. chinensis and Q. serrata, to which last Miquel reduces the Georgian Q. castaneifolia, C. A. M., which both C. Koch and Grise- bach consider identical with Q. vallonca. DR. H. F. HANCE ON CHINESE SILEWORM-OAKS. 9 books, and, though searcely more than outline sketches, in some instances very characteristic. The fifth chapter of this work is devoted to a description of the 7s’ing-kang tree, for the following translation of which I am indebted to Mr. Mayers. I would remark that all Chinese notices of this kind contain, mixed with trustworthy statements, inaccuracies and puerilities, but in this respect are not worse than the Natural History of Pliny *, and immeasurably superior to J. B. Porta’s * Phytognomonica. Chi- nese scientific literature is, and has been for ages, in that stage which the late Dr. Whewell characterized as the “commentatorial.” “The work called * Kiu Hwang Pén Ts’ao’ (or ‘ Description of Plants Available for use as Food,’ published during the Ming dynasty) did not specify the locality where the 75'ng-kang tree grows ; but at present it is found everywhere. The larger trees, producing acorns, are called Stang-li (the generic names for the oak); whilst those that are smaller, and do not produce acorns, are called Zs'zng-kang. In its stem and foliage the Z5'ng-kang is altogether similar to the Siang-li; but the colour of its leaves is a lighter green, and its flowers less abundant. Their flavour is also bitter, but their properties are mild and devoid of noxious effect. The young leaves are gathered, and after being dried, are steeped in water until they become of a yellow colour. The water is then changed; and after being strained they are eaten with salt and oil ” t. * I may here note, as bearing on the subject of this paper, that Pliny has a curious passage about a moth in the Island of Cos, the worm of which feeds, amongst other plants, on oak-leaves, and from whose cocoons silk was woven. Though fabulous and inaccurate in many particulars, as his statements so fre- quently are, owing to their being compiled without any critical spirit from an immense variety of sources, the allusions to the domestic rearing of this worm, to the softening of the cocoons in water, and their subsequent reeling, and to the lightness of the tissue as adapting it for summer wear, all clearly point to a wild silkworm. “Fieri autem primo papiliones parvos, nudosque: mox frigorunr impatientia villis inhorrescere [a very Lamarckian idea!], et adversum hie- mem tunicas sibi instaurare densas, pedum asperitate radentes foliorum lanu- ginem vellere: hanc ab his cogi unguium carminatione, mox trahi inter ramos tenuari ceu pectine. Postea apprehensam corpori involvi nido volubili. Tum ab homine tolli, fictilibusque vasis tepore et furfurum esca nutriri: atque ita subnasci sui generis plumas, quibus vestitos ad alia pensa dimitti. Qus vero cepta sint lanificia, humore lentescere mox in fila tenuari junceo fuso. Nec puduit has vestes usurpare etiam viros, levitatem propter zstivam " (Hist. Nat. xi. 23). f Although oak-leaves are still used in Italy as fodder for cattle, as was 10 DR. H. F. HANCE ON CHINESE SILKWORM-OAKS. * Additional Note.—The Ts'ing-kang tree grows on hills inter- spersed with the Tsiang-li oak trees, being, in fact, of the same kind, but devoid of flowers and fruit. A green ball is frequently found developed at the extremity of its twigs, consisting of hairs as fine as the silky fibres of the Tsung tree (qu. a palm ?), but somewhat tougher. The native silk products of Kweichow pro- vince are woven from the silkworm cocoons of this tree; but they are of mediocre value, on the same principle that mulberry trees producing fruit, and Siang trees producing chestnuts, are all adapted to the growth of silkworm." The plate annexed to the above description, a copy of which is here given, represents an oak with leaves like those of a shallow-lobed form of Q robur, and with three fruits (unless they are intended for the *'oak-apple" mentioned in the text), one distinctly stalked, the dense squam:e of the cupule entirely con- cealing the acorn, and looking like those of Q. dentata, Thunb., though closely appressed, instead of being more or less re- flexed. Baron Léon d'Hervey-Saint-Denys, in his ‘Recherches sur P Agriculture et 1 Horticulture des Chinois’ *, published in 1850, and alluded to in Dr. M‘Cartee’s article, has quoted from the * Annales Forestiéres’ a note, by the Pére Julien Bertrand, on the management of the wild silkworms in the province of Qwei- chow, here referred to, one of the two oaks employed being, in all probability, that figured; and as the subject is of great in- terest, and the book, I believe, not very well known, I have thought it well to translate this missionary's observations in full. * Thong-kin-foo, July 19, 1842. “T think I told you, some years since, that we have here a wild silkworm which feeds on oak-leaves, and which seems to be an object of much interest to the French Government. I imagine you will be glad to have some account of this, and I only regret that my ignorance of natural history prevents the case in the time of the Romans (Colum. De re Rust. vi. 3), they can scarcely be supposed to afford, even under the hands of a Celestial cordon-bleu, a spe- cially luxurious or appetizing salad. The Flora cibaria of the Chinese, however, is so alarmingly comprehensive that it is more difficult to say what is noć than what ¿s eaten. * Page 162. DR. H. F. HANCE ON CHINESE SILKWORM-OAKS. 13 my treating a subject of such importance in a more worthy manner. “These worms are met with in the more mountainous districts v! Kweichow, and also in some parts of the province of Szchuen, 12 DR. H. F. HANCE ON CHINESE SILKWORM-OAKS. as, for example, Ki-kiang, San-tehouen, and Pa-hien. Although they may be reared advantageously in various localities, their favourite loeation is in Kweichow, on the highest mountains, where the air is cooler and purer than elsewhere. As the mulberry-silkworm succeeds best in warm countries, you will doubtless be surprised, as was M. Hébert, one of the French delegates in China, to hear that these worms succeed better on the mountains than in the plains, where the climate is milder. Such is the case, however, as proved by long native experience, and by the fact that in the mountains there are two crops of silk annually, whilst in the low grounds the worm yields only one, and that much inferior to the first one from the higher regions— showing that the oak-silkworm requires a cold rather than a warm temperature. *'The management of these worms is altogether different from that of the ordinary silkworm. They are reared on the trees, not in houses. As soon as they are hatched they are taken to the hills and placed on the trees. If it were attempted to raise them at home, by supplying them with oak-leaves, in the same way as mulberry-leaves are given to the common silkworm, they would die without tasting the leaves, which they require to eat on the tree, and to pick out for themselves. The oaks on which they feed require no particular care; they are in their natural state. I may here say a few words regarding these trees. There are two of them in China, one called Ts'ing-kang, the other Fu-li: these two species differ but little; and in order to distinguish them it is necessary to examine them very closely. The only difference consists in the leaves, and the hardness of the wood. The Ts'ng-kang is harder than the Fu-li; its leaves are long and toothed, and somewhat like those of the chestnut. The Fu-li has shorter and broader leaves; so far as I can judge, it is iden- tical with our French oak, at least that of Le Velay; for I have never examined the oaks of other provinces. Although the worms eat both kinds, they prefer the Zs’ing-kang to the Fu-li. The oaks are never allowed to grow old here; every eight or nine years they are cut down to the ground ; the subterranean trunks throw up new shoots, which are again cut down after the lapse of another eight or nine years, so that the oak woods are merely copses. All the mountains hereabouts are covered with these trees. “Ten or eleven days after the oak-moths have laid their eggs, DR. H. F. HANCE ON CHINESE SILKWORM-OAK5. 13 thousands of little black larvee are seen moving about the basket, which are immediately carried to the hills, and deposited on the trees, the leaves of which (for it is only the end of March or the beginning of April) are only half expanded. Here they are left night and day, in rain or wind. There is no need to watch them by night; but some one keeps at hand during the day to frighten away the birds, and to help the worms to pass from one tree to another, and pick up such as may have been blown down or have fallen to the ground. * The worms change colour four times: at first they are black, later violet, some afterwards yellow, and finally a blackish violet. It takes forty or fifty days to attain this last stage; and they are then as thick as a man’s little finger. They show a special in- stinct in protecting themselves against the weather; when it rains they move to the under surface of a leaf, and when cold winds prevail they shelter themselves on the least-exposed side of the leaf. Towards the end of March 1840 I was staying amongst a Christian community where these worms were reared in great numbers; on the 28th the newly hatched worms were on the trees, on the 30th snow fell, and the three succeeding days were so piercingly cold that within doors it was impossible to leave the fire. I said to my Christians, * Your silkworms will surely all die with this weather.’ ‘Oh! no, they answered, *they are a little numbed by the cold, naturally, but they won't die.” Nor did they; for, happening on the 3rd of April to pass by the place where the worms were on the trees, I saw them eating greedily. * After devouring the leaves for forty or fifty days, they begin to spin cocoons, which are rather more than an inch long, as thick as a walnut, and of a somewhat whitish-yellow colour. As some worms are always stronger than the others, so also there are cocoons of a larger size than usual. They are spun on a leaf rolled round into a hollow cone or funnel; and if one is not sufficiently large, two are fastened together. The worms com. mence by weaving the outside of the cocoon, in which they then shut themselves up and spin the remainder, the whole task occupying only three days. The date of collection of these cocoons varies according to locality; thus in the plains and at slight elevations, they are gathered in from the 20th to the 24th May, or a few days later, whilst on the mountains of Qweichow this is not done until the 15th to 20th June. Here 14 DR. H. F. HANCE ON CHINESE SILKWORM-OAKS. also, vegetation being in arrear, the silkworms are hatched later. * In the mountainous districts of Qweichow, and even in some parts of Szchuen, all the grubs are not killed, but a small quan- tity of the cocoons are kept, to raise at once a second crop. At low elevations, however, but one crop is raised, as a second would not repay the time and trouble of rearing, owing to the heats of July and August, which would prove fatal to the greater propor- tion of the worms. * On the higher mountains, where the nights are always cool, and the heat is tempered by the constant breezes, and where insect enemies are rare, the worms grow as vigorously as the first brood. The second crop is gathered about the lst Oc- tober. “The produce of these oak-worms, although inferior to that of the ordinary silkworm, is nevertheless very fine and very strong, and when woven yields a very pretty bright fabric. I be- lieve a great deal might be done with this silk in France, and that our Government is quite right in looking at the acclimatization of the oak-silkworm as a matter of considerable importance." The preceding extracts are suggestive in several respects. Thus I suppose that the somewhat obscure statement of the Chinese Cyclopedia as to the identity, except in size and bearing of fruit, of the Ts'ing-kang and Siang-li, may be elucidated by the information given to and recorded by Dr. M‘Cartee, that only young trees are used for rearing silkworms on, and by Pére Bertrand’s statement that, in Qweichow, young trees, or the new growths sent up from the stumps of felled ones, are alone employed for this purpose. This would show a great unifor- mity in one particular in very distant parts of China; and it is quite possible that, as with animals amongst our stock-breeders and fanciers at home, so here, one species of oak may, at different ages, and according to whether it is available or not for a special purpose, receive two distinct substantive names. Another point worthy of attention is the affinity of the several oaks employed (as far as these have been scientifically determined) to those of Europe and Western Asia. To this I have already adverted in my first paper, mentioned at the commencement of these notes ; and if, as I think likely, the oak here figured is the Fu-li (sup- posed to be identical with Q. robur by Father Bertrand), there ON THE FLORAL CONFORMATION OF THE GENUS BYRSANTHUS. 15 is little doubt of the tolerably close relationship of the two. Having now for several years past paid more than ordinary at- tention to the genus Quercus, I may say (I hope without in- curring the charge of presumption) that I believe the specific limits of the Lepidobalani of the Eastern Hemisphere are most vaguely and unsatisfactorily defined, and am by no means satisfied of the distinctness of several of those of north-eastern Asia. Similar doubts have been expressed by several European botanists; but the total disagreement of writers as to which are good species, even amongst well-known forms, is perhaps the most convincing proof of the uncertain status of many members of the group. I do not, I think, in fact, and certainly not in intention, wrong French advocates of acclimatization, when I assume that they are, as a rule, rather too enthusiastic as to the results of their pet projects; but I see no reason to alter the opinion I expressed formerly (long before I was cognizant of M. Bernard’s excellent notice), that all circumstances would seem to conspire to render the culture of the oak-silkworm in Europe a sure matter of success, if properly set on foot and fostered. On its importance, if successful, there is no need to enlarge. It is curious to know, from independent sources both native and foreign, that the rearing of a silkworm on oaks is carried on in the south-west as well as the north-east of this vast empire ; and it will be a matter of great interest to ascertain whether any oak is employed for the same purpose in northern Burma or Assam, in both of which the genus Quercus is a characteristic type of vegetation, and from which Kweichow is only separated by the province of Yunnan. British Vice-Consulate, Whampoa, Sept. 1870. Note on the Genus Byrsanthus (Gill) and its Floral Confor- mation. By MaxweLL T. MasTzns, M.D., F.R.S., F.L.S. [Read January 19, 1871.] Tmurs genus, which was established by Guillemin *, is closely allied to Homalium, if not indeed identical with it. Endlicher rejected Guillemin’s name, and substituted that of Anetia t, on the ground that a genus of Lobeliacee had received at the hands * Deless. Ic. Select. Plant. iii. 30. t. 52. - T Gen, Pl. p. 923. n. 5088. 16 DR. M. T. MASTERS ON THE FLORAL CONFORMATION of Presl the very similar designation of Byrsanthes. Bentham and Hooker*, however, keep up the original designation of Guillemin, expressly stating that they have no personal ac- quaintance with the genus in question, and copying the charac- ters assigned to it from Endlieher, who, in his turn, probably availed himself of the original description of Guillemin. Guille- min founded the genus on a specimen which he considered iden- tical with a plant briefly alluded to, but not named or fully described, by Robert Brown. This conjecture I believe to be in- correct; and it is with a view of rectifying this error that I ven- ture to lay before the Society the present communication, in which I shall also allude to the structural arrangement of the andreecium. Brown’s original notice of the plant was founded on a speci- men gathered in Congo by Christian Smith, and is couched in the following terms +:—“ In the collection from Congo a plant occurs, evidently allied, and perhaps referable to, Homalium, from which it differs only in the greater number of glands alter- nating with the stamina, whose fasciculi are in consequence de- composed, the inner stamen of each fasciculus being separated from the outer by one of the additional glands, This plant was first found on the banks of the Gambia by Mr. Park, from whose specimens 1 have ascertained that the embryo is enclosed in a fleshy albumen." Here, then, we have the plant first noted by our illustrious compatriot, and also the circumstance that the plant in question is indigenous in the region of the Congo and the Gambia, two widely separated districts. Guillemin's figure and description were taken from a plant collected in Senegal by Heudelot. Act- ing on the conviction that the specimen gathered by the last- mentioned collector belonged to the same species as that alluded to by Brown, Guillemin described it fully, figured it, and gave it the name of Byrsanthus Brownii. The examination of the specimens collected by Chr. Smith, and comparison with the figure and description given by the French botanists, lead me to infer that we have to deal with two distinet species, and, further, that the plant of Guillemin is nob the same as the one mentioned incidentally by Brown. The prin- cipal points of distinction are to be found in the circumstances that * Gen. Pl. i. p. 800. + R. Brown in Tuckey’s Congo, Miscell. Works, ed. Bennett, vol. i. p. 120. OF THE GENUS BYRSANTHUS. 17 in Brown's plant the ovary is wholly inseparate from the flower- tube, while in the species described by Guillemin the ovary is half superior; moreover the petals in the specimen collected by Chr. Smith are patent, while in the description and figure given by Guillemin the petals are connivent *. But the great interest attaching to these plants resides in their curious morphological structure. Owing to the great thickness and opacity of the parts of the flower in the dried plant, it is not altogether an easy matter to determine accurately the rela- tionship of the several parts. An examination of recent spe- cimens in all stages of development is, indeed, necessary for their full comprehension; nevertheless from the evidence before me, and from the analogies offered by allied genera, I believe the structure to be correctly interpreted as follows :— From the summit of the top-shaped flower-tube emerge five (rarely six) ovate leathery valvate sepals. Within, and alter- nating with the sepals, originate the petals, of the same number and form as the sepals, but larger, thicker, and more indu- plicate at the margins, and hence somewhat concave. Within the petals are two or three series of fleshy cushion-like glands intermingled with stamens. The difficulty before alluded to con- sists in ascertaining the precise relative position of the glands and stamens, and in consequence their exact morphological sig- nificance. If my interpretation be correct, the arrangement is as follows :— First row of 5 or 6 sepals; second of 5 or 6 petals; third of glands and stamens thus disposed :—in front of each sepal one gland; in front of each petal one gland, flanked on either side by a fertile stamen: fourth row of glands and stamens, one gland before each sepal, one perfect stamen before each petal ; * The two species may briefly thus be characterized ; a fuller description will appear in the forthcoming (second) volume of the ‘ Flora of Tropical Africa,’ while certain structural details are alluded to in the context :— Byrsantuus Brownsu, Guill. in Ic. Deless. iii. p. 30, t. 52, Foliis ovali-ob- longis; petalis conniventibus; ovario basi tantum cum floris tubo coad- unato; stylis in tubum fusiformem inferne coalitis. Hab. Senegal, Coll. Heudelot. B. grrovsvs, Mast., sp. nov. Foliis oblongis; petalis patentibus ; ovario omnino cum floris tubo coadunato; stylis divergentibus omnino liberis vel imo basi tantum coalitis. Hab, Gambia, Park; Congo, Chr. Smith. LINN. JOURN.—BOTANY, VOL. XIII. C 18 DR. M. T. MASTERS ON THE FLORAL CONFORMATION this latter stamen is therefore separated from the centre of the petal by a gland belonging to the outer series. Fifth row or gy- necium, ovary connate with the flower-tube, or more or less free from from it, one-celled, with 5 parietal placentas supporting numerous ovules. Styles coherent at the base, or free. The arrangement of the stamens and glands may be more clearly made evident by the following diagram, where $ repre- sents the sepals, P the petals, x the glands, st the stamens. S S S S S P P P P P x st x st x st X st x st X st x st X st x st X st x st x st x st x st x st Guillemin describes the position of the stamens and glands almost precisely in the same way :—“ Stamina duplici serie dis- posita numero triplici petalorum, duobus exterioribus ad latera cujusque petali, glandula interjecta; uno interiore ante glandu- lam petalo oppositam glandula hine et illine comitato." In the diagram given in Delessert’s ‘Icones, the artist com- mitted an error, to which Guillemin thus alludes :—“ In serie interna glandularum et staminum due glandule collaterales inter stamina errore gravi delineate fuerunt dein unica glandula reipsa exstet inter stamina." It is the more necessary to call attention to this correction, as Lindley, in the * Vegetable King- dom,’ p. 742, has repeated the diagram without making the ne- cessary alteration. Adverting now to the arrangement of the androecium in some nearly allied genera, we find the most simple condition in Black wellia (Homalium § Blackwellia, Benth.). Here there is a single stamen in front of each petal, and no glands at all. Visa (re ferred to Homalium by Bentham and Hooker) is described having a similar structure, but with a gland between each pair of stamens. In Casearia, which is not quite so closely allied to Byrsanthus as the preceding, we have a single series of fertile stamens alternating with petaloid staminodes. In Bivinia the stamens are grouped in fascicles between the sepals, and there are five glands opposite to the sepals. The petals are absent in this genus. In Dissomeria and in Homalium $ Racoubea, the stamens are numerous, and arranged in fascicles before the petals. In My riantheia (referred by Bentham and Hooker to Homalium) 9 OF THE GENUS BYRSANTHUS. 19 similar arrangement is described. On either side of the fertile stamens of Lunania are two glands, or probably the glands are bilobed, as in Osmelia. As my examination of the genera just alluded to has been but superficial (except in the case of Homa- lium and Casearia), I would not express myself too confidently as to the exact disposition of the andrecium and the glands in those particular genera. It suffices for my present purpose to show that the general tendency is to have one fertile stamen or a group.of such in front of each petal, alternating usually with glands. In some cases there is but one row of such organs; while in other instances there are two, and possibly even three. From the circumstances just alluded to, I think I am justified in inferring that the groups of glands and stamens in front of the petals of Byrsanthus are strictly homologous with the pha- langes of stamens in Bivinia &c. ; and I would further propose the following interpretation of the morphology of the andreecium in the genus under consideration. The outer row of the androecium consists of ten stamens— five opposite to the sepals, and existing in the form of glands, and five compound stamens opposite to the petals, and consisting of a central barren portion or gland, and two fertile filaments, one on each side of the gland. The inner row of the andreecium con- sists likewise of ten stamens—five represented by glands placed in front of the sepals, five by fertile stamens superposed to the petals. . An examination of the course of development can alone de- monstrate whether or not this interpretation is correct; in the mean time I may state that my inference is based on the exami- nation both of mature flowers and of young unopened buds. Before passing from the consideration of the andreecium of Byrsanthus, it may be well to allude to the estivation of the stamens, as that is a point on which some stress has been laid of late by Clos and others. The filaments of Byrsanthus epigynus are slender, widely spreading, and straight in the ex- panded flower, but incurved and even spirally twisted in the bud. The anthers are small, subglobose, didymous, 2-celled, bursting outwardly by two longitudinal chinks, and surmounted by a short thickened process continuous with the connective. With reference to the position of the ovary of Byrsanthus, the difference in the case of the two species herein mentioned would be sufficient, in many cases, for the establishment not c2 20 MR. D. HANBURY ON RADIX GALANGE. only of a new genus, but would justify the position of the two plants in two different orders or even subclasses. But such differences in Homalinee are, as remarked by Brown *, of com- paratively little moment. “The cohesion [adhesion] of the ova- rium with the tube of the perianthium, though existing in various degrees in all the genera above enumerated, is probably only a character of secondary importance in Homaline»; for an un- published genus, found by Commerson in Madagascar, which in every other respect agrees with this family, has ovarium su- perum." Brown also pointed out the circumstance that the sta- mens are opposite to the sepals in Passifloree, while in Homalinee they are opposite to the petals. As to the question whether Byrsanthus, asa genus, is suffi- ciently distinct from Homalium, opinions will naturally differ. The main differences consist in the petals and the increased number of glands in the andrecium. Now, if we adopt Ben- tham's views t and combine Blackwellia with Homalium, we have equal reasons for including Byrsanthus also in the same generic group as a well-marked section. Nevertheless, in the absence of ripe fruit and of better materials than have been at my dis- posal, it seems better to retain Guillemin's genus as distinct. Historical Notes on the Radix Galange of Pharmacy. By Dante, Hanpvry, Esq., F.R.S. and F.L.S. [Read January 19, 1871.] Ix discovering and describing the plant which yields the Radis Galange minoris of pharmacy, Dr. Hance has added an interest- ing chapter to the history of a substance which for many cen- turies has been an object of trade between Europe and the East. Galangal does not, indeed, possess properties which can claim for it the rank of an important medicine, being simply 4 pungent aromatie of the nature of ginger; but it has so long held a place in the pharmacope@ias of Europe, and enters into s0 many ancient receipts, that I need hardly apologize for offer- ing to the Linnean Society a few notes on its pharmacological history. Galangal was apparently unknown to the ancient Greeks and * Loc. sup. cit. + Journ. of Linn. Soc., Botany, vol. v. Suppl. p. 87. MR. D. HANBURY ON RADIX GALANGE, 21 Romans; at least no mention of it can be found in the classical authors. Its introduetion into Europe was due to the Arabians, in whose writings it is noticed at a very early period. Thus Ibn Khurdádbah, an Arab geographer who served under the Khalif Mutammid, a.D. 869-885, has left some information respecting China, after which he speaks of the country of Sila, whieh exports ..... musk, aloes [;. e. aloes-wood |, camphor, s porcelain, satin, cinnamon [cassia], and galangal *. The celebrated geographer Edrisi, who wrote a.D. 1154, ob- serves of Aden, that it is the port for Scinde, India, and China, from which last country are brought musk, aloes-wood, pepper, cardamoms, cinnamom, galangal, mace, myrobalans, cam- phor, nutmegs, cloves and cubebs f. The Arabian physicians, from Rhazes and Alkindi in the tenth and eleventh centuries downwards, make frequent reference to galangal as an ingredient of the complicated medicines then in use. Among the later Greeks I cannot find any mention made of this drug prior to Myrepsus, who probably resided as physician at the court of the Greek Emperors at Nicea in the thirteenth cen- tury; though several authors ,declare it is referred to much earlier. It is constantly named by Actuarius, who may have been contemporary with Myrepsus. In a work published some years ago in Paris, entitled ‘ As- sises de Jérusalem, ou Recueil des Ouvrages de Jurisprudence composés pendant le xiii? siécle dans les Royaumes de Jérusalem et de Chypre’ 1, there is a remarkable list of commodities liable to to duty during the twelfth century at the port of Acon in Syria (the modern Akka), at that period a great emporium of Medi- terranean trade, in which many Indian spices and drugs, in- cluding galangal, are enumerated. We find galangal also noticed, together with ginger and zedoary, as productions of India imported into Palestine, by Jaques de Vitri, Bishop of Acon in the early part of the thirteenth century $; and * “Le Livre des Routes et des Provinces, par Ibn Khordadbeh, traduit et annoté par C. Barbier de Meynard," Journ. Asiatique, sér. vi. tome v. (1865), p. 294. t ‘ Géographie d'Edrisi, traduite par A. Jaubert, Paris, 1836-40, 4to, tome i. 51. t Paris, 1841—43, fol. tome ii. chap. 142. $ Vitriaco (Jac. de), * Historia Orientalis et Occidentalis, 1597, 8vo, | p. 172. p- 22 MR. D. HANBURY ON RADIX GALANGE. in the * Romance of Godefroi de Bouillon, a poem written in the twelfth century, it is named as one of the rarities of the East, which the Crusaders were deluded into believing would be found in plenty in the Holy Land *. Marco Polo, in his travels in Asia in the thirteenth century, observed galangal to be produced in Southern China (Province of Foochow ?), as well as in Java t. About this period it was also known in Western Europe. St. Hildegard, Abbess of Bingen, who died in A.D. 1179, names it as galgan, and comments upon its medicinal virtues t. Galangal is catalogued with other spices (as ginger, cinnamon, cloves, and nutmegs) in the tariff of duties levied in the port of Colibre (Collioure), in Roussillon, in a.p. 1252 $. A more interesting notice of the drug is contained in the jour- nal of expenses of John, King of France, from July 1, 1359, to July 8, 1360, during his residence in England, preserved in the ‘Comptes de 1'Argenterie des Rois de France.’ Besides purchases of sugar, mace, ginger, cloves, pepper, cardamoms, calamus aroma- tieus, and many other drugs, we find three entries for galangal, namely, for 4 lb. 184., for 2 lbs. 6s., and for 1 Ib. 22d.|| As the price of gold happens to be also mentioned in one part of the ac- count, it is easy to form an estimate of the relative value of galangal. This shows the price of 3s. per pound to be equivalent to about 10s. of our present money—not extravagant for a commo- dity transported from the remotest Asia to the centre of England. In Professor J. E. Thorold Rogers's *History of Agriculture and Priees in England, there are eleven entries indicating the price of galangal in England between A.D. 1264 and 1376. The highest was in 1307, when 2 lbs. of the spice purchased for the Crown * * Bibliothèque de l'Ecole des Chartes, tome ii, (1840-41), p. 437. + ‘Le Livre de Marco Polo’ (éd. Pauthier: Paris, 1865), pp, 522, 561. t ‘S. Hildegardis Abbatisse Opera omnia, accurante J. P. Migne, Paris 1855, p. 1134. $ Capmany, * Memorias Historicas sobre la Marina, Comercio y Artes de la Ciudad de Barcelona,' 1779, tomo ii. p. 20. || The original entries are as follows :— “ Lundy VII* jour d'octobre. Jehan Kelleshulle, espicier à St. Boutoul, pour espices prises de li pour le Roy....Galingal, demie livre 18d. Jeudy XIIe jour de février....Galingal, 2 livres, 6s. Samedy XXVII* jour de juing.--- Berthélemi Mine, espicier.. . . Galingal, une livre, 22d... . ." L. Douet D'Arcq, ‘Comptes de l'Argenterie des Rois de France au XIV* siècle. Paris, 1851, 8vo. pp. 218, 232, 265, 266, MR. D. HANBURY ON RADIX GALANG.E. 23 were paid for at the rate of 6s. 8d. The other entries indicate the price as from 1s. 6d. to 3s. per lb. In the fifteenth century galangal was evidently in common use ; for Saladinus, physician to one of the Princes of Tarentum, circa A.D. 1442-1458, reckons it among the things necessaria et usitata which should be found in the shop of every aromatarius *. As might be expected, it is included in all the older pharmacopceias and antidotaria. Garcia D'Orta, first physician to the Portuguese Viceroy of India at Goa, and a resident in India for thirty years, is, I think, the first writer to point out (1563) that there are two sorts of galangal—the one, as he says, of smaller size and more potent virtues brought from China, the other a thicker and less aromatic rhizome produced in Java f. This distinction is perfectly correct. The Greater Galangal, which is termed Radix galange majoris, is yielded by Alpinia Ga- langa, Willd., a plant of Java t; the lesser, called Radix galange minoris or simply Radix galange, is derived, as we now know, from the plant which Dr. Hance has described as A. officinarum. It is the latter drug alone that is at present found in European commerce §. The name galangal, galanga or garingal, Galgant in German, is derived from the Arabic khalanján ; whether that word may be a corruption of the Chinese name liang-kiang, signifying mild ginger, I must leave it to others to decide. Let me say a few words regarding the uses of galangal. Asa medicine, the manifold virtues formerly ascribed to it must be ig- nored ; the drug is an aromatic stimulant, and might take the place of ginger, as indeed it does in some countries. That it is still in use in Europe is evident from the exports from China and from the considerable parcels offered in the publie drug sales of London ||. * ‘Compendium Aromatariorum,’ Bonon. 1488, fol. t *Colloquios dos Simples e drogas he cousas medicinais da India,' Goa, 1563, Colloquio 24. i Maranta Galanga, Linn., Sp. Pl. and Swartz, Obs. Bot. $ Moodeen Sheriff, in his learned ‘Supplement to the Pharmacopoeia of India’ (Madras, 1869), states that in the bazars of Hyderabad and in some other parts of India the rhizome of Alpinia calcarata, Rosc., is sold as a sort of galangal ; and that a species of Alpinia growing in gardens about Madras, which, conceiv- ing it to be new to science, he has described and named as 4. Khulinjan, has a rhizome much resembling the Lesser Galangal of China. l| Three hundred bags, each 112 lbs., imported from Whampoa were offered 24 MR. D. HANBURY ON RADIX GALANGJE. The chief consumption, however, is not in England, but in Russia *. It is there used for a variety of purposes, as for flavouring the liqueur called aastoita. The drug is also employed by brewers, and to impart a pungent flavour to vinegar, a use noticed by Pomet t so long ago as 1694. As a popular medicine and spice, it is much sold in Livonia, Esthonia, and in Central Russia; and by the Tartars it is taken with tea. It is also in requisition in Russia as a cattle-medicine; and all over Europe there is à small consumption of it in regular medicine. There is doubtless some quantity of galangal of both sorts used in India. By a ‘Report on the External Commerce of the Presi- dency of Bombay for the year 1865-66’ I find that there was imported into the port of Bombay of *Gallingall" from China 520 cwt., from Penang, Singapore, the Straits of Malacca, and Siam 70 ewt., and from ports in Malabar 834 ewt. Of the total quantity (1424 cwt.), 716 cwt. was reshipped to the Arabian and Persian Gulfs. According to Rondot, writing in 1848, the trade in this drug is on the decline +; and the statistics which I have examined tend strongly to show that this is the fact. The foregoing notes may be thus summarized :— l. Galangal was noticed by the Arab geographer Ibn Khur- dádbah in the ninth eentury as a production of the region which exports musk, camphor, and aloes-wood. 2. It was used by the Arabians and later Greek physicians, and was known in northern Europe in the twelfth century. 3. It was imported during the thirteenth century with other eastern spices by way of Aden, the Red Sea, and Egypt, to Akka, in Syria, whence it was carried to other ports of the Mediter- ranean. 4. Two forms of the drug were notieed by Garcia d'Orta in for sale by Messrs. Lewis and Peat, 27 Oct., 1870. The quantity was not thought remarkable; and I am assured that a single buyer will sometimes purchase such a lot at one time for shipment to the continent. * Professor Regel, of St. Petersburg, and A. v. Bunge, of Dorpat, and Mr. Justus Eck, of London, have all obligingly supplied me with information as to the use of galangal in Russia. My thanks are also due to my friend Professor Flückiger, who on this, as on other occasions, has kindly offered me valuable suggestions. + ‘Histoire des Drogues, Paris, 1694, fol., part 1, p. 64. 1 ‘Commerce d'Exportation de la Chine,’ Paris, 1848, p. 98. REV. 8. MATEER ON THE TAMIL POPULAR NAMES OF PLANTS. 25 1568 ; these are still found in commerce and are derived respec- tively from Alpinia Galanga, Willd., and A. officinarum, Hance. 5. Galangal is still used throughout Europe, but is consumed most largely in Russia. It is also used in India, and is shipped to ports in the Persian Gulf and Red Sea. Remarks on the Tamil Popular Names of Plants. By The Rev. S. Matter, F.L.S. [Read March 2, 1871.] Havine had the opportunity during a residence of some years in Travancore of becoming familiar with the popular names of plants in two of the vernacular languages, Tamil and Malayalim, I have drawn up for my own use somewhat complete lists of such names. In examining these vocabularies I have often noticed with interest traces of some of the mental peculiarities and notions of the Hindus, and obtained many little items of curious information. In offering a few remarks on this subject, I shall refer almost ex- clusively to the Tamil language, which is much more complete and highly cultivated than the Malayalim. Considerable difficulty is experienced by Europeans in the at- tempt to ascertain from the natives the correct vernacular names of plants. Many of the people pretend to be acquainted with the indigenous plants, but are utterly incapable of recognizing or identifying the different species with any thing like certainty. Repeated questioning, careful cross-examination, and persevering inquiries in different parts of the Tamil country are necessary, in order to obtain a reasonable degree of assurance on many points connected with this subject. And it may be observed in passing that even in English works the transliteration into our own cha- racters of the vernacular designations of plants is often executed with most discreditable carelessness and want of method or uni- formity. , - Great inaccuracy and a general absence of scientific system ob- tains in the Tamil botanical nomenclature. Some attempt, how- ever, at a rough natural system of classification may be discerned in the native mode of naming plants which might, perhaps, to some little extent be availed of in the composition of botanical works in Tamil. Certain general terms are applied to classes of plants bearing some apparent resemblance or external similarity 26 REV.S. MATEER ON THE TAMIL POPULAR NAMES OF PLANTS. to each other; and the particular species is defined by an adjective or another noun preceding. “Tutti,” for instance, appears to signify mallows in general. Then there are ottututti = adhering or sticking mallows = Urena sinuata, so called from its having fruits which cling to clothing or other surfaces with which they come into contact. Kattututti= jungle- or wild tutti, is Sida hirta, Nila tutti = ground-tutti, is Sida cordifolia, and Perun tutti = great tutti, is Sida asiatica, and so on. A similar use is made of such general terms as tali = Convol- vulus, kalli = Euphorbia, atti = Ficus, malligei = jasmin, tulasi = mint or sweet basil, korei = Cyperus or sedge, mulli = thorn, pasi = mosses or aquatic plants, and others. These are qualified and defined as in many of our own popular names of plants by such nouns or adjectives as river, water, white, black, mountain, jungle or wild, dog, elephant, sea or maritime, foreign, &c. For example, attumulli = river-thorn is Dilivaria ilicifolia, a pretty flowering acanthaceous plant, with leaves strikingly resembling those of the holly, which grows by the river-sides. Neer nochi = water nochi, is the small tree Vitex trifolia; pey karumbu = devil sugar-cane, is Saccharum arundinaceum, a wild and useless sugar- cane. It is well known that the South-Indian or Dravidian languages are not Sanskritic in character or origin, though some of them have adopted a considerable proportion of Sanskrit words. Of the common names of plants, some are, as might be expected, Sanskrit; but the great majority are Dravidian. Many wide- spread and useful plants, such as the Plantain, Mango, Tamarind, and others, have names derived from both sources. In seeking, however, to turn these popular names to account for a vernacular systematic nomenclature of plants according to the European natural system of classification, we find serious diffi- culties in the way. In the first place, great deficiencies in the list of general terms are at once apparent. There is no word, for example, applicable to the class or order of Palms, only proper names of each species; nor are there terms for Ferns and other important orders of plants. Again, as might be expected, no really proper and distinctive names exist in these languages for plants which have been merely introduced, though now naturalized; Alamanda, for instance, 1$ merely known as “ yellow flower ;” the Cashew, the Feringhee (or REY.S. MATEER ON THE TAMIL POPULAR NAMES OF PLANTS. 27 Frank) or the Quilon Mango; Poinciana pulcherrima is Peacock Cassia; the Bixa or Arnotta shrub, monkey-yellow ; the Bapaw is the Feringhee or the Curry Castor-oil ; the Casuarina, wind-trem- bler or whip-tree, names suggested by the waving of the branches and the strange melancholy whisper of the wind through these trees, and by the use of their tough pliant branches as whips or rods. The potato is denominated “round root," and liquorice “very sweet root.” Lantana indica also is simply “ hedge- plant,” being generally used for fences. Not only are there many deficiencies, but serious and absurd errors occur in this rude mode of natural classification. * Arasu,” for example (which appears to mean “ the royal tree ””), is the correct designation of the noble and sacred Ficus religiosa ; but Thespesia populnea, a malvaceous tree, is called “ flowering arasu," Puvarasu, which has come to be written in English Portia. And even Trewia nudiflora, a Euphorbiaceous tree, is dignified with the name of “attu puvarasu,” the river flowering arasu. Here we have the same term applied to plants of three distinct and strikingly dissimilar orders. Again, *varei" is the common name of the various species of Musa (plantains or bananas) ; but Canna indica,.or Indian shot, certainly not unlike a plantain in habit, if not in size, is “ pū- varei,” the flowering plantain ; and even Commelina benghalensis, a mere straggling herbaceous plant, is called kanavarei, forest- plantain. * Tarei" (from “tar” to be low, to recline) denotes the Pandanus or screw-pine; but “ kattarei,” rock- or stone-tarei, is applied to Aloé indica. * Tamarei " is the Lotus, Water-Lily ; but the Euphorbiaceous tree, Macaranga indica, is actually named ^ vattatamarei," round or orbicular tamarei, possibly fro m the slight resemblance which its large peltate leaves bear to the orbicular leaves of the Velumbium. Then the fragrant labiate Orthosiphon bracteatum is called “ rock- tamarei,” and Pistia stratiotes ** agasa tamarei,” or sky-lotus, per- haps from some comparison of the water in which it floats while growing to the sky above. All this is just what we might expect from a people so inexact and fanciful as the Tamilians, seizing the most trifling and acci- dental points of resemblance, and bestowing names upon plants almost at haphazard, without caring to investigate their essential and distinctive characters. The popular names of British plants, 28 REV. S. MATEER ON THE TAMIL POPULAR NAMES OF PLANTS. however, are often no better in this respect, as, for example, in calling Epilobium “the flowering Willow.” It is interesting to note the view taken by the Tamil people of many of the indigenous plants, as seen in the names given to those plants. A few of these may be mentioned. The Cocoa-nut is called “tenkay,” honey-nut, showing the high estimation in which this valuable fruit is held. The Plantain is * vàrei," from the root var, to flourish, prosper, from its remarkable exuberance of growth and fertility. Tts stalks and fruit are.therefore used on occasions of rejoicing as an emblem of prosperity and happiness. The Tamarind is simply called * puli," the “acid” or “sour” tree. * Tutti,” used for Malvaceous plants, is from “tu,” to eat, imply- ing that some of these are edible, and intimating that it was early observed that plants of this order are wholly destitute of noxious qualities. Cassia alata, “the ringworm-shrub," used to cure cutaneous eruptions, is called vandukolli = beetle- or bee-killer ; and Michelia champaca is likewise known as “ vanduna malar," the flower-tree untouched by beetles or bees. Perhaps the strong scent of the flower does repel bees, but I have not examined whether this be the fact. Ipomea tridentata is called “ horse-killer.”’ Asclepias prolifera, the juice of which is used as an antidote to cutaneous disease, is named “ nach aruppàn "— the poison-killer or antidote to poison. Names are bestowed upon plants, just as in our own language, from the resemblance which they are supposed to bear to some common objects. For example, the Fern Hemionitis cordata, which we compare in the trivial name to a heart, they call * dog's ear." The pretty little golden Aster plant, Inula indica, is mukkotti = a gold ring set with a jewel, or a nose-jewel. The long cylindrical Gourd which we compare to a snake, Tri- chosanthes anguina, they liken to the proboscis of an elephant (aneikodan surei). Utricularia stellaris is the “egg pási,” and Zannichellia indica the “ hair pàsi." Other names are given from some obvious characteristic or sen- sible quality of the plant. ; E Y REV. S. MATEER ON THE TAMIL POPULAR NAMES OF PLANTS. 29 Demia extensa, a twining Asclepiad with white cottony hairs around the seeds and furnishing a fine fibre from the stem, is named “ hedge-cotton." Mussenda frondosa is “ white leaf," from its curious and beautiful white calycine leaf. Crotalaria laburnifolia is * kilukiluppei" the “ rattler,” from the ripe dry seeds rattling in the legume, the very 1dea of our own botanical name from the Greek “ krotalon ” a rattle. “Adutinna palei,” the * milk-plant which the goat will not eat," is Aristolochia bracteata, every portion of which is intensely bitter. A few plants take their popular names from legendary tales or personages, like our St. John's Wort. Thus Spinifex squarrosa, a sea-side grass with spherical bristly inflorescence, is called * Ravana’s whiskers.” It may be interesting also to trace botanical terms belonging to the Tamil language now included in our own scientific nomencla- ture. Many of these Indian words are very oddly Latinized, and some of them quite misapplied. The following are some of the Tamil words thus adopted by us. 1. The Malabar nut, an Acanthaceous shrub, is Adhatoda vasica, in Tamil adatodei. Todei is used for various species of Citrus ; ada means immoveable or not shaking. 2. A genus of Xanthoxylacew is called Ailanthus. This word does not occur in Tamil; but in the cognate language, Malayalim, aranthal is the name of Ailanthus excelsus. 3. The Pineapple is Ananassa. This is said to be derived from *nanas," the name in Peru, where the plant is indigenous. In Tamil, however, it is commonly called ‘‘annasi” or * annatarei,”’ the food or edible tarei or Pandanus. 4. A genus of lofty and graceful palms is called Aréca, which I should be inclined to pronounce 4A'reca. In Tamil, “kay ” means fruit, and “adeikáy " pickled fruits. In Malayalim, however, * adekka ” is the appropriate name of the Betel-nut Palm, adopted in our botanical works. 5. The beautiful Papilionaceous genus Agati is simply in Tamil * agatti.” 6. Mangifera indica is so named from * mà," the Tamil name of the tree, which, compounded with * kay," becomes * mankay,”’ cor- rupted by Europeans into “ mango." 7. Odina in Anacardiacez is * uthiyam." Pavetta in Cinchones is * Pávattei," which I should therefore pronounce Pávetta. ” -80 REV.8. MATEER ON THE TAMIL POPULAR NAMES OF PLANTS. 8. Pongamia is simply * pungu ” or “punga,” which, however, is said by Dr. Winslow to be applied to Dalbe: ‘gia arborea. | : 9. Zingiber and Ginger, I believe, come from “ inji,” the green - ginger root. 10. Moringa pterygosperma, the horse-radish tree, is in Tamil l * murungei," from * murungu ” to perish, dissolve, the wood being - extraordinarily soft and perishable. In the trivial names of plants also many Tamil names make : their appearance. 1. Of this character I believe is the term galanga or galangal, | applied to a species of Alpinia. This appears to be simply the | South-Indian word * kirangu" or “kilangu,” root, plural “ kir- . angugal." 2. Cassia Tora is from “ tagarei," the native name of the plant. ij Cerbera Odallam is a poisonous Apocynaceous tree, named in Ma- - layálim * uthalam." Euphorbia Tirukalli is from “ tirugukalli," twisted spurge, but appears to be misapplied to an unarmed species. The native name applies to Euphorbia tortilis. 3. Chavica Betle is the betel-leaf plant. The word betel is ont sometimes applied by English writers to the Areca-nut, which is | chewed along with the leaf, but it properly denotes the latter — only. The word in the original is “ vettilei,’ from verum bare OF — plain, and lez, leaf, which, according to the laws of Tamil euphony, | becomes vettilei, and is corrupted by Europeans into * betel." * Kiriattu" is, I believe, the native name of Andrographis panici- — lata; while Agathotes Chirayta is commonly spoken of as the | Chiretta herb. I think, however, I have observed that the natives | call the former “ country kiriattu." “ Brinjal ” is said in Prior's * Popular Names of British Plants’ to be a Tamil word; but this is certainly an error. I know not i from what language the word really comes. It may be worth noting, in connexion with this subject, that . many of the names of places in South India are derived from the | names of plants which abound in the respective localities. Such | | are Anjengo or Anjutenga=the five Cocoa-nut trees, and Calli- mere point or Kalli medu = Euphorbia-hill. ON THE TREMELLINEOUS FUNGI AND THEIR ANALOGUES. 31 New Notes upon the Tremellineous Fungiand their Analogues. By L. R. and C. TunásneE. I. WuEN a special organization or a singular anatomical structure is common to a very great number of different beings, itis evident that the slightest modification of this organization or structure deserves the observer's attention; and this is particularly true if the modification affects an important apparatus, such as the re- produetion in the basidiophorous fungi. This is the reason why, in this great family of plants, the little group of the Tre- melle and analogous species excites a peculiar interest. The Fungi Tremellinei, indeed, are not distinguished from other Basidiomycetes merely by their mucous consistency; for this occurs also in certain Hydna, in Merulius, &c.; they more- over present in the construction of their hymeniwm some pe- culiar characters which make it easy to recognize them. Their basidia, or hymenial and sporophorous cells, are shaped, as we have shown formerly *, after two distinct types. Some (for in- stance, those of the Dacryomycetes and Guepinia Peziza Tul.) are at first narrowly claviform, then they extend in two thick and divaricated arms or processes, each of which bears one reniform and divided spore. By this forked appearance, the basidia of the Daeryomycetes are easily distinguished from the similarly dispo- rous, but obtuse, basidia of certain Hypochni and other non- mucilaginous Hymenomycetes. The second type of basidia occurs amongst the genuine Tre- melle, where these sporophorous cells are subglobose, or quite spherical, and usually divided from the top to the bottom, into four equal parts. These segments become divergent from each other or remain united; but all grow, in the same way, into the form of long threads or tubes, which reach the periphery or super- ficies of the fungus, and there produce reniform and generally undivided spores. If, notwithstanding M. Fuckel’s contrary opinion, we admit, with the ancient mycologists, that an undoubted analcgy unites the Auricularie with the Tremelle, then we must mention a third kind of basidia, very distinct from the former ones; we mean the sporophorous tubes that M. De Bary has seen and described in Auricularia sambucina Mart. (Hirneola Auricula Jude Berk.), consisting of upright and thick threads, each one * See the * Ann. des Sc. Nat., 3rd ser., t. xix. (1853), pp. 193-231, pls. x.-xiii. 32 MM. L.R. AND C. TULASNE ON THE TREMELLINEOUS FUNGI being divided by transverse walls into four cells or stumps which throw out above long and fertile spieules. (See De Bary's ‘Morph. u. Physiol. der Pilze’ &c., p. 116, fig. 47,a-c.) It is a condition common to these three types that the basidium and its appendices are filled with a granular and not very trans- parent plasma, which is wholly employed in forming the spores, so that when the latter have become perfect and mature the organs that have produced ‘them are quite empty and transpa- rent. It is moreover observed that the spicules are generally far thicker than is mostly the case amongst other Hymeno- mycetes. IL An hymenium resembling that of the Dacryomycetes is found also in the Calocere, which, on account o? their linear shape, have for a long time been mixed up with the Clavarie ; but the real or legitimate affinities of the former are now understood and interpreted by the most able mycologists, such as MM. de Bary and Fuckel, in the same manner as that already adopted by our- selves. (See ‘Ann. des Sc. Nat., 3rd ser., t. xix.) In the Zremella helvelloides DC., a beautiful species, the struc- ture of which has hitherto remained unknown, we have met with basidia that are forked almost like those of the Dacryomycetes, but nearly as globose as those of genuine Tremelle. This fungus, which M. Fries has justly removed from the Tremelle to his group of the Guepiniz, differs nevertheless from several of the latter in being fertile on the under surface. In Guepinia Peziza Tul. -the hymenium is only spread on the superior and cup-like face of the plant, exactly as is usual amongst the Pezize. We give here a short description of Guepinia helvelloides Fr. GUEPINIA HELVELLOIDES Fr. El. Fung. parte alt. p. 81.—Tremella helvelloides DC. Fl. Fr. t. ii. p. 93.—Tremella rufa Jacq. Miscell. Austr. t. i.—Gyrocephalus juratensis Pers. in Actis Soc. Linn. Par. ad ann. 1824, p. 77.— Fungus gregarius, terrestris, gelatinoso-carno- . sus, tenax, totus carnei s. purpurascentis coloris, cujus pileus tenuius- culus, primo quasi spathuliformis, adultus autem semiorbicularis V. obovatus, diametro 1-2-pollicaris, maxime repandus aut saltem con- chatus, et circumcirca attenuato-recurvus, insistit in stipite laterali omnino sibi consubstantiali, compresso, ssepius canaliculato, verticali aut obliquo, altitudine vario, szpius vero circiter pollicari basique velutino ; superna pilei istius pagina tandem ob papillas lineares dense erectas velutina, albida et quasi farinosa, vulgo plane sterilis nec nisi rarissime parcissimeque sparsim sporophora deprehenditur ; adverse e contrario pagina, que scilicet deorsum spectat, tota glabra et fertilis; AND THEIR ANALOGUES. 33 interdum venis paucis prominulis, Meruliorum more, instructa; ba- sidia obovato-globosa tandemque bipartita, sterigmata duo linearia, longa, divaricata exserunt quibus singulis spora ovata v.'brevissime oblonga, sepe quadantenus curvula, utrinque tandem obtusissima, brevissima, 0°01 millim. longa et 0:004-0:007 crassa, debito tempore insistit. Crescit sero autumno in pinetis et fagetis montanis seepissimeque series describit lineares. Frequentem vidimus in sylvis Carthusianis prope Gratianopolim Delphinensium, exeunte Septembri a D. 1857; longe rariorem eontra a. 1859, mense eodem. Sporz forma et crassitudine variant; pulveris instar albidi posticam s.in- feram pileifaciem humidz,cumulatz velant, quam cultelli ope si raderis, hanc farinam non zgre colliges. Spore recentes, basi acutiusculz, quid- quam sterigmatis aliquando retinent, ocellumque pallidum sub medio tegmine monstrant; endochroma seu plasma contentum in guttam oleosam crassam mediamque tandem pro maxima parte vertitur. Spore ex alterutro apice, rarius e latere, germen filiforme pro- trudunt. Fungus truncis, quisquiliis foliisve putridis vulgo sedet, ejusque my- eelio involvi quandoque suspicati sumus lapillos illos seu calcareas concretiones quas trunci isti corrupti suis in penetralibus fovebant ; exterum hi lapilli albidi, licet nucis avellanz et quod excedit crassitu- dine, oblongi autem et varie compressi, ejusdem omnino nature vide- bantur atque nuclei longe minores quos in Tremella v. Nematelia quadam mauritaniea olim videramus. (Cfr. Ann. Sc. Nat. ser. 3, t. xix.) III. Since our first paper upon the Tremellini, we have found in Tremella frondosa Bull., and T. albida Huds., an hymenial or fertile apparatus quite similar, so far as regards the tetrasporous basidia, to that of Tremella mesenterica Retz. The fine groups of Tremellg frondosa which we gathered in January 1863, on the dried stump of an oak in the forest of Meudon, near Paris, did not measure less than 15-20 centim. in diameter; they were of a very pale flesh-colour, inclining to yellowish; the membrane of the fungus is thin, corrugated, or like crumpled stuff, and dis- solves into aqueous mucilage. Each basidium is formed of three or, more frequently, four globose cells which. at last become almost detached from one another and terminate in flexuous spicules 0:03—05 millim. long; the spores are shortly ovoid, and in ger- mination become very nearly spherical. In Tremella albida Huds., which we met with in December 1861 on the bark of a Sycamore (Acer Pseudoplatanus), the ele- ments or parts of the basidia remained united, and its spores were crescent-shaped and blunted at each end. LINN. JOURN.—BOTANY, VOL, XIII. D 34 MM. L. R. AND C. TULASNE ON THE TREMELLINEOUS FUNGI After having attentively examined many ZTremelle, we cannot help suspecting that M. Fuckel is mistaken in imagining that Tremella foliacea Pers. is a conidiophorous form of Bulgaria in- quinans Fr.* Such an opinion, indeed, would lead one to affirm that this so-called Zremella foliacea is not a genuine Tremella at all, if M. Fuckel’s plant is the same as that described by Persoon ; but, on the other hand, we can hardly suppose that the epithet foliacea would have been applied to the genuine conidiophorous Bulgaria inquinans, such as we have seen and described it in the ‘Annales des Sciences.’ We will here mention a peculiar Zremella which grows as a parasite on Spheria strumella Fr., and which, most probably on account of this, has acquired its strange name, although nobody, as far as we know, has ever remarked the habitual presence of this parasite on the old pulvinuli of the Spheria above mentioned. We do not mean to say that parasitic Tremelle have never been mentioned ; but several of those which have been so called seem rather of an ambiguous nature. No such doubt can exist about our Zremella neglecta; for by its organization it resembles com- pletely T. frondosa and T. albida. However, we must confess that we have never as yet seen its globose or quadripartite basidia producing sporophorous spicules. This gap will certainly be filled up some day, and there is therefore no reason why we should not give the following description of the new fungus :— TREMELLA NEGLECTAT, pulvinulis natura de more gelatinosis, glo- bosis, perexiguis, sordide albidis, strato mucoso crasso superne vestitis; basidiis (supra dicto strato retectis) breviter pedicellatis, globosis aut nonnihil pyriformibus, primum integris et obtusissimis, postea autem ad basim usque quadripartitis, divisuris tandem plus minus divaricatis, quadantenus sursum acutatis et verisimiliter in sterigma sporophorum singulatim productis. Parasitatur hiemali tempore in cespitibus obsoletis Spherie strumelle ribicole cujus rostris porrectis vulgo asperatur; quum autem am- plius crevit, eadem rostra immersa aut plus minus latitantia arcte fovet. Sexcenties vidimus in ramis et sarmentis emortuis Ribis rubri sylvestris circa Cadvillam Versaliorum, locis sylvarum opacis udisque. Fungillus, nisi adoleverit et jove pluvio favente intumuerit, «egre dignos- citur, szepissime vero latens adest ; semper matricem suam quasi stru- mosam efficere videtur. Hujusce conceptacula solita, non obstante Tremella insita, rite informantur, paucas autem sporas, ni fallimur, pleraque maturare valent. * See Fuckel, Symb. Mycol. p. 286. AND THEIR ANALOGUES. 35 IV. Basidia very analogous to those of the Tremelle have been observed by Mr. Fred. Currey in a peculiar Hydnum, scarcely less gelatinous than the former, and therefore named by Scopoli Hydnum gelatinosum * (see the Proceedings of the Linn. Soc. for 1860). But this observation does not seem to invalidate the close analogy of this fungus with the most genuine Hydna. The case is different with two fungi hitherto ascribed to the group of the T'helephore, and which are found to possess exactly the hymenial structure of the Tremelle. We will first mention Corticium incrustans Pers. (Obs. Mycol. part i. p. 39. n. 82), which modern authors have placed amongst the Thelephore under the names of Thelephora incrustans (Fr. Syst. Myc. t. i p. 448; Elench. part i. p. 214) and Thele- phora sebacea (Fr. Elench. l. c.; Berk. Outl. of Brit. Fung. p. 440, pl.17.fig.6). Itisafungus which grows not unfrequently about Versailles in damp and dark woods at the end of summer; it creeps upon the earth and clings to its surface, adhering to every thing, dead or living; in spreading itself out it clothes stones and stems of herbs and shrubs, here and there even rising without interior supports in little columns or rough clavules, free or partly soldered to one another, thick and glabrous, or ending in bristling tufts, whence it has received the names of Clavaria laciniata Pers. (in herb. Thuill.) and Thelephora clavarioides Thuill. (in suopte herb.). The fertile hymeniwm of this singu- lar vegetable is to be found chiefly where it becomes more deeply yellow or yellowish, and where its surface is more smooth; else- where the filamentous network of which it is composed is alone usually found. The threads of its tissue are very ramified, thin, and seldom provided with partitions ; the hymeniwm is formed from the terminal cells, which proceed from the ultimate and thinnest branches of those threads. The basidia are ovato-globose, as in the Tremelle ; they measure about 0'013 millim. one way and 0:01 millim. the other way; and each one divides, according to its length, into two or three equal parts, which extend into as many fertile threads or spicules. These are but little divergent, and attain a length of 0°015 millim. to 0°03 millim. The spores are slightly reniform and simple, scarcely differing in form and dimensions from those of Tremella mesenterica Retz. ; and they do not much exceed in length the hundredth of a millimetre; * See Fries, Syst. Mycol. t. i. p. 407. D2 86 MM. L. R. AND C. TULASNE ON THE TREMELLINEOUS FUNGI frequently in germinating they produce secondary spores, after the manner of the spores of Tremella violacea Relh. (See the Ann. des Se. Nat. 8rd ser. t. xix. pl. 12. fig. 12.) Another Corticium, from the same author as the former, the Corticium cesium Pers. (Obs. Mycol. part i. p. 15, pl. 3. fig. 6) or Thelephora (Phyllacteria) cesia Fr. (Syst. Myc. t. i. p. 449, n. 2), presents a very analogous structure. It also is found on damp sand in forests; but it is far thinner and more brittle, and its dark colour is ash-blue. It does not, as far as we have ever seen, invade (like Corticium incrustans) the stocks of herbaceous plants and shrubs, or shoot forth from its surface any kind of clavules. Its outside is exactly like that of an applied or resupinate Thelephora. Nevertheless its hymenium is composed of globose basidia, which divide into two, three, or four equal parts, which produce afterwards as many linear sterigmata, four or five times as long as themselves. The spores of the fungus are just like those of Corticium incrustans. The two plants of which we are speaking, and which have been placed amongst the Thelephore, are evidently very far from their natural allies; we propose to unite them to the Tremellineous fungi under the generic name of Sebacina. Here are, then, two new diagnoses for our mycological cata- logues. 1. SEBACINA INCRUSTANS nob.=Corticium incrustans Pers. Obs. Myc. part i. p. 39, n. 82.=Thelephora incrustans ejusd. in sua Syn. Fung. p. 577.=Thelephora sebacea Pers. Myc. Europ.; Fr. El. Fung. parti. p. 214; Letell. Suppl. à Bull. tab. 607; Berk. Outl. of Brit. Fung. p. 440, tab. 17. fig. 6; Fuck. Symb. Mycol. p. 29.= Thelephora clavarioides Thuill. MSS. in suopte herb., nunc e the- sauris Musei Bot. Parisini.=Clavaria laciniata Pers. MSS. in herb. Thuilleriano (minime autem TAelephora laciniata Pers., neque Cla- varia cristata aliorum). Fungus recens natus Atheliam vel Hi- mantiam niveam fingit, tenuis est ambituque fimbriatus s. laciniatus, pedetentim autem incrassatur, longius circumcirca protenditur et in medio sordide luteolus fit; simul etiam in conulos vel clavulas irre- gulares, obtusas et integras alias, laciniatas seu cristatas alteras, sparsim assurgit; nunc solo tenus undique applicatus obrepit, nunc sæpiusque quisquilia, herbas lapillosque simul investit. Basidia fer- tilia in partibus saturatius fucatis tantummodo generantur; hyphis exilibus de more terminalia ovatoque globosa insistunt, 0:013 millim. hine, 0°01 millim. illinc equant, adultaque state sterigmata quatuor; rarius duo tantum, filiformia, crassa, 0:025-0:035 millim. longa €t Wien AND THEIR ANALOGUES. 37 initio subfasciculata ex apice agunt ; simul preterea in partes totidem coadunatas ipsa longitrorsum vulgo dividuntur. Spore ovato-ob- ` longz et nonnihil reniformes plasma granosum et lacunam vel guttam ocelliformem mediam fovent, que, ubi deciderint, in germen cras- sum, loco temporeque faventibus, protrahuntur, unde etiam, si bre- vius constiterit, spora nova s. sporidium enascitur. Haud infrequens provenit, jove pluvio, in udis et opacis sylvarum, secus ambulacra quidem et fossas, sera estate autumnoque, et in agro Cadvillensi Versaliorum jampridem quotannis nobis occurrit. Hyphe quibus tela fungina constat ubique similique modo tenuissime sunt, ramosz et parcissime septiferze; earumdem brachia in strato fertili peculiariter crebra et contorto-intricata reperiuntur; ubi bra- chium quodcunque ibidem in basidium desinit, plerumque surculum seu brachium alterum sub basidio ipso agere satagit. Cl. Fuckelius, loco sup. citato, candide fatetur se sporas T'he- lephore sebacee nondum vidisse; fungus enim sepissime totus sterilis aut parcissime sporophorus invenitur. 2. SEBACINA C#SIA nob. = Corticium cesium Pers. Obs. Myc. parte i. p. 15, tab. iii. fig. 6 (saltem ut videtur). =Thelephora ceesia ejusd. in sua Synopsi Meth. Fung. et Myc. Europ.=Thelephora (Phyllacteria) ceesia Fr. Syst. Myc. t. i. p. 449. Fere tota byssina est et coloris cinereo-cesii, arene inter muscos repens heret, ac pas- sim etiam in pulvinulos obtusos et deformes incrassata prominet; ceterum de basidiorum (que bipartita vel quadripartita item depre- henduntur) forma et crassitudine Sebacinam incrustantem prorsus imitatur; sporz paulo minores et contractiores plerzque videntur. Nascitur in sylvis arenosis et umbrosis Octobrique (anno 1861) Fonte- bellaqueo (loco dicto Butte à Guay) nobis primum obvia est. Utrum fungillus noster admodum ipse sit atque Persoonianus modo citatus, necne, egre decernendum est, licet ejusdem omnino nature videatur. Habitu saltem et structura fertili precedenti plane congener est. Merisma fastidiosum Pers. ad Sebacinas jure spectare, ob so- litum et fastidibilem habitum, libenter suspicamur; quod quidem vivum olim, nee semel, Caville et Modoni agri Versaliensis va- gantes reperimus, minime autem attentis oculis et vitrorum ope scrutatos fuisse recordamur. Quasi grandis Sebacina incrus- tantis, forma, ni nos fallit memoria, omnino videtur. V. Those who have observed the fertile apparatus or hyme- nium of the Auricularie, must have noticed the similarity of the linear multipartite basidia which compose it, to the spori- diophorous promycelium of the Puccinie and Podisomata. These basidia call to mind the sporophorous threads of our Hypochnus 38 MM. L. R. AND C. TULASNE ON THE TREMELLINEOUS FUNGI purpureus (in Ann. des Se. Nat. 5th ser. t. iv. [1865], p. 295), except that these threads are bent in the shape of a crosier, instead of being rectilinear and upright. If we consider the bys- soid nature and effuse or indeterminate form of that Hypochnus, by which characters it comes near to the Sebacine above men- tioned, it will perhaps, appear to stand in the same relation to the Auricularia as these pretended Thelephore to the true Tre- melle. As one analogy often leads to another, we may remark that the fertile crosiers of Hypochnus purpureus would be like those of Ptychogaster albus Cord., and of Pilacre Fr., if their spores were not borne upon such long pedicels. In order to enable the reader to make this comparison, we should have had much pleasure in adding to this text the analy- tical drawings we made, some years ago, of the two genera of Gasteromycetes and of the Hypochnus just mentioned *; but these drawings are now, alas! in the hands of our enemies the Prussian soldiers, and perhaps already destroyed or burnt. VI. The hymenial structure of the Tremelle and their ana- logues is moreover complicated by the habitual presence of a spermatophorous apparatus, the elements of which are sometimes mixed with those of the sporophorous hymenium, at other times, on the contrary, separately congregated on certain spots of the surface of the fungus. This duplicate arrangement may be met with in the same individual, as was observed by us in Tremella mesenterica Retz. 'The spermatia of this species are little cor- puscles, spherical and innumerable. Those of Exidia spiculosa Sommerf., and of Dacryomyces deliquescens Dub., are rather ovoid, and their presence is often not easy to verify. Since our first investigations, we have found a splendid sper- matophorous apparatus in a rosy Tremella which grows on the dead trunks ot Cherry-trees. Here the spermatophorous spots are orbicular, concave, and marginate, so that they resemble the cups of grouped Pezize placed on the inferior lobes of the plant. The spermatia themselves are cylindrical, bent in the shape of a bow, and, united three or four together, they form little groups or capitules on the slightly enlarged tops of the fertile threads. This fungus is sufficiently interesting to deserve a description. * With regard to the same fungi, see what we have formerly written in the Annales des Sc, Nat. 5th sér. t. iv. (1865), pp. 290-296. AND THEIR ANALOGUES. 39 TREMELLA CERASI Schum. Mollis, levis, dilute rosea, pulvinulis dif- fusis, valde repandis et corrugatis, pollices 1-2 crassis, saepeque in imo ambitu sparsim vel congestim mirum in modum foveolatis, fo- veolis autem seu urceolis orbicularibus vel oblongis et inzequalibus, velo fugaci limboque seu margine incrassato ac saturatius fucato do- natis, in disco contra abunde spermatiophoro dilutius purpurascenti- bus; basidiis in omni fungi pariete, preter foveolas spermatiophoras, confertim solitoque ordine instructis, globosis aut nonnihil ovatis, 0:01-0:013 millim. crassis, sterigmata filiformia quatuor de more ex apice enitentibus; sporis oblongis, curvulis 0*01—0:02 millim. longis et 0:005-0:01 millim. crassis ; spermatiis cylindricis, lunulatis, utrin- que obtusis, 0:006-0-01 millim. longitudine zquantibus, nec 0:002 millim. crassioribus, in summis hyphis vix incrassatis congestim in- sistentibus, nec nisi in urceolis supra dictis generatis. Oritur in cortice putrescente Cerasi vulgaris, autumno, nobisque jam- pridem occurrit in agro Lugdunensi, prope pagum dictum Pomey, haud procul a castello S. Symphorien-sur- Coise, Octobri ineunte (A.D. 1855). Fungus e corticez matricis fissuris erumpit et liber expanditur. Modo dicti urceoli, omni attentione digni, discos fertiles Urceolaria seru- pose, e lichenum gente, habitu quodammodo mentiuntur; nunc fungi parieti plano sessiles imponuntur, nunc contra prominent; quocumque autem modo se habeant, margine incrassato definiuntur, initioque velo purpurascente tenui et continuo obducuntur; id veli postea medio dirumpitur ejusque residua in margine, propterea quasi limbato, aliquantulum persistunt. Hyphz spermatiophore, tenues et subdichotome ramosæ, in corymbos densos abire videntur quorum extrema brachia, brevissima ac vix ac ne vix capitata, denso sperma- tiorum fasciculo singula coronantur. Spermatia mire exigua seepius quaternatim coadunantur ; nonnulla tamen, ni fallimur, sparsim soli- taria deprehenduntur que sic dictarum Tuberculariarum conidia de origine semulantur. Spermatia cxterum omnia, sporis longe minora et multo magis incurvata, quum origine, tum forma et exiguitate, ab iisdem sine negotio discriminantur. Germina quelibet ex his sper- matiis exire nunquam vidimus. Spore, e basidiis rite enate, plane genuinz seu Tremellinez sunt, itemque, quum germinaverint, sporas secundarias aliquando enituntur. VII. As we are now recapitulating briefly all the kinds of organs that may be available for the reproduction of Tremel- lineous fungi, we cannot forget that one of them, Dacryomyces deliquescens Dub., sometimes resolves itself into a multitude of gemme or conidia, and that the specimens in which this occurs are at once recognized by an attentive observer from their pecu- liar appearance; and if any one should be inclined to think that 40 MM. L. R. AND C. TULASNE ON THE TREMELLINEOUS FUNGI those singular specimens might belong to a different type, he would give up his conjecture on seeing the conidiophorous nature and the typical or sporophorous structure sometimes united in the same subject (see the Ann. des Sc. Nat. 3rd ser. t. xix. pp. 216- 219, pl. xiii.). If we are not mistaken, we must consider very analogous to conidiophorous Dacryomyces a Tremellineous production that grows on the bark of the dead branches of Salix caprea. It appears also under the shape of little pulvinules, globose or irregular, scarcely as large as the seeds of the sweet pea, and of a red carmine colour. The whole mass of these pulpous cor- puselés is composed of ovoid, smooth, transparent cells soldered to each other by the ends, forming ramified monilia, and recalling to memory the Hormiscia and other well-known ferments. Such cells or conidia separate from one another very easily. We have only once found this fungus, in the damp woods of Chaville, near Versailles, in January 1865 ; we will here give a short technical description of it. DACRYOMYCES PURPUREUS f, pulvinulis exiguis, purpureis, erumpen- tibus, paucis gregariis, imo solitariis, primum compressis et acutis seu eristatis, postea autem deformibus et jove pluvio collabentibus, na- tura pulposis totisque fere e conidii seu cellulis ovatis, levibus, simplicibus ac primum catenatis, catenis vero seu monilibus abunde ramosis. Nascitur, hiberno tempore, ex emortuo fissoque cortice Salicis ca- pree et nobis semel hactenus obvius estin sylva umbrosa Cadville ad Versalias, anno S. 1865, mense Januario imeunte. Nuperrimo tempore eundem fungillum, ni nos omnia fallunt, iterum reperimus, nec quidem infrequentem, in asseribus pineis jamdiu sub dio degentibus, apud Venetos Armorice australis, Novembri mense (1870). VIII. Besides the characters before mentioned, the spores of the Tremellineous fungi have the power of sending forth, in ger- mination, either simple threads or secondary spores (sporidia). These are sometimes solitary, and resemble the mother spores except in being slightly smaller ; sometimes they are much smaller, of a peculiar shape, and very numerous, like those which occur in various Discomycetes, such as Pezize, Bulgarie, Dothidee, &c. (See our Selecta Fung. Carpol. t. 2 & 3.) IX. The Tremellineous fungi being so abundantly provided with reproductive bodies of various kinds (being in fact in this respect AND THEIR ANALOGUES. 41 more richly furnished than all the other families of Basidiomycetes), it is difficult to understand why they have been placed among the Jungi imperfecti—that is, amongst the fungi of which the repro- ductive cycle is ineomplete, or, in other words, of which the most perfect fertile form is still unknown. M. Fuckel, indeed, says that the indisputable connexion between some of the Tremelli- neous fungi and the Ascomycetes leads to a suspicion as to the autonomy of the remainder*. This reasoning would be specious if the matter in question were true Tremellineous plants; but although, for instance, the Coryne sarcoides assumes, under one of its forms, the appearance of a Tremella, and therefore has been, for a long while, considered a legitimate Tremella, it by no means follows that all the Tremellineous fungi are to be assi- milated to this production, which has been so long doubtful. If the organization of the gemmiparous or spermatophorous Coryne sarcoides be compared with that of Tremella mesenterica and its congeners, it will certainly be impossible to deny various similari- ties; but we may justly refuse to admit a real analogy until one or other, at least, of the two following circumstances has been positively observed, viz. the existence in Coryne sarcoides of a system of bisporous or tetrasporous basidia, like that of the Fungi tremellinet, or the presence in Tremella mesenterica, or in any other true Tremella, of ascophorous disks, such as the cups of the Pezize, and especially of Peziza sarcoides. Solong as nothing of this kind has been seen, it is better to consider the Tremelle and their analogues to be complete or perfect fungi, and as well known, at least, as the Agaricineous or any group whatever of Basidiomycetes. X. Before we bring these notes to a close, it will be, perhaps, expedient to mention a fungus thoroughly Tremellineous in its consistency, but which, by its ramified shape and fructification, seems to be completely separated from the legitimate Tremelle. We mean Ceratium hydnoides, an almost ephemeral production which is well known to all mycologists, although hitherto no incon- testable affinities have been found for it. Its small branches, simple or ramified, seem to be destroyed and to disappear at the least touch, and they are all bristling with little monosporous spicules: its organization reminds one almost exactly of Rhopalomyces Berk. ; and as these spicules are probably merely a secondary apparatus of fructification of fungi of higher order, may we not suppose * Seo Fuckel, Symb. Mycol. pp. 4, 5, 10, 402, 403, e? passim. 42 MR. J. P. M. WEALE ON A SOUTH-AFRICAN DISPERIS. that the same is the case with the Ceratium? We have, how- ever, at present, nothing more to put forward by way of ar- gument in support of this hypothesis. The ovoid spores of the Ceratium increase very much in size, and become nearly spherical in germination ; the germs themselves are thick, very obtuse, and send out little branches very rapidly. Kerleano near Auray (Morbihan), December 1870. Notes on a Species of Disperis found on the Kagaberg, South Africa. By J. P. Maxset WEALE, B.A. Oxon. (Commu- nicated by C. Darwin, Esq., F.R. & L.SS.)* [Read November 3, 1870.] Tuts pretty little white and green flower is found in boggy places at the mouths of springs on the Kagaberg, in the months of February and March. As far as I am aware, no descriptions have been as yet published of the mode of fertilization of either the genus Disperis or Corycium; and as their structure offers many peculiarities strikingly divergent from the better-known genera of Orchidez, I shall endeavour to describe the curious contrivances exhibited in this plant. Each spike contains from 1 to 8 blossoms. The back sepal is adnate to the petals, and forms a very inflated dome-like galea having a very distinct keel-like crest. The two petals spread outwardly about halfway from their base, and are marked with green glandular ridges. They then lap forwards and taper up- wards to a point like the opening to a tent. The two lateral sepals have each a short nectary about the centre of the blade, in front of, and at each side of the column. The column forms, with the adnate labellum, a broad fluted pillar surmounted by a cup, whose long lip tapers upwards be- tween the adhering apices of the two petals. On each side of the labellum project the two step-like processes of the rostellum considerably in front of the face of the flower. Having thus described the general aspect of the flower, a more * [This and the three following papers by Mr. Weale were accompanied by drawings, which remain in the keeping of the Society, and which may be con- sulted by any one who may wish to make a special study of the subjects re- ferred to, —Szc. L.S.] MR. J. P. M. WEALE ON A SOUTH-AFRICAN DISPERIS. 43 minute account of its separate parts will be requisite for a com- plete understanding of its structure and the adaptation of its parts. Viewed laterally, the labellum presents in its front part a deep cup, the lip of which curves upwards at almost a right angle to the posterior portion, in such a manner that no insect could possibly reach its interior from the front—a difficulty which is increased by the outspread edges of the middle portion of the petals. Behind, the two sides of the cup rise backwards to a level with the top of the lip, and then spread behind and over the anther in an elongated somewhat oval appendage. On each side of the centre of the labellum two large membranous shield- like expansions, convex anteriorly and concave posteriorly, spread out and completely envelope the pollinia. The two narrow trans- verse divaricate stigmas, close to, immediately behind and on each side of the cup of the labellum, lie on the fleshy base of these ex- pansions, while immediately in front of them project the legs of the step-like rostellar processes. I will describe the relative situation of the anther and ros- tellum, on the supposition that the face of the flower is placed vertically—a position, however, which it does not assume in na- ture. The two steps of the rostellum would then lie in a plane parallel to the face of the flower, and nearly at right angles to the legs. The legs, again, would be placed at right angles to the pollen-masses ; and as the long slender caudicle follows the direc- tion of the leg-like processes, it is evident that in its natural po- sition it is doubly bent, somewhat like a reversed S or a Z. The foot or step of the rostellum is broad in front and tapers behind, somewhat like an isosceles triangle with rounded angles. The membranous edges curve slightly inwards, and resemble a lady's stirrup with very broad toes. Each is curved slightly outwards; and the edges of its mem- brane are also curled in, to retain the caudicle in its place. The viscid disk is an elongated thin membrane of nearly the same shape as the stirrup or step, and viscid on its upper surface; the caudicle arises from its narrow inner edge. I examined this flower for some time before I could make out how it could possibly be fertilized by an insect; and the difficulty was increased by my holding the flower vertically as above de- scribed. In nature the flower is bent down in a nearly hori- zontal position; the blades of the two sepals extend, on the contrary, in nearly opposite planes. 44. MR. J. P. M. WEALE ON A SOUTH-AFRICAN DISPERIS. Let us suppose an insect to visit these flowers, as is proved by the fact of their producing numerous fertile capsules. It would most probably alight on the lateral sepals and suck some nectar from their little nectaries. It would then probably try to reach the cup-like labellum in the interior of the flower, in order to drink the more copious supply of nectar which lies in the cup. If it crawled up towards the base of the sepals, it would pro- bably, unless very small, find no footing on the narrow blade, nor could it enter the galea at this point, as the petals, as before stated, taper away to their junction with the broad fluted column and labellum which fill up the whole of the centre of the galea. The stirrup-like processes of the rostellum, however, stand out; and their broad feet would form a convenient landing-place especially suitable for a small Hymenopter or Dipter. In trying to enter thence the two chambers which open on either side of the labellum, its feet would stick to the viscid disk, and the long elastic caudicle would prevent its forward progress. In trying to rid itself of the incumbrance it would doubless with- draw the pollinium or pollinia; and were it again to attempt an entrance, their projecting faces would strike against the column and prevent ingress. Under these circumstances the insect would probably fly away to another flower, and whilst doing so, the pollinia would assume another position. In about a minute after removal the caudicle bends backwards, so that the pollen- mass lies above and behind the front part of the long viscid disk. Were the insect to alight then on a plant whose pollinia had already been withdrawn, it could easily enter the lateral cham- bers already mentioned, and could thence reach the nectar through the hollowed sides of the cup. In turning round to do this, the pollen-grains would almost certainly become attached to the stigma; or, as they adhere together but slightly, some would be knocked off against the dome, and would almost certainly fall on the stigma. I tried to fertilize this plant artificially by inserting a needle; but it was only by a good deal of awkward twisting that its ferti- lization could thus be effected ; and the utility of the movement of the caudicle was not apparent, as it was almost as easy to do it before the contraction had taken place. At the same time it was evident that no insect of considerable size could enter the side chamber; and I presume the fertilization is OBSERVATIONS ON THE FERTILIZATION OF DISA MACRANTHA. 45 effected by some of the small bees or beetles which frequent flowers. The position of the two viscid disks (which are so prominently situated, and which have their upper surface viscid, instead of the under surface as is usually the case in Orchids) shows plainly their office as steps to the forum of the galeatic chamber, one to each receptacle. This is the only instance that I know of amongst Orchides» in which the adaptation suggests that the tarsi of insects are the agents of fertilization, although in Asclepiads this would appear to be in some instances the normal method. e | Some Observations on the Fertilization of Disa macrantha. By By J. P. Mansit Weate, B.A. Oxon. (Communicated by C. Darwin, Esq., F.R. & L.SS.) [Read November 23, 1870.] TuE diminution in size and simplicity of structure of the la- bellum in some species of the genus Disa would seem to indi- cate its little service as an attractor of insects and as a necessary appendage to the fertilization of the flowers, its office being re- placed by the large and often gaily-coloured posterior sepal. In the adjoining genus Brownleea, this reduction is so con- siderable, that it may be said to be in a merely rudimentary con- dition, and to have altogether lost the important function it usually holds in the order. Compared with some species of Disa, it is comparatively large in D. macrantha ; certainly it is in rela- tive proportions to D. grandiflora. The back spurred sepalis, on the other' hand, proportionally large. The two lateral sepals, the petals, and labellum spread out- wardly from the column, so as to form a salver-like opening to the funnel-shaped posterior sepal. The colour of the blossom varies much, from nearly pure white with a few pale mauve spots on the petals and labellum, to a bright rich purple; sometimes the spots are small and indistinct, sometimes in large blotches, scarlet and almost orange. It emits, especially towards night, an overpowering, heavy perfume, almost too strong to be agreeable. 46 OBSERVATIONS ON THE FERTILIZATION OF DISA MACRANTHA. = a eal In the normal position of the open flower the anther lies back, - behind and above the stigma, nearly at right angles to the stig- matic surface. The caudicles rise upwards and forwards from the anther-case to their junction with the viscid disks. These are placed parallel to each other in the clefts of the lofty turret- like rostellum perpendicularly to, a little behind, and consider- ably above the stigma. The disks are exceedingly viscid, and take some time to harden. On withdrawal the large heavy pollinium hangs down by its own weight, and freely dangles in the air, suspended by the long and flexible caudicle. The plants generally grow in open gullies at the base of the * Kagaberg." I have, however, met with them on the mountain itself. They are generally surrounded by high grasses and her- baceous plants, and seem to prefer moderately sheltered and moist situations. Nothing can be simpler than the fertilization of the flower. The brillant colouring, the heavy perfume, the conspicuous size of the plant and flowers, are sufficient to attraet both by day and night flying insects; and although I have never detected any in the act of fertilizing, nor seen any with the pollinia attached, I feel assured it must be frequently visited. The plants bear abundance of seed—in which respect they differ from Disa cornuta, so far as my observations go, and from Disa grandiflora, according to Mr. Trimen (vide Journ. Linn. Soc. vol. vii. no. 27, p. 144). There is one point, however, to which I would wish to draw attention, viz. £o the frequency of self-fertilization in this species. A very slight jerk, when the flower is fully expanded, suffices to eject the pollinia from their widely open anther-cases, and to bring them into contact with the stigma. This in nature is not un- seldom the case, as I have repeatedly found many flowers thus fertilized. Iam inclined also to think that an insect withdrawing the pollinia, attached as they must be to the lower portion of it$ body, would probably also fertilize the same flower. The anther in this species is supine; in D. cornuta it lies still further back ; but in D. grandiflora it assumes a position slightly more erect. i ' In the bud the anther is nearly quite erect, and the rostellum projects right in front of and over the stigma. Both D. grandiflora and cornuta are comparatively sterile, L4 MR. J. P. M. WEALE ON SOUTH-AFRICAN HABENARLE. 47 although apparently so well adapted for being fertilized by in- sects; yet in this species, where self-fertilization cannot be rare, the flowers produce an abundance of seed. Were the anther erect, as in the bud, or still more supine, as in D. cornuta, self-fertilization would be impossible, and other contrivances for fertilization would be necessary. This is the ease in D. cor- nuta, where the pollinium undergoes an upward movement after removal. I mention these circumstances with the view of seeing them worked out in other species. At present it seems strange that out of three very conspicuous species the most fertile should be one frequently liable to self- impregnation. Mr. Trimen, in the paper referred to, has aptly observed that Disa grandiflora seems to be a correlative case to that of Ophrys muscifera: it is curious that we should find the parallel carried out in Disa macrantha, an instance almost corresponding to Ophrys apifera, in which self-fertilization would appear to be the rule instead of the exception, and whose fertility is considerably greater than that of O. muscifera. Notes on some Species of Habenaria de, in South Africa. Abs- tract of a paper by J. P. Mansexn Weitz, B.A. Oxon. (Com- municated by Charles Darwin, Esq., F.R. & L.SS.) [Read November 3, 1870.] In a species of Habenaria found in December 1869 and January 1870 on my farm “ Brooklyn,” nine miles from King William’s Town, the contraction of the caudicle takes place principally at the end attached to the viscid disk, which is seen to be very much thicker than the portion attached to the pollinium when removed from the rostellum. So great is the tension when zz situ that it is sur- prising the pollinia are not often dragged from the anther, or the disk from the rostellum. The fertilization of the plant is simple in the extreme, as any inseet settling on the bridge must almost certainly deposit one or both of the pollen-masses on the stigmata. The plant does not appear to be visited by diurnal insects, but must be very attractive to nocturnal ones, as, although each spike bears many flowers, and the plant itself grows in considerable 48 MR. J. P. M. WEALE ON THE FERTILIZATION OF abundance on the open grassy flats, almost every flower that I have examined has been fertilized, In another species of Habenaria found on the Kagaberg, in February 1869, and on my farm * Brooklyn” in February and March 1869, the whole caudicle, when ¿n situ, is relatively much shorter than in any of the preceding species, and does not con- tract on withdrawal, but is nearly rigid. The viscid disk is seen to be oval on its outside, with a slight extension laterally. The caudicle at its juncture with the disk is somewhat trian- gular, the outer angle joining the projecting portion of the disk. This triangular appearance is produced by its being folded over on itself, something like a T-hinge ; at the same time, as if this fold had not produced a sufficient shortening of the caudicle, a thin tail-like portion projects beyond. I was, at the time when I first examined it, inclined to think that the thickened fold was the homologue of the drum-like pedicel of H. chlorantha mentioned by Mr. Darwin; but the structure under the microscope appeared to indicate that it is really a thickened portion of the caudicle corresponding to the discal ex- tremities of the caudicle in the two former species. I watched very carefully to see whether any movement took place on removal, and was at first inclined to think so: but -on more careful examination I found that I was mistaken; in fact the incurved portion of the pollen-masses is quite sufficient to place them in a proper position for the fertilization of the flower. I found considerable difficulty in removing the viscid disk, although its prominent position seemed to offer as great fa- cilities as in the other species; and the constant fertilization of the flowers throughout a whole spike leads me to suspect that my pin had too smooth a surface for the viscid disk to adhere to. Observations on the Mode in which certain Species of Asclepiadew are fertilized. Abstract of a paper by J. P. Mansen WEALE, B.A. Oxon. (Communicated by CuHartzs Darwin, Esq. F.R. & L.SS.) [Read November 3, 1870.] Ox placing the blossoms of Gomphocarpus physocarpus in water, 1 noticed that numbers of flies, attracted by the sweet nectar con- CERTAIN SPECIES OF ASCLEPIADEX. 49 tained in the cucullate folioles, got attached to the stigmatie glands and appeared unable to release tbemselves. On allowing the flowers to remain until completely withered, 1 ascertained that the flies had not sufücient strength to extricate themselves and eventually perished with the flowers. That other insects also frequented the flowers to their own de- triment was abundantly visible from the remains of legs, belong- ing to small moths and other insects, detached and adhering to the stigmatic glands. After leaving Port Elizabeth my researches were for some time abandoned, and were resumed partly at the Koonap and partly at . Bedford. At the latter place I found that the same insects fre- quented G. pAysocarpus, with the important addition of several large wasps. It was here that I first observed that the pollen-masses were inserted in the fissures of the anthers ; but in most instances they seemed to have been pushed down instead of being regularly in- serted after withdrawal. In two other species of GompAocarpus common at Bedford I found the pollen-masses removed, and in some instances inserted, and I also captured several species of winged Hymenoptera with pollen-masses attached to their tarsi. Besides these, I may mention a beetle belonging to the genus Lycus, some moths, and Pyrameis cardui. At Ettrick I observed some plants of Gomphocarpus fructicosus and G. physocarpus, the first a true denizen of the Karroo, the latter confined to the grass-country, which two kinds of soil join hereabouts. Both plants were visited by large Hymenoptera ; and their flowers and pollen-masses resemble each other very closely in structure. I gathered in the neighbourhood specimens which appeared to partake of the characteristies of both plants in a modified degree, such as the distribution of tomentum, the colouring of the foliage and flowers, and the shape of the fol- licles. Iregret to state that, owing to several days’ incessant rain, these specimens were destroyed by mould during desic- cation, yet I am almost convinced that they partook of a hybrid nature, and am inclined to think that, under favourable cir- cumstances, they may be fertile and be established as permanent varieties. Later observations have tended to confirm this opinion ; for it is LINN. JOURN.— BOTANY, VOL. XIII. E MISSOURI 50 MR. J. P. M. WEALE ON THE FERTILIZATION OF not uncommon for plants of most dissimilar genera in this Order to become fertilized by alien pollen, although in such instances the fruit is not matured. ? Xyomalobium lingueforme ? Harv. MSS. This plant grows abundantly near my house, and flowers from November to January. The calyx is small. The corolla is large, and the segments curve upwards over the edge of the large stigma. The segments are parted nearly to their base, so that there are large openings between them. Each segment is directly oppo- site the fissure of the anthers, so as to prevent access from the -sides of the flower. The folioles are 3-lobed and fleshy. The lobes are stout and pointed; they curve outwards and inwards, like the corolla. The central lobe is the longest, and stands up on one side of the stigmatic gland between each anther. The smaller lobes of each foliole point upwards beneath each stigmatic gland. The bases of the folicles within secrete a sweet juice very attractive to Hymenoptera. The stigma is large and flat, and stands high up in the flower, so that the ale of the anthers project forwards between the cen- tral lobes of the folioles. The lobes are very convex and widely open at their base. It will thus be easily understood that while access to the stig- matie glands and fissures of the anthers is extremely difficult from the sides, from above the flower the same is tolerably easy, as the long central lobes of the folioles stand up between each gland, while a pair of the smaller lobes of two folioles meet immediately below the rounded projecting ale of the anthers. The folioles and stigma are pale yellowish green, nearly white, with a few brownish markings; and the corolla is green. The flowers are consequently inconspicuous although of moderate size. They have no scent, and do not secrete a large quantity of nectar at a time, although I imagine the flower continues its secretion for a long time until fertilized. This species is constantly visited by a large black-and-yellow wasp, which, from the neuration of the wings, belongs apparently to De St.-Fargeau’s genus Pallosoma, one of the Pepside. I have observed as many as six of these insects on one plant busily sucking the drops of nectar from the base of the tongue- shaped folioles. When thus engaged, they are exceedingly rest- less and active, straddling with their long legs across the flower CERTAIN SPECIES OF ASCLEPIADES. 51 and pushing their proboscides eagerly into the flower. While thus scrambling over an umbel of flowers many of the pollen- masses are extracted, by the claws of their tarsi catching in the notched stigmatic glands. In the prime of their flowering most of the plants have their masses withdrawn before withering, and sometimes as many as two pollen-masses inserted in one fissure, although it is seldom that pollen-masses are inserted between all the al of one flower. The stigmatic glands* and arms to which each pair of pollinia are attached are edged by a delicate pale yellowish transparent membrane, which I am inclined to think is viscidulous. The stigmatic gland is deeply furrowed in the centre, narrow at the apex, and widely open towards the base. The arms, which are rather short, are bent upwards at their junction with the gland, then again downwards in a rather deep curve, and again upwards. The edging membrane is carried beyond to the point where the pollen-mass is attached. The pollen-masses themselves are somewhat truncated and quadrate. In the following descriptions the margins of each pollen-mass will be named in reference to their position in the anther-cells ; viz. that which faces the inner extremity of the cell will be called the inner margin, and that facing the fissure of the ale the outer margin. From the above description it is evident that if an insect, while scrambling over the plant, inserted the claws of its tarsus or any other hooked portion of its body, such as its mandibles &c., be- neath the gland, or if the claw got inserted towards the base and was then drawn upwards in the contracting channel of the gland, it would become firmly attached and easily withdraw the pollen- masses. Although the pollen-masses are often withdrawn by the tarsi of insects belonging to the Coleoptera, Hymenoptera, Hemi- ptera, and Lepidoptera, I am much disposed to doubt whether in most instances this mode would ensure the replacement of the gland in the fissures formed by the ale of the anthers. * In the description of the stigmatic gland and arms the latter are described relatively to the gland, and not to their position in theanther-case. Thus when they are spoken of as extending outward, it is in reference to the position of the gland, and not to the anther-case. I mention this in order to render the de- scription plainer, the margins of the pollen-masses themselves being spoken of in an inverse manner. +) ~ 52 MR. J. P. M. WEALE ON THE FERTILIZATION OF On the wasp already mentioned I have found pollen-masses at- tached to the tarsi, to the long hairs of the sternum and coxe, and to the spines of the leg; but I have never found more than a single pair thus attached, and have never found glands separated from their masses. On the other hand, it is by no means uncommon to find several combinations of the glands attached to the unremoved pollen-masses, as also to those inserted in the fissures of the an- there. I once found and figured a portion of the head of an insect attached to a pollen-mass, but I unfortunately lost it in removing it from the pollen-mass to which it was fixed. I have twice observed these wasps with several pollen-masses attached to some portion of the head, but failed in capturing the specimens. I noted this especially as occurring on the 23rd of January of this year, when I observed numbers of these wasps frequenting the plants. I am myself thoroughly convinced of the correctness of this view, as without it there ean be no explanation of the structure of these plants. I have repeatedly watched wasps with the pollen-masses attached to their tarsi; and although they have visited many flowers, I have in no instance seen the masses inserted, although it is not uncommon to see them thus withdrawn. Although numbers of pairs of pollinia are withdrawn, very few in comparison are inserted. I have already stated that in Gomphocarpus fruticosus the pol- len-masses close together inwardly on withdrawal, so as to clasp tightly the leg of the insect. In this species there is a similar movement, only in a slighter degree. The arms bend inwardly, so that the two pollen-masses are nearly parallel, but somewhat apart from each other. I have not ascertained the structure which produces these and other movements in the pollen-masses of Asclepiads, as, owing to their diminutive size, they require close and careful examination, for which I have not had sufficient leisure. It is a curious matter for observation that in some Asclepiads this movement never takes place, in some only slightly, while in others, as in Periglossum and Aspidoglossum, the movement takes place in an entirely different direction. In this species I have found most curious combinations of the pollen-masses. I-have seen five glands attached together, in the first of which an arm was still inserted, showing that the com- CERTAIN SPECIES OF ASCLEPIADEZ. 53 bination was originally greater. From these, four pollen-masses had been detached, eight still remaining attached to the glands. In another instance four glands were attached together, from which six pollen-masses had been removed, two only remaining attached to the unremoved gland. In this case also an arm was left disruptured in the upper gland, showing that the combination had been formerly larger. Towards the close of their flowering-season a careful examina- tion of these plants would probably result in many more curious combinations. As an illustration of the probable correctness of this species not being fertilized by the tarsi, I have never met with a gland attached by itself to the tarsus; yet in most cases the presence of a disruptured arm would show that one gland at least remains attached. When the Pallosoma visits these flowers, as I have frequently observed, it plunges its head in between the middle lobes of the folioles to their base, busily sucking the nectar; but in so doing the smaller lobes, projecting upwards, interfere with it in some mea- sure, and, as I have noticed, caused considerable annoyance to the insect. To this and the exciting influence of the nectar I, in some measure, attribute its restlessness; although it is, apart from these, an active and watchful insect. I have also noticed that it some- times sucks round the gland itself; and as Robert Brown states that the gland, in the species which he examined, continues to secrete after the opening of the flower, I am disposed to think that this secretion may be of essential service to the flower in attracting the wasps when the more abundant store of nectar at the base of the folioles is exhausted. It is probable, too, that the adherence of so bulky an object as any of the combinations already mentioned to any part of the bead would cause much discomfort to so agile an insect, to release itself from which the fissures of the anthers offer a ready means. The attentions of this insect are paid to several other Ascle- piads, such as Periglossum, as also to a Cissus and a Eucomis. These flowers are, most of them, dull-coloured and of very different size, but afford, apparently, a quality of nectar peculiarly pleas- ing to this wasp; for there were in blossom at the same time Ascle- piads quite as conspicuous and more so than Periglossum, affording, too, an abundance of nectar, but which I have never seen it visit, although they appeared attractive to some other Hymenoptera. 54 MR. J. P. M. WEALE ON THE FERTILIZATION OF Lastly I have to advert to the singular fact of the pollinia of other Asclepiads being often inserted in different species—a fact the more strange, as in Periglossum I have always found the es- terior pellucid margin of the pollen-mass inserted instead of the inner, as in most other Asclepiads; and it is on this margin that the rupture and protrusion of the pollen-cells takes place, which is the normal mode of fertilization in this plant. I have on many occasions found the pollen-masses of Periglos- sum inserted in the fissures of the anthers of this species; and not merely are they inserted, but the masses are disrupted and give out tubes which appear to penetrate the ovarium of the flower. This noticeable fact, combined with my remarks on Gomphocar- - pus fruticosus and G. physocarpus lead me to suspect that such accidents may occasionally lead to results of which we are at pre- sent but faintly aware, but to elucidate which a series of repeated and careful experiments would be requisite. In the first plant, out of 6 flowers open, 4 had pollen-masses inserted, and 2 had 4 pairs of pollen-masses extracted. If it be reckoned that each flower has 5 pairs of pollen-masses and 5 fissures, it will be seen that this plant has very few withdrawn, —much fewer than is usual, so far as my observations go. Thus 6 x 5=30 pairs pollen-masses, 4 , withdrawn, 20 , 1n situ, 6 x 5=30 fissures, 6 with masses inserted, 24 unfertilized. In the second plant, not reckoning that destroyed by insects, out of 8 flowers, 7 had pollen-masses inserted, and one had foreign pollen inserted, and 7 had 19 pairs of pollen-masses ex- tracted. Thus, not reckoning the foreign pollen, S x 5 —40 pairs pollen-masses, 19 5 withdrawn, 21 . in M. CERTAIN SPECIES OF ASCLEPIADEX. 55 8 x 5=40 fissures, 16 with masses inserted, 24 unfertilized. As I have collected seed of this and other Asclepiads, I hope next year to supply the Society with some statistics more satis- factory than those inserted in this paper. Pachycarpus. This plant, which is almost as abundant on my farm as the last, produces many and larger flowers, but very seldom fruit, and then generally only one follicle on a plant. The corolla is brown, and widely open. The folioles are horizontal and ex- panded, as in the genus generally, and contain a good deal of nectar in the furrows. The stigma does not project as in the last species. The ale of the anthers project outwards, are widely open, and acute at the base, where they turn slightly upwards. The pairs of pollinia are widely expanded and the masses oblong. The stigmatic gland is large and channelled, very narrow in the centre, and broadly open at the apex and base. The arms are curved downwards, and outwards, upwards, and downwards, at their junction with the masses. Where they join the gland there are two small expansions of membrane, and the arms themselves are slightly edged with membrane. On removal the arms and masses are never inflexed, but remain as rigid as when in the anthers. I have only very seldom found masses inserted in the fissures, although I have examined many plants on different occasions. Periglossum. The flowers of this plant are arranged in dense umbels, and are greenish and very inconspicuous. The plant itself has much the aspect of a Carex, and grows among rank herbage by the banks of streams. The corolla is not very widely open. The folioles adhere closely to the stigma, are broad above, and rounded, somewhat like a half moon, with the horns bent downwards like hooks. Below the half-moon expansion the foliole is much contracted. The pollen-masses are remarkable for the minute size of their glands and the length of their arms. The arms are bent down- wards, outwards, upwards, and downwards from the junction with their gland to the pollen-masses. The pollen-mass itself is very small compared with the length of the arms. The arms are 56 MR. J. P. M. WEALE ON THE FERTILIZATION OF expanded where the masses join. The masses bend outwards, are elongated and considerably curved. Their last third is pel- lucid and much attenuated. It is on this pellucid outer margin that the rupture and protrusion of the pollen-tubes take place. The ale of the anthers project but little, and extend low down in the flower. They are widely open at the base, and then sud- denly contract, so as to form a sort of sharp notch. The whole gynostege is closely enveloped by the folioles and corolla. On withdrawing the pollen-masses with a pin, the movement, which is very curious, can be easily seen. On lifting up the gland a short distance, it bends inwards towards the centre of the stigma, and the arms outwards and away from it. On withdrawing it entirely, the long arms bend out completely, and hang loosely from the small gland. The pollen-masses somewhat resemble the long curved-up hoofs which sheep acquire when feeding on marshy soft ground. The disruption of the pollinia generally takes place where the arms are joined to the gland, and not at the junction of the pollen-masses to the arms, as is usually the case. I have, on one occasion, seen them disrupted in this place. I think it not improbable that the long thin dangling arms of the pollen-masses may render the fertilization of this flower somewhat diffieult, and that the hooked folioles delay an insect by the long upcurved pollinia being caught in them, in which case it would push downwards in order to extricate itself, and the end of the pollen-mass would almost certainly be caught in the sharply notched fissure of the anther, and become detached. The hooked portion of the foliole is widely eurved, and thus the pollen-mass would be prevented from becoming detached as it does in the bootjack-like fissure. The Pallosoma, already mentioned, is a very constant visitor of this plant; and had it not been for its conspicuous visitor I should have often missed a plant. I think this affords a useful hint to collectors, as many small and inconspicuous Asclepiads would otherwise have often escaped my notice. I have never observed pollen-masses attached to the tarsi of this Wasp, nor have I on any occasion seen combinations of the glands ; nor do I see how such a combination could be possible, as the arm is so attenuated at its junction with the gland that it 18 very easily ruptured, whereas, at the other extremity, the width of the arm is greater than that of the pollen-mass itself. The wasp, when visiting the plant, greedily sucks nectar from the CERTAIN SPECIES OF ASCLEPIADEE. 57 base of the folioles, although it is apparently secreted in very small quantities. Like the other species already described it is destitute of scent. Cordylogyne. The pollen-masses are oblong and bend outwards. On removal the arms bend slightly in. Combinations take place between the glands and pollen-masses in the usual manner, viz. by the insertion of the arms in the furrows of the gland. The flowers are very extensively visited by some insects unknown to me. This concludes my present examination of the order; but although imperfect, my observations lead me to expect great results from future investigations, and that we shall find in the other subdivisions of the order, viz. the Tribes Periplocee, Seca- moree, and Stapeliee, as wonderful contrivances graduating into each other as in the different tribes of Orchids. In conclusion, I would remark that sufficient evidence has been obtained to show that insects extensively fertilize these plants, and that there appears to be an adaptation between the form and movements of the pollinia, the ale of the anthers, and the position and shape of the folioles and corolla. The secretion of nectar, and the powerful scent, so offensive in some, so sweet in others, are probably by no means the least im- portant functions of the plants. As a rule these plants are dull-coloured ; but there are many conspicuous exceptions among the Stapeliez and others. From the very small quantity of mature follicles produced in each plant, so disproportionate to the number of flowers, it is highly probable that these plants require to be impregnated by several pollen-masses ; and when we take into consideration how many pollen-masses are removed in proportion to those inserted, it would seem that in most species nature is less economical in her adaptation of means to ends than she is in most instances. It is also a singular fact that although all the flowers have two carpels, one is almost universally abortive. A general view of all these facts would lead to the conclusion that the adaptation of the several parts of the flower, in such Asclepiads as have been examined, is inferior in its perfection to that attained in Orchids. Perhaps some sort of compensation is obtained by the relatively large size of the seeds and the tufts of silken hair with which they are provided, by means of which they are wafted away by the slightest breeze. 58 MR. G. BENTHAM ON AUSTRALIAN PROTEACEZ. Orchids are manifestly inferior in this respect. They pro- duce, as remarked by Mr. Darwin, a prodigious quantity of fine seed which rarely germinates. Notes on the Styles of Australian Proteacee. By Grorer BentHdm, F.R.S., P.L.S. [Read April 6, 1871.] (Prarzs I. & 11.) In the Proteaces, as in the Composite and some other Orders, it had been observed that the anthers in most cases open and dis- charge their pollen upon an enclosed pubescent papillose or glu- tinous portion of the style, usually described as the stigma, before the flower expands ; and it was therefore concluded that fecunda- tion then and there took place. This has now long been shown to be a fallacy in the case of Composite; for, as Lessing and others have pointed out, the really stigmatie portion of the style is always on the inner face and often only at the base ofthe style- branches, which remain hermetically closed until the flower has opened and they are protruded beyond the anthers. Then, and then only, do these branches open so as to render the stigmatic surface accessible to any pollen which may be shed upon them. In the Proteace: the case is different; the style is undivided, the stigmatic surface is superficial even in the bud, and the contri- vances to screen it more or less from the action of the pollen which is then being scattered around it, reserving it for the pollen of other flowers after it has been released from the enclosing pe- rianth, are very various. Those which I have observed in the course of my examination of the Order for the Australian flora are chiefly the following. These observations, however, are made almost exclusively on dried specimens, as I was only able to exa- mine a very few Grevillea and Hakea flowers in a living state, and the notes I could collect from previous observers were but very few. They will require, therefore, to be supplemented and probably in several instances corrected by those who can watch the process of ripening and mutual action of the anthers and stigma on the living plants. As a general rule, the anthers in the bud form a close cylinder round the papillose portion of the style, which has probably some stimulating influence on them ; for immediately before the open- ing of the flower we find the anthers open inside and the pollen- MR. G. BENTHAM ON AUSTRALIAN PROTEACEA, 59 grains crowded on the style, whilst the real stigma, whether within or above or below the anther-cylinder, is as yet imma- ture, dry, and evidently incapable of absorbing pollen. When the flower is quite ready to expand, the force which overcomes the cohesion of the valvate perianth-segments and anthers gene- rally bursts them asunder with more or less elasticity so as to promote the scattering of the previously loosened pollen, after which the liberated style matures its stigma and becomes ready to receive any pollen that may reach it from neighbouring flowers. The cohesion of the perianth-segments is generally stronger in their limb or antheriferous portion than in the tube, and more particularly so either at the tip or at the base of the limb at the point of insertion of the stamens, immediately under the base of the anthers. The ripe anthers, with the immature stigma buried in a mass of pollen, are thus kept in close confinement and inacti- vity, in some instances for a lengthened period. When the style by its growth at last succeeds in liberating itself, it is chiefly in two ways. In many of the straight regular-flowered species it will force its way straight through the end, separating the tips of the segments, which then roll back with more or less of elasticity to the base of the stamens or lower down. In many curved- flowered species the resistance opposed by the cohesion of the limb is greater, the style by its growth becomes more and more bent like a bow, breaks out through a lateral slit, and finally draws out its stigmatie end from the limb either by splitting it from the base upwards or by slightly opening it at the base only. It would appear that the extra force required by the style for this final effort is sometimes influenced by meteorological or other ex- ternal eonditions whieh may not always occur; for we often see in Banksia cones which have been long in flower, and have even ripened their seeds, the majority of the withered flowers with the anthers and end of the style still closely imprisoned in the closed limb. But then in those cones the majority of the ovaries have not ripened into fruits; and I believe that the perfect fruits always correspond to liberated styles; but this point requires further observation on the living plant. One thing appears certain, that there is no genus in the order where the stigma is longer kept smothered in a bed of pollen, whilst there is none where effective fecundation is proportionally more rare. In a cone of about a thousand flowers we often find not more than two or three dozen, and sometimes not one dozen, fully formed fruits. 60 MR. G. BENTHAM ON AUSTRALIAN PROTEACEAX. As a good example of the style of the regular straight flower we may take that of Petrophila longifolia (Plate I. fig. 1), the upper portion of which is usually described as a biarticulate stigma. But, in the first place, there is no articulation ; the upper portion, which may here well be called the brush, does not separate from the lower, and neither portion is stigmatic on the surface. The brush, a dense mass of short papillose hairs, corresponds exactly in length with the anthers, and collects their pollen ; at the end and shut out from all contact with the pollen by the closed ends of the anthers, is the small stigmatic surface, which, besides its immaturity, is further protected by being closely applied to the inflected tips of the perianth-segments. The part played by the so-called lower article, the expanded turbinate end of the smooth part of the style, is not very clear; it probably acts as an impediment to the premature escape of the pollen through the perianth-tube, or possibly by its expansion assists in the forcing open of the segments. This distinction of the brush and turbinate base, how- ever, is not so marked in all Petrophile; the brush is some- times very thin and slightly papillose, tapering at the base into the slender style, the papillose portion extending always only to the base of the anthers. In the very natural genus Persoonia there are three sections, showing two very different stigmatic arrangements. In : € . : > O_; 522; Kunth, Enum. iv. 358. Bulbus ovoideus sordide virescens 2-3 MR. J. G. BAKER ON SCILLEXZ AND CHLOROGALEZ. 271 poll. erassus. Folia 5-6 carnoso-herbacea glabra lorata 1-2 pedes longa, 1-2 poll. lata deorsum concava extrorsum planiuscula ad apicem - acutum sensim attenuata. Scapus validus 12-2-pedalis. Racemus deusus 1-13 pedes longus, 50-100-florus vel ultra, expansus 15-18 lin. latus. Pedicelli ascendentes, infimi 5-6 lin. longi. Bracteæ lan- ceolato-setacez, infime 6-9 lin. long. Perianthium 5-6 lin. longum segmentis lanceolatis 13 lin. latis dorso viridi vittatis. Filamenta line- aria subzequalia 23-3 lin. longa. Styluslilin.longus. Cap. B. Spei. 48. O. LONGEBRACTEATUM, Jacq. Hort. Vind. t. 29; Red. Lil. t. 190; Schult. fil. Syst. vii. 521; Kunth, Enum. iv. 357.—0O. bracteatum, Thunb. Prodr. 62; Fl. Cap. 314. Bulbus ovoideus 3-4 poll. crassus. Folia 5-6 carnoso-herbacea glabra lanceolata viridia 13-2 pedes longa, 9-18 lin. lata ad apicem acutum sensim attenuata. Scapus validus strictus 13-2-pedalis. Racemus dense 30—60-florus vel ultra, expan- sus 6-9 poll. longus, 15-18 lin. latus. Pedicelli ascendentes, infimi 9-12 lin. longi. Bractew 9-12 lin. long: setacez basi lineares ante florescentiam conspicue. Perianthium 43-5 lin. longum segmentis oblongis obtusis albidis 13 lin. latis dorso viridi-vittatis. Filamenta 1-3 lin. longa alterna linearia et lanceolata. Stylus 1 lin. longus. Cap. B. Spei, Thunberg! Dr. Gill! etc. 49. O. NARBONENSE, Linn. Sp. Plant. 440; Gouan, Ill. 26; Bot. Mag. t.9510; Kunth, Enum. iv. 355; Gren. Fl. France, iii. 183; Parl. Fl. Ital. ii. 447; Reich. Icones, t. 1029.—O. stachyoides, Ait. Kew. i. 441; Kunth, Enum. iv. 356 ; Reich. Icones, t. 1030.—Beryllis sta- chyoides, Salisb. Gen. 33.—O. lacteum, Vill. Delph. ii. 272, non Jacq. —O pyrenaicum, Desf. Atl. 293, non Linn. Bulbus ovoideus 9-15 lin. erassus. Folia 4-6 firmiora quam in pyrenaico et ante finem flo- rationis haud evanida carnoso-herbacea anguste lorata 12-18 poll. longa, 3-6 lin. lata, glabra glauco-viridia facie canaliculata. Scapus 1-1i-pedalis strictus. Racemus sublaxe 20-50-florus, expansus 4-8 poll. longus, 11-2 poll. latus. Pedicelli primum subpatentes, deinde ascendentes, inferiores 9-15 lin. longi. Bractez lanceolate cuspidate 6-9 lin. longe. Perianthium 5-6 lin. longum segmentis oblanceo- latis subobtusis 1-13 lin. latis albidis distincte viridi-vittatis 4—6-ner- vatis. Filamenta 2 lin. longa basi lanceolata. Stylus filiformis 13 lin. longus. Ad insulis Fortunatis, Lusitania, et Gallia ad Algeriam, Graeciam, et Caucasum. O. trigynum, Red. Lil. t. 4V7 ; Kunth, Enum. iv. 357, est forma monstrosa hortensis carpellis discretis; O. brachy- stachys, C. Koch, Linnea, xxii. 248, ex descriptione non potui segre- gare; O. densum, Boiss. et Blanche, Diagn. ii. 4, 107, est forma ori- entalis gracilis densiflora; O. pyramidalis, Linn. Sp. Plant. 441, Red. Lil. t. 422, Jacq. Ic. t. 425, Kunth, Enum. iv. 355, est forma ro- busta grandiflora hortensis. Var. B. FUSCESCENS, Baker.—O. fuscescens, Boiss. et Gaill. Diagn. ii. 4, 278 MR. J. G. BAKER ON SCILLEX AND CHLOROGALEJE. 107. Perianthium fuscescens, filamentis segmentis quadruplo breviori- bus. Bractez utrinque lacinia setacea aucte. Antilibanus, Gaillardot. 50. O. CRENULATUM, Linn. Suppl. 198; Thunb. Prodr. 62; Fl. Cap. 315; Kunth, Enum. iv. 371. Bulbus ovoideus 5-6 lin. crassus. Folia 2 oblonga obtusa vel subacuta crassa carnoso-herbacea 12-15 lin. longa, 4-5 lin. lata marginibus prominulis papillosis glabris. Seapus 2—1-pollicaris. Racemus 4-8-florus, 1-2 poll. longus, 5-6 lin. latus. Pedicelli diutine ascendentes, infimi 5-4 lin. longi. — Bractez lanceo- late acuminate 3-4 lin. longi. Perianthium 3-33 lin. longum, seg- mentis ovato-lanceolatis 14 lin. latis albis distincte viridi-vittatis. Fi- lamenta subzqualia lanceolata 2 lin. longa. Stylus 1 lin. longus. Cap. B. Spei, ''hunberg. 51. O. ovarum, Thunb. Prodr. 62; Fl. Cap. 315; Kunth, Enum. iv. 359. Folia *2"' rigide coriacea oblonga 14-2 poll. longa, 4-6 lin. lata subobtusa cuspidata marginibus incrassatis scariosis glabris. Scapus pollicaris. Racemi dense 6-20-flori, 9-15 lin. longi, 10-12 lin. lati. Pedicelli infimi 12-2 lin. longi. Bractez lanceolate acuminate 2-3 lin. longæ. Perianthium 3 lin. longum, segmentis ligulatis ob- tusis l lin. latis albis distincte viridi-vittatis. Filamenta Janceolata subzequalia 2-23 lin. longa. Stylus 1 lin. longus. Cap. B. Spei (vidi exempla originalia in Herb. Thunb., sed an folia ad scapum recte per- tineant dubito). Subgenus V. OsmYNE (Salisb. extens.). Racemi 6-40-flori, expansi ovoideo-deltoidei vel lanceolati. Peri- anthii segmenta flava dorso distincte viridi carinata 2~4-nervata, flore expanso rotata vel leviter recurvata. Filamenta segmentis duplo breviora vel interdum longiora. Stylus elongatus filifor- mis ovario :equilongus vel interdum duplo longior.— Osmyne et Teniola, Salisb.— Ornithogalum, $ Albucoides, Kunth, ex parte. 52. O. virratum, Kunth, Enum. iv. 368.—Albuca vittata, Gawl. Bot. Mag. t. 1329; Schultes fil. Syst. vii. 500.—Teniola vittata, Salisb. Gen. 35. Bulbus ovoideus 1 poll. crassus. Folia 5-6 subteretia car- noso-herbacea 6-8 poll. longa 1-14 lin. lata glabra facie canaliculata, glaucescenti-viridia. Scapus strictus 6-9-pollicaris. Racemus laxe 6-12-florus expansus, 3-4 poll. longus, 13 poll. latus. Pedicelli in- feriores 6-9 lin. longi subpatentes, leviter cernui. Perianthium 5-6 lin. longum inodorum segmentis oblanceolatis 1-13 lin. latis leviter reflexis luteis dorso distincte anguste viridi-carinatis 4-5-nervatis. Filamenta 23-3 lin. longa basi lanceolata, alterna latiora bicuspidata. Stylus filiformis 3-53 lin. longus gracillimus. Cap. B. Spei. 53. O. BARBATUM, Jacg. Hort. Schoen. t. 91 ; Kunth, Enum. iv. 366. Bulbus ovoideo-subglobosus 1 poll. crassus. Folia 1-2 carnoso-her- MR. J. G. BAKER ON SCILLEJE AND CHLOROGALEA. 279 bacea subulata glabra 6-9 poll. longa facie canaliculata. Scapus sub- pedalis folio crassior teres erectus. Racemus laxe 5-6-florus expansus 23-3 poll. longus. Pedicelli erecto-patentes, infimi 5-6 lin. longi. Bractez lineares purpurascentes 3-4 lin. longa. Perianthium 5-6 lin. longum segmentis oblongis obtusis 13-2 lin. latis luteis viridi- carinatis omnibus apice leviter cucullatis interioribus pubescentibus. Filamenta subsequalia, segmentis subduplo breviora, basi complanata. Stylus filiformis 2-3 lin. longus. Cap. B. Spei. 54. O. TUBEROSUM, Mill. Dict. no. 10; Kunth, Enum. iv. 372.—0O. polyphyllum, Jacq. lc. t. 430; Kunth, Enum. iv. 366.—O. consan- guineum, Kunth, Enum. iv. 368. Bulbus ovoideus 12-15 lin. crassus fuscescens. Folia plura, 8-12, lineari-subulata, carnoso-herbacea, 6-8 poll. longa, 1-13 lin. lata, facie canaliculata, basi dilatata. Scapus 6— 18-pollicaris. Racemus laxe 6-12-florus, expansus 3-4 poll. longus, 27-3 poll. latus. Pedicelli subpatentes, infimi 12-15 lin. longi. Bractez lanceolate acuminatz pedicellis subequilonge. Perianthium odorum 9-10 lin. longum segmentis oblongis obtusis | 1-2 lin. latis luteis dorso distincte late viridi-carinatis 4-5-nervatis. Filamenta subzqualia 3- 4 lin. longa basi applanata. Stylus rectus 3 lin. longus. Cap. B. Spei, Ecklon! Zeyher! Mund et Maire. 55. O. BorusraNuM, Baker. Bulbus depresso-globosus 1 poll. crassus. Folia 2-1 superposita basi subpetiolata basin scapi amplectentia la- mina horizontaliter patente lanceolata 15-18 poll. longa, 4-5 lin. lata acuta margine ciliata. Scapus gracilis semipedalis. Racemus laxe 6-10-florus, expansus 3-4 poll. longus. Pedicelli ascendentes 17-3 lin. Jongi. Bractez parvz deltoidez acuminate. Perianthium 43-5 lin. longum segmentis lanceolatis subacutis flavicanti-viridibus flore expanso reflexis indistincte carinatis. Filamenta lanceolata, segmentis duplo breviora, alterna latiora. Stylus filiformis ovario brevior. Cap. B. Spei, Bolus, 96! Vidi tab. ex exemplo in hort. Kew. anuo 1823 cultum ab Bowie missum. 56. O. Kinkrr, Baker.— Folia plura carnoso-herbacea glabra linearia basin scapi amplectentia ad pedem et ultra longa 3-4 lin. lata. Scapus gracilis semipedalis. Racemus laxissime 6-10-florus expansus 5-6 poll. longus, pedicellis ascendentibus infimis 13-2 poll. longis superi- oribus sensim brevioribus. Bracteze lanceolate scariose 3-4 lin. longz. Perianthium 5-6 lin. longum segmentis 13 lin. latis omni- bus apice cucullatis luteis dorso late distincte viridi-carinatis. Filamenta 3-4 lin. longa subzequalia linearia. Stylus cylindricus 2- 23 lin. longus. Africa tropicalis austro-orientalis in ditione fluminis Zambesi ad Shupanga, Dr. Kirk! 57. O. SUAVEOLENS, Jacq. Ic. t. 43! ; Willd. Sp. ii. 122; Schult. fil, Syst. Veg. vii. 523 ; Kunth, Enum. iv. 366,—0O. albucoides, Thun. Fl, 280 MR. J. G. BAKER ON SCILLE.E AND CHLOROGALEZ. Cap. edit. Schult. 314 ; Kunth, Enum. iv. 367.—Antherieum albucoides, Ait. Hort. Kew. i. 449.— Phalangium albucoides, Poir. Ency. v. 249. Bulbus globosus 1-13 poll. crassus. Folia 5-6 carnoso-herbacea linearia 6-9 poll. longa, 2-2! lin. lata acuta basin scapi amplectentia. Scapus pedalis et ultra strictus teres modice validus. Racemus laxe 6-12-florus, expansus 4-6 poll. longus. Pedicelli ascendentes, infe- riores 13-23 poll. longi. Bractez lanceolate 8-12 lin. longze scariose persistentes. Perianthium 6-8 lin. longum suaveolens segmentis ob- lanceolatis obtusis 13 lin. latis luteis dorso late carinatis distincte 4-5- nervatis. Filamenta 3 lin. longa subzqualia basi leviter applanata. Stylus filiformis 21-3 lin. longus leviter declinatus. Cap. B. Spei, Thunberg ! Hort. Kew. anno 1788! Drege, 1511! (Exemplum origi- nale Aitonianum in herb. Banks. et Thunbergianum in herb. suo vidi.) Var. 8, ODORATUM, Baker.—O. odoratum, Jacq. Icones, t. 432; Andr. Bot. Rep. t. 260; Kunth, Enum. iv. 366.—Osmyne odorata, Salisb. Gen.35. Robustior foliis latioribus, racemo interdum pedali, bulbo 2 poll. crasso. Cap. B. Spei. Vix ultra formam robustam hortensem. 58. O. PRASINUM, Lindl. Bot. Reg. t. 158; Kunth, Enum. iv. 360. Bulbus ovoideus 13-2 poll. erassus. Folia 6-8 carnoso-herbacea line- aria pedalia glaucescenti-viridia 4-5 lin. lata ad apicem sensim angus- tata, facie canaliculata. Seapus pedalis et ultra. Racemus dense 20- 40-florus, expansus 3-5 poll. longus, 3-4 poll.latus. Pedicelli sub- patentes stricti, infimi 2 poll. longi. Bractez lineari-subulate, infimae 6-9 lin. longz. Perianthium inodorum 6-7 lin. longum segmentis oblongis obtusis 13-2 lin. latis luteis dorso late viridi-vittatis distincte 4-5-nervatis. Filamenta 3-4 lin. longa subzqualia basi lanceolata. Stylus gracilis filiformis 3 lin. longus. Cap. B. Spei, Burchell, 1966! Ad ripas fluminis Aapages, Burke! 59. O. MELLERI, Baker. Bulbus ovoideus 12 poll. crassus. Folia 4- 5 lineari-lorata carnoso-herbacea basin scapi amplectentia sesquipedalia glabra 5-6 lin. lata ad apicem sensim angustata. Scapus 13-2-pedalis strictus modice validus. Racemus laxe 6-15-florus, expansus 6-9 poll. longus, 15-18 lin. latus. Pedicelli subpatentes vel ascendentes flori- feri distincte cernui, infimi 4-6 lin. longi. Bractez lineares membra- nace 3-4 lin. longe. Perianthium 8-9 lin. longum segmentis ob- longo-oblanceolatis 2-23 lin. latis obtusis luteis dorso late viridi-cari- natis distincte 5—6-nervatis. Filamenta 5-6 lin. longa subzequalia basi deltoidea valvata. Stylus filiformis 5-6 lin. longus. Africa tropicalis austro-orientalis ad montes Manganja, Dr. Meller. In Hort. Kew. anno 1862 cult! 60. O. MACRANTHUM, Baker. Folia carnoso-herbacea lorata glabra flaccida 14-2 pedes longa, 4-6 lin. lata. Scapus * * * . Race- MR. J. G. BAKER ON SCILLEE AND CHLOROGALE.E. 281 mus laxe 20-25-florus, expansus ad pedem et ultra longus 31-4 poll. latus. Bractez lineares 3-4 lin. longe ante anthesin evanescentes. Pedicelli 13-2 poll. longi ascendentes apice cernui turbinatim incras- sati. Perianthium 15-16 lin. longum segmentis teneris ad basin tertii superioris 23 lin. latis flavescentibus dorso anguste viridi-carinatis distincte 2-3-nervatis. Filamenta 8-9 lin. longa ex apice anguste pe- taloideo-complanata. Stylus rectus gracillimus 6-7 lin. longus. Cap. B. Spei, Drege, 2204! 3531! Subgenus VI. Caruissa (Salisb. extens.). Racemi 3-100-fori expansi lanceolati vel subcylindriei. Perianthii segmenta albida concoloria venis flabellatis inconspicuis, dorso haud viridi-carinatis, flore expanso patentia vel leviter recurvata. Filamenta hypogyna segmentis subduplo breviora. Stylus ovario «equilongus vel brevior.— Cathissa et Eustachys, ex parte, Salisb. 61. O. cnisEUvM, Baker. Bulbus fuscescens 13-2 poll. crassus. Folia 5-6 lineari-subulata glabra ascendentia 6-8 poll. longa haud 1 lin. lata. Scapus gracilis subpedalis. Racemus circiter 20-florus, expansus 3-4 poll. longus. Pedicelli erecto-patentes 2-3 lin. longi. Bractez line- ares 11-2 lin. longz. Perianthium 3 lin. longum segmentis subcon- coloribus albido-griseis 1 lin. latis obtusis alternis apice leviter cucul- latis. Filamenta subzqualia l lin. longa basi leviter applanata. Stylus ovario «equilongus. Cap. B. Spei (e tab. ex exemplo in Hort. Kew. anno 1823 cult. a Bowie misso). 62. O. ZEYHERI, Baker. Bulbus ovoideus 5-6 lin. crassus. Folia 5-6 carnoso-herbacea lineari-subulata glabra 6-12 poll. longa, ¿-1 lin. lata. Seapus 3-12-pollicaris. Racemus laxe 12-30-florus expansus, 3-6 poll. longus, 15-18 lin. latus. Pedicelli ascendentes, infimi 9-12 lin. longi. Bractez lineares membranacex 3-4 lin. longe. Perian- thium 21-3 lin. longum segmentis albidis subconcoloribus oblanceo- latis subacutis haud 1 lin. latis. Filamenta subzqualia 13 lin. longa basi linearia. Stylus ovario turbinato brevior haud 1 lin. longus. Cap. B. Spei, Drege, 8/85! Cooper, 601! Zeyher, 1686! 63. O. DELTOIDEUM, Baker. Folia 4-5 basin scapi longe involventia superposita parte libera subulata coriacea persistentia 1 poll. longa facie canalieulata prorsus setis flavidis conspicuis vestita. Scapus gracilis 3-4-pollicaris. Racemus laxe 3-6-florus. Pedicelli graciles ascendentes, infimi 3-4 lin. longi. Bracteæ membranace: late cor- dato-deltoidez cuspidate } lin. longz. Perianthium 4-43 lin. lon- gum segmentis lanceolatis albidis concoloribus acutis 1j lin. latis. Filamenta 3 lin. longa lanceolata subxqualia. Stylus ovario æqui- longus. Cap. B. Spei, Drege, 2664! 282 MR. J. G. BAKER ON SCILLEZE AND CHLOROGALE.E. 64. O. PUBESCENS, Baker. Bulbus ovoideus 3-4 lin. crassus. Folia 2-3 basin scapi longe involventia superposita, lamina horizontaliter patentia lanceolata membranacea 6-12 lin. longa, 2 lin. lata undique pilis setosis conspicuis vestita. Scapus gracilis 4—5-pollicaris flexuo- sus. Racemus laxe 5—6-florus primum nutans, expansus | poll. longus. Pedicelli ascendentes, infimi 2-3 lin. longi. Bractez lineari-subulate li-2lin. longe. Perianthium campanulatum 13-2 lin. longum seg- mentis lanceolatis acutis concoloribus. Filamenta | lin. longa fili- formia subxqualia. Stylus 1 lin. longus. Cap. B. Spei (Albany), Williamson in herb. Trin. Coll. Dub. ! 65. O. viLosuM, Linn. Suppl. 199; Thunb. Prodr. 61; Fl. Cap. 315; Kunth, Enum. iv. 3/0.—O. Mundianum, Kunth, Enum. iv. 351. Bulbus ovoideus 6-9 lin. crassus castaneo-tunicatus. Folia 3-4 basin scapi amplectentia rigide coriacea linearia 4-6 poll. longa, 14-2 lin. lata, distincte nervata, persistentia, marginibus distincte ciliatis. Seapus gracilis flexuosus 6-12-pollicaris. Racemus sublaxe 6-20- florus, expansus 2-3 poll. longus, 1-13 poll. latus. Pedicelli ascen- dentes spe arcuati, infimi 6-12 lin. vel ultra longi. Bractez persis- tentes lanceolatz cuspidate, infime 4-6 lin. longz. Perianthium 4-5 lin. longum segmentis lanceolatis acutis albidis concoloribus 1 lin. latis. Filamenta subequalia lanceolata 23-3 lin. longa. Stylus 13 lin. longus ovario zquilongus. Cap. B. Spei, Masson! Thunberg ! Burchell, 6802! Cooper, 1670! 1695! etc. 66. O. nisPiDUM, Hornem. Hort. Hafn. 331; Kunth, Enum. iv. 350. —Anthericum pilosum, Jacq. Ic. t. 416; Willd. Sp. Plant. ii. 140.— Phalangium pilosum, Poir. Ency. v. 244. Bulbus depresso-globosus albidus 9-12 lin. crassus. Folia 3-4 basin scapi longe involventia superposita linearia plana flaccida carnoso-herbacea patentia 5-6 poll. longa, 5-6 lin. lata ad oras pilosa. Scapus gracilis 8-10-pollicaris. Racemus laxe 8-9-florus, expansus 3-4 poll. longus, 2 poll. latus. Pedicelli ascendentes, infimi 9-12 lin. longi. Bractez lanceolata 4- 6 lin. longe. Perianthium 8-9 lin. longum inodorum segmentis albi- dis subconcoloribus lanceolatis acutis flore expanso stellatis. Filamenta eequalia subulata segmentis subduplo breviora. Stylus filiformis ovario equilongus. Cap. B. Spei. . 67. O. uNIFoLIUM, Gawl. Bot. Mag. t. 935; Schultes fil. Syst. vii. 529; Kunth, Enum. iv. 359.— Scilla unifolia, Linn. Sp. Plant. 443. —Cathissa uniflora, Salisb. Gen. 34.—O. nanum, Brot. Fl. Lusit. 529! Phyt. Lus. t. 46. fig. 1, non Smith. Bulbus ovoideus subfuscus 4-6 lin. crassus. Folium solitarium, basin scapi longe involvens, lineare, facie canaliculatum, glabrum, carnoso-herbaceum, 3-6 poll. longum, 13-2 lin. latum. Seapus 3-6-pollicaris. Racemus subspi- catus 3-6-florus. Bractew lanceolate cuspidate 4-6 lin. longs. 3 MR. J. G. BAKER ON SCILLEA AND CHLOROGALE. 283 Perianthium 5-6 lin. longum, segmentis oblongis obtusis 1 4-2 lin. latis albidis concoloribus. Filamenta zqualia filiformia 3 lin. longa. Stylus gracilis 2 lin. longus. Hispania, Lusitania, Mauritania. Var. 8, CONCINNUM, Gawl. Bot. Mag. t. 953.—O. concinnum, Salisb. Prodr. 240.—Cathissa concinna, Salish. Gen. 34.—O. roccense, Link in Schrad. Journ. 1799, iv. 320.— Robustior elatior, foliis 2 raro 3-5 interdum subpedalibus, racemo 8-20-floro. Hispania, Bourgeau, 2543! Lusitania, Welwitsch ! 68. O. rENELLUM, Jacq. Ic. t. 427 ; Schultes fil. Syst. vii. 526; Kunth, Enum. iv. 358. Bulbus globosus brunneus 6-9 lin. crassus. Folia 4-5 carnoso-herbacea glabra linearia 8-10 poll. longa, 13-2 lin. lata facie canaliculata. Scapus gracilis subpedalis viridi-fuseus. Race- mus sublaxe 6-12-florus, expansus lanceolatus 3-4 poll. longus. Pe- dicelli ascendentes, inferiores 6-8 lin. longi. Bracteæ lanceolate acuminate pedicellis subzequilongz. Perianthium 6-3 lin. longum inodorum segmentis lanceolatis acutis 13-2 lin. latis lacteis dorso ob- scure virentibus. Filamenta 23-3 lin. longa subequalia deorsum leviter applanata. Stylus 1} lin. longus ovario lanceolato duplo bre- vior. Cap. B. Spei, Burchell, 7216 ! 69. O. LATIFOLIUM, Linn. Sp. Plant. 440; Mill. Dict. no. 3; Jacq. Icones, t. 424; Gawl. Bot. Mag. t. 876; Bot. Heg. t. 1978; Kunth, Enum. iv. 354.—Eustachys latifolia, Salisb. Gen. 33.—O. arcuatum, Stev. Act. Mosq. vii. 75; Kunth, Enum. iv. 355; Led. Flor. Ross. iv. 158. _Bulbus globosus 13-2 poll. crassus. Folia 5-6 carnoso- herbacea lorata glabra flaccida viridia 12-15 poll. longa, 6-12 lin. vel in planta culta 13-2 poll. lata. Scapus validus erectus 1-2-pedalis. Racemus 50-100-florus vel ultra, expansus 12-18 poll. longus, 3-4 poll. latus. Pedicelli 13-2 poll. longi diutine arcuato-ascendentes. Bractez lanceolatz acuminate 6-9 lin. longe. Perianthium 5-6 lin. longum segmentis lacteis concoloribus oblongo-oblanceolatis subobtu- sis 11-2 lin. latis. Filamenta 13-2 lin. longa subequalia lanceolata acuminata. Stylus 1 lin. longus. Tauria, Steven! Caucasus, Ho- henacker! Kurdistan, Brant! 70. O. LAcTEUM, Jacq. Ic. t. 434 ; Andr. Bot. Rep.t. 274; Bot. Mag. t. 1134; Red. Lil.t. 418; Lodd. Bot. Cab. t. 1159; Kunth, Enum. iv. 354, non Villars.—Aspasia lactea, Salisb. Gen. 34. Bulbus sub- globosus albidus 1-13 poll. crassus. Folia 9-10 carnoso-herbacea ascendentia rosulata basin scapi amplectentia Iciata 8-12 poll. longa, 1-] poll. lata margine subtiliter ciliata. Scapus validus erectus 1-2- pedalis. Racemus dense 20-50-florus vel ultra expansus semipedalis vel ultra longus 2-23 poll. latus. Pedicelli erecto- patentes, infimi 9- 12 lin. longi. Bractez lanceolate 6-9 lin. longe. Perianthium 7-9 lin. longum, inodorum, segmentis oblongis subobtusis 3-4 lin. latis LINN. JOURN.— BOTANY, VOL. XIII. X 284 MR. J. Œ. BAKER ON SCILLEJE AND CHLOROGALEZ. lacteis concoloribus, ima basi virescentibus. Filamenta 23-3 lin. longa subzequalia vix applanata. Stylus 13 lin. longus. Cap. B. Spei, Ecklon, 569! Whitehead. Var. 8, conicum, Baker.—O. conicum, Jacq. Ic. t. 428; Bot. Mag. t. 3538; Kunth, Enum. iv. 354. Gracilior, foliis bracteis et perianthii segmentis angustioribus, racemo laxiore. Cap. B. Spei. Subgenus VII. LEDEBOURIOPSIS, Baker. Species anomale ab omnibus reliquis recedentes filamentis di- stincte perigynis perianthii segmentis flore expanso supra basin faleatis. Stirps inter Ornithogalum, Scillam et Urgineam am- bigua. 71. O.? ANoMALUM, Baker in Saund. Ref. Bot. t. 178. Bulbus glo- bosus viridis 2 poll. crassus. Folia sepissime solitaria, raro geminata, perfecte teretia 12-2 pedes longa, 2-23 lin. crassa carnoso-herbacea flaccida glaucescentia ad apicem sensim attenuata. Scapus gracilis teres pallide glauco-viridis sesquipedalis. Racemus sublaxe 30-40- florus, expansus 6-9 poll. longus vix 1 poll. latus. Bractez minutis- sime deltoidez. Pedicelli erecto-patentes stricti apice articulati, in- fimi 2-3 lin. longi. Perianthium 3 lin. longum segmentis basi con- natis lanceolatis obtusis luteis dorso viridi-vittatis flore expanso supra basin falcatis. Filamenta perigyna subzqualia vix applanata segmen- tis duplo breviora. Stylus ovario zquilongus. Cap. B. Spei (v. v. in hort. Saund. Cooper legit !), MacOwan, 1853! 72. O.? Coorznr, Baker. Folia 2-linearia planiuscula glabra crasse carnoso-coriacea acuta basin scapi amplectentia 7-8 poll. longa, 3-4 lin. lata marginibus leviter incrassata. Scapus 1-13-pedalis. Race- mus laxe 40-100-florus expansus ad pedem longitudinis attingens vix l poll. latus. Pedicelli ascendentes vel cernui 3—4 lin. longi. Bracteze lanceolatæ vel deltoideæ albidæ membranaceæ 1-13 lin. longæ. Pe- rianthium campanulatum 3 lin. longum segmentis lanceolatis subob- tusis luteis dorso viridi-vittatis uninervatis flore expanso supra basin faleatis. Filamenta perigyna subzqualia basi linearia. Stylus 13 lin. longus ovario zquilongus. Cap. B. Spei, Cooper cult. hort. Kew. anno 1864! ad oram orientalem, Mrs. Barber in herb. T. C. D.! 73. O.? caPrrATUM, Hook. fil. Bot. Mag. t. 5388. Bulbus globosus 3-2 poll. erassus. Folia post anthesin matura plura carnoso- herbacea lineari-lorata demum ad pedem longitudinis attingentia 4-6 lin. lata. Scapi 1-2 validi erecti teretes 6-9-pollicares. Racemi dense 20-30-flori capitati globosi 15-18 lin. lati. Pedicelli 3-6 lin. longi. Bractez minutz saccate ovato-deltoideze membranacez albide. Pe- rianthium 2-23 lin. longum segmentis marcescentibus lanceolatis sub- MR. J. G. BAKER ON SCILLEA AND CHLOROGALE E. 285 obtusis facie albidis dorso uninervatis purpureo-vittatis flore expanso supra basin faleatis. Filamenta clavato-filiformia subzqualia seg- mentis subduplo breviora. Stylus ovario sequilongus. Kaffraria britannica, Cooper, 208 ! 34. ALBUCA, Linn. Linn. Gen. no. 416; Dryand. Act. Holm. 1784, 296; Endlich. Gen. 1133; Kunth, Enum. iv. 372; Harvey, Cap. Gen. 2nd edit. 397; Salisb. Gen. 36.—Falconera, Pallastema, et Branciona, Salisb. Gen. 36. Perianthium basi 6-partitum segmentis 3 exterioribus oblongis obtusis apice cucullatis patulis late viridi-vittatis multinervatis, 3 interioribus paulo brevioribus latioribus diutine calyptratim conniventibus apice tuberculo flavo glanduloso cristatis. Sta- mina 6 hypogyna segmentis interioribus «quilonga antheris ob- longis versatilibus alternis (segmentis exterioribus oppositis) abortivis vel minoribus. Ovarium sessile oblongum ovulis in loculo numerosis; stylus ovario subeequilongus triquetro-clavatus inverse pyramidalis vel in § Pallastema filiformis ovario 2-3-plo longior; stigmata 3 sepissime deltoidea papillosa. Capsula magna chartacea loculicide trivalvis, seminibus in loculo pluri- bus (sepe ultra 20) oblongis confertis. Testa membranacea nigra. Herbe bulbose foliis synanthiis carnoso-herbaceis loratis vel lineari-filiformibus, floribus albo-viridibus vel flavo-viridibus spectosis. $ Evarsuca. Stamina exteriora castrata. Stylus brevis obconicus. Albiflora foliis loratis .... +e eee ee eee eee eee 1. altissima. Flaviflore foliis glabris. Bulbus squamis truncatis coronatus. Folia lorata 9-12 lin. lata ..............-. 2. major. Mone linearia 4-6 lin. lata |... eere 3. minor. Bulbus fibris solutis coronatus .............. 4. flaccida. Viridiflora foliis glanduloso-villosis ............ 5. viridiflora. $$ FarcowEna (Salisb. extens.). Stamina omnia fertilia. Stylus brevis obconicus. Racemus deltoideus floribus diutine erectis. Albiflorz. Eolia lorata 6-12 lin. lata ............-. ys 6. fastigiata. Folia linearia 4-6 lin. lata................ 7. caudata. x2 286 .MR. J. G. BAKER ON SCILLEA AND CHLOROGALEZ. Flaviflorz. Folia lorata planiuscula. Bulbus squamis integris coronatus ...... 8. aurea. Bulbus setis fibrosis coronatus .......... 9. setosa. O 10. tenuifolia. Racemus lanceolatus floribus primum cernuis. Albiflora foliis lineari-subulatis viscosis ...... 11. viscosa. Flaviflorz. Eolia linearia glabra ............- ere 12 fragrans. Folia subteretia glandulosa .............. 13. spiralis. $$$ PanrasrEMA (Salisb.). Stamina omnia fertilia. Stylus fili- Jormis elongatus. Perianthii segmenta 2 -21 lin. lata. Ferianthium 8-9 lin. longum |..-.........- 14. abyssinica. Perianthium 10-14 lin. longum ............ 15. angolensis. Perianthii segmenta 4-4) lin. lata ............ 16. Bainesi. l. A. ALTISSIMA, Dryand. Act. Holm. 1784, 292; Jacq. lc. t. 63; Kunth, Enum. iv. 373.—A. alba, Lam. Encyc. i. 76.—A. cornuta, Red. Lil. t. 70; Kunth,Enum. iv. 373. Bulbus depresso-globosus 2 poll. vel ultra erassus. Folia 6-9 lorata planiuscula glauca carnoso- herbacea glabra 11-2 pedes longa medio 1-2 poll. lata. Scapus teres validus glaucus 13-2-pedalis. Racemus laxe 12-30-florus ex- pansus 12-18 poll. longus, 3-4 poll.latus. Pedicelli infimi 2-23 poll. longi, floriferi apice cernui. Bractez lanceolate 9-12 lin. longs. Perianthium inodorum 9-12 lin. longum seginentis exterioribus ob- longo-lanceolatis 23-3 lin. latis albis late viridi-vittatis. Filamenta alterna sterilia, fertilia 5-6 lin. longa. Stylus 3-4 lin. longus. Cap- sula 8 9 lin. longa seminibus in loeulo 20-30. Cap. B. Spei, Hort. Lee, 1782! Hort. Jacquin! Burchell, 6197! Hort. Luxembourg, 1820! ete. 2. A. MAJOR, Linn. Sp. Plant. 438 ; Jacq. Ic. t. 443; Bot. Mag. t. 804; Red. Lil. t. 69; Bot. Cab. t. 1191; Kunth, Enum. iv. 3/4. Bulbus globosus 13-2 poll. crassus proliferus. Folia 6-10 lorata pla- niuscula 1-11-pedalia, 9-12 lin. lata glabra earnoso-herbacea. Scapus teres validus 1—2-pedalis. Racemus laxe 6-15-florus, expansus 6- 15 poll. longus, 3-4 poll. latus. Pedicelli infimi 2-3 poll. longi, primum apice cernui demum ascendentes. Bractez lanceolate 1-14 poll. longe. Perianthium inodorum 9-12 lin. longum segmentis ex- terioribus oblongo-lanceolatis 21—3 lin. latis flavis late viridi-carinatis. Filamenta alba filiformia, alterna sterilia. Stylus 34 lin. longus. . Capsula 9-10 lin. longa seminibus in loculo 20-30. Cap. B. Spei, MR. J. G. BAKER ON SCILLEZE AND CHLOROGALER. 287 (Herb. Ait. ex hort. Kew. 1779!), Zeyher, 4190! Harvey, 887 ! Alex- ander ! etc. - A. MINOR, Linn. Sp. Plant. 438; Bot. Mag. t. 720; Red. Lil.t. 21; Kunth, Enum. iv. 374; Ref. Bot. t. 9339. Bulbus ovoideus 12-15 lin. crassus proliferus squamis paucis latis coronata. Folia 5-6 linearia viridia glabra carnoso-herbacea 6-12 poll. longa medio 4-6 lin. lata facie deorsum concava. Scapus firmus teres 12-18-pollicaris. Race- mus laxe 6-12-florus, expansus 6-9 poll. longus, 3-4 poll. latus. Pe- dicelli subpatentes apice cernui, inferiores 15-18 lin. longi. Bracteæ lanceolate 4-6 lin. longe. Perianthium inodorum 9-12 lin. longum segmentis exterioribus oblongo-lanceolatis 21—3 lin. latis flavis late viridi-carinatis, interioribus paulo brevioribus, glandula reniformi pa- pilosa apice instructis. Filamenta alba filiformia, alterna sterilia. Stylus ovario «equilongus. Capsula ovoidea 8-9 lin. longa seminibus in loculo 20. Cap. B. Spei (Herb. Aiton. ex hort. Kew. 1784!) Ecklon et Zeyher, 98! Wright, 222! Macgillivray, 477 ! (v. v. in hort. Kew.). Ab A. minore, A. coarctata, Dry. Act. Holm. 1784, p. 293, Kunth, Enum. iv. 375, ex descriptione, solum differt bracteis pedicellis ultrauncialibus paulo brevioribus. Non vidi. - A. FLACCIDA, Jacq. Ic. t. 444; Kunth, Enum. iv. 374; Ref. Bot. t. 334. Bulbus globosus 6-9 lin. crassus fibris pluribus setosis coro- nata. Folia 5-6-linearia pallide viridia glabra carnoso-herbacea facie canaliculata flaccido-decurvata, 6-12 poll. longa, 3-6 lin. lata. Scapus glaber flexuosus 6-12 poll. longus. Racemus laxe 6-8-florus, expansus 6-8 poll. longus subsecundus. Pedicelli 12-18 lin. longi, floriferi patentes apice cernui, fructiferi ascendentes. Bractez lan- ceolate, infime 6-9 lin. longe. Perianthium inodorum 9-10 lin. longum segmentis exterioribus flavis late viridi-carinatis. Filamenta alterna sterilia. Stylus ovario xquilongus. Capsula ovoidea 7-8 lin. longa, seminibus in loculo 20 vel ultra. Cap. B. Spei, Zeyher, 1714! Harvey, 812! Cooper! ad sinum Delagoa, Forbes! . A. vIRIDIFLORA, Jacq. Ic. t. 446; Bot. Mag. t. 1656; Kunth, Enum. iv. 374. Bulbus globosus 1-13 poll crassus. Folia 6-9-line- aria subteretia erectiuseula subpedalia 2-3 lin. lata dense persistenter villoso-glandulosa. Scapus subpedalis deorsum villosa. Racemus laxe 6-10-forus. Pedicelli infimi 13-2 poll. longi. Bracaew lanceo- late 9-12 lin. longz. Perianthium inodorum pendulum ¿-1 poll. longum segmentis exterioribus oblongo-lanceolatis omnino viridibus, intimis sursum flavo tinctis. Filamenta alterna sterilia. Stylus ovario zquilongus. Cap. B. Spei. . A. FASTIGIATA, Dry. Act. Holm. 1784, 296 ; Bot. Rep. t. 450; Bot. Reg. t. 277; Red. Lil. t. 474; Ref. Bot. t. 44.— Falconera fastigi- ata, Salisb. Gen. 36. Bulbus 2-3 poll. crassus, globosus, apice squa- 288 7. MR. J. G. BAKER ON SCILLEÆ AND CHLOROGALE.F. mosus. Folia 5-6, lorata planiuscula carnoso-herbacea glabra saturate viridia 12-18 poll. longa, 6-12 lin. lata marginibus obscure ciliatis. Seapus validus subpedalis. Racemus subdeltoideus 12-30-florus, ex- pansus 12-18 poll. longus. Pedicelli ascendentes, infimi 4-6 poll. longi, superiores multo breviores. Bractez lanceolata 9-12 lin. longze. Perianthium erectum inodorum 9-12 lin. longum segmentis exteriori- bus oblongis 3-4 lin. latis albis late viridi-vittatis, interioribus di- stinete brevioribus. Stamina omnia fertilia antheris alternis minori- bus. Stylus 3-4 lin. longus. Cap. B. Spei herb. Aiton. ex hort. Kew. 1779! Burchell! Cooper! etc. (v. v. in hort. Saunders). A. CAUDATA, Jacq. Ic. t. 442; Kunth, Enum. iv. 375; Ref. Bot. t. 45. Bulbus globosus 2-3 poll. erassus viridis apice squamosus. Folia 4-6 linearia glabra carnoso-herbacea viridia subpedalia 4-6 lin. lata facie canaliculata. Scapus 1-13-pedalis. Racemus subdeltoideus 6-15-florus, expansus 6-9 poll.longus. Pedicelli diutine ascendentes, infimi 3-4 poll longi. Bractez lanceolate 9-12 lin.longz. Perian- thium erectum inodorum 9-10 lin. longum segmentis exterioribus albis late viridi-vittatis. Stamina omnia fertilia. Stylus 3-4 lin. longus. Capsula ovoidea 8-9 lin. longa seminibus in loculo 20 vel ultra. Cap. B. Spei, Burke! (v. v. in hort. Saunders, legit Cooper). . A. AUREA, Jacq. Ic. t. 441; Kunth, Enum. iv. 376. Bulbus glo- bosus 15-18 lin. crassus squamis haud solutis coronatus. Folia 6-9 lorata glabra carnoso-herbacea planiuscula sesquipedalia vel ultra, medio 6-12 lin. lata. Scapus validus pedalis vel ultra. Racemi sub- deltoidei 10-30-flori, expansi 12-18 poll. longi. Pedicelli ascendentes, infimi 3-6 poll. longi. Bractez lanceolate 12-18 lin. longz. Peri- anthiam erectum inodorum 12-15 lin. longum segmentis exterioribus oblongis obtusis flavis late viridi-vittatis. Stamina omnia fertilia. Stylus 5-6 lin. longus. Cap. B. Spei, hort. Burchell, 1817! hort. Saunders, legit Cooper ! . A. serosa, Jacq. Ic. t. 440; Bot. Mag. t. 1481; Kunth, Enum. iv. 375.—Branciona setosa, Salisb. Gen. 36. Bulbus emersus 2-3 poll. crassus fibris setosis pluribus coronatus. Folia 6-9 lorata carnoso- herbacea glabra 12-18 poll. longa medio 6-9 lin: lata. Scapus strictus validus 12-18-pollicaris. Racemi laxe 10-30-flori, pedicellis infimis 2-3 poll. longis. Bractez lanceolate 9-12 lin. longe. Perianthium erectum leviter odorum 10-12 lin. longum segmentis flavis late viridi- vittatis. Filamenta omnia fertilia. Stylus 4-5 lin. longus. Cap. B. Spei, Burchell, hort. Saunders ! hort. Kew. ! 10, A. TENUIFOLIA, Baker in Saund. Ref. Bot. t. 335. Bulbus ovoi- deus 6-8 lin. erassus membranaceo-tunicatus. Folia 6-9 filiformia 5-6 poll. longa vix ultra 4 lin. crassa facie leviter canaliculata saturate viridia carnoso- herbacea inconspicue glandulosa. Scapus gracilis 4-6- MR. J. G. BAKER ON SCILLEJE AND CHLOROGALER. 289 pollicaris. Flores 3-4 laxe corymbosi. Pedicelli erecto-patentes, in- fimi 2-3 poll. longi. Bractez lanceolate 6-8 lin. longa. Perian- thium erectum 8-9 lin. longum flavum. Stamina omnia fertilia. Stylus ovario equilongus. Cap. B. Spei, hort. Saunders, legit Mac- Owan, v. v. 11. A. viscosa, Linn. Suppl. 126; Dryand. Act. Holm. 1784, 297 Jacq. Ic. t. 445; Kunth, Enum. iv. 377.—Falconera viscosa, Salisb. Gen. 36. Bulbus ovoideus 1-14 poll. crassus apice squamosus. Folia plura lineari-subulata erectiuscula 6-12 poll. longa, 1-2 lin. lata facie profunde canaliculata dense persistenter viscosa. Scapus teres visco- sus subpedalis. Racemus laxe 6-12-florus. Pedicelli infimi 1-2 poll. longi viscosi apice cernui. Bractez lanceolate 6-12 lin. longe. Pe- rianthium cernuum inodorum 8-10 lin. longum, segmentis exteriori- bus oblongis albidis late viridi-carinatis intimis flavo-tinctis. Stamina omnia fertilia. Stylus ovario «equilongus. Cap. B. Spei, hort. Fo- thergill, 1781! Burchell, 6222! 12. A. FRAGRANS, Jacq. Hort. Schoen. t. 84; Kunth, Enum. iv. 376. Bulbus globosus 2 poll. erassus proliferus. Folia 6-8 linearia glabra carnoso-herbacea ]i-2 pedes longa deorsum 6-9 lin. lata extrorsum planiuseula. Scapus teres sesquipedalis. Racemus sublaxe 15-20- florus, expansus ad pedem longus, 3 poll. latus. Pedicelli infimi 12- 15 lin. longi, floriferi apice cernui, fructiferi ascendentes. Bractex lan- ceolatz 3-4 lin. longze. Perianthium fragrantissimum, 8-9 lin. lon- gum, segmentis exterioribus flavis late viridi-vittatis. Stamina omnia fertilia. Stylus ovario equilongus. Cap. B. Spei. 13. A. SPIRALIS, Linn. Suppl. 196; "Thunb. Act. Holm. 1806, 54. t. 2. fig. 2; Jacq. Ic. t. 439; Kunth, Enum. iv. 377.—Falconera spiralis, Salisb. Gen. 36. Bulbus ovoideus 8-9 lin. crassus, membranaceo- tunicatus. Folia 10-12 subteretia 6-9 liu. longa antice canaliculata carnoso-herbacea dense glanduloso-pubescentia apice circinata. Sca- pus gracilis 6-8-pollicaris. Racemus laxe 4-6-florus. Pedicelli in- fimi 8-12 lin. longi, floriferi apice cernui. Bracteæ lanceolatze 8-9 lin. longz. Perianthium cernuum inodorum 7-9 lin. longum flavum. Stamina omnia fertilia. Cap. B. Spei, Harvey! etc. 14. A. ABYSSINICA, Dryand. Ait. Holm. 1784, 297 ; Jacq. Ic. t.64; Red. Lil. t. 195 ; Kunth, Enum, iv. 376.—Pallastema abyssinica, Salisb. Gen. 36. Bulbus globosus 13-2 poll. crassus. Folia 6-8 carnoso- herbacea glabra 2-3 pedes longa 6-12 lin. lata lineari-lorata ad apicem sensim angustata. Scapus validus strictus teres crassitie calami 2-4- pedalis. Racemus subdense 30-50-florus vel ultra, expansus 6-12 poll. longus 2 poll. latus. Pedicelli ascendentes 3-6 lin. longi. Bractez lineari-subulate 1-2 poll. long:e ante florescentiam conspicu:e. Perianthium inodorum 8-9 lin. longum, segmentis lanceolatis luteis 290 MR. J. G. BAKER ON SCILLEX AND CHLOROGALES. dorso late viridibus 8-9-nervatis 2-23 lin. latis. Filamenta 6-7 lin. longa omnia fertilia, basi lanceolata, alterna latiora. Stylus filiformis 6-7 lin. longus. Abyssinia, Plowden ! Schimper, 756! etc. 15. A. ANGOLENSIS, Welw. in Saund. Ref. Bot. t. 336. Bulbus maxi- mus. Folia 6-8 carnoso-herbacea glabra pallide viridia 14-2 pedes longa lineari-lorata 1 poll. lata ad apicem sensim angustata. Scapus validus teres strictus 4-6-pedalis. Racemus cylindricus densiflorus 1-14 pedem longus expansus 23-3 poll. latus. Pedicelli 3-6 lin. longi. Bracteze lineares vel lanceolate-cuspidate floribus expansis eequilonge velsuperantes. Perianthium 10-14 lin. longum segmentis luteis late viridi-vittatis multinervatis. Filamenta omnia fertilia, basi lanceo- lata, dimidio superiore filiformia. Stylus filiformis 6-7 lin. longus. Angola, Welwitsch ! (v. v. in hort. Saund.). 16. A. BarwEstr, Baker. Folia glabra carnoso-herbacea 13-2 pedes longa lineari-lorata 8-12 lin. lata. Scapus * * * Racemus ex- pansus 15-18 poll. longus, 23-3 poll. latus, rachi valida sulcata deor- sum 5-6 lin. crassa, subdense 40-60-florus. Pedicelli infimi 12-15 lin. longi demum stricti erecto-patentes. Bractew lineares 1 poll. longz ante florescentiam conspicuz. Perianthium 12-15 lin. longum segmentis oblongis 4-44 lin. latis flavis dorso late viridibus distincte 8-12-nervatis. Filamenta 9-10 lin. longa basi lanceolata. Stylus filiformis 8-9 lin. longus. Africa austro-tropicalis centralis prope Koobie, Baines ! Species dubia. A. PARVIFLORA, Don, Dietr. Gart. Lex. edit. 2,1. 237; Kunth, Enum. iv. 377. Bulbus subrotundus. Seapus erectus 6-10-pollicaris. Flores parvi flavescenti-virides subsessiles. Bractew parvae anguste acuminate. Cap. B. Spei (non vidi). Speeies exclus. A.? GARDENI, Hook. Bot. Mag. t. 4842 (planta chinensis, non natal- ensis, ut dicitur) est verisimiliter Theropogonis species nova; sed fructum non vidi. A. physodes, fugaz, filifolia, et exuviata, Gawl., sunt Urginee species. 35. Bowiea, Harv. Harv. in Bot. Mag. sub t. 5619; Cap. Gen. edit. 2, 401, non Haworth in Phil. Mag. Oct. 1824, 299. Perianthium basi connatum segmentis 6 æqualibus marcescentibus anguste ligulatis viridibus dorso distincte 1-nervatis flore ex- panso rotatim patentibus velleviter reflexis. Stamiaa 6 unise- MR. J. G. BAKER ON SCILLEA AND CHLOROGALEZ. 291 riata perigyna filamentis filiformibus segmentis subduplo brevi- oribus, antheris oblongis versatilibus. Ovarium sessile ovoideum triloculare ovulis in loculo pluribus; stylus rectus brevis fili- formis; stigma capitatum. Capsula ovoidea membranacea locu- licide trivalvis, seminibus in loculo 5-6 parvis oblongis com- pressis. Testa nitida nigra membranacea. 1. B. voLugriLis, Harv. Bot. Mag. t. 5619. Bulbus viridis squamosus globosus 4-5 poll. crassus. Folia 2 erecta linearia parva carnoso- herbacea cito evanescentia. Caulis more Asparagi vage volubilis copiose irregulariter ramosus ramulis floriferis racemosis et sterilibus copiosis viridibus carnosis dichotome furcatis. Pedicelli 6-18 lin. longi haud articulati. Bracteæ minutæ lanceolatæ membranaceæ basi calcaratæ. Perianthium 3-4 lin. longum segmentis 3 lin. latis. Cap- sula 4-6 lin. longa. Natalia, Cooper! M‘Ken! Buchanan, ete. Kaf- fraria, Hutton! Barber, 892! (v. v. in hort. Kew. et Saundersii). 36. CHLOROGALUM, Kunth. Kunth, Enum. iv. 681; Torrey, Bot. Mex. Bound. 218, t. 60.— Ornithogalum, $ Chlorogalum, Lindl. Bot. Mag. xxviii. sub t. 28.—A ntherici, Phalangii, et Scille sp., duct. Perianthium basi 6-partitum segmentis subzqualibns lineari- ligulatis albidis dorso vittatis confertim 3-6-nervatis flore ex- panso falcatis. Stamina 6 e basi segmentorum subuniseriata filamentis filiformibus segmentis brevioribus ; antheris lineari- oblongis versatilibus. Ovarium sessile globosum triloculare loculis biovulatis; stylus filiformis rectus vel leviter curvatus ; stigma tricuspidatum. Capsula turbinata parva sessilis loculi- cide trivalvis, seminibus in loculo 1-2 obovoideo-triquetris. Testa nigra membranacea. Herbe bulbose foliis duris synanthiis caule bracteato floribus laxe racemoso-paniculatis teneris albidis anthericoideis pedicellis apice articulatis. Bulbus apice fibris duris coronatus........ l. pomeridianum. Bulbus apice fibris nullis coronatus ...... 2. angustifolium. 1. C. POMERIDIANUM, Kunth, Enum. iv. 682; Torrey, Bot. Mea. Bound. 218, t. 60.—Anthericum pomeridianum, Gaw/. Bot. Reg. t. 564.— Scilla pomeridiana, DC. in Red. Lil. t. 421.—Phalangium po- - meridianum, D. Don in Sweet, Flow. Gard. ii. t. 381.—C. divarica- tum, Kunth, loc. cit.—Ornithogalum divaricatum, Lindl. Bot. Reg. xxvii. t. 28. Bulbus elongato-ovoideus, tunicis extimis apice in fibras copiosas setosas persistentes solutis. Folia radicalia plura 29 € 2. C. ANGUSTIFOLIUM, Kellogg, Proc. Calif. Acad. ii. 104, t. 30. 2 á MR. J. G. BAKER ON SCILLES AND CHLOROGALEZ. ineari-ligulata 12-18 poll. longa medio 8-10 lin. lata glabra multi- nervata marginibus valde undulatis. Scapus (panicula inclusa) 2-3- pedalis. Racemi 6-18 poll. longi, 12-30-flori, expansi 13-2 poll. lati. Pedicelli ascendentes 6-9 lin. longi, raro geminati. Bractez lanceo- late 1-14 lin. long. Perianthium 8-9 lin. longum, segmentis 1-14 lin. latis dorso 4-6-nervatis flore expanso revolutis. California, Fremont! W. Lobb! etc. Ubi vocatur ** Soap-plant ” ob succum copi- osum saponaceum bulborum. Bulbus breviter ovoideus, tunicis apice nullo modo setosis. Folia minora angustiora 4-8 poll. longa medio 3-4 lin. lata, margine vix undulata. Scapus, panicula inclusa, 2-3-pedalis. —Panicule rami inulti erecto-patentes. ^ Pedicelli semper solitarii 1-3 lin. longi. Bractex lanceolate 2-3 lin. longs. Perianthium 6-7 lin. longum segmentis dorso trinervatis flore expanso haud revolutis. California, Dr. Kellogg (non vidi). 37. NOLINA, Rich. Rich. in Michx. Flor. Amer. i. 208; Endlich. Gen. no. 1064; Kunth, Enum. iv. 656; Chap. Flor. S. U. S. 483.—Nolinea, Pers. Ench. i. 399.—Phalangii sp., Lam. Perianthium 6-partitum segmentis 6 equalibus oblongo-lanceolatis albidis marcescentibus dorso l-nervatis flore expanso reflexis. Stamina 6 perigyna uniseriata filamentis filiformibus segmentis brevioribus antheris oblongis versatilibus. Ovarium sessile ovoideum triloculare ovulis in loculo geminis collateralibus ; stylus subnullus; stigma peltatum tricuspidatum. Capsula parva membranacea ovoidea septicide trivalvis, seminibus in loeulo solitariis parvis obovoideo-oblongis haud compressis, loculis 1-2 szpe abortivis. 1. N. GEorGinaA, Rich. in Miche. Flora, 1. 208; Pursh, Flora, i. 240, ii. 746; Kunth, Enum. iv. 656; Chapman, Flora, 483.— Phalangium virgatum, Lam. Encyc. v. 246. Bulbus magnus tunicatus. Folia ra- dicalia 1-2 pedes longa 13-2 lin. lata anguste ligulata coriacea di- stincte paucinervata margine denticulata. Caulis copiose bracteatus, panicula inclusa, 2-3-pedalis. Panicule rami multi erecto-patentes floribus copiosis superne subdensis inferne laxioribus. Pedicelli erecto-patentes apice articulati 1-3 lin. longi, inferiores geminati. Bractez lanceolate persistentes 1-13 lin. longe. Perianthium 14 lin. longum. Georgia, Dr. Boykin! etc. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 293 New Species of Musci collected in Ceylon by Dr. Thwaites. Described by W. Mirren, A.L.S. [ Read June 20, 1872.] SINCE the “ Enumeration of the Mosses of the East Indies ” in the ‘Journal of the Proceedings of the Linnean Society’ in 1859, con- derable additions bave been made to the flora of Ceylon through the species described by Dr. Carl Miller (Linnea, Band 36) from the collections made by Dr. Neitner; and by the species described in the following pages Ceylon appears to have a rich moss-flora, approaching to that of the Indian archipelago by the presence of species of Garovaglia, numerous forms of Syrrhopodon, and Calym- peres, of those species of Macromitrium which have curved cells in the lower portion of the leaf, and of species of Chetomitrium. Among the species recently added to the flora of Ceylon by Dr. Thwaites there are some whose presence is very remarkable; of these, one is a Leptostomum, a genus of few species before con- fined to Australia, Tasmania, New Zealand, and the western coast of South America, the new species differing only in small parti- culars from those previously known. ‘The presence of this austral form in Ceylon seems analogous to that of the moss which was enumerated in the * Musci Indici" as PAyllogonium elegans, and supposed to be identical with the species so named and figured in that part of the Antarctic flora containing the flora of New Zea- land, and for which Reichardt has proposed the genus Orthorhyn- chium, “ perist. simpl. dent. ext. 16 «equidistantes sepe conflu- entes cartaliginei. Calyptra mitriformis campanulata multoties laciniata. Columella exserta." And C. Müller, Linn. Band xxxvi. p. 28 (from whom I have taken the above characters), in the same place, points out the distinctions by which the original species O. elegans differs from the Ceylon moss which he describes as O. Neitneri, the most important difference being derived from the imbrication of the terminal bud at the apex of the shoots, which is said to be open in O. elegans and closed in O. Neitneri, a distinc- tion which I have failed to verify. Two other species are also at the same time described as belonging to this genus :—O. Hampeanum from Sealer's Cove, Australia Felix, Dr. Mueller, the description of which agrees so well with Acrocladium politum (Hook. fil. et Wils.), a moss whieh Dr. Mueller has several times sent from other localities, that it may be feared it has been again mistaken 294 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. as it was by Dr. Montagne, when he described it as PAyllogonium callichroum ; indeed, so close is the resemblance of the foliage of this moss to that of the Phyllogonia, that the barren stems might well be supposed to belong to a species of that genus. The remaining spe- cies, ascribed to Orthorhynchium, is a barren moss from the Philip- pine Islands (O. pAilippinense) ; and its description offers no positive difference whereby it may be distinguished from the Neckera phyl- logonioides, Sull. United St. Exped. t. 17, which was also obtained in the same islands. Of the genns Orthorhynchium itself, it ap- pears that the chief distinction exists in the striation of the ca- lyptra, which is really like that of Orthotrichum, and covers a great portion of the capsule; whereas in thé other Phyllogonia this organ is small and smooth; in P. angustifolium it covers only the short beak of the operculum, and in this particular is similar to that of some Fissidentes, in which the calyptra, being too short to be affected by the growth of the capsule, remains unsplit. The occurrence of another species of the genus Spherothectum in Ceylon is a curious circumstance, which prompts the inference that among the Campylopus-like barren mosses of the European and Indian mountains some further species may be found with the curious inconspicuous phascoid fruit. All the descriptions of the species about to be mentioned have been drawn up by me; and for any errors or oversights in their definition I am alone respon- sible; but as many of the species had already been clearly distin- guished by Dr. Thwaites, it has been arranged that our joint names should be attached to them, which appears to me a very small tribute to the energy with which he has investigated the flora of Ceylon. A few species nearly allied to several Ceylon mosses, but hitherto undescribed, have been added from various sources. Hurstpierpoint, June 1872. SPHAROTHECIUM, Hamp. in Triana et Planchon, Fl. Nov. Granat Musci, p. 25. S. RECONDITUM, Thw. et Mitt. Czespitosum, humile; caulis superne ramulis fertilibus pluribus in comam congestis divisus; folia inferiora erecto-patentia anguste lanceolata subulato-acuminata canaliculata, margine integerrimo, levia, nervo inferne 2 folii latitudinis occupante, usque fere ad apicem a pagina distincto, cellulis basalibus rubro-fuscis inde oblongis rectangulis pel- lucidis superne in parvas ob longas obliquas obscuras densas transeunti- MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 295 bus; folia comalia perichetialia subduplo longiora, interiora basi latiora, ovalia, convolutacea, exinde in subulam canaliculatam e nervo fere omnino constitutam producta, margine versus apicem serrulata ; theca in pedunculo brevi pallido cygnicolli-flexo globoso-ovalis, oper- culo parvo acumine obliquo, peristomio dentibus pallidis brevibus fugacibus annulo lato fere obtectis; calyptra parva ad thece medium descendens, fimbriata. Hab. Ins. Ceylon, in collibus altioribus ad rivulorum margines et in de- clivibus humidis haud rarum. Dr. Thwaites. Caulis 1-3 lineas altus. Folia comalia bilinearia stramineo-viridia rigi- diuscula subnitida. Pedunculus theca duplo longior. Very much resembling S. comosum, Hampe, from the Bogo- tan Andes, but with leaves which, when flattened, are more nearly Janceolate, with laxer and more pellucid cells. Another specimen, without fruit, has been sent by Dr. Thwaites; but it does not offer sufficient characters to distinguish it as a distinct species, although it has much the appearance of being so. DicrANELLA, C. Müll. D. EDENTATA, Thw. et Mitt. Caulis fertilis brevis, innovationibus infra perichetium oriundis ramo- sus; folia laxe disposita patentia, a basi caulis sensim angustata, acuta, nervo crassiusculo fere ad apicem usque a pagina distincto carinata, canalieulata, margine angusto implano integerrimo, cellulis basi ad angulos latioribus oblongis rectangulis, reliquis elongatis an- gustis; perichetialia duplo longiora, a basi latiora ovata, subula fere omnino e nervo formata angustiore attenuata; theca in pedunculo brevi pallido, globoso-ovalis erecta gymnostoma, operculo longe ob- lique rostrato. Hab. Ins. Ceylon. Dr. Thwaites. Plantula vix lineas tres alta, cujus pedunculus bilinearis est. Folia caulina semilineam longa, perichetialia interna sesquilineam longa, omnia luteo-viridia fuscescentia nitida. Theca intense fusca ore satis parvo, operculo a basi depresso subito rostro constricto. D. iNFUsCATA, Thw. et Mitt. Monoica ; caulis humilis; folia patentia a basi sensim angustata, anguste elongata canaliculata iutegerrima, nervo angusto ad apicem usque a pagina distincto, cellulis angustis elongatis areolata, perichzetialia a basi subovata convoluta sensim attenuata; theca in pedunculo brevi- usculo recto, ovali-cylindracea erecta zequalis pallida levis, operculo subulato obliquo ; peristomii dentes breves integri. Hab. Ins. Ceylon. Dr. Thwaites. 2960 MR. W. MITTEN ON NEW SPECTES OF MUSCI FROM CEYLON. Caulis 2-3 lineas altus. Folia vix sesquilineam longa viridia. Pedun- culus 3-4 lineas longus fuseus. Similar to D. Perrottetii (Mont.) and unlike any other Indian species. D. suBANGULATA, Thw. et Mitt. Dioica, humilis; folia patentia, comalia secunda subfalcata, inferiora a basi sensim subulato-angustata integerrima, nervo lato fere usque ad apicem distincto, cellulis elongatis angustis areolata ; pericheetialia a basi obovata convolutacea in subulam elongatam fere omnino enervo constitutam producta ; theca in pedunculo brevi, sicca et madida rec- tinscula, elliptica ovalisve, erecta, subangulata, operculo subulato ob- lique curvato ; peristomium brevissimum, dentibus brevibus irregu- laribus incompletis. Hab. Ins. Ceylon. Dr. Thwaites. Caulis trilinearis. Folia inferiora lineam longa, superiora bilinearia, fusco-viridia. Pedunculus lineas duas longus pallidus. CYNONTODIUM, Hedw. C. AMCENUM, Tw. et Mitt. Monoicum ; caulis humilis ; folia sparsa, erecto-patentia, anguste lanceo- lata, sensim angustata, nervo vix a pagina distincto carinata, apice parce denticulata, cellulis inferne angustis, superne sensim abbreviatis, omnibus concoloribus; comalia longiora; perichetialia interna basi ovalia convolutacea, nervo in subulam excurrente ; theca in pedunculo pallido, oblongo-cylindracea, operculo conico acuminato obliquo; pe- ristomium parvum dentibus brevibus valde irregularibus angustissimis rubris, annulo lato circumdatis; flos masculus gemmiformis, ad basin perichzetii in ramulo brevissimo impositus. Hab. Ins. Ceylon! Newera Ellia etc. in declivibus. Dr. Thwaites. Caulis 4 lineas altus, gracillimus. Folia inferiora lineam longa, fusces- centia, superiora seu comalia bilinearia, pallide viridia, sicca parum curvata. Pedunculus 4-5 lineas altus. Theca lineam longa pallide castanea, sicca subplicata, C. stricto Americe australis simile. Dicranum, Hedw. D. DECUMBENS, Thw. et Mitt. Dioicum; caulis decumbens divisus; folia faleato-secunda, a basi subovata sensim longe angustata, canaliculata, nervo versus apicem angustum : pagina indistincto, margineque incurvo integerrima subintegerrimave, cellulis alaribus pluribus quadratis, superioribus angustis elongatis, altioribus abbreviatis ; perichetialia longiora latioraque, apicibus mar- MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 297 gine dorsoque serrulatis; theca in peduneulo elongato, oblonga, cur- vata, sicca et evacua pluries plicata, leptodermis, basi callosa, ore obliquo. : Hab. Ins. Ceylon. Dr. Thwaites. Caulis semiunciam altus. Folia bilinearia, fusco-viridia absque nitore, siccatione vix mutata. Pedunculus 9 lineas longus, ruber. Theca evacua pallide fusca, membranacea. This moss nearly resembles in size and general appearance D. virens, Wedw.; but its foliage is less curled, and the leaves are more narrow towards their points. CAMPYLOPUS, Brid. C. SUBULIFOLIUS, Thw. et Mitt. Caulis rigidus, radicellis rubrofuscis tomentosus ; folia sparsa patentia, inferne lanceolata, superne subulato-angustata, stricta, canaliculato- concava, nervo inferne 2 folii latitudinis, superne fere totam partem subulatam occupante, dorso apicem versus lamellis paucis humilibus parce denticulatis alato, margine erecto apice serrulato, cellulis alari- bus pluribus quadratis rubro-fuscis, inde viridibus angustis sensim in- ovales minutas transeuntibus; perichetialia basi ovalia convoluta, nervo in subulam angustiorem longiorem magis serrulatam educto ; theca ovalis gibba, ore obliquo, operculo rostrato, calyptra fimbriata. Hab. Yns. Ceylon. Caulis unciam altus. Folia lineas duas longa, superiora luteo-viridia, in- feriora fuscescentia, omnia apicibus concoloribus. PacrroruyrrvM, Mitt. Muse. Austr. Amer. 92. P. NITENS, Thw. et Mitt. Ciespitosum; caulis gracilis subsimplex fusco-ruber; folia erecto-patentia subsecunda, apicalia curvata secunda, sicca medio inflexa plana laxe contorta, elliptico-lanceolata, subulato-attenuata, nervo hyalino nitido usque fere ad summum apicem a pagina planiuscula distincto, mar- gine apice minute denticulata, cellulis alaribus pluribus quadratis ob- longisque flavis, superioribus ad marginem angustissimis hyalinis, in limbum latum nitidum usque ad $ folii longitudinis inferne 7 folii latitudinis oceupantem productum conflatis, interioribus minutissimis rotundis obscuris areolata. Hab. Ins. Ceylon Prov. central. in collibus altioribus, ad arborum trun- cos late cespitosum, vulgatissimum. Dr. Thwaites. Caulis subuncialis. Folia sesquilineam longa, pallide luteo-viridia, sub- glauca, nervo limboque marginali nitidis. Nearly resembling P. amene-virens, Mitt. Musc. Ind. Or. (Leucoloma), but with a much broader band of hyaline ceils on the margins of its leaves. 298 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. OCTOBLEPHARUM, Hedw. O. RADULA, Thw. et Mitt. Caulis gracilis ruber furcatim divisus; folia a basi oblonga, erecta, parum latiora, superne margine denticulis ciliata integerrimave, cellulis hya- linis areolata, exinde pateutia subsemitorta sensim angustata, nervo crasso obscuriuseulo subtrigono-tereti intus extusque setulis brevibus papilloso, apicem versus excurrente ibique denticulato percursa, et prz- terea laminam angustam pellucidam e cellula singula latam, utroque la- tere levem, foliorum partem superiorem totam constituente, limbo mar- ginali angustissimo subserrulato circumdata nec nisi lesione cernenda. Hab. Ins. Ceylon. Dr. Thwaites. Caulis unciam altus. Folia sesquilineam longa, cujus 2 erecta, reliqua patentia, albida, siccatione immutata. O. seabro, Mitt. Linn. Soc. Jl. vol. x. p. 178, simillimum, neque prater structuram partis foliorum superioris distinguendum. SYRRHOPODON, Schwegr. Suppl. ii. 110. Eusyrrhopodon, Mitt. Musc. Austr. Amer. 113. S. ALBIDUS, Thw. et Mitt. Caulis gracilis elongatus; folia sparsa tertia parte inferiore eorum longi- tudinis erecta, cauli appressa suboblonga, superne latiora, cellulis hyalinis areolata, inde in folium patens lineare planiusculum apice obtusum denticulatum producta, margine limbo angusto cartilagineo ubique circumdata, integerrima, nervo levi percursa, cellulis rotundis obscuris przcipue extus protuberantibus papillosa. Hab. Ins. Ceylon. Dr. Thwaites. Caulis 1-13 unciam altus. Folia vix bilinearia albida dilutissime viridi tincta, sicca vel humida patentia immutata. In habit this species resembles S. tristichus, Nees, but is more slender and of a paler colour. S. caspirosus, Thw. et Mitt. Dense cespitosus, mollis ; caulis elongatus furcatim ramosus ruber ; folia dense inserta erecto-patentia laxe imbricata anguste lanceolata acuta obtusiusculave planiuscula inferne canaliculata, nervo angusto pallido percursa, dorso in folii medium papillis scabro, ceteroquin levi, mar- gine limbo angusto pallido, a folii medio minute, apice evidentius ser- rulato, cellulis hyalinis ultra folii medium longitudinis ascendentibus, superioribus minutis obscuris intus extusque minute papilliferis. Hab. Ins. Ceylon, regio subtropica. Dr. Thwaites. Caulis unciam altus. Folia sesquilineam longa, inferne albida, superne pallidissime glauco-viridia, exsiccatione directione parum mutata. Larger than S. rufescens, Hook. et Grev., but with narrower leaves. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 299 S. srRICTUS, Thw. et Mitt. Caulis furcatim divisus; folia patentia, stricta, a basi suboblonga erectiora, superne margine denticulis pluribus brevibus ciliata, inde in folium lineare elongatum rigidum angustata, apicem versus angustiora sub- abrupta, iuferiora nonnulla aliquantulum infra apicem subspathu- latim dilatata, omnia margine incrassato tereti, remote denticulato et scabroso, nervo supra partem folii erectiorem intus extusque minutis- sime denticulato scabrosulo, cellulis superioribus minutis obscuris, in- ferioribus hyalinis fere totam partem erectiorem folii occupantibus : pericheetia ad apicem caulis sepe plura; folia parte inferiore latiora, longiora, sensim angustata, margine incrassato plus minus obsoleto : theca in pedunculo brevi rubro cylindracea, operculo subulato sube- quilongo; peristomii parvi dentes breves irregulares angusti; calyptra superne subscabra. Hab. Ins. Ceylon. Dr. Thwaites. Caulis semiuncia parum altior. Folia lineas tres longa, pallidissime luteo-viridi-glauca, statu sicco vel humido vix mutata. Pedunculus trilmearis. Theca fusca. Thyridium, Mitt. Muse. Austr. Amer. 114. S. PARVULUS, Thw. et Mitt. Caulis repens, ramis brevibus ramulosis in stratum humile congestis ; folia dense inserta, a basi ad insertionem cunestim angustata, erecta, subovata, cellulis anguste oblongis elongatisque hyalinis areolata, inde patentia ligulata apice lata apiculo brevissimo terminata, margine flex- uoso ineurvo, limbo pellucido supra folii basin latiore ibique dentieulis brevibus serrato, infra apicem evanescente marginata, nervo concolori percursa, cellulis superioribus minutis rotundis obscuris levibus; pericheetialia basi latiora, cxteroquin caulinis similia; theca in pedun- culo gracili rubro, ovalis, collo sensim angustato, operculo subulato, thecz longitudine; peristomii dentes aurantiaci; calyptra superne scabra. Hab. Ins. Ceylon. Dr. Thwaites. Rami circiter bilineares. Folia lineam longa, inferiora fusca, apicalia luteo-viridia, sicca incurvo-crispata. Systectum, Schimper. S. ABBREVIATUM, Thw. et Mitt. Subacaulis, simplex; folia inferiora patula incurva inferne parum latiora cellulis inferioribus oblongis rectangulis hyalinis, superne marginibus usque ad apicem inflexis canaliculata cellulis rotundatis obscuris pa- pillosis areolata, nervo flavescente in mucronem excurrente carinata, margine ob papillas crenulata; perichzetialia erectiora, duplo longiora, LINN. JOURN.—BOTANY, VOL. XIII. Y 300 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. dimidio inferiore subelliptico concavo pellucide areolato, exinde ad apicem caulinis similia; theca in pedunculo brevissimo subsessilis, globoso-ovalis, rostello brevi erecto subobliquove terminata, calyptra parva plus minus profunde fissa; flos masculus foliis internis elongatis marginibus haud inflexis. Hab. Ins. Ceylon, ad terram. Dr. Thwaites. Plantule cum foliis lineam parum superantes. Folia perichetialia cir- citer lineam longa, omnia siccitate crispata viridia. Theca ob brevitatem pedunculi fere immersa. TortuLa, Hedw. Fund. 11. 92. Helicopodon, Mitt. Musc. Austr. Amer. 142. T. coNsANGUINEA, Thw. et Mitt. Dioica, humilis ; folia a basi parum erectiora, patentia, oblonga, obtuse acuta, integerrima, nervo percurrente dorso superne scabro, cellulis superioribus parvis rotundatis, inferioribus oblongis pellucidis, partem quartam folii longitudinis occupantibus, omnibus limitibus angustis distinctis; perichetialia inferne latiora, pellucidiora, ovato-oblonga, exteroquin conformia; theca in pedunculo rubro, oblongo-cylindracea, operculo subulato; peristomium thecæ longitudine equale, bis tor- tum; flos masculus satis magnus, foliis basi dilatatis cinctus. Hab. Ins. Ceylon, ad terram. Dr. Thwaites. Caulis 3-6 lineas altus. Folia semilineam longa, pallide glauco-viridia, sicca appressa curvata. Pedunculus sexlinearis. Macromitrium, Brid. i. 306. Goniostoma, Mitt. Muse. Austr. Amer. 194. M. ANGULOSUM, Thw. et Mitt. Rami humiles, ramosi; folia patentia, incurva, ligulata, obtusa, nervo percurrente carinata, integerrima, cellulis rotundis obscuriusculis, basalibus paucissimis oblongis pallidioribus; perichztialia breviora, lanceolata, acuta; theca in pedunculo breviusculo, sicca tetragono- ovalis, per longitudinem plicata, collo sensim angustato; calyptra nuda. Hab. Yos. Ceylon. Dr. Thwaites. Rami 2-3 lineas longi. Folia lineam longa, viridia, siccitate crispata, humida apicalia sepe apicibus incurvis. Pedunculus trilinearis, cras- siusculus. Theca humida leevis, sicca ore plicis contractis clausa. This nearly resembles M. goniopodum, Mitt. Musc. Aust. Amer. 198, but has its leaves wider at their apices, and the capsule hasa longer neck. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 301 Leiostoma, Mitt. Musc. Austr. Amer. 194. M. RAMENTOSUM, Thw. et Mitt. Rami crassiusculi, ramosi ; folia patentia, elongate lanceolata, apice lati- uscula acuta, nervo in apiculum brevem excurrente carinata, integer- rima, cellulis superioribus minutis rotundis distinctis, inferioribus basin versus oblongis elongatisque angustissimis, intus parce papilli- feris, basalibus ad nervum precipue uno latere majoribus oblongis di- latatis pellucidis ; pericheetialia vix breviora, subovata acuta; theca in peduneulo breviusculo ovata, plicata, operculo acuminato rostrato ; peristomii dentes truncati, geminati; calyptra infra basin thecz de- scendens, ramentis flavis rufisve obtecta. Hab. Ins. Ceylon. Dr. Thwaites. Rami 3-1 unciam alti. Folia bilinearia, apicalia viridia, inferiora fus- cescentia, omnia subnitida, siccitate laxe crispato-contorta. Pedun- culus 4 lineas longus. Calyptra straminea, fulvescens. In size and appearance like M. sulcatum (Hook.), but with leaves narrowed above their middle. M. CONTORTUM, Thw. et Mitt. Rami abbreviati; folia densa, patentia, sicca in spiram apice acutam con- torta, subovato-lanceolata obtusa, nervo in mucronem pungentem ex- currente carinata, margine inferne uno latere recurvo, integerrima, cellulis inferioribus pellucidis fere ad folii longitudinis medium usque ascendentibus, basalibus oblongis, sensim superne abbreviatis rotundis, supra foli medium rotundatis obscuris; perichztialia «equilonga, ovata sensim acuminata; theca in peduneulo angulato, breviter ovalis, levis, sicca sub ore rotundo contracta, operculo conico rostrato; peri- stomium membrana brevissima indivisa; calyptra nuda. Hab. Ins. Ceylon. Dr. Thwaites. Rami bilineares. Folia juniora viridia, seniora fusca, longitudine lineam dimidiam parum excedentia, tenaci-nervia, difficillime a caule sepa- randa. Pedunculus quadrilinearis, ruber. Allied to M. Blumei, Nees ab E., and to M. concinnum, Mitt. Bry. Javan. t. 110, but with the nerve of the leaf more excurrent. M. hispidulum and M. seminudum, together with M. angulatum (Mitt. Linn. Soe. Journ. vol. x. p. 167) and the species named by Nees von Esenbeck JM. orthostichum, as well as a few others, form a small group which does not very closely approach any other sec- tion of the extensive genus Dacromitrium. In habit, these spe- cies resemble some of those like AL gracile, which belong to the section with curved cells in the lower part of the leaf; but the areo- lation of the leaves is very different, the narrow cells being nearly or quite absent. The following is an outline of this division :— x2 302 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. Macromitrium, sectio nov. Cometium. Rami graciles, dichotome fasciculatimve ramosi. Folia seriatim dispositi, e cellulis rotundatis areolata. Theca in pedunculo breviusculo quadrangulo scabro, brevis, sicca 4-6-plicata levisve. Calyptra ramentis inferioribus divaricatis. | * Folia perichetialia vaginula breviora. M. orthostichum, Nees ab E. M. appressifolium, sp. nov. M. hispidulum, sp. nov. M. angulatum, Mitt. ** Folia perichetialia vaginule «equalia. M. seminudum, sp. nov. *** Folia perichetialia vaginula longiora. M. minutum, sp. nov. M. ORTHOSTICHUM, Nees ab E. ; Schwaegr. Supp. t. 316. Folia dense inserta, a basi brevi erecta squarroso-patula, semitorta, sicca curvata recurvave subsquarrosa, elongate lanceolata, apice obtusiuscula, nervo in mucronem brevem excurrente carmata, margine crenulata, basin versus subrecurva, cellulis rotundatis, basalibus infimis paucis- simis oblongis; perichztialia conformia, vaginula breviora; calyptra ramentis latis copiosis, superioribus appressis, inferioribus divergen- tibus decurvisque dentatis. Hab. Java (ex herb. Nees) in mont. Megamendong et Paroerongo reg. sup. alt. 7000-10,000’, Motley. M. APPRESSIFOLIUM, Mitt. Folia approximata, a basi brevissima vix erectiore patula, recurva, subse- mitorta, sicca contorta appressa, oblongo-lanceolata obtusa, nervo in mucronem brevem excurrente, margine crenulata, cellulis omnibus rotundo-quadratis; perichetialia similia, vaginula breviora; calyptra ramentis superioribus appressis, inferioribus decurvis denticulatis.— M. orthostichum, Bryol. Javanica, t. 107. Hab. Java (ex herb. Dozy et Molkenboer). M. orthosticho simile, foliis autem siccis appressis brevioribus, ramen- tisque calyptre minus copiosis, M. HISPIDULUM, Thw. et Mitt. Rami graciles, in ezespites luteos aggregati, ramosi; folia in series quinas erectas disposita, a basi brevi erectiore recurva, patentia, oblongo-lance- olata, apice acuta obtusiusculave breviter acuminata, nervo rufescehte carinata, planiuseula, margine ob cellulas prominulas minute apice di- stinctius crenulata, cellulis ubique rotundatis obscuris papillosis areo- MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 303 lata; pericheetialia longiora, erecta, angustiora; theca in pedunculo brevi scabro, ovali-pyriformis, ore parvo sicco rotundo, operculo rostra- to; peristomium membrana brevissima truncata; calyptra ramentis crassiusculis denticulatis, superioribus appressis, inferioribus divergen- tibus, basalibus dependentibus obtecta. Hab. Ins. Ceylon, Dr. Thwaites; Central Prov. Ettangwala, alt. 3000', on trees, Mr. Beckett. Rami semiunciales uncialesve. Folia iu ramorum apicibus viridia, re- liqua ferruginea, sicca contorta, semilineam longa. Pedunculus lineam longus. M. SEMINUDUM, Thw. et Mitt. Rami ezspitosi ramosi; folia in series quinas rectas disposita, a basi brevi erectiore patentia, oblongo-ligulata, apice acuta brevissime acu- minatave, nervo concolori subexcurrente, margine minute serrulato planiuseulo, cellulis eonformibus papillosis rotundis obseuriuseulis areolata; perichetialia longiora, conformia ; theca in pedunculo brevi scabro, late ovalis, sicca ovata, ore parvo siccitate plicis quatuor clauso ; opereulum rostratum; peristomium membrana brevissima truncata ; calyptra supra medium nuda, infra ramentis recurvis dentieulatis vestita. i: Hab. Ins. Ceylon, Dr. Thwaites. Rami semiunciales, Folia $ lineam longa, superiora viridia mox ful- vescentia, sicca contorta. Pedunculus sesquilineam longus, crassius- culus. Theca satis magna, cinnamomea. Leaves longer than in M. hispidulum, papille less prominent, capsule larger. M. MINUTUM, Mitt. Folia approximata, a basi brevi erecta patentia patulave, sicca appressa contorta, elongate lanceolata, apice latiusculo, nervo fusco in acumen longiusculum denticulatum excurrente carinata, margine crenulata, apice serrulata, cellulis rotundatis, basalibus infimis paucis oblongis ; perichzetialia elongata, vaginula duplo longiora, conformia ; calyptra ramentis superioribus appressis, inferioribus decurvis. Hab. Jàva, in Mont. Megamendong, alt. 4000'-7000', Motley. M. orthosticho M. appressifolioque minus, foliis quoad latitudinem longioribus acutioribusque. Zxaopow, Hook. et Tayl. Muse. Brit. 70. - Z. PERPUSILLUS, Thw. et Mitt. Monoicus ; caulis brevissimus; folia patentia, apice subrecurva, late spa- thulata, acutiuscula, canaliculata, nervo infra apicem evanescente, margine apicis integerrimo subcrenatove ; comalia majora, obtusiora, cellulis mollibus carnosulis, superioribus rotundis obscuriusculis, infe- rioribus oblongis rectangulis hyalinis; perichzetialia tria erecta, intima 304 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. ovato-lanceolata obtusiuseula, inferne convoluta, apicibus basin thecæ ovalis brevicollis octoplicatz attingentibus ; operculum convexo-acumi- natum ; peristomium dentibus latis 8, ciliis nullis. Hab. Ins. Ceylon, ad corticem, Dr. Thwaites. Caulis linea brevior, inferne radiculis rubris parce obtectus. Folia semilineam longa, mollia, viridissima, haud papillosa. Pedunculus linea brevior, pallidus. Theca fusca, collo brevissimo. Z. HUMILIS, Thw. et Mitt. Dioieus, parvulus; folia erecto-patentia, sublanceolato-ligulata, acumi- nata, nervo lato in acumen excurrente carinata, margine integerrimo, cellulis superioribus minutis rotundis distinctis lzevibus subpellucidis, inferioribus oblongis pellucidis; perichetialia pauca, subduplo longiora, erecta; theca in pedunculo brevi gracili, cum ejus collo ovali oblonga, opereulo conico rostrato ; peristomium dentibus latis 8, ciliis brevibus angustis 8. Hab. Ins. Ceylon, ad corticem, Dr. Thwaites. Caulis 4 lineas altus. Folia inferiora fusca, superiora viridia, linea breviora, strictiuscula. Pedunculus sesquilinearis. Theea fusca, pro plantula satis longa. Flos masculus non visus. Z. BREVICILIATUS, Thw. et Mitt. Synoicus ; caulis inferne fusco-tomentosus ; folia a basi brevi angustiore erecta, cellulis paucis oblongis pellucidis areolata, patentia, ligulata, flexuosa, acuta, nervo infra apicem evanido carinata, margine apicem versus denticulato, cellulis parvis rotundatis obscuris minutissime pa- pillosis; perichztialia intima brevia, lanceolata acuta, integerrima ; theca longe pedunculata, ovalis, in collum sensim angustata, ore parvo, operculo subulato obliquo; peristomium simplex, internum, ciliis bre- vissimis 16. Hab. lus. Ceylon, Dr. Thwaites. Caulis semiunciam altus. Folia pallide viridia, lineam longa. Pedun- culus semiuncialis, gracilis, pallidus. Theca longiuscula. This resembles Z. subdenticulatus, Hampe, from the Equatorial Andes, especially in its elongate capsule. ENTOSTHODON, Schwegr. Amphoritheca, Hampe. E. PLANIFOLIUS, Thw. et Mitt. Monoicus, humilis; folia patula, obovata, planiuscula, acumine brevi semi- torto pungente, integerrima, nervo debili supra medium evanido, margine e seriebus cellularum angustissimarum tribus in limbum an- gustum callosum conflatis, cellulis amplis pellucidis; theca in pedun- culo recto, sporangio globoso, collo pyriformi breviore, operculo plano ; peristomium nullum. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 805 Hab. Ins. Ceylon, Dr. Thwaites. Caulis lineam altus. Folia sesquilineam longa, pallide viridia, limbo dilute flavo. Pedunculus quadrilinearis ruber. In its apiculate leaves this somewhat resembles E. physcomi- trioides, Mont. (Funaria); but their margins are distinctly lim- bate, and the capsule is gymnostomous. TAxronia, Hook. T. IMBRICATA, Thw. et Mitt. Monoica; caulis humilis, inferne radiculosus ; folia imbricata, obovata, obtusa, excavata, nervo in mucronem patulum recurvumve sublevem exeurrente, margine apicis denticulis paucis brevibus serrato, cellulis oval-hexagonis areolata; perichetialia conformia; vaginula pilis paucis elongatis exsertis hirta; pedunculus crassus, pallidus ; theca cylindracea, operculo convexo acuminato; peristomii dentes carnosuli, 8. Calyptra tenuimembranacea, nitida, pilis elongatis divaricatis de- pendentibusque hirsuta. Hab. Ins. Ceylon, Dr. Thwaites. Folia lineam longa, pallide luteo-viridia. LEPTOSTOMUM, Brown, Trans. Linn. Soc. x. 130. L. DENSUM, Thw. et Mitt. Caules in ezspitem densum radicellis rufis intertextum congesti; folia patentia, sicca appressa subcontorta, oblonga, obtusissima, nervo in pilum leevem excurrente carinata, margine recurvo integerrimo, cellulis rotundis densis; perichetialia conformia, pilo longiore flexuoso, apice minute denticulato; theca in pedunculo elongato flavo, oblongo- ovalis, erecta, operculo obtuso subhemispheerico. Hab. Ins. Ceylon, Dr. Thwaites. Caulis unciam altus. Folia inferiora radicellis obvelata, pallide fusca, superiora innovationum viridia. Pedunculus uncialis. Theca sesqui- linearis. DrEPANOPHYLLUM, Rich., Hook. Muse. Exot. 145. D. oPPOSITIFOLIUM, Thw. et Mitt. Caulis procumbens, gracilis, viridis, dein fuscescens, innovationibus paucis e latere oriundis, apicem versus foliis magnitudine decrescentibus mi- nusque asymmetricis ramosus, radicellis brevibus rubris corpusculisque tubulosis haustellariiformibus fuscis ex axillis foliorum egredientibus sparsim vestitus ; folia opposita, plana, ambitu suboblongo-ovata, inz- quilatera, margine superiore arcuato, limbo e cellulis angustis pluribus conflato, apice tantum dentieulis paucis eroso ibique evanescenter 306 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. marginato, margine inferiore recto subsinuatove, limbo obsoleto, apicem versus crenulato, nervo ad marginem inferiorem propinquo, apice in mucronem ad caulis apicem spectantem excurrente, cellulis parvis rotundis pellucidis, interstitiis crassis obscuris, areolata. Hab. Ins. Ceylon, Dr. Thwaites. Caulis circiter semiuncialis, latitudine cum foliis lineari. Folia pallide viridia absque nitore. Ervopium, Brid. Euerpodium, Witt. Muse. Austr. Amer. 403. E. CEYLONICUM, Thw. et Mitt. Monoicum ; caulis ramis irregularibus subpinnatim ramosus, prostratus ; folia complanata, superiora (= dorsalia) imbricata, directione divergen- tia, oblongo-orbiculata, basi ad insertionem angusta asymmetrica, mar- gine superiore arcuato, inferiore rectiusculo, integerrima, obtusissima, cellulis parvis rotundis obscuris areolata, inferiora ( — ventralia) multo minora, directione patentia, oblonga, obtusa, symmetrica; perichzetium foliis imbricatis erectis convolutaceis, apicibus obtusis parum recurvis, internis ovalibus margine papillis minutis eroso ; theca in pedunculo brevi pallido, ovali-oblonga, operculo conico acumine obliquo; calyptra infra medium thecse descendens, basi truncata, uno latere breviter fissa, superne tumida subscabra in apieulum brevem producta; flos masculus triphyllus, foliis ovatis obtusis. Hab. Ceylon, ad corticem, Dr. Thwaites. Rami cum foliis vix tertiam quartamve partem linex lati. Folia intense viridia. Ramus fructifer eum capsula completa quam linea brevior. Pedunculus theca parum brevior, pallidissimus. Compared with Æ. domingense, Brid., Schwegr. Supp. t. 267 (Anectangium), the following differences present themselves :— In E. domingense the leaves are compressed and rather densely inserted; those on the underside of the branches not obviously distinct or different in form from those of the upper, which are of an oblong-ovate outline, their apices obtuse, their base being slightly asymmetric, from one margin being more rounded than the other. In Æ. ceylonicum the upper or dorsal leaves, almost orbicular in outline, are distinctly bifarious and so inserted that each one slightly overlaps the inferior edge of the one next above it, completely covering the leaves inserted on the ventral side, which are of considerably less than half the size of the dorsal ones, so that the branches viewed with a lens on the dorsal side may be very easily overlooked as belonging to some species of Lejeunia allied to the common L. serpyllifolia. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 307 Stephanostoma. Folia compressa, subdistiche imbricata. Fructus ex apicibus ramorum oriundus. Theca sessilis, exserta, ore demum plicis inflexis, cornuta. E. BeLL11, Mitt. Monoicum; caulis, repens, ramosus; folia compressa, imbricata, patentia, dorsalia orbiculari-ovalia breviter acuminata, ventralia late ovali- ovata aeuta, cellulis parvis rotundis obscuris areolata ; ramus fructifer elongatus, foliis inter se remotiusculis ovatis acutis, superioribus pa- tentibus; theca in pedunculo brevissimo fere sessilis, oblonga, cylin- _ dracea, pallide albo-fusca, chartacea, operculo convexo rostellato clausa, levis, post operculi delapsum autem ore pallide fusco, plicis quatuor in- flexis clauso et brevissime quadricornuto, infra os omnino levis, inferne magis ventricosa ; calyptra conica acuta, basi infra os thecze paululum descendente, brevissime plurifida, uno latere autem profundius fissa ; flos masculus parvus, e foliis tribus ovatis obtusis compositus. Hab. India orientalis, ad corticem, Bell. The appearance of the specimens seems to denote that they have grown in a situation liable to inundation, or that the species is riparian in its habitats. This very curious moss creeps over bark in a prostrate manner; its foliage is of a dull green colour, and when dry is loosely ap- pressed to the stem and erecto-patent ; when wet it resembles the species of Erpodium and Aulacopilum so closely that there would be no hesitation in placing it with them; the fruit, however, differs considerably, and comes nearer to Leptocalpe, yet dif- fering greatly in the calyptra; the capsule is large, and, after the fall of the operculum, becomes almost colourless, excepting the plicate mouth, which is pale straw-colour. AULACOPILUM, Wiis. A. TUMIDULUM, Thw. et Mitt. Monoicum; folia compressa, superiora (=dorsalia) divergentia ovato-lan- ceolata asymmetrica, basi obliqua, margine superiore arcuato, inferiore rectiusculo, inferne parum incurva, apice cellula unica clongata apicu- lato, cellulis parvis rotundis obscuris minute papillosis areolata; ven- tralia minora, directione patentia, lanceolata; perichetialia interna, erecta, convolutacea, ovalia, subacuminate acuta; pedunculus brevis, pallidus; theca oblonga, ore satis magno, operculo conico acuminato ; calyptra tumida, lobis elongatis appressis, apice acuto subscabro. Hab. Ins. Ceylon, ad corticem, Dr. Thwaites. Habitu statura adspectuque A. glauco, Wils., simile, foliis autem densius insertis, dorsalibus apicibus latioribus minus acuminatis. 308 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. A. ABBREVIATUM, Mitt. Monoicum; folia compressa, dorsalia imbricata, directione patula, ob- longo-ovalia, symmetrica, obtusinscule acuta, ventralia angustiora di- rectione vix diversa, apicem versus latiuscula, lanceolata acuta, cellulis parvis rotundis obscuris areolata, papillis minutis ubique erosa ; peri- cheetialia convolutacea, ovalia acuta ; theca in pedunculo brevi pallido, brevis, poculiformis, operculo breviter conico acuminato; calyptra lobis breviuseulis appressis, apice acuto subscabro ; flos masculus gem- miforinis. Hab. India orient. ad caudices tomentosos filicis cujusdam repens, JW. Bell. A. glauco A. tumiduloque statura simillimum, foliis autem dorsalibus apices versus latis obtusiusculis, angulo obtuso terminatis, et theca et calyptra brevioribus. A. INCANUM, Mitt. Monoicum, pinnatim ramosum ; folia compressa, dorsalia patentia ovata apice in acumen angustum diaphanum producta, ventralia dimidio mi- nora ovato-lanceolata, cellulis rotundatis obscuris, papillis minutis ; perichactialia erecta, interna, convolutacea, ovata subacuminata; theca in pedunculo pallido subzquilongo, ovalis, operculo breviter conico- acuminato ; calyptra ad perichztium usque descendens, apice sub- scabra; flos masculus gemmiformis, foliis acutis. Hab. Africa australis, East London, ad corticem, beatus Rooper. This closely resembles 4. glaucum, but has the points of its leaves diaphanous. Erpodium Beccarii, C. Muller, Venturi in Nuovo Giornale Bo- tanico Italiano, fasc. i. p. 18, is clearly, from its description, an Aulacopilum, and is stated to have leaves with an elongate hyaline bair-point, by which it much differs from 4. incanum. Anomopon, Hook. A. FILIFORMIS, Thw. et Mitt. Caulis procumbens, ramis ramosis subpinnatus ; folia caulina erecto-pa- tentia, undique disposita, basi subovata, acumine lato ligulato obtuso, nervo tenui medio evanescente, ramea compressa patula ovato-oblonga obtusa, nervo supra medium evanido, ob prominentiam papillarum margine crenulato, cellulis parvis rotundis papillosis obscuris. Hab. Ins. Ceylon, Dr. Thwaites. Caulis unciam longus; rami bilineares. Folia intense viridia subglauca. This appears to be a species less than A. tristis, Cesati; and, so far as can be seen by the specimens yet examined, the habit seems more depressed. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 309 A. Hurronu, Mitt. Caulis procumbens, ramis ascendentibus subpinnatim divisus ; folia cau- lina patentia, ovata, acumine lato apice acutiusculo, nervo indistincto supra medium evanido, ramea subcompressa, basi subovata, ligulata, acutiuscula, omnia inferne canaliculato-concava, apicem versus plana, cellulis rotundis obscuris papillosis areolata, margineque crenulata. Hab. New Zealand, Great Barrier Island, ZZutton et Kirk. Similar to A. tristis, Cesati, in stature, but more depressed and forming matted tufts. A. EXILIS, Mitt. Caulis procumbens, ramis irregularibus subpinnatim ramosus; folia compressa, caulina basi ovata ligulata obtusiuscula, rameaque patentia conformia vel basi minus ovata magis compressa, inferne canaliculato- concava, apice plana obtusissima, cellulis rotundis obscuris minutis- sime papillosis areolata, nervo indistincto supra medium evanido. Hab. Africa australis, Natal, Attercliffe, J. Sanderson. A. Huttoni minor, foliis rameis obtusioribus magis compressis. CYATHOPHORUM, Brid. C. SUBLIMBATUM, Tw. et Mitt. Caules apice szepe attenuati, foliis magnitudine decrescentibus, simplices; folia lateralia divergentia, ovalia acuminata, margine a medio ad apicem usque denticulis angustis serrata, limbo obsoleto, nervo ad tertiam folii longitudinis partem producto simplici furcatove; foliastipuliformia late ovalia acuminata, enervia, margine superne denticulata. Hab. las. Ceylon, Dr. Thwaites. C. Hookeri statura adspectuque simile, foliis autem limbo indistincto cellulisque paululum majoribus et foliis stipuliformibus enervibus den- tieulatis diversum. HyPOPTERY.GIUM, Brid. Euhypopterygium, Mitt. Musc. Aust. Amer. H. APICULATUM, Thw. et Mitt. Dioicum ; folia in ramorum primariorum medio latissime oblongo-ovata, in acumen breve producta, nervo infra apicem evanido, limbo superne denticulato circumdata, cellulis parvis rotundis obscuriusculis ; folia stipuliformia orbiculata, nervo in acumen angustum excurrente, inte- gerrima; perichetialia late ovata, nervo excurrente acuminata ; theca in pedunculo elongato, oblonga, horizontalis, operculo longirostri. Hab. Ins. Ceylon, Dr. Thwaites. Stipes semiuncialis, in frondem circiter semiunciam longam apice divi- 310 MR. w. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. sus. Folia pallide glauco-viridia, sicca deflexa. Pedunculus semiun- ciam longus. H. aristato, Dryol. Javan. t. 141, valde similis; folia autem cellulis fere dimidio minoribus areolata. Hooxertra, Sm. Trans. Linn. Soc. ix. 275. Cyclodietyon, Mitt. Journ. Linn. Soc. vii. 163. H. CEYLANICA, Thw. et Mitt. Monoica synoicaque; caulis ramosus ; folia compressa, media late ovalia, intermedia oblongo-ovalia, lateralia angustiora, uno latere a medio ad basin usque inflexo, omnia obtusa, apiculo brevi terminata, margini- bus limbo angusto e duplici serie cellularum composito circumdata, apice serrulata, nervis ad $ folii longitudinis productis, dorso apicem versus denticulatis, cellulis hexagonis pellucidis; perichztialia parva, lanceolata ; theca in pedunculo elongato, ovalis, horizontalis decurvave, collo vix ullo distincto, operculo conico acuto; peristomium dentibus lamina externa per lineam mediam divisa, internum processibus den- tium longitudine, in membrana ad j ejus longitudinis exserta im- positis. Hab. Ins. Ceylon, rarissima, Dr. Thwaites. Caulis pallidus, depresse cespitosus, cum foliis lineam latus. Folia pal- lide viridia. Pedunculus uncialis, ruber. H. blumeano similis, foliis autem denticulatis. LzrrpoPrnvw, Brid. ii. 267. Eulepidopilum, Mitt. Muse. Austr. Amer. 267. L. FURCATUM, Thw. et Mitt. Monoicum; caulis dichotome divisus; folia compressa, nitida, media ovalia acuminata, intermedia ovali-lanceolata parum acuminata, late- ralia divergentia elliptico-lanceolata uno latere inferne inflexo, omnia medio tenus binervia, margine apicem versus serrulato, limbo angusto e duplici serie cellularum angustarum composito circumdata, cellulis oblongis areolata; perichetialia parva, lanceolata; theca in pedunculo duplo longiore, erecta, ovalis, evacuata infra os siccitate valde con- stricta; peristomii dentes e stratis laminæ externe divisis, interne processus zquilongi angusti in membrana usque ad ¿ dentium lon- gitudinis exserta impositi; ealyptra pilis paucis inspersa. Hab. Ins. Ceylon. “Dr. Thwaites. Caulis biuncialis, latitudine cum foliis vix bilinearis. Folia superiora viridia, seniora fuscescentia interdum purpurascentia. Pedunculus subbilinearis, lzevis. . MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 311 DisricuoruvrruM, Dozy et Molk. Muse. Archip. Ind. 99. Mniadelphus, C. Muller. D. LIMPIDUM, Thw. et Mitt. Caulis elongatus, subsimplex, flaccidus; folia compressa, media late ovali- oblonga, in apiculum brevem acuminata, intermedia patentia latera- liaque patula longiora ceteroquin conformia, planiuscula, mollia, mar- gine parum undulata, angustissime limbata, integerrima, cellulis ro- tundo-hexagonis limpidis limitibus angustis areolata, nervo angusto sub apice evanido. Hab. Ins. Ceylon, Dr. Thwaites. Caulis biuncialis, fuscescens, cum foliis 24 lineas latus. Folia tenera pallide glauco-viridia, seniora pallide fusca, siccitate corrugata. Cells of the leaves towards the apex and margin twice as large as in D. succulentum, Mitt. Muse. Ind.; and partly from this cause, although the outline of the leaf 1s nearly the same, the pre- sent species appears more pellucid. D. MUCRONATUM, Thw. et Mitt. Synoicum ; caulis breviusculus; folia superiora pallide viridia, seniora fusca compressa, omnia fere conformia, elongate oblonga, lateralia pa- tentia basi parum angustiora oblonga subspathulata obtusa cum api- culo elongato angusto pungente, margine flexuoso limbo angusto cir- cumcincto integerrima, nervo usque ad f folii longitudinis producto, cellulis parvis rotundis; perichzetialia minuta, ovalia acuta acuminatave ; theca in peduneulo breviuseulo rubro gracillimo, ovalis, horizontalis, colo sensin attenuato; peristomium dentibus lamina externa per lineam mediam divisa, internum zequilongum processibus in membrana usque ad 3 ejus longitudinis exserta impositis. Hab. Ins. Ceylon, Dr. Thwaites. Caulis subsemiuncialis, cum foliis 2 lineas latus. Folia siccitate margi- nibus flexuosis. Pedunculus bilinearis. Larger than D. cuspidatum, Dozy et Molk. Muse. Archip. Ind. t. 33; leaves more oblong, less spathulate, and with longer nerves. Eniorvus, Brid. E. Lvcrpus, Thw. et Mitt. Caulis gracilis, inferne nigro-fuscus, simplex vel superne divisus; folia, apicem versus caulis minora, compressa, plana, media intermediaque late ovalia, lateralia ovalia subduplo majora, omnia apiculo brevi termi- nata, nervis inzequalibus brevibus inconspicuis, limbo angusto e duplici serie cellularum angustarum composito marginata, superne denticulis parvis spinuloso-serrata, cellulis ovalibus hexagonis post resorptionem chlorophylli parietibus angustis areolata ; perichetialia ovato-lanceolata 312 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. acuminata, immarginata, subserrata ; theca in pedunculo brevi ubique setulis brevibus obtecto, breviter ovalis collo attenuato, horizontalis, operculo conico tenuirostri; peristomium dentibus lamina exteriore per lineam mediam divisa, latioribus, internum equilongum, pro- eessibus in membrana usque ad 3 dentium longitudinis exserta im- positis; calyptra pilis hyalinis obtecta basique fimbriata. Hab. Ins. Ceylon, Dr. Thwaites. Caulis unciam altus, cum foliis lineas duas latus. Folia pallide viridia, sicca scariosa nitida, cellulis sub lente iridicoloribus. Pedunculus bi- linearis, pallidus, setulis ad flexuram, ubi in collum thece transit, parum longioribus surrectis. Smaller than Æ. remotifolius, C. Muller, Bryol. Javan. t. 157 but the peduncle less setulose. GAROVAGLIA, Endl. G. DENSIFOLIA, Thw. et Mitt. Caulesdense congesti, rigidi, crassi, teretes; folia dense inserta, imbricata, oblonga, infra medium parum latiora, apice obtusa cum apiculo brevi recurvo, concava, plicis 2-4 levibus exarata, haud excavata, brevis- sime fere obsolete binervia, margine basis late implana, apicem versus minute serrulata, cellulis oblongis oblique seriatis, alaribus parvis quadratis flavis areolata ; perichzetia exserta, obconica, foliis orbiculari- obovatis vaginantibus, apicibus acuminatis patentibus integerrimis ; theca oblonga, immersa, operculo acuminato parum obliquo, peristo- mium infra thecx os impositum, dentibus parvis horizontalibus inflex- isve rubris teneris cohzrentibus liberisve. Hab. Ins. Ceylon, Dr. Thwaites. Caulis 1-3 uncias altus, cum foliis 2 lineas crassus. Folia superiora straminea, inferiora fuscescentia, perichetialia supra orificium thee cupulatim expansa. This agrees with G. plicata (Nees ab E.) in its immersed cap- sule, but is immediately distinguished by its wide, short, pointed leaves. From G. aristata, V. d. Bosch et Lac. Bryol. Javan. t. 185, it differs in the more nearly oblong form of its leaves, and also in the form of those belonging to the perichetium. The following sketeh of the species certainly known to belong to the genus Garovaglia will show the positions assignable to the additions here proposed :— MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 313 A. Theca immersa. * Peristomium incompletum. plicata (Nees). Powellii, Mitt. aristata, V. d. Bosch et Lac. levifolia, sp. nov. obtusifolia, sp. nov. densifolia, sp. nov. tortifolia, sp. nov. ** Peristomium duplex, completum. setigera (Sull.). cuspidata, Mitt. B. Theea exserta. elegans, Dozy et Molk. moluccensis, V. d. Bosch et Lac. samoana, Mitt. angustifolia, Mitt. compressa, sp. nov. carinata, sp. nov. Some other species have been described by C. Müller as having been found in Ceylon by Nietner; but, from the descriptions, they must be different in habit, for they are stated to have creeping stems, a character observable in none of the species above tabu- lated, in all which the habit appears to be as in Orthotrichum. G. OBTUSIFOLIA, Thw. et Mitt. Caulis basi divisus; foliadense inserta, imbricata, patentia, latissime ovalia obtusissima, cum mucrone minuto inflexo reflexove, sub apice excavata, circiter quadriplicata, plicis dorso leevibus, breviter binervia, margine integerrimo minutissimeve serrulato, cellulis densiusculis areolata; pe- richeetialia erecta, latissime ovalia, cuspide brevi patente ; theca emer- gens, ovato-oblonga, sieca sub ore contracta; peristomium depressum, dentibus carnosis rubris cohxrentibus siccitate mvolutis, internum zquilongum processibus subcarnosis angustis flavis liberis. Hab. Ins. Ceylon, Dr. Thwaites. G. densifolie simillima, foliis autem sub apice profundius concavis, api- culo inconspicuo et perichztio peristomioque diversa. G. LAEVIFOLIA, Thw. et Mitt. Dense cespitosa; caulis subsimplex ; folia dense inserta, tereti- et tumidi- imbricata, oblongo-ovalia, apice excavata, cochleariformi-concava, lzvia, apiculo subplano brevi recurvo, margine apicem versus serrulato, enervia, cellulis oblique seriatis areolata ; theca oblonga, in perichzetium 314 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. foliis erectis lanceolatis integerrimis convolutis immersa, operculi rostro brevi obliquo; peristomium depressum, dentibus brevibus co- herentibus, vestigiis tenerrimis processus adh:zrentibus. Hab. Ins. Ceylon, Dr. Thwaites. Crassitudine caulium G. plicate similis, primo visu tamen ex omnibus speciebus hucusque descriptis foliis haud plicatis distincta. G. TORTIFOLIA, Mitt. Caulis eurvatus, ruber, inter folia conspicuus ; folia subcompressa, diver- gentia, semitorta, ovali-elliptica, sensim in acumen longiusculum an- gustum leve producta, plicis tribus dorso levibus exarata, parum excavata, margine late implano, superne denticulis angustis serrato, à medio ad basin usque revoluto, breviter binervia, cellulis angustis areo- lata; pericheetialia late ovalia, sensim acuminata, acumine erecto, inferne dentieulis paucis remotis inconspicuis serrata, apice integra; theca immersa, cylindracea, opereulo acuminato obliquo. Hab. Ins. Borneo, Sarawak, Everett. G. eleganti quoad formam foliorum similis, directione autem eorum di- varicata et theca immersa statim dignoscitur. G. compressa, Mitt. Caulis elongatus, curvatus; folia compressa, patentia patulave, stricta, elongate lanceolata, apice latiuscule acuta, basi parum latiora, plicis tribus statu sicco distinctioribus exarata, planiuscula, enervia, margine apicem versus denticulato, cellulis angustis areolata; perichetialia parva, late ovalia, acumen versus breviter denticulata; theca in pedun- culo zequilongo, evlindracea, operculo oblique rostrato. Hab. Ins. Borneo, Sarawak, Everett. More slender than G. angustifolia, and with narrower leaves, whieh are not acuminate. Unlike any other species in its narrow, flat, compressed leaves. G. cARINATA, Mitt. Caulis inferne decumbens, rarius divisus; folia patentia, interdum subse- cunda, ovata, vel ob margines basi reflexos latissime lanceolato-ovata, in acumen pungens dentatum sensim producta, concava, infra acumen parum excavata, tri- quadriplicata, breviter binervia, dorso denticulis paucis scabriuscula, margine denticulis serrato, cellulis parvis angustis oblique seriatis areolata ; perichzetialia parva, erecta, ovata, apicedentato- lobata subintegrave acuta, convoluta ; theca in pedunculo breviore, sub- curvata, oblongo-cylindracea; peristomium depressum, dentibus sub- horizontalibus apicibus incurvis inferne cohxerentibus rubris. Hab. Bootan (herb. Griffith). A robust species, with leaves less densely inserted than they are in G. plicata and less spreading than in G. elegans; there are present on the back of the leaf a few scattered teeth of the MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 315 same kind as those so prominent on the folds of the leaves of G. plicata. From the figure in ‘ Bryologia Javanica,’ t. 195, G. aristata in size and appearance resembles G. carinata; but the fructification is as represented in the same work, t. 196 (G. moluccensis). The internal peristome, as seen in old capsules, appears to exist only in a rudimentary state, aud is perhaps adherent to the teeth. PrEnoBRYUM, Hornsch. P. INVOLUTUM, Thw. et Mitt. Rami inferne nudi, superne in frondem subpinnatim ramosam divisi; folis patentibus ovali-oblongis brevi-apiculatis marginibus incurvis conniventibus integerrimis, nervo angusto infra apicem evanido, cellu- lis angustis, basalibus ad angulos densioribus aurantiacis areolatis; pe- richztia e ramo primario vel e ramulorum latere oriunda, foliis lon- gioribus convolutaceis erectis subulato-angustatis integerrimis; theca in pedunculo brevi cylindracea, operculo conico-rostrato ; peristomium breve, depressum, dentibus teneris pallidis, annulo lato cinctis ; calyptra dimidiata. Hab. Ins. Ceylon, prov. central., alt 4000'-6000', Dr. Thwaites. Rami fertiles biunciales, steriles elongati ramulis nonnullis filiformi- attenuatis. Folia lineam longa, stramineo-viridia, interdum fulves- centia, subnitida; sicca immutata. Pedunculus bilinearis, fuscus. ‘Theca fusca, pedunculo longitudine zqualis. This species, in habit and appearance, closely resembles P. convolutum (Dozy et Molk.), *Musc. Archip. Ind.’ t. 50 (Sym- physodon), and P. cylindraceum (Mont.), but is a little more robust; the foliage, in the incurvation of the margins of the leaves, resembles that of the first-named moss. The exserted capsule and dimidiate calyptra, so different from those of spe- cles otherwise closely related, may suffice for the placing of this species in a different section, which appears to be the most na- tural way of disposing of such differences. Many similar instances are presented amongst the mosses referred to the genus Neckera, the capsule being immersed in the perichetial leaves with a short calyptra, or exserted and having a longer calyptra decidedly split on one side. P. CEYLANICUM, Thw. et Mitt. Rami clongati, curvati, simplices vel parum divisi, ramulos angustos sto- loniformes hic illic emittentes ; folia patentia divergentiave, densius- cula, sed laxe imbricata, late ovalia obtusa et muer ne angusto subu- lato, excavata, valde concava, margine basi inflexo caulem amplexante, LINN. JOURN.— BOTANY, VOL. XIII, Z 316 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. apicem versus incurvo, integerrima, nervis binis basi conflatis, cellulis angustissimis, basalibus nullis diversiformibus areolata ; perichzetialia erecta, exteriora brevia, interiora elongate oblonga concava, versus acumen subulatum parum excavata, ad thecam oblongo-ovatam brevi- pedunculatam emergentem appressa; peristomium et humidum et siccum in conum acutum prominens e dentibus angustis firmis dorso articulationibus prominentibus denticulato-asperis pallide fuscis com- positum.—Meteorium crassicaule, Mitt. Musci Ind. Or. Hab. Ins. Ceylon, Rambodde : Gardner, no. 9; Thwaites, no. 181. This fine species in size and appearance very nearly resembles the Meteorium crassicaule (C. Müller), Bryol. Jav. t. 210, but differs in its more elongated branches, which produce but few flagelliform shoots. The leaves in all the specimens are of a rich glossy brown colour, except towards the apices of the stems, where they are straw-coloured and less glossy ; they are destitute of any longitudinal folds. The Meteorium flexipes, Mitt. Muse. Ind. Or., in foliage and appearance nearly resembles P. ceylanicum ; but its capsule is exserted. All these species present characters which are rather those of Pterobryum than of Meteorium ; for they are not species having a pendulous habit. METEORIUM, Brid. ii. 264. M. ATTENUATUM, Thw. et Mitt. Rami elongati, graciles, ramulis divergentibus irregulariter pinnatim ra- mosi; folia patentia divergentiaque, ovato-lanceolata, sensim angustis- sime attenuata, ramulina angustiora, tenuiter capillari-attenuata, omnia basi cordata, alis inflexis, marginibus minute serrulatis, acu- mine levi, nervo tenui medio evanescente, cellulis angustis papilla unica dorso prominula punctulatis. Hab. Ins. Ceylon, Dr. Thwaites. Rami semipedales, ramulis semiuncialibus ramosi. Folia seniora nigri- cantia, juniora stramineo-viridia, nitida, sicca parum mutata. M. RUFIFOLIUM, Thw. et Mitt. Monoicum! caulis pendulus, ramis dissitis divaricatis; folia ramea ap- pressa vel cum ramulinis patula subcompressa, ovato-lanceolata sen- sim angustata, apice anguste acuta, basi uno latere inflexa, plani- uscula, nervo tenui supra medium evanido, margine ubique serrulato, cellulis angustissimis elongatis, papillis singulis minutis dorso promi- nulis; perichetialia parva, conformia, erecta; theca in pedunculo brevi lxvi, ovalis, operculo subulato obliquo ; peristomium internum proces- sibus angustis cum dentibus subzquilongis, in membrana brevi exserta impositis; calyptra parva, basi plurifida, sed uno latere profundius fissa. Hab. Ins. Ceylon, Dr. Thwaites. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 317 Rami semipedales, graciles, ramulis semiuncialibus. Folia sesquilineam longa, flavo-fulvescentia subnitida, sicca directione immutata. Pedun- culus lineam longus. In appearance similar to M. aureum, Griff., but with shining leaves, singular in its inflorescence. M. ENERVE, Thw. et Mitt. Rami elongati, penduli, ramulis divaricatis pinnati; folia ramea appressa, imbricata, late ovata, apice in subulam elongatam angustam producta, integerrima, enervia, cellulis angustissimis leevibus, ad angulos basali- bus paucis latioribus suboblongis concoloribus areolata; ramulina pa- tentia oblonga obtusa, in acumen subulatum angustum elongata, margine superiore acumineque subserrulata, canaliculato-concava, enervia, eellulis superioribus papillis singulis difficillime definiendis punctata. Hab. Ins. Ceylon, Dr. Thwaites. Rami pedales, ramulis uncialibus vel brevioribus. Folia lineam longa, ramea cauli appressa, ramulina plumose patentia; juniora viridia, deinde fulvo-fuscescentia. Cn xTOMITRIUM, Dozy et Molk. C. CONFERTUM, Thw. et Mitt. Dense c:zspitosum ; rami ascendentes, ramosi ; folia late oblonga, integer- rima, basi subcordata, apice subito in acumen subulatum denticulatum contracta, brevissime binervia, ramea patula imbricata ovato-lanceolata acuta subacuminatave excavata margine apiceque implana serrulata enervia, cellulis angustis elongatis areolata ; perichetialia plura, cau- linis majora, subconformia, apice integra lacerave, marginibus ciliato- dentatis; theca in pedunculo rubro breviusculo basi levi superne brevissime setuloso-aspero inclinata, ovali-oblonga, ore obliquo, oper- culo curvirostro, peristomii dentibus latis, interni aquilongi processi- bus latiusculis in membrana usque ad 3 ejus longitudinis exserta im- positis; calyptra setulis aspera, basi lacera. Hab. Ins. Ceylon, Dr. Thwaites. C. venusto, Reinw. et Nees ( Hypnum), simillimum. . MACROHYMENIUM, C. Muller. M. Lave, Thw. et Mitt. Monoicum ; rami breves; folia patentia, imbricata, elliptico-lanceolata, concava, margine anguste plana, integerrima, cellulis oblongis oblique seriatis, alaribus pluribus limitibus incrassatis aurantiacis; perichz- tialia erecto-patentia, majora, magis acuminata; pedunculus ruber, levis, ubi in thece erecte ovalis collum transit subscaber; operculum conicum, rostro subulato ; peristomium dentibus brevibus supra me- z2 318 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. dium subulatis, internum processibus quam dentes duplo longioribus, in membrana brevissima insidentibus. Hab. Ins. Ceylon, Dr. Thwaites. M. rufo, Dozy et Molk. Musc. Archip. Ind. t. 54, similis, foliis autem longioribus. Rhegmatodon Nietneri, C. Mueller, Linn. Band xxxvi. p. 20, which must also be included in the genus Macrohymenium, is described with a capsule * undique asperula,” and must therefore be suffi- ciently distinct. SEMATOPHYLLUM, Mitt. Journ. Linn. Soc. 1868. Sect. Acroporium, Mitt. 1. c. S. PUNCTULIFERUM, Thw. et Mitt. Dioicum ; caulis procumbens, ramis assurgentibus laxius confertis; folia laxe inserta, patentia, recta, apicalia m gemmam teretem angustam elongatam convoluta, ovali-elliptica acuta, marginibus superne conni- ventibus tubulosa, integerrima, basi subcordata, cellulis alaribus con- spicuis, superioribus angustis subconflatis dorso limitibus prominenti- bus; perichetialia ovata, in acumen elongatum serrulatum products ; peduneulus elongatus, ruber, seaber; theca ovalis, horizontalis; peri- stomium dentibus latis, lamina interna quam externis angustiore, in- ternum processibus, ciliis brevioribus interpositis, in membrana ad ¿ dentium longitudinis exserta insidentibus. Hab. Ins. Ceylon, Dr. Thwaites. Folia semilineam parum excedentia. Pedunculus sexlinearis. Sect. Rhaphidorrhynchum, Schimp. S. CAPILLIFERUM, Thw. et Mitt. Dioicum? caulis procumbens, ramis assurgentibus ; folia laxe et tereti- uscule imbricata, patentia, ovalia, basi constricta, apice in acumen an- gustum rectum subulato-attenuata, integerrima, excavata, margine superne subplana, cellulis angustis alaribus distinctis areolata, subu- lata, integerrima ; pedunculus elongatus, levis; theca ovalis, horizon- talis, operculo subulirostro ; peristomii interni processus ciliis singulis interpositis in membrana usque ad 3 ejus longitudinis exserta impositi. Hab. Ins. Ceylon, Dr. Thwaites. Rami sesquilineam crassi. Folia lineam longa, tenera, nitida, levia. Pedunculus semiuncialis. Theca parva. S. cirrhifolio, Schwegr., statura similis, folia autem haud compressa, integerrima. S. MONOSTICTUM, Thw. et Mitt. Monoicum, pusillum, cespitosum ; folia compressa, lateralia divergentia, apicibus decurvatis subsecundis, ovali-lanceolata subulato-acuminata, MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 319 eoncava, basi constricta, margine erecto superne subflexuoso serrulata, cellulis elongatis apicibus in papillas exstantibus perspicuis, alaribus distinctis fuscis areolata; perichztialia longiora, conspicue serrata ; pedunculus ruber, versus collum thecz ovalis pendulz scaber; oper- culum longissime subulatum; peristomium dentibus latis, lamina interna quam externa angustiore, internum zequilongum processibus cilii nullis in membrana ad dentium dimidiam longitudinem exserta impositis. Hab. Ins. Ceylen, Dr. Thwaites. Rami cum foliis lineam lati. Folia viridia, fuscescentia. Pedunculus trilinearis, apicem versus arcuatus. Theea minuta. S. RUFICAULE, Thw. et Mitt. Dioicum; caulis procumbens, ramis assurgentibus rufis; folia sub- compressa, laxe disposita, patula, recta, lanceolata subulato-angu- stata, apice autem latiusculo, concava, margine tenuiter sed argute serrulato, cellulis superioribus elongatis apicibus in papillas grossi- usculam exstantibus, basin versus lzvioribus, alaribus majusculis oblongis externis hyalinis areolata ; perichzetialia erecta, longiora, ar- gutius serrata, subplicata ; pedunculus validus, ruber; theca ovalis, inclinata, operculo conico subulirostri; peristomii dentibus validis intus lamelliferis, processibus in membrana usque ad j dentium longi- tudinis exserta impositis. Hab. Ins. Ceylon, Dr. Thwaites. Caulis 1-13 unciam longus, gracillimus. Folia semilineam parum exce- dentia, plumose disposita, glauco-pallido-viridia. Pedunculus 4-5 li- neas longus. S. ASPERIFOLIUM, Thw. et Mitt. Dioicum? caulis procumbens, ramis ascendentibus laxifolüs; folia pa- tentia, stricta, sicca directione immutata, apicalia erecta appressa, anguste elongate lanceolata, sensim angustata, parum concava, mar- gine apiceque implana, subintegerrima vel apice denticulis inconspicuis, cellulis obscuriusculis oblique seriatis, omnibus finitibus prominentibus dorso scabris, cellulis alaribus externis oblongis conspicuis hyalinis; pe- richztialia e basi subovata superne lobato-dentata, in acumen loriforme producta, marginibus argute denticulatis; theca in pedunculo brevius- culo superne scabro, ovalis, inclinata, subhorizontalis, operculo subu- lato; peristomii dentes lati, lamina interna multo angustiore, interni processibus ciliis nullis in membrana ad ¿ dentium longitudinis ex- serta impositis. Hab. Ins. Ceylon, Dr. Thwaites. Rami bilineares, rufi. Folia lineam longa, flavo-viridia, nitoris destituta, Pedunculus 4-linearis. S. RAMULINUM, Thw. et Mitt. Dioieum ? cespitulosum, ramis subpinnatim divisis ; folia subcompressa 320 MR. w. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. imbrieata, inferiora oblonga, apice in acumen angustum producta, superiora et ramulina obtusiuscula, apiculo brevi latiusculo, basi seepe subsaccata, infra apicem excavata apice margineque plano sub- integerrimo, cellulis angustis elongatis apicibus elevatis, preecipue ad dorsum partis excavate prominulis, basalibus flavis, alaribus distinctis hyalinis; pericheetialia longiora, erecta, lanceolata acuminata, flexuosa, margine plana, superne serrulata; pedunculus superne papulosus ; theca ovalis, horizontalis deflexave, operculo subulirostri ; peristomium inter- num processibus apice tenuissime capillari-attenuatis quam dentes lon- gioribus ciliisque singulis in membrana usque ad mediam dentium lon- gitudinem exserta impositis. Hab. Ins. Ceylon, Dr. Thwaites. Rami 3-4 lineas longi, rubri, cum foliis lineam lati. Folia pallide viridia, subnitida. Pedunculus 4-5 lineas longus, ruber. T. fequendamensi (Hampe), Muse. Aust. Amer., simile, sed caulibus longioribus crassioribusque, statura Chionostomo rostrato (Griff.) similis. Levcomium, Mitt. Journ. Linn. Soc. 1868. L. LIMPIDUM, Thu. et Mitt. Synoicum ; caulis debilis; folia compressa, subdecurva, lateralia patentia ovato-lanceolata subacuminata integerrima, cellulis angustis limpidis areolata ; pericheetialia parva, latiora, subito acuminata, subdenticulata ; theca in pedunculo rubro, ovalis, horizontalis, operculo rostrato ; peri- stomium dentibus laminis externis per lineam mediam divisis, internum zquilongum, processibus in membrana usque ad 4 ejus longitudinis exserta impositis. Hab. Ins. Ceylon, Dr. Thwaites. Caulis semiuncialis, cum foliis sesquilineam latus. Folia pallide viridia, sicca vix mutata. Pedunculus semiuncialis. EcTROPOTHECIUM, Mitt. in Journ. Linn. Soc. 1868. - E. LEVIGATUM, Thw. et Mitt. Monoicum, depresse ceespitosum ; caulis ramis approximatis, pinnatus; folia compressa, complanata, lateralia patentia subacuminata, media intermediaque ovali-elliptica acuta, margine serrulata, basi constricta, cellulis alaribus parvis inconspicuis, reliquis angustis lzvibus ; periche- tialia erecta, ovato-oblonga, subulato-angustata, serrulata; theca in pedunculo ruberrimo, parva, ovalis, inzequalis, operculo rostro brevi subcurvulo; peristomium normale. Hab. Ins. Ceylon, Dr. Thwaites. Rami cum foliis vix lineam lati. Folia albo-viridia, subnitida, siccitate varum mutata. Pedunculus 6-7 lineas longus. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 321 E. suBRETUSUM, Thw. et Mitt. Dioicum? caulis procumbens, pinnatus; folia complanata, lateralia patula complicata, omnia ovato-ligulata, apice oblique obtusa subretusave, brevissime binervia, margine apicem versus (ibique denticulis pluribus approximatis apicibus divisis) crebre serrata, cellulis angustissimis, su- perioribus apicibus in papillas aculeiformes prominentibus areolata ; perichzetialia erecta, ovato-lanceolata, longius binervia, serrulata; theca in pedunculo elongato levi, horizontalis, ovalis, collo sensim attenuato, ore satis magno, serius hiante; peristomium internum processibus ciliisque in unum coalitis «equilongis, in membrana usque ad 3 longitu- dinis dentium exserta insidentibus. Hab. Ins. Ceylon, Dr. Thwaites. Caulis circiter trilinearis. Folia subnitida, sicca decurva viridia. Pe- dunculus 8-9 lineas longus. PLAGIOTHECIUM, Schimp. Bryol. Europ. P. suBGLAUCUM, Thw. et Mitt. Caulis prostratus, elongatus, simplex ; folia complanata, lateralia patenti- divergentia, ovata, basi inzequalia, apice parum acuminata acuta, inte- gerrima, angustissime areolata, nervis brevibus inzequalibus. Hab. Ins. Ceylon, Dr. Thwaites. Caulis 2 uncias longus, cum foliis sesquilineam latus. Folia glauco- viridia, siccitate transversim subundulata. Hyexum, Dill. Sec. Helicodontium, Schwegr. iii. ii. 2. H. APPLANATUM, Thw. et Mitt. Monoicum, pusillum, depresse caespitosum, subpinnatim ramosum ; folia caulina assurgentia, ovato-lanceolata subacuminata, subserrulata, ramea compressa serrulata, nervo concolori tenui medio evanido, cellulis ad angulos abbreviatis, reliquis oblongis, omnibus obscuriusculis areolata ; perich:zetialia erecta, ovata acuminata, enervia ; theca in pedunculo bre- viusculo rubro, breviter ovalis; peristomium internum processibus quam dentes brevioribus, membrana basali fere nulla. Hab. Ins. Ceylon, Dr. Thwaites. Rami cum foliis circiter semilineam lati. Folia viridia, obscuriuscula, sicca cauli appressa stricta. Pedunculus bilinearis. MNIODENDRON, Lindb. M. DELTOIDEUM, Thw. et Mitt. Stipes tomentosus, folis basi deltoideis angulis subauriculatis inde in subulam productis divaricatis, nervo percurrente dorso remote denti- 3229 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. culato, margine basi crenato superne remote serrulatis, cellulis ad angulos latioribus, reliquis angustis areolatis. Hab. Adam's Peak, Gardner. Sent also by Dr. Thwaites. Habit and appearance entirely those of M. divaricatum, for which it was passed over in the Musc. Ind., but different in the base of its leaves. M. ARISTINERVE, Mitt. Stipes tomentosus, foliis divaricatis basi deltoideis inde in subulam elongatam aristiformem productis, nervo tereti longe excurrente dorso remote dentato, margine remotiuscule serrulato, cellulis ad angulos suboblongis, reliquis angustis areolatis; rami abbreviati, foliis densis basi subdeltoideo-ovatis inde subulatis argutins serrulatis ; folia peri- cheetialia e basi tenuiore ovata, longe tenereque subulata, subserrulata ; theca in pedunculo elongato rubro, horizontalis decurvatave, cylin- dracea, pluries plicata, operculo a basi conica sensim subulato. Hab. Ins. Borneo in acumine montis Kina-Balloo, H. Lowe. Stipes subuncialis ramis semiuncialibus. Pedunculus sesquiuncialis. M. comoso habitu valde simile, foliis autem aristinerviis satis diversum. FissrpENs, Hedw. F. TERMINIFLORUS, Thw. et Mitt. Dioicus, humilis; folia circiter 20-juga, elongate oblongo-lanceolata acuta, nervo concolori percursa, lamina vera ultra medium producta apice quali, margine integerrimo, limbo nullo, cellulis minutis ob- scuris minutissime papillosis areolata ; theca in pedunculo breviusculo crassiusculo rubro, ovalis, inclinata, suberecta, subzqualis, operculo conico acuminato ; flos masculus terminalis foliorum laminis basi sinu- atis vaginantibus inclusus. Hab. Ins. Ceylon, Dr. Thwaites. Caulis trilinearis, cum foliis lineam latus. Folia glauco-viridia, superiora fere lineam longa, subflaccida, siccitate crispato-decurva. F. BICOLOR, Thw. et Mitt. Humilis; folia 6-] 2-juga, erecto-patentia, superiora conniventia incurvata ovato-oblonga acuta, nervo pellucido in apiculum breve excurrente, margine minutissime crenulato, lamina vera ultra medium producta apice subzequali, limbo hyalino a basi usque ad medium ascendente marginata, cellulis minutis opacis; pedunculus pallidus ; theca ovalis, erecta. Hab. Ins. Ceylon, ad terram, Dr. Thwaites. Caulis 1-13 lineam altus. Folia j lineam longa, intense viridia, nervo pellucido latiusculo. | Pedunculus lineam longus. F. semilimbato similis. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 323 F. PAPILLOSUS, Thw. et Mitt. Monoicus; folia circiter sexjuga oblongo-lanceolata acuta, nervo palli- diore in apice brevissime excurrente mucronata, lamina vera ad me- dium usque producta apice x quali, limbo pallido ubique marginata et in laminam apicalem breviter ascendente, marginibus reliquis crenulatis, cellulis parvis ob papillas valde prominentes subobscuris ; pedunculus elongatus, pallidus; theca ovalis, æqualis, inclinata. Hab. Ins. Ceylon, Dr. Thwaites. Caulis lineam altus. Folia semilineam longa, intense viridia, siccitate decurva nervoque limbi marginalis laminz vere pallido, haud pellu- cido, prominente. F. bicolori similis, sed major et substantia foliorum et nervo siccitate prominente diversus. F. FUSCOVIRIDIS, Thw. et Mitt. Caulis elongatus, ramulis brevibus ramosus; folia 10-20-juga, patentia, elongato-oblonga acuta, nervo concolori pallidiore in apice evanido, margine crenulato subserrulato, cellulis rotundis ob papillas promi- nulas subobscuris limitibus pellucidis, lamina vera ad medium pro- ducta apice uno latere rotundato-inzquali; pedunculus fuscus; theca ovalis, suberecta, equalis ; flos masculus in ramulo brevi terminalis. Hab. Ins. Ceylon, ad corticem, Dr. Thwaites. Caulis 2 lineas altus, cum foliis lineam latus. Folia semilineam longa, pallide fusco-viridia, sicca parum mutata. F. CRASSINERVIS, Thw. et Mitt. Folia patentia, 6-9-juga, oblongo-lanceolata acuta, nervo crasso luteo ex- eurrente, margine subintegerrimo, cellulis rotundis limitibusque levibus pellucidis, lamina vera usque ad medium producta apice uno latere rotundato ; pedunculus pallidus ; theca ovalis, zequalis, erecta. Hab. Ins. Ceylon, Dr. Thwaites. Caulis 2 lineas altus, cum foliis $ lineam latus. Folia semilineam longa, pallidissime luteo-fusca, sicca parum mutata. Pedunculus lineam altus. F. minutus, Thw. et Mitt. Folia 4-8-juga, patentia, ovali-oblonga, obtusiuscule acuta, nervo pellu- cido pallido sub apicem evanido, lamina vera usque ad medium pro- ducta, apice parum inzequali, in caulibus sterilibus ubique limbo desti- tuta, in caulibus autem fructiferis presertim in foliis superioribus limbo usque ad medium laminx vere producto marginata, cellulis minutis ob- scuris areolata; pedunculus, crassiusculus, ruber ; theca ovalis, erecta. Hab. Ins. Ceylon, ad corticem, Dr. Thwaites. Caulis fructifer semilineam altus. Folia $ lineam longa, viridissima. Pedunculus lineam altus. F. semilimbato similis, sed minor. 324 MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. F. MICROCLADUS, Thw. et Mitt. Caulis inferne nudus, ramosus; folia 6-12-juga, superiora patentia, ob- longo-ligulata obtusa, nervo pallido subpellucido in medio laminz apicalis evanescente, marginibus minute crenulatis, lamina vera ultra medium ascendente, apice valde inzquali, latere ad terram spectante apice rotundato, altero latere cum lamina apicali continuo, ubique limbo destituta, cellulis minutis obscuris subopacis; pedunculus breviusculus, pallidus; theca erecta, ovalis, operculo rostrato. Hab. Ins. Ceylon, Dr. Thwaites. Caulis lineam altus, inferne sæpe foliis parvis equitantibus acutis remo- tiusculis, superne longioribus 1 lineam metientibus intense viridibus fuscescentibus, nervo in junioribus pellucido, in senioribus flavicante, infra apicem longe desinente. F. Gardneri Brasilize simillimus. F. FLABELLULUM, Tw. et Mitt. Caulis simplex ; folia 5—7-juga, oblongo-lanceolata acuta, nervo concolora in apiculum brevem excurrente, margine crenulato, cellulis rotundis parvis pellucidis limitibus latiusculis areolata, lamina vera ad medium producta apice inzequali ; pedunculus elongatus, pallidus ; theca ovalis, «equalis, suberecta; flos masculus terminalis, foliis basi sinuatis, alis laminz vere vaginantibus. Hab. Ins. Ceylon, Dr. Thwaites. Caulis semilineam altus, cum foliis semilineam latus. Folia vix lineam dimidiam longa, pallida leete viridia, sicca vix curvata. Pedunculus 2-3 lineas longus. Theca minuta. Habitus F. exilis, Hedw. F. ANGUSTUS, Thw. et Mitt. Caulis basi divisus; folia erecto-patentia, arcte approximata, imbricata, anguste et elongate lineari-lanceolata acuta, integerrima, nervo latius- culo concolori pallidiore in apice evanido, lamina vera ultra medium ascendente, apice subzequali, cellulis rotundis obscuris areolata. Hab. Ins. Ceylon, Dr. Thwaites. Caulis semiuncialis, cum foliis lineam latus. Folia lineam longa, sicca directione immutata, flavo-fusco-viridia. F. petrophilo, Sulliv. in Proceed. Amer. Acad. Aug. 1861, similis, sed haud nitidus. F. virens, Thw. et Mitt. Folia 3-6-juga, patentia, elliptico-lanceolata acuta, nervo brevissime ex- currente pellucido, margine crenulato, cellulis obscuris parvis papilli- feris, lamina vera usque ad medium producta apice :equali, limbo pallido marginata; pedunculus fuscus ; theca parva, ovalis, inclinata, basi subinzequalis, operculo rostrato. Hab. Ius. Ceylon, Dr. Thwaites. MR. W. MITTEN ON NEW SPECIES OF MUSCI FROM CEYLON. 3265 Caulis 3-1 lineam altus, cum foliis semilineam latus. Folia 1-3 lineam longa, intense viridia. Pedunculus subbilinearis. F. AxiLLiFLORUS, Thw. et Mitt. Folia 6-9-juga, patentia, lanceolata acuta, nervo subpellucido excurrente, margine crenulato, cellulis parvis obscuris papillosis, lamina vera ad medium producta apice «equali, limbo angusto cartilagineo marginata ; pedunculus elongatus, ruber; theca ovalis, subzequalis, inclinata ; flores masculi plures, axillares. Hab. lns. Ceylon, Dr. Thwaites. Caulis lineam altus, cum foliis latitudinem semilinez parum excedens, Folia semilineam longa, viridia, siccitate curvata. Pedunculus lineam longus. F. PLUMULA, Thw. et Mitt. Folia circiter 12-juga, patentia, late oblonga acuta, breviter apiculata, nervo concolori in apiculo evanescente, margine integerrimo, cellulis parvis rotundis limitibusque pellucidis lzevibus, lamina vera ad medium producta apice inzquali, limbo angusto pellucido marginata, pedun- culus breviusculus, apice curvatus; theca ovalis, horizontalis, zqualis. Hab. Ins. Ceylon, Dr. Thwaites. Caulis 3 lineas altus, cum foliis 14 lineam latus. Folia $ lineam longa, dilute lucideque viridia, sicca parum flexa. Pedunculus lineam longus. F. MULTIFLORUS, Thw. et Mitt. Caulis fertilis foliis 20-jugis patentibus oblongis acutis, nervo concolori in apice evanido, margine crenulato, cellulis parvis obscuris, lamina vera ad medium producta apice subzquali, limbo tenui cartilagineo marginata; pedunculus brevis, apice curvatus; theca ovalis, zequalis, horizontalis, opereulo rostrato. Hab. lus. Ceylon, Dr. Thwaites. Caulis 4 lineas altus, cum foliis lineam latus. Folia 3 lineam longa, viridia, sicca parum deflexa. Pedunculus lineam longus. F. plumule simillimus, foliis autem obscuris. F. PENNATULUS, Thw. et Mitt. Caulis elongatus, interdum ramosus, fertilis foliis 20-jugis patentibus approximatis elongate oblongis acutis, nervo pallido breviter excur- rente apiculatis, lamina vera ultra medium ascendente, apice subinze- quali, limbo fere ad apicem continuo marginata, marginibus reliquis minutissime crenulatis, cellulis minutis obscuris ; pedunculus pallidus ; theca breviter ovalis, suberecta, inzequalis. Hab. Ins. Ceylon, ad terram, Dr. Thwaites. Caulis 3-4 lineas altus. Folia semilineam longa, viridia fuscescentia. Pedunculus linea parum longior. F. Becxerril, Mitt. Caulis fertilis folis 4-5-jugis patulis, sterilis gracilior foliis 6-S-jugis minoribus, omnibus lanceolatis acutis subacuminatis, nervo crassius- 326 REV. W. A. LEIGHTON ON TWO NEW SPECIES OF culo viridi in mucronem excurrente, margine limbo angusto viridi cir- cumdatis, lamina vera valde inzquali, uno latere (ad terram spectante) minore, apice in novam descendentem angustato, cellulis satis magnis ovalibus et rotundo-ovalibus subpellucidis; pedunculus elongatus, rufus ; theca minuta, ovalis, basi gibba, insequalis, inclinata. Hab. Ins. Ceylon (Maanagalla, prov. centr. ad terram), Beckett, no. 9. Caules fertilis sterilisque lineam alti. Folia | ongiora, in caule fertili 2 lineam longa, in caule sterili dimidiam breviora, omnia pallide viridia, siccitate vix curvata. Pedunculus 3 lineas altus, apice pallidus. Theca abrupte in pedunculo contracta. F. exili habitu similis. EXPLANATION OF PLATE V. Erpodium ceylonicum. A. Fig. 1, plant, of the natural size. 2, portion of stem with leaves, as seen from above; 3, a leaf from the upper and one from the under side s the stem ; 4, perichætium with capsule and perichætial leaf expanded ; 5, calyptra ; 6, male flower and perigonial leaf: all magnified. E. Belli. B. Fig. 1, plant, of the natural size. 2, portion of the stem with leaves, as seen from the upper side; 3, leaves; 4, fertile branch with perichxtial leaves, capsule, and calyptra; 5, capsule after the emission of the pont 6, perigonium : all magnified. Aulacopilum tumidulum. C. Fig. 1, plant, of the natural size. 2, portions of stem as seen from above and from beneath ; 3, a leaf detached ; 4, pericheetium and capsule; 5, calyptra covering the same; 6, perigonium : all magnified. A. abbreviatum. D. Fig. 1, plant, of the natural size. 2, portion of the stem with leaves, as seen from above; 3, leaves detached; 4, perichetium and capsule; 5, calyptra; 6, perigonium : all magnified. On two new Species of the Genus Mycoporum, Flotow. By the Rev. W. A. LzicHtox, B.A., F.L.S., F.B.S. Ed. (Prate VI.) [Read December 5, 1872.] Mr. Currey has forwarded to me for examination and deter- mination two lichens on Bamboo, collected in Yomah, Pegu, THE GENUS MYCOPORUM. 827 Ar by Dr. S. Kurz in 1871, and sent to Mr. Currey with some fungi from Calcutta. They are very singular and remark- able plants, apparently referable to the genus Mycoporum of Flotow, although they possess peculiarities which lead one to doubt whether they ought not to constitute a separate new genus. Both are, so far as my present knowledge extends, undescribed, and may be characterized and designated as follows :— 1. MvcoPoRUM MELASPILON, Leight., n. sp. Thallus albidus, tenuis, subfurfuraceus, continuus aut subevanescens, linea nigro-fusca limitatus; apothecia atra vel fusco-nigra, irregula- riter congregata, 8-12- vel plurinodulosa, plano-depressa, excipulo communi involuta; noduli rotundati, plano-depressi, interne gru- mosi, arthonoidei, virescentes, opaci, continentes 7 ascos pyriformes incolores; paraphyses nulle ; spore 8, incolores, cylindraceo-fusi- formes, 13-septate, maxim. On bamboo, Yomah, Pegu, Dr. S. Kurz. The thallus appears to be composed of pale gonidia, and is either persistent or more or less evanescent. The apothecia look like irregular black spots or macule—under a microscope, com- posed of eight or twelve or more round, plane, depressed nodules clustered together and covered with a brownish black continuous membrane or excipulum, which is depressed or sunken in the in- terstices between the nodules. On viewing a single nodule hori- zontally with transmitted light, it appears of an opaque green hue, and contains seven colourless pyriform asci, six of which are ar- ranged at intervals around a central seventh. A vertical section shows the nodule to be composed of a thickish green grumous mass, darker on the outer surface, in which are imbedded the asci. 2. MycoporuM CIRCULARE, Leight., n. sp. Thallus albidus, tenuis, continuus, subfurfuraceus, linea nigro-fusca limitatus; apothecia consistentia e 6-7 nodulis, atris, rotundatis, subconvexis, verruciformibus, minutis, in circulum dispositis, primo aspectu nudis at realiter coopertis, excipulo communi interstitialiter depresso; noduli interne grumosi, virescentes, opaci, continentes 6 ascos rotundatos; paraphyses nulle ; spore 8, incolores, maxime, oblong aut sublineari-oblonge, muraliloculares, 13-septatz, loculi 3-4-5 in singulis seriebus. On bamboo, Yomah, Pegu, Dr. S. Kurz. A still more singular plant than the preceding. The thallus consists of pale gonidia. The nodules are arranged in minute gem-like circles, in number from six to seven, and are like minute 328 MR. W. T.v HISELTON DYER ON THE DETERMINATION OF subconvex verruce, but without any apical pore or epithecium, and are apparently separate, but in reality covered with a con- tinuous membrane or excipulum, concealed partially by the thallus, but seen in a vertical section to be a continuous green gru- mous mass. Each of the nodules, when viewed horizontally under the microscope, appears as a green grumous mass, in which the asci are circularly arranged. ` EXPLANATION OF PLATE VI. Fig. 1: a, lichen, natural size; 4, apothecium of clustered nodules, magnified 40 times; c, nodule, viewed horizontally, magnified 170 times ; d, vertical sec- tion of nodule, magnified 170 times ;“e, spore, magnified under pressure 1000 times. The small circular bodies to the left of the spore (e) are gonidia. Fig. 2: a, lichen, natural size ; 5, apothecia, magnified 40 times ; c, nodule viewed horizontally, magnified 170 times; d, vertical section of apothecium, mag- nified ; e, spore, magnified 1000 times. On the Determination of three imperfectly known S cies of Indian Ternstreemiacee. By W. T. Tutserron Drér, B.A., B.Se., F.L.S. [Read April 18, 1872, and January 16, 1873.] Tur synonymy of the species of Ternstroemiacee contained in the catalogue of Dr. Wallich’s herbarium has become in many cases involved. This has been probably in a great measure the result of the misplacement of labels in the sets of specimens origi- nally distributed. It was almost inevitable that such a displace- ment should take place to some extent, inasmuch as the specimens corresponding to the several numbers in the catalogue were not kept separate in the distribution; on the contrary, they were often laid one upon the top of the other, the numbers being written on loose labels, the juxtaposition of which with the specimens might easily become disturbed. No great harm would result from this in cases where the type collection is available for correcting mistakes. It so happens, however, that Ternstroemiacee is one of the orders in the Walli- chian herbarium in which many of the type specimens are want- ing. They appear to have been sent abroad for description, and THREE IMPERFCTLY KNOWN INDIAN TERNSTR@MIACER. 329 to have been lost or never returned. In addition to this, there is occasional (though not perhaps serious) confusion anongst the remaining type specimens themselves. One of Wallich's names, for example, is represented by an Zxonanthes and a Gordonia, both of which are fastened to the same sheet. It would not have been needful to call attention to these diffi- culties, except for the purpose of pointing out that the determina- tions of the Wallichian types given by Choisy in his * Mémoire sur les Ternstroemiacées? cannot be relied upon. They appear, in fact, to have been based upon the specimens in the Hookerian herbarium, with no reference to the type collection. Choisy fell consequently into very great errors, to two of the most important of which I propose to call attention. Camellia? Scottiana, Wall. Cat. 3668. This plant is represented in Dr. Wallich's type collection by merely a couple of leaves, which had been sent from Munipur in a letter addressed by Mr. D. Scott to Mr. James Kidd, of Cal- eutta. The postscript is pinned to the sheet on which the leaves are fastened ; it states the belief of the writer that they belong to the Tea-plant, and supports this belief by the testimony of a Chinaman staying in the place. It is further suggested that “ it is perhaps the same species that Mr. Gardner sent down from Nypal.” Mr. Gardner appears to have sent specimens both of the true Tea from an introduced Chinese plant, and also of Ca- mellia Kissi, Wall., which, though quite distinct from the Tea- plant, appears to have been distinguished from it at first with some difficulty *. After careful examination I feel satisfied that Mr. Scott’s leaves belong to the Assam Tea-plant ; the late Dr. Anderson appears, from a MS. note in the Kew Herbarium, to have arrived at the same conclusion. The documents relating to the discovery of the Tea-plant in Assam are to be found in the *Journal of the Asiatie Society, vol. iv. The first is a letter from the Committee of Tea-culture, in which the Committee state that they “were acquainted with the fact that so far back as 1826 the late ingenious Mr. David Scott sent down from Munipore specimens of the leaves of a shrub which he insisted upon was the real Tea" (Z. c. p. 42). There is no date attached to the leaves in Dr. Wallich’s herbarium ; but * Asiat. Res. xiii. pp. 428 ef seq. 330 ONIMPERFECTLY KNOWN INDIAN TERNSTR(OMIACE X. it seems probable that they are the same as those to which the Committee of Tea-culture refer. In the Kew Herbarium the lithographed label from Dr. Wal- lich’s catalogue belonging to Camellia? Scottiana is affixed to a sheet on which are specimens of Adinandra dumosa. This is of course the result of accident; but Choisy, without looking any further into the matter, published Camellia? Scottiana as a syno- nym of that plant. It is remarkable also that Dr. Seemann, in his paper on the genera Camellia and Thea*, which contains evi- deuce of his having consulted Dr. Wallich’s herbarium, is yet satisfied with remarking that Camellia Scottiana “is held to be Adinandra dumosa, Jack." I am sure that no person, botanist or not, comparing the leaves of these two plants, could acquiesce in their identity. Ternstremia? coriacea, Wall. Cat. 1453. On p. 158 of Dr. Wallich’s catalogue, Camellia axillaris, Roxb., is given as a synonym of this plant, on the strength of a specimen labelled with that name in Dr. Roxburgh’s handwriting. On the other hand, the plant now known as Gordonia anomala, Spreng., 18 figured in the ‘ Bot. Reg.’ t. 349, as Camellia axillaris, Roxb., and in the text a brief diagnosis, stated to have been taken from a MS. of Roxburgh’s in the Banksian collection, describes Rox- burgh’s plant as having serrulate leaves and asilky calyx. Ihave tried to trace this MS.in the Banksian Library, but without success. It is clear, however, that the plant from Roxburgh’s herbarium, having entire leaves and a smooth calyx, cannot be the plant he in- tended for Camellia axillaris. What that really was, there is perhaps now no material for certainly knowing. If it was identical with Gordonia anomala (which is not improbable), there must have been some mistake as to its original source of introduction into the Calcutta Botanic Garden. For the latter plant is only known from South China ; and if, as stated in the ‘ Bot. Reg., it was obtained from Penang, it can only have been from a garden. Don placed the plant figured in the ‘ Bot. Reg.’ in his genus Polyspora; and Choisy, guided apparently merely by the synonyms, and evidently without comparing the plants, determined Zernstramia ? coriacea, Wall., to be the same as Polyspora axillaris, Don. Now, whatever doubt there might be as to Roxburgh’s plant, there can be none as to Don's; and therefore this determination of Choisy's is certainly * Trans, Linn. Soc. vol. xxii, MAJOR-GENERAL MUNRO ON THE BOTANY OF JAMAICA. 331 wrong. I agree with Dr. Anderson in identifying Wallich's plant with Adinandra acuminata, Korth. Ternstremia? khasyana, Choisy. In his memoir, Choisy describes (p. 20) a specimen in the herba- rium of M. Boissier which had been collected by Griffith in Khasya, and distributed by him with the number 422. Choisy states that he had seen the same plant in the Paris herbarium ; and he refers it as a doubtful species to Ternstreemia under the name of T.? khasyana. Some doubt at once arises as to the correctness of this determination, from the small number of stamens (10) observed by him. The examination of the type specimen, which was most kindly intrusted to me by M. Boissier, proved that Choisy was in error as to the number of stamens, since they are really indefinite. Although, however, the plant possessed a facies remarkably simi- lar to that of many Ternstroemiacee, it certainly did not belong to any Indian species. With the aid of Professor Oliver, I finally succeeded in identifying the rather fragmentary type specimen with numerous examples of Illicium Griffithii, H. f. & T., in the Kew herbarium, which had been collected in Khasya by Grif- fith, and bore the same distribution-number (422). There is a marked agreement of “ adaptive" characters between Illicium Griffithii and Ternstremia japonica, Thb. In fact, acute-leaved specimens of the latter, also collected in Khasyah, closely resemble, in the dried state, the Illicium, both as regards the leaves and the external aspect of the flowers. Extract of a letter from Major-General Mou; C.B., to Mr. BeNTHAM, dated May 29, 1872. My work anent botanizing has been in Jamaica, where I recently spent six weeks, and pretty fairly explored about two thirds of the island towards the west. Generally speaking, it is not so interesting a botanieal country as Trinidad or St. Vincent. The same plants prevail over large areas, caused, I presume, by the general homogeneity of the geological formation, principally lime- stone. There are few showy flowers and creepers ; and I scarcely recollect seeing any variegated plants in the woods. A large extent of country is almost devoid of water, except when the rains are actually falling; and several of the rivers disappear under- ground for miles through some of the crevices and between the great beds ofthe white limestone ; so that dripping rocks and moist LINN. JOURN.— BOTANY, VOL. XIII. 2A 332 MAJOR-GENERAL MUNRO ON THE BOTANY OF JAMAICA. dells, where Flora loves to dwell, are rare. Iam now speaking of the western half of the island. Portlandias, Solandras, and Bro- meliacex (rather numerous species) and the really only very com- mon Orchid Broughtonia sanguinea, together with Amaryllis equestris, in clumps of a hundred, are the principal native hand- some flowering plants that are to be seen. The roadsides are margined for miles with the pretty berries of the Abrus precatorius and the showy yellow flowers of two species of Stigmaphyllum and one species of Echites. The ordinary roadside weeds follow- ing you everywhere are the red and white varieties of Asclepias curassavica and Bidens bipinnatus. I have never seen a single handsome Compositous plant in Jamaica wild. Down the lovely valleys of Acton river the ferns are very grand in size, but not nu- merous in species. I measured a frond of Dicksonia dissectá, which I found to be 14 feet long. Several of the ferns are dimor- phous; and I was surprised to find one species of Anemia the most abundant of all the ferns. I omitted to mention amongst the ornamental native flowering plants Bletia purpurea, of which I have seen five hundred at a time growing within a space of very few yards. Two or three of the Melastomacee are orna- mental for their foliage; and on the mouutains of Newcastle Blakea trinervis, often called Mountain-rose, is very ornamental in flower. Foreign plants contribute greatly to the beauty of Jamaica. "The varied hues of the leaves of the Mangoes are beau- tiful in the extreme ; and I never recollect observing the same effect in India. The Mangoe is spreading itself everywhere, and in fifty years will be the principal tree in the island. The soil of Jamaica is very well suited for many foreign plants; and the hills about Newcastle are beautifully gay with PAaius Tankerville in thousands, Cuphea, Agapanthus, Morea, Tephrosia purpurea, Alpinia nutans, the variegated form of Abutilon Thompsoni spread- ing abundantly from seed, Calla ethiopica, Amaryllis, myriads of Brugmansia, Roses (very fine), Furze, Strawberries, &c. They are beautiful to look at; but you can scarcely recognize among them a native plant. In the Blue Mountains there is always moisture ; and there the vegetation of foliage is very glorious. On the fern- walk above Newcastle, in an hour’s time, I collected seven or eight different species of Tree fern, including Pteris aculeata, or an allied species ; Hymenophyllum and Trichomanes are there also in abund- ance. MR. F. CURREY ON A NEW GENUS IN THE ORDER MUCEDINES. 333 On a new Genus jn the Order Mucedines. By FREDERICK Curkzr, M.A., F.R.S., Sec. L.S. Puate VII. [Read June 20, 1872.] Tue plant here described was brought to my notice by Dr. R. O. Cunningham, F.L.S., who had received an account of it from Dr. D. D. Cunningham, of Calcutta, where the fungus is found in the rainy season covering the flowers of Hibiscus rose sinensis. I have since corresponded with Dr. D. D. Cunningham upon the subject, and received from him further information, accompanied by some admirable drawings; and I am thus enabled to lay before the Society the following account of the fungus in question. The mycelium traverses the tissue of the fading corolla of the Hibiscus, and is only scantily jointed. The fertile threads are erect, unbranched, and continuous. These threads are swollen at the apex; and from the swollen apex proceed numerous shortly stalked pyriform cells, which ultimately form the funnels here- after to be noticed. The spores originate from the above-men- tioned cells in the manner shown in fig. 6 (Pl. VIL), which represents a cell very highly magnified, with the young spores attached. At this stage of growth there is no differentiation of one part of the cell from the other; but shortly afterwards a line appears cutting off the upper portion of the cell, to which the spores are attached, as shown in fig. 5. This upper portion ultimately collapses and sinks down, or falls inwards, as it were, carrying with it the ripe spores. The subsequent fate of this collapsing membrane is not very clear; but it seems probable that it decays, and thus sets free the ripe spores. The result, however, is that the spores eventually become detached, and the cells, which were originally pyriform, assume the shape of stalked funnels still attached to the capitate portion of the fertile thread. This condition of the fungus is shown in fig. 3, where the spores have almost disappeared, a few detached ones being still visible in some of the funnels, and one of the latter being still quite full of spores. It happens occasionally that the margins of the funnels, instead of being entire and sharply defined, exhibit irregular torn fragments of membrane, being obviously the re- mains of the collapsed upper portion. On one occasion Dr. Cun- ningham observed a funnel from which the spores had escaped, leaving a bladder-like protrusion, which had shrunk away from the funnel, but had not become detached. From the appearance 384 MR. F. CURREY ON A NEW GENUS IN THE ORDER MUCEDINES. presented in this instance, and which is shown in fig. 8, it would seem that the bladder-like protrusion must be the inner membrane of the capitate cell, from which the upper portion of the outer membrane has become detached -by a sort of circumscissile dehis- cence. The ripe spores of the fungus are obovate and of a deep . madder-brown colour. With regard to the classification of this interesting plant, it is not easy to find a fitting place forit. It belongs, of course, to the order Mucedines ; but there is not in that order, as far as I know, any genus to which it is at all nearly related. The peculiar cha- racter is the differentiation of the upper and lower portions of the sporiferous cells and the formation of the funnels by the col- lapse of the upper portion. I propose for its reception a new genus, to be called Cunninghamia, which may be thus characterized :— CUNNINGHAMIA, novum genus. Mycelium repens, sparsim septatum : flocci fertiles simplices, erecti, ad apices clavato-incrassati, sporophoris ornati: sporophore primum pyriformes, dein, parte superiore subsidente, infundibu- liformes : spore juniores hyaline, mature rubro-fusce, capitulis densis stipate, dein in infundibulis acervatim cumulate. Unica species :— CUNNINGHAMIA INFUNDIBULIFERA, Curr. Hab. Ad petala marcescentia Hibisci rose sinensis, in horto botanico apud Calcuttam tempore pluvio. EXPLANATION OF THE PLATE. Fig. 1. A spore germinating, and a portion of the mycelium from the tissues of the corolla of Hibiscus rosa sinensis, magnified 330 times. Fig. 2. A fertile thread, viewed by reflected light, magnified 65 times. Fig. 3. A fertile thread past maturity, most of the spores having fallen and the sporophores having collapsed and assumed the shape of funnels, magnified 103 times. Fig. 4. Three funnels,with a small portion of the clavate apex of the fertile thread. The fragments of membrane adherent to the edges of the funnels appear to be the remains of the upper, collapsed portion of the sporophores. Fig. 5. Animmature sporophore, with young spores, showing the line of demar- cation at which the sporophore eventually collapses, magnified 330 times. Fig. 6. A similar sporophore before the formation of the line of demarcation, with five young spores attached, magnified 700 times. Fig. 7. A ripe spore, magnified 330 times, Fig. 8. A sporophore past maturity, exhibiting a bladder-like protrusion, being apparently the inner membrane of that portion of the REN TR: which was aboye the line of demarcation. MR. G. BENTHAM ON COMPOSITA. 335 Notes on the Classification, History, and feographical Distri- bution of Composite. By Grorar Bentham, F.R.S., P.L.S. [Prares VITI. to XT.] CONTENTS. Page Page F Introduction... eas 336 | 5. Helianthoidee ............ 431 II. Comparative value of generic | 6. Helenioidez ............... 446 characters ese 344 | 7. Anthemides e 450 1. Sexual differences......... 344 | 8. Senecionideze. .. ............ 455 2. Di- and trimorphism ... 348 | 9. Calendulaceæ ............ 463 3. Differences in the Pistil 348 | 10. Arctohdem .- — aa 464 4. in the Fruit......... 352 | 11. Cynaroidem ............... 466 5. —— in,the Andrecium 358 | 12. Mutisiacem ............... 71 6. in the Corolla ...... 363 | 13. Cichoriacese ............ + 475 7. —— in the Calyx ...... 366 | B. Comparative antiquity of 8. in the ultimate In- | races in Composite ...... 481 florescence and bracts... 966 | C. Present regions or chief 9. in the Foliage ...... 310 centres or areas of the 10. —— in the Habit, sta- principal races of Com- ture, and general inflo- POSER. os eas 484 FESCENCE .......... sss. Y | 1. General repartition of 11. in the Geographical Composite between the distribution ............ 373 New and the Old World 486 TIT. Sketch of the primary divi- Table 1. Repartition of Gen- sions of the order ...... 374 era in the two divisions 487 1. Vernoniacese ............... 374 Table 2. Ditto of Species ... 487 2. Eupatoriacem ............ 375 Table 3. Tropical connexion 3. Asteroidem .:............. 376 of the two divisions as 4. Tuulpidess e 377 indicated by identical or 5. Helianthoidee ............ 379 closely allied genera ... 490 6. Helenioidez............... 381 Table 4. Ditto as indicated by 7. Anthemidese ............... 382 identical species ......... 493 8. Senecionidem ............ 383 Table 5. Extratropical north- 9. Calendulaces ............ 385 | ern eonnexion between 10. Arctotidex ............... 385 the two divisions......... 495 11. Cynaroidesm ............... 387 Table 6. Connexions between 19. Mutisiaces ............... 388 | South Americaand South 13. Cichoriaces .............. 389 | AM ioo alb 501 IV. History and geographical dis- | Table 7. Ditto between S. tribution ceee e 390 | America and Australia 503 A. Distribution of the tribes i Table 8. Cosmopolitan genera 504 and principal genera ... 393 | 2, Separate distribution of 1. Vernoniacez ............... 393 | Compositz into regions 506 2. Eupatoriacesm ............ 400 Table 9. Number of Compo- 3. Asteroidem .............-- 402 sitze in the several Ame- 4, Inicios |... ee 414 | rican regions ............ 507 LINN. JOURN.—BOTANY, VOL. XIII. 25 336 MR. G. BENTHAM ON COMPOSITA. Page | Page Table 10. Summary.of Ame- 7. Connexions be- rican distribution ...... 517 tween distant Notes: 1. Mexican Region... 517 Old-World R ... 551 2. United-States R... 519 Insular Regions ............... 554 3. West-Indian R.... 521 | 1. Sandwich Islands (tab. 4. Andine Region... 522 IS 555 5. Brazilian Region. 524 2. Galapagos (tab. 14) ... 556 6. Chilian Region ... 526 3. Juan Fernandez, Masa- 7. Connexions be- fueram, &e...... es 557 tween distant American 4. South-Sea Islands...... 558 Fegions feces 527 5. Atlantie Isl. (table 15) 559 Table 11. Number of Com- | 6. St, Helena and Tristan posite in the several | d'Acunha o 563 Old-World Regions ... 529 7. Mascarene Isl.(tab. 16) 564 Table 12. Summary of Old- "8. New Caledonia......... 566 World distribution...... 539 9. New Zealand (tab. 17) 567 Notes: 1. Mediterranean R. 539 D. Colonizing Composite or 2. Europxo-AsiaticR. 542 Introduced species...... 568 3. Tropical African R. 544 Table 18. Escapes from cul- 4. Tropical Asiatic R. 546 UTI E1816) ors a 569 5. South-African R... 547 Table 19. Weeds of cultiv... 570 6. Australian Region 549 Conclusion (csc. re cee 576 I. IxTrRODUCTION. THE Composite are at once the largest, the most distinct, and the most uniform, and therefore the most natural, of all orders of Pheenogamous plants. Nearly ten thousand known species are separated from each other by characters most of which are usually considered as only of secondary importance; and I cannot recall a single ambiguous species as to which there can be any hesita- tion in pronouncing whether it does or does not belong to the order. The very few cases where species have been erroneously referred to or excluded from it have been the result of conjec- tural determinations of imperfect specimens, or of gross ignorance on the part of the observer. The andrecium, gynecium, and fruit, as to all the essential characters of number of parts and rela- tive position, the seed and its embryo in every particular, are absolutely uniform throughout the order; or in the very few cases of a slight variation (as, for instance, in the shape of the cotyledons) the differences are no more than specifie, varying in one and the same small genus. To distribute, therefore, these ten thousand species into thirteen tribes and above seven hundred INTRODUCTION. 337 and sixty genera, we are compelled to derive our characters from inflorescence and its rhachis and bracts, from the pappus or abnor- ma] development of the rudimentary calyx, from the shape of the corolla, from sexual abortions, from appendages to the anthers, from the external form or appendages of the style-branches, and from very slight variations in the external form of the fruit, many of which, in other orders, are scarcely reckoned of more than specific value. In Linneus’s artificial sexual system the sexual characters were necessarily taken as of primary importance, and the Composite, forming the chief portion of the class Syngenesia, are divided into four orders :— Polygamia «qualis, with all the flowers (or, as it is more convenient to call them, the florets) of each head herma- phrodite and fertile ; Polygamia superflua, with the floréts of the circumference female, those of the disk hermaphrodite, and all fertile; Polygamia frustranea, with the florets of the circumference barren, those of the disk hermaphrodite and fertile ; and Polygamia necessaria, With the florets of the circumference female and fertile, those of the disk hermaphrodite but barren. To these Linnzus added a fifth order, which, notwithstanding its analogous name Polygamia segregata, is not founded on any sexual distinction, but on inflorescence only, being characterized by the numerous uni- florous heads crowded on a common receptacle. Such an arrangement proved, however, to be purely artificial ; the strikingly different groups, for instance, of the Hawkweeds' and Thistles, are amalgamated with a few others with which they have evidently no other connexion than as members of the same family, into the first Linnean order; and notwithstanding the endeavours of most subsequent synantherologists to maintain as much as possible the value originally attached to these sexual distinctions, they have felt repeatedly compelled to unite iato single or closely allied genera, species which would, on these Linnean principles, be distributed over most or all of his orders. Jussieu and Ventenat, following up the ideas broached by Vaillant, considered the Syngenesia polygamia of Linnsus as a class, under the name of CowPosrr x, dividing it into three families or natural orders—Cichoriacee, Cynarocephale, and Corymbifere,— an arrangement which, up to the date of De Candolle's *Prodromus,' was generally followed by the latter, as well as by most other French botanists—considering, however, the clase as a natural order 222 338 MR. G. BENTHAM ON COMPOSITE. under the Jussieuan name of Composite, and his three families as so many suborders, to which Lagasca and De Candolle added a fourth under the name of Labiatiflore. Of these suborders the Cichoriacee, with the corollas all ligulate, have ever since maintained their ground as the most natural as well as the most accurately defined group in the whole order; the three others have been very variously modified or subdivided into tribes, none of which can as yet be considered as separated by any absolute characters. Henri Cassini was the first who undertook a general revision and redistribution of Composite. Leaving the Cichoriacee un- disturbed as a distinct tribe (changing the name, however, to Lactucee), he rearranged the great mass of tubuliflorous Composite upon new principles. His long series of articles, some of them first sketched out in the ‘ Bulletin des Sciences de la Société Phi- lomathique,’ were distributed over the sixty volumes of the ‘ Dic- tionnaire des Sciences Naturelles; some of the more general ones collected in his ‘Opuscules Botaniques; the whole published between the years 1816 and 1834. These papers include a large number of very valuable observations, the result of the study of as many species as he could obtain in a living state, or could examine in the herbaria of Jussieu and others. He was the first to make use of the modifications of the style and anthers in the general systematic arrangement of the order; and he clearly exhibited the funetions of the collecting, or, as he not inaptly terms them, sweeping hairs and papille (poils collecteurs, poils ba- layeurs, papilles balayeuses of Cassini, Fegehaare of Hildebrand). His table, or, rather, map, ofthe tribes (plate 1 of the * Opuscules ") shows a just appreciation of the natural affinities of the order, and of the principal groups of which it is composed, and in some respects, as in the tribe of Inulee, gives a better arrangement than those of Lessing and De Candolle. Unfortunately, however, in working out the details of the genera in the ‘ Dictionnaire,’ he indulged in an enormous and useless multiplication of generic names, which only tended to throw the nomenclature into confusion, and cast a slur upon all his labours. Wherever he observed a slight difference in the involucre, pappus, or general aspect, or could not readily identify an imperfect specimen, an engraved figure, or a description often incorrect, he at once set it down as a new genus, and has thus, more than any other botanist of equal ability overloaded the science with useless synonyms. So recklessly, indeed, did he give way to this mania of coining new names, that INTRODUCTION. 339 he on many occasions proposed two, or even three, for the same genus, leaving future botanists to take their choice. Robert Brown, in his memoir in the twelfth volume of our Transactions, with his usual accuracy, elucidated many obscure points in the structure and affinities of several genera of the order, here as elsewhere indicating even much more than he expressec in detail, but did not touch upon the general arrangement or distribution of the genera. David Don, in the years 1828 to 1832, published in the Edin- burgh New Philosophical Journal and other periodicals, as well as in our own Transactions, various monographie papers on Com- posite, chiefly on Cichoriacee and on South-American Mutisiacew and a few others. In his new genera he made use of some neglected characters derived from the venation of the corolla &e. but he seems to have consulted but very little the works of his predecessors, and to have been quite unaware of the important peculiarities of the style pointed out by Cassini. Lessing, after some preliminary papers in the ‘ Linnea,’ published, in his ‘ Synopsis Generum Compositarum,’ a new arrangement of the order, founded, still more than Cassini’s, upon the modifications of the style, proposing many alterations in Cassini’s groups, which cannot always be considered as improvements, except in so far as they resulted from a command of more ample materials. Cassini’s tribes and subtribes are generally natural, although his genera are often species only. Lessing's tribes and subtribes are very technical, whilst the main characters are not always sufficiently absolute to give them the advantages of an artificial classification ; and in his genera he often relies too much on the pappus, the variations of which are less in conformity with general differences than those of almost any other organ. De Candolle, when he arrived at the Composite for the * Pro- dromus,' had before him far more extensive collections than any of his predecessors. He had in former years worked out mono- graphs of some portions of the order ; and he now set to work, with his usual ardour and methodical mind, to the diagnosis and systematic arrangement of a mass of species nearly equal in num- ber to those of the whole vegetable kingdom known in the days of Linnsus. But a severe attack of illness came on shortly after he had commenced ; and on resuming work after the lapse of a couple of years, although his ardour and perseverance remained, he was no longer so well able to grapple with complicated difficulties. 340 . MR. G. BENTHAM ON COMPOSITE. He followed Lessing too closely where his own more perfect specimens might have shown errors; and in the numerous new genera he proposed, he, in too many cases, neglected the verification of the tribual characters which from their aspect he presumed them to possess. Thus it is that identical species, for instance, are not unfrequently repeated in different tribes, and that the most closely allied genera are often widely separated without cross references to indicate their connexion, the general result being that further researches have occasioned greater changes in proportion in this than in any other part of the ‘ Prodromus.’ C. H. Schultz Bipontinus, devoting many years of his life exclu- sively to this order, considered himself, and was regarded by many others as the great synantherologist of his day; but he did not live to work out any general system. In his numerous detached papers he modified the circumscription of many genera, corrected errors, and consigned to print many valuable observations ; but he seemed always too much in haste to bring out something new, to divide or to consolidate old genera, changing long lists of names of species of which he had only examined one or two, to alter upon slight grounds the scale of relative value previously given to generic characters, without, however, relinquishing the idea that there exists in nature such a scale possessing a high degree of fixity, and in general to affix his own stamp upon all future synantherological labours. Where he has taken time to work out his monographs in detail his observations have appeared to me reliable for their accuracy; but where he has proceeded hastily it is difficult to follow him. His determinations, for in- stance, of Mandon’s Bolivian, of Riedel’s and Langsdorff’s Brazi- lian, of Liebmann’s Mexican Composite are replete with misnomers. There are other points also in which either I cannot quite agree with him, or from which I should differ widely. His multiplica- tion of species is sometimes carried very far. His reliance chiefly upon the form of the achene for generic distinctions is in some cases a great improvement, in others carried so far as to become purely artificial. I must agree with him in his high estimation of the labours of Cassini; but that does not appear a sufficient reason for adding one more to the numerous names already given to the order. Even for those who maintain that all natural orders must be named after some one of their genera, with the affix of acee, there is Lindley's name of Asteracee, which has the right of priority over that of Cassiniacee proposed by Schultz, besides that INTRODUCTION. 341 the genus Aster, being so much better known and (slightly modified) so much more universally distributed than the little-known local genus Cassinia, is much better suited for a so-called ordinal type. Of all the modern contributions to the study of Composite none are more important for the accuracy of observation and the due appreciation of characters and affinities than those of Asa Gray. His views (first exemplified in the 2nd volume of the ‘Flora of North America,’ further carried out in a long series of memoirs or detached papers on the Composite of various collec- tions published in the * Smithsonian Contributions,’ in the reports of various American exploring or surveying expeditions, in the Proceedings of the American Academy, of the Boston Society, in Hooker's Journals of Botany, and other periodicals) may always be implicitly followed without any danger of being led into error, although sometimes a difference of opinion may exist upon such minor points as the generic or subgenerie value to be given to a group. The only real difficulty in his case arises from the dis- persion of his papers in such a large number of publications, not always within reach of the generality of botanists, and some of which it is scarcely possible not to overlook. There are many others whom I might mention as having con- tributed more or less beneficially or prejudicially to our know- ledge of Composite; but that would require the enumeration of the greater number of modern systematic botanists. Composite are so easy to examine, that botanists of very little experience are readily led into the belief that they have mastered every thing relating to a specimen they are examining; and if they discover some point not quite agreeing with the technical characters of the genus it is presumed to belong to, they are at once ready to establish a new one ; and there are few orders which have been so universally dabbled with, or where there have been so many repetitions of species, of genera, or of observations. The litera- ture of Hieracium and of Aster, for instance, is almost, if not quite, as bulky as that of Rubus and of Rosa; and generally the synonyms of this part of our‘ Genera Plantarum’ are nearly twice as numerous as the adopted generic names. There are, however, among the numerous more speculative naturalists whom the promulgation of the Darwinian theories have called into action, two who require notice as having specially taken up the subject of Composite with reference to those pro- 342 MR. G. BENTHAM ON COMPOSIT X. visions for cross fertilization originally pointed out by Cassini, and with more or less of recognition forming the basis of the systems of classification of Lessing and De Candolle, but more or less neglected or ignored by Don, by Schultz Bipontinus, and by most of the minor more specific synantherologists. Of the two theorists I allude to, Prof. F. Hildebrand, of Bonn, and Federigo Delpino, of Florence, the former has published, in the last volume of the ‘Nova Acta Nature Curiosorum,’ elaborate observations, accom- panied by excellent illustrations, of the fecundating apparatus of thirty species of Composite belonging to most of the principal groups of the order. In the general considerations which follow, he endeavours to show that in Composite, at least, all observations indicate that unisexuality where it exists has proceeded from hermaphroditism, and that the primitive parent of the order had capitula consisting entirely of protandrous hermaphrodite florets— a conclusion which may be a correct one, but for which the data at hand are wonderfully few. The same writer has a paper in the last number of the ‘ Botanische Zeitung’ on the meaus of dis- persion supplied by the fruits of Composite, upon which I may have to make some observations when speaking of geographical distribution. Delpino, in his * Studi sopra un legnaggio anemofilo delle Com- poste, amidst many shrewd and instructive observations on the dichogamic arrangements of Composite in the comparatively few species he has had the opportunity of studying, supplements them rather largely from the sources of imagination. Attaching the greatest importance to the Darwinian distinction he has on various occasions worked out, between anemophilous and zoidiophi- lous plants (those which effect cross fertilization by the agency of winds or by the agency of insects), he considers that evidence derived from this character alone is suffieient to prove descent and affinity, without taking into account the numerous cases alluded to by Darwin, and some of which are mentioned by Delpino himself in this very paper, where important variations in this respect occur in different species of one and the same genus. His genealogical tree of Artemisiacee, from Campanulacez down to Xanthium spinosum, his statements, as proved facts, that Campanulacee transmitted hermaphroditism through Lobeliacee to their descendants the Composite, that Composite inherited pro- terandry from Lobeliacee, but acquired in many cases unisexuality during subsequent generations, are mere conjectures. So in pro- INTRODUCTION. 343 posing his tribe of Artemisiacee to include Ambrosiee and exclude Tanacetex, a tribe which he characterizes as irrefutably natural, it appears to me that by relying for it solely on one character (the anemophilous fertilization) his classification becomes as artificial as when Linneus relies solely on the repartition of the sexes, Lessing on the shape of the style-branches, Schultz Bipontinus on the shape of the achene, or others on the pappus alone. As to his idea that heterogamous capitula with fertile female ray-florets and sterile disk-florets may be regarded as simple hermaphrodite proterogynous flowers descended from inflores- cences of proterandrous flowers, this is surely but little more than a play upon words. With regard to the part I have myself taken in the elaboration of this great order, I may observe that, long before I was called upon to undertake it for our ‘Genera Plantarum,’ I had had repeated occasion to test the value of the labours of my prede- cessors, and for various publications had examined in detail the Composite of Europe, of British Guiana, of tropical Africa, of China and allied Indian forms, and especially of Australia; and I have now, with the aid of the rich stores and extensive library collected together at Kew, and liberal assistance in the way of loan of specimens or notes on typical specimens in Continental herbaria, communicated by my friends M. Decaisne and M. Cosson of Paris, Dr. Ascherson of Berlin, and Dr. Fenzl of Vienna, been enabled to examine specimens and compare the original descriptions of an immense majority of the published genera, subgenera, or sections, as well as of numerous species which have been supposed to present some anomaly, or which by their aspect seemed to suggest some peculiarities which might affect the genericcharacter. Notwithstanding, however, the lengthened period which I have devoted to this tedious labour, I feel that there is yet much left to be done to future synantherologists who can undertake throughout a specific monograph, such as I have only been able to do with regard to the Composite of some countries or of a limited number of genera. There are also cer- tain characters, and some of them evidently important, which our herbarium specimens often do not supply. Perfectly ripe achenes are often wanting in whole genera; and microscopical characters, such as the form of the pollen, upon which much stress has been laid of late years, have been observed in too small a number of species to ascertain their real connexion with general affinity. In LINN, JOURN.—BOTANY, VOL. XIII. 20 344 MR. G. BENTHAM ON COMPOSITR. general my chief labour has been the testing and verifying or reconciling the observations of others, although this has always been preceded by the examination of specimens and drawing up my own generic character, afterwards modifying it when neces- sary in points elicited by others which I had at first overlooked. The principal changes I have proposed in the general methods of Lessing and De Candolle were determined upon and worked out long before I was aware that they were in a great measure a return to that of Cassini. The confusion which his multiplication of names had produced, and the unusual terminology of his descrip- tions, had excited in my mind a prejudice against him, until, after completing my work of detail, I came to study his generalizations, which showed how much better his views of affinities coincided with mine than those of his successors ; and I have since had the satisfaction to learn that the principal of these changes I have made have also met with the approval of such careful observers as Asa Gray and Ed. Boissier. It is scarcely necessary to add that in this, as in other parts of our ‘Genera Plantarum,’ any important changes which either Dr. Hooker or myself have proposed in the orders we have respectively undertaken, have always been after consultation and in concert with the other. . II. Comparative VALUE OF GENERIC CHARACTERS. Before entering into a general outline of the main divisions we have adopted, some explanation is required of the principles upon which we have conformed to or departed from the systems . of our predecessors ; we must form to ourselves some idea of the comparative value of the various characters put forward by different synantherologists ; and for this purpose it will be necessary sever- ally to consider them in some detail. We may take them in the following order :—1. Sexual differences ; 2. Di- and trimorphism ; 3. differences in the Pistil, 4. in the Fruit, 5. in the Androecium, 6. in the Corolla, 7. in the Calyx, 8. in the ultimate Inflorescence and bracts (2. e. in the capitulum and its receptacle, involucre, and palez), 9. in Foliage, 10. in Habit, stature, and general inflores- cence, and, 11. in Geographical distribution. l. Sexual Differences. Characters derived from this source were, as already observed, considered of the highest importance by Linneus, who founded on them his primary divisions of the order. Subsequent syste- matists have gradually placed them lower in the scale, but yet YALUE OF CHARACTERS. 845 have, generally speaking, regarded them as absolute for the di- . stinction of genera. Schultz Bipontinus, however, in some of his later generic changes, has shown a disposition to neglect them, apparently from having observed their little accordance in certain cases with generic groups he was disposed to form on other grounds. But I have been unable to ascertain how far he generalized this degradation of the character. My own observa- tions would lead me to conclude that, like other characters, sexu- ality varies in value in different tribes and in different genera, in a few cases absolute even in tribes, often of considerable impor- ` tance in genera, but often also specific only, or at most available for sections or artificial groups of species. These sexual differences relate to those of the individual flowers or florets within the head, and those of the flower-heads taken generally. With regard to the individual florets, it is usual to distinguish four kinds—hermaphrodite, male, female, and neuter; but the sterility of the pistil is often so uncertain or variable in the anther-bearing flowers, in which it is never absolutely deficient, that I have found it much more convenient to designate as hermaphrodite all florets having perfect anthers, whether their pistil be susceptible of fertilization or not, distinguishing them as fertile or sterile—and as female all florets in which the anthers are abortive or deficient and the style is present; the neuter florets, reduced to a corolla with a rudimentary scarcely distinet ovary, might be classed in the same category as the females, as the abortion of the style is sometimes gradual or uncertain. Thus reducing the kinds of florets to two when both occur in the same head, the hermaphrodite ones invariably occupy the centre, the females being placed in one or more concentric rows in the cir- cumference. Nuttall bad indeed published a genus which he characterized from the supposed singular inversion of this position, the females being, as he believed, in the centre, surrounded by males, and gave it, therefore, the name of Parastrephia; but upon inspecting the original specimen in Nuttall’s herbarium (a mere fragment gathered by Curson near Arequipa) it appeared to me that he had been deceived by some degree of unisexuality in the flower-heads, some having nearly or perhaps all the florets female and therefore central as well as cireumferential, whilst one of the heads appeared to be entirely hermaphrodite. There are not heads enough on the specimen to verify the fact; but I at 346 MR. G. BENTHAM ON COMPOSIT E. once recognized the plant as one of which we have good speci- mens from the same locality, and which Meyen placed in Baccharis (B. phyliceformis, Meyen) on account of its general affinity, but which Walpers invita natura transferred to Vernonia on account of the pappus. In these specimens the relative posi- tion of the hermaphrodite and female flowers is normal, although Nuttall’s genus can be sustained as distinct from Baccharis on other grounds. Sterility may be more or less perfect either in the central florets, extending outwards more or less to the greater portion or to the whole of the hermaphrodite ones, or in the female florets, but extending never, as far as I am aware, within the outermost row; or, in one and the same head, both the outer female and the innermost of the hermaphrodite florets may be sterile. In three of the principal tribes, Vernoniacee, Eupatoriaces, and Cichoriacez, uniform hermaphroditism of the florets is, I believe, quite constant, and the sterility of any of them (other than accidental) very rare and exceptional. So far, therefore, Linneus’s order of Polygamia æqualis is maintainable, the pre- sence of any female florets at once excluding from either of these tribes any plant supposed to belong to them; but the character goes no further as a tribual one. In Cynaroidex an outer row of female or neutral florets occurs in a few genera, but is not con- stant even in the same genus; in Mutisiacex it is more frequent and more constant. In the remaining eight orders the presence of one or more rows of female florets is the rule, but with ex- ceptions, sometimes in single species of large genera where it is usually constant, sometimes in the majority of species or in whole genera, and in two subtribes of Inuloidez very prevalent or quite constant. In allthese cases we are therefore obliged to be very eautious in making use of the homogamous or heterogamous flower-heads as an absolute generic distinction. Still less value can usually be attached to the sterility of the inner or outer florets, although in some cases it appears to be positively generic or even subtribual. The central hermaphrodite and outermost female row of florets are constantly sterile in some Calendulaces, all the hermaphrodite florets constantly sterile in the subtribes Milleriew, Melampodiex, and Ambrosiew, of Helianthoidex, and in some genera of other subtribes ortribes; in other genera the greater or less sterility of the central florets is of no more than specific value. The sterility of the circumferential florets (reduced in that case YALUE OF CHARACTERS. 847 to a corolla with a rudimentary or small abortive ovary without style or stamens, and therefore called neuter) has been considered by many synantherologists an absolute generic indication; and although now generally abandoned in the case of the enlarged neutral florets of some Cynaroidee, it has been strictly adhered to by Lessing, De Candolle, and others in the case of Asteroidee, Helianthoidew, and others with radiate flower-heads. But in many genera it appears to establish a purely artificial distinction ; it separates from large and natural genera a few species or a single one without any other peculiarity but what every one admits is purely specifie; and besides it is in these cases not always strictly constant. I would therefore, with Schultz Bipon- tinus, reunite Galatella with Aster, Delucia with Bidens, Leptosyne and others with Coreopsis. In some Helianthoidez, however, it prevails through so large a number of species, otherwise evidently congeners, that I have adopted it among the essential generic characters of Helianthus, Viguiera, Tithonia, Oyedea, Gymnolomia, Rudbeckia, £c.—and even as the sole generic distinction between Aspilia and Wedelia, Actinomeris and Verbesina—but more as a matter of convenience than of conformity to nature, in large groups of species where no better principle of subdivision has yet been proposed. This sterility of the ray-florets is yet more con- stant in the majority of the genera of the tribe Arctotidez, though here, again, there are a few, such as Arctotis itself, where they are fertile. Unisexuality of flower-heads is sometimes a constant generic or subtribual character; the heads are, for instance, constantly moncecious in Ambrosiew, constantly dicecious in Petrobiex, Tarchonanthese, Baccharis, Lycoseris, Moquinia, &c. In these cases the male heads always consist of hermaphrodite sterile florets, with the anthers perfect, the style always present, but the ovary abortive and usually reduced to a mere rudiment; in the female heads the pistil is perfect, the anthers often present, but free and without pollen, or reduced to small rudiments or entirely deficient. In several Inuloidez (Plucheines or Gnaphalies) the unisexuality of the flower-heads is less perfect and more incon- stant, the male heads having occasionally a few female fertile florets in the circumference, the females one, two, or more her- maphrodite and sterile ones in the centre, and the proportions varying sometimes from species to species or from individual to individual. 348 MR. G. BENTHAM ON COMPOSIT X. 2. Di- and Trimorphism. Di- and trimorphism in individual florets is usually connected with unisexuality, the female corollas being either more slender than the hermaphrodite, or with a more expanded irregular or ligulate limb, or those of the outer rows ligulate and of the inner rows slender; or they may be reduced to a mere rudiment or entirely wanting ; and the pappus may also differ, that of the female florets being often reduced from that of the hermaphrodites or entirely wanting; but the value of these differences 1s usually no greater than as an indication of sexuality. The reduction or alteration of the pappus in the female florets, so often taken advantage of as a generic distinction, very frequently produces a purely artificial separation of a very few or of a single species from a genus other- wise natural, and is now generally abandoned by A. Gray and others. On the other hand, with regard to dimorphism in the corol- las, although the presence or absence of female florets is often not even of generic value, the form they assume when present, whether tubular like the hermaphrodites but more slender, or ligulate, is sometimes of subtribual importance, as in the case of the separa- tion of the last four subtribes of Inuloides from the first six, of which this is one of the principal characters. In unisexual flower-heads the female corollas are always more or less reduced, never having the expanded or ligulate limbs so frequently observed in heterogamous heads. Dimorphism in the flower-heads independent of the sexes is very rare in Composite, and, as far as hitherto observed, only of specific value. Linnzus had already informed us that in Gerbera (Anandria) bellidiastrum the vernal flower-heads are heteroga- mous and radiate, and the autumnal ones homogamous and dis- coid; and analogous dimorphisms have since been observed in a very few scattered species. Delpino has also observed a certain degree of dimorphism in the female flower-heads of two species of Franseria, which on that account he establishes as two distinct genera—a separation in which it is very difficult to concur. 9. Differences in the Pistil. The ovary and ovule, as to their structure, insertion, position, and other relations to other parts of the flower, are absolutely uniform throughout the order, and afford no clue to generie or tribual distinctions. Nor is the style itself more available in the case of the female florets, where its sole functions are those de- YALUE OF CHARACTERS. 849 volving upon it as a portion of the female organ—the receiving the pollen and conveying it to the ovary. These styles of the female florets are uniformly divided into two equal more or less stigmatie branches, glabrous without and papillose inside, which may occasionally vary in length or thickness, but only slightly so, and very rarely, as far as I have been able to observe, so as to give any but a very slight generic clue. It is the style of the hermaphrodite florets, in its usually principal and often sole function of sweeping the pollen out of the antheral tube, that presents those external differences which by Lessing and De Candolle have been taken as absolute tribual characters, and which are in fact, generally speaking, important and useful, but which are also liable to numerous exceptions. As an instance of the confusion resulting from the use made of this character, I may point out that in the illustrations of the styles of eight of the principal tribes given in Lindley's ‘ Vegetable Kingdom,’ p. 703, two (n. 2 and 3) have not the forms characteristic of the tribes they represent, and, indeed, are taken from genera which, although included by De Candolle in Eupatoriaces and Asteroidew respec- tively, have been erroneously there placed, whilst three others (n. 5, 6, and 7) are far from being typical of the majority of their cotribuals. Referring to the diagrams which I have here given, Plate X., there is no doubt that the styles of the large genera Vernonia (fig. 2), Eupatorium (fig. 3), Aster (fig. 4), Senecio (fig. 6), and Carduus (fig. 8) are prevalent also in a considerable number of genera closely connected with them on other accounts ; but, on the other hand, some of these forms are to be met with in genera naturally far removed from them, or are not in closely allied genera, or, again, are so connected with each other by intermediate forms as to render them in some cases useless even as artificial characters. The Vernonia style (fig. 2), with its long, slender, almost acute branches, nearly equally hirsute throughout, with the stigmatic series scarcely prominent on the inner surface towards the base, is, I believe, constant in the 35 genera and near 500 species of the tribe; but it is also to be met with in a few genera which on every other account must be placed either in Asteroidee (e. g. Chrysopsis), in Inuloidex, or in Senecionidez (e. g. Liabum and Gynura). The Eupatorium style (fig. 3), with long, obtuse or club-shaped branches only minutely papillose instead of being hirsute, but with the slightly conspicuous stigmatic series of 350 MR. G. BENTHAM ON COMPOSITE. Vernoniaces, is not only constant in the tribe of 33 genera and above 700 species (although slightly modified in Mikania, where it is less obtuse and more slender), but is also, I believe, exclusive, or there is only a slight approach to it in a few Inuloideæ or Mutisiacex. The Asteroid style (fig. 4), with flattened branches, the marginal stigmatic series very prominent but not reaching the extremity, which consists of a so-called appendage, long or short, broad or narrow, acute or obtuse, and papillose or shortly hirsute all over, is very prevalent in the tribe as now limited; but occa- sionally the appendage is so short as to bring the style nearly to that of the Senecionidez or of the Inuloides, and sometimes a truly Asteroid style, as well as numerous approaches to it, may be ob- served in various genera of Helianthoidew, Helenioides, and Senecionidem. The Jnula style (fig. 7), with the stigmatic series reaching to the end of the branches, or nearly so, without append- ages as in Senecio, but the branches rounded, not truncate, at the end, prevails in several subtribes of Inuloidee, but in others passes into the true Senecio style. It is also to be occasionally met with in Arctotidee and Mutisiacez, as well as in genera closely allied to Senecio. The Senecio style (fig. 6) has flattened branches like that of Aster, but narrower and usually recurved; and the stigmatic series reach the extremity, which is truncate and fringed with hairs. This style is uniform in nearly the whole of the 900 species of Senecio and in several allied genera; but even in Senecio itself the extremities of the branches are occasionally rounded, or form an exceedingly short appendage (in this tribe called cone), which, in other genera very closely allied to Senecio, lengthens out (as in fig. 5) even into the Vernonia form (fig. 2), the stigmatic series also gradually becoming less conspicuous; and the true Senecio style is observable in numerous genera which on other accounts must be placed in Inuloidee, Helianthoidex, Helenioidew, Anthemide:e, or Mutisiaceew (Nassaviee). The Carduus style (fig. 8), with a so- called articulation or change of texture, and abrupt thickening or ring of hairs, above the middle, at or usually much below the branching, the branches, if long, erect, but usually very short and erect or spreading, is general, but not universal, in Cynaroidee, and is also observable in some genera of Mutisiaceee and Arctotidee. In all tribes which admit of central sterile hermaphrodite florets (Asteroidee, Inuloidez, Helianthoideze, Helenioidez, Anthemidese, Senecionidez, Calendulacee, and Arctotidew) the styles of these sterile florets are for the most part filiform or slightly clavate, YALUE OF CHARACTERS. 351 undivided or with slender erect branches, very papillose or hirsute outside, without stigmatic series inside, and very similar in all the different tribes where they occur. The Cichoriaceous style (fig. 1), with slender, almost filiform, papillose branches, varying but slightly in being more or less acute or obtuse, and very rarely some- what shortened and flattened, is uniform in the tribe, but is also precisely the one most general in the female florets of the order. To recapitulate, the style-branches of the hermaphrodite florets afford one of the most useful characters for the determination of genera and some tribes; but all attempts to take it as absolute have hitherto miserably failed, and it must always be considered in combination with other characters. An increase in the number of style-branches from two to three has been occasionally observed, but appears never to be even of specific importance ; for I have met with it in genera otherwise far removed from each other, and never found it to be constant in all the flowers of the same specimen. The epigynous disk has been much made use of for the distinc- tion of genera, by a few botanists who have specially studied a small number of Composite forms. When present it varies much in form : generally speaking, it is a fleshy or glandular, thick, annular projection round the base of the style, either free from it or more orless connected with it, and passing, as it were, into a bulbiform base to the style itself, besides other modifications. In some Cen- taurec,for instance, it forms a membranous or paleaceous 5-toothed cup or tube, having the appearance of an inner pappus; in many genera there is no trace of it. I had formerly, in concurrence with Steetz, Schultz Bipontinus, and others, thought that these differences might be useful at least for generic distinction; but when I came to observe it in several large natural genera, I found it so variable, that I felt compelled to omit it from the generic characters whilst unable to verify it in every species. Delpino connects the presence or absence of the epigynous disk with the entomophilous or anemophilous character of the fertilization ; and he may be right, although the observations hitherto made are far too few to assume such to be the case; and even if it be so, the character seems of comparatively inferior systematic value, ento- mophilous and anemophilous fertilization occurring sometimes, in Composite as in other orders, in plants otherwise closely allied—as well stated by Delpino in the above-quoted pamphlet, p. 34, almost in contradiction to the generic and even tribual importance he attaches to the character in the same memoir. LINN. JOURN.—BOTANY, VOL. XIII. 2D 352 MR. G. BENTHAM ON COMPOSITE. 4. Differences in the Fruit, i. e. in the Achene and its Pappus. The fruit, which in many large and natural orders, such as Le- guminose, Umbellifere, Rubiaceew, Myrtacee, &c., has furnished many valuable characters for the distinction of genera or tribes, has naturally been looked to for similar purposes in Composite ; and Schultz Bipontinus, for instance, has prided himself on esta- blishing his genera on carpological principles. But throughout this vast order, so far as the achene itself is concerned, the structure is absolutely the same; there remain only outward form and consistence, which, however useful in the case of many genera, are, if too closely relied upon, apt in Composite, as in Leguminose and others, to break up very natural genera, especially when the modifications are the result of development in the course of growth from the ovary, and not discernible at the period of fertilization of the flower. The achene (always without its pappus) is very rarely of any use in determining the tribe of a Composite plant, beyond a few vague forms appreciable perhaps to the eye, though difficult to describe, which are only to be found in some one or two of the thirteen tribes; and, moreover, there are so large a number of species in which the ripe fully formed achene is as yet unknown, that its absolute value in large genera where it is usually so uni- form, such as Vernonia, Aster, Senecio, &c., is as yet very doubtful. Outward form, however, when ripe achenes are obtainable, is so prominent a character that much use has been made of it, and often to great advantage, in the definition of genera otherwise natural, and sometimes for that of subtribes or smaller groups of genera. The principal modifications are:—(1) the shape acquired in lateral development, (2) in longitudinal development, (3) the development of the ribs, (4) the consistence, (5) the surface and indumentum. In lateral development the achene as it ripens becomes angular, terete, or flattened, with a considerable degree of constancy in genera or even in subtribes, although in some cases, where it has been the most relied upon for the separation of large nearly allied genera (as in Asteroide:, for instance), there are usually a few ex- ceptional or intermediate species which forbid any absolute reliance upon it. When flattened it becomes important, as in Umbellifere, to consider whether the flattening is lateral (that is, in the direc- tion of the radius of the flower-head) or dorsal (that is, at right angles to the radius); this forms a good distinction, for instance, between the subtribes Verbesinew and Coreopside of Helian- thoidem. This character, however, applies only to the acheues of the disk ; those of the outer row, even when the others are laterally YALUE OF CHARACTERS. 353 flattened, are often dorsally so, or, if the inner edge protrudes, they become triquetrous. In longitudinal development the achene is either equally grown at both ends, so as (if terete) to form a regular eylinder, or tapers at the upper end beyond the seed into what was formerly called a stipes to the pappus, but is now recognized as a neck or beak to the achene, or tapers at the lower end into more or less of a stipes, all which forms variously combined, being readily observed (if ripe achenes are obtained ), have been made great use of in the fixation of genera or even of subtribes, especially in Cichoriacez ; but one of the most prominent differences, between the achene merely con- tracted at the top, and the distinctly beaked achene, has now been generally abandoned in such genera as Crepis, Leontodon, &c., where there occurs every grade from the one to the other; and in many genera, of which it is still the most marked character, it is necessary to be very cautious in its use. The development of the ribs of the achene would at first sight seem to hold the same position in the classification of Composite as in that of Umbelliferz, the ribs being precisely of the same nature ; they represent the ribs or nerves of the calycine leaves which form the adnate calyx-tube; their systematic value, however, is more varied in Composite than in Umbellifere. These ribs are equal to the number of the calycine leaves (usually 5, or sometimes 4) when the primary nerve alone or midrib of each leaf is prominent, double that number when an intermediate rib is formed between each two by the combination of the lateral nerves of each two adjoining calyeine leaves, three times (15-12) if these lateral nerves are separately developed, four or more times if four or more parallel veins are developed from each leaf. This distinction between the development and non-development of the secondary ribs has been found constant, as far as hitherto observed, for the separation of the two principal subtribes of Eupatoriacez, but has quite failed, even for genera, in Vernoniacez. In other cases the reduction of the ribs to the two marginal ones of flattened achenes, the pro- trusion of one, two, three, or more of the primary ribs into acute angles, or their expansion into wings, or their regular equidistance, or the contraction of some of the intervals between them, are often the characters of good genera, but have in many instances (as in Chrysanthemum) been too much relied upon, producing generic combinations or disseverances far from natural. The consistence of the ripe achene is occasionally, but rarely, 2p2 854 MR. G. BENTHAM ON COMPOSIT X. useful as a generic character. The thick achene, whether hard and bony or fleshy, is only to be met with in Cynaroidez, Arcto- tides, Calendulacee (Osteospermum), and a few Helianthoidee ; but even there, although pretty constant in genera, is unavailable for tribualor even subtribual distinctions. The pericarp is never, I believe, truly crustaceous and fragile, but from the ordinary con- sistence it passes in a few genera into thinly membranous. The smooth pitted or muricate surface has been made use of in the case of some Cichoriacee especially, but cannot be implicitly relied on. In some genera of Helianthoidez, for instance, as also in Villanova, Adenostemma, Brachycome, and some others, strongly muricate and perfectly smooth achenes are met with in different capitula of the same specimen, or proceeding from different florets of one and the same capitulum. The difference between the densely silky-hairy and the glabrous or slightly hirsute achene, and in some cases the woolly indumentum, has been found a good generic character in some Helenioides, Cynaroides, Arctotide:e, Mutisiaceee, and Cichoriacee. The pappus may be best considered under the head of the fruit ; for although homologically it is generally admitted to be an altered or semiabortive calyx-limb, and although when present it is always already to be met with at the time of flowering, yet it is on the ripe achene that it has attained its fullest development in those innumerable variations which strike the eye of the most super- ficial observer, and which have been eagerly seized upon to cha- racterize a large proportion of the thousand and one petty genera with which synantherology has been encumbered. Constant or nearly so in each species, with very few exceptions, the pappus will often, in a most natural genus, so vary from species to species, as to make it a most difficult task to decide whether it should be neglected altogether, or, if taken into account, what modifications may be taken as generic, subtribual, or tribual. The presence or absence of a pappus or its degree of develop- ment is always of much less importance than its nature when present ; for there are frequently exceptional species or varieties- where it is wanting in genera or species where it is usually present ; and therefore it is, that where we have a specimen with no pappus, we must be very careful to determine its affinities by other cha- racters. In some cases, however, the absence of pappus has proved a really constant generic character, and is often a clue at least to the tribe of a Composite. It is, for instance, almost VALUE OF CHARACTERS. 355 always present in Senecionide, and very frequently deficient in Anthemides. There is not, however, a single tribe in which there is not at least one genus deprived of all pappus. Among the various peculiarities by which the pappus is diver- sified, it is a very difficult question to determine which are and which are not of importance in classification, whether we reason à priori from the presumed homology and functions of the organ, or whether we confine ourselves to the experience of its conformity or non-conformity to variations of other organs. Asa general rule, the greater the consolidation and contraction of the inner and more important floral organs, and the more the outer compara- tively vegetative organs are called in to assist in some part of the reproduetive functions, the more constant and important in classification are the latter. In Composite we have alluded to the great condensation of the floral organs; and, amongst other reduetions, the calyx-limb, in the form of a pappus, appears to have its functions annihilated or confined to the sole purpose of assisting in the dissemination of the seed, whilst its ordinary part of protecting the young flower is here supplied by the bracts collected in an involucre. The modifications under which this pappus is acted on by wind, or adheres to extraneous substances, ought therefore to be systematically important, as well as those which are indications of its homologies. On the other hand, the pappus being in so many cases a very much reduced or rudimén- tary organ, those differences depending on the degree of develop- ment must, as in the case of other rudimentary organs, be very little relied on. T Of these modifications, we may first consider the indications of homologies. If the pappus is a reduced calyx-limb, then in all cases where it consists of a single ring of bristles or palex it is easy to suppose that these may represent the parallel ,veins, nerves, or ribs of the calyx-lobes or calycine leaves—more especi- ally as they, or at least the most valid among their number, often correspond in number, and are even in direct continuation of the ribs of the achene ; but where these bristles are very numerous and crowded in a dense tuft, scarcely separable into series, or where they form several concentric series, very distinct from each other and often differing in nature, their homology is not so easy to settle. Most probably the longest or principal series repre- sent the calyx-ribs, which may sometimes, owing to their great number and crowded state, become forced, as it were, into two or 356 MR. G. BENTHAM ON COMPOSIT. three apparent series, when they are organically in a single one ; and these, therefore, should be taken chiefly into account in using the pappus as a systematic character. The gradually diminishing outer rows, so much less constant in their presence, absence, shape, or degree of development than the principal row, may be analogous only to those extra teeth or appendages of some Mela- stomacee, Verbenacez, &c., which have been called epicalyces. The occasionally present reduced innermost row (as in some Cen- taureas), which, as above mentioned, may be only a modification of the epigynous disk, is of scarcely more than specific value. The four principal modifications of the pappus which may be made available in the tribual characters are:—1, the setose pappus, where the ring consists of a number, often indefinite, of bristle-like rays, sometimes very slender and hair-like, sometimes thicker or flattened and passing into pales; 2, the paleaceous pappus, consisting of a usually smaller and often definite number of flattened chaff-like or transparent scales; 3, the aristiform pappus, where one, two, three, or more of the ribs of the achene are produced into rigid awns or teeth; and, 4, the coroniform pappus, where the very short pale: are united in a ring or cup. These different forms give very useful and general characters without being absolute; for in every tribe there are exceptions to the normal pappus, besides that the different forms may pass so gradually one into the other as to make it difficult to decide to wnich class a given pappus should be referred. The pales may be very obtuse or produced into a point which may lengthen into an awn, whilst the flattened base may gradually shorten or be obliterated, or the pales may gradually increase in number and diminish in breadth, till they come better under the designation of sete, or they may gradually shorten and unite more or less into a corona. Generally speaking, the pappus may be said to be setose or slightly paleaceous in Vernoniaces, Eupatoriaces, Asteroides, and most subtribes of Inuloideæ ; aristiform or truly paleaceous in the subtribe Buphthalmee of Inuloide* and in Helianthoides, paleaceous in Helenioides, coroniform or none in Anthemidesze, setose again in Senecionidew, none in Calendulacese, paleaceous or none in Arctotide:e, setose or slightly paleaceous and usually very copious in Cynaroideze, setose or paleaceous in Mutisiacee and Cichoriacez. In all tribes there are a few genera or species, and in Asteroides, Inuloidew, Helianthoidez, and Helenioidese several genera without any pappus ; and in almost all tribes there are a YALUE OF CHARACTERS. 857 very few striking exceptions to the normal form, although perhaps specific only. Of the further modifications of the pappus, there is one upon which great stress is often laid, as being of supposed absolute generic importance—the difference between the simply setose (where the setz are denticulate or scabrous only) and the plu- mose (where the sete are bordered by fine cilia like the plumes of a feather). But the value of this character has been much overrated. The plumose pappus occurs most frequently in Cichoriaceee, Mutisiace:, and Cynaroidez; in the two former tribes it is often constant in otherwise good genera ; in Cynaroideee it is also not uncommon, but rarely accompanied by other marked differences ; it is little more than a specific character in Onopordon, Jurinea, Tricholepis, Centaurea, dic. ; and if we have maintained it for the technical separation of Cnicus from Carduus, it is partly from convenience, on account of the large number of species it separates—partly on geographical grounds; for the American species, which are now numerous and pass into a distinct type in other respects, have always the pappus plumose. In the other tribes the plumose pappus is rare, occasionally constant in small but natural genera, in other instances passing gradually through allied species into the simply setose. The difference is indeed but one of degree; the lateral denticulations or cilia are always the same in relative position. When they are shorter than the diameter of the pappus-ray or main seta, the latter is termed simply scabrous or denticulate, when they shortly exceed that diameter it is said to be barbellate, and plumose only when they are considerably longer. The intermediate barbellate stage, how- ever, is much more rare than the simply denticulate or the decidedly plumose state. Another distinction often of some importance, and much insisted upon of late, especially by Weddell, is, in those very frequent cases where the setose pappus parts with its achene— whether it falls off altogether, the setee being united in a ring at the base, or whether each seta falls off separately. This character is often of much avail in several genera of Cichoriacee ; and Weddell made use of it chiefly in aid of the difficult discrimination of some Gnaphalioid genera or subgenera, and at first sight very success- fully; but a closer investigation of a much larger number of species than he had at his command has in some measure lowered again the "— value of the character. 358 MR. G. BENTHAM ON COMPOSITJE. Generally speaking, the modifieations of the pappus, however inconstant in genera or even sections, are very constant in species ; the exceptions are chiefly in the coroniform and other very much reduced pappi, whieh may be present or absent in different indi- viduals or varieties of one and the same species—as, for instance, in those forms of Chrysanthemum leucanthemum so elaborately worked out by Fenzl, as well as in several Ma£ricarie and some species of Centaurea, several Helenioides, &c. 5. Differences in the Andrecium. In all essential points the androecium of Composite is as uniform throughout the order as the pistil; it offers none of those differ- ences in number, symmetry, position, divection, or structure which in so many orders are called in aid of the discrimination of genera. The anthers, equal in number to the corolla-lobes, are united, or perhaps, in a few cases, only closely connivent in a cylinder round the style, and their cells open inwardly and longitudinally without any variation ; so also in regard to the filaments, they are always in- serted in(adnate to)the corolla-tube, and attached to the base ofthe connective. These filaments vary slightly in the height to which they are adnate to the corolla-tube (a question of degree affording no available characters), and also in their being free or monadel- phous after quitting the tube, and glabrous or papillose-hairy. Both the latter characters have been made use of in Cynaroides as generie—the former to separate rather too artificially three monotypic genera from Carduus and Cnicus, the hairy or non- hairy filaments rather more successfully applied to the distinction of a few large genera where it proves constant. There is also, in many Senecionides, for instance, an abrupt dilatation or change of texture, and almost an articulation at some little distance below the insertion of the anther. But it remains to be ascertained how far this is constant even in the genus Senecio; and my own observations are insufficient to establish it as a generic character. The anthers, however, are sometimes provided with certain appendages apparently of little or no functional or homological importance, but which nevertheless, from the remarkable con- stancy of their presence or absence in whole tribes, supply one of the most valuable characters in Composite if applied with proper caution. These appendages are either apical or basal. At the top of the anther-tube each connective is produced into a thinly cartilaginous erect or incurved membrane or point, which may YALUE OF CHARACTERS. 359 possibly act some part in influencing the dissemination of the pollen, although nothing in that respect has as yet been ascer- tained. These appendages are uniform throughout the order, ex- cept as to length or breadth, and except as to two subtribes whieh they assist in characterizing. [In the subtribe Piqueriee of Eupa- toriacez they are deficient, the anthers being truncate on the top ; and in the subtribe Ambrosice of Helianthoidee they taper into an incurved point. In one or perhaps a very few species of Helianthoidex, they are reduced to a small point or possibly deficient. At the base of the anthers the appendages usually called tails are much more systematically valuable, as affording by their pre- sence or absence almost absolute tests of several of the largest tribes. The difficulty is, in a few cases, to decide whether the anthers are or are not to be properly designated as tailed; and allowance must be made for a few, although rare, real exceptions. First, as to determining what is a tail to an anther, the accom- panying figures (Plate 1X.) will readily explain the principal differ- ences. In figs. 1 and 2 the anthers are truncate at the base; in fig. 3 produced into obtuse auricles; in fig. 4 sagittate with the auricles of adjoining anthers connate to the point, in fig. 5 sagit- tate with the auricles free; and in all these they are inappendi- culate or tailless, although the auricles in figs. 4 and 5 may be very acute, for the points are not produced beyond the pollen- bearing cells; in figs. 6 and 7 these auricles, united in fig. 6, distinct in fig. 7, are very shortly produced into what may be termed rudimentary tails, and occasion one of the great diffi- culties, as different botanists have described them as tailed or tailless ; in figs. 8 and 9 the tails are decidedly setiform, those of adjoining anthers united so as to show five setze to the pentamer- ous andreecium in fig. 8; setiform but distinct in fig. 9, showing ten sete either approximate in pairs or equidistant as represented in the plate, or lying close to the filament; the tails are ciliate at the end or ciliately fringed in figs. 9 and 10, dilated and fringed in fig. 11. But the observation of some of these differences requires eonsiderable care and some experience. If not well soaked out, the acute auricles of figs. 4 and 5 may be mistaken for the really pointed ones of figs. 6 and 7 ; and the tails of fig. 7, and even of fig. 9, sometimes lie so close to the filament, that when very fine they are frequently overlooked. To show the degree of constancy of these various forms we may take the thir- 360 MR. G. BENTHAM ON COMPOSIT E. teen tribes in succession, as far as known; and I may state that there are very few genera in which I have not examined the anthers, and usually verified them in a considerable number of species of the larger genera. Vernoniacez (near 500 species in 35 genera) have normally the anthers of figs. 3, 4, or 5, and strictly so in the great majority of species. They are, therefore, usually characterized as tailless ; but there are some four or five genera, one of at least twenty species, the others monotypic or nearly so, which pass into fig. 7, and have therefore been described as exceptionally tailed, although they never, as far as I am aware, have the elongated setiform or fringed appendages of figs. 8, 9, and 10. Several species of Vernonia itself | have also been described as having shortly tailed anthers ; but that is owing to careless observation of the acute auricles not sufh- ciently soaked out so as to show that they are polliniferous to the end. Vernoniacez, therefore, may be safely characterized as having the anthers sagittate at the base, with the auricles obtuse, acute, or rarely produced into short points. Eupatoriacez (above 700 species in 35 genera) have the anthers of fig. 1 and 2, passing occasionally into fig. 3, but never, as far as I am aware, beyond that, and are therefore characterized as having the anthers truncate, emarginate, or rarely sagittate at the base, and strictly tailless. Asteroidez (nearly 1500 species in 88 genera) have the anthers as strictly tailless (varying only from truncate to sagittate) as in Eupatoriacese, with the exception of some very few, perhaps not above a dozen, species of Australasian Olearias and Celmisias, in which the auricles have been observed to have minute very fine points, similar to the tails of figs. 7 and 8 when in a reduced form. In these rare cases the other tribual characters are so decided as to leave no doubt as to the tribe to which the plant should be referred. Inuloidez (above 1100 species in 138 genera) have the tailed anthers of figs. 6 to 10, not absolutely without exception; but having myself examined all the genera, except six monotypie ones unknown to me, and by far the greatest number of species, I have only found the few species of Laggera, three or four of Phagnalon, and perhaps as many more scattered over other genera in which these tails are wanting. There are, however, many species, especially among the Gnaphalioid genera, in which they are so fine, and so apt when wet to lie close to the filament, YALUE OF CHARACTERS. 361 as to be difficult to observe, besides that they have been neglected or carelessly overlooked in the characters given by authors to various genera, which have therefore been placed in wrong tribes. In all these doubtful cases there are characters derived from the style and other organs ready to be called in aid. Helianthoidese (nearly 1100 species in 140 genera) vary in their anthers as in their styles. Like the Vernoniacex, the base of the anthers passes from fig. 2 to fig. 7, but never beyond that ; they are usually sagittate with obtuse, acute, or more or less pointed-acu- minate auricles, but not properly speaking tailed, although they are on some occasions so described by Grisebach and some others. They are often also as obtusely truncate at the base as the Eupa- toriacese. Helenioidez (nearly 300 species in 60 genera) are, in respect of anther-bases, like Helianthoidez, but with less variation, ranging from fig. 2 to fig. 5, rarely if ever passing into figs. 6 or 7. Anthemides (about 650 species in 41 genera) appear to be always quite tailless, and show more frequently fig.2 than figs. 3, 4, or 5, and never, as far as I am aware, go beyond that. Senecionide (about 1350 species in 42 genera) have again the range of Helianthoides in their anther-bases, which are almost always sagittate from fig. 3 to fig. 5, rarely truneate as in fig. 2, or shortly pointed as in figs. 6 or 7, never, properly speaking, tailed as in figs. 8 to 10; but, as in the case of Helianthoidee, some of them are occasionally described as tailed. Calendulacee, the smallest of all the tribes (not quite 120 species in 8 genera), isless definite than any in respect of the anther-tails, which are more or less decided from fig. 6 to fig. 8, rarely so obsolete as to show figs. 4 or 5, nor yet so prominent or fringed as to represent figs. 9 or 10. Arctotidez (about 250 species in 16 genera), usually connected with Cynaroidee on account of their styles, but more nearly allied to Anthemidez in their involueres and flowers, show an approach to the latter tribe in their anther-bases, which are never more tailed than in figs. 6 and 7, and usually ranging from figs. 2 to 5. Cynaroidez (nearly 900 species in 36 genera) belong essentially to the tailed-anthered division. Their tails are usually long and fringed, as in figs. 9 and 10, or dilated as in fig. 11; but there are exceptions. In most Serratulas, and in some species of Xeran- themum, Centaurea, and a few others, the tails are more and more 362 MR. G. BENTHAM ON COMPOSITAE. shortened or quite obsolete; but these exceptional species have other characters which leave no doubt as to their affinities, and after all they are but very few in proportion to the number of species in the tribe. Mutisiacez (about 450 species in 49 genera) belong also to the tailed class, showing the anther-bases of figs. 8 to 10, without any gradual reduction of the tails as in Cynaroidese, but with two re- markable exceptions. Schlechtendahlia and Barnadesia have no tails at all, even rudimentary, to their anthers. The former, a single anomalous species, may possibly hereafter be connected with some other tribe; but the ten species of Barnadesia are so decidedly Mutisiaceous in every other respect, that they must be accepted ' as a striking exception. Lastly, Cichoriaceee (above 700 species in 56 genera) are as uni- form in their anthers as in their styles and corollas, although there may be some ambiguity as to the class of anther-bases to which they should be referred. They have been described as tailed and as tailless ; and after examining hundreds of species either in the dry or in the fresh state, I have been left in doubt as to which is the best designation. The anthers are always sagittate at the base with pointed auricles ; but how far the fine point is produced beyond the end of the cells is a matter of uncertainty ; it is gener- ally so produced, although never to any considerable length, and seems to vary in that respect (within very narrow limits) in one and the same species ; but these niceties are difficult to appreciate, and I may not always have been sufficiently careful in my dissections. There are some other differences in the anthers, such as the proportion occupied by the polliniferous portion, the degree in which the anther-tube is exserted from or included in the corolla- tube, &c., in which I have been hitherto unable to discover any good generic indications. Steetz and some others have also esta- blished genera on characters derived solely from the shape of the pollen-grains ; but this character has been accurately observed in by tar too small a number of species to be as yet made available for systematic purposes. It would require the close observation and study of years to decide upon its value ; and if really suffici- ently connected with other characters to establish it as a natural one, it never would practically be very useful,as requiring a high microscopical power to verify it. It has been generally said that the pollen is angular (usually dodecaedrous) and scabrous in Cichoriaces, globular or elliptical and smooth in Mutisiacese, YALUE OF CHARACTERS. 363 globular and echinulate in the remaining tribes ; but several excep- tional genera or species have already been noted, which, if they had been confined to genera otherwise anomalous, such as Stokesia, might have tended to establish the value of the character. But Steetz has also separated, on this character alone, plants which are, on every other account, evidently congeners ; and it remains to be observed whether some of the differences noted may not be indi- vidual only or even dependent on age or degree of development. The presence or absence of staminodia, or abortive or imperfect stamens in the female florets, has been regarded as a character of some importance; and it is, in a few cases, perhaps generic, but never much to be relied on. These staminodia are frequently to be met with, and perhaps constant in some genera, in Mutisiacee, in Petrobiee, and a few other Helianthoides, and a very few Senecionideæ, rare, if ever observed, in Asteroides, Inuloideze (except one or two species of Buphthalmex), Helenioidez, and Anthemide:e. 6. Differences in the Corolla. The corolla of Composite is superior and gamopetalous, with a valvate estivation throughout the order with as much uniformity as has been observed in the essential characters of the andreecium and pistil. It is usually pentamerous, but not unfrequently tetramerous, and occasionally trimerous—differences which are sometimes generic, frequently specific or sexual only, or variable in the same species, never tribual. The available differences consist in the varied development of the limb, whether regular or irregular. The first and most obvious distinction which strikes the eye of the most careless observer is that between the tubular and the ligulate limb, giving three principal forms of flower-heads— the discoid (where the corollas of all the florets are tubular), the radiate (where the external ones are ligulate and the central ones tubular),and the liguliform (where all the corollas areligulate). But on further investigation this distribution requires to be modified. The ligulate corollas of the Cichoriacez do not correspond to those of the ray in other tribes. In the Cichoriacez, or true Liguliflorz, the ligula is 5-merous; it consists of the whole of the five united petals forming a flattened lamina, truncate and shortly 5-toothed at the end (Plate VIII. fig. 1), and this with the utmost uniformity throughout the tribe; whilst in all other Composite, collectively distinguished by Weddell and others as Tubuliflore, when the 364 MR. G. BENTHAM ON COMPOSIT X. florets of the circumference, or, in a few cases, all or nearly all are ligulate, the ligula is trimerous only, consisting of three only of the united petals, the two inner ones being reduced to minute teeth, or, more generally, entirely deficient (fig. 7). The only instances among Tubuliflore of an approach to the Cichoriaceous corolla are in Stokesia, a monotypic genus of Vernoniacer, where the corollas are irregularly expanded into a 5-lobed lamina, and in a very few Mutisiacee and two or three species of Cynaroide:e, where the lamina is 4-merous, or even, in a few species, 5-merous, almost as equally so as in Cichoriacee. In Cichoriacex, there- fore, the corolla gives us an absolute tribual character, but no generic ones within the tribe, beyond a few indications derived from size or colour. It is the diversity in the corolla of Tubuli- flore alone that we have further to consider. The corollas of the female florets always differ from the herma- phrodite ones of the same species, sometimes only in being more slender, very frequently in the shape of the limb; and, in so far as the difference is owing to sex alone, the characters to be derived from the presence or absence and relative number and position of these diverging forms have been already considered under the head of sexual differences; but there are other diversities of form to which some importance is attached. The most remarkable is the so-called bilabiate form of most Mutisiacesz, which had induced many to class that tribe as a distinct suborder, under the name of Labiatiflore. In that tribe there is in the same head, or in dif- ferent species or genera, a gradual passage from the regular 5-lobed to the ligulate limb :—first, a slight irregularity, owing to the inner- most lobe being more deeply separated than the others; then two of the inner ones are more deeply separated or more erect than the three outer ones (Plate VIII. fig. 5); then, again, the two inner ones shorten, whilst the three outer lengthen and become gradu- ally consolidated into a ligula; sometimes the two inner and three outer ones are respectively united, the former into a short inner lip, the latter into a longer outer one (ñg. 6), or the inner lobes disappear altogether, leaving the truly ligulate trimerous female corolla of ordinary radiate heads (fig. 7). All these forms are to be met with in Mutisiacez, which can therefore no longer be absolutely characterized by their corollas. In tbe subtribe Goch- natiee, for instance, they are all tubuliform, and as regular as in any discoid genus or tribe; and although the well-developed bilabiate form is almost limited to Mutisiacew, yet there are a YALUE OF CHARACTERS. 365 few genera among radiate Senecionidee or Inuloideee, and per- haps some others, where the ray-florets have at the base of their lamina one or two small inner lobes. In other respects the form of the florets may afford some slight indication of the genera or tribes. Long narrow corolla-lobes to the hermaphrodite regular floret (fig. 4) are characteristic of most Ver- noniacee and Cynaroidez, rare in Asteroidez, Inuloidex, Helian- thoideæ, and Senecionidew. The female florets when present are almost always slender, but regular or nearly so (fig. 8) in some genera or subtribes of Asteroidex, Inuloidee, Anthemidex, and Senecio- nidex, always ligulate (fig. 7) in others. Ina few genera or sec- tions of genera of Asteroidex, and in one of Mutisiacez, the outer rows of female florets are ligulate, the inner filiform ; and in some genera or subgenera of Helianthoides or Anthemidee the female florets have only a rudimentary corolla, or are absolutely without any. In all these respects differences in form of the corolla are more important than its absolute presence or absence, or than the degree of development when present. The general shape of the limb (that is, of the dilated portion above the insertion of the stamens) ofthe regular corolla, whether campan- ulate (fig. 3) or gradually dilated, or scarcely thicker than the tube and cylindrical (fig. 2), is sometimes characteristic of genera, but very frequently specific only. Colour is also, in some measure, characteristic of some tribes. The corollas are, I believe, never yellow in Vernoniacez or Eupatoriacex, and not very frequently so in Cynaroideee, in all of which the prevailing colour is purplish, varying from pink to blue, although pure blue is not frequent. In Cichoriacex yellow is the common colour, although some species, groups of species, or even genera are blue; pink and purple rare. In all the other tribes yellow is the prevailing and, in some tribes, the constant colour in the hermaphrodite florets ; pink, purple, and blue exceptional, the latter very rare. The female florets when expanded into a ray are either of the colour ofthe disk, or pink; purple, blue, or white, with a yellow disk. This distinction between homochromous and heterochromous flower-heads, although specific only in some genera, e. y. Senecio, is generic in Asteroides, where it has served indeed to characterize two of the subdivisions of the tribe—somewhat artificial ones it is true, but yet the best that have been proposed. White flowers are not common in the order, but are to be met with in all the tribes, whether the normal colour be yellow or not; they have even been made to serve as a generic 366 MR. G. BENTHAM ON COMPOSIT X. character to distinguish Cacalia from Senecio, and Nabalus from Prenanthes. This, however, only holds good for North-American species. The constancy of the venation of the corolla-lobes in Composite pointed out by Brown has since been frequently alluded to, and is, as it were, the natural consequence of the equally constant valvate estivation. The chief minor difference observable in venation is in the undivided part of the limb, and results from the prominence of the central vein alone of each petal, or from the more or less conspicuousness of the lateral veins, those of two adjoining petals combined into one, or distinct, and all similar to the central vein. The latter arrangement is conspicuous, for instance, in the ligule of Trichocline, and comes in aid of the generic character. 7. Differences in the Calyx. The calyx in Compositz is so absolutely deprived of its ordinary function (the protection of the young flower), so completely absorbed in the ovary, or so absolutely rudimentary at the time of flowering, that the only available distinctive characters it affords result from the varied impress it gives to the external surface of the fruit, or from the forms it acquires as a portion of that fruit. These characters, therefore, have been considered under the head of the fruit, as those of its ribs and pappus. 8. Differences in the Ultimate Inflorescence and Bracts (the Capitulum, its Receptacle, Involucre, and Palee). Inflorescence and its bracts are most frequently classed among the vegetative organs, and consequently assigned a low grade in the scale of generie characters, too low a one, probably, in many cases. In Composite, however, where they are called upon to take a part as accessories to the reproductive operation, to supply in some respects, the place of some of the reproductive class that have been disabled by consolidation, they acquire an unwonted degree of fixity, and thence of generic value. Thus it is, that the diversities in the flower-head or capitulum, its involucre, receptacle, and palez, are properly taken account of in the essential character of every genus of the order. Whatever be the primary inflorescence of Composite (solitary, cymose, paniculate, capitate, or any other form), the ultimate rami- fication is invariably a capitulum— several flowers (often very many, sometimes very few, or a single one) sessile, or nearly so, on a com- VALUE OF CHARACTERS. 367 mon receptacle, within an involucre of several, often very many, rarely only two or three, closely packed bracts, which act more or less the part of the suppressed calyces in protecting the buds or the young fruits. Notwithstanding the special names given to these organs by various synantherologists, there is nothing to distinguish them from the corresponding organs in divers genera of Dipsacew, Umbelliferze, Cornaees, Myrtaceæ (Darwinia), Pro- teacex, Thy melec, and many others—nothing whatever except their constancy. In order correctly to understand the minor modifiea- tions to which these organs are subject, it is necessary to keep their homology in view ; and therefore it is that we have thought it better to retain the intelligible terminology of involucral bracts and receptacle, than to encumber it with such special terms as periclinium, phyllaries, clinanthium, &c., which only serve to give unnecessary trouble and convey false notions. The capitulum characterized by the involucre exists throughout the order; the involuere may in some compact compound inflores- cences be reduced to two or three bracts only, never, I believe, to a single one, even when uniflorous. The number (taken gene- rally within certain limits, rarely as absolutely precise) and arrangement of the bracts, their general form and consistency, and the general form the involucre itself or the capitulum (including the florets) assumes, afford generic characters in most cases excel- lent from their constancy, although, from their ready perceptibility and the aspect they give to the plant, they are apt to be too hastily observed and too implicitly relied upon. They are also far from absolute as tribual characters, although they may give good general indications. Thus the single row of inner equal erect involucral bracts, with or without much smaller or differently shaped or much looser outer ones, so common in Senecionidex, Helenioidew, and: Helianthoidew, are seldom, if ever, to be met with in Vernoniacesm, Eupatoriacex, Asteroides, Inuloidex, Anthemidex, Arctotidex, or Cynaroides. The ovoid or globular involucre with many rows of closely imbricate bracts is chiefly characteristic of Vernoniacerm, Cynaroidez, the subtribe Gorteriez of Arctotidee, a portion of Mutisiacex, and exists only in isolated genera in other tribes. The broadly hemispherical involucre with scarious-tipped or bordered inner bracts is general in Anthemidez, in the subtribe Euarctotex of Arctotidez, and in some Asteroides. Involucres, however, are sometimes deceptive, and precisely the same forms may be occa- sionally met with in two genera belonging to widely distant tribes, LINN. JOURN.—BOTANY, VOL. XIII. 2 E 368 MR. G. BENTHAM ON COMPOSITE. The receptacle, which is homologically the dilated apex of the peduncle, has in its general shape (concave, flat, convex, conical, or elongated) oceasionally supplied generie charaeters, rarely of much value, and sometimes deceptive. Where the involucral bracts have been in many rows and very deciduous, they leave after falling off an apparently ovoid, obovoid, turbinate, or obconical receptacle, which has been used generically to separate species from their allies described as having flat or convex receptacles, when really there has been no difference at all, except that in the one case the portion of the peduncle on which the bracts had rested is taken into account, and in the other the apex only within the bracts has been considered. The extreme forms of the receptacle, either concave or elongated into the rhachis of a cylindrical spike, have been on some rare occasions observed in single species, which have on that account been at once raised into monotypic genera; but I do not believe that these differences have ever been found accompanied by other peculiarities, or to be prevalent through two or more otherwise allied species, and must therefore be regarded as specifie only. The so-called pale: of the receptacle within the involucre have been made great use of, by their presence or absence, for charac- terizing genera, subtribes, or even tribes. But this use, however appropriately adopted in some cases, has in others degenerated into abuse. Homologically, these pales are the same as the involucral bracts. There is not even the difference that lies between ordinary bracts and bracteoles, for both are on the same axis, but is solely that which distinguishes the lower empty bracts of an ordinary spike or other simple inflorescence from those which subtend the individual flowers. In Composite, however, these inner subtending bracts amongst the florets are so frequently different from the outer empty orinvolucral ones in form, size, consistence, constancy, &c., that their designation by the distinct name of palee is of great convenience in systematic descriptions. There is sometimes, however, a difficulty in regard to the inner- most row of outer bracts, either when these are gradually modified so as to pass, as it were, into the flowering bracts or palex, or when each one of that innermost row of outer bracts exactly subtends one of the florets of the outermost row, or even embraces or encloses it, and is then different in form &e. both from the outer involucral bracts and from the inner palez ; and some controversy has been carried on as to whether this is an inner row of involucral bracts VALUE OF CHARACTERS. 369 or an outer row of receptacular palee. But this is a mere war of words. Homologically the two are one and the same thing; but, as a matter of convenience in description, it has been a general, and ought to be an universal, rule to call all that are outside of all the florets involucral bracts, and those only which are within the outer rows of florets receptacular palez. With regard to the use made of the presence or absence of these pale: not only as a generi¢ but as a subtribual or even tribual cha- racter, it proves with certain restrictions to be a good one, although in some cases it is very artificial or uncertain. Thus the rigid, usually persistent, receptacular paleze are constant, or nearly so, in Helianthoidez and the subtribe Buphthalmee of Inuloidez, and never occur in Helenioidese; whilst the thinner more deciduous ones of a very few small genera of Vernoniacesw, Eupatoriacee, Asteroide, and Senecionides are sometimes inconstant in the same genus. In Anthemidee this character conveniently, although somewhat artificially, separates the Huanthemee from the Chrysan- themes. In Cichoriacez, where, from the absolute uniformity of the florets, there is such a dearth of distinctive characters, these pales have been eagerly seized upon for the separation of some subtribes or other divisions; but a very little examination will show that they are at most of generic value for the separation of such closely allied groups as Hypocheris from Leontodon, Rodigia from Crepis, &e. The peculiar shape assumed by the pales is often of much more absolute generic importance, as, for instance, those of Scolymus, which are accompanied by so many other characters in habit, involucre, receptacle, &c., although still with the uniform florets of Cichoriace:e. There is a state of the receptacle which has been variously described as paleaceous or naked, or neither, and which certainly sometimes passes from one to the other. On the naked receptacle each floret, after the fruit has fallen off, leaves a more or less marked scar, either on a small protuberance which may be said to be a rudimentary pedicel, or in a depression in which the achene was seated. In the former case the receptacle is said to be scrobiculate or furrowed by the depressions round the protuberances; in the latter it is described as foveolate or pitted. When there is neither much protuberance nor depression, but the area of each achene is marked by a more or less distinctly raised line round its circum- ference, the receptacle is said to be areolate. When this line is more raised and jagged on the edge, or broken into short sete or 2£2 370 MR. G. BENTHAM ON COMPOSITAE. pales, the receptacle is fimbrillate, setiferous when these sete are elongated; and if the margins of the pits rise considerably (some- times enclosing the whole achene) it is termed alveolate or honey- combed. All these varieties in the non-paleaceous receptacle are usually noticed in generic characters, and they are sometimes constant in good genera; but in general they have been too much insisted on, and have produced purely artificial combinations. The Cynaroidex, however, are remarkable for the setose character of the receptacle throughout the tribe, with the exception of Ono- pordon and very few other species rather than genera, where the sete become very short or disappear altogether, and a few others where the sete are more or less combined into true pales. The alveolate receptacle is most remarkable in some small or mono- typie genera belonging to very different tribes, e. g. Albertinia in Vernoniaces, and Baldnina in Helianthoidez. It is also charac- teristic of several genera of Arctotidez. When the involucral bracts of the innermost row precisely sub- tend the florets of the outermost row, and more or less enclose or become adnate to them, or assume more or less of the character of receptacular pales, they often acquire a generie or even a sub- tribual importance, as in several Helianthoider, Helenioides, or Cichoriacez, although occasionally the difference may be little more than specific. 9. Differences in the Foliage. The foliage in Composite is, within certain limits, as variable as in other large orders. It has no one peculiar character which cannot be matched in many other orders; and the only two fea- tures of importance which it does not possess are (1) that there are no stipules (for the auricular expansions at the base of the petiole in some species of Liabum, a very few Helianthoidex, dc. cannot be properly designated as such), and (2) that the leaves, though often much divided, are never compound with articulate leaflets. Amongst all variations to which it is liable, there is one only of any systematic importance—the difference between the opposite (including the rare instances of strictly verticillate) and the alternate leaves, which sometimes constitutes a good, though not quite absolute, tribual character, although also in other in- stances itis not even generic. Thus the leaves are alternate, with few exceptions, in Vernoniacez, Asteroides, Inuloidese, Anthe- mides, Calendulacewm, Arctotides, Cynaroidew, Mutisiacem, and VALUE OF CHARACTERS. 871 Cichoriaceee—the exceptions being 2 species in 465 in Vernoniaces, 1 genus and some 10 or 12 stray species out of 1470 in Aste- roidez, about 20 species out of 1150 in Inuloideæ, about 6 in 650 in Anthemidee, 10 in 115 in Calendulacez, 2 in 450 in Muti- siacez, and no exceptional opposite leaves known in Arctotidez 250 species, Cynaroidew 880, or Cichoriaces 720 species. Alter- nate leaves are also prevalent in Senecionidez, but with nearly 10 per cent. exceptional opposite ones (usually in distinet genera). In Helenioidez the opposite and alternate are rather more equally divided, the former being the more frequent ; and opposite leaves, at least in the lower part of the stem, are the rule in Eupatoriacee and Helianthoidex, although with several striking exceptions. Entire, toothed, or divided leaves may be respectively prevalent in tribes or genera, but rarely in a degree to be much relied on for a general character. The much-divided leaves, for instance, so common in Anthemidee are exceedingly rare in Vernoniacee, Eupatoriaces, and Inuloidee. The prickly-lobed foliage of so many Cynaroidez is scarcely to be met with elsewhere, except in a few Arctotidez (Gorteriex), one small genus of Cichoriacez (Scolymus), and here and there in monotypic genera scattered through other tribes. 10. Differences in Habit, Stature, and General Inflorescence. In habit and stature, Composite are as variable as other large orders, without these differences being often even of generic im- portance, although they may sometimes give useful indications. Arborescent Composite are rare; and frutescent ones prevail only in comparatively few genera ; but in some cases these habits may assist in the discrimination of groups where more absolute cha- racters fail, especially when the differences are connected with geographical distribution. Thus in the great mass of Hetero- chromous Asteroidee and Conyres, including the large genera Aster, Erigeron, and Conyza, taken in their widest sense with a multitude of smaller genera around them, there is no one positive character to separate the groups, large or small, into which the six or seven hundred species have been distributed,—nothing absolute to separate Aster from Conyza, which nevertheless no experienced synantherologist would dream of uniting. Here, therefore, stature, combined with geographical distribution, have been appropriately called in aid by Weddell, excluding all the southern, usually frutescent, species from the northern, constantly 372 MR. G. BENTHAM ON COMPOSITE. herbaceous, Asters ; and there are many instances in other tribes where a shrubby or herbaceous habit may similarly assist in distinguishing natural genera. Beyond this, habit is systemati- cally as vaguea character in Composite as in other orders, acqui- ring importance only in geographical botany for the distinction of those plant-forms of which Grisebach has made so much use in his studies of geographical distribution with reference to climatology. The general inflorescence in Composite (that is, the successive development of the flower-heads) is invariably centrifugal, whilst the ultimate inflorescence (the successive development of the florets within the head) is as constantly centripetal; and this circum- stance is very important in characterizing the compound inflores- cence of the subtribes Lychnophores in Vernoniaces, Lagasceæ in Helianthoides, Angianthes in Inuloidex, and of several genera of Filaginee, Relhaniee, Buphthalmes, Cynaroidex, £c. In these cases numerous capitula, closely sessile on a common receptacle, are collected into a globular, oblong, or depressed cluster often surrounded by a common involucre, the whole assuming the ap- pearance of a single flower-head. This is more especially the case when each individual head is reduced to a single floret with two, three, or very few involucral bracts. Such compound clusters have been sometimes confounded with single flower-heads like those of Albertinia, where the very deep and fringed alveoli of the receptacle resemble the separate involucres of the compound cluster. In such cases the difference between the centripetal inflorescence of Albertinia and the centrifugal one of Eremanthus has been well pointed out by Schultz Bipontinus. I am not aware of any other very marked generie character to be derived from the inflorescence of Composite. If in a few instances the heads are apparently axillary and sessile, it is from the abbrevia- tion of the flowering branch. In Liatris the inflorescence is spicate or racemose, but always with the terminal head first developed. In this case the genera Trilisa and Carphephorus are chiefly distinguished by the corymbose inflorescence, accompanied, however, by some other characters. Among miscellaneous vegetative characters must be included indumentum, to which we have not perhaps paid sufficient atten- tion. Mr. Archer, in a paper published in the fifth volume of our Journal, has shown its value in the classification of the numerous species of Olearia. The stellate hairs have also been shown to be a good generic character for the separation of Hieracium and - VALUE OF CHARACTERS. 373 Andryala from Crepis, of Bedfordia from Senecio, &c.; a certain glandular odoriferous indumentum is characteristic of most Madiee, Buphthalmes, and a few others; and many other in- stances of the generic value of differences in the nature of hairs might be adduced, although the subject has not as yet been suffi- ciently worked out. The presence or absence, abundance or paucity, greater or less development of any description of indu- mentum appears never to be of more than specific value. ll. Differences in Geographical Distribution. I purpose entering into the geographical distribution of Com- posite under a separate head. I only mention it now to show the importance of taking it into consideration in the demarcation of generic groups. If the two theories be admitted, that allied species and genera have a common origin, and that the descen- dants of a common stock placed in different regions having no in- tercommunication will vary in these different regions with different combinations of characters, it will be seen how much geographical distribution may be made to check the value given to generic or other groups founded upon technical distinctions. But the use of geographical distribution as a generic character is liable to many errors. There are especially two great difficulties to overcome :— first to determine upon how far geographical distribution is due to origin, and how far to climatological influences; and, secondly, to decide upon the all-important distinctions between what we may, with Hewett Watson, designate as true denizens of a country, colonists, and aliens: and even among the denizens a still more difficult, although important, point for consideration is the remote- ness of the period at which the common stock has been connected with the flora of other countries. For instance, the Helichrysee, as well as the Arctotidew (Cymbonotus), of Australia are ciosely allied to the corresponding South-African groups without having a single species in common, excepting such as are known to be modern colonists. In the case of Helichrysee the forms have multiplied exceedingly in both regions in the same or in different directions, and have every right to be designated as true native races. But Arctotideæ, numerous in South Africa, have assumed only one form in Australia. Is that form to be regarded as of contemporaneous origin with others common to the two regions, but modified into a local generic type, or is it some old colonist still to be found, or perhaps now extinct, in South Africa? isa question which remains undecided. 374 MR. G. BENTHAM ON COMPOSITA. Til. SKETCH or THE Primary Divisions or THE ORDER. In determining upon the sequence of tribes and subtribes to be adopted in our ‘ Genera Plantarum,’ we have here, as efsewhere, en- deavoured to avoid all useless changes in the Candollean method so generally followed; and we have abandoned some transpositions which I had proposed in the ‘ Flora Hongkongensis’ and ‘ Flora Australiensis.? I had there, for instance, commenced with the Cynaroidee in order to place them in juxtaposition with Verno- niacee; but on working out all the allied genera which I had not previously examined, I have found that they were as well, if not better, placed in the position De Candolle had assigued them. As, however, there are a few other important alterations above alluded to which appeared to me absolutely essential in order to give more definiteness, as well as a more natural character, to some of the great divisions of the order, it may be useful to re- view shortly the modified characters I would now assign to the thirteen tribes we have adopted, referring for the technical synopsis to our ‘ Genera Plantarum.’ 1. Vernoniacee. We have reduced the Vernoniacee to those which have uni- formly homogamous capitula with hermaphrodite florets never yellow. This removes several genera which in minor characters also were very exceptional in the tribe. The Pectidez, which have not even the style of Vernoniacex, resume their place with the Tagetinez under Inuloidee ; the Liabex are transferred to Sene- eionidez, notwithstanding their style, which, however, is not abso- lutely without example in other genera of that tribe; and Gun- delia and Platycarpha form a small subtribe of Arctotidez next to Cynaroider. Vernoniaceew thus modified are usually perennial herbs or shrubs, rarely trees, or very rarely annuals. Their leaves are, with the exception of two or perhaps three species, alternate, en- tire, toothed, or very rarely lyrately pinnatifid; the involucral bracts imbrieate in several rows, except in a very few small ano- malous genera, where they are reduced to two equal rows or to a small definite number. The capitula are uniformly homogamous, with the florets all hermaphrodite, and equally fertile or rarely slightly dicecious; they are sometimes reduced to a single floret ; and in several genera they are closely clustered, forming a dense globular or oblong compound head with or without a common invo- PRIMARY DIVISIONS OF THE ORDER. 375 lucre. The receptacle is naked or slightly fimbrillate, except in two monotypic genera where it bears palew subtending the florets, and one where it is deeply alveolate. The corollas are tubular, regular, with five narrow lobes to the limb, and varying from a reddish purple to nearly white, except in one monotypic genus where they are blue aud expand into a five-lobed ligula, approach- ing that of Cicboriacez, in one small genus in which they are more deeply split on the inner side, nearly as in some Mutisiacee, in one species of Veronia where they are somewhat bilabiate, and in two monotypic genera where they are very small with the lobes reduced to short teeth ; their colour is never yellow. The anthers are never without the terminal appendage to the connec- tive; they are more or less emarginate or sagittate at the base; the auricles of contiguous anthers are usually connate to the end, obtuse or acute, or very rarely produced into short combined or distinct points or rudimentary tails. The style-branches are slender, acute, or scarcely obtuse, uniformly and shortly hirsute, the stigmatic series towards the base on their inner surface not very conspicuous ; and in two monotypic (otherwise anomalous) genera the style is exceptionally almost entire. The achenes are usually terete or slightly flattened and equally ten-ribbed; but in some genera or species they are five- or four-angled, occasionally also with more than ten ribs, and in one monotypie genus per- fectly smooth and shining. The pappus is usually setose and copious ; the sete in some genera flattened into pales, and in a very few others very much reduced or absolutely wanting. 2. Eupatoriacee, We have reduced De Candolle's Eupatoriacez, as we have done his Vernoniacez, by withdrawing the genera with heterogamous capitula, which we have referred to Senecionidex, leaving the tribe one of the best-defined by its style, as well as by several secondary ebaracters. Eupatoriacee are herbs or shrubs, rarely trees, very rarely an- nuals. The leaves, as a rule, at least the lower ones, are opposite and entire or toothed ; but iu a few genera they are all alternate, and exceptionally so in several others; and in very few species they are divided. The involucral bracts are imbricate in several rows or nearly equal in about two rows, in some genera reduced to four, five, or six. The capitula are always homogamous, with all the flowers hermaphrodite and fertile, and are very rarely re- 376 MR. Œ. BENTHAM ON COMPOSITX. duced to a single floret. The receptacle is without pales, except in three or four genera where deciduous pale: subtend the florets in some or all of the species. The corollas are tubular and regular, the limb variously shaped, with five or rarely four short teeth, very rarely (in Liatris and its allies) with the longer lobes of Vernoniaces; they vary in colour from pink or purplish to white or pale yellowish, never truly yellow. The anthers in one sub- iribe are truncate at the top without the normal appendages which are present in the other subtribes, as in the generality of Composite; they are uniformly obtuse at the base, without pro- minent pointed auricles or tails. The style-branches are elon- gated, terete or somewhat flattened, obtuse or club-shaped towards the end, minutely papillose, but not hairy, with the stigmatic serles rather obscure on the inner surface towards their base. The achenes in the first two subtribes are four- or five-angled, or rarely flat; in the third subtribe more terete and ten-ribbed: the pappus usually of fine but rigid sete in one or two rows, in a few small genera plumose or nearly so, in some others reduced to a small definite number, or intermixed with short scales or palez, or the whole pappus reduced to these small pale» or entirely deficient. 3. Asteroidee. We limit the Asteroides to the genera with tailless anthers and with appendages to the style-branches. The subtribes or di- visions Spheranthex, Tarchonanthez, Plucheinez, Inuleze, and Buphthalmex of De Candolle form part of our great tribe of Inuloidee; and the Ecliptez pass into the Helianthoidex, most of the genera enumerated by De Candolle under the former name having nearly related or identical genera among the latter. The Asteroidez thus limited are for the most part readily distin- guished, by the above characters of the style and anthers, from all other tribes; but there are a few Inuloide* and Senecionidese where the involucre and other secondary characters must be called in aid. Asteroidee are mostly herbaceous perennials, or sometimes annuals ; but a few southern or insular genera are shrubby, rarely growing into small trees. The leaves are, with very few excep- tions, alternate, entire or toothed or occasionally divided, but much less frequently so than in Anthemidez. The involucral bracts are usually imbricate in several rows, in a few genera all nearly equal in about two rows. The capitula are usually heterogamous, PRIMARY DIVISIONS OF THE ORDER. 377 with one or more outer rows of female florets rarely sterile; the disk-florets also most frequently fertile, although in a few genera constantly sterile, and in a few others the capitula are homogamous from the deficiency of the female florets. The receptacle is usually naked, pitted, or shortly fimbrillate, rarely bearing deciduous pales subtending the florets. The corollas of the female florets are some- times produced into a trimerous entire or toothed ligula, the two inner lobes of the limb entirely deficient; or the corollas are slender, shorter than the style and truncate at the end, or with a small two- or three-toothed limb; those of the disk are regular, the limb more or less dilated or campanulate, with five, rarely four, teeth or short lobes; they are usually yellow, whilst those of the ray are in some genera homochromous, in others heterochromous, and white or variously coloured. The authers are never without the normal terminal appendages to the connective, and either ob- tuse at the base or rarely sagittate with acute auricles, in a very few species mucronate, or almost produced into minute fine tails. The style-branches of the fertile disk-florets are more or less flat- tened, the marginal stigmatic series usually conspicuous, and beyond them a terminal papillose or hirsute appendage, sometimes very short and obtuse, more frequently triangular or lanceolate, occasionally narrow and elongated, almost as in Vernoniacez. Where the disk-florets are sterile, the style-branches are very narrow, or the style remains undivided. The achenes are usually small, flat, with nerve-like margins, or more or less five- or more ribbed, and becoming terete; rarely produced into a beak, still more rarely, if ever, winged. The pappus is usually setose and copious in one or more rows; in a few genera the sete are plu- mose, in others much reduced or very few, or absolutely none, rarely replaced or accompanied by small thin palee. 4. Inuloidee. The tribe of Inuloidez, as we propose to restore them, are nearly the same as the Inulées of Cassini. They consist of De Candolle’s subtribes Spheranthee, Tarchonanthes, Pluchei- nes, Inulew, Cesulinez, and Buphthalmex taken from his tribe of Asteroides, and of the Angianthex, Cassiniew, Helichrysex, Seriphiee, Antennariex, Leysseriee, and Relhaniee subtracted from the Senecionidez. Their chief distinction from the two tribes they are thus withdrawn from consists in the basal appendages of the anthers, and in the absence of those terminal appendages to 378 MR. G. BENTHAM ON COMPOSITE. the style-branches of the fertile disk-florets which are almost uni- versal in Asteroides, and occasionally, if not very frequently, observable in Senecionidez. The leaves of Inuloidex are, with very few exceptions, alternate and entire, occasionally decurrent, very rarely opposite or lobed, the involueral bracts usually imbricate in several rows, scarcely ever showing that equality of a single or inner row, with or without small outer ones, so frequent in Senecionidez and Helenioidez. The capitula are most frequently heterogamous, with one or more rows of female florets almost, if not quite, universally fertile, the disk-florets sometimes also fertile, but in not a few genera sterile 5 in some genera, however, the capitula are dicecious; and in a con- siderable number of the Helichrysum group of the subtribe Gnaphaliez, in all Angianthez, and in a few genera of other sub- tribes they are homogamous from the total absence of female florets. The receptacle is generally without pales, except in the subtribes Buphthalmex and Filaginez, and a very few scattered genera of other subtribes, where it is wholly or partially palea- ceous. The corollas of the female florets are either short, slender, and minutely toothed at the summit without any expanded limb, or produced into a trimerous entire or toothed ligula, the two inner lobes of the limb either entirely deficient or very rarely ap- pearing in the shape of one or two short slender appendages at the base of the ligula. The corollas of the disk are generally’ those of Asteroidez, Senecionidez, and allied tribes, with four or five short teeth or lobes, very rarely more deeply lobed, those of both sexes almost, though not quite, universally homochromous and usually some shade of yellow. "The anthers are never without the terminal appendage to the connective, which is normal in its shape; they are always more or less sagittate at the base; the auricles of adjoining anthers usually connate to the end and pro- duced beyond the polliniferous part into tails or fine hair-like appendages, either simple or fringed with long hair-like branches. These appendages or tails may be all free, or those of adjoining anthers connate, so as to form ten or five only to the whole an- drecium. In the former case they are sometimes (e. y. in many Gnaphalieze) so fine and short, and lie so close to the filament, that they may be readily, and have been frequently, overlooked ; aud they are, as above mentioued, absolutely wanting in some twelve to twenty species out of 1100. The style-branches of the fertile disk-florets may be more or less flattened, slender, or PRIMARY DIVISIONS OF THE ORDEB. 379 somewhat dilated towards the end, rounded at the tip, or truncate and penicillate, as in Senecio, but always without terminal appen- dages, the stigmatic lines reaching quite, or very nearly, to the end. Where the disk-florets are sterile, their styles are almost always undivided and strongly papillose or hirsute. The achenes are various, usually very small, flat, terete, or angular in the sub- tribes with filiform female florets, longer in the radiate subtribes, very rarely rather large and black, as in so many Helianthoidee, rarely produced into a beak, and never either winged or thick and hard or fleshy. The pappus is usually setose with simple or plumose setze, but in the subtribe Buphthalmex more frequently paleaceous, and in a few genera or species of various subtribes reduced to a corona or very scanty or wholly deficient. 5. Helianthoidee. Our Helianthoidez consist chiefly of De Candolle's subtribes Melampodinex and Helianthes, and of his subdivisions Euga- linsoge: and Madiex, to which we have added the Eclipteze, placed by De Candolle under Asteroides on account of a supposed con- formity of the style to that of the latter and not of the former tribe. But a closer examination has entirely put an end to this artificial distinction, and the Ecliptee do not even remain a di- stinct group of Helianthoidee. De Candolle's nine genera have .to be distributed into different subdivisions. Thus Blainvillea eannot be removed from its closely allied Wedelia; Salmea is very near Verbesina; Dahlia comes next to Coreopsis, Siegesbeckia to Jageria, Sabazia to Gymnolomia; and Cryphiospermum is identical with Enhydra. The Helianthoidee thus modified differ from Asteroidee in their usually combining opposite leaves, a rigidly paleaceous receptacle, and a rigid pappus of few awns or pales. All these characters, however, have exceptions, and many minor cireumstances must be taken into account in fixing the limits of the tribe. Helianthoidez are usually rather coarse herbs or shrubs, their indumentum generally more or less scabrous or hirsute, or some- timessilky with simple hairs, rarely intricately tomentose or woolly. The leaves, at least the lower ones, are most frequently opposite and entire or toothed ; in some genera the upper ones, and in a few the whole of them are alternate; and much divided leaves occur in some of the subtribes. The involucral bracts are very often biseriate, with a difference between the outer and inner rows, 380 MR. G. BENTHAM ON COMPOSITE. but sometimes imbricate in several rows, very generally more or less herbaceous, but sometimes dry and rigid or membranous, very rarely the inner ones scarious. The capitula are most fre- quently heterogamous, with a single row of female and fertile or neutral and sterile florets in the circumference, the disk-florets fertile or in some subtribes uniformly sterile. In some genera also the capitula are (by the imperfection rather than by the absence of the male or female organs) strictly unisexual, either monoecious or dicecious ; and sometimes the capitula are homoga- mous from the deficiency of the ray-florets. The fertile florets are uniformly subtended, and sometimes embraced by or enclosed in the pales of the receptacle or innermost bracts of the involucre. These pale: are deficient only in the centre of the capitulum in the subtribe Milleriez,, where the disk-florets are always sterile, and in some Madiez, where they are frequently so. The corollas of the outer female or neutral florets are usually ligulate and tri- merous, very rarely with an irregularly campanulate limb, or shortly tubular; in the subtribe Ambrosiez they are reduced to a short conical tube or entirely deficient ; the corollas of the disk with five, or rarely four, short lobes or teeth; in a very few small genera there is a slight tendency to the bilabiate form, either by the development of one or two small fine upper lobes to the ray- florets, or by an irregularity in the disk-florets. These disk-florets are generally yellow, sometimes white, rarely purple; the rays usually homochromous. The anthers have the normal terminal appendage, except in Eleutheranthera and possibly a few species of allied genera; and in the Ambrosiee these appendages termi- nate in inflected points; the basal auricles sometimes very short and obtuse, are more frequently acute, and sometimes produced into short points, which have been termed tails. The anther-tube, as a whole, is in many genera much exserted and black. The styles of the fertile disk-florets vary in different genera, from the trun- cate tips of Senecio to the appendiculate branches of Asteroidez, or the subulate hispid branches of Vernoniaces. As in other tribes, the style remains undivided in the disk-florets of most genera where they are constantly sterile. The achenes are often rather large, either thick and hard or sometimes even succulent, or laterally or dorsally flattened and sometimes winged. The pappus most frequently consists of two or three rigid awns or scales corresponding, and often continuous, with the principal ribs or angles of the achene, with or without smaller interme- PRIMARY DIVISIONS OF THE ORDER. 381 diate scales; in most Galinsoge:x the scales are more numerous and often equal, as in Helenioidez», and in a considerable number of genera the pappus is entirely wanting; any tendency towards the setose pappus of Senecionidee is exceedingly rare, the nearest approach to it being in some species of Calea. 6. Helenioidee. Our Helenioidex, as we propose to circumscribe them, may perhaps not be so readily admitted as a distinct tribe as most of the others ; for the genera we have grouped under that name pass respectively into Helianthoidex, from which they differ chiefly in the absence of pales to the receptacle, or into Anthemidez, from which they are separated by the involucre, habit, and in great measure by geographical distribution ; and some species or small genera have almost the pappus, as well as other characters, of Senecionidew. Yet we believe the tribe to be not unnatural, and that in thus uniting the groups of genera it consists of we leave the tribes from which they are withdrawn much better defined. These groups are De Candolle's division Gaillardiez of his subtribe Helenies, his subtribe Tagetinez, the few-flowered epaleaceous genera of his Flaveriez, all included by him in his large tribe of Senecionidezx, with his Pectidex, taken from Vernoniacee, of which, as above observed, they have neither the style nor the habit, involucre, or pappus. Helenioidez are all herbaceous or very shortly shrubby at the base, with the exception of the somewhat anomalous small genus Cacosmia. They are not so coarse nor so roughly hirsute as the majority of Helianthoidee, often glabrous ; the indumentum, when present, is usually eottony or soft, sometimes glandular or viscid. The leaves are often opposite, but in some genera even the lower ones are alternate, and vary from entire or toothed to the divided form so prevalent in Anthemidew. The involucral bracts are generally either in about two rows and more or less herbaceous, as in Helianthoidez, but usually thinner, or in a single row with or without a calyculus, as in most Senecionidee ; but in a very few genera they are imbricate in several rows, or approach those of An- themidez, with the inner ones scarious. The capitula are heteroga- mous in most genera, with a single row of female or rarely neutral florets in the circumference ; the disk-florets fertile or rarely sterile, homogamous in some genera by the absence of the female florets, The receptacle is uniformly naked or slightly pitted, except in Gail- 382 MR. G. BENTHAM ON COMPOSITA, lardia, where it is fimbrillate. The corollas of the female or neu- tral florets are ligulate and trimerous, usually forming a promi- nent ray, but sometimes small ; those of the disk-florets with five, or rarely four, short lobes or teeth ; all usually yellow and homo- chromous, or those of the disk purple, and the same colour ex- tending sometimes to the base of the ray. The anthers are like those of Helianthoides with the normal terminal appendage, the basal auricles obtuse or acute, but scarcely pointed ; the style- branches vary in different genera as much as in Helianthoidee, from the truncate tips of Senecio and of the Anthemidez to the appendiculate branches of Asteroidew or the subulate hispid branches of Vernoniacezs. The undivided style has been only ob- served in the sterile disk-florets of Blennosperma. The achenes ave frequently longer and narrower than in the adjoining tribes, angular or terete, rarely flattened or winged, but in ihe Euhele- nie: usually turbinate and hairy or woolly. The pappus is nor- mally paleaceous; the scales definite or indefinite, obtuse or acute, or, when numerous, occasionally attenuated into sete, almost like those of Senecionidez ; in some genera very short, rarely united in a cup, and in many genera liable to disappear altogether in one or more species. 7. Anthemidee. Our tribe of Anthemidex is the same as De Candolle's subtribe of that name, after deducting two or three small genera which had been inadvertently placed there. It is closely connected as to a few genera with the subtribe Euheleniew of Helenioidez, as to one or two others with Senecionidex ; and the group of Cotulez almost pass into some of the epappose Asteroide: ; but, generally speaking, the involucres, the habit, the styles, and the want of any setose or aristiform pappus readily distinguish the tribe. Anthemidex are often odoriferous in their herbage; the great majority are herbaceous; but they also include shrubby species or genera; their indumentum is rather woolly, glutinous, or soft than coarsely hispid. Their leaves are, with very few exceptions, alternate, and most frequently lobed, much divided, or at least toothed; in a very few small genera opposite and entire. The involueral bracts are usually imbricate in several rows, dry or, the inner ones at least, scarious at the end; in several genera, however, of the Cotula group they are nearly equal in about two rows and thinly herbaceous. The capitula are most frequently PRIMARY DIVISIONS OF THE ORDER. 383 heterogamous, with one or, in some Cotulez, several rows of female fertile, or rarely neutral and sterile, florets in the circumference ; the disk-florets fertile, or in a few small genera sterile; but there are also several genera in which the capitula are quite homogamous, the ray-florets being deficient. The receptacle is with or without pales»; when present, they are usually deciduous. The corollas of the outer florets are either short, slender, and minutely toothed at the summit, without any expanded limb, or more frequently produced into a trimerous, entire, or toothed ligula; the two inner lobes entirely deficient, or the whole corolla of the female florets reduced in some Cotuleæ to a small rudiment or entirely wanting: those of the disk are generally those of Asteroidex and Senecionidez, with four or five, rarely three, short teeth or lobes ; these disk-florets almost always yellow, the rays either homochro- mous or heterochromous (white or pink) in one and the same genus. The anthers have the normal terminal appendage, and are usually obtuse at the base, always without tails or distinct points. The style-branches of the disk-florets are more constant in their shape than in most tribes, truncate and usually penicillate at the end, except in the sterile florets of two or three somewhat anomalous monotypie genera, where they are slender; and whether truncate or not, they remain connate to the end in the sterile florets of a few genera. The achenes are usually rather small, often angular and truncate at the top, or those especially of the ray dorsally flattened or triquetrous and sometimes winged. The pappus is very commonly deficient; when present, reduced to a paleaceous ring or cup or oblique auricle, very rarely consisting of small distinct pales, passing, in one genus, almost into the sete of Senecionidez. 8. Senecionidee. Our tribe of Senecionidez consists of the subtribe Senecionez of De Candolle, with the addition, first, of the Liabese and Tussila- gineze, taken from his Vernoniacez and Eupatoriacee respectively, on account of their heterogamous capitula with the florets most frequently yellow; and, secondly, of the Othonne:e, removed from Cynarez as not having the habit or characters of that tribe, the undivided style of the disk-florets being that of similarly sterile florets of most tribes. The anthers distinguish them from Inu- loidez, the pappus and habit from Helianthoidez and Helenioidee, the pappus and involucre from Anthemidez. The leaves of Senecionidez are very various—alternate in the LINN. JOURN.—BOTANY, VOL. XIII, 2r 384. MR. G. BENTHAM ON COMPOSITA. great majority, but strictly opposite in a few genera, entire or variously toothed or divided. The indumentum is usually cottony and whitish or soft, sometimes glutinous, rarely coarsely hirsute, in one genus stellate. The involuere very generally consists of a single or almost simple inner row of equal, more or less herbaceous or membranous, rarely almost fleshy, bracts, united or free, with or without more or less of smaller outer ones, com- monly called a calyculus; in a few genera, however, they gradu- ally increase from the outer to the inner ones; they are never pungent, and rarely appendiculate. The capitula are usually hete- rogamous; the female florets in a single row almost always fertile ; the disk-florets fertile in most genera, but all, or mostly, sterile in some others ; and in a few genera, and several species of others, the female florets are deficient, leaving the capitula homogamous. The receptacle is usually without pale: ; these, however, are present, subtending the florets and usually deciduous, in two small genera (one of them monotypic). The corollas of the female florets are usually ligulate, with a trimerous spreading lamina, entire or toothed, the two inner lobes entirely wanting, or rarely present in the shape of one or two fine teeth or short filiform lobes ; the disk-florets regular, with five, or rarely four, short lobes or teeth, the latter yellow or rarely white or purple; the ray-florets homochromous or, in a few species only, heterochromous. The anthers have the normal terminal appendage, and are usually sagittate at the base, with the auricles usually aeute and some- times produced into small points, but never into the hair-like tails of most Inuloidew. The ordinary form of the style-branches of the disk-florets is with dilated truncate penicillate tips, as in Anthemidex ; but these in some genera bear obtuse or acute ter- minal appendages, hairy, as in Asteroidez, but perhaps rather less flattened, and therefore called cones, and their hairs usually shorter than those which surround the base of these appendages ; but in a few cases the styles pass into those of Asteroidew, and in others the branches are so narrow and so much more equally hairy or papillose that they become in some genera almost, in others quite, like those of Vernoniacew or Eupatoriaces. Where the disk-florets are sterile, the style-branches are usually filiform and connate to the end. The achenes are various, usually angular or terete and striate, truncate or shortly contracted at the end, not beaked, flattened only in two or three small genera, and never winged. The pappus almost always setose and copious ; the sete usually fine PRIMARY DIVISIONS ON THE ORDER. 385 and soft, simple or rarely plumose, in a few small genera rigid, in a very few small genera or species of larger ones deficient on the achenes of the ray, and in Gamolepis on all the achenes. 9. Calendulacee. Our small tribe of Calendulacee consists of De Candolle's subtribes Calendulee and Osteosperme:e, which he classes under Cynaroidew, but which we consider much more closely con- nected with Senecionidesm. They have the style of the latter tribe, and differ from it chiefly in their involucre, the constant deficiency of pappus, and usually by the forms assumed by the ripe achenes of the outer, or of the next to the outer, row of florets. Calendulacez are herbs or small shrubs, usually much branched or spreading from the base ; the leaves are alternate or very rarely opposite, entire or toothed or lobed, very rarely much divided. The involucral bracts are usually narrow, more open than in Senecio- nidez, in one, two, or rarely three rows, herbaceous or membra- nous, often with scarious margins. The capitula are heteroga- mous, with the female florets in a single row, either fertile or sterile; the central disk-florets almost always, and sometimes the whole of them, sterile; the outer row of disk-florets often the most perfect of the capitulum. The receptacle is naked, or rarely bears a few setze amongst the florets. "The corollas of the female florets are ligulate, with a trimerous spreading or sometimes re- duced lamina, entire or toothed; the two inner lobes entirely deficient; the disk-florets regular and usually with five short lobes or teeth, usually yellow. The corollas of the ray homochromous. The anthers have the normal terminal appendage, and are sagit- tate at the base, with acute, or rarely obtuse, auricles, often pro- duced into small points, which might be regarded as short tails. The styles of the disk-florets, when these are fertile, have their branches usually truncate and penicillate at the tips, as in Sene- cionidez ; in the sterile florets the style is often undivided, as in many other tribes. The fertile achenes, or at least the outer row, often acquire an unusual development and various irregular shapes, much curved or arched or winged, or sometimes thick and hard, and they have never any pappus. 10. Arctotidee. Our tribe of Arctotidex comprises, first, the subtribes Arctoter 2F2 886 MR. G. BENTHAM ON COMPOSITE. and Gorteriexz of De Candolle’s Cynaroidee, differing from those - we leave in the latter tribe in their heterogamous usually radiate capitula, in the absence of anther-tails, and, in some measure, in their styles ; and, secondly, of the genera Gundelia and Platy- carpha, withdrawn from Vernoniacez, of which they have neither the habit nor the styles nor any other important characters. Thus constituted, the Arctotidez: appear to form a connexion between Anthemides and Calendulacez on the one hand and Cynaroideze on the other. Arctotides are sometimes stemless or spreading herbs, more or less cottony-white or glabrous and Anthemis-like, sometimes with the prickly aspect of many Cynaroidex, occasionally assuming the common South-African small shrubby form. The leaves are always alternate or radical, often much lobed and sometimes prickly-toothed. The involucral bracts are imbricate in several series, free with the inner ones often scarious at the end in the subtribe Euarctotez, usually connate at the base, and hardening after flowering in the Gorteriex, pungent or prickly-toothed in many Gorteriez and in Gundeliew. The capitula are (with very few exceptions where the ray-florets are wanting) heterogamous and radiate, the florets of the ray in a single row and often sterile, and the disk-florets either all fertile or the inner ones, rarely all, sterile. In Gundeliee the ray-florets are wanting and the capi- tula collected in close clusters or compound heads, with more or less of a common leafy involucre. The receptacle is usually without pale, but in several genera very deeply honeycombed, the raised cells almost or quite enclosing the achenes, or, in one genus, truly paleaceous. The corollas of the female florets are always ligulate, with a trimerous spreading lamina, entire or toothed, the two inner lobes usually deficient, but occasionally appearing as two short or slender teeth; the disk-florets regular, with five or rarely four short lobes or teeth (as in Anthemidez); in most Euarctotew the limb deeply divided into narrow, often erect, lobes (as in Cynaroidezx) ; in most Gorteries and Gundeliex the disk-florets yellow, the rays homochromous or reddish or purple, especially outside. The anthers have the normal terminal appendage ; they are more or less sagittate at the base; the auricles sometimes obtuse, more frequently acute or produced into short points, without the distinct tails of most Cynaroidez. The style-branches of the perfect disk-florets are linear or oblong, long or short, usually obtuse and papillose outside, the papillz PRIMARY DIVISIONS OF THE ORDER. 387 extending frequently below the ramification, but not ending in an , abrupt ring of hairs or swelling as in most Cynaroidew. In the sterile disk-florets the style is usually undivided. The achenes are usually rather thick, often angular, and sometimes winged, and never beaked, occasionally densely hairy or woolly, without any or with a coroniform or paleaceous pappus. 11. Cynaroidee. We reduce the Cynaroidee to the last nine of De Candolle’s subtribes, which never have the capitula normally radiate, and in which the styles have usually an abrupt thickening or change of texture or ring of hairs below the ramification, at a point where the external papille commence; the branches are also usually very short or reduced to mere teeth. These characters, how- ever, are not quite constant; and it is not easy to express in words any definite limitation of the tribe. Some Mutisiacew (Gochnatiez), indeed, approach it very nearly; but otherwise I believe, there is never any ground for hesitation as to including or excluding any genus in or from it. The Cynaroidex are, with very few exceptions, herbaceous, and often assume that peculiar prickly habit which gives them the common name of Thistles. Their indumentum is usually loosely cottony or woolly, rarely silky or hispid. The leaves are always alternate, often sinuate or lobed or divided and prickly-toothed, but sometimes entire and rigidly ciliate or quite unarmed and soft. The involucral bracts are always imbricate, in several, often numerous, rows. The capitula are usually homogamous, but have occasionally an external row (sometimes only a very few) of sterile or female florets. The receptacle, often thick and hard or fleshy, is usually densely covered with rigid, almost paleaceous, sete, longer or shorter than the achenes, sometimes more or less united into pales at the base, in a very few small genera reduced to mere fimbrille or minute teeth bordering the slight pits or areoles of the receptacle. The corollas of the circumferential, neutral or rarely female, florets have the limb usually enlarged and regular or slightly irregular, sometimes small and distinctly bilabiate; or if expanded into a ligula, it is pentamerous, as in Cichoriaceze, only more deeply lobed, like that of Stokesia in Vernoniacez, never showing the trimerous ligula of the normally radiate tribes. The corollas of the hermaphrodite florets have the deeply and narrowly lobed limb of Vernoniacex, but often 388 MR. G. BENTHAM ON COMPOSITS. oblique, still longer and narrower, and more abruptly dilated from the tube. The anthers, with the normal terminal appendages, are sagittate at the base; and, with very few exceptions, the auricles are produced into fringed appendages or tails. The style is usually, as above mentioned, an abrupt thickening or change of texture or ring of hairs (called by Cassini an articulation) at a greater or less distance below the ramification, and is more or less papillose (not hairy) from that to the end: the branches are most frequently short, rather obtuse, and erect or at length spreading, sometimes reduced to a scarcely perceptible notch, sometimes longer and linear ; and in a few cases the external ring disappears and the branches are elongated and slightly dilated upwards so as to bring the style nearer to that of the Arctotidesw and of some Inuloidew. The achenes are usually thick and often hard, rarely flat or winged, and never beaked. The pappus most frequently consists of several, often numerous, rows of rigid sete, inereasing in length from the outer to the inner, or to the next to the inner, row, with sometimes an innermost row more definite in numbers and more paleaceous ; or the whole pappus may consist of this single row of definite or indefinite pales or sets, with very few or none at all of the outer sete; or the pappus may be reduced to short deciduous sete or palese or be entirely deficient. 12. Mutisiacee. Our Mutisiacee correspond to the Labiatiflore, considered a suborder by several synantherologists, and comprise the Muti- siacee and Nassauviacee ot De Candolle, excepting, however, from the former the small group of Facelidee, which had been inserted by Lessing and retained by De Candolle for reasons not very. intelligible, and since correctly referred by Weddell to Gnaphaliee. They differ generally from Cynaroidee in the corolla of some or all the florets being more or less bilabiate, and the want of the rigid setz of the receptacle characteristic of most Cynaroidese ; but both characters have exceptions, and the precise limits of the tribe are difficult to fix in general terms. The habit of Mutisiacez is most variable, but very frequently shrubby, or almost stemless except the radical scapes ; the leaves alternate or radical, except in two monotypic genera, entire or toothed or pinnatifid, very rarely much divided or prickly. The involucral bracts usually imbricate in several rows, rarely forming a single row of equal bracts, with or without small outer ones, as PRIMARY DIVISIONS OF THE ORDER. 389 in Senecionidem. The capitula are homogamous, heterogamous, or sometimes dicecious ; when heterogamous, often more or less * radiate, but not always in the contrasted manner of the preceding tribes, the corollas being gradually modified or enlarged from the centre to the circumference, as in some Cichoriaceze, or the ray- florets are irregular or 4-5-merous. The receptacle is naked or shortly alveolate or fimbrilliferous, not rigidly setose as in Cynaroidez, but in a few small genera bearing simple pale subtending the florets. The corollas exhibit every variation of the 5-merous limb, regular, bilabiate in various proportions, or expanded into a ligula, in a very few exceptional species becoming trimerous by the suppression of the inner lobes. The colour is very various; but the capitula appear to be always homochromous. The anthers have the normal terminal ap- pendage ; and the basal auricles have conspicuous, usually fringed, appendages or tails, except in two genera, where there is no trace of them. The style-branches when long are nearly those of Inuloides ; but they are more frequently short or erect and con- nivent, or very shortly spreading at the tips, usually obtuse or truncate, always without terminal appendages, and papillose outside or penicillate at the tip in a few genera. The achenes vary much in form; their pappus is usually setose, simple or plumose, or formed of narrow pales, very rarely deficient. 13. Cichoriacee. The Cichoriacez remain within the same definite limits ori- ginally assigned to them by Jussieu; they are at once known by their homogamous capitula, with the corallas all expanded into a 5-merous, truncate, 5-toothed ligula; and the only approach to them is traceable in the Vernoniaceous genus Stokesia, or in the Mutisiaceous genus Catamizis. Cichoriacez are usually herbs, very rarely growing into shrubs or small trees, and are as rarely thistle-like and prickly. The leaves are always alternate or radical, entire or toothed or pinnati- fid, rarely much divided. The involucral bracts are imbricate in several rows, or equal in a single row, with or without external smaller ones, and are usually membranous or herbaceous, rarely scarious or rigid and prickly. The capitula are always homoga- mous, the outer row of florets sometimes rather longer, forming a sort of ray. The receptacle is naked, or in a few genera bears pale: or sete subtending the florets. The corollas are uniformly 990 MR. G. BENTHAM ON COMPOSITA. ligulate, with a pentamerous spreading lamina, truncate and 5- toothed at the end. The anthers have the normal terminal appen- dage, and are sagittate at the base, with acute auricles often pro- duced into a short point. The style-branches are those of the female florets of other tribes, slender, acute or almost obtuse, papillose but not hirsute, sometimes rather broader and slightly flattened. The achenes are various, usually narrow or flat, and sometimes produced into a slender beak. The pappus usually with one or more rows of simple or plumose set ; but sometimes it con- sists of thin pales or of few sete or awns, or is entirely wanting. IV. HISTORY AND GEoanAPHICAL DISTRIBUTION. The ancient history of Composite must be more purely con- jectural than that of many other large groups of plants. The geological record is remarkably scanty ; but in the case of the members of this order, the absence of their remains is no proof of their non-existence at various geological periods. They are very rarely aquatic; and a comparatively small number only are to be met with on the borders of such waters as are wont to ac- cumulate stores of organic remains ; nor yet do they shed a profu- sion of leaves likely to be carried to any such hoarding-places. The great mass of them live, die, and are thoroughly consumed, without leaving a single fragment to serve as evidence or indication to future generations. It is only here and there that the winds appear to have carried an achene, by means of its pappus, to some place of deposit; and thus it is that Oswald Heer found in the upper miocene tertiary deposits of central Europe various im- pressions which he refers, on plausible grounds, to Composite. He is also probably justified in his conjecture that the great majority of them belong to Cichoriaces, two or three to Cyna- roidez, and that one is probably the achene of an aquatic Bidens. All this, if well founded, would show that at that tertiary epoch Composite existed in Europe of the same general character as those which are now to be met with. It would seem to prove that they had then already attained that highly differentiated character they now possess, and consequently must have been already of very old date, although they had left no previous record of their existence which has as yet been exposed to our observa- tion. I can find no further reliable notice of fossil Composite ; for Isay nothing of Massalonghi’s Silphidiwm-leaves; their reference HISTORY AND GEOGRAPHICAL DISTRIBUTION. 391 to Composite is only a wild guess without a particle of evidence in support of it; large herbaceous Composite are not in the habit of casting their leaves unwithered, so as to have become encased in mud unaltered in shape. Still less need I take notice of Martius’s Lychnophorites, which have since been more plausibly referred to large monocotyledons of the Vellozia type. And even Heer’s above-mentioned Miocene Composite achenes are doubted by some paleontologists, who contend that they are seeds of Apocynexe. Some, indeed, of Heer’s figures show the pappus not to be strictly terminal, but to proceed from an oblique or some- what lateral notch, which is unusual though not unknown in Composite (e. g. Tourneuxia) ; but many of the figures might be identified with more than one recent achene and pappus. In the absence of all direct evidence we are left to judge of the antiquity and origin of Composite from their comparative struc- ture and from their geographical distribution, as to both of which we have still much to learn, and in both which respects several of the boldest of modern hypothesists have neglected or been ignorant of much that is known. A general notion is prevalent, especially among French botan- ists, that Composite are at the summit of the scale of progression in thevegetable kingdom—that De Candolle's idea that the greatest perfection was to be sought for where, as in some Thalamiflorz, the essential parts of the flower, the petals, stamens, and carpels are the most distinct from each other, is altogether erroneous—that these Thalamiflore are, in fact, the nearest to the Monochlamydes, which commence from the base of the Dicotyledonous scale—and that the high degree of consolidation in the floral organs of Compositz is a strong proof of perfection and thence of a com- paratively recent origin. It seems very probable that these views may be correct ; yet, on the other hand, we must bear in mind that the numerous monotypic or oligotypic highly distinct genera con- fined respectively to the widely distant centres of preservation of the Mediterranean region, tropical and Southern Africa, Southern and Western Australia, Chili, the Mexican region, &c., point to a very wide dispersion of the original stock of the order at a very early period, when the physical configuration of the surface of the globe must have been very different from what it is now,—that this dispersion appears, indeed, to have been so early as to give time for the absolute fixation of secondary characters, which in most orders are very inconstant—and that, moreover, 392 MR. G. BENTHAM ON COMPOSITE. previously to this dispersion the stock must have existed long enough to give absolute permanence and an otherwise unexampled constancy to those essential characters of primary importance which I shall recur to in detail. The presence of an involucre, the symmetry of the floral organs, the abortion of the calyx-limb, the estivation of the corolla-lobes, the syngenesy of the anthers, and, above all, the structure of the pistil, fruit, and seed are not known to offer a single exception throughout the ten thousand species of the order. But although Composite must thus have existed, in some shape or other, but yet with all these essential characters, at an early geological period, the differentiation of the larger groups proba- bly took place after the isolation of the actual centres of preser- vation. Of the thirteen tribes: adopted, two only, Asteroideze and Senecionidez, may be said to be cosmopolitan or nearly so. Cichoriacew, Cynaroidexe, and Anthemidez belong to the northern hemisphere with chief centres in the Mediterranean and Central Asiatic regions ; a few, but those (except some Cichoriacez) either forming part of or closely allied to Europeo-Asiatic genera, have spread over North America and even down the Andes to extra- tropical South America. Calendulacee and Arctotides are African, extending sparingly into Europe. Vernoniaces, Eupa- toriaceze, Helianthoidee, Helenioidez, and Mutisiaces are essen- tially American, but with a few types which may have arisen in the tropical and subtropical regions of Africa and Asia. The great tribe of Inuloidex is for the most part Old- World, although the subtribes Plucheinez and Gnaphaliez have been long enough in America to have there formed a very few generic types. Be- fore, however, drawing any further. general hypothetical conclu- sions as to the early history of the order, and the course of its present distribution, it will be necessary to recapitulate succes- sively the data hitherto supplied to us by the several tribes, sub- tribes, and principal genera it is composed of. For this purpose I propose taking the several tribes successively in their systematic order, although in the further details under each tribe I shall endeavour to take genera and subordinate races as much as possible in their natural rather than in their technical limits, I shall then proceed to consider the chief centres or regions occupied by the present races of Composite, the limits to be assigned to them, their distinctive characters and mutual con- nexions. I must now, however, observe, to prevent misunderstand- HISTORY AND GEOGRAPHICAL DISTRIBUTION. 393 ings in the following detailed notes, that these regions are founded solely on the distribution of Composite, and would require much modification as to those orders of plants especially which take a part in forest vegetation. I must also explain that the Mexican region, here often referred to, includes California, W. Texas, and a portion of Central America, and that the Mediterranean region, in the Composite sense, must include the Levant and Persia. A. DISTRIBUTION OF THE TRIBES AND PRINCIPAL GENERA. 1. Vernoniacee. This tribe, as limited for systematic purposes, consists princi- pally of one large genus with a number of smaller ones closely connected with it, forming altogether one subtribual or generic group of a higher order; and to this are added afew small genera so distinct as to leave some doubts as to their real affinities. I shall here, therefore, commence with the principal genus. Vernonia, a genus, as we now propose to limit it, of nearly 400 species, and closely connected with about 25 smaller genera com- prising above 120 more species, has its chief centres in tropical America and tropical Africa, forming in both countries more or less divergent groups, but in different directions, the species more numerous in America, the forms more varied in Africa. From tropical America it spreads more sparingly into North America and extratropical South America, and from tropical into Southern Africa, and eastward into tropical and subtropical Asia, forming in each of these outlying districts more or less local groups. More than three fourths of the genus belong to the section Lepidaploa, which, rather from its wide geographical range and connexions than from its happening to include the species first taken as the type, may be conjectured to be nearest to the original form. At least four fifths of its species are tropical American ; but it includes also the North-American ones, a portion of those from Africa, and five or six Asiatic species. In this section the achenes are equally 10-ribbed, with the inner sete of the pappus long and fine, rarely slightly dilated, the outer numerous and short, more frequently flattened and almost scale-like than fine. This great multitude of species has to be methodized and distinguished by foliage, general inflorescence, size, and shape of the capitula, and by the obtuse, acute, or aristate, appressed, or squarrose invo- 394 MR. G. BENTHAM ON COMPOSITAE. lucral bracts; but even these secondary characters are most frequently differently combined in the New and in the Old World and in the different regions of each. No other section or group of Vernonia can be strictly said to be common to the New and the Old World. Aberrant modifications of some one of the more important characters may indeed be observed in both the regions, but differently combined in the two with other characters. 1. In the two Brazilian species of Hololepis and the single Mexican Leiboldia there is not the usual disproportion in the pappus, the outer sete being but little distinguishable from the inner in length, numbers, or rigidity. The Brazilian Hololepis is, moreover, remarkable for: the large involucres, of which the outer foliaceous bracts conceal the inner ones. The Mexican .Leiboldia has also rather large involucres; but its bracts are all acute and normally accrescent from the outer to the inner. In the Old World we have in the Vernonia calycina, Wall, from Prome, a close representative of the Brazilian Hololepis as to habit and involucre, and the pappus also abnormal, but in a direction pecu- liar to the Old World; the outer sete, though longer than in Lepidaploa, are fewer and shorter than the inner ones. In the tropical-African V. purpurea, Sch. Bip. the outer bracts are sometimes enlarged and foliaceous as in Y. calycina; but the pappus is normal, and the plant is in other respects a true Lepidaploa. 2. The tropical-American Critoniopsis, containing five or six species, with a normal pappus, is distinguished by its ample panicles of small few-flowered capitula, the involucral bracts obtuse, the inner ones frequently very deciduous, as in the allied genus Piptocarpha. This section is not identically present in the Old World, but is there represented by the section Strobocalyx of about a dozen species, with similar inflorescence and capitula ; but the pappus has the tendency, so frequent in Old-World Vernonia, to the attenuation, reduction, or almost total disappear- ance of the small outer sete of the pappus. The aberrant forms peculiar either to the New or to the Old World, which, however, we have thought not sufficiently distinct to retain as separate genera, are the following :— 1. In tropical America the three small sections Stenocephalum, Trianthea, and Eremosis have the normal pappus and other essential characters of the original Vernonia; but the narrow involucres contain but very few florets. In Stenocephalum the DISTRIBUTION OF TRIBES. 395 inflorescence approaches that of the Lychnophoree, a truly Ame- rican type ; but the section is closely connected with the ordinary Lepidaplog through V. obtusata, Less., and others. Trianthea is more normal in general inflorescence as well as in the involucre, except that it contains usually only three florets; and Hremosis only differs in these florets being reduced to a single one. Triana’s Lherasia, from the mountains of Columbia, must be included in the section Trianthea, with a habit still nearer to that of Oliganthes. In the Old World when the Vernonia-capitulum becomes 3-flowered or only single-flowered (Monosis Wightiana, DC.) the species assume the habit, inflorescence, and involucres of Strobocalyx above referred to. Vernonia complicata, V. bahamensis, and V. lepidota, Griseb., from Cuba or the Bahamas, are insular aberrant forms which require further study and comparison with the S.- American genus Piptolepis, or the San-Domingo Piptocoma, of which last I have seen no specimen. 2. In the Old World the section Stengelia, including Ascaricida, with the inflorescence, pappus, &c. of Lepidaploa, diverges in the involucres, of which the inner bracts end more or less in membra- nous coloured appendages, a character not carried further in any allied genera. In some species the sete of the pappus become flattened and more rigid, showing a tendency towards the palea- ceous pappus of the American Stilpnopappus and its allies—a tendeney which, however, does not appear to be carried any further in any Old-World genus allied to Vernonia; for Herderia is probably more nearly connected with Ethulia. The sections Gymnanthemum and Xipholepis are chiefly charac- terized by the reduction of the outer pappus to fine sete, few in number, of variable length, and sometimes disappearing altogether —a circumstance towards which American forms show no tendency, except in Lachnorhiza, a Cuban species, which, on account chiefly of its peculiar habit and the form of its involucre, very different from those of the Old- World Gymnanthema, has been admitted as a distinct insular genus. The sections Cyanopis, Lepidella, and Tephrodes exhibit a divergence which would at first sight appear important enough to raise them into a distinct genus, but that the main character, the 4- or 5-angled, or almost terete and nerveless, not equally 10-ribbed achene, is so variously combined with others in species which in other respects would belong to Lepidaploa, and is, moreover, so 396 MR. G. BENTHAM ON COMPOSITS. vague in some species where the secondary ribs are present but much less prominent than the primary ones, that no tolerably fair line of demarcation can be drawn between the groups. The pappus in the three sections is very nearly the normal one of Vernonia: that of Tephrodes usually white, copious, and rather soft; that of Cyanopis more fragile, approaching that of Centratherum ; that of Lepidella with the outer short row more paleaceous, as in many American Lepidaploe. No American species, as far as I am aware, shows the 4- or 5-augled or the nerveless achenes of these Old-World groups. The Mascarene section Distephanus (two species), with a some- what different habit, a compact inflorescence, and a remarkably fimbrillate receptacle, might perhaps be maintained as a distinct genus; but the extreme vagueness of the character derived from the receptacle (a tendeney to which is observable in several African species) will probably justify its reduction to a section, chiefly geographical, of Vernonia. Another supposed Mascarene genus, Bechium (a single species), appears to me to be a true Lepidaploa, with a slight tendency to the coloured tips to the bracts of the involucre of Stengelia, and a somewhat peculiar habit, the leaves being almost radical. The numerous Vernonioid genera, most of them small or mono- typic, which have been maintained around Vernonia, partly from habit, but chiefly on account of more or less marked divergences in characters regarded as essential, but which are yet connected with the main group by small gradations, may be classed as follows according to the nature of those divergences :— Ist. In the mucronate or subcaudate anther-auricles.— This is exhibited in one American and three Old- World genera. The American Piptocarpha, with nearly twenty species, is closely connected in involucre, and in one species in inflorescence, with the American section Critoniopsis and the Old- World Strobocalys ; most species, however, have a peculiar inflorescence of a character much more American than Asiatic or African. The pappus is sometimes the normal Vernonian double one ; sometimes the outer series is reduced to a few fine sete, or disappears altogether, as in the Old-World Strobocalya. In the Old World the few cases of Vernoniaces with subcaudate anthers show a different combinationof other characters. Thethree genera Centauropsis from Madagascar with two species, Adenoon from the E.-Indian peninsula, and Pleurocarpea from N. Aus- DISTRIBUTION OF TRIBES. 397 tralia, both monotypic, have each a much reduced pappus ; in Cen- tauropsis it is copious, but shorter than the achene, in Pleurocarpea reduced to the few exceedingly caducous sete of the American Centrathere, in Adenoon entirely deficient. In all three the involucral bracts are very persistent, quite different from those of the American Piptocarpha. 2nd. Divergences in the receptacle.—In the Old World the receptacle rather deeply foveolate and fimbrilliferous, which has served to distinguish the above-mentioned Mascarene section Distephanus and the genus Bothriocline, a single tropical African species, is but little more than may be observed occasionally in several Vernoniacee of other genera, both in the Old and the New World, and even in Vernonia itself; in Distephanus it seems to be accompanied by some slight difference in habit, in Bothriocline by a peculiar involucre, as also by opposite or whorled leaves, a cireum- stance otherwise unknown in the whole tribe of Vernoniaces except in a single Brazilian, otherwise normal, Vernonia. In America, among the genera otherwise closely allied to Vernonia, the Bra- zilian monotypic Albertinia has the receptacle so deeply alveolate as completely to envelop the achenes; and the Mexican Bolanosa, also monotypic, has separate caducous pales embracing?each floret, with the habit, involucre, and other characters quite those of a Vernonia (Lepidaploa). Neither of the above characters, however, is of more than generic value, occurring in groups otherwise very far distant from each other. . The deeply alveolate receptacle is met with, for instance, in the Helianthoid Balduina from North America, and in several South-African Arctotidee. The deci- duous palez of Bolanosa occur here and there in species or genera of Eupatoriacee, Asteroides, Senecionidew, and other tribes usually deprived of them. 3rd. Divergences in the pappus.—The indefinite sete of the inner or principal pappus of Vernonia assume a paleaceous character in several tropical American genera, otherwise nearly connected with Vernonia (Stilpnopappus and others); whilst in the Old World there is only a slight tendency in this direction, as already observed in some species of the section Stengelia, with- out being there carried further. On the other hand, the same paleaceous pappus prevails generally, though not universally, through the strictly American genera with glomerate capitula to be mentioned presently. This is, therefore, a type much developed in Ameriea, but early arrested in the Old World, or an old type 398 MR. G. BENTHAM ON COMPOSITAE. which has retained more permanency in the former than in the latter region. The reduced pappus is common to the New and the Old World, but more frequent and more regularly connected with that of Ver- nonia in the latter than in the former. Centratherum, with the sete exceedingly caducous and usually a peculiar involucre, is common to both regions, but much more marked in America, where it is limited also to one or two species passing into Oiosper- mum without any pappus at all, and never having any near approach to the ordinary pappus of Vernonia; whilst in Asia and Africa it is connected by various gradations with Vernonia, especially with the section Cyanopis, in which also the pappus is frequently deci- duous, and does not seem to pass gradually into the African pappus- less genera Ethulia and Gutenbergia. Thelatter,indeed,' especially Ethulia, seem to be more nearly connected with Herderia, also African, in which the developed pappus has a very different character from that of Vernonia and its modifications. 4th. Divergences in the involucre.—These are so great within the genus Vernonia itself, that they are scarcely admissible as distinctive generic marks, except when combined with other characters as in the above-mentioned Centratherum and Oiosper- mum, or with habit and geographical isolation, as in the Cuban Lachnorhiza, to which also I have already referred. 5th. Divergence in the general inflorescence.—A large group of American Vernoniacew diverge gradually from the typical Vernonie in their compound inflorescence—the flower-heads, either 1- or few-flowered, being closely sessile several or many together on a common receptacle, so as to give the whole mass the general appearance of a single capitulum. In Lychnophora and most of its allies this change is accompanied by an alteration in the pappus, which is paleaceous as in Stilpnopappus; but in Eremanthus it passes sometimes almost into that of Vernonia, and in Vanillosmopsis into that*of Centratherum, whilst in some Vernonia and Piptocarphe the few-flowered capitula, sessile in small clusters, further connect the true Vernonie with the Lych- nophoree. All these modifications are American ; the Old- World Vernonie and groups immediately connected with them show no tendency to the compound inflorescence, although it is there exhibited in several genera belonging to other tribes. The mere reduction of the number of florets to very few or to a single one, which induced the establishment of the Separate genus DISTRIBUTION OF TRIBES. 399 Monosis, is now found to take place in so many and so various genera of Composite, that it is no longer held to be more than a specific character. The tribe of V ernoniacez is made to include several small groups or isolated species, which, although they have the general characters of the tribe, are not, as far as at present known, connected with the central Vernonia group by any gradation of intermediate forms; these are :— 1. Elephantopus, somewhat related to Lychnophorez, will be further mentioned under the head of American genera with one cosmopolitan species. 2. Sparganophorus, Pacourina, and Heterocoma, three mono- typic tropical American genera, connected with each other in in- florescence and in the development on the top of the achene of a prominent ring or cup, either alone constituting the pappus or encircling a pappus of small caducous bristles. The essential characters are truly Vernonian; and some approach to their peculiar inflorescence may be seen in a few species of Vernonia, Piptocarpha, Stilpnopappus dec. ; but of the terminal cartilaginous ring the only trace I have observed is in the Old- World Ethulia, very different in general character. 3. Stokesia, a monotypic North-American genus, stands alone without any near relations. Its remoter affinities are on the one hand with Vernoniacex, of which it has the style and anthers, and under which it is therefore classed, and on the other with Cichoriacex, which it approaches in its blue almost ligulate and 5-merous although deeply lobed corollas. It is the only genus not Mutisiaceous which shows any such connexion with that very marked tribe. 4. Corymbium, a small South-African genus with the Vernonian style; but the habit, involucre, silky-villous achenes, and the pappus are very different from any thing else known in the tribe. No other nearer connexion, however, has occurred to me. 5. Rolandra and Spiracantha, two monotypic tropical American genera, with very numerous minute 1-flowered capitula, collected in globular head-like axillary clusters, as in some Lychnophoreex, and perhaps on the whole best placed, as hitherto, amongst Ver- noniacee, although they have neither the characteristic style nor the corolla of the tribe, showing possibly some connexion, although a distant one, with some American small-flowered Helianthoidee or Helenioidez. LINN. JOURN.— BOTANY, VOL. XUI, 28 400 MR. G. BENTHAM ON COMPOSITE. 2. Eupatoriacee. Eupatoriacese may be regarded as one large and natural essen- tially American group or genus in an extended sense of the term ; for, multifarious and distinct as it is, some of the last mentioned small groups or monotypic genera of Vernoniacez rank as high in the latter respect. Eupatoriacee must, therefore, either not be so ancient as some other groups of Composite, or some other reason must have interfered with their early dispersion; for although abundant and evidently early established over the whole of temperate and tropical America, from California to Chili and Buenos Ayres, as evidenced by the distinct local generic groups they have formed in North America, in the Mexican region, in the tropical Andes, in Brazil, and in Chili, they are either wholly absent or have not, with the single possible exception to be pre- sently mentioned, produced any distinct species in tropical or transtropical Africa, Asia, or Australia. They are indeed entirely absent from the Australian flora, and would be also wanting in the floras of tropical and southern Africa and tropical Asia, but for three essentially American genera, Adenostemma, Ageratum, and Mikania, which have each one cosmopolitan species, and will be considered hereafter, and for one or two species of Eupatorium itself, which from Northern Asia may have penetrated within the tropical limits. This leads us to the northern and sole connexion of Eupatori- acese with the Old World, which may be observed in two genera: —1st, in the wide-spread American genus Eupatorium itself, whieh has in that continent above 400 species, and is represented in the Old World by about 8 or 10 rather variable species, all of one North-American type, although not exactly identical with any one species of that country, most of them from Eastern Asia, one of which extends over the whole of Europe, another (by some considered an extreme variety of an Asiatic one) is scattered over the Mediterranean region, and reappears almost in the same form in the Canary Islands, and another, also near the Asiatic ones, has been found on the Zambesi in south tropical Africa. There is, indeed, such a general family likeness between these Old-World forms and some of their E. North-American congeners, that they may well be imagined to have sprung from some parent race that may have passed over from America, and in the various vicissitudes of their career through the lapse of ages, spreading gradually over a vast extent of territory, disappear- DISTRIBUTION OF TRIBES. 401 ing in some places and flourishing and further migrating in others, have become gradually differentiated into species which may scarcely yet be taken as perfected. 2nd, in the genus Adenostyles, which has one Californian and two European species, without any known representative in the vast intervening regions of Asia on the one side, or of central and eastern North Ameriea on the other. "There is no doubt of the generic identity ; the Californian species, though quite distinct from either of our European ones, is yet more nearly allied to one of them (4. alpina) than the two European ones are to each other. Like many others now con- fined to America, the genus probably once ranged over a great part of the temperate northern hemisphere, and instead of dis- appearing from the whole of the Old World, has kept its ground in Europe, becoming extinct in Asia. It is this same genus Adenostyles which alone supplies some sort of link connecting Eupatoriacex with any other tribe ; for in habit and involucre it shows some approach to some species of Senecio ( Cacalia), although the style and other characters leave no doubt as to its being a true Eupatoriacea. . Eupatoriacee, within American limits, have formed a number of genera, groups of genera, or subgenera, more or less local or general, but most of them passing so nearly one into the other as to require little notice for the present purpose. The most re- markable is perhaps the group or subtribe of Piqueriex, charac- terized by the anthers truncate at the top, without that appen- dage to the connectivum so universal in the rest of Composite. It comprises 7 genera, with about 80 species, chiefly Western, ranging from Chili to Mexico, with two or three South-Brazilian and three Cuban species; it includes also the above-mentioned Adenostemma, to which I shall refer under the head of genera with one cosmopolitan species. This remarkable deviation from the almost absolute uniformity of Composite is probably, there- fore, of West-American origin, and not ancient enough to have spread into other continents now severed from America. I have not observed it in any other group of Compositie, although the appendage may be very much reduced iu a very few tropical American Helianthoidez. The other deviation I would mention is not so important, nor very strictly defined, but purely local. The three genera (or sub- genera) Liatris, Trilisa, and Carphephorus form a little North- American group, almost limited to the regions east of the Andes or 2a2 402 MR. G, BENTHAM ON COMPOSITE. Rocky Mountains. Their constantly alternate leaves, their general habit, and, to a certain degree, their corollas, may indicate some approach to the Vernoniaceæ; but their styles and other most important characters are essentially Eupatoriaceous. 3. Asteroidee. The vast tribe of Asteroides is neither so well marked as a whole as Vernoniacese and Eupatoriacee, nor yet is it well divisible into distinct groups. Nearly the whole of the 90 genera, comprising above 1400 species, pass into each other through ex- ceptional or intermediate forms ; and there are species if not genera closely connecting Asteroides with Inuloides, with Anthemidee, and with Senecionides. Of these four intimately connected tribes, comprising about half the known Composite, two, Asteroidew and Senecionidex, are cosmopolitan, the two others chiefly Old- World. The Asteroidew not being divisible into distinct subtribes, we may for geographical purposes consider a certain number of types with the various divergences from them, and then take up a few comparatively isolated forms. The principal of these types are Grangea, Bellis, Solidago, Aster, Erigeron, Conyza, and Baccharis. Six of them are eommon to the New and the Old World, the first two chiefly belonging to the Old World, the next four more numerous in the New, and Baccharis entirely American. The whole tribe affects chiefly temperate or mountain regions of both hemispheres, tropical Asteroides being comparatively rare. 1. The Aster type.—Aster, taken in its most extended sense, ranges over the whole area of the tribe; but isolation has been ancient enough to admit of its having established special forms in different countries, which are now admitted as genera by most botanists. Aster itself, as we have limited it, forms a group of about 250 species belonging to the temperate regions of the northern hemisphere, chiefly North American, a very few species rather abnormal descending along the western regions into South America, and a mountain form connected with a European and Asiatic one crossing the equator in Eastern Africa. These true Asters are herbaceous, usually perennial, often tall, though some mountain species are quite dwarfed or almost stemless ; the invo- lueres are usually broad, the heads heterochromous, the achenes flat, and the pappus copious. Among the modificationsjobserved within the genus as we now retain it which some asterologists DISTRIBUTION OF TRIBES, 403 have considered to be in themselves of generic importance, the following are more or less connected with geographical distribu- tion :— (1) The section Alpigenia, ranging over the mountain-regions of the northern hemisphere, and including the above-mentioned transtropical African species, with the achenes and pappus of the typical Asters, acquires the habit and involucres and, in some measure, the increased numbers of ray-florets of a corresponding mountain group of Erigeron inhabiting the same regions, and thus to a certain degree connects the two genera. (2) The Andine section Noticastrum, consisting, as far as known, of two species only, with a slight modification in the ribs of the achenes, but with the pappus and other essential characters of Aster, closely connects the genus with the Andine group of Erigeron; whilst the nearly related section Heterastrum, also South American and chiefly extratropical, approaches another South- American set of Erigeron, even in the pappus, and has been alternately placed in both genera. So, also, the section Oxytripo- lium, with very few species dispersed over North and South America, connects the same set of Erigerons with the section Orthomeris of Aster. This latter section, chiefly differing in the involucre, although present in North America, has been much more developed and become much more varied in Asia. (3) Tripolium and Galatella are Europseo-Asiatie forms charac- terized chiefly by the reduction, sterility, or disappearance of the florets of the ray, towards which there is very little tendency in N. America, although Galatella itself is there represented by one species. Tripoliumis a single very variable species with a peculiar involuere, the ray-florets sometimes abundant and conspicuous, sometimes very few or absolutely none. Galatella has the invo- lucre of the section Orthomeris (the one most abundant in Asia) ; but the female florets are usually sterile or have not even a rudi- mentary style, or in some species are usually entirely deficient. For the Jatter case the genus Linosyris was established; but it has been proved that the Old- World species sometimes acquire the ray and become true Asters of the section Galatella. The American species associated with Linosyris on account of the absence of ray- florets belong rather to the Solidago group to be presently referred to. Close around Aster are two small genera belonging geographi- eally to parts of the main region of that genus, and which, even 404 MR. G. BENTHAM ON COMPOSITAE. on structural grounds, might perhaps be restored to it as sec- tions,—the N.-American Sericocarpus of 5 species, and the Asiatic monotypic Callistephus. Six other genera, not diverging more or perhaps even so much as the above two in essential characters, but with a very different geographical range and apparent origin as well as a distinct habit, have been by several botanists reunited with Aster, because they have no distinct character which is not liable to exception. But the species are numerous and the exceptions few ; and, geographical considerations coming in aid, it would seem to be more in confor- mity with the evidences of affinities thus obtained to maintain them as distinct genera. These are :— (1) The South-African genus Felicia (including the greater por- tion of Agathea), about 45 species, much branched shrubs, or, if herbaceous, small annuals branching from the base, all very unlike any American true Asters, retaining the achenes but not the pappus of that genus, which is more like that of Erigeron, but different from either in its fragility. The very few exceptions in their case consist in one or two eastern species approaching in some respects the alpigenous Asters represented in the same country by. A. natalensis. Closely connected with these Felicias, and perhaps not separable, is the monotypic extratropical S.-American Sommer feltia, : (2) The Australian genus Olearia, about 85 species, mostly shrubby, like the S.-African Felicias, but larger and retaining, not the achene but the pappus of the American Asters. This achene is no longer flattened, but terete or nearly so. There is also an occasional tendency to extreme acuteness or even fine points to the auricles of the anthers, never observable in any true Aster, but traceable sometimes in another Antarctic or Australasian genus, Celmisia, to be presently referred to. The exceptions in this case are a very few Australian herbaceous Olearie, but with a very different habit from any American Asters, and a very few N. American Asters (Biotie) with subterete achenes, but accompanied by a typical American involucre and other characters not to be met with in Olearia. These circumstances taken together consti- tute a much greater difference between Aster and Olearia than is observable between the alpine Asters and Erigerons. (3) The antarctic or subantarctic American genusChiliotrichium, a genus of three species, one of the numerous connecting links between the Australian and the extratropical or AndineS.-American DISTRIBUTION OF TRIBES. 405 floras. It is closely allied as well to Olearia of the former as to Diplostephium of the latter region, differing slightly from both in the presence of a few palez on the receptacle. (4) The South-American Andine genus Diplostephium, of about 18 shrubby species, more nearly connected, perhaps, with the Australian Olearia than with the American Asters, but with a habit and foliage of their own. The achenes are terete or nearly so, not flattened as in Aster; and the pappus tends much more than in that genus to the shortening of an outer row of sete. (5) & (6) Commidendron and Melanodendron, two insular genera, the former of three, the latter of a single species, all confined to the island of St. Helena, where they may have been originally diffe- rentiated from the ancient type of the group, and, like a few others of the most ancient St.-Helena plants, appear to have retained more of a S.-American (Andine or Western) than of a South- African character. Their nearest connexion is with Diplostephium, not with Felicia. 2. The Erigeron type.—Evrigeron, taken in an extended sense, has nearly the same geographical range as Aster, but without so great a tendency to develop local forms, geographical subgenera, sections, or species. It is also, in point of structural characters, very closely allied to and blending in with Aster, touching it at various points, and passing into it more gradually than the above- mentioned semi-geographical genera Felicia, Olearia, Kc. ; and yet synantherologists are unanimous in its admission into the ranks of genera of the first order. It passes, indeed, quite as gradually into Conyza, and thence into other equally large groups, which, unless we give up all idea of methodizing, must be admitted as very different, although they cannot be strictly defined. "This greater blending into allied forms may also not be due to any nearer genealogical affinities, but possibly to greater inherent facilities for propagation, dispersion, and original intercrossing of breeds. The species of Erigeron are in general less distinct from each other than those of the Aster type ; and most of them are far more widely dispersed ; a few also as annual weeds multiply exceedingly wherever they are carried with cultivation, in this respect also agreeing more with Conyza than with Aster. To distinguish, however, Oonyza and Erigeron from Aster, we have but little besides the inerease in number and reduction in size of the female florets, which in Zrigeron, although they have still the corollas produced into a ligula, have that ligula always very narrow, and 406 MR. G. BENTHAM ON COMPOSITE. often short ; whilst in Conyza these corollas are still further reduced to a filiform tube, shorter than the style, toothed or truncate at the top, the ligula remaining undeveloped. The achenes, also, both in Erigeron and Conyza, are usually much smaller than in Aster, with the pappus much less copious, of finer setæ, usually but not always in a single row. The same great facilities enjoyed by many species of Erigeron for propagation and dispersion increase the difficulty of fixing the geographical origin of some of the sections of which it is composed. Euerigeron, ranging from FE. uniflorus and E. alpinus to E. acris, belongs to the northern hemisphere, and is chiefly mountainous, passing into the section Alpigenia of Aster, and may be as much, or nearly as much, of Old-World as of American extraction. Cenotus, the section which passes into Conyza, is now pretty nearly cosmopolitan, and, like Conyza, overruns tropical as well as temperate regions, the preponderance of local species being African as well as American. The other sections appear to be chiefly or entirely American, and perhaps all of American origin. Phenactis, however, belonging to the northern hemisphere, has two genuine Asiatic species. Phalacroloma consists of American annuals, two of which have overrun a great part of the Old World as weeds, like the. well-known Æ. canadensis, which is almost inter- mediate between Cenotus and Euerigeron. | .Erigeridium is a single N.-American species of Euerigeron, somewhat aberrant in the form of the achene, a deviation which does not appear to go further in any genera of the Erigeron group. The South-American sections are rather more distinct and local. Leptostelma is a purely Bra- zilian form, resembling some of the large-flowered North-American species of Plenactis, but with an exceptionally fimbrilliferous receptacle. This also does not connect itself with neighbouring genera. Oritrophium, from the Andes, aberrant both from Aster and Erigeron in the form of the style-branches, in other respects approaches the Andine section Voticastrum of Aster nearly as much as the northern Euerigeron approaches Alpigenia; Oritrophium assumes also often the habit of the more southern Celmisia, but differs in the achene and other characters. Terranea (E. fruticosus, DC.), from the island of Juan Fernandez, approaches Aster in a third direction, being closely connected in many respects with the South-American chiefly maritime Oxytripolia. Around Erigeron may be grouped the following slightly dum gent genera, designated as much by their geographical areas as by DISTRIBUTION OF TRIBES. 407 any structural characters, and yet natural enough to be readily recognized :—1. The Antarctic and Australian genera Pleurophyl- lum and Celmisia, closely connected with the Andine secüon Oritrophium of Erigeron, differ from it chiefly in the shape of the achene, which is more that of Olearia, a genus bearing the same geo- graphical relation to Aster that Celmisia does to Erigeron, 2. The Hawaian Zetramolopium, an insular group of about half a dozen species, which during its long isolation has, like so many other insular forms, assumed more or less of a shrubby habit. In this respect it is still connected with Erigeron through the similarly insular (Juan Fernandez) Terranea above mentioned. As to structural characters, Tetramolopium has, on account of its subulate style-appendages, been connected by A. Gray with Vittadinia; but the latter genus appears to me to be further removed from Erigeron. A similar style is observable in the Andine section Oritropium of Erigeron. 3 &4. The small Asiatic extratropical genera Brachyactis and Lachnophyllum, the former with one species extending into North America, both nearly allied to Aster, Erigeron,and Conyza, the species bandied about from the one to the other, difficult tech- nically to distinguish from them, but with a peculiar habit justifying their maintenance as distinct genera, unless the three great types be reunited into a single one. 5. The Asiatic and African Microglossa, which, however, may be best considered, with its other relations, under the Conyza type. There are yet two genera connected with Aster and Erigeron, but rather more clearly distinguished on structural grounds, and to a certain degree supported geographically, both with elongated although still flattened achenes, with the pappus almost that of Aster, but with a habit more approaching Erigeron: one is Podo- coma, with five South-American and one Australian species in which the achene is attenuated into a beak, whilst the style is that of Aster, but the ligule more numerous, like those of the Aster-like Erigerons; the other, Vittadinia, has the flowers of Erigeron, or almost of Conyza, elongated but beakless achenes, the pappus nearly of Aster, and the subulate style-appendages of the section Oritropium of Erigeron ; and with these characters we find species scattered over South America and Australia, as in the case of Podocoma, but with the addition of one from the Sandwich Islands. There are, again, a number of small or monotypic genera ranging geographically around the Aster or Erigeron groups, which have 408 MR, G. BENTHAM ON COMPOSITE. been in a great measure technically distinguished by slight struc- tural differences, and which in other respects generally correspond to the Asteroid sections or diverging genera of the same respective regions, Thus we have the Asiatic and North-American Boltonia, the Asiatic Heteropappus, and the N.-American Zownsendia, Monoptilon, and Psilactis diflering from the typical Asters of the same regions in scarcely any thing but the pappus. The Californian Corethrogyne differs rather more in the style and other characters, and may possibly have other connexions. Zremiastrum is nearer to Erigeron. The Andine Hinterhubera, with remarkable irregular corollas to the external female florets, approaches in other respects the South-American Vittadinia. The same peculiarity in the corollas characterizes also the genus Lessingia, referred, perhaps erroneously, to the homochromous subtribe, and also West- American, though limited to the northern hemisphere. In South Africa, Amellus, characterized by the pales of the receptacle, and Mairia and Gymnostephium, by the plumose or reduced pappus, partake in other respects of the Asteroid form characteristic of that region, that of Felicia; so also does the monotypic Charieis, which in some respects connects the tribe with the Senecionidex. Finally, three small genera, Distasis, Chetopappa, and Minuria, the two former from W. North America, the latter from Australia, have an abnormal pappus nearly similar in the three, but not much else in common, except as members of the heterochromous group of Asteroidee ; but each will be found to be nearly allied to Asteroid genera of its own country, Distasis being comparable with the section Orthomeris of Aster, Minuria approaching in many respects the Australian Calotis. 3. The Bellis type has a somewhat unusual geographical dis- tribution. It belongs to extratropical western regions of the Old World and to Australia, with one or two representatives in North America; but its principal seat is in Australia. Indeed the European and North-African Bellis, with four or five species, the American Astranthium, one, or perhaps two, species, the Azorean Seubertia, one species, and the South-African Steirodiscus, two species, are but very imperfectly distinguished as genera from the Australian Brachycome, which has forty species; and amongst slightly diverging genera the European .Bellium has three species, the North-American (Mexican) Keerlia two, the South- African Garuleum three, whilst the Australian Calotís has fifteen. Ail these genera may be regarded as more, nearly allied to each DISTRIBUTION OF TRIBES. 409 other than to any other ones of their own respective countries, although in some respects Calotis may be compared to Minuria, and Bellis itself may be said to approach some Anthemidez in the Old World, and in the New World, through Aphanostephus, to pass into Egletes and the more tropical Grangea type. The distinct genus Lagenophora, allied on the one hand to the Bellis type of Asteroidez, on the other to the Cotula group of Anthemidez, has a more normal extratropical geographical area, having its chief seat in Australia, but one species extending here and there into S.E. Asia, and allied but distinct representatives being found in Antarctic America and the Sandwich Islands. The scattered distribution of the Bellis type of Asteroidez and of the closely allied Cotula type of Anthemidez, and the local endemic generic as well as sectional or specific types or races they have both or one of them left in each of the great centres of preservation of Composite, the Mediterranean, South-African, Australian, Chilian, and Mexican regions, might suggest the idea of comparative antiquity ; and if so, combining its consideration with that of the Helianthoidez, we might conjecture that in Com- positee the annihilation of the calyx-limb, or its reduction to a small cup or to a definite number of teeth or awns continuous with the ribs of the tube, preceded its development into a setose pappus. Under this view the parent type of the Bellidee and Cotules would in the Old World have become further differentiated into the tribe of Anthemidez, whilst in America its development would have been limited to the few nearly allied Grangeoid genera, being otherwise replaced by the more anciently separated Helenioid genera. 4. The Grangea type, with usually more numerous and smaller female florets, less compressed achenes, dic. than the Bellis type, is much more tropical. It spreads over the warmer regions of Asia and Africa with a few American forms, which latter connect it more immediately with the preceding types. Inthe Old World it remains as distinct in geographical range as in structural character. A few of the genera, especially Alyriactis and Rhynchospermum, both of which extend rather further north than the others, have the achenes flattened, with nerve-like borders so common in the preceding types, but with the beak of Zagenophora. In the re- maining genera the achenes are more like those of the Anthemides of the Cotula type ; and, still more than the Belliex, the Grangex, by their pappus reduced to a small cup or corona or entirely de- 410 MR. G. BENTHAM ON COMPOSIT®. ficient, connect the Asteroides with the Anthemides. The genera are all small; seven are from the Old World, of which three (Myriactis 5 species, Grangea 4 species, and Dichrocephala 5 spe- cies) have a wide range over Asia and Africa; whilst four monotypic ones are local, or nearly so, Rhynchospermum and Cyathocline being exclusively Asiatic, and Ceruana and Microtrichia limited to Africa. The two American genera Aphanostephus, 3 species, and Egletes, 5 species, have not thrown out any immediately divergent genera around them; the connexion of the former with the Bellis group and thence with typical Asters has been already mentioned. Eyletes is remarkable in its close resemblance in every respect to the African and Asiatie Grangeas, excepting in the single generic cha- racter the presence of the ray-flowers, which in Africa are always deficient—a character which in other genera is frequently variable even in one species, but here apparently constant and geogra- phical. 5. The Solidago type.— We have here about 320 species in 24 genera, all nearly allied to each other and only distinguished technically from Aster and its immediate allies by the homochro- mous florets, the ray-florets, when present, being yellow, like the disk—a character in general of so little value that it cannot, in Senecio for instance, be admitted as of more than specific import- ance, and yet is here accompanied by so much of habit and certain prevalent, although not absolute, peculiarities, that it is univer- sally acquiesced in, notwithstanding the frequent difficulty in as- certaining it. The erroneous appreciation of colour in dried spe- cimens has led to many mistakes, and there are several groups where the rays are deficient ; in these cases it is only by compli- cated affinities in other respects that the place of a plant can be determined. In general, the rayless species and genera belong to the Solidago group; but this absence of ray occurs sometimes even in Aster itself; and experience has now shown that the European Linosyris and American Bigelowia, united by some of the most eminent synantherologists, belong, the one to the Aster, the other to the Solidago group. And here geographical distribution may be called in aid. The great seat of the group is extratropical America, North and South, with a few intermediate Andine species. The large genus Solidago itselfis almost entirely North American and extratropical, but represented by one vari- able species in extratropical South America, and by another at least as variable in the north temperate regions of the Old World, DISTRIBUTION OF TRIBES. ALL both these outlying species being identical with, or closely allied to, genuine North-American ones. The other genera (excepting 5 to be presently mentioned) range from California to Chili and Patagonia, with a North-American and chiefly western preponder- ance, and in South America scarcely extending eastward of the great mountain-chain until the continent narrows in the extreme south. Not one of the group is to be met with in tropical Brazil, Guiana, or the West Indies ; not one, besides Solidago virga aurea, in Europe, Asia, Northern or Tropieal Africa, or in Australia. The 5 excepted genera comprise 58 South-African species. Of these, 2 continental genera, Pteronia with 51 and Fresenia with 3 species, are rayless ; and one might therefore at first suspect that they may not be correctly associated with a group so different geographically. But the involucre and general aspect are much nearer those of many homochromous than of any heterochromous genera (excepting in respect of the shrubby character so fre- quently assumed in the region by most Asteroid genera); and there is one monotypic continental genus, Homochroma, and two Mascarene ones, Glycideras of one and Rochonia of two species, in which the ray is, according to all accounts, homochromous and yellow. Moreover the geographical connexion between South Africa, extratropical South America, and the Mexican region is not so anomalous as would at first appear. I shall have to quote instances of it under the general head of extratropical southern connexions between America and the Old World. With regard to the distribution of the American genera within the general limits assigned to the group, Solidago, including its offset Brachycheta, with about 80 species, has, as I have already mentioned, only one immediate southern representative. The large shrubby Bigelowie, which might make a good genus of 4 species, are exclusively Andine. The remaining Bigelowie, 16 species, Haplopappus, about 60 species, Grindelia and Guttier - rezia, about 20 species each, range from Chili to California, but all with more diversified, as well as more numerous, forms in the north than in the south. Hysterionica, 5 species, may be consi- dered the southern representative of the northern Chrysopsis, 21 species; Nardophyllum and Lepidophyllum, each of 5 or 6 species, re- presenting the northern Ericameria of 4 species, are the only ones which show any southern preponderance. Xanthocephalum, with 7 Mexican species, has one in the Andes of Columbia. Hetero- theca, 5 species, Pentacheta, 2 species, and the monotypie genera 412 MR. G. BENTHAM ON COMPOSITE. Bradburia, Aphantochate, and Xanthisma, are exclusively North- American. Philippi’s two monotypic Chilian genera, Steriphe and Chiliophyllum, are unknown to me; and without having exa- mined them I cannot feel certain that they are really forms of the Solidago group distinet from any of the above. 6. The Conyza type.—This is as intimately blended with the Erigeron series as that is with the Asters, preserving generally the small flat or narrow achenes and slender uniseriate pappus of Erigeron, but with a great reduction in size and increase in number of the female florets, these being shorter than the style, filiform, and truncate, or two- or three-toothed, rarely producing a small scarcely spreading ligula. But all these characters are vague. Conyza itself passes, on the one hand, into Erigeron, on the other into Laggera and Blumea, and in a third direction comes very near to Baccharis. The geographical range of the Conyza type, however, is some- what different from that of Erigeron ; it is much more tropical and chiefly Old-World. Conyza itself, with 50 species, ranges over the warmer regions of Asia, Africa, and America ; and one or two species, as ready colonizers or weeds of cultivation, extend over the whole area. The allied more local species of this, which may be called the typical form of the genus, are some of them African, some American, without any particular local physiognomy ; but amongst the more divergent species in America, C. triplinervia and its allies assume a form approaching that of some species of the American Baccharides, and C. gnaphalodes (Lennecia, Cass.) has the pappus of an American Erigeron. Inthe Old World the principal divergent forms are Fimbrillaria and Dimorphanthes, each with several species, the former tending towards the Old-World genus Nidorella, the latter with an abnormal inyolucre unknown in this group in the New World. None of the genera closely diverging from Conyza are American. Haastia, with 3 species, is its New-Zealand representative. Thespis is a monotypic East-Indian. Nidorella, Heteromma, Chrysocoma, and Nolletia, comprising together about 28 species, are chiefly extratropical, but South-African, Volletia alone having a repre- sentative in North Africa; and Vidorella passes into some tropical African and Asiatic species which may almost equally well be placed in Conyza. There remain three or four genera connected in many respects with the Conyza group, but also giving indications of other affini- DISTRIBUTION OF TRIBES. 413 ties; these are:—1. Psiadia, an African and Mascarene genus of about 14 species, with something of the involucres and yellow ray- florets of some of the Solidago group; but the ligule are as nume- rous and almost as small as in some of the Conyzoid genera. The glutinous inflorescence of several species recalls some of the South-American homochromous genera, whilst the shrubby habit brings it nearer to an African type; and the constant sterility of the hermaphrodite florets is also chiefly to be found in some South-African and Australian genera of the tribe. 2. Microglossa, a tropical African and Asiatic genus, as to which it still remains a doubtful point whether it should be associated with the Erigeron or the Conyza type. 3. Adelostigma, one, or perhaps two, tropical species of a truly African character, but seemingly connect- ing the Conyza type of Asteroidez, of which it has the tailless anthers, numerous filiform female florets, small achenes, &c., with the true Zaulee, which it approaches in habit and involucre. 4. Parastrephia, a single Peruvian species which has been de- seribed as a Baccharis and as a Vernonia, and which Nuttall con- sidered to be anomalous in the whole order in having the female florets in the centre of the head surrounded by the hermaphrodite ones. But in this he was misled by insufficient specimens ; and the examination of more perfect ones shows it to be a connexion, as it were, between Conyza and Baccharis, technically belonging to the former group, but in habit, geographical station, and pro- bably in real relationship much more closely allied to Baccharis, where Meyen first placed it. 7 (and lastly). The Baccharis type —This consists of two genera, including above 250 species, which may be loosely defined as more or less dicecious Conyzas. Although the florets in the capitula of Asteroides show so frequently sexual differences in the same capitulum, it is only in these two Baccharoid genera that there is any unisexuality in the capitula or the individual plants. In this respect they correspond with several Inuloide* (Plu- cheinez, Gnaphaliez, &e.), but have not the anther-appendages of that tribe ; and their geographical positionis different. Baccharis and Heterothalamus are exclusively American and chiefly South- American, where they accommodate themselves to every soil and climate, ranging over the tropical plains, dispersed over the moun- tain-regions in great abundance, and extending to the extreme south, although not accompanying other Andine and Magellanic genera over to the Antarctic or the Australasian region. They 414 MR. G. BENTITAM ON COMPOSITE. extend also into North America, but in somewhat diminished numbers and varieties of form, and have not there diffused them- selves generally enough, or far enough northward, or early enough, to have spread into temperate Asia. The species are in general comparatively local; and none have shown any of the Conyza dis- position to become introduced into foreign lands. There is a small oriental plant, the Gymnarrhena of Desfon- taines, which technically, from its anthers and style, might be referred to the Baccharis group of Asteroidex ; but in habit and natural affinities, as well as in geographical station, it is so near to Geigeria, that we are compelled, as it were, to place it among Buphthalmez, although exceptional in what we reckon the most essential characters of Inuloidez. 4. Inuloidee. The tribe Inuloidex, not quite so numerous as Asteroidex, is more varied, the 1150 to 1200 species being easily distributed into 138 genera, or, according to some botanists, nearly double that number; and these again may be collected into 9 fairly distinct subtribes—Turchonanthee, Plucheinee, Filaginee, Gna- phaliee, Angianthee, Relhaniee, Athrixiee, Euinulee, and Buph- thalmee, all more or less geographical as well as structural. Asa whole, Inuloidez belong for the most part to the Old World, and several of the subtribes exclusively so; and the tribe is fairly limited (among heterogamous tribes) by the double character of style-branches without terminal appendages, and anthers with ap- pendiculate or so-called tailed auricles. The exceptional species are very few, and the frontier-lines not very difficult to trace, although in some measure the subtribe Plucheinex may be said to pass into Asteroidex of the Conyza group, Gnaphaliex into Sene- cionidez, Euinuleæ into Mutisiacese, and Buphthalme:e into Heli- anthoidee. Of the above subtribes, the second, third, and fourth, though most numerous in the Old World, range also over the New; the other six are limited to the Old World. The first and sixth are exclusively South-A friean, the fifth almost exclusively Australian, the seventh South-African with one or two more northern species, the eighth and ninth Afriean, European, and rather more sparingly Asiatic. We will take the subtribes, however, rather in their systematic than in their geographical sequence, commencing with the tribes Ld f L Y DISTRIBUTION OF TRIBES. 415 in which the female florets, when present, are filiform, not ligulate. 1. The TancHoNANTHEX are placed first, as having their diœ- cious character in common with the Baccharidew (the last sub- tribe of Asteroides), which they may in that respect be said to represent in South Africa; and the hermaphrodite florets being constantly sterile, their styles are the same in both cases; but there is no direct connexion between the two. The habit and the anthers are as absolutely different as the geographical situation, besides that, in Tarchonanthez, the form of the style-branches of the female florets is a greater departure in shortness and breadth from the almost uniform slender shape than I have met with else- where in the order. The subtribe consists of two genera and nine species, all strictly South-African, to which should probably be added the Mascarene Syachodendron, of which the sterile flowers alone are known. 2. PnLvcurrNEX. The genus Pluchea in an extended sense might include Blumea, Sachsia, Rhodogeron, and Tessaria; and, indeed, the whole subtribe show the filiform female florets, small achenes, and slender pappus of the Conyza group of Aste- roides, but with the Inuloid anther-tails, the styles without ter- minal appendages, and usually drier involueral bracts, and a few other features rather different from those of Conyza. The various genera proposed or adopted depend chiefly upon differences in the pappus and other individual characters, aud are more or less artificial and not always geographical. The nearest to Conyza is Blumea, strictly confined to the tropical and subtropical regions of the Old World and mainly Asiatic. It consists of about 60 species of varied habit, ranging from that of Conyza to Pluchea itself. Several species, widely diffused tropical Asiatic weeds, very difficult to distinguish from each other, only differ from Conyza in their anthers, whilst in the closely allied African and Asiatic genus Laggera, of about ten species, these anthers even lose their tails; but these species are in their styles and other respects quite removed from Conyza, and belong to a group inter- mediate, as it were, between Blumea aud Pluchea. Notwithstand- ing, therefore, any such exceptional forms which interfere with accurate technical characters, I believe there is never any difi- culty in distinguishing at once any Inuloid Laggera, Blumea, or ‘other Plucheinea from any Asteroid of the Conyza group. Pluchea, less varied in form than Blumea, and far less numerous LINN. JOURN.—BOTANY, VOL, XIII, 2H 416 MR. G. BENTHAM ON COMPOSIT E. in species, is yet far more extended in geographical range; and as it has established endemic species in different parts of its area, several of these have been raised into monotypic genera, sometimes with so much of the common habit and character that we have thought it illogical to adopt them; others are retained with some hesitation. The genus, as a whole (about twenty-five species), differs from Blumea in the corymbose inflorescence, in the disk-florets generally (though not always) sterile with undi- vided styles, and usually in the broader more coriaceous involu- eral bracts. It is nearly equally distributed over America, Africa, and Asia, chiefly within the tropics, but extending somewhat northwards both in America and Asia, and sparingly to the south of the tropic in Australia. No one species is common to the New and the Old World ; but the commonest species of the two hemi- spheres are about as near to each other as each is to the other species of its own region. The Old-World genera, chiefly d separated by charac- ters which we would consider as no more than specific, are:—1. Berthelotia, an East-Indian species which had been removed from the neighbourhood of Pluchea, chiefly on the mistaken sup- position that the anthers were not tailed. It has also been cha- racterized by the sete of the pappus being more or less connate at the base; but that is also observable, though in a less degree, in the common Pluchea indica. 2. Karelinia, a Caspian species with a narrower involucre and a long white pappus, the sets of the disk-florets mostly thickened at the apex, as is so often the case where the florets are sterile. This also had been erroneously presumed to have tailless anthers. 3. Oligocephalum may be given as a sectional name for P. pinnatifida, Hook. f., from tropical Africa, and P. frutescens, Benth., from Scinde, small shrubby species which in some respects form, as it were, a passage into Blumea, but are, on the whole, nearest allied to Pluchea; whilst the Laggeras above mentioned, intermediate in another direction, have the inflorescence and chief characters of Blumea, with the involucre rather of Pluchea—and having, in addition, no tails to the anthers, are retained as a distinct genus. 4. Spiropodium is an Australian species with the capitula generally, but not quite constantly, dicecious, a tendency to which arrangement is also ob- servable occasionally in the African P. Dioscoridis and some others. The Australian P. tetranthera, which is also nearly dicecions, diverges further in its tetramerous sterile florets and DISTRIBUTION OF TRIDES. 417 in the inflorescence tending towards that of Monarrhenes. 5. Eyrea, with three species, all Australian, is a form diverging in another direction, in its broader or hemispherical, often solitary, capitula and narrower involucral bracts, but apparently better placed as a section of Pluchea than as a distinct genus. In America three small genera, scarcely more divergent from Pluchea than some of the foregoing, but each with a very special geographical range, may still be kept up as distinct ; these are :— 1, Sachsia, three Cuban species more distinct from Pluchea in habit than in character; 2, Rhodogeron, a single species, also from Cuba, with the female fiorets almost ligulate, an exception to the whole subtribe; and, 3, Zessaria, five species limited to western temperate or Andine America from Chili to California, has not the tropical geographical character of Pluchea, but is elosely allied to the less tropical Asiatic forms of Pluchea both in habit and in eharaeter, whilst in indumentum and in the con- sistence of the involucral braets it shows some approach to the Gnaphaliee. Stenachentum, consisting of two or three Brazilian species, although hitherto included in Pluchea, is much more distinct than any of the foregoing, especially in its long achenes, exceptional in the subtribe. Two Old-World forms of limited geographical range diverge rather more prominently from Pluchea, to which they bear, nevertheless, much general resemblance—Pterigeron, five Aus- tralian species, and Vanothamnus, one East-Indian species, both with anomalous corollas, the former showing an approach to those of some Athrixiez, the latter to those of Mutisiacem. This Asiatic monotypic genus, with two equally monotypic Australian genera, Thespidium and Coleocoma, unite the Pluchean involucre and flowers with a much modified, reduced, or evanescent pappus; and the latter two Australian plants have, moreover, a very pecu- liar habit and inflorescence. Epaltes, a genus of about nine species, is spread over the same wide tropical and subtropical area as Pluchea itself. With the essential characters of the subtribe Plucheine:e, it bears a general resemblance, in habit and pappusless achenes, to the Vernonia- eeous Ethulia, with which genus it has often been confounded through a total neglect of the principal floral characters. The genus as a whole is a very natural one, although the species of each region have been raised into separate genera upon characters which scarcely deserve more than a specific rank. Thus Pachy- 2n2 418 MR. G. BENTHAM ON COMPOSIT F. thelia, Stectz, comprises two American species (of which one, JE. brasiliensis, is also African), Litogyne four African species, the original Epaltes a single Asiatic one, and Spheromorphea (S. petiolaris, DC.) and Ethuliopsis ( Gynophanes, Steetz) each a single Australian species. In the last-named species the capitula, as has been observed as to some Australian Pluchee, are almost or quite dicecious. Denekid is a curious little South-African genus of two, or perhaps three, closely allied species, with the anthers of Plu- cheinee, but in habit and some other characters approaching rather Nidorella amongst Conyzoid Asteroide, and forming one of the strongest links between the two tribes. The pappus is unlike that of any genus of either subtribe. Spheranthus, Pterocaulon, and Monarrhenus form a small tropical or subtropical group with the main characters of Plucheinex, but with the small glomerate capitula so prevalent in Filaginez, An- gianthez, and Relhaniez. Spheranthus (ten species) belongs exclusively to the Old World; Pterocaulon (eleven species) is also American ; both are more prevalent in Asia and tropical Austra- lasia than in Africa; Monarrhenus (three species) is exclusively Mascarene. The three genera, though closely allied, are fairly di- stinguishable. From the American Pterocaulon the Australasian species have been usually considered generically distinct, under the name of Monenteles, characterized by the solitary disk-florets ; but the two genera, established without reference to each other at about the same time, had never been fairly compared, and the sup- posed differential character is now no longer in accord with geogra- phical distribution; for the Brazilian P, spicatus has the solitary disk-florets and glabrous receptacle of Monenteles, whilst the Aus- tralian M. sphacelata has two or three disk-florets, as in the majority of American Pterocaulons. 3. FinaGINEX. The majority of the genera here included form a very natural group closely allied to Gnaphaliex as to the prin- cipal characters, but with the disk-florets most frequently sterile with undivided styles as in so many Plucheinex, and specially distinguished from both by their capitula usually small and glomerate almost as in Angianthez, and by the paleæ subtending or enclosing the female florets, or at least the outer ones. The seven genera we have adopted, comprising about forty species, range over the temperate regions of the northern hemisphere, scarcely penetrating within the tropics, several of them common DISTRIBUTION OF TRIBES. 419 weeds in the Old World; the American species chiefly western and extratropical both in North and South America. ‘The three principal genera, Evax, Micropus, and Filago, are chiefly Old- World, but represented also by a few species in Western America; the sections Acantholena and Bombycilena of Micropus belong, how- ever, exclusively to the Old World; and the section Stylocline of Micropus, with the nearly allied genera Psilocarphus, Diaperia, and Micropus, are limited to California or Chili, or both. The somewhat more divergent genus /floga (Trichogyne, DC.) has a very different geographical area, being chiefly South-African, with one representative in the Mediterranean region. Three genera with many of the technical characters of Filagineze, but with a very different habit and geographical range, are not, perhaps, properly included in the subtribe, or at the most should only be considered its tropical considerably modified represen- tative. These are Cylindrocline, with one Mauritian species, and Blepharispermum and Athroisma, which together include four from tropical Asia or Africa. Their larger black achenes and reduced paleaceous or deficient pappus seem to bring them nearer to the Buphthalmex ; but the receptacular pales covering the filiform female florets are characteristic of Filaginee. They thus form an isolated group which would be still more out of place in any other tribe or subtribe with which they might be compared. Petalacte, a single South-African species, is another ambiguous genus which, technically, would be placed among Filaginee, but, from its habit, geographical, and some other characters, may be better considered as an exceptional Helichrysea next to Anaveton, which it so closely resembles. 4, GNAPHALIEX, a very large subtribe, distinguished from Plucheineee chiefly by the scarious, or coloured, and radiating involucres, may be divided into two groups or large natural genera, Eugnaphaliee and Helichrysee, the one of about 200, the other of about 400 species, closely allied to and, as it were, passing into each other, insomuch that some botanists have united into one the two typical genera Gaaphalium and Helichrysum, but distinguished with few exceptions (chiefly tropical African) by a general character of some importance, the female florets outnum- bering the others and usually exceedingly numerous in the Eugnaphaliee, few, or disappearing altogether,in the Helichryses; and the geographical distribution is different. ^ Eugnaphalieze range over the whole world, forming only small distinct groups in 490 MR. G. BENTHAM ON COMPOSITAE. local areas; Helichrysez are, with few exceptions, limited to South Africa, Australia, and the Mediterranean region of the Old Wortd: The delimitation and subdivision of the genus Gnaphalium has much puzzled all synantherologists, and is still in a very unsatis- factory state. Among the species now generally admitted to belong to it are some which by Lessing and De Candolle were established as a genus of Mutisiaee: ; and the differences observable, whether in habit or structure, are not much in accord with the geographical distribution. The genus as a whole is cosmopolitan, though rather more temperate or mountainous than tropical ; and the two prin- cipal groups into which it might be divided as to habit inflorescence and involueres, represented, for example, by G. polycephalum and G. uliginosum respectively, are both to be found over the same area; whilst G. luteo-album, which closely connects the two groups, is ubiquitous. Weddell proposed the division into two genera founded on the pappus, of which the sete are quite free and sepa- rately caducous in Gnaphalium, united at the base in a ring and falling off altogether in Gamocheta: and this appears at first sight very plausible, and accompanied by some difference in the involucre; but upon further investigation it is found to separate species like G. purpureum and G. indicum, so closely allied that they are found mixed together in most collections, whilst it unites into one genus G. polycephalum, G. lavandulaceum, and G. uligi- nosum, species evidently the most remote in affinity of the whole series. It would seem, therefore, that Weddell’s divisions can only be taken as somewhat artificial sections, and that if we attempt any more natural although vaguely characterized groups we must recognize three as very generally diffused, those above- mentioned as typified by G. polycephalum and G. uliginosum with theintermediate G. luteo-album under Eugnaphalium, and that exem- plified by G. purpureum and G. sylvaticum under Gamocheta, and about five other groups confined to special geographical areas :— 1, two or three Andine species (Mexican or South-American) belonging to Eugnaphalium, but remarkable for their involucral bracts radiating as in Chionolena, which they also resemble in habit; 2, Lucilia (including Bellos), a Chilian or south Andine group, differing from other Gamochete chiefly in their longer nar- row involucres; 3, Merope, dwarf tufted or prostrate Andine plants, with the involucral bracts more spreading after flowering than in Lucilia, to which A. Gray unites them ; 4, Omatotheca, a Europxo- Asiatic and North-American Alpine plant, dwarf like the Meropes, DISTRIBUTION OF TRIBES. 421 with the involucre of a Gamocheta, but the pappus of a Eugna- phalium; and, 5, Anaphalioides, a remarkable New-Zealand group of two or three species, with the spreading involucres of Anaphalis, and in sexual arrangements connecting that genus with Gna- phaliwn. With regard to the groups so much further divergent from Gnaphalium as to have been retained as distinct genera, one, Achyrocline (fifteen species), is common to the New and the Old World; like many tropical-American (chiefly Brazilian) genera it has two or three representatives in tropical Africa. It is nearest, on the one hand, to the corymbose group of the section Eugna- phalium, but with narrow few-flowered capitula more densely corymbose, a Helichrysum-like involucre, &c.; and, on the other hand, to Stenocline-among Helichrysez, of which it has the aspect, but with the Gnaphalioid sexual proportions. Nearly allied to it is the monotypic tropical-African Chiliocephalum without any pappus. Chevreulia and Facelis are two small South-American Andine or extratropical genera connected with the Lucilia series of Gna- phaliwm of the same region, but differing, the one in the long beak into which the achene is produced, the other in the plumose pappus. Lasiopogon, from extratropical Africa south and north, and from the latter extending into the Levant, is another small genus, differing from some of the smaller Zugnaphaliee and Filagines of the same region in the plumose pappus. Phagnalon has a dozen species from the Mediterranean region more isolated in character. The anthers are often almost or quite tailless ; and the species were indeed formerly included in Conyza; but their involucres, habit, styles, and the anther-tails of some of the species justify their having been removed to Gnaphaliez. All the above genera have the disk-florets usually fertile as well as the females; in the following eight or nine genera, so nearly connected with Gaaphalium that they have most of them been united with it by some authors, the disk-florets are almost univer- sally sterile and often with undivided styles. Chionolena is Brazilian, allied in other respects to the American Gnaphalia with radiate involueres. —Luciliopsis is Andine, near the Andine Gna- phalia of the section Lucilia. Tafalla is also Andine, but more tropical, and in its habit and dicecious capitula connected with some species of Baccharis of the same country, but with the Gnaphalioid inyolucre, anthers, £c. Mniodes, again, another high- 422 MR. G. BENTHAM ON COMPOSITA,. Andine small genus with dicecious capitula, corresponds in other respects with Merope, also high-Andine. Antennaria and Leow topodon are mountain genera connected with the smaller Eu- gnaphalia, dispersed over Europe, Asia, North Ameriea, and the Andes of South America. Anaphalis, allied on the one hand to Antennaria, on the other to the corymbose Eugnaphalia, is inter- mediate between the two as to sexual characters, and differs from both in the usually radiating involucres. It is more Asiatic than any of the other genera, being represented in North America by only a single species out of about twenty-five. Stuartina, Demi- dium, and Amphidoxa, three monotypic genera from Australia, Madagascar, and South Africa respectively, are allied in habit to the small glomerate Eugnaphalia; but, besides the. frequent sterility of the disk-florets, they differ in the pappus reduced to a very few sete in Amphidoxa, entirely deficient in the other two. There remains only Raoulia, an Australasian mountain genus of fourteen species, chiefly from New Zealand, which may be said to be almost strictly intermediate between Eugnaphaliex and Heli- chryses, the proportion of the female and disk-florets being variable and often nearly equal, but certainly with a Eugnaphalioid tendency both in that and in habit. Helichrysez present one of those instances (such as Proteacee, Restiaces, &c.) in which a large very natural group of plants had spread over two regions, South Africa and Australia, now quite isolated, but then possibly in connexion with each other, in times sufficiently remote for them to have diverged in each region into different forms, and have multiplied greatly in both, without having preserved a single species in common. The Helichrysex, how- ever, have retained a closer affinity than the larger groups above mentioned. The South-African Ericacez have only a representa- tive order or suborder in Australia (Epacridee); Proteacew and Restiacee have tribes but no genera in common; among Helichryses there are common genera and even sections, but no species. Helichrysum itself, the largest genus of the subtribe, has, out of about 260 species, 137 South-African and about 60 Australasian (chiefly Australian with a few from New Zealand), and in each country has established distinct sectional races. The subgenus Lepicline is wholly South-African, as are also several sections of Euhelichrysum ; others of these are exclusively Australian ; but the sections Xerochlena and Ozothamnus, although chiefly Australian, DISTRIBUTION OF TRIBES. 423 have afew African representatives ; for many groups of the former ‘ave been found in South Africa, and a few species referable to the latterin Madagascar. The genus extends also into North Africa, Europe, and Central Asia—that is, into the Mediterranean region taken in avery extended sense. All these northern species belong to the section Stechas, also represented in South Africa, with two monotypic exceptions: one is Cladocheta, which, however, might well be included in Stechas, although De Candolle generically re- moved it on account of the sete of the pappus being more or less united in bundles; but the little value of this character is shown by its occurrence likewise in other species, as, for instance, . in the Australian monotypic section Acanthocladium, which has the habit of the tropical-African H. spinosum and H. horridum, neither of which has the same pappus. The other exception in- cludes H. frigidum, Willd., and H. virgineum, DC., dwarf alpine species, only known, the one from the mountains of Corsica, and perhaps of Lebanon, the other from Mount Athos in Greece, both very unlike any other species growing north of the equator. The radiating involueres are those of the southern Xerochlene; and the whole plant has much outward resemblance to the New-Zealand H. (Gnaphalium, Hook. f.) prostratum and bellidioides, but with the densely silky-villous achenes of several Heliptera. Helipterum cannot well be called a good natural genus, but rather a collection of local South-African or Australian subgenera or sections, retained as a distinct group rather for convenience’ sake, and solely founded on the artificial character of a plumose pappus; andeven tbat fails to draw a distinct line separating it from some species of Helichrysum. Ytis much more Australian than African; for of 42 species, 30 belong to the former region, and 12 to South Africa;it is in Australia also that it blends most with Helichrysum ; in South Africa the two genera are more distinct. We have, in the ‘Genera Plantarum,’ characterized four sections, of which one small one, Syncarpha, with two species, is South African, two, Pteropogon ten species and Monencyanthes seven species, are Australian; the principal one, Euhelipterum, is common to both regions, although even here there is a slight difference; the majority of the Australian Euheliptera have radiating involueres, which are exhibited only by very few of the South- African species. Stenocline is a small genus of a more tropical character than any others of the Helichrysez, and the only one common to the New and the Old World. 1t has eight species, of which six are Mascarene 424 MR. G. BENTHAM ON COMPOSIT E. and two Brazilian. It is very nearly allied to Helichrysum itself but still nearer, in many respects, to the tropical African ash American Achyrocline, although the sexual relations place the two in different divisions of Gnaphalium. Gardner’s genus Leucopholis has two Brazilian species very closely allied to Stenocline, with the habit, however, of the American Chionolena, and differing from both chiefly in inflorescence, the few-flowered heads being closely sessile, and collected in a globular cluster or compound head, as in the Angianthez. The remaining genera of Helichrysex require but little notice to their geographical distribution ; they are all limited to South Africa or to Australia, without any of the forms sufficiently similar in their divergences in both countries to be considered representatives ; for the diverging characters are differently com- bined in the two regions. They are chiefly characterized by the involucres and pappus. South Africa has six of these genera, comprising sixteen species ; South Australia sixteen genera, com- prising forty-eight species. Among the latter I may particularly mention Millotia and Quinetia, both monotypic, as connecting in some measure the Helichrysex with the Senecionidez, having the peculiar almost uniseriate involucre so rare in Inuloidee, gener- ally so frequent in Senecionidem. Among the South-African genera the monotypic Phenocoma is remarkable for the foliage, which is that characteristic of the Relhaniee; but the filiform female florets and the broadly radiating involucres are rather those of Helichryse:e. 5. The ANGIANTHEE proper constitute a group of eight genera and about sixty-four species, exclusively Australian, and, with the exception of two New-Zealand species of Craspedia, limited to Australia itself. With a Gnaphalioid habit and connexion, they are further removed from Gnaphalium itself than the Helichrysex, the capitula being always homogamous without any female florets whatever; the small closely aggregate capitula are also to be met with in some Filaginez ; but in the latter tribe there are always female florets embraced by the receptacular or involueral palez. And, geographically, the Filaginez are entirely wanting in Aus- tralia, the fatherland of the Angianthee. ‘esulia, a monotypic East-Indian genus, stands alone. Its essential characters are indeed those of Angianthes; but its habit, its country, and several points of structure show but a re- mote affinity with any Australian genus of that subtribe. DISTRIBUTION OF TRIBES. 425 Eriosphera, a monotypic South-African genus which I only know from Harvey’s figure and description, might also be techni- cally referred to Angianthes, where, indeed, it may be regarded - as a South-African representative of the Australian Gnaphalodes. It also bears much resemblance in outward aspect to the South-African Lasiopogon, but has no female florets and a very different pappus. 6. RELHANIEEX. The Inuloidew have now taken leave of America; the remaining subtribes with the capitula radiate (when heterogamous) are all African, European, or Asiatie, very sparingly and anomalously represented in Australia. "We have seen that the Inuloid genera with the greatest profusion of fili- form female florets with exserted styles, Plucheineæ and Eugna- phaliez, supply not only the most widely diffused genera and species, but also those which have most readily established local sub- genera or diverging genera in both the New and the Old World. The Helichrysez with few female florets have scarcely spread into or maintained themselves in America, whilst the Angianthex, without any, have remained within their own limited areas. Nor do the tribes with ligulate female florets appear to have been more successful than those which are strictly homogamous. Possibly the greater facilities enjoyed by the disciform over the radiate races for fertilization and for dispersion, resulting from the peculiar structure and mutual arrangement of the male and female florets and fruits, may have had some effect on their extended distribu- tion; but it is difficult to appreciate the effects of each one of the numerous more or less counteracting influences which have at various times acted on the dispersion, establishment, restriction, or extinction of genera and species in different regions. Strictly dicecious Inuloideæ and other Composite certainly appear to have been less successful in spreading than those with androgynous capitula, where structure &e. is otherwise similar. The African Tarchonanthee cannot well be considered close representatives of the American Baccharidez ; the smaller dicecious genera have not a wide range, with the exception of Antennaria, where other influences, resulting from alpine station, may have come into play. Whatever, therefore, may have been the cause, these three Inuloid subtribes, in which the female florets, when present, are always ligulate, are absent from America, and are mostly, although not entirely, extratropical, and, generally speaking, of very 426 MR. G, BENTHAM ON COMPOSITE. limited areas. The Relhaniez (fourteen genera and near one hun- dred species)are exclusively South African, or sparingly represented in the Mascarene Islands. They are in many respects so closely connected with the Helichrysee of the same region as to make it sometimes difficult to determine to which subtribe a genus should belong. Thus JMefalasia and Lachnospermum, having no female florets in any of the species known, might equally well belong to either subtribe, but that they have the peculiar foliage of Rel- haniez, concave or tomentose on the upper instead of the under side, unknown in any other subtribe of Inuloidex, excepting perhaps Phenocoma, another genus, like Metalasia, rendered am- biguous by the absence of female florets, but which, notwith- standing its Relhanieous foliage, seems, in involucre and other characters, to have more affinity with Helichrysee. Some Relhaniew have the one- or few-flowered aggregate capitula of the Australian Angianthes, but accompanied by a strictly South-African, not Australian, combination of habit, foliage, and other characters, showing the affinity between the two subtribes to be distant. The individual genera of Relhaniex, distinguished chiefly by the aggregate or separate flower-heads, or by the various pappus- forms, afford nothing special to remark upon, as far as hitherto observed, in respect of geographical distribution, all being confined to the same limited area. 7. ATHRIXIEX. This subtribe, although still chiefly South- African, is not so local in geographical distribution, and more varied in structure than the Relhanieew. The Athrixiee are at once distinguished from the Relhaniee by the foliage, from Euinulex by the style, and generally from both in habit. The genera, however, require separate consideration; for they are not so blended into each other as most of those of the preceding subtribes. Athrixia itself, with fourteen species, is represented in South Africa, Madagascar, Abyssinia, and Australia, although in no case by identieal species in any two of these regions. Four of the five Australian species form a local section distinct from the African ones, and which has indeed been raised into two genera, but properly reduced to Athrizia by Asa Gray. The fifth Australian species, however (4. aculeata, Steetz, or Asteridia, Lindl.), is nearer in structure and habit to the typical 4. capensis than to any of its own fellow-citizens ; and the single Madagascar species DISTRIBUTION OF TRIBES. 427 seems in some measure to connect the two Australian types. Antithrixia and Arrowsmithia may be considered as somewhat divergent forms of or offsets from Athrixia, both of them South- African—Antithrixia represented also by one species in Abyssinia, Arrowsmithia monotypic and local. Leyssera, with three South-African species, is the only genus of Athrixiee which has a representative (not, however, specifically identical) in North Africa. Although of a perfectly distinct structural type from that of Athriwia, it is not, nevertheless, at all more intimately connected with any other Inuloid genera, whether Euinulee or Buphthalmex, which have a similar geo- graphical distribution. Macowania, of a single species, and Heterolepis, with three spe- cies, are both limited to South Africa, and form very distinct genera, although generally connected with Athrixiez. The last- mentioned, Heterolepis, has hitherto been referred to Arctotide:e, of which it has neither the habit northe achenes, pappus, or style, perhaps from an undue appreciation of the value of the scarious involucral bracts, which, however, is more or less observable in Leyssera and other truly Athrixious genera. Podolepis is a very distinet Australian genus of a dozen species, remarkable for the irregular and varied development of the corolla of the female florets. It establishes in this respect, as also in the habit and involueres of some of its species, a connexion between some Australian forms of Athrixia and Helichrysum respectively. We have here a relationship, established by structural peculiari- ties and confirmed by origin as presumed from geographical dis- tribution, between species such as Athrixia australis and Podolepis rutidochlamys, which might readily, on a hasty inspection, be re- ferred the one to Asteroidee, the other to Helichrysum. 8. EvrNULEZ, 19 genera and about 120 species, are so nearly connected with each other, that, with the exception of two or three rather more distinct monotypic forms, they might be considered as constituting a single large genus. Nearly half the species are still retained in the genus Inula; and the genus Pulicaria, for in- stance, including one half of the remaining species, although constant in its character derived from the pappus, is probably really less distant from some Jnule than are the two sections Bubonium and Cappa from each other, although these are re- garded by all botanists as congeners. Taken, therefore, as a whole, Euinulee differ from Athrixiee 428 MR. G, BENTHAM ON COMPOSITE. t in the style, and generally in habit, involucres, and other minor characters, and in a more northern geographical distri- bution. They have some outward resemblance to some American Asteroide of the Solidago group (Haplopappus inuloides, Chry- sopsis, &c.), but with very different styles and anthers and a wide geographical severance, and are rather more nearly con- nected in structure and station with a few genera of Mutisi- aces, amongst which one genus (Printzia) has been recently placed; but their real affinity, structural and geographical, and therefore presumedly genetic, is with Athrixiee onthe one hand and Buphthalmee on the other. They belong exclusively to Europe, Asia, and Africa; their chief centre appears to be the great Mediterranean and Oriental region; but they extend south- ward into South Africa, where they have established a few small local genera, and eastward to the tropical and subtropical extreme east of Asia, although not enough to the north-east to have passed into North America. In Asia they may be said to be par- tially replaced by allied Mutisiaceous genera; and the South- African genus Printzia, five species, has been, as above-mentioned, hitherto actually referred to Mutisiace:, although without the essential characters of that tribe, and to our eyes having a close affinity with the Inuloid genus Jphiona. Neither in Australia nor in America is there any genus of Euinulex, nor yet of any nearly allied Mutisiacew. The North-west American Luina, with a deceptive aspect of some species of Inula, proves, when examined, to be as different in structure as the Znula-like Asteroidew above mentioned, but in this instance to belong to Senecionide:ze. Of the separate genera, the two principal ones above mentioned, Inula and Pulicaria, range generally over the greater part of the area of the subtribe, and are, besides the two species of the Me- diterranean Jasonia, which may almost be regarded as a section of Inula, the only ones which extend to Europe. The next nume- rous genus, /phiona (12 species), has still a wide range from the Levant to east tropical and South Africa. Codonocephalum and Amblyocarpum, two nearly allied monotypic genera, are limited to the Levant; Grantia, with four species, is in that region and in Algeria. Allagopappus and Vierea, both monotypic, belong to the Canary Islands, where are also some rather peculiar species of Inula. Vicoa, with five species, is more tropical both in Asia and Africa. Calostephane and Porphyrostemma, both monotypic, are also tropical, but African only. Pegolettia has one tropical and DISTRIBUTION OF TRIBES. 429 two South-African species; Bojeria one Mascarene and one South-African; and, lastly, Printzia, with five species, Homo- chete, with one species, and the two monotypic genera unknown to me, Alinurothamnus amd Cypselodontia, are South-African. Among all the above forms Porphyrostemma is the most excep- tional, having the purple, narrow, linear, very numerous ligulate corollas of Erigeron, with all the essential characters of Euinule:e. All the other genera are homochromous. There remains the somewhat anomalous genus Carpesium, with four or five species, which in its tubular female florets connects Euinulee with Plucheinee, but, upon the whole, is best placed in the former subtribe. Its geographical area is within the chief range of Euinulez, South Europe and temperate and tropical Asia. 9. Bupnrustmes. The subtribe Buphthalmes, sixteen genera, but scarcely above fifty species, allied to Euinulee, has a nearly similar geographical distribution, somewhat more restricted east- ward, and offers some exceptions. Buphthalmex are chiefly African, European, and Oriental. Their structural connexions are more general than those of the other subtribes of Inuloidea, their styles the same as in Euinulee in the North-African and European genera, like those of the Athrixiez in some exclusively South-African forms ; and their tailed anthers and alternate leaves leave no doubt as to their place in this tribe; but in their rigidly paleaceous receptacle, the nature of the pappus in several genera, and some other respects they point to some connexion with Helianthoidex. The supposed affinity to Asteroidew appears more remote, and can only have been suggested by the numerous narrow yellow ligule of some genera. Amongst themselves, the genera, although small, are more distinct than most of those of the preceding subtribes. Of the seven most nearly connected with each other, three were long united under Buphthalmum; but two of the others have been hitherto placed in very different tribes, owing to inattention to the anther-tails, and also to the supposition that there were no Buphthalmes in South Africa, As it now stands, Buphthalmum is reduced to four species exclu- sively European: Odontospermum seven species, and Pallonis one species, formerly included in Buphthalmum, belong to the Medi- terranean region generally, the former extending to the Canarian and Cape-Verd Islands; but Callilepis, two South-African spe- cies, Sphacophyllum, one Mascarene species, and Anisopappus, two 430 MR. G. BENTHAM ON COMPOSITE. tropical species, appear to me to be quite as near, if not nearer, to Buphthalmum than Odontospermum or Pallenis, although Calli- lepis, for the reasons above stated, had been placed by De Candolle in his subtribe Heleniez of Senecionidex, and Sphacophyllum, in- cluded by him in the Helianthoid genus Epallage, referred by him to Anthemidez. The two species of Anisopappus are both tropical African—one from western Africa, published by J. D. H oker as a Buphthalmum ; the other, only recently found by Colonel Grant in east tropical Africa, and undoubtedly indigenous, proves to be identical with a long-known South-Chinese plant published under the present generic name by Hooker and Arnott. It has never been found in any intervening district, and the most careful exa- mination can detect no difference in the specimens from those widely distant regions—a case analogous to, but more remarkable than, that of the Zupatorium which connects east tropical Africa with north-eastern India. The monotypic Oriental Chrysophthalmum is also very near Buphthalmum ; but the female florets are deficient, and the habit is more that of Amblyocarpum among Euinulee. Rhanterium, an Algerian genus of two species, and Anvillea, also of two species, one Algerian the other Oriental, are remark- able for their involucres becoming subglobose and often spines- cent, like those of so many Cynaroidee ; but their other cbarac- ters, and even the habit, on a closer investigation point out their close connexion with Buphthalmez. Ondetia, one species, from southern subtropical Africa, is another form of Buphthalmese, di- vergent in its involucre; but whilst the two previously named Mediterranean genera assume in that respect the Mediterranean Cynaroid type, the southern Ondetia takes the scariose involucre of the southern Arctotidese. Geigeria, with eight South-African and two Arabico-Nubian species, is an anomalous genus, rather puzzling as to its affinities; the styles, the anthers, the colour of the flowers, and, to a certain degree, the habit are those of Buphthalmee, whilst the densely setose receptacle points to Cynaroidez, and the deeply lobed co- rolas to that tribe or to Mutisiacee ; but upon the whole its nearest affinity seems to remain with Buphthalmesm. Geogra- phical distribution does not here afford much assistance; but at least it is as much in favour of the Buphthalmoid affinity as of any other. Gymnarrhena, a single Oriental species, is still more puzzling as DISTRIBUTION OF TRIBES. 431 toits affinities. It has the peculiar habit of the dwarf species of Geigeria and similarly deeply lobed yellow corollas ; the receptacle, paleaceous under the female florets, but not under the sterlie florets of the centre, is that of Rhanterium ; but the female florets are in many rows, with short tubular corollas, as in the Conyzoid and many Grangeoid Asteroidew, and as in the Plucheines amongst Inuloides. The tailless anthers would exclude it from all Inuloid subtribes and place it technically, as proposed by De Candolle, among Asteroidez, where we might insert it after Heterothalamus in the Baccharis group. But it is in all other cha- racters, as well as in a geographical point of view, so perfect a stranger there that it seems preferable to class it next to Geigeria, as an exceptional form, such as is Barnadesia amongst Mutisiacez. The style does not help us; for, the disk-florets being sterile, it is undivided, as is the case so generaily in all tribes when similarly circumstanced. Two South-African genera, Osmites, with six species, and Os- mitopsis, with a single one, take their place among Buphthalmez in respect of almost all their general characters, as well as in habit (of Odontospermum) and the peculiar odour of the foliage; but the style is rather that of the South-African Athrixiew than of the more northern typical Buphthalme:e. There remains a small anomalous plant from a very different region, which, after being attached to various tribes, must perhaps find its resting-place next to Buphthalmesz. This is the Vablo- nium of Tasmania. Cassini placed it amongst Anthemidew. In working up the Australian flora, I had trusted perhaps rather too much to Bauer’s and Fitch’s elaborate drawings and analysis, and referred it, after some hesitation, to Helianthoidesm. A more careful examination shows that we had all overlooked the long setiform appendages or tails to the anther-auricles. This places it technically among Buphthalmex, with which also, notwith- standing its reduced size, the foliage and indumentum agree better than with Helanthoides. Anthemidee are quite out of the question. 5. Helianthoidee. The tribe Helianthoidez is, again, one of the large ones. Not quite so numerous as Inuloides, it is still more varied. The species, rather under 1100, are distributed into about 140 genera not so easily classed into distinct subtribes as Inuloidex, rather more scattered geographically, and many of the smaller genera LINN. JOURN.—-BOTANY, VOL. XIII. A1 432 MR. G. BENTHAM ON COMPOSITS. remarkably distinct. As a whole the tribe is essentially American, and chiefly tropical or subtropical; but some genera have been early enough in the warmer regions of the Old World to have there established distinet species or sections, a very few small or monotypie ones from Afriea or Madagascar or East India suffi- ciently differentiated to be maintained as genera (Micractis, Epallage, Guizotia, Glossocardia, Microlecane, Glossogyne). The development and usual persistence of the pale: of the receptacle or bracts subtending the individual florets, the tailless anthers, and the pappus, when present, consisting of few rigid awns or pales, some of them more directly corresponding to the primary ribs of the achenes, are its principal characters. The style is very variable. The order is slightly connected through Lagascea with Vernoniacez, through Ambrosiee with Anthemides, through Madiew with Helenies, and through a few Verbesinee with Inu- loidew (Buphthalmex); but the delimitation is rarely doubtful. Some of the structural characters, as well as the dispersion of several genera over the warmer regions of both the New and the Old World, seem to point this out as containing some of the most ancient forms of the order. Among the numerous subdivisions which have been proposed, we have thought that ten might be maintained as subtribes, although very unequal in point of numbers and geographical range. As in the Inuloidex, we will take them in detail in their systematic sequence. 1. Lagascea, a small genus of about seven species, is so distinct from all others as to require separation as a subtribe, and even the tribe it should be classed under is uncertain. Its style is that of Vernoniaces, where it has been technically associated with Zle- phantopus amongst genera with glomerate uniflorous capitula ; but the habit, the mostly opposite leaves, theindumentum, and especially the corolla and pappus are so different from any thing observable in that tribe, that it has appeared to me to be better placed as a somewhat anomalous Helianthoid. Its geographical area is that of a large proportion of the American anomalous oligotypic genera, the Mexican region, to which all its species are limited except one, which, apparently as a weed of cultivation, has spread over many parts of South America, and has been also carried into the tropical regions of the Old World. 2. Under the name of MILLERIE E are collected a number of small or monotypic genera, some of which may not really prove to have DISTRIBUTION OF TRIBES. 433 claims to consanguinity sufficient to be thus associated, and may possibly be, like Lagascea, isolated remnants of old, almost extin- guished races or local divergent forms, whose connexions have not been properly appreciated. In the mean time they are somewhat technieally associated by their mostly Helianthoid habit, foliage, and involucres, by their usually small capitula with a few female fertile florets ; the disk-florets always sterile with undivided styles ; the achenes of the fertile florets Helianthoid, rather large, often dorsally flattened or thick, subtended by or enclosed in the inner involucral bracts (or outer receptacular palez), without any or with a short coroniform or Helianthoid pappus; the abortive disk-achenes without any pappus, their florets crowded in the centre of the capitulum without intervening pales, or, at most, with a few reduced subtending sete. Geographically they are somewhat scattered ; the majority are American and tropical Heptanthus, three species, Pinillosia, three species, and the monotypic Lanta- nopsis, are restricted to Cuba; Tetranthus, two species, is also insular, restricted to San Domingo ; Elvira has three species, which some may regard as so many distinct monotypic genera, of which two are insular, limited to the Galapagos, the third dispersed from Central over many parts of tropical South America, as is also the monotypic Milleria ; Stachycephalum, also monotypic, is limited to Mexico,—the above seven (or nine) genera having, perhaps, sufficient characters in common to be united into one genus of a higher grade; all have opposite or radical leaves and few-flowered small capitula. Technically allied to them, with small few-flowered capitula and radical or alternate leaves, and, perhaps, really of very distant affinities, is a genus of a much wider geographical distribution. Adenocaulon, with five species, belongs to the tem- perate or mountain W.-American region, two species being Chilian, one N.W.-American, and two, possibly varieties of the N.- Ame- rican one, inhabitants respectively of Japan and the Himalaya. This genus had hitherto been classed with Tussilago, where, to my mind, it would be a more perfect stranger than amongst Milleriez. The three remaining genera, Riencourtia, five or six species, from eastern tropical South America, Desmanthodium, two Mexican species, and Clibadiym, with fourteen species more generally dis- persed over tropical and subtropical America, establish the connect- ing link between Milleriew and Melampodinezx and the still more normal Helianthoid subtribe Verbesinez. 212 434 MR. G. BENTHAM ON COMPOSIT f. 3. MELAMPODINES, nineteen or twenty genera and nearly 100 species, constitute a more definite subtribe, intermediate between Milleriee and the great mass of Helianthoidee (Verbesineg), differing from the former in their completely paleaceous receptacle, from the latter in the constant sterility and undivided styles of the disk-florets. They are, with two exceptions, exclusively American and chiefly tropical, but occasionally extending further both north- ward and, in a less degree, southward. Most of them require here but a very short mention. Ichthyothere, eight species, Baltimora, two species, and Acanthospermum, two species, are strictly tropical and chiefly eastern ; Melampodium, eighteen species, is also tropical, but with a much wider range, extending northwards over the Mexicano-Texan region, with one species (probably introduced) found also in various parts of the warmer regions of the Old World, and another, to all appearance really indigenous, in the Philippine Islands. —Espeletia, eleven species, Philoglossa, one or two species, and Schizoptera, one species, are also South-American and tropical, but limited to the Andes; Parthenium, six species, and Polymnia, twelve species, are tropical, but also extratropical and chiefly, but not entirely, western, the former extending from Chili to the Mexican region, the latter from the Argentine Republic to Canada. Eight genera are limited to the Mexican region, — Berlandiera, with five species, Guardiola, with four, Tri- gonospermum, with two species, and the monotypic Lindheimera, Engelmannia, Dicranocarpus, Aiolotheca, and Parthenice. Another monotypic genus, Lecocarpus, is limited to the Galapagos Islands; Silphium, with eleven species, is exclusively North-American. This geographical arrangement is, however, not in conformity to struc- tural affinities ; the natural divisions of the subtribe have generally a wide range. The northern Silphium, the Andine Schizoptera, and theintermediate Berlandiera, Lindheimera, and Engelmannia might form one genus, which would then have as extended an area as Polymnia. The east tropical Acanthospermum and the Galapagian Lecocarpus might be included in the widely spread Melampodium. There remains the genus Chrysogonum, which, although it is undoubtedly nearly related to Silphium, and has even one species belonging to the same area in North America, is yet more strongly represented in a widely distant region; two species are East-Indian and three are tropical-Australian. "These Old-World species have been hitherto known under the name of Moonia; but, on attempting to draw up comparative characters in the subtribe, DISTRIBUTION OF TRIBES. 435 I could find nothing in habit or structure to distinguish the American from the East-Indian genera. The species are distinct enough; but one of those of the Old World is more nearly aliied to the American one than to its co-citizens. 4. AMBROSIEX, ten genera and about forty species, form so distinct and natural a subtribe that it has been repeatedly pro- posed to raise them to the rank of an independent tribe, suborder, or even of a distinct order; and regarding the characters of Xan- thium and Ambrosia alone, as they are usually expressed with a slight exaggeration, the separation would seem justified ; I have, I believe, myself somewhere assented to it; but after a detailed examination of all the surrounding genera, I have felt compelled to admit that the majority of synantherologists are correct in placing them under the Helianthoidee. They are, without doubt, connected with Artemisia as well as with Melampodinee, having much of the habit of the former and passing into the latter through Parthenice; but geographically, as well as structu- rally, the relationship to Melampodinex appears to me to be the closest. The Ambrosiez are strictly American, although three or four species, as in the case of Elephantopus, Eclipta, &c., may be widely dispersed also over the Old World, whereas the Artemisie belong to an Old-World series, and are themselves of the Old World, although they may have some identieal or representative species in the extratropical or mountain-regions of America. Del- pino, however, in his above-mentioned memoir, as well as in his private letters, insists on the close connexion of Artemisia with Ambrosiezx, forming his subfamily of Artemisiacee, which he thinks he has established on irrefutable grounds. But to me it appears this is only a very natural attaching of undue importance to a class of characters the study of which he has specially carried out with so much success. He proposes two distinct * families," Senecionide and Helianthace:, the one with the style truncate at the end with a terminal tuft or marginal ring of hairs destined to scrape or push the pollen out of the anther-tube, the other with the hairs descending along the outside of the style so as to sweep out the pollen. This is an old distinction which experience has prevented from being made generally available. If Delpino had not confined himself to the examination of so small a proportion of the varied style-forms, if he had gone through any considerable number of our Senecionidez and Helianthoides, he would soon have been stopped in his endeavours to class them according to his 496 MR. G. BENTHAM ON COMPOSITE. views. To gonofurther than the few figured by Hildebrand in his above-mentioned memoir in the ‘ Nov. Act. Nat. Cur.,’ what would he have done with the series Doronicum, t. 2. f. 23-26 (an undoubted truncate penicillate Senecionida), Bidens, t. 1. f. 30, 31, Emilia, t. 1. f£. 11-13, Dahlia, t. 1. f. 26, 27, and Gaillardia, t. 3. f£. 1, 2, which last is as undoubted a Heli- anthaceous style according to his views, and passes into the Asteroid Solidago style, t. 2. f. 7-9? Where would he, where could he draw the line? And if he had gone through many of the common tropical genera, he would have been obliged to re- move Gynura far from Senecio, Dahlia from Coreopsis, Spilanthes from Verbesina, de. Delpino next divides Senecionide into two subfamilies, Sene- cionee, with zoidiophilous, and Artemisiacee, with anemophilous fertilization—according to him, a very remarkable and constant distinction, accompanied also by a difference in the position of the capitula, erect in one case, nodding in the other. As to the two modes of cross fertilization, or rather of the conveyance of pollen, it has hitherto been observed in so very small a number of species, that I must refrain from expressing any opinion as to their constancy as generic characters; but I would only refer to his own note, p. 34, as to the occurrence of the two modes in dif- ferent species of one and the same genus. Erect and nodding capitula occur not unfrequently (e. g. Lactuca and Prenanthes) in different species of the same genus; they are not constantly nodding in Artemisia; and the Melampodineous genus Parthe- nice, too closely allied to Parthenium to be widely separated from it, has the habit and nodding capitula of the Ambrosieous genus Cyclachena. Ambrosiex are remarkable for their anthers less perfectly con- nate than in any other Composite, although closely approximate, forming the usual cylinder and often slightly cohering ; they are also distinguished by their terminal appendages inflected or hooked at the end, as observed by A. Gray*. The anther- bearing florets are as constantly sterile as in Melampodinee and Milleriew, and the styles of these sterile florets as constantly * In a recent part of the ‘Proceedings of the Academy of Natural Sciences of Philadelphia, 1869, p. 189, Mr. T. Meehan observes that in Ambrosia artemisiefolia this inflexed setiform appendage is only to be found on anthers which do not present perfect pollen; the abundantly polliniferous anthers are broad, without horns. DISTRIBUTION OF TRIBES. 437 undivided, but with the different termination pointed out by various synantherologists and insisted on by Delpino. The genera may be distributed into two groups. In the first, the Ives, the capitula are heterogamous, as in Melampodines, with female florets in the circumference and sterile antheriferous ones in the disk. In Jva itself, with seven or eight Northern or Central American species, the corollas of the female florets are much re- duced ; in the two species of Cyclachena and in the monotypic genera Euphrosyne, Dicoria, and Oxytenia, all Mexican or Califor- nian, the female corollas entirely disappear, the style proceeding from the summit of the naked ovary, or at most surrounded by a rudimentary ring. In the second group the capitula are strictly moneecious, the males usually placed in a different part of the plant from the females. The female florets are again apetalous ; but each one is completely enclosed in an involucral bract conso- lidated with more or less of the outer ones in a close utricle, from the beaked apex of which issues the style. These female capitula are sometimes one-flowered and distinct; sometimes there are two to four female flowers, each in a separate beaked and closed divi- sion of acommon mass. Whether this mass is an aggregate of two to four consolidated one-flowered capitula or a single capitu- lum with the inner involucral bracts closed round the achenes, as in Sclerocarpus or Melampodium, and connate with each other as well as with the outer bracts, is a disputed point, the advocates of each side of the question being certain that they are right. To me it appears that the inflorescence may be explained either way, the florets not being numerous enough to supply any such proof as we have in the case of Albertinia. The four genera of this second group are all American, or, at least, as it would appear, of American and probably western origin. Hymenoclea, two species, is Mexican or Californian ; Franseria, ten species, ranges from Chili to California, extending also eastward in North America; Ambrosia, twelve species, belongs to the same regions, but one or two of its species are also spread over a great part of the Old World, as in the above-mentioned cases of Elephantopus &c. Xanthium has two or three species, but too well known over almost all warm or temperate regions of the globe. The genus is probably of American origin, although the common species X. strumarium had evidently made its way into the Old World long before the discovery of America, and has established both in Asia and Europe many so-called species, none 438 MR. G. BENTHAM ON COMPOSITE. of which, however, appears to have acquired much of a local cha- racter, and some are repeated in America. X. macrocarpum, DC., first described as a Mediterranean plant, is believed to be a more modern introduction from America; X. spinosum, Linn., which has quite recently extended its range over new countries (e. g. Australia), was originally said to be Chilian. 5. PETROBIEX form a small subtribe of three genera compri- sing 4 species, very distinct, by their strictly dicecious capitula, and remarkable for the flowers of the two sexes being more nearly similar than in most heterogamous Composite. The co- rollas of both sexes are regular, though still different in propor- tions ; the stamens in the females are more developed than usual, having well-formed anthers, although small, free, and without pollen; the styles of the males are undivided in one genus, branched in the two others. They are trees or shrubs, with the other characters of Helianthoidew. Geographically they are all of limited range. Podanthus, with two species (or varieties ?), is Chilian; Astemma, a single species as yet only known from Hum- boldt’s original specimen with female capitula, is from the Andes of Quito; the third, Petrobium, also monotypic, is insular, being limited to the island of St. Helena. Several of the above circum- stances suggest the probability of these genera exhibiting the nearest approach to the primitive form of Composite. 6. ZINNIEZ are a group of five or six genera, comprising twenty- five species, only separated from the great mass of Verbesinew by the ligulate corollas of the female florets sessile or nearly so, and persistent on the ripe achene, without any external pappus or border at the base, and so deceptively continuous with the achene as in some instances to have given rise to a query whether this corolla did not really stand in the place of the pappus and represent the calyx-limb—a query, however, to which a careful examination will at once give a negative answer. These genera are all West- Ameriean and chiefly of the Mexican region ; Tragoceros four species, Zinnia twelve species, Sanvitalia three or four species, and Aganippea two species limited to that region; Heliopsis has three species, of which one extends southwards along the Andes and the two others eastwards in North Ameriea. One or two species of Zinnia, long in cultivation for ornament, have established themselves as colonists in some parts of the Old World. Phil- actis, a single species, also Mexican, is unknown to me, but pro- bably belongs to the Zinnia group. DISTRIBUTION OF TRIBES. 439 7. The main subtribe of VERBESINEA comprises about 570 spe- cies distributed amongst 57 genera, many of them natural enough, but distinguished by characters of comparatively small import- ance, sometimes passing into each other, and often very technical and very difficult to group together except into very artificial series. The great majority are American, many of them restricted to that continent or to limited areas within it; but some are well represented in tropical Africa, or in a less degree in Asia or Aus- tralia, and two small genera are exclusively Mascarene. The geographical distribution of the two following subtribes (8. Conz- OPSIDEZ,17 genera with 150 species; and 9. GALINSOGE®, 7 genera with 80 species) is nearly the same, the one comprising also 4 small exclusively Old- World tropical genera, and the latter 1 small Sandwich-Island endemic genus. As the differences which distinguish them from Verbesines are also of somewhat minor importance (chiefly the shape of the achene in the one case, the nature of the pappus in the other), and as the real value of the generie distinctions is often as yet uncertain in all three subtribes, we may consider the whole as one group, taking the principal genera rather in the order of their geographical distribution, com- mencing with those American ones which are also represented in the Old World by distinct species. Wedelia, about forty species, Blainvillea, ten species, slightly differing in the pappus, and Aspilia, thirty species, with neutral ray-florets, may be regarded as one large genus, chiefly Ame- rican, of which each of these divisions includes several Old- World species, the whole group also scarcely distinct from several other purely American genera. Wedelia itself, as now limited, com- prises three tolerably well-marked sections: (1) Stemmodon has three or four tropical-American species and one in tropical Asia, W. calendulacea, a maritime plant very closely allied to the simi- larly maritime West-Indian W. gracilis, although not identical; (2) Cyathophora, with numerous tropical A merican, has one insular (Galapagos) species, W. frutescens, Hook. f., which appears quite distinct from Jacquin's plant of that name, one East-Indian species, JV. urticifolia, and one in east tropical Africa ; (3) Wol- lastonia, with the pappus very much reduced or disappearing altogether, appears to be an Old- World deviation, and is limited to tropical and subtropical Asia and Australia. Although usually regarded as a genus, there is really nothing but this reduced pappus to separate it from Cyathophora; one species, indeed, so 440 MR. G. BENTHAM ON COMPOSITE. closely resembles the above-mentioned Asiatic W. (Cyathophora) urticifolia, as to be frequently mixed with it in collections. Blainvillea is represented in tropical Africa and Asia by three or four species, one of them proposed as a distinct genus, but which are all closely allied to a common east tropical American weed, the typical B. rhomboidea, Cass.; they seem, however, to be rather representative than identical species. Two other common American weeds (both nearly allied to, but sectionally or, according to some, generically diverging from B. rhomboidea), Blainvillea biaristata, DC., and Eleutheranthera ruderalis, Sch. Bip., are not represented in Africa. Aspilia was the generic name originally given to a Madagascar plant, which, on a comparative examination, has ap- peared to me to be strictly congener as well with the African Co- ronocarpi as with a number of American, chiefly Brazilian, plants, referred by different botanists to various genera, including the whole of the genus Anomostephium, DC. Amidst these several names, Dupetit Thouars's Aspilia has the right of priority. The genus thus formed is divisible into three not very well-defined series, of which two are exclusively Brazilian; the third, extend- ing in Ameriea from Brazil to Mexico, also ineludes some of the African species, though no one species is identical in the two continents ; the Masearene and one or two Afriean ones eannot be exactly included in either of the American series. The tropical American genera Zexmenia, twenty species, and Oyedea, twenty- two species, neither of them represented in Africa, are very closely allied to Wedelia, as is also the insular Lipocheta, consist- ing of ten Sandwich-Island and one Galapagos species (Macrea, Hook. f., united with Lipocheta by A. Gray). Sclerocarpus was originally established for a tropical African plant now known to extend into tropical Asia, remarkable for the receptacular palez completely enclosing the disk-achenes and hardened round them, so as to appear to form part of them. Pre- cisely the same structure was observed in some tropical American species, never compared with the African one, but distributed into various genera, although one of them (Gymnopsis uniserialis, Hook.), if not exactly identical with S. africanus, is so closely allied to it as to be strictly representative. The genus thus con- solidated comprises one tropical African and nine American spe- cies, chiefly from the Mexican region, but extending also into tropical South America. A similar structure is observable in the American genus Montanoa, of about fourteen species, ranging from Columbia to the Mexican region. DISTRIBUTION OF TRIBES. 441 Melanthera is a small natural group of about eight species, di- stinct in the form of its achenes as well as in habit and minor characters, and common to tropical America and Africa. It has been divided into four purely artificial genera :—Melanthera proper, four American species without any female florets ; Lipotriche, one African species with fertile ray-florets ; JEchinocephalum, one American, and Wurmschnittia, one Abyssinian species with neutral ray -florets. Spilanthes is another natural genus, readily distinguished from its nearest allies by the truncate style-branches and other charac- ters, and widely distributed over the tropical world. It is diffi- cult without a detailed study to fix either the number of species, ranging between twenty and forty, or rightly to appreciate the geographical distribution of some of them. The greater number appear to be American; and one or two of these, as in Elephan- topus, range over the Old World; but a few also appear to be really of Old-World origin, especially two extending from the Indian archipelago into Australia. Coreopsis, in the extended sense in which we have taken it, neg- lecting, as in Melanthera, the differences between the neutral and the fertile ray-florets, contains nearly sixty species, and, although chiefly American, has established distinct forms in tro- pical Africa and in the Sandwich Islands. In America the range of the genus is wide, but chiefly northern, western, or Andine, and consequently not quite of so tropical a character as that of most American genera represented in tropical Africa. Several local American species or groups of species have been separated at various times as distinct genera characterized by the fertility or by the reduction of the pappus of the ray-florets, or by slight modifications in the margins of the achenes, dc. The African species were by some singular misconception of characters re- ferred originally to Verbesina ; they have since been established by Schultz Bipontinus as a distinct genus under the name of Presti- naria; but they correspond too closely to the Peruvian (shrubby) or Californian (herbaceous) Agariste to be generically separated from them. The Sandwich-Island Campylothece, united by A. Gray with Coreopsis and by Schultz Bipontinus with Bidens, must be regarded as an insular group almost as near to the one as to the other, although technically referrable to Coreopsis. The species are so varied, however, in habit and in some minor points of structure, that they could scarcely be kept together had not their geogra- 442 MR. G. BENTHAM ON COMPOSITE. phieal isolation indieated a relationship which would not other- wise have been so clear. Bidens, a genus of about fifty species (nearly doubled by some botanists), although technically distinguished from Coreopsis by a somewhat trivial charaeter, the asperities of the awns of the achenes directed downwards instead of upwards, is nevertheless a natural genus ; and although geographically it may have as wide a range as Coreopsis, its distribution has a different character. The genus has two natural sections; one (Platycarpea) is so generally diffused over the temperate regions of the northern hemisphere that it would be difficult to determine whether its origin is Ame- rican or Europxo-Asiatic; and the representative species in the two regions differ but little from each other. "There are, however, two or three American species unrepresented in the Old World; and Cassini’s St.-Domingo genus Narvalina (a single species) may be considered as a divergence only from the Platycarpea, thus confirming some other evidences of the American ancestry of the group. One of our common species, B. tripartita, Linn., repre- sented in America by B. frondosa, Linn., and B. connata, Muehl., reappears in the southern mountain-ranges of Australia. The other section, Psilocarpea, is more tropical and essentially Ame- rican. Two species are indeed amongst the commonest weeds all over the warmer regions of the Old World; but that is a case similar to that of the Elephantopus, if, indeed, the presence of these species in some districts be not due to modern importations, won- derfully facilitated by the prehensile nature of the awns of the achenes. A Sandwich-Island Bidens, in its reduced pappus, shows an anomalous insular form, and may possibly be derived rather from the Coreopsides (Campylothecas) of the same islands. At any rate, this group shows the connexion of Coreopsis with Bidens, and is an example of divergence, with different combinations of characters in the isolated islands, from those which have become established in the general continental area of the genera. Various groups, further diverging from Coreopsis and Bidens, have arisen in various portions of the extended area of the genera :— in west tropical America, extending more or less from Bolivia to the Mexican region, Dahlia with four or five species and Cosmos with ten; in east tropical America Isostigma, five species; in subtropical America, north and south (Mexican and Bonarian regions), Thelesperma, four or five species; and in east tropical Asia and Australia, Glossogyne, five species. DISTRIBUTION OF TRIBES. 448 Amongst the epappose Verbesinew there are two small and very distinct genera which have a very wide tropical distribu- tion. Enhydra, with about a dozen species, is well represented in tropical Asia, Africa, and America, and has no near connexions in either country to indicate its origin, unless perhaps the Andine genus Aphanactis, two species, prove to be really allied toit. The most distinct species of Enhydra are also American; those of the Old World may be varieties of a single one. The other genus, Eclipta (three or four species), has likewise one cosmopolitan tropical species, to which the nearest allied local one is Australian; but beyond that there are no further connexions in the Old World. The remaining one or two species, forming a slightly distinct section, are in extratropical South Ameriea, where also is to be found the next nearest monotypic genus, Leptocarpha. Chrysanthellum is a small annual weed dispersed, under various names, over tropical Asia, Africa, and America, without affording any clue as to its original country, except the faint one supplied by a second species which has established itself in the Galapagos, tending to indicate an American origin. Affinities with other genera give no further assistance; for the nearest to it (though quite distinct from it) appear to be Heterospermum and Glosso- cardia, both monotypic, the one tropical-American, the other East-Indian. Synedrella, which is a nearer approach to the true Verbesinez, has two American species, of which one, like Zle- phantopus, is dispersed over tropical Africa and Asia. With regard to the Verbesinee strictly limited to America, the North-American genera take a great part, although not displaying any proportionate diversity of form except in the Mexican region. Rudbeckia, which, taken as a whole, is a natural and distinct genus of twenty-five species, is limited to North America and almost to the United-States region; so also are Balsamorhiza (ten species), Wyethia (four species), Helianthella (six species), all more or less diverging from Helianthus, but geographically rather more western. Tetragonotheca (three species) is likewise strictly North-American, but more distinct. Helianthus itself is by far the largest North-American Helianthoid genus; for about forty out of its fifty-two species are spread over that continent without having any special western character. It is, however, represented in Central and Southern America not only by a few species, which cannot well be generically distinguished from it, descending along 444 MR. G. BENTHAM ON COMPOSIT X. the Andes of Peru as far as Chili, but far more numerously by the Central-American genus Tithonia, three or four species, and the general tropical-American Viguiera, of about sixty species, both of which are, on the one hand, somewhat artificially distinguished from Helianthus, and, on the other hand, pass almost into the already-mentioned tropical Wedelia group, or into a few of the smaller Mexican or tropical genera which I shall presently refer to. Confined to the Mexican region we have nine or ten genera of one or two species each :— Rumfordia, Selloa, Aximiphyllum ( Aba- saloa ?), Varilla, Chromolepis, Mirasolia, lostephane, Otopappus, and Podachenium, to which we may add the Coreopsideous genus Coreocarpus, also of two species only. Small as they are, I do not think that any of these genera are sufficiently connected with any of their large cotribuals to be incorporated with them, unless these again be much more consolidated ; nor do they form of themselves a separate group in the subtribe. Like so many others of the same region, they may be considered as the scattered remnants of various ancient races. The distinct genus Encelia, which, taken in its natural extended limits, comprises twenty-two species, is also Mexican, but extends southwards and northwards from Chili to California. In the insular genus Scalesia, eighteen or ten Galapagian species, may be traced a connexion with the above mentioned Mirasolia, which belongs to the southern or Central-American portion of the Mexican region. In the South-American Andes we have again four genera of one or two species each :—Monactis, Stemmatella, Aphanactis, and Garcilassa, as much if not more isolated than those of the Mexican region, none of them having any nearer connexions than the general affinity to the whole subtribe. Pascalia, a monotypic genus of the same region but more southern, and quite or nearly extratropical, is generally allied to the Wedelia group. Among tropical-American Verbesinex, besides those already mentioned as connected with the North, the most important is Verbesina itself, with about fifty species, dispersed over the whole region, and represented by several species in North America, and one or two extending beyond the tropies to the south. One species, distinguished by most authors on very trifling characters under the name of Ximenesia, is met with in tropical Africa and some other warm countries, but evidently introduced from America, DISTRIBUTION OF TRIBES. 445 where alone true Verbesine and their immediate connexions are indigenous. The nearest slightly diverging genera are Actino- meris, nine species, ranging from the Mexican region eastward in N. America, only distinguished from Verbesina by the sterility of the ray-florets, and Salmea, twelve W.-Indian and Mexican species. The Brazilian monotypic Salmiopsis appears to be a connecting link between Salmea and some Brazilian Viguiere ; and the above- mentioned Mexican Otopappus may also be considered as a diver- gent form of Salmea. The tropical Wulifia, eight species, and the tropical and Mexi- ean Perymenium, ten species, have their nearest connexions pro- bably with Wedelia and with Melanthera. Gymnolomia, eighteen species, Zaluzania, seven species, and Sabazia, eight species, might perhaps be considered as a single genus ranging over the Mexiean region, Central America, the W. Indies, and Columbia, but not, as far as I am aware, extending into E. tropical S. America. The above-mentioned Mexican Varilla might perhaps be included with them. Jegeria, six tro- pical-A merican species, ranging from Bonaria to Mexico, in some respects approaches the same group, being evidently very near Sabazia, but really perhaps more nearly connected with Stemma- tella and Siegesbeckia, the chief geographical centre of all three being apparently the Andine region. The West-Indian Borrichia, three species, and Chenocephalus, one species, and the tropical-A merican Zrichospira, also monotypic, stand each of them isolated as it were in the great subtribe of Verbesine. There are two genera of the subtribe Coreopside, bordering upon Verbesinez, that are limited to tropical Africa (and, indeed, both of them have been hitherto observed as indigenous in Abyssinia only )}— Guizotia, with three species (one of them spread by culti- vation over East India), and Microlecane, one species. The nearest connexion of both may be with some of the African forms of Coreopsis ; but it is not very close. Two Mascarene genera, Micractis, one species, and Epallage, two species, evidently belong to Verbesinee ; but I am unable at pre- sent to trace out the genera they are most nearly connected with. The foregoing genera belong to the subtribes Verbesinew and Coreopside ; the third subtribe Galinsogee, which I have grouped with them, is entirely American. The genera of which it is com- 446 MR. G. BENTHAM ON COMPOSIT. posed have been usually classed under Helenioidex, on account of their scaly pappus ; but their affinity appears to me to be much greater with Verbesinex, of which they have the habit, the palea- ceous receptacle, &c., so much so that some species of Calea where the pappus is occasionally or constantly deficient, are diffi- cult to distinguish from Sabazia and its allied Verbesineous genera. Of the seven genera composing the subtribe, all well defined, if taken in their extended sense, three (Balduina, two species, Mar- shallia, three species, and the monotypic Blepharispermum) are ex- clusively N. American, three ( Galinsoga, three species, Calea, about sixty species, and Tridax, six species) are widely dispersed over the tropical and even subtropical regions of America, one species of Galinsoga and one of Tridax having become extensively spread as introduced weeds, the former in the temperate and tropical regions of the Old World, the latter within the tropics only. The seventh genus, Dubautia, four species, is insular, limited to the Sandwich Islands. 10. The subtribe Maprex forms a very natural group, con- nected, it is true, with the Helenioidez as well as with the Helian- thoidew, but most nearly so with the latter, and with a very limited geographical range. Five genera (Madia eight species, Hemizonia about twenty-five, Lagophylla three, Layia about twelve, and Achyrachena one species) belong to western N. America from Mexico to British Columbia, one of the species re- appearing in Chili. These might all be easily regarded as a single genus. The two other genera of the subtribe, differing more perhaps in habit and the large size of the capitula than in any important structural characters (Wilkesia, one species, and Argyroxiphium, two species), are insular, limited to the Sandwich Islands. 6. Helenioidee. The essentially American Helenioides connect the American Helianthoidese with the Old- World Anthemides on the one hand, and with the cosmopolitan Senecionidez on the other. The tribe is generally considered as forming three subtribes, or divisions of Senecionidez or Helianthoidez ; but it appears to me that the cir- cumscription of these large groups is more natural if they are kept distinct. The Helenioidee are not numerous in species, but varied in form; the species (not quite half as many as those of Anthemidex, under one third of those of Helianthoidex) average DISTRIBUTION OF TRIBES. 447 five to a genus, whilst in Helianthoidex the average is about eight, in Anthemidee above sixteen. It differs from Helianthoidee in the absence of any pales to the receptacle, and most frequently in the involucre approaching more to that of Senecionide: or of Anthemidez, in the pappus either of distinct equal scales or re- duced as in Anthemides, or passing almost into the sete of Sene- cionidez, in the shape of the achenes and in the greater fertility of the florets. Sterile disk-florets with undivided styles, charac- teristic of three considerable subtribes of Helianthoidez, are only known in the somewhat anomalous genus Blennosperma among Helenioidee ; and sterile ray-florets, not unfrequent in the former, have only been observed in Gaillardia among the latter, The geographical distribution is nearly that of Helianthoidez, but more strictly American, and chiefly western or extratropical ; only three species out of near 300 are known in the Old World, of which two are S.-Afriean, and one, identical with a S.-American one, is Australian. The tribe consists of four very natural subtribes, Beriee, Fla- veriee, Tagetinee, and Euhelenica, besides four more distinct genera technically united as a fifth subtribe under the name of Jaumiee. The subtribe Beriee, about 110 species in 30 genera, is the most characteristic of the tribe (although for the latter the name of Helenioidez has been adopted as having the right of priority) in structure as well as in geographical range. The Beriez are indeed throughout so eminently W.-American, that very little special mention need here be made of separate genera: 26 out of the 29 are found in the Mexican region (if we include California) ; three of these (Chenactis, Hymenopappus, and Pala- foxia) extend rather more eastward in N. America; one only (Schkuhria) extends into E. tropical S. America ; four ( Lasthenia, Bahia, Villanova, and Blennosperma) are represented in Chili by identical or nearly allied species. Of the three genera not yet observed in the Mexican region, one (Zhymopsis) is not far removed from it, being an insular form limited to Cuba, the two others (Closia and Amblyopappus) are Chilian, and may yet appear in the northern hemisphere. The genus Flaveria, seven species, with the closely allied mono- typie Sartwellia, belongs to the same W.-American region as the Beriex, extending from Chili to Mexico and Florida ; one species, either identical with or closely representative of the commonest LINN. JOURN.—BOTANY, VOL. XIII. 2 K a 448 MR. G. BENTHAM ON COMPOSITA. of the American ones, is also found in Australia. If only a colonist there, it must be so ancient a one as to have undergone some slight modifications in form. Asa subtribe these Flaveriex are rather further removed from Helianthoides, and approach the Tagetinem ; their involucre is the most prevalent one m Senecionides, the style that of Anthemidez and the larger por- tion of Senecionidez, the achenes such as are prevalent in Bwries and Tagetinez. The S.-African monotypic genus Cadiscus is anomalous, but appears to me to be much more nearly connected with the Hele- nioid Flaveriex than with any S.- African type. The Tagetinee, 13 or 14 genera and above 100 species, form as a whole a very natural group, which will, moreover, very naturally divide into three, Porophyllum, Tagetes, and Pectis, taken each in the most extended sense. All three have their principal seat in the Mexiean region, but extend in a few species all over the warmer parts of South America; very few species reach Cali- fornia, none extend far eastward in N. America. A monotypic form diverging from Porophyllum (Lescaillea) is insular, limited to Cuba; another monotypic, Schizotrichia, is Peruvian. None are known from the Old-World except as introduced weeds, one or two species of Tagetes itself, long cultivated for ornament, having almost naturalized themselves in some parts of tropical Asia and Africa. As a whole the subtribe connects Helianthoide; with Sene- cionidez, Porophyllum and some species of Pectis having almost the pappus of the latter tribe. The whole, or nearly the whole, are re- markable for the large oily receptacles or glands scattered on their foliage and involucres. Pectis (40 species) has the style-branches much shorter than in any other genus of Helenioide:e, or of any of the nearly connected tribes, and is, moreover, marked by the rigid cilia at the base of the leaves or petioles. Syncephalanthus, a monotypic form included among those which diverge from Tagetes, has a very curious inflorescence ; the capitula are collected in clusters which assume precisely the aspect of the single radiate capitula of Bebera, the central capitulum of the cluster having no ray, and the ray-florets of the surrounding ones being only on the outer side, so as to form one continuous ray for the whole cluster. This peculiarity occurs also in the S.-African genus Cdera, and appears to have no special significance, syste- matical or geographical The genus Clappia, two species, so DISTRIBUTION OF TRIBES. 449 closely connects Tagetines with Senecionidez, that it is difficult to determine to which it should be referred. The aspect and most of the characters, as well as geographical considerations, tend towards Tagetinee; but there appear to be no oleaginous glands. Both the species require further investigation from more perfect specimens. -Euheleniee, seven genera and about forty-five species, all American, are chiefly extratropical; they form in some instances a near approach to Anthemides, and may be generally con- sidered as the American representatives of that Old- World tribe, although in a very few species they also show an approach to Senecionidex. Structurally they differ from the preceding sub- tribes, chiefly by their shorter silky-villous achenes, and by their broader, more open, and sometimes Anthemoid involucres. The principal genera are not in N. America so specially Mexican as most Helenioides, but spread more equally to the east- ward. They generally, if taken with their most natural limits, include in one genus species with fertile or sterile ray-florets, or without any at all. Under this view Helenium, with about six- teen N.-American species, may be said to be represented in extratropical S. America by Cephalophora, four or five species ; Gaillardia, with five N.-American species, has a sixth extra- tropical southern one (Gúntheria). Actinella with ten species is confined to N. America; but Hymenoxys, four species, which is nearly related to it, but with a more Anthemoid aspect, is both in extratropical S. America and in the Mexican region. Psathyrotes, a Mexican genus of three species, has much of the involucre and pappus of a Senecionidea; but the achenes and some other characters are those of the Euheleniew, and the closely allied monotypic Trichoptilium, from California, connects it with the latter in the pappus also. There remain four genera, which, on account of their involu- cral bracts, imbricate in several rows, increasing from the outer to the inner, are anomalous in Helenioidex, and are artificially placed in a separate subtribe, Jaumiee. Two of them, Cacosmia four or five species, and Geissopappus two species, are from tropical Ameriea, and correspond in many respects to the Helianthoid genera Calea and Galinsoga from the same region, but have the naked receptaculum and the achenes of Helenioidese. Jaumea isa small genus which appears to me as distinct in habit and character as it is remarkable for its scattered geographical distri- 2x2 450 MR. G. BENTHAM ON COMPOSIT X. bution. Its five species have been published as so many separate genera, but never appear to have been compared with each other. The original typical species is a creeping maritime plant from Buenos Ayres with rayless capitula ; Coinogyne, a maritime Cali- fornian plant, is scarcely to be distinguished from it except by its radiate capitula; Espejoa is a Mexican species, with an erect branching stem and radiate capitula; Chethymenia is another erect branching Mexican species with radiate capitula, but with much less obtuse involucral bracts. Hypericophyllum, from $. tropical Africa, closely resembles Chethymenia; but the capitula are rayless as in the typical Jaumea, and the leaves, though entire glabrous and rather thick as in the other four, are broader. If all these had been found in the same district, no one would have doubted their being congeners; and had any of them, in its own special locality, diverged into allied forms different from those of the distant species, we might have admitted them, as distinct genera upon very slight characters ; but as none have any near connexions in their own district, we must conclude that they are all really congeners with the scattered distribution, hitherto un- accounted for, of Melasma, Alectra, and others. Venegazia, a monotypic Californian genus thus associated with Jaumiex, appears in some respects to approach Anthemides in structure; but the involucre, the achenes, and the pappus, as well as the geographical position, are those of Helenioidex. Olivea is another monotypic genus of the Mexican region, but rather more nearly connected with normal Helenioidez. 7. Anthemidee. Anthemides, with very few exceptions, are essentially of the Old World, chiefly extratropical, and far less varied than the two preceding tribes. About 650 species are contained in forty genera ; and several of these seem to pass into each other. It is not easy, either, to group them into well-marked subtribes ; and, as in the case of Asteroides, it will be necessary to consider a few of the principal genera as centres of groups round which others are more or less divergent. As a whole, Anthemidew are remarkably constant in their tailless anthers and truncate style-branches ; and their pappus, either very shortly paleaceous or coroniform or entirely wanting, has but very few exceptions. Their habit and involucre often connect them with Asteroidee on the one hand and Arctotidee on the other ; but their style readily distinguishes DISTRIBUTION OF TRIBES. 451 them from both. Between Anthemidex and Helenioidee (Euhe- leniez) there is, perhaps, no very definite boundary ; but generally the habit, involucre, or geographical distribution, accompanied by various minor indications, do not leave much doubt as to the position of a genus. In the ‘Genera Plantarum’ we have, for convenience’ sake, classed the genera somewhat artificially. In considering their geo- graphical distribution we must adopt a different sequence, taking :— first the great northern genera Anthemis and its allies, Chrysan- themum, Tanacetum, and Artemisia, some of which extend a few species into North America; then the South-African Athanasia, Hippia, and their allies, all restricted to that region; thirdly, the Cotulee, which are more cosmopolitan, belonging in a great measure to the southern hemisphere ; and, lastly, a few isolated local genera. Anthemis, above eighty species, Anacyclus, about ten species, Achillea, near 100 species, Santolina, about eight species, and the monotypic Cladanthus, Diotis, and, perhaps, Lonas and Mecomiscus; separated from each other by characters of no great importance, besides habit, are distinguished from Chrysanthemum by their paleaceous receptacle. The two larger genera, Anthemis and Achillea, range over Europe, North Africa, and extratropical Asia, their chief centre being the Mediterranean region and the Levant. One species, the common Achillea millefolium, appears to be spread over the whole of Europe, northern and central Asia, and a great part of North America, where are also mountain species of the same genus. But no Anthemis is to be met with in America or in the southern hemisphere except as weeds of cultivation. Santolina and Anacyclus are limited to the Mediterranean region taken in arather wide sense. Cladanthus is a West-Mediterranean plant; Diotis a maritime species extending along the greater part of the European and African coasts, around the Mediterranean, and along the Ocean from the Cape to the British Islands. Meco- miscus is an Algerian plant, exceptional in the tribe on account of its leaves opposite and entire as in the southern Eumorphia and GZdera. Chrysanthemum, taken in the extended sense we have given it in the ‘Genera Plantarum,’ includes above eighty good species, and has nearly the same range as the Anthemis group. It has, however, fewer mountain species than Achillea, and extends only into the extreme north of America ; southward it reaches much further than 452 MR. G. BENTHAM ON COMPOSITA. Anthemis or Achillea. Several species, usually of a somewhat shrubby growth, are natives of the Canary Islands, and two or three, also shrubby and somewhat anomalous, are South-African. But the circumscription of the genus is somewhat uncertain: some botanists divide it into about twenty genera, which may readily be distributed into five series; others, again, remove the greater part of the perennial species into Tunacetum. This, how- ever, does not much affect the group geographieally considered. Excepting the Canary-Island Argyranthema, the two or three larger series into which the genus might be divided have nearly the same range as the whole genus; and the numerous monotypic genera proposed belong chiefly toits great centre, the Mediterranean region. Richteria alone belongs to the mountain-region of Central Asia, where are also found two small genera, Addardia, four or five species, and the monotypic Cancrinia, which might almost have been included in Chrysanthemum, but for their pappus, which in both is exceptional in the tribe, showing an approach in the one case to that of Senecionide:, in the other to Helenioidew. One of the few South-African CArysanthema is, perhaps, a still further deviation from the ordinary type than the Canary-Island Argy- ranthema; but it has not been generically distinguished by the botanists who have worked out the Cape flora, and is not, perhaps, sufficiently known properly to appreciate its affinities. Matricaria, with about twenty species, has the wide range of Chrysanthemum, with, however, a southern preponderance—the perennial species with restricted areas belonging chiefly to South Africa, the northern species, chiefly annuals, having a very general distribution (partly as weeds of cultivation), two of them occurring in North, especially North-west America; two or three only of the more restricted species belong to the Mediterranean region. The twenty species have long been in a very unsettled state as to their systematic arrangement. Distributed into half a dozen small genera, or united in two only, severally associated some with Chrysanthemum, others with Tanacetum or Cotula, they are now generally recognized as forming one generie group, con- necting, as it were, Chrysanthemum with the Cotulex, differing from the former chiefly in the ribs of the achenes not equidistant round the achene, but more or less approximate towards the inner face, leaving a broader dorsal interval, and generally by their conical or elongated receptacle and the involucre approaching that of the Cotulee. In one species there is also a tendency to DISTRIBUTION OF TRIBES. 453 the reduction of the female corollas. This approach to Cotules is thus traceable both in geographical distribution and in structure. Tanacetum, about thirty species as we now propose to limit it, belongs exclusively to the northern hemisphere; for the South- African species retained by Harvey in the genus appear to be much better placed in Schistostephium. It has been found diffi- cult to draw up definite structural characters constantly to distinguish Tanacetum from Chrysanthemum ; for the most impor- tant difference, the female florets, short and tubular or filiform in Tanacetum, ligulate in Chrysanthemum, is unavailable in the few species or varieties where the female florets are deficient; and therefore Schultz Bipontinus and some others have brought a large number of the common Chrysanthema into Tanacetum. But this appears to me to-be a very unnatural combination ; and in the few cases which might otherwise have been doubtful, habit comes in aid of the distinction. Tanacetum also, on the other hand, runs as much into Artemisia; and in order to maintain some order in the tribe we must here, as in Asteroidez, admit as genera large and prominent groups, although they may be confluent on their borders. Tanacetum has a more Eastern range than Chrysanthemum; there are but few in Europe or in the West-Mediterranean region, more numerous in the Levant and Central Asia; some species extend into the far north, and thence into North America, where, in the mountains of the western regions, are two endemic species with some slight structural peculiarities which induced Nuttall to propose them as a distinct genus, Spheromeria. Artemisia, to which some botanists ascribe near 200 species, with the same general centre as Tanacetum, Asiatic rather more than European, has a wider range. Abundant in the temperate regions of the northern hemisphere throughout the Old World, it has also many Alpine and Arctic species, and spreads not only over the greater part of North America, but also down the western ranges of mountains to extratropical South America. Geographically Artemisia thus meets there with the genus Ambrosia, possessing a somewhat similar foliage, nodding capitula, a style in some respects similar, and, according to Delpino, a similar anemo- philous fertilization, to which characters I have already alluded under Ambrosiez ; but here the affinity ceases. There is nothing in Artemisia of that perfect separation of the sexes, of that free- dom or very slight connexion of the anthers, of their peculiar 454 MR. G. BENTHAM ON COMPOSIT X. inflected appendages, of the enclosure of the achenes in their subtending bracts, or of the many other features characteristic of the Ambrosiex ; and in the style it is only in a very few species of Artemisia that there is any approach to the consolidation of its branches in the sterile flowers, which is constant in Ambrosiez, and which, moreover, is of common occurrence in the sterile flowers of many other Composite belonging to very different groups: geographically, also, it is only as the outskirts of the wide range of some generally diffused species that the Mexican region, the great centre of preservation of the Ambrosiex, possesses one or two Artemisiee ; this genus has not there produced a single endemic form. Artemisia is, on the other hand, very closely con- nected with Tanacetum, and has intermediates in the true fatherland of the two genera: Artemisia fasciculata, Bieb., for instance, has the habit and inflorescence of Tanacetum, with the characters of Artemisia ; and the monotypie genus Cronostephium has the habit and inflorescence of Artemisia, with the characters of Tanacetum. There are other Asiatic species also which have given no small irouble to determine to which of the two genera they should be referred. The majority of the South-African genera (excluding Cotulee) require but little comment, although distributed with the northern genera into different series of the tribe according as their recep- tacle is with or without pale, or their female florets ligulate, tubular, or deficient. A family likeness may be traced between Athanasia, forty species, and some fifty species distributed amongst ten or eleven small genera; but no common character can be assigned them. Gonospermum, three or four species, from the Canary Islands, forms the nearest approach to Athanasia in the northern hemisphere ; and Sehistostephium and Pentzia may be compared with Tanacetum. -Hippia, four species, is in some respects an approach to Cotulee. But upon the whole these South-African Anthemides show a much more remote affinity to the northern ones than would have been supposed, from the genera being not only intermixed in our artificial classifications, but species of the two areas united by some in the same genera. South-Africa has also some small genera quite isolated, although technieally, as to structure as well as geographically, included in Anthemidez, such as:— GZdera, four species, with small opposite leaves and a peculiar inflorescence already alluded to; ZwumorpAia, one species, with small opposite decussate leaves, but with a DISTRIBUTION OF TRIBES. 455 normal inflorescence ; Eriocephalus, seventeen species, with some of the characters of the distant American subtribes Melampodinex and Ambrosiex; and Lasiospermum, four species, with the densely woolly achenes of some Arctotideze. The Cotulec form a rather more distinct group of Anthemidee ; and, geographically, they have been so long and so widely dispersed as to have established local genera or subgenera in very distant regions. Generally they belong to the southern hemisphere, and are mostly extratropical; but a few range over the temperate regions of the northern hemisphere of the Old World, or are within the tropics of both the Old and the New. They are gene- rally small annuals or dwarf prostrate perennials with small capitula, the involueral bracts nearly equal in about two rows, the female florets usually numerous, with short, regular or irregular corollas, not strictly ligulate, and sometimes very much reduced or entirely wanting. Cotula itself, with about forty species, has the wide range of the whole group; it has been variously subdi- vided into sections or distinct genera, without any of them (except when monotypical) having any distinct. geographical area. Wa- nanthea, the only allied genus exclusively northern, consists of a single species from the mountains of Corsica, showing some approach to the Chrysanthema and other Anthemidex of the same Mediterranean region. The slightly diverging genus Cenia, eight species, and Otochlamys, a single species, are limited to South Africa. Centipeda, three species, is more tropical in the Old World, although in America it is only in the southern extratro- pical regions. Plagiocheilus, six species, is limited to extratropical or Andine South America; so would also be Soliva (four species ?) but for one of them which has established itself in Australia, and another in Portugal and South Carolina, perhaps as ancient, pos- sibly as more modern colonists. Abrotanella is yet more southern, ranging from Antarctic America to New Zealand and the southern mountains of Australia. The three remaining monotypic genera, Ceratogyne, Elachanthus, and Isoetopsis, all from extratropical Australia, are somewhat anomalous in their styles and some other structural characters, but can scarcely be so well placed in any other tribe, and certain!y with none having similar geographical connexions. 8. Senecionidec. The tribe of Senecionidez, next to Asteroides the most numerous 456 MR. G. BENTHAM ON COMPOSITE. in species, has, owing to the overgrown proportions of one vast. genus, by far the greatest average number of species to a genus: nearly 1400 species are comprised in about 40 genera; but of these species, two thirds belong to Senecio itself, reducing the average of the remainder to more ordinary proportions. The tribe is divi- sible, according to structure, into four somewhat artificial sub- tribes, which, although generally confirmed by geographical dis- tribution, yet in this respect show some embarrassing exceptions. The principle subtribe, Euseneciones, is truly cosmopolitan ; Liabee are American, with one tropical-African exception ; Tussilaginee belong to the temperate northern regions, with one South-African exception; Othonnes to South Africa, with the exception of one widely spread high mountain genus. As a whole, the tribe is distinguished, amongst those which have tail- less anthers and a setose pappus, from Vernoniacesz and Eupa- toriacew by their yellow disk-florets and frequently heteroga- mous capitula, and from Asteroideew by their involucres, habit, and generally, though not always, by their styles. Senecio itself is not only the largest genus among Composite, but one of the largest, if not the largest, among Phenogamous plants, and certainly the most widely spread; truly cosmopolitan and ubiquitous, abounding in local species in almost every region of the globe, in the Old and the New World, from the equator to the arctic regions and the extreme south, on Alpine summits, in stony wastes or sandy deserts, in swamps, on sea-coasts, on the borders of streams, everywhere are Senecios to be met with ; and yet individually the species have not wide areas. No species is common to the New and the Old World, except in the far north; no one has, I believe its range interrupted by any consi- derable interval ; and notwithstanding the facilities for transport afforded by the proportions of the pappus and the achenes, few have a very wide area, or, as weeds of cultivation, establish them- selves in a new country with that readiness so marked in the Conyzoid Erigerons, for instance. It is, moreover, not easy to give any definite centre for the genus. It is less abundant, however, in the tropics, and most varied in temperate and cool or moun- tain-regions; so that some centre may be vaguely traced in the mountain regions and high latitudes of the northern hemisphere down the Andes from California to Chile, in Antarctic America, Southern Australia, and especially in South Africa. It is not easy, either, to divide it into sections or series by any combi- DISTRIBUTION OF TRIBES. 457 nation of structural characters and geographical distribution. Although more than forty genera have been proposed for species which we now include in Sezecio, I have failed in all my endea- vours to fix upon even a single group which I could definitely mark out. Besides the vegetative organs, there are differences, it is true, in the achenes, anthers, and styles; but these have been as yet observed in too small a proportion of the known species, or have been found too little in accordance with each other or with habit or geographical relation, to be made available for sec- tional distinctions. I had observed, for instance, that a number of East-Indian species, erect herbs with entire leaves and nume- rous few-flowered paniculate capitula, had the auricles of the anthers acute or subeaudate (Pl. IX. fig. 4,6, or 7), whilst in the majority of species they are truncate, obtuse, or somewhat acute, and then free and approximate to their own filament (figs. 2, 3, and 5); and I thought I had established a good section, to which I gave the name of Synotios. I found again the same foliage, and the anthers still more decidedly subcaudate, in some rather tall climbers of the same country, one of which had been generically distinguished by Miquel under the name of Cissampelopsis, and I added these to my section, although they differed in the larger many-flowered capitula. In S. buimala, Ham., however, another climbing species from the same country with still larger capitula, the character of the anthers failed. In the Canary-Island S. pal- mensis, DC. (the genus Bethencourtia, Chois.), the anthers, and in a great measure the habit, were found to be again those of my proposed section Synotios, which still might have been kept up; but when I came to the American species, I found the same anthers with pointed connate auricles exemplified here and there in West-Indian or Andine species, which had in other respects no connexion whatever with the above-mentioned East-Indian ones. Again, some North-American species have been retained under the old name of Cacalia, characterized mainly by their white-flowered homogamous capitula, with the style-branches produced into short cones and a somewhat distinct habit. The same flowers and styles occur in the South-African Aleinias with a totally different habit ; and these, again, agree in habit with the Kleinioid Senecios of the same country, although the style passes into the common Senecio form with truncate tips. The short appendages to the style are more or less distinctly observable in various species, which have on that account been placed in the genera Cacalia, Ligularia, 458 MR. G. BENTHAM ON COMPOSITAE. Gynoxys, or retained by all botanists in Senecio itself. These ap- pendages, however, are very short and often obscure, and scarcely more than rounded instead of truncate tips, or, as observed by Weddell, deceptions caused by the inequalities in the length of the hairs which form the terminal tuft. I had also long retained the North-Asiatie genus Ligularia, in which the collecting-hairs or papille descend more or less down the back of the style- branches, accompanied by a peculiar habit, with large subracemose radiating capitula. But here, again, I had ultimately to abandon the separation; sometimes the characters, sometimes the habit and geographical relations were at fault; and at present we are obliged to follow De Candolle in making our primary divisions of the genus purely geographical, and in each country subdivide them according to characters which have locally acquired relative im- portance. I think, however, that if any experienced monogra- phist were carefully to study the 900 odd species of Senecio, and especially to compare the various forms the ripe achenes assume, with the characters derivable from the styles, the anthers, the vegetative organs, and the apparent geographical origins, he might succeed in bringing out sectional combinations which have escaped me, and might even reestablish as independent genera some of those Cacalioid or Ligularioid groups which in the present state of our knowledge I have felt compelled to unite with Senecio. A. number of small genera, more or less divergent from Senecio, have a much more local character. One only, Erechthites, a genus of about a dozen species in a great measure tropical, distin- guished chiefly by the filiform female florets, has a wide range. Its great centre is South America; but it is found northwards as far as Carolina; and in Australia and New Zealand it has established several endemic species. In tropical Asia the single species observed is probably a recent introduction from America. In Africa it is, I believe, unknown; the Euseneciones which there assume the above-mentioned main character of Erechthites are connected with Senecio through different channels. The other American genera closely connected with Senecio are Culcitium, about 14 species, and Gynoxys (from which, following Weddell, we exclude the alternate-leaved scandent species), about 12 species, both genera Andine. Culcitium is very near Senecio, differing from some of the genuine species of that genus from the same country only in the involucre; and even in that respect there are intermediate species which have been alternately referred DISTRIBUTION OF TRIBES. 459 to the one or the other. Gynoxys, with more prominent appendages to the style-branches than in the aberrant Senecios above alluded to, is also definitely distinguished by the uniformly opposite leaves, which bring it near to some Liabeæ of the same region. In the tropical regions of the Old World, Gynura, about 20 species, including one from Australia, has diverged considerably in the style, which is an approach to that of the Liabes; but in all other respects the genus is close to Senecio: although enjoying a wide range, its chief centre is Eastern Asia. One African spe- cies, the genus Cremocephalus of Cassini, has a tendency to ex- hibit the deviation observed in Erechthites, the reduction of the female florets to the slender tubular form; but this character does not here appear to be sufficiently marked or constant to justify the retaining it as amonotypic genus. In the Mascarene Islands there are two genera ( Faujasia, three species, and Eriothrix, one species), a third (Stilpnogyne) in South Africa and a fourth ( Me- lalema) in Antarctic America, both monotypic, all differring from Senecio in the same character, the female filiform florets, but di- verging from such very different points of that great genus, that they cannot well be united on this ground, and Eriothrix and -Melalema especially have each a very peculiar habit. In Africa also Cineraria, as now reduced to twenty-five species, chiefly southern, with, however, three Abyssinian ones, differs from Senecio in the flattened achenes, to which there is no tendency in Senecio or its allies in any other country. The New-Zealand monotypic Brachyglottis and the Australian Bedfordia, two species, are both so near Senecio that they have been sometimes merged init; but they appear at least as distinct as several of the other divergent groups; and Bedfordia especially is exceptional in the tribe, and approaches the Australian Asteroides (Olearia) in habit and stel- late indumentum. The genera next in order of divergence from Senecio are all extratropical; four are N. W.-American (Mexican or Califor- nian)— Tetradymia with three or four species, Raillardella, Cro- cidium, Bartlettia, and Haploesthes, all monotypic, these last three showing perhaps some approach towards Asteroidez, but on the whole much nearer to Senecionidezm. Arnica, about ten species, a mountain genus, extends generally over the central and northern regions of Europe, Asia, and North America; and is distinct, especially in its opposite leaves and its involucre. Doronicum (as now modified so as to include Aronicum and exclude Pericallis) 460 MR. G. BENTHAM ON COMPOSIT. X. has ten or twelve species, and approaches Arnica in involucre and other characters, but with alternate leaves. Its geographical range in Europe and Asia is nearly the same, but less mountainous or northern, and it does not reach North America. Lopholena, a single South-African species with its singular involucral bracts and long style-appendages, is somewhat isolated in its relationships. There are three insular genera. Ist. Raillardia has nine Sandwich-Island species with an insular shrubby habit; A. Gray unites with it as a section the above-mentioned Californian mono- typie Raillardella, for both have long narrow style-appendages and a plumose pappus ; but their habit is so widely different as to suggest their connexion with the Euseneciones having been quite separate; and I have availed myself of some differences in the achenes and pappus to maintain the two as distinct genera. 2nd. Robinsonia, four species, and, 3rd, Balbisia, one species, are from the isle of Juan Fernandez, where they form small trees of a very pecu- liar habit. Although their connexion with Euseneciones seems greater than with any other subtribe or tribe, yet in their dicecious capitula, in the presence of small free anthers without pollen in the female florets, and some other points they approach the sub- tribe Petrobiee of Helianthoidee. Their convolute cotyledons have been pointed out as distinguishing them from all other Com- posite ; but, as already observed, that character is not constant in Robinsonia. In R. Gayana the embryo is usually, if not always, precisely that of the great mass of Composite. There remain two genera which show the great difficulty of giving technical characters to what appear to be natural groups, Werneria and Othonnopsis, the former with the characters of Othonnex, but evidently more naturally connected with Senecio, and Othonnopsis as evidently connected with Othonna, but with the characters of Eusenecionez. Werneria is a high mountain genus of about seventeen species, differing from Senecio in the involucral scales strictly uniseriate, united to near the middle or higher up in a regular smooth ribless lobed cup, and with a pecu- liar habit rare in Senecio. Its great centre is in the Andes of South America; but one species, unknown to me, has been de- scribed from Mexico; and I am unable to separate from Werneria generically, either in habit or in character, Senecio nanus, Sch. Bip., from the mountains of Abyssinia, nor the Ligularia nana, Dcne., from the Himalayas. Othonnopsis is an Old- World genus of eight species, chiefly South-African, but with one North-African, one DISTRIBUTION OF TRIBES. 461 Arabian, and one Persian species, all nearly allied to the South- Afriean Othonna (from which they were first separated by Spach), but with the styles and involucres of Eusenecionez. Othonnez, excluding the two last-named genera, form a small subtribe of about 130 species in 5 genera, entirely S.-African, with the exception of Huryops, which has established one species in Abyssinia and another in Arabia. The three principal genera, -Euryops, 27 species, Gamolepis, 12 species, and Othonna (including Doria), about 80 species, appear to me to have a very close natural connexion, although usually placed in three widely distinct tribes. Euryops is generally acknowledged as a Senecionidea ; Gamolepis, only differing from it in the want of any pappus, has on that account been referred to Anthemidez, where it is evidently a perfect stranger; and an occasional absence of pappus in groups usually provided with it has now been observed in too many parts of the system to retain its formerly supposed importance. The third genus, Othonna, has been referred to Cynaroides, through a false appreciation of the style, which has neither the termina- tion nor the external ring of hairs or so-called articulation of that tribe, but is a genuine Senecionid style with a trun- cate penicillate tip; only, as the disk-florets are sterile, it remains undivided as in most other tribes under similar cir- cumstances. Liabezx is a small subtribe characterized by its imbricate invo- lucre and Vernonioid style. The principal genus, Liabum, of about 40 species, is S.-American, chiefly Andine, but extending in a few species into the W. Indies and northward to Mexico. It has been almost universally classed under Vernoniacese on account of its style; but the yellow heterogamous usually radiate capitula, as well as the habit, are very foreign to that tribe, whilst there is much that connects it with Senecionidew. The opposite leaves, though not common in the latter tribe, are to be met with in Arnica, Haploesthes, and Gynoxys; and the style is scarcely so far removed from that of Gynura as the latter from the ordinary truncate style of Senecio. The W.-Indian and Columbian genus Neurolena, two species, admitted on all sides to be a Senecionid, is very nearly allied to Liabum, and, indeed, closely connected with it through Schistocarpha, a Mexican and Peruyian genus of four species, with the opposite leaves of. Liabum and the paleace- ous receptacle of Veurolena. The small genera we have included in Liabum, differing from each other more in habit than in 462 MR. G. BENTHAM ON COMPOSITE. character, have no separate geographical areas; they are all Andine. There is, however, a tropical-African monotypic genus which technically would belong to Liabex rather than to any other sub- tribe, but appears to be almost as much isolated in its natural affinities as in its geographical position. This is Gongrothamnus from E. tropical Africa, enumerated by De Candolle as a species of Vernonia, but differing from the whole of that tribe in its yellow flowers and triplinerved leaves, besides that the style-branches, being minutely papillose and not hairy, are not strictly those of Vernoniacez. Its nearest affinities remain yet to be traced out. The subtribe Tussilagines, which, as well as the Liabex, we consider as more closely connected with Senecionidew than with any other tribe, belongs in its normal genera to the mountain or temperate regions of the northern hemisphere. These genera are usually placed amongst Eupatoriacee on account of a slight re- semblance in the style-branches ; but their heterogamous capitula, frequently yellow, remove them as far from Eupatoriaces as Liabes are from Vernoniacee; and here we have, moreover, the prevalence of a truly Senecionid involucre. The subtribe com- prises four genera of undoubted afinity— Tussilago a single species, Petasites about twelve, Homogyne three, and Cremanthodium four or five species; the first three, constituting the old genus Tussilago, are all European, and Homogyne exclusively so; the other two extend over Asia and N. America. Cremanthodium is Himalayan ; only one of its species has as yet been published, and has been re- ferred to Ligularia, of which, however, it has neither the habit nor the style; and its affinity to Tussilago is confirmed by other species. Here, again, we have three isolated genera, which we can only class artificially as connected with Tussilaginew». One is Luina, a single N.W.-American species, with something of the habit of an Inula, and, indeed, some approach to that genus in the almost setose points to the auricles of the anthers ; but these points or sete are exceedingly short, and scarcely more than observable in some Senecios. The involucre is that of Senecio, the style- branches between those of Inula and Tussilago, the geographical position very far removed from that of Inula, but quite within the range of Tussilaginem. Peucephyllum, another monotypic N.W.- American genus, referred by A, Gray with doubt to Eupatoriacee, DISTRIBUTION OF TRIBES. 463 appears to me,on account of its yellow flowers and other characters to be referable rather to the present group, although it has even there no near connexions. The other extraneous genus is the S.-African Alciope, with two species, placed by De Candolle amongst Asteroidex, with which it seems to have much less connexion than with Senecionidezm. Its style is that of Cremanthodium ; the habit resembles that of Arnica, to which Thunberg referred it. It is not nearly allied to any genus of its own country. The anthers are those of the Sene- cionidez generally, and remarkable in that those of one species have the contiguous auricles of adjoining anthers connate, in the other the auricles are quite free and closed upon their own filaments. 9. Calendulacee. Calendulaces constitute the smallest and most compact of the tribes we have adopted, and might almost have been enumerated amongst the subtribes of Senecionidez (with which it has much more affinity than with Cynaroidez, under which it is usually classed), but that there is a tendency to produce appendages or tails to the anther-auricles, and there is never any pappus. The sterility of the inner disk-florets, sometimes accompanied by a similar sterility in the ray, and the large size acquired by some or all the perfect achenes are also peculiarities, which justify the maintenance of the group as a distinct tribe. Itisalmost entirely African. Of the three largest genera, two (Dimorphotheca, twenty species, and Osteospermum, thirty-eight species) are exclusively S.-African ; Tripteris, twenty-eight species, is also S.-African, but has likewise a North-African subtropical or tropical species. Olt- gocarpus has three S.-African species and one in the island of St. Helena, whether aboriginal there or whether an introduction from S. Africa, and being yet to be discovered there, remains doubtful. Calendula, with scarcely ten species, although double that number have been described, belongs to the Mediterranean region, extend- ing from the Canary Islands to Persia. Dipterocome is a curious anomalous monotypic Persian genus, evidently allied to Oligo- carpus, but thus placed on the limits of the tribe both structurally and geographically. Eriachenium is another monotypic genus, which Schultz has correctly referred to this African tribe, although it comes from a distant land, Antarctic America. It is anomalous in babit, but nearer to Oligocarpus than to any other genus. LINN. JOURN.—BOTANY, VOL. XIII. 2L 464 MR. G. BENTHAM ON COMPOSIT E. Ruckeria, two species, is a true S.- African, and appears closely to connect Calendulacez with Senecionidese (Othonnes); but the specimens preserved are as yet insufficient to make us fully aequainted with its characters and affinities. 10. Arctotidee. The Arctotidex, although twice as numerous, both in genera and species, as the Calendulacez, form still a small Old- World tribe, with their chief area in S. Africa, where, however, they have no immediate connexions. They pass on the one hand rather gradually into Cynaroidew, an Old-World tribe it is true, but almost exclusively of the northern hemisphere ; and at the other end they seem in some measure connected with some of the Anthemideous genera of the Northern, not-of the South-African type. On the whole they may perhaps be considered the southern representatives of the Cynaroidesw, with which Lessing and De Candolle associated them, but from which they differ es- sentially in their usually radiate heterogamous capitula, to a con- siderable degree in their styles, in the constant deficiency of tails to the anthers, and, as above, in their geographical distribu- tion. They consist of three or, perhaps, rather four subtribes, which must be reviewed separately. 1. The genus Ursinia (including Sphenogyne) forms a distinct group of about 54 species, all S.-African, although one of them reappears in (or extends into) Abyssinia, differing from Arctotidee generally in their truncate style-branches, their paleaceous recep- tacle, and glabrous foliage. It appears to me, however, to be more nearly connected with Euarctotez than with any other tribe or subtribe. De Candolle placed it among Heleniew, where it has certainly no connexions, structural or geographical. The peculiar palee of the pappus, distinctly convolute-contorted in their arrangement, are much more those of Arctotis itself than of any Helenioidee. The habit and involucre connect them with some Anthemides of the northern type, as well as with several of the true S.-African Euarctotex. 2. The Euarctotee comprise seven genera and about fifty species, with the broad involucres of which the inner bracts are scarious at the end of some of the northern genera of Anthemidese. They approach that tribe also in their pappus reduced to scales or en- tirely wanting, but differ in the styles approaching those of DISTRIBUTION OF TRIBES. 465 Cynaroides, in their achenes usually larger and thicker, and in other points. Their main area is S. Africa. Arctotis itself, out of thirty species, has one in Abyssinia; Landtia, four species, has two S.-African and two Abyssinian ; Haplocarpha, four species, Arcto- theca, one species, and Venidium, eighteen species, are exclusively S.-African, although one species of the first extends rather within the tropical limits. Cryptostemma has three species, of which one has become perfectly naturalized in Portugal ; and Cymbonotus, one species (the only one known of the tribe which is not African), has hitherto been only gathered in Australia, where the earliest ex- plorers found it fully established and apparently indigenous. Itis totally disconnected from any Australian genus, and it diverges much less from the S.-African genera of Euarctotes than the Magellanic Eriachentum above mentioned does from the Calendu- lacez ; the origin of both is as yet inexplicable. 3. The Gorteriee, with the chief characters of Euarctotez, differ from them in the involucres and some other points, which bring them nearer to Cynaroidew, of which they may be consi- dered the S.-African representatives, differing in their radiate capitula, and more or less in their styles and other points. We have here, therefore, among these Old- World tribes, the Anthe- mides of the Anthemis and Chrysanthemum type, all belonging to the northern hemisphere, connected with the Cynaroidex, also all northern, not by any northern groups, but through the almost exclusively southern Arctotidex ; whilst the intermediates between the southern Anthemidezx of the Athanasia group and the southern Arctotidez are to be sought for exclusively among the northern Anthemideex. The Gorteriese comprise about 120 species in seven genera, which do not appear to require any separate mention here ; for they have all the same S.-African range, with here and there a species ex- tending to within the tropics, but none, I believe, passing the equator. None have established themselves, even as introduced weeds, into distant lands. 4. Gundelia, a single Persian species, and Platycarpha, two S.- African species, are two very distinct anomalous forms, which, from some mistaken observation of their styles, had hitherto been placed amongst Vernoniacee, next to Elephantopus. The only connexion with the latter genus appears to be that of numerous few-flowered capitula being collected in a close gencral cluster or compound head. But that character exists in Asteroidez, in 212 - 466 MR. G. BENTHAM ON COMPOSITAE. Inuloidese, in Helianthoidee, and in Cynaroides; and our two genera are in other respects totally at variance with Vernoniacez. They appear to me to have the style and several other characters of Arctotidez, and really almost to close up the gap between the S.-African Arctotidee (Gorteriex) and the northern Cynaroidez (Echinopsidez); and one genus has the geographical position of the former, the other of the latter, both included in the general range of the two tribes. ll. Cynaroidee. The Cynaroide form the largest tribe in the northern hemi- sphere of the Old World, where they do not, with the exception of one very widely spread eastern species, cross the tropics south ward, their only extension, and that a sparing one, being into N. America and thence down the western ranges of mountains to Chili, with one Australian species. They comprise near 900 species in about 36 genera ; the subtribes into which they are divided are not well marked out, or, in some respects, perhaps too artificial, although there are some very distinct genera. Asa whole, taking geographi- cal distribution into account as a check upon structural characters, the tribe is definite in itslimits; their habit, involuere, receptacle, corolla, anthers, and styles are all characteristie, and though each one may show exceptions, these exceptions never occur in all the organs at once. Their nearest connexions are with the Muti- siacese on the one hand, and the Arctotidee (Gorteriee) on the other ; but the nearest connecting genera belong to these tribes respectively, and not to Cynaroidew. Not following precisely the subtribes of the * Genera Plantarum,’ we will consider successively Six prominent genera—Centaurea, Saussurea, Cnicus, Carlina, Xeranthemum, and Echinops, taking under each the smaller genera more or less diverging from them—the first three of which are the only ones of the tribe which extend into America. The great centre of the whole tribe is the Mediterranean region, taking it in its extended area so as to include Persia; and many genera are limited to its eastern portion. Centaurea, as most generally understood, is a genus of about 320 species, having the geographical range of the tribe, most abundant in the Mediterranean region and the Levant, but extending in America and Africa to the utmost limits of the general area of Cynaroidem. Although very fairly defined as a whole, it presents such infinite variety in the tips or appendages of its involucral DISTRIBUTION OF TRIBES. 467 scales and in the pappus, that it has been split up by Cassini and others into above fifty genera. Cassini’s, founded chiefly on the involucral scales, are, perhaps, the most natural; those of De Candolle, Spach, and Boissier, derived mainly from the pappus, may be rather more definite, but are very artificial, often widely separating species otherwise closely allied ; and very few of the groups formed on either grounds have any local character. The very few American species (North-western or Chilian) are con- nected with each other by their large capitula with broad fringed or torn scarious appendages to their involucral bracts ; but the nearest approach to these are from the diametrieally opposite limits of the general range—two Abyssinian species, which Boissier has even proposed to add to the genus Plectocephalus, founded on the American ones. The Abyssinian ones, however, are really, notwithstanding their pappus, more nearly connected with some of the European or Asiatic species of the Lopholoma group. A species still more remarkable for its distant outlying station is the Australian Leuzea australis, Gaudich., which we now find it necessary to associate with the section Rhaponticum of Centaurea. It is in some measure allied to the Abyssinian and West-American large-headed species above-mentioned ; but its closest affinity is with a Spanish species, the Leuzea rhaponticoides of Graells. A considerable number of the species have a rather wide range within the general area; and some appear to hybridize readily. C. nigra, belonging chiefly to temperate regions, extends over the greater part of Europe and extratropical Asia ; and two species, C. calcitrapa and C. melitensis, are frequently carried out in baliast or as weeds of cultivation to distant lands. A large number, how- ever, of the species are restricted to small areas. The small genera Crupina, two (or, aceording to some, five) species, Volutarella, four or five species, Zoegea, two species, and Leuzea, one species, all slightly diverging from Centaurea, belong to the same Mediterranean region taken in an extended sense E. and W., but do not spread northwards. Carbenia, one species, Carthamus, about twenty, and Cardun- cellus, about fourteen species, belong still to the Centaurea group and Mediterranean region, more abundant in the west than in the east, the chief character connecting all the above genera consisting in the very oblique or lateral scar at the point of attachment of the achenes. The same character of the achenes, though perhaps usually no 468 MR. d, BENTHAM ON COMPOSITAE. so much pronounced, prevails in the allied genera Serratula, about thirty species, and Zricholepis, seven species; but a very percep- tible difference in the involucral scales gives them a different habit, and their geographical range is not quite the same; they are more northern and eastern, and not quite so Mediterranean. Serratula itself is, moreover, somewhat exceptional in the tribe by the great reduction or almost total suppression of the tails or appendages to the anther-auricles. This genus extends over the whole of Europe and temperate Asia, but is not in America. Zricholepis is exclusively Asiatic. Myopordon is a monotypic Persian genus of which the affinities are as yet very uncertain. It is evidently near. the Centaurea with spinous involucres ; but the areola at the point of attachment of the achene seems to be quite basal; the specimens, however, are imperfect. The Saussurea group of genera have the basal scar to the achenes of the Cnicus series; but their filaments are always glabrous and free, and their involucral bracts, having neither the prickles of the majority of the Onicus series nor the scarious appendages so frequent in Centaurea, give to the plants a very different facies. Saussurea itself, about sixty species, is also distinguished by the pappus, either with a single row of sete or the external sete comparatively few, fine, and short. Geographically it is of a much more mountainous character, with some species consequently of a much wider range than most Cynaroidee. It has several high- Alpine or Arctic species, and extends over Europe, extratropical Asia, and rather high northern America: two or three species descend in Asia to within the tropics; and one of these Asiatic species has extended itself into eastern Australia to the utmost limits of, and even beyond the tropical region. Stehelina, six species (including possibly the monotypic -Kechlea), is, as it were, the Mediterranean representative of Saussurea, and has also a single-rowed pappus, but of a somewhat different texture, and the setze mostly united in pairs or in bundles. It is limited to the Mediterranean region. Jurinea, about forty species, has much of the aspect of Saussurea, and is divisible, chiefly according to inflorescence, into sections corresponding to those of Saussurea, some of them, perhaps, rather more distinct; and one, ZEgopordon of Boissier, with the set» of the receptacle almost as much reduced as in Onopordon and Be- rardia, might, perhaps, be retained as a monotypic genus. As a DISTRIBUTION OF TRIBES. 469 whole, Jurinea differs from Saussurea chiefly in the multiseriate pappus. It is less mountainous than that genus, and more Medi- terranean in its character; it does not extend to the Arctic regions or to America, although abundant in Asia. The monotypic Goniocaulon, from East India, and Berardia, also monotypic, from the western extremity of the Alps, though distant from each other, are very near to different sections of Jurinea. Warionia, from the Sahara of Africa, and therefore from the extreme limits of the Cynaroid-area, is a very distinct form, although still referable to the Jurinea group. Of the Carduinz proper, or true Thistles, with the leaves and involucres usually prickly, the filaments hairy or monadelphous, the areola or scar of the achenes basal, and the pappus-sete in several rows, Cnicus, above 150 species, is the largest as well as the widest-spread genus. Like Centaurea it is diffused over the whole of the Mediterranean region, Europe, and extratropical Asia, from the Canary Islands to Japan, and extends also into North America and down the western mountain-range to the tropics, but scarcely beyond; and two or three species are readily carried with cultivation into the tropics and beyond them. Like other large Cynaroid genera it has been divided; but none of the genera proposed to be dismembered from it among the great mass of Old-World species have any natural structural character or special geographical range. In America, however, it appears to have been very early established in the Mexican region, and there to have diverged more or less into a special group with large beads and peculiar, often highly coloured, involucres, culminating in the Erythrolena of Don, which, however, is too closely connected through a long series of intermediates with some of the Old- World forms to be maintained as a genus. The whole genus Cnicus is often merged in Carduus, the two differing only in the pappus, plumose in the one, simply setose in the other, and naturally forming but one group. The geographical range would not be materially affected by the union, except that Carduus in the limited sense, with between thirty and forty species, has a much more restricted area than that of Cnicus, being unknown in America. Onopordon, twelve species, Cynara, six, Silybum, one, Galactites, two, and Tyrimnus, one species, are all forms very slightly diverging from Carduus and Cnicus; and all belong to the Mediterranean region taken in a wide sense, Cynara extending to the Canary 470 MR. G. BENTHAM ON COMPOSITA. Islands; and one species carried out with man to extratropical $. America, has there found circumstances so congenial to its consti- tution that, as a successful eolonist, it has overpowered the native vegetation over vast tracts of country. None are truly Àmerican or high-northern. Cousinia is another genus of true Thistles, differing from Carduus and Cnicus in the slender, fragile, usually small and exceedingly caducous setze of the pappus, and remarkable for the large number of species confined to a small area. Above 100 well-marked species have been described ; and many more exist in our herbaria, although the whole genus is limited to Western and Central Asia, the chief centre of its range being in Persia; and at its western extremity it barely reaches the Mediterranean. Arctium is a small European and temperate Asiatic genus, of which the number of species, whether two or seven or eight, is a matter of contention, and which, though not prickly, is but a slight divergence from Carduus, with a pappus nearly that of Cousinia. The Carlina group includes three closely allied genera which might be treated as one—Carlina itself, fourteen species, Atrac- tylis, about the same number, and Thevenotia, two species. They connect the Carduus and Cnicus group with the Xeranthema. With the prickly thistle-like aspect of the former, they have the densely villous achenes and simple series of more or less paleaceous pappi of the latter. Geographically they are widely spread, although they do not reach America. Their chief seat is the Mediterranean region; and one or two species extend over the greater portion of Europe and extratropical Asia. They have also, always within their general Old-World range, established some local forms distinct enough to have been often considered genera. These are Carlowitzia (belonging to Carlina), two species in the Canary Islands, Thevenotia, two species in Persia, and Atractylodes (now reduced to Afractylis), two species in Japan and China. Atractylis Preauaii is another Canary-Island form, which might almost be regarded as sui generis. The corollas of the sterile florets at the circumference of the capitula in Atrac- tylis assume the palmate or 5-merous ligulate form which, as in Stokesia, indicates an approach to the Cichoraceous corollas, not to those of the rays of ordinary radiate capitula. Xeranthemum, four or five species, and the closely allied mono- typic Chardinia and Siebera form a small very distinct group of Cynaroidez, limited to their great centre the Mediterranean DISTRIBUTION OF TRIBES. 471 region and Levant. In habit, achenes, and pappus they show some approach to Catananche in Cichoriacee ; in involucre a dis- tant resemblance, but no affinity, to Helichrysee ; their external female or neutral florets tend towards the bilabiate form of Mu- tisiaceee ; but their main characters are so essentially those of Cynaroidex, that they cannot be really considered much in the light of connecting links with either of the above outlying tribes. Amphoricarpus is a monotypic Dalmatian genus, with which I am not sufficiently acquainted to form any opinion on its sup- posed relationship to Xeranthemum. Echinops, including the small almost monotypic genus or sec- tion Acantholepis, forms an exceedingly distinct group of nearly seventy species, ranging over the Mediterranean region, the Levant, and Central Asia, which, in the numerous uniflorous capitula collected in dense globular clusters or compound heads, have the same relation to the true Cynaroides that the Gun- deliex have to the Arctotidee. To a certain degree also there is here some approach of the two tribes to each other; but the gap is still wide. On the other hand, Cardopatum, of two spe- cies (one from Algeria, the other from the Levant, and therefore from the same region), may be said to form a connecting link between Echinops and Carlina. 12. Mutisiacee. The Mutisiacez are varied in form and widely scattered in geo- graphical position. About 450 species are contained in 50 genera, the chief centre of which is far distant from that of the tribes they are most nearly connected with, although there is some over- lapping of their respective areas. They are most nearly allied in structure, though most opposed geographically, to Cynaroi- dex; some genera (Gochnatiex) have almost the characters of that tribe; and there is scarcely a very definite line between the same Gochnatiee and Inuloidee (Euinulez), whilst there are some genera also which (with Stokesia in Vernoniaces) exhibit the nearest approach in the order, though still but a distant one, to Cichoriacee. The chief centre of Mutisiacez may be said to be Western, and especially South-western, America, where Cyna- roidez are not represented by any endemic genus and are very slightly so by a few outlying species of large European genera, where Euinulee are unknown and Cichoriacee few; whilst the Mediterranean region, the great centre of Cynaroidew 472 MR. G. BENTHAM ON COMPOSITE. and Inuloideæ, and in a considerable degree that of Cichoriacee, is entirely deprived of Mutisiacez, or at most has a single species on its extreme southern limits. The five subtribes into which the tribe is divided on structural grounds are not very strictly geo- graphical: Barnadesiee are South-American; Onoseridee also South-American, with the exception of two or three tropical- African forms; Gochnatiex and Gerberes have several represen- tatives in Asia and Africa; Nassauviex are exclusively American and chiefly, but not entirely, western and extratropical. It is better, however, here to consider the genera of the first four sub- tribes rather more in the order of their geographical areas. Barnadesia, ten species, Mutisia, thirty-six species, Onoseris, about twelve species, Chuguiragua, above thirty species, all very distinct and well-defined genera, are all South-American and Andine, but each one extending eastward in one or more Brazi- lian species (in Chuquiragua nearly half the species often sepa- rated under the name of Plotovia) without any connexion with the Old World. Round the above may be grouped three smali Andine genera, Plagia, three species, and Aphyllocladon and Chie- nopappus, both monotypie, as well as three very distinct small genera from east tropical America (Brazil or Guiana), Schlechten- dahlia and Wunderlichia, both monotypic, and Stifftia, four spe- cies, which might almost be considered as two or three distinct genera. Three tropical-African forms also (Pletotaxis, one spe- cies, Erythrocephalum, two or three species, and Phyllactinia, one species) appear to be more nearly connected with the Ame- rican Onoseris group than with any genera of their own country. Gochnatia, ten South-American tropical or extratropical spe- cies, Moguinia, twelve species, all Brazilian except one from Mexico, together with Seris, two Brazilian species, and Hyalis and Cyclolepis, both monotypic and extratropical South-American, all closely allied to each other, form a rather natural group approach- ing in many respects some Cynaroidez (of the Saussurea group), and more remotely connected with some Euinulez. This group is represented in Cuba by the genus Azastraphia, four species ; in southern and in a less degree in tropical Africa by Dicoma, thir- teen species, and Hochstetteria, one species ; and still more closely in the Himalaya by the monotypic Leucomeris, scarcely distin- guishable from Gochnatia itself, except by the corymbose inflor- escence. The African genus Dicoma above mentioned, of which one of DISTRIBUTION OF TRIBES. 478 the tropieal species extends into East India, includes various forms with great diversity in habit, involucre, and pappus, but so connected with each other as to render it difficult to distribute them even into well-marked sections; the most marked form, some species of the section Pterocoma, DC., offer the only instance of some approach in outward aspect to the above-mentioned Bra- zilian Schlechtendahlia. Lrichocline, about twenty South-American extratropical, subtro- pical, or Andine, with one Australian species, Chaptalia, eighteen South-American tropical or extratropical species, represented also in Mexico and the southern United States of America, and Gerbera, twenty species, chiefly South-African, but with a few tropical or mountain species dispersed over tropical Africa, East India, and Hastern Asia as far Japan, form one natural group, divided by some into about sixteen genera, but fairly separable into the three above mentioned; for I think there are structural characters fully suflicient to separate the American Chaptalias from the Old-World Gerberas, with which Schultz Bipontinus unites them. All three genera bave a uniform habit, the leaves all radical, usually white underneath, and monocephalous scapes. The single Australian species, which I had once described as a genus under the name of Amblyspermum, I now find to be inseparable from the South- American Zrichoclines. Lycoseris, ten species, all South-American and chiefly Andine, extending from Bolivia to Central America, Chetanthera, twenty- six species, and the monotypic Brachyclados and Iobaphes, all extratropical or high Andine, belong to the same Gerbera subtribe, but are very distinct from the three last-mentioned genera, and are unrepresented in the Old World. Four small genera with a considerable family likeness, although each with well-marked structural characters, 4insliea, ten spe- cies, and the monotypic Macroclinidium, Pertya, and Myripnois, are Asiatic and Eastern Asiatic, and mostly extratropical or Chino-Japanese. The last-named three genera have much of the character of some of the Gochnatia group, especially of the Cuban Anastraphia, but also show an approach to some Cynaroidez, as, for instance, to the (similarly Japanese) section Atractylodes of Atractylis. Ainslica, on the other hand, which descends sparingly to within the tropies, has somewhat of the habit, though not much of the characters, of some Cichoriacem ; and one species was ori- ginally described as a Hieracium (H. silhetense, DC.). A very 474 MR. G. BENTHAM ON COMPOSIT X. curious monotypic genus, however, from the Siwalik hills of East India, Catamixis of Thomson, has really the pentamerous regu- larly 5-toothed corollas of Cichoriaceze, but with the anthers, styles, achenes, and involucres of Mutisiacee, thus forming a real point of connexion between the two tribes. The habit is an uncommon one in either tribe, being rather that of a Baccharis or a Pluchea. There are two very exceptional South-African genera which can only be referred to Mutisiacez, but for which 1 can suggest no near connexions :—Anisocheta, a single species, which, probably from some vague resemblance to Mikania in its climbing habit, pa- nieulate inflorescence, and few-flowered capitula, had been placed by De Candolle in Eupatoriacez, of which it has neither the anthers nor the style, nor the corolla, nor the opposite leaves. Though far distant, yet it appears to me to be better placed near the Gochnatia group than in any other position I can assign to it. The other is Oldenburgia, three species, perhaps not strictly con- geners, with all the essential characters of the Gerbera group, but in their singular habit coming nearer to the Brazilian Wunder- lichia, belonging to the Gochnatia group. The only American plant of the Gerbera group approaching it in habit is perhaps the Chilian monotypic Pachylena, Hesperomannia is an insular (Sandwich-Island) monotypic genus of the Gochnatia group, of arborescent habit, with the large capitula, achenes, and pappus of the Brazilian typical Stifftia, from which it differs in the involucre and style. The fifth subtribe, Nassauviez, with the corollas more con- stantly and distinctly bilabiate than several of the preceding subtribes, has also rather different connexions, having gene- rally the styles and occasionally the involucres of Senecionidee. It is exclusively American, chiefly southern and western, extending into North America only along the Mexican region to California. The genera, as consolidated by the most recent synantherologists (without going quite so far as Schultz Bipon- tinus) are natural and well defined, some of them quite iso- lated except as to the general tribual characters. Perezia, forty, and Trixis, thirty species, extend from the Argentine States and Chili to Mexieo, with.one or two more eastern tropical species in Brazil or the West Indies. Proustia, six or seven species, very nearly allied to Perezia, with the habit of some Gochnatie, is generally South-Andine, but is also represented in Mexico. Jungia, with twelve species, very near Trizis, does DISTRIBUTION OF TRIBES. 475 not reach further north than Columbia. Leuceria, twenty-five species, Polyachyrus, seven, Nassauvia, twenty-five, T'riptilion, six, Pamphalea, four or five species, and the monotypic genera Osy- phyllum and Moscharia are limited to extratropical South Ame- rica, chiefly Chili, or they advance very little northward along the Andes. Macrachenium is a single Magellanic species with the habit of a Chaptalia, and the characters nearer those of Trixis ; and Cephalopappus is a single and rare Brazilian species, with the characters nearly of the Chilian Pamphalea, but a totally different habit. 13. Cichoriacee. The tribe of Cichoriaces is an extensive one, and, as already observed, is the most definitely marked out in the whole order. It has also a wide geographical range. Its chief seat is in the northern hemisphere and more especially the Old World, where most of the larger genera have the great majority of their species. Most of the genera of limited areas belong to the Mediterranean region; yet several are also located and have been apparently developed in Western America, especially in the Mexican region. The number of species known is above 700, distributed into nearly 60 genera, not always very clearly defined, yet we believe rather better marked out than the very numerous smaller ones into which they have sometimes been divided. It is very difficult to arrange these genera into subtribes ; and those we have adopted are in a great degree artificial, and have little or no connexion with geo- graphical distribution; we must therefore now consider the principal genera separately. Crepis (including Barkhausia and Youngia), about 130 species, to which might be added about a dozen more contained in the small slightly divergent genera Pterotheca, Phecasium, Phalacro- deris, and Rodigia, is essentially of the Old World. The few N.-American species, although proposed by Nuttall as two di- stinet genera, Psilachenia and Crepidium, belong to the typical group of Eucrepis. The genus is divisible into twelve to fourteen sections, not all very distinct, but each marked by some peculiari- ties. Most of them, as well as the four small divergent genera above mentioned, belong to the Mediterranean region; two, Eucrepis and Barkhausia, range over Europe, N. Africa, and extratropical Asia ; and the former, Eucrepis, extends also into N. America; the sections Soyeria and Intybellia belong to the 476 MR. G. BENTHAM ON COMPOSIT x. mountain-regions of Central Europe and Asia, and are replaced in the Mediterranean region by numerous species of the section Omalocline, some of which are also Alpine, but only in the southern ranges of mountains. Youngia is more Asiatic and especially eastern ; one species is tropical and extends to the northeru dis- tricts of Australia. The Anisoramphus proposed by De Candolle as a genus, but which we have with some hesitation reduced to a section, is a single tropical-African mountain species. Hieracium, about 150 species, increased by some botanists to between 200 and 300 or even more, bas a geographical range un- usual in Composite, upon which climatological influences may have had some effect. It belongs chiefly to mountainous or tem- perate regions; and in the Old World it is essentially western, from Scandinavia to the Spanish Peninsula and the Western Alps; in Eastern Europe the species are more rare, and in Asia very few. In America they reappear in the extratropical regions N. and S., and down the range of the Andes; and two species are in the southern hemisphere in the Old World, one in S. Africa, the other in Madagascar ; none are known in Australia. The genus is divided by Fries into three principal and natural sections—Pilo- sella, Archieracium, and Stenotheca. The first two comprise nearly the whole of the Old- World species of the northern hemisphere. Stenotheca is represented in the Old World by one or two species of the Western Alps, and by the two southern species above men- tioned, all evidently nearly allied. The American species, of which Fries enumerates forty-five, but which are probably reducible to little more than half that number, have been, with the exception of two or three high northern Archieracia, referred by that writer to Stenotheca, to which they appear certainly for the most part nearer than to Pilosella, to which Schultz Bipontinus refers them; they have, however, to a certain degree a facies of their own, passing, perhaps, from the one to the other,but not representing the Archieracia excepting as congeners. The small genus Andryala, variously estimated at from half a dozen to above a dozen species, is a slight divergence from the European Hieracia, with the same western character but more southern, from the Mediterranean region to the Canary Islands. Picris, about twenty-four species, has its chief seat in the Medi- terranean region, especially its western portion, extending also down to the Azores: two species are generally spread over Europe and Western Asia; and one is to be found in most parts of DISTRIBUTION OF TRIBES. 477 the world, especially the extratropical world north and. south, and has received a separate name in. almost every country, although the distinetive characters given might be generally found in Euro- pean specimens. It is so readily carried, however, with cultivation, that it is difficult to say how far it is a denizen or a colonist only in the distant regions where it is found. Sonchus, with about the same number of species (twenty-four) as Picris, is similarly cireumstanced as to geographical distribution, the chief seat being the Mediterranean region and the Canary Islands, with one or two species everywhere accompanying culti- vation and possibly true denizens in more than one distant extra- tropical region ; but neither the one nor the other has established any endemic groups or distinct species beyond the main area of the genus. Lactuca, sixty species, nearly allied to Sonchus, has a much wider range, and is generally moreeastern. It is, however, divisible into five or six sections, somewhat different in their geographical distri- bution. Brachyrhamphus and Phenixopus belong specially to the Mediterranean region. Scariola, containing the typical Lettuces of many botanists, is more generally spread over Europe and a great part of Asia; Cicerbita and Mulgedium, especially the nu- merous and showy blue-flowered species, are frequent in mountain- districts, and extend over Europe, central and temperate Asia, and N. America, with endemice species in each country. Zzeris, again, is yet more eastern Asiatic, with one European species, and bears much the same relation to Lactuca generally which Youngia does to Crepis. Chorisma, allied to Zxeris, consists of a few species scattered over various parts of Asia, from Asia Minor to Japan. Prenanthes, sixteen species, nearly related to the section Cicerbita or Mulgedium of Lactuca, has the same range, chiefly mountain- ous, over Central Europe, the mountains of Asia and N. America. The other genera diverging from Lactuca and Sonchus are limited to the northern hemisphere of the Old World. Chondrilla, fifteen species, extends over the Mediterranean region and a great part of temperate Asia. Picridium, variously estimated at from five to ten species, and Microrhynchus, about twenty species, belong to the Mediterranean region, the latter extending into north tropical Africa and the Canary Islands. Heterachena, one species, is Abyssinian and Arabian ; and the monotypic Dianthoseris, allied to Lactuca, but with the habit of a Werneria, is limited to the mountains of Abyssinia. Five small rather more distinct genera, 478 MR. G. BENTHAM ON COMPOSIT X. comprising amongst them only eleven species, connected in some measure both with Crepis and Lactuca (viz. Zacintha, Acanthoce- phalus, Heteracia, Rhagadiolus, and Kelpinia), are also limited to the Mediterranean region. Leontodon, about forty species, is another of the Mediterranean genera which has a few species widely spread over Europe and extratropical Asia, and two or three are now to be met with in various distant regions, but probably as colonists only, except in North America, where Torrey and Gray’s monotypic genus Apar- gidium may be considered as an endemic species of Leontodon, nearly related to one of the European mountain species. The genus has been broken up into ten or twelve sections or genera; but they are either monotypie or have no special area, and all belong to the same general Mediterranean region. Taraxacum, nearly allied to Leontodon, and variously estimated at from four or five to above forty species, has a very wide distribu- tion, accommodating itself to every variety of station (thus account- ing for the intricate variability of its forms) and readily colonizing. The extratropical regions of the northern hemisphere comprise its chief centre; and it may be more universal in the Old World than in North America ; but it appears to be also a true denizen of the far south, both in America and Australia, and is to be met with even in warmer regions. The genera Troximon, sixteen species, Pyrrhopappus, three or four species, Calycoseris, two species, and the monotypic Glyptopleura are all American and almost exclusively north-western, with their chief centre in the Mexican region (including California). Troximon reappears in one or two species in extratropical South America, and Pyrrhopappus extends somewhat eastward in North America. These genera are quite absent from the Old World; but they may in some measure be considered as West-American representatives of Leontodon and Taraxacum. Hypocheris, about thirty species, allied to Leontodon in habit and structure, has a wider general distribution, and a rather more American character. Common to both the New and the Old World, with two species so generally distributed and so readily colonizing as to make it difficult to say where they are most at home, Hypocheris bas perhaps most species in the mountain and temperate regions of America, especially South America, but the most diversified forms in the Mediterranean region of the Old World. Minute differences in the pappus have induced its general DISTRIBUTION OF TRIBES. 479 aivision into three or more very artificial genera. Taking more natural, though not always very well-defined, sections, Achyro- phorus and Serioloides are common to the New and the Old World, Oreophila is exclusively Andine, Amblachenium is a single Siberian species, Euhypocheris and Porcellites are the two. cosmo- politan species, Seriola, Metabasis, Robertia, and Arachnites form a small group exclusively Mediterranean. Malacothrix, about nine species, and the monotypic Anisocoma allied to it are Californian genera, in some respects allied to Hypocheris and perhaps to Hieracium, but forming a very distinct local group. Tragopogon, about thirty species, and Scorzonera, about one hundred, are widely distributed over the temperate and even sub- tropical regions of the northern hemisphere in the Old World, with their chief seat again in the Mediterranean region, to which belong also a few small local genera slightly diverging from them :— Uro- spermum, two species, of which one reappears in South Africa, but probably as a rather old colonist ; Epilasia, five species or varie- ties, from the Persian region; and Zouwrneuxia, one species, from the deserts of South Algeria. None of these genera extends to America; but the group or subtribe may be said to be in some measure represented in North-west Ameriea and the Mexican region by Lygodesmia, five or six species, Stephanomeria, about eight species, Scorzonella, two or three species, and the monotypic Pinaropappus, and in extratropieal South Ameriea by the mono- typie Picrosia. Tolpis, fifteen to eighteen species, belongs to the Mediterranean region, more especially the western portion, and extends to the Canary and Azores Islands, where it has established endemic spe- cies; the monotypic Spanish Hispidella is also allied toit ; and no nearly related form is known from any distant region. Hyoseris, four species, is also peculiar to the Mediterranean region and nearly surrounding districts. Arnoseris, a single spe- cies diverging from Hyoseris, is more generally spread over Europe, especially the western districts, and is a ready colonist in Australia and some other countries. These genera may be said to be in some measure represented in America by Microseris, twelve species, Krigia, four or five species, and the monotypic Phalacroseris, all northern and chiefly north-western, except one species of Microseris, which, from extratropical South America, extends to Australasia. LINN. JOURN.—BOTANY, VOL, XIII. 2M 480 MR. G. BENTHAM ON COMPOSITA. Lapsana, belonging to the extratropical regions of the northern hemisphere in the Old World, has one European species very widely spread, and is found asa colonist in North America as well as in other distant parts of the globe. The genus, however, is truly represented in North America by the nearly allied mono- typie Apogon, a Japanese species of Lapsana being almost inter- mediate between the common one and this Apogon. Cichorium is a very distinct genus, of which one species is widely distributed over the northern hemisphere in the Old World, readily colonizing in many other districts; and a second is limited to the eastern and southern portions of the Mediterranean region. The cultivated Chicory, often given as a third East-Indian species, is probably only a cultivated modification of the common C. Zntybus. The genus is wholly unrepresented by any endemic American form. Catananche, five species, and the monotypic Hymenonema and Henseleria form a very distinct Mediterranean group unknown elsewhere; Henseleria is western, Hymenonema eastern; Cata- nanche is both. Scolymus, three species from the same region, is still more isolated in habit and foliage, which is almost that of a thistle, and in struc- ture, which, except in those invariable characters which place it among Cichoriaces, is unlike that of any other known Composite. We have finally two very remarkable arborescent insular genera; Dendroseris, seven species, from the island of Juan Fer- nandez, and Fitchia, one species, from the Pacifie islands. Both are truly Cichoriaceous in their corollas, anthers, and styles, and Dendroseris, at least, in the milky juice of its bark; but their achenes are different from those of Cichoriacex generally, as well as their involucres and habit; and Fitchia, in its receptacular pales, awned achenes, &c., recalls the Helianthoidee. Having thus rapidly sketched out the principal facts which have struck me in the investigation of the geographical distribu- tion of the genera of Composite, as compared with their struc- tural characters, we may proceed to the inquiry as to how far they can assist us in the solution of the two great problems :— Which, amongst the numerous types or generie forms now exhibited by the order, represent the most ancient races, the nearest to the primitive form of the order? And what are the principal centres where the greater number of the present races appear to have been differentiated, and whence they have spread over the areas DISTRIBUTION OF TRIBES. 481 they now occupy, thus establishing more or less distinct regions ? —these centres of individual races to be carefully distinguished from the supposed centres of creation or centres of vegetation, from whieh whole diversified floras are supposed to have radiated, and to the fallacy of which I specially alluded in my Address of 1869. B. COMPARATIVE ANTIQUITY OF RACES IN CoMPOSIT. If we are justified in observing that races, like individuals, have successive periods of progressive growth, of full vigour, and of gradual decay, we may, in the one case as in the other, conclude generally that those which we find to be in the latter stage are the oldest—a conclusion, however, which, in the one case as in the other, must of course be very much modified by the consideration of the numerous constitutional or external circumstances which bring on premature decay and extinction. A preliminary inquiry, however, is necessary into what con- stitutes, what are the evidences of, progress, vigour, or decay in a genus, species, or other race of plants, all of which may be entirely independent of the evident vigour or decrepitude of the individuals the race is composed of. The result of the best-founded opinions on this subject which to my knowledge have been propounded is that a race of plants, be it tribe or genus or species, in its period of full vigour, is widely dispersed, accommodates itself to a great variety of climatological, physical or other external influences, is numerous and varied in subordinate races as well as individuals, these subordinate races, especially those immediately subordinate, not being separated by wide structural gaps, and not having acquired any very marked local characters, but for the most part passing, as it were, into each other, their respective distinctive characters not having yet acquired any marked degree of correlation. On the other hand, a race in a state of decay is represented by subordinate races very distinct in structural characters, of restricted areas, and requiring for their preservation special climatological or other physical conditions, and consequently comparatively few in in- dividuals. Of the former, Vernonia, Eupatorium, Aster, Senecio, &c.; of the latter, the first five or six subtribes of Helianthoidee appear to be good examples among Composite. Old decaying and apparently expiring races may, however, in some of their branches, owing perhaps to a slight change in 2x2 482 MR, G. BENTHAM ON COMPOSITA. constitution, habit, or external circumstances, start into new life. Young progressive races, which, like the vigorous young indivi- duals which we see rise from the rotten remains of an aged plane, or olive, or fig-tree, may be rising before our eyes from some branch of an old race which has passed its prime, or whose origin may already be so remote as to be concealed from us. These young progressive races will be very prolific, ready colo- nizers; and their subordinate races will be generally numerous and so blended together as to defy all positive determination of their limits, and be variously estimated as subgenera, sec- tions, species, subspecies, or varieties. Most of the Cichoria- ceous genera may perhaps in this respect be considered as less ancient than most other tribes and still in a state of progress. The six Asteroid types above mentioned and the subtribe Gna- phatiew of Inuloidez may perhaps be regarded as races still vigorous, but breaking up into subordinate races of a local cha- racter, many of which already give indications of future diminu- tion and extinction, but some of which, as yet of a very low grade, exhibit a great susceptibility of extension and progress. Some confirmation of the hypothesis that some of the oldest of the primary or tribual and subtribual types of Composite are to be sought for in Helianthoides, and some of the most recent (of those dating from geological periods antecedent to the present one) among Cichoriaees, may perhaps be derived from their struc- ture. The great consolidation and uniform structure of the essen- tial organs of fructifieation of Composite has, as already men- tioned, been adduced as evidence of their comparatively recent origin; and this consolidation and uniformity is least marked in Helianthoides, most so in Cichoriacem. In many Helianthoides we find, for instance, the outer bracts of the involucre more folia- ceous, the bracts subtending the flowers (or receptacular palez) more normally developed and more firmly attached, the calyx- limb (or pappus) less transformed, consisting frequently of per- sistent teeth or ariste directly continuous with the ribs of the T and thus showing their really calycine nature, the anthers in some genera less firmly united and perhaps sometimes quite free; and in the female flowers of the Petrobiew we have an or- dinary campanulate regular corolla with the anthers (although small and sterile) well formed on short filaments alternating with the corolla-lobes and far from each other. In Cichoriacew the uniformity of the organs of fructification is more absolute than in COMPARATIVE ANTIQUITY. 483 any other tribe, neither the pistil nor the andreecium, nor even the corolla affording the slightest structural or sexual distinctive character throughout the eight hundred odd species; the pappus, though more variable, is always amongst those which show the least resemblance to a calyx-limb; and the receptacular palex, in the few cases where present, are the least like ordinary bracts. We may thus, perhaps, be led to conjecture that the primitive form of Composite had regular gamopetalous flowers with an inferior ovary, the calyx, corolla, and uniseriate stamens isomerous and probably 5-merous, and the pistil 2-carpellary as in several Rubiacez and allied orders, but the ovary internally already reduced to a single cell with a single erect ovule, and the seed exalbuminous, enclosed in an indehiscent pericarp, and containing a straight embryo with an inferior radule—and that it is in the gradual course of subsequent consolidations that the bracts have crowded round the condensed flowers and usurped the functions of the calyx-limb, which has become obliterated or transformed so as to be better adapted to its new duties; the corollas have become contracted, or the outer ones variously developed in forms and colours adapted to assist in the process of cross fertilization (vexillary functions of Delpino); the anthers, brought into close contact by the compression of the flowers, have become united and their styles gradually modified so as to assist them in dis- charging their pollen; and the conversion from hermaphroditism to unisexuality may in various races have variously preceded or followed some or all of these changes, and produced those nume- rous variations observed in the order. We might further be led to imagine that several of these changes had taken place at avery early period, previously to the disruption or stoppage of communication between what are now the tropical regions of the globe, that, besides the parent form above supposed, Composite existed showing several important modifications, such as, Ist, the regular and uniform tubular development of the corolla, accompanied by more or less of suppression of the inner bracts and of the normal calyx-limb and substitution of a pappus; 2nd, the reduction of the corolla-limb, attended frequently by a sexual dimorphism, and occasional oblique development of the outer female corollas; and, 3rd, perhaps at a later period, the uniform unilateral development of the whole of the corollas, accompanied usually by a suppression of the inner bracts and conversion of the ealyx-limb into the pappus. From the first.of these modifica- 484 MR. G. BENTHAM ON COMPOSITA. tions would have sprung the Eupatoriacese in America, the Vernoniacez in the New and the Old World, the Cynaroides in the northern and the Mutisiacese in the southern hemisphere. From the second modifieation would have arisen, first, the more slightly altered Helianthoidesw in both the New and the Old World, but chiefly in the former; 2nd, the Helenioidese in America, and Anthemides in the Old World, with the thinly paleaceous modification or total suppression of the inner bracts and calyx-limb ; and, 3rd, the cosmopolitan Asteroidex, Senecio- nidez, and the majority of the Inuloides, with an almost universal ‘suppression of the inner bracts and conversion of the calyx-limb into a setose pappus. The third general modification, with a very few slight exceptions, has settled down into those Cichoriacee whose absolute uniformity has already been observed upon. Some further remarks bearing upon the above points may be elicited in the investigation of the principle present centres or regions of Composite to which I shall now proceed. C. Present REGIONS OR CHIEF CENTRES OR AREAS OF THE Prrvorpan RACES or Composit. The position of the great centres of the order is evidently in some measure influenced by its prevalent constitution and the consequent effects of climatological and other physical causes upon the gradual migrations of its species. Rarely arborescent and gregarious, still more rarely aquatic, Composite are in a great measure excluded from the vast forest-clad lowlands of the Amazon region of America or of east tropical Asia. In the swampy bogs of the northern hemisphere they may not be so rare, but the species are few. Their favourite haunts are treeless or thinly clad mountain-regions, and especially the lower but broken grounds, rocky ridges, or open campos of warm extratropical or subtropical districts. They may be met with, indeed, at the highest altitudes or latitudes which will bear phenogamic vege- tation, as well as in the warmest tropical deserts, anda few species as ready colonists are perfectly ubiquitous in the traces of man; but there are tracts of country, such as the Mediterranean region, South Africa, extratropical America, both Mexican and South- Andine, especially abounding in highly differentiated races of very limited areas, others, again, such as the more temperate or mountain districts of the northern hemisphere, where Composite genera and species are as numerous and ill-defined in their sub- REGIONS AND AREAS OF DISTRIBUTION. 485 ordinate races as wide and vague in their geographical range. These severally constitute the centres of differentiation or areas of preservation, which I shall endeavour to define as Regions of Composite. Besides, however, the difficulty of assigning limits to adjoining regions, owing to the mutual interchange of races across their frontiers, even the most distant regions are sometimes connected by races which, owing sometimes to ready colonization, sometimes, perhaps, to antiquity of origin, are now found to occupy very wide or broken and interrupted areas; and a few races may be said to be truly cosmopolitan, affecting no one region more than another. I propose, therefore, to follow up the distinction of regions by a sketch of their connexions and of such evidences as we may trace of the supposed origin of the connecting races. Conjectures, however, as to the original centres or birth-places of all these widely dispersed and interrupted races must, of course, be very hazardous; for if we still hold to the axiom that affinity means consanguinity, we must suppose some preexisting physical conditions and configurations of the globe very different from the present ones, upon the precise nature of which those geologists who admit them at all, seem to be by no means agreed. Into these supposed conditions there would be no advantage in entering now; I only advert to them for the purpose of explaining that, when I speak of ancient connexions between regions now sepa- rated by impassable barriers, such as tropical America and Africa, Mexico and the Argentine States, South Africa and Australia, &c., I by no means take it as decided whether that connexion was by contemporaneous continuity of land-elevation or climate now broken off, or by successive connexions with some common land now destroyed, or by means of transport now no longer existing, or by any other facilities afforded by ancient conditions of the globe as yet unknown to us. In sketching the principal regions of the globe as marked out more or less distinctly by the different races of Composite which inhabit them, I shall commence with the primary division into the New and the Old World, and then detail the principal regions in each of these great divisions. I should observe, how- ever, that in using the terms New and Old World rather than those of Western and Eastern continents adopted by Grisebach, it is merely because the former appear to me to be more familiar and more readily understood. Neither term is strictly correct ; for it is not intended by the words New and Old to indicate any 486 MR. G. BENTHAM ON COMPOSIT X. comparative antiquity of existence (a geological question, upon whieh I have no right to form any opinion), nor yet the novelty or antiquity of our knowledge of them ; for Australia, the most recent of our important discoveries, must in phytogeography be included in the Old World ; and the terms Western and Eastern Conti- nents, as applied by inhabitants of Western Europe or Eastern America, must be reversed by the inhabitants of Eastern Asia or Western America. I must observe also that in the Tables given in the following pages the numbers of genera and especially of species must never be taken as absolute; they are at best approxi- mative only, and in some instances may be purely conjectural : they are, however, the best I have been able to arrive at without a careful working-out of the whole of the species known, which would be too many years’ labour for me to undertake. Further discoveries would likewise require considerable modifications * ; and to those who do not agree with me as to the cireumseription of genera and species, the absolute numbers might be very different. 1 have endeavoured, however, to keep as much as possible to a uniform standard in this respect; aud if the same course be adopted by those who inultiply or reduce distinetions, the comparative results will probably remain nearly the same. 1. General Repartition of Composite between the New and the Old World. In these Tables are included, in the American or New- World division, the West-Indian Islands, and in the Old- World division the Eastern Archipelago and Australia. The Sandwich and South- Sea Islands, the Galapagos, Juan Fernandez, St. Helena, the Atlantic Islands, the Mascarene group, and New Zealand, not- withstanding the American character of the Composite of the first groups, and the Old-World connexions of those of the last three, are here omitted ; for their endemic races affect very little the general repartition between the two great divisions of the globe, and their geographical peculiarities appear to require con- sideration under distinct heads. The numbers given, both of genera and species, are intended to apply to natives only, or races which may have been anciently established without the interven- tion of man, to the exclusion of modern colonists. * Two or three new genera and a few new species received since this paper was placed in the printers’ hands, would already require some slight changes in a few of the figures of some of the following Tables. REGIONS AND AREAS OF DISTRIBUTION. 487 Table 1. Repartition of Genera in the New and Old Worlds. | | | All chiefly American. 1 chiefly American, 7 chiefly Old-World. a 4 almost entirely American. b 1 almost entirely American, and 1 chiefly Old-World. c Nearly all chiefly Old-World. d Most of them chiefly American. € All almost entirely American. f 1 chiefly American, 7 chiefly Old-World. g h America. Old World. ose aa bd Tribes. mon to | Total. Taslak. Total Pode | Total Endes Total. the two. Vernoniacer...... 25 | 29 10 14 4 39 1 40 Eupatoriacex 30 35 $e 5 5a | 35 E 35 Asteroidex ...... 39 49 34 44 105 | 83 8 91 Inuloidez ......... 16 97 | 107 | 118 lle | 134 8 | 142 Helianthoidez ... 107 | 125 4 22 18d | 129 9 | 138 Helenioidez ...... 57 59 3l 3 2e 60 a 60 Anthemidex ...... 3 12 32 41 Of | 44 1 45 Senecionidese 14 19 19 24 5 38 6 44 Calendulaceze 1 1 fi 7 8 fie 8 Arctotider ...... Vi A 17 17 17 ; 17 Cynaroidex ......| ... 3 34 37 3 37 37 Mutisiacem ...... | 37 38 14 15 lg | 52 1 53 Cichoriacee ...... 14 | 24 30 40 104 | 54 2 56 Total. 343 | 421 309 | 387 78 | 730 96 | 766 Table 2. Repartition of Species in the New and Old Worlds. The numbers in the following Table, especially as to the larger genera, are often but roughly estimated, and may on a elose scrutiny require in some instances considerable modification. The common species do not include the weeds of cultivation intro- duced in modern times from one division to the other. | America Old Word. | ub ` z Add | Grand Tribes. mon to| Total. | Tasular.! Total. Ende- | Toal MS | Total ME Vernoniacese ...... 319. | 376 | 1021106 4 528 | 12 540 Eupatoriacem . 740 743 13 16 3 756 1 OT Asteroider ...... 822 | 830 | 434 | 442 8 1264 | 94 | 1358 Inuloidez ......... 146 157 939 | 950 11 1096 | 93 1189 Helianthoidem ... 953 963 67 TE 10 1030 47 1077 Helenioidez ...... 304 | 304 8 9. ou 307 1 308 Anthemidee ...... 48 63 593 608 | 15 656 28 684 Senecionidez 497 | 502 | 682 | 687 | 6 1184 | 50 | 1234 Calendulaceæ 1 1 102. [1102 | —.. 103 3 106 Arctotidez ...... ix m 237 | 237) ... 29-31 — 237 Cynaroider ...... 41 42 971 972 |. 1 1013 13 1026 Mutisiacezm ...... 383 | 383 60 ert eee 443 1 444 Cichoriacex ...... 156 161 605 610 5 76 48 809 [-.- poo. 4463 | 4525 | 4858 | 4920 | 63 | 9383 | 386 |9769 488 MR. G. BENTHAM ON COMPOSITAE. Upon the whole it would appear from the above Tables that Composite are not unfairly distributed between the New and the Old World as well as to numbers as to variety of forms, although with a balance in both respects rather in favour of America, the numbers being 4525 species in 421 genera in the New World, against 4920 species in 387 genera in the Old. Further discoveries may also, probably, inerease this disparity ; for there are many tracts in the great mountain-chain extending from California to Chili, so rich in Composite over the whole of its vast extent, which are as yet but little known or wholly unexplored, whilst in the two richest Composite regions in the Old World, the Mediterrrnean and South- African, the Composite forms as yet unknown must be comparatively few. From tropical Africa we may expect rather more, especially as to generic forms ; but these may be compensated by fresh discoveries from the interior of Brazil and La Plata. Were the insular floras included in the enumeration, the disparity would again be slightly diminished ; for the imperfectly known Mascarene and the well-explored Atlantie Islands would add 18 genera and species to the Old World, whilst the Sandwich and Galapago Islands, now pretty fairly investigated, only add 8 genera and 70 species to the New. St. Helena, New Zealand, and the South-Sea Islands are wholly excluded from these caleulations, as not being specially referable to either of the two great divisions. With regard to the comparative diversity and distinctness of forms in the two divisions, it may be observed that the number of species to a genus is about 107 in America, and 1277 in the Old World, showing in the former more numerous remnants of old types, in the latter a greater luxuriance of flourishing and in- creasing genera. It will be seen, however, that in each division there are regions remarkably characterized in both respects. ‘When we come to consider the tribes into which Composite have been divided, we at once see a great disparity in their repar- tition between the two great divisions. Two tribes are almost en- tirely American, the Eupatoriacee and Helenioidez. In the former 30 genera, including above 200 species, are exclusively American, 3 genera, containing together above 80 species, have each one of their species extending into the Old World. Eupatorium itself, of above 400 American species, is represented in the Old World by about 10 species; and the small genus Adenostyles alone has an Old- World preponderance, having 2 European and 1 Californian REGIONS AND AREAS OF DISTRIBUTION. 489 species. Helenioidewm have 57 genera, containing nearly 300 species, exclusively American; one genus, Flaveria, of 7 American species, has one of them extending into Australia, another, of four American species, is represented by 1 in South Africa, and a third, Cadiscus, is a monotypic South-African plant. Three more tribes have a strong American Propondoanpes both in their firey and in their subtribes : RRAK bees = ? genera and * pus species z E^ * species in the Old World; 125 963 Helianthoidee ¡3% bcd genera and j;; Species, against B5 Old- World deum and i7. 105 L genera and Y == 3 species, against =, 15 O1d-World genera and 9 qg Species. Three more large tribes are more ed distributed : Asteroides American, against n genera and 1 k S species; and Mutisiacez 5 -; American LI . 44 have $ E ? New-World genera and È Ed species, against ;; genera and d species in the Old World, showing still an American a bas: is slightly diee in des with as genera and =~ species, against 2 * genera and $ EI 7 Species in the ils Old World, "d still more so in cag edet with 5 American genera and 15 265 species, against 5; ° genera and 255 E * species in the Old World. The remaining tribes are much more decidedly charac- teristic of the Old World. Of Inuloidee the largest, the nine subtribes are all of the Old World, three only of them being A dieat h America in companya few numbers ; p the genera are ij, 3; and the species on in the Old ur to E and 157 ; 1096 In the New. a bares zi genera and $5 oe Species in the Old World, to E 1 genera and 2 E 5 T in the New. Cyna- roideæ, with all the 37 genera and“ species in the Old World, are represented in America by only s species, belonging to 3 of those genera. Calendulacew with 7 of its 8 genera, including 102 species, in the Old World, has a single monotypie American genus; aud the 17 genera and 237 species of Arctotidece are exclusively of the Old World. Tosum up in a few words, Composite as a whole are not very unequally distributed between the two divisions of the globe; and of the 13 tribes, 12 are common to the two, and only one restricted to one of them. But out of 730 genera 78 only are common to the two, showing already a far greater difference 490 MR. G. BENTHAM ON COMPOSIT X. in the character of the Composite of the two divisions than can be attributed to any climatological or other physical causes ; and the difference is still more striking in the species, of which only 63 out of a total number of nearly $400 are common to the New and the Old World; and these common ones are chiefly either Alpine or high northern, where the general flora is more continu- ous, or ready colonizers, although their presence in the two divi- sions cannot be attributed to recent colonization. These genera and species common to the two divisions require separate consideration, according as they are tropical, northern, or southern. Table 3. Tropical connexion between America and the Old World as indicated by identical or closely allied Genera and Sections without identical Species. Genera. Tropical-American representatives. Tropical Old-World representatives. CENTRATHERUM Connexions VERNONIA. Representative Sections. Diverging sections Connexions ERIGERON Conyza. Identical groups Diverging groups Connexions PLUCHEA. Representative 2 widely spread species, extending southward of the tropics, and one of them reappearing in Australia. Oiospermum, a closely allied mono- type, otherwise those of Vernonia. Hololepis 2 species. Ef ve Sy about 200 species, a few of which extend beyond the tropics both north and south. Critoniopsis 6 species. hal, Leiboldia, St , Trianthea, and Eremosis, all small Columbian or Mexican sections. Ha yy 24 species, Piptolepis 8 Sp. and Albertinia, Vanillosmopsis, Blanchetia, Lachnorhiza, and Bola- nosa, all monotypic, immediately connected with Vernonia; and 17 genera with above 80 species rather more remote. See northern connexions, Table 5, Dimorphanthes, Cass., or genuine Co- nyze, represented by C. chilensis and ies. Lennecia ( C. gnaphalioides and allies), C. triplinervia and allies, Erigeron (Cenotus) more northern, and in the tropics Nidorella and Psiadia, on the one hand, Laggera, Blumea, Pluchea, and smaller ones on the other. P. purpurascens and allies. 4 Asiatic species, differing perhaps sectionally from the American ones. None nearer than Vernonia. V. calycina, Wall, in Asia, and V. purpurea, Sch. Bip., in Africa. | Gymnanthemum about 20 species, Asiatie and African, and perhaps a few Asiatic true Lepidaploe. — Strobocalyx about 20 species, Asiatic E Qe : ipholepis about 12 species, Cyanopts about "0 eet Pioda cbout 20 Species, all with the majority Asiatie, but also several African | species. Lepidella 15 species and Stengelia 14 species, exclusively ican, or nearly so. Bothriocline in Africa, Adenoon and Lamprospermum in Asia, Pleuro- carpea in Australia, and Centaurop- sis in Madagascar, all — and immediately connected, and 4 genera with about 9 species rather more remote. Dimorphanthes, represented by C. egyptiaca and allies. Marginate (C. Gouani and allies), Fimbrillaria ( C. ivefolia and allies). Erigeron (Cenotus), more northern ; and in the tropics Baecharis on the one hand, and Pluchea and allies on the other. P. tomentosa and allies. REGIONS AND AREAS OF DISTRIBUTION. 491 Genera. Tropieal-Ameriean representatives. | Tropical Old-World representatives. PLUCHEA. Diverging groups P. Quitoc and allies, several of them | P. indica, Berthelotia, DC., P. pinna- extratropical. tifida, Hook. f., and allies; the spe- cies more Asiatic than African, and a few more divergent extratropical : Asiatic or Australian species. Connexions ......| Stenachenium 3 species (Brazil), | Laggera 10 species, Blumea 55 spe- ACHYROCLINE ... Connexions ...... GNAPHALIUM ... STENOCLINE ...... Connexions ...... GRANGEINEZ. Representative genera ......... Connexions MELAMPODIUM... Connexions CHRYSOGONUM ... Connexions ENHYDBRA rsi: Connexions SCLEROCARPUS ... Connexions BLAINVILLEA...... Connexions WEDELIA. Seet. Stemmodon Sect. Cyathophora Sect. Wollastonia Connexions ...... MELANTHERA. Sect. Wurm- echmidtia or Echinocephalum. Sect. Eumelan- thera, Sect. Lipotriche Connexions .. . Connexions ...... Sachsia 3 species, and Rhodogeron l species (Cuba), Tessaria, 5 species (Western). 10 species, Brazilian and Western. Gnaphalium and Stenocline. See cosmopolitan genera, Table 8. 2 species, Brazilian. Gnaphalium and Achyrocline. Egletes 6 species. Aphanostephus 3 species, and, more istant, several north-western Aste- roidee. 17 e (tropical or north subtropi- cal). Acanthospermum 2 species (tropical), Lecocarpus 1 species (Galapagos). 1 species, North-American, subtro- pical, Silphium 11 species (North-Ameri- can), Berlandiera, Lindheimera, and Engelmannia 7 species (Mexico and Texas), and Schizoptera 1 species (Andine). 3 or 4 species, aquatic and tropical. Aphanactis 2 species, Andine. 9 species, tropical or north subtro- pical. Montanoa 14 species, Mexican and Andine. lspecies cosectional with the African, and 2 or 3 species sectionally differ- ent. Wedelia, Aspilia, &c. 4 species, mostly maritime. : About 25 species, Brazilian, Andine, &e. None. Besides Blainvillea and Aspilia, Zex- menia 20 species, Oyedea 22 species, | About 30 s ecies, 2 or 3 cosectional with 5 African; the others diver- ent. ‘Besides Wedelia and Blainvillea, Zexmenia, Oyedea, &c., as above. 1 species, Brazil. 4 species, Brazil, Indian Ocean, &c. None. None very near, but less distant than in Old World. About 20 species, tropical, None very near, but Sabazia nearer than any Old-World. cies, and, in another direction, Conyza. 6 species, all African. Gnaphalium, Stenocline, and others. 6 species, Mascarene. : Helichrysum, Gnaphalium, Achyro- cline, and others. Grangea 2 species. Ceruana 1 species and several small tropical genera and, more distant, several north-eastern Asteroides, and, in another direction, Cotulee. 1 species (Philippines) and a colonist from America. None near. 1 species, tropical Australian, closely representing the North-American one; 2 species Australian and 3 species Asiatic, all tropical and somewhat divergent. None near. 1 species, African and Asiatic. None near. l species, African, closely represen- tative of one of the American. None near. 3 or 4 species African, 1 also Asiatic, all cosectional with 1 American. Wedelia and Aspilia. l1 species, East-Asiatic. 1 species African, 1 species Asiatic. About 10 species, Asiatic and Austra- lian. None but Blainvillea and Aspilia. 8 species, of which 5 African cosec- tional with 2 or 3 American, 2 Afri- can and 1 Mascarene divergent. None but Wedelia and Blainvillea, 1 species, African. None. 1 species, African. None near. 20r 3 species, Asiatic, Australian, and chiefly East-African; all tropical. None. 492 MR. G. BENTHAM ON COMPOSIT E. Genera. Tropical-American representatives. | Tropical Old-World representatives. COREOPSIS ......... 3 or 4 Peruvian species near the | About 8 species, East-African ( Presti- African, and nearly 40 diverging, | naria), near the Peruvian. chiefly North-American eat Mexican. Connexions ...... Bidens (Psilocarpea) nearly 40 spe- | Bidens (Psilocarpea) 2 species (colo- cies, Cosmos 10 species, Dahlia 4, | nists?), Glossogyne 5 species (East- Hidalgoa 2, Isostigma 5, Thelesperma | Asiatic and Australian), Guizotia 3, 4 or > species, all tropical or subtro- | and Microlecane 1 (African). pical. : CHRYSANTHEL- |1 species tropical-American, 1 Gala- | 1 species, African and Asiatic, pro- LUM. pagos. bably the same as the American. Connexions ...... Heterospermum 1 species. Glossocardia 1 species, Asiatic. A first rapid glance over the above Table shows the general American character of the whole. For the most part the African and Asiatic species sections or genera are few and disconnected, the corresponding American ones numerous and closely connected on all sides with American allies. And yet the endemic Old- World races from species to genera are too numerous and varied to admit of the supposition that they can ever have migrated from America and become extinct in their birth-place. It would seem rather that whatever may have been the cause of the parent Wedelie, Sclerocarpi, Melanthere, &c. having once been esta- blished both in Africa and in America, or in some land at one or different times in connexion with the two continents, they had by long isolation become more and more differentiated in the two— that in America they have as races prospered and multiplied in every direction and possibly retained many of their very early forms ; whilst in the Old World they have found less genial cir- cumstances, they have for the most part dwindled away, a far greater proportion than in America have become extinct, and the few local representatives we now see are probably in the course of extinction. And this will, I believe, be found to be more particu- larly the case with the African races, Here, more perhaps than in any other part of the globe, in Composite as in so many other orders, we may fancy we see the scattered remains of ancient races dwindling down to their last representatives. It is not so, however, with a few of the races included in the above Table, especially some of those which have rather more of an Asiatic than an African character. The sections Tephrodes, Cyanopis, and Gymnanthemum of Vernonia, the section Wollas- tonia of Wedelia, the genera Ethulia and Chrysanthellum may be flourishing and increasing races, which have already been much differentiated in the Old World and are likely to become more so; the two last-named, as well as Vernonia (Tephrodes) conyzoides, are REGIONS AND AREAS OF DISTRIBUTION. 493 ready colonizers, although neither the Ethulia nor the V. conyzoides has as yet appeared in America. The African Vernonie (section Stengelia) and Coreopsides (section Prestinaria) and the eastern Chrysogona and Glossogynes, although showing each several species, have probably already seen their best days. Table 4. Tropical connexion between the New and the Old World as indicated by identical Species. This Table includes the species of those genera only which have an essentially tropical character, although some may, perhaps, extend northwards or southwards beyond the tropics, excluding the common species as well of the cosmopolitan genera enumerated in Table 8 as of the northern or southern connecting genera contained in Tables 5, 6, and 7, which may also be found within the tropics, as in both cases the original connexion or communi- cation was probably extratropical. Species. Area. Congeners. Connexions. Sparganophorus Vaillantii East tropical America, Elephantopus scaber ...... Adenostemma viscosum ... Ageratum conyzoides sssr.. Mikania scandens Epaltes brasiliensis Ambrosia maritima ......... Xanthium strumarium Siegesbeckia orientalis...... Eclipta alba, Hassk. (E. erecta, E. prostrata, &e.). West tropical Africa, (semiaquatic). America, Asia, and Africa, tropical and subtropical. Cosmopolitan, tropical and subtropical (ready colonist). Cosmopolitan, tropical and subtropical (ready colonist). America, Asia, and Africa, tropical and subtropical. East tropical America, West tropical Africa (semiaquatic ?). America, Asia, Africa, and South Europe, tropical and subtro- pical (maritime). and South Europe, tropical and subtro- les pem America, Afriea, tropieal and subtropical (ready colonist). America, Asia, Africa, and Australia, tropi- eal and subtropical (ready colonist). | None. .| America, Asia, Africa, . 8 American and 1 African species. 4 American species. 15 American species. 60 American species. 1 American cosec- tional species, 6 African, Asiatic, or Australian, tropi- cal or subtropical, forming separate sections. 11 American species. 2 or 3 American spe- cies (all ready colo- nists). 1 American species. 1 cosectional species, East - Asiatic and Australian, 1 or 2 American, forming a separate section. (Paceurina) American. None immediate, but the nearest American. All American. All American. All American. Chiefly African. All American. All American. All American. All American. 494 MR. G. BENTHAM ON COMPOSITE. Species. Area. Congeners. Connexions. Bidens pilosa and B. bi- pinnata. Synedrella nodiflora........ Chrysanthellum.............. Enhydra (several names, probably one species). Cosmopolitan, tropical | and subtropical (ready colonists). America, Asia, and Africa, tropical (ready colonist). America, Asia, Africa, tropical. and America, Asia, Africa, and Australia, tropi- cal (semiaquatic). | 1 tropical-American About 40 cosectional Chiefly American, with & species, all tropical | few African. American; a distinct section,amphigeous and extratropical. species. 1 tropical-American (Galapagian). 2 or 3 tropieal-Ame- rican species. Chiefly American, with a few African, besides the following. ltropical-Asiatic mono- type (Glossocardia) and I tropica] - American (Heterospermum), _the further ones chiefly American. None immediate; nearest American. the rio, Cotula coronopifolia.........| Europe, South Africa, | About 35 Old-World | A few small allied genera Australia, extratropi- | species, chiefly chiefly South-African, cal South America. South - African, 2| Australian, New Zea- > South-American. land, and extratropical or Andine, 1 Ameri- can. Centipeda orbicularis(My-| South Africa, Aus- | 1 Australian (and 1 | Same as Cotula. e minuta and M. tralia, New Zealand, elatinoides). extratropical South erica, African ?) species. We have here, as in the first list, a marked American prepon- derance; for, with the exception of Chrysanthellum, Cotula, and Centipeda, the connexions are exclusively, or almost exclusively, American. It is possible, indeed, that the Ageratum, the Sieges- beckia, the two Bidens, and even the Synedrella may be of comparatively modern introduction, and may belong therefore rather to the class of species of which the interchange between various regions is now going on. The Ambrosia also as a maritime plant may have been brought over at any time. The Spargano- phorus, the Enhydra, and perhaps the Epaltes belong to the class of semiaquatic plants whose wide diffusion and ready dispersion to great distances have been frequently observed, although, perhaps, not yet satisfactorily accounted for. Xanthium spinosum, long known in the Old World, but which has only very recently taken possession of Australia, is here omitted, as being believed by many to be only a modern colonist in the Old World. But the .Elephantopus, Adenostemma, Mikania, Eclipta, Xanthium strumarium, and Chrysanthellum, besides the above-mentioned Enhydra and Epaltes, give strong presumptive evidence of a prehistoric establishment in the Old World, not, perhaps, ancient enough to have settled down into distinct species, but having most of them already produced more or less marked varieties, which may be considered as incipient species, to be further REGIONS AND AREAS OF DISTRIBUTION. 495 differentiated should any change in physical conditions tend to isolate them. It will be observed that these American species supposed to be prehistorically established in the Old World are in their present distribution mostly rather of an Asiatic and chiefly Eastern than of an African character; and this, taken in conjunction with the Wedelia (Stemmodon), Melampodium, and some other E.-Asiatic types of the first list, and with many instances that might be taken from other orders, might induce a belief that, as far as plants are concerned, the connexion or communication (whatever may have been its nature) between America and east tropical Asia was of an antiquity less remote than that between tropical America and Africa. Table 5. Extratropical Northern connexion between America and the Old World as indicated by identical or nearly allied Species, Sections, or Genera. This Table includes all the genera whose connecting sections or. species are northern and extratropical, although some of them may extend into the tropics in one or both divisions of the globe, but generally with a greater divergence in character as well as in geographical position than in the north. Some may reappear in the south, and even may there cross again from the New to the Old World (e. g. Centaurea), and may therefore be repeated in Table 7, although the primary connexion was probably northern. Genera. | American races. Common races. | Old-World races. About 24 North-Ameri- ean species, but about 400 more dispersed over Central and South EUPATORIUM. or 3 extending west- ward to the Mediterra- nean region and Europe generally, and lin east tropical Africa; all North-American ones | nearer to the Old- | World forms than to | America. the mass of American | None; but some of the 7 or 8 Asiatic species, 2 | ones. i | elosely allied to each | other and to some | North-American 5 | forms. Connexions ...| Several American genera |... eene ! None very near. ADENOSTYLES | 1 Californian species. None; but the American | 2 European species. species nearer to one of | i the European than | : | they are to each other. | Connexions | None very near; bub|.—————————— ar | None nearer than Eupa- Brickellia perhaps | torium, a more remote nearer than Eupato- | one with some Senecio- | rium. | nidege. SOLIDAGO .... Nearly 80 species, chiefly | S. virga-aurea, L. | None besidesthe common North-American ; a few one and a few colonists. West-American or : | southern extratropical. | Connexions ...| Several North or West |.......... cen None in the north; a LINN. JOURN.— BOTANY, VOL. XIII. | American or south ex- | tratropical genera. few extratropical | South-African genera. 2N | | | | | | | 496 MR. G. BENTHAM ON COMPOSITE. Genera, American races. Common races. Old-World races. BELLIS.......... m Connexions .. BOLTONIA Connexions BOTER SL uerus. Connexions ERIGERON .... Connexions . BRACHYACTIS. Connexions ... FILAGO group. Connexions GNAPHALIUM. Connexions ... ANTENNARIA ... Connexions ... 2 North-American (Southern States) spe- cies. None very near. 7 North-American or Mexican species. Townsendia and others, connecting with Aster. Nearly 100 species, be- longing to the common sections in North Ame- rica and the Mexican region. Several south- ern species of divergent sections. Numerous extratropical or Alpine, both north and south. Nearly 40 species, belong- ing to the common sec- tion in North or West America, Several di- verging sections in North and South Ame- rica. Extratropical ; Aster and allies tropical. Bee Conyza, Table 3. None besides the com- mon species. Erigeron and Conyza, both very close. Filago 1 Californian spe- cies cosectional with the Old-World species, the common species in Chili; Evax 1 species, Micropus 4 species, of divergent sections. Psilocarpus, ^ Diaperia, and Micropsis (12 spe- cies), all very close and North-American, chiefly western. Gna- halium more remote. umerous species,north- ern and southern, chiefly extratropical or mountain, but also tro- pe The following three |.... genera and a few others, chieflySouth-American, and not numerous in species. 8 North-American spe- cies, including the 3 common species and 1 southern extratropical. The two following ge- nera and,through them, Gnaphalium. No common species, wessascassossqesecroovosetosssoseeoe No common species. rr rn rro rr rr reso 4common sections— Al- pigenia, Euaster, Ortho- meris, and Galatella; 1 species, A.(Alpigenia) alpinus. III $esesssssosoostosossone E. alpinus, L.; E. uniflo- rus, L.; E. glabratus, Hoppe (E. acris, L.?); all in section Trimor- hea, besides colonists 1n other sections. f9issssootesososoesttsceeosesesesons B. ciliata, Ledeb. Filago gallica, L. (colo- nist?); the Californian Filago closely represen- tative. a9ssssasotosotasesetosseeostascasonn G. luteo-album, L.; G. purpureum, L.; G. uligi- nosum, L.; G. sylvati- cum, L.; G. supinum, Vill. ; Osrsssesosesessossessosseseossos A. dioica, Gærtn.; A. al- pina, Gærtn.; A. car- pathica, Br. 6 Mediterranean-region species. Bellium in the Mediter- ranean region, Brachy- come and Steirodiscus 1n Australia and South rica. 5 East-Asiatic species, not strictly cosectional with the American. Heteropappus and others, connecting with Aster. About 40 species, belong- ging to the common sections; all extratropi- cal, chiefly Asiatic, few European, and 1 South-east African. Few north, but numerous south; all extratro- pical. About 8 or 9 species, be- longing to the common section, chiefly north- ern. Several diverging sections, northern, tro- pical, or southern. Extratropical ; Aster and allies tropical. See Conyza, Table 3. 6 Asiatic species. Erigeron and Conyza, both very close. Filago 7 species, Evax 9, a Mieropus 3 species, all Mediterranean or Europxo-Asiatic. Ifloga 8 species, Medi- terranean and South- African, nearly con- nected. Gnaphalium more remote. Numerous species, north and south, chiefly ex- tratropical ormountain, but also tropical. Helichrysum, with seve- ral northern and very numerous southern species, and several other southern genera, with numerous species, very close. The fol- lowing three northern genera and seve others, chiefly southern, rather more distant. species, Europeo- Asiatic, including the 3 common species. 5 The two following genera and, through them, Gnaphalium. REGIONS AND AREAS OF DISTRIBUTION. 497 Old-World races. Genera, American races. Common races. LEONTOPODIUM| 2 Andine species. No common species. 3 species, Europso- : siatic. Connexions ... et and Gnapha- |... ees eseeereei iius Antennaria and Gaapha- ium. lium. ANAPHALIS...... None besides the com- | A. margaritacea (Gna- | About 24 species, all mon species. phaliwn, Linn.). Asiatic, northern or tro- pical, chiefly mountain- ous, besides the com- mon species, which is Norta ean Asiatic and j ^ perhaps European. Connexions —| GnapAaiumn. Pie ereet eee daa AA. d ADENOCAULON.| 1 North-west American, | No common species; but | 1 or 2 species, Japan and Connexions ... ACHILLEA Connexions ... CHRYSANTHE- MUM, Connexions ... MATRICARIA ... Connexions ... TANACETUM ... Connexions ... ARTEMISIA Connexions ... PETASITES Connexions ... ARNOR LE Suo 2 Chilian species, Milleriess generally, but none very close. 3 North-American spe- cies, including the com- mon species. one, None besides the com- mon species. None besides Matricaria, None besides the com- mon species. None besides Chrysan- themum. 5 North-American spe- cies. Artemisia only. About 30 N.-American species, including the 9 common species, and 2 or 3 south extratropi- eal or Andine species. Tanacetum and a distant one with Ambrosiez. 4 North-American spe- cies. ! None very near; Luina 1 species, and Peuce- hyllum 1 species, both Noh rest American, and rather distant. 9 North-American spe- cies, including the com- mon species. the North-west Ameri- can and the Asiatic species closely repre- sentative. C. leucanthemum, L. (2 or 3 varieties or species). D qeossossossotessosotosoetos M. inodora, L., and M. discoidea, DC. A. borealis, Pall. ; A. dra- cunculoides, Pursh ? (=A. dracunculus ?); A. vulgaris, L.; A. glo- merata, Ledeb. ; A. glo- bularia, Cham.; A. arc- fica, Less.; A. frigida, Willd.; A. absinthium?, L. A. angustifolia, Vahl? (A. montana, var. ?). Himalaya. None. About 80 Europæo- Asiatic and Mediterra- nean species. Santolina, Anthemis, &c., numerous species. About 100 species, Euro- pæo-Asiatic and a few South-African ; several cosectional, the others divergent. Matricaria and several other Anthemidee. About 20 species, above half Mediterranean and Europso-Asiatie, the remainder South-Afri- can. Chrysanthemum, on the one hand, and, through Nunanthea, Cotula on the other, numerous species. About 30 species, Medi- terranean and Euro- peo-Asiatic, but chiefly Asiatie, On the one hand Arte- misia, through Crosso- stephium, on the other several South-African enera. About 120 species, chiefly Europso-Asiatic and Mediterranean, and most abundant in Asia. A few South-African or tropical species. . Tanacetum,through Cros- sostephium. 8 Europeeo-Asiatic spe- cies. Tussilago, Homogyne, and Cremanthodium, toge- ther 9 species, all very near and Europeo- Asiatic. y The common species in North Asia, and the typical A. montana, Europe and Asia. 252 | | 498 MR. G. BENTHAM ON COMPOSITAE. Genera. American races. Common races. Old-World races. ARNICA. Connexions ...| None close; Crocidium, l |. Doronicum, 10 species, species, rather distant. Mediterranean and Europeeo-Asiatic. BENECIO ......... Very numerous species, S. palustris, Hook.; S. Very numerous species, north and south, igidus, Less.; S. rese- | north and south. difolius, L.; S. pseudo- arnica, Less. Connexions ...| Culcitium, Gynoxys, and |. HH Notonia, Emilia, Gynura, other divergent genera. Cineraria, and other » divergent genera. WERNERIA ...... 16 Andine species. No common species. 1 Himalayan, 1 Abyssi- nian species. Connexions ...| Senecio, very neam, == AE Senecio, very near. : ONIGUS. o... 35 North-American and | No common species be- | About 130 species, Eu- Mexican species, sides colonists. ropeo-Asiatic, Medi- terranean, with a few tropical species. Connexions ...| None very near; the |HMMHMs Carduus, 60 species, very two followin; enera close, and many others the only apo, A slightly divergent. | SAUSSUREA...... None besides the com- | S. alpina, DC About 60 species, chiefly mon species. Asiatic, a few Euro- pean. Oonnenrons -. None near. = -PPa aeres ssori ar SS Serratula, Jurinea, and several others, nume- rous in species. CENTAUREA ...|1 North-American, 5 | No common species, but | Above 300 species, Eu- Chilian species. the erican ost | ropzo-Asiatic and Me- cosectional with a few | diterranean, 6 tropical Mediterranean, tropi- | and 1 Australian spe- cal African, and the | cies. Australian species. HYOSERIS group! Microseris, Krigia, and | No common genera. Hyoseris and Arnoseris, Phalacroseris, 16 spe- 5 Mediterranean and cies,chiefly North-west Europzo-Asiatie spe- American. cies. Connexions ...| Cichoriacese generally. A 00 Cichoriacee generally. LAPSANA group.| Apogon 1 North-Ameri- | No common genera. Lapsana, 4 species, 1 can species. orth-east Asiatic (the nearest to Apogon), the others Europseo-Asiatic : and Mediterranean. Connexions ...| Cichoriaces generally. |.................... eese Cichoriacese generally. CREPIB ............ 6 species, North-Ame- | C. nana, Richards.; C.bi- | About 90 species, Euro- rican; all Eucrepis,in- | ennis, Linn, pzeo-Asiatic and Medi- cluding the 2 common terranean, 1 South-Af- species, rican, and 1 Australian. Eucrepis and several di- E verging sections. Connexions ...| None nearer than Hie- |... ees Phecasium and a few racium, other small nearly allied genera. Hiera- | cium rather more dis- | > tant. PORE aos | None besides the com- | P. hieracioides, Linn. About 28 species, Euro- mon species, pæo-Asiatic and Medi- terranean, including the common species, which is almost cos- x mopolitan. Connexions ...| None nearer than Crepis. |........... esee None nearer than Crepis. HIERACIUM ....| About 25 northern and | Sections Archhieracium | Above 100 species, many 15 southern species, | and Stenotheca. H.mu- | Archhieracium, few belonging to the com- | rorum, Linn.? Stenotheca besides mon sections; mostly Pilosella, which is not | extratropical or moun- American. | tain, and chiefly Steno- | | theca. Connexions ... Malacothrir, 9 Andryala, 6 Mediterra- | | Anisocoma, 1 species ; neither very close. nean species, very close. | | } REGIONS AND AREAS OF DISTRIBUTION. 499 Genera. American races. Common races. | Old-World races. | HYPOCH(ERIS...| About 25 species, South | Sections Achyrophorus | About 10 species, Medi- American, chiefly An- | and Serioloides, but no | terranean and Euro- dine or extratropical. common species except | peeo-Asiatic, including H. glabra, L., and H. | the common sections radicata, L., probably | and small endemic sec- colonists in most sta- | tions. tions. Connexions ...| Leontodon the nearest. |... eene Leontodon the nearest. LEONTODON ...| Section Apargidium, 1 | No common section or | Near 40 species of sec- species. species. tions distinct from the American. Connexions ...| Troximon, 16 species |... None nearer than Hypo- North-American, with cheris and Taraxacum. 1 South-American spe- cies, besides Hypoche- ris and Taraxacum. TARAXACUM ...| None besides the com- | T. officinale (often colo- | About 5 species besides mon species. nist ?). the common one, Eu- ropso-Asiatic and Me- diterranean, the com- mon one cosmopolitan. Connexions | Leontodon. TE bon Paen, aega — and Chon- T . LACTUCA...:..... 8 species North-Ame- | Sections Cicerbita and | About 60 species, chiefly rican, of the two com- | Mulgedium, no com- Europzso-Asiatie and mon sections, and 1 or mon species. Mediterranean; the 2 tropical-American. common sectionschiefly Europso-Asiatie, the Mediterranean often divergent, and a few tropical species. Connexions ... None nearer than Pre- |... sse Chondrilla and some nanthes, small genera, Europæo- Asiatic and Mediterra- nean. Prenanthes rather further. PRENANTHES ...| 10 North-American spe- | No common section or About 6 Europeo-Asiatic cies of a distinct (?) species. and Mediterranean spe- section. cies, near to but distinct from the American section. Connexions ...| Lactuca and Sonchus, not |............. eee Lactuca and Sonchus, not very near. very near. |. SONCHUS......... None besides the com- | S. oleraceus, L. (incl. S. | About 24 species, Euro- mon species. asper, Hoffm.), frequent | peeo-Asiatic and Medi- colonist. | terranean, including the common (cosmopo- | litan) species. Connexions ...| None nearer than Pre- |... Microrhynchus, Hetera- nanthes. chena, and Picridium about 27 species, chiefly Mediterranean. Pre- nanthes more distant. SCORZONERA | Pinaropappus, Lygodes- | No common genera. T jn aen ^ dg cron group. mia, Scorzonella, and d Tourecidid Vail Stephanomeria, about ola De y 20 North-west Ame- ) species, Suropeo- rican, and Picrosia 1 Asiatic and Mediterra- South-American species. nosi.. Connexions ...| Cichoriaceæ poes. ES E AR n Cichoriacee generally. The above Table would appear to give ample evidence of a (geologically) modern interchange of Composite vegetation. Thirty of the thirty-eight generic groups have a general range over North-eastern Asia and Northern America, and may therefore be considered as comparatively continuous. They include about four-and-twenty species absolutely identical in the two divisions ; 500 MR. G. BENTHAM ON COMPOSITE. and many others are closely represented in both, indicating a common origin in one of the two; but whether the flora, or, indeed, any individual race, has travelled eastward or westward it would be diffieult to decide; the evidence is different as to different genera, and in almost any case may be explained both ways. Eupatorium, Solidago, and Aster (Euaster) are very large American genera, numerous in species throughout North America, diminish- ing, however, in numbers in the north-west, reduced to very few in East Asia, and dwindling down to a single one or two in West Europe. Most of the Anthemidee and Cichoriacez, numerous in the Mediterranean region and West Asia, as well as Tanacetum and Artemisia, whose chief seat is, perhaps, Central Asia, all diminish eastwards, and are reduced to very few in North America, chiefly at high latitudes or along mountain-ranges. The primá facie conclu- sion would be that the former, of American origin, had struggled to extend themselves westward with less and less of success as the distance from home increased, and the latter, of Old- World origin, had met with a similar fate in their progress eastwards. But, on the other hand, it might also be argued (perhaps, however, with less plausibility) that both had once ranged over the whole region in a small number of specific races, but that, the one set finding the west and the other the east more congenial, the circumstances more favourable to their preservation and development, they had in course of time multiplied in the one division not only in indi- viduals but in differentiated races, whilst in the other they had more or less succumbed to adverse influences and gradually become extinct, barring the few representatives still capable of accommodating themselves to the circumstances among which they are placed. Eight or ten, however, of the generic groups enume- rated have no such apparent continuity ; their widely disconnected areas seem to imply an ancient very wide range, early broken up through the greater part of its extent, leaving here and there a few isolated remnants which have lasted long enough to produce endemic races at the opposite extremes, the common parent races having become extinct in their typical forms. Adenostyles, with one species in California and two in Central Europe, Filaginez (Evax, Filago, and Micropus), with seven North-west-American and nineteen Mediterranean species, Bellis, with two in the southern states of North America, six in the Mediterranean region, Werneria, with one Abyssinian, one Himalayan, and about sixteen Andine species, Oentaurea, of a type unknown in Central REGIONS AND AREAS OF DISTRIBUTION. 501 or Northern Asia, with one species in the United States, four or five in Chili, one in the mountains of South-east Australia, three or four in the Mediterranean region, and two in Abyssinia, would, under the theory that affinity indicates consanguinity, imply an immense and luxuriant dispersion in early times, such as we now see in Sonchus, Picris, or Hypocheris, followed by an almost equally general destruction from causes as to which it seems at present vain to offer any conjecture. Adenocaulon, Leontopodium, Cnicus, and some sections of Aster, Hyoseris, &c. show a dispersion somewhat intermediate in character, less disjointed than the second series of genera, but of an area more broken than the first series. Extratropical southern connexion between America and the Old World, as indicated by identical or nearly representative Species, Sections, Genera, or marked groups of Genera. These connexions are twofold—first, between South America and South Africa, and, secondly, between South America and Australia. The two are quite distinct from each other, the former very few and remote, the latter rather marked, and all quite independent of the connexions between South Africa and Australia enumerated below at the end of the notes on Old- World distribution. Taking, first, the South-African races, we have Table 6. Connexions between South America and South Africa. (There are no species common to the two.) Genera. | FLAVERIEX........ | Exopical South-American South-African representatives. notypes, and, through them, the whole tribe of Helenioidex. l'laveria, 8 species, South and West America; the nearest genus to Cadiscus, but not closely repre- sentative. representatives. ASTER type.........| Sommerfeldtia, 1 species (South-east | Felicia, 45 species: 2 America). Connexions ...... Aster (Seth and West America), | Aster (Europe, North Asia, and 1 nearer than Diplostephium (Andes), Bouth-east African speciet), nearer or Chiliotrichium (extratropical than Olearia (Australia). H South America). OMOCHROMOUS " ASTEROIDE E ...| Hysterionica, Nardophyllum, and | Pteronia, Fresenia, and Homochroma, E epidophyllum, about 20 species. about 55 species. Connexions......! Solidago, Haplopappus, &c., above | None except the Mascarene Rochonia 150 North- and West-American | and Glycideras, 3 or 4 species. species. : JAUMEÀ a — 2 Soath- American species, 1 Mexican | 1 south-east tropical species (Hyperi- and 1 Californian species. cophyllum), nearer to the Mexican (Espejoa) than to the South-east American (Jaumea) and North- : west American species (Coinogyne). Connexions ...... Venegasia and Olivea, Mexican mo- | None except the remote one with the monotypic Cadiscus. Cadiscus, 1 species. 502 MR. G. BENTHAM ON COMPOSITE. 3 Extratropical South-American Ate : Genera. representatives. Bouth-African representatives. FLAVERIEA. Connexions ...... Tagetinee, near 100 species, not very |None. closely allied. CoTULE:x............ Plagiochei Connexions CALENDULACE.E lus, 6 species, chiefly south- western. Soliva and Abrotanella, about 5 spe- cies, further ones very remote. Antarctic Cotula and Cenia, about 30 species. Gradual with the whole tribe of An- themideee. Eriachenium, 1 species, America. None. Oligocarpus?, 3 species. Connexions ...... 7 genera of Calendulacee, including 112 species, and, through them, South-African Senecionidew on the one hand, and Arctotides on the other. Gerbera, 20 species, South- and West- African, with a few Asiatic. None very near; a marked gap be- tween Gerbera and the other African and Asiatic Mutisiaceee. GERBERA type ...| Chaptalia, 18 species, southern and tropical. Trichocline, 20 species, southern and tropical, and, through them, with several other Mutisiacee. Connexions ...... Out of the above list Felicia and Sommerfeldtia, Pteronia and allies with Hysterionica and allies, Plagiocheilus and Cotula, Chaptalia and Gerbera, may be regarded as the results of the partial break-up of four great cosmopolitan or very widely spread southern or extratropical races—the Aster type, the Solidago type, the Cotula type, and the Gerbera type. But Jaumea, section Hypericophyllum, and Cadiscus are very singular in their geogra- phical position. Both monotypic, they are the unique represen- tatives in the Old World of the great American tribe of Helenioidez. One of them, Hypericophyllum, is so near in structure to the South- American Jaumea, to the Californian Coinogyne, and especially to the Mexican Espejoa, that I have ventured to unite them generi- cally—to which course I fear 1 may meet with many objectors, chiefly on account of the very great geographical discrepancy. But I may observe that a similar distribution, although rare, has been noted in other parts of the vegetable kingdom, as, for instance, in the two Scrophularineous genera Melasma and Alectra, which have each a South-African, a South-American, and a Mexican species. Cadiscus is less closely allied to Flaveria and Porophyllum, but is much nearer to them than to any Old-World forms. As a pendant to these two cases we have the monotypic Eriachenium, from the southern extremity of South America, the sole American representative of the eminently African tribe of Calendulacez, whose connexions with other tribes are also exclu- sively of the Old World. REGIONS AND AREAS OF DISTRIBUTION. 503 Table 7. Connexions between South America and Australia. Genera. eee Australian representatives. CENTRATHERUM | 2 South-American species. 1 species identical with one of the : American ones. Connexions ...... Oiospermum, 1 species, and very nu- | Pleurocarpea, 1 species, not very near, merous Vernoniacez. and no other Vernoniacese. ASTER type ......... Chiliotrichium, 3 species, quite Olearia about 80 species, Australia 2 Southern. and New Zealand. Connexions ...... Diplostephium 18 Andine species, | None nearer than Erigeron. ERIGERON type _ (for Erigeron itself see Table 8) Connexions PTEROCAULON Connexions ECLIPTA (see also Table 4)......... Connexions ...... FLAVERIA Connexions ...... ERECHTHITES ... Connexions CENTAUREA Connexions TRICHOCLINE...... Connexions MICROSERIS , Connexions ...... and, through them, the northern and western Asters. Vittadinia, 1 species, Podocoma, 5 species. Erigeron about 14 species, and, through them, with Sommerfeldtia and the northern and western Asters. 4| 6 species, mostly tropical or even northern, Tropical, with Epaltes, Pluchea, and allies. 2 Chilian species (besides the cos- mopolitan one). Leptocarpha, 1 Chilian species, and eee Verbesinese generally. "f erican species, chiefly southern and western. Sartwellia, 1 North-west American species, and more distant with the agetines, all American. Plagiocheilus, 6 South-American spe- cies. Soliva, 4 South-American extratropi- cal and tropical species. Abrotanella, 1 Antarctic American species. None nearer than Artemisia. 7 tropical or western species, 1 ex- tending to North America. The cosmopolitan Senecio. 5 Chilian species. None except 1 North-American spe- cies; the numerous other species and allied genera all northern Old World. About 20 American, chiefly southern, species. Chaptalia, 18 American species, and, through them, many American Mu- tisiacese. 1 American extratropical and several North-west American species. None southern nearer than Hypo- cheris; in the north Krigia, and, through it, other Cichoriacez. Vittadinia, 4 species, Podocoma, 1 spe- cies (none identical). Erigeron, 4 species, but no other nearer than Olearia. 1 spor om only 2 Australian, the others North - Caledonian, and 1 South-Asiatic. Epaltes, 2 species, but no others near. 1 Australian and South-east Asiatic species (besides the cosmopolitan one). None near, and very few more dis- tant Verbesinez. l species closely representative of the commonest American one. None. Cotula, section Leptinella, 3 Austra- lian species and of various sections, 1 ew-Zealand and Australian species. Soliva, 1 species, identical with one of the South-American, Abrotanella, 3 Australian and 4 New- Zealand species. None. 6 Australian and New-Zealand spe- cies, 1 (colonist ?) in South Asia. The cosmopolitan Senecio. 1 Australian species. None. 1 Australian species. None. 1 Australian and New-Zealand spe- cies. i None except colonists. The amphigeous races enumerated in the above Table are of a very mixed geographical character. The Vittadinia, Podocoma, Pterocaulon, Eclipta, and Erechthites are subtropical ; and although they have no representative, except presumed colonists, in tropical Africa and Asia, and therefore did not find a place in Table 3, 504 MR. G. BENTHAM ON COMPOSITE. yet their geographical connexions and presumable origin may be assimilated to the few genera of that Table whose Asiatic character approaches or even exceeds the American. The single species of Centratherum, Flaveria, Soliva, and Trichocline, identical with, or closely representative of, corresponding American species of essen- tially American genera, suggest doubts whether they are ancient (or even modern ?) colonists from America or really remains of an ancient common flora. The Chiliotrichium and Olearia, the Plagiocheilus and Leptinella, the Abrotanella and the Microseris, and even the Centaurea form part of that general Antarctic flora which in so many orders shows a striking connexion between Australia (especially South-east Australia and Tasmania), New Zealand, and the southern extremity of South America, the Micro- seris and Centaurea showing in Australia the extreme end of an area extending from the northern extratropieal Old World over North Ameriea down the western backbone of the New World to the extreme south, and thence to Australasia. Australia, therefore, in regard to America, would appear once to have had in the south an antarctic or mountain connexion or communication sufficient for the interchange of races, to have received in the north in ancient times, as part of the Indo-Australian region, a few tropical or subtropical American races, and in ancient, as in recent, times to have readily admitted and favoured the spread of colonists from America as well as from South Africa, and more recently from Europe. Table 8. Endemic Species of cosmopolitan or very widely spread Genera. I include under this head those genera or groups of genera which have endemic representatives both in the New and the Old World, and both in the northern and in the southern hemisphere. The numbers given are necessarily very vaguely estimated for the northern and tropical regions, and must be taken rather as rela- , tive than as absolute; those for South Africa and Australia, founded on already worked-up floras, will be found more accurate. Such cosmopolitan or widely spread species as Gnaphalium luteo- album, Erigeron linifolium, Pluchea indica, Cotula coronopifolia, &c. are omitted, as having nearly the area of the genus, at least in their own primary division of the globe. Senecio has no such cosmopolitan species. REGIONS AND AREAS OF DISTRIBUTION. 505 Endemic Species in Euro North .__, |Extratro- pe : Genera and groupe of America, OPC) pical |, end, Tropical) South | Austra- senem. ingluding america. |, S082, mia" die | Africa. da. Aster type ......... 120 28 11 50 2 53 63 Erigeron ......... 45 35 18 8 1 1 3 Conyza cisco 3 4 3 2 20 8 2 Pluchea ............ 7 5 aoe 1 10 sae + Gnaphalium ...... 18 20 18 10 12 8 6 Coll 1 z: RA 6 1 20 8 Senecio ............ 105 265 100 160 80 190 28 Gerbera type...... 1 26 11 1 4 15 1 Hieracium ......... 26 10 4 110 m 2 sa Of the above genera, Conyza, Pluchea, and, in a less degree, Gerbera have a rather tropical character; the others are more prevalent in temperate or mountain-regions. All, except Cotula, are endowed with means of dispersion which we should, primá Jacie, qualify as ready, the pappus spreading and light in propor- tion to the achene; but the ready colonizers (one, two, or three to a genus) belong to three or four only (Erigeron, Conyza, Gna- phalium, and perhaps Pluchea) of the eight pappose genera, or to Cotula, which has no pappus. Senecio is remarkable for the enor- mous number of locally restricted species, no one of them common to any two of the above regions, and not yet satisfactorily distri- buted into sections at once geographical and structural. Its only colonizer, S. vulgaris, is not classed here amongst the ready colo- nists ; for, as far as I can learn, although carried out into some distant lands with cultivation, it does not, like Erigeron canadense and others, establish itself over the country in waste and unculti- vated localities. The sections or divisions of Gnaphalium, Erige- ron, and Conyza are more marked than those of Senecio; but the principal ones are not geographical. Aster, Cotula, and Gerbera have established subordinate races, geographical as well as struc- tural, sufficiently distinct for us to have adopted them as genera. To the above genera might be added a few of very wide distri- bution, which, from Europe and northern Asia, spread round by North America and the Andes down to extratropical South Ame- rica and even to Australia, such as Centaurea, Hypocheris, Cnicus, &c.; but they appear to be better placed, as instances of extratro- pical northern connexions, in Table 5. To the same Table belongs also Hieracium, which I have added also to the present one on 006 MR, G. BENTHAM ON COMPOSITE. account of the two South-African species belonging to a section more abundant in America than in the Old World, and indicating perhaps an ancient area much more extended than the present one. 2. Separate distribution of Composite into Regions. Coming now to the consideration of the separate distribution of Composite in America and in the Old World, we may observe one striking difference in this respect in the two divisions of the globe with regard to the extratropical or subtropical races which form the great bulk of the order. In America the northern and southern tribes are the same, although in different proportions ; and there are a considerable number of identical genera and even species in the north and in the south. In the Old World, on the contrary, two large northern tribes (Cynaroidew and Cichoriacez) are absent, or very sparingly represented, in the south ; whilst the southern Arctotidez, as well as several subtribes of other tribes, are wanting in the north ; and the genera common to the Mediter- ranean and S.-A frican regions (excepting cosmopolitan genera) are very few. This great difference in the two divisions of the globe may be due in a great measure to the direction of the great chain of mountains which in Ameriea, running north and south, facili- tates, or has facilitated, means of intercommunication to races of the constitution of Composite, to which the east and west moun- tain-ranges, plains, and deserts of the Old World only oppose obstacles. In both divisions, omitting the comparatively few Alpine and cosmopolitan races, we have three great specially composite regions which may be at once centres of differentiation of races and areas of preservation of mixed floras, having more of the former character in the Old World and of the latter in Ame- rica. The Mediterranean, the South-African, and the Australian Composite are respectively far more distinct than the Mexican, the Chilian, and the Brazilian, which are, moreover, further con- nected by what may be termed a fourth intervening region, the Andine; whilst in the Old World the only intermediate connect- ing-region between the north and south is a very partial one in éastern Africa. I shall now, however, enter into some further de- tails as to each of these regions, as well as in regard to a few others less defined—that is to say, the United-States region in America, and, in the Old World, the western, or African, and eastern, oT Asiatic, tropical regions. I add also to the American regions the West-Indian insular group, as being enclosed, as it were, betwee? REGIONS AND AREAS OF DISTRIBUTION. 507 North and South America, and closely connected botanically with the united continents, giving altogether six American and six Old-World regions. It must, however, be remembered that these regions have reference to Composite alone, and would require great modifications for orders rich in forest-trees, or in paludose or aquatie races, &c., which are, as already observed, so very rare in Composite. It must also be borne in mind, in making use of the two following Tables of the distribution of Composite in America and in the Old World respectively, that the limits of the regions are in some instances as vague as the absolute numbers of species are uncertain, owing to our insufficient acquaintance with the data on which they should be founded. Such limits as are here had in view will be specially explained in the notes which follow each of the Tables. The insular regions severally con- nected with the two divisions will be separately considered, and are not included in the totals given in the following Tables 9 to 12. Table 9. Number of Species of Composite in each of the American Regions. Number of Species. | | A | a | ED .| Genera Genera. E E E y LE in Old FHEHPHEHEHEHETHB GS Mud LE-EE-E- 2 Ee E 3 Sales Siar seca EE | i | i | 1 parganophorus ............... cos Ee e O. w. as M D Uca 5x | : Su pondere Ed E retro onn oen ecl p | ml] EINRLOE Oicepermum. eoe ei cad ids wits A e | 1 Centratherum .................- - 1, 2| 2|..| 2| ow. Blanchetia...... 1 eb ie ee ib tad, bas peek be Vanillosmopsis .................. eee MAN (PSY Albertina.. -eoir A A Rd I LI MOON n- ei 20 | 8 | 24 | 32 |180 | 12 250| o. w. Lachnorhiza 7. 0. ERE LlSLbp eed Piptocarpha 220. epee ts 1| 4] 20/... | 24 catar o ene ae ME LpI.pere pr + Stilpnopappus .................. e xb : PIN. ocn o decer ird uE a H 8 Oliganthes A oen FIIIT] I 8 Piptocoda sao a 4 besa Neon: 1i econ | cose 1 Proteopeis. i... oe eos lua A E GAO dE 1 o AA dors | oe | ipee | ooo | 1 Haplostephium |... np. pedem | 2L. 2 Iyelmopbopa cirio cocinas pedem m | 17 | 1H . | — MM AA (A | Carried forward ............ | 23 | 9 |30 |44 |258 | 12 |344 Genera. Mexican region, MR. G. BENTHAM ON COMPOSITA. Number of Species. Btates reg. West United VERNONIACER. Brought forward ......... Lychnophoriopsis............... Eremanthus ..................... Chronopappus .................. E Telmatophile DIL Elephantopus .................. E A Lice oer cioe Ac Indies. Andine region. Brazilian region. America. Total in iy 9 :8 > most e Ophryosporus .................. Helogyne 9 E Gymnocoronis .................. Adenostemma ................ Rclerolepia--.— sss dn NOIR A EE M NU 'l'überostylis -. o. ae Aperabum ;...5.—...- 0e DLP Hofmeisteria eerie REDE. cesan coc rire Trichogonia Brachyandra..................... Leptoclinium DIDI Arrianilius otero esco j^ Adenostyles Brickellia Lee ise Carpochzte Kuhnig:- e o ese Luar. ss Carphephorus eechtottsossostose Ie Total Eupatoriacex 10 5 AE L4 20 te 3-% pl OO pd pudo pul pl O pd . pt . . D: 2i adam: al | MIT O z [er] - Om Ob Tbh OOO [wr] pá Di nh OD RO mm totom Sm ES , — O bo 0 C$ 189 41 743 e] REGIONS AND AREAS OF DISTRIBUTION. 509 Number of Species. G g P 3 ag ama ee EFIEHIPHEHEFIEFIESIBE 21 S| Bs) 2S | SER So | Se) Se Sia lea dE Aros lad ASTEROIDEE. Gymnosperma ................-- Di a decode» Xanthocephalum ............... TE a 1 Bh oo e Guüerezia.-— ee MESE! E .| 6,20 Grindelia 4. ie 8] 6 1 «9220 i«Pentacheeta t... 25 d sdb Aphantocheta ............. zl pueri Steriphe rers Adone sol tee xe wl Vela Bradburia se.. cose cose dla SOME EMI IHeterotheca ees $: l a 5 Chrysopsis 5. ee AI GS tee 1 B ua -Hysterioniea. ——......-.—..- ducc pecie 2| 41 5 EXANGHISMA f.e ereo ricas ESSI ceeds mde Haplopappus ..............-.-- 164/25. | Loe | 24 1202 Chrysothamnus ...... cna 10 9s | 20 Ericameria lona 0 4 cb ese Soldago |... terrere 7.972. 9 151 1:2 801 0 w. «Brachychieta....--..... e eee sd A E a regi Lessing eee 4 p ies «coli Nardophyllum .................. sd poc EE 5| 6 Lepidophyllum ........ Sones e eue 2. os. 3| 5 Chiliophyllum ............. edge bed Lj Læstadis - 0... AT cu pese i.c Egletes —.. oo: os 9X rr 92 11.20 Aphanostephus................-- nde cere Koerlia ;.........-- e pedem Lagenophora ................-- "a tee o. dee 2-125100: Ww. Bellis a oaeeo iere Ii f T OS. Ww. Monoptilon ........ . eie sd eoa Ties ie due. xr a Townsendia .......... ER 8] 7 cde dpa uere T TET Xp s CES : hetOpappa -s::<--sr:--+2s--: <- One e coc one Piet. Peste RE EA ix SPEI dee 3 Boltonis es- posers ardin 94 Ajajaj ‘ Ti oW, Corethrogyne ....... XECCOGC Ol iis lodos ce eee Rremiastrum ——.. 1 : bo plene cet Sericocarpus...... A A O eee, eres LE A EA El Y | OO A 4| ee O W Sommerfeldtia .................- roc etes emt Hinterhubera .......... ed. A e 3 sd Diplostephium ...............-.. 2 oe [IS lo RÍO Chiliotrichium .......... BUE EE a A [aces [coal oO Podocoma......... eee eaten es lore e 11.91.21 51:0-W. Brachyactis ....... UE qu E qe cs apto Beas Ena E LOA e Erigeron -ooie ses scot Peep aD 331 11 | 16.) 5| 14 109% | 0o we Vittadinia <....c.sc5:-5 2-0: sepe eee eed OE Grae Conyzi — ...—. E P osc e 6! -1 2|] 5| 31 € | Ley 0w: Parsstoplos. E AAN ty fice | we DE see eae Baccharis -1o eres Carle 30| 5| 7| 70 100 | 40 |250 Heterothalamus ............... - berpesog oJ Total Asteroidex ...... ...1203 |269 | 27 |137 [119 |131 ¡830 10 MR. G. BENTHAM ON COMPOSIT X, Number of Species. s | E E so Genera. E à i - M E J E d 3 g q World. ; 32 | 23g . EEES -EJEFISEMEEJE EIER AERE: INULOIDES. | Stenachznium .................. Le. Ss 3 Pluchea............. E oes 6i 4) 3) 3) 3) 1:121 0 W: RETT R E rere E SERERE o 3 Rhodogeron ..................... aes 1:5... e 1 Tessaria.......... Seren seece JE oos bane 4i l 1 5 Hpaltes -— sess p) et teas a ee Pterocaulon ....................- ve 2 1 4i.. 6| ow. Hyvax 0 oem El: sew os : E o. W. Psilo AU E E E 1 1 396 Tauris as ce | onc e ud. oW. PDiaperia EP A A 2| 4 was ved MIGFODBSIS -- 2 o eee n “ss ces JE Filago oye eck esses: I| es ees 141552. | 0V. 'TTafalla 3. sacro oidos ELI TL is 4 Mniodes: —.....-.« eee : ae 2: F 2 Antennaria ................sse 9: esl 9 ow «Duciliopsis.. eere eee PRESSES ES Olpandra- eee e DLL ao tp Chionolena ..................... : Se: 2: 2. 4 Leontopodium .................. ot 2. | aso Ue Anaphalis- 5. eere : 1 dens S 1 0. W. Ghevreula eue EET 2.19414. Hace. ee PECES Id 21 3 Achyrocline AAN an le re Gee Bere eee Gnaphalium ....................- 10/1; 3¡20| 3| 20] 60} o. w. Bienocine.— E aie es PEE 2 | Masc. Leucopholis ....... I T E 2: 2 Total Inuloidee ............ 30 | 30 | 12 | 51 | 28 | 31 |157 HELIANTHOIDES. i QU IUE 7 1 1 1 7 Heptanthus -eoii A 4: 5 2 mila o sico k gd 3 Tetrantihus onene renate Fe PA ut 2 Eria M ER 1 1 1 1 1 Tentanopeis ...........—.. e 2 I: 1 Stachycephalum ............... T pee : 1 Milleria....... E CU 1 1 soeur Adenocanlon.......— ace. Ses 1 Vas 21-3 o. W. BiencourlHià -o ecu lO eee Eus Desmanthodium ............... 2 [4 ue wee 2 Olbsdium-. S T Biol Gli 0f 4 14 Jhiyoihen..........— eee i ien 3] 6 8 Trigonospermum ............... El es 2 Polymma pernil: 4| 9 81 1 12 DOLÍA eroi ee SI 11 11 Philoglosa -no me bed wes be 2 Dicranocarpus .................. 1 un 1 Carried forward............ 22| 8.16 131 -18 | 2 | 78 REGIONS AND AREAS OF DISTRIBUTION. 511 Number of Species. tib Genera Genera. a & ; > 3 A PE in Ol EHEFIFHEHEHEHEHBE Z7 = 25 So SERE E EEE E ELIANTHODELE. Brought forward | ......... 22| 3]|16,|31]|18| 2| 78 Guardiola... E a see 4 Baltimore pip | ok 1 1 Melampodmm-. <5. ..5..: 15) || 41 4 18] o.w Acanthospermum............... edes ee Pale 2 2 Sohizoptora o.i. seven ences seo Paces on pe eee) es Coal A A oe nienie. o ns Jose beo dE iB! E o AS ve | E werd oe beel 1 bloww Berláandiera — aa 5 5 Imdheimera-. 6 ec . Ei. oe 1 Bnselmanmua . € — — III. mob Parthenium... ae ijo 1 1 6 'Añolotheca 9. IE uel 1 Parthenice 5 1 See ee d EM 1 CR C RU T DI 1. - |. ri Oyentes coo 1 Seele eec deoa DICO re esos Iul cte 1 Cyelachupa e 1 Job ecdesia -«Euphrosyne... 5.1 L|. bcd Hymenodlea V Zee cce 2 AM DONA 2 46. e 41.06] 9l 2| 4|. Oo ow. Pra nsona —4 2 or 04s. 6 9 e 2-[ 3 | 10 Kant hM oo rue 1 oT 1 } 2| 4| o.w. Podanthus |... eem cec. 93-9 Astemma cio ES IR oes 1 vic 1 Eragocëros noe see 4 aah vont t Philol 22 oa Se 1 O ere ree bee ae 1 ZihnB8ilc.. c. ee aA zd 12 Sanvala — ee, 7. 4 rxdge d b. eee 4 Helio pais 5... estes ee 2 1 woe [eco Aganippe soii itio ieiti 2 x EP MOnaGtis= i. ee eoe : 2 oe 2 Rumtordia .— e 1 m ue E Siegesbeckia -iig ciones Jon Peco te 2131 € W. Stemmatella ........... Pag om c [eee 1 ds 1 JEFA o c HUE Tru loe] ES II PD Bnhydràe o a I 1 l 41 4| o.w. AphanneliS socne 6:0 Pu VE 2 Belipta eoo 1 1 jl 1 1 2| 2] ow. Deptocarpha-- es cudbend ces | cee ec dE dede Seloa | ASKS IIb p 1 Axinipbyllum occ raro 2 ee 2 Anasaloa ose eae 1 2p i ZAaluzanià o 9 {i Li 7 Sabazia eel sen 4 4 8 Mari o ue 2 p. p 2 Gymnoloma -o n ai BII: 21 4 16 SClerocarpus. et 6 JE! 1 10 | ow. Montano as 2I 3 14 Carried forward............ 152 | 39 | 26 | 68 | 40 | 18 1280 LINN. JOURN,—BOTANY, VOL. XIII. 20 12 MR. G. BENTHAM ON COMPOSITAE. Number of Species. Eb | a , Genera Genera. E aly E E 2d E: 4j g E £i World. -HIISEACSCEEPAEPESSEIEE: HELIANTIOIDEX. | | Brought forward ......... 152 | 39 | 26 | 68 | 40 | 18 ¡280 Doc A AA PES ve po Budbeckia i e es PARERE E ec pap Chromolepis AAA aarp El E Balsamorhiza 565r 341.9 «s. fF 0 "Tetragonotheca.................- SUIS bue x. dw IU Mar o n ere ae pat ek 8T1-9 P4 v9 Mirasolia Lee eren SI Lo O eaa 2 lKostenhane elt 2 nee tet POR Bornehia ns Ri PI ae ae oe Pascalia.. cotos a ee 1 1 Bisinvilloð e 1 1 9| eT Ww Weedon err erem 1 8| 10 tE] 30] ww Eleutheranthera ............... E 1 1 pales 1 ABDINA cien deed eere Cisl 4 981 bd oe 5. Zexnienià 522 5-2 5 rm 161 Fi 21 4t 3| 1115 ENODEM Se eene derer tents 2 91H 22 hi eerie ed cat dee 71.2 we 4 "hon 50:50 a bille 5 hil o St I| 41164241 160 Eehanthus is 10 | 32 Oi | 3-5 Dimerostemmna.................. Bee leaped a oto de cae: 1 o AAA 81-7243 4 PPS Melanthera ciación 84d bt Sp Bs | 5| 6x COR o osado 15i E415] i 20 A AAA D A a o O 6 "AGUnOIDeris je 34 6 ee 9 Verbesina S i e duos 271 31 5118) 9| 2150 ROUEN clic ied ae se pere 1 Podacbenium: icono 02s. iT I wu 1 Spilanthes ick pcs cee leash Bi 21:41 5| T ]-419 |] vw BUDE. S eee Bi. | Sd tod a PER ri seid a O Gg eee 1 Hymenostephium ....... qued EI nt : 2 Chænocephalus ..............- ee Ge e m.p ed 1 x 1 BERGER, on an DRED ie SN 114 ca poe | OW A ecc ee Oe ee ee ae eg Heterospermum .............-- 2 cb egi. T5 Doreocsarpus 0.005. ocurre ea Bee ds ie ae 2 COPSODSIS o ore root en I 27] 7-6 | | ow DEN is he dae S cee | vis es XE D Hrdaleoa eee ale eee eee Zi io] ur. "Fhelesperni aair gu obe Ja [90 Cosmos eve. tetas 9 1 2:591 Esp 10 Bidens” +... uc ER 161 7| 512012] 3| 45] ow Narvalns.. aos 5 0. sepes iioo 1 ASOMAN eL Dd Bl ee Chrysanthellum ............... II II dh di 3p IE Carried forward ........ -...941 [166 | 76 [192 |175 | 39 [836 | REGIONS AND AREAS OF DISTRIBUTION. 518 Number of Species. 50 Genera. Bur Poe 88 Os OM EHEHIPHEHEH EE 35| Word, Salas | Bg = . z ERE EE HELIANTHOIDER. Brought forward ......... 341 [166 | 76 [192 175 | 39 (836 A Lieu ruoeoporcete: SII II 1| DB Caleaz Ex. edueulneov 14: 1 1] I8 | 30. |... 1:60 Balduina —-. . 5. Ad 2 s deese o Moe EIE E S f wo epharipappus ...........006+ ae ae 1 Tridax CH RU ed. 5 EE Ert Madia coo econ 812 Leas Hemizon ia Coc ccreccecvccscceces E ... c.. 25 «Easophylla. jiin . 3 pute EC Ele ee 12:| I 12 Achyrachæna e l|. 1 Total Helianthoides...... 411 |175 | 79 |217 |206 | 41 [963 HELENIOIWES. Cacosmia :....... 1 9 00e pot. aem 3 p - S Geissopappus >»: sses: -siste qo cuc Coe se Tae TUAE Occ ds 1 4| ow. Venogadia «a. is debvexcs es curis 1 1 Olivan eee ESSE Ete 1 Rosllas«... eiie ees Jewels. d.d Taphamia n- -ors ve e IQ: eb l.l 2 Pertyle$.— 5e coe : : Orypappus ves Burke R E 2 2 Beria oo aaan s D 5 Actinolepis s.e a n bl.ixcka-le O Whitney ET ee org ce ee ee s Lasthenia 26.0257... 5.acsesc ees Dees weed | EE Monolopia RA 2 2 Hecübsa eere reete I5. 1 Euddelha ico o9: l 3 Hulse eaa oer 6 |... T 6 Chznactil!..: e eec Hi3 mp Hymenopappus ............... It 6 e Tua Syntrichopappus ............... I sob Bahi |. S eo 16; 4 1 | 20 Behkührig A 6... 3 21 8 Hymenolhnx asa 2:55 2 Amaire 9. ns 1 1 Closia 5 n eere s 5| 5 Blennosperma 11. oe vines eren i4. 11.4 Milanoya . o 24 2 4 FBlorestüng-...... ue 2-4 2 Palafona — .. 95 5s : : T Huriopasppuas.i sees Pee 0n 3 2 Galéeana. als Ira 1 Carried forward ...........| 99 | 21 | ... | 6| 2/11 [131 202 514 MR. G. BENTHAM ON COMPOSIT X. Number of Species. Genera. E E 3 ag Caola EHEFIPEAEHIERIERIEHIUCEE! He | 3 joale S RITmRISS -HEHEHEHEREHEE HELENIOIDEs. Brought forward ......... 9041211... 6| “Beery HS Amblyopappus ...............--- dedos esu p Thymopsis A neo bot meet Mierospermum.................. H vos ecbecp F Sartwellia .. ......... les. bz A O wee a conuertere 4| T5] Lip 61 vw TOROS ceu 55-000, Maira | L1 841-06 H- L4 15 Wescaillea esee Acc ee ee Adenophyllum .................. 3 3 MOOI PA A 2. 2 Dysodia.. entes Eel 10; 2| 1 10 Syncephalanthus ........ deiade oe ee A 1 Bchizotrichis error ess qu 1 Hymenatherum ....... ies MESS ee be Poh ME TREBOBDVIS cmoccinccanetanes» 3 3 Adenopappus .................- Lio pelis 1 EONS oaks icf eibacsisanddeseccki BO) eb O 1O j4 20 Ciryielinia.......- eed eere eerie l2ueebpeebe 1 NE EA SIUE UNI QUE 16 | ..| 8| 16+ 10 40 Clappia ooo ae ass: Sh oP aah ao Piet Sa ee Cephalophora .................. at V l4 Helani Louie edet STRIS. 18 Gaillardia PEA 2| 6 2, 8 o Loon cer iret 2| 9 10 Mare ed E IA TC 23e 2| 4 Triehoptilium. ................«. 3 3 Poathyrotos no. serere: 1 1 Total Helenioidez......... 201 | 51 | 15 | 32 | 25 | 23 [304 ANTHEMIDEA. ABIS | irae css ai esoseveskes: le Spa ss] SF 00 We LeueiiÉDYE 1... ene ecito. Elio edu di 1 LA, AAA Medie de cis 5d A us P. 3 Chrysanthemum ............... uer i sa] adele Os We Matricaria. . 550202335 sc swssces se LO es tee wl 2 | On UDIN EA Paneer ere 2: eb nec c Salt On W- Centipede eere x Ib I] ow. Plapiocheilus | .........—-.e e: expe e d0sse feel ap GO Olivas eee se ccc cick se stecese es SS oe) oe l S ew Abreotanella iaci SF e: è 1 1 o. W. Tanaecbum a a li 5 iel De ON Atona rai 3 | 30 3 | 32 | o.w. Total Anthemidez......... 12142] ..| 8| 2]. 9| 63 SENECIONIDEE. A eee Gual 9 1.32 hi ode | 40 IM A cuss vo ee ee a ae yen [ick Carried forward............ 7 3 | 32 1 | 41 REGIONS AND AREAS OF DISTRIBUTION. 515 Number of Species. a te é pes G A : See : i = 63133183158 a a|gal£2| Won a o T -FAEEIDCIEEIEFIERIEE SENECIONIDER. Brought forward .........] 7 |. 3 | 32| 1 | 41 Schistocarpha ............- codd | ere | al sce ees ds Neurolsena co Vie 1 d]. Peucephahim —..... eae n je dde ws Tuinaz 5.2 eos nacida pw | eal sd eco 1 Petasites ....... m ane à 4 undo use 4| o.w. Armi o 20. 7 on nS ii ves 7 O. W. Crocdiumm eee roc 1 sides aco acc 1 Bartlettia ....... peers SOF ee ee ore ee e 1 Haploesthos seules. *lseleln.]e]ul Raillardella ........ Bocas eee IAP cd oda 1) Gap dE ode |) oc 2 Melaléma all pro | oes 1 1 Erechthites . 5... 06e es Oo) 11 144141 |. ON Quictium a er e o a quo 1) 14 Senecio acacia ds cobos 80 | 38 | 10 ¡130 | 25 ¡100 ¡380 | o.w. OYhoxyu O E REE qe REA moll y A SA E Tetradymia siesti ie 3| 2 "e a| 3 Werneria .............-- sese 17 18 ¡ ow. Total Senecionidez......... 102 | 54 | 15 211 | 29 ¡103 |502 CALENDULACES. Eriachenium ......... rc 3 I CYNAROIDEZ, : Cnicus a. Bases edu esse oes 20 | 15 n 10D, [2 Oats Saussurea ............- seven stack: os 1 w 1 0. W. Centaurea? ..... o ooo e t rd I 5:61 0 Total Cynaroidee ......... 20 | 17 5 | 42 Motistace#. Schlechtendahlia ............... A D EI. 1 Barnadegas 2... o : es ; Id. "7 Chionopappus ......... dT t e te dese. Muüggs eS. RARO UD : 18| 3| 15 | 36 Hyalis 35:3: erisera NIIT. [I4 4 Plaza 17-0 a mm -F dis = : Gypothamnum -ois 22 HR a ias aee Onoseria.::.......... E MM E : TIO f ELED Urmenetea A Se] ura ML I 1 Chuquiragua ......... ERREUR ; =o 6 [10 D S Doniophytón. PS : kell T Wunderlichia cuco E Sop xad Euge LE 1 Gochnatin —. aa A TII TL 31] 3 11 Moqumnia a ue ies rada a LIE da Cyclolepis coles caos rod e VUES Du ew 1 Ber 66 5 0s RE A gs 2 2 SUIS: ue usus Sede vases cess is 5 5 Carried forward ...... ouo 1 | 49 | 43 | 39 [135 516 . MR. G. BENTHAM ON COMPOSITAE. Number of Species. ; j|. wm Genera. $alvasl.s 24|l3a|Sal E World EHEHFEEIEHEHEHEE SeipaibSs (ae ngja MUTISIACEZ. Brought forward ......... 1|..| 1 | 49 | 43 | 39 [135 ADAG AA E do bobada Pachyleng 4... ect wel ue Patas] ll] o E eicere E O A ueldi bed. Brachyclados .................. sclcebeebeso eer fi d Chetanthera ............ 222 een OE elus 1 | .../| 95 | 26 Trichocline ecccccsacscacgecosece bec eee HESS 2 12 8 20 0. W Chaptalia ............ A EC €«| Li 6[191-8] 2| 18 Macrachznium................-- Scc [see IE: Leuemria .........:::. ener e 25 2 23 | 25 Oxyphyllum oie: dus ui 14-3 Polyachyrus ......... eene d 1 6| 7 Proust eves coche cess eee 1 2 5. Perena 4. oec 19 10| -2 | 22 | 46 IP TAT c s enses ev] 1 1 Jb b.e c oss E 6 1 6 | 18 3 | 30 A A oes 92 1 B 'INASSAUVÍA: onses osede ca cde0sccts : ee] eee] 25" | 23 ri A AR m. E 6; 6 Moscliaria-. 5: unen : Dc a A A di ca P IST. plor Cephalopappus................-- Pe ee en rere A Pre a: Total Mutisiacesm ......... 22 | 1| 12 |102 | 81 [176 |383 CICHORIACE. MICFOSePIB |... roce tec 9| 2 J | | ew RPM ee tested moros aap PhalacrosBerlS o a a Ei a Supra o A O eaae: 1 1 TI a cgi 1 1 1| ew. DEDE o decent PENA I 8 c bl: Lm Theracium.- 5. ose ee 16 | 12 10| 21 4] 40] 0. w- Malacothrix ota 00: 9-1 wid ec pec Ing Jt rewar tu qe T T Pe- 1 - Hypochaeris ones Do «I 9f 3110/201 ee Toontodon c. ster eee fees Pies 1 0. W. TETORIN eooo eaa 9-F-IO- byes : 6 | 16 Ühtaxacum atodos «TT k 1 1| 0. w. Pyrrhopappus ......-........--- EZ a Ab 2; CalyooBériS. eere gui pedi oe 92: Glyptopleurum.................. Dr pue 1 ECHUCA lus eie 1 Sel Qu ol 10 | o.w. Prenanthes ,..... eire 10 zc ses 10 | o.w. Nonebts. o eU ace: T 1 1 ree 1 1 O. W. Pinaropsppus ci-ti neiss: e a E O O! Lygodesmia ...................-. VUES arp T 3 PSCTORIE oc DO O PII 1 1 BSOFFOHOMA. Lore 3| 2 dad 3 Stephanomeria .......-.......... 6| 4 8! Total Cichoriacez ......... 56 |71| 3|20| 5|29 [161 | REGIONS AND AREAS OF DISTRIBUTION. 517 Table 10. Summary of American distribution as compared with the Old World. Regions. | i Total | Total [Common Tribes. A in in a o is er! . . PELRA Mexican. Qnited pes | andino. | Brazi | Ohilian [aa e World. a; visions, ON PS EE NM | : G. | Sp.| G. |Sp.| G. | Sp.| G. | Sp.| G. | Sp.) G. |Sp.| G. | Sp. G G. | Sp. Vernoniacese in 4 bs 3 10 9 3i 8 E 24 | 290| 1 5 29 | 376| 14! 156| 4| 3 Eupatoriac:se ... 21 | 246| 9 | 53| 6 | 94] 111156| 15 | 189| 7| 41/35, 743| 5| 16 5| 8 Asteroideze -| 29 | 203| 20 |269| 6 | 27 | 16 |137| 8 |119| 19 |131| 49 | 830| 44| 444| 10 | 8 Inuloideze......... {11} 32) 71 81| 5|12|18| 51| 12| 28| 8| 81| 27 | 157] 118|-950| 11 | 11 Helianthoides | 90 | 411| 38 |175| 36 | 79 | 57 |217| 39 |208| 23 | 411125 | 963, 22| 77| 18 | 10 Helenioidezm ...| 5 201| 12 | 51| 6|15| 8| 32 23| 13 | 23] 59 | 304; 3| 3| 2]... Anthemidee ..| 7| 12) 5| 42|... 3| 8| 1| 2| 5| 9|11| 63| 41, 608| 9 | 15 Senecionides ... 11 | 102) 7 | 54| 4|15| 8/211] 2| 29| 4|103| 19 | 502| 24| 687| 5| 6 Calendulaceg ...| .., |......| ... SEEN ROS TENERE stia d nidis d ot DU PS Es Arctotidem ......] ... ]...... Piero cic c E A e e Cynaroidez...... 1| 29 3| 17 we. lio dumm | -e [ane | 1] 5] 8 | 42 37| 972 3: l Mutisiacere ...... 5-22 1| E 4|12| 18| 81| 161101 | 27 1176) 38 |. 883| 151 e0 Pr Cichoriacee ... 16 | 56| 17 | 711 2| 3| 3| 201 2| 5| 7| 29| 24 | 161| 40| 610 10 | 5 Total ene BAB E 122 | 114 | 78 |291 |145 dod un: | 972 116 | 602 |420 epis e 18 | 62. Nores. 1. Mexican Region. I propose the name of Mexican region for a tract of country with a very remarkable Composite character, extending along the great north and south chains of American mountains, and chiefly on its western declivities, from California to Central America, in- cluding the greater portion of both. I am unable at present to give it any more definite limits. I have no doubt that materials from which a general line could be drawn might be extracted from the numerous Boundary Surveys of the United-States Expedi- tions; but that would be a labour which we could only hope to see achieved within a reasonable time by Americans themselves. It is chiefly from a general estimate of the recorded areas of the species we possess, that I would exclude to the north the Oregon territory and northern Rocky Mountains, as characterized by the Asteroid Gnaphaloid Artemisioid and other mountain-races con- necting North-American with Asiatic Composites; and to the west the greater part of Texas, as showing, in their Rudbeckioid Helianthoid and other races the characters of the United States rather than of the Mexican region. In the intermediate Salt- Lake region, the limits between the western and eastern Compo- site floras would seem, by Mr. Serene Watson's ‘ Introduction to the Nevada and Utah Flora, to lie in the Washoe Mountains of Western Nevada. To the south, the Composite of the tierra fria 518 MR. G. BENTHAM ON COMPOSITE, and tierra templada, at least of Central America, are decidedly Mexican in character ; whilst the hot moist southern provinces show a much greater proportion of Columbian races and woukl therefore fall into the Andine region ; the transition, however, from the one to the other is here evidently gradual and com- plieated. It is possible, when better known, that this region may be sub- divided into two or more, for there are many genera peculiar either to the northern or to the southern districts; but at present the whole appear to maintain the same general character. The more distinet of the monotypie genera are often confined to limited areas as well in the north as in the south, aud those which con- tain several species either range over the whole area, or the northern and southern species blend very much together ; and in the superficial sketch which alone I am able now to give I feel compeiled to regard the whole as a single region. The first peculiarity that strikes one in glancing down the Mexican columns of the above lists is the great diversity of forms showu by the large number of genera, both absolute and in pro- portion to the species, as compared with most other regions. Those of the Mexican region are about 100 more than in either of the other three most varied regions (216 against 143 in the Andine, 145 in the Mediterranean, and 149 in the South-African regions). Nearly half of these Mexican genera are endemic, with an average of about threespecies to a genus, and one half of these are quite mono! y pic—several of them with but few eonnexions,although not so many perhaps absolutely isolated as in some of the Old- World or insular regions. If we deduct the larger genera which have their chief seat without the region, Vernonia, Eupatorium, Aster, Senecio, and Cnicus, the average number of species in the whole region is about four to the genus, and even with the addi- tion of the above five it is not much above five to the genus; whilst in the adjoining United-States and Andine regions it is above six, and in the Old World, in the corresponding Mediterranean region, the average is more than thirteen species to a genus. The larger genera of a specially Mexican type which have flourished and established a large number of species and varieties are Stevia, also abundant in the Andes, Brickellia, scarcely extending beyond the regiou except in a single tropical species, the homochromous Asteroidex, some of them repeated in the south, and the Madies and Tagetinez, almost endemic. REGIONS AND AREAS OF DISTRIBUTION. 519 The characteristic tribes of the Mexican region are, in the first place, the Helianthoidez, Helenioides, and homochromous Aste- roidex ; and, secondly, the Eupatoriaces, the former including most ofthe small endemic genera which mark the region as a great centre of preservation of ancient forms, the latter exhibiting the genera which appear to be now in the greatest degree of pros- perity. In the remaining Asteroides, Vernoniaces, Gnaphalioid Inuloidez, and even Senecionides the endemic genera are few, and those which have numerous endemie species are still richer in the more southern Andine region; the Cichoriaces are below those of the adjoining United-States region, although considerably more numerous than those of the adjoining Andine, and even of the more congenial, though distant, Chilian region; the few Mu- tisiaceze are the outlying representatives of a South-American and eminently Chilian tribe. A few Old- World, and especially Euro- pean aud Mediterranean, races are here represented more strongly than in the intermediate United States, or in some cases to their exclusion. Adenostyles californica is intermediate, as it were, between the two European species. Leucampyz is a close repre- sentative of the European Anthemis; Baileya is also nearer to some Old- World Chrysanthema than to any American genus ; and the Old-World Cnicus has more numerous and much more marked Mexican than United-States species. It is probable that many additions will be made to the Compo- site flora of the Mexican region by future explorations, and more especially in well-marked endemie monotypes, which, from the severe struggle they have sustained, are generally confined in small numbers to limited localities. The above-mentioned Leucampyz, so remarkable in its European connexions, is one of the most recent discoveries, and only reached me, in fact, at the moment when these notes are undergoing the last revision for press. 2. United-States Region. Under this name I would, so far as Composite are con- cerned, include the whole of North America east and north of the Mexican region. It is true that this may bea combination of two floras of separate origin which may appear at first to be very distinet, the Rocky-Mountain British-Columbian and Canadian flora connected with the Asiatie, and the more strictly American flora characterized in Composite by such genera as Helianthus, Coreopsis, Rudbeckia, Solidago, Liatris, &c.; but the two are so 520 MR. G. BENTHAM ON COMPOSITE. blended together that, with the data now at my eommand, I have been unable to separate them; and both present some general features which may justify the considering the two regions, for the present purpose, as one. The United-States region presents in its endemic, or almost endemic, genera a great contrast to the Mexican; the strictly endemie genera, indeed, which do not cross into Mexico or Cali- fornia are but very few ; but above thirty are almost confined to the region; and if we include in them the Euaster section of Aster, Solidago, and Helianthus, which, with Rudbeckia, Liatris, Sil- phium, Helenium, &c., are so characteristic of the region, we have an average of twelve or thirteen species, instead of about four, to a genus. The averages are brought much nearer together if we take into account the representatives of the genera belonging chiefly to other regions ; for whilst the Mexican region has a very large number of the southern genera Vernonia, Eupatorium, &c., the United States have but few species either of specially Mexican or southern genera or of those which belong to the northern regions of the Old World. The general average of species to a genus is thus brought to a little above six (774 species in 122 genera) in the United-States region, against a little over five (1330 species in 246 genera) in the Mexican. This remarkable deve- lopment ofthe endemic genera in the United States, compared with the paucity of their species in the Mexican region, may be taken as evidence of the prosperity of progressive races in the former, whilst the Mexiean region affords greater protection for the preservation of expiring races. There are, however, two or three of the United-States mono- types as remarkable in their isolation or distant connexions as any ofthe Mexican genera. Stokesia and Sclerolepis have no imme- diate affinities, the former being intermediate, as it were, between the otherwise distant tribes of Vernoniacez and Cichoriacez ; aud Sclerolepis, although technically placed among Eupatoriacc®, differs in habit and foliage from all other Composite. Chrys0- gonum is also remarkable from its connexion with the tropical- Asiatic Moonia, so close as to force us to unite them in one genus. The other endemic or prevailing genera belong to American tribes subtribes or genera, but with a greater proportion than in any other American region of the Old- World Anthemideze Cynaroidez and Cichoriacee ; whilst the South-American Mutisiacex are only represented by a single widely spread species, and the Old- REGIONS AND AREAS OF DISTRIBUTION. 521 World southern Cotuleæ Calendulaces and Arctotidesx are wholly absent. It is probable that some modifications in the above relative numbers may ensue from a rectification of the limits now so vaguely assigned to the Mexican and United-States Composite regions; but the flora of the latter has now been so generally in- vestigated that the future discoveries of endemie monotypes are likely to be but very few in proportion to those we may expect from Mexico and California. 3. West-Indian Region. Although the West Indies in their Composite show a generally close connexion with the continent of Central and Southern America, yet they are sufficiently separated to exhibit many of the characteristics of insular floras, and to require treating as a separate region. Among the larger islands, Cuba shows more of the character of the Mexican, Jamaica of the Andine, Trinidad of the North-east Brazilian or Guiana region, Porto Rico and San Domingo, which, as far as known, may be considered an ex- tension of the Cuban chain, have as yet been but very imper- fectly explored. The connexion of any of the islands with the opposite coast of North America appears in Composite to be confined to a few maritime species or to such as have a very wide American range. Among the characteristic genera of the West Indies may be reckoned Salmea, which out of twelve species has eight West- Indian and four Columbian or Mexican, Veurolena, with one West-Indian and one Columbian species, and Borrichia, which from the islands extends round the coasts north, south, and west. Each of the larger islands has also its monotypic or small endemic genera, ten in Cuba, Lachnorhiza, Phania, Sachsia, Rhodogeron, Heptanthus, Pinillosia, Lantanopsis, Thymopsis, Lescaillea, and Anastraphia, averaging two species each, one in Jamaica, the monotype Cheno- cephalus, and three in San Domingo, Piptocoma, Narvalina, and Te- tranthus,the former two monotypic, the latter of two species. These islands partake also of some of the generally dispersed large South- and Central-American genera. Cuba, for instance, has twelve species of Vernonia, twenty-four of Eupatorium, twelve of Mikania, and five or six each of a few others ; and in Grisebach’s Flora of the British West Indies (slightly modified to reduce the genera to the standard we have adopted) are included twelve species of 522 MR. G. BENTHAM ON COMPOSITA, Vernonia, thirty-one of Eupatorium, ten of Mikania, and six or seven each of a few others ; but the number of Mexican or South- American genera represented in the islands by single or only by two or three species is sufficient to give, as a general average, not quite three to a genus in Cuba, and rather more than three in the British Islands; or if the whole of the islands, as far as known, are taken into account, the average is brought up nearly to four species to a genus. The endemie genera of the islands consist generally of herba- ceous aud often small species; where shrubby they belong to groups which are elsewhere shrubby, and the species of genera common to other countries are not more shrubby than their con- tinental congeners. Narvalina, however, may be exceptionally regarded as a shrubby representative of the herbaceous genus Bidens. There is no tendency to the arborescent forms of the more isolated islands of the Pacific and Atlantic Oceans. The mountainous islands of Cuba and San Domingo have more the character of detached fragments of a continental mountain-chain which have preserved the remains of a very varied flora, than of really isolated islands that have through a long course of ages modified such races as may have been casually brought to them under former physical conditions. 4, Andine Region. The Andine or west tropical region of South America is but vaguely defined for our present purpose. It is a mountain-tract, connected in the north with the Mexican, to the east with the Brazilian, and to the south with the Chilian region, including generally the Columbian, Ecuadorian, Peruvian, and Bolivian States; but the bordering districts on each frontier are among _those of which the vegetation is, perhaps, the least known to us, thus depriving us of the data necessary for determining not only what are the precise limits of the region, but even whether any such can be assigned. The statistics of its Composite are thus, as yet, very uncertain. With a marked general character, it contains also numerous species of the great and widely diffused American genera Vernonia, Eupatorium, Stevia, Mikania, and Baccharis, as well as of the cosmopolitan Senecio. Its own genera are connected sometimes with the Mexican ones to the north, sometimes with the Chilian ones to the south, or with the Bra- zilian to the east; and some of those common to Mexico aud REGIONS AND AREAS OF DISTRIBUTION. 623 Chili run also along the western declivity of these intermediate Andes. All numbers given in the preceding Tables are in this case, therefore, particularly uncertain, notwithstanding the valu- able data supplied by Weddell’s * Chloris Andina.’ Accurate as his details are, he followed for the limits of his region chiefly altitude, which affects what Grisebach terms plant-forms and species, rather than genera or other races more indicative of origin, which are mostly very different in the Chilian and Columbian Andes. In comparing the Composite of this region with those of the adjoining Chilian, Brazilian, and especiglly the Mexican region, a striking peculiarity is the small number of endemic monoty pes. They do not exceed ten ; and not more than half a dozen endemic genera of two or three species could be added. The physical conditions are not adapted for the preservation of isolated races of varied idiosyncracies; they are too generally uniform to afford the necessary protection against luxuriant races which can freely range over large districts. As in the temperate and mountain regions of the north, this comparative uniformity of physical con- ditions has given at once a wide range to species and a large average of species to the characteristic genera. These compara- tively uniform conditions are also evidently such as to favour the development of Composite; and, moreover, the region itself is probably one which was very early inhabited by the order. The total number now known, very nearly the same as that of the Brazilian region, far exceeds that of any other American one except the Mexican. A few of the endemic or nearly endemic genera (such as Astemma, some of the Mutisiacez, dic.) may be supposed to bear evidences of great antiquity; others appear to be in the height of prosperity and luxuriance ; and the region exhibits more arborescent Composite than any other, except insular ones. Among the characteristic tribes of the Andine region the heterochromous Asteroidez, the Senecionidez, and Eupatoriacee, which take a second place in the Mexican region, may be here placed in the first rank on a par with the Helianthoidex ; the Mexican Helenioidez and homochromous Asteroidex arereduced to very few species; and the only endemic races of any higher value are a small Andine section of Chrysothamnus* (four species), Cacosmia (three species), which is almost as near to Calea as to its technical cotribuals, and Schizotrichia, a single species as yet, perhaps, imperfectly known. Vernoniacez, rather more numerous than in * Sincethis paper has been in the printer's hands it has been pointed out to me by Asa Gray that De Candolle's name Bigelovia has the right of priority for this genus, ^ 524 MR. G. BENTHAM ON COMPOSITE. the Mexican, are still far below those of the Brazilian region ; the Gnaphalioid Inuloidex, generally mountain plants, are rather more at home than in either. Mutisiacex, characteristic of the Chilian region and very rare in the Mexican, have several large and flourishing genera endemie, or, at any rate, with their chief seat, in the Andine region. The Old- World orders partially represented both in the north and south are reduced in the intermediate Andes to fewer numbers than in any American region except the Bra- zilian: such as are to be met with (Cotulee, Hieracium, Hypo- choris) may be regarded as remnants of those ancient migrations from north to south or from south to north to which I shall allude under Table 10, these remnants having alone survived the altered physical conditions so as to establish a few subordinate endemie races. Future discoveries may probably add much to the Composite of the Andine region, the eastern valleys of the great mountain- range being, as yet, but little explored; but these additions we must expect rather to consist of new species to the prevailing genera, than of such new forms, especially monotypes, as the three adjoining regions may be more likely to supply. : 9. Brazilian Region. The Brazilian or east tropical region of South-American Com- positz is that vast traet of country extending from the eastern declivity of the Andes to the Atlantic, southward to the Rio Grande do Sul, and northward to the valley of the Orenoco. It might be supposed to be naturally divided into two, the northern or Guiana and the southern or true Brazilian region, separated by the broad forest-valley or plain of the Amazon, so poor in Com- posite; but, as far as known, the Composite peculiar to the mountains east of the Orenoco are generally of a Brazilian type, or, at least, rather Brazilian than Andine; and the few that are characteristic of the low moist valleys of the great rivers spread over too wide an area and are too much interwoven with the others to be made use of for the distinction of separate districts. The statistics of the order in the whole region, however, are as yet more unsettled even than those of the Andine region. This portion of the Brazilian flora is now being worked up, and, when completed, will no doubt give many corrections to the numbers given in the above Table, which are necessarily often little more than conjectural, founded on a hasty turning over of specimens and reference to publications. REGIONS AND AREAS OF DISTRIBUTION. 525 We have here, again, as in the Mexican region, a large proportion of endemic monotypes or small genera; but the larger wide-spread American genera, especially Vernonia, Eupatorium, Mikania, and Baccharis, are so copiously represented, and some of the charac- teristie ones so rich in species, that the total average of the region is above seven to the genus, at least one more than in the Andine and two more than in the Mexican. Notwithstanding the large extent of the forest-plains above mentioned, almost as unfavourable for the development and preservation of Composite races as those of east tropieal Asia, the physical conditions of the hilly districts appear to be suited both for the preservation of expiring types in limited stations and for the luxurious development of others in the prime of life. The greater number of the monotypic or small endemic genera of the region belong to the southern portion, the campos and sierras of the Upper Rio San Francisco, and thence to Mattogrosso and Chiquitos, separating the great valleys or plains of the Amazon and the Parana. Itis there also that are to be found the most re- markable forms, unrepresented in any other part of America, the Lychnophoree, Schlechtendahlia, Wunderlichia, &c., the two latter having no very near connexions anywhere. A few,such as Pacou- rina, Sparganophorus, Riencourtia, Trichospira, &c., belong more specially to the northern or intermediate portion, but extend more or less into Brazil proper, and some of them have crossed over into tropical Africa; others, again, like St¿ftia, belong to both north and south divisions, with endemic species of limited areas in each. As characteristic tribes of the region, Vernoniacex undoubtedly take the first rank in the number of species, both in relation to the total Composite of the region (nearly one third) and in relation to the total number of the tribe in America (more than three fourths, including eighteen endemic genera). Helianthoidex, mostly of the subtribe Verbesineze, are also dominant, Eupatoriacee rather more numerous than in the Andine region, and Mutisiaceze about the same, several of the latter having their chief seat or being quite endemie in Brazil. Senecionidez, especially the genus Senecio, and Helenioidee are fewer than in any other American continental region; so also Asteroides, with the exception of Baccharis and the Gnaphalioid Inuloidez. The Old-World orders have but very few species or are quite unrepresented. Additions to the Composite flora of the region are chiefly to be expected from Mattogrosso and other western districts, and per- 526 MR. G. BENTHAM ON COMPOSITE. haps from the Upper Rio San Francisco, and possibly a few from the unexplored regions of Guiana bordering on Venezuela; but the collections of Schomburgk, Spruce, and others from Guiana and North Brazil have hitherto shown fewer remarkable Com- posite than of several other orders. 6. Chilian Region. The Chilian or extratropical South-American region of Compo- site comprises the whole of that continent south of the Andine and Brazilian ones. It is in some degree a mixed region: the elevated ridge partakes of the general Andine character as to its Composite ; the extreme south might perhaps be separated as a portion of a general Antarctic region ; and many of the strictly Chilian genera, confined to the Cordilleras, do not reach the plains of Buenos Ayres to the east. Yet, on the whole, it is a general area of preservation of Composite races sufficiently distinct from the Andine and Brazilian, which immediately border it on the north, to be regarded as one general region—the more so, as in its repetitions or representations of distant northern races the eastern districts take their proportionate share with the western ones. The flora of Chili proper is perhaps better known than that of Brazil or the Andes; but still there is a large tract of country in the interior, especially where it borders on Bolivia, as well as the provinces of Tucuman and others of La Plata, in regard to which the data are too scattered to enable us to judge readily to which region they should be referred. The Atacama plants described by Philippi evidently belong to the Chilian, and are included under that head in the preceding Table. The small or monotypic endemic genera, about as numerous as the Brazilian ones, bear a much greater proportion to the total number of Composite in the region, although they are still con- siderably fewer than in the Mexican region. The large American genera Vernonia, Eupatorium, and Baccharis have much fewer representatives than within the tropics, and Mikania has but à single species, a deficiency partly compensated by a greater specific luxuriance in Senecio and some specially Chilian genera; and the average of species to a genus is brought to a little above fivej'the same as in the Mexican region. Extratropical South America appears to have afforded physical conditions favourable at once for the preservation of locally limited types, either the remnants of very aucient introductions or differentiated in the region itself, REGIONS AND AREAS OF DISTRIBUTION. 527 and for the development and differentiation of numerous species in a considerable number of genera. Among the characteristic tribes Mutisiacee evidently come first. This is the sole region in which they are dominant, and so much so as to form nearly one third of the total number of its Composite, and to include nearly one half of the whole number of American species of the tribe, fifteen of the twenty-seven genera of the region being almost or quite absolutely endemic. Asteroide:, especially the homochromous genera corresponding to the Mexican ones, besides Baccharis and Erigeron, the Gnaphalioid Inuloidew and Eupatoriacew are fairly represented; the more tropical or northern Helianthoidew and Vernoniace: are but very few; Helenioidee are also few, but more in proportion than in the tropical regions. The European Cichoriacee and Anthemidex also reappear ; and even Cynaroidee and Calendulaces are amongst the evidently ancient inhabitants, the former represented by five endemic species of the specially Mediterranean genus Centaurea, the latter by an endemic monotype of an otherwise African and Mediterranean tribe. 7. Connexions between distant American Regions. Under this head we can only refer to north and south; for although there is a great difference between the Composite of the east and west coasts,both in North and in South A merica, the regions are continuous; there is no special eastern centre or area of pre- servation separated by a broad interval from the great western ridge, which might render the appearance of the same genera in both a noteworthy circumstance. The eastern genera, except a few local and monotypic, either extend to the western limits of the region or penetrate continuously into the western region ; some of the western genera send a few species into or all over the eastern region, but do not reappear abruptly after a broad interval. But the case is very different with respect tonorth and south. A considerable number of genera and even some species are esta- blished in extratropical North and South America, completely separated by a long tropical interval. The physical or other cendi: ‘ons which have in ancient or modern times admitted of the gradual extension of certain Composite races from east to west or from west to east have not been interrupted by the interposition of impassable barriers; whilst any such continuity between north and south is in the present geological period absolutely broken by LINN. JOUBN.— BOTANY, VOL. XIII, 2 P 528 MR. @. BENTHAM ON COMPOSITA. the intervention of tropical ungenial regions, far too wide to be crossed by the ordinary gradual progress of plant-races. The reappearance of identical races, generic or specific, in the north and south can therefore only be attributed either to a preexistence of different physical or other extraneous conditions allowing of gradual migrations, or to ancient colonizations through channels which now appear no longer to exist. These remarks may be illustrated by the following list of genera, identical or representative, in extratropical North and South America, without any, or rarely with only a single, species in the intermediate tropical regions :— Species in || Species in extratropical | extratropical Genera. A > A A Genera. ————— North | South | North | South Amer. Amer. | Amer. | Amer. Gymnocoronis ...... | ld d D MEA | 8 E Grutierrezia ............ BM 6! Jaumea ni 3 1 | Qrimda da ET PON | DABIS n. osse 2 1 Chrysopsis ............ It Habia: soas cunis 19 1 Hysterionica ......... -.. | 5 | Blennosperma......... 1 1 Haplopappus ......... 38 26. Ií Vallanoya. a-st 4 Sobdago sss 80 2 le Closia: atado pad E 5 Ericameria ............) 4 | Hymenatherum ...... ms | Nardophyllum g Gaillardia ............ | z Lepidophyllum } || Hymenoxys............ 9-4 2 IIE Gs |... Bo | Artemisia ..........-. 30 3 Antennaria ............ 8^ 1 | Centaura 1 5 Adenocaulon ......... 1o 2 i| rs n.. e ies oe, Eransetia es sese 6 3 | Troximon ........ ... 10 i1 6 NCAA o 15 3 | J Scorzonella ............ 3 | Thelesperma ......... 4 L| Pisos... odii le | 1 il | REGIONS AND AREAS OF DISTRIBUTION. 529 Table 11. Number of Species of Composite in each of the Old-World Regions. Number of Species. Genera. $ E & 2 E | yg eri JIFPIEPIEMEPIEFIEPIE SL PAPA OA FOIE der pP Be arise E D PRIOR IR u AP EA lee 3^ ES VERNONIACER. | [Une em atl AE Id xe 1 | Amer. Lhulis 5... cne silane 1} 2 Lv dre ME E ” E o | aed Gutenbergia 4... oense: EA DUE ahs E 5 Centratherum ............ pal n 1 414-1 195) Amer: Lampracheenium ............... made ice Toed e 1 «Botlirioclmess... eren esee Iced sorte 1 Adenoon 4... eo eats i ax pda o al EPleurocarpæa O ii A osas did Vernonia s... 1 NE see . | 70 | 46 | 15 | 1/125 | Amer. Hoplophyllum ................61| +. teekond ses um Heérdena eerie idi 2. haea e | seu cou Corymbium. ........... 5 ene a sedes: ceto A a Elephantopus .................. A A A O r4] dPD MM Total Vernoniacez......... 84 | 55 | 24 | 4 |156 EvPATORIACEE. | Adenostemma .................. Er Hi 1 1 1 1 | Amer. Ageralum eere sce Pole llli ds ad Ame Eupatorium... S 24 51.L4-515-]1:2] 19] ame Mikanis 2. cass cesses x d Sd 1 1 11. 1 | Amer. Adenostyles .:.2...05...s0ssessees- 2| 85s ec 3 | Amer. Total Eupatoriacez......... 41 8| 41 8|^9| 2110 ASTEROIDES. Boldago ose Ti 155 elec E [Amer Homochroma .................. eux I 1 Eresenia.. cnica. ve dee fen an 3 Plerónia. orn ec a E olo OE Microtrichia ....—.. e eene sub a wp Dicrocephala............. o I 3| 4 Hf D Oyathoeline, E 2. 2-3: ene bes ee er A TANDA oe coso ses secre wil. tee ie Sr ee 2 Coruna... coro Dake. att a eet l Myriacüs .........: eee c 1s pil 5 o S Rhynehospermum .......... sdqueluwwlTdiw]e ti lagenophora.....——— o]. Seba po Slip Ame Garuleum ...... yay : Suite d E BUBIFÜdISCUS co uersus DONC ET 2: 4 bee o ARA lw 1 | 36 | 37 Carried forward............ 4| 1| 9|15]61 | 40 120 |] SP se 530 MB. G. BENTHAM ON COMPOSITÆ. Number of Species. T. E o G 3 B enera Genera. 1 E E E E "d in 8 li a. |S E., a ¡03 America. EHEHEHEHEHEHEE mm if lá ja |4 je ASTEROIDES, Brought forward ......... 4| 1| 9/15/61 | 40 [120 «Bells Ses aeieea eane 6| 1 eld wc 6 | Amor: «Belbum cies eck Sotessoscese itses 3 3 'Amellus a sccscacstececsstescce sss fa eee eee tS be eee Gymnostephus .................- us use lux. © bendi ChaFlel8 ever E R nta pe (ceo d 1 Mamia o... eie oerte sita Por Late l0 10 Minuria 00. coa eo ee nes Ses eb Bee Calotis Ye TITVTTETTITT 000000000000 ecc eee .. 15 15 Heteropappus ................-- Al CWiclbu D aedueedod Soo A a a do i. 5 | Amer. CGalhsiephus veo e| bul. 1 LS race 8| 8| e 1 50 | Amer. HOC o T es le tdg i do leaning s E a acd EIE E ERSIPOE: Camisa ener. Ves si RI PPodocoma o0 one los eer 1, 1 | Amer. Lechuophyllum.................. Kl 1l ese 2 SODIUM oeeie 2| 4|. et 6 | Amer. RE. oos cold e cipe coh 6] 9| 23) 31.11 4/16 | Amia: BERND AS sleleisl di t AM MOGFOBIOSEA. |. esee £s 3| 4|. Naidorélla aoee: 24.14 15 CODVER ooo seu Eger edes 4 14| 7] 9 2 | 32 | Amer. PuRdim ato E S E dos) dodo 2-1 l 3 AOP OSUEN AA Ce ets ares 2 AUE E con entier denen sense . eccL X Dus H ieteromms: A ahne acion be beber de 1 A ee e ien Selle abuse m Noo eo s 1 vs 31 4 Total Asteroides ......... 34 | 75 | 36 | 31 167 135 444 InvLorpEE. Brachylena AI a a E Y pe M Farchonanthus o-s eeen co ES xu ua ion 3 Blumea& —. eee 1 3|48| 4| 7 | 55 E&ggera o. oko. 14 ake "l3. lim A Lu. cerne eere il 4l 6 6 | 16 | Amer. a e CO du TI E hospi. ess Ree totes os 11.3 Coledcoma con. e or ulocleb ipu Nanothamnus .............-.. lato Prats Denekm eoe peek} Sool ke es |e 2E 2 sete icine cee UP) 311:3] 2| B Spheeranthus .........00.........| 1 a1 5 ..] 2| 8 Carried forward............ | 2| 2/24/62 | 17 | 26 [118 | | | | REGIONS AND AREAS OF DISTRIBUTION. 531 Number of Species. $e g 3 3 E 3 Genera Genera. E d & E 8 as} in S $.|*9.|49.[8.|8.|O .| America. HAEHEHEHEHEHES -E1E3C as E TESI: SE/BRI SE | SR BR 82) SE a ja já ja la 4 le INULOIDEE. Brought forward ......... 2|, 2124/62/17 | 26 118 Pterocaulon ............... esce eee uae 5 | Amer. Blepharispermum ............... eeu 25 x. 3 ALTlTOISDAR S oca eee reese ED 3 m qe 1 Symphyllocarpus ............ .. P. re dr 1 uc E Ar i DAI D Done Oo |: scu wg Amer Micropus tocadas oriri. [stg Amer: AO o aae TE TI succ Amor: Hogar iio se ee lie ii o 8 Stuarting: ...— 5 i. ee fusien tatal led Amphidoxa oiae ae a e pee a aa Ereg 1 Chiliocephalum.................. E 1 sedes Antennaria - n e T £1. ].-|—1 E Bi BN Leontopodium .................. ise fF OS bee Pie | ap TA Anaphalis ase g: sets 8 2519555 E32 || oe | 2b | Amer: Pterygopappus .....-..:...-..--- eee o a coco Bcn ed DT LAODO oer Iu Id 2 Phagnalon e.a 1.6: +. Er Elton Achyrochine n- tinain secu, rg Amer Gnaphallum -ooe <= 8| 6/10; 6/10| 8,30 | Amer. IBaoühs E wort [eee ne reu pm. Leptorhynchus .................- ndum eT 81-8 MEA ci E dE «c [ete Or O Hehpterüm = oe woe | love beso Ev Helichrysum ).-. oan i 30 | 3 | 23 | 2 |137 | 52 |2385 Meontony®: -ensi Lupe OJ uri D Fachyrhynohüs -nn meet ec Ij DT Cassmia s.s ISDN xt 2 ec. 1 | 13 | 14 Phenocoma coca tee [cep 14 1 Sehonia eerte m eus ii 1 Anaxeton 5... 0.5. cee: e Mu bb Petalaco" reete s He Ip 1 Iuolena-.—. ns CIS. = sel 504 B Podotheca ........... erred ese : oes 51 5 Millotis |. lee : s . 21 32 Quinchao reos ele : 1 1 Rutüdods 0n i pum á S repe Ammobium «eee : : scel 231-9 SOY phocoronis -i-si nsee Sea Ae Seok seep ok 1 Toxanihus...:. cess, Mu. : es dicas E -rrochlamys::.-.... eue coe | ooo ere [a dale E Humea ..... UR T eire. a : Sd Sel ek ACORN otero 0er at .. 2:2 PiUocArDa 00r SE s 1 1 FROGS | eene ainia de 25: eie LI Cesulia ........... Kr ge uen d li bep Carried forward............ 71 | 41 | 66 | 89 201 ¡188 ae 532 MR. G. BENTHAM ON COMPOSITE. Number of Species. lo f=] j E E & 3 a Genera | Genera. El E a ‘a $ 3 in ois O A A Eslúclas $e SE 48) eS JE pE Ae Loam Ew qe E A E R E Š & n 2 n Z H 2 ÍNULOIDES. | | Brought forward ......... 71 | 41 | 66 | 89 201 188 597 Myriocephalus .................. ded en =| 8) 8 Jea a a er] e | 22 122 Gnephosis 1e hr ie ve wo ER | 12 | 12 Calocephalus.....................| 00» | eee see | 19. 10 Cephalipterum .................. eee Fees 2] Et Gnaphalodes ...........-—.-.1..e s X | 3) 3 Sr ds, Ber Dereon tr eer - 1 | EE Oraspedig |. tuu ace e ME 4 Chthonocephalus ............... oot sss | 3| 3 IBERO LIC esoo teasee Di HON E ee a 19 |... | 19 A nes reri : 51. 5 Elytropappus....................- are 6 | - 6 PBierothrz ioio. oo 94 8 Amphiglossa < co.. ps. sec a e 3 Bryomorphe roscado nacen Soe lector aca ce 1 Métaladm oso ees M Core 20 20 Lachnospermum ............... : 1 1 Nestleràp 0e. eoe tere 10 10 Annplypha- — —— 5 2 2 POOR cere e reete i81. | 18 Hosen aos I A eov 1 delas. 4 BOWEN. oaeiai oee e ee E a E Eodolepik a.u err: s enee accede 2 REINS oaen: Fe 9l. 3 PEU Cone eene Rodeo 6! OH ERN Lido cernere eei raa 1 3 C ATFOWBDHBIA 1... eee. seed bee JE Bee 1 1 Codonocephalum ............... 33. eid rei LIE BO e D ues A |. |52 Homocheta -s-re e pe elec pol Bojen 5:59. e vo duse ecl e SIE Minurothamnus ............... c zr €ypselodonüa s. Ec e e a ok Ginta oa a ce E eae Be ee ole tr cll A NM E 2 ys E. A sho coos noma Soy fg UNDE ubberisebebt Pl] Éphions o one ee eoe Sit Fill] 9L. 13 the, ean erc tne eletbb.r lta 3 VICOS ACA Li asia sa casein nera 4 EE 6 | Culo to plano .,...L-..eeree eee a) A E Et 1 Pulena- 13 uod cid i68: 2 41: 41 I 22 Porphyrostemma .......... Www tfiba 1 Amblyocarpum ..............«. 1 no lu In 1 Carried forward ............ 132 | 72 | 85 | 99 ¡322 (268 903 | | | | | | | REGIONS AND AREAS OF DISTRIBUTION. 533 Number of Species. E s E E a Genera Genera, 8 lą q à E yg in Pelega aliala] America E8|88|$8|$8|15| 38/57 CEIEEIEHIEHIEEJEEIET: S m m je ie 4 ie INULOIDEA. Brought forward .........192 | 72 | 85 | 99 [522 [268 [903 Carpesium ..... aos dl wr DA A 5 Geigeri: O rossi 2 kat 8 19 Gymnarrhena ........ pyre 1 ipse levies ese: [a ok Rhantertum :.....5.. ee 2. sese A Anvillea e6s0000009000020099099009 .. 2 ... oo . s.. . 2 Ondetia osorno ee o eee Ehe : 1 1 Buphthalmum. cta... 4 gue 4 Callilepis eren clu pu 2 2 Anisopappas A ccs lai T 2 Odontospermum ............... Dips 5 Pallenis*- DOS ETE E: psv e 1 Chrysophthalmum ............ Grip den dere depo 3 Osmites — a AL se Rare 6 6 Osmitopsia-. - È Galanga . . (UE officinarum. . . . S> 6 stachyodes . + . . . . 126 Amomum angustifolium, note on 154 Andreecium, on the development of, in Cochliostema, Lem. . . 204 Anomodon exilis; p o oo r 909 —— filiformis- . . . . . +. 308 Huttoni s. . . . . = 900 Anthemides, distribution of . . 450 Antiquity, comparative, of races in Composite . . . . 481 Arctotidez, distribution of Asclepiadese, fertilization of . . 48 Aspergillus Mülleri . . . . . 175 Aspidium angulare and aculeatum, varieties of . . : D Asteroide:, d.stribution of . . 402 Atylosia geminiflora . ve 95 —— glandulosa . . . . . . 185 Aulacopilum abbreviatum . 308, 326 incanum . . E . 308 —— tumidulum . . . . 307, 326 Auricularia albicans. . . . . 170 Australian Fungi. . . 155 — — Proteaces, notes on ‘the styles of. . . 58 Baker, J. G. Revision of the. ge- nera and species of Scillee and Chlorogales . 2... . . 209 LINN. JOURN.—BOTANY, VOL. XIII. Page Banksia. . . NTC OO) Bauhinia reola UE . 188 Bennett, A. W. On the floral structure of Impatiens fulva, Nuttall, with especial reference to the imperfect self-fertilized flowers . . 147 Bentham, G. N otes. on the styles of Australian Proteacee . . 58 Notes on the classification, history,and geographical distri- bution of Composite . . 335 Berkeley, M. J. Australian Fungi, received principally from Baron F. von Mueller and Dr. Schom- Eis burpk > se Bovista Mülleri . . . . +. .171 Bowiea . ee 290 Bryological notes ‘by s. O. Lind berg Buxus Harandi ; Up AE NIE . 123 Byrsanthus, note on the genus, and its floral conformation «~ 1 Byrsanthus Brown. . , 4 P H epigynus «ME Calendulacese, distribution of . 463 Camassia, species of. . . 256, 267 Camellia Scottiana, Wall. . . 329 Campylopus subulifolius . . . 297 Capparis galeata and C. pde i note on . +. - : 72 Carex Fabri . . +. +. + + 90 siderosticta . . Centres, chief, of the principal races of Compositse s Ceylon, new species of Musei from . 0.7. 208 Cheetomitrium- ‘confertum S, Sli China, notes on some pa from Northern. + . Chinese silkworm-oaks, supple mentary note On . +. + Chlorogales, revision of genera and species of. . + + + .Chlorogalum . + . +» +. 291 Choanephora Cunninghamiana . 578 Cichoriaces, distribution of . . 475 Clasmatodon Bertrami (Schimp. ) 71 —— parvulus (Hamp.) . . + 70 —— perpusillus (DeN)... FO Clavaria lorithamna . . +. . 169 2T 580 INDEX. Page Page Cochliostema, development of the Ectropothecium levigatum . . 320 andreciumom: : > 67s : : 204 subrtusum > . +. +. : 321 Colonizing Composite . . . . 468 | Elsholtzia Stauntoni . . +. +. 85 Cometium angulatum . . . . 302 | Entosthodon planifolius . . 904 appressifoium , . . . . 302 | Eriopus lucidus . . +. +. +. . 311 ——hispidulum. . . . . . 302 | Erpodium Belli . . . . 307, 326 —— minutum . . . . . . 808 ceylonieum . . . . 306, 326 orthostichum . . . . . 302 | Eucomis, species of . . 224, 225 seminudum . . . 808 | Euonymus Bungeanus . . AT gi Composite, notes on the classifi- cation, history, and sepa cal distribution of . . . 335 ; comparative value of generic characters in . . . 844 ; Sketch of the primary divi- sionsof, . . 974 , history and geographical distribution of. . . 390 Conospermum, fecundating appe- ratusof . . . 63 Cordylogyne, fertilization of . . 57 Crepidotus globiger . . . . . 158 Cunninghamia infundibulifera 334,578 Currey, F. On a new genus in the order Mucedines . 333, 578 Cyathophorum sublimbatum. . 309 Cynaroidee, distribution of . . 466 peta amoenum .. . . 296 acryomyces purpureus . . . 40 Deedalea aulacophylla . . . . 166 —r-Bowmani . . . . . . 166 Hobsoni. . . . 165 Dalzell, N. A. Note on. Capparis galeata (Fresen.) and C. Mur- rayi,J. Graham . . 72 ——. Remarkson Dolichos u uni- Horus, Lamarck . . . , 145 ——.. New anu ues from Western India. . . . 185 Determination of three imper- fectly known species of Indian Ternstroemiacee . . 328 Dickie, G. On the marine Alge of the Island of St. Helena. . 178 Dicranella edentata . . . . . 295 mfuscata +... 9295 —Wmbanguhlà -. . . . .3296 Dieranum deeumbens . . . . 296 Diervillaflorida . . . . 81 Disa macrantha, fertilization of . 45 Disperis, notes on a South- African 42 Distichophyllum limpidum . . 311 mucronatum . . . . . 311 Dolichos uniflorus, remarkson . 145 Drepanophyllum oppositifolium . 305 rimiopsis QU ca qua issu, 2A —, species of . . . 227, 228 Dyer, W, T. T. On three imper- fectly known TTernstroemiacee . 328 Eupatoriacee, distribution of . 400 Fabronia pusilla. . . . . 72 Schimper jS 1 0:295 64 Favolus csespitosus . . . . . 167 squamiger . . +. 166 Fertilization of certain species of Asclepiadee . . . 0 48 of Disa macrantha, observa tions on the . .- ao 45 Fissidens angustus . . . . . 324 I axlhflorus : i273) 5 820 —— Beckettii . . . . +... 329 bicolor . ro oo e 922 ——crassinervis . . . . + 323 —— Flabellulum . . . . . 324 == fuscoviridis’. |. . 929 microcladus s 324 — e minutus. > . a - .- 929 === mulüflorus. . —. . . . 929 papillosus . . . . e 2928 —— pennatuus . . . . . . 325 2c Plumua o . . . . » 32D terminiflorus . . . +. . 322 virens . . . . 924 Floral conformation of Byrsan- thus,noteon . . «15 Floral structure of Impatiens Jules, Nuttall o. .« HA Fraxinus Bungeana. . . . . 83 Fungi Australian . . . . . 155 Tremellinei ~ = : < 9 Garovaglia carinata . . . +. + 314 compressa . — . . . . did densifoha. . . 4 . . 912 === ImvifoHa..¢ o. = 919 —— obtusifolia . . . . . + 313 torüfolia > e= . «t. 914 Generie eharacters, comparative value of, in Composite . . .3 Geographical distribution of Com- posite . . E . 390 Geopyxis aluticolor . . . . . 176. Grevillea buxifolia . . . . + 61 Guepinia helvelloides . . . . 32 Helenioides, distribution of . . 446 Habenaria, South-African species of . Hanbury, D. Historical notes on, the Radix Selenga of pharmacy E NU e 20 INDEX. Page Hanbury,D. Note on 4momum angustifolium, Sonnerat . . 154 Hance, H. F. On the source of the Radix Galange minoris of pharmacologists . . . . . 1 ——. Supplementary note on Chinese Silkworm-oaks . . . 7 Notes on some plants from Northern China . . . +. + 74 ——. Flore Hongkongensis Sup- 95 plementum . . . . . +. Helianthoides, distribution of . 431 Hexagona decipiens. . 166 m Muler , . . Fira. 108 Historical notes on the Aadix Galange of pharmacy . . . 20 Hongkongensis, Flore Supple- mentum. -a o vow. 98 Hookeria ceylanica . . . . . 310 Hygrophorus flammans . . . 160 nigricans . - ~: 160 Hypnum applanatum . 321 zx Qurvisebum . a 9 0000: 68 ——"Teedali . . . . 0% Hypocrea cerebriformis . . + 177 Hypopterygium apiculatum . . 309 Hypoxylon scleropheum . . . 177 Impatiens fulva, floral structure DS. o 1 4 India, Western, new Leguminose Bh... vo 5 4 4 032 188 Inuloides, distribution of . . . 414 Isaria fuciformis . . . . + . 175 Jamaica; botany of . . . . . 331 Kup Miller. 9. . 2 107 Laschia micropus . . . +. . 170 Leguminose, new, from Western Tada. . 185 Leighton, W. A. On two new species of Mycoporum, Flotow 326 Lenzites Beckleri. . . . . . 161 Lepidopilum furcatum . . . . 310 Lepiota Beckleri . . . . . +. 156 v bubalina : . . ~- 4158 , Leptosomum densum . . . . 305 Leucomium limpidum . . . . 320 Lindberg, ‘S. O. Bryological Mites. s 4 . 66 ——. Remarks on Mesotus, Mit. 182 On Zoopsis, H. f. & T. . 188 Macrohymenium leve . . . . 317 Macromitrium appressifolium . 302 ——neueoumm 0. 02 9 a4 . 900 -— econtorlum.- « 4. . . «X 801 == epa. sy v 4 DON inom . . ... «4 908 —— orthostichum . .. > . . 302 ——— ramentosum . $ . 901 —— seminudum . . . . 903 581 Page Masters, M. T. Note on the genus Byrsanthus (Guill.) and its floral conformation . . . 15 On the development of the andrecium in Cochliostema, Lem... o a n o Soa e 5204 Mateer, S. Remarks on the Tamil popular names of plants . . 25 Mesotus, Mitt., remarks on . 182 Meteorium attenuatum. . . . 316 enerve s ime idi . 317 rufifolium . . 2x 6/2916 . 187 Millettia pallida. . .. Mitten, W. New species of Musci collected in Ceylon by Dr. Thwaites . DON Mniodendron aristinerve . . . 322 deltoideum.. .. .. . +. . 321 Mucedines, new genus of . 333,578 Munro, C. B., extract of a letter from, to Mr. Bentham . . 331 Musci, new, from Ceylon . . 293 Mutisiacese, distribution of . 471 Mycena tuberigena . . . . . 156 Mycoporum circulare E rdi melaspilon . . . 327 Naucoria Bowmanni . 158 frustioola —. 4 xw wee MB Nolina georgina . . . +. . . 292 Northern China, plants from. . 74 Octoblepharum Radula . . 298 Ornithogalum . . . . . . 297 , Species of . . + 258-284 Orobanche ombrochares , . . 84 pyenostachya . . 84 Oxytropis psammocharis . 78 78 subfaleata — mt ds Pachycarpus, fertilization of . . 55 Panax, gender of. =~ (note) 104 Panus coriaceus . 2d eon Patella Adamsoni . . + + .176 Paxillus Muelleri . . - ‘ er Peganum nigelastrum . . +. »- Periglossum, fertilization of . . 55 Persoonia « « « (mac Peziza Adamsoni are enhe aluticolor 176 —— hirneoloides . 176 Bile... woe MIN Phlebia hispidula : . 167 Pileostigma foveolatum. . + + 188 Pinus Bungeana . . . - » . 87 Pistaeia chinensis .. . - + + 77 Plagiothecium.subglaucum . 321 Plethiosphace pogonocalyx + 85 Pleurotus candescens . . + + 157 Guilfoylei . . . « . + 158 illuminans . oo. o MAE . 157 scabriusculus . + 582 INDEX. Page Page Poicilophyllum nitens . . . . 297 | Syrrhopodon albidus . 298 Polyporus ascoboloides . 162 cespitosus . . . +. 298 Beckleri . : . 162 parvulus . . . +. 299 cireus 4 3 v0... e) 102 strictus . . . (oe x 299 corrivalis . . . . 162 | Systegium abbreviatum . . 299 endapalus . . 163 | Tamil popular names of plants, : Libum . . . oc. MN remarka on A o 25 Polysaccum marmoratum . . . 171 Tayloria imbricata 305 Proteacem, Australian, styles Ternstroemia coriacea, Wall. . 330 of. 0. +. 58 | —— khasyana, Choisy. . 331 Psathyra Sonderiana 159 | Ternstromiacese, determination Pterobryum ceylanicum 315 of three Indian . . . . . 328 —— involutum . . +. .918 | Thelephora congesta 168 Puccinia acuta v... .. 12 exsculpta 168 dichondre . í . 173 luteo-cincta 168 Quercus castaneifolia . . 8 | Tortula consanguinea . 300 chinensis . . . . . , 8 | Trametes acupunctata . . + . 164 dentata =: ucc co X devera 3 . 165 —— mongolica . . . . . . 7 | —~epitephra iw 165 —— vallonea . . . 8 pholma. . v 164 Radix Galange minoris, source —— pyrrhocreas . . 164 of the , em | ungulata . 165 — : historical notes on . 20 | Tremella cerasi 39 Reeks, H. On the varieties of neglecta . . 34 Aspidium angulare and aculea- Tremellineous Fungi and their fue . . i 65 analogues, new notes upon 31 Rhinotrichum pulchrum : 175 | Tropidia grandis. . . +. . . 128 Rhynchosia mollissima . 186 | Tulasne, L. R.and C. New notes Rhynchostegium curvisetum , 68 upon the Tremellineous Fungi Reestelia polita . 174 and their quce Sok o) St. Helena, on the marine Alge Urginea . . wu . a of the Island of . . . 178 species of . . . . 216-224 Salvia pogonocalyx . . 85 | Uromyces puccinoides . . 173 Scleroderma pandanaceum . . 171 | Ustilago marmorata . . . . 174 Scilla Kk bes cH . 228 | Vernoniacez, distribution of . . 393 , Species of . . . . 229-256 | Verticillum eximium . . . 175. Scillere, revision of genera and spe- Weale, J. P.M. Notes on a spe- Guo . v... . 209 cies of Disperis found, on the MERE WEM > . a... M Kagaberg, South Africa, . . 42 incrustans . . . 36 Some observations on the Sematophyllum asperifolium . 319 fertilization of Disa macrantha. 45 capilliferum . . . 318 On some species of Habe- —— monostictum , + . 918 naria found in South Africa . 47 —— punctuliferum . +. . 318 Observations on the mode —— ramulinum . + . 319 in which certain species of uj ruficaule . . + . 919 clepiades are fertilized . . 48 Senecionidese, distribution of. 45b | Whiteheadia . . . . < . > . 226 Silkworm-oaks, supplementary Xylaria ectogramma . . +. + 177 note on . . E phosphorea. . . . . . 177 Spherothecium reconditum . . 294 Xylopodium australe . . . + 171 Stephanostoma Belli . . . . 307 | Xyomalobium lingueforme . 50 Stereum Schomburgkii . . . 168 | Zoopsis, H. f. 4 T., on the genus 188 sparsum. . . » » « »« 109 | Zygodon breviciliatus Se £ . 04 Stirlingia . . LI P . . . LI 64 TEE is . . . . . . s 304 flynaphen.. . ..» 2 © » « 68 | ——perpusillus, . . , . + 908 END OF THE THIRTEENTH VOLUME. Voi XIII tab. oc .Journ.Bot Q A Linn J-N Fitch imp. i Fitch deLetiith Y W linn Soc Journ Rot Vol XIILtab.2. \ We T. d v E E x. ^5 ) MY ? 1 1 W.H Fiteh.del ethth. J N Fitch imp. Zann. Soc. Journ. Bot. VOL XI tab3 2 Kath. Bennett. del. Garra 8c. Tom Socdourn Bot Vol Xft tab. F : 4 ] l e Ob OY Linn. Soc Torn Bot Vol XI bak 5. G Jarman oc. W. Ag; En del. Linn.Soc Journ Bat. Vol. XM tab E. E ^ GJarman se WALeghton del Zinn SocJourn. Bot Vol. XM tab "V GJarman sc. Limn.Soc.Journ.Bot Vol. XIlI tab.8. JN Fitch imp WHEitch hth. Linn.Soc. Journ.BotVol XIILtab.9. JN Fitch 10 W.H.Fitch hth, inn.Soc.Journ T BotVolXIII tab 10 XY Se E RE E TC EET EE a ma JAN Fitch imp Linn. Soc. Journ Bat. Vol XIILTab11 (^ y 3l Je ES Cal sala Lobeliaceás V. DS XE ue Vernoniace de Lachoriaceae Anthem (1-2 1 liL.